No title

HAZ IFC 195(C).qxd:INNER COVER

10/11/11

2:31 PM

Page 1

SUBMISSION OF ARTICLES Articles must be written in good English. We recommend that authors whose native language is not English have their manuscripts checked before submitting. A high standard of English is a criterion of acceptance for the review process. Articles not written in acceptable English will be returned to the author without review. Language services: Authors who require information about language editing and copyediting services pre- and post-submission please visit http://webshop. elsevier.com/languageediting or our customer support site at http://support.elsevier.com.

Aims and Scope The Journal of Hazardous Materials publishes full length research papers, reviews, case studies and short communications which improve our understanding of the hazards and risks certain materials pose to people and the environment or deal with ways of controlling these hazards and associated risks. The Editors particularly welcome proposals for review articles. To limit the scope the following areas are excluded: standard municipal and waste and wastewater treatment processes; manufacturing of explosives. The Journal publishes high-impact contributions on: • Characterization of the harmful effects of chemicals and materials • Measurement and monitoring of hazardous materials • Transport and fate of HM in environment • Risk assessment and management • Physico-chemical and separation processes • Biological processes • Advanced oxidation processes • Safer and cleaner technologies The Journal publishes papers with significant novelty and scientific impact. The Editors reserve the right to decline, without external review, papers that do not meet these criteria, including papers that: • Are very similar to previous publications, with changed target substrates, employed materials, analyzed sites or experimental methods, • Deal with parameter optimization of known processes without new concepts and/or interpretations, • Report the environmental analysis and monitoring of specific geographic areas without presenting new insights and/or hypothesis testing, or • Do not focus on the environmental relevance and significance of the studied systems or materials. Editors Professor Diana S. Aga, Department of Chemistry, State University of New York (SUNY) at Buffalo, 359 Natural Sciences Complex, Buffalo, NY 142603000, USA, Tel: (716) 645-4220, E-mail: [email protected] Professor Wonyong Choi, School of Environmental Science & Engineering Pohang University of Science and Technology (POSTECH) Pohang, Korea 790-784 Phone: +82-54-279-2283 Fax: 82-54-279-8299 E-mail:[email protected] Professor Andrew Daugulis, Department of Chemical Engineering, Queen’s University, Dupuis Hau, Kingston, K7L 3N6, Canada, Tel.: (613) 533-2784; Fax: (613) 533-6637; E-mail: [email protected] Professor Gianluca Li Puma, Department of Chemical Engineering, Loughborough University, Loughborough LE11 3TU, United Kingdom. Tel: +44 (0) 1509 222 510; Fax: +44 (0) 1509 223 923; E-mail: [email protected] Professor Gerasimos Lyberatos, School of Chemical Engineering, National Technical University of Athens, Zografou Campus, Athens 15780, Greece. Tel.: +302107723256; Fax: +302107723155; E-mail: [email protected] Professor Tay Joo Hwa, Division of Environmental & Water Resources Engineering, School of Civil & Environmental Engineering, Nanyang Technology University, Nanyang Avenue, Singapore 639798, Tel: +65 6790 6228, Fax: +65 6791 0676, E-mail: [email protected] Associate Editors Jaesang Lee, Water Environment Center, National Agenda Research Division, Korea Institute of Science and Technology (KIST), 39-1 Hawolgok-Dong, Seongbuk-Gu, 136-791 Seoul, South Korea, Tel: +82-2-958-6947, E-mail: [email protected] Constantina Marazioti, Laboratory of Biochemical & Environmental Technology, Dept. of Chemical Engineering, University of Patras, Patras, Greece, Tel: +30 6976 128157, E-mail: [email protected]

Online Submission of Articles Submission to this journal proceeds entirely on-line via http://ees.elsevier.com/hazmat. You will be guided stepwise through the creation and uploading of the various files. The system automatically converts source files to a single Adobe Acrobat PDF version of the article, which is used in the peer-review process. Please note that even though manuscript source files are converted to PDF at submission for the review process, these source files are needed for further processing after acceptance. All correspondence, including notification of the Editor’s decision and requests for revision, takes place by e-mail and via the Author’s homepage, removing the need for a hard-copy paper trail.

MANUSCRIPT PREPARATION Authors are requested to refer to the detailed guide for authors and artwork instructions available via the Journal’s homepage at http://www.elsevier. com/locate/hazmat

AUTHOR ENQUIRIES For enquiries relating to the submission of articles (including electronic submission) please visit this journal’s homepage at http://www.elsevier.com/locate/jhazmat. Contact details for questions arising after acceptance of an article, especially those relating to proofs, will be provided by the publisher. You can track accepted articles at http://www.elsevier.com/trackarticle. You can also check our Author FAQs at http://www.elsevier.com/authorFAQ and/or contact Customer Support via http://support.elsevier.com.

For a full and complete Guide for Authors, please go to: http://www.elsevier.com/locate/ jhazmat

OFFPRINTS The corresponding author, at no cost, will be provided with a PDF file of the article via e-mail. The PDF file is a watermarked version of the published article and includes a cover sheet with the journal cover image and a disclaimer outlining the terms and conditions of use. Additional paper offprints can be ordered by the authors. An order form with prices will be sent to the corresponding author.

FUNDING BODY AGREEMENTS AND POLICIES Elsevier has established agreements and developed policies to allow authors whose articles appear in journals published by Elsevier, to comply with potential manuscript archiving requiremens as specified as conditions of their grant awards. To learn more about existing agreements and policies please visit http://www.elsevier.com/fundingbodies

INFORMATION FOR ADVERTISERS Advertising information: If you are interested in advertising or other commercial opportunities please e-mail [email protected] and your enquiry will be passed to the correct person who will respond to you within 48 hours. USA mailing notice: Journal of Hazardous Materials (ISSN 0304-3894) is published semi-monthly by Elsevier B.V. (Radarweg 29, 1043 NX Amsterdam, The Netherlands). Periodical postage paid at Rahway NJ and additional mailing offices. USA POSTMASTERS: Send change of address to Journal of Hazardous Materials, Elsevier Customer Service Department, 3251 Riverport Lane, Maryland Heights, MO 63043, USA. AIRFREIGHT AND MAILING in the USA by Mercury International Limited, 365, Blair Road, Avenel, NJ 07001. The paper used in this publication meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper).

Printed by Polestar Wheatons Ltd, Exeter, UK

Editorial Board R. Andreozzi (Naples, Italy) J. Baeyens (Tremelo, Belgium) A. Ballester (Madrid, Spain) E. Brillas (Barcelona, Spain) V. Cozzani (Bologna, Italy) G. de la Rosa (Guanajuato, Mexico) D. Dong (Changchun, China) M. Erdem (Elazig˘, Turkey) S. Esplugas (Barcelona, Spain) M. Fingas (Alberta, Canada) V. K. Gupta (Roorkee, India) O. Hamdaoui (Annaba, Algeria)

W.-J. Lee (Tainan, Taiwan, ROC) P. Lianos (Patras, Greece) E. Lima (Porto Alegre, Brazil) T. Mangialardi (Rome, Italy) A. Mittal (Bhopal, India) D. Mohan (New Delhi, India) C. Mulligan (Montreal, Canada) A. Nzihou (Albi, France) J.M. Park (Pohang-si, Republic of Korea) H.J. Pasman (Delft, The Netherlands) S. Pavlostathis (Atlanta, GA, USA) H. Potgieter (WITS, South Africa)

P. Randall (Cincinnatti, OH, USA) K. Reddy (Chicago, IL, USA) J. Sneddon (Lake Charles, Louisiana, USA) M. Soylak (Kayseri, Turkey) M. Talawar (The Hague, Netherlands) T. Townsend (Gainesville, FL, USA) S. Tunali Akar (Eskis¸ehir, Turkey) S. Wang (Perth, WA, Australia) H.-Q. Yu (Hefei, China) J. Yu (Wuhan, China) S. Zeman (Pardubice, Czech Republic) A.I. Zouboulis (Thessaloniki, Greece)

Publication schedule and subscription information

Journal of Hazardous Materials (ISSN 0304-3894). For 2011 volumes 182–192 are scheduled for publication. Subscription prices are available upon request from the Publisher or from the Elsevier Customer Service Department nearest you or from this journal’s website (http://www.elsevier.com/locate/jhazmat). Further information is available on this journal and other Elsevier Products through Elsevier’s website (http://www.elsevier.com). Subscriptions are accepted on a prepaid basis only and are entered on a calendar year basis. Issues are sent by standard mail (surface within Europe, air delivery outside Europe). Priority rates are available upon request. Claims for missing issues should be made within six months of the date of dispatch. Orders, claims, and journal enquiries: please contact the Elsevier Customer Service Department nearest you: St. Louis: Elsevier Customer Service Department, 3251 Riverport Lane, Maryland Heights, MO 63043, USA; phone: (877) 8397126 [toll free within the USA]; (+1) (314) 4478878 [outside the USA]; fax: (+1) (314) 4478077; e-mail: [email protected] Oxford: Elsevier Customer Service Department, The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK; phone: (+44) (1865) 843434; fax: (+44) (1865) 843970; e-mail: [email protected] Tokyo: Elsevier Customer Service Department, 4F Higashi-Azabu, 1-Chome Bldg, 1-9-15 Higashi-Azabu, Minato-ku, Tokyo 106-0044, Japan; phone: (+81) (3) 5561 5037; fax: (+81) (3) 5561 5047; e-mail: [email protected] Singapore: Elsevier Customer Service Department, 3 Killiney Road, #08-01 Winsland House I, Singapore 239519; phone: (+65) 63490222; fax: (+65) 67331510; e-mail: [email protected] doi:10.1016/S0304-3894(11)01250-7

Journal of Hazardous Materials 195 (2011) 1–10

Contents lists available at ScienceDirect

Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat

Review

The modern paradox of unregulated cooking activities and indoor air quality Ki-Hyun Kim a,∗ , Sudhir Kumar Pandey a,1 , Ehsanul Kabir a , Janice Susaya a , Richard J.C. Brown b a b

Department of Environment & Energy, Sejong University, Seoul, Republic of Korea Analytical Science Division, National Physical Laboratory, Teddington TW11 0LW, UK

a r t i c l e

i n f o

Article history: Received 17 June 2011 Received in revised form 10 August 2011 Accepted 11 August 2011 Available online 17 August 2011 Keywords: Food Cooking Hazardous pollutants Cancer Human health Emission factor

a b s t r a c t Pollutant emission from domestic and commercial cooking activities is a previously neglected area of concern with respect to human health worldwide. Its health effects are relevant to people across the globe, not only those using low quality food materials in lesser-developed countries but also to more affluent people enjoying higher quality food in developed countries. Based on the available database of pollutant emissions derived from fire-based cooking, its environmental significance is explored in a number of ways, especially with respect to the exposure to hazardous vapors and particulate pollutants. Discussion is extended to describe the risk in relation to cooking methods, cooking materials, fuels, etc. The observed pollutant levels are also evaluated against the current regulations and guidelines established in national and international legislation. The limitations and future prospects for the control of cooking hazards are discussed. © 2011 Elsevier B.V. All rights reserved.

Contents 1. 2. 3. 4. 5. 6. 7.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pollutants released from cooking fuels and their effect on indoor air pollution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The effect of different cooking methods and ingredients on IAQ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Emission inventories of pollutants released from cooking activities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The level of cooking-related emission in relation to regulations and guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Health and environmental impacts of cooking activities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Future directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction With the progress of healthcare science, human life expectancy has increased gradually over the years. Likewise, with the increasing pervasiveness of advanced civilization and urbanization many primitive risks that threatened human life previously have been reduced or eliminated. As such, the pattern of risks to livelihood and their relative magnitudes have also been altered dramatically. Among many risks in our normal everyday life, not many people are aware of the risks associated with cooking activities. As the use of fire became part of human culture, all populations have become

∗ Corresponding author. Tel.: +82 2 3408 3233; fax: +82 2 3408 4320. E-mail addresses: [email protected], [email protected] (K.-H. Kim). 1 Present address: Department of Botany, Guru Ghasidas Central University Bilaspur (C.G.), 495 009, India. 0304-3894/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jhazmat.2011.08.037

1 2 2 2 5 7 9 9 9

prone to this cooking-related risk, regardless of their race, age, wealth, cultural food preferences, etc. The selection of fire-based cooking approaches such as frying, roasting, and grilling can exert a significant impact not only on the quality of the food but also on pollutant emissions [1,2]. The extent of the latter can be controlled by the combined effects of different recipes, cooking procedures, food materials (and ingredients), fuel types, extraction/ventilation equipment, etc. [3]. The style of these cooking activities and their impact can also be affected by macroscale variables like population, culture, climate, and geographical location. Thus assessment of these cooking related risks becomes a delicate issue with sociological sensitivities, if certain cooking types with deeply ingrained traditional methods are labeled as increased risk. Humans can be subject to cooking-related risks via various intake routes either directly (overcooked foodstuffs) or indirectly (fumes). Our emphasis in this review is directed mainly

2

K.-H. Kim et al. / Journal of Hazardous Materials 195 (2011) 1–10

towards indoor air pollutants (IAP) liberated from fire-based cooking activities, the ubiquitous risks of their exposure, and guidelines/regulations for maintaining indoor air quality (IAQ) from such activities. 2. Pollutants released from cooking fuels and their effect on indoor air pollution As most fire-based cooking cannot be carried out without fuels, the effect of fuel combustion can add to the risks of cooking activities. In fact, cooking fuels are one of the most important causes of IAP, particularly in developing countries [4]. Nearly half of the world’s population use solid fuels and biomass (e.g., coal, wood, animal dung, and crop residues) as their primary energy source [5,6]. Solid fuel is used commonly among the poor, particularly in rural areas in developing countries [7]. In urban areas comparatively clean fuels (like liquid petroleum gas (LPG), biogas, natural gas, and kerosene) are available for cooking purposes, along with electricity [7]. Biomass combustion produces a large number of harmful pollutants including respirable particulate matter (PM10 ), carbon monoxide (CO), nitrogen oxides (NO2 ), formaldehyde, benzene, poly-cyclic aromatic hydrocarbons (PAH), and many other toxic compounds [8,9]. Furthermore, because of the relatively confined nature of indoor spaces with low air turnover rates, pollutants liberated inside will not disperse quickly to sustain low IAQ. Pollutants emitted from fuel consumption (for cooking purposes) have been studied in relation to cooking fuel type as part of IAP research (Tables 1 and 2). Traditional open-fire cooking stoves, used extensively in rural households in many developing countries, generally release high quantity of particles and harmful pollutants in smoke [10–12]. Indoor PM10 concentrations from cooking via biomass combustion were measured as 1545 ␮g m−3 in Kenya [13] and 1200 ␮g m−3 in Mozambique [14]. In rural Bolivia, the 6h mean levels of PM10 , when cooking from indoor and outdoor kitchens by cow dung, were 1830 and 250 ␮g m−3 , respectively [15]. Additionally, common volatile organic compounds (VOCs) such as benzene, toluene, and xylene (commonly called BTX) were significantly higher with the use of biomass fuel relative to natural gas [16]. Likewise, considerable emissions of VOCs and metals have also been detected from combustion of BBQ charcoals produced from several countries [17,18]. As cooking activities occur in the close proximity of people, they are usually exposed to high levels of these pollutants [19]. 3. The effect of different cooking methods and ingredients on IAQ Cooking ingredients vary widely, reflecting local environmental, geographical, economic, and cultural factors. Moreover, cooking

methods are diverse enough to encompass baking, roasting, frying, grilling, barbecuing, smoking, boiling, steaming, microwaving, braising, etc. As such, the cooking of each food reflects its own combinations for the above factors. The data on air pollutant emissions as a function of different food ingredients are summarized in Table 3. Moreover, as the pollutant type and levels are also influenced by the cooking methods, the data for air pollutant emissions is also compiled in relation to cooking style in Table 4. For instance, stir-frying in a wok is the most common cooking practice in China through which many HAPs are released [20]. Schauer et al. [21] estimated emission rates of gas-phase, semi-volatile, and particle-phase organic compounds (C1 to C27) from commercial-scale food cooking operations using seed oils. In Korean-style barbecue restaurants using hot steel pan and broiling steel bars (above a charcoal burner), a list of 99 pollutants (including respirable suspended particulates (RSP), CO, and VOCs) were detected [22]. Lee et al. [20] investigated the IAP at four restaurants in metropolitan Hong Kong and found high concentrations of formaldehyde (177 ␮g m−3 ) and benzene (18.4 ␮g m−3 ) in the dining areas of the Korean-style barbecue restaurant. Furthermore, meat charbroiling was thus identified as one of the previously unconsidered sources of heavy aldehydes in urban air [23]. Acrolein is also released from heated oils during domestic cooking. Heated canola, extra-virgin olive, and olive oils, when heated at 180 ◦ C, were reported to emit acrolein at 52.6, 9.3, and 9.6 mg h−1 L−1 [24], respectively. Moreover, indoor acrolein levels are found to persist for a considerable time (a half-life of 14.4 ± 2.6 h) after cooking under poor ventilation [25].

4. Emission inventories of pollutants released from cooking activities Cooking related emissions can be important sources of major airborne pollutants (e.g., PM, SO2 , and CO) as well as trace-level pollutants (e.g., secondary organic aerosols (SOA), organic carbon (OC), and elemental carbon (EC)). The latter also represents important constituents of the global carbon balance. Emissions inventories for cooking activities are generally lacking, and considered a “missing or unaccounted fraction of the area source category”, regardless of pollutant type. Several researchers estimated emissions inventories for many cooking activities based on various statistical approaches (Table 5), and some limited national inventories exist. In the UK, the National Atmospheric Emissions Inventory (NAEI – www.naei.org.uk) estimated VOC emissions for commercial food production activities such as animal feed manufacture, biscuit, cake and cereal production, coffee roasting, sugar production, and margarine and vegetable oil production.

Table 1 Comparison of indoor air pollutant concentrations measured from different fuels used for cooking (concentrations in ␮g m−3 ). Fuels used for cooking

CO

SO2

Natural gas

4170

185

Coal Charcoal LPG

Biomass

6550

PM10

– 56.2 –

247

710 50 147 744 1545

Benzene 13.7

4.24

30.3

27.3 6.03

70.9 36.3

959

61.5

3322

60.1 135

Toluene 2.70

Xylene 3.81

Formaldehyde 17.2

708 315

– 115 284 –

Wood Dung

436

NO2

54.2 353 1200 300 1830

625

34.2

337

18.5

9.92

Country

References

China Bangladesh China Korea India Malaysia India India Kenya Bangladesh India Mozambique Malaysia Bolivia

[34] [16] [34] [17] [11] [12] [35] [11] [13] [16] [35] [14] [12] [15]

Table 2 Comparison of emission concentrations between different cooking activities in various indoor (and outdoor) environments. CO2 CO (ppm) (ppb)

Environment/ emission source

Hongkong (Lee et al. [20])

1648 15.7 Korean BBQ-style restaurant (indoor) 2344 Chinese hot-pot 8.11 restaurant (indoor) 1031 2.23 Chinese dim sum restaurant (indoor) 636 0.01 Western canteen (indoor) 512 Korean BBQ-style 1.92 restaurant (outdoor) 780 1.21 Chinese hot-pot restaurant (outdoor) 4.2 Chinese dim sum 1.26 restaurant (outdoor) 435 1.13 Western canteen (outdoor) Chinese restaurant (Hunan cooking style emissions) Chinese restaurant (Cantonese cooking style emissions) Chinese restaurant (ambient concentrations) Average of different kitchen environments Average of different Living room environments 13,100 Indoor with traditional stoves Outdoor 1800 Indoor with improved stoves Outdoor Personal (using traditional stoves)a Personal (using improved stoves)a

China (He et al. [49])

Bangladesh (Beghum et al. [50])

Honduras (Clark et al. [51])

PM10 (ppb)

PM2.5 (ppb)

HCHO THC Benzene (ppb) (ppm) (ppb)

Toluene Methyl (ppb) Chloride (ppb)

Chloroform (ppb)

1.44

1.17

0.18

11.40

0.02

0.16

0.001

0.015

0.11

0.08

0.04

8.50

0.01

0.09

0.020

0.010

0.03

0.03

0.02

5.60

0.01

0.08

0.007

0.003

0.04

0.02

0.02

4.00

0.00

0.02

0.001

0.001

0.08

0.06

0.13

6.78

0.01

0.21

0.011

0.004

0.08

0.07

0.06

6.50

0.00

0.03

0.002

0.003

0.08

0.06

0.02

5.10

0.01

0.04

0.014

0.002

0.10

0.07

0.02

4.10

0.01

0.04

0.001

0.001

1.41

0.67

n-Fatty acids (ppm)

Dicarboxylic acids (ppm)

PAH (ppb)

287

8.15

25.0

250

5.14

41.7

357

0.04

65.9

129

97.6

0.31

0.60 0.19

n-Alkanes Total (ppb) carbonyls (ppb)

K.-H. Kim et al. / Journal of Hazardous Materials 195 (2011) 1–10

Country

1.00 0.36 0.27 0.22 0.14 0.07

3

4

Table 2 (Continued) Environment/ emission source

Hongkong (Ho et al. [52])

General Chinese Restaurant Large Medium 1 Medium 2 Small Sinchuan Spicy Food Restaurant Hongkong style Fast food Demonstration Kitchen Single dish 1 Single dish 2 Chinese barbeque kitchen A B C Korean BBQ Western fast-food chain stops A B Western small fast-food chain stops A B Western restaurant

a

CO2 CO (ppm) (ppb)

PM10 (ppb)

PM2.5 (ppb)

HCHO THC Benzene (ppb) (ppm) (ppb)

Toluene Methyl (ppb) Chloride (ppb)

Chloroform (ppb)

n-Fatty acids (ppm)

Honduras. Personal PM2.5 was assessed by attaching the sampler to the participant’s clothing nearest breathing zone and placing the pump in a pack worn around waist.

Dicarboxylic acids (ppm)

PAH (ppb)

n-Alkanes Total (ppb) carbonyls (ppb)

831 96.6 277 289 715 152

226 81.8

179 188 414 473

762 350

113 149 160

K.-H. Kim et al. / Journal of Hazardous Materials 195 (2011) 1–10

Country

K.-H. Kim et al. / Journal of Hazardous Materials 195 (2011) 1–10

5

Table 3 Differences in pollutant emissions between different oil types under varying cooking conditions (mean: mg h−1 L−1 ). Ordera

Oil type/treatments

1

24.7 Canola oil (heated at 180 ◦ C 33 50.5 for 15 h) 25.1 Extra virgin olive oil 46.7 8.6 (heated at 180 ◦ C for 15 h) 30.2 51 7.41 Olive oil (heated at 180 ◦ C for 15 h) ◦ 38.5 Canola oil (heated at 240 C 84.7 51.9 for 7 h) 64.5 124.3 7.09 Extra virgin olive oil (heated at 240 ◦ C for 7 h) 69.2 138.1 7.61 olive oil (heated at 240 ◦ C for 7 h) Safflower oil heated at 210 ◦ C for different intervals (h) 0 317 653 28.7 1 408 947 48.1 349 949 43.7 2 365 489 45.4 3 400 447 45 4 5 375 635 43.3 6 369 611 42.4 Safflower oil heated for 6 h at different temperatures (◦ C) 388 799 47.5 180 369.36 611.1 42.4 210 1658.5 969.2 25.2 240 1371.4 1942.4 80.7 270 ◦ Coconut oil heated for 6 h at different temperatures ( C) 180 370.1 198.7 15.61 210 951.3 255.4 13.88 240 1853.5 493 34.6 3429.6 420.7 18.38 270 Canola oil heated for 6 h at different temperatures (◦ C) 303.07 304 110.7 180 607.6 158.2 210 335.4 240 339.07 806.1 195.3 270 1691.1 1549.1 419 Extra virgin olive oil heated for 6 h at different temperatures (◦ C) 481.1 310.8 27.3 180 210 855.6 939.1 162.5 240 1288.8 1562.5 637.2 270 1392.1 1699.7 736

2

3

4

5

a

Total akanals

Total alkenals (mg h−1 L−1 )

Total alkadienals (mg h−1 L−1 )

Total aldehydes (mg h−1 L−1 )

Total oleic acid derivatives (mg h−1 L−1 )

Total linoleic acid derivatives (mg h−1 L−1 )

Total linolic acid derivatives (mg h−1 L−1 )

108.1

46.3

26.5

34.3

80.4

61.4

13.5

4.1

88.5

70.7

13.8

2.48

175

112

30.7

31.7

195.9

171.1

18.8

2.73

214.9

156.2

44

2.81

999 1403 886 900 892 1053 1022 1235 1023 2653 3395 584 1221 2381 3869 718 1101 1340 3659 819 1957 3489 3828

Source of the data: For order 1 (Fullana et al. [53]) and orders 2 through 5 (Katragadda et al. [54]).

In 2009, these estimates ranged from 70 (for coffee roasting) to 10,400 tonnes (animal feed production). The US EPA [26] estimated emissions from cooking beef and chicken by street vending cooking devices (charcoal grilling). It revealed that marinated meat resulted in higher pollutant emissions than non-marinated meat, while no significant differences exist in emission strengths between meat types. It was also pointed that charcoal did not contribute significantly to the pollutant emissions relative to the meat stuffs. In order to acquire a broader understanding of the emission profile of different cooking activities, Roe et al. [27] developed a national emissions inventory for commercial cooking in the USA (Table 5). Apart from these examples, there are relatively few efforts to develop emission inventories of the range of cooking activities in domestic and commercial settings. As the extent and nature of cooking activities can vary considerably, there is a pressing need to establish emission inventories across a much wider range of geographies, cultures and cooking styles. Such efforts will help us assess both short and long-term health impacts of human exposure to cooking related pollution and accelerate the implementation of regulation to govern safe levels of emission (especially in commercial settings).

5. The level of cooking-related emission in relation to regulations and guidelines To protect the public from the possible health effects of cooking emissions, various regulations and guidelines have been issued by various authorities (Table 6). The pollutants measured from cooking fuels and food smoke were compared based on the literature survey (Table 1). As health criteria for IAP are generally limited, this focuses on CO, BTX, and formaldehyde. Note that CO and xylene however did not exceed any of their regulations and guidelines (Table 6). Kabir et al. [17] reported levels of toluene (625 ␮g m−3 ) that exceeded the chronic-duration inhalation MRL (300 ␮g m−3 ) and the EPA reference air concentration (400 ␮g m−3 ). Their benzene data (315 ␮g m−3 ) likewise exceeded the chronic (10 ␮g m−3 ), intermediate (20 ␮g m−3 ), and acute (30 ␮g m−3 )-duration inhalation MRLs set by ATSDR, and the 30 ␮g m−3 reference air concentration set by EPA. Similarly, formaldehyde levels greatly exceeded the chronic, intermediate, and acute-duration inhalation MRL set by ATSDR, the REL (8-h TWA), and the 15-min ceiling limit set by NIOSH. Aside from charcoal, other cooking fuels can yield considerable emissions. Khalequzzaman et al. [16] reported 17.2 ␮g m−3 of

6

K.-H. Kim et al. / Journal of Hazardous Materials 195 (2011) 1–10

Table 4 Comparison of pollutant emissions between different food/cooking methods. [A] Microwaving popcorn (Rosati et al. [55]) Concentration range in chamber (ng mL−1 )

Compound name Butyric acid Diacetyl Acetoin Propylene glycol 2-Nonanone Triacetin Acetic acid 2-Butoxy-1-methyl-2-oxoethyl ester butanoic acid p-Xylene Pentanal Toluene Hexanal 2-Methyl propanoic acid 2-Octanone Heptanal Benzaldehyde 2-(2-Hydroxypropoxy) 1-propanol Acetophenone Siloxanes 2-Tridecanone 3-Methyl butanal 2-Methyl butanal Furfural 4-Methyl-3-penten-2-one 2-Pentyl furan 2-(2-Ethoxyethoxy) ethanol 2-Ethyl 1-hexanol 3-Hexanone Ethyl ester butanoic acid Butyl ester 2-propenoic acid 2,3-Butanedioldiacetate Cyclotetrasiloxane Decamethyl cyclopentasiloxane Octanoic acid Dodecamethyl cyclohexasiloxane Dodecamethyl pentasiloxane Dihydro-5-pentyl-2(3H)-furanone Octanal Styrene 1-Ethoxy-2-methyl propane Methyl ester octanoic acid Ethyl ester octanoic acid Tridecane 2-(Perfluorooctyl)ethanol 8:2-telomer

0.1–8.6 0.02–5.8 0.01–4.2 0.005–1.3 0.005–1.4 0.01–1.2 0.005–0.5 0.005–0.7 0.01–0.4 0.01–0.02 0.01–0.04 0.01–0.05 0.01–0.27 0.01–1.28 0.01–0.02 0.01–0.02 0.01–0.5 0.015–0.01 0.01–0.03 0.01–0.16 0.01–0.01 0.01–0.03 0.01–0.37 0.01–1.20 0.01–0.01 0.01–0.3 0.01–0.06 0.01–0.17 0.01–0.05 0.01–0.04 0.01–0.33 0.01–0.09 0.01–0.02 0.01–0.16 0.01–0.05 0.01–0.03 0.01–0.08 0.015–0.01 0.01–0.02 0.01–0.02 0.01–0.01 0.01–0.05 0.01–0.05 0.0005–0.009

[B] Emission from combination of food and cooking style (all concentration in ppb: Kabir et al. [2]) Compound

Steamed cabbage

Boiled clam

Hydrogen sulfide Methane thiol Dimethyl sulfide Dimethyl sulfide Acetaldehyde Propionaldehyde Butyraldehyde Isovaleraldehyde Styrene Toluene para-xylene Methylethyl ketone Methylisobutylketone Butylacetate Isobutylalcohol Propionic acid Butyric acid Isovaleric acid Valeraldehyde

0.86 0.15 9.44 1.2 12 0.39 0.39 0.44 0.37 26.3 1.62 3.21 0.04 0.44 0.09 2.27 0.06 3.46 0.06

0.2 0.15 0.26 0.06 18.7 2.81 0.39 0.44 0.31 19.8 1.51 5.45 0.48 0.04 0.09 2.5 0.2 5.75 0.06

Brewed coffee 0.2 13.5 16.9 4.32 153 31.8 77.6 0.44 0.36 24 1.95 52.6 0.04 0.04 3.08 5.84 0.06 15.9 0.06

Fried cabbage

Grilled clam

0.2 63.8 25.6 9.34 12.5 5.4 15.3 0.44 0.07 51.2 1.57 3.21 0.04 0.04 0.09 4.39 0.06 0.05 0.14

39.6 0.15 31.3 35.5 253 8.65 12.9 0.44 0.2 51.1 1.99 28.2 0.04 0.04 3.91 36.1 5.11 1.97 0.12

Roasted coffee 2398 2070 98.7 24.5 5233 366 458 600 8.36 123 0.03 964 0.04 0.04 0.09 695 67 132 8.39

K.-H. Kim et al. / Journal of Hazardous Materials 195 (2011) 1–10

7

Table 5 Emission inventory of air pollutants released from different cooking activities. [A] Emission factors considering a mass balance approach Emission rate (␮g kg−1 )

Cooking activity

Meat charbroiling

Reference

Air pollutant category

Gas phase

Alkanes Olefins Carbonyls Aromatics and napthenes Unidentified organic compounds Aliphatic aldehydes Ketones Alkanoic acids Alkenoic acids Unresolved mixture

1,470,000 2,450,000 5,480,000 200,000 4,590,000

Alkanes

15.8

Saturated n-aldehydes Ketones n-Alkanoic acids Others Olefinic n-aldehydes n-Alkenoic acids Non-extractable organics

138.6 12.9 25.2 25.6

Particle phase [23]

260,000 220,000 480,000 320,000 1,300,000

Stir frying vegetables 0.96

[21]

4.9 0.89 0.72 7.98 0.34

[B] Total emission rate estimated for USA (based on Roe et al. [27]). Air pollutant category

Emission rate (tonnes year−1 ) Conveyorized charbroiling

VOCs PAHs CO PM10 PM2.5

2113 43 7401 8460 8201

Under-fired charbroiling 7234 122 23,662 60,304 58,295

formaldehyde from Bangladeshi cooking fuels which slightly exceeded its chronic-duration inhalation MRL (10 ␮g m−3 ). The same research team measured 13.7 ␮g m−3 of benzene from natural gas which exceeded the 10 ␮g m−3 chronic-duration inhalation MRL set by ATSDR. They also measured 54.2 ␮g m−3 of benzene from biomass which exceeded the chronic, intermediate, and acute-duration inhalation MRLs and the reference air concentration set by the EPA. According to Qing [34], very high levels of SO2 from natural gas (185 ␮g m−3 ) and coal (436 ␮g m−3 ) exceeded not only the acute-duration inhalation MRL (30 ␮g m−3 ) but the air quality guidelines of the WHO (40–60 ␮g m−3 ) and NAAQS (80 ␮g m−3 ). The SO2 emission from coal (436 ␮g m−3 ) further exceeded the NAAQS (24-h exposure limit: 365 ␮g m−3 ) of the EPA and 1-h exposure limit (350 ␮g m−3 ) set by the WHO. Its emission from biomass fuel (61.5 ␮g m−3 ) [35] also exceeded the acute-duration inhalation MRL (40–60 ␮g m−3 ) of the WHO. In addition, acrolein levels reported in Section 3 commonly exceeded many exposure guidelines. Its indoor concentrations (26.4–64.5 ␮g m−3 ) during cooking exceeded not only the inhalation reference concentration (RfC: 0.02 ␮g m−3 ) of the EPA but also the intermediate (0.09 ␮g m−3 ) and acute-duration inhalation MRLs (6.88 ␮g m−3 ) [25]. Considering the frequent exceedance of the IAP due to cooking, one could easily extrapolate its effect on human health. In this regard, it is worth assessing the carcinogenic potentials of the pollutants discussed above. It should be noted that benzene is a known human carcinogen for all routes of exposure based on convincing evidence from both human and animal studies by IARC, EPA, and NTP. Furthermore, formaldehyde has been classified as a probable human carcinogen based on limited (human) and sufficient (animal) evidence [36]. As such, formaldehyde (and toluene) are regulated as hazardous air pollutants (HAPs) by the U.S. Congress [37] and are subject to the regulations for various manufacturing

Deep fat frying 1173

Flat griddle frying 39 41 1941 15,679 11,916

Clamshell griddle frying 940

1073 909

processes and operations [38]. However, as to the carcinogenicity, various regulatory agencies have not yet firmly assigned cancer classifications for xylene, toluene, CO, SO2 , and acrolein or assessed their carcinogenic potential due to inadequacy of data or evidence. 6. Health and environmental impacts of cooking activities There is a line of evidence that cooking related emissions can cause severe health problems. For instance, Yang et al. [39] demonstrated that cooking fume is a major cause of lung cancer in Chinese women. Based on an epidemiological study, Yu et al. [40] also concluded that cumulative exposure to cooking emissions by means of any form of frying could increase the risk of lung cancer for nonsmoking women in Hong Kong. Despite a low smoking rate, these subjects recorded one of the highest non-smoking lung cancer rates worldwide which was ascribed to cumulative exposure to cooking fume rather than to the peak concentrations experienced during cooking [40]. Furthermore the risk of active tuberculosis increased in Indians (aged 20 years and older) cooking with biomass fuel relative to cleaner fuels [41]. This estimate is comparable to the report made by WHO [42] based on non-clinical measures [43]. In addition, chronic exposure to biomass fuel combustion products was also suspected to cause chromosomal and DNA damage and upregulation of DNA repair mechanisms in premenopausal women in rural areas [44]. Evidence also indicates an etiological link between indoor coal burning and lung cancer. For instance, high lung cancer rates in Chinese women were closely associated with the combustion of smoky coal emitting submicron particles with mutagenic organics, especially aromatic and polar compounds [45]. There have been many attempts to estimate the global burden of disease due to the use of solid fuels by applying disease specific

8

K.-H. Kim et al. / Journal of Hazardous Materials 195 (2011) 1–10

Table 6 International and national regulations, advisories, and guidelines issued by various agencies and published by the Agency for Toxic Substances and Diseases Registry (ATSDR).a Agency

Description

International guidelines (Air) ATSDR Acute-duration MRL Intermediate-duration MRL Chronic-duration MRL Cancer classificationb IARC National regulations and guidelines (Air) TLV – 8 h TWA ACGIH STEL Ceiling Limit for Occupation Exposure (TLV-STEL) Cancer classificationc Hazardous Air Pollutants EPA Cancer classificationd Inhalation reference concentration OSHA PEL (8-h TWA) for general, construction, and shipyard industries 15-min STEL Acceptable ceiling concentration Acceptable max. peak above the acceptable ceiling conc. for an 8-h shift for a max. duration of 10 min NIOSH REL (8-h TWA) REL (10-h TWA) REL (15-min ceiling) IDLH STEL NTP Cancer classificatione Agency

Formaldehyde (Ref. [28]) (mg m−3 )

Benzene (Ref. [29]) (mg m−3 )

Toluene (Ref. [30]) (mg m−3 )

0.05 0.04 0.01 Group 2A

0.03 0.02 0.01 Group 1

3.8

1.6f 7.99f

188

0.3 Group 3

0.37 Yes B1

Yes A 0.03 3.19

0.92

A4 Yes D 0.4 754

2.46 1130 1884 0.02

377 0.32g

0.12 24.6

1597g 3.19g A

B

Description

International guidelines Inhalation MRL ATSDR Acute-duration Intermediate-duration STEL (occupational exposure) Cancer classificationb IARC Air quality guidelines WHO 10-min exposure limit 1-h exposure limit 24-h exposure limit Annual arithmetic mean TWA based on effects other than cancer or odor/annoyance: 15 min-TWA 30 min-TWA 1 h-TWA 8 h-TWA National regulations and guidelines (Air) TLV (TWA) ACGIH TLV (ceiling limit) Carcinogenicity classificationc Hazardous Air Pollutants EPA Cancer classificationd Inhalation reference concentration National Ambient Air Quality Standards (NAAQS) 24-h exposure limit Annual arithmetic mean 3-h exposure limit 8-h averaging time 1-h averaging time OSHA PEL (8-h TWA) for general industry REL TWA NIOSH REL (10-h TWA) IDLH STEL Ceiling Cancer classificatione NTP

SO2 (Ref. [31]) (mg m−3 )

CO (Ref. [32]) (mg m−3 )

Acrolein (Ref. [33]) (mg m−3 )

0.003 4.0E−05

0.03 10 Group 3

565

No data

Group 3

0.5 0.35 0.10–0.15 0.04–0.06 100 60 30 10 5.2

29

No No No

0.23f A4 Yes ID 2.00E−05

10h 40h 55

0.23

0.365h 0.08 1.3h

13 5

40 1375 13

0.23 4.59 0.69

229 None

a Definitions: ACGIH = American Conference of Governmental Industrial Hygienists; EPA = Environmental Protection Agency; IARC = International Agency for Research on Cancer; IDLH = immediately dangerous to life or health; MRL = inhalation Minimum Risk Level; NIOSH = National Institute for Occupational Safety and Health; NTP = National Toxicology Program; OSHA = Occupational Safety and Health Administration; PEL = permissible exposure limit; REL = recommended exposure limit; STEL = short-term exposure limit; TLV = threshold limit values; TWA = time-weighted average; WHO = World Health Organization. b IARC cancer classification: Group 1 (carcinogenic to humans); Group 2A (probable human carcinogen), Group 3 (Not classifiable as to carcinogenicity to humans). c ACGIH cancer classifications: A4 (not classifiable as a human carcinogen). d EPA cancer classification: A (known human carcinogen); B1 (probable human carcinogen); D (substances are unclassifiable as to their carcinogenicity); ID (data are inadequate for an assessment of the carcinogenic potential). e NTP cancer classifications: A (substance known to be carcinogenic); B (reasonably anticipated to be a human carcinogen). f Refers to the potential significant contribution to the overall exposure by the cutaneous route, including mucous membranes and the eyes, either by contact with vapors or, of probable greater significance, by direct skin contact with the substance. g NIOSH potential occupational carcinogen. h Not to be exceeded more than once per year.

K.-H. Kim et al. / Journal of Hazardous Materials 195 (2011) 1–10

9

Fig. 1. Global distribution estimates of deaths caused by indoor smoke from solid fuels by WHO sub-region for 2000. Source: Ref. [47].

relative risks or odds ratios to global estimates of the household number relying on such fuels. IAP from solid fuel use and urban outdoor air pollution are in fact estimated to cover 3.1 million premature deaths worldwide every year and 3.2% of the global burden of disease expressed in disability-adjusted life years (DALYs) [46]. If this burden is evaluated on a regional basis, it varies significantly due to many contributing factors (Fig. 1). For instance, almost 80% of ill health effects occur in Africa and Southeast Asia. Globally, indoor smoke from solid fuels ranks as the eighth risk factor. This rises to fourth (after (1) childhood and maternal underweight, (2) unsafe sex, and (3) unsafe water, sanitation, and hygiene) in developing countries (∼40% of the world population) [47]. Indoor smoke risks present with a high mortality thus ranks higher than micronutrient deficiencies and tobacco risks. Hence, to minimize the exposure to cooking related emissions, efforts should be directed to improve cooking devices, development of alternate energy sources (such as sun light), living environments, and cooking behavior. It is important to note that cooking emissions can also exert influences on climate change. For instance, solid fuel dependency exacerbates deforestation which indirectly contributes to the build-up of greenhouse gases (e.g., CO2 ). Deforestation can also cause soil erosion, pollution of streams with sediment and debris, loss of biodiversity, and alteration of vector-borne disease transmission patterns [48]. Moreover, these emissions themselves contain a wide array of pollutants that can contribute to the global climate change.

understanding of the toxic effects of IAP is critical in establishing the safety of indoor air. It is an interesting modern paradox that most legislation in developed countries governs the allowed concentration of HAPs in ambient outdoor air, and yet the citizens which this legislation is designed to protect spend the majority of their time in indoor locations, especially at home, where the concentrations of these regulated pollutants are often much higher. While it is unlikely that regulatory air quality legislation will ever penetrate the threshold of the domestic private home, it seems increasingly important to educate the public of the potential dangers of IAP so that they may take informed choices about their behaviors, and if desired install abatement and extraction technologies to improve their IAQ. Despite the recognition of the potent role of cooking emission, it has scarcely been investigated in relation to possible human health impact. Most studies were directed to emission estimates for selected cooking procedures under certain conditions, while epidemiological studies are lacking for particular cooking procedures. Future research should thus be directed towards a comprehensive survey of the most common cooking procedures with respect to HAPs emission and their direct or indirect impacts on human health and the surrounding atmosphere. Acknowledgements This work was supported by a Grant from the National Research Foundation (NRF) of Korea funded by the Ministry of Education, Science and Technology (MEST) (No. 2010-0007876).

7. Future directions The results of our review confirm that a wide array of HAPs is released during food preparation using common solid fuels and the main materials (and ingredients) for cooking. The level of pollutants released via such activities can pose serious threats to human health, especially to those performing the cooking and their household members. The health effects of such exposure are not restricted to the respiratory tract but can readily cross the alveolar–capillary barrier to reach vital body organs through the circulatory system. It is thus important to monitor IAP in residential settings, restaurant kitchens, and dining areas. In addition, a better

References [1] K. Skog, Cooking procedures and food mutagens: a literature review, Food Chem. Toxicol. 31 (1993) 655–675. [2] E. Kabir, K.-H. Kim, An investigation on hazardous and odorous pollutants emission during cooking activities, J. Hazard. Mater. 188 (2011) 443–454. [3] Q. Zhang, R.H. Gangupomu, D. Ramirez, Y. Zhu, Measurement of ultrafine particles and other air pollutants emitted by cooking activities, Int. J. Environ. Res. Public Health 7 (2010) 1744–1759. [4] D. Behera, G. Aggarwal, Domestic cooking fuel exposure and tuberculosis in Indian women, Indian J. Chest Dis. Allied Sci. 52 (2010) 137–138. [5] A.K.N. Reddy, R.H. Williams, T.B. Johansson, Energy after Rio: Prospects and Challenges, United Nations Publications, New York, 1996.

10

K.-H. Kim et al. / Journal of Hazardous Materials 195 (2011) 1–10

[6] N. Bruce, R. Perez-Padilla, R. Albalak, Indoor air pollution in developing countries: a major environmental and public health challenge, Bull. WHO 98 (2008) 1078–1092. [7] K.-H. Kim, S.A. Jahan, E. Kabir, A review of diseases associated with household air pollution due to the use of biomass fuels, J. Hazard. Mater. 19 (2011) 425–431. [8] K.R. Smith, J.M. Samet, I. Romieu, N. Bruce, Indoor air pollution in developing countries and acute respiratory infections in children, Thorax 55 (2000) 518–532. [9] A.J.S. Jan, Risk assessment of acute health and comfort effects of indoor air pollution, Ann. N. Y. Acad. Sci. 641 (1992) 56–61. [10] D. Schwela, Cooking smoke: a silent killer, People Planet 6 (1997) 24–25. [11] R. Kumar, J.K. Nagar, N. Raj, P. Kumar, A.S. Kushwah, M. Meena, S.N. Gaur, Impact of domestic air pollution from cooking fuel on respiratory allergies in children in India, Asian Pac. J. Allergy Immunol. 26 (2008) 213–222. [12] M.A.K. Lodhi, A. Zain-al-Abdin, Indoor air pollutants produced from fossil fuel and biomass, Energy Convers. Manage. 40 (1999) 243–248. [13] M. Ezzati, D.M. Kammen, Indoor air pollution from biomass combustion and acute respiratory infections in Kenya: an exposure-response study, Lancet 358 (2001) 619–624. [14] A. Ellegard, Cooking fuel smoke and respiratory symptoms among women of low income areas of Mapato, Environ. Health Prospect. 104 (1996) 980–985. [15] R. Albalak, G.J. Keeler, A.R. Frisancho, M. Haber, Assessment of PM10 concentrations from domestic biomass fuel combustion in rural Bolivian highland villages, Environ. Sci. Technol. 33 (1999) 2505–2509. [16] M. Khalequzzaman, M. Kamijima, S. Kiyoshi, B.A. Hoque, Indoor air pollution and the health of children in biomass- and fossil-fuel users of Bangladesh: situation in two different seasons, Environ. Health Prev. Med. 15 (2010) 236–243. [17] E. Kabir, K.-H. Kim, J.W. Ahn, O.F. Hong, J.R. Sohn, Barbecue charcoal combustion as a potential source of aromatic volatile organic compounds and carbonyls, J. Hazard. Mater. 174 (2010) 492–499. [18] J. Susaya, K.-H. Kim, J.-W. Ahn, M.-C. Jung, C.-H. Kang, BBQ charcoal combustion as an important source of trace metal exposure to humans, J. Hazard. Mater. 176 (2010) 932–937. [19] K.R. Smith, Indoor air pollution in developing countries: recommendations for research, Indoor Air 12 (2002) 1–7. [20] S.C. Lee, W.M. Li, L.Y. Chan, Indoor air quality at restaurants with different styles of cooking in metropolitan Hong Kong, Sci. Total. Environ. 279 (2001) 181–193. [21] J.J. Schauer, M.J. Kleeman, G.R. Cass, B.R.T. Simoneit, Measurement of emissions from air pollution sources: 4. C1–C27 organic compounds from cooking with seed oils, Environ. Sci. Technol. 36 (2002) 567–575. [22] S.O. Baek, Y.S. Kim, R. Perry, Indoor air quality in homes, offices and restaurants in Korean urban areas – indoor/outdoor relationships, Atmos. Environ. 31 (1997) 529–544. [23] J.J. Schauer, M.J. Kleeman, G.R. Cass, B.R.T. Simoneit, Measurement of emissions from air pollution sources: 1. C1 through C29 Organic compounds from meat charbroiling, Environ. Sci. Technol. 33 (1999) 1566–1577. [24] A. Fullana, A.A. Carbonell-Barrachina, S. Sidhu, Comparison of volatile aldehydes present in the cooking fumes of extra virgin olive, olive, and canola oils, J. Agric. Food Chem. 52 (2004) 5207–5214. [25] V.Y. Seaman, D.H. Bennett, T.M. Cahill, Indoor acrolein emission and decay rates resulting from domestic cooking events, Atmos. Environ. 43 (2009) 6199–6204. [26] United States Environmental Protection Agency (US EPA), Emissions from street vendor cooking devices (charcoal grilling) EPA/600/Sr-99/048, June 1999. [27] S.M. Roe, M.D. Spivey, H.C. Lindquist, P. Hemmer, R. Huntley, National Emissions inventory for commercial cooking, 2004. Available at: www.p2pays.org/ref/43/42612.pdf. [28] Agency of Toxic Substance and Disease Registry (ATSDR), Toxicological profile for formaldehyde, US DHHS, Public Health Service, 1999. [29] Agency of Toxic Substance and Disease Registry (ATSDR), Toxicological profile for benzene, US DHHS, Public Health Service, 2007. [30] Agency of Toxic Substance and Disease Registry (ATSDR), Toxicological profile for toluene, US DHHS, Public Health Service, 2000. [31] Agency of Toxic Substance and Disease Registry (ATSDR), Toxicological profile for sulfur dioxide, U.S. Department of Health and Human Services (US DHHS), Public Health Service, 1998.

[32] Agency of Toxic Substance and Disease Registry (ATSDR), Draft toxicological profile for carbon monoxide, US DHHS, Public Health Service, 2009. [33] Agency of Toxic Substance and Disease Registry (ATSDR), Toxicological profile for Acrolein, US DHHS, Public Health Service, 2007. [34] Y.H. Qing, Effect of indoor air pollution on respiratory diseases in children, in: Proceeding (in Chinese) of China–Japan environment health conference, Zhengzhou, China, 1992, pp. 53–55. [35] B.K. Padhi, P.K. Padhy, Domestic fuels, indoor air pollution, and children’s health, Ann. N. Y. Acad. Sci. 1140 (2008) 209–217. [36] IRIS, Online. U.S. Environmental Protection Agency, Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati, OH, April 1999. Available at: http://www.epa.gov/ngispgm3/iris/subst. [37] U.S. Congress, Clean air act amendments, Hazardous air pollutants, Section 112, national emissions standards for hazardous air pollutants, Public Law 101-549, U.S. Congress. November 15, 1990. [38] US EPA, Standards of performance for volatile organic compounds (VOC) emissions from synthetic organic chemical manufacturing industry (SOCMI) distillation operations, U.S. EPA, Code of Federal Regulations, 40 CFR 60, Subpart NNN, 1995. [39] S.C. Yang, S.N. Jenq, Z.C. Kang, H. Lee, Identification of Benzo[a]pyrene 7,8Diol 9,10-Epoxide N2-Deoxyguanosine in human lung adenocarcinoma cells exposed to cooking oil fumes from fFrying fish under domestic conditions, Chem. Res. Toxicol. 13 (2000) 1046–1050. [40] I.T.S. Yu, Y.L. Chiu, J.S.K. Au, T.W. Wong, J.L. Tang, Dose–response relationship between cooking fumes exposures and lung cancer among Chinese nonsmoking women, Cancer Res. 66 (2006) 4961–4967. [41] V.K. Mishra, R.D. Retherford, K.R. Smith, Biomass cooking fuels and prevalence of tuberculosis in India, Int. J. Infect. Dis. 3 (1999) 119–129. [42] WHO, Tuberculosis Programme Review, World Health Organization Geneva, India, 1992. [43] B.N. Gupta, N. Mathur, P.N. Mahendra, A.K. Srivastava, A study of the household environmental risk factors pertaining to respiratory diseases, Energy Environ. Rev. 13 (1997) 61–67. [44] N.K. Mondal, B. Mukherjee, D. Das, M.R. Ray, Micronucleus formation, DNA damage and repair in premenopausal women chronically exposed to high level of indoor air pollution from biomass fuel use in rural India, Mutat. Res. 697 (2010) 47–54. [45] IARC, Environmental carcinogens methods of analysis and exposure measurement, Indoor air, Vol. 12, IARC Scientific Publications No. 109, Lyon, France, 1993. [46] World Health Organization (WHO), Global Health Risks: Mortality and Burden of Diseases Attributable to Selected Major Risks, Geneva, 2009. Available at: http://www.who.int/healthinfo/global burden disease/GlobalHealthRisks report full.pdf. [47] World Health Organization (WHO), World Health Report, 2002. Available at: http://www.who.int/whr/en/. [48] Emerging challenges – new findings: emerging and re-emerging infectious diseases: links to environmental change, in: P. Harrison (Ed.), GEO Yearbook 2004/5: An Overview of our Changing Environment, United Nations Environment Programme, Nairobi, 2005. [49] L.-Y. He, M. Hu, X.-F. Huang, B.-D. Yu, Y.-H. Zhang, D.-Q. Liu, Measurement of emissions of fine particulate organic matter from Chinese cooking, Atmos. Environ. 38 (2004) 6557–6564. [50] B.A. Begum, S.K. Paul, M.D. Hossain, S.K. Biswas, P.K. Hopke, Indoor air pollution from particulate matter emissions in different households in rural areas of Bangladesh, Build. Environ. 44 (2009) 898–903. [51] M.L. Clark, S.J. Reynolds, J.B. Burch, S. Conway, A.M. Bachand, J.L. Peel, Indoor air pollution, cookstove quality, and housing characteristics in two Honduran communities, Environ. Res. 110 (2010) 12–18. [52] S.S.H. Ho, J.Z. Yu, K.W. Chu, L.L. Yeung, Carbonyl emissions from commercial cooking sources in Hong Kong, J. Air Waste Manage. Assoc. 56 (2006) 1091–1098. [53] A. Fullana, A.A. Carbonell-Barrachina, S. Sidhu, Volatile aldehyde emissions from heated cooking oils, J. Sci. Food Agric. 84 (2004) 2015–2021. [54] H.R. Katragadda, A. Fullana, S. Sidhu, Á. Carbonell-Barrachina, Emissions of volatile aldehydes from heated cooking oils, Food Chem. 120 (2010) 59–65. [55] J.A. Rosati, K.A. Krebs, X. Liu, Emissions from cooking microwave popcorn, Crit. Rev. Food Sci. Nutr. 47 (8) (2007) 701–709.

Journal of Hazardous Materials 195 (2011) 11–29

Contents lists available at SciVerse ScienceDirect

Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat

Review

A review on techniques to enhance electrochemical remediation of contaminated soils Albert T. Yeung a,∗ , Ying-Ying Gu b,1 a

Department of Civil Engineering, The University of Hong Kong, Pokfulam Road, Hong Kong Department of Environmental Science & Engineering, College of Chemical Engineering, China University of Petroleum (East China), 66 West Changjiang Road, Qingdao 266555, People’s Republic of China b

a r t i c l e

i n f o

Article history: Received 20 June 2011 Received in revised form 15 August 2011 Accepted 15 August 2011 Available online 22 August 2011 Keywords: Electrochemical remediation Soil remediation Enhancement techniques Contaminant solubilization Soil pH control Coupling of remediation technologies

a b s t r a c t Electrochemical remediation is a promising remediation technology for soils contaminated with inorganic, organic, and mixed contaminants. A direct-current electric field is imposed on the contaminated soil to extract the contaminants by the combined mechanisms of electroosmosis, electromigration, and/or electrophoresis. The technology is particularly effective in fine-grained soils of low hydraulic conductivity and large specific surface area. However, the effectiveness of the technology may be diminished by sorption of contaminants on soil particle surfaces and various effects induced by the hydrogen ions and hydroxide ions generated at the electrodes. Various enhancement techniques have been developed to tackle these diminishing effects. A comprehensive review of these techniques is given in this paper with a view to providing useful information to researchers and practitioners in this field. © 2011 Elsevier B.V. All rights reserved.

Contents 1. 2. 3.

4.

5.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Classification of enhancement techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Techniques to solubilize contaminants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Lowering of soil pH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Introduction of enhancement agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1. Chelants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2. Complexing agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.3. Surfactants and cosolvents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.4. Oxidizing/reducing agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.5. Cation solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Soil pH control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Electrode conditioning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Use of ion exchange membrane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Coupling with other remediation technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1. Oxidation/reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2. Bioremediation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3. Permeable reactive barriers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3.1. Lasagna process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3.2. Zero-valent iron (ZVI) PRB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3.3. PRBs of different reactive media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

∗ Corresponding author. Tel.: +852 28598018; fax: +852 25595337. E-mail addresses: [email protected] (A.T. Yeung), [email protected] (Y.-Y. Gu). 1 Tel.: +86 532 86984668; fax: +86 532 86984668. 0304-3894/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jhazmat.2011.08.047

12 12 12 12 13 13 14 14 15 15 16 16 18 18 18 20 21 21 22 23

12

6.

A.T. Yeung, Y.-Y. Gu / Journal of Hazardous Materials 195 (2011) 11–29

5.3.4. PRB – summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4. Phytoremediation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.5. Ultrasonication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.6. Other remediation technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction Electrochemical remediation is a promising technology to remediate fine-grained soils contaminated by inorganic, organic, and mixed contaminants. A direct-current (dc) electric field is imposed on the contaminated soil. The contaminants are migrated by the combined mechanisms of electroosmosis, electromigration, and/or electrophoresis. Detailed description of these fundamental electrokinetic phenomena in soil is given by Yeung [1] and Yeung and Gu [2]. As a dc electric field is a much more effective force in driving fluid through fine-grained soils than a hydraulic gradient [3], electrochemical remediation is particularly applicable to fine-grained soils of low hydraulic conductivity and large specific area. Milestone developments and future research directions of the technology are given in Yeung [4]. However, as the soilchemical fluid system is an electrochemical system [5], many electrochemical reactions are occurring simultaneously during electrochemical remediation of contaminated soil [6]. Moreover, the large specific area of the fine-grained soil provides numerous sites for soil-contaminant interactions. These interactions are soil specific, contaminant specific, dynamic, reversible, and pHdependent. The coupling of electrochemical reactions with the soil-contaminant interactions makes the electrochemical remediation process extremely complex. Similar to most remediation technologies, electrochemical remediation can only extract mobile contaminants from soil [7,8]. Contaminants can exist as sorbed species on soil particle surfaces, sorbed species on colloidal particulates suspended in soil pore fluid, dissolved species in soil pore fluid, or solid species as precipitates. Only contaminants exist as dissolved species in the soil pore fluid or sorbed species on colloidal particulates suspended in soil pore fluid can be extracted by most remediation technologies, and electrochemical remediation is no exception [7]. Therefore, enhancement techniques are developed to solubilize contaminants in soil and to keep them in a mobile chemical state. Electrolytic decomposition of electrolytes occurs at the electrodes, generating H+ ions at the anode (the positive electrode) and OH− ions at the cathode (the negative electrode). These ions are migrated into the contaminated soil, resulting in changes in soil pH as a function of time and space. The change in soil pH can change the chemical states of contaminants, rendering them immobile. It can also change the magnitude and direction of electroosmotic flow, affecting the advective transport of contaminants in soil pore fluid by electroosmosis. Moreover, these ions can polarize the electrodes and reduce the effectiveness of the dc electric field imposed. Therefore, controlling soil pH is very important for the success of electrochemical remediation. In many cases, application of electrochemical remediation alone is not adequate to remediate the contaminated soil to the required acceptance level. Therefore, the technology is enhanced by coupling with other remediation technologies as part of a remediation train of processes. The synergy can achieve results that are better than the sum of technologies applied individually. Many techniques to enhance the extraction efficiency of electrochemical remediation of contaminated soil have been developed throughout the years. A comprehensive review of these techniques

24 24 25 25 26 26 26

is given in this paper to facilitate effective applications of these enhancement techniques by researchers and practitioners in the field of electrochemical remediation of contaminated fine-grained materials such as clay, sediment, and sludge. 2. Classification of enhancement techniques The primary objectives of these enhancement techniques are: (1) to solubilize contaminants in soil and to keep them in mobile states; (2) to control the soil pH within a range of values favoring the application of electrochemical remediation; and (3) to destruct, breakdown, or transform the contaminants simultaneously or sequentially. Therefore, the enhancement techniques are broadly classified into three groups: (1) techniques that solubilize contaminants and keep them in mobile states; (2) techniques that control soil pH; and (3) remediation techniques that can be coupled with electrochemical remediation synergistically to destruct, breakdown, or transform the contaminants simultaneously or sequentially. However, these three groups of techniques are inter-related. Detailed classification of these techniques is presented in Fig. 1. 3. Techniques to solubilize contaminants Contaminants in soil can be sorbed on soil particle surfaces or exist as precipitates in soil pores under certain environmental conditions, rendering them immobile. These contaminants may go into dissolved phases again when the environmental conditions change. Therefore, the temporary immobility of contaminants cannot be considered as permanent containment. However, it does create a difficult hurdle for the remediation process. Enhancement techniques have been developed to solubilize contaminants during electrochemical remediation including: (1) lowering of soil pH; and (2) introduction of enhancement agents. 3.1. Lowering of soil pH Most metals can be solubilized in a low pH environment. During the electrochemical remediation process, H+ ions are generated at the anode and migrated towards the cathode, an acid front is thus developed. A low pH environment can be generated in soil of low acid/base buffer capacity and extraction of metals can be achieved with a reasonable degree of success. For natural soils of high acid/base buffer capacity, strong acids and weak acids have been used as enhancement agents to neutralize the OH− ions generated at the cathode and to lower of the soil pH. Weak acids, such as acetic acid CH3 COOH and citric acid, can also serve as a complexing agent and a chelant, respectively. Strong acids are observed to be more effective than weak acids in many studies. However, it should be noted that when the soil pH is lower than the point of zero charge (PZC) [9], the direction of electroosmotic flow is reversed, i.e., from the cathode towards the anode. The advective transport of contaminant by electroosmosis would diminish the electromigration of cations towards the cathode. Moreover, a very low pH environment developed during the

A.T. Yeung, Y.-Y. Gu / Journal of Hazardous Materials 195 (2011) 11–29

13

Fig. 1. Classification of enhancement techniques for electrochemical remediation.

remediation process may impact the environment adversely and render the remediated soil not readily arable afterwards. More details on lowering of soil pH are presented in Section 4 later in this review paper. 3.2. Introduction of enhancement agents When the acid/base buffer capacity of soil is high, i.e., the resistance of soil to pH change is high, it is very difficult to lower the soil pH by the H+ ions generated by electrolysis or introduction of acid to the soil. Therefore, other enhancement agents have to be utilized to desorb contaminants sorbed on soil particle surfaces and to keep them in the dissolved phase. These enhancement agents include: (1) chelants or chelating agents; (2) complexing agents; (3) surfactants and cosolvents; (4) oxidizing/reducing agents; and (5) cation solutions. 3.2.1. Chelants Chelation is the formation or presence of two or more separate bonds between a bi-dentate or multi-dentate ligand, i.e., the chelant, and a single metal central atom or ion. Chelants can thus desorb toxic metals from soil particle surfaces by forming strong water-soluble complexes which can be removed by the chelant-enhanced electrochemical remediation. An example on how ethylenediamine-N,N -disuccinic acid (EDDS), a biodegradable chelant, solubilizes sorbed Pb from soil particle surfaces is illustrated in Fig. 2 [10]. The chelant-enhanced electrochemical remediation is thus a four-step process: (1) injection of the chelant into the contaminated soil by electroosmosis and/or electromigration; (2) formation of soluble Pb–EDDS complex on soil particle surfaces; (2) dislodgement of Pb–EDDS complex from soil particle surfaces to soil pore fluid; and (3) extraction of Pb as Pb–EDDS complex by electroosmosis and/or electromigration. Chelants, such as carboxylates, organophosphonates, polyamines, and industrial wastewaters [11], have been used or investigated as enhancement agents in electrochemical remediation. Among all the chelating agents, aminopolycarboxylates, such as ethylenediaminetetraacetic acid (EDTA) and (diethylenetriamine)pentaacetic acid (DTPA), and hydroxycarboxylates, such as citric acid, have been most frequently used in electrochemical

remediation. A detailed review of use of chelants in electrochemical remediation is given by Yeung and Gu [2], and will not be repeated in this review paper. In addition to solubilizing sorbed contaminants from soil particle surfaces, chelants also change the zeta potential of soil particle surfaces. In general, chelants lower (becomes more negative) the zeta potential of soil particle surfaces [12]. The lowering of the zeta potential of soil particle surfaces increases the positive electroosmotic volume flow rate of soil pore fluid, i.e., from the anode towards the cathode, facilitating the advective transport of contaminants by electroosmosis towards the cathode. Post-remediation of treatment and disposal of the used extraction fluid is a problem as it is rich in metal-chelant complexes.

Fig. 2. Solubilization of sorbed Pb from soil particle surfaces by EDDS.

14

A.T. Yeung, Y.-Y. Gu / Journal of Hazardous Materials 195 (2011) 11–29

Some chelants, such as EDTA, are toxic, especially in their free forms [13,14], and are poorly photo-, chemo-, or biodegradable in the environment [15]. Different methods to handle used extraction fluids are presented by Lestan et al. [16]. The methods currently available to recover chelants from used extraction fluids still encounter operational difficulties and they work well only for a few contaminants and soil types [2,16]. The development of more robust recycling methods for used chelants would greatly increase the economic value of chelant-enhanced electrochemical remediation. 3.2.2. Complexing agents Complexing agents are chemicals which form coordination complexes with metal ions. Coordination complexes differ from chelate complexes by the formation of only a single bond between the metal central atom or ion and the complexing agent. Some complexing agents, such as I− , Cl− , NH3 − , and OH− , are introduced into soil as conditioning acids or bases during electrochemical remediation process. These ligands can form soluble complexes with metals such as [HgI4 ]2− , [CuCl2 ]− , [CuCl4 ]2− , [Cu(NH3 )4 ]2+ , [Zn(OH)4 ]2− , [Cr(OH)4 ]− , and [Cr(OH)3 ]2− . It has been demonstrated by many researchers [17–20] that mercury could be efficiently extracted by iodide-enhanced electrochemical remediation as soluble complex HgI4 2− . Sulfate reducing bacteria were shown to be a viable tool for treatment of the acidic and oxidative Hg-contaminated iodide waste solution resulting from the enhanced electrochemical remediation [21]. Acetic acid, CH3 COOH, is a complexing agent frequently utilized to enhance electrochemical remediation [22–24]. Although it is not as effective as strong acid such as HNO3 , it is preferred in soil remediation. It can neutralize the electrolysis product at the cathode to reduce energy consumption, and keep the electrolyte pH within a certain range by its acid/base buffer capacity. Moreover, it is relatively cheap, biodegradable, and environmentally safe. Similarly, lactic acid was used to enhance electrochemical remediation of Cu-contaminated soil [25]. Cyclodextrins are nontoxic, biodegradable, and have low affinity of sorption onto the soil particle surfaces in a wide pH range [26]. Moreover, they have the ability to form inclusion complexes with many substrates in aqueous solutions. Hydroxypropyl-␤cyclodextrin, carboxymethyl-␤-cyclodextrin, ␤-cyclodextrin, and methyl-␤-cyclodextrin have been utilized to enhance electrochemical remediation of soils and sediments contaminated with organic compounds and heavy metals [26–32], with varying degrees of success. Ammonium acetate, CH3 COONH4 , was used as anolyte by Chen et al. [33] in their bench-scale experiments on electrokinetic removal of Cu from soil using a constant electrical current density of 1.33 A/m2 . Their results reveal that a concentration of CH3 COONH4 of higher than 0.1 M was needed to sustain the electroosmotic flow. The apparent electrical conductivity of the specimen was controlled by the 10-mm thick layer of soil close to the cathode. The high pH condition in the vicinity of the cathode favors copper–ammonia complex reactions, thus increasing the solubility and removal rate of Cu during electrochemical remediation. The extraction efficiency of Cu increased with the concentration of CH3 COONH4 used. When 0.5 M CH3 COONH4 was used, the proportion of soil containing Cu was less than 10% after treatment. 3.2.3. Surfactants and cosolvents Cationic, anionic, or non-ionic surfactants are amphiphilic compounds containing both hydrophilic groups (heads) and hydrophobic groups (tails). There are both synthetic and natural surfactants. Natural surfactants are also known as biosurfactants, as they are biologically produced from yeast or bacteria from various substrates including sugars, oils, alkanes, and wastes [34].

Fig. 3. Variation of surface tension, interfacial tension, and contaminant solubility with surfactant concentration (after Mulligan et al. [35]).

Surfactants can lower the surface tension of a liquid to allow easier spreading, and the interfacial tension between two liquids, or between a liquid and a solid. Therefore, they may act as adhesives, flocculating agents, wetting agents, foaming agents, detergents, de-emulsifiers, penetrants, and dispersants. Typical desirable functions of surfactants include solubility enhancement, surface tension reduction, critical micelle concentration, wetting ability, and foaming capacity [35]. Surfactant monomers form spheroid or lamellar structures with organic pseudo-phase interiors, which lowers surface or interfacial tensions The minimum concentration at which any added surfactant molecules appear with high probability as micellar aggregates is called the critical micelle concentration (CMC) [36]. The variation of surface tension, interfacial tension, and contaminant solubility with surfactant concentration is schematically shown in Fig. 3. Both synthetic surfactant and natural surfactants can be used as additives in the phase separation processes for remediation of organic compound-contaminated soils by enhancing the aqueous solubility and mobility of organic contaminants [35,37–40]. Moreover, surfactants have been observed by many researchers to be feasible in enhancing heavy metal extraction from soil and sludge [41]. Several factors can adversely affect the efficiency of soil flushing using surfactants including: (1) hardness of groundwater; (2) sorption of surfactants onto clay particle surfaces; (3) inactivation of surfactants due to rapid biodegradation; and (4) difficulties in recovering the surfactant from used flushing solution [42]. Therefore, factors that need to be considered in the selection of surfactants in electrochemical remediation include: (1) efficiency and effectiveness of the surfactant in remediating the contamination; (2) biodegradability of the surfactant and degradation products; (3) toxicity of the surfactant and its degradation products to humans, animals, plants, and the ecology; (4) ability to be recovered, recycled, and reused; (5) public perception and regulatory restrictions; (6) functionality of the surfactant at different pHs; (7) electrical charges, if any, carried by the surfactant; and (8) cost. Overall, desirable surfactant characteristics for soil remediation include biodegradability, low toxicity, solubility at groundwater temperatures, low sorption onto soil particles, effective at concentrations lower than 3%, low soil dispersion, low surface tensions, and low CMC. Anionic and non-ionic surfactants are less likely to be sorbed onto soil particle surfaces but anionic surfactants may precipitate. However, co-injection of an anionic surfactant with a non-ionic surfactant can reduce precipitation and also CMC values [43].

A.T. Yeung, Y.-Y. Gu / Journal of Hazardous Materials 195 (2011) 11–29

Biosurfactants or natural surfactants are known for their biodegradability, reduced toxicity, and environmental-friendliness [44], and they may be less expensive in some cases [45]. Moreover, their efficiency is often higher than those of synthetic surfactants, i.e., a similar surface tension reduction can be achieved by introduction of a smaller quantity of biosurfactant [46]. They are proven to be more tolerant to extreme variations in temperature, ionic strength, and pH [47–49]. Moreover, they may potentially be produced in situ using the organic contaminants as substrates for their production. Biosurfactants are essentially classified either as low- or high-molecular-mass. High-molecular-mass biosurfactants consist of particulate and polymeric amphiphiles. Low molecular-mass biosurfactants can broadly be classified into three groups: (1) glycolipids or lipopolysaccharides, such as rhamnolipids, trehalolopids [50], and sophorolipids [51,52]; (2) lipoproteins–lipopeptides, such as acyclic [53], and cyclic ones (cyclolipopeptides) [54,55]; and (3) hydroxylated cross-linked fatty acids (mycolic acids) or phospholipids [46]. Both synthetic surfactant and natural surfactants have been reported to be efficient in mobilizing organic contaminants during electrochemical remediation of soil contaminated by organic compounds. The feasibility of using synthetic surfactants such as alkyl polyglucoside (APG), sodium dodecyl sulfate (SDS), sodium dodecylbenzenesulfonate (SDBS), Pannox 110, Brij 30, Triton X-100, Griton ALM 100, Calfax 16L-35, Igepal CA-720, Tergitol 15-S-7, Tergitol NP-10, and Tween 80 have been studied by many researchers to enhance electrochemical remediation of soils contaminated by petroleum hydrocarbons [56], polycyclic aromatic hydrocarbons (PAHs) [27,57–64], 1,2-dichlorobenzene (1,2-DCB) [65], hexachlorobenzene (HCB) [29,63], dichlorodiphenyltrichloroethane (DDT) [66], ethylbenzene [67], chlorobenzene [68], trichloroethylene (TCE) [68], and diesel oil [69,70]. The viability of these synthetic surfactants in enhancing electrochemical remediation of soils contaminated by various organic compounds has been established. However, few studies have been carried out on the use of natural surfactants to enhance electrochemical remediation. Nonetheless, rhamnolipid is the most frequently used biosurfactant as an enhancement agent for electrochemical remediation of soils contaminated by organic contaminants. Chang et al. [61] compared the performance of rhamnolipid with Triton X-100, a synthetic surfactant, in enhancing the extraction of phenanthrene from unsaturated soils by electrochemical remediation. Their results indicate that rhamnolipid was more efficient in removing phenanthrene from soil than Triton X-100. Moreover, the electroosmotic flow rate in the rhamnolipid system was higher than that in Triton X-100. In addition to the higher electroosmotic flow rate, the higher remediation efficiency may also be attributed to the promotion of microbial growth in the soil-water system in the presence of rhamnolipid. Gonzini et al. [71] also studied the effects of rhamnolipid on enhancing electrochemical remediation of a gasoil-contaminated soil. Their results indicate that the remediation efficiency of gasoil could be increased up to 86.7% by increasing the dose of rhamnolipid. Moreover, the lower concentration of the gasoil in the liquid phase at the higher concentration of the biosurfactant demonstrated evidently that rhamnolipid could enhance gasoil biodegradation, possibly through two mechanisms: (1) increasing the aqueous solubility of hydrocarbons and thus their bioavailability to microorganisms; and (2) interacting with microorganisms to make their cell surfaces more hydrophobic and thus easier to associate with hydrophobic substrates. They also identified the need for future development on surfactant production by autochthonous microorganisms, so as to reduce the surfactant cost for field application of the technology. Groboillot et al. [72] studied the feasibility of using amphisin, a biosurfactant, to enhance electrochemical remediation of dredged

15

harbor sediments contaminated by PAHs. Their results indicate pure amphisin from Pseudomonas fluorescens DSS73 was more effective in solubilizing and mobilizing PAHs strongly sorbed to sediments than a synthetic anionic surfactant. Amphisin production by bacteria in natural environment was also considered. Although the growth of P. fluorescens DSS73 was weakened by the three model PAHs above saturation, amphisin was still produced. Kaya and Yukselen [73] studied the effects of anionic, cationic, and non-ionic surfactants on the zeta potential of soil particle surfaces of kaolinite, montmorillonite, and quartz powder in the presence of Li+ , Ca2+ , Cu2+ , Pb2+ , and Al3+ . Understanding the variations of zeta potential of soil particle surfaces with the introduction of surfactants is important because the zeta potential controls the direction and rate of electroosmotic flow which impact the contaminant extraction efficiency of electrochemical remediation [1,2]. Their results indicate that the presence of cationic surfactant significantly increases (becomes less negative) the zeta potential of soil particle surfaces in an acidic environment (pH ∼4). The presence of the anionic surfactant makes the zeta potential of soil particle surfaces more negative. However, the non-ionic surfactant has little effect on the zeta potential of soil particle surfaces. They recommended the determination of zeta potential of soil particle surfaces prior to electrochemical remediation to maximize the remediation efficiency of the technique. However, the results of using surfactants to enhance the extraction efficiency of metal contaminants from soil by electrochemical remediation are mixed. Some researchers reported positive results [74,75], while other researchers reported insignificant enhancement [64,76,77]. Cosolvent is a second solvent added in small quantity to the primary solvent to form a mixture that may greatly enhance the solvent power of the primary solvent due to synergism. They can enhance the aqueous solubility of many organic contaminants through cosolvent effect. Several cosolvents, such as ethanol [78,79], n-butylamine [80–82], n-propanol [70], acetone [80], and tetrahydrofuran [80], have been examined for their ability to enhance the solubilization of organic compounds such as PAHs and diesel oil in soil during the electrochemical remediation process. 3.2.4. Oxidizing/reducing agents Oxidizing or reducing agents can be injected into contaminated soil to manipulate the in situ chemistry and microbiology, so as to enhance extraction of contaminants or to reduce their toxicity through oxidation or reduction reactions. Oxidizing agents may include air or oxygen, or chemical oxidants, such as hydrogen peroxide H2 O2 , potassium permanganate KMnO4 or sodium permanganate NaMnO4 , ozone, chlorine, or oxygen releasing compounds. Contaminants are chemically or microbially oxidized. Similarly, reducing agents such as Fe2+ , Fe0 , calcium polysulfide, or sodium dithionite can be used to reduce contaminants in soil. The injection of oxidizing/reducing agents during electrochemical remediation of contaminated soil is equivalent to coupling electrochemical remediation with oxidation/reduction to remediate contaminated soil. Therefore, the subject will be treated in Section 5. 3.2.5. Cation solutions Coletta et al. [83] used natural solutions containing clay extracts and synthetic solutions with varying concentrations of Al3+ , Ca2+ , and Na+ as anodic flushing solutions to investigate the feasibility of enhancing electrochemical remediation of Pb-contaminated clay of initial Pb concentration of 340–410 mg/kg dry clay (dry) and moisture content of 80–83%. Natural flushing solutions were prepared by mixing water and clay in ratios varying from 2:1 to 40:1 by weight, and the supernatant was used as an anodic flushing solution. The 7:1 natural solution was observed to be most effective

16

A.T. Yeung, Y.-Y. Gu / Journal of Hazardous Materials 195 (2011) 11–29

for Pb removal. The solution was composed of 2.289, 0.314, 1.464, 0.803, and 0.908 ppm of Ca2+ , A13+ , Na+ , Mg2+ , and K+ , respectively. Synthetic solutions were prepared using AlCl3 , Ca(NO3 )2 , and NaCl solutions. The Pb extraction efficiency was highest when the solution ionic strength was approximately 0.001 M for each element group, and with trivalent Al3+ and divalent Ca2+ ions at concentrations of 0.064 and 0.31 mM, respectively. Moreover, the 0.31 mM Ca synthetic solution exhibited the highest overall Pb extraction efficiency due to its high ionic mobility, large hydrated ionic radius, and near optimum ionic strength. Energy requirement was determined to be 8–31 kWh/m3 of soil. Reddy and Chinthamreddy [84] investigated the feasibility of enhancing electrochemical remediation of glacial till spiked with Cr6+ , Ni2+ , and Cd2+ of concentrations of 1000, 500, and 250 mg/kg, respectively by simultaneous injection of 0.1 M NaCl from the anode and 0.1 M EDTA from the cathode. Their experimental results indicate that the presence of NaCl sustained the electric current and electroosmotic flow. The remediation efficiency of Cr was increased considerably to 79%. Ni and Cd were migrated significantly towards the anode but eventually accumulated in the soil near the anode. The accumulation of these metals was attributed to the preferential complexation of EDTA with H+ ions in an acidic environment. The thickness of the diffuse double layer around soil particles 1/ (m) is given by 1 = 



εRT 2000 × cF 2 z 2

(1)

where ε is the permittivity of soil pore fluid (F/m); R is the universal gas constant (8.314 J/mol K); T is the absolute temperature (K); c is the concentration of cations in the diffuse double layer (mol/L); F is the Faraday constant (96,485 C/mol); and z is the valence of cations in the diffuse double layer. The electroosmotic volume flow rate is given by Q = ke ie A

(2) (m3 /s);

ke is where Q is the electroosmotic volume flow rate the coefficient of electroosmotic conductivity (m2 /V s); ie is the electrical gradient (V/m); and A is the total cross-sectional area perpendicular to the direction of flow (m2 ). It should be noted that the polyvalent cations injected into contaminated soil may replace contaminant ions or H+ ions in the diffuse double layer of the soil. Cation exchange in clay follows a replaceability series that favors the adsorption of cations of higher valence. If two atoms have the same valence, the larger cation is favored. The order of adsorption is shown with corresponding ionic radii in A˚ as follows [3], Al3+ 0.57

>

Pb2+ 1.19

>

Ca2+ 1.06

>

Mg2+ 0.78

>

K+ 1.33

>

Na+ 0.98

>

Li+ 0.7

In addition to valence and ionic radius, ion concentration is another important factor affecting cation exchange. A cation of high concentration and low replacing power may be preferred to a cation of low concentration and high replacing power [3]. Adsorption of cations of high valence in the diffuse double layer of clay particle surfaces causes a decrease in the thickness of the diffuse double layer around clay platelets as predicted by Eq. (1) [85,86]. Moreover, a high concentration of cations in the soil pore fluid causes further decrease in the thickness of the diffuse double layer and an increase in ionic strength of the system. The decrease in thickness of the diffuse double layer decreases the repulsive forces among clay particles and allows the van der Waals attractive forces among clay platelets to dominate, resulting in flocculation of clay particles. The flocculated structure in the clay fabric causes an increase in porosity and decrease in tortuosity of flow paths, leading to an increase in tortuosity factor, and resulting in increases in the hydraulic conductivity and coefficient of electroosmotic conductivity of the clay, and the effective

ionic mobilities of the ionic species in the soil pore fluid. However, an increase in ionic strength (electrolyte concentration) of the soil pore fluid increases the electrical conductivity of soil and energy consumption of the process. Conversely, a decrease in ionic strength (electrolyte concentration) increases the thickness of the diffuse double layer, leading to a decrease in the coefficient of electroosmotic conductivity and a reduction of electroosmotic flow rate and electromigration of ions. However, the electric current flowing through the soil is reduced, leading to lower energy consumption. As the electroosmotic flow rate and electromigration are the dominant transport mechanisms in electrochemical remediation, there is an optimized ionic strength to maximize the overall remediation efficiency of electrochemical remediation. 4. Soil pH control As a result of electrolytic decomposition of electrolytes at the electrodes, H+ and OH− ions are generated at the anode and the cathode, respectively during the electrochemical remediation process as follows: Oxidation at the anode : Reduction at the cathode :

2H2 O − 4e− → 4H+ + O2 ↑ −

−

4H2 O + 4e → 4OH + 2H2 ↑

(3) (4)

The generated H+ and OH− ions are migrated into the soil by the dc electric field imposed on the soil. As a result, the soil pH near the anode is lowered and that near the cathode is raised. Different techniques have been developed to condition the electrode reservoir solutions, i.e., the anolyte and catholyte, so as to eliminate the adverse impacts of electrode reactions. The primary purpose of electrode reservoir conditioning is to maintain the pHs of anolyte and/or catholyte within appropriate ranges specific to the contaminants being remediated. In most cases, the pH of anolyte is raised and that of catholyte is lowered. The conditioning is particularly important for electrochemical remediation of soils of low acid/base buffer capacity, as the resistances to pH change of these soils are low. Specific objectives of reservoir conditioning include [87]: (1) precipitation of metal contaminants should be avoided and/or precipitates should be solubilized and mobilized; (2) electrical conductivity of the specimen should not be increased excessively in a short duration so as to avoid diminishing of the advective transport of contaminant by electroosmosis prematurely; (3) the electrolysis reaction at the cathode should possibly be depolarized to avoid the generation of OH− ions and their transport into the specimen; (4) the depolarization would also assist in decreasing the electrical potential difference across the specimen and reduce energy consumption of the process; (5) if any chemical is used, the metal precipitate with this new chemical should be soluble within the pH ranges maintained by reservoir conditioning; (6) any special chemicals introduced should not result in any increase in toxic residue in the soil; and (7) the additional cost of chemicals and/or equipment for reservoir conditioning should not increase the overall cost of the electrochemical remediation process significantly. The most frequently used reservoir conditioning techniques in electrochemical remediation are: (1) electrode conditioning by conditioning agents; and (2) use of ion exchange membranes. 4.1. Electrode conditioning Weak acids may be introduced to neutralize the OH− ions generated at the cathode during the electrochemical remediation process. However, improper use of some acids in the process may pose a health hazard. For example, the use of HCl may pose a health hazard as: (1) it may increase Cl− concentration in groundwater; (2) it may promote the formation of some insoluble chloride salts,

A.T. Yeung, Y.-Y. Gu / Journal of Hazardous Materials 195 (2011) 11–29

for example, PbCl2 ; and (3) Cl2 gas will be generated by electrolysis if it reaches the anode. Organic acids, such as CH3 COOH or citric acid, are weak acids that undergo partial dissociation in water. There are several advantages in using these weak acids to depolarize the OH− ions generated at the cathode: (1) they are environmentally safe and biodegradable; (2) they possess certain acid/base buffer capacities so that they can maintain the electrolyte pH to some extent; (3) they are complexing agents that can form soluble complexes with metals to enhance solubilization of heavy metals sorbed on soil particle surfaces and to maintain mobility of heavy metals in soil; (4) the concentration of ions generated by acid dissociation is very low as their pKa values are relatively high, the resulting increase in the electrical conductivity of soil and thus the power consumption are small; and (5) these weak acid ions prevent the formation of other insoluble salts in the vicinity of the cathode, preventing the development of a low electrical conductivity zone and dissipation of excessive electrical energy in the soil near the cathode [2]. Experimental results on removal of Pb from kaolinite by Lee and Yang [88] indicate that external circulation of the electrolyte solution from the cathode reservoir to the anode reservoir could control pore fluid pH, and prevent excessive H+ from decreasing electroosmotic flow rate and excessive OH− from increasing heavy metal precipitation. Saichek and Reddy [89] demonstrated that the use of NaOH to control pH at the anode could improve the extraction efficiency of phenanthrene from kaolin by electrochemical remediation. Experimental results of Hicks and Tondorf [90] on removal of Zn from Georgia kaolinite, a soil of low acid/base buffer capacity, reveal that problems related to isoelectric focusing could be prevented by rinsing away the OH− ions generated at the cathode, achieving an extraction efficiency of 95%. The experimental results of Puppala et al. [87], Rødsand et al. [91], and Reed et al. [92] indicate that the addition of CH3 COOH to the cathode reservoir prevented the development of alkaline conditions in the soil. The technique could improve the extraction efficiency of Pb, as the soil pH nearest to the cathode was lowered to prevent precipitation of Pb(OH)2 . Zhou et al. [93] studied the performance of electrochemical remediation of the low pH Chinese red soil contaminated by Cu and Zn enhanced by catholyte conditioning. Without catholyte conditioning, the soil pH near the cathode was increased from 4.2 to above 6, resulting in accumulation of large quantities of Cu and Zn precipitates in the vicinity of the cathode. Application of lactic acid as catholyte pH conditioning agent improved the extraction efficiency of Cu and Zn from the soil. Increasing the ionic strength of the conditioning agent by adding 10 mM CaCl2 further enhanced Cu removal, but did not cause a significant improvement for Zn extraction. The feasibility of using reservoir conditioning to enhance electrochemical remediation of heavy Cd-contaminated soil was investigated by Gidarakos and Giannis [94]. 0.01 M CH3 COOH or 0.01 M citric acid was used as catholyte to prevent Cd from precipitating as hydroxide. Their results reveal that when the catholyte pH was controlled to be lower than 4, significant amounts of H+ ions produced at the anode could be migrated throughout the specimen, resulting in desorption of Cd from soil particle surfaces and a very high extraction efficiency. Ryu et al. [95] studied the performance of laboratory-scale electrochemical remediation on Cu-, As-, and Pb-contaminated soil enhanced by electrolyte conditioning. Their results reveal that catholyte conditioning using HNO3 increased the removal of Cu and Pb from the soil, and the maximum removal was 60.1% for Cu and 75.1% for Pb. Anolyte conditioning using NaOH enhanced the migration of As which exists in an anionic form and 43.1% of As was removed.

17

Fig. 4. The NEOCHIM electrode (after Leinz et al. [100]).

Genc et al. [96] used CH3 COOH to keep both the anolyte and catholyte at pH ≤4 in their laboratory study on electrochemical remediation of contaminated sediment from Cuyahoga River, OH, USA. The river sediment was contaminated by Mn, Cu, Zn, and Pb. However, the low pH of catholyte generated reverse electroosmotic flow, i.e., from the cathode towards the anode. As a result, they observed the accumulation of Mn near the cathode. However, other metals, such as Cu, Zn, and Pb were mostly in the middle section of the specimen. Moreover, as a result of reverse electroosmotic flow, the extraction efficiencies of metals were low. The highest extraction efficiencies of Mn, Cu, and Pb observed were 18%, 20% and 12%, respectively, and no removal of Zn was observed in all their experiments. Buffer solutions, such as CH3 COOH and NaHCO3 , have also been successfully used to control the pH of electrode reservoir electrolytes so as to control the electroosmotic flow direction and to maintain the electroosmotic volume flow rate during the electrochemical remediation of Pb- or Cd-contaminated Milwhite kaolinite, a natural clay of high acid/base buffer capacity [97,98]. The NEOCHIM technology was developed by the U.S. Geological Survey on the foundation of Russian scientists’ research results on CHIM, a method of electrogeochemical sampling for use in the exploration of buried mineral deposits. A schematic of the NEOCHIM electrode is shown in Fig. 4. The technology solves the problems associated with the presence of H+ and OH− ions in the vicinity of electrodes by using an electrode made of two compartments linked by a salt bridge [99]. The power electrode is immersed in a conducting fluid in the inner compartment where H+ and OH− ions produced by electrolysis are retained and prevented from reaching the outer compartment by the salt bridge. The salt bridge is retained by a semipermeable parchment membrane at the base of the inner compartment. A further conducting fluid is retained by the outer compartment. Electrical contact of the electrode with soil is made through a semipermeable parchment membrane at the base of the outer compartment. The membrane allows the passage of ions from the conducting fluid into the soil and from the soil into the fluid, while retaining the fluid in the compartment. The experimental results of Leinz et al. [100] on electrochemical remediation also indicate the high potential of the NEOCHIM process for the monitoring and remediation of hazardous waste sites.

18

A.T. Yeung, Y.-Y. Gu / Journal of Hazardous Materials 195 (2011) 11–29

4.2. Use of ion exchange membrane Another technique of reservoir conditioning is the use of ion exchange membranes or ion-selective membranes to isolate specific ions generated by electrode reactions from the contaminated soil. Cation exchange membranes essentially allow only cations to pass through, and anion exchange membranes allow only anions to pass. Therefore, a cation exchange membrane installed between the cathode and contaminated soil can prevent the OH− ions generated at the cathode from migrating into the contaminated soil and precipitating with metal contaminants as hydroxides. The metal contaminant cations can be migrated from the soil through the cation exchange membrane into the catholyte to precipitate with the OH− ions on the membrane surface or in the catholyte. However, precipitation with the OH− ions on the membrane surface causes fouling of the membrane. The deterioration of the membrane performance is essentially caused by deposition of foulants on the membrane surface, resulting in an increase in flow resistance of the membrane and decreases in fluxes flowing through the membrane. Nonetheless, the technique is promising as it does not introduce any additional chemicals into the system [101]. However, membrane fouling remains one of the most crucial factors limiting the use of ion exchange membranes in electrochemical remediation [102]. Moreover, the experimental results of Rødsand et al. [91] indicate that the membrane extraction technique did not enhance the extraction of Pb from spiked Norwegian marine clay by electrochemical remediation as expected. Puppala et al. [87] studied the use of NafionTM membrane to limit the transport of OH− ions into soil during electrochemical remediation of an illitic deposit contaminated by Pb. The advantage of the membrane technology is that it would not be necessary to neutralize the cathode by continuous introduction of acid, resulting in considerable saving of acid cost. However, the energy consumption of the process was increased by the electrical resistance of the membrane. Lower expenditures are anticipated if: (a) the membrane is changed periodically and cleaned to prevent fouling, and/or (b) the post-membrane catholyte is removed frequently for precipitation. Moreover, the relative high cost of NafionTM membrane may increase the cost of in situ electrochemical remediation unless the system can be engineered and optimized to decrease the cost during real-life field implementation. Therefore, cost-efficient field techniques should be devised. The results of Li et al. [103] indicate the use of a cation selective membrane installed at the front of the cathode to prevent OH− ion migration towards the anode could greatly enhance the extraction efficiency of Cu by electrochemical remediation. However, they observed that very little Cu ions could penetrate the cation-selective membrane to precipitate in the cathode compartment. Although a cation selective membrane should ideally not permit anions, such as OH− , to enter, most of the Cu precipitated as hydroxides in the compartment between the soil and the membrane, indicating the membrane was not 100% effective and some OH− ions still entered the compartment and precipitated the Cu ions there. Kim et al. [104] installed an anion exchange membrane between the anode and contaminated soil specimen and a cation exchange membrane between the cathode and the soil specimen to enhance electrochemical remediation of a Cd- and Pb-contaminated kaolinite. Moreover, an auxiliary solution cell was installed between the cation exchange membrane and the contaminated soil. Small holes were punched in the membrane to allow OH− ions to move into the auxiliary solution cell from the catholyte so that metal contaminants were precipitated in the auxiliary solution cell instead of at the catholyte. Their results indicate the overall extraction efficiencies of membrane-enhanced electrochemical remediation were improved tremendously due to the prevention of hydroxide

precipitation in the soil and increase in electric current efficiency. Moreover, the installation of the auxiliary solution cell could nullify the fouling problem within the cation exchange membrane and thus improve the overall effectiveness of the electrochemical remediation process. 5. Coupling with other remediation technologies There are many technologies available for remediation of contaminated soil and groundwater [8,105]. They all have their advantages and disadvantages. Some of these technologies can be coupled with electrochemical remediation synergistically so that the coupled remediation efficiency is higher the sum of the individual technologies applied individually. Some of the remediation technologies with feasibility of coupling with electrochemical remediation are presented here. However, the feasibility of many other remediation technologies coupling with electrochemical remediation has yet to be investigated and it should be noted that there are numerous opportunities of coupling these remediation technologies with electrochemical remediation to improve the remediation efficiency of contaminated soil and groundwater drastically for the benefit of mankind and the environment. 5.1. Oxidation/reduction The oxidation/reduction remediation technologies focus on modifying the chemistry and microbiology of the environment by injecting selected reagents into the subsurface to enhance degradation and extraction of contaminants by in situ chemical oxidation/reduction reactions [8]. The technologies are applicable for a wide range of inorganic, organic, and mixed contaminants. The most widely studied and utilized oxidation technology in environmental engineering is probably the Fenton process. All processes that involve catalytic reaction between hydrogen peroxide H2 O2 and Fe2+ ions can be denoted as Fenton processes [106]. The Fenton process involves two major steps: (1) oxidation of Fe2+ ions to Fe3+ ions with decomposition of H2 O2 and generation of hydroxyl radicals, as illustrated in Eq. (5); and (2) degradation of organic contaminants by hydroxyl radicals through oxidation as illustrated in Eqs. (6) and (7), Fe2+ + H2 O2 → Fe3+ + OH• + OH−

(5)

RH + HO• → H2 O + R •

(6)

R•

(7)

+ Fe

3+

→ Fe

2+

+ products

By-products of the chemical reactions presented in Eq. (7) can be further degraded by radical mechanism to complete mineralization. Although Eq. (5) is often referred as the Fenton reaction, other important reactions, such as the occurrence of the Fenton catalytic cycle, also occur: Fe2+ + HO• → Fe3+ + HO− 3+

+ H2 O2 → Fe

2+

2+

•

+ HO2 → Fe

3+

3+

+ HO2 • → Fe

2+

Fe Fe Fe

(8) •

+ HO2 + HO + OH2

−

+ O2 + H

−

(9) (10)

+

(11)

The presence of Fe is catalytic. The hydroxyl radicals so generated are strong and relatively unspecific oxidants that react with most organic contaminants. Therefore, the Fenton process is widely used for the destruction of biorefractory organic contaminants such as benzene, phenols and chlorophenols in wastewater or drinking water. The radicals oxidize the organic molecule by abstracting hydrogen atoms as illustrated in Eq. (6) or by adding themselves to double bonds and aromatic rings. The hydroxyl radicals are only active in aqueous form and thus cannot attack contaminants sorbed

A.T. Yeung, Y.-Y. Gu / Journal of Hazardous Materials 195 (2011) 11–29

on soil particle surfaces [107]. However, it has been demonstrated that it is technically feasible to use high concentration H2 O2 to oxidize contaminants sorbed on soil particle surfaces [108,109], as high concentration H2 O2 favors the generation of highly reactive species, such as HO2 • (hydroperoxyl radicals), O2 •− (superoxide anions), and HO2 •− (hydroperoxide anions), other than hydroxyl radicals. The generation of these non-hydroxyl and highly reactive radicals in the presence of high concentration H2 O2 leads to aggressive reactions that ultimately oxidize the contaminants sorbed on soil particle surfaces [110,111]. However, the Fenton process is effective only at low pHs of 3–5. Therefore, pH adjustment may be required during the remediation process. Recently, there are many investigations into the Fenton-like processes for the degradation of organic contaminants. These processes can be broadly classified into three groups: (1) processes that use ferric salts as catalyst to incite the Fenton reaction, i.e., Eq. (5); (2) processes that use heterogeneous Fenton type catalysts such as iron powder, iron-oxides, iron-ligands, or iron ions doped in zeolites, pillared clays or resins; and (3) processes that use other metal ions, e.g., copper, manganese or cobalt, as catalyst. The major advantages of the Fenton type processes are: (1) they are able to degrade many organic contaminants to harmless or biodegradable products; (2) they use relatively cheap reagents; and (3) the reagents are safe to handle and environmentally benign [106]. The bench-scale laboratory experimental results of Yang and Long [112] and Yang and Liu [113] indicate that it is technically feasible to couple the Fenton-like process with electrochemical remediation using a permeable reactive barrier of granular scrap iron powder to extract and degrade phenol and trichloroethylene (TCE) in situ, respectively. The overall contaminant remediation efficiency is contributed by two mechanisms: (1) destruction of organic contaminants by the Fenton-like process; and (2) extraction of contaminants by electrochemical remediation. Their experimental results also reveal that the percentage of organic contaminant destruction increased with the quantity of scrap iron powder used in the process. However, a larger quantity of scrap iron powder embedded in soil would decrease the coefficient of electroosmotic conductivity, resulting in lower efficiency of advective transport of the contaminant by electroosmosis and thus lower contaminant extraction efficiency. Moreover, the smaller was the size of the scrap granular iron powder, the higher was the destruction efficiency, but the lower was the overall contaminant remediation efficiency. Kim et al. [114] explored the feasibility of coupling the Fenton process with electrochemical remediation to remediate phenanthrene-contaminated EPK kaolinite, using the iron minerals on soil particle surfaces as catalyst. Their results reveal that the intermediate anions, i.e., HO2 − and O2 •− , generated by the Fentonlike reactions changed the electrical current intensity significantly. The addition of 0.01 N H2 SO4 to the anode reservoir improved the stability of H2 O2 and treatment efficiency of phenanthrene in the soil specimen. More than a half of the spiked phenanthrene was destructed or extracted after 21 days of treatment. Therefore, the use of H2 O2 and dilute acid, as an anode purging solution, is a feasible technology for the remediation of halogenated organic compound-contaminated soil of low hydraulic conductivity, low acid/base buffer capacity, and high iron content. Kim et al. [115] attempted to remediate phenanthrene-contaminated Hadong clay similarly, however, the acid/base buffer capacity of Hadong clay is high due to its high carbonate content. Their results reveal that the presence of carbonates of high acid/base buffer capacity reduced the stability of H2 O2 and treatment efficiency of phenanthrene, and confirmed that the Fenton reaction is effective only at low pHs of 3–5.

19

Different methods have been attempted to overcome the problem of high acid/base buffer capacity of soil. Kim et al. [116] studied the stabilizing effects of phosphate and sodium dodecyl sulfate (SDS) on H2 O2 during electrochemical remediation of phenanthrene-contaminated Hadong clay coupled with the Fenton-like process. Both stabilizers decreased (becomes more negative) the zeta potential of soil particle surfaces due to complexation of phosphate and SDS with oxides, resulting in increase of electroosmotic volume flow rate. Complexation with phosphate hindered the migration of dissolved Fe ions towards the cathode significantly. However, SDS could dissolve the Fe ion from the Fe oxide of soil and transport the dissolved Fe ions towards the cathode. Nonetheless, transition metal complexation with phosphate and SDS improved the stability of H2 O2 , in particular, in the high pH region near the cathode by SDS. The increase of H2 O2 stability allowed more reaction time for the Fenton-like process, resulting in better treatment efficiency of phenanthrene. Kim et al. [117] studied the performance of H2 SO4 and HCl injected from the anode for pH control in the remediation of phenanthrene-contaminated Hadong clay by electrochemical remediation. When H2 SO4 was utilized, the reduced species of sulfate may increase the decomposition rate of H2 O2 near the anode significantly as follows: SO4 2− + 2e− + 2H+ → SO3 2− + H2 O −

(12)

−

HSO3 + H2 O2 → SO2 OOH + H2 O

(13)

SO2 OOH− + H+ → SO4 2− + 2H+

(14) HS− ,

Moreover, reduced sulfur species, such as H2 S and accumulated in the region near the cathode due to the reducing environment of the region. The generation of these sulfur species is accompanied by a significant stoichiometric decrease of H+ ions in the soil pore fluid, SO4 2− + 8e− + 10H+ → H2 S + 4H2 O

(15)

resulting in a sharp increase in soil pH, rapid decomposition of H2 O2 , and generation of O2 gas. Such decomposition of H2 O2 was not observed in experiments using HCl as the pH control agent. Moreover, H2 O2 may be re-generated near the cathode by the reaction, O2 + 2H+ + 2e− → H2 O2

(16)

The remediation efficiency of phenanthrene-contaminated soil by the Fenton-like process is dependent on both the extent of degradation and migration by electroosmosis. Alcantara et al. [118] studied the electrochemical remediation of phenanthrene-contaminated kaolinite of initial concentration of 500 mg/kg of soil. Electrochemical remediation alone resulted in negligible remediation of phenanthrene. Fenton-like reaction was thus generated in kaolinite which was also contaminated by Fe. When both the anode and cathode reservoirs were filled with 10% H2 O2 , an overall extraction and destruction efficiency of phenanthrene of 99% was obtained in 14 days by applying an electrical gradient of 300 V/m across the soil specimen. It should be noted that the soil pH was maintained at approximately 3.5 without pH control, favoring the Fenton-like processes. Reddy and Karri [119] applied electrochemical remediation enhanced by the Fenton-like process to kaolin contaminated with a mixture of Ni and phenanthrene each at a concentration of 500 mg/kg of dry soil. The objective of the coupled remediation processes was simultaneous oxidation of phenanthrene and extraction of Ni. Experiments were conducted using H2 O2 solution in concentrations of 5%, 10%, 20%, and 30% and deionized water as control. Native Fe was used as catalyst for the Fenton-like process. A dc electrical gradient of 1 V/cm was applied and H2 O2 solution was

20

A.T. Yeung, Y.-Y. Gu / Journal of Hazardous Materials 195 (2011) 11–29

introduced at the anode for 4 weeks. The volume of electroosmotic flow was substantial in all the experiments, approximately one pore volume in the control experiment and 1.2–1.6 pore volumes in the H2 O2 experiments. Oxidation of phenanthrene increased with concentration of H2 O2 and a maximum of 56% oxidation was observed with 30% H2 O2 . Nickel was migrated from the anode towards the cathode but it was precipitated near the cathode as a result of the high pH environment. They concluded that optimization of H2 O2 /catalyst concentration and electrical gradient applied, and control of soil pH are required to improve the efficiency of oxidation of phenanthrene and extraction of Ni simultaneously. Oonnittan et al. [120] studied the feasibility of electrochemical remediation of hexachlorobenzene (HCB)-contaminated kaolin enhanced by the Fenton process with and without using ␤cyclodextrin to enhance the solubility of HCB in the soil pore fluid. The initial concentration of HCB in kaolin was 100 mg/kg of soil. The native iron in kaolin was utilized to catalyze the Fenton-like reaction and no soluble iron was added during the process. After 15 days of treatment, a maximum remediation efficiency of 76% was observed when 30% H2 O2 was used in the absence of ␤cyclodextrin. However, the introduction of ␤-cyclodextrin as an enhancing agent led to a slower rate of oxidation. Tsai et al. [121] studied the feasibility of electrochemical remediation of diesel-contaminated soils enhanced by the use of 0.1 M NaCl as purging solution and corroded iron electrodes. Their experimental results indicate the concentration of total petroleum hydrocarbon diesel in the contaminated soil was reduced from 10,000 to 300 mg/kg by electrokinetically enhanced oxidation in the presence of both 8% H2 O2 and Fe3 O4 (corroded iron electrodes), i.e., remediation efficiency of 97%. However, individually applied electrochemical remediation and Fenton oxidation can only yield remediation efficiencies of 55% and 27%, respectively. The synergistic effect of the two remediation technologies is thus evident. Oonnittan et al. [122] identified the importance of efficient oxidant delivery methodologies for effective contaminant oxidation to occur. The success of electrochemical remediation coupled with the Fenton process depends heavily on the good contact between the contaminant and the oxidant facilitated under optimized reaction conditions. Isosaari et al. [123] coupled persulfate oxidation with electrochemical remediation to cleanup creosote-contaminated soil for 8 weeks. Their results reveal that electrokinetically enhanced oxidation with sodium persulfate Na2 S2 O8 resulted in remediation efficiency of creosote removal of 35% which is better than that of electrochemical remediation of 24% or persulfate oxidation of 12% individually. The oxidant generated more positive redox potential than electrochemical remediation alone. Moreover, the persulfate treatment decreased the electroosmotic volume flow rate. The results of elemental analyses indicate decrease in the natural Al and Ca concentrations, increase in Zn, Cu, P, and S concentrations, and migration of several metal cations towards the cathode. The effectiveness of electrokinetically enhanced persulfate oxidation for destruction of TCE spiked in a sandy clay soil was evaluated by Yang and Yeh [124]. Their experimental results indicate that electroosmosis could greatly enhance the transport of the injected Na2 S2 O8 from the anode reservoir to the cathode reservoir via the contaminated soil, enhancing the in situ chemical oxidation of TCE. Moreover, the injection of nano-scale Fe3 O4 was observed to have a profound impact in the activation of persulfate oxidation. Reddy and Chinthamreddy [125] studied the electromigration of Cr6+ , NI2+ , and Cd2+ in clayey soils containing different in situ reducing agents in bench-scale experiments. Two different clays, kaolin and glacial till, were used with or without a reducing agent. Kaolin is a soil of low acid/base buffer capacity and glacial till is a soil of high acid/base buffer capacity. The reducing agent used was humic acid, ferrous sulfate, or sodium sulfide of concentration of

humic acid, Fe2+ , and S− of 1000 mg/kg soil. The soils were then spiked with Cr6+ , Ni2+ , and Cd2+ in concentrations of 1000, 500, and 250 mg/kg, respectively, and treated by an electrical gradient of 1 V/cm for more than 200 h. The reduction of chromium from Cr6+ to Cr3+ was completed prior to electrochemical remediation. Their results indicate that the extent of Cr6+ reduction was dependent on the type and quantity of reducing agent in the soil in the order of sulfide > ferrous iron > humic acid. Moreover, electromigration of Cr6+ was significantly retarded in the presence of sulfide because of: (1) the opposite directions of migration of Cr6+ and Cr3+ ; (2) sorption and precipitation of Cr3+ in high pH regions near the cathode in kaolin and throughout the glacial till; and (3) sorption of Cr6+ in low pH regions near the anode in both soils. Both Ni2+ and Cd2+ were migrated towards the cathode in kaolin. However, the migration was significantly retarded in the presence of sulfide due to the pH increase throughout the soil. The initial high pH conditions within the glacial till caused Ni2+ and Cd2+ to precipitate, so the effects of reducing agents were inconsequential. The study demonstrated evidently that the reducing agents, particularly sulfide, in soils may affect the redox chemistry and pH of the soil, ultimately affecting the remediation efficiency of electrochemical remediation. Weeks and Pamukcu [126] conducted a study to demonstrate the feasibility of in situ reduction of Cr6+ to Cr3+ by introducing ferrous iron Fe2+ , a reducing agent, to the contaminated soil electrokinetically. Their results indicate that the Cr6+ in soils could be effectively reduced to Cr3+ by electrochemical remediation. Moreover, they demonstrated that the Nernst equation may be applicable to model the soil-water system to estimate the concentrations of different Cr species after electrochemical remediation.

5.2. Bioremediation Bioremediation is the use of microorganisms (mainly bacteria) to decompose hazardous contaminants, transform them to less harmful forms, and/or immobilize them under suitable environmental conditions [8]. The success of bioremediation requires the simultaneous existence of microorganisms, contaminants (food for the microorganism), electron acceptors, and essential nutrients for the microorganisms to grow. In fine-grained soils of low hydraulic conductivity, it is difficult to supply microorganism and the required electron acceptors or nutrients to the contaminants, or to supply the contaminants to natural occurring microorganisms. Electrokinetics-enhanced bioremediation or bioelectrokinetics is the technology that couples bioremediation with electrochemical remediation by supplying the microorganisms, electron acceptors, or nutrients to the contaminants, or migrating the contaminants to the microorganisms by electrokinetic flow processes. The ability to directionally transport bacteria from injection points into zones of contamination is a distinct advantage of electrokinetics-enhanced bioremediation for in situ remediation [127]. Electroosmosis and/or electrophoresis have been utilized successfully to inject a Pseudomonas strain (bacterial cell capable of degrading diesel) into diesel-contaminated soil [128]; Sphingomonas sp. L138 and Mycobacterium frederiksbergense LB501TG (polycyclic aromatic hydrocarbon-degrading bacteria) into model aquifers made of glass beads, alluvial sand from Lake Geneva, and historically polluted clayey soil in the laboratory [129]; Pseudomonas putida, Bacillus subtilis, and Klebsiella pneumoniae to stimulate bacterial cell migration and biodegradation of crude oil in soil [130]; Sphingomonas sp. LB126 (fluorene-degrading bacteria) into a laboratory model aquifer [131]; Bacillus spp. (nitrate reducing bacteria) to remove nitrate from soil [132]; B. subtilis LBBMA 155 and nitrogen-starved cells of Pseudomonas sp. LBBMA 81 into a residual

A.T. Yeung, Y.-Y. Gu / Journal of Hazardous Materials 195 (2011) 11–29

soil [133]; and Acidithiobacillus thiooxidans (sulfur-oxidizing bacteria) into tailing soil contaminated by Cd, Cu, Pb, Zn, Co, and As [134]. However, it was observed that electrokinetic transport of strongly charged and highly adhesive cells of M. frederiksbergense LB501TG in different model aquifers was poor [129]. Lee and Kim [135] injected A. thiooxidans (sulfur-oxidizing bacteria) into shooting range soil contaminated by Cu, Zn, and Pb. The bioleaching process improved the extraction efficiencies of Cu and Zn by electrochemical remediation. However, PbSO4 , a byproduct of sulfur oxidation, existed as precipitates and was immobile. Nonetheless, the problem was overcome by subsequent injection of EDTA. Electrokinetics was used successfully to inject ammonium nitrogen into fine-grained soil [136], benzoic acid cometabolite into TCE-contaminated soil [137], acetate and phosphate amendment into Cr6+ -contaminated soil [138], KH2 PO4 and triethyl phosphate into kaolin soils [139], oxygenated and nutrient-rich liquid into creosote-contaminated soil [140], and nitrate to toluenecontaminated soil under denitrifying conditions [141]. The results of Schmidt et al. [142] indicate the feasibility of injecting nitrate and ammonium into a very humid clayey silt of high plasticity, high electrical conductivity, low hydraulic conductivity, low density, high acid/base buffer capacity, and high cation exchange capacity. However, injection of phosphorous into this type of soil did not prove to be successful. Lohner et al. [143,144] studied the distributions of microbial electron acceptors nitrate and sulfate and of the nutrients ammonium and phosphate by electrokinetics in a model sandy soil. Their results reveal that the ion distribution in the soil was significantly influenced by the pH profile and the imposed electrical gradient. The results of Xu et al. [145] reveal that ammonium and nitrate ions could be distributed more uniformly in phenanthrene contaminated-soil by reversal of electrode polarity. The results of Jackson et al. [146] indicate electrokinetics could enhance the bioremediation of 2,4-dichlorophenoxyacetic acid-contaminated soil by increasing the bioavailability of the contaminant to microorganism. Similar observation was made by Fan et al. [147] during their study on in situ bioremediation of 2,4dichlorophenol-contaminated soil. Wu et al. [148] demonstrated experimentally that electrokinetic injection of lactate, a negatively charged biodegradable organic, in sand was dependent on electric current density. However, the increase in electric current intensity did not result in a proportional increase in lactate transport due to development of an appreciable electroosmotic flow from the anode to the cathode. Tiehm et al. [149] observed that the microbial activities of vinyl chloride degrading microorganisms were inhibited by electrochemical reaction products when stainless steel electrodes and titanium electrodes with mixed oxide coating type DN201 were used. However, when the electrodes were separated from the microorganisms by bipolar membranes, no inhibition by the electric field was observed. Li et al. [150] demonstrated that a dc electric current could stimulate microbial activities and accelerate the biodegradation of petroleum, and there is a strong positive correlation between the electric intensity and the bioremediation efficiency of petroleum. The results of Wick et al. [151] suggest that the presence of an electric field, if suitably applied, would not influence the composition and physiology of soil microbial communities and hence would not affect their potential to biodegrade subsurface contaminants. The results of Kim et al. [152] also suggest that the application of electrokinetics could be a promising soil remediation technology if soil parameters, electric current, and electrolyte were suitably controlled based on the understanding of interaction between electrokinetics, contaminants, and indigenous microbial community. Moreover, the increase in soil temperature during electrochemical remediation promotes microbial activities in general.

21

However, the microbial activities can also be inhibited if the soil temperature is higher than 45 ◦ C. 5.3. Permeable reactive barriers A permeable reactive barrier (PRB) is an engineered barrier made of reactive treatment media placed across the flow path of a contaminant plume in aquifer that removes or degrades contaminants in the groundwater flowing through it [153]. It relies on the flow of contaminants through the barrier. The use of granular zero-valent iron (ZVI) mixed with soil to construct PRBs for entrapping or decomposing contaminants in the subsurface has gained widespread acceptance by the environmental remediation and regulatory communities in recent years. Considerable investigation has been conducted to understand the interfacial chemistry of granular iron, and the sorption and degradation mechanisms of contaminants. When a PRB is coupled with electrochemical remediation, the flow of contaminants through the barrier is not provided by the advective transport of contaminants driven by the natural hydraulic gradient of groundwater. It is driven by the electroosmotic flow of soil pore fluid, electromigration of charged species, and/or electrophoresis of charged particulates. In most cases, particularly in fine-grained soils, these transport mechanisms are far more significant than that driven by the natural hydraulic gradient of groundwater. The sorption characteristics of most solid particle surfaces are pH-dependent. The degradation reactions of many contaminants are also pH-dependent. As a result, the pH gradient generated by the electrochemical remediation process in the PRB may affect the sorption and degradation mechanisms of the reactive medium in the PRB. The use of enhancement agents in electrochemical reaction would further complicate the situation. Moreover, it is possible to construct a PRB in the subsurface by electrokinetic flow processes. Therefore, there are many additional aspects that need to be considered when a PRB is coupled with the electrochemical remediation process to improve the remediation efficiencies of organic, inorganic, and mixed contaminants. 5.3.1. Lasagna process Electrochemical remediation is coupled with sorption/degradation of contaminants in treatment zones installed directly in contaminated soils in the Lasagna process. The Lasagna process is an in situ remediation technique that applies the concept of Integrated In situ Remediation [154]. A dc electric field is applied to migrate the contaminants from soil into treatment zones where the contaminants are removed by sorption, immobilization, or degradation as shown in Fig. 5. The technique is called “Lasagna” because of the layered appearance of electrodes and treatment zones. Theoretically, it can remediate organic, inorganic, and mixed contaminants. Electrodes and treatment zones can be of any orientation depending upon the emplacement technology used and the characteristics of the site and contaminant. The treatment process is composed of these key steps [154]: (1) Highly permeable zones in close proximity of the contaminated soil are created by hydrofracturing or similar technologies. Appropriate materials such as sorbents, catalytic agents, microbes, oxidants and buffers are introduced to these highly permeable zones to transform them into treatment zones. (2) Electrokinetic flow processes are utilized to migrate contaminants from soil into treatment zones. Since these zones are located close to each other, the time taken for the contaminants to move from zone to zone can be very short. (3) For highly non-polar contaminants, surfactants can be introduced into the fluid or incorporated into the treatment zones

22

A.T. Yeung, Y.-Y. Gu / Journal of Hazardous Materials 195 (2011) 11–29

5.3.2. Zero-valent iron (ZVI) PRB Chew and Zhang [161] investigated the feasibility of electrochemical remediation coupled with a ZVI PRB installed at the anode to treat a nitrate-contaminated soil according to the chemical reaction, 5Fe0 + 2NO3 − + 12H+ → 5Fe2+ + N2 ↑ + 6H2 O

Fig. 5. Principle of Lasagna Process (after Ho et al. [154]).

to solubilize the organics. For a mixture of organics and metals, the treatment zones can contain sorbents for binding the metals and/or microbes or catalysts for degrading the organics. (4) If needed, the fluid flow direction can be reversed periodically by switching the electrical polarity. The operation would enable multiple passes of the contaminants through the treatment zones for complete sorption/destruction. The polarity reversal also serves to minimize complications associated with longterm operation of uni-directional electrokinetic flow processes. The high pH cathode effluent can be re-circulated through the contaminated soil when the polarity of the electric field applied is reversed, i.e., the cathode has been reversed to become the anode and vice versa. The recycling of effluent provides a convenient means for pH neutralization of the contaminated soil and minimization of wastewater generation.

The technique has been proved to be technically feasible in bench-scale laboratory experiments on the degradation of paranitrophenol in kaolinite [154] and field-scale experiments on remediation of TCE-contaminated soils at various sites [155–158]. Jackman et al. [159] demonstrated the feasibility of migrating 2,4-dichlorephenoxyacetic acid in contaminated silt soil by electrokinetic flow processes into microorganism active treatment zone for biodegradation of organic contaminants. A bench-scale experiment was conducted by Ma et al. [160] to investigate the simultaneous removal of 2,4-dichlorophenol (2,4DCP) and Cd from a sandy loam by the Lasagna process using a newtype of bamboo charcoal as sorbent and periodic polarity reversals at different intervals. Their results indicate that the Lasagna process was effective in the simultaneous extraction of 2,4-DCP and Cd from sandy soil. Moreover, the extraction efficiencies were higher when the electrical polarity was reversed at 24-h intervals.

(17)

The amount of nitrate–nitrogen transformed by electrochemical remediation was increased significantly by coupling with a ZVI PRB. The major transformation products were ammonia–nitrogen and nitrogen gas. Moon et al. [162] investigated the mechanisms of TCE degradation during electrochemical remediation coupled with a ZVI PRB. Their results indicate the rate of reductive dechlorination of TCE was improved 1.3–5.8 times of that of a ZVI PRB alone. The most effective configuration of electrode and ZVI PRB for TCE removal was with the cathode installed at the hydraulic down-gradient. The enhancement was attributed to the availability of more electron sources including: (1) the dc power supply; (2) electrolysis of water; (3) oxidation of ZVI; (4) oxidation of dissolved Fe2+ ; (5) oxidation of molecular hydrogen at the cathode; and (6) oxidation of Fe2+ in mineral precipitates. Each of these electron sources was evaluated for their potential influences on the TCE removal capacity through the electron competition model and energy consumption. A strong correlation between the quantity of electrons generated, removal capacity, and energy-effectiveness was identified. Yuan [163] investigated the effect of ZVI PRB position and ZVI quantity on the efficiency of electrochemical remediation of tetrachloroethylene (PCE)-contaminated clay coupled with a ZVI PRB. The PRB was composed of 2–16 g of ZVI mixed with Ottawa sand in a ratio of 1:2 by weight. Her results indicate that the best position of the PRB was at the cathode and the remediation efficiency of PCE was 2.4 times that of electrochemical remediation alone. The remediation efficiency also increased with the quantity of ZVI in the barrier. The highest remediation efficiency of 90.7% was observed when the quantity of ZVI in the barrier was increased to 16 g. Moreover, it was observed that the more was ZVI in the barrier, the higher was the electroosmotic flow rate, and the lower was final soil pH after treatment. The effectiveness of a ZVI PRB barrier installed at the middle of the soil specimen during electrochemical remediation of hyperCr6+ -contaminated clay (2497 mg/kg) was investigated by Weng et al. [164]. The barrier was composed of 1:1 ratio of granular ZVI and sand by weight. Their results indicate that the migration of H+ ions was greatly retarded by the strong opposite migration of anionic CrO4 2− ions, resulting in a reverse electroosmotic flow and development of alkaline zone across the specimen. The alkaline environment promoted the release of Cr6+ from the clay. Chromium removal was indicated by the high Cr6+ concentration in the anolyte and the presence of Cr3+ precipitates in the catholyte. The reduction efficiency of Cr6+ to Cr3+ was increased by the ZVI PRB. The electrochemical remediation coupled with a ZVI PRB has transformed the contaminant in the hyper-Cr6+ -contaminated soil to the less toxic form of Cr3+ . Yuan and Chiang [165] investigated the removal mechanisms of As from soil by electrochemical remediation coupled with a PRB made of ZVI and FeOOH. The extraction efficiency for As was increased by 60–120% by the PRB. The best performance was achieved when a FeOOH layer was installed at the middle of the soil specimen. The improvement was attributed to higher surface area of FeOOH and the migration of HAsO4 2− towards the anode by electromigration. The presence of As on the surface of the reactive media of the PRB was confirmed by results obtained by SEM coupled with energy dispersive spectroscopy. Moreover, the extraction of As contributed by surface sorption/precipitation on the PRB reactive media was much more than that by the electrokinetic flow

A.T. Yeung, Y.-Y. Gu / Journal of Hazardous Materials 195 (2011) 11–29

23

voltage. The hydraulic conductivity and unconfined compressive strength of the iron-rich band so produced were 1 × 10−9 m/s or less and 10.8 N/mm2 , respectively. The barrier may function as a PRB to degrade contaminants or an impervious barrier to contaminant transport. By monitoring the dc electric current intensity passing through the barrier, the integrity of the Fe-rich band may be assessed. Moreover, the barrier may ‘self-heal’ by continuing application of a dc electric current.

Fig. 6. Principle of iron-rich barrier generation by electrokinetics (after Faulkner et al. [169]).

processes. However, electromigration was a more dominant contaminant migration mechanism than the advective transport by electroosmosis. Cang et al. [166] investigated the feasibility of treating a Crcontaminated soil by electrochemical remediation coupled with a ZVI PRB. The reactions between Cr6+ in groundwater and the ZVI in the PRB are as follows: 2Fe0 + Cr2 O7 2− + 7H2 O → 2Cr(OH)3 ↓ + 2Fe(OH)3 ↓ + 2OH− (18) Fe0 + CrO4 2− + 4H2 O → 2Cr(OH)3 ↓ + Fe(OH)3 ↓ + 2OH− ZVI is ultimately oxidized to Fe3+

(19)

which precipitates as Fe(OH)3 , while Cr6+ is reduced to Cr3+ and also precipitates in the PRB. During the reactions, the OH− ions released increases the soil pH and decreases the sorption capacity of Cr6+ on soil particle surfaces. Their results indicate that the technique was feasible for the remediation of Cr-contaminated soil. The maximum remediation efficiency of Cr achieved was 72%. The quantities of Cr in the anolyte and catholyte with a PRB were smaller than those without. The position of the PRB affected both the direction and rate of electroosmotic flow. The optimum positions of the PRBs are between the contaminated soil specimen and the electrodes. Wan et al. [167] investigated the feasibility of surfactantenhanced electrochemical remediation coupled with a PRB composed of microscale Pd/Fe for the treatment of a HCBcontaminated soil. The reduction kinetics of HCB by nanoscale Pd/Fe bimetallic particles was faster than that by nanoscale Fe particles. The degradation products of HCB using nanoscale Pd/Fe bimetallic particles have less chloro substituents than those using nanoscale Fe particles. The effects can be attributed to the catalytic effect of Pd on the Fe surface [168]. The nonionic surfactant Triton X-100 was selected as the solubility-enhancing agent. Their results indicate that HCB removal was generally increased by a factor of 4 as HCB was removed from soil through several sequential processes: (a) advective transport of HCB from the anode towards the cathode by electroosmosis; (b) complete sorption/degradation by the reactive Pd/Fe particles in the PRB; and (3) probable electrochemical reactions near the cathode. ZVI PRBs may be constructed in situ by electrokinetics. Faulkner et al. [169] have successfully generated subsurface barriers of continuous Fe-rich precipitates in situ by electrokinetics in their laboratory-scale experiments. Continuous vertical and horizontal Fe-rich bands up to 2 cm thick have been generated by applying a voltage of less than 5 V over a period of 300–500 h, using sacrificial iron electrodes 15–30 cm apart as shown in Fig. 6. The Fe-rich barrier is composed of amorphous iron, goethite, lepidocrocite, maghemite, and native iron. The applied dc electric field dissolved the sacrificial anode and injected the Fe ions into the soil. The Fe ions then re-precipitated in an alkaline environment to form the barrier. The thickness of the Fe-rich band increased with the applied

5.3.3. PRBs of different reactive media Chung and Lee [170] investigated the potential use of atomizing slag as an inexpensive PRB reactive medium coupled with electrochemical remediation for simultaneous treatment of soil contaminated by TCE and Cd by laboratory-scale experiments. Their results indicate that the TCE concentration of the effluent through the PRB during electrochemical remediation were much lower than that of electrochemical remediation alone. Some of the TCE passing through the PRB would have been dechlorinated by the atomizing slag as indicated by the higher chloride concentration of the effluent. In general, both the remediation efficiencies of TCE and Cd achieved approximately 90%. The removal rate of Cd from the soil specimen was higher than that of TCE as a result of the additional transport by electromigration due to its positive charge. Kimura et al. [171] investigated the possibility of coupling electrochemical remediation with a ferrite treatment zone (FTZ) to treat Cu-contaminated kaolinite. The FTZ was constructed between the cathode and contaminated kaolinite of soil containing polyferric sulfate solution so that the concentration of ferrite in the FTZ was 1000 ppm (mg/kg). Their results indicate 92% of Cu ions in contaminated kaolinite were migrated into the FTZ by electrochemical remediation and ferritized by the alkaline environment generated by the process after 48 h of treatment. The Cu ions were insolubilized by the ferrite reagent in the FTZ and accumulated as copper-ferrite through these chemical reactions [172], nCu2+ + (3 − n)Fe2+ + 6OH− → Cun Fe(3−n) (OH)6

(20)

Cun Fe(3n−1) (OH)6 + (1/2)O2 → Cun Fe(3−n) O4 + 3H2 O

(21)

Barrado et al. [173] suggested the co-precipitation mechanism for Fe2+ and divalent or polyvalent metal ions as follows, xCu2+ +FeSO4 +6NaOH + (1/2)O2 → Cux Fe(3−x) O4 + 3Na2 SO4 + 3H2 O + x[Fe2+ ]

(22)

The copper-ferrite precipitates are magnetic and can be separated from solution easily. Therefore, the advantages of coupling a FTZ with electrochemical remediation include: (1) it is possible collect the extracted heavy metals in a specific FTZ; (2) the treatment of a large quantity of Cu-rich wastewater produced by electrochemical remediation can be avoided; and (3) there is a possibility that copper-ferrite can be recovered by magnetic separation. It is envisaged in field implementation that the FTZ can be constructed near the cathode by injecting ferrite reagent into the soil, and the contaminated water is migrated to the FTZ by the electrochemical remediation process. Afterwards, the FTZ is excavated and the Cu is recovered by appropriate processes, such as soil washing using acid and magnetic separation. The feasibility of electrochemical remediation of Crcontaminated clay enhanced by a PRB made of transformed Red Mud (TRM) was investigated by De Gioannis et al. [174] in bench-scale experiments. The TRM is primarily composed of micron-sized NaOH etched aggregates of (hydrated) Fe oxides (hematite and ferrihydrite 35% by weight) and hydrated alumina (boehmite and gibbsite 20% by weight). These are impregnated by newly formed and more or less soluble alkaline minerals, including sodalite (15% by weight), Ca(OH)2 , hydroxycarbonates

24

A.T. Yeung, Y.-Y. Gu / Journal of Hazardous Materials 195 (2011) 11–29

and calcium alumino silicates (portlandite, calcite, cancrinite, hydrocalumite and aluminohydrocalcite 15 wt%), Mg(OH)2 and magnesium alumino silicates (brucite and hydrotalcite 4 wt%). Their results reveal that the remediation efficiency of Cr6+ was proportional to treatment duration. The acidic environment near the anode generated by the electrochemical remediation process improved the sorption capacity of TRM for metal-oxyanions. Therefore, the PRB made of TRM installed near the anode improved the remediation efficiency of metal-oxyanions by electrochemical remediation. Yuan et al. [175] investigated the feasibility of surfactantenhanced electrochemical remediation of 1,2-dichlorobenzene (1,2-DCB)-contaminated soil coupled with a carbon nanotube (CNT) PRB installed at the middle of the specimen. CNT is becoming a prominent material being applied in the removal of aqueous and gaseous pollutants due to its high specific area, high reaction ability, and high electron transfer capacity [176–179]. It is highly expected that CNT will become an effective reactive medium in the PRB for removal of organic contaminants from the subsurface. Their results indicate the remediation efficiency of electrochemical remediation could be significantly improved by the introduction of SDS and coupling with a CNT PRB. Removal of 1,2 DCB was primarily contributed by surface sorption of the contaminant on CNT rather than by electrokinetic flow processes. However, electrophoresis of anionic SDS micelles towards the anode became a more critical contributor when the surfactant was used as processing fluid. An enhanced electrochemical remediation process coupled with a PRB made of carbon nanotube coated with cobalt (CNT-Co) was investigated for As5+ removal by Yuan et al. [175]. Their experimental results indicate the PRB made of CNT did not contribute much to the remediation efficiency of As5+ . However, the PRB made of CNTCo increased the remediation efficiency from 35% to 62%. The better remediation efficiency of electrochemical remediation enhanced by the PRB made of CNT-Co was attributed to the higher sorption of As5+ onto CNT-Co surfaces than CNT surfaces. Removal of As5+ was thus primarily contributed by the surface sorption of As5+ onto CNT-Co instead of the electrokinetic flow processes. The surface characteristics of CNT-Co, as revealed by SEM coupled with energy dispersive spectroscopy, evidently confirmed that As was adsorbed on the passive layer surface. The results of an investigation using sequential extraction revealed that the binding between As5+ and soil particles was shifted considerably from strong binding forms, i.e., Fe–Mn oxide, organic, and residual, to weak binding forms, i.e., exchange and carbonate, after electrochemical remediation. Han et al. [180] investigated the feasibility of enhancing the electrochemical remediation of Cu-contaminated kaolinite by coupling with a PRB made of carbonized foods waste (CFW). The CFW is composed of more than 85% oxygen, calcium and carbon. The size range of the CFW is 75–150 ␮m within porous structures. The specific area, total pore volume, and average pore diameter of ˚ CFW were determined to be 14.16 m2 /g, 46.9 mm3 /g, and 132.4 A, respectively. The sorption efficiency of CFW used as a PRB reactive medium was found to be 4–8 times more efficient than that of zeolite. Throughout the experiment, an electrical gradient of 1 V/cm was implemented and acetic acid was injected from the anode to improve the remediation efficiency. Their results indicate the installation of a CFW PRB did not influence the electroosmotic flow. However, the electroosmotic flow was increased by the injection of CH3 COOH with time. The majority of Cu2+ extracted from kaolinite was sorbed by CFW. 5.3.4. PRB – summary The remediation efficiency of electrochemical remediation can be enhanced by coupling with a PRB. Depending on the type of contaminant to be treated, different reactive media of the PRB can be utilized. However, experimental results from different researchers

to date indicate the remediation efficiency is primarily contributed by the sorption capacity of the reactive medium of the PRB. The role of electrochemical remediation lies in the migration of contaminants towards the PRB, generation of an acidic environment near the anode, and generation of an alkaline environment near the cathode. However, the sorption capacity of the reactive medium of the PRB can be promoted by the acidity or alkalinity of the environment to improve the remediation efficiency. 5.4. Phytoremediation Phytoremediation is the use of plants to remove, degrade, or sequester inorganic and organic contaminants from soil and/or groundwater [8]. It is an emerging cost-effective alternative to conventional remediation technologies. However, contaminants may have limited bioavailability in the soil, methods to facilitate its transport to the shoots and roots of plants are thus required for successful application of phytoremediation. O’Connor et al. [181] investigated the use of coupling phytoremediation with electrochemical remediation to decontaminate soils contaminated by Cu, Cd, and As. It can be observed in their results that the dc electric field could transport metal contaminants from the anode towards the cathode, and generate significant changes in soil pH. Moreover, perennial ryegrass could be grown in the treated soils to take up a proportion of the mobilized metals into its shoot system. In their bench-scale studies, Lim et al. [182] demonstrated the effectiveness of Indian mustard (Brassica juncea) grown in contaminated soil in accumulating high tissue concentration of Pb, with the addition of EDTA in the soil and the application of a dc electric field around the plants. The accumulation of Pb in the shoots using EDTA and a dc electric field was increased by two- to fourfold that of using EDTA only. Similarly, the shoot Cu concentrations of ryegrass in the phytoremediation of contaminated soil enhanced by EDTA and EDDS was increased by 46% and 61%, respectively when coupled with electrochemical remediation [183]. Aboughalma et al. [184] studied the use of potato tubers to decontaminate soils polluted with Zn, Pb, Cu, and Cd in their laboratory-scale experiments using: (1) a dc electric field; (2) an alternating-current (ac) electric field; and (3) no electric field, i.e., the control. Their results reveal that metal accumulation in plant roots treated with electrical fields was generally higher than the control. The overall metal uptake in plant shoots treated with a dc electric field was lower than those treated with an ac electric field and the control, although there was a higher accumulation of Zn and Cu in the plant roots treated with electrical fields. The Zn uptake in plant shoots treated with an ac electric field was higher than that treated with a dc electric field and the control. Zn and Cu accumulation in plant roots treated with a dc electric field and an ac electric field were similar and higher than that of the control. Bi et al. [185] studied the growth of rapeseed (Brassica napus) plants and tobacco (Nicotiana tabacum) plants under a dc electric field and an ac electric field and their abilities to decontaminate a soil contaminated by Cd, and a soil contaminated by Cd, Zn, and Pb. Their results reveal that the biomass production of rapeseed plants was enhanced by the ac electric field. However, the ac electric field has no effect on the biomass production of tobacco plants and the dc electric field even has a negative effect. Moreover, metal uptake by the rapeseed plant shoot was enhanced by the application of the ac electric field. Cang et al. [186] studied the effects of dc electric current on the growth of Indian mustard (B. juncea) and speciation of soil heavy metals in pot experiments for 35 days. The soil was contaminated by Cd, Cu, Pb, and Zn. Their results indicate that plant uptake of metals was increased by the electrokinetics-assisted phytoremediation. Moreover, electrical gradient was identified to be the most

A.T. Yeung, Y.-Y. Gu / Journal of Hazardous Materials 195 (2011) 11–29

important factor in affecting the plant growth, soil properties, and metal concentrations in the soil and plant. 5.5. Ultrasonication Acoustic waves or acoustic energy can enhance migration of contaminants in soil and facilitate their subsequent remediation and/or removal by these effects: (1) increase in kinetic energy of soil pore fluid, causing rise in temperature, and increase in volume and pressure of the soil pore fluid; (2) decrease in viscosity of soil pore fluid, increasing the volume flow rate of the soil pore fluid; (3) increase in molecular motion of contaminants, inducing disintegration and mobilization of contaminants sorbed on soil particle surfaces; and (4) cavitation (forming bubbles) in soil pore fluid, causing increase in porosity and hydraulic conductivity of soil [187]. Chung and Kamon [187] studied the performance of electrochemical remediation coupled with ultrasonication on simultaneous remediation of Pb and phenanthrene from contaminated natural clay. Their bench-scale experiments were conducted using specially designed and fabricated devices. Their experimental results reveal that both the fluid outflow rate and remediation efficiencies for both heavy metal and PAH were increased by the coupled remediation technologies in comparison to electrochemical remediation alone. The average outflow rate was increased from 120 mL/h to 143 mL/h, an increase of 19% by the coupling effects of electrokinetic and ultrasonic phenomena. The average remediation efficiency for Pb was increased from 88% to 91%, an increase of 3.4%; and the average remediation efficiency for phenanthrene was increased from 85% to 90%, an increase of 5.9%. Chung [188] evaluated the performance of four remediation technologies, i.e., soil flushing, electrochemical remediation, ultrasonication, and electrochemical remediation coupled with ultrasonication, in the remediation of river sand from Korea contaminated by diesel fuel and Cd. His results indicate the coupled remediation technologies increased both the volume flow rate and contaminant extraction efficiencies. After 100 min, the final accumulated flow volume was 2200 mL, 2400 mL, 3800 mL, and 4000 mL by soil flushing, electrochemical remediation, ultrasonication, and electrochemical remediation coupled with ultrasonication, respectively. The final accumulated flow volume was thus increased by 9%, 73%, and 82% by electrochemical remediation, ultrasonication, and electrochemical remediation coupled with, ultrasonication, respectively. The remediation efficiencies for diesel fuel were 65%, 67%, 85%, and 87% by soil flushing, electrochemical remediation, ultrasonication, and electrochemical remediation coupled with ultrasonication, respectively, Similarly, the remediation efficiencies for Cd were 62%, 76%, 65%, and 83%, respectively. It is evident that electrochemical remediation coupled with ultrasonication is the most effective technique to extract heavy metal and hydrocarbon simultaneously from the contaminated sandy soil. Moreover, electrochemical remediation was observed to be the most effective method for the treatment of heavy metal, e.g., Cd, while ultrasonic remediation was the most effective for hydrocarbon, e.g., diesel fuel. As a result, the coupled techniques can be used effectively to extract both the heavy metal and hydrocarbon from contaminated soils simultaneously. Pham et al. [189] studied the performance of electrochemical remediation enhanced by ultrasonication in the cleanup of kaolin contaminated by a mixture of three persistent organic pollutants: HCB, phenanthrene, and fluoranthene. Their bench-scale experimental results conclude that the remediation efficiencies for these three persistent organic pollutants by electrochemical remediation coupled with ultrasonication was higher than those of electrochemical remediation alone. Although the ultrasonic enhancement could increase both the electric current intensity and electroos-

25

motic volume flow rate, it could only increase the remediation efficiency for less than 10%. HCB is the most difficult contaminant to extract because of its high stability, while fluoranthene is the easiest contaminant to extract. Enhancement of electrochemical remediation by ultrasonication can be considered as one of the feasible technology to extract PAHs from contaminated soil. Pham et al. [162] studied the feasibility of using electrochemical remediation enhanced by ultrasonication or the surfactant 2-hydroxylpropyl-␤-cyclodextrin to remediate soil contaminated by the hydrophobic compounds of HCB and phenanthrene. Their results indicate that both contaminants could be mobilized by electrochemical remediation enhanced by either ultrasonication or surfactant. However, it is more difficult to extract HCB because of its stability and low water-solubility. Moreover, remediation of phenanthrene enhanced by ultrasonication was more efficient than that by surfactant, as ultrasound can degrade the contaminant through oxidation by free radicals. Shrestha et al. [190] utilized electrochemical remediation coupled with ultrasonication to treat kaolin contaminated by chrysene of concentrations of 25, 50, 75, and 100 mg/kg. Their results indicate the coupled technologies could improve the remediation efficiency of electrochemical remediation. Moreover, the remediation efficiency decreased with increase in the initial concentration of chrysene.

5.6. Other remediation technologies There are many mature remediation technologies for contaminated soil and groundwater [8,105], and they can potentially be coupled with electrochemical remediation to enhance their individual remediation efficiencies synergistically. However, these possibilities have yet to be investigated. For example, production of electrochemical oxidation equivalents in situ by inserting anodes in contaminated soil appears to be a promising idea, but the approach is proven to have poor remediation efficiency and the effects are much localized in soil. However, in situ production of oxidants has many advantages: (1) oxidants of short lifetimes can be used in the remediation process; (2) no stabilization of peroxides is necessary; (3) the hazard of storing large quantities of chemicals is avoided; and (4) logistics of handling chemicals is much simpler. A new approach is being investigated by Wesner et al. [191] to separate the in situ production of tailored oxidants and the transport of the oxidants by electrokinetics. Thermal desorption is a technology that heats contaminated soil or sludge in situ or ex situ to volatize the contaminants and remove them from soil [8]. Volatile and semi-volatile organics are removed from contaminated soil in thermal desorbers at 100–300 ◦ C for low-temperature thermal desorption, or at 300–550 ◦ C for hightemperature thermal desorption [192]. When a dc or ac electrical current is flowing through a contaminated soil, resistive or ohmic heating occurs. The heating can be used to accelerate many chemical and biological reactions occurring in the contaminated soil, and to modify many physical properties of contaminants. For example, heating can be used to increase desorption of many organic contaminants from soil particle surfaces and to remove dense non-aqueous phase liquids. Increased temperature may increase the aqueous solubility, decrease the density, decrease the viscosity, and increase the volatilization of organic contaminants, facilitating their transport in soil. Elevated temperatures not exceeding the temperature tolerance of microbial consortia can increase their metabolic activity and bioavailability, resulting in enhancement of biodegradation of organic contaminants. However, these thermal effects during electrochemical remediation have not been well studied to date [193].

26

A.T. Yeung, Y.-Y. Gu / Journal of Hazardous Materials 195 (2011) 11–29

6. Conclusions Electrochemical remediation is a promising technology for the remediation of fine-grained soil contaminated by inorganic, organic, and mixed contaminants. However, enhancement techniques are often required to improve the remediation efficiency of the technology. A comprehensive review on techniques to enhance electrochemical remediation of contaminated fine-grained materials is given in this paper. A comprehensive and updated list of references is also provided for the reader who is interested in a particular enhancement technique to perform further study. Acknowledgements The sabbatical leave of the first author in California supported by the Faculty of Engineering of The University of Hong Kong to complete these review papers is appreciated. The financial support for the doctoral study of the second author provided by The University of Hong Kong is gratefully acknowledged. References [1] A.T. Yeung, Electrokinetic flow processes in porous media and their applications, in: M.Y. Corapcioglu (Ed.), Advances in Porous Media, vol. 2, Elsevier, Amsterdam, 1994, pp. 309–395. [2] A.T. Yeung, Y.-Y. Gu, Use of chelating agents in electrochemical remediation of contaminated soil, in: D.C.W. Tsang, I.M.C. Lo (Eds.), Applications of Chelating Agents for Land Decontamination Technologies, ASCE Press, Reston, 2011. [3] J.K. Mitchell, K. Soga, Fundamentals of Soil Behavior, 3rd ed., John Wiley & Sons, Hoboken, 2005. [4] A.T. Yeung, Milestone developments, myths, and future directions of electrokinetic remediation, Sep. Purif. Technol. 79 (2011) 124–132. [5] A.T. Yeung, Fundamental aspects of prolonged electrokinetic flows in kaolinites, Geomech. Geoeng.: Int. J. 1 (2006) 13–25. [6] A.T. Yeung, Geochemical processes affecting electrochemical remediation, in: K.R. Reddy, C. Cameselle (Eds.), Electrochemical Remediation Technologies for Polluted Soils, Sediments and Groundwater, John Wiley & Sons, Hoboken, 2009, pp. 65–94. [7] A.T. Yeung, Contaminant extractability by electrokinetics, Environ. Eng. Sci. 23 (2006) 202–224. [8] A.T. Yeung, Remediation technologies for contaminated sites, in: Y. Chen, X. Tang, L. Zhan (Eds.), Advances in Environmental Geotechnics, Zhejiang University Press, Hangzhou, 2009, pp. 328–369. [9] G. Sposito, On points of zero charge, Environ. Sci. Technol. 32 (1998) 2815–2819. [10] D. Lestan, B. Kos, Soil washing using a biodegradable chelator, in: B. Nowack, J.M. VanBriesen (Eds.), Biogeochemistry of Chelating Agents, ACS Symposium Series 910, American Chemical Society, Washington, DC, 2005, pp. 383–397. [11] Y.-Y. Gu, A.T. Yeung, Desorption of cadmium from a natural Shanghai clay using citric acid industrial wastewater, J. Hazard. Mater. 191 (2011) 144–149. [12] Y.-Y. Gu, A.T. Yeung, A. Koenig, H.-J. Li, Effects of chelating agents on zeta potential of cadmium-contaminated natural clay, Sep. Sci. Technol. 44 (2009) 2203–2222. [13] M. Sillanpaa, A. Oikari, Assessing the impact of complexation by EDTA and DTPA on heavy metal toxicity using Microtox bioassay, Chemosphere 32 (1996) 1485–1497. [14] N. Dirilgen, Effects of pH and chelator EDTA on Cr toxicity and accumulation in Lemma minor, Chemosphere 37 (1998) 771–783. [15] B. Nortemann, Biodegradation of EDTA, Appl. Microbiol. Biotechnol. 51 (1999) 751–759. [16] D. Lestan, C.L. Luo, X.D. Li, The use of chelating agents in the remediation of metal-contaminated soils: a review, Environ. Pollut. 153 (2008) 3–13. [17] C.D. Cox, M.A. Shoesmith, M.M. Ghosh, Electrokinetic remediation of mercurycontaminated soils using iodine/iodide lixiviant, Environ. Sci. Technol. 30 (1996) 1933–1938. [18] K.R. Reddy, C. Chaparro, R.E. Saichek, Removal of mercury from clayey soils using electrokinetics, J. Environ. Sci. Health A – Tox. Hazard. Subst. Environ. Eng. 38 (2003) 307–338. [19] P. Suèr, T. Lifvergren, Mercury-contaminated soil remediation by iodide and electroreclamation, J. Environ. Eng., ASCE 129 (2003) 441–446. [20] Z.M. Shen, J.D. Zhang, L.Y. Qu, Z.Q. Dong, S.S. Zheng, W.H. Wang, A modified EK method with an I− /I2 lixiviant assisted and approaching cathodes to remedy mercury contaminated field soils, Environ. Geol. 57 (2009) 1399–1407. [21] T. Hakansson, P. Suer, B. Mattiasson, B. Allard, Sulphate reducing bacteria to precipitate mercury after electrokinetic soil remediation, Int. J. Environ. Sci. Technol. 5 (2008) 267–274. [22] O. Hanay, H. Hasar, N.N. Kocer, O. Ozdemir, Removal of Pb from sewage sludge by electrokinetics: effect of pH and washing solution type, Environ. Technol. 30 (2009) 1177–1185.

[23] G.N. Kim, Y.H. Jung, J.J. Lee, J.K. Moon, C.H. Jung, An analysis of a flushing effect on the electrokinetic-flushing removal of cobalt and cesium from a soil around decommissioning site, Sep. Purif. Technol. 63 (2008) 116–121. [24] L.M. Rajic, B.D. Dalmacija, J.S. Trickovic, M.B. Dalmacija, D.M. Krcmar, Behavior of zinc, nickel, copper and cadmium during the electrokinetic remediation of sediment from the Great Backa Canal (Serbia), J. Environ. Sci. Health A – Tox. Hazard. Subst. Environ. Eng. 45 (2010) 1134–1143. [25] D.M. Zhou, L. Cang, A.N. Alshawabkeh, Y.J. Wang, X.Z. Hao, Pilot-scale electrokinetic treatment of a Cu contaminated red soil, Chemosphere 63 (2006) 964–971. [26] K. Maturi, K.R. Reddy, Simultaneous removal of organic compounds and heavy metals from soils by electrokinetic remediation with a modified cyclodextrin, Chemosphere 63 (2006) 1022–1031. [27] S.O. Ko, M.A. Schlautman, E.R. Carraway, Cyclodextrin-enhanced electrokinetic removal of phenanthrene from a model clay soil, Environ. Sci. Technol. 34 (2000) 1535–1541. [28] A.P. Khodadoust, K.R. Reddy, O. Narla, Cyclodextrin-enhanced electrokinetic remediation of soils contaminated with 2,4-dinitrotoluene, J. Environ. Eng., ASCE 132 (2006) 1043–1050. [29] S.H. Yuan, M. Tian, X.H. Lu, Electrokinetic movement of hexachlorobenzene in clayed soils enhanced by Tween 80 and beta-cyclodextrin, J. Hazard. Mater. 137 (2006) 1218–1225. [30] T.P. Li, S.H. Yuan, J.Z. Wan, L. Lin, H.Y. Long, X.F. Wu, X.H. Lu, Pilot-scale electrokinetic movement of HCB and Zn in real contaminated sediments enhanced with hydroxypropyl-␤-cyclodextrin, Chemosphere 76 (2009) 1226–1232. [31] T.P. Li, S.H. Yuan, J.Z. Wan, X.H. Lu, Hydroxypropyl-␤-cyclodextrin enhanced electrokinetic remediation of sediment contaminated with HCB and heavy metals, J. Hazard. Mater. 176 (2010) 306–312. [32] C.L. Wan, M.A. Du, D.J. Lee, X. Yang, W.C. Ma, L.N. Zheng, Electrokinetic remediation and microbial community shift of ␤-cyclodextrin-dissolved petroleum hydrocarbon-contaminated soil, Appl. Microbiol. Biotechnol. 89 (2011) 2019–2025. [33] J.L. Chen, S.F. Yang, C.C. Wu, S. Ton, Effect of ammonia as a complexing agent on electrokinetic remediation of copper-contaminated soil, Sep. Purif. Technol. 79 (2011) 157–163. [34] S.C. Lin, Biosurfactants: recent advances, J. Chem. Technol. Biotechnol. 66 (1996) 109–120. [35] C.N. Mulligan, R.N. Yong, B.F. Gibbs, Surfactant-enhanced remediation of contaminated soil: a review, Eng. Geol. 60 (2001) 371–380. [36] E. Ruckenstein, R. Nagarajan, Critical micelle concentration – transition point for micellar size distribution, J. Phys. Chem. 79 (1975) 2622–2626. [37] W. Chu, C.Y. Kwan, Remediation of contaminated soil by solvent/surfactant system, Chemosphere 53 (2003) 9–15. [38] C.N. Mulligan, F. Eftekhari, Remediation with surfactant foam of PCPcontaminated soil, Eng. Geol. 70 (2003) 269–279. [39] D. Fabbri, A. Crime, M. Davezza, C. Medana, C. Baiocchi, A.B. Prevot, E. Pramauro, Surfactant-assisted removal of swep residues from soil and photocatalytic treatment of the washing wastes, Appl. Catal. B – Environ. 92 (2009) 318–325. [40] C.N. Mulligan, Recent advances in the environmental applications of biosurfactants, Curr. Opin. Colloid Interface Sci. 14 (2009) 372–378. [41] R.A. Doong, Y.W. Wu, W.G. Lei, Surfactant enhanced remediation of cadmium contaminated soils, Water Sci. Technol. 37 (1998) 65–71. [42] J.H. Harwell, Factors affecting surfactant performance in groundwater remediation applications, in: D.A. Sabitini, R.C. Knox (Eds.), Transport and Remediation of Subsurface Contaminants: Colloidal, Interfacial, and Surfactant Phenomena, ACS Symposium Series 491, American Chemical Society, Washington, DC, 1992, pp. 124–132. [43] D.A. Sabatini, R.C. Knox, J.H. Harwell, Emerging technologies in surfactantenhanced subsurface remediation, in: D.A. Sabitini, R.C. Knox, J.H. Harwell (Eds.), Surfactant-enhanced Subsurface Remediation: Emerging Technologies, ACS Symposium Series 594, American Chemical Society, Washington, DC, 1995, pp. 1–9. [44] D.G. Cooper, Biosurfactants, Microbiol. Sci. 3 (1986) 145–149. [45] C.C. West, J.H. Harwell, Surfactant and subsurface remediation, Environ. Sci. Technol. 26 (1992) 2324–2330. [46] J.D. Desai, I.M. Banat, Microbial production of surfactants and their commercial potential, Microbiol. Mol. Biol. Rev. 61 (1997) 47–64. [47] E.Z. Ron, E. Rosenberg, Biosurfactants and oil bioremediation, Curr. Opin. Biotechnol. 13 (2002) 249–252. [48] S. Mukherjee, P. Das, R. Sen, Towards commercial production of microbial surfactants, Trends Biotechnol. 24 (2006) 509–515. [49] K. Muthusamy, S. Gopalakrishnan, T.K. Ravi, P. Sivachidambaram, Biosurfactants: properties, commercial production and application, Curr. Sci. 94 (2008) 736–747. [50] G. Soberón-Chávez, F. Lépine, E. Déziel, Production of rhamnolipids by Pseudomonas aeruginosa, Appl. Microbiol. Biotechnol. 68 (2005) 718–725. [51] Y. Uchida, R. Tsuchiya, M. Chino, J. Hirano, T. Tabuchi, Extracellular accumulation of mono- and di-succinoyl trehalose lipids by a strain of Rhodococcus erythropolis grown on n-alkanes, Agric. Biol. Chem. 53 (1989) 757–763. [52] Y. Hirata, M. Ryua, Y. Odaa, K. Igarashia, A. Nagatsukaa, T. Furutaa, M. Sugiuraa, Novel characteristics of sophorolipids, yeast glycolipid biosurfactants, next term as biodegradable low-foaming surfactants, J. Biosci. Bioeng. 108 (2009) 142–146. [53] G.S. Kiran, T.A. Thomas, J. Selvin, B. Sabarathnam, A.P. Lipton, Optimization and characterization of a new lipopeptide biosurfactant produced by marine

A.T. Yeung, Y.-Y. Gu / Journal of Hazardous Materials 195 (2011) 11–29

[54]

[55]

[56]

[57]

[58]

[59]

[60]

[61]

[62] [63]

[64]

[65]

[66] [67]

[68]

[69]

[70]

[71]

[72]

[73] [74] [75]

[76]

[77]

[78]

[79]

[80]

[81]

Brevibacterium aureum MSA13 in solid state culture, Bioresour. Technol. 101 (2010) 2389–2396. T.H. Nielsen, J. Sorensen, Production of cyclic lipopeptides by Pseudomonas fluorescens strains in bulk soil and in the sugar beet rhizosphere, Appl. Environ. Microbiol. 69 (2003) 861–868. J.M. Raaijmakers, I. De Bruijn, M.J.D. De Kock, Cyclic lipopeptide production by plant-associated Pseudomonas spp.: diversity, activity, biosynthesis, and regulation, Mol. Plant-Microbe Interact. 19 (2006) 699–710. C.S. Jeon, J.S. Yang, K.J. Kim, K. Baek, Electrokinetic removal of petroleum hydrocarbon from residual clayey soil following a washing process, Clean-Soil Air Water 38 (2010) 189–193. K.R. Reddy, R.E. Saichek, Effect of soil type on electrokinetic removal of phenanthrene using surfactants and cosolvents, J. Environ. Eng., ASCE 129 (2003) 336–346. R.E. Saichek, K.R. Reddy, Surfactant-enhanced electrokinetic remediation of polycyclic aromatic hydrocarbons in heterogeneous subsurface environments, J. Environ. Eng. Sci. 4 (2005) 327–339. J.W. Yang, Y.J. Lee, J.Y. Park, S.J. Kim, J.Y. Lee, Application of APG and Calfax 16L-35 on surfactant-enhanced electrokinetic removal of phenanthrene from kaolinite, Eng. Geol. 77 (2005) 243–251. J.Y. Park, H.H. Lee, S.J. Kim, Y.J. Lee, J.W. Yang, Surfactant-enhanced electrokinetic removal of phenanthrene from kaolinite, J. Hazard. Mater. 140 (2007) 230–236. J.H. Chang, Z.M. Qiang, C.P. Huang, A.V. Ellis, Phenanthrene removal in unsaturated soils treated by electrokinetics with different surfactants-Triton X-100 and rhamnolipid, Colloids Surf. A 348 (2009) 157–163. M.T. Alcántara, J. Gómez, M. Pazos, M.A. Sanromán, Electrokinetic remediation of PAH mixtures from kaolin, J. Hazard. Mater. 179 (2010) 1156–1160. T.D. Pham, R.A. Shrestha, M. Sillanpaa, Removal of hexachlorobenzene and phenanthrene from clayey soil by surfactant- and ultrasound-assisted electrokinetics, J. Environ. Eng., ASCE 136 (2010) 739–742. K.R. Reddy, C. Cameselle, P. Ala, Integrated electrokinetic-soil flushing to remove mixed organic and metal contaminants, J. Appl. Electrochem. 40 (2010) 1269–1279. C. Yuan, C.H. Hung, W.L. Huang, Enhancement with carbon nanotube barrier on 1,2-dichlorobenzene removal from soil by surfactant-assisted electrokinetic (SAEK) process – the effect of processing fluid, Sep. Sci. Technol. 44 (2009) 2284–2303. A. Karagunduz, A. Gezer, G. Karasuloglu, Surfactant enhanced electrokinetic remediation of DDT from soils, Sci. Total Environ. 385 (2007) 1–11. C. Yuan, C.H. Weng, Remediating ethylbenzene-contaminated clayey soil by a surfactant-aided electrokinetic (SAEK) process, Chemosphere 57 (2004) 225–232. A.Y. Kolosov, K.I. Popov, N.A. Shabanova, A.A. Artemeva, B.M. Kogut, A.S. Frid, V.Y. Zelvenskii, E.M. Urinovich, Electrokinetic removal of hydrophobic organic compounds from soil, Russian J. Appl. Chem. 74 (2001) 631–635. J. Kim, K. Lee, Effects of electric field directions on surfactant enhanced electrokinetic remediation of diesel-contaminated sand column, J. Environ. Sci. Health A – Tox. Hazard. Subst. Environ. Eng. 34 (1999) 863–877. H. Han, Y.J. Lee, S.H. Kim, J.W. Yang, Electrokinetic remediation of soil contaminated with diesel oil using EDTA-cosolvent solutions, Sep. Sci. Technol. 44 (2009) 2437–2454. O. Gonzini, A. Plaza, L. Di Palma, M.C. Lobo, Electrokinetic remediation of gasoil contaminated soil enhanced by rhamnolipid, J. Appl. Electrochem. 40 (2010) 1239–1248. A. Groboillot, F. Portet-Koltalo, F. Le Derf, M.J.G. Feuilloley, N. Orange, C.D. Poc, Novel application of cyclolipopeptide amphis feasibility study as additive to remediate polycyclic aromatic hydrocarbon (PAH) contaminated sediments, Int. J. Mol. Sci. 12 (2011) 1787–1806. A. Kaya, Y. Yukselen, Zeta potential of soils with surfactants and its relevance to electrokinetic remediation, J. Hazard. Mater. 120 (2005) 119–126. C.H. Weng, C. Yuan, Removal of Cr(III) from clay soils by electrokinetics, Environ. Geochem. Health 23 (2001) 281–285. P.W. Chang, Y.C. Tan, I.S. Wang, On-farm remediation and fertility enhancement of cadmium-contaminated silt loam by electrokinetic treatment process in Taiwan, Irrigation Drain. 57 (2008) 187–201. A. Giannis, E. Gidarakos, A. Skouta, Application of sodium dodecyl sulfate and humic acid as surfactants on electrokinetic remediation of cadmiumcontaminated soil, Desalination 211 (2007) 249–260. R.K. Srivastava, R.P. Tiwari, P. Bala Ramudu, Removal of cadmium using electrokinetics – effect of enhancement agents, in: Proc. 1st Sri Lankan Geotech. Soc. Int. Conf. Soil Rock Eng., Colombo, 2007. R.E. Saichek, K.R. Reddy, Effects of system variables on surfactant enhanced electrokinetic removal of polycyclic aromatic hydrocarbons from clayey soils, Environ. Technol. 24 (2003) 503–515. J.Z. Wan, S.H. Yuan, J. Chen, T.P. Li, L. Lin, X.H. Lu, Solubility-enhanced electrokinetic movement of hexachlorobenzene in sediments: a comparison of cosolvent and cyclodextrin, J. Hazard. Mater. 166 (2009) 221–226. A. Li, K.A. Cheung, K.R. Reddy, Cosolvent-enhanced electrokinetic remediation of soils contaminated with phenanthrene, J. Environ. Eng., ASCE 126 (2000) 527–533. K.R. Reddy, P.R. Ala, S. Sharma, S.N. Kumar, Enhanced electrokinetic remediation of contaminated manufactured gas plant soil, Eng. Geol. 85 (2006) 132–146.

27

[82] K. Maturi, K.R. Reddy, Cosolvent-enhanced desorption and transport of heavy metals and organic contaminants in soils during electrokinetic remediation, Water Air Soil Pollut. 189 (2008) 199–211. [83] T.F. Coletta, C.J. Bruell, D.K. Ryan, H.I. Inyang, Cation-enhanced removal of lead from kaolinite by electrokinetics, J. Environ. Eng., ASCE 123 (1997) 1227–1233. [84] K.R. Reddy, S. Chinthamreddy, Enhanced electrokinetic remediation of heavy metals in glacial till soils using different electrolyte solutions, J. Environ. Eng., ASCE 130 (2004) 442–455. [85] R.J. Hunter, Zeta Potential in Colloid Science: Principles and Applications, Academic Press, London, 1981. [86] A.T. Yeung, Diffuse double layer equations in SI units, J. Geotech. Eng., ASCE 118 (1992) 2000–2005. [87] S.K. Puppala, A.N. Alshawabkeh, Y.B. Acar, R.J. Gale, M. Bricka, Enhanced electrokinetic remediation of high sorption capacity soil, J. Hazard. Mater. 55 (1997) 203–220. [88] H.H. Lee, J.W. Yang, A new method to control electrolytes pH by circulation system in electrokinetic soil remediation, J. Hazard. Mater. 77 (2000) 227–240. [89] R.E. Saichek, K.R. Reddy, Effect of pH control at the anode for the electrokinetic removal of phenanthrene from kaolin soil, Chemosphere 51 (2003) 273–287. [90] R.E. Hicks, S. Tondorf, Electrorestoration of metal contaminated soils, Environ. Sci. Technol. 28 (1994) 2203–2210. [91] T. Rødsand, Y.B. Acar, G. Breedveld, Electrokinetic extraction of lead from spiked Norwegian marine clay, in: Y.B. Acar, D.E. Daniel (Eds.), Characterization, Containment, Remediation, and Performance in Environmental Geotechnics, Geotech. Spec. Publ. No. 46, vol. 2, ASCE, New York, 1995, pp. 1518–1534. [92] B.E. Reed, M.T. Berg, J.C. Thompson, J.H. Hatfield, Chemical conditioning of electrode reservoirs during electrokinetic soil flushing of Pb-contaminated silt loam, J. Environ. Eng., ASCE 121 (1995) 805–815. [93] D.M. Zhou, C.F. Deng, L. Cang, A.N. Alshawabkeh, Electrokinetic remediation of a Cu–Zn contaminated red soil by controlling the voltage and conditioning catholyte pH, Chemosphere 61 (2005) 519–527. [94] E. Gidarakos, A. Giannis, Chelate agents enhanced electrokinetic remediation for removal cadmium and zinc by conditioning catholyte pH, Water Air Soil Pollut. 172 (2006) 295–312. [95] B.G. Ryu, G.Y. Park, J.W. Yang, K. Baek, Electrolyte conditioning for electrokinetic remediation of As, Cu, and Pb-contaminated soil, Sep. Purif. Technol. 79 (2011) 170–176. [96] A. Genc, G. Chase, A. Foos, Electrokinetic removal of manganese from river sediment, Water Air Soil Pollut. 197 (2009) 131–141. [97] A.T. Yeung, C. Hsu, R.M. Menon, EDTA enhanced electrokinetic extraction of lead, J. Geotech. Eng., ASCE 122 (1996) 666–673. [98] A.T. Yeung, C. Hsu, Electrokinetic remediation of cadmium-contaminated clay, J. Environ. Eng., ASCE 131 (2005) 298–304. [99] R.W. Leinz, D.B. Hoover, D.L. Fey, D.B. Smith, T. Patterson, Electrogeochemical sampling with NEOCHIM – results of tests over buried gold deposits, J. Geochem. Explor. 61 (1998) 57–86. [100] R.W. Leinz, D.B. Hoover, A.L. Meier, NEOCHIM: an electrochemical method for environmental application, J. Geochem. Explor. 64 (1998) 421–434. [101] Z.M. Shen, X.J. Chen, J.P. Jia, L. Qu, W.H. Wang, Comparison of electrokinetic soil remediation methods using one fixed anode and approaching anodes, Environ. Pollut. 150 (2007) 193–199. [102] H.J. Lee, J.M. Choi, J.W. Cho, S.H. Moon, Characterization of anion exchange membranes fouled with humate during electrodialysis, J. Membr. Sci. 203 (2002) 115–126. [103] Z. Li, J.-W. Yu, I. Neretnieks, Electroremediation: removal of heavy metals from soils by using cation selective membrane, Environ. Sci. Technol. 32 (1998) 394–397. [104] W.S. Kim, S.O. Kim, K.W. Kim, Enhanced electrokinetic extraction of heavy metals from soils assisted by ion exchange membranes, J. Hazard. Mater. 118 (2005) 93–102. [105] F.A. Caliman, B.M. Robu, C. Smaranda, V.L. Pavel, M. Gavrilescu, Soil and groundwater cleanup: benefits and limits of emerging technologies, Clean Technol. Environ. Policy 13 (2011) 241–268. [106] N. Koprivanac, H. Kusic, AOP as an effective tool for the minimization of hazardous organic pollutants in colored wastewater; chemical and photochemical processes, in: A.A. Lewinsky (Ed.), Hazardous Materials and Wastewater: Treatment, Removal and Analysis, Nova Science Publishers, New York, 2007, pp. 149–200. [107] E. Neyens, J. Baeyens, A review of classic Fenton’s peroxidation as an advanced oxidation technique, J. Hazard. Mater. 98 (2003) 33–50. [108] R.J. Watts, S. Kong, M. Dippre, W.T. Barnes, Oxidation of sorbed hexachlorobenzene in soils using catalysed hydrogen peroxide, J. Hazard. Mater. 39 (1994) 33–47. [109] R.J. Watts, P.C. Stanton, Mineralization of sorbed and NAPL-phase hexadecane by catalyzed hydrogen peroxide, Water Res. 33 (1999) 1405–1414. [110] F.J. Rivas, Polycyclic aromatic hydrocarbons sorbed on soils: a short review of chemical oxidation based treatments, J. Hazard. Mater. 138 (2006) 234–251. [111] E. Ferrarese, G. Andreottola, I.A. Oprea, Remediation of PAH-contaminated sediments by chemical oxidation, J. Hazard. Mater. 152 (2008) 128–139. [112] G.C.C. Yang, Y.W. Long, Removal and degradation of phenol in a saturated flow by in-situ electrokinetic remediation and Fenton-like process, J. Hazard. Mater. 69 (1999) 259–271.

28

A.T. Yeung, Y.-Y. Gu / Journal of Hazardous Materials 195 (2011) 11–29

[113] G.C.C. Yang, C.Y. Liu, Remediation of TCE contaminated soils by in situ EKFenton process, J. Hazard. Mater. 85 (2001) 317–331. [114] S.S. Kim, J.H. Kim, S.J. Han, Application of the electrokinetic-Fenton process for the remediation of kaolinite contaminated with phenanthrene, J. Hazard. Mater. 118 (2005) 121–131. [115] J.H. Kim, S.J. Han, S.S. Kim, J.W. Yang, Effect of soil chemical properties on the remediation of phenanthrene-contaminated soil by electrokinetic-Fenton process, Chemosphere 63 (2006) 1667–1676. [116] J.H. Kim, S.S. Kim, J.W. Yang, Role of stabilizers for treatment of clayey soil contaminated with phenanthrene through electrokinetic-Fenton process – some experimental evidences, Electrochim. Acta 53 (2007) 1663–1670. [117] J.H. Kim, J.Y. Kim, S.S. Kim, Effect of H2 SO4 and HCl in the anode purging solution for the electrokinetic-Fenton remediation of soil contaminated with phenanthrene, J. Environ. Sci. Health A – Tox. Hazard. Subst. Environ. Eng. 44 (2009) 1111–1119. [118] T. Alcantara, M. Pazos, S. Gouveia, C. Cameselle, M.A. Sanroman, Remediation of phenanthrene from contaminated kaolinite by electroremediation-Fenton technology, J. Environ. Sci. Health A – Tox. Hazard. Subst. Environ. Eng. 43 (2008) 901–906. [119] K.R. Reddy, M.R. Karri, Effect of oxidant dosage on integrated electrochemical remediation of contaminant mixtures in soils, J. Environ. Sci. Health A – Tox. Hazard. Subst. Environ. Eng. 43 (2008) 881–893. [120] A. Oonnittan, R.A. Shrestha, M. Sillanpaa, Effect of cyclodextrin on the remediation of hexachlorobenzene in soil by electrokinetic Fenton process, Sep. Purif. Technol. 64 (2009) 314–320. [121] T.T. Tsai, J. Sah, C.M. Kao, Application of iron electrode corrosion enhanced electrokinetic-Fenton oxidation to remediate diesel contaminated soils: a laboratory feasibility study, J. Hydrol. 380 (2010) 4–13. [122] A. Oonnittan, P. Isosaari, M. Sillanpaa, Oxidant availability in soil and its effect on HCB removal during electrokinetic Fenton process, Sep. Purif. Technol. 76 (2010) 146–150. [123] P. Isosaari, R. Piskonen, P. Ojala, S. Voipio, K. Eilola, E. Lehmus, M. Itavaara, Integration of electrokinetics and chemical oxidation for the remediation of creosote-contaminated clay, J. Hazard. Mater. 144 (2007) 538–548. [124] G.C.C. Yang, C.F. Yeh, Enhanced nano-Fe3 O4 /S2 O8 2− oxidation of trichloroethylene in a clayey soil by electrokinetics, Sep. Purif. Technol. 79 (2011) 264–271. [125] K.R. Reddy, S. Chinthamreddy, Electrokinetic remediation of heavy metalcontaminated soils under reducing environments, Waste Manage. 19 (1999) 269–282. [126] A. Weeks, S. Pamukcu, Electrokinetically enhanced reduction of [Cr(Vl)] in contaminated soils, in: Proc. 15th Int. Conf. Soil Mech. Geotech. Eng., vols. 1–3, Istanbul, 2001, pp. 1935–1938. [127] M.F. DeFlaun, C.W. Condee, Electrokinetic transport of bacteria, J. Hazard. Mater. 55 (1997) 263–277. [128] H.S. Lee, K. Lee, Bioremediation of diesel-contaminated soil by bacterial cells transported by electrokinetics, J. Microbiol. Biotechnol. 11 (2001) 1038– 1045. [129] L.Y. Wick, P.A. Mattle, P. Wattiau, H. Harms, Electrokinetic transport of PAHdegrading bacteria in model aquifers and soil, Environ. Sci. Technol. 38 (2004) 4596–4602. [130] A. Olszanowski, K. Piechowiak, The use of an electric field to enhance bacterial movement and hydrocarbon biodegradation in soils, Polish J. Environ. Stud. 15 (2006) 303–309. [131] L. Shi, S. Muller, H. Harms, L.Y. Wick, Factors influencing the electrokinetic dispersion of PAH-degrading bacteria in a laboratory model aquifer, Appl. Microbiol. Biotechnol. 80 (2008) 507–515. [132] J.H. Choi, S. Maruthamuthu, H.G. Lee, T.H. Ha, J.H. Bae, Nitrate removal by electro-bioremediation technology in Korean soil, J. Hazard. Mater. 168 (2009) 1208–1216. [133] U.N. Da Rocha, M.R. Totola, D.M.M. Pessoa, J.T. Araruna, J.C.L. Neves, A.C. Borges, Mobilisation of bacteria in a fine-grained residual soil by electrophoresis, J. Hazard. Mater. 161 (2009) 485–491. [134] K.Y. Lee, H.A. Kim, B.T. Lee, S.O. Kim, Y.H. Kwon, K.W. Kim, A feasibility study on bioelectrokinetics for the removal of heavy metals from tailing soil, Environ. Geochem. Health 33 (2011) 3–11. [135] K.Y. Lee, K.W. Kim, Heavy metal removal from shooting range soil by hybrid electrokinetics with bacteria and enhancing agents, Environ. Sci. Technol. 44 (2010) 9482–9487. [136] M. Elektorowicz, V. Boeva, Electrokinetic supply of nutrients in soil bioremediation, Environ. Technol. 17 (1996) 1339–1349. [137] M.F. Rabbi, B. Clark, R.J. Gale, E. Ozsu-Acar, J. Pardue, A. Jackson, In situ TCE bioremediation study using electrokinetic cometabolite injection, Waste Manage. 20 (2000) 279–286. [138] K.R. Reddy, S. Chinthamreddy, R.E. Saichek, T.J. Cutright, Nutrient amendment for the bioremediation of a chromium-contaminated soil by electrokinetics, Energy Sources 25 (2003) 931–943. [139] G. Lee, H. Ro, S. Lee, S. Lee, Electrokinetically enhanced transport of organic and inorganic phosphorus in a low permeability soil, Geosci. J. 10 (2006) 85–89. [140] S. Suni, E. Malinen, J. Kosonen, H. Silvennoinen, M. Romantschuk, Electrokinetically enhanced bioremediation of creosote-contaminated soil: laboratory and field studies, J. Environ. Sci. Health A – Tox. Hazard. Subst. Environ. Eng. 42 (2007) 277–287. [141] A. Tiehm, T. Augenstein, D. Ilieva, H. Schell, C. Weidlich, K.M. Mangold, Bioelectro-remediation: electrokinetic transport of nitrate in a flow-through

[142]

[143]

[144]

[145]

[146]

[147]

[148]

[149]

[150]

[151]

[152]

[153] [154]

[155]

[156]

[157]

[158]

[159]

[160]

[161]

[162]

[163] [164]

[165]

[166]

[167]

[168]

system for enhanced toluene biodegradation, J. Appl. Electrochem. 40 (2010) 1263–1268. C.A.B. Schmidt, M.C. Barbosa, M.D.S. de Almeida, A laboratory feasibility study on electrokinetic injection of nutrients on an organic, tropical, clayey soil, J. Hazard. Mater. 143 (2007) 655–661. S.T. Lohner, D. Katzoreck, A. Tiehm, Electromigration of microbial electron acceptors and nutrients: (I) Transport in synthetic media, J. Environ. Sci. Health A – Tox. Hazard. Subst. Environ. Eng. 43 (2008) 913–921. S.T. Lohner, D. Katzoreck, A. Tiehm, Electromigration of microbial electron acceptors and nutrients: (II) Transport in groundwater, J. Environ. Sci. Health A – Tox. Hazard. Subst. Environ. Eng. 43 (2008) 922–925. W. Xu, C.P. Wang, H.B. Liu, Z.Y. Zhang, H.W. Sun, A laboratory feasibility study on a new electrokinetic nutrient injection pattern and bioremediation of phenanthrene in a clayey soil, J. Hazard. Mater. 184 (2010) 798–804. W.A. Jackson, M.A. Jeon, J.H. Pardue, T. Anderson, Enhanced natural attenuation of perchlorate in soils using electrokinetic injection, in: Bioremediation of Inorganic Compounds, Proc. 6th International In Situ On-Site Bioremediation Symposium, vol. 9, San Diego, 2001, pp. 273–280. X.Y. Fan, H. Wang, Q.S. Luo, J.W. Ma, X.H. Zhang, The use of 2D non-uniform electric field to enhance in situ bioremediation of 2,4-dichlorophenolcontaminated soil, J. Hazard. Mater. 148 (2007) 29–37. X. Wu, A.N. Alshawabkeh, D.B. Gent, S.L. Larson, J.L. Davis, Lactate transport in soil by DC fields, J. Geotech. Geoenviron. Eng., ASCE 133 (2007) 1587– 1596. A. Tiehm, S.T. Lohner, T. Augenstein, Effects of direct electric current and electrode reactions on vinyl chloride degrading microorganisms, Electrochim. Acta 54 (2009) 3453–3459. T.T. Li, S.H. Guo, B. Wu, F.M. Li, Z.X. Niu, Effect of electric intensity on the microbial degradation of petroleum pollutants in soil, J. Environ. Sci. – China 22 (2010) 1381–1386. L.Y. Wick, F. Buchholz, I. Fetzer, S. Kleinsteuber, C. Hartig, L. Shi, A. Miltner, H. Harms, G.N. Pucci, Responses of soil microbial communities to weak electric fields, Sci. Total Environ. 408 (2010) 4886–4893. S.H. Kim, H.Y. Han, Y.J. Lee, C.W. Kim, J.W. Yang, Effect of electrokinetic remediation on indigenous microbial activity and community within diesel contaminated soil, Sci. Total Environ. 408 (2010) 3162–3168. P.L. Palmer, Permeable treatment barriers, in: In situ Treatment Technology, 2nd ed., Lewis Publishers, Boca Raton, 2001, pp. 459–482. S.V. Ho, C.J. Athmer, P.W. Sheridan, A.P. Shapiro, Scale-up aspects of the LasagnaTM process for in situ soil decontamination, J. Hazard. Mater. 55 (1997) 39–60. S.V. Ho, P.W. Sheridan, C.J. Athmer, M.A. Heitkamp, J.M. Brackin, D. Weber, P.H. Brodsky, Integrated in situ soil remediation technology: the Lasagna process, Environ. Sci. Technol. 29 (1995) 2528–2534. S.V. Ho, C. Athmer, P.W. Sheridan, B.M. Hughes, R. Orth, D. Mckenzie, P.H. Brodsky, A. Shapiro, R. Thornton, J. Salvo, D. Schultz, R. Landis, R. Griffith, S. Shoemaker, The Lasagna technology for in situ soil remediation. 1. Small field test, Environ. Sci. Technol. 33 (1999) 1086–1091. S.V. Ho, C. Athmer, P.W. Sheridan, B.M. Hughes, R. Orth, D. Mckenzie, P.H. Brodsky, A.M. Shapiro, T.M. Sivavec, J. Salvo, D. Schultz, R. Landis, R. Griffith, S. Shoemaker, The Lasagna technology for in situ soil remediation. 2. Large field test, Environ. Sci. Technol. 33 (1999) 1092–1099. M. Roulier, M. Kemper, S. Al-Abed, L. Murdoch, P. Cluxton, J.-L. Chen, W. Davishoover, Feasibility of electrokinetic soil remediation in horizontal LasagnaTM cells, J. Hazard. Mater. 77 (2000) 161–176. S.A. Jackman, G. Maini, A.K. Sharman, G. Sunderland, C.J. Knowles, Electrokinetic movement and biodegradation of 2,4-dichlorephenoxyacetic acid in silt soil, Biotechnol. Bioeng. 74 (2001) 40–48. J.W. Ma, F.Y. Wang, Z.H. Huang, H. Wang, Simultaneous removal of 2,4-dichlorophenol and Cd from soils by electrokinetic remediation combined with activated bamboo charcoal, J. Hazard. Mater. 176 (2010) 715–720. C.F. Chew, T.C. Zhang, In-situ remediation of nitrate contaminated groundwater by electrokinetics iron wall processes, Water Sci. Technol. 38 (1998) 135–142. J.W. Moon, H.S. Moon, H. Kim, Y. Roh, Remediation of TCE-contaminated groundwater using zero valent iron and direct current: experimental results and electron competition model, Environ. Geol. 48 (2005) 805–817. C. Yuan, The effect of Fe(0) on electrokinetic remediation of clay contaminated with perchloroethylene, Water Sci. Technol. 53 (2006) 91–98. C.H. Weng, Y.T. Lin, T.Y. Lin, C.M. Kao, Enhancement of electrokinetic remediation of hyper-Cr(VI) contaminated clay by zero-valent iron, J. Hazard. Mater. 149 (2007) 292–302. C. Yuan, T.S. Chiang, The mechanisms of arsenic removal from soil by electrokinetic process coupled with iron permeable reaction barrier, Chemosphere 67 (2007) 1533–1542. L. Cang, D.M. Zhou, D.Y. Wu, A.N. Alshawabkeh, Coupling electrokinetics with permeable reactive barriers of zero-valent iron for treating a chromium contaminated soil, Sep. Sci. Technol. 44 (2009) 2188–2202. J.Z. Wan, Z.R. Li, X.H. Lu, S.H. Yuan, Remediation of a hexachlorobenzenecontaminated soil by surfactant-enhanced electrokinetics coupled with microscale Pd/Fe PRB, J. Hazard. Mater. 184 (2010) 184–190. Y.H. Shih, Y.C. Chen, M.Y. Chen, Y.T. Tai, C.P. Tso, Dechlorination of hexachlorobenzene by using nanoscale Fe and nanoscale Pd/Fe bimetallic particles, Colloids Surf. A 332 (2009) 84–89.

A.T. Yeung, Y.-Y. Gu / Journal of Hazardous Materials 195 (2011) 11–29 [169] D.W.S. Faulkner, L. Hopkinson, A.B. Cundy, Electrokinetic generation of reactive iron-rich barriers in wet sediments: implications for contaminated land management, Mineral. Mag. 69 (2005) 749–757. [170] H.I. Chung, M. Lee, A new method for remedial treatment of contaminated clayey soils by electrokinetics coupled with permeable reactive barriers, Electrochim. Acta 52 (2007) 3427–3431. [171] T. Kimura, K.I. Takase, N. Terui, S. Tanaka, Ferritization treatment of copper in soil by electrokinetic remediation, J. Hazard. Mater. 143 (2007) 662–667. [172] S.S. Mandaokar, D.M. Dharmadhikari, S.S. Dara, Retrieval of heavy-metal ions from solution via ferritisation, Environ. Pollut. 83 (1994) 277–282. [173] E. Barrado, F. Prieto, M. Vega, F. Fernandez-Polanco, Optimization of the operational variables of a medium-scale reactor for metal-containing wastewater purification by ferrite formation, Water Res. 32 (1998) 3055–3061. [174] G. De Gioannis, A. Muntoni, R. Ruggeri, J.J.P. Zijlstra, Chromate adsorption in a transformed red mud permeable reactive barrier using electrokinesis, J. Environ. Sci. Health A – Tox. Hazard. Subst. Environ. Eng. 43 (2008) 969–974. [175] C. Yuan, C.H. Hung, K.C. Chen, Electrokinetic remediation of arsenate spiked soil assisted by CNT-Co barrier – the effect of barrier position and processing fluid, J. Hazard. Mater. 171 (2009) 563–570. [176] X.J. Peng, Y.H. Li, Z.K. Luan, Z.C. Di, H.Y. Wang, B.H. Tian, Z.P. Jia, Adsorption of 1,2-dichlorobenzene from water to carbon nanotubes, Chem. Phys. Lett. 376 (2003) 154–158. [177] Y. Yu, J.C. Yu, C.Y. Chan, Y.K. Che, J.C. Zhao, L. Ding, W.K. Ge, P.K. Wong, Enhancement of adsorption and photocatalytic activity of TiO2 by using carbon nanotubes for the treatment of azo dye, Appl. Catal. B – Environ. 61 (2005) 1–11. [178] K. Yang, X.L. Wang, L.Z. Zhu, B.S. Xing, Competitive sorption of pyrene, phenanthrene, and naphthalene on multiwalled carbon nanotubes, Environ. Sci. Technol. 40 (2006) 5804–5810. [179] K. Yang, L.Z. Zhu, B.S. Xing, Adsorption of polycyclic aromatic hydrocarbons by carbon nanomaterials, Environ. Sci. Technol. 40 (2006) 1855–1861. [180] J.G. Han, K.K. Hong, Y.W. Kim, J.Y. Lee, Enhanced electrokinetic (E/K) remediation on copper contaminated soil by CFW (carbonized foods waste), J. Hazard. Mater. 177 (2010) 530–538. [181] C.S. O’Connor, N.W. Lepp, R. Edwards, G. Sunderland, The combined use of electrokinetic remediation and phytoremediation to decontaminate metalpolluted soils: a laboratory-scale feasibility study, Environ. Monit. Assess. 84 (2003) 141–158.

29

[182] J.M. Lim, A.L. Salido, D.J. Butcher, Phytoremediation of lead using Indian mustard (Brassica juncea) with EDTA and electrodics, Microchem. J. 76 (2004) 3–9. [183] D.M. Zhou, H.F. Chen, L. Cang, Y. Wang, Ryegrass uptake of soil Cu/Zn induced by EDTA/EDDS together with a vertical direct-current electrical field, Chemosphere 67 (2007) 1671–1676. [184] H. Aboughalma, R. Bi, M. Schlaak, Electrokinetic enhancement on phytoremediation in Zn, Pb, Cu and Cd contaminated soil using potato plants, J. Environ. Sci. Health A – Tox. Hazard. Subst. Environ. Eng. 43 (2008) 926–933. [185] R. Bi, M. Schlaak, E. Siefert, R. Lord, H. Connolly, Influence of electrical fields (AC and DC) on phytoremediation of metal polluted soils with rapeseed (Brassica napus) and tobacco (Nicotiana tabacum), Chemosphere 83 (2011) 318–326. [186] L. Cang, Q.Y. Wang, D.M. Zhou, H. Xu, Effects of electrokinetic-assisted phytoremediation of a multiple-metal contaminated soil on soil metal bioavailability and uptake by Indian mustard, Sep. Purif. Technol. 79 (2011) 246–253. [187] H.I. Chung, M. Kamon, Ultrasonically enhanced electrokinetic remediation for removal of Pb and phenanthrene in contaminated soils, Eng. Geol. 77 (2005) 233–242. [188] H.I. Chung, Treatment of contaminated groundwater in sandy layer under river bank by electrokinetic and ultrasonic technology, Water Sci. Technol. 55 (2007) 329–338. [189] T.D. Pham, R.A. Shrestha, J. Virkutyte, M. Sillanpaa, Combined ultrasonication and electrokinetic remediation for persistent organic removal from contaminated kaolin, Electrochim. Acta 54 (2009) 1403–1407. [190] R.A. Shrestha, T.D. Pham, M. Sillanpaa, Electro ultrasonic remediation of polycyclic aromatic hydrocarbons from contaminated soil, J. Appl. Electrochem. 40 (2010) 1407–1413. [191] W. Wesner, A. Diamant, B. Schrammel, M. Unterberger, Electrosynthesis of oxidants and their electrokinetic distribution, in: K.R. Reddy, C. Cameselle (Eds.), Electrochemical Remediation Technologies for Polluted Soils, Sediments and Groundwater, John Wiley & Sons, Hoboken, 2009, pp. 473–482. [192] J.A. Soesilo, S.R. Wilson, Site Remediation Planning and Management, Lewis Publishers, Boca Raton, 1997. [193] G.J. Smith, Coupled electrokinetic-thermal desorption, in: K.R. Reddy, C. Cameselle (Eds.), Electrochemical Remediation Technologies for Polluted Soils, Sediments and Groundwater, John Wiley & Sons, Hoboken, 2009, pp. 505–535.

Journal of Hazardous Materials 195 (2011) 30–54

Contents lists available at SciVerse ScienceDirect

Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat

Review

Non-thermal plasmas for non-catalytic and catalytic VOC abatement Arne M. Vandenbroucke, Rino Morent ∗ , Nathalie De Geyter, Christophe Leys Research Unit Plasma Technology, Department of Applied Physics, Faculty of Engineering, Ghent University, Jozef Plateaustraat 22, 9000 Ghent, Belgium

a r t i c l e

i n f o

Article history: Received 27 February 2011 Received in revised form 19 August 2011 Accepted 22 August 2011 Available online 27 August 2011 Keywords: Non-thermal plasma Plasma–catalysis Volatile organic compounds Waste gas treatment

a b s t r a c t This paper reviews recent achievements and the current status of non-thermal plasma (NTP) technology for the abatement of volatile organic compounds (VOCs). Many reactor configurations have been developed to generate a NTP at atmospheric pressure. Therefore in this review article, the principles of generating NTPs are outlined. Further on, this paper is divided in two equally important parts: plasmaalone and plasma–catalytic systems. Combination of NTP with heterogeneous catalysis has attracted increased attention in order to overcome the weaknesses of plasma-alone systems. An overview is given of the present understanding of the mechanisms involved in plasma–catalytic processes. In both parts (plasma-alone systems and plasma–catalysis), literature on the abatement of VOCs is reviewed in close detail. Special attention is given to the influence of critical process parameters on the removal process. © 2011 Elsevier B.V. All rights reserved.

Contents 1. 2.

3.

4.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Non-thermal plasmas for VOC abatement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Without catalyst . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.1. What is a non-thermal plasma? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.2. Reactor concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.3. VOC abatement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Combined with catalyst . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.1. What is plasma–catalysis? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.2. Different types of catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.3. VOC abatement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Critical process parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Initial VOC concentration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Humidity level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Oxygen content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5. Gas flow rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction Environmental protection is becoming an issue of growing concern in our globalized world. The industrialization of many economies has led to the emission of various kinds of substances that danger both human and ecological life [1]. Since World War II,

∗ Corresponding author. E-mail address: [email protected] (R. Morent). 0304-3894/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jhazmat.2011.08.060

30 31 31 31 31 34 38 38 40 40 47 47 47 47 48 48 48 49

governments have become aware that emission legislation needs to become increasingly severe in order to ensure the protection of our environment for future generations. International treaties like the Kyoto protocol (1997) and the protocol of Gothenburg (1999) are important examples. Exhausts from mobile (e.g. cars) and stationary sources (e.g. plants) pollute the air with a variety of harmful substances that threat human and ecological life [2]. Next to NOx , SOx , H2 S,. . ., volatile organic compounds (VOCs) are a large and important group of pollutants. Their high volatility causes them to rapidly

A.M. Vandenbroucke et al. / Journal of Hazardous Materials 195 (2011) 30–54 Table 1 Typical VOCs and their health effects. VOC

Formula

Effects

O Acetone

H3C

C

CH3

Carcinogen

O Formaldehyde

H Dichloroethane

C

Sore throat, dizziness, headache

H

H

H

H

H

Cl

Cl

H

Cl

Cl

Cl

Cl

Cl

Cl

Cl

Trichloroethylene

Tetrachloroethylene

Paralysis of nerve center

Liver and kidney disease, paralysis of nerve center

Probable heart and liver disease, skin irritation

Benzene

Carcinogen

Toluene

Headache, dizziness

Xylene

Headache, dizziness

Styrene

Probable carcinogen

31

non-selective, creating a chemical reactive environment in which harmful substances are readily decomposed. Although NTP for end-of-pipe application has been frequently proposed in literature for the removal of VOCs, NOx and SO2 ,. . . [14–16] formation of unwanted byproducts, poor energy efficiency and mineralization are serious roadblocks towards industrial implementation. To overcome these problems, researchers are combining the advantages of NTP and catalysis in a technique called plasma–catalysis. This innovative method has become a hot topic over approximately the last ten years. The primary idea is that, by placing catalysts inside or in close vicinity of the discharge zone, retention time can be increased through adsorption of target molecules, favoring complete oxidation to CO2 and H2 O [17]. Interestingly, combining both techniques creates a synergism that is caused by various mechanisms [18–20]. This review gives an overview of the literature on plasmaassisted decomposition of VOCs with a special focus on plasma–catalysis. In the first part, an overview of different NTP reactors is given. In the second part, based on a large number of papers, an extended review is presented dealing with the treatment of VOCs with plasma-alone as well as with plasma–catalytic systems. Particular attention is paid to the most studied target compounds, i.e. trichloroethylene, benzene and toluene. Also general mechanisms that govern plasma–catalysis are summarized. In the third section, special attention is given to the influence of critical process parameters on the removal process. In the final section, future trends for this promising hybrid technique are discussed.

2. Non-thermal plasmas for VOC abatement 2.1. Without catalyst

evaporate and enter the earth’s atmosphere. Depending on their chemical structure and concentration, they can cause various effects such as the creation of photochemical smog, secondary aerosols and tropospheric ozone [3]. They also have an effect on the intensification of global warming and on the deterioration of the stratospheric ozone layer. Some of them are toxic and cause odour nuisance while others have carcinogenic effects, proving their adverse effects on human health [4]. Table 1 provides an overview of typical VOCs that have been studied for removal with non-thermal plasma (NTP) and plasma–catalysis along with their related health effects. Conventional methods to control VOC emissions are wellestablished technologies such as adsorption [5], thermal and catalytic oxidation [6], membrane separation [7], bioreaction [8] and photocatalysis [9]. The disadvantage of these methods is that they become cost-inefficient and difficult to operate when low concentrations of VOC need to be treated [10]. With the increased severity of emission limits in mind, this creates the need for an alternative technology that overcomes these weaknesses. For the abatement of VOCs, NTP technology has attracted growing interest of scientists over the last 2 decades [11,12]. The energy that is delivered to the system, is almost completely consumed for accelerating electrons. They gain a typical temperature of 10,000–250,000 K (1–20 eV) [13], while the background gas remains at room temperature. This non-equilibrium state makes it unnecessary to heat the entire treated gas flow. Accelerated primary electrons collide with background molecules (N2 , O2 , H2 O,. . .) producing secondary electrons, photons, ions and radicals. These latter species are responsible for the oxidation of VOC molecules, although ionic reactions are also possible. This process is highly

2.1.1. What is a non-thermal plasma? Non-thermal plasmas are generated by applying a sufficiently strong electric field to ensure the discharge of a neutral gas. This creates a quasi neutral environment containing neutrals, ions, radicals, electrons and UV photons. Due to their light mass, electrons are selectively accelerated by the field and gain high temperatures while the heavier ions remain relatively cold through energy exchange by collisions with the background gas. The bulk gas molecules (e.g. N2 , O2 ) are bombarded by the electrons, typically having temperatures ranging from 10,000 K to 250,000 K (1–20 eV). This produces excited gas molecules (N2 *, O2 *) which lose their excess energy by emitting photons or heat. Next to excitation, other processes like ionization, dissociation and electron attachment occur in the discharge zone. Through these reaction channels, unstable reactive species like ions and free radicals are formed. Free radicals, such as OH• and O• , are highly reactive species which are ideal for the conversion of environmental pollutants to CO2 , H2 O and other degradation products at uncharacteristic low temperatures. The generation of NTP at atmospheric pressure and ambient temperature has been the subject of many research papers during the last two decades. This has lead to great advances, mainly on laboratory scale. However, large-scale demonstrations of NTP technology for waste gas cleaning are also currently operative [13,21].

2.1.2. Reactor concepts Researchers have investigated a variety of NTP reactors for environmental purposes. The classification of these different reactors is rather complex and depends on multiple characteristics, such as:

32

A.M. Vandenbroucke et al. / Journal of Hazardous Materials 195 (2011) 30–54

Fig. 1. Illustrations of various NTP reactor configurations.

- type of discharge: (DC or pulsed) corona discharge, surface discharge, dielectric barrier discharge, ferro-electric packed bed discharge,. . . - type of power supply: AC, DC, pulse, microwave, RF,. . . - other characteristics: electrode configuration, voltage level, polarity, gas composition,. . . For the conventional NTP reactors that are employed in laboratory experiments, only the main characteristics will be briefly discussed here. A more detailed discussion can be found in literatures [22–27]. A dielectric barrier discharge (DBD) or silent discharge, typically has at least one dielectric (e.g. glass, quartz or ceramic) between the electrodes. DBDs are generally operated in one of the planar or cylindrical configurations shown in Fig. 1. When the local electron density at certain locations in the discharge gap reaches a critical value, a large number of separate and short-lived current filaments are formed, also referred to as microdischarges. These bright, thin filaments are statistically distributed in space and time and are formed by channel streamers with nanosecond duration [28]. When a microdischarge reaches the dielectric, it spreads into a surface discharge and the accumulation of the transferred charge on the surface of the dielectric barrier reduces the electric field. As the electric field further reduces, electron attachment prevails over ionization and the microdischarges are extinguished. When the polarity of the AC voltage changes, the formation of a microdischarge is repeated at the same location if the

electron density again reaches a critical value necessary for electrical breakdown. Therefore, the use of the dielectric in the discharge zone has two functions: (1) limiting the charge transferred by an individual microdischarge, thereby preventing the transition to an arc discharge, and (2) spreading the microdischarge over the electrode surface which increases the probability of electron–molecule collisions with bulk gas molecules [28]. This type of arrangement is often referred to as a volume discharge [29]. Another type of arrangement to generate NTP in a DBD is the surface discharge [29] (Fig. 1(c)). Here for example, a series of strip electrodes are attached to the surface of a high-purity alumina ceramic base. A film like counter electrode is embedded in the inside of the alumina ceramic base and functions as an induction electrode. The ceramic can be either planar or cylindrical [30,31]. When an AC voltage is applied between the strip electrodes and the embedded counter electrode, a surface discharge starts from the peripheral edges of each discharge electrode and stretches out along the ceramic surface. The surface discharges actually consist of many nanosecond surface streamers. In another configuration to generate a surface discharge, strip electrodes can be placed on the inner surface of a cylindrical surface discharge reactor [32]. In this set-up, a DBD discharge is also formed between the central rod electrode and the surface electrodes. A pulsed corona discharge (Fig. 1(d)) applies a pulsed power supply with a fast voltage rise time (tens of nanoseconds) to enable an increase in corona voltage and power without formation of sparks, which can damage the reactor and decrease the process efficiency.

A.M. Vandenbroucke et al. / Journal of Hazardous Materials 195 (2011) 30–54

The required voltage level to energize the discharge depends on the distance between the electrodes, the pulse duration and the gas composition [33]. The duration of a pulse voltage is typically in the order of 100–200 ns to ensure that spark formation is prevented and that the energy dissipation by ions is minimal. The latter is important to enhance the energy efficiency of the system. The electrode configuration of a pulsed corona discharge reactor can be either wire-to-cylinder [34–36] or wire-to-plate [37,38], although the former allows a better spatial distribution of the streamers and a higher energy density deposition in the gas [34]. The pulsed corona discharge usually consists of streamers, for which the ionization zone fills the entire electrode gap (e.g. 10 cm). This is favorable in terms of up-scaling and reducing the pressure drop. However, upscaling is hampered by the high demands on the electronics of large pulsed power voltage sources. A ferroelectric pellet packed-bed reactor (Fig. 1(e)) is a packed-bed reactor filled with perovskite oxide pellets. These reactors can have a parallel-plate or a coaxial configuration. Barium titanate (BaTiO3 ) is the most widely used ferroelectric material for environmental purposes, owing to its high dielectric constant (2000 < ε < 10,000). Other used ferroelectric materials are NaNO2 [39], MgTiO4 , CaTiO3 , SrTiO3 , PbTiO3 [40] and PbZrO3 –PbTiO3 [41]. Application of an external electric field leads to polarization of the ferroelectric material and induces strong local electric fields at the contact points between the pellets and between the pellets and electrodes. This enables the production of partial discharges in the vicinity of each contact point between pellets. The presence of ferroelectric pellets in the discharge zone is beneficial for a uniform gas distribution and electrical discharge but causes an increase in pressure drop over the reactor length. Ferroelectric packed-bed reactors could serve as an alternative approach to enhance the energy efficiency, because the increase of the electric field will lead to a higher mean electron energy. Hence, the energetic electrons tend to form active species through dissociation and ionization, rather than forming less useful species through rotational and vibrational excitation. This leads to a more favorable consumption of the energy delivered, because electron-impact reactions are mainly

33

responsible for the plasma chemistry that destroys environmental pollutants. A DC corona discharge is generated at atmospheric pressure when sharp points, edges or thin wires are subjected to a sufficiently large electric field. This causes a local increase of the electric field in the vicinity of the sharp curvature of the electrode. This is e.g. the case for a point-to-plate or for a wire-to-cylinder configuration. The corona discharge is initiated by acceleration of free electrons and subsequent electron collision processes. Due to formation of electron/positive-ion pairs and their separating process, an electron avalanche is created which sustains the corona discharge. Visually, this discharge is characterized by a weak glow region around the sharp electrode. Depending on the polarity of this electrode, the formation mechanism of the electron avalanche physically differs [22]. When the electrode with the strongest curvature is connected to the positive output of the power supply, a positive DC corona discharge is generated. Propagation of the discharge mainly depends on secondary photo-ionization processes around the sharp tip. The positive corona is characterized by the presence of streamers, i.e. numerous thin current filaments which are chaotically distributed in the gap. At a certain threshold voltage the discharge transitions from the stable corona mode to an unstable spark discharge regime. In the case that the sharp electrode is connected to the negative output, a negative DC corona discharge is formed. Here, impact ionization of gas molecules is generally responsible for the propagation of the discharge. As the applied voltage increases, the negative corona will initially form Trichel pulse corona, followed by pulseless corona and spark discharge [22]. However, certain research groups [42–48] have succeeded in generating a glow discharge at atmospheric pressure before the negative corona shifts to a spark discharge. Akishev et al. [47] applied a special electrode geometry and a fast gas flow to stabilize the discharge, hence delaying it from creating sparks. Vertriest et al. [49] successfully tested the multipin-to-plate reactor concept for VOC abatement. Antao et al. [50] recently reviewed the operating regimes of atmospheric pressure DC corona discharges and their potential applications.

Table 2 Overview of published papers on TCE removal with NTP. Plasma type

Carrier gas

Flow rate (mL/min)

Concentration range (ppm)

Maximum removal efficiency (%)

DBD

Ar/O2 Ar/O2 /H2 O

104

500

>99 90

DBDa DBDb

Air Humid air

700 500

250 150–200

DBD

Dry air

400

DBD DBD DBD DBD

Dry air Dry air Dry air Humid air

DBD Surface discharge

Energy density (J/L)

Energy yield (g/kWh)

References

50 150

193.5d 58d

[51]

>99 >99

140 480

34.6d 8.1d

[115] [133]

1000 100

95 >99

150 135

122.5d 14.3d

[138]

400 2000 510 200–510

100 250 430 750

99 98 >99 98–99

200 120 350 2400

9.6d 39.5d 23.8d 6d

[161] [171] [172] [229]

Dry air

400

1000

99 95–99

1400 1150

13.7d 16.3d

[143]

DBD Pulsed corona

Dry air

2 × 104

160

85 90

100 50

26.3d 55.7d

[53]

Pulsed coronac Pulsed corona Positive corona DC negative glow discharge Capillary tube discharge

Humid air Dry air Dry air Humid air Dry air

– – 1500 106 1000

1000 100 100 120 452

90 80 67 47 80

100 50 580 37 –

174.1d 30.9 2.2d 29.5d –

a b c d

Copper rod inner electrode. Inner electrode made of sintered metal fibres. Pulsed corona discharge with reticulated vitreous carbon electrodes. Calculated from data retrieved from reference.

[52] [72] [118] [49] [78]

34

A.M. Vandenbroucke et al. / Journal of Hazardous Materials 195 (2011) 30–54

Table 3 Overview of published papers on benzene removal with NTP. Plasma type

Carrier gas

Flow rate (mL/min)

Concentration range (ppm)

Maximum removal efficiency (%)

Energy density (J/L)

Energy yield (g/kWh)

References

DBD DBD DBD DBD DBD DBD DBD DBD DBD DBD DBD Pulsed corona Positive DC coronaa DC glowb BaTiO3 packed-bed BaTiO3 packed-bed BaTiO3 packed-bed BaTiO3 packed-bed BaTiO3 packed-bed BaTiO3 packed-bedc

Air (0–90% RH) Dry air Humid air Dry air (5% O2 ) Ar/2–40% O2 Dry air Dry air Dry air Dry air Dry air Dry air Dry air Dry air Dry air Dry air Humid air (0.5% H2 O) Dry air Dry air Dry air Humid air (0.5% H2 O)

500–2000 200 104 4000–5000 275 250 500 400 4000 1667 35 × 103 100 – 100 200 200 203–210 – 200 200

500–2700 100 276 200 500–104 300–380 105 200 203–210 407 250 300 300 296 200 200 200 110 200 200

>99.9 90 >99 75 30 11 35 70 40 50 50 75 >99 90 >99 75 65 98 60 95

2000–3000 680 810 305 – 170 360 3150 370 – 230 30 – 4000 3000 1800 400 130 600 1600

– 1.5d 3.9d 5.7d – 2.5 1.2d 0.5d 2.6d – 6.3d 86.1d – 0.9d 0.8d 1d 3.7d 9.4d 2.3d 1.4d

[69] [70] [71] [75] [77] [108] [129] [163] [175] [179] [230] [90] [73] [72] [40] [163] [174] [176] [212] [231]

a b c d

Corona reactor is sealed after addition of benzene/air mixture. Microhollow cathode. Glass layer between two concentric electrodes. Calculated from data retrieved from reference.

2.1.3. VOC abatement Tables 2–5 give an overview of published papers on VOC removal with NTP. For each reference, experimental conditions are given, along with the maximum removal efficiency and the corresponding energy yield in g/kWh calculated as followed: C ×  × M × 0.15 Energy yield = in ε where Cin is the initial concentration (ppm) of the VOC with molecular weight M (g/mol),  the maximum removal efficiency and ε the corresponding energy density (J/L), i.e. the energy deposited per unit volume of process gas. Each calculation is based on the fact that one mole of a gas occupies 24.04 L volume at standard ambient temperature and pressure (293 K and 101325 Pa). In what follows, particular attention is paid to the most studied target compounds, i.e. trichloroethylene, benzene and toluene. In

Table 5, a selection has been made of other relevant, but less frequently studied VOCs. For more details about operating conditions and results, the reader can consult the corresponding references. 2.1.3.1. Trichloroethylene. As can be seen from Table 2, TCE is a chlorinated olefin which has attracted a lot of attention because it can be relatively easy removed by NTP without the addition of considerable energy. This results from the fact that reactive radicals, produced in the plasma discharge, easily add to the carbon-carbon double bond thereby initiating the oxidation process. Evans et al. [51] carried out an experimental and computational study of the plasma remediation of TCE in dry and wet Ar/O2 mixtures using a silent discharge plasma. They found that the ClO radical is an important intermediate which oxidizes TCE. In wet mixtures, ClO is partially consumed by OH radicals, resulting in a lower decomposition rate of TCE. They suggest a diagram

Table 4 Overview of published papers on toluene removal with NTP. Plasma type

Carrier gas

Flow rate (mL/min)

Concentration range (ppm)

Maximum removal efficiency (%)

Energy density (J/L)

Energy yield (g/kWh)

References

DBD

Dry air (5% O2 )

4000–5000

200

75

310

6.6b

[75]

DBD

N2 dry air

2000

400

21 23

240

4.7b 5.2b

[81]

DBD DBD DBD packed with glass pellets DBD packed with glass pellets DBD packed with glass beads Multicell DBD packed with glass beadsa DC back corona Pulsed corona Positive corona BaTiO3 packed-bed Dielectric capillary plasma electrode discharge Capillary tube discharge

N2 /5% O2 (0.2% RH) Humid air (55% RH) Dry air Humid air (95% RH) Dry air Dry air

100 1000 600 500 315 1000

50 100 1100 500 240 110

73 46 75–80 91 36 72

600 2100 1000 18.5 172 2502

0.8b 0.3b 11.5b 11.5 6.8b 0.4b

[82] [203] [167] [232] [202] [224]

Dry air Dry air Humid air (26% RH) Dry air Air

100–750 450 104 – –

5–200 500 0.5 101 266.5

93 >99 80 95 >99

2400 1000 65 125 3500

0.4b 6.7b 0.1b – 1b

[79] [84] [83] [176] [233]

Dry air

350

1246

86

–

–

[78]

a b

Three cells. Calculated from data retrieved from reference.

A.M. Vandenbroucke et al. / Journal of Hazardous Materials 195 (2011) 30–54

35

Table 5 Published papers on removal of other VOCs with NTP. Target compound

Reference

Acetaldehyde Acetone Acetylene Dichloromethane Formaldehyde Methane Methanol Propane Propene Styrene Tetrachloromethane Xylene

[215,234] [31,235–240] [241,242] [210,235,243–246] [191,194] [72,218,219,247] [53,61,210,248–250] [34,190,193,251–253] [34,241,252,254,255] [256,257] [61,216,235,245,258–260] [35,192,233,238,261]

of the dominant reaction pathways in plasma remediation of TCE giving CO, CO2 , COCl2 and HCl as main byproducts. According to the authors, the toxic byproduct phosgene (COCl2 ) can easily be removed from the exhaust stream by placing a water scrubber downstream of the plasma discharge. This is a cost effective posttreatment because removal of phosgene with their DBD reactor requires high energy. The authors propose the following reactions as the dominant degradation pathway for TCE in dry Ar/O2 mixtures: C2 HCl3 + O• → CHOCl + CCl2

(1)

ClO•

(2)

CCl2 + O2 →

+ COCl

CHOCl +

O•

→ COCl +

OH•

(3)

CHOCl +

Cl•

→ COCl + HCl

(4)

COCl +

O•

→ CO +

COCl +

Cl•

→ CO + Cl2

COCl + O2 → CO2

ClO•

+ ClO•

(5) (6) (7)

The ClO radical rapidly back-reacts with TCE leading to the formation of phosgene and methylchloride by the following reaction: C2 HCl3 + ClO• → COCl2 + CHCl2

(8)

Methylchloride then quickly reacts with oxygen in the subsequent reaction: CHCl2 + O• → CHOCl + Cl

(9)

In wet mixtures, two additional species can be produced by reaction of OH with TCE, CHCl2 –COCl (dichloroacetylchloride; DCAC) and CHCl2 . DCAC is detected as main byproduct of TCE decomposition with a pulsed corona discharge by Kirkpatrick et al. They suggest the reaction of TCE with ClO radicals leading to the formation of DCAC under dry conditions, as follows [52]: C2 HCl3 + ClO• → CHCl2 COCl + Cl

(10)

C2 HCl3 + OH• → CHCl2 COCl + H

(11)

Under humid conditions the formation of DCAC is suppressed, suggesting that ClO radicals are quenched by OH radicals by the reaction: ClO• + OH• → HCl + O2

(12)

Cl radicals can further attack DCAC, leading to the formation of CO, HCl, CCl4 , CHCl3 and COCl2 as final products. The effect of temperature on the removal chemistry and byproduct formation of TCE is studied by Hsiao et al. [53]. Experiments, carried out with a pulsed corona and a DBD reactor, have shown that the removal of TCE and the formation of COx depend on temperature but not on reactor type. Moreover, higher temperatures

Fig. 2. TCE decomposition mechanism. Reprinted from Ref. [54], with permission from Elsevier.

cause a decrease in energy yield for TCE. The formation of byproducts (CO, CO2 , COCl2 , HCl and DCAC) is almost the same as found by Evans et al. [51]. Prager et al. [54] report the degradation of TCE with electron beam treatment. They found CO, HCl, COCl2 , DCAC and CHCl3 as main byproducts next to traces of CCl4 and CCl3 –COCl (trichloroacetylchloride; TCAA). In the proposed degradation mechanism (Fig. 2), OH radicals add to the double bond of TCE forming OH adducts. These adducts decompose and produce chlorine radicals or to a minor extend dichloromethyl radicals. Next, chlorine radicals add to the double bond and in a subsequent reaction with oxygen, the corresponding peroxyl radical is formed. In a bimolecular reaction step, molecular oxygen and alkoxy radicals are formed, which fragmentate to DCAC and chlorine which in turn re-enters the first chain reaction. In a second chain reaction, DCAC is further decomposed to HCl, COCl2 and CO. To minimize the formation of chloroacetic acids and phosgene, a wet scrubbing system is installed downstream of the electron beam system. Hakoda et al.

36

A.M. Vandenbroucke et al. / Journal of Hazardous Materials 195 (2011) 30–54

[55,56] also conclude that TCE decomposition with electron beam proceeds via a Cl radical addition chain reaction induced by OH radicals via the following reactions: C2 HCl3 + OH• → C2 HCl3 OH•

(13)

C2 HCl3 OH• + O2 → OOC2 HCl3 OH•

(14)

2OOC2 HCl3 OH∗ → 2OC2 HCl3 OH∗ + O2

(15)

OC2 HCl3 OH• → C2 HCl2 OH(O) + Cl•

(16)

Vitale et al. [57] examined the effect of a carbon-carbon double bond on electron beam treatment of TCE. The primary decomposition products found in their study are CO, CO2 , COCl2 , DCAC and HCl. Chloroform and TCAA were found as minor decomposition products. These researchers propose a reaction pathway in which dissociative electron attachment is believed to be the primary initiation step. This reaction produces chlorine radicals and a doubly chlorinated ethylene anion: C2 HCl3 + e− → CHClCCl− + Cl

(17)

In a study by Yamamoto and Futamura [27], the pronounced decomposition of TCE in dry nitrogen also strongly argues for dissociative electron attachment as the first stage in the decomposition of TCE. Vitale et al. propose that the chlorinated ethylene anion will likely decompose by direct oxidation: CHClCCl− + O2 → CHClCClOO− −

CHClCClOO → CHOCl + COCl

(18) −

CHOCl → CO + HCl −

COCl → CO + Cl

−

(19)

electrons seems to be the likely primary decomposition mechanism. These reactions initiate the detachment of Cl radicals, which in turn decompose more TCE molecules by Cl radical addition to the carbon-carbon double bound causing a chain reaction as proposed by Vitale et al. [57]. Futamura and Yamamoto [62] have applied a pulsed corona and a ferroelectric (BaTiO3 ) packed-bed reactor for TCE removal. In wet nitrogen, dichloromethane, chloroform, pentachloroethane, carbon tetrachloride, 1,1,2,2,- and 1,1,1,2-tetrachloroethanes and tetrachloroethylene are detected as major byproducts for the packed-bed reactor by using GC–MS (gas chromatography–mass spectrometry). Chloro- and dichloroacetylenes, (Z)- and (E)-1,2dichloroethylenes, and 1,1,2-trichloroethane are obtained as minor byproducts. With a pulsed corona reactor, 1,1,2-trichloroethane is the main byproduct along with tetrachloroethylene, (Z)-1,2dichloroethylene and negligible amounts of polychloromethanes. When air is used as carrier gas for the decomposition with the packed-bed reactor, only phosgene could be detected. For both reactors and for both carrier gases CO, CO2 , NOx and N2 O are also formed as byproducts. Formation of DCAC is, however, not observed in aerated conditions, which is in contrast with previous mentioned studies. The authors propose a plausible reaction mechanism under deaerated conditions. In the presence of O2 , they suggest that triplet oxygen molecules scavenge intermediate carbon radicals derived from TCE decomposition in an autoxidation process. Unstable alkylperoxy radicals are generated and further oxidatively decompose to render CO and CO2 , as shown in the following general reaction:

(20)

R • + O2 → ROO• → intermediates → CO + CO2

(21)

Urashima and Chang suggest that electron impact processes produce C, H, N radicals and negative ions. According to the authors the oxidation processes will take place directly by radicals or via oxidation of negative ions [63]. They propose a mechanism of TCE destruction based on 162 reactions [64]. In a study performed by Han and Oda [65], the effect of oxygen concentration on byproduct distribution is examined. TCE decomposition efficiency improves with decreasing oxygen content except for 0% oxygen. The formation of DCAC is maximal for 2% oxygen, while TCAA formation decreases with decreasing oxygen concentration. They suggest that oxygen species, like O(1 D) or other states in the discharge, react more strongly with the precursor of DCAC (CHCl2 –CCl2 • ) than that of TCAA (CCl3 –CH• ). When nitrogen is used as carrier gas, the GC–MS could detect HCl, Cl2 , C2 H2 Cl2 , CHCl3 , CCl4 and C2 HCl5 as byproducts. The authors suggest that collisions between TCE and electrons and (or) N2 excited species (N2 *) generate chlorine radicals. The main decomposition mechanism is considered to be the chlorine radical chain reaction as mentioned before by other authors.

Then, in a secondary autocatalytic radical reaction, chlorine radicals add to the least substituted carbon atom of the double bond of TCE resulting in the start of a chlorine radical chain reaction [58,59]. Bertrand et al. [60] suggest that addition to the least chlorinated site is favored over addition at the more chlorinated site by at least a factor 8. A possible chlorine addition reaction mechanism for the favored reaction is as follows: CHCl2 CCl2 + O2 → CHCl2 CCl2 OO

(22)

2 CHCl2 CCl2 OO → 2 CHCl2 CCl2 O + O2

(23)

CHCl2 CCl2 O → CHCl2 COCl + Cl

(24)

CHCl2 CCl2 O → CHCl2 + COCl2

(25)

CHCl2 + O2 → CHClO + Cl + O

(26)

CHClO → CO + HCl

(27)

DCAC decomposes to form HCl, COCl2 and chlorinated radicals through the following reaction: CHCl2 COCl + Cl (or O2 ) → CCl2 COCl + HCl (or H2 O)

(28)

CCl2 COCl + O2 → CCl2 OOCOCl

(29)

2 CCl2 OOCOCl → 2 CCl2 OCOCl + O2

(30)

CCl2 OCOCl → COCl + COCl2

(31)

COCl → CO + Cl

(32)

Phosgene may further decompose through Cl abstraction by chlorine, oxygen or other radicals forming CO and Cl2 or Cl radicals. The TCE removal rate is reduced by the presence of reaction products such as phosgene, HCl and DCAC through scavenging of electrons in the plasma which could otherwise initiate more dissociative electron attachment reactions of TCE. The study of Penetrante et al. [61] shows that for small initial concentrations of TCE in dry air, the reaction with O radicals and

(33)

2.1.3.2. Benzene. Benzene has attracted attention for NTP removal because it is a carcinogenic compound that has detrimental effects on human health. Table 3 summarizes published papers on benzene removal with NTP. In order to minimize operation costs for NTP removal it is important to optimize the operation conditions. Ogata et al. [40] investigated the effects of properties of ferroelectric materials, AC frequency, initial concentration of benzene and the concentration of O2 in the background gas for the removal of benzene in air using a ferroelectric packed-bed reactor. Under dry conditions benzene removal results in a low CO2 -selectivity and in the formation of various byproducts, such as CO, C2 H2 , N2 O, NO and NO2 . To improve this technique for practical applications, Ogata et al. [66] have studied the effect of water vapor on the removal of benzene with a ferroelectric packed-bed reactor. They suggest that a portion of the lattice oxygen species in BaTiO3 pellets is deactivated

A.M. Vandenbroucke et al. / Journal of Hazardous Materials 195 (2011) 30–54

37

efficiency of benzene to a large extent and proves that DBD treatment is competitive with other technologies although formation of solid residues and aerosol particles are issues that must be solved to secure an effective operation. In a study by Jiang et al. [72], a DC microhollow cathode glow discharge is applied to remove 300 ppm benzene from dry air. The authors use a zero-dimensional plasma chemistry code (KINEMA) [73] to model the benzene dissociation mechanism in a benzene/dry air mixture plasma. Important dissociation reactions predicted by the model are: C6 H6 + O• → C6 H5 + OH

(34)

C6 H6 + O2 → C6 H5 + HO2 −

C6 H6 + e → C6 H5 + H + e

Fig. 3. Effect of input power on benzene removal efficiency and energy yield. Reprinted from Ref. [71], with permission from Elsevier.

by adsorption of H2 O on the surface of the pellets. This results in a suppressed formation of CO and N2 O, a higher CO2 -selectivity and a lower decomposition of benzene. These observations are confirmed by Kim et al. [67]. The negative effect of humidity on benzene removal is ascribed to the interaction of water vapor with the surface of BaTiO3 pellets which may alter the surface state. As a consequence, plasma properties may be negatively affected [68], resulting in slower chemical destruction pathways. Cal and Schluep [69] investigated the decomposition of benzene in a DBD reactor as a function of the relative humidity (RH) without the presence of ferroelectric pellets. In both dry and wet gas streams, near complete destruction (>99.9%) of benzene is achieved and no intermediate hydrocarbons are observed with GC–MS. However, in wet gas streams the mineralization degree is greatly improved compared to dry air. Unfortunately, at high RH, a polymeric film is produced on the dielectric plates which slowly decreases the removal efficiency of benzene through time. Lee et al. [70] also used a DBD discharge to decompose 100 ppm of benzene in air. The authors suggest a plausible reaction mechanism that includes the formation of all byproducts detected by GC–MS and FT-IR (Fourier transform-infrared) spectroscopy. According to the authors, the plasma can produce O radicals from O2 , which can react with benzene to form CO2 , H2 O and benzene cation by a series of reactions. Benzene could directly be decomposed by the plasma to form phenol and benzenediol. The plasma is also capable of decomposing stable CO2 to form CO radicals that would add to phenol. This leads to the formation of secondary products such as benzaldehyde and benzoic acid. Finally, decomposition of H2 O by the discharge forms H and OH radicals which lead to the formation of benzene, phenol and benzenediol. Also, Ye et al. [71] have investigated the feasibility of benzene destruction with a DBD discharge. Experiments are carried out with a laboratory scale and a scale-up DBD reactor. With the former reactor, high removal efficiencies are obtained with lower flow rates, lower initial concentrations and higher input power (Fig. 3). In contrast, higher initial concentration and input power provide a high-energy efficiency for benzene removal. For the scale-up reactor, adding DBD systems in series can enhance the decomposition efficiency to a large extent. However, after a certain treatment time brown polymeric deposits are formed on the inside wall of the reactor which can finally lead to mechanical failure of the dielectric due to thermal energy built up. The deposit can be removed by passing air through the reactor at 6 kV for several minutes. GC–MS analysis revealed that phenol, hydroquinone and nitrophenol are the main products contained in the deposition. The feasibility study shows that multiple DBD systems in series can enhance the removal

(35) −

(36)

Modeling results reveal that the dominant dissociation reactions for benzene destruction in the DC glow discharge are atomic oxygen impact reactions. They suggest that the benzene destruction rate and efficiency are limited due to atomic oxygen losses in the boundary layer of the dielectric walls, which confine the discharge in the direction perpendicular to the gas flow direction. Satoh et al. [73] have applied a positive DC corona discharge between a multi-needle and a plane electrode for the removal of 300 ppm benzene in different N2 /O2 mixtures. Analysis of the exhaust stream is performed with FT-IR and shows C2 H2 , HCN, NO and HCOOH as intermediate products and CO2 as an end product. At low oxygen concentrations (0.2%) benzene is primarily converted into CO2 via CO, whereas at high oxygen concentrations (20%) benzene is converted into CO2 via CO and HCOOH. After treatment, benzene fragments are deposited on the plane electrode and discharge chamber at low oxygen concentrations. It is found that an increase in the oxygen concentration inhibits the decomposition of benzene, which is also the case with a DBD discharge [74]. However, with a packed-bed reactor, a higher N2 /O2 ratio improves the decomposition of benzene [40] which indicates that the effect of the amount of oxygen in the background gas depends on the type of discharge. Kim et al. [75] also investigated the influence of oxygen and found an optimum O2 concentration of 3–5% for benzene removal with a DBD discharge. Further increase of the oxygen concentration drastically decreases the decomposition efficiency. They suggest that higher benzene destruction at lower O2 partial pressure is due to the contribution of N radicals and excited N2 molecules. Comparison of the reaction rate constants indicates that the reaction with N2 (A3 +u ) is more plausible and is even faster than the reaction with O radicals (k = 1.6 × 10−14 cm3 molecule−1 s−1 ). However, as O2 partial pressure increases, quenching of N2 (A3 +u ) becomes significant [76] and the rate of reaction slows down. In addition, more O atoms are produced due to direct electron-impact dissociation and collision dissociation by N2 (A3 +u ), but at the same time O atoms are also consumed in the formation of O3 . Because the gas-phase reaction between ozone and benzene is very slow (k = 1.72 × 10−22 cm3 molecule−1 s−1 ), it does not contribute to the decomposition of benzene. N2 (A3 +u ) + C6 H6 → products k = 1.6 × 10−10 cm3 molecule−1 s−1 N + C6 H6 → products k < 10−15 cm3 molecule−1 s−1 N2 (A3 +u ) + O2 → N2 + O2

(37) (38)

k = 2.4 × 10−12 cm3 molecule−1 s−1 (39)

N2 (A3 +u ) + O2 → N2 + 2 O k = 2.5 × 10−12 cm3 molecule−1 s−1 (40)

38

A.M. Vandenbroucke et al. / Journal of Hazardous Materials 195 (2011) 30–54

C6 H5 CH3 + O3 → C6 H5 CHO2 + H2 O k = 1.5 × 10−22 cm3 molecule−1 s−1

Fig. 4. Gas chromatogram of surface washed ethanolic solution from 10 min DBD discharge. Inset: Gas chromatogram of ethanolic solution of various standards employed. Reprinted from Ref. [77], with permission from Elsevier.

In a recent study by Dey et al. [77] the formation of byproducts of benzene oxidation in a Ar/O2 flow with a DBD reactor is carefully analyzed with GC–FID (flame ionization detector) and GC–MS. A plausible sequential reaction mechanism is given to rationalize the observation of the various byproducts. In the gas phase, only phenol and biphenyl are detected at a maximum conversion of 3%. GC analysis of an ethanolic solution of the polymeric deposit on the dielectric surface reveals the presence of substituted phenols besides phenol and biphenyl (Fig. 4). It is suggested that the intermediate phenyl radical plays the role of the primary precursor. 2.1.3.3. Toluene. Table 4 tends to give an overview of published work on toluene removal with the aid of NTP. Toluene can be regarded as the most studied VOC for abatement on laboratory scale. Therefore only a selection of papers will be discussed here. Other references can be found in Table 4. Kohno et al. [78] have applied a DC capillary tube discharge reactor and investigated the effect of gas flow rate, initial toluene concentration and reactor operating conditions. According to the authors, the following destruction process can be expected in a NTP environment: C6 H5 CH3 + e− → products

k = 10−6 cm3 molecule−1 s−1

(41)

C6 H5 CH3 + O+ , O2 + , N, N2 + → C6 H5 CH3 + + O, O3 , N, N2 k = 10−10 cm3 molecule−1 s−1 C6 H5 CH3 + + e− → C6 H5 + CH3

(42) k = 10−7 cm3 molecule−1 s−1 (43)

C6 H5 CH3 + OH → C6 H5 CH3 OH k = 5.2 × 10−12 cm3 molecule−1 s−1

(47)

FT-IR spectroscopy detects CO2 , CO, NO2 and H2 O as gaseous byproducts and a significant amount of brown particles are deposited at the exit of the reactor. It is suggested that CO2 and CO mainly form carbon and nitrogen hydride bonded aerosol particles and tars. CO2 and H2 O are observed as main reaction products by Mista and Kacprzyk [79]. They also detect a thin polymeric film (brown residues) covering the discharge electrode and dielectric layer. Operation at higher energy densities can successively be applied to oxidize the condensed polymeric species to CO2 . Machala et al. [80] suggest that formation of aerosols including peroxy-acetylnitrates species (PANs) may be possible during toluene removal through a mechanism that is similar to formation of photochemical smog in the atmosphere. In pure nitrogen [81], GC–MS analysis showed that N2 plays a major role in the polymerisation process through the formation of C–N C and C–(NH)–bonds. A proposal of the polymerisation process is given to explain the formation of micrometric sized particles in the plasma reactor. In Ref. [82] a wire plate DBD has been used to examine the humidity effect on toluene decomposition. A maximum removal efficiency of 73% was achieved in a gas stream containing 0.2% H2 O in N2 with 5% O2 . This controlled humidity is governed by two opposite effects: as humidity increases, more H2 O molecules collide with high-energy electrons and form OH radicals, resulting in a higher removal efficiency. On the other hand, the electronegative characteristic of H2 O limits the electron density in the plasma and quenches activated chemical species, as concluded by Van Durme et al. [83]. Kim et al. [75] have confirmed that 5% O2 is the optimum oxygen partial pressure in a dry nitrogen stream, as is the case for benzene. Recently, Schiorlin et al. [84] have tested three different corona discharges (positive DC, negative DC, positive pulsed) for toluene removal and have observed that process efficiency increases in the order positive DC < negative DC < positive pulsed. By investigating the effect of humidity on the removal efficiency, it is concluded that for both negative DC and positive pulsed corona, OH radicals are involved in the initial stage of toluene oxidation. When the RH was greater than 60%, removal efficiency slightly drops due to saturation and inhibition of the OH radical forming reactions, i.e. dissociation of H2 O molecules induced by interaction with electrons or by reaction with O(1 D). A positive DC corona discharge has been applied by Van Durme et al. [83] in order to abate toluene from indoor air and to unravel the degradation pathway. The removal of toluene is achieved with a characteristic energy density of 50 J/L. Fig. 5 shows that partially oxidized intermediates are formed under the applied conditions. By determining the effect of humidity, the authors find out that OH radicals play a major role in the oxidation kinetics due to initiation by H-abstraction or OH-addition. The byproducts detected by GC–MS consist of benzaldehyde, benzylalcohol, formic acid, nitrophenols and furans.

(44) 2.2. Combined with catalyst

C6 H5 CH3 + OH → C6 H5 CH2 + H2 O k = 7 × 10−13 cm3 molecule−1 s−1

(45)

C6 H5 CH3 + O → C6 H5 CH2 O + H k = 8.4 × 10−14 cm3 molecule−1 s−1

(46)

2.2.1. What is plasma–catalysis? Many studies have shown that NTP is attractive for the removal of NOx , SOx , odours and VOCs. There is, however, a consensus among researchers that application of NTP for VOC abatement suffers from 3 main weaknesses, i.e. incomplete oxidation with emission of harmful compounds (CO, NOx , other VOCs), a poor energy efficiency and a low mineralization degree.

A.M. Vandenbroucke et al. / Journal of Hazardous Materials 195 (2011) 30–54

39

Fig. 5. Chromatogram of GC–MS analysis for the identification of toluene degradation products. Reprinted from Ref. [83], with permission from Elsevier.

The combination of NTP with heterogeneous catalysts can be divided in two categories depending on the location of the catalyst: in-plasma catalysis (IPC) and post-plasma catalysis (PPC). The latter is a two-stage process where the catalyst is located downstream of the plasma reactor while the former is a single stage process with the catalyst being exposed to the active plasma. In literature, several different terms and corresponding abbreviations have already been proposed to represent IPC and PPC. For in-plasma catalysis, one can find among others: plasma-driven catalysis (PDC) [13], inplasma catalysis reactor (IPCR) [85], single-stage plasma–catalysis (SPC) [86], plasma and catalyst integrated technologies (PACT) [87] or combined plasma catalysis (CPC) [88,89]. For PPC, the following terms have been proposed: plasma-enhanced catalysis (PEC) [13], post-plasma catalysis reactor (PPCR) [85], two-stage plasma catalysis (TPC) [86]. In plasma–catalysis, synergetic effects are related to the activation of the catalyst by the plasma. Activation mechanisms include ozone, UV, local heating, changes in work function, activation of lattice oxygen, adsorption/desorption, creation of electron–hole pairs and direct interaction of gas-phase radicals with adsorbed pollutants [75]. The plasma–catalyst interactions described in the following paragraphs contribute to one or more of these catalyst activation mechanisms. The presented experimental findings, applying to specific working conditions, may appear as scattered pieces of information. Indeed further research is needed to connect the loose ends and unravel the detailed mechanisms. However, it is meaningful to try and extract some general pathways at this stage. 2.2.1.1. Influence of the catalyst on the plasma processes. Discharge mode: The physical properties of a discharge will be affected if a catalyst is introduced into the discharge zone. When for example a dielectric surface is introduced in the gap of a streamer-type discharge, the discharge mode at least partially changes from bulk streamers to more intense streamers running along the surface (surface flashover) [90]. Similar field effects can lead to higher average electron energies when the discharge zone is filled with ferroelectric pellets, leading to a more oxidative discharge [91]. Parameters that influence the effect of the packed bed on the discharge are the dielectric constant of the pellet material and the size and shape of the pellets. The dielectric constant affects the electric field in the void between the pellets and thereby the mean electron energy. With increasing pellet size the number of microdischarges decreases, but the amount of charge that is transferred per microdischarge increases [92]. Reactive species production: Obviously, introducing a heterogeneous catalyst changes the physical characteristics of the discharge, so the chemical activity will be affected as well. Roland et al. [18]

studied the oxidation of various organic substances immobilized on porous and non-porous alumina and silica catalysts and concluded that short-living active species are formed in the pore volume of porous materials when exposed to NTP. On the other hand, introducing a catalyst can reduce the concentration of ionic species [93]. However, this effect did not impair the catalyst’s role in reducing the emissions of ozone and carbon monoxide for this particular application (indoor air control).

2.2.1.2. Influence of the plasma on the catalytic processes. Catalyst properties: Non-thermal plasmas are used for catalyst preparation [94–99]. Plasma treatment of the catalyst enhances the dispersion of active catalytic components [100,101] and influences the stability and catalytic activity of the exposed catalyst material [102]. The oxidation state of the catalyst can also be altered by NTP. For instance, when a Mn2 O3 catalyst is exposed for a long time to a DBD plasma, X-ray diffraction spectra reveal the presence of Mn3 O4 , a lower-valent manganese oxide with a larger oxidation capability. Due to plasma–catalyst interactions, less parent Ti–O bonds are found on TiO2 surfaces after several hours of discharge operation [103]. Even new types of active sites with unusual properties may be formed [104], such as stable Al–O–O* with a lifetime exceeding more than two weeks, as observed in the pores of Al2 O3 in IPC experiments [104]. Plasma exposure can result in an increase or decrease of the specific surface area or in a change of catalyst structure [100,102,105]. Adsorption: Adsorption processes play an important role in plasma–catalytic reaction mechanisms. If the catalyst has a significant adsorption capacity for pollutant molecules, it prolongs the pollutant retention time in the reactor. In the case of IPC, the pollutant concentration in the discharge zone is increased. The resulting higher collision probability between pollutant molecules and active species enhances the removal efficiency. Adsorption of VOC and active species increases with the porosity of the catalyst [106]. Under conditions where plasma-generated ozone is not effective in itself to destroy pollutants, high decomposition rates are obtained due to the adsorption of ozone on the catalyst surface and the subsequent dissociation into atomic oxygen species [107]. Humidity is a critical parameter in plasma–catalytic processes. The adsorption of water on the catalyst surface results in a decrease of the reaction probability of the VOC with the surface and therefore reduces the catalyst activity [82]. Thermal activation: Although gas heating will result in higher catalyst surface temperatures [108], the heating effect is in general too small to account for thermal activation of the catalyst. However, hot spots can be formed in packed-bed reactors as a result of localized heating by intense microdischarges that run between

40

A.M. Vandenbroucke et al. / Journal of Hazardous Materials 195 (2011) 30–54

sharp edges and corners of adjacent pellets. Increased catalyst temperatures can promote catalytic VOC removal [109]. Plasma-mediated activation of photocatalysts: In photocatalysis, VOCs are adsorbed on the surface of a porous semiconductor material that is exposed to UV radiation. The UV photons generate electron–hole pairs, inducing the subsequent oxidation of the adsorbed VOC by valence band holes. In a final step the oxidation products are desorbed. Among other photocatalysts (e.g. ZnO, ZnS, CdS, Fe2 O3 , WO3 ), TiO2 is one of the most efficient for the decomposition of a wide range of VOCs. Moreover, the combination of TiO2 with NTP results in higher oxidation efficiencies and better selectivity to CO2 . For the anatase phase of TiO2 , having a bandgap of 3.2 eV, it takes a photon with a wavelength shorter than 388 nm to create an electron–hole pair. Although there are excited nitrogen states that emit light in this wavelength range, there is experimental evidence that photocatalysis induced by UV light from the plasma cannot explain the observed synergy in several hybrid plasma/TiO2 systems reported in literature. For instance, Sano et al. [110] has detected no enhancement in acetylene conversion when the reactor walls are coated with TiO2 . Emission spectra of the surface discharge plasma with and without catalyst coating reveal that UV light from the plasma is absorbed by TiO2 , but the intensity is too weak for photoactivation. This observation has been confirmed by Huang et al. [111], who employed a wire-cylinder DBD reactor with a photocatalyst sheet stuck along the inner wall of the tube. Kim et al. have tested a DBD reactor packed with Ag/TiO2 for benzene removal [112]. When O2 -benzene mixtures are diluted with argon, significantly higher decomposition efficiencies are observed compared to N2 dilution. This result suggests that the role of UV light for photoactivation is negligible because light emission from excited argon ranges in the visible range (400–850 nm). However, other groups report that UV light emitted from the plasma can act as a source for activation of TiO2 [70,113,114]. Subrahmanyam et al. [115] suggest that the increased activity with sintered metal fibres modified with TiO2 might be related to activation as well as to photocatalytic action in the presence of UV light emitted by the plasma discharge. In some cases, TiO2 shows plasma-induced catalytic activity under conditions where there is no or very little UV emitted by the plasma [112,116]. Direct plasma activation has been observed when TiO2 is exposed to an atmospheric pressure argon discharge at room temperature [117]. The question then arises how the plasma-exposed TiO2 is activated, if not by UV photons. Different mechanisms to bridge the TiO2 band-gap by plasma-driven processes can be envisaged, but to date there is insufficient information to elaborate on the relative importance of electrons, ions, metastables, charging effects, surface recombination, etc. 2.2.2. Different types of catalysts As in classical heterogeneous catalysis, the catalyst material can be introduced in the hybrid system in different ways for both IPC (Fig. 6) and PPC: in the form of pellets (a so-called packed-bed configuration) [91,118–121], foam [82,100,122,123,102] or honeycomb monolith [93,124–127], as a layer of catalyst material [128] or as a coating on the reactor wall [110,129] or electrodes [115,130–135]. Many catalysts have been tested for VOC abatement with IPC and PPC. Historically, the first materials tested were porous adsorbents placed inside the discharge region as in references [41,136]. The idea is that, by introducing these materials, the retention time of VOC molecules would increase along with the probability of surface reactions with active chemical plasma species (electrons, radicals, ions, photons). Adsorbents that were used to achieve a more complete oxidation are ␥-Al2 O3 [17,18,136,137] and zeolites or molecular sieves [17,138–142]. Furthermore, these materials are coated or impregnated with (noble) metals such as silver, palladium, platinum, rhodium, nickel, molybde-

Fig. 6. Most common catalyst insertion methods for IPC configuration. Reprinted from Ref. [228], with permission from Elsevier.

num, copper, cobalt or manganese to provide catalytic activity [75,105,107,109,142–151]. Adsorbents also function as support for metal oxides [20,149,152–158]. Extensive attention has been given during recent years to the use of photocatalysts, in particular to TiO2 . In most studies, TiO2 is inserted in the discharge region in order to achieve activation through different mechanisms. This catalyst has also been coated with (noble) metals [67,75,107,112,116,159,160] and metal oxides [138,161–163]. Additionally, it has been used as a coating on activated carbon filter [111] or fibre [164], on glass fibres [165,166] or beads [70,113,167], nickel foam [168], silica gel pellets [129] and on UV lamp [169]. 2.2.3. VOC abatement Tables 6–9 give a summary of literature on VOC removal with plasma–catalysis. For each paper catalyst information and operating conditions are presented along with the maximum removal efficiency and energy density. In this section, particular attention is again paid to the most studied target compounds, i.e. trichloroethylene, benzene and toluene. Table 9 presents a list of other relevant, but less frequently studied VOCs that have been examined in plasma–catalytic studies. For more details about operating conditions and results, the reader can consult the corresponding references. 2.2.3.1. Trichloroethylene. Table 6 presents published papers regarding TCE abatement. Oda et al. [162] have investigated the effect of TCE initial concentration, pellet size and sintering temperature for TiO2 catalysts on the TCE decomposition performance. When the barrier type reactor was filled with TiO2 sintered at 673 K, the breakdown voltage to generate NTP greatly reduces in comparison to the empty reactor and the reactor filled with TiO2 sintered at 1373 K. They suggest that the nonuniform geometrical distribution of the disk-like dielectric pellets sintered at 673 K disturbed the electric field and generated an electric field concentration at the contacting area of the pellets. This results in the formation of contacting point discharges or surface discharges on the pellet surfaces, lowering the breakdown voltage and improving the decomposition energy efficiency. Moreover, they indicate that too fine TiO2 particles disturb the gas flow and cause insufficient filling of the discharge area with plasma. In another study by Oda et al. [170], MnO2 is used as a postplasma catalyst in a direct (contaminated air is directly processed by the plasma) and an indirect process (plasma-processed clean air is mixed with the contaminated air). Manganese oxide is very effi-

A.M. Vandenbroucke et al. / Journal of Hazardous Materials 195 (2011) 30–54

41

Table 6 Overview of published papers on TCE removal with plasma–catalysis. Plasma type

Catalyst

Position

Tcat (K)

Carrier gas

Flow rate (mL/min)

Concentration Maximum range removal (ppm) efficiency (%)

Energy density (J/L)

References

DBD DBD

MnO2 TiO2 /SMF CoOx (3 wt%)/SMF MnOx (3 wt%)/SMF TiO2 /MnOx /SMF MnOx (3 wt%)/SMF MnO

PPC IPC

293 –

Air Air

500 700

250 250

95–99 >99

240 –

[65] [131]

IPC IPC

293 293

Air Dry air

500 10002000

150–200 250

95–99 >99>99

550 120

[133] [144]

DBD

TiO2 V2 O5 (0.7 wt%)/TiO2 V2 O5 (4.6 wt%)/TiO2 WO3 (4.2 wt%)/TiO2

IPC

293

Dry air

400

100

>99 95–99 90–95 >99

180 140 140 180

[161]

DBD

TiO2 sintered at 1373 K 0.5–1 mm 1–2 mm 2–3 mm TiO2 sintered at 673 K MnO2 Au/SBA-15 TiO2 Pd(0.05 wt%)/Al2 O3

IPC

293

Dry air

400

1000

DBD DBD

DBD DBD DC positive corona DC negative glow Surface discharge

V2 O5 /TiO2 Cu–ZSM-5

[162] >99 >99 >99 >99

200 120 120 120

IPC PPC IPC PPC

293 – – 373

Dry air Dry air Dry air Humid air

400 510 1500 2000

1000 430 100 600–700

>99 >99 85 80

120 670 600 300

[170] [172] [118] [147]

IPC

293

Dry air

400

1000

>95 >95

50 50

[143]

cient in enhancing the decomposition efficiency for both processes. The catalysts effectiveness to dissociate ozone generates oxygen radicals which are excellent oxidizers for TCE removal. Han et al. [171] have further examined the effect of the manganese dioxide post-plasma catalyst for the direct and indirect process. For the direct process oxygen species, generated from collisions between excited species (or electrons) with O2 , mainly oxidize TCE into DCAC. The increased decomposition efficiency for the direct process is ascribed to the oxidation of the remaining TCE into trichloroacetaldehyde (CCl3 –CHO, TCAA) by oxygen species produced during ozone decomposition at the surface of MnO2 . The COx yield increases from 15% to 35% at an energy density of 120 J/L when MnO2 is present. When the energy density is raised to 400 J/L, a COx yield of 98% is established. For the indirect process, similar

Fig. 7. Selectivity to CO and CO2 as a function of input energy for inner electrodes made of SMF and MnOx /SMF. Reprinted from Ref. [133], with permission from Elsevier.

conclusions are made although the COx yield is not as good as for the direct process. Magureanu et al. [133] have tested a plasma–catalytic DBD reactor with an inner electrode made of sintered metal fibres (SMF) coated by transition metal oxides. Fig. 7 shows the CO and CO2 selectivity over the range of energy densities used. The selectivity to CO2 reaches 25% with the SMF and showed a significant improvement with MnOx /SMF, up to 60%. The use of MnOx /SMF does, however, not substantially lower the selectivity to CO. Thus, as compared to the reactor with SMF electrode, TCE conversion and CO2 selectivity were significantly enhanced using MnOx /SMF. The ability of MnO2 to decompose ozone in situ, produces strong oxidizing atomic oxygen species on the catalyst surface. These species may lead to an enhanced oxidation of TCE resulting in a high CO2 selectivity [115,131,133]. After reaction, XPS (X-ray photoelectron spectroscopy) analysis of the catalyst has revealed that both manganese and iron have preserved their initial oxidation state. The used catalyst, however, shows an enrichment of iron on the catalyst surface suggesting a redispersion of manganese on the surface during reaction. Finally, XPS also reveals some chlorine deposition on the catalyst surface after reaction. In another study conducted by Magureanu et al. [172], gold nano-particles embedded in SBA-15 have been tested for PPC. The catalyst with the least amount of Au (0.5 wt%) seems to enhance the COx selectivity the most and has the best catalytic performance. As for MnO2 , the Au/SBA-15 can dissociate ozone, produced in the plasma, to oxygen radicals that decompose TCE. They suggest that in the presence of ozone generated in the plasma, isolated gold cations are the active sites that elucidate the catalytic behaviour. To achieve a more complete oxidation of TCE at a reduced energy cost, Morent et al. [118] have used a hybrid plasma–catalyst system with cylindrical TiO2 pellets for IPC. They suggest that the increased removal fraction for the plasma–catalytic system can be explained through adsorption on and/or photoactivation of TiO2 . Adsorption of TCE molecules on the surface of TiO2 increases the residence time of TCE in the discharge.

42

A.M. Vandenbroucke et al. / Journal of Hazardous Materials 195 (2011) 30–54

Table 7 Overview of published papers on benzene removal with plasma–catalysis. Flow rate (mL/min) Concentration range Maximum removal (ppm) efficiency (%)

Energy density (J/L)

References

Dry air

100 500 500

100 105 105

60 54 50

900 360 320

[93]

293

Dry air

250

300–380

12 16

170 170

[108]

IPC

–

Dry air

400

200

90 >99 >99

3150

[163]

Ag(2 wt%)/TiO2 Kr/I2 (KrI* excimer UV radiation) TiO2 TiO2 /Al2 O3 silica gel

IPC IPC

373 293

Dry air Dry air

4000 13 × 103 –130 × 103

110 30–940

>99 66.5

125 –

[176] [230]

IPC IPC IPC

293 – –

Dry air Dry air Dry air

– 200 100

188 100 300

98 50 85

– 140 –

[262] [70] [90]

Multistage corona a

TiO2 Sol–gel TiO2 Pt/Sol–gel TiO2

IPC

293

Dry air

60

1500

92.7 91.7 >99

–

[222]

Surface discharge

Ag(1 wt%)/TiO2

IPC

373

Dry air Humid air

200–3000

200–210

89 86

383 391

[67]

Surface discharge

Ag(4 wt%)/TiO2 IPC Ni(2 wt%)/TiO2 Ag(0.5/5 wt%)/Al2 O3 Pt(0.5 wt%)/Al2 O3 Pd(0.5 wt%)/Al2 O3 Ferrierite Ag(2 wt%)/H–Y

373

Dry air

4000–104

200

>99 >99 >99 >99 >99 >99 >99

89–194

[75]

Packed-bed DBD

TiO2 Pt(1 wt%)/TiO2 V2 O5 (1 wt%)/TiO2

IPC

373

Dry air

2000

203–210

82 80 90

388 391 383

[174]

BaTiO3 packed-bed

TiO2 Ag(0.5 wt%)/TiO2 Al2 O3 Ag(0.5 wt%)/Al2 O3 TiO2 Ag(0.5 wt%)/TiO2 Al2 O3 Ag(0.5 wt%)/Al2 O3

IPC

292

Dry air

1000

500

60

[180]

PPC

292

66 60 52 49 34 46 28 39

Plasma type

Catalyst

Position Tcat (K) Carrier gas

DBD

TiO2 MnO2 TiO2 –silica

IPC

–

DBD

TiO2 MnO2

IPC

DBD

TiO2 Pt(1 wt%)/TiO2 V2 O5 (1 wt%)/TiO2

DBD DBD DBD DBD glow discharge Pulsed corona

a

Four stages in serie.

To confirm the presence of excited species of nitrogen, Subrahmanyam et al. give an UV–vis emission spectrum of the DBD plasma discharge in the wavelength range 250–500 nm. It is proven that emission of excited nitrogen molecules (N2 *) is in the range of the band gap of the TiO2 /SMF catalyst [115]. They suggest that the increased activity of TiO2 /SMF might be due to photocatalytic action in the presence of UV light as well as activation of TiO2 by the plasma discharge. Vandenbroucke et al. [147] have investigated the use of a DC glow discharge combined with Pd/␥-Al2 O3 located in an oven downstream. When the catalyst temperature was set at 373 K, the combined system showed synergistic effects on the removal of TCE. By comparing the experimental removal efficiency of the hybrid system with the removal calculated by multiplying the individual effects (plasma and catalyst alone), 12–22% additional TCE was decomposed. A more elaborated review on plasma–catalytic abatement of TCE can be found in [173]. 2.2.3.2. Benzene. Ogata et al. [136] have performed much research on the removal of benzene with plasma–catalysis. In a first study, they test an adsorbent hybrid reactor packed with a mixture of BaTiO3 and Al2 O3 pellets and compare the results with a BaTiO3

packed reactor and a two-stage reactor (BaTiO3 packed reactor with Al2 O3 downstream). The hybrid reactor shows the best performance, owing to its better energy efficiency, CO2 -selectivity and suppressed N2 O formation. The combined effect of benzene concentration on Al2 O3 followed by surface decomposition and gas-phase reaction is thought to be responsible for the enhanced decomposition. Cyclic operation of adsorption and plasma discharge is suggested to further improve the energy efficiency. In Ref. [137] they continue examining a catalyst hybrid reactor with metal supported Al2 O3 and have found that Ag-, Co-, Cu and Ni-supported Al2 O3 shows a slightly better CO/CO2 ratio and a lower N2 O formation than the adsorbent hybrid reactor. Next, a zeolite hybrid plasma reactor (mixture of zeolite and BaTiO3 ) has been applied for dilute benzene decomposition [139]. The higher adsorption capacity of zeolite structures compared to alumina allows a higher decomposition efficiency and CO/CO2 ratio if the micropore surface area is large enough for accommodation of benzene molecules. The authors have also found that benzene adsorbed outside of a zeolite crystalline pore decomposed more easily than that inside a zeolite pore. In Ref. [145] they expand the study and examine the effect of BaTiO3 pellet size and mixing ratio of BaTiO3 and adsorbent, catalyst or zeolite. Plasma energy is found to be almost independent of the pellet size. However, with pellets larger than 2 mm in diam-

A.M. Vandenbroucke et al. / Journal of Hazardous Materials 195 (2011) 30–54

43

Table 8 Overview of published papers on toluene removal with plasma–catalysis. Plasma type

Catalyst

Position

Tcat (K)

Carrier gas

Flow rate (mL/min)

Concentration range (ppm)

Maximum removal efficiency (%)

Energy density (J/L)

References

DBD

Al2 O3

IPC

Air (18% RH)

104

220

[17]

Fe2 O3 /MnO honeycomb SMF CoOx (3 wt%)/SMF MnOx (3 wt%)/SMF

PPC

Dry air

2500

85

60–65 80 65

110

DBD

293 373 293

72

[127]

IPC

–

Air

500

250

60 70 65

160

[131]

[134]

DBD

DBD

SMF CoOx (3 wt%)/SMF MnOx (3 wt%)/SMF Cu

IPC

–

Air

500

500

90 92 95 90

298

DBD

MnPO4 Mn–APO-5 Mn–SAPO-11

PPC

673

Air

50–150

560

70 65 70

900–2700 [142]

DBD

Ag/TiO2

IPC

373

Air

4000

101

95

125

[176]

DBD

Ti–MPS Mn(5 wt%)–Ti–MPS Mn(10 wt%)–Ti–MPS

PPC

–

Air

200

1000

45 58 75

300

[263]

Wire-cylinder DBD

TiO2 /activated carbon filter Al2 O3 TiO2 /Al2 O3 MnO2 (5 wt%)/Al2 O3 MnO2 (10 wt%)/Al2 O3 MnO2 (15 wt%)/Al2 O3

IPC

–

Air (0.5% H2 O)

200

100

55

–

[111]

IPC

–

Dry air

2000

186

75 86 84 96 96

700

[20]

TiO2 /glass pellets TiO2 /Al2 O3 /Ni foam MnO2 /Al/Ni foam Mn-1 Mn-2 Mn-3 N150 (MnO2 –Fe2 O3 ) Al2 O3 MnO2 (9 wt%)/Al2 O3 Activated carbon (AC) MnO2 (3 wt%)/AC

IPC PPC IPC PPC

293 – – 300

Dry air Dry air 5% O2 /N2 Air

600 200 100 300

1100 50 50 200

80 95 >95 90–95

1000 900 750 1400

[167] [122] [100] [264]

PPC

–

Air

588

240

76 74 88 98.5 99.7

172

[202]

Wire-cylinder DBD

Wire-cylinder DBD Wire-plate DBD Wire-plateDBD DBD (pulsed)

DBD packed with glass beads

Multistage packed-bed DBD

MnO2 MnO2 –CuO

PPC

–

Air

104

70

>99 >99

340

[265]

BaTiO3 packed-bed

Al2 O3 Ag2 O(7 wt%)/Al2 O3 MnO2 (7 wt%)/Al2 O3 Al2 O3 Ag2 O(7 wt%)/Al2 O3 MnO2 (7 wt%)/Al2 O3

IPC

673 573 603 698 573 603

Dry air

1000

500

95 >99 >99 78 >99 >99

60

[153]

TiO2 Al2 O3 Ag(0.5 wt%)/Al2 O3 TiO2 Al2 O3 Ag(0.5 wt%)/TiO2 Ag(0.5 wt%)/Al2 O3

IPC

753

Dry air

1000

500

91

[180]

PPC

886

60 >99 >99 95 >99 95–99 99

PPC IPC

513 433

Air Dry air

2 × 104 1000

330 200

90–95 85

142 140–150

[126] [130]

IPC

–

Air

100

300

>95

5.43.5

[90]

Pulsed corona

Pt-honeycomb Reticulated vitreous carbon Pt/Rh coated electrodes AlO2 Silica gel Al2 O3

IPC

–

Air

400

1100

>99

1100

[181]

DC positive corona

TiO2

IPC PPC

–

Air

1000

80–100

75 70

160 330

[182]

DC positive corona

TiO2 CuO–MnO2 /TiO2

IPC PPC

293

Dry air

104

0.5

82 78

17 2.5

[186]

DC positive corona

Cu–Mn/TiO2 (a) N140 N150 Pd(0.5 wt%)/Al2 O3 Cu–Mn/TiO2 (b)

PPC

293

Air (50% RH)

104

0.5

40 47 34 47 62

14 16 16 10 20

[107]

BaTiO3 packed-bed

Pulsed corona Pulsed corona

Pulsed corona

PPC

44

A.M. Vandenbroucke et al. / Journal of Hazardous Materials 195 (2011) 30–54

Table 8 (Continued) Plasma type

Catalyst

Position

Tcat (K)

Carrier gas

Flow rate (mL/min)

Concentration range (ppm)

Maximum removal efficiency (%)

Energy density (J/L)

References

Wire-cylinder corona

TiO2 (3 wt%)/glass beads TiO2 (3 wt%)/Al2 O3

IPC

–

10% O2 /N2

4000

1000

70–75

–

[113]

Positive DC streamer Surface discharge

Surface discharge Pulsed wire-cylinder DBD

Cu–Mn/Al2 O3 Ni/cordierit honeycomb Mn–Cu/cordierit honeycomb V/cordierit honeycomb zeolites Pt/Al2 O3

80–85 PPC PPC

573 –

Air Air

133 × 10 666 × 103

45 30

96 40–45

20 –

[146] [93]

IPC IPC

– 468

Air (0.5% H2 O) Air

500 2 × 104

200 300

– 92

– –

[140] [266]

eter sparking occurs earlier. For the catalyst hybrid reactor (with metal supported Al2 O3 ), larger BaTiO3 pellets in comparison to catalyst pellets, are beneficial because high-energy plasma is formed around the contact points of the BaTiO3 pellets (Fig. 8). This shows the importance of the combination method to effectively induce catalytic properties. Kim et al. have also tested various catalyst formulations and reactor types to enhance the decomposition of benzene with NTP. A BaTiO3 packed-bed reactor has been modified by replacing the ferroelectric material with TiO2 , Pt/TiO2 or Ag/TiO2 pellets [174]. The reactor is placed in an oven that controls the temperature at 373 K. Experiments reveal that the catalytic activity for benzene decomposition is in the order Ag/TiO2 > TiO2 > Pt/TiO2 . The silver catalyst also improves the CO2 -selectivity with 15% compared to the BaTiO3 packed-bed reactor. Beside CO2 and CO, no other byproducts are formed, which is confirmed by good carbon balances. Results indicate that the energy density is the governing factor for benzene decomposition rather than the amount of Ag/TiO2 (grams of catalyst) in the reactor [175] or the gas residence time [176]. However, larger amounts of Ag/TiO2 slightly reduce the formation of N2 O. In this study, formic acid is found as minor byproduct at lower energy density. While a pulsed corona and surface discharge reactor form aerosols during benzene removal, negligible amounts are detected in the reactor packed with Ag/TiO2 [175]. In a subsequent study [177], Ag-loading amount on TiO2 (percentage of Ag on catalyst) confirms to have no effect on the benzene removal. This parameter,

3

however, plays an important role for the oxidative decomposition of intermediates on the TiO2 surface, indicated by the carbon balance. Larger Ag-loading seems to benefit the carbon balance and CO2 -selectivity. Further work has examined the activation mechanism of the Ag/TiO2 catalyst in the hybrid reactor [112]. Thermal catalytic experiments and comparison of the effects of dilution gases (Ar, N2 ) on benzene removal respectively reveal that temperature is not an important parameter and contribution of plasma generated UV light to the photoactivation of the catalyst is negligible. The authors therefore suggest that in situ decomposition of ozone over Ag/TiO2 and plasma-induced catalysis at higher energy density play a dominant role. The observed zero-order kinetics to benzene concentration supports the latter assumption. The catalyst shows good durability against catalyst deactivation for over 150 h of continuous operation tests [176]. Finally, Kim et al. [75] have tested a cycled system of adsorption and oxygen plasma as earlier proposed by Ogata et al. [136] and Song et al. [17]. Benzene oxidation is examined as function of oxygen partial pressure (0–80% O2 ) and different catalyst types (TiO2 , ␥-Al2 O3 , zeolites) inside the reactor. An increase of O2 partial pressure improves both the decomposition and CO2 -selectivity of benzene regardless of the catalyst used. Tests with the cycled system demonstrate that the regeneration mode must be done with pure oxygen to fully suppress harmful Nx Oy formation. The authors suggest a plausible reaction mechanism where removal of benzene mainly proceeds on the surface of the main catalysts (Fig. 9).

Fig. 8. Image of plasma discharge in a (a) BaTiO3 packed-bed DBD and in hybrid reactors with mixtures of (b) BaTiO3 > catalyst and (c) BaTiO3 < catalyst. Reprinted from Ref. [145], with permission from Elsevier.

A.M. Vandenbroucke et al. / Journal of Hazardous Materials 195 (2011) 30–54

45

Fig. 9. Plausible mechanism for IPC for VOCs on various catalysts. Reprinted from Ref. [75], with permission from Elsevier.

Fig. 10. Mechanism for MnO2 -catalyzed oxidation of benzene. Reprinted from Ref. [129], with permission from Elsevier.

Recently, Fan et al. [178] have also investigated a cycled system with a storage and a discharge stage packed with a metal supported zeolite (Ag/HZSM-5). High oxidation rate of adsorbed benzene as well as low energy cost (3.7 × 10−3 kWh/m3 ) are achieved at a moderate discharge power. Additionally, Ag/HZSM-5 exhibited good stability during cycled operation. In a study by Futamura et al. [129], a DBD discharge is applied to investigate the synergistic effect of filling the plasma reactor with different catalysts (TiO2 , MnO2 and TiO2 -silica gel). They suggest a mechanism for MnO2 -catalyzed oxidation of benzene (Fig. 10). Apparently, adsorption of ozone forms oxygen atoms on the MnO2

surface which partially desorb as O(3 P) in the gas phase, acting as possible oxidants for benzene decomposition. Park et al. [163] have attached sheet type catalysts (TiO2 , Pt/TiO2 and V2 O5 /TiO2 ) on the dielectric barrier of a DBD discharge. Benzene decomposition efficiency decreases in the order V2 O5 /TiO2 > Pt/TiO2 > TiO2 . Suppression of N2 O formation and improved mineralization degrees are obtained with all catalysts. Results indicate that high-energy electrons along with UV light generated from DBD plasma excite the TiO2 catalysts. A hybrid plasma-photocatalyst system has also been tested by Lee et al. [70]. Comparison of OES (optical emission spectroscopy) spectra of the DBD glow discharge and an UV lamp confirms that the discharge emits UV light with an energy corresponding to 3–4 eV. The authors assume that photocatalysis could be possible using plasma as a photoactivation source, as proposed by Park et al. [163]. Titanium dioxide is coated on glass beads and on three types of ␥Al2 O3 with different surface area, pore volume and pore diameter. High porous alumina dramatically enhances the benzene conversion and mineralization degree.

Table 9 Published papers on removal of other VOCs with plasma–catalysis. Target VOC

References

Acetaldehyde Acetone Acetylene Dichloromethane Formaldehyde Methane Methanol Propane Propene Styrene Tetrachloromethane Xylene

[110,267,268] [91,160,167] [106,165,166,241,269] [105,144,270,271] [220,272] [90,148,273] [249] [17,149,157] [149] [151,176,274,275] [150,198] [159,176,276–282]

Fig. 11. FT-IR spectra showing the plasma–catalytic destruction of benzene with Ag/␥-Al2 O3 , as a function of temperature in a two-stage configuration. Reprinted from Ref. [180], with permission from Elsevier.

46

A.M. Vandenbroucke et al. / Journal of Hazardous Materials 195 (2011) 30–54

Fig. 12. Effect of energy density on energy yield of different catalysts (RH: 20%; initial toluene concentration: 105 ppm; gas flow rate: 450 mL/min). Reprinted from Ref. [183], with permission from Elsevier. Fig. 13. Influence of SMF modification and energy density on the conversion of 100 ppm toluene.

In Ref. [179] the influence of humidity on benzene removal is investigated with a DBD packed with Raschig rings coated with nano TiO2 films. Humidity negatively affects decomposition of benzene for three reasons: deactivation of high-energy electrons, inhibition of ozone formation and suppression of the catalyst activity of TiO2 for benzene oxidation with ozone. Harling et al. [180] have examined the effect of temperature (293–886 K) and catalyst position (IPC/PPC). Fig. 11 shows IR spectra of the plasma–catalytic destruction of benzene as a function of temperature with Ag/␥-Al2 O3 in a two-stage configuration. When compared to thermal catalysis, NOx formation is detected and increasing amounts are produced at elevated temperatures. Additionally, higher levels of destruction are observed at lower temperatures for plasma–catalysis.

2.2.3.3. Toluene. Table 8 gives an extensive overview of the papers that have been published on the plasma–catalytic abatement of toluene. A concise discussion is given on selected papers. Other references can be found in Table 8. Song et al. [17] have applied a DBD packed with macro-porous ␥-Al2 O3 and investigated the effect of adsorption and elevated temperature. Higher operating temperatures (373 K) cause a reduction in adsorption capability. However, toluene removal is more favorable under these conditions in comparison with the use of non-adsorbing glass beads. The use of ␥-Al2 O3 beads proves to reduce some of the gas-phase byproducts, such as O3 and HNO3 , generated by the NTP process. Malik and Jiang [181] have also indicated that selecting the alumina packing with higher overall surface area can lower ozone generation without affecting the destruction efficiency of toluene. In a study by Li et al. [182], a DC streamer corona discharge is employed in combination with TiO2 pellets. Positioning the photocatalyst between the needle and mesh electrodes benefits the plasma discharge due to a higher streamer repetition rate. This configuration shows the best performance for decomposition (76%) and energy efficiency (7.2 g/kWh). This is attributed to the simultaneous decomposition of gas phase and adsorbed toluene and to possible TiO2 activation by plasma inducing catalytic reactions. In absence of the TiO2 layer, both the decomposition (44%) and efficiency (3.2 g/kWh) significantly drop. The authors claim that intermittent operation can improve the efficiency due to the regeneration of the catalyst surface through desorption during the discharge.

Reprinted from Ref. [134], with permission from Elsevier.

Guo et al. [183] have applied a DBD to study the effect of MnOx /Al2 O3 /nickel foam for IPC. Earlier results have confirmed that MnOx /Al2 O3 /nickel foam is the most effective for toluene removal among different catalysts tested [102]. Fig. 12 shows that the MnOx catalyst greatly improves the energy yield as compared to the plasma-alone system. A sampling method has been developed to detect OH radicals in the gas phase and on the catalyst surface [184]. The catalyst can enhance the toluene removal efficiency due to efficient reactions of OH radicals with toluene on the surface or the active sites and other active species on the catalyst. With the plasma–catalytic system toluene removal decreases with increased humidity. It is suggested that water molecules cover the catalyst surface, resulting in a lower reaction probability [185]. Indeed, Van Durme et al. [186] have concluded that water molecules adsorb on the catalyst surface to form mono- or multilayers that block active sites and create an extra diffusion layer for toluene to reach the catalyst surface. This hypothesis has also been confirmed by Huang et al. [122,123]. In a recent paper [187], Huang et al. have investigated the effect of water vapor on toluene removal efficiency, carbon balance, CO2 selectivity and outlet ozone concentration. A wire-plate DBD filled with MnOx /Al2 O3 /nickel foam or TiO2 /Al2 O3 /nickel foam is used to perform experiments. The results show that increased humidity lowers the formation of ozone through quenching of energetic electrons. Also, catalytic decomposition of ozone is depressed by the presence of water vapor due to competitive adsorption causing deactivation of the catalyst and suppression of catalytic ozonation. The carbon balance and CO2 selectivity reach maximum values when RH is in the range 25–75%. Van Durme et al. [107] have also studied the effect of humidity on PPC removal of toluene. As for IPC, PPC is less efficient when RH increases. With Pd/Al2 O3 as PPC removal efficiencies are >90% and 37% at dry air and air with 74% RH (298 K), respectively. The negative humidity effect is mostly attributed to changing Van der Waals interactions. In a recent study by Huang et al. [168], NTP has been combined with a photocatalyst located downstream. Experimental results indicate that catalytic ozonation plays a vital role in toluene decomposition. The dominant active species in the NTP-driven photocatalyst system are active oxygen species formed from ozone catalytic decomposition. The decomposition pathway of toluene

A.M. Vandenbroucke et al. / Journal of Hazardous Materials 195 (2011) 30–54

has been elucidated in subsequent work [188]. Detected byproducts for IPC removal of toluene with TiO2 /Al2 O3 /nickel foam include benzene, benzaldehyde, formic acid and small amounts of acetic acid and 2-methylamylene. Subrahmanyam et al. [134] modified a sintered metal fibre filter, which acts as inner electrode, with MnOx and CoOx . Fig. 13 shows the influence of this modification and energy density on the conversion of 100 ppm toluene. At an energy density of 235 J/L, nearly 100% conversion has been achieved with both MnOx and CoOx /SMF. Whereas SMF only shows 50% CO2 selectivity, MnOx /SMF reaches 80% even at 235 J/L. Interestingly, no polymeric carbon deposits are detected. All the catalytic electrodes maintain the same activity during almost 3 h of toluene decomposition. This proves that the electrodes maintain their stability during VOC destruction. Magureanu et al. [142] have tested MnPO4 , Mn-APO-5 and Mn-SAPO-11 as PPC catalysts in an oven for temperatures up to 673 K. Even at low temperature, a remarkable synergetic effect has been observed while the catalysts alone are not active at that temperature level. The authors expect a further increase in the plasma–catalytic synergy by placing the catalyst in the discharge region, where short-lived species produced in the plasma will most likely contribute to oxidation on the catalyst surface. 3. Critical process parameters Various process parameters determine the initial condition of the feeded gas stream. In the following section these parameters are discussed which are critical for an effective operation of both catalytic and non-catalytic NTP systems. For each parameter, the different influences on the removal performance of the configuration will be discussed and compared if possible. 3.1. Temperature In most cases, the NTP process removes VOCs more effectively as the process temperature increases. This is ascribed to an increased reaction rate of O and OH radicals with VOCs due to the endothermic behaviour of these reactions [53,127,189–195]. This is, however, only the case for VOCs that are primarily decomposed through radical reactions. When electron impact is thought to be the primary decomposition step (e.g. CCl4 ), no temperature dependence on the removal is observed because the electron density is not really influenced hereby [196,197]. However, in Ref. [198] CCl4 destruction is greatly improved at high temperature. This can be explained by the fact that the maximum energy density is also significantly higher than in [196], which might lead to higher decomposition. The improved removal rate and energy efficiency can also be explained by an increase in the reduced electric field (E/n) with increasing temperatures. The reduced electric field, being the ratio of the electric field (E) and the gas density (n), is an important factor that determines the electron energy in the plasma. Since the gas density decreases as the gas temperature increases at constant pressure, NTP systems tend to operate at a higher reduced electric field [192,199]. When the catalyst is located downstream, NTP produced ozone can be decomposed by reaction with molecular oxygen in the gas phase: O3 + O2 → O + O2 + O2

(48)

The rate constant of this reaction is accelerated at elevated temperatures (5 times higher at 573 K compared to 373 K). However, the lifetime of the produced oxygen atoms in the gas phase is too short to react with VOCs adsorbed on the catalyst surface. At the same time, reactions at the catalyst surface between adsorbed oxygen atoms and VOCs are also accelerated. The net-result of these two competing effects is the most likely explanation for

47

the different temperature dependencies found in literature: with increasing temperature, VOC decomposition efficiency can remain almost constant [142], can increase [149,153,180] or can decrease [127]. 3.2. Initial VOC concentration Generally, the VOC concentration of actual industrial exhaust streams strongly varies. Therefore the effect of VOC concentration on the removal process has been abundantly studied. When the initial concentration rises, each VOC molecule shares fewer electrons and reactive plasma species. Consequently, numerous research papers have pointed out that higher initial VOC concentrations are detrimental for the removal efficiency in catalytic and non-catalytic NTP systems. Some papers also indicate that the characteristic energy [200–204] (i.e. the energy density needed to decompose 63% of the initial VOC concentration) and the energy yield [49,199,205] are an increasing function of the initial VOC concentration. For some halogenated carbons, the initial concentration barely seems to affect the decomposition efficiency. This is the case for HFC-134a [206], CFC-12 [207], HCFC-22 [208] and TCE [62]. This may be partly attributed to secondary decomposition induced by fragment ions and radicals produced by primary destruction steps [62]. Another plausible explanation may be that for these compounds the primary destruction by reactive plasma species is the rate-determining step, leading to similar decomposition efficiencies regardless of the initial concentration [208]. 3.3. Humidity level The effect of humidity is of great interest for practical applications in industry since process gas consists of ambient air that usually contains water vapor at fluctuating concentrations. It appears that the effect of water vapor strongly depends on its concentration as well as on the type of the target VOC and the type of discharge. Water plays an important role in the plasma chemistry since it decomposes into OH and H radicals in a NTP environment as follows: H2 O + e− → OH• + H• + e− 3

+u )

→ N2

H2 O + O( D) →

2 OH•

H2 O + N2 (A 1

+ OH•

(49) + H•

(50) (51)

The oxidation power of OH is generally much stronger than those of other oxidants such as oxygen atoms and peroxyl radicals. The introduction of water vapor can induce changes in the electrical and physical properties of the discharge. The effect of water vapor has been mostly studied with (packed) DBD reactors. For this type of discharge, the presence of water vapor is known to reduce the total charge transferred in a microdischarge which ultimately decreases the volume of the reactive plasma zone [68]. The plasma characteristics of corona discharges are also affected by the presence of water vapor. At higher RH, lower currents are observed for a given voltage [83]. This is attributed to a higher probability of the plasma attachment processes resulting in a reduced OH production [209]. Water also has an adverse effect on VOC removal due to its electronegative characteristic which limits the electron density and quenches activated chemical species [82]. The effect of humidity has been tested for several VOCs. It seems that the addition of water negatively influences the properties of the discharge irrespective of the VOC chemical structure. However, the enhanced production of OH caused by higher water vapor content competes with the latter effect, depending on the VOC chemical structure [210]. The influence on the removal process is designated as an enhancement, a suppression or a neutral effect depending on the chemical structure of the target VOC. Table 10

48

A.M. Vandenbroucke et al. / Journal of Hazardous Materials 195 (2011) 30–54

Table 10 Influence of humidity on VOC removal with NTP.

Table 11 Optimal oxygen content for VOC removal with NTP.

Target VOC

Plasma type

Influence

References

Target VOC

Optimal O2 content (%)

References

Acetylene Benzene Benzene Benzene Butane Chloroform Dichloromethane Formaldehyde Methane Methanol Propane TCE TCE TCE Tetrachloromethane Tetrachloromethane Toluene Toluene Toluene Toluene Toluene 1,1,1-Trichloroethane p-Xylene

DBD Packed DBD DBD Packed DBD DBD Gliding arc Packed DBD DBD DBD Packed DBD DBD DBD Surface discharge SPCPa Packed DBD Gliding arc Corona Packed DBD Pulsed corona Gliding arc Surface discharge DBD DBD

Suppression Suppression Suppression Suppression Suppression Suppression Suppression Neutral Enhancement Neutral Neutral Suppression Suppression Neutral Suppression Suppression Enhancement Neutral Suppression Neutral Neutral Suppression Enhancement

[166] [66] [69] [283] [213] [284] [205] [194] [247] [205] [193] [213] [285] [286] [216] [284] [169] [205] [244] [199] [140] [287] [192]

Acetaldehyde Benzene Benzene Carbon tetrafluoride Dichloromethane 2-Heptanone HCFC-22 HFC-134a Methylbromide TCE Toluene Toluene Trichloromethane p-Xylene

3–5 0.2 3–5 1 1–3 2–3 0.5 0.5 2 2 3–5 2 0.5 5

[215] [73] [75] [288] [243] [200] [208] [206] [289] [290] [75] [102] [217] [192]

a

Surface discharge induced plasma chemical processing.

gives an overview of research results concerning the effect of humidity on the decomposition efficiency in various plasma reactors. Some studies have shown that an optimal water vapor content exists for achieving a maximum VOC removal efficiency. Interestingly, this optimum is around 20% RH for both TCE [211] and toluene [82,83]. Furthermore, addition of water counteracts the formation of ozone due to consumption of O(1 D) (reaction (51)) which is the most important origin of ozone formation [179]. It has also been shown that water vapor decreases the formation of CO and enhances the selectivity towards CO2 [166,212,213]. In case of a PPC system, catalytic ozonation will play a minor role due to the inhibition of ozone formation by humidity. Secondly, the catalyst surface can be covered with layers of H2 O preventing the adsorption of ozone and VOCs and consequently minimalizing direct catalyst/VOC intermolecular interactions [107,122,186,214]. In this context, the morphology and chemical composition of the catalyst are important factors that influence the interactions with H2 O. Therefore, it is desirable to choose a catalyst that is less susceptible to H2 O adsorption. Finally, increased humidity can poison catalytic active sites and lower the catalysts activity [122,214].

level because process gas depends on its industrial environment and in a lot of situations it consists of ambient air. Similar effects are observed for IPC systems. Additionally, direct reactions between oxygen radicals and VOC molecules adsorbed on the catalyst surface add to the positive effect of a moderate O2 addition to N2 [220]. However, in [75], several catalysts (TiO2 , ␥-Al2 O3 , zeolites) have been tested at varying oxygen content for the removal of toluene and benzene with a cycled system of removal and adsorption. For all catalysts tested, the removal efficiency increased with oxygen content ranging from 0% to 100%. Operation at higher oxygen content is also able to reduce the formation of N2 O and NO2 . As for the influence of the oxygen content on the performance of PPC configurations, no studies were found in literature. 3.5. Gas flow rate The gas flow rate applied in laboratory experiments generally ranges from 0.1 L/min to 10 L/min. The effect of decreasing the gas flow rate logically implies an increase in residence time of the VOC in the system. Hence, the collision probability for electron-impact reactions and for reactions between VOCs and plasma generated radicals and metastables is enhanced which increases the decomposition efficiency. When the NTP system is combined with a catalyst, the same argumentation can be made. In that case, the increased probability of surface reactions is beneficial for the removal process. In the interest of practical operation, some groups have studied multistage NTP reactors with the aim to increase the residence time without decreasing the gas flow rate [116,221–224].

3.4. Oxygen content 4. Future trends Similar to the presence of water vapor, the oxygen content in the gas stream affects the discharge performance and plays a very important role in the occurring chemical reactions. A small increase in oxygen concentration generally leads to an enhanced generation of reactive oxygen radicals, resulting in a higher removal efficiency. However, due to its electronegative character, higher oxygen concentrations tend to trigger electron attachment reactions. Consequently, this limits the electron density and changes the electron energy distribution functions [215,216]. Also, oxygen and oxygen radicals are able to consume reactive species such as excited nitrogen molecules and nitrogen atoms, which are otherwise used for destroying VOCs [217–219]. Collectively, the phenomena described above ensure the existence of an optimal oxygen content for VOC removal with NTP (Table 11). It appears that the optimal oxygen content ranges between 1% and 5%. For practical application in industrial waste gas treatment, it is, however, in most cases difficult to control the oxygen content to this

The extensive literature review regarding NTP and plasma–catalytic decomposition of VOCs demonstrates that there are still challenges that have to be adressed in future work. For plasma-alone systems, incomplete oxidation of VOCs leads to the formation of various intermediate and unwanted byproducts. Although for certain compounds decomposition mechanisms are proposed, there is still need to expand the knowledge on plasma-chemical kinetics. The derived information about e.g. the distribution of byproducts can be very useful to choose an appropriate catalyst or to enhance existing catalytic formulations in order to increase the efficiency of the hybrid system. In the case of plasma–catalytic systems, synergistic effects on the overall removal efficiency are often observed. The mechanisms that contribute to this synergy are thoroughly investigated and different elucidations are proposed in literature often showing discrepancies between them. Therefore, better understanding of

A.M. Vandenbroucke et al. / Journal of Hazardous Materials 195 (2011) 30–54

which mechanisms have the most important contributions and which chemical species play a dominant role in the decomposition of VOCs is still of great interest. From this point of view, development of well-designed instruments specialized in in situ measurements is crucial. The research of plasma material interactions for VOC treatment is recently looking at the opportunity to regenerate VOC saturated surfaces with the aid of NTP systems [225–227]. This process of alternate adsorption and desorption can convert flue gases with a large flow rate and low VOC concentration into that with a low flow rate and high concentration. Further developments could yield an economical VOC removal process for small- and medium-sized facilities that emit diluted VOC waste gases. Therefore, further progress on this subject is suspected in the nearby future.

References [1] P.S. Monks, C. Granier, S. Fuzzi, A. Stohl, M.L. Williams, H. Akimoto, M. Amann, A. Baklanov, U. Baltensperger, I. Bey, N. Blake, R.S. Blake, K. Carslawn, O.R. Cooper, F. Dentener, D. Fowler, E. Fragkou, G.J. Frost, S. Generoso, P. Ginoux, V. Grewet, A. Guenther, H.C. Hansson, S. Hennew, J. Hjorth, A. Hofzumahaus, H. Huntrieser, I.S.A. Isaksen, M.E. Jenkin, J. Kaiser, M. Kanakidou, Z. Klimont, M. Kulmala, P. Laj, M.G. Lawrence, J.D. Lee, C. Liousse, M. Maione, G. McFiggans, A. Metzger, A. Mieville, N. Moussiopoulos, J.J. Orlando, C.D. O’Dowd, P.I. Palmer, D.D. Parrish, A. Petzold, U. Platt, U. Pöschl, A.S.H. Prevot, C.E. Reeves, S. Reimann, Y. Rudich, K. Sellegri, R. Steinbrecher, D. Simpson, H. ten Brink, J. Theloke, G.R. van der Werf, R. Vautard, V. Vestreng, C. Vlachokostas, R. von Glasow, Atmospheric composition change—global and regional air quality, Atmospheric Environment 43 (2009) 5268–5350. [2] R. Atkinson, Atmospheric chemistry of VOCs and NOx , Atmospheric Environment 34 (2000) 2063–2101. [3] J.R. Odum, T.P.W. Jungkamp, R.J. Griffin, H.J.L. Forstner, R.C. Flagan, J.H. Seinfeld, Aromatics, reformulated gasoline, and atmospheric organic aerosol formation, Environmental Science & Technology 31 (1997) 1890–1897. [4] R.A Larson, E.J. Weber, Reaction Mechanisms in Environmental Organic Chemistry, Lewis Publishers, 2010. [5] J. Chorria, M. Jaroniec, M. Kloske, Ordered siliceous and carbonaceous nanomaterials in environmental protection, Ochrona Srodowiska 29 (2007) 3–12. [6] W.B. Li, H. Gong, Recent progress in the removal of volatile organic compounds by catalytic combustion, Acta Physico-Chimica Sinica 26 (2010) 885–894. [7] A. Kumar, J. Dewulf, H. Van Langenhove, Membrane-based biological waste gas treatment, Chemical Engineering Journal 136 (2008) 82–91. [8] S. Mudliar, B. Giri, K. Padoley, D. Satpute, R. Dixit, P. Bhatt, R. Pandey, A. Juwarkar, A. Vaidya, Bioreactors for treatment of VOCs and odours—a review, Journal of Environmental Management 91 (2010) 1039–1054. [9] K. Demeestere, J. Dewulf, H. Van Langenhove, Heterogeneous photocatalysis as an advanced oxidation process for the abatement of chlorinated, monocyclic aromatic and sulfurous volatile organic compounds in air: state of the art, Critical Reviews in Environmental Science and Technology 37 (2007) 489–538. [10] F.I Khan, A.K. Ghoshal, Removal of volatile organic compounds from polluted air, Journal of Loss Prevention in the Process Industries 13 (2000) 527–545. [11] T. Yamamoto, K. Ramaanathan, P.A. Lawless, D.S. Ensor, J.R. Newsome, N. Plaks, G.H. Ramsey, Control of volatile organic compounds by an AC energized ferroelectric pellet reactor and a pulsed corona reactor, IEEE Transactions on Industry Applications 28 (1992) 528–534. [12] C.M. Nunez, G.H. Ramsey, W.H. Ponder, J.H. Abbott, L.E. Hamel, P.H. Kariher, Corona destruction: an innovative control technology for VOCs and air toxics, Air and Waste 43 (1993) 242–247. [13] H.H. Kim, Nonthermal plasma processing for air-pollution control: a historical review, current issues, and future prospects, Plasma Processes and Polymers 1 (2004) 91–110. [14] J.Y Park, J.G. Jung, J.S. Kim, G.H. Rim, K.S. Kim, Effect of nonthermal plasma reactor for CF4 decomposition, IEEE Transactions on Plasma Science 31 (2003) 1349–1354. [15] T. Yamamoto, B.S. Rajanikanth, M. Okubo, T. Kuroki, M. Nishino, Performance evaluation of nonthermal plasma reactors for NO oxidation in diesel engine exhaust gas treatment, IEEE Transactions on Industry Applications 39 (2003) 1608–1613. [16] H.B. Ma, P. Chen, M.L. Zhang, X.Y. Lin, R. Ruan, Study of SO2 removal using non-thermal plasma induced by dielectric barrier discharge (DBD), Plasma Chemistry and Plasma Processing 22 (2002) 239–254. [17] Y.H. Song, S.J. Kim, K.I. Choi, T. Yamamoto, Effects of adsorption and temperature on a nonthermal plasma process for removing VOCs, Journal of Electrostatics 55 (2002) 189–201. [18] U. Roland, F. Holzer, E.D. Kopinke, Combination of non-thermal plasma and heterogeneous catalysis for oxidation of volatile organic compounds Part 2. Ozone decomposition and deactivation of gamma-Al2 O3 , Applied Catalysis B: Environmental 58 (2005) 217–226.

49

[19] C. Subrahmanyam, Catalytic non-thermal plasma reactor for total oxidation of volatile organic compounds, Indian Journal of Chemistry Section A: Inorganic Bio-Inorganic Physical Theoretical & Analytical Chemistry 48 (2009) 1062–1068. [20] T. Zhu, J. Li, W.J. Liang, Y.Q. Jin, Synergistic effect of catalyst for oxidation removal of toluene, Journal of Hazardous Materials 165 (2009) 1258–1260. [21] A. Mizuno, Industrial applications of atmospheric non-thermal plasma in environmental remediation, Plasma Physics and Controlled Fusion 49 (2007) A1–A15. [22] J.S. Chang, P.A. Lawless, T. Yamamoto, Corona discharge processes, IEEE Transactions on Plasma Science 19 (1991) 1152–1166. [23] A. Fridman, A. Chirokov, A. Gutsol, Non-thermal atmospheric pressure discharges, Journal of Physics D: Applied Physics 38 (2005) R1–R24. [24] U. Kogelschatz, Dielectric-barrier discharges: their history, discharge physics, and industrial applications, Plasma Chemistry and Plasma Processing 23 (2003) 1–46. [25] V. Nehra, A. Kumar, H.K. Dwivedi, Atmospheric non-thermal plasma sources, International Journal of Engineering 2 (2008) 53–68. [26] R. Hackam, H. Akiyama, Air pollution control by electrical discharges, IEEE Transactions on Dielectrics and Electrical Insulation 7 (2000) 654–683. [27] T. Yamamoto, S. Futamura, Nonthermal plasma processing for controlling volatile organic compounds, Combustion Science and Technology 133 (1998) 117–133. [28] B. Eliasson, M. Hirth, U. Kogelschatz, Ozone synthesis from oxygen in dielectric barrier discharges, Journal of Physics D: Applied Physics 20 (1987) 1421–1437. [29] G.J. Pietsch, Peculiarities of dielectric barrier discharges, Contributions to Plasma Physics 41 (2001) 620–628. [30] S. Masuda, S. Hosokawa, X.L. Tu, K. Sakakibara, S. Kitoh, S. Sakai, Destruction of gaseous-pollutants by surface-induced plasma chemical process (Spcs), IEEE Transactions on Industry Applications 29 (1993) 781–786. [31] S. Masuda, S. Hosokawa, X. Tu, Z. Wang, Novel plasma chemical technologies—Ppcp and Spcp for control of gaseous-pollutants and air toxics, Journal of Electrostatics 34 (1995) 415–438. [32] K. Urashima, J.S. Chang, T. Ito, Reduction of NOx from combustion flue gases by superimposed barrier discharge plasma reactors, IEEE Transactions on Industry Applications 33 (1997) 879–886. [33] R. McAdams, Prospects for non-thermal atmospheric plasmas for pollution abatement, Journal of Physics D: Applied Physics 34 (2001) 2810–2821. [34] J. Jarrige, P. Vervisch, Decomposition of three volatile organic compounds by nanosecond pulsed corona discharge: study of by-product formation and influence of high voltage pulse parameters, Journal of Applied Physics 99 (2006). [35] M. Nifuku, M. Horvath, J. Bodnar, G.Y. Zhang, T. Tanaka, E. Kiss, G. Woynarovich, H. Katoh, A study on the decomposition of volatile organic compounds by pulse corona, Journal of Electrostatics 40–1 (1997) 687–692. [36] E.H.W.M. Smulders, B.E.J.M. van Heesch, S.S.V.B. van Paasen, Pulsed power corona discharges for air pollution control, IEEE Transactions on Plasma Science 26 (1998) 1476–1484. [37] K.F. Shang, Y. Wu, J. Li, G.F. Li, D. Li, N.H. Wang, Reduction of NOx /SO2 by wire-plate type pulsed discharge reactor with pulsed corona radical shower, Plasma Chemistry and Plasma Processing 26 (2006) 443–454. [38] Y. Shi, J. Ruan, X. Wang, W. Ll, T. Tan, Decomposition of mixed malodorants in a wire-plate pulse corona reactor, Environmental Science & Technology 39 (2005) 6786–6791. [39] W.J. Liang, J. Li, Y.Q. Jin, Abatement of toluene from gas streams via ferroelectric packed bed dielectric barrier discharge plasma, Journal of Hazardous Materials 170 (2009) 633–638. [40] A. Ogata, N. Shintani, K. Mizuno, S. Kushiyama, T. Yamamoto, Decomposition of benzene using a nonthermal plasma reactor packed with ferroelectric pellets, IEEE Transactions on Industry Applications 35 (1999) 753–759. [41] U. Roland, F. Holzer, F.D. Kopinke, Improved oxidation of air pollutants in a non-thermal plasma, Catalysis Today 73 (2002) 315–323. [42] S. Okazaki, M. Kogoma, M. Uehara, Y. Kimura, Appearance of stable glowdischarge in air, argon, oxygen and nitrogen at atmospheric-pressure using a 50-Hz source, Journal of Physics D: Applied Physics 26 (1993) 889–892. [43] Y. Akishev, A. Deryugin, V. Karal’nik, I. Kochetov, A. Napartovich, N. Trushkin, Numerical simulation and experimental study of an atmospheric pressure direct-current glow discharge, Plasma Physics Reports 20 (1994) 511–524. [44] D. Trunec, A. Brablec, F. Stastny, Experimental study of atmospheric pressure glow discharge, Contributions to Plasma Physics 38 (1998) 435–445. [45] E.E. Kunhardt, Generation of large-volume, atmospheric-pressure, nonequilibrium plasmas, IEEE Transactions on Plasma Science 28 (2000) 189–200. [46] Y.S Akishev, M.E. Grushin, I.V. Kochetov, A.P. Napartovich, M.V. Pan’kin, N.I. Trushkin, Transition of a multipin negative corona in atmospheric air to a glow discharge, Plasma Physics Reports 26 (2000) 157–163. [47] Y. Akishev, O. Goossens, T. Callebaut, C. Leys, A. Napartovich, N. Trushkin, The influence of electrode geometry and gas flow on corona-to-glow and glowto-spark threshold currents in air, Journal of Physics D: Applied Physics 34 (2001) 2875–2882. [48] T. Callebaut, I. Kochetov, Y. Akishev, A. Napartovich, C. Leys, Numerical simulation and experimental study of the corona and glow regime of a negative pin-to-plate discharge in flowing ambient air, Plasma Sources Science & Technology 13 (2004) 245–250. [49] R. Vertriest, R. Morent, J. Dewulf, C. Leys, H. Van Langenhove, Multi-pinto-plate atmospheric glow discharge for the removal of volatile organic

50

[50]

[51]

[52]

[53]

[54]

[55]

[56]

[57]

[58] [59] [60]

[61]

[62]

[63]

[64]

[65]

[66]

[67]

[68]

[69]

[70]

[71]

[72]

[73] [74]

[75]

A.M. Vandenbroucke et al. / Journal of Hazardous Materials 195 (2011) 30–54 compounds in waste air, Plasma Sources Science & Technology 12 (2003) 412–416. D.S. Antao, D.A. Staack, A. Fridman, B. Farouk, Atmospheric pressure DC corona discharges: operating regimes and potential applications, Plasma Sources Science & Technology 18 (2009) 035016 (035011 pp.). D. Evans, L.A. Rosocha, G.K. Anderson, J.J. Coogan, M.J. Kushner, Plasma remediation of trichloroethylene in silent discharge plasmas, Journal of Applied Physics 74 (1993) 5378–5386. M.J. Kirkpatrick, W.C. Finney, B.R. Locke, Chlorinated organic compound removal by gas phase pulsed streamer corona electrical discharge with reticulated vitreous carbon electrodes, Plasmas and Polymers 8 (2003) 165–177. M.C. Hsiao, B.T. Merritt, B.M. Penetrante, G.E. Vogtlin, P.H. Wallman, Plasmaassisted decomposition of methanol and trichloroethylene in atmosphericpressure air streams by electrical-discharge processing, Journal of Applied Physics 78 (1995) 3451–3456. L. Prager, H. Langguth, S. Rummel, R. Mehnert, Electron-beam degradation of chlorinated hydrocarbons in air, Radiation Physics and Chemistry 46 (1995) 1137–1142. T. Hakoda, S. Hashimoto, Y. Fujiyama, A. Mizuno, Decomposition mechanism for electron beam irradiation of vaporized trichloroethylene-air mixtures, Journal of Physical Chemistry A 104 (2000) 59–66. T. Hakoda, G. Zhang, S. Hashimoto, Decomposition of chloroethenes in electron beam irradiation, Radiation Physics and Chemistry 54 (1999) 541–546. S.A. Vitale, K. Hadidi, D.R. Cohn, P. Falkos, The effect of a carbon-carbon double bond on electron beam-generated plasma decomposition of trichloroethylene and 1,1,1-trichloroethane, Plasma Chemistry and Plasma Processing 17 (1997) 59–78. E. Sanhueza, I.C. Hisatsune, J. Heicklen, Oxidation of haloethylenes, Chemical Reviews 76 (1976) 801–826. G. Huybrech, L. Meyers, Gas-phase chlorine-photosensitized oxidation of trichloroethylene, Transactions of the Faraday Society 62 (1966) 2191. L. Bertrand, J.A. Franklin, P. Goldfing, G. Huybrech, Point of attack of a chlorine atom on trichloroethylene, Journal of Physical Chemistry 72 (1968) 3926. B.M. Penetrante, M.C. Hsiao, J.N. Bardsley, B.T. Merritt, G.E. Vogtlin, A. Kuthi, C.P. Burkhart, J.R. Bayless, Identification of mechanisms for decomposition of air pollutants by non-thermal plasma processing, Plasma Sources Science & Technology 6 (1997) 251–259. S. Futamura, T. Yamamoto, Byproduct identification and mechanism determination in plasma chemical decomposition of trichloroethylene, IEEE Transactions on Industry Applications 33 (1997) 447–453. K. Urashima, J.S. Chang, Removal of volatile organic compounds from air streams and industrial flue gases by non-thermal plasma technology, IEEE Transactions on Dielectrics and Electrical Insulation 7 (2000) 602–614. J.S. Chang, K. Urashima, T. Ito, Mechanism of non-thermal plasma treatment of volatile organic compounds in dry air, in: D.W. Tedder, F.G. Pohland (Eds.), Emerging Technologies in Hazardous Waste Management, ACS Press, Atlanta, 1994, pp. 203–206. S.B. Han, T. Oda, Decomposition mechanism of trichloroethylene based on byproduct distribution in the hybrid barrier discharge plasma process, Plasma Sources Science & Technology 16 (2007) 413–421. A. Ogata, N. Shintani, K. Yamanouchi, K. Mizuno, S. Kushiyama, T. Yamamoto, Effect of water vapor on benzene decomposition using a nonthermaldischarge plasma reactor, Plasma Chemistry and Plasma Processing 20 (2000) 453–467. H.H. Kim, H. Kobara, A. Ogata, S. Futamura, Comparative assessment of different nonthermal plasma reactors on energy efficiency and aerosol formation from the decomposition of gas-phase benzene, IEEE Transactions on Industry Applications 41 (2005) 206–214. Z. Falkenstein, J.J. Coogan, Microdischarge behaviour in the silent discharge of nitrogen-oxygen and water-air mixtures, Journal of Physics D: Applied Physics 30 (1997) 817–825. M.P. Cal, M. Schluep, Destruction of benzene with non-thermal plasma in dielectric barrier discharge reactors, Environmental Progress 20 (2001) 151–156. B.Y. Lee, S.H. Park, S.C. Lee, M. Kang, S.J. Choung, Decomposition of benzene by using a discharge plasma—photocatalyst hybrid system, Catalysis Today 93–5 (2004) 769–776. Z.L. Ye, Y.N. Zhang, P. Li, L.Y. Yang, R.X. Zhang, H.Q. Hou, Feasibility of destruction of gaseous benzene with dielectric barrier discharge, Journal of Hazardous Materials 156 (2008) 356–364. C.Q. Jiang, A.A.H. Mohamed, R.H. Stark, J.H. Yuan, K.H. Schoenbach, Removal of volatile organic compounds in atmospheric pressure air by means of direct current glow discharges, IEEE Transactions on Plasma Science 33 (2005) 1416–1425. K. Satoh, T. Matsuzawa, H. Itoh, Decomposition of benzene in a corona discharge at atmospheric pressure, Thin Solid Films 516 (2008) 4423–4429. N. Goto, H. Kurimoto, S. Kudo, Y. Wanatabe, Decomposition of benzene by barrier discharge in N2 at low concentrations of O2 , IEEJ Transactions on Fundamentals and Materials 123 (2003) 900–906. H.H. Kim, A. Ogata, S. Futamura, Oxygen partial pressure-dependent behavior of various catalysts for the total oxidation of VOCs using cycled system of adsorption and oxygen plasma, Applied Catalysis B: Environmental 79 (2008) 356–367.

[76] M.F. Golde, Reactions of N2(A3-Sigma-U+), International Journal of Chemical Kinetics 20 (1988) 75–92. [77] G.R. Dey, A. Sharma, K.K. Pushpa, T.N. Das, Variable products in dielectricbarrier discharge assisted benzene oxidation, Journal of Hazardous Materials 178 (2010) 693–698. [78] H. Kohno, A.A. Berezin, J.S. Chang, M. Tamura, T. Yamamoto, A. Shibuya, S. Hondo, Destruction of volatile organic compounds used in a semiconductor industry by a capillary tube discharge reactor, IEEE Transactions on Industry Applications 34 (1998) 953–966. [79] W. Mista, R. Kacprzyk, Decomposition of toluene using non-thermal plasma reactor at room temperature, Catalysis Today 137 (2008) 345–349. [80] Z. Machala, E. Marode, M. Morvová, P. Lukáˇc, DC glow discharge in atmospheric air as a source for volatile organic compounds abatement, Plasma Processes and Polymers 2 (2005) 152–161. [81] S. Ognier, S. Cavadias, J. Amouroux, Aromatic VOC removal by formation of microparticles in pure nitrogen discharge, Plasma Processes and Polymers 4 (2007) 528–536. [82] Y.F. Guo, D.Q. Ye, K.F. Chen, Y.F. Tian, Humidity effect on toluene decomposition in a wire-plate dielectric barrier discharge reactor, Plasma Chemistry and Plasma Processing 26 (2006) 237–249. [83] J. Van Durme, J. Dewulf, W. Sysmans, C. Leys, H. Van Langenhove, Abatement and degradation pathways of toluene in indoor air by positive corona discharge, Chemosphere 68 (2007) 1821–1829. [84] M. Schiorlin, E. Marotta, M. Rea, C. Paradisi, Comparison of toluene removal in air at atmospheric conditions by different corona discharges, Environmental Science & Technology 43 (2009) 9386–9392. [85] F. Holzer, U. Roland, F.D. Kopinke, Combination of non-thermal plasma and heterogeneous catalysis for oxidation of volatile organic compounds Part 1. Accessibility of the intra-particle volume, Applied Catalysis B: Environmental 38 (2002) 163–181. [86] H.L. Chen, H.M. Lee, S.H. Chen, M.B. Chang, S.J. Yu, S.N. Li, Removal of volatile organic compounds by single-stage and two-stage plasma catalysis systems: a review of the performance enhancement mechanisms, current status, and suitable applications, Environmental Science & Technology 43 (2009) 2216–2227. [87] X. Chen, J. Rozak, J.C. Lin, S.L. Suib, J. Hayashi, H. Matsumoto, Oxidative decomposition of chlorinated hydrocarbons by glow discharge in PACT (plasma and catalyst integrated technologies) reactors, Applied Catalysis A: General 219 (2001) 25–31. [88] M.B. Chang, H.M. Lee, Abatement of perfluorocarbons with combined plasma catalysis in atmospheric-pressure environment, Catalysis Today 89 (2004) 109–115. [89] S.J. Yu, M.B. Chang, Oxidative conversion of PFC via plasma processing with dielectric barrier discharges, Plasma Chemistry and Plasma Processing 21 (2001) 311–327. [90] M.A. Malik, Y. Minamitani, K.H. Schoenbach, Comparison of catalytic activity of aluminum oxide and silica gel for decomposition of volatile organic compounds (VOCs) in a plasmacatalytic reactor, IEEE Transactions on Plasma Science 33 (2005) 50–56. [91] F. Holzer, F.D. Kopinke, U. Roland, Influence of ferroelectric materials and catalysts on the performance of non-thermal plasma (NTP) for the removal of air pollutants, Plasma Chemistry and Plasma Processing 25 (2005) 595–611. [92] H.L. Chen, H.M. Lee, S.H. Chen, M.B. Chang, Review of packed-bed plasma reactor for ozone generation and air pollution control, Industrial & Engineering Chemistry Research 47 (2008) 2122–2130. [93] J.O. Chae, V. Demidiouk, M. Yeulash, I.C. Choi, T.G. Jung, Experimental study for indoor air control by plasma–catalyst hybrid system, IEEE Transactions on Plasma Science 32 (2004) 493–497. [94] J.G. Wang, C.J. Liu, Y.P. Zhang, K.L. Yu, X.L. Zhu, F. He, Partial oxidation of methane to syngas over glow discharge plasma treated Ni-Fe/Al2 O3 catalyst, Catalysis Today 89 (2004) 183–191. [95] Z.H. Li, S.X. Tian, H.T. Wang, H.B. Tian, Plasma treatment of Ni catalyst via a corona discharge, Journal of Molecular Catalysis A: Chemical 211 (2004) 149–153. [96] X.L. Zhu, P.P. Huo, Y.P. Zhang, C.J. Liu, Characterization of argon glow discharge plasma reduced Pt/Al2 O3 catalyst, Industrial & Engineering Chemistry Research 45 (2006) 8604–8609. [97] C.J. Liu, J.J. Zou, K.L. Yu, D.G. Cheng, Y. Han, J. Zhan, C. Ratanatawanate, B.W.L. Jang, Plasma application for more environmentally friendly catalyst preparation, Pure and Applied Chemistry 78 (2006) 1227–1238. [98] C. Ratanatawanate, M. Macias, B.W.L. Jang, Promotion effect of the nonthermal RF plasma treatment on Ni/Al2 O3 for benzene hydrogenation, Industrial & Engineering Chemistry Research 44 (2005) 9868–9874. [99] Y.R. Zhu, Z.H. Li, Y.B. Zhou, J. Lv, H.T. Wang, Plasma treatment of Ni and Pt catalysts for partial oxidation of methane, Reaction Kinetics and Catalysis Letters 87 (2005) 33–41. [100] Y.F. Guo, D.Q. Ye, K.F. Chen, J.C. He, W.L. Chen, Toluene decomposition using a wire-plate dielectric barrier discharge reactor with manganese oxide catalyst in situ, Journal of Molecular Catalysis A: Chemical 245 (2006) 93–100. [101] Y.P. Zhang, P.S. Ma, X.L. Zhu, C.J. Liu, Y.T. Shen, A novel plasma-treated Pt/NaZSM-5 catalyst for NO reduction by methane, Catalysis Communications 5 (2004) 35–39. [102] Y.F. Guo, D.Q. Ye, K.F. Chen, J.C. He, Toluene removal by a DBD-type plasma combined with metal oxides catalysts supported by nickel foam, Catalysis Today 126 (2007) 328–337.

A.M. Vandenbroucke et al. / Journal of Hazardous Materials 195 (2011) 30–54 [103] A.E. Wallis, J.C. Whitehead, K. Zhang, Plasma-assisted catalysis for the destruction of CFC-12 in atmospheric pressure gas streams using TiO2 , Catalysis Letters 113 (2007) 29–33. [104] A.S. Pribytkov, G.N. Baeva, N.S. Telegina, A.L. Tarasov, A.Y. Stakheev, A.V. Tel’nov, V.N. Golubeva, Effect of electron irradiation on the catalytic properties of supported Pd catalysts, Kinetics and Catalysis 47 (2006) 765–769. [105] A.E. Wallis, J.C. Whitehead, K. Zhang, The removal of dichloromethane from atmospheric pressure nitrogen gas streams using plasma-assisted catalysis, Applied Catalysis B: Environmental 74 (2007) 111–116. [106] A. Rousseau, O. Guaitella, J. Ropcke, L.V. Gatilova, Y.A. Tolmachev, Combination of a pulsed microwave plasma with a catalyst for acetylene oxidation, Applied Physics Letters 85 (2004) 2199–2201. [107] J. Van Durme, J. Dewulf, K. Demeestere, C. Leys, H. Van Langenhove, Postplasma catalytic technology for the removal of toluene from indoor air: effect of humidity, Applied Catalysis B: Environmental 87 (2009) 78–83. [108] B. Lu, X. Zhang, X. Yu, T. Feng, S. Yao, Catalytic oxidation of benzene using DBD corona discharges, Journal of Hazardous Materials 137 (2006) 633–637. [109] H.H. Kim, A. Ogata, S. Futamura, Effect of different catalysts on the decomposition of VOCS using flow-type plasma-driven catalysis, IEEE Transactions on Plasma Science 34 (2006) 984–995. [110] T. Sano, N. Negishi, E. Sakai, S. Matsuzawa, Contributions of photocatalytic/catalytic activities of TiO2 and gamma-Al2 O3 in nonthermal plasma on oxidation of acetaldehyde and CO, Journal of Molecular Catalysis A: Chemical 245 (2006) 235–241. [111] H.B. Huang, D.Q. Ye, M.L. Fu, F. Da Feng, Contribution of UV light to the decomposition of toluene in dielectric barrier discharge plasma/photocatalysis system, Plasma Chemistry and Plasma Processing 27 (2007) 577–588. [112] H.H. Kim, S.M. Oh, A. Ogata, S. Futamura, Decomposition of gas-phase benzene using plasma-driven catalyst (PDC) reactor packed with Ag/TiO2 catalyst, Applied Catalysis B: Environmental 56 (2005) 213–220. [113] M. Kang, B.J. Kim, S.M. Cho, C.H. Chung, B.W. Kim, G.Y. Han, K.J. Yoon, Decomposition of toluene using an atmospheric pressure plasma/TiO2 catalytic system, Journal of Molecular Catalysis A: Chemical 180 (2002) 125–132. [114] M. Kang, Y.R. Ko, M.K. Jeon, S.C. Lee, S.J. Choung, J.Y. Park, S. Klim, S.H. Choi, Characterization of Bi/TiO2 nanometer sized particle synthesized by solvothermal method and CH3 CHO decomposition in a plasma-photocatalytic system, Journal of Photochemistry and Photobiology A: Chemistry 173 (2005) 128–136. [115] C. Subrahmanyam, M. Magureanu, D. Laub, A. Renken, L. Kiwi-Minsker, Nonthermal plasma abatement of trichloroethylene enhanced by photocatalysis, Journal of Physical Chemistry C 111 (2007) 4315–4318. [116] S. Chavadej, K. Saktrakool, P. Rangsunvigit, L.L. Lobban, T. Sreethawong, Oxidation of ethylene by a multistage corona discharge system in the absence and presence of Pt/TiO2 , Chemical Engineering Journal 132 (2007) 345–353. [117] A. Ogata, H.H. Kim, S.M. Oh, S. Futamura, Evidence for direct activation of solid surface by plasma discharge on CFC decomposition, Thin Solid Films 506 (2006) 373–377. [118] R. Morent, J. Dewulf, N. Steenhaut, C. Leys, H. Van Langenhove, Hybrid plasma–catalyst system for the removal of trichloroethylene in air, Journal of Advanced Oxidation Technologies 9 (2006) 53–58. [119] K. Takaki, K. Urashima, J.S. Chang, Ferro-electric pellet shape effect on C2F6 removal by a packed-bed-type nonthermal plasma reactor, IEEE Transactions on Plasma Science 32 (2004) 2175–2183. [120] K. Urashima, K.G. Kostov, J.S. Chang, Y. Okayasu, T. Iwaizumi, K. Yoshimura, T. Kato, Removal of C2F6 from a semiconductor process flue gas by a ferroelectric packed-bed barrier discharge reactor with an adsorber, IEEE Transactions on Industry Applications 37 (2001) 1456–1463. [121] C. Leys, D. Neirynck, R. Morent, E. Temmerman, DC-excited cold atmospheric pressure plasmas, Czechoslovak Journal of Physics 56 (2006) B896–B902. [122] H.B. Huang, D.Q. Ye, Combination of photocatalysis downstream the nonthermal plasma reactor for oxidation of gas-phase toluene, Journal of Hazardous Materials 171 (2009) 535–541. [123] H.B. Huang, D.Q. Ye, X.J. Guan, The simultaneous catalytic removal of VOCs and O-3 in a post-plasma, Catalysis Today 139 (2008) 43–48. [124] N. Blin-Simiand, P. Tardiveau, A. Risacher, F. Jorand, S. Pasquiers, Removal of 2-heptanone by dielectric barrier discharges—the effect of a catalyst support, Plasma Processes and Polymers 2 (2005) 256–262. [125] K. Krawczyk, B. Ulejczyk, H.K. Song, A. Lamenta, B. Paluch, K. SchmidtSzalowski, Plasma–catalytic reactor for decomposition of chlorinated hydrocarbons, Plasma Chemistry and Plasma Processing 29 (2009) 27–41. [126] V. Demidiouk, J.O. Chae, Decomposition of volatile organic compounds in plasma–catalytic system, IEEE Transactions on Plasma Science 33 (2005) 157–161. [127] K. Hayashi, H. Yasui, M. Tanaka, S. Futamura, S. Kurita, K. Aoyagi, Temperature dependence of toluene decomposition behavior in the discharge-catalyst hybrid reactor, IEEE Transactions on Industry Applications 45 (2009) 1553–1558. [128] K. Hensel, S. Katsura, A. Mizuno, DC microdischarges inside porous ceramics, IEEE Transactions on Plasma Science 33 (2005) 574–575. [129] S. Futamura, H. Einaga, H. Kabashima, L.Y. Hwan, Synergistic effect of silent discharge plasma and catalysts on benzene decomposition, Catalysis Today 89 (2004) 89–95. [130] M.J. Kirkpatrick, W.C. Finney, B.R. Locke, Plasma–catalyst interactions in the treatment of volatile organic compounds and NOx with pulsed corona

[131]

[132]

[133]

[134]

[135]

[136]

[137]

[138]

[139]

[140]

[141]

[142]

[143]

[144]

[145]

[146]

[147]

[148]

[149]

[150]

[151]

[152]

[153]

[154]

[155]

[156]

51

discharge and reticulated vitreous carbon Pt/Rh-coated electrodes, Catalysis Today 89 (2004) 117–126. C. Subrahmanyam, A. Renken, L. Kiwi-Minsker, Novel catalytic non-thermal plasma reactor for the abatement of VOCs, Chemical Engineering Journal 134 (2007) 78–83. C. Subrahmanyam, A. Renken, L. Kiwi-Minsker, Novel catalytic dielectric barrier discharge reactor for gas-phase abatement of isopropanol, Plasma Chemistry and Plasma Processing 27 (2007) 13–22. M. Magureanu, N.B. Mandache, V.I. Parvulescu, C. Subrahmanyam, A. Renken, L. Kiwi-Minsker, Improved performance of non-thermal plasma reactor during decomposition of trichloroethylene: optimization of the reactor geometry and introduction of catalytic electrode, Applied Catalysis B: Environmental 74 (2007) 270–277. C. Subrahmanyam, A. Renken, L. Kiwi-Minsker, Catalytic non-thermal plasma reactor for abatement of toluene, Chemical Engineering Journal 160 (2010) 677–682. X. Chen, J. Rozak, J.C. Lin, S.L. Suib, Y. Hayashi, H. Matsumoto, Oxidative decomposition of chlorinated hydrocarbons by glow discharge in PACT (plasma and catalyst integrated technologies) reactors, Applied Catalysis A: General 219 (2001) 25–31. A. Ogata, K. Yamanouchi, K. Mizuno, S. Kushiyama, T. Yamamoto, Oxidation of dilute benzene in an alumina hybrid plasma reactor at atmospheric pressure, Plasma Chemistry and Plasma Processing 19 (1999) 383–394. A. Ogata, K. Yamanouchi, K. Mizuno, S. Kushiyama, T. Yamamoto, Decomposition of benzene using alumina-hybrid and catalyst-hybrid plasma reactors, IEEE Transactions on Industry Applications 35 (1999) 1289–1295. T. Oda, K. Yamaji, Dilute trichloroethylene decomposition in air by using nonthermal plasma—catalyst effect, Journal of Advanced Oxidation Technologies 6 (2003) 93–99. A. Ogata, D. Ito, K. Mizuno, S. Kushiyama, T. Yamamoto, Removal of dilute benzene using a zeolte-hybrid plasma reactor, IEEE Transactions on Industry Applications 37 (2001) 959–964. S.M. Oh, H.H. Kim, A. Ogata, H. Einaga, S. Futamura, D.W. Park, Effect of zeolite in surface discharge plasma on the decomposition of toluene, Catalysis Letters 99 (2005) 101–104. S.M. Oh, H.H. Kim, H. Einaga, A. Ogata, S. Futamura, D.W. Park, Zeolitecombined plasma reactor for decomposition of toluene, Thin Solid Films 506 (2006) 418–422. M. Magureanu, N.B. Mandache, P. Eloy, E.M. Gaigneaux, V.I. Parvulescu, Plasma-assisted catalysis for volatile organic compounds abatement, Applied Catalysis B: Environmental 61 (2005) 12–20. T. Oda, T. Takahashi, S. Kohzuma, Decomposition of dilute trichloroethylene by using nonthermal plasma processing-frequency and catalyst effects, IEEE Transactions on Industry Applications 37 (2001) 965–970. A.E. Wallis, J.C. Whitehead, K. Zhang, The removal of dichloromethane from atmospheric pressure air streams using plasma-assisted catalysis, Applied Catalysis B: Environmental 72 (2007) 282–288. A. Ogata, H. Einaga, H. Kabashima, S. Futamura, S. Kushiyama, H.H. Kim, Effective combination of nonthermal plasma and catalysts for decomposition of benzene in air, Applied Catalysis B: Environmental 46 (2003) 87–95. H. Grossmannova, D. Neirynck, C. Leys, Atmospheric discharge combined with Cu-Mn/Al2 O3 catalyst unit for the removal of toluene, Czechoslovak Journal of Physics 56 (2006) B1156–B1161. A. Vandenbroucke, R. Morent, N. De Geyter, M.T. Nguyen Dinh, J.M. Giraudon, J.F. Lamonier, C. Leys, Plasma–catalytic decomposition of TCE, International Journal of Plasma Environmental Science and Technology 4 (2010) 135–138. P. Da Costa, R. Marques, S. Da Costa, Plasma catalytic oxidation of methane on alumina-supported noble metal catalysts, Applied Catalysis B: Environmental 84 (2008) 214–222. T. Blackbeard, V. Demidyuk, S.L. Hill, J.C. Whitehead, The effect of temperature on the plasma–catalytic destruction of propane and propene: a comparison with thermal catalysis, Plasma Chemistry and Plasma Processing 29 (2009) 411–419. T. Yamamoto, K. Mizuno, I. Tamori, A. Ogata, M. Nifuku, M. Michalska, G. Prieto, Catalysis-assisted plasma technology for carbon tetrachloride destruction, IEEE Transactions on Industry Applications 32 (1996) 100–105. C.L. Chang, H.L. Bai, S.J. Lu, Destruction of styrene in an air stream by packed dielectric barrier discharge reactors, Plasma Chemistry and Plasma Processing 25 (2005) 641–657. M.B. Chang, H.M. Lee, Abatement of perfluorocarbons with combined plasma–catalysis in atmospheric pressure environment, Catalysis Today 89 (2004) 109–115. V. Demidyuk, J.C. Whitehead, Influence of temperature on gas-phase toluene decomposition in plasma–catalytic system, Plasma Chemistry and Plasma Processing 27 (2007) 85–94. V. Demidiouk, S.I. Moon, J.O. Chae, D.Y. Lee, Application of a plasma–catalytic system for decomposition of volatile organic compounds, Journal of the Korean Physical Society 42 (2003) S966–S970. C.L. Chang, T.S. Lin, Elimination of carbon monoxide in the gas streams by dielectric barrier discharge systems with Mn catalyst, Plasma Chemistry and Plasma Processing 25 (2005) 387–401. H.L. Chen, H.M. Lee, L.C. Cheng, M.B. Chang, S.J. Yu, S.N. Li, Influence of nonthermal plasma reactor type on CF4 and SF6 abatements, IEEE Transactions on Plasma Science 36 (2008) 509–515.

52

A.M. Vandenbroucke et al. / Journal of Hazardous Materials 195 (2011) 30–54

[157] J. Jarrige, P. Vervisch, Plasma-enhanced catalysis of propane and isopropyl alcohol at ambient temperature on a MnO2 -based catalyst, Applied Catalysis B: Environmental 90 (2009) 74–82. [158] X. Fan, T.L. Zhu, M.Y. Wang, X.M. Li, Removal of low concentration BTX in air using a combined plasma catalysis system, Chemosphere 75 (2009) 1301–1306. [159] T. Hakoda, K. Matsumoto, A. Mizuno, K. Hirota, Role of metals loaded on a TiO2 surface in the oxidation of xylene in air using an electron beam irradiation/catalytic process, Applied Catalysis A: General 357 (2009) 244–249. [160] A.S. Besov, A.V. Vorontsov, Acceleration of acetone destruction process under synergistic action of photocatalytic oxidation and barrier discharge, Plasma Chemistry and Plasma Processing 27 (2007) 624–634. [161] T. Oda, K. Yamaji, T. Takahashi, Decomposition of dilute trichloroethylene by nonthermal plasma processing—gas flow rate, catalyst, and ozone effect, IEEE Transactions on Industry Applications 40 (2004) 430–436. [162] T. Oda, T. Takahahshi, K. Yamaji, Nonthermal plasma processing for dilute VOCs decomposition, IEEE Transactions on Industry Applications 38 (2002) 873–878. [163] D.W. Park, S.H. Yoon, G.J. Kim, H. Sekiguchi, The effect of catalyst on the decomposition of dilute benzene using dielectric barrier discharge, Journal of Industrial and Engineering Chemistry 8 (2002) 393–398. [164] R.B. Sun, Z.G. Xi, F.H. Chao, W. Zhang, H.S. Zhang, D.F. Yang, Decomposition of low-concentration gas-phase toluene using plasma-driven photocatalyst reactor, Atmospheric Environment 41 (2007) 6853–6859. [165] O. Guaitella, F. Thevenet, E. Puzenat, C. Guillard, A. Rousseau, C2 H2 oxidation by plasma/TiO2 combination: influence of the porosity, and photocatalytic mechanisms under plasma exposure, Applied Catalysis B: Environmental 80 (2008) 296–305. [166] F. Thevenet, O. Guaitella, E. Puzenat, C. Guillard, A. Rousseau, Influence of water vapour on plasma/photocatalytic oxidation efficiency of acetylene, Applied Catalysis B: Environmental 84 (2008) 813–820. [167] C.L. Chang, T.S. Lin, Decomposition of toluene and acetone in packed dielectric barrier discharge reactors, Plasma Chemistry and Plasma Processing 25 (2005) 227–243. [168] H.B. Huang, D.Q. Ye, D.Y.C. Leung, Removal of toluene using UV-irradiated and nonthermal plasma-driven photocatalyst system, Journal of Environmental Engineering: ASCE 136 (2010) 1231–1236. [169] M. Kogoma, Y. Miki, K. Tanaka, K. Takahashi, Highly efficient VOC decomposition using a complex system (OH Radical, Ozone-UV, and TiO2 ), Plasma Processes and Polymers 3 (2006) 727–733. [170] T. Oda, T. Takahashi, K. Yamaji, TCE decomposition by the non-thermal plasma process concerning ozone effect, IEEE Transaction on Industrial Applications 40 (2004) 1249–1256. [171] S.B. Han, T. Oda, R. Ono, Improvement of the energy efficiency in the decomposition of dilute trichloroethylene by the barrier discharge plasma process, IEEE Transactions on Industry Applications 41 (2005) 1343–1349. [172] M. Magureanu, N.B. Mandache, J.C. Hu, R. Richards, M. Florea, V.I. Parvulescu, Plasma-assisted catalysis total oxidation of trichloroethylene over gold nanoparticles embedded in SBA-15 catalysts, Applied Catalysis B: Environmental 76 (2007) 275–281. [173] A.M. Vandenbroucke, R. Morent, N. De Geyter, C. Leys, Decomposition of trichloroethylene with plasma–catalysis: a review, Journal of Advanced Oxidation Technologies 14 (2011) 165–173. [174] H.H. Kim, Y.H. Lee, A. Ogata, S. Futamura, Plasma-driven catalyst processing packed with photocatalyst for gas-phase benzene decomposition, Catalysis Communications 4 (2003) 347–351. [175] H.H. Kim, A. Ogata, S. Futamura, Y.H. Lee, Decomposition of gas-phase benzene using hybrid systems of a non-thermal plasma and catalysts, Journal of the Korean Physical Society 44 (2004) 1163–1167. [176] H.H. Kim, A. Ogata, S. Futamura, Atmospheric plasma-driven catalysis for the low temperature decomposition of dilute aromatic compounds, Journal of Physics D: Applied Physics 38 (2005) 1292–1300. [177] H.H. Kim, S.M. Oh, A. Ogata, S. Futamura, Decomposition of benzene using Ag/TiO2 packed plasma-driven catalyst reactor: influence of electrode configuration and Ag-loading amount, Catalysis Letters 96 (2004) 189–194. [178] H.Y. Fan, C. Shi, X.S. Li, D.Z. Zhao, Y. Xu, High-efficiency plasma catalytic removal of dilute benzene from air, Journal of Physics D: Applied Physics 42 (2009) 22510, 225105. [179] T. Zhu, J. Li, Y. Jin, Y. Liang, G. Ma, Decomposition of benzene by non-thermal plasma processing: photocatalyst and ozone effect, International Journal of Environmental Science and Technology 5 (2008) 375–384. [180] A.M. Harling, V. Demidyuk, S.J. Fischer, J.C. Whitehead, Plasma–catalysis destruction of aromatics for environmental clean-up: effect of temperature and configuration, Applied Catalysis B: Environmental 82 (2008) 180–189. [181] M.A. Malik, X.Z. Jiang, Catalyst assisted destruction of trichloroethylene and toluene in corona discharges, Journal of Environmental Sciences 12 (2000) 7–11. [182] D.A. Li, D. Yakushiji, S. Kanazawa, T. Ohkubo, Y. Nomoto, Decomposition of toluene by streamer corona discharge with catalyst, Journal of Electrostatics 55 (2002) 311–319. [183] Y.F. Guo, X. Liao, J.H. He, W. Ou, D.Q. Ye, Effect of manganese oxide catalyst on the dielectric barrier discharge decomposition of toluene, Catalysis Today 153 (2010) 176–183. [184] Y.F. Guo, X.B. Liao, D.Q. Ye, Journal of Environmental Science: China 20 (2008) 1429.

[185] X.B. Liao, Y.F. Guo, J.H. He, W.J. Ou, D.Q. Ye, Hydroxyl radicals formation in dielectric barrier discharge during decomposition of toluene, Plasma Chemistry and Plasma Processing 30 (2010) 841–853. [186] J. Van Durme, J. Dewulf, W. Sysmans, C. Leys, H. Van Langenhove, Efficient toluene abatement in indoor air by a plasma catalytic hybrid system, Applied Catalysis B: Environmental 74 (2007) 161–169. [187] H.B. Huang, D.Q. Ye, D.Y.C. Leung, Plasma-driven catalysis process for toluene abatement: effect of water vapor, IEEE Transactions on Plasma Science 39 (2011) 576–580. [188] H. Huang, D.Q. Ye, D.Y.C. Leung, F. Feng, X.J. Guan, Byproducts and pathways of toluene destruction via plasma–catalysis, Journal of Molecular Catalysis A: Chemical 336 (2011) 87–93. [189] B.M. Penetrante, M.C. Hsiao, J.N. Bardsley, B.T. Merritt, G.E. Vogtlin, A. Kuthi, C.P. Burkhart, J.R. Bayless, Decomposition of methylene chloride by electron beam and pulsed corona processing, Physics Letters A 235 (1997) 76–82. [190] V. Demidyuk, S.L. Hill, J.C. Whitehead, Enhancement of the destruction of propane in a low-temperature plasma by the addition of unsaturated hydrocarbons: experiment and modeling, Journal of Physical Chemistry A 112 (2008) 7862–7867. [191] N. Blin-Simiand, S. Pasquiers, F. Jorand, C. Postel, J.R. Vacher, Removal of formaldehyde in nitrogen and in dry air by a DBD: importance of temperature and role of nitrogen metastable states, Journal of Physics D: Applied Physics 42 (2009). [192] H.M. Lee, M.B. Chang, Abatement of gas-phase p-xylene via dielectric barrier discharges, Plasma Chemistry and Plasma Processing 23 (2003) 541–558. [193] O. Aubry, J.M. Cormier, Improvement of the diluted propane efficiency treatment using a non-thermal plasma, Plasma Chemistry and Plasma Processing 29 (2009) 13–25. [194] D.G. Storch, M.J. Kushner, Destruction mechanisms for formaldehyde in atmospheric-pressure low-temperature plasmas, Journal of Applied Physics 73 (1993) 51–55. [195] J. Li, S.P. Bai, X.C. Shi, S.L. Han, X.M. Zhu, W.C. Chen, Y.K. Pu, Effects of temperature on benzene oxidation in dielectric barrier discharges, Plasma Chemistry and Plasma Processing 28 (2008) 39–48. [196] B.M. Penetrante, M.C. Hsiao, J.N. Bardsley, B.T. Merritt, G.E. Vogtlin, P.H. Wallman, Electron beam and pulsed corona processing of volatile organic compounds in gas streams, Pure and Applied Chemistry 68 (1996) 1083–1087. [197] A.M. Harling, H.H. Kim, S. Futamura, J.C. Whitehead, Temperature dependence of plasma–catalysis using a nonthermal, atmospheric pressure packed bed; the destruction of benzene and toluene, Journal of Physical Chemistry C 111 (2007) 5090–5095. [198] L.W Huang, K. Nakajyo, T. Hari, S. Ozawa, H. Matsuda, Decomposition of carbon tetrachloride by a pulsed corona reactor incorporated with in situ absorption, Industrial & Engineering Chemistry Research 40 (2001) 5481–5486. [199] C.M. Du, J.H. Yan, B. Cheron, Decomposition of toluene in a gliding arc discharge plasma reactor, Plasma Sources Science & Technology 16 (2007) 791–797. [200] A.S. Chiper, N. Blin-Simiand, M. Heninger, H. Mestdagh, P. Boissel, F. Jorand, J. Lemaire, J. Leprovost, S. Pasquiers, G. Popa, C. Postel, Detailed characterization of 2-heptanone conversion by dielectric barrier discharge in N2 and N2 /O2 mixtures, Journal of Physical Chemistry A 114 (2010) 397–407. [201] N. Blin-Simiand, F. Jorand, L. Magne, S. Pasquiers, C. Postel, J.R. Vacher, Plasma reactivity and plasma-surface interactions during treatment of toluene by a dielectric barrier discharge, Plasma Chemistry and Plasma Processing 28 (2008) 429–466. [202] S. Delagrange, L. Pinard, J.M. Tatibouet, Combination of a non-thermal plasma and a catalyst for toluene removal from air: manganese based oxide catalysts, Applied Catalysis B: Environmental 68 (2006) 92–98. [203] J.H. Byeon, J.H. Park, Y.S. Jo, K.Y. Yoon, J. Hwang, Removal of gaseous toluene and submicron aerosol particles using a dielectric barrier discharge reactor, Journal of Hazardous Materials 175 (2010) 417–422. [204] L.A. Rosocha, R.A. Korzekwa, Advanced oxidation and reduction processes in the gas phase using non-thermal plasmas, Journal of Advanced Oxidation Technologies 4 (1999) 247–264. [205] S. Futamura, M. Sugasawa, Additive effect on energy efficiency and byproduct distribution in VOC decomposition with nonthermal plasma, IEEE Transactions on Industry Applications 44 (2008) 40–45. [206] Y.S. Mok, V. Demidyuk, J.C. Whitehead, Decomposition of hydrofluorocarbons in a dielectric-packed plasma reactor, Journal of Physical Chemistry A 112 (2008) 6586–6591. [207] H.C. Kang, Decomposition of chlorofluorocarbon by non-thermal plasma, Journal of Industrial and Engineering Chemistry 8 (2002) 488–492. [208] J.H. Oh, Y.S. Mok, S.B. Lee, Destruction of HCFC-22 and distribution of byproducts in a nonthermal plasma reactor packed with dielectric pellets, Journal of Korean Physical Society 54 (2009) 1539–1546. [209] L. Fouad, S. Elhazek, Effect of humidity on positive corona discharge in a 3electrode system, Journal of Electrostatics 35 (1995) 21–30. [210] M. Sugasawa, T. Terasawa, S. Futamura, Additive effect of water on the decomposition of VOCs in nonthermal plasma, IEEE Transactions on Industry Applications 46 (2010) 1692–1698. [211] Y. Nakagawa, H. Fujisawa, R. Ono, T. Oda, Dilute trichloroethylene decomposition by using high pressure non-thermal plasma-humidity effects, in: 2010 IEEE Industry Applications Society Annual Meeting, 2010.

A.M. Vandenbroucke et al. / Journal of Hazardous Materials 195 (2011) 30–54 [212] A. Ogata, D. Ito, K. Mizuno, S. Kushiyama, A. Gal, T. Yamamoto, Effect of coexisting components on aromatic decomposition in a packed-bed plasma reactor, Applied Catalysis A: General 236 (2002) 9–15. [213] S. Futamura, A.H. Zhang, T. Yamamoto, The dependence of nonthermal plasma behavior of VOCs on their chemical structures, Journal of Electrostatics 42 (1997) 51–62. [214] H. Einaga, T. Ibusuki, S. Futamura, Performance evaluation of a hybrid system comprising silent discharge plasma and manganese oxide catalysts for benzene decomposition, IEEE Transactions on Industry Applications 37 (2001) 1476–1482. [215] H.M. Lee, M.B. Chang, Gas-phase removal of acetaldehyde via packed-bed dielectric barrier discharge reactor, Plasma Chemistry and Plasma Processing 21 (2001) 329–343. [216] R.G. Tonkyn, S.E. Barlow, T.M. Orlando, Destruction of carbon tetrachloride in a dielectric barrier/packed-bed corona reactor, Journal of Applied Physics 80 (1996) 4877–4886. [217] Y.S. Mok, S.B. Lee, J.H. Oh, K.S. Ra, B.H. Sung, Abatement of trichloromethane by using nonthermal plasma reactors, Plasma Chemistry and Plasma Processing 28 (2008) 663–676. [218] A. Ogata, K. Mizuno, S. Kushiyama, T. Yamamoto, Methane decomposition in a barium titanate packed-bed nonthermal plasma reactor, Plasma Chemistry and Plasma Processing 18 (1998) 363–373. [219] K.J. Pringle, J.C. Whitehead, J.J. Wilman, J.H. Wu, The chemistry of methane remediation by a non-thermal atmospheric pressure plasma, Plasma Chemistry and Plasma Processing 24 (2004) 421–434. [220] H.X. Ding, A.M. Zhu, F.G. Lu, Y. Xu, J. Zhang, X.F. Yang, Low-temperature plasma–catalytic oxidation of formaldehyde in atmospheric pressure gas streams, Journal of Physics D: Applied Physics 39 (2006) 3603–3608. [221] A.M. Harling, D.J. Glover, J.C. Whitehead, K. Zhang, Industrial scale destruction of environmental pollutants using a novel plasma reactor, Industrial & Engineering Chemistry Research 47 (2008) 5856–5860. [222] S. Chavadej, W. Kiatubolpaiboon, P. Rangsunvigit, T. Sreethawong, A combined multistage corona discharge and catalytic system for gaseous benzene removal, Journal of Molecular Catalysis A: Chemical 263 (2007) 128–136. [223] Y. Shi, J.J. Ruan, X. Wang, W. Li, T.E. Tan, Evaluation of multiple corona reactor modes and the application in odor removal, Plasma Chemistry and Plasma Processing 26 (2006) 187–196. [224] A.M. Harling, D.J. Glover, J.C. Whitehead, K. Zhang, Novel method for enhancing the destruction of environmental pollutants by the combination of multiple plasma discharges, Environmental Science & Technology 42 (2008) 4546–4550. [225] T. Kuroki, T. Fujioka, R. Kawabata, M. Okubo, T. Yamamoto, Regeneration of honeycomb zeolite by nonthermal plasma desorption of toluene, IEEE Transactions on Industry Applications 45 (2009) 10–15. [226] H.H. Kim, S. Tsubota, M. Date, A. Ogata, S. Futamura, Catalyst regeneration and activity enhancement of Au/TiO2 by atmospheric pressure nonthermal plasma, Applied Catalysis A: General 329 (2007) 93–98. [227] T. Kuroki, T. Fujioka, M. Okubo, T. Yamamoto, Toluene concentration using honeycomb nonthermal plasma desorption, Thin Solid Films 515 (2007) 4272–4277. [228] J. Van Durme, J. Dewulf, C. Leys, H. Van Langenhove, Combining non-thermal plasma with heterogeneous catalysis in waste gas treatment: a review, Applied Catalysis B: Environmental 78 (2008) 324–333. [229] M. Magureanu, N.B. Mandache, V.I. Parvulescu, Chlorinated organic compounds decomposition in a dielectric barrier discharge, Plasma Chemistry and Plasma Processing 27 (2007) 679–690. [230] Z.L. Ye, Y. Shen, R.X. Zhang, H.Q. Hou, Destruction of benzene in an air stream by the outer combined plasma photolysis method, Journal of Physics D: Applied Physics 41 (2008). [231] A. Ogata, K. Miyamae, K. Mizuno, S. Kushiyama, M. Tezuka, Decomposition of benzene in air in a plasma reactor: effect of reactor type and operating conditions, Plasma Chemistry and Plasma Processing 22 (2002) 537–552. [232] C.H. Lin, H. Bai, Energy effectiveness of nonthermal plasma reactors for toluene vapor destruction, Journal of Environmental Engineering 127 (2001) 648–654. [233] A.D. Koutsospyros, S.M. Yin, C. Christodoulatos, K. Becker, Plasmochemical degradation of volatile organic compounds (VOC) in a capillary discharge plasma reactor, IEEE Transactions on Plasma Science 33 (2005) 42–49. [234] K. Faungnawakij, N. Sano, D. Yamamoto, T. Kanki, T. Charinpanitkul, W. Tanthapanichakoon, Removal of acetaldehyde in air using a wetted-wall corona discharge reactor, Chemical Engineering Journal 103 (2004) 115–122. [235] T. Oda, A. Kumada, K. Tanaka, T. Takahashi, S. Masuda, Low-temperature atmospheric-pressure discharge plasma processing for volatile organiccompounds, Journal of Electrostatics 35 (1995) 93–101. [236] T. Oda, R. Yamashita, I. Haga, T. Takahashi, S. Masuda, Decomposition of gaseous organic contaminants by surface discharge induced plasma chemical processing SPCP, IEEE Transactions on Industry Applications 32 (1996) 118–124. [237] R. Rudolph, K.P. Francke, H. Miessner, Concentration dependence of VOC decomposition by dielectric barrier discharges, Plasma Chemistry and Plasma Processing 22 (2002) 401–412. [238] H.C. Wang, D. Li, Y. Wu, J. Li, G.F. Li, Removal of four kinds of volatile organic compounds mixture in air using silent discharge reactor driven by bipolar pulsed power, Journal of Electrostatics 67 (2009) 547–553.

53

[239] M.N. Lyulyukin, A.S. Besov, A.V. Vorontsov, The influence of corona electrodes thickness on the efficiency of plasmachemical oxidation of acetone, Plasma Chemistry and Plasma Processing 31 (2011) 23–39. [240] M.G. Sobacchi, A.V. Saveliev, A.A. Fridman, A.F. Gutsol, L.A. Kennedy, Experimental assessment of pulsed corona discharge for treatment of VOC emissions, Plasma Chemistry and Plasma Processing 23 (2003) 347–370. [241] J. Hoard, T.J. Wallington, R.L. Bretz, A. Malkin, R. Dorai, M.J. Kushner, Importance of O(P-3) atoms and OH radicals in hydrocarbon oxidation during the nonthermal plasma treatment of diesel exhaust inferred using relative-rate methods, International Journal of Chemical Kinetics 35 (2003) 231–238. [242] A. Rousseau, O. Guaitella, L. Gatilova, M. Hannemann, J. Ropcke, Measurement of the C2 H2 destruction kinetics by infrared laser absorption spectroscopy in a pulsed low pressure dc discharge, Journal of Physics D: Applied Physics 40 (2007) 2018–2025. [243] C. Fitzsimons, F. Ismail, J.C. Whitehead, J.J. Wilman, The chemistry of dichloromethane destruction in atmospheric-pressure gas streams by a dielectric packed-bed plasma reactor, Journal of Physical Chemistry A 104 (2000) 6032–6038. [244] R.A. Korzekwa, M.G. Grothaus, R.K. Hutcherson, R.A. Roush, R. Brown, Destruction of hazardous air pollutants using a fast rise time pulsed corona reactor, Review of Scientific Instruments 69 (1998) 1886–1892. [245] E. Marotta, G. Scorrano, C. Paradisi, Ionic reactions of chlorinated volatile organic compounds in air plasma at atmospheric pressure, Plasma Processes and Polymers 2 (2005) 209–217. [246] M. Sugasawa, G. Annadurai, S. Futamura, Reaction behavior of toluenedichloromethane mixture in nonthermal plasma, IEEE Transactions on Industry Applications 45 (2009) 1499–1505. [247] A. Mfopara, M.J. Kirkpatrick, E. Odic, Dilute methane treatment by atmospheric pressure dielectric barrier discharge: effects of water vapor, Plasma Chemistry and Plasma Processing 29 (2009) 91–102. [248] M. Derakhshesh, J. Abedi, H. Hassanzadeh, Mechanism of methanol decomposition by non-thermal plasma, Journal of Electrostatics 68 (2010) 424–428. [249] R. Taranto, D. Frochot, P. Pichat, Combining cold plasma and TiO2 photocatalysis to purify gaseous effluents: a preliminary study using methanol-contaminated air, Industrial & Engineering Chemistry Research 46 (2007) 7611–7614. [250] T. Sato, M. Kambe, H. Nishiyama, Analysis of a methanol decomposition process by a nonthermal plasma flow, JSME International Journal Series B: Fluids and Thermal Engineering 48 (2005) 432–439. [251] S.L. Hill, J.C. Whitehead, K. Zhang, Plasma processing of propane at hyperatmospheric pressure: experiment and modelling, Plasma Processes and Polymers 4 (2007) 710–718. [252] S.L. Hill, H.H. Kim, S. Futamura, J.C. Whitehead, The destruction of atmospheric pressure propane and propene using a surface discharge plasma reactor, Journal of Physical Chemistry A 112 (2008) 3953–3958. [253] N. Moreau, S. Pasquiers, N. Blin-Simiand, L. Magne, F. Jorand, C. Postel, J.R. Vacher, Propane dissociation in a non-thermal high-pressure nitrogen plasma, Journal of Physics D: Applied Physics 43 (2010), -. [254] Y.S. Mok, C.M. Nam, M.H. Cho, I.S. Nam, Decomposition of volatile organic compounds and nitric oxide by nonthermal plasma discharge processes, IEEE Transactions on Plasma Science 30 (2002) 408–416. [255] I. Orlandini, U. Riedel, Oxidation of propene and the formation of methyl nitrate in non-thermal plasma discharges, Catalysis Today 89 (2004) 83–88. [256] G.K. Anderson, H. Snyder, J. Coogan, Oxidation of styrene in a silent discharge plasma, Plasma Chemistry and Plasma Processing 19 (1999) 131–151. [257] A. Ostapczuk, A.G. Chmielewski, V. Honkonen, J. Ruuskanen, J. Tarhanen, B. Svarfvar, Preliminary test in decomposition of styrene by electron beam treatment, Radiation Physics and Chemistry 56 (1999) 369–371. [258] B.M. Penetrante, M.C. Hsiao, J.N. Bardsley, B.T. Merritt, G.E. Vogtlin, P.H. Wallman, A. Kuthi, C.P. Burkhart, J.R. Bayless, Electron-beam and pulsed corona processing of carbon-tetrachloride in atmospheric-pressure gas streams, Physics Letters A 209 (1995) 69–77. [259] A. Lamenta, S. Jodzis, K. Krawczyk, K. Schmidt-Szalowski, Carbon tetrachloride decomposition in spark discharge plasma, Polish Journal of Chemistry 83 (2009) 169–174. [260] S.J. Rubio, M.C. Quintero, A. Rodero, Application of microwave air plasma in the destruction of trichloroethylene and carbon tetrachloride at atmospheric pressure, Journal of Hazardous Materials 186 (2011) 820–826. [261] H. Kohno, M. Tamura, S. Honda, A. Shibuya, T. Yamamoto, A.A. Berezin, J.S. Chang, Generation of aerosol particles in the process of xylene and TCE decomposition from air streams by a ferroelectric packed-bed barrier discharge reactor, Journal of Aerosol Science 28 (1997) S413–S414. [262] T. Zhu, J. Li, Y.Q. Jin, Y.H. Liang, G.D. Ma, Gaseous phase benzene decomposition by non-thermal plasma coupled with nano titania catalyst, International Journal of Environmental Science and Technology 6 (2009) 141–148. [263] J.Y. Ban, Y.H. Son, M. Kang, S.J. Choung, Highly concentrated toluene decomposition on the dielectric barrier discharge (DBD) plasma-photocatalytic hybrid system with Mn-Ti-incorporated mesoporous silicate photocatalyst (Mn-TiMPS), Applied Surface Science 253 (2006) 535–542. [264] M. Magureanu, N.B. Mandache, E. Gaigneaux, C. Paun, V.I. Parvulescu, Toluene oxidation in a plasma–catalytic system, Journal of Applied Physics 99 (2006). [265] A.M. Harling, D.J. Glover, J.C. Whitehead, K. Zhang, The role of ozone in the plasma–catalytic destruction of environmental pollutants, Applied Catalysis B: Environmental 90 (2009) 157–161.

54

A.M. Vandenbroucke et al. / Journal of Hazardous Materials 195 (2011) 30–54

[266] V. Demidiouk, S.I. Moon, J.O. Chae, Toluene and butyl acetate removal from air by plasma–catalytic system, Catalysis Communications 4 (2003) 51–56. [267] A. Mizuno, Y. Kisanuki, M. Noguchi, S. Katsura, S.H. Lee, U.K. Hong, S.Y. Shin, J.H. Kang, Indoor air cleaning using a pulsed discharge plasma, IEEE Transactions on Industry Applications 35 (1999) 1284–1288. [268] T. Ohshima, T. Kondo, N. Kitajima, M. Sato, Adsorption and plasma decomposition of gaseous acetaldehyde on fibrous activated carbon, IEEE Transactions on Industry Applications 46 (2010) 23–28. [269] A. Rousseau, O. Guaitella, L. Gatilova, F. Thevenet, C. Guillard, J. Ropcke, G.D. Stancu, Photocatalyst activation in a pulsed low pressure discharge, Applied Physics Letters 87 (2005). [270] A.M. Harling, A.E. Wallis, J.C. Whitehead, The effect of temperature on the removal of DCM using non-thermal, atmospheric-pressure plasma-assisted catalysis, Plasma Processes and Polymers 4 (2007) 463–470. [271] A. Ogata, K. Saito, H.H. Kim, M. Sugasawa, H. Aritani, H. Einaga, Performance of an ozone decomposition catalyst in hybrid plasma reactors for volatile organic compound removal, Plasma Chemistry and Plasma Processing 30 (2010) 33–42. [272] H.X. Ding, A.M. Zhu, X.F. Yang, C.H. Li, Y. Xu, Removal of formaldehyde from gas streams via packed-bed dielectric barrier discharge plasmas, Journal of Physics D: Applied Physics 38 (2005) 4160–4167. [273] R. Marques, S. Da Costa, P. Da Costa, Plasma-assisted catalytic oxidation of methane—on the influence of plasma energy deposition and feed composition, Applied Catalysis B: Environmental 82 (2008) 50–57. [274] J. Kim, B. Han, Y. Kim, J.H. Lee, C.R. Park, J.C. Kim, J.C. Kim, K.J. Kim, Removal of VOCs by hybrid electron beam reactor with catalyst bed, Radiation Physics and Chemistry 71 (2004) 429–432. [275] K.J. Kim, J.C. Kim, J.Y. Kim, Y. Sunwoo, Development of hybrid technology using E-beam and catalyst for aromatic VOCs control, Radiation Physics and Chemistry 73 (2005) 85–90. [276] T. Hakoda, A. Shimada, A. Kimura, M. Taguchi, Y. Sugo, K. Arak, E.B. Dally, K. Hirota, An electron-beam irradiation/catalytic oxidation system for purification of aromatic hydrocarbons/air mixture under practical gasflow condition, Industrial & Engineering Chemistry Research 49 (2010) 5517–5522. [277] T. Hakoda, K. Matsumoto, A. Shimada, T. Narita, T. Kojima, K. Hirota, Application of ozone decomposition catalysts to electron-beam irradiated xylene/air mixtures for enhancing carbon dioxide production, Radiation Physics and Chemistry 77 (2008) 585–590.

[278] T. Hakoda, K. Matsumoto, A. Mizuno, T. Kojima, K. Hirota, Catalytic oxidation of xylene in air using TiO2 under electron beam irradiation, Plasma Chemistry and Plasma Processing 28 (2008) 25–37. [279] T. Hakoda, K. Matsumoto, A. Mizuno, T. Narita, T. Kojima, K. Hirota, Oxidation process of xylene in air using Ag/TiO2 under electron beam irradiation, IEEE Transactions on Industry Applications 44 (2008) 1950–1956. [280] T. Hakoda, K. Matsumoto, A. Mizuno, K. Hirota, Oxidation of xylene and its irradiation byproducts using an electron-beam irradiating gamma-Al2 O3 bed, Journal of Physics D: Applied Physics 41 (2008). [281] K.P. Francke, H. Miessner, R. Rudolph, Cleaning of air streams from organic pollutants by plasma–catalytic oxidation, Plasma Chemistry and Plasma Processing 20 (2000) 393–403. [282] T. Kuroki, K. Hirai, R. Kawabata, M. Okubo, T. Yamamoto, Decomposition of adsorbed xylene on adsorbents using nonthermal plasma with gas circulation, IEEE Transactions on Industry Applications 46 (2010) 672–679. [283] S. Futamura, A.H. Zhang, H. Einaga, H. Kabashima, Involvement of catalyst materials in nonthermal plasma chemical processing of hazardous air pollutants, Catalysis Today 72 (2002) 259–265. [284] K. Krawczyk, B. Ulejczyk, Influence of water vapor on CCl4 and CHCl3 conversion in gliding discharge, Plasma Chemistry and Plasma Processing 24 (2004) 155–167. [285] S. Futamura, H. Einaga, A.H. Zhang, Comparison of reactor performance in the nonthermal plasma chemical processing of hazardous air pollutants, IEEE Transactions on Industry Applications 37 (2001) 978–985. [286] R. Yamashita, T. Takahashi, T. Oda, Humidity effect on non-thermal plasma processing for VOCs decomposition, in: IAS’96—Conference Record of the 1996 IEEE Industry Applications Conference, Thirty-First IAS Annual Meeting, vol. 1–4, 1996, pp. 1826–1829. [287] S. Agnihotri, M.P. Cal, J. Prien, Destruction of 1,1,1-trichloroethane using dielectric barrier discharge nonthermal plasma, Jorunal of Environmental Engineering 130 (2004) 349–355. [288] Y. Kim, K.T. Kim, M.S. Cha, Y.H. Song, S.J. Kim, CF4 decompositions using streamer- and glow-mode in dielectric barrier discharges, IEEE Transactions on Plasma Science 33 (2005) 1041–1046. [289] S. Futamura, A.H. Zhang, T. Yamamoto, Mechanisms for formation of inorganic byproducts in plasma chemical processing of hazardous air pollutants, IEEE Transactions on Industry Applications 35 (1999) 760–766. [290] Z. Falkenstein, Effects of the O2 concentration on the removal efficiency of volatile organic compounds with dielectric barrier discharges in Ar and N2 , Journal of Applied Physics 85 (1999) 525–529.

Journal of Hazardous Materials 195 (2011) 55–61

Contents lists available at ScienceDirect

Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat

Emulsion stabilization using ionic liquid [BMIM]+ [NTf2 ]− and performance evaluation on the extraction of chromium Rahul Kumar Goyal, N.S. Jayakumar, M.A. Hashim ∗ Department of Chemical Engineering, University of Malaya, Malaysia

a r t i c l e

i n f o

Article history: Received 7 December 2010 Received in revised form 23 February 2011 Accepted 9 March 2011 Available online 15 March 2011 Keywords: Emulsion liquid membrane [BMIM]+ [NTf2 ]− TOMAC Chromium Removal

a b s t r a c t This study focuses on the role of a hydrophobic ionic liquid 1-butyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide, [BMIM]+ [NTf2 ]− in the preparation of emulsion liquid membrane (ELM) phase containing kerosene as solvent, Span 80 as surfactant, NaOH as internal phase and TOMAC (tri-n-octylmethylammonium chloride) a second ionic liquid as carrier. The first time used [BMIM]+ [NTf2 ]− in ELM was found to play the role of a stabilizer. The emulsion prepared using [BMIM]+ [NTf2 ]− has a long period of stability of about 7 h (at 3% (w/w) of [BMIM]+ [NTf2 ]− ) which otherwise has a brief stability up to only 7 min. The stability of the emulsion increases with the increase in concentration of [BMIM]+ [NTf2 ]− up to 3% (w/w). Nevertheless, with further increase in concentration of [BMIM]+ [NTf2 ]− , a reduction in the stability occurs. The extraction experiments were carried out after holding the ELM for 2 h after the preparation and a removal efficiency of approximately 80% was obtained for Cr. The destabilization of the emulsion was studied by observing the change in the interface height. An empirical correlation for the stability of the emulsion has been proposed. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Liquid membrane technology is well known for its wide application in extraction processes to separate hydrocarbons [1–3], heavy metals [4–6], amin o acids [7,8] and biological compounds [9,10]. The potential advantages of liquid membrane techniques are low capital and operating costs, low energy and extractant consumption, high concentration factors and high fluxes. This technology has an edge over solvent extraction because it requires less energy and operates in a single stage for extraction and stripping. The main types of liquid membrane systems include emulsion liquid membrane, supported liquid membrane and bulk liquid membrane. However, the liquid membrane techniques have not been adopted for large scale industrial processes primarily due to problems in maintaining its stability. Emulsion liquid membrane (ELM) was invented by Li [11] to separate hydrocarbons and this technique has been utilized for many other applications such as metal extraction [5,6], wastewater treatment [6] and bio-medical separation [9]. Stability of emulsion is a major concern in the effective use of ELM either in laboratory scale or industrial scale. The resistance to rupture of liquid membrane at high shear stress defines the stability of the emulsion liquid membrane. Repeated coalescence of

∗ Corresponding author. Fax: +60 3 79675319. E-mail address: [email protected] (M.A. Hashim). 0304-3894/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jhazmat.2011.03.024

the internal droplets on the interface, creaming due to density difference, Ostwald Ripening and flocculation cause the instability of the emulsion. Several techniques to overcome the stability problem have been proposed, and these include the use of aliphatic solvent instead of aromatic solvent [12], the increment of the carbon chain length of the aliphatic solvent [13], the increment of the surfactant concentration [14], the increment of membrane viscosity [13,15], the use of co-surfactants [16], non-Newtonian conversion of the membrane phase [17], the use of Janus particles as stabilizers in emulsion polymerization [18] and the use of functionalized silica particles for high internal phase emulsion [19]. All of the remedies have their own tradeoffs and compromises with the overall extraction efficiency. Room temperature ionic liquids (RTILs) possess unique and exceptional properties such as negligible vapor pressure, inflammability, thermal stability even at high temperatures, highly polar yet non-coordinating solvent and application based adjustable miscibility/immiscibility in chemical processes [20–25]. These properties have made them potentially useful in a wide range of applications in industries as well as in research. Ionic liquids possess a very negligible vapor pressure that has enabled them to be used as a “green solvent” in synthesis [23,24,26–28], separation and purification [29–34], and electrochemical applications [35]. RTILs being stable and in the liquid form at room temperature, are made of organic cation and organic/inorganic anion. The physical and chemical properties of RTILs can be altered by changing the cation or anion or both to facilitate a particular task, hence they are

56

R.K. Goyal et al. / Journal of Hazardous Materials 195 (2011) 55–61

(a)

(b) C8H17 C8H17

+ N C8H17

[Cl-]

CH3

Table 1 Physical and thermodynamic properties of [BMIM]+ [NTf2 ]− [36].

N N C4H13 O O NS S CF3 CF3 O O +

[NTf2]-

Fig. 1. Structures of TOMAC and [BMIM]+ [NTf2 ]− .

sometimes referred as “task specific” ionic liquids [24]. However, very few studies have been reported on the application of ionic liquids in emulsion liquid membrane. Hence, an effort to incorporate ionic liquids with emulsion liquid membrane has been made by investigating the stability and % removal efficiency of emulsion liquid membrane in the presence of ionic liquids. Chromium metal was selected to investigate the removal efficiency of emulsion liquid membrane. The present study focuses on enhancing the stability of ELM, identification of the role of a hydrophobic ionic liquid [BMIM]+ [NTf2 ]− in it and extraction efficiency of the ELM. Ionic liquid [BMIM]+ [NTf2 ]− was chosen over other ionic liquids due to its hydrophobicity, minimum toxicity, relatively less viscosity and density. 2. Materials and methods 2.1. Chemicals Ionic liquids [BMIM]+ [NTf2 ]− and TOMAC, with structural formulae illustrated in Fig. 1, were directly obtained from Merck (Germany) and used without any further purification while kerosene of boiling point ranged from 180◦ C to 280 ◦ C was received from ACROS (USA). Span 80 (Sorbitan oleate or Sorbitan (Z)mono-9-octadecenoate) a non-ionic surfactant with a ratio of 4:3 of hydrophilic to lipophilic (HLB), was purchased from Merck (Malaysia). Sodium hydroxide pellets, potassium dichromate and hydrochloric acid were procured from R&M Chemicals (UK). The solution of sodium hydroxide of desired normality was prepared by dissolving appropriate weight of pellets in de-ionized water. Similarly, Cr solution of 500 mg/L was prepared by mixing suitable amount of potassium dichromate in de-ionized water. The prepared Cr solution was diluted with de-ionized water according to the required concentration.

Property

Temperature (◦ C)

Value

Density (g/ml)

25

1 .43

Viscosity (cP)

25

52

Surface tension (dyne/cm) (Water equilibrateda )

25

36 .8

Thermal decomposition temperature (◦ C) (Water equilibrateda ) Water content (mg/l) (Water equilibrateda ) Melting point (◦ C) (Driedb )

394 25

3280 4

Water equilibrated denotes that [BMIM]+ [NTf2 ]− was kept in contact with water. Dried stands for water equilibrated [BMIM]+ [NTf2 ]− that was dried at 70 ◦ C for 4 h on a vacuum line. a

b

bilizer) were added to the mixture. The mixture was homogenized for up to 5 min by the homogenizer at 8400 rpm. NaOH was added drop-wise into the mixture, keeping the whole mixture homogenized for the next 5 min. The ratio of internal to organic phase (I/O) was kept at 1:3 for all the experiments. The surfactant concentration (Span 80) (wherever applicable) was kept 3% (w/w) which is an optimized concentration in order to avoid swelling and to provide sufficient stability. Photographs of the beaker containing the emulsion were taken at regular intervals to analyze its stability. The photographs were analyzed by AUTOCAD to determine the phase separation rate of the emulsion. 2.3.2. Extraction of chromium The prepared emulsion was poured into another 250 mL beaker containing the Cr solution of 100 mg/L. The ratio of emulsion to feed phase (E/F) was kept at 1:2 for all the extraction experiments. The pH of the feed phase was maintained below 1.5 to establish a pH difference between the internal and external phases, hence maintaining a driving force for Cr to diffuse through the membrane. The whole mixture was gently stirred by a mechanical stirrer, and an agitation speed of 300 rpm was found to be the best to generate fine globules of emulsion with lowest possible breakage. Samples were taken at a regular interval using disposable syringes and the syringes were kept left undisturbed for some time until the emulsion and the feed phase were separated. The feed phase was then taken out, filtered and analyzed using ICPspectrophotometer.

2.2. Analytical instruments 3. Results and discussion An ICP-spectrophotometer (Perkin Elmer, model: Optima 7000 DV) was used for the measurement of the Cr concentration. The emulsion was prepared using a high speed homogenizer (IKA, model: T25 digtal Ultra Turrax) and the dispersion of the emulsion in the feed phase was carried out by a stirrer (IKA, model: RW11 Lab Egg). pH values were measured using a CyberScan 510 pH meter while photographs were taken using a digital camera (NIKON, model: DSLR D3000). Surface tension was measured by a tensiometer (Fisher Scientific, model: Tensiomat 21® ) using a Pt/Ir Du Noüy ring.

Physico-chemical properties of this ionic liquid are as shown in Table 1. 3.1. Identification of the role of [BMIM]+ [NTf2 ]− in emulsion without TOMAC

2.3. Procedure

As the first stage of this study, the role of ionic liquid [BMIM]+ [NTf2 ]− is conjectured to behave as one of the following: either as a carrier, surfactant, solvent or stabilizer. In order to substantiate its role, the following experiments were conducted and discussed below.

2.3.1. Preparation of emulsion and stability analysis The emulsion was prepared in a 100 mL unbaffled beaker by mixing organic solvent and an appropriate amount of non ionic surfactant Span 80. Subsequently, the carrier and ionic liquid (sta-

3.1.1. Consideration of [BMIM]+ [NTf2 ]− as a carrier In order to identify the role of [BMIM]+ [NTf2 ]− as a carrier, emulsion was prepared by taking kerosene as solvent, Span 80 as surfactant, NaOH (0.1 N) as internal phase and varying amount

R.K. Goyal et al. / Journal of Hazardous Materials 195 (2011) 55–61

57

Fig. 2. % removal of Cr with time as a function of [BMIM]+ [NTf2 ]− concentration (% w/w) of Cr without any TOMAC. Span 80 = 3% (w/w); I/O = 1:3. Emulsion to feed phase ratio was taken at 1:2.

of [BMIM]+ [NTf2 ]− . I/O ratio was maintained at 1/3 and Span 80 concentration was kept at the value of 3% (w/w). The initial Cr concentration in the feed phase was 100 mg/L and the pH of the feed was maintained below 1.5. The effect of ionic liquid on the percentage removal of Cr in an emulsion liquid membrane is, as shown in Fig. 2. Generally, the role of a carrier in emulsion liquid membrane extraction is to enhance the final removal and to increase the rate of extraction. In the absence of TOMAC (carrier), Cr extraction is facilitated by type-I mechanism if [BMIM]+ [NTf2 ]− cannot play the role of a carrier. If the ionic liquid was presumed to act as a carrier in the emulsion, then the final % removal of Cr and the rate of % removal of Cr should have increased with the increase in the ionic liquid concentration. On the contrary, an insignificant decrease in the final removal and a significant decrease in the rate of removal can be seen upon increasing the concentration of [BMIM]+ [NTf2 ]− from 0 to 2% (w/w), as shown in Fig. 2. Moreover, both the parameters keep on decreasing with the increase in the concentration of ionic liquid up to the value of 4% (w/w). This indicates that the [BMIM]+ [NTf2 ]− is not involved in making complex and transporting the metal from the feed phase to the internal phase. Hence, it’s a type-I facilitation where no carrier is present. The decrease in the % removal of chromium with an increase in concentration of [BMIM]+ [NTf2 ]− can be explained as the increased mass transfer resistance caused by [BMIM]+ [NTf2 ]− during the time of extraction and stripping. It was believed that the diffusion of Cr was hindered by the big size of [BMIM]+ [NTf2 ]− . Electrostatic and Van der waal’s attraction also slowed down the transport of Cr. A sudden decrease in the % extraction of chromium was observed at 6% (w/w) of [BMIM]+ [NTf2 ]− after 50 min. This discrepancy can be explained by the aggregated sedimenting tendency of [BMIM]+ [NTf2 ]− due to its high density after a long time and at higher concentration of the ionic liquid. From these observations and facts, it can be concluded that [BMIM]+ [NTf2 ]− cannot act as a carrier for this operation.

3.1.2. Consideration of [BMIM]+ [NTf2 ]− as a surfactant The role of a surfactant in ELM is to minimize the interfacial energy (interfacial tension) between the organic and the aqueous phase. There is no literature available regarding the use of [BMIM]+ [NTf2 ]− as a surfactant, so is for their HLB number. [BMIM]+ has the properties that can make this ionic liquid to act as a surfactant. Interfacial tension of kerosene and NaOH interface and CMC (Critical Micelle Concentration) of [BMIM]+ [NTf2 ]− were experimentally determined. Emulsion preparation with kerosene as solvent, NaOH as internal reagent and [BMIM]+ [NTf2 ]− as surfactant (assumed) was also tried out to check the feasibility of [BMIM]+ [NTf2 ]− acting as a surfactant. [BMIM]+ [NTf2 ]− is a hydrophobic ionic liquid. It has a density more than kerosene and NaOH. Several combinations of the concentration of organic phase (kerosene); aqueous phase (NaOH) and [BMIM]+ [NTf2 ]− were tested. No emulsion was yielded even if the mixture was homogenized at 15,000 rpm. The concentration of [BMIM]+ [NTf2 ]− was also varied from in the range of 0.40–7% (w/w) for all of the above combinations of phases but no emulsion was observed. The failure of micelle formation of [BMIM]+ [NTf2 ]− in kerosene can be explained by two possible reasons [37]. The first is the small hydrocarbon tail (butyl) attached to the cationic group of [BMIM]+ [NTf2 ]− that does not interact well enough with the kerosene hydrocarbon chain to yield the micelles of the ionic liquids. The other reason that may be attributed is the big size of [NTf2 ]− anion which is hard to fit the micelle surface region (Stern Layer). Therefore, it can be concluded that [BMIM]+ , the cationic part of [BMIM]+ [NTf2 ]− does not behave as a surfactant for the above mentioned solvent and the internal phase. 3.1.3. Consideration of [BMIM]+ [NTf2 ]− as a solvent Ionic liquids have been proved to be the solvents of future, primarily based on their unique properties over organic solvents. In order to verify the feasibility of ionic liquid [BMIM]+ [NTf2 ]− as a

58

R.K. Goyal et al. / Journal of Hazardous Materials 195 (2011) 55–61

Fig. 5. The effect of the concentration of [BMIM]+ [NTf2 ]− on the stability time of ELM. Span 80 = 3% (w/w); TOMAC = 0.29% (w/w), I/O = 1/3. Fig. 3. Emulsion prepared with 3% (w/w) Span 80, TOMAC = 0.29% (w/w); [BMIM]+ [NTf2 ]− = 0% (w/w) (after 7 min); scale bar = 1 cm.

solvent for emulsion liquid membrane the emulsion was prepared using [BMIM]+ [NTf2 ]− as a solvent, Span 80 as a surfactant and NaOH as the receiving phase while keeping the proportions of each component the same as it was prepared with kerosene as a solvent in the previous sections. The concentrations and volumes of [BMIM]+ [NTf2 ]− , Span 80 and NaOH and homogenizing speed and time were varied in order to get a stabilized membrane but the stability lasted only for 5 min with the best composition. The density difference between [BMIM]+ [NTf2 ]− and NaOH solution is higher than the difference between kerosene and NaOH therefore, proclivity of sedimenting of the former emulsion is much higher than the latter. Hence, emulsion formed with [BMIM]+ [NTf2 ]− as solvent lasted only for a short duration. The adsorbed amount of Span 80 on the surface of [BMIM]+ [NTf2 ]− was found to be very small and hence interfacial tension was not too reduced to make fine internal droplets of NaOH. The formation of small droplets of Span 80 took place on the surface of ionic liquid upon increasing the Span 80 concentration above 0.3% (w/w). Hence, another reason for the reduced stability may be explained as the insufficient reduction in the interfacial tension of the solvent by Span 80. From the above discussion, it could be concluded that [BMIM]+ [NTf2 ]− cannot be used as a solvent for the above mentioned surfactant and the internal phase to make a stable emulsion liquid membrane.

3.1.4. [BMIM]+ [NTf2 ]− as a stabilizer when TOMAC is used as a carrier TOMAC is a very good phase transfer catalyst which is relatively less expensive, easily available and less toxic. Hence, TOMAC was selected as a carrier to study the effect of ionic liquid ([BMIM]+ [NTf2 ]− ) on the stability of emulsion and subsequent removal efficiency of the same. The problem with TOMAC when it is used as a carrier in the emulsion liquid membrane with no concentration of [BMIM]+ [NTf2 ]− , stability lasted only for 7 min, as shown in Fig. 3 which is not sufficient for the extraction to take place and for subsequent demulsification. From Fig. 4(a), it can be observed clearly that the emulsion was stable for up to 7 h when 3% (w/w) of the ionic liquid [BMIM]+ [NTf2 ]− is added. On the other hand, Fig. 4(b) depicts the separated organic and aqueous phases after 5 h when there is no concentration of [BMIM]+ [NTf2 ]− present in the emulsion. In fact, the phase separation started only after 7 min without [BMIM]+ [NTf2 ]− in the emulsion. The stability time with the varying concentration of [BMIM]+ [NTf2 ]− is, as shown in Fig. 5. The stability time increases with an increase in the concentration of [BMIM]+ [NTf2 ]− for up to 3% (w/w) of [BMIM]+ [NTf2 ]− . After 3% (w/w) of [BMIM]+ [NTf2 ]− onwards, the stability time decreased which consolidated the fact that the [BMIM]+ [NTf2 ]− helped to stabilize the emulsion liquid membrane when it was present up to a certain maximum concentration. If the emulsion liquid membrane contained more than this concentration then emulsion sedimentation took place due to the higher density of ionic liquid. Each experiment was

Fig. 4. (a) Emulsion prepared with [BMIM]+ [NTf2 ]− = 3% (w/w); TOMAC = 0.29% (w/w) (after 7 h), (b) Emulsion prepared with [BMIM]+ [NTf2 ]− = 0% (w/w); TOMAC = 0.29% (w/w) (after 5 h); scale bar = 1 cm.

R.K. Goyal et al. / Journal of Hazardous Materials 195 (2011) 55–61

59

Fig. 6. % removal of Cr with time as a function of [BMIM]+ [NTf2 ]− concentration (% w/w) with TOMAC. TOMAC = 0.29% (w/w); Span 80 = 3% (w/w); I/O = 1:3. E/F =1:2.

conducted twice and the results for the stability time were reproducible with minor difference between two corresponding values. The increased stability of the emulsion liquid membrane by addition of [BMIM]+ [NTf2 ]− may be explained by Coulombic interactions of the charges on the ions of ionic liquids [BMIM]+ [NTf2 ]− and TOMAC. The other interactions present in the emulsion are between other chemical complex groups such as Span 80 and TOMAC; TOMAC and NaOH; [BMIM]+ [NTf2 ]− and NaOH. These strong interactions help to avoid the coalescence of the internal droplets but they also cause the hindrance to Cr–TOMAC complex diffusion through the membrane. Apart from strong interactions between ions, there is a possibility of hydrogen bonding present between [BMIM]+ [NTf2 ]− and [OH]− group of NaOH. The hydrogen bonding may cause a strong protection surrounding the internal droplets to avoid coalescence. [BMIM]+ [NTf2 ]− is capable of developing a polymeric structure with large cavities [38] when it is used for different kinds of reactions. These polymeric structures of ionic liquid may also help to understand the cause for the enhanced stability. The polymeric structure of ionic liquid [BMIM]+ [NTf2 ]− may behave like polymeric surfactant of the A–B, A–B–A or (BA)n graft type to generate a repulsive barrier to prevent the collapse of the emulsion liquid membrane.

3.1.5. The removal efficiency of the emulsion liquid membrane stabilized by ionic liquid [BMIM]+ [NTf2 ]− with TOMAC The prepared emulsion was kept for 2 h to verify its stability then it was poured into an unbaffled beaker containing Cr feed phase. The emulsion was prepared with varied concentration of [BMIM]+ [NTf2 ]− . The samples were taken at regular intervals. The effect of ionic liquid [BMIM]+ [NTf2 ]− on the removal efficient of the emulsion liquid membrane having TOMAC as a carrier is, as shown in Fig. 6. TOMAC concentration was kept at a constant value of 0.29% (w/w) for all the experiments. On the contrary, the percentage removal of the emulsion liquid membrane prepared with TOMAC and [BMIM]+ [NTf2 ]− decreases due to the hindrance caused by both of the compounds. From Fig. 6, the time taken for 70% of the removal of chromium is only 5 min.

3.1.6. Effect of [BMIM]+ [NTf2 ]− concentration on phase separation rate of the stabilized emulsion The emulsion was prepared by taking kerosene as solvent, Span 80 as surfactant, NaOH (0.1 N) as internal phase, TOMAC as carrier and varying amount of [BMIM]+ [NTf2 ]− . I/O ratio was maintained at 1/3 and Span 80 concentration was kept 3% (w/w). TOMAC concentration was kept at a constant value of 0.29% (w/w). The stabilized membranes started to separate into organic and aqueous phases after their maximum time of stability which is dependent on the concentration of [BMIM]+ [NTf2 ]− . Creaming and coalescence are the main causes for emulsion sedimentation for the current composition of emulsion. The sedimentation of the emulsion due to Ostwald ripening is insignificant, since aqueous NaOH and kerosene are almost insoluble in each other. The stabilized membrane was held for the next 3 h after it started to destabilize to analyze the stability with respect to the concentration of [BMIM]+ [NTf2 ]− . The calculation of the phase separation was done by noting the height of the interface from the bottom of the beaker at a regular interval. The normalized height of the emulsion is a ratio of the height of the interface from the ground level and the total height of the emulsion. Therefore, it’s a dimensionless quantity. The stabilized membrane stability time and their phase separation with respect to time are shown in Fig. 7. It can be observed from Fig. 7 that an increment in the concentration of [BMIM]+ [NTf2 ]− to 3% (w/w) increases the final (after 3 h) interface height of the destabilized emulsion. It implies that the sedimentation rate decreases upon increasing the concentration of [BMIM]+ [NTf2 ]− up to 3% (w/w). The decreased sedimentation rate may be explained by the effective electrostatic interactions between the two ionic liquids, [BMIM]+ [NTf2 ]− and TOMAC, over the density of ionic liquids and NaOH. Fig. 7 illustrates that an increment in the concentration of [BMIM]+ [NTf2 ]− above 3% (w/w) decreases the interface height of the destabilized emulsion. It means that the sedimentation rate of the destabilized emulsion increases upon increasing the concentration of [BMIM]+ [NTf2 ]− above 3% (w/w). The increased sedimentation could be understood by the dominance of the density of [BMIM]+ [NTf2 ]− and NaOH over electrostatic interactions between ionic liquids [BMIM]+ [NTf2 ]− and TOMAC. However, the complete phase separation of the stabilized membrane into its original phases took place only after 2–3 days.

60

R.K. Goyal et al. / Journal of Hazardous Materials 195 (2011) 55–61

Fig. 7. The effect of ionic liquid concentration (% w/w)) [BMIM]+ [NTf2 ]− on the phase separation of the stabilized membrane. TOMAC = 0.29% (w/w); Span 80 = 3% (w/w); I/O = 1:3.

The experimental data of stability time and initial rate of sedimentation of the emulsion prepared with TOMAC = 0.29% (w/w); Span 80 = 3% (w/w); I/O = 1:3 has been given in Table 2. The stability time of the emulsion as a function of [BMIM]+ [NTf2 ]− concentration (% w/w) (x1 ) and initial rate of sedimentation (x2 ) is correlated to be as ts, pred = ax1b + x2c + 7

(1)

where a, b and c are parameters. The term 7 min in the above correlation takes care of stabilized time when concentration of [BMIM]+ [NTf2 ]− is equal to zero in emulsion liquid membrane containing TOMAC as carrier, NaOH as stripping phase and kerosene as solvent. The parameters are estimated using data in Table 1, by nonlinear parameter estimation scheme with the help of MATLAB 7.0.4 software and the predicted stabilized time relationship is given as ts, pred = 1446.6x10.54549 x20.38057 + 7

Table 3 Comparison between experimental stability time and predicted and stability time. Concentration of [BMIM]+ [NTf2 ]−

ts,exp (min)

ts,pred (min)

% deviation

0.6 1.5 2.0 3.0 4.0

220 330 400 425 465

221.4 335.3 365.9 430.5 457.7

0.6 1.6 8.5 1.3 1.5

(2)

The stability time of the emulsion is dependent more on the concentration of [BMIM]+ [NTf2 ]− than the initial rate of sedimentation of the emulsion, as observed from Eq. (2). The comparison between the predicted values and experimental values has been reported in Table 3. The low values of deviations as summarized in Table 3 imply the accuracy of the correlation. The predicted and experimental stabilized times are in good agreement within ±9% deviation as shown in Fig. 8. However, the

Fig. 8. Experimental stability time versus predicted stability time.

Table 2 Experimental data on stabilized time with ionic liquid concentration [BMIM]+ [NTf2 ]− and initial rate of sedimentation height with TOMAC = 0.29% (w/w); Span 80 = 3% (w/w); I/O = 1:3.

correlation is only applicable for lower range of the ionic liquid concentrations. It does not hold the accuracy for the higher concentrations of ionic liquid [BMIM]+ [NTf2 ]− .

Concentration of [BMIM]+ [NTf2 ]− (% w/w)

Initial rate of sedimentation time (min−1 )

Experimental stability (min)

4. Conclusion

0.6 1.5 2.0 3.0 4.0

0.0150 0.0120 0.0100 0.0086 0.0067

220 330 400 425 465

The present work focuses on the stability aspects of emulsion liquid membrane and in this context, the experimental investigation identifies the use of hydrophobic ionic liquid 1-butyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide, [BMIM]+ [NTf2 ]− as a stabilizer with the preparation of the emul-

R.K. Goyal et al. / Journal of Hazardous Materials 195 (2011) 55–61

sion liquid membrane containing TOMAC as a carrier. The enhanced stability of the emulsion liquid membrane caused by the addition of [BMIM]+ [NTf2 ]− could be explained by the strong interactions such as coulombic, dipolar and ionic interactions among the ionic liquids and NaOH. It was observed that the stability of the emulsion liquid membrane could be enhanced for a duration of up to 7 h. Experiments had proved that the stability of the emulsion liquid membrane can be enhanced for a duration of up to 7 h. 80% removal of chromium could be achieved even after keeping the emulsion for 2 h before the extraction experiments were carried out. The sedimentation rate of the stabilized membrane for the next 3 h after its maximum stability time was found to be decreasing with the increase in concentration of [BMIM]+ [NTf2 ]− up to 3% (w/w). It starts to increase with further increase in the concentration of [BMIM]+ [NTf2 ]− . An empirical correlation relating stability time of emulsion as a function of [BMIM]+ [NTf2 ]− concentration (% w/w) and initial rate of sedimentation of the emulsion is proposed and the predicted stability times are in good agreement with the experimental stability times. Ultimately, this paper reflects the potential use of “task specific” ionic liquids as a stabilizer in the field of emulsion liquid membrane. List of symbols a, b, c constant E/F emulsion to feed phase ratio I/O internal to organic phase ratio ts, exp experimental stability time of the emulsion (min) ts, pred predicted stability time of the emulsion (min) x1 concentration of [BMIM]+ [NTf2 ]− (% w/w) initial rate of sedimentation of the emulsion (min−1 ) x2 Acknowledgement The authors acknowledge funding from University of Malaya, Malaysia. References [1] C.M. Das, G. Rungta, S.D. Arya, S. De, Removal of dyes and their mixtures from aqueous solution using liquid emulsion membrane, J. Hazard. Mater. 159 (2008) 365–371. [2] C.C. Lin, R.L. Long, Removal of nitric acid by emulsion liquid membrane: experimental results and model prediction, J. Membr. Sci. 134 (1997) 33–45. [3] P. Venkateswaran, K. Palanivelu, Recovery of phenol from aqueous solution by supported liquid membrane using vegetable oils as liquid membrane, J. Hazard. Mater. 131 (2006) 146–152. [4] S. Chowta, P.K. Mohapatra, B.S. Tomar, K.M. Michael, A. Dakshinamoorthy, V.K. Manchanda, Recovery of americium(III) from low acid solutions using an emulsion liquid membrane containing PC-88A as the carrier extractant, Desalination Water Treat. 12 (2009) 62–67. [5] R.A. Kumbasar, I. Sahin, Separation and concentration of cobalt from ammoniacal solutions containing cobalt and nickel by emulsion liquid membranes using 5,7-dibromo-8-hydroxyquinoline (DBHQ), J. Membr. Sci. 325 (2008) 712– 718. [6] H.R. Mortaheb, H. Kosuge, B. Mokhtarani, M.H. Amini, H.R. Banihashemi, Study on removal of cadmium from wastewater by emulsion liquid membrane, J. Hazard. Mater. 165 (2009) 630–636. [7] H. Itoh, M.P. Thien, T.A. Hatton, D.I.C. Wang, Water transport mechanism in liquid emulsion membrane process for the separation of amino acids, J. Membr. Sci. 51 (1990) 309–322. [8] M. Matsumoto, T. Ohtake, M. Hirata, T. Hano, Extraction rates of amino acids by an emulsion liquid membrane with tri-n-octylmethylammonium chloride, J. Chem. Technol. Biotechnol. 73 (1998) 237–242.

61

[9] T. Kaghazchi, A. Kargari, R. Yegani, A. Zare, Emulsion liquid membrane pertraction of L-lysine from dilute aqueous solutions by D2EHPA mobile carrier, Desalination 190 (2006) 161–171. [10] J.-Q. Shen, W.-P. Yin, Y.-X. Zhao, L.-J. Yu, Extraction of alanine using emulsion liquid membranes featuring a cationic carrier, J. Membr. Sci. 120 (1996) 45–53. [11] N.N. Li, Separating Hydrocarbons with Liquid Membranes, U.S. Patent 3,410,794 (1968). [12] F. Nakashio, M. Goto, M. Matsumoto, J. Irie, K. Kondo, Role of surfactants in the behavior of emulsion liquid membranes—development of new surfactants, J. Membr. Sci. 38 (1988) 249–260. [13] T. Kinugasa, K. Watanabe, H. Takeuchi, Effect of organic solvents on stability of liquid surfactant membranes, J. Chem. Eng. Japan 22 (1989) 593–597. [14] E.C. Hsu, N.N. Li, Membrane recovery in liquid membrane separation processes, Sep. Sci. Technol. 20 (2–3) (1985) 115–130. [15] R.E. Terry, N.N. Li, W.S. Ho, Extraction of phenolic compounds and organic acids by liquid membranes, J. Membr. Sci. 10 (1982) 305–323. [16] Y.S. Mok, K.H. Lee, W.K. Lee, Control of swelling in liquid emulsion membranes employed for lactic acid separation, J. Chem. Tech. Biotechnol. 65 (1996) 309–316. [17] A.H.P. Skelland, X. Meng, Non-Newtonian conversion solves problems of stability, permeability, and swelling in emulsion liquid membranes, J. Membr. Sci. 158 (1999) 1–15. [18] H.E. Muller, A. Walther, M. Hoffmann, Emulsion polymerization using janus particles as stabilizers, Angew. Chem. Int. Ed. 47 (2008) 711–714. [19] A. Bismarck, V.O. Ikem, A. Menner, High internal phase emulsions stabilized solely by functionalized silica particles, Angew. Chem. Int. Ed. 47 (2008) 8277–8279. [20] J.D. Holbrey, K.R. Seddon, Ionic liquids, Clean Prod. Processes 1 (1999) 223–237. [21] M.J. Earle, K.R. Seddon, Ionic liquids. Green solvents for the future, Pure Appl. Chem. 72 (7) (2000) 1391–1398. [22] K. Mikami, Green Reaction Media in Organic Synthesis, first ed., Blackwell, UK, 2005. [23] T. Welton, Room-temperature ionic liquids. Solvents for synthesis and catalysis, Chem. Rev. 99 (1999) 2071–2083. [24] P. Wasserscheid, T. Welton, Ionic Liquids in Synthesis, Second ed., WILEY-VCH, Weinheim, 2008. [25] M. Freemantle, New horizons for ionic liquids, Chem. Eng. News 79 (2001) 21–25. [26] C.J. Adam, M.J. Earle, G. Robert, K.R. Seddon, Friedel–crafts reactions in room temperature ionic liquids, Chem. Commun. 19 (1998) 2097–2098. [27] P.J. Dyson, D.J. Ellis, D.C. Parker, T. Welton, Arene hydrogenation in a roomtemperature ionic liquid using a ruthenium cluster catalyst, Chem. Commun. 1 (1999) 25–26. [28] C.E. Song, E.J. Roh, Practical method to recycle a chiral (salen) Mn epoxidation catalyst by using an ionic liquid, Chem. Commun. 10 (2000) 837–838. [29] A.E. Visser, R.P. Swatloski, W.M. Reichert, R. Mayton, S. Sheff, A. Wierzbicki, J.H. Davis Jr., R.D. Rogers, Task-specific ionic liquids for the extraction of metal ions from aqueous solutions, Chem. Commun. 1 (2001) 135–136. [30] A.P. de los Rios, F.J. Hernandez-Fernandez, L.J. Lozano, S.S. Sanchez, J.I. Moreno, C. Godlinez, Removal of metal ions from aqueous solutions by extraction with ionic liquids, J. Chem. Eng. Data 55 (2010) 605–608. [31] A.E. Visser, R.P. Swatloski, W.M. Reichert, S.T. Griffin, R.D. Rogers, Traditional extractants in nontraditional solvents: group 1 and 2 extraction by crown ethers in room-temperature ionic liquids, Ind. Eng. Chem. Res. 39 (10) (2000) 3596–3604. [32] S. Dai, Y.H. Ju, C.E. Barnes, Solvent extraction of strontium nitrate by a crown ether using room temperature ionic liquids, J. Chem. Soc. Dalton Trans. (1999) 1201–1202. [33] L.C. Branco, J.G. Crespo, C.A.M. Afonso, High selective transport of organic compounds by using supported liquid membranes based on ionic liquids, Angew. Chem. Int. Ed. 41 (2002) 2771–2773. [34] A.G. Fadeev, M.M. Meagher, Opportunities for ionic liquids in recovery of biofuels, Chem. Commun. 3 (2001) 295–296. [35] I.W. Sun, C.L. Hussey, Electrochemistry of niobium chloride and oxide chloride complexes in the basic aluminum chloride-1-methyl-3-ethylimidazolium chloride room-temperature ionic liquid, Inorg. Chem. 28 (1989) 2731–2737. [36] J.G. Huddleston, A.E. Visser, W.M. Reichert, H.D. Willauer, G.A. Broker, R.D. Rogers, Characterization and comparison of hydrophilic and hydrophobic room temperature ionic liquids incorporating the imidazolium cation, Green Chem. 3 (2001) 156–164. [37] M. Blesic, M.H. Marques, N.V. Plechkova, K.R. Seddon, L.P.N. Rebelo, A. Lopes, Self-aggregation of ionic liquids: micelle formation in aqueous solution, Green Chem. 9 (2007) 481–490. [38] H. Weingrtner, Understanding ionic liquids at the molecular level: facts, problems, and controversies, Angew. Chem. Int. Ed. 47 (2008) 654–670.

Journal of Hazardous Materials 195 (2011) 62–67

Contents lists available at ScienceDirect

Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat

The removal of some rare earth elements from their aqueous solutions on by-pass cement dust (BCD) O.I.M. Ali a , H.H. Osman a , S.A. Sayed a,b , M.E.H. Shalabi c,∗ a b c

Faculty of Science, Helwan University, Ain Helwan, Cairo, Egypt Ha’el University, Saudi Arabia1 CMRDI, Tabbin, Cairo, Egypt

a r t i c l e

i n f o

Article history: Received 20 October 2010 Received in revised form 30 July 2011 Accepted 4 August 2011 Available online 25 August 2011 Keywords: BCD Rare earth elements Uptake Sorption

a b s t r a c t The sorption behavior of yttrium (Y3+ ), neodymium (Nd3+ ), gadolinium (Gd3+ ), samarium (Sm3+ ) and lutetium (Lu3+ ) from their aqueous solutions by by-pass cement dust (BCD) has been investigated using a batch technique. The sorption on BCD was studied as a function of pH, shaking time, initial concentration, mass of BCD and temperature. It was found that the sorption capacity of BCD had the order of Lu3+ > Sm3+ > Y3+ > Gd3+ ≈ Nd3+ following Freundlich isotherm at the determined optimum conditions. The results also demonstrated that the sorption data fit well the pseudo-second-order kinetic model. Thermodynamic parameters such as H◦ , S◦ and G◦ indicated that the sorption of the investigated REEs on BCD was endothermic, favored at high temperature and spontaneous in nature. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Recently, more and more rare earth elements (REEs) enter the environment through various pathways as a result of rapid increase of the utilization of REE resources and their applications in many fields of modern industry and daily life [1–3]. Rare earth elements are used as petrogenetic tracers in internal geodynamic studies of the earth [4]. Moreover, millions of tons of fertilizers containing REEs are also used worldwide for increasing agricultural productivity [5]. The increased demand for REEs means increased public exposure to the REEs both from various commercial products and from production wastes/effluents. In regions with high levels of REEs, elevated levels of REEs are found in human [6] and other organisms [7]. REEs are entering the human body due to exposure to various industrial processes can affect metabolic processes. Trivalent ions, especially La(III) and Gd(III) can interfere with calcium channels in human and animal cells. They can also alter or even inhibit the action of various enzymes and when they found in neurons can regulate synaptic transmission, as well as block some receptors (for example, glutamate receptors) [8]. Numerous techniques are available for the separation and recovery of REEs such as chemical precipitation, ion exchange, solvent extraction and adsorption [9–12]. Adsorption represents the most

∗ Corresponding author. Tel.: +20 223591108. E-mail address: [email protected] (M.E.H. Shalabi). 1 On leave. 0304-3894/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jhazmat.2011.08.014

efficient and widely applied technique. During the last decade, research efforts have been directed towards using low cost alternative techniques or development of low cost sorbents applicable for the removal and separation of undesirable metal ions from an aqueous phase [13–15]. A variety of low cost sorbents such as fly ash [16], rice husk [17], peat [18], peat moss [19], red mud [20], teawaste [21], olive stones [22], date bits [23] have been tried. Recently, research efforts have been directed towards the use of wastes as sorbent materials in an attempt to minimize the processing costs and to protect the environment and public health [24]. By-pass cement dust (BCD) is the by-product of the manufacture of Portland cement. It is a fine grained material that is generated during the calcination process in the kiln. CaO is the major constitutes of BCD composition. Other constituents include SiO2 , Al2 O3 , Fe2 O3 , K2 O, Na2 O, Cl, etc. Most of cement companies generate high quantities of BCD every year [25]. Potential effectiveness of BCD as a low cost sorbent for some pollutants has been demonstrated. For instance, BCD was used for removal; some heavy metals from textile industrial effluents [26], Cu(II), Ni(II), Zn(II), Fe(III), Co(III), U(VI) and Th(IV) from aqueous solutions [27–29], Cr(III) from tanning wastewater effluents [30] and in wastewater treatment [31]. Therefore, the objective of this research is to investigate the influence of various experimental parameters on sorption of some rare earth elements (including yttrium, neodymium, gadolinium, samarium, and lutetium) on BCD as an effective and low cost sorbent material in aqueous solutions. Moreover, the kinetics, isotherms and thermodynamics characteristics of the sorption

O.I.M. Ali et al. / Journal of Hazardous Materials 195 (2011) 62–67

63

complexing agent at 655 nm against reagent blank [34]. Uptake percentage (%E) and distribution constant KD (mL/g) were calculated from the equations: %E =

Co − Ce × 100 Co

KD (mL/g) =

(1)

C − C  V o e Ce

m

(2)

where Co and Ce are the initial and equilibrium REE concentration in the solution (mg/L), respectively. V is the volume of the aqueous solution (mL) contacted with BCD and m is the mass of BCD in grams. 2.3. Sorption kinetics

Fig. 1. X-ray diffraction of by-pass cement dust.

process of the investigated elements from their aqueous solutions on BCD will be discussed.

To examine the controlling mechanism of the sorption process, two kinetic models were used to test experimental data. The Lagergren pseudo first-order equation is a simple kinetic analysis of sorption in the following form [35]: log(qe − qt ) = log qe −

2. Experimental 2.1. Materials 2.1.1. By-pass cement dust (BCD) By-pass cement dust (BCD) was brought from National Cement Co., Egypt. The received BCD was placed in a glass container which in turn was kept in a desiccator all the time of the experiments. BCD chemical composition (Table 1) was identified using X-ray fluorescence (PANalytical Axios advanced, The Netherlands). The constituents’ phases of BCD were identified by X-Ray diffraction analysis (XRD brucker axs D8 advance, Germany) with CuK␣ radia˚ with a typical scanning begin at 2 equal to 20–80◦ tion (1.5406 A) and scan rate of 20 min−1 . X-Ray diffraction pattern of BCD is shown in Fig. 1. It indicates that BCD mainly consists of calcite, calcium sulfate, mono calcium silicates, calcium carbonate, quartz and sodium chloride (where CaCO3 = 38.34%, NaCl = 3.77%, KCl = 5.54%, CaSO4 = 4.93%, SiO2 free = 7.08%, CaSiO3 = 5%, CaO free = 12.20%). The cation exchange capacity (CEC) of BCD was determined using ammonium acetate saturation method [32] and it was 6.53 meq/g. 2.1.2. Reagents Standard individual REE solutions (1000 mg/L) for Y(III), Nd(III), Gd(III), Sm(III) and Lu(III) were prepared from the corresponding oxides (Aldrich Co., Germany) by dissolving 0.254, 0.162, 0.231, 0.116 and 0.26 g, respectively in 10 mL of perchloric acid. The solutions were heated till complete evaporation and another 5 mL of perchloric acid was added, then the solutions were diluted to volume with 0.1 N perchloric acid. Arsenazo (III) was obtained from Aldrich Co, Germany. All other reagents used were analytical reagent grade. In all experiments, doubled distilled water was utilised for preparation and dilution of solutions.

 k t  1 2.303

(3)

where qe and qt are the sorbed concentration of REE at equilibrium and at any time t, respectively and k1 is the overall rate constant of first-order sorption. The slope of the plot of log(qe − qt ) as a function of t can be used to determine the first-order rate constant k1 . The activation energy of the sorption process can be determined using Arrhenius equation. It can be calculated from the slope of a plot of log k and 1/T, since the slope equal to (−Ea /2.303R). In addition, the pseudo second-order equation based on sorption equilibrium capacity may be expressed in the following form [35]: 1 1 = + k2 t qe − qt qe

(4)

where k2 is the rate constant of the second-order sorption. Eq. (4) can be rearranged to: t 1 t = + 2 qt q k2 qe e

(5)

Similarly, the slope of the plot of t/qt as a function of t was used to determine the second order rate constant k2 that is used to determine the activation energy of the sorption process using Arrhenius equation. 2.4. Thermodynamic parameters Determination of thermodynamic parameters were based on experiments that carried out by shaking 0.02 g BCD with solutions of each REE of concentration (100 mg/L) adjusted at pH 7 ± 0.1 for 5 min at different temperatures. The thermodynamic parameters (H◦ , G◦ and S◦ ) were calculated from the sorption results. 3. Results and discussion

2.2. Sorption experiments 3.1. Sorption experiments The sorption experiments were studied by a batch technique. In the experiments, BCD was separately shaken with each REE solution at various experimental conditions. Separating of solid phase from liquid was done by centrifuging at 4000 rpm for 15 min. The pH of the solutions was maintained by thiel buffer [33] in the range (2–7). After equilibration, the REEs’ concentrations were determined spectrophotometrically employing Shimadzu UV–Vis160A Spectrophotometer using Arsenazo III (0.05%, w/v) as a

The parameters which may affect the uptake of REEs by pass cement dust, such as shaking time, temperature, pH, initial concentration of REE and sorbent mass were investigated. The results showed that the equilibrium reached to its maximum within 9 min of shaking, while uptake percentage only slightly increased on raising temperature up to 60 ◦ C. Therefore, the sorption experiments were carried out at room temperature for 9 min.

64

O.I.M. Ali et al. / Journal of Hazardous Materials 195 (2011) 62–67

Table 1 Major chemical constituents of BCD. Constituents

Mass (%)

CaO

SiO2

P2 O5

Na2 O

Fe2 O3

MgO

K2 O

Al2 O3

Cl−

SO3

LOI

44.24

10.54

6.35

2.00

1.92

1.45

3.5

4.7

4.93

2.9

17.47

Table 2 Variation of the KD with the REEs’ initial concentrations for their sorption on BCD. [REE]

50 100 200 300 400 500

KD × 104 Y3+

Nd3+

Gd3+

Sm3+

Lu3+

6.12 0.61 0.08 0.03 0.02 0.02

6.12 0.68 0.06 0.04 0.02 0.02

6.12 6.12 6.12 6.12 0.07 0.03

6.12 6.12 6.12 6.12 0.92 0.28

6.12 0.74 0.11 0.05 0.03 0.03

Fig. 3. Variation of the meq (REE)/g BCD with the REEs’ initial concentrations for their sorption on BCD. Operating conditions: 50 mL solution, 0.04 g BCD, pH = 7, and shaking time = 9 min.

Fig. 2. Variation of the uptake percentage with pH for REEs sorption on BCD. Operating conditions: 50 mL solution, [REE] = 100 mg/L, 0.02 g BCD, and shaking time = 9 min.

3.1.1. Effect of pH Fig. 2 shows the influence of pH on the sorption of the investigated REEs. The data reveals that the percentage of sorption steeply increases with increasing pH up to 7 ± 0.1. Consequently, in the subsequent work, the sorption experiments were carried out at pH 7. Generally, below pH 2 dissolution of BCD was occurred. Above pH 7 the formation of a precipitate was observed. Low REEs uptake at low pH values is most probably due to the protonation of the active sites in BCD, which inhibits their binding ability towards the REEs [36]. In addition, as pH increases, surface positive charge decreases, this would result in lower columbic repulsion of the adsorbed REE ions [37]. In aqueous solution, the hydrolysis of trivalent lanthanides begins at pH as low as 6 and various species can be formed, such as Ln(OH)2+ , Ln(OH)2 + , Ln(OH)3 , Ln(OH)4 − [38]. Thus, as pH increases, hydrolysis precipitation most probably would start due to the formation of various hydrocomplexes in aqueous solution. The observed reduction in the percentage of REE uptake on BCD at low pH by the sorbent indicates that the interaction between them is most probably due to an ion exchange process [39]. 3.1.2. Effect of initial concentration of REE The sorption of each REE as a function of their initial concentrations was studied at room temperature by varying the REE initial concentration from 50 to 500 mg/L. Table 2 shows the calculated

distribution coefficients (KD ) at different REE initial concentrations. Over the studied range, the distribution coefficients vary by more than 6 orders of magnitude. The inverse correlation between REE initial concentration and distribution coefficient reflects the greater partitioning of REEs into the solid phase (up to 100 mg/L for Y3+ , Nd3+ , Gd3+ and 300 mg/L for Sm3+ and Lu3+ ). With increasing REEs concentration, the distribution coefficients stabilize, owing to saturation of the BCD surface. These results indicate that energetically less favorable sites become involved with increasing REEs concentration in the aqueous solution. Additionally, the results were expressed in terms of meq of REE sorbed per unit mass of BCD (Fig. 3) to investigate the difference in affinity of BCD towards each REE. It is obvious from this figure that the sorption capacity of BCD has the order of Lu3+ > Sm3+ > Y3+ > Gd3+ ≈ Nd3+ . The amount of REE sorbed per unit mass of BCD increased with the initial REE concentration as expected. The plateau values for Lu3+ , Y3+ , Gd3+ , and Nd3+ represent saturation of the active sites on BCD available for interaction. The decrease in Sm3+ sorption at high concentration may be due to the Sm3+ affinity to sorbate–sorbate interaction than sorbate–sorbent one. This is more or less in agreement with the sorption of Sm3+ by natural clinoptilolite-containing Tuff [40]. The data from Fig. 3 reveals that the sorption capacity of Sm3+ and Lu3+ was higher than CEC of BCD which reflects again that sorption mechanism of these elements is a mixed mechanism of ion exchange and hydrolysis precipitation as well. 3.1.3. Effect of sorbent mass Effect of sorbent mass on the sorption process of the investigated REEs is represented in Fig. 4. The experimental results reveal that the sorption efficiency of REEs increases up to the optimum mass of 0.04 g BCD for Y3+ , Gd3+ , Nd3+ and 0.01 g for Sm3+ and Lu3+ beyond which the sorption efficiency does not change with the sorbent mass.

O.I.M. Ali et al. / Journal of Hazardous Materials 195 (2011) 62–67

65

Fig. 5. Langmuir sorption isotherm of REEs sorption on BCD.

3.4. Sorption isotherms models Fig. 4. Variation of the uptake percentage with the sorbent mass for REEs sorption on BCD. Operating conditions: 50 mL solution, [REE] = 100 mg/L, pH = 7, and shaking time = 9 min.

The increase in uptake percentage with increasing the BCD mass may be due to increasing number of sorbent particles in the solution that allows more REEs ions to interact with more binding sites. 3.2. Sorption kinetics For simplicity, the kinetics experiments were carried out for Gd3+ and Sm3+ . The studies shows that the pseudo-first-order kinetic model did not fit the data for the sorption process since the values of correlation factor R2 were small. Therefore, the pseudosecond-order kinetic model was applied. The calculated values of k2 and Ea with the values of the linear correlation coefficients (R2 ) are represented in Table 3. The values of correlation factor R2 of the pseudo-second-order were better than those of pseudo-first-order model indicating second-order kinetics of the sorption process on BCD. The natural logarithms of the rate constants (k) were used according to the Arrhenius equation to calculate the activation energy of sorption process. It was found that the activation energy was within the range of 0–40 kJ/mol, which means that the sorption process is a physical one [41]. 3.3. Thermodynamic parameters For better understanding the mechanism of the sorption process of REEs on BCD, thermodynamic parameters were determined. They were calculated using the equation: ln KD =

H ◦ S ◦ − R RT

(6)

where KD is the distribution coefficient (mL/g), S◦ is standard entropy (J/mol K), H◦ is standard enthalpy (kJ/mol), T is the absolute temperature (K) and R is the gas constant (8.314 J/mol K). The standard Gibbs free energy G◦ values (kJ/mol) were calculated from the equation: G◦ = H ◦ − TS ◦ H◦ ,

(7) S◦

G◦

The values of and were calculated from the slopes and intercepts of plots of ln KD versus 1/T. The values are presented in Table 4. The positive value of enthalpy change H◦ for the process confirms the endothermic nature of the process, while the positive entropy of sorption S◦ reflects the affinity of BCD towards these elements. The obtained values of G◦ point to that the feasibility of the sorption process of the investigated elements on BCD and its spontaneous nature without an induction period.

The sorption data of Sm3+ , Gd3+ and Nd3+ have been subjected to different sorption isotherms models namely Langmuir and Freundlich over a range of 100–300 mg/L for Nd3+ and 100–400 mg/L for Gd3+ and Sm3+ . The goodness-of-fit between experimental data and the model predicted values was expressed by the correlation coefficient (R2 , values close or equal 1). When the value of R2 is close to 1, it does not mean that the fit is necessarily good [42]. Therefore, the conformity between experimental data and the model predicted values was expressed by the total mean error (ε%), which is the discrepancy between the experimental data and the predicted values [43]:

n

ε% =

i=1

|qe(exper.) − qe(calc.) |

n

q i=1 e(exper.)

(8)

A relatively low (ε%) value indicates which model can be successfully used to describe the sorption equilibrium on BCD. 3.4.1. Langmuir isotherm The Langmuir isotherm was applied for the sorption equilibrium on BCD according to the equation: Ce 1 Ce = + qe Qe bQe

(9)

where Ce is the equilibrium concentration of REE in solution (mg/L), qe is the amount of solute sorbed per unit mass of BCD at equilibrium (mg/g) and Qe (mg/g) and b (L/mg) are the Langmuir constants related to monolayer sorption capacity and free energy of sorption, respectively. From Table 5, it can be concluded that Qe parameter for Sm3+ was 8.32 meq/g, which was higher than the CEC of BCD. These data reflects again the mixed mechanism of sorption on BCD of ion exchange process and hydrolysis precipitation as well. According to the Langmuir model, sorption occurs uniformly on the active sites of the sorbent and once a sorbate occupies a site, no further sorption can take place at this site. A plot of Ce /qe versus Ce would result in a straight line with a slope of (1/Qe ) and intercept of 1/bQe as shown in Fig. 5. The values of the slopes, intercepts of the plots and total mean errors are presented in Table 5. The values of total mean error (ε%) between this model and the experimental data for Sm3+ , Gd3+ and Nd3+ were 3.76, 1.38 and 6.31, respectively. 3.4.2. Freundlich isotherm The simple Freundlich isotherm is often used for heterogeneous surface energy systems according to the equation: log qe = log k +

1 log Ce n

(12)

where qe is the amount of solute sorbed per unit mass of BCD at equilibrium (mg/g) and Ce is the equilibrium concentration of REE

66

O.I.M. Ali et al. / Journal of Hazardous Materials 195 (2011) 62–67

Table 3 The calculated parameters of the pseudo-second order kinetics model for sorption of Gd3+ and Sm3+ on BCD at different temperatures. Gd3+

Temperature (K)

Sm3+ −3

k2 (×10 303 313 323 333

2

g/mg min)

R

3.22 5.51 7.04 8.99

Ea (kJ/mol)

0.9927 0.9843 0.9813 0.9908

27.96

k2 (×10−3 g/mg min)

R2

10.23 11.31 12.04 13.98

0.9987 0.9928 0.999 0.9941

Ea (kJ/mol)

9.97

Table 4 Thermodynamic parameters for sorption of Y3+ , Nd3+ , Gd3+ , Sm3+ and Lu3+ on BCD. H◦ (kJ/mol)

REE

3+

3+

3+

Y , Gd , Nd Sm3+ , Lu3+

S◦ (kJ/mol K)

50.72 19.99

G◦ (kJ/mol)

0.250 0.124

303 K

313 K

323 K

−24.90 −17.60

−27.03 −18.82

−30.03 −20.06

333 K −32.53 −21.30

Table 5 Isotherms constants and values of correlation factors (R2 ) for sorption of Sm3+ , Nd3+ and Gd3+ on BCD. REE

Freundlich isotherm

3+

Sm Nd3+ Gd3+

Langmuir isotherm

1/n

log k

R2

ε%

Qe (meq/g)

b

R2

RL

ε%

0.599 0.220 0.260

1.828 2.130 1.860

0.9388 0.9722 0.9541

2.83 0.50 0.68

8.32 4.44 5.87

0.123 0.089 0.096

0.9835 0.9892 0.9945

0.010 0.0270 0.075

3.76 2.31 1.38

Fig. 6. Freundlich sorption isotherm of REEs sorption on BCD.

in solution (mg/L). K and n are constants characteristics of the system. Log k and 1/n are the Freundlich constants related to sorption capacity and sorption intensity of the sorbent, respectively. A plot of log qe as a function of log Ce would result in a straight line with a slope of (1/n) and intercept of log k as shown in Fig. 6. The values of 1/n, log k and ε% are presented in Table 5. The values of 1/n < 1 correspond to a heterogeneous surface with an exponential distribution of energy of the sorption sites [44]. The total mean error (ε%) between this model and the experimental data represents the best fit of experimental data than Langmuir one. The fit of the data to Freundlich isotherm indicates that the sorption process is not restricted to one specific class of sites and assumed surface heterogeneity [45]. This trend is due to the high surface area of the sorbent and multilayer of sorption on the BCD. This trend was also investigated by M. Al-Meshragi et al. [46] for sorption of Cr(III) on BCD. 4. Conclusions • The efficiency of BCD for the sorption of Y3+ , Nd3+ , Gd3+ , Sm3+ and Lu3+ from their aqueous solutions was investigated. It was

found that maximum sorption capacity was achieved at pH 7 using thiel buffer and the sorption capacity of BCD has the order of Lu3+ > Sm3+ > Y3+ > Gd3+ ≈ Nd3+ . The sorption of these elements on BCD was found to follow pseudo-second-order kinetics. • The thermodynamic parameters H◦ , S◦ and G◦ values of the REEs sorption onto BCD show endothermic heat of sorption, favored at high temperatures. The positive entropy value is an indication of the probability of favorable nature of sorption and the process is spontaneous. • The equilibrium data have been analyzed using Langmuir and Freundlich isotherms. The Freundlich isotherm was demonstrated to provide the best correlation and the lowest total error for sorption of the studied elements on BCD. • BCD may be successfully used as effective, low cost and abundant source for the removal of Y3+ , Nd3+ , Gd3+ , Sm3+ and Lu3+ from their aqueous solutions and may be used as alternative to more costly materials. References [1] J. Kalinowski, J. Mezyk, F. Meinardi, R. Tubino, M. Cocchi, D. Virgili, Electric field and charge induced quenching of luminescence in electroluminescent emitters based on lanthanide complexes, Chem. Phys. Lett. 453 (2008) 82–86. [2] G.A. Molander, J.A.C. Romero, Lanthanocene catalysts in selective organic synthesis, Chem. Rev. 102 (2002) 2161–2185. [3] K. Kuriki, Y. Koike, Y. Okamoto, Plastic optical fiber lasers and amplifiers containing lanthanide complexes, Chem. Rev. 102 (2002) 2347–2356. [4] K.H. Johannesson, K.J. Stetzenbach, V.F. Hodge, Rare earth elements as geochemical tracers of regional groundwater mixing, Geochim. Cosmochim. Acta 61 (1997) 3605–3618. [5] W. Bremmer, Rare earth applications in Chinese agriculture elements, Rare Earths Spec. Metals Appl. Technol. 3 (1994) 20–24. [6] S.-L. Tong, W.-Z. Zhu, Z.-H. Gao, Y.-X. Meng, R.-L. Peng, G.-C. Lu, Distribution characteristics of rare earth elements in children’s scalp hair from a rare earths mining area in southern China, J. Environ. Sci. Health A: Toxic/Hazard. Subst. Environ. Eng. 39 (2004) 2517–2532. [7] L. Weltje, H. Heidenreich, W.Z. Zhu, H.T. Wolterbeek, S. Korhammer, J.J.M. de Goeij, B. Markert, Lanthanide concentrations in freshwater plants and molluscs, related to those in surface water, pore water and sediment. A case study in The Netherlands, Sci. Total Environ. 286 (2002) 191–214. [8] A. Pałasz, P. Czekaj, Toxicological cytophysiological aspects of lanthanides action, Acta Biochim. Pol. 47 (2000) 1107–1114. [9] F. Wiberg, N. Wiberg, A.F. Hollemann, Inorganic Chemistry, Academic Press, San Diego, CA.

O.I.M. Ali et al. / Journal of Hazardous Materials 195 (2011) 62–67 [10] M. Atanassovaa, Synergistic solvent extraction and separation of lanthanide (III) ions with 4-benzoyl-3-phenyl-5-isoxazolone and the quaternary ammonium salt, Solvent Extr. Ion Exch. 27 (2009) 159–171. [11] S.D. Faust, O.M. Aly, Chemistry of Water Treatment, Butter Worth Publishers, NY, USA, 1976. [12] H.A. Omar, H. Moloukhia, Use of activated carbon in removal of some radioisotopes from their waste solutions, J. Hazard. Mater. 157 (2008) 242–246. [13] M. Grover, M.S. Narayanswamy, Removal of hexavalent chromium by adsorption on fly ash, Institute of Engneers Indian, J. Environ. Eng. 63 (1982) 36–39. [14] N. Kannan, A. Vanangamudi, A study on removal of chromium Cr(VI) by adsorption lignite coal, Ind. J. Environ. Pollut. 11 (1991) 241–245. [15] S. Chand, V.K. Agarwal, C. Pavankumar, Removal of hexavalent chromium from wastewater by adsorption, Ind. J. Environ. Health 36 (1994) 151–158. [16] A. Sen, A. De, Adsorption of mercury (II) by coal fly ash, Water Res. 21 (1987) 885–888. [17] K. Srinivasan, N. Balasubramaniam, T.V. Ramakrishna, Studies on chromium removal by rice husk carbon, Ind. J. Environ. Health 30 (1988) 376–387. [18] P.A. Brown, S.J. Allen, Metal removal from wastewater using peat, Water Res. 34 (16) (2000) 3907–3916. [19] D. Humelnicu, L. Bulgariu, M. Macoveanu, On the retention of uranyl and thorium ions from radioactive solution on peat moss, J. Hazard. Mater. 174 (2010) 782–787. [20] V.K. Gupta, M. Gupta, S. Sharma, Process development for the removal of lead and chromium from aqueous solutions using red mud—an aluminum industry waste, Water Res. 35 (2001) 1125–1134. [21] A.H. Mahvi, D. Naghipour, F. Vaezi, S. Nazmara, Teawaste as an adsorbent for heavy metal removal from industrial wastewaters, Am. J. Appl. Sci. 2 (1) (2005) 372–375. [22] C. Kütahyali, M. Eral, Sorption studies of uranium and thorium on activated carbon prepared from olive stones: kinetics and thermodynamics aspects, J. Nucl. Mater. 396 (2010) 251–256. [23] E.M. Saad, R.A. Mansour, A. El-Asmy, M.S. El-Shahawi, Sorption profile and chromatographic separation of uranium (VI) ions from aqueous solutions onto date pits solid sorbent, Talanta 76 (2008) 1041–1046. [24] A. Mackie, S. Boilard, M.E. Walsh, C.B. Lake, Physicochemical characterization of cement kiln dust for potential reuse in acidic wastewater treatment, J. Hazard. Mater. 173 (2010) 283–291. [25] H.Y. Ahmed, A.M. Othman, A.A. Mahmoud, Effect of using waste cement dust as a mineral filter on the mechanical properties of hot asphalt, Ass. Univ. Bull. Environ. Res. 9 (2006) 51–59. [26] E.K. Mahmoued, Cement kiln dust and coal filters treatment of textile industrial effluents, Desalination 255 (2009) 175–178. [27] N.G. Zaki, I.A. Khattab, N.M. Abd El-Monem, Removal of some heavy metals by BCD leachate, J. Hazard. Mater. 147 (2007) 21–27. [28] M.H. El-Awady, T.M. Sami, Removal of heavy metals by cement kiln dust, Bull. Environ. Contam. Toxicol. 59 (1997) 603–610. [29] O.M. Ali, H.H. El-Kashif, S.A. Sayed, M.E.H. Shalabi, The removal of uranium and thorium from aqueous solutions onto by-pass cement dust (BCD), Eurasian Chem. Tech. J. 11 (2009) 105–113.

67

[30] O.A. Fadali, E.E. Ebrahiem, Y.H. Magdy, A.A.M. Daifullah, M.M. Nassar, Removal of chromium from tannery effluents by adsorption, J. Environ. Sci. Health A: Environ. Sci. Eng. 9 (2005) 465–472. [31] B. Hegazy, M.A. El-Khateeb, A.A. El-adly, M.M. Kamel, Low-cost wastewater treatment technology, J. Appl. Sci. 7 (2007) 815–819. [32] M. Horshall, A.I. Spiff, Studies on the effect of pH on the sorption of Pb2+ and Cd2+ ions from aqueous solutions by Caladium bicolor (Wild Cocoyam) biomass, J. Biotechnol. 7 (3) (2004) 313–323. [33] H.T.S. Britton, Hydrogen Ions, fourth ed., Chapman and Hall, London, 1952, pp. 354–358. [34] Z. Marczenko, Spectrophotometric Determination of Elements, third ed., Ellis Horwood Chichester, UK, 1986. [35] Y.S. Ho, G. Mckay, Pseudo-second order model for sorption processes, Water Res. 34 (5) (1999) 578–584. [36] M.S. El-Shahawi, M.A. Othman, M.A. Abdel-Fadeel, Kinetics, thermodynamic and chromotographic behavior of the uranyl ions sorption from aqueous thiocyanate media onto polyurethane foams, Anal. Chim. Acta 546 (2005) 221–228. [37] K.E. Laintz, J. Yu, C.M. Wai, Separation of lanthanides with sodium bis-(trifluoroethyl) dithiocarbamate chelation and supercritical fluid chromatography, Anal. Chem. 64 (3) (1992) 311–315. [38] E. Bentouhami, G.M. Bouet, J. Meullemeestre, F. Vierling, M.A. Khan, Physicochemical study of the hydrolysis of rare-earth elements (III) and thorium (IV), Comptes Rendus Chimie 7 (5) (2004) 537–545. [39] E.A. El-Sofany, Sorption of Cd(II) and Se(IV) from aqueous solution using modified rice husk, J. Hazard. Mater. 147 (2007) 546–555. [40] N.M. Kozhevnikova, E.P. Ermakova, A study of sorption of samarium(III) ions by natural clinoptilolite-containing tuff, Russ. J. Appl. Chem. 81 (2008) 2095–2098. [41] A. Mellah, S. Chegrouche, M. Barkat, The removal of uranium (VI) from aqueous solution onto activated carbon: kinetic and thermodynamic investigation, Colloid Interface Sci. 169 (2006) 146–152. [42] S.C. Chapra, R.P. Canale, Numerical Methods for Engineers, third ed., McGrawHill Co., Singapore, 1998. [43] W.H. Press, B.P. Flannery, S.A. Teukolsky, W.T. Vetterling, Numerical Recipes in Pascal: The Art of Scientific Computing, Cambridge University Press, Cambridge, 1989. [44] M.M. Saeed, Adsorption profile and thermodynamic parameters of the preconcentration of Eu(III) on 2-thenoyltrifluoroacetone loaded polyurethane (PUR) foam, J. Radioanal. Nucl. Chem. 256 (2003) 73–80. [45] E.A. El-Sofany, Removal of lanthanum and gadolinium from nitrate medium using Aliquat-336 Impregnated Onto Amberlite XAD-4, J. Hazard. Mater. 153 (2008) 948–949. [46] M. Al-Meshragi, H.G. Ibrahim, M.M. Aboabboud, Equilibrium and kinetics of chromium adsorption on cement kiln dust, in: Proceedings of the World Congress on Engineering and Computer Science, San Francisco, USA, October 22–24, 2008.

Journal of Hazardous Materials 195 (2011) 68–72

Contents lists available at ScienceDirect

Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat

Silver nanotoxicity using a light-emitting biosensor Pseudomonas putida isolated from a wastewater treatment plant R.I. Dams a,∗ , A. Biswas a , A. Olesiejuk a , T. Fernandes b , N. Christofi a a b

Centre for Nano Safety, Edinburgh Napier University, Scotland, UK Nano Safety Research Group, Heriot-Watt University, Edinburgh EH14 4AS, Scotland, UK

a r t i c l e

i n f o

Article history: Received 8 November 2010 Received in revised form 29 July 2011 Accepted 4 August 2011 Available online 10 August 2011 Keywords: Pseudomonas putida Silver nanoparticles Nanotoxicity

a b s t r a c t The effect of silver ions, nano- and micro-particles on a luminescent biosensor bacterium Pseudomonas putida originally isolated from activated sludge was assessed. The bacterium carrying a stable chromosomal copy of the lux operon (luxCDABE) was able to detect toxicity of ionic and particulate silver over short term incubations ranging from 30 to 240 min. The IC50 values obtained at different time intervals showed that highest toxicity (lowest IC50 ) was obtained after 90 min incubation for all toxicants and this is considered the optimum incubation for testing. The data show that ionic silver is the most toxic followed by nanosilver particles with microsilver particles being least toxic. Release of nanomaterials is likely to have an effect on the activated sludge process as indicated by the study using a common sludge bacterium involved in biodegradation of organic wastes. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Engineered nanomaterials, that have seen significant increases in use recently, are mainly composed of metallic nanoparticles. The knowledge of the ecotoxicology of nanoparticles (NP) to bacteria and other microbes is still limited, even though some manufactured nanoparticles (NP; materials with three dimensions between 1 and 100 nm) [1] such as silver and titanium, which are constantly released into the environment, are known as antibacterial agents. As such it is essential that technology that fully assesses their effects on natural microbial organisms, and on biogeochemical cycling in the environment, is available. Clearly there is a concern that these novel materials could be released into the environment. Whole microbial cell biosensors are now widely used as research tools in the testing of substances likely to elicit cytotoxic and genotoxic events, and, in the determination of bioavailability of chemicals [2]. They embrace genetically engineered bacteria that have a toxicant detecting gene that is coupled with a reporter gene (e.g. luminescence gene such as lux or luc) capable of producing a detectable response on activation. Wiles et al. [3] argue that autochthonous microorganisms would be appropriate in toxicity testing with the potential for in situ relevance. Silver, a metal used extensively in various consumer products because of its effective antimicrobial properties, is subject to release to sewer the sewerage system. Therefore it is important

∗ Corresponding author. Tel.: +44 0131 455 2291; fax: +44 0131 455 2291. E-mail address: [email protected] (R.I. Dams). 0304-3894/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jhazmat.2011.08.013

to determine whether has an effect on activated sludge microorganisms. Silver nanoparticles are well known for their antibacterial activity [4–7] and have become one of the nanomaterials mostly used in consumer product [8]. Being considered as the most prevalent of engineered materials [9] is likely to enter the wastewater treatment plants as nanowaste. Blaser et al. [10] pointed out that silver released to wastewater is incorporated into sewage sludge and may spread further on agricultural fields where will mainly stay in the top layer of soils [11]. This work aims to assess the effect of silver nanoparticles on the activated sludge process by examining the response of a common bacterial member of the sludge consortium involved in sewage organic degradation. In this study ecotoxicological was performed using a genetically-modified Pseudomonas isolated from a polluted, phenolic-rich, wastewater treatment system by transposon mutagenesis [3]. The Pseudomonas carries a stable chromosomal copy of the lux operon (luxCDABE) derived from Photorhabdus luminescens with continuous output of light. The bioluminescence bioassay performed in this study has the advantage of allowing monitoring the presence of these nanoparticles, specially silver nanoparticles, the object of this study. 2. Materials and methods 2.1. Silver chemicals All of the chemicals used were analytical grade. Silver powder nanoparticles (average size: 35 nm, 99.5% metal basis, spherical morphology and cubic crystallographic structure) and

R.I. Dams et al. / Journal of Hazardous Materials 195 (2011) 68–72

69

microparticles (average size: 0.6–1.6 ␮m) were obtained from Nanoamor (Nanostructured and Amorphous Materials Inc., TX, USA). Stock suspensions of silver ion (AgNO3 ), silver nanoparticles (Ag-NP) and silver microparticles (Ag-MP) were freshly made in 20 ml universal flasks and placed into an ultrasonic bath (XB6 Grant Instruments Cambridge Ltd., UK) at 25 kHz, 25 ◦ C for 30 min. Two fold dilutions were externally prepared of these suspensions and added to the 96 well black microtitre plates (Sterling, Caerphilly, UK) to give final toxicant concentration in the range to be tested. Two different stabilisers were added to the nanoparticles and microparticles working solutions: 0.1% citric acid, 0.1% BSA (bovine serum albumin) and comparisons were also made with preparations without stabilisers. Previous studies have used BSA for stabilisation of ZnO nanoparticles [12] and of carbon black nanoparticles [13]. Thus improving the stability of particle suspension reducing particle agglomeration and settling over time [13]. Citric acid is known to act as a chelating agent and may thus be able to also act to quench toxicity of dissolved metal toxicants. All working solutions were light protected and used within 30 min of preparation. In order to assess any shading effects of the particles on the bacterial cells, the light output was measured before and after addition of particles. No shading effects were determined.

For error analysis, all of the experiments were conducted 3 times on different plates. Data from eight wells were used for one concentration and coefficients of variation (CV) between independent assays were calculated using Microsoft Excel 97. Differences among treatments were tested using a two-way analysis of variance (ANOVA) to determine which treatments were statistically different (P < 0.05).

2.2. Media and growth conditions

3. Results and discussion

The bacterial strain used in this study, Pseudomonas putida BS566::luxCDABE was constructed based upon chromosomal expression of the luxCDABE operon derived from an entomopathogenic nematode symbiont, P. luminescens [14]. Originally isolated from the treatment system [15], this reporter organism encompassing a dynamic xenobiotic sensing range, is suitable for placement around an industrial processing system to monitor remediation in multiple compartments. Cultures were grown in Luria–Bertani (LB) broth (10.0 g L−1 of Tryptone; 5.0 g L−1 of yeast extract; 5.0 g L−1 of sodium chloride), containing 100 mg L−1 Kanamycin, and overnight at 26 ◦ C in shaking conditions (200 rpm) to late exponential stage. Prior to the assay, cultures were diluted to approximately 107 cells/ml and regrown under the same conditions for two to three generations without Kanamycin. When OD (optical density) reached 0.2 (approximately 108 cells/ml) toxicity tests were carried out.

IC50 values following challenge with AgNO3 , Ag-NP and Ag-MP (with and without dispersant) are shown in Table 1. The experimental uncertainty of these bioluminescence bioassays is within the coefficients of variation. Calculated coefficients of variation (CV) between independent assays were found to be between 1 and 15%. Fig. 1 shows the light output reduction by P. putida BS566::luxCDABE when challenged with different concentrations up to 2500 ␮g L−1 silver ion. Among the silver species tested, AgNO3 is by far the most toxic to P. putida highlighting the action of ionic Ag+ (P < 0.05). It is well known that silver ion and silver based compounds are highly toxic to micro organisms and have strong biocide effects to many bacteria species [17–20]. Figs. 2 and 3 show the light output reduction by P. putida BS566::luxCDABE when challenged with Ag nanoparticles and Ag microparticles. In the presence of BSA as stabilizer, Ag nanoparticles showed to be statistically more toxic than Ag microparticles (P < 0.05). The same was observed when using citric acid as stabilizer. The highest toxicity (Table 1) was observed after 90 min incubation and indeed this was the case for all silver species tested with or without stabilisation. The order in toxicity was Ag+ > AgNP (35 nm) > Ag-MP (0.6–1.6 ␮m). In this current study we tested three different silver species, and ionic Ag+ being the most toxic to the biosensor tested. Similar results were found by Choi et al. [4] when testing the toxicity of silver species. In their studies, the

2.3. Biosensor assay Luminescence measurements were undertaken using a 96well plate luminometer (FLUOstar Optima, BMG Labtech, UK) in 96 well black microtitre plates (Sterling, Caerphilly, UK) whereby each well contained bacterial inoculum and toxicant at the required concentration in 100 ␮L volumes, using an integration time of 1 s at a temperature of 28 ◦ C. Readings were taken every 30 min for 240 min. Control wells containing LB broth with P. putida BS566::luxCDABE were run and changes in toxicity for the test systems expressed as percentages of the control. Luminescence values were expressed in the instrument’s arbitrary relative light units (RLU). The maximal response ratios were the highest ratios of luminescence in the samplecontaining wells to luminescence in wells containing untreated cells determined during a specified period: 30, 90, 180 and 240 min [16]. Inhibitory concentration which represents 50% inhibition of light output (IC50 ) in relation to the control was assessed for all toxicants tested at each time point. All experiments were run at least 3 times (most were ran 4 times) at different dates with different batches.

2.4. Calculation of IC50 (IC, inhibitory concentration) values The IC values were calculated using a statistical program developed in-house. The program fits a three parameter logistic model to the logarithm of the concentration by weighted least squares. The parameters are the initial response, the slope and the intercept. It is assumed that the response would decline to zero at sufficiently high concentrations. The initial response effectively uses the information from both the controls (if present) and low concentrations. The weights used are taken to be proportional to the fitted response but with adjustments for high and low responses; this is to protect against bias due to “hormesis” effects (stimulatory effects causing increased light output when challenged with low toxicant concentrations) and the effective omission of data respectively. 2.5. Data analysis

Table 1 IC50 (mg L−1 ) values for light emission reduction by P. putida BS566::luxCDABE after 30, 90, 180 and 240 min. Values presented are an average of at least 3 independent experiments carried out with different batches, standard deviation between 1 and 15%. IC50 values (mg L−1 ) Time (min)

30

90

180

240

AgNO3 Ag-NP Ag-NP BSA Ag-NP CA Ag MP Ag MP BSA Ag MP CA

0.44 88 102 147 715 375 700

0.18 81 35 126 530 256 240

0.25 91.5 43 136 765 308 300

0.30 184 50 149 1075 330 337

70

R.I. Dams et al. / Journal of Hazardous Materials 195 (2011) 68–72

Fig. 1. Light output reduction by P. putida BS566::luxCDABE when challenged with 2500 ␮g L−1 silver ion after 30, 90, 80 and 240 min. Mean of three replicates with different batches.

ionic form Ag+ was more toxic to heterotrophic Escherichia coli than 16 nm silver nanoparticles, while to autotrophic nitrifying bacteria Ag nanoparticles were more toxic than the ionic form. It has been pointed out that the bactericidal effect of nanoparticles is dependent on the concentration of nanoparticles and the initial bacterial concentration [19]. In our study we used 35 nm silver nanoparticles which were dispersed in liquid cultures with an initial bacterial concentration of 108 UFC cells/ml for the toxi-

Fig. 3. Light output reduction by P. putida BS566::luxCDABE when challenged with 1500 mg L−1 Ag micro particles: (a) without stabilizer, (b) with 0.1% citric acid, and (c) with BSA after 30, 90, 80 and 240 min. Mean of three replicates with different batches.

Fig. 2. Light output reduction by P. putida BS566::luxCDABE when challenged with 200 mg L−1 Ag nanoparticles: (a) without stabilizer, (b) with 0.1% citric acid, and (c) with BSA after 30, 90, 80 and 240 min. Mean of three replicates with different batches.

city tests. Using E. coli as a model, Sondi and Sondi-Salopeki [18] showed that the bacterial growth in LB liquid medium (with an initial bacterial concentration of 107 UFC cells/ml) was delayed by nanoparticles size 12 nm. However, other studies using the Microtox system [21] for bioluminescence testing have shown no toxicity on exposures to silver nanoparticles. In their study, however, these authors used a marine microrganism Photobacterium phosphoreum which requires a sodium chloride concentration of 22% and is not indicated for ecotoxicity of terrestrial systems [21]. Whole cell biosensors are of greater environmental relevance for luminescence-based testing in terrestrial systems compared to the Microtox system testing. Furthermore, whole cell biosensor have shown a greater sensitivity than Microtox system [16]. In this current study we used a whole cell biosensor P. putida BS566::luxCDABE, a terrestrial bacterium for the bioluminescence testing. This methodology has been used in many studies formerly [2,15,16,22–25]. Beaton et al. [22] showed that the biosensor E. coli HB101 pUCD607 is a sensitive indicator of changes in toxicity in a soil system spiked with 2,4-dichlorophenol; Shaw et al. [23] used the biosensor lux marked Burkholderia RASC c2 in bioluminesce inhibition studies and Boyd et al. [24] used Burkholderia RASC c2 and Pseudomonas fluorescens 10586 to assess the toxicity of chlorophenols. Sinclair et al. [25] showed the toxic response of the lux-marked biosensors as Pseudomonas fluorencens and E. coli to 2,4-dichlorophenol.

R.I. Dams et al. / Journal of Hazardous Materials 195 (2011) 68–72

Our biosensor responded within 90 min (as see in Table 1) to the presence of toxicants. Thus, this strain can be used for the rapid and sensitive detection of potentially toxic silver compounds. Overall, the toxicity of silver was found not to be dependent. Overtime, the IC50 values obtained over 90-min for ionic silver, Ag-NP and Ag-MP was about 1.5–2.0 times lower than the 240-min test. Therefore, according to our findings, a 90-min test should be taken when monitoring and evaluating wastewater treatment plants for silver toxicity. If we consider 90 min assay results, which shows the highest toxicity, to compare the data, it can be noticed that Ag-NP are nearly 200 times less toxic that Ag+ , with Ag-MP ∼3000 times less toxic in non-stabilised systems (P < 0.05). The effect with Ag-NP was more pronounced in BSA systems with a calculated IC50 value ∼4 times lower than Ag-NP with citric acid (P < 0.05). Addition of BSA seems to result in higher toxicity perhaps through better dispersion of the NP providing more surface area for Ag to have an effect. This however may not be the case as citric acid may be just as effective in preventing agglomeration. Citric acid is an effective chelating/complexing agent for metals in solution. The effect of citric acid on the toxicity value might be due to immobilization of Ag+ released in solution or those on surfaces. On the one hand one may need to show a worst case toxic effect of well dispersed NP systems but on the other hand real systems will have a range of cheating/complexing agents that will affect the toxicity of substances released into its environment. Dissolution is likely to be a critical step for some metallic nanoparticles in determining fate in the environment and within the organism. In this study when particles were well dispersed using BSA as stabilizer, a higher toxicity was observed. Solubility strongly influences the toxicity and when no stabilizer was used the IC50 values were higher, probably indicating that less soluble compounds were available for the bacterial cells. Brunner et al. [26] observed that nanoparticles with a low solubility such as TiO2 showed no toxicity to mammalian cells while more soluble nanoparticles like ZnO showed a higher toxicity. The use of different stabilisers and their effect on toxicity values must be assessed particularly if laboratory bioassay results are to be used to derive wastewater discharge consents. There were no significant differences between the IC50 values obtained for Ag-MP with BSA and citric acid stabilisation (P > 0.05) and both values were ∼2 times statistically lower than Ag-MP tested without stabilisation (P < 0.05). Particle size does cause a toxicity difference. In the present study, if we compare the toxicity between nanoparticles and their micro size counterparts, we noticed that when Ag-NP were stabilized showed a higher toxicity when compared with the micro size counterparts (P < 0.05). Other studies have also showed no toxicity of micro scaled particles when compared with nano size counterparts. For instance, Jiang et al. [20] observed a higher toxicity of nanoparticles of Al2 O3 , TiO2 and ZnO than their micro size scale counterparts which showed no or lower toxicity. Sinha et al. [27] have noticed that ZnO nanoparticles disintegrate Gram negative bacteria cell membrane and accumulate in cytoplasm, while when these cells were grown in micro particle counterparts the cell membrane and cytoplasm were intact. In this study we used 35 nm spherical nanosilver particles. Smaller nanosilver particles are more active than larger ones because of their higher surface area. However, in this study we used relatively large size of silver nanoparticles (35 nm) which proved to have a higher level of toxicity against the biosensor tested. Silver nanoparticles surface area plays quite an important role for antibacterial activity which depends on its exposed surface area concentration. This dependency is originated from the released Ag+ from the nanosilver surface. Recent studies [28,29] indicated that when nanosilver particles are small and release many Ag+ ions, the antibacterial activity is dominated by these

71

ions rather than the nanosilver particles. However, when larger sizes are used (as the ones used in this study) which have a low release of Ag+ ions, the nanosilver particles themselves also influence the antibacterial activity as indicated by these results here presented. Metal nanoparticles as silver have a tendency to attach on the cell wall [20]. So, toxicity is not dependent only on release of Ag+ , but also it depends on other factors such as the attachment of particles on the cell surface, disruption of cell membrane and consequently accumulation of nanoparticles in the cytoplasm. Surprisingly, the counterpart micro sized particles showed a lower toxicity if we consider that smaller particles release more Ag+ ions that the larger ones. However, one should consider other factors that affect toxicity such as the initial concentration of particles and bacteria, and particle bioavailability, among many other factors. So, although silver NPs have a higher surface area and probably a higher release of Ag+ , there was probably not enough contact in order to cause damage to cell membrane and cytoplasm since toxicity is dependent on contact and/or bacterial attachment to the particle as pointed by Jiang et al. [20] and Sinha et al. [27]. Furthermore, it has been demonstrated that small particles as the counterpart microsized are in suspension only in small numbers and are not able to attach to the bacterial surface as the nanoparticles do [20]. Then, in this case of the micro sized particles as the ones used in this study, they caused less harm to the bacterial cells than the nanoparticles ones due to their inability to attach on cell surface. The antimicrobial effect is related to the amount and the rate of silver released by NP. Severe structural changes occur in the bacterial cell wall when ionised silver binds to cell membrane proteins which leads to protein distortion and cell death [30,31]. Silver is classified as the “soft” metal group [10] and it complexes with many organic or inorganic materials such as chloride, sulphide, thiosulphate [32]. In order to evaluate the impact of silver discharge in the environment it is important to understand the fate and transport of silver in wastewater treatment plants. The applicability of our sensor in wastewater treatment plants has been previously demonstrated [3,33]. The biosensor here tested, P. putida BS566::luxCDABE had accurately predicted toxicity shifts in wastewater treatment plants, with a high tolerance to a phenolic cocktail, thus demonstrating an effective biosensing in all treatment compartments [3]. Philp et al. [33] have tested immobilised P. putida BS566::luxCDABE in phenolic wastewater treatment plant. The biosensor tested proved to be able to discriminate toxicity of various zones within the waste treatment plant [33]. Further studies in our laboratories using wastewater samples affected by silver toxicity have being carried out using this biosensor and results will be published somewhere else. Nanowaste is likely to increase and therefore enter the wastewater treatment plants which are the final step to control silver discharge. Estimation of silver load in sewage sludge and its microorganisms growth inhibition has been predicted. Blaser et al. [10] predicted that an expected silver concentration in sewage treatment plant range from 2 ␮g L−1 to 18 ␮g L−1 . Shafer et al. [34] reported a range of ∼2–4 ␮g L−1 of silver in sewage treatment plants treating common wastewater and a much higher load from industrial discharges (from 24 to 105 ␮g L−1 ). The removal of silver ion by chloride free sludge is dependent on the silver-sludge loading, the solution pH and the concentration of dissolved organic matter. Studying the interactions of silver with wastewater constituents, Wang et al. [35] showed that silver ion can be removed through chloride precipitation and sludge adsorption. However, the authors [35] pointed out that the formation of silver-ion-dissolved organic matter complexes, which is increased in alkaline conditions, reduces the silver ion adsorption by sludge.

72

R.I. Dams et al. / Journal of Hazardous Materials 195 (2011) 68–72

4. Conclusions In conclusion, our results demonstrate that the use of a bacterial biosensor as P. putida BS566::luxCDABE provides a robust, early warning system of acute toxicity which could lead to process failure. This strain is suitable for toxicity monitoring in a highly polluted industrial wastewater treatment streams. The information regarding the inhibition of microbial growth by different Ag compounds, especially in wastewater treatment systems, is valuable for operating planning and control. The presence and activity of microorganisms in biological wastewater treatment are vital to the process. One of the projected applications of such strains is its combined use as analytical panel for toxicant detection. An important advantage of using these organisms is that a positive response will not only indicate the presence of a toxicant but will also provide some idea as to its character. Acknowledgements This work was part of a project funded by the Brazilian National Research and Development Council (CNPq). Also we would like to acknowledge the Environmental Microbiology Research Laboratory at Edinburgh Napier University, Edinburgh, Scotland. References [1] J. Bridges, W. Jong, T. Jung, D. Williams, T.F. Fernandes, J.-P. Marty, T. Butz, Opinion on the scientific aspects of the existing and proposed definitions relating to products of nanoscience and nanotechnologies, SCENIHR (Scientific Committee on Emerging and Newly-Identified Health Risks), European Commission, Brussels, Belgium, November 2007, . [2] S. Belkin, Genetically engineered microorganisms for pollution monitoring, in: Soil and Water Pollution Monitoring, Protection and Remediation, NATO Science Series: IV: Earth and Environmental Sciences, vol. 69, 2007, pp. 147–160. [3] S. Wiles, A.S. Whiteley, J.C. Philp, M.J. Bailey, Development of bespoke bioluminescent reporters with the potential for in situ deployment within a phenolic-remediating wastewater treatment system, J. Microbiol. Methods 5 (2003) 667–677. [4] O. Choi, K.K. Deng, N-J. Kim, L. Ross, Z.R.Y. Surampalli, The inhibitory effects of silver nanoparticles, silver ions, and silver chloride colloids on microbial growth, Water Res. 42 (2008) 3066–3074. [5] J.P. Ruparelia, A.K. Chatterjee, S.P. Duttagupta, S. Mukherji, Strain specificity in antimicrobial activity of silver and copper nanoparticles strain specificity in antimicrobial activity of silver and copper nanoparticles, Acta Biomater. 4 (2008) 707–716. [6] Q. Chang, L. Yan, M. Chen, H. He, J. Qu, Bactericidal mechanism of Ag/Al2 O3 against Escherichia coli, Langmuir 23 (2007) 11197–11199. [7] S. Ghosh, R. Kaushik, K. Nagalakshmi, S.L. Hoti, G.A. Menezes, B.N. Harish, H.N. Vasan, Antimicrobial activity of highly stable silver nanoparticles embedded in agar–agar matrix as a thin film, Carbohydr. Res. 345 (2010) 2220–2227. [8] A.D. Maynard, E. Michelson, The Nanotechnology Consumer Product Inventory (2006), . [9] D. Rejeski, D. Lekas, Nanotechnology field observations: scouting the new industrial waste, J. Cleaner Prod. 16 (2008) 1014–1017. [10] S.A. Blaser, M. Scheringer, M. MacLeod, K. Hungerbuhler, Estimation of cumulative aquatic exposure and risk due to silver: contribution of nanofunctionalised plastics and textiles, Sci. Total Environ. 390 (2008) 396–409. [11] H. Hou, T. Takamatsu, M.K. Koshikawa, M. Hosomi, Migration of silver, indium, tin, antimony and bismuth and variations in their chemical fractions on addition to uncontaminated soils, Soil Sci. 170 (2005) 624–639.

[12] R. Brayner, R. Ferrari-Iliou, N. Brivois, S. Djediat, F. Marc, M.F. Benedetti, F. Fieˇıvet, Toxicological impact studies based on Escherichia coli bacteria in ultrafine ZnO nanoparticles colloidal medium, Nano Lett. 86 (2006) 866–870. [13] L. Foucaud, M.R. Wilson, D.M. Brown, V. Stone, Measurements of reactive species production by nanoparticles prepared in biologically relevant media, Toxicol. Lett. 174 (2007) 1–9. [14] M. Winson, S. Swift, P.J. Hill, C.M. Sims, G. Griesmayr, B.W. Bycroft, P. Williams, S.A.B. Gordon, G.S.A.B. Stewart, Engineering the luxCDABE genes from Photorhabdus luminescens to provide a bioluminescent reporter for constitutive and promoter probe plasmids and mini-Tn5 constructs, FEMS Microbiol. Lett. 163 (1998) 193–202. [15] A.S. Whiteley, S. Wiles, A.K. Lilley, J. Philp, M.J. Bailey, Ecological and physiological analyses of pseudomonad species within a phenol remediation system, J. Microbiol. Methods 44 (2001) 79–88. [16] S. Belkin, D.R. Smulski, A.C. Vollmer, T.K. Van Dyk, R.A. Larossa, Oxidative stress detection with Escherichia coli, Appl. Environ. Microbiol. 62 (1996) 2252–2256. [17] G. Zhao, E. Stevens, Multiple parameters for the comprehensive evaluation of the susceptibility of Escherichia coli to the silver ion, Biometals 11 (1998) 27–32. [18] I. Sondi, B. Salopek-Sondi, Silver nanoparticles as antimicrobial agent: a case study on E. coli as a model for Gram-negative bacteria, J. Colloid Interface Sci. 275 (2004) 177–182. [19] S. Pal, Y.K. Tak, J.M. Song, Does the antibacterial activity of silver nanoparticles depend upon the shape of the nanoparticle? A study of the Gram-negative bacterium Escherichia coli, Appl. Environ. Microbiol. 73 (2007) 1712–1720. [20] W. Jiang, H. Mashayekhi, B. Xing, Bacterial toxicity comparison between nanoand micro-scaled oxide particles, Environ. Pollut. 157 (2009) 1619–1625. [21] R. Barrena, E. Casals, J. Colón, X. Font, A. Sánchez, V. Puntes, Evaluation of the ecotoxicity of model nanoparticles, Chemosphere 75 (2009) 850–857. [22] Y. Beaton, L.J. Shaw, L.A. Glover, A. Mehard, K. Killkam, Biosensing 2,4dichlorophenol toxicity during biodegradation by Burkholderia sp. RASC c2 in soil, Environ. Sci. Technol. 33 (1999) 4086–4091. [23] L.J. Shaw, Y. Beaton, L.A. Glover, K. Killkam, Development and characterization of a lux-modified 2,4-dichlorophenol-degrading Burkholderia sp. RASC, Appl. Environ. Microbiol. 1 (1999) 393–399. [24] E.M. Boyd, K. Killkam, A.A. Mehard, Toxicity of mono-, di- and tri-chlorophenols to lux marked terrestrial bacteria Burkholderia species RASC c2 and Pseudomonas fluorescens, Chemosphere 43 (2001) 157–166. [25] G.M. Sinclair, G.I. Paton, A.A. Meharg, K. Killham, Lux-biosensor assessment of pH effects on microbial sorption and toxicity of chlorophenols. FEMS (Federation of European Microbiological Societies), Microbiol. Lett. 174 (1999) 273–278. [26] T.J. Brunner, P. Wick, P. Manser, P. Spohn, R.N. Grass, L.K. Limbach, A. Bruinink, W.J. Stark, In vitro cytotoxicity of oxide nanoparticles: comparison to asbestos, silica, and the effect of particle solubility, Environ. Sci. Technol. 40 (2006) 4374–4381. [27] R. Sinha, R. Karan, A. Sinha, S.K. Khare, Interaction and nanotoxic effect of ZnO and Ag nanoparticles on mesophilic and halophilic bacterial cells, Bioresour. Technol. (2010), doi:10.1016/j.biortech.2010.07.117. [28] G.A. Sotiriou, S.E. Pratsinis, Antibacterial activity of nanosilver ions and particles, Environ. Sci. Technol. 44 (2010) 5649–5654. [29] G.A. Sotiriou, A. Telekia, A. Camenzinda, F. Krumeicha, A. Meyerb, S. Pankeb, S.E. Pratsinis, Nanosilver on nanostructured silica: antibacterial activity and Ag surface area, Chem. Eng. J. (2011), doi:10.1016/j.cej.2011.01.099. [30] A.B.C. Landsdown, Silver I: its antibacterial properties and mechanism of action, J. Wound Care 11 (2002) 125–138. [31] J.J. Castellano, S.M. Shafii, F. Ko, G. Donate, T.E. Wright, R. Mannari, Comparative evaluation of silver-containing antimicrobial dressings and drugs, Int. Wound J. 4 (2007) 114–122. [32] H.T. Ratte, Bioaccumulation and toxicity of silver compounds: a review, Environ. Toxicol. Chem. 18 (1999) 89–108. [33] J.C. Philp, S. Balmand, E. Hajto, M. Bailey, S. Wiles, A. Whiteley, A.K. Lilley, J. Hajto, S.A. Dunbar,.Whole cell immobilised biosensors for toxicity assessment of a wastewater treatment plant treating phenolics-containing waste, Anal. Chim. Acta 487 (2003) 61–74. [34] M.M. Shafer, J.T. Overdier, D.H. Armstong, Removal portioning and fate of silver and other metals in wastewater treatment plants and effluent-receiving streams, Environ. Toxicol. Chem. 17 (4) (1998) 630–641. [35] J. Wang, C.P. Huan, D. Pirestan, Interactions of silver with wastewater constituents, Water Res. 37 (2003) 4444–4452.

Journal of Hazardous Materials 195 (2011) 73–81

Contents lists available at ScienceDirect

Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat

Utilization of different crown ethers impregnated polymeric resin for treatment of low level liquid radioactive waste by column chromatography M.F. Attallah a,∗ , E.H. Borai a , M.A. Hilal a , F.A. Shehata a , M.M. Abo-Aly b a b

Analytical Chemistry and Control Department, Hot Laboratories and Waste Management Center, Atomic Energy Authority, Post Code 13759, Abu Zaabal, Cairo, Egypt Chemistry Department, Faculty of Science, Ain Shams University, Cairo, Egypt

a r t i c l e

i n f o

Article history: Received 2 February 2011 Received in revised form 2 August 2011 Accepted 4 August 2011 Available online 25 August 2011 Keywords: Liquid radioactive waste treatment Impregnated polymeric resin Separation Radionuclides Crown ether derivatives

a b s t r a c t The main goal of this study was to find a novel impregnated resin as an alternative for the conventional resin (KY-2 and AN-31) used for low and intermediate level liquid radioactive waste treatment. Novel impregnated ion exchangers namely, poly (acrylamide-acrylic acid-acrylonitril)N,N’-methylenedi-acrylamide-4,4’(5 )di-t-butylbenzo 18 crown 6 [P(AM-AA-AN)-DAM/DtBB18C6], poly (acrylamide-acrylic acid-acrylonitril)-N,N’-methylenediacrylamide-dibenzo 18 crown 6 [P(AM-AA-AN)DAM/DB18C6], and poly (acrylamide-acrylic acid-acrylonitril)-N,N’-methylenediacrylamide-18 crown 6 [P(AM-AA-AN)-DAM/18C6] were prepared and their removal efficiency of some radionuclides was investigated. Preliminary batch experiments were performed in order to study the influence of the different derivatives of 18 crown 6 on the characteristic removal performance. Separation of 134 Cs, 60 Co, 65 Zn and (152+154) Eu radionuclides from low level liquid radioactive waste was investigated by using column chromatography with P(AM-AA-AN)-DAM/DtBB18C6 and metal salt solutions traced with the corresponding radionuclides. Breakthrough data was obtained in a fixed bed column at room temperature (298 K) using different bed heights and flow rates. The breakthrough capacities were found to be 94.7, 83.3, 58.7, 43.1 (mg/g) for 60 Co, 65 Zn, 134 Cs, and (152+154) Eu, respectively. Pre-concentration and separation of all radionuclides under study have been carried out using different concentration of nitric and/or oxalic acids. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Frequently used methods for treatment of liquid radioactive waste include chemical precipitation, evaporation, solvent extraction and ion exchange processes. Among the ion exchanger materials, impregnated polymeric resins have been used for the selective removal and separation of some radionuclides from radioactive liquid waste [1–6] as well as for the pre-concentration of metal species [7–10]. Crown ethers are effective extractants due to their ability to form stable complexes with metal ions. This property of crown ethers has led to the elaboration of new processes to extract radioactive elements from radioactive waste solutions [11–14]. Among the crown ethers, which are selective for alkali metal ions, derivatives of 21 crown 7 (21C7) have been extensively used for cesium extraction [15]. The key for the extraction is the good match between the cavity of the crown ether and the ionic radius of the metal ion. However, there are also several reports

∗ Corresponding author. Tel.: +20 2 446 20 806; fax: +20 2 446 20 784. E-mail addresses: [email protected], mohamedfathy [email protected] (M.F. Attallah). 0304-3894/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jhazmat.2011.08.007

involving the extraction of cesium by 18-membered crown ethers (18C6) [16–21]. The new cesium-selective macrocycle calix[4]arene-bis[4-(2ethylhexyl)benzo-crown-6] (“BEHBCalixC6”) has been studied by Engel et al. The other calixcrown extractant, calix[4]arene-bis[4tert-octyl-benzo-crown-6] (“BOBCalixC6”) was used to synthesis of this new extractant “BEHBCalixC6”. It was found that replacement of the tert-ocytl alkyl chains on the benzo-crown portion of the calixcrown by 2-ethylhexyl chains improves the equilibrium solubility of the free calixcrown in aliphatic diluents, while not affecting the cesium extraction strength [3]. Development of the chromatographic partitioning of cesium and strontium utilizing two impregnated polymeric composites was also studied by Zhang et al. A novel, specific macro porous silica-based 4,4 (5 )di-t-butylcylohexano 18 crown 6 (DtBuCH18C6) chelating polymeric material was synthesized by impregnating DtBuCH18C6 molecule into Si-polymer particles that was prepared by a series of polymerization reactions. They found that, DtBuCH18C6/Si-polymer is highly selective for Sr2+ , where as DtBuCH18C6 acts as chelating agent for Sr2+ [4]. In acidic HLW, Cs(I) has been separated by a novel silica-based polymeric adsorption material, Calix[4]arene-R14/SiO2 -P, which is an excellent

74

M.F. Attallah et al. / Journal of Hazardous Materials 195 (2011) 73–81

molecular recognition reagent for Cs(I) from Sr(II) and other fission products [5]. The present work was oriented to study the effect of different derivatives of 18 crown 6 based on poly (acrylamide-acrylic acid-acrylonitrile) resin as a novel impregnated polymeric material which could be used to remove some key radionuclides from low level liquid radioactive waste. 2. Experimental 2.1. Chemicals and reagents All chemicals and reagents used in this study were of analytical grade purity and were used without further purification. Cesium chloride and europium nitrate, were obtained from Prolabo (England). Acrylic acid, acrylonitrile purity (99%) and cobalt chloride were obtained from Merck (Germany). Acrylamide was supplied from BDH (England), 4,4 (5 )di-t-butylbenzo 18 crown 6 (DtBB18C6) were purchased from Fluka (Switzerland) while 18 Crown 6 (18C6) and dibenzo 18 Crown 6 (DB18C6) were supplied from Aldrich (USA). N,N’-methylenediacrylamide (DAM) was obtained from Aldrich (USA). Oxalic acid and sodium hydroxide were purchased from Adwic (Egypt). Nitric acid and zinc chloride were obtained from Winlab (England). A radioactive waste sample containing mixed radionuclides 134 Cs, 65 Zn, 60 Co and 152+154 Eu was collected from various laboratory research activities in Hot Laboratories Center, Egypt. 2.2. Synthesis of impregnated polymeric material (acrylamide-acrylic acid-acrylonitrile) N,N’Poly methylenediacrylamide P(AM-AA-AN)-DAM was prepared to use ␥-radiation induced template copolymerization and reported in our previous work [19,20]. In order to convert the polymer from the H+ -form to the Na+ form, the polymer (H+ ) was soaked in 0.1 M NaOH for 24 h. The solid material was separated from the solution by decantation and dried in electric oven (at ∼100 ◦ C). P(AM-AA-AN)-DAM (particle size 1.0–0.5 mm) was mixed individually with different concentration of each of 18 crown 6 (18C6), dibenzo 18 crown 6 (DB18C6) and 4,4 (5 )di-t-butylbenzo 18 crown 6 (D-t-BB18C6) that were dissolved in nitrobenzene and soaked overnight, decanted, then dried at ∼50 ◦ C for 24 h in an electric oven. The obtained three types of impregnated polymeric ion exchangers were subsequently used for batch and/or column experiments.

and 152+154 Eu radionuclides from radioactive liquid waste. Different forms of P(AM-AA-AN)-DAM such as H+ and Na+ were prepared to test the effect of counter ion on the removal efficiency. Furthermore, the effect of crown ether loading (10, 20 and 40%, w/v) on the polymers was studied in order to find the optimal impregnation of the polymeric resin used for the removal of radionuclides. For this, 5 mL of radioactive liquid waste (at pH 8) was mixed with 50 mg of desired impregnated polymeric materials. The mixture was contacted on the thermostatic shaker at room temperature for 24 h to attain equilibrium. The activity concentration of the radionuclides in solution was determined radiometrically using the HPGe detector. The sorption percent (S %) of the impregnated ion exchange resin is calculated according to the following equation: S (%) =

Ci − Cf Ci

× 100

(1)

where Ci and Cf are the initial and final counting rates per unit volume for the radionuclide, respectively; C0 is the initial concentration (mg/L) of metal ions used. 2.5. Column chromatography studies Fixed bed sorption studies were conducted to evaluate the column performance for Cs, Co, Zn and Eu ions removal on P(AMAA-AN)-DAM (Na+ )/DtBB18C6. Experiments were carried out in column of 0.8 cm inner diameter and 12.0 cm length packed with prepared P(AM-AA-AN)-DAM (Na+ )/DtBB18C6 at pH of 5.0 that was selected from our previous batch experiments [19]. Sampling of effluent was done at predetermined time intervals in order to investigate the breakthrough point. The effects of inlet eluent flow rate (1.0, 3.0 and 5.0 mL/min) and the resin bed height (2.0, 4.0, 6.0 and 8.0 cm) on the performance of the breakthrough curves for each ion were studied. The initial concentration of all inactive ions was kept at 100 mg/L. The break-through capacity (Q0.5 ) of the impregnated ion exchange resin is calculated according to the following equation: Breakthrough capacity (Q0.5 ) =

V(50%) × C0 m

(2)

where, V(50%) is the volume to break through at 50% uptake in L and m is the weight of the impregnated polymeric material (g). Set of experimental trials has been performed in order to elute and/or separate radionuclides that retained on the impregnated polymeric materials. In this respect different eluent reagents such as oxalic and nitric acid were used.

2.3. Instruments

2.6. Characterization of real radioactive liquid waste

The impregnated polymers were investigated using a FT-IR spectrometer (Bomen, Hartman & Braun, and model MB-157, Canada). The sample was ground into fine powder and dried to eliminate the moisture content. Representative amount of the impregnated polymer (2.0 mg) was then mixed with (98.0 mg) of potassium bromide (KBr). The mixture was compressed into the disc of 5 mm diameter and 1 mm thickness. The IR spectra of the prepared disc was then measured and recorded. Measurements of the gamma radioactivity of the different radionuclides in the samples were carried out using a non-destructive ␥-ray spectroscopic technique with high purity germanium (HPGe) detector model 2201-Oxford (USA).

Two types of liquid wastes were collected. The first waste sample, including mono-radionuclide (137 Cs only) was collected from the storage tank in Egyptian plant for treatment of radioactive liquid waste. While the second waste sample, including mixed radionuclides 134 Cs, 65 Zn, 60 Co and 152+154 Eu was collected from various laboratory research activities in Hot Laboratories Center, located at Abu Zaable site, Cairo, Egypt. Characterization of liquid radioactive waste used in this work has been done in our previous work [19,22], and are reported in Table 1. 3. Results and discussion

2.4. Batch experiments

3.1. Sorption percentage of radionuclides using different derivatives of 18 crown 6 based on polymeric resin

Preliminary batch experiments were performed to investigate the efficiency of different crown ethers impregnated into P(AMAA-AN)-DAM particles towards the removal of 134 Cs, 65 Zn, 60 Co

In order to investigate the sorption of some hazardous radionuclides from radioactive liquid waste, various types of impregnated polymeric resin were prepared based on some derivatives of 18

M.F. Attallah et al. / Journal of Hazardous Materials 195 (2011) 73–81

75

Fig. 1. Structure formula of 18 crown 6 (18C6), dibenzo 18 crown 6 (DB18C6) and 4,4 (5 )di-t-butylbenzo 18 crown 6 (DtBB18C6).

Table 1 Some chemical and radiochemical properties of liquid radioactive wastes [19,22]. Name

Individual radionuclide waste

pH 8 1844 TDS (mg/L) 6460 Conductivity (␮S/cm) Anions and cations (mg/L) − 199.0 Cl 443.0 NO3 − SO4 2− 1063 40.0 PO4 3− 33.5 Li+ 8.50 Na+ 2+ 9.0 Ca 2+ 25.0 Pb 25.0 Zn2+ ND Cu2+ ND Ni2+ Activity concentration (Bq/L) 137 Cs 25,397 ± 715 134 Cs ND 60 Co ND 152 Eu ND 154 Eu ND 65 Zn ND

Mixed radionuclide waste 7.5 809 1716 400.0 90.0 200.0 75.0 20.0 ND ND ND 12.5 2.50 15.0 ND 128,549 ± 1399 8320 ± 971 71,921 ± 2667 9707 ± 489 5175 ± 200

ND: not detected.

crown 6, including 18 crown 6 (18C6), dibenzo 18 crown 6 (DB18C6) and 4,4 (5 )di-t-butylbenzo 18 crown 6 (DtBB18C6) that were impregnated individually into P(AM-AA-AN)-DAM polymeric resin. Structural formula of 18C6, DB18C6 and DtBB18C6 are presented in Fig. 1. The obtained sorption percentages have been calculated and reported in Table 2. It appeared that the P(AM-AA-AN)-DAM poly-

meric resin in the Na+ form was more efficient for the sorption of the radionuclides than in the H+ form. Significant enhancement from 9, 13.3, 33 and 56.3% to 55.1, 54, 87.6 and 88.4% was obtained for the sorption of 137 Cs, 134 Cs, 60 Co and 65 Zn, respectively. This is attributed to ion exchange process that takes place for the radionuclides with Na+ ions more favorably than with H+ ions. This characteristic behavior was reported in other investigations. In extraction chromatography, many authors activate the ion exchange resin by NaCl solution to improve the uptake percentage especially for monovalent cations [19,20,22–26]. For example, Borai et al. [24] showed this idea for the impregnated zeolite materials that used for Cs removal. Based on their results, they demonstrated that the distribution coefficients and the corresponding uptake percentages of Cs-134 are highly affected (decreases) by potassium rather than sodium ions in the waste solution. This may be due to the close similarity in ionic radii between Cs+ and K+ rather than that between Cs+ and Na+ . Therefore, K+ ion could compete more with Cs+ ion during the sorption process. This finding has a typical explanation to our results that showed significant improvement in the uptake percentage of Cs ion due to the activation of the resin with sodium rather than H+ form. This phenomena is clear in uni-univalent cations exchange rather than divalent cases. Therefore, interesting high uptake values for Co(II) and Zn(II) with P(AM-AA-AN)-DAM were obtained even without impregnation. Better sorption of radionuclides on Na+ form was attributed to the consistence ionic radius of Na+ ion with the radionuclides rather than H+ ion [19,20,23]. Total dissolved salts and electric conductivity were found to be 1844 and 809 mg/L and 6460 and 1716 ␮S/cm (as reported in Table 1) for individual 137 Cs and mixed radionuclides radioactive

Table 2 Uptake percentage of some radionuclides from LLLRW using different crown ether derivatives based on P(AM-AA-AN)-DAM ion exchanger. Crown ether

a

+

R(H ) R(Na+ ) R(H+ ) + 10% 18C6 R(H+ ) + 20% 18C6 R(H+ ) + 40% 18C6 R(H+ ) + 10% DB18C6 R(H+ ) + 20% DB18C6 R(H+ ) + 40% DB18C6 R(H+ ) + 10% DtBB18C6 R(H+ ) + 20%DtBB18C6 R(H+ ) + 40% DtBB18C6 R(Na+ ) + 10% DtBB18C6 R(Na+ ) + 20% DtBB18C6 R(Na+ ) + 40% DtBB18C6

Mono

Mixed radionuclide

137

134

Cs

9.0 55.1 10.0 18.8 17.0 11.8 18.7 29.2 56 53.6 54.7 82.7 82.5 83.2

Cs

13.3 54 8.9 15.1 9.5 8.9 25.8 17.4 32.8 32.0 41.4 81.6 82.0 82.7

Experimental condition: real waste samples, contact time 24 h at room temperature. a R: means resin (P(AM-AA-AN)-DAM).

60

Co

33 87.6 11.7 23.2 10.5 10.0 16.2 24.6 33.1 33.3 38.4 84.7 85.3 86.7

65

Zn

56.3 88.4 45.4 62.5 55.7 54.1 44.3 45.1 65.2 68.2 66.7 81.1 82.2 82.5

152+154

63.3 42.3 72.2 85.8 77.3 83.9 91.6 84.7 89 93.4 92.5 83.8 80.7 84.3

Eu

76

M.F. Attallah et al. / Journal of Hazardous Materials 195 (2011) 73–81

Table 3 Distribution behavior of hazardous radionuclide using different crown ether. Impregnated resin or extractant Impregnated resin SiO2 -P/DtBuCH18C6 PVA/DCH18C6 P(AM-AA-AN)-DAM/DtBB18C6

Extractant DB18C6 DAB18C6 DHB18C6 DNB18C6 DtBB18C6 DtBB18C6 BEHBCalixC6 BOBCalixC6 BOBCalixC6 Calix[4]arene-BC6 a

Radionuclide

Distribution coefficient (mL/g)

References

Sr Cs Sr Cs Co Zn Eu

946 P(AM-AA-AN)-DAM (H+ )/ DB18C6 > P(AM-AA-AN)-DAM (H+ )/ 18C6 Based on these results, P(AM-AA-AN)-DAM (Na+ )/DtBB18C6 (10%) as impregnated ion exchanger was selected for the subsequent investigations. The obtained distribution coefficients of radionuclides under study have been compared with different impregnated crown ethers, derivative of crown ether and calix crown ether as extractant. It was observed the impregnated materials with different crown ether that used to remove and separate radionuclides are limited in literature. As shown in Table 3, the prepared resin (P(AM-AA-AN)-DAM/DtBB18C6 provided significantly high distribution coefficient for Cs radionuclides compared to the other derivative crown ether and calixarene.

sites for sorption of metal ions beside various factional groups in the polymer support materials. 3.3. Chromatographic column studies The operation and performance of a column are known to be influenced by it a number of parameters such as type, concentration and flow rate of the feed solution as well as column bed height. In this respect optimization of some variables is essential to evaluate the column performance. Fixed bed column experiments were carried out to study the sorption dynamics. The shape of the breakthrough curve and the time for the breakthrough appearance are the predominant factors for determining the operation and the dynamic response of the sorption column. The general position of the breakthrough curve along the volume/time axis depends on the capacity of column with respect to bed height, the feed concentration and the flow rate [40–43]. In this concern different flow rates as well as various bed heights were tested at fixed initial ion concentration of 100 mg/L for all ions under investigation. 3.3.1. Effect of bed height As shown in Fig. 3 (see also Table 4), the breakthrough capacity (Q0.5 ), breakthrough time was increased with increasing bed height. The increase in the ion sorption with bed height was due to the increase of the sorbent mass in larger beds, which provide greater sorption sites for the metal ions. The obtained results are agree with the same trend by other authors [19,43–45]. Based on the obtained result it could be found that breakthrough capacity (Q0.5 ) obeyed the following sequence at the same corresponding bed height: Co > Zn > Cs > Eu

3.2. Impregnated polymer structure The infrared spectra of the polymeric material and impregnated polymeric material (P(AM-AA-AN)-DAM (Na+ )/DtBB18C6) show (Fig. 2) that there are many vibrationally absorption bands, characterized mainly to carboxylate, carboxylic, ester, ether and nitrile groups. The broad absorption band at ∼3443, 3427 cm−1 is characterized to stretching vibrations of CONH2 related to amide group content in polymeric resin and impregnated polymeric material. This band was confirmed by the appearance of another band at 1520 cm−1 . Moreover, there is a strong absorption band at ∼2922 and 2875 cm−1 , attributed to stretching vibrations of CH2 group, which is confirmed by another band at 1076, 1165, and 1146 cm−1 . Two characteristic absorption bands at ∼1428, 1418 cm−1 are related to carboxylate group as well as absorption bands at ∼1710, 1428, 1418, 962 cm−1 are attributed to the carboxylic group. The absorption bands at 2242, 2240 cm−1 , are due to the nitrile group, as well as a band at 1655, 1597 cm−1 are attributed to C O bond. The new other two bands at 2957, 1366 cm−1 are attributed to t-butyl in impregnated polymeric material as well as a band at 1459 cm−1 is characterized to nitro-aromatic group [34–37]. The presence of the carboxylate and ester groups in P(AM-AAAN)-DAM indicated the interaction of DAM with carboxylic groups of acrylic acid of the polymeric chain. It was found that DAM acts as a crosslinker in the polymerization of acrylamide, acrylamide–acrylic acid and acrylic acid–acrylonitrile [38,39]. This implies the presence of acrylamide, acrylic acid, acrylonitrile and ether units in the impregnated polymeric chains, as shown in Fig. 2b. The spectroscopy revealed that the resin, including DAM is linked between the polymeric chains according to the mechanism for the template copolymerizeation of AA–AN on P(AM) in the presence of DAM while DtBB18C6 may be linked to the polymeric chains according to hydrogen bond [19,39]. Crown ethers were providing the active

3.3.2. Effect of flow rate The effect of flow rate on 134 Cs, 60 Co, 65 Zn and (152+154) Eu sorption by P(AM-AA-AN)-DAM (Na+ )/DtBB18C6 was studied by varying the flow rate for 1.0, 3.0 and 5.0 mL/min at the fixed bed height (4.0 cm) and initial concentrations (100 mg/L) for all ions under study. The plots of the breakthrough curves of 134 Cs, 60 Co, 65 Zn and (152+154) Eu at various flow rates are shown in Fig. 4. As shown from Fig. 4, an increase in flow rate reduces the effluent breakthrough volume and thereby decreases the retention time of the elements. This is due to the decrease in the residence time of the 134 Cs, 60 Co, 65 Zn and (152+154) Eu within the bed at higher flow rates. Much sharper breakthrough curves for 134 Cs, 60 Co, 65 Zn and (152+154) Eu sorption onto P(AM-AA-AN)-DAM (Na+ )/DtBB18C6 were obtained at higher flow rates. The breakthrough time and the amount of total 134 Cs, 60 Co, 65 Zn and (152+154) Eu sorbed also decreased with increasing flow rate, as presented in Table 4. This is attributed to the reduced contact time causing a weak distribution of the liquid inside the column, which leads to a lower diffusively of the solute among the particle of the P(AM-AA-AN)DAM (Na+ )/DtBB18C6 [41]. 3.4. Separation of 134 Cs, 60 Co, 65 Zn and (152+154) Eu from radioactive liquid waste Based on the previous results, removal and separation of 134 Cs, and (152+154) Eu radionuclides from low level liquid radioactive waste was investigated using column containing P(AMAA-AN)-DAM (Na+ )/DtBB18C6 at flow rate 3.0 mL/min and 4.0 cm bed height. The loading process was carried out by passing an appropriate volume of the radioactive waste solution. Some set of an experiment were preferred for removal and separation process towards 134 Cs, 60 Co, 65 Zn and (152+154) Eu radionuclides using

60 Co, 65 Zn

M.F. Attallah et al. / Journal of Hazardous Materials 195 (2011) 73–81

1.0

1.0

0.8

0.8

79

0.6

2+

0.4

C / Co

C / Co

+

Cs

Co

0.6

5.0 mL 3.0 mL 1.0 mL

0.4

5.0 mL 3.0 mL 1.0 mL

0.2

0.2

0.0 0

200

400

600

800

1000

1200

1400

0.0

0

200 400 600 800 1000 1200 1400 1600 1800

Effluent volume, mL

Effluent volume, mL

1.0

1.0

0.8

0.8 2+

0.6 5.0 mL 3.0 mL 1.0 mL

0.4

3+

Eu

0.6

5.0 mL 3.0 mL 1.0 mL

0.4

0.2

0.2

0.0

C / Co

C / Co

Zn

0

200 400 600 800 1000 1200 1400 1600 1800

0.0

0

100 200 300 400 500 600 700 800 900

Effluent volume, mL

Effluent volume, mL

Fig. 4. Breakthrough curve of Cs+ , Co2+ , Zn2+ and Eu3+ sorbed onto P(AM-AA-AN)-DAM (Na+ )/DtBB18C6 at different flow rate and at fixed bed height (4.0 cm) as well as initial concentration (100 mg/L) at pH 5.

80

Concentration, (mg/L)

nitric and oxalic acids as eluent reagents. Separation and removal of radionuclides under study are presented in Figs. 5–7. Different concentrations of nitric acid such as 0.1, 0.5 mol/L, and 0.12 mol/L of oxalic acid at pH 4.5 were investigated for removal and/or separation process. As shown in Fig. 5, 0.5 M of nitric acid was used as eluent for removal and separation process. It was found that preconcentration and removal of 134 Cs, 60 Co, 65 Zn and (152+154) Eu radionuclides was done by 60 mL with recovery percent >98%, as presented in Table 5. It can be inferred that 0.5 mol/L nitric acid is a good eluent for preconcentration and removal of all radionuclides under study, but it is not capable of the separation of radionuclides from each other. The second trial was carried out using 0.1 mol/L nitric acid as eluent as shown in Fig. 6. Separation of 134 Cs from 60 Co, 65 Zn and (152+154) Eu radionuclides was obtained by 50 mL, with recovery percent 98%. No release of any other radionuclides on elution by 0.1 mol/L nitric acid took place. Therefore, higher concentration of nitric acid (0.5 mol/L) was applied to elute 60 Co, 65 Zn and (152+154) Eu radionuclides.

0.5M HNO3 Cs Co Zn Eu

60

40

20

0

0

20

40

60

80

100

Effluent volume, mL Fig. 5. Elution curves of 134 Cs, 3.0 mL/min, 4.0 cm bed height.

60

Co,

65

Zn and

(152+154)

Eu by 0.5 mol/L nitric acid at

80

M.F. Attallah et al. / Journal of Hazardous Materials 195 (2011) 73–81

//

0.1M HNO 3

0.5M HNO 3

60 Cs Co Zn Eu

50

Concentration, (mg/L)

40

30

20

10

0

0

20

40

60

80

100

120

140

160

Effluent volume, mL Fig. 6. Gradient separation and removal of 4.0 cm bed height.

134

Cs,

60

Co,

65

Zn and

(152+154)

Eu from radioactive liquid waste using 0.1 mol/L followed by 0.5 mol/L nitric acid at 3.0 mL/min,

Table 5 Recovery percent of 134 Cs, 60 Co, 65 Zn and (152+154) Eu using different eluent. Eluent

Recovery, % Effluent volume, mL 134

0.12 M Oxalic acid (Fig. 7) 0.1 M nitric acid (Fig. 6) 0.5 M nitric acid (Fig. 6)

60

Cs

99 (60 mL) 98.3 (50 mL) 98.7 (50 mL)

65

Co

96.8 (70 mL) – 97.6 (50 mL)

The third trial was carried out using 0.12 mol/L of oxalic acid at pH 4.5, and it appeared that 0.12 mol/L of oxalic acid is highly efficient eluent for the preconcentration of both 134 Cs and 60 Co from 65 Zn and (152+154) Eu within about the first 50 mL. On the 0.12 M of oxalic acid

(152+154)

Zn

90 (80 mL) – 99 (60 mL)

Eu

– – 99.5 (60 mL)

other hand, 65 Zn was separated from (152+154) Eu within the second 50 mL, while (152+154) Eu is not eluted by oxalic acid. Then the last stage was performed for the separation of (152+154) Eu successfully by gradient elution of 120 mL of 0.5 mol/L nitric acid as 0.5 M of nitric acid

//

60 Cs Co Zn Eu

Concentration, (mg/L)

50

40

30

20

10

0

0

20

40

60

80

10 0

120

140

16 0

18 0

200

220

240

Effluent Volume, mL Fig. 7. Gradient separation and removal of 134 Cs, 60 Co, 65 Zn and (152+154) Eu from radioactive liquid waste using 0.12 mol/L of oxalic acid at pH 4.5 followed by 0.5 mol/L nitric acid at 3.0 mL/min, 4.0 cm bed height.

M.F. Attallah et al. / Journal of Hazardous Materials 195 (2011) 73–81

depicted in Fig. 7. The recovery percent of 134 Cs, 60 Co, and 65 Zn were 99, 96.8, and 90% using 0.12 mol/L of oxalic acid as well as 99.5% for (152 + 154) Eu using 0.5 mol/L of nitric acid, as shown in Table 5. 4. Conclusions Impregnated polymeric materials, namely P(AM-AA-AN)DAM/D-t-BB18C6, P(AM-AA-AN)-DAM/DB18C6 and P(AM-AAAN)-DAM/18C6 were prepared and the removal properties for some hazardous radionuclides from radioactive liquid waste were investigated. P(AM-AA-AN)-DAM/D-t-BB18C6 exhibits promising high sorption characteristics for the removal of 137 Cs and other mixed radionuclides such as 60 Co, 65 Zn and (152+154) Eu. Therefore, it can be used as an alternative sorbent material to the Egyptian plant for treatment of radioactive liquid waste. Separation and removal of some hazardous fission products were successfully applied from low level liquid radioactive waste (LLLRW) containing mixed radionuclides (134 Cs, 65 Zn, 60 Co and 152+154 Eu) using extraction column chromatography packed with P(AM-AAAN)-DAM/D-t-BB18C6. The separation process was done with high recovery by gradient elution of nitric and oxalic acids. References [1] S.A. Ansari, P.N. Pathak, M. Husain, A.K. Prasad, V.S. Parmar, V.K. Manchanada, Talanta 68 (2006) 1273. [2] B.S. Shaibu, M.L.P. Reddy, A. Bhattacharyya, V.K. Manchanda, J. Magn. Magn. Mater. 301 (2006) 312. [3] N.L. Engel, P.V. Bonnesen, B.A. Tomkins, T.J. Haverlock, B.A. Moyer, Solv. Extr. Ion Exch. 22 (4) (2004) 611. [4] A. Zhang, E. Kuraoka, H. Hoshi, M. Kumagai, J. Chromatogr. A 1061 (2004) 175. [5] A. Zhang, E. Kuraoka, M. Kumagai, Sep. Purif. Technol. 50 (2006) 35. [6] A. Zhang, E. Kuraoka, M. Kumagai, J. Chromatogr. A 1157 (2007) 85. [7] T. Honjo, H. Kitayama, K. Terada, T. Kiba, Fresenius Z. Anal. Chem. 330 (1988) 159. [8] A.K. Kostad, P.Y.T. Chow, F.F. Cantwell, Anal. Chem. 60 (1988) 1569. [9] J.P. Bernal, E. Rodriguez De San Miguel, J.C. Aguilar, G. Salazar, J. De Gyves, Sep. Sci. Technol. 35 (10) (2000) 1661. [10] K.A.K. Ebraheem, M.S. Mubarak, Z.J. Yassien, F. Khalil, Sep. Sci. Technol. 35 (13) (2000) 2115. [11] I.H. Gerow, J.E. Smith Jr., M.W. Davis Jr., Sep. Sci. Technol. 16 (1981) 519. [12] K.L. Nas, Solv. Extr. Ion Exch. 11 (1993) 729. [13] V.S. Talanov, G.G. Talanova, M.G. Gorbunova, R.A. Bartsc, J. Chem. Soc., Prekin Trans. 2 (2002) 209. [14] V.V. Yakshin, V.I. Zhilov, S.V. Demin, G.A. Pribylova, I.G. Tananaev, A.Y. Tsivadze, B.F. Myasoedov, C. R. Chim. 10 (2007) 1020.

81

[15] Y. Takeda, Topics Curr. Chem. 121 (1984) 1. [16] M.L. Dietz, E.P. Horwitz, M.P. Jensen, S. Rhoads, R.A. Bartsch, A. Palka, J. Krzykawski, J. Nam, Solv. Extr. Ion Exch. 14 (1996) 357. [17] A. Kumar, P.K. Mohapatra, V.K. Manchanda, J. Radioanal. Nucl. Chem. 229 (1998) 169. [18] Y. Kikuchi, Y. Sakamoto, Anal. Chim. Acta 403 (2000) 325. [19] M.F. Attallah, Radiochemical studies on separation and removal of some radionuclides from radioactive wastes, Ph.D. Thesis, Chemistry Department, Faculty of Science, Ain Shams University, Cairo, Egypt, 2009. [20] M.F. Attallah, E.H. Borai, K.F. Allan, Radiochemistry 51 (6) (2009) 622. [21] P.K. Mohapatra, D.S. Lakshmi, D. Mohan, V.K. Manchanda, J. Membr. Sci. 232 (2004) 133. [22] E.H. Borai, M.A. Hilal, M.F. Attallah, F.A. Shehata, Radiochim. Acta 96 (2008) 441. [23] F.A. Shehata, M.F. Attallah, E.H. Borai, M.A. Hilal, M.M. Abo-Aly, Appl. Radiat. Isot. 68 (2010) 239. [24] E.H. Borai, R. Harjula, L. Malinen, A. Paajanen, J. Hazard. Mater. 172 (2009) 416. [25] T. Siyam, M.M. Abdel-Hamid, I.M.M. El-Naggar, Macromol. Rep. A 32 (Suppls. 5 & 6) (1995) 871. [26] T. Siyam, I.M. El-Naggar, H.F. Aly, Intern. Topical Meeting on Nuclear and Hazards Waste Management Spectrum 96, 18–23 August, Seatle, Washington, 1996, p. 66. [27] T.M. Letcher, J.D. Mercer-Chalmers, R.L. Kay, Pure Appl. Chem. 66 (3) (1994) 419. [28] J.M. Caridade Costaa, P.M.S. Rodrigues, Port. Electrochim. Acta 20 (2002) 167. [29] H.J. Buschmann, H. Dong, E. Schollmeyer, J. Coord. Chem. 30 (1993) 311. [30] A. Zhang, Y.-Z. Wei, M. Kumagai, T. Koyama, J. Radioanal. Nucl. Chem. 262 (3) (2004) 739. [31] O.A. Zakurdaeva, S.V. Nesterov, N.A. Shmakova, G.K. Semenova, E.O. Sozontova, V.I. Feldman, Nucl. Instr. Meth. Phys. Res. B 265 (2007) 356. [32] L.H. Delmau, P.V. Bonnesen, B.A. Moyer, Hydrometallurgy 72 (2004) 9. [33] M.G. Gorbunova, P.V. Bonnesen, N.L. Engle, E. Bazelaire, L.H. Delmau, B.A. Moyer, Tetrahedron Lett. 44 (2003) 5397. [34] J.A. Dean, Lange’s Handbook of Chemistry, thirteenth ed., McGraw-Hill Inc., USA, 1985. [35] B. Staurt, Modern Infrared Spectroscopy, John Wiley and Sons, Ltd, West Sussex, PO 19 IUD, England, 1996. [36] R.A. Nyquist, R.O. Kagel, Infrared and Raman Spectra of Inorganic Compounds and Organic Salts, Academic Press Inc, 1997. [37] K. Nakamoto, Infrared and Raman Spectra of Inorganic and Coordination Compounds, John Wiley and Sons Publications, New York, 1978. [38] K.F. Allan, T. Siyam, W.A. Sanad, 6th Arab Intern. Conf. on Polym. Sci. and Technology, 1–5 September, Sharm El Sheikh-Sinaa, Egypt, 2001, p. 121. [39] K.F., Allan, Gamma radiation-induced preparation of some polymeric resins and their use for the treatment of waste water, Ph.D. Thesis, Chemistry Department, Faculty of Science, Suez Canal University, Egypt, 2004. [40] S. Netpradit, P. Thiravetyan, S. Towprayoon, Water Res. 38 (2004) 71. [41] T.S. Singh, K.K. Pant, Sep. Purif. Technol. 48 (2006) 288. [42] E. Malkoc, Y. Nuhoglu, J. Hazard. Mater. 135 (2006) 328. [43] M. Sarkar, A.R. Sarkar, J.L. Goswami, J. Hazard. Mater. 149 (2007) 666. [44] S.S. Metwally, Ion exchange characteristics of poly-acrylamide and polyacrylonitrile based Ce (IV) phosphate for removal and separation of some radioactive nuclides from waste solutions, Ph.D. Thesis, Chemistry Department, Faculty of Science, Ain Shams University, 2008. [45] A.M. El-Kamash, J. Hazard. Mater. 151 (2008) 432.

Journal of Hazardous Materials 195 (2011) 82–91

Contents lists available at ScienceDirect

Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat

Characteristics and source apportionment of PM1 emissions at a roadside station Y. Cheng a,b,c,∗ , S.C. Zou d , S.C. Lee c , J.C. Chow a,b,e , K.F. Ho c , J.G. Watson a,b,e , Y.M. Han b , R.J. Zhang f , F. Zhang a , P.S. Yau c , Y. Huang c , Y. Bai a , W.J. Wu a a Department of Environmental Science and Technology, School of Human Settlements and Civil Engineering, Xi’an Jiaotong University, No.28 Xianning West Road, Xi’an, Shaanxi, 710049, China b SKLLQG, Institute of Earth and Environment, CAS, Xi’an, Shaanxi, 710075, China c Department of Civil and Structural Engineering, Research Center for Environmental Technology and Management, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong d School of Marine Sciences, Sun Yat-sen University, Guangzhou 510275, China e Division of Atmospheric Sciences, Desert Research Institute, Reno, NV, USA f Key Laboratory of Regional Climate-Environment Research for Temperate East Asia, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, 100029, China

a r t i c l e

i n f o

Article history: Received 22 March 2011 Received in revised form 4 August 2011 Accepted 4 August 2011 Available online 25 August 2011 Keywords: PM1 Chemical composition PMF

a b s t r a c t The mass concentrations of PM1 (particles less than 1.0 ␮m in aerodynamic diameter), organic carbon (OC), elemental carbon (EC), water-soluble ions, and up to 25 elements were reported for 24 h aerosol samples collected every sixth day at a roadside sampling station in Hong Kong from October 2004 to September 2005. Annual average PM1 mass concentration was 44.5 ± 19.5 ␮g m−3 . EC, OM (organic matter, OC × 1.2), and SO4 = were the dominant components, accounting for ∼36%, ∼26%, and ∼24% of PM1 , respectively. Other components, i.e., NO3 − , NH4 + , geological material, trace elements and unidentified material, comprised the remaining ∼14%. Annual average OC/EC ratio (0.6 ± 0.3) was low, indicating that primary vehicle exhaust was the major source of carbonaceous aerosols. The seasonal variations of pollutants were due to gas-particle partitioning processes or a change in air mass rather than secondary aerosol produced locally. Vehicle exhaust, secondary aerosols, and waste incinerator/biomass burning were dominant air pollution sources, accounting for ∼38%, ∼22% and ∼16% of PM1 , respectively. Pollution episodes during summer (May–August) which were frequently accompanied by tropical storms or typhoons were dominated by vehicle emissions. During winter (November–February) pollution episodes coincided with northeasterly monsoons were characterized by secondary aerosols and incinerator/biomass burning emissions. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Adverse health effects of exposure to particles have been proven in numerous toxicological and epidemiological studies [1–5]. Fine particles, like PM1 (particles with an aerodynamic diameter of less than 1.0 ␮m), are those most harmful to human beings, as they are able to penetrate into the human respiratory and circulation system, resulting in adverse health effects [4,6]. The mechanism of these adverse health effects is unclear; however, previous research indicates that toxic elements and compounds carried in fine particles may play an important role. About 70–80% of toxic trace

∗ Corresponding author at: Department of Environmental Science and Technology, School of Human Settlements and Civil Engineering, Xi’an Jiaotong University, No.28 Xianning West Road, Xi’an, Shaanxi, 710049, China. Tel.: +86 29 83395078; fax: +86 29 83395078. E-mail address: [email protected] (Y. Cheng). 0304-3894/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jhazmat.2011.08.005

elements, like lead (Pb), cadmium (Cd), arsenic (As) and nickel (Ni), as well as Polycyclic Aromatic Hydrocarbon compounds (PAHs), were found in PM1 , with the majority of sulfur (S), vanadium (V), selenium (Se) and zinc (Zn) found in the submicron particle size range [7,8]. Fine particles, mainly arising from vehicle exhaust, comprised the majority of airborne total suspended particles (TSP) in the atmosphere of Hong Kong [9–12]. Lee et al. [11] found that PM1 constituted ∼70% of the PM2.5 mass at the Roadside Air Quality Monitoring Station on the campus of Hong Kong Polytechnic University (i.e., the PU Roadside Station). They suggested using PM1 as an indicator for vehicular emissions at the PU Roadside Station due to less influence from non-vehicle sources [11]. The concentrations of the elements Cr, Fe, Co, Cu, As, and Ba at the roadside location were a factor of two higher than those measured at the background Hok Tsui (HT) station in Hong Kong [7]. In a roadside microenvironment, increasing particles counts were found when vehicles accelerate (e.g., after stopping at a

39,214 41,268 41,747 38,066 39,955 631 543 401 538 536 ± ± ± ± ±

(m)

1040 1116 1053 1044 1069 6 7 6 3 7 ± ± ± ± ± 11 14 15 10 12 3 4 6 5 5 ± ± ± ± ± 4 5 7 5 5 10 9 16 9 10 ± ± ± ± ± 21 22 23 24 23 43 74 83 27 82 ± ± ± ± ± 67 172 92 55 102 4 55 33 2 46 81.6 83.5 72.5 76.0 78.9 20.2 28.0 26.6 18.5 23.3

± ± ± ± ±

4.2 1.6 2.0 3.6 5.2

(%) (◦ C)

± ± ± ± ±

9.0 6.3 11.4 11.4 10.6

(mm)

± ± ± ± ±

(km hr−1 )

Total bright sunshine (h) Wind speed

Prevailinig wind (◦ ) Rainfall RH

Seven days with mixing height higher than 3000 were not counter here.

The Teflon-membrane filters were analysed for the presence of 51 elements (from Na to U) by X-ray fluorescence (XRF, Watson et al. [21]) at the Environmental Analysis Facility of the

a

2.3. Chemical analyses

Temperature

The 24 h PM1 sampling was performed once every sixth day from 8 October 2004 to 23 September 2005 and forty valid sample sets were obtained. Approximately 5% of additional field blanks were collected for blank subtraction and error propagation. A URG3000ABC sampler (URG corporation, Chapel Hill, USA) with one PM1 inlet (Teflon® coated aluminum, URG corporation, Chapel Hill, USA), operated at 16.7 L/min, was used to collect samples. The sampling inlet was about 1.5 m above street level. The PM1 sampler was equipped with two parallel channels containing 47 mm Teflonmembrane and quartz-fiber filters at the flow rate of 8.3 L/min for each channel. Both the Teflon-membrane and quartz-fiber filters were weighed twice before and after sampling, respectively, using a Sartorius Model MC5 Microbalance (Göttingen, Germany) with a sensitivity of ±1 ␮g in the 0–250 mg range. Before weighing, filters were equilibrated for 24 h in a desiccator at 20–30 ◦ C and a relative humidity of 30–40%. Prior to sampling quartz-fiber filters were preheated in an electric furnace at 900 ◦ C for 3 h to remove carbonaceous contaminants. Collected quartz fiber samples were stored in a refrigerator at about 4 ◦ C to prevent the evaporation of volatile components prior to chemical analysis.

Season

2.2. Sampling method

Table 1 Summary of daily meteorological parameters and vehicle numbers in four seasons from October 2004 to September 2005.

The climate in Hong Kong is sub-tropical, influenced by the Asian monsoons. The cooling of the great Asian land mass during winter and its heating during summer give rise to monsoonal winds on a very large scale, which leads to four seasons of unequal duration in Hong Kong [18]. The four seasons in this study are defined as prolonged summer (May–August) and winter (November–February), and transitional, short spring (March–April) and autumn (September–October), as listed in Table 1. An examination of the historical climatology records shows that meteorological characteristics during the study period did not deviate from the norm. Daily meteorological data was obtained from the Hong Kong Observatory. PM1 samples were collected at the PU Roadside Station (22.30◦ N, 114.17◦ E), located in a residential and commercial area near Victoria Harbour. The sampling site [11] is about 1–2 m away from the curb of Hong Chong Road, which is approximately 30 m wide with four lanes for each direction leading to the busiest crossharbour tunnel in Hong Kong. During the sampling period, daily traffic flow remained at roughly 120,000 vehicles per day [19,20]. Traffic data were obtained from the toll data maintained for Victoria Harbour.

Daily global solar radiation (MJ m−2 )

2.1. Climate and sampling location

Spring Summer Autumn Winter Average

2. Experimental method

3 35 13 1 15

Mixing heighta

Daily diesel vehicle (#)

Daily total vehicle (#)

signal light or a bus stop), especially for diesel-fueled vehicles [13]. Previous studies [e.g., 11,14,15,16,17] have provided limited information on PM1 in Hong Kong. For this study, 24 h sampling of PM1 was conducted every sixth day at the PU Roadside Station from 8 October 2004 to 23 September 2005. Study objectives were to: (1) characterize the chemical composition and seasonal variation in speciated PM1 ; (2) quantify the source contributions to PM1 by the Positive Matrix Factorization (PMF) receptor model; and (3) investigate meteorological characteristics that may affect the occurrence and strength of each air pollution source.

83

123,130 121,997 123,517 122,690 122,580

Y. Cheng et al. / Journal of Hazardous Materials 195 (2011) 82–91

84

Y. Cheng et al. / Journal of Hazardous Materials 195 (2011) 82–91

PM1 massQuartz (μg m-3 )

100

was present as soluble K+ in the atmosphere. These physical consistencies validate the effectiveness of the sampling and chemical analysis methods used in this study.

y=1.01(±0.07)x-4.76(±2.17) R=0.92 n=40

80 60

3.2. PM1 mass concentration and mass closure

40 20 0

50

0

100

PM1 massTeflon (μg m-3 ) Fig. 1. PM1 gravimetric mass concentrations from Teflon-membrane and quartzfiber filters at the PU Roadside Station from October 2004 to September 2005.

Desert Research Institute (Desert Research Institute, DRI, Reno, NV, USA). Half of the quartz-fiber filters were extracted with distilled-deionized water and the extracts were analysed for chloride (Cl− ), nitrate (NO3 − ), sulfate (SO4 = ), water-soluble sodium (Na+ ), potassium (K+ ), calcium (Ca++ ) and ammonium (NH4 + ) ions by ion chromatography (DIONEX, USA), following the methodology described by Chow and Watson [22], in the Air Laboratory of the Hong Kong Polytechnic University. OC and EC were measured on a 0.5 cm2 quartz-fiber filter punch from the remaining half of the filters using a DRI Model 2001 carbon analyzer with a thermal-optical reflectance (TOR) method following the Interagency Monitoring of Protected Visual Environments (IMPROVE) protocol [23,24]. Field blanks were analysed for blank subtraction and error propagation. Major mass constituents, including OC, EC, SO4 = , NO3 − , NH4 + , and elements (e.g., Al, Si, Ca, Fe, and Zn), were detected in almost every sample. Concentrations of only 25 elements were reported here because elements, such as Sc, Co, Ga, Se, Y, Nb, Mo, Pd, Ag, Cd, In, Sb, Cs, La, Ce, Sm, Eu, Tb, Hf, Ta, Wo, Ir, Au, Hg, Tl, and U, seldom showed concentrations higher than three times their respective minimum detectable limits (MDLs, [25]). 3. Results and discussion 3.1. Measurement validation Because PM1 samples were acquired on two different substrates and chemical analyses were performed in different laboratories, consistency tests are needed as part of the quality assurance process. Fig. 1 shows the comparison of PM1 gravimetric mass measurements from the collocated Teflon-membrane and quartz-fiber filters. Good agreements (slope close to unity) and high correlation (R = 0.92) demonstrated the consistency of sample and gravimetric analyses. Similar comparisons were also carried out by Engelbrecht et al. [26] in the USA and Louie et al. [14] in Hong Kong for PM2.5 samples. Because quartz-fiber filters are known to have positive sampling artifacts due to absorption of gaseous organic compounds and water [27–31] and known to have a tendency to shred and fragment during sample handling, the following discussion refers to Teflon-membrane mass unless otherwise specified. Regarding different chemical analysis methods, Fig. 2 shows reasonable agreement for SO4 = versus S (R = 0.97), and K+ versus K (R = 0.98), sampled on different sampling substrates. The ratio of SO4 = to S was 2.57 with a small intercept (−0.05 ␮g m−3 ), indicating that more than ∼85% of S was present as soluble SO4 = in the atmosphere and that both XRF and IC measurements were valid. The scatter plot of K+ and K also showed a slope of 0.92 and close to zero intercept (−0.02 ␮g m−3 ) suggesting that over ∼90% of total K

The annual average PM1 mass concentration from October 2004 to September 2005 was 44.0 ± 19.4 ␮g m−3 (Table 2). This level was higher than that measured at the urban Chung Shan site in Taiwan (17.1, 13.1, 9.7 ␮g m−3 in spring, autumn and winter, respectively, [32]) and at the Virolahti background station in Finland (4.3 ± 3.8 and 3.8 ± 3.6 ␮g m−3 in summer and winter, respectively, [33]), but much lower than that (127.3 ± 62.1 ␮g m−3 ) at an urban site in Xi’an, China [34]. Mass balance measurements of PM1 showed that EC, OM (OM = OC × 1.2, [35]), and SO4 = were the major components of PM1 , accounting for ∼36%, ∼26%, and ∼24% of the PM1 mass, respectively. Low abundances were found for NO3 − (∼5%), NH4 + (∼3%), geological material and trace elements (∼3%), and unidentified material (∼3%). The statistical summary of PM1 mass concentration in Table 2 lists maximum PM1 mass concentration in winter (52.9 ± 20.1 ␮g m−3 ), followed by autumn (48.7 ± 24.8 ␮g m−3 ), spring (41.3 ± 7.5 ␮g m−3 ) and summer (34.8 ± 17.9 ␮g m−3 ). Hong Kong is located at south edge of East Asia and China. Monsoon winds exert a profound influence on the air quality of Hong Kong, as previously reported [11,14–16,36,37]. In the summer, prevailing southerly winds, with the resultant vector of 172◦ (Table 1), brought clean marine air masses to Hong Kong. In the autumn, winter, and spring, with the resultant vector of 71◦ (Table 1), prevailing northeasterly winds transported continental emissions from interior Asia to Hong Kong and the South China Sea. This explains the higher PM1 mass concentrations in winter, autumn, and spring compared to summer. Seasonal variation in mixing height, daily diesel vehicle numbers, and total vehicle numbers on the Hong Chong road (Table 1) are not significant factors that explain the seasonality of air pollution. 3.3. Carbonaceous aerosols During the sampling period, PM1 OC ranged from 3.2 to 29.8 ␮g m−3 and EC ranged from 8.3 to 26.8 ␮g m−3 . Annual average OC and EC were 9.6 ± 4.9 and 15.8 ± 5.1 ␮g m−3 , respectively. Compared to one of the most polluted inland cities (Xi’an) in China, average OC from this study was a factor of two lower than OC (21.0 ␮g m−3 ) reported by Shen et al. [34], and average EC was three times higher than in Xi’an (5.1 ␮g m−3 ), suggesting different source categories for carbonaceous aerosols between the two cities. The PU Roadside Station was dominated by fresh vehicle emissions, with a low average OC/EC ratio of 0.6 ± 0.3, and Xi’an was dominated by coal combustion emissions, with a high OC/EC ratio of 4.4 [34]. It has been reported that fresh vehicle emissions accounted for more than 60% of OC at a typical roadside Mong Kok (MK) station in Hong Kong after examining detailed organic species in PM2.5 using gas chromatography–mass spectrometry (GC/MS) method [38]. Average wintertime OC was 12.2 ± 6.0 ␮g m−3 (Table 2), approximately 60% higher than summer. High OC concentrations at the PU Roadside Station in winter were not due to the nearby on-road primary vehicle exhaust because the daily percentage of diesel-fueled vehicles (∼32%) and total traffic numbers on the Hong Chong road is consistent throughout the entire sampling period, as shown in Table 1. In addition, measurements of OC are sensitive to ambient and sampling conditions because gas-particle partitioning of OC are impacted by surrounding meteorological parameters. The seasonal data showed an inverse relationship between OC concentrations (Table 2) and temperature (Table 1) in line with the dynamic equi-

Y. Cheng et al. / Journal of Hazardous Materials 195 (2011) 82–91

b

45

Sulfate (μg m-3 )

36

Soluble potassium (μg m-3 )

a

y=2.57(±0.10)x-0.05(±0.52) R=0.97 n=40

27 18 9 0

0

5

10

Sulfur (μg

15

85

3

y=0.92(±0.03)x-0.02(±0.02) R=0.98 n=40

2

1

0

0

1

2

Total potassium (μg

m-3 )

3

m -3 )

Fig. 2. Physical consistency tests of PM1 measurements for (a) sulfate (SO4 = ) by ion chromatography on quartz-fiber filters versus total sulfur (S) by X-ray fluorescence (XRF) on Teflon-membrane filters; (b) soluble potassium (K+ ) versus total potassium (K).

librium between particle and gas phase OC, which supports high OC concentrations observed in winter. Except for primary sources, OC can be formed in the atmosphere as secondary aerosol that is typically enhanced when solar intensity is higher and daylight hours are longer. Table 1 shows that seasonal sunshine hours are highest in autumn when the OC/EC ratio is low indicating that it is not secondary aerosol produced locally but perhaps gas to particle partitioning processes or a change in air mass that is responsible for the seasonal changes. Pollutants transported from mainland China have been reported to increase the PM2.5 OC concentrations at the PU Roadside Station [11] and other ambient monitoring stations [14–16,36,37]. A previous study observed high OC concentrations (17.8 ± 10.2 ␮g m−3 ) and OC/EC ratios (2.9) in an upwind

area of Guangdong province [39], compared to those in Hong Kong. The seasonal EC concentrations showed highest value in summer (17.5 ± 5.8 ␮g m−3 ), followed by autumn (16.0 ± 8.0 ␮g m−3 ), winter (15.3 ± 3.8 ␮g m−3 ), and spring (12.2 ± 3.5 ␮g m−3 ). Good relationships between daily EC and wind speeds were found from November 2004 to April 2005 (winter and spring), with a correlation coefficient (R) of 0.72. EC decreased from 20.1 to 8.6 ␮g m−3 as northeasterly wind speeds increased from 8 to 39 km h−1 . In addition, the upwind area has less impact on EC levels at the PU Roadside Station when winter monsoons prevail because significantly lower EC (6.0 ± 3.2 ␮g m−3 ) has been reported over there [39]. Above evidence suggests that winter monsoons had a dispersive effect on EC concentrations.

Table 2 Statistical summary of 24 h PM1 measurements at the PU Roadside Station from October 2004 to September 2005. Totala

PM1

Spring

␮g m−3

Average

SDb

Average

SDb

Average

SDb

Average

SDb

Average

SDb

Mass (Teflon) Organic carbon (OC) Elemental carbon (EC) Chloride (Cl− ) Nitrate (NO3 − ) Sulfate (SO4 2− ) Soluble sodium (Na+ ) Ammonium (NH4+ ) Soluble potassium (K+ ) Sodium (Na) Magnesium (Mg) Aluminum (Al) Silicon (Si) Phosphorus (P) Sulfur (S) Chlorine (Cl) Potassium (K) Calcium (Ca) Titanium (Ti) Vanadium (V) Manganese (Mn) Iron (Fe) Nickel (Ni) Copper (Cu) Zinc (Zn) Arsenic (As) Bromine (Br) Rubidium (Rb) Strontium (Sr) Zirconium (Zr) Tin (Sn) Antimony (Sb) Barium (Ba) Lead (Pb)

41.3 8.9 12.2 0.4 2.8 10.0 1.2 3.0 0.42 0.43 0.12 0.14 0.18 0.18 4.4 0.41 0.50 0.06 0.0048 0.015 0.018 0.14 0.0048 0.011 0.11 0.0078 0.011 0.0031 0.0051 0.0081 0.025 0.027 0.029 0.040

7.5 0.4 3.5 0.3 0.9 5.3 0.1 1.1 0.20 0.17 0.04 0.05 0.09 0.08 2.0 0.60 0.26 0.02 0.0032 0.011 0.006 0.04 0.0036 0.002 0.06 0.0024 0.005 0.0036 0.0042 0.0011 0.016 0.010 0.023 0.022

34.8 7.3 17.5 0.2 0.8 6.7 1.3 1.4 0.18 0.39 0.11 0.12 0.14 0.12 2.7 0.10 0.20 0.08 0.0095 0.017 0.015 0.25 0.0054 0.012 0.18 0.0050 0.007 0.0017 0.0041 0.0082 0.015 0.038 0.025 0.017

17.9 3.3 5.8 0.2 0.7 5.0 0.2 1.3 0.17 0.36 0.07 0.11 0.17 0.10 2.4 0.03 0.21 0.11 0.013 0.019 0.013 0.38 0.0051 0.013 0.21 0.0042 0.005 0.0009 0.0022 0.0049 0.008 0.007 0.023 0.022

48.7 7.9 16.0 0.1 1.2 15.6 1.1 3.4 0.39 0.69 0.20 0.12 0.17 0.29 6.2 0.12 0.44 0.06 0.0058 0.016 0.016 0.18 0.0035 0.014 0.24 0.0147 0.005 0.0039 0.0043 0.0087 0.019 0.042 0.044 0.043

24.8 2.6 8.0 0.0 0.6 14.5 0.1 2.7 0.31 0.43 0.09 0.06 0.09 0.23 5.1 0.10 0.34 0.02 0.0025 0.006 0.005 0.08 0.0025 0.004 0.11 0.0000 0.001 0.0024 0.0011 0.0036 0.011 0.007 0.026 0.029

52.9 12.2 15.3 0.4 2.8 13.9 1.3 3.3 0.80 0.53 0.15 0.15 0.30 0.21 5.1 0.19 0.89 0.08 0.0085 0.016 0.029 0.24 0.0055 0.016 0.26 0.0150 0.017 0.0070 0.0046 0.0062 0.040 0.039 0.025 0.077

20.1 6.0 3.8 0.2 2.0 6.0 0.2 1.4 0.62 0.21 0.05 0.08 0.17 0.08 2.0 0.15 0.58 0.05 0.0067 0.018 0.021 0.17 0.0061 0.008 0.18 0.0103 0.013 0.0049 0.0031 0.0043 0.021 0.007 0.011 0.049

44.0 9.6 15.8 0.3 1.9 10.7 1.3 2.5 0.47 0.48 0.13 0.13 0.21 0.18 4.2 0.21 0.54 0.07 0.0082 0.016 0.021 0.23 0.0052 0.013 0.21 0.0127 0.012 0.0045 0.0045 0.0074 0.027 0.037 0.027 0.047

19.4 4.9 5.1 0.2 1.6 7.1 0.2 1.6 0.49 0.29 0.06 0.09 0.17 0.11 2.6 0.26 0.51 0.07 0.0091 0.017 0.017 0.26 0.0051 0.009 0.18 0.0093 0.010 0.0041 0.0027 0.0043 0.019 0.008 0.019 0.044

a

Summer

Total number of samples are equal to 40; b standard deviation.

Autumn

Winter

Y. Cheng et al. / Journal of Hazardous Materials 195 (2011) 82–91

3.4. Water-soluble ions and elements Table 2 shows that SO4 = (10.7 ± 7.1 ␮g m−3 ) was by far the major PM1 ion, followed by NH4 + (2.5 ± 1.6 ␮g m−3 ), and NO3 − (1.9 ± 1.6 ␮g m−3 ). Abundances of other ions were low: Na+ (1.3 ± 0.2 ␮g m−3 ), K+ (0.47 ± 0.49 ␮g m−3 ), and Cl− (0.3 ± 0.2 ␮g m−3 ). Crustal elements (i.e., Fe, Si, Mg, Al, and Ca in decreasing concentrations) were low, in the range of 0.07–0.23 ␮g m−3 . Trace elements (i.e., Sb, Sn, Ba, Mn, V, Cu, As, Br, Ti, Zr, Ni, Rb, and Sr) were in the range of 0.0045 ± 0.0027–0.037 ± 0.008 ␮g m−3 , with the exception of Zn (0.21 ± 0.18 ␮g m−3 ) and P (0.18 ± 0.11 ␮g m−3 ). On average, watersoluble ions, crustal elements, and the remaining elements accounted for ∼39%, 1.8%, and 0.6% of the PM1 mass, respectively. Out of all water-soluble ions and elements, eight species, such as SO4 = , NH4 + , NO3 − , Cl− , K+ , As, Br and Pb, exhibited significant seasonal differences at the 0.05 level using one-way analysis of variance (ANOVA) [41]. The concentrations of these pollutants were generally lowest during summer and highest during autumn and winter periods, as shown in Table 2. SO4 = and NH4 + had the highest concentrations in autumn, with average values of 15.6 ± 14.5 and 3.4 ± 2.7 ␮g m−3 , respectively. Fine-mode SO4 = , NO3 − , and NH4 + are secondary aerosols, arising from oxidization of gaseous precursors in air. Strong correlation (R = 0.96) between SO4 = and NH4 + suggests their co-existence in the atmosphere. The relationships between NO3 − and NH4 + were moderate (R = 0.61). Comparisons between the calculated and observed NH4 + concentrations were conducted to evaluate the formation of ions. NH4 + concentration can be calculated based on the stoichiometric ratios of the major compounds (i.e., ammonium sulfate [(NH4 )2 SO4 ], ammonium bisulfate [NH4 HSO4 ] and ammonium nitrate [NH4 NO3 ]); assuming that NO3 − is in the form of NH4 NO3 and that SO4 = is in the form of either (NH4 )2 SO4 or NH4 HSO4 . Fig. 3 shows the good correlation (R = 0.98) between calculated and measured NH4 + concentrations. The slope was 1.7 when (NH4 )2 SO4 was

-3

Calculated ammonium (µg m )

14

y=1.74(±0.06)x+0.21(±0.17) R=0.98 n=40

12 10

PM1

(NH4)2SO4+NH4NO3 NH4HSO4 +NH4NO3

8 6 4

y=0.97(±0.03)x+0.14(±0.08) R=0.98 n=40

2 0

0

2

4

6

8

10

12

14

-3

Measured ammonium (µg m ) Fig. 3. Comparison between calculated and measured ammonium in PM1 (calculated NH4 + = 0.38 × [SO4 = ] + 0.29 × [NO3 − ]) or NH4 HSO4 (i.e., NH4 + = 0.192 × [SO4 = ] + 0.29 × [NO3 − ]).

assumed and 1.0 when NH4 HSO4 was assumed. This suggests that aerosol is acidic (i.e., not fully neutralized with available NH4 + ) and in the form of NH4 HSO4 . The anion and cation balance in Fig. 4 also shows high correlation (R = 0.98). A deficiency of 11% in cations was found, especially at high loading concentrations, confirming the existence of acid aerosol. The seasonal anion-to-cation equivalent ratios (A/C) were 1.2 ± 0.04, 1.0 ± 0.2, 1.2 ± 0.3 and 1.2 ± 0.1 in spring, summer, autumn, and winter, respectively. Most samples had an A/C ratio higher than unity, especially during cold seasons. Increasing industrial activities in mainland China and prevailing northeasterly winds during winter may have contributed to the elevated SO4 = concentrations. Only ∼27% of the PM1 samples, mostly from summer, gave A/C ratios less than unity. Similar to those reported by Lin et al. [8], concentrations of crustal elements (i.e., Si, Al, Ca, Ti) were correlated with each other (R > 0.75), while Fig. 2 shows that over 90% of K is in the form of K+ , showing poor correlation with crustal elements. Abundant K in the form of K+ suggests the influence of biomass burning and waste incinerator in the Macao Special Administrative Region (SAR) [14]. This is confirmed by good correlations (R = 0.8) of K+ with Rb and Pb. The annual average Pb level (47 ± 44 ng m−3 ) at the PU Roadside Station is three times lower than the annual USA standard of 150 ng m−3 . Annual average PM1 V (16 ± 17 ng m−3 ), Mn (21 ± 17 ng m−3 ), and Pb (47 ± 44 ng m−3 ) are much lower than the World Health Organization [42] guideline values of 1, 0.15, and 0.5 ␮g m−3 , respectively. Two carcinogenic substances, As 0.8 -3

In summer, a higher fraction of OC exists in the vapor phase as temperatures increase, which results in low particle OC concentration (7.3 ± 3.3 ␮g m−3 ) and low OC/EC ratio of 0.4 ± 0.1. Among all potential sources (e.g., vehicle exhaust, cooking, and vegetative burning), vehicle exhaust is the most likely to produce such low OC/EC ratios [40]. Secondary organic aerosols were insignificant at the PU Roadside Station also in the summer because low OC/EC ratios and a moderate correlation between OC and total sunshine hours (R = 0.42) and solar radiation (R = 0.06) were observed. The summer daily OC and EC concentrations correlated well with prevailing wind directions, with correlation coefficients (R) of 0.76 and 0.84, respectively. Concentrations increased from 5.8 to 20.7 ␮g m−3 for OC, and from 9.4 to 27.1 ␮g m−3 for EC as the direction of the vector of prevailing winds changed from 50◦ to 300◦ . However, OC/EC ratios did not follow the changes in wind directions, with the minimum and maximum values of 0.31–0.62, respectively. The evidence above suggests the existence of primary sources southwest of the Hong Chong Road. Vehicle exhaust from the Victoria Harbour tunnel may contribute to observed carbonaceous aerosols, because the exit of the tunnel is about 800 m southwest of the PU Roadside Station in summer. In addition to vehicle exhaust from the tunnel, ship/container terminal emissions may contribute as well. Several container ports are distributed to the southwest of the PU sampling station, stretching for miles along the south coast of Kowloon Peninsula, Hong Kong. Elevated pollution levels around the Victoria Harbour area (near the sampling site) due to the influences of local vehicle exhaust and ship emissions have been reported by the Institute for the Environment of the Hong Kong University of Science and Technology (http://envf.ust.hk) and Civic Exchange (http://www.civic-exchange.org/).

Anion equivalence (ueq m )

86

0.6

y=1.45(±0.05)x-0.06(±0.01) R=0.98 n=40

0.4

0.2

0.0 0.0

0.2

0.4

0.6

0.8

-3

Cation equivalence (ueq m ) Fig. 4. Scatter plots of PM1 anion versus cation measurements from PM1 data. The anion equivalence was calculated from Cl− , NO3 − , SO4 = and the cation equivalence was calculated from Na+ , NH4 + , K+ , Mg++ , and Ca++ .

70

Factor1_Vehicle

60

50

50

40

40

30

30

10

0

0

Factor3_ Secondary aerosol

50 40

60 50 40 30

20

20

10

10

0

0

30

80 60

Factor 6_ Residul oil combustion

40

20

20

10

0 NO3SO4= NH4+ K+ Ca++ OC EC Na Mg Al Si P S K Ca Ti V Mn Fe Ni Cu Zn Br Pb

0 NO3SO4= NH4+ K+ Ca++ OC EC Na Mg Al Si P S K Ca Ti V Mn Fe Ni Cu Zn Br Pb

Factors' contribution (%)

40

Factor5_Waste incinerator/ biomass burning

Factor4_Resuspended road dust

NO3SO4= NH4+ K+ Ca++ OC EC Na Mg Al Si P S K Ca Ti V Mn Fe Ni Cu Zn Br Pb

30

NO3SO4= NH4+ K+ Ca++ OC EC Na Mg Al Si P S K Ca Ti V Mn Fe Ni Cu Zn Br Pb

Factors' contribution (%)

60

50

Factor2_Industrial exhaust

NO3SO4= NH4+ K+ Ca++ OC EC Na Mg Al Si P S K Ca Ti V Mn Fe Ni Cu Zn Br Pb

10

60

87

20

20

NO3SO4= NH4+ K+ Ca++ OC EC Na Mg Al Si P S K Ca Ti V Mn Fe Ni Cu Zn Br Pb

Factors' contribution (%)

Y. Cheng et al. / Journal of Hazardous Materials 195 (2011) 82–91

Fig. 5. Factor loadings obtained from positive matrix factorization (PMF) analysis of chemical constituents of PM1 .

(13 ± 9 ng m−3 ) and Ni (5 ± 5 ng m−3 ), and one toxic substance, Cu (13 ± 9 ng m−3 ), have concentrations lower than their California chronic exposure limits. Evidence indicates that all elements in PM1 related to human health were lower than corresponding guideline values over the duration of four seasons at the PU Roadside Station. 3.5. Source apportionment by the PMF model

Percentage of PM (%)

PMF has been shown to be a powerful tool for source identification [43,44] and has been used to assess PM2.5 and PM10 source

contributions in the Arctic [45], Hong Kong [12,46], Thailand [47], Vermont [48], and cities in the USA [44,49–51]. For this study, measured concentration values and uncertainties (sampling and chemical analytical errors) were used as input data for the PMF3.0 model [52–54]. Species with a signal-to-noise ratio less than 0.2 were excluded from the analysis [54]. Finally, only 25 species were included in the PMF3.0, including mass, NO3 − , SO4 = , NH4 + , K+ , Ca++ , OC, EC, Na, Mg, Al, Si, P, S, K, Ca, Ti, V, Mn, Fe, Ni, Cu, Zn, Br, and Pb. Forty samples were involved in the calculation.

Chemical Profile of PM1 from the PMF

10

Chemical Profile of PM2.5from the tunnel

measurement

1

NO3SO4= NH4+ K+ OC EC Na Mg Al Si P S K Ca Ti V Mn Fe Ni Cu Zn Br

0.1

Chemical profile for mixed-vehicles Fig. 6. Comparison of PMF-calculated and tunnel-measured vehicle chemical profiles in Hong Kong.

Fig. 7. Average source contributions of each factor to PM1 mass.

88

Y. Cheng et al. / Journal of Hazardous Materials 195 (2011) 82–91

25 20

EC_Vehicle

10

5

5

0

0

0.07 0.06

=

SO4 _Secondary aerosol

15

10

V_Residul oil combustion

0.05

2.0

K_Waste incinerator / biomass burning

1.5

0.04 1.0

0.03 0.02

0.5

0.01 0.00

0.0

0.7

0.8

0.6

Si_Suspended road dust

0.5

0.7 0.6

Zn_Industrial exhaust

0.5

0.4

0.4

0.3

0.3 0.2

0.1

0.1

0.0

0.0 2004-10-8 2004-10-14 2004-10-20 2004-10-26 2004-10-29 2004-11-1 2004-11-4 2004-11-13 2004-11-19 2004-11-24 2004-12-22 2004-12-28 2005-1-4 2005-1-10 2005-1-22 2005-1-30 2005-2-3 2005-2-11 2005-4-4 2005-4-12 2005-4-19 2005-4-27 2005-5-4 2005-5-21 2005-6-7 2005-6-14 2005-6-21 2005-6-28 2005-7-5 2005-7-13 2005-7-20 2005-7-22 2005-7-28 2005-8-4 2005-8-9 2005-8-12 2005-8-31 2005-9-10 2005-9-16 2005-9-23

0.2

2004-10-8 2004-10-14 2004-10-20 2004-10-26 2004-10-29 2004-11-1 2004-11-4 2004-11-13 2004-11-19 2004-11-24 2004-12-22 2004-12-28 2005-1-4 2005-1-10 2005-1-22 2005-1-30 2005-2-3 2005-2-11 2005-4-4 2005-4-12 2005-4-19 2005-4-27 2005-5-4 2005-5-21 2005-6-7 2005-6-14 2005-6-21 2005-6-28 2005-7-5 2005-7-13 2005-7-20 2005-7-22 2005-7-28 2005-8-4 2005-8-9 2005-8-12 2005-8-31 2005-9-10 2005-9-16 2005-9-23

-3

25 20

15

Species concentration (µg m )

30

Fig. 8. Temporal patterns of marker species from the six source categories during the time period from October 2004 to September 2005.

Six factors were generated by the PMF model and the contributions of each factor are shown in Fig. 5. The first factor with high loading of OC, EC, and Ca++ , are characteristic of vehicle exhaust [40,55–57] and lube oil additives [58], respectively. As shown in Fig. 6, the species abundances in this PMF-derived vehicle source profile are comparable to those measured in an urban tunnel in Hong Kong [55], especially for the major species (i.e., OC, EC, NO3 − , SO4 = , NH4 = , Na, Mg, and K). PMF-derived vehicle source profiles underestimate crustal elements (i.e., Al, Si, Ca, Fe), which is reasonable because these elements mainly occur in PM2.5 but not in PM1 . The second factor, loaded with Mn and Zn, represents exhaust from industry [59,60]. The third factor is identified as secondary aerosols, based on the high abundances of NO3 − , SO4 = , and NH4 + . The fourth factor, enriched with Al, Si, Ca, Ti, and Fe, is best explained as geological material or resuspended road dust [61,62]. The fifth factor with abundant K+ , Br, and Pb, is indicative of waste incinerator/biomass burning emissions, as previously found by Louie et al. [14]. The sixth factor, loaded with V and Ni, is consistent with residual oil combustion, likely from ship emissions or utilities at container terminals [12,17]. Fig. 7 shows that vehicle exhaust is the largest contributor, accounting for ∼38% of the PM1 mass. This is consistent with sampling in a roadside vehicle exhaust-dominated environment. Secondary aerosols are the second largest contributor, accounting for ∼22% of PM1 . It has been shown through previous analyses that long-range transported secondary aerosols impact the air quality at the sampling site when the air mass has traveled over China before reaching Hong Kong. Waste incinerator/biomass burning and residual oil combustion account for ∼16% and ∼12% of PM1 , respectively. Industrial exhaust and resuspended road dust account for ∼7% and ∼5% of PM1 , respectively.

Marker species (i.e., EC, Zn, SO4 = , Si, K+ , and V, respectively) were selected as representative components of their six respective source categories. Temporal patterns of marker species (Fig. 8) and meteorological characteristics on episode days (Table 3) were examined in order to identify essential meteorological parameters that may affect the occurrence and intensity of certain types of air pollution. The highest five concentrations for each marker species were considered to air pollution episodes, as shown in Table 3. Meteorological characteristics were obtained from the Hong Kong Observatory. The temporal patterns of SO4 = and K+ suggest that secondary aerosols and waste incinerator/biomass burning frequently occurred in autumn, winter, and spring. The prevailing wind directions were 70◦ and 54◦ on the SO4 = and K+ episode days, respectively. Moreover, elevated SO4 = , K+ , and PM1 concentrations often existed simultaneously. Vehicle exhaust episodes mainly occurred in summer, accompanied by typhoon or tropical storm events. Elevated PM1 concentrations were often observed at the same time. The temporal variation in V showed weak seasonality for residual oil combustion sources. V and Ni correlated well with each other (R = 0.97) and did not change with the changes in wind directions throughout the entire sampling periods. Therefore, ship emissions may also influence the PU Roadside Station as a regional pollution source. Wind direction was the most important meteorological parameter for industrial exhaust sources as they are stationary emitters. Resuspended road dust episodes predominantly occurred in winter, with less rainfall and dry northeasterly winds prevail. Throughout the entire sampling period, air pollution on 4 January 2005 was the most serious, with all marker species showing elevated concentrations in the atmosphere, which could be explained by the extremely low mixing height of 477 m on that

Table 3 Summary of air pollution episodes (highest five concentrations of each marker species) from October 2004 to September 2005. Episode day

Vehicle Secondary exhaust aerosols

a

EC ∼38%

b

8October2004 14 October 2004 20 October 2004 26 October 2004 1 November 2004 4 November 2004 19 November 2004 4 January 2005 10 January 2005 19 April 2005 21 June 2005

67.0 55.9 84.4 67.1 62.4 40.4 41.4 93.6 76.2 48.9 32.6

13 July 2005 20 July 2005

40.9 64.2

4 August 2005

25.1

12 August 2005

74.5

26.8

31 August 2005 10 September 2005 23 September 2005 Annual average 44.0

60.4 72.4 50.6 15.7

22.5

a

SO4 = ∼22%

Waste incinerator/bio moss burning K+ ∼16%

Resuspended road dust

Residual oil combustion

Si ∼5%

V ∼12%

Air temperature (◦ C)

Relative humidity (%)

Rainfall (mm)

Prevailing wind direction (◦ )

Wind speed (km h−1 )

Weather

904 1268 1022 967 663 620 733 477 732 660 641

25.9 25.1 24.8 25.5 23.1 23.1 20.4 16.5 15.2 23.3 28.5

45% 70% 74% 58% 70% 73% 51% 70% 73% 84% 83%

0.0 0.1 5.8 2.3 2.2 0.0 0.0 1.2 0.0 4.7 9.7

10 90 80 20 80 80 90 70 90 70 230

17 36 31 37 28 29 21 26 24 19 22

842 573

29.8 28.7

75% 81%

0.2 16.2

230 230

20 23

836

28.5

82%

14.9

240

20

0.5

738

28.3

83%

16.4

230

21

29.4 28.1 28.6

1067

728 898 838 23.0

76% 78% 67% 75%

0.3 6.6 Trace 10.8

270 90 10 87.8

18 18 42 23

Fine and dry Fine Haze Fine Some haze Sunshine Fine Haze Haze Sunshine Heavy rains and Thunderstorms Sunny and hot Sunny and hot; Thunderstorms Sunny and hot; Thunderstorms Haze; Thunderstorms Fine Haze Hot and hazy Annual average 44.0

Industrial Mixing exhaust height (m) Zn ∼7%

1.3 19.2 24.1 18.4

26.5

1.6 1.6

2.2 1.4

0.7 0.5 0.4 0.5

0.054

0.7 0.4

0.067

0.5

0.031 0.7

25.2 22.8 0.7 0.070 0.050 32.2 25.2 10.7

0.5

0.2

0.016

0.2

Y. Cheng et al. / Journal of Hazardous Materials 195 (2011) 82–91

PM1 mass

Marker species; b percentage of PM1 .

89

90

Y. Cheng et al. / Journal of Hazardous Materials 195 (2011) 82–91

day. Overall, average concentrations of each marker species, except for EC, on episode days were around two to three times higher than the annual average values. The average concentration of EC was 24.5 ± 1.8 ng m−3 on episode days, approximately 1.6 times the annual average value. Meteorological data shows that low mixing heights (averaged at 786 m) occurred on all air pollution episode days and relative humidity ranged from 70 to 85%. Except for vehicle exhaust episodes, all air pollution episodes were associated with hazy, reduced visibility conditions, similar to those found in Beijing [63] and Guangzhou, China [64]. 4. Conclusions Twenty-four-hour PM1 samples were collected every sixth day at an urban roadside monitoring station in Hong Kong from October 2004 to September 2005. Concentrations of OC, EC, water-soluble ions, and up to 25 elements were reported. The seasonal average PM1 concentrations were 41.3 ± 7.5, 34.8 ± 17.9, 48.7 ± 24.8, and 52.9 ± 20.1 ␮g m−3 for spring, summer, autumn, and winter, respectively. PM1 and major component species (e.g., OC, SO4 = , NO3 − , NH4 + , etc.) showed distinct seasonal patterns with elevated concentrations typically in autumn, winter and spring, which were associated with northeasterly winds that transport from the continental Asian interior. Except for long-range transported regional pollutants, OC were impacted by gas-particle partitioning as well. EC was an exception with elevated concentrations in summer associated with primary emissions from local nearby tunnel and container ports. The low OC/EC ratio ( dermal contact > inhalation. Inhalation presented the lowest cancer risks (EW (DW): 0.000873 × 10−6 ; EW (OBS): 4.59 × 10−11 ) because PCBs possess a low volatility [56].

97

Ingestion is a significant exposure pathway for PCBs usually due to involuntary consumption of soil. As mentioned above, cancer risks for PCBs through ingestion (EW (DW): 20.8 × 10−6 ; EW (OBS): 1.09 × 10−6 ) were relatively high when compared to dermal contact and inhalation. Consequently, an in vitro system that simulated the human digestive system, was adopted in the current study in order to investigate the amount of bioaccessible carcinogenic PCBs in the soil samples. According to Table 7, it can be observed that the cancer risk for bioaccessible PCBs from EW (DW) was significantly higher than the other types of soil (mean = 1.71 × 10−6 ± 2.96 × 10−6 ), implying that the bioaccessible PCBs concentrations in EW (DW) may pose potential cancer risk in humans. The major exposure pathway of PBDEs and PCBs to workers or farmers in OF, A, EW (S) and EW (DW) was discovered to be dermal contact. But the pathway in OBS and EW (OBS) in contrast was via inhalation, as the combustion activities in these land use types tend to generate ultra fine particles less than PM0.1 , which can penetrate deeply into the lungs and cause adverse health effects [57]. The estimation of health risks via inhalation should be based on pollutants adsorbed onto respirable particles of soils (less than PM10 ) [31]. Only the inhaled soil particles with a size of less than PM10 can be deposited in the upper part of the respiratory tract or penetrate deeply into the lungs [58,59]. Fine soil particles (less than PM10 ) with organic pollutants (such as PAHs) and inorganic pollutants (such as Cu, Cd and Zn), may be able to cause oxidative stress and inflammation after penetrating into the lungs [60,61]. This study used soil particles with a diameter of less than 2 mm to estimate cancer health risks in humans via the exposure route of inhalation, implying that not all soil particles were able to penetrate into the lungs. In addition, the concentrations of pollutants in soil particles with a diameter of less than 2 mm should be lower than the particles smaller than PM10 . Consequently, the human health risks based on pollutant concentrations would more than likely be underestimated. Furthermore, the absent IUR of BDE209 causes this evaluation of cancer risks to be underestimated. There is a need to derive pollutant toxicity values based on the inhalation pathway from experimental data in order to fill the gap of risk assessment via this means especially in the OBS and EW (OBS). Evaluating the health risks by using bioavailable pollutant concentrations is commonly regarded as the most accurate way, because only the bioavailable portion of the contaminants will ultimately reach our bloodstream and exert adverse effects on our body [20]. However, this method usually brings along ethical concerns due to the involvement of animal experiments, therefore, assessing bioaccessible fractions of pollutants may be a suitable alternative in portraying the reality [20]. In this study, bioaccessible fractions of PCBs were used to estimate the health risks via ingestion using an in vitro digestion model. No bioaccessible fractions of pollutants were used in the cancer risk estimations of the other two studied pathways of the present

Table 7 Cancer risks via ingestion of soils in humans from different types of agricultural land uses based on bioaccessible PCB toxic equivalent concentrations at minimum, medium, maximum and mean. Sampling sites

Organic farm Agricultural E-waste storage Open burning site E-waste dismantling workshop E-waste open burning site

Cancer risks via ingestion Min

Median

Max

Mean

N.D. N.D. N.D. N.D. N.D. N.D.

N.D. N.D. N.D. 0.0203 0.00136 0.00611

N.D. 0.0152 0.00359 0.0365 5.13 0.139

N.D. 0.00507 0.00120 0.0189 1.71 0.0483

Note: cancer risk are in bold, N.D. means not detected and the values of cancer risk are in the unit of 10−6 .

± ± ± ± ±

0.00879 0.00207 0.0183 2.96 0.0784

98

Y.B. Man et al. / Journal of Hazardous Materials 195 (2011) 92–99

study, consequently, the cancer risks from PCBs may be underestimated. 4. Conclusion Inhalation of soil particles is the major exposure pathway of PBDEs and PCBs to humans from OBS and EW (OBS). Whereas, the major exposure pathway of other land use types including (OF, A. EW (S) and EW (DW)) is via dermal contact of soils. Soils from EW (DW) and EW (OBS) were of the greatest concern in terms of threatening human health as they contained the highest concentrations of PCBs and PBDEs, resulting in relatively high cancer risks amongst the 6 types of land use. The burning and dismantling activities in e-waste sites may still potentially pose cancer risks to humans. Although the cancer risks of PBDEs via the exposure pathways of ingestion and dermal contact of soils in EW (DW) and EW (OBS) were still very low, these two pathways were not the major exposure pathways in EW (OBS). Hence, regular monitoring is required as these pollutants may be continuously deposited on soils and eventually accumulated to hazardous levels. Acknowledgements The authors would like to thank the Public Policy Research Grants (2002-PPR-3), Special Equipment Grant (SEG HKBUO09) of the Research Grants Council of Hong Kong and the Mini-AoE (Areas of Excellence) Fund from Hong Kong Baptist University (RC/AoE/08–09/01) for financial support. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.jhazmat.2011.08.010. References [1] H. Liu, Q. Zhou, Y. Wang, Q. Zhang, Z. Cai, G. Jiang, E-waste recycling induced polybrominated diphenyl ethers, polychlorinated biphenyls, polychlorinated dibenzo-p-dioxins and dibenzo-furans pollution in the ambient environment, Environ. Int. 34 (2008) 67–72. [2] M.H. Wong, S.C. Wu, W.J. Deng, X.Z. Yu, Q. Luo, A.O.W. Leung, C.S.C. Wong, W.J. Luksemburg, A.S. Wong, Export of toxic chemicals – a review of the case of uncontrolled electronic-waste recycling, Environ. Pollut. 149 (2007) 131–140. [3] F. Rahman, K.H. Langford, M.D. Scrimshaw, J.N. Lester, Polybrominated diphenyl ether (PBDE) flame retardants, Sci. Total Environ. 275 (2001) 1–17. [4] US EPA (United States Environmental Protection Agency), Polychlorinated Biphenyls (PCBs) Update: Impact on Fish Advisories, EPA-823-F-99-019, Environmental Protection Agency, Washington, 1999. [5] J.C.W. Lam, R.K.F. Lau, M.B. Murphy, P.K.S. Lam, Temporal trends of hexabromocyclododecanes (HBCDs) and polybrominated diphenyl ethers (PBDEs) and detection of two novel flame retardants in marine mammals from Hong Kong, South China, Environ. Sci. Technol. 43 (2009) 6944–6949. [6] C.K.C. Wong, K.M. Leung, B.H.T. Poon, C.Y. Lan, M.H. Wong, Organochlorine hydrocarbons in human breast milk collected in Hong Kong and Guangzhou, Arch. Environ. Contam. Toxicol. 43 (2002) 364–372. [7] J. De Boer, P.G. Wester, H.J.C. Klamer, W.E. Lewis, J.P. Boon, Do flame retardants threaten ocean life? Nature 394 (1998) 28–29. [8] J.E. Goodman, Neurodevelopmental effects of decabromodiphenyl ether (BDE209) and implications for the reference dose, Regulat. Toxicol. Pharmacol. 54 (2009) 91–104. [9] ATSDR (Agency for toxic substances and disease registry), Toxicological Profile for Polybrominated Diphenyls and Polybrominated Diphenyl Ethers-draft for Public Comment, Agency for Toxic Substances and Disease Registry, US Department of Health and Human Services, Public Health Service, Atlanta, 2002. [10] T.A. McDonald, A perspective on the potential health risks of PBDEs, Chemosphere 46 (2002) 745–755. [11] L.S. Birnbaum, D.F. Staskal, Brominated flame retardents: a cause for concern? Environ. Health Perspect. 12 (2004) 9–17. [12] US EPA (United States Environmental Protection Agency), Polychlorinated Biphenyls (PCBs), Basic Information, 2009. [13] C.R. Bryant, L.H. Russwurm, A.G. McLellan, The City’s Countryside: Land and its Management in the Rural-Urban Fringe, Longman, New York, 1982. [14] T. Lindstrom, E. Hansen, H. Juslin, Forest certification: the view from Europe’s NIPFs, J. Forest. 97 (1999) 25–30.

[15] C.Y. Jim, Rural blight and land use planning in Hong Kong, Environment 17 (1997) 269–281. [16] Y.W. Li, Agricultural Land in Hong Kong: A Solution Space for Urban Development, MSc. Dissertation. The University of Hong Kong, Hong Kong, 1998. [17] J.P. Wu, X.J. Luo, Y. Zhang, Y. Luo, S.J. Chen, B.X. Mai, Z.Y. Yang, Bioaccumulation of polybrominated diphenyl ethers (PBDEs) and polychlorinated biphenyls (PCBs) in wild aquatic species from an electronic waste (e-waste) recycling site in South China, Environ. Int. 34 (2008) 1109–1113. [18] J. She, A. Holden, M. Sharp, M. Tanner, C. Williams-Derry, K. Hooper, Polybrominated diphenyl ethers (PBDEs) and polychlorinated biphenyls (PCBs) in breast milk from the Pacific Northwest, Chemosphere 67 (2007) S307–S317. [19] A. Mazdai, N.G. Dodder, M.P. Abernathy, R.A. Hites, R.M. Bigsby, Polybrominated diphenyl ethers in maternal and fetal blood samples, Environ. Health Perspect. 111 (2003) 1249–1252. [20] Environment Agency, UK In-vitro Methods for the Measurement of the Oral Bioaccessibility of Selected Metals and Metalloids in Soils: A Critical Review, Environment Agency, Bristol, Britain, 2002. [21] B.N. Lopez, Y.B. Man, Y.G. Zhao, J.S. Zheng, A.O.W. Leung, J. Yao, M.H. Wong, Major pollutants in soils of abandoned agricultural land contaminated by e-waste activities in Hong Kong, Arch. Environ. Contam. Toxicol. 61 (2011) 101–114. [22] S.E. Allen, Chemical Analysis of Ecological Materials, second ed., Blackwell Scientific Publications, Oxford England, Boston, 1989. [23] US EPA (United States Environmental Protection Agency), Method for Determination of Total Organic Carbon (TOC) in Soils and Sediments Method, Environmental Protection Agency, Washington, 2002. [24] US EPA (United States Environmental Protection Agency), Method 3540C: Soxhlet Extraction, Environmental Protection Agency, Washington, 1996. [25] US EPA (United States Environmental Protection Agency), Method 3620B: Florisil Cleanup, Environmental Protection Agency, Washington, 1996. [26] US EPA (United States Environmental Protection Agency), Method 8270C: Semivolatile Organic Compounds by Gas Chromatography/Mass Spectrometry (GC/MS), Environmental Protection Agency, Washington, 1996. [27] US EPA (United States Environmental Protection Agency), Office of Water, Office of Science and Technology, Engineering and Analysis Division (4303T), EPA Method 1614, Brominated Diphenyl Ethers in Water, Soil, Sediment and Tissue by HRGC-HRMS, Environmental Protection Agency, Washington, 2007. [28] M.H. Zheng, Progress and prospect for dioxins analysis in China, in: The China International Symposium on Persistent Toxic Substances, Beijing, China, 2004, pp. 7–11. [29] US EPA (United States Environmental Protection Agency), Exposure Factors Handbook. EPA/600/P-95/002F, Environmental Protection Agency, Washington, 1997. [30] US EPA (United States Environmental Protection Agency), Risk Assessment Guidance for Superfund, Vol. I: Human Health Evaluation Manual (F, Supplemental Guidance for Inhalation Risk Assessment) EPA/540/R/070/002, Environmental Protection Agency, Washington, 2009. [31] US EPA (United States Environmental Protection Agency), Supplemental Guidance for Developing Soil Screening Levels for Superfund Sites, OSWER 9355. 4-24, Environmental Protection Agency, Office of Solid Waste and Emergency Response, Washington, 2001. [32] E.J. Calabrese, P.T. Gilbert, C.E. Kostecki, How much dirt do children eat? An emerging environmental health question, Comment. Toxicol. 1 (1987) 229–241. [33] M.M. Lee, A. Wuwilliams, A.S. Whittemore, S. Zheng, R. Gallagher, C.Z.et al. Teh, Comparison of dietary habits physical-activity and body-size among Chinese in North-America and China, Int. J. Epidemiol. 23 (1994) 984–990. [34] US EPA (United States Environmental Protection Agency), Risk Assessment Guidance for Superfund, Vol. I: Human Health Evaluation Manual, EPA/540/189/002, Environmental Protection Agency, Office of Solid Waste and Emergency Response, Washington, 1989. [35] US EPA (United States Environmental Protection Agency), Mid Atlantic Risk Assessment, Regional Screening Level (RSL) Summary Table, Environmental Protection Agency, Washington, 2010. [36] ATSDR (New York State Department of Health under cooperative agreement with the Agency for Toxic Substances and Disease Registry), Public Health Assessment. Johnstown City Landfill, Johnstown, Fulton Country, (CERCLIS NO. NYD980506927), Department of Health and Human Services, Public Health Service, Atlanta, 1995. [37] M.V. Ruby, A. Davis, R. Schoof, S. Eberle, C.M. Sellstone, Estimation of lead and arsenic bioavailability using a physiologically based extraction test, Environ. Sci. Technol. 30 (1996) 422–430. [38] X. Tang, L. Tang, Y. Zhu, B. Xing, J. Duan, M. Zheng, Assessment of the bioaccessibility of polycyclic aromatic hydrocarbons in soils from Beijing using an in vitro test, Environ. Pollut. 140 (2006) 279–285. [39] A.O.W. Leung, W.J. Luksemburg, A. Wong, A. Wong, Spatial distribution of polybrominated diphenyl ethers and polychlorinated dibenzo-p-dioxins and dibenzofurans in soil and combusted residue at guiyu, an electronic waste recycling site in Southeast China, Environ. Sci. Technol. 41 (2007) 2730–2737. [40] S.C. Lema, J.T. Dickey, I.R. Schultz, P. Swanson, Dietary exposure to 2,2,4,4tetrabromodiphenyl ether (PBDE-47) alters thyroid status and thyroid hormone–regulated gene transcription in the pituitary and brain, Environ. Health Perspect. 116 (2008) 12. [41] J. Cheng, J. Gu, J. Ma, X. Chen, M. Zhang, W. Wang, Neurobehavioural effects, redox responses and tissue distribution in rat offspring developmental exposure to BDE-99, Chemosphere 75 (2009) 963–968.

Y.B. Man et al. / Journal of Hazardous Materials 195 (2011) 92–99 [42] H.B. Zhang, Y.M. Luo, M.H. Wong, Q.G. Zhao, G.L. Zhang, Concentration and possible sources of polychlorinated biphenyls in the soils of Hong Kong, Geoderma 138 (2007) 244–251. [43] N. Ren, M. Que, Y.F. Li, Y. Liu, X. Wan, D. Xu, E. Sverko, J. Ma, Polychlorinated biphenyls in Chinese surface soils, Environ. Sci. Technol. 41 (2007) 3871–3876. [44] W. Wilcke, W. Amelung, Persistent organic pollutants in native grassland soils along a climosequence in North America, Soil Sci. Soc. Am. J. 64 (2000) 2140–2148. [45] S.N. Meijer, W.A. Ockenden, A. Sweetman, K. Breivik, J.O. Grimalt, K.C. Jones, Global distribution and budget of PCBs and HCB in background surface soils: Implications for sources and environmental processes, Environ. Sci. Technol. 37 (2003) 667–672. [46] CCME (Canadian Council of Ministers of the Environment), Canadian Environmental Quality Guideline, CCME, Canada, 1991. [47] VROM (Volkshuisvesting, Ruimtelijke Ordening en Milieubeheer) (Ministry of Housing, Spatial Planning and Environment), Circular on Target Values and Intervention Values for Soil Remediation, Spatial Planning and Environment, Ministry of Housing, Netherlands, 2000. [48] G. Zhao, Z. Wang, M.H. Dong, K. Rao, J. Luo, D. Wang, J. Zha, S. Huang, Y. Xu, M. Ma, PBBs, PBDEs, and PCBs levels in hair of residents around e-waste disassembly sites in Zhejiang Province, China, and their potential sources, Sci. Total Environ. 397 (2008) 46–57. [49] United Nations Environment Programme (UNEP), Draft Guidelines on BAT for Smouldering Copper Cable, UNEP/POPS/EGB.2/INF/12, UNEP, Kenya, 2003. [50] S.A. Boyd, S. Sun, Residual petroleum and polychlorobiphenyl oils as sorptive phases for organic contaminants in soils, Environ. Sci. Technol. 24 (1990) 142–144. [51] B.N. Zegers, W.E. Lewis, K. Booij, R.H. Smittenberg, W. Boer, J.P.B.J. deBoon, Levels of polybrominated diphenyl ether flame retardants in sediment cores from Western Europe, Environ. Sci. Technol. 37 (2003) 3803–3807. [52] ATSDR (Agency for toxic substances and disease registry), Polybrominated Diphenyl Ethers, Agency for Toxic Substances and Disease Registry, US Department of Health and Human Services, Public Health Service, Atlanta, 2004. [53] US EPA (United States Environmental Protection Agency), Mid Atlantic Risk Assessment, Frequently Asked Questions, Environmental Protection Agency, Washington, 2011.

99

[54] Q. Luo, Z.W. Cai, M.H. Wong, Polybrominated diphenyl ethers in fish and sediment from river polluted by electronic waste, Sci. Total Environ. 383 (2007) 115–127. [55] J. Boon, W. Lewis, A. Tjoen, C. Allchin, R. Law, J. de Boer, Levels of polybrominated diphenyl ether (PBDE) flame retardants in animals representing different trophic levels of the North Sea food web, Environ. Sci. Technol. 36 (2002) 4025–4032. [56] US EPA (United States Environmental Protection Agency), Polychlorinated Biphenyls (PCBs), CASRN 1336-36-3, Environmental Protection Agency, Washington, 1997. [57] G.R. Cass, L.A. Hughes, P. Bhave, M.J. Kleeman, J.O. Allen, L.G. Salmon, The chemical composition of atmospheric ultrafine particles, Philos. Trans. R. Soc. London A: Math. Phys. Eng. Sci. 358 (2000) 2581–2592. [58] J. Ferin, Pulmonary retention and clearance of particles, Toxicol. Lett. 72 (1994) 121–125. [59] C.S. Kim, S.C. Hu, Regional deposition of inhaled particles in human lungs: comparison between men and women, J. Appl. Physiol. 84 (1998) 1834–1844. [60] P.T. Scheepers, R.P. Bos, Combustion of diesel fuel from a toxicological perspective II, toxicity, Int. Arch. Occup. Environ. Health 64 (1992) 163–177. [61] K. Donaldson, W. MacNee, Mini-review potential mechanisms of adverse pulmonary and cardiovascular effects of particulate air pollution (PM10), Int. J. Hyg. Environ. Health 203 (2001) 411–415. [62] EPD (Hong Kong Environmental Protection Department), Press Release: Test Results of Soil Samples Collected from e-Waste Workshops, Environmental Protection Department, Hong Kong, 2005. [63] Greenpeace China, Test report for analysis for polychlorinated biphenyls (PCBs), polybrominated biphenyls (PBBs), polybrominated diphenyl ether (PBDEs) and selected metal compounds, 2005. [64] W.J. Luksemburg, R.S. Mitzel, R.G. Peterson, M.M. Maier, M. Schuld, J.M. Hedin, H.D. Zhou, A.S. Wong, Polychlorinated dibenzodioxins and dibenzofurans (PCDDs/PCDFs) and polybrominated diphenylether (PBDE) levels in environmental and human hair samples around an electronic waste processing site in Guiyu, Guangdong Province, China, in: Poster presented at 22nd International Symposium on Halogenated Environmental Organic Pollutants and Persistent Organic Pollutants (POPs) (Dioxins 2002), Spanish Council for Scientific Research Laboratory for Dioxins, Barcelona, Spain, 2002.

Journal of Hazardous Materials 195 (2011) 100–106

Contents lists available at ScienceDirect

Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat

Critical assessment of suitable methods used for determination of antibacterial properties at photocatalytic surfaces ´ b,∗ , Eva Musilová b , Jiˇrí Zita a Josef Krysa a b

Institute of Chemical Technology Prague, Department of Inorganic Technology, Technická 5, CZ-166 28 Prague, Czech Republic Institute of Chemical Technology Prague, Department of Water Technology and Environmental Engineering, Technická 5, CZ-166 28 Prague, Czech Republic

a r t i c l e

i n f o

Article history: Received 21 April 2011 Received in revised form 29 July 2011 Accepted 4 August 2011 Available online 10 August 2011 Keywords: TiO2 Antibacterial activity ISO 27447:2009(E) E. coli E. faecalis

a b s t r a c t This work describes the development of methods necessary for antibacterial effect evaluation on irradiated TiO2 layers. Two methods using bacteria suspensions and the glass adhesion method (based on ISO 27447:2009(E)) were critically assessed and compared. As test bacteria gram negative Escherichia coli and gram positive Enterococcus faecalis were employed. The method using 50 cm3 of bacteria suspension is convenient for testing layers with strong antibacterial effect (prepared from powder photocatalysts). For the evaluation of the antibacterial effect of sol gel layers, the glass adhesion method based on the ISO is more appropriate than the method with 3 cm3 of bacteria suspension. The reason is that the later does not allow a distinction between the inhibition effect of TiO2 and UV light itself. Some improvements of the ISO method were suggested, namely the use of gelatinous pills (CCM) of bacteria, using saline solution instead of nutrient broth for bacteria suspension preparation and the application of selective media for bacteria cultivation. Decreasing the light intensity from 0.6 mW cm−2 to 0.2 mW cm−2 (fulfilling the requirements of the ISO) results in almost negligible effect of UV light itself, thus enabling proper testing of the antibacterial properties of TiO2 thin films. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Titanium dioxide (TiO2 ) is one of the most popular photocatalysts [1]. In the presence of ultraviolet light (UV-A), TiO2 in anatase form is capable of decomposing organic compounds and microorganisms on its surface. Due to this ability TiO2 has high potential in many fields of application, such as medicine [2], architecture and water and air purification [3,4]. So far chlorine is the most common agent for water disinfection. Inhibition of bacteria by chlorine is very fast and efficient. However, it is well-known that chlorine reacts with organic materials (humic substances) producing chloroorganic compounds (e.g. trihalomethanes (THMS)) which are considered to be carcinogenic [5,6]. This has led to the development of alternative methods for water treatment based on the interaction of a photocatalyst with UV light [6–8]. Among the photocatalysts investigated TiO2 is the most suitable because it is stable, non-toxic and relatively cheap [9–13]. Many different microorganisms are used for antibacterial tests on photocatalytic surfaces namely, Pseudomonas aeruginosa [14,18], Enterococcus faecim [14], Candida albicans [14], Staphylococcus aureus [14,19–21], Bacillus pumilus [22] and Bacillus

∗ Corresponding author. Phone +420 220 444 112; Fax: +420 220 444 410. ´ E-mail address: [email protected] (J. Krysa). 0304-3894/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jhazmat.2011.08.009

megaterium [23] but the most commonly used is Escherichia coli [14,16,17,23–30]. E. coli belongs to the group of Gram negative thermo tolerant coliform bacteria. Usually it appears in the digestive tract of humans and warm-blooded animals, where it is useful for the host (synthesising vitamins and supporting the overall balance of microorganisms in the intestines by suppressing the growth of harmful bacteria) [31]. E. coli usually remains harmlessly confined to the intestinal lumen; however, in a debilitated or immunosuppressed host, or when gastrointestinal barriers are violated, even normal “nonpathogenic” strains of E. coli can cause infection. Infections due to pathogenic E. coli may be limited to the mucosal surfaces or may disseminate throughout the body. Three general clinical syndromes result from infection with inherently pathogenic E. coli strains: (i) urinary tract infection, (ii) sepsis/meningitis, and (iii) enteric/diarrheal disease [32]. E. coli is considered as an indicator of faecal contamination and is widely used, not only as a model microorganism for physiological, biochemical and genetic experiments, but also for antibacterial tests of different chemical substances and materials. There are many papers describing the antibacterial testing of photocatalytic surfaces, but methods and conditions are often different. The most common arrangement is an experimental setup where a drop of bacterial suspension is laid on a glass support covered by TiO2 layer [14–17]. Another experimental set-up consists

J. Kr´ ysa et al. / Journal of Hazardous Materials 195 (2011) 100–106

in the bacterial suspension placed in Teflon ring placed on titania thin film [18]. The variety of the test conditions and parameters described in the literature requires the creation of a unified system for antibacterial testing. The main reason for such system is so that results obtained on different photocatalytic surfaces and in different laboratories can be easily and clearly compared. The ISO 27447:2009(E) [33] standard (Fine ceramics (advanced ceramics, advanced technical ceramics) – test method for antibacterial activity of semiconducting photocatalytic materials) was introduced in 2009, because all currently described methods for antibacterial tests of TiO2 photocatalytic layers have different procedures and conditions. The standard includes selection of suitable microorganisms and determines the conditions of the testing methods such as light intensity, amount of microorganisms and design of apparatus. The aim of the present work was the critical assessment of several methods used in laboratories worldwide for the determination of antibacterial properties of TiO2 thin films and their comparison with the ISO standard method. The special attention was given to the question of whether it is really necessary to follow all the conditions mentioned in the ISO standard or if it is possible to adjust some conditions according to the experience and facilities of each laboratory. 2. Experimental 2.1. Chemicals For bacterial suspension preparation and for bacteria cultivation NaCl (Penta, p.a.), m-FC Agar Base (Himedia), Rosolic acid (Himedia), Slanetz and Bartley Medium (Himedia), NaOH (Penta, p.a.) were used. Titanium(IV) isopropoxide (97%; Sigma-Aldrich) and tetraethyl orthosilicate TEOS (purity 98%; Fluka) were used to prepare the titania and silica in the TiO2 /SiO2 /glass films. Absolute ethanol (p.a. Penta) and ethyl acetoacetate (purity p.a. 99%; Fluka) were utlised as solvents and hydrochloric acid (p.a. 36%; Penta) and nitric acid (p.a. 65%; Penta) were employed as sol–gel catalysts. Evonik-Degussa P25 TiO2 powder was used for particulate layer preparation. 2.2. Preparation of TiO2 thin films The microscope (75 × 25 × 1 mm2 ) soda-lime glass substrates were first dip-coated (withdrawal speed: 60 mm min−1 ) into the SiO2 sol [34] to form the necessary SiO2 barrier against metal ion (mainly Na+ ) diffusion from the glass substrate into the titania film [35]. The SiO2 interlayer was calcined at 530 ◦ C for 3 h. The titania layer was then produced by subsequent dip-coating in the titania sol [34]. After the dip-coating process, the titania films were calcined at 530 ◦ C for 3 h. The resulting layers were around 250 nm thick and the amount of titania in each layer was around 0.04 mg cm−2 . Particulate layers were prepared by sedimentation of ultrasonically pretreated suspensions of P25 TiO2 (75 % of anatase, 25 % of rutile, crystalline size around 30 nm, BET surface area around 50 m2 g−1 ) on the same glass substrate as for sol–gel layers followed by calcination for 2 h at 300 ◦ C. The amount of photocatalyst deposited on the glass supports was 0.1, 0.2, 0.5 and 1.0 mg TiO2 cm−2 . 2.3. Microorganism used The tested microorganisms were Gram negative (G−) bacterium E. coli (CCM 3954) and Gram positive (G+) bacterium Enterococcus faecalis (CCM 4224). The pure cultures of bacteria were

101

obtained as gelatinous pills from the Czech Collection of Microorganisms (CCM), Masaryk University, Brno. The pills consist of the lyophilisated form of preserved bacteria (cca 108 CFU/ml) and the main composition of the protecting medium is gelatine. The pills must be stored at low temperature (+2 to +8 ◦ C) and used within 3 years. Before each test, it was necessary to dissolve the pill of bacteria for each culture in 9 cm3 of sterile saline solution (8.5 g dm−3 NaCl) and cultivate it for 24 h at 37 ◦ C. The bacterial suspension was than diluted with saline solution (10-times dilution method) to obtain the required concentration (CFU/ml) for each test. For the purpose of analysis, the bacterial suspension was diluted several times (10-times dilution method) to obtain the count of 30 colonies to 300 colonies in each Petri dish. To avoid contamination, selective medium m-FC agar for E. coli [36] and Slanetz-Bartley for E. faecalis [37] were used (selective media were chosen according to the water quality standards). Petri dishes with E. coli were than incubated for 24 h at 43 ◦ C and with E. faecalis for 48 h at 37 ◦ C. The number of colonies was counted and the results were expressed as the number of colony-forming units per millilitre (CFU/ml). 2.4. 50 cm3 test Particulate layers of P25 and sol–gel layers were placed in 50 cm3 of E. coli suspension. The scheme of the reactor is shown in Fig. 1A. The incident light intensity was 1.0 mW cm−2 (SYLVANIA Lynx CFS BLB, maximum at 365 nm) and the initial bacteria concentration was around 1 × 104 CFU/ml. During irradiation a 1 cm3 sample from the reaction solution was taken every 30 min. The sample was then diluted, cultivated and the results of the experiment were recorded as dependence of log(CFU/ml) versus time. To observe the effect of UV light itself, the clear glass substrate (blank) was also tested in 50 cm3 of E. coli suspension. 2.5. 3 cm3 test In this case the sol–gel TiO2 sample (25 mm × 30 mm) was placed in the small Petri dish (diameter 45 mm). Then the 3 cm3 of bacterial suspension (3.3 × 106 and 2.5 × 104 CFU/ml) was added and the dish was covered by a glass lid to minimize the vaporization (Fig. 1B). The whole system was placed on a platform shaker to insure mixing of the bacterial suspension in contact with TiO2 surface. In this test, the light intensity was 0.6 mW cm−2 (BLB Philips TL-D 15W, 300–400 nm, broad maximum at 365 nm). At regular time intervals, 0.1 cm3 of the irradiated cell suspension were taken, diluted and analysed. To see the effect of UV light itself, the clear glass substrate (blank) was also tested in another Petri dish. 2.6. Glass adhesion test This method is based on the ISO 27447:2009(E) standard, which works with the bacteria (E. coli) spread on the test surface (25 × 30 mm2 ) and covered by adhesive glass (24 × 24 mm2 ). This so called “sandwich” was then put in the Petri dish (diameter 45 mm) with wet paper filter and the dish was covered with the cap (Fig. 1C). The volume of the cell suspension was 0.05 cm3 and the concentration of bacteria was within the interval 2.0 × 106 to 8.0 × 106 CFU/ml. In this experimental set-up, the irradiation conditions were the same as in the 3 cm3 test (0.6 mW cm−2 ). After a given interval of time, the cap was removed and the cover glass together with the TiO2 /glass sample were shaken out in 10 cm3 of saline solution, diluted and analysed. The effect of light intensity on E. coli and E. faecalis degradation was studied for light intensities in the range 0.2 to 0.6 mW cm−2 . Different intensities of the incident light were achieved by changing the distance of the sample from the light source and also by placing the stainless steel grid in front of the light source. For the

102

J. Kr´ ysa et al. / Journal of Hazardous Materials 195 (2011) 100–106

1 2 4 5 8 B

A 25 ml method

3 ml method

1 2 3 4 5 6 7

C

Fig. 1. Schematic diagram of the antibacterial tests arrangement: (A) 50 cm3 method, (B) 3 cm3 method, (C) thin film method – according to the ISO standard [33]; 1-light source, 2-moisture preservation glass, 3-cover slide glass, 4-bacterial suspension, 5-TiO2 sample under test, 6-glass rod, 7-wet filter paper, 8-magnetic stirrer.

distances 24 and 36 cm the light intensities were 0.6 and 0.4 mW cm−2 , respectively. When the grid was used the light intensity for the distance of 36 cm was 0.2 mW cm−2 . 3. Results and discussion 3.1. Test method For the determination of the photocatalytic activity of titania layers, photochemical reactors where the TiO2 layer is placed in the solution of model dye or organic compound are often used. Such types of reactor have been commonly employed in our laboratory for the determination of photoactivity using dye Acid Orange 7 [38]. Our first approach to the antibacterial test was simply to replace the dye solution with bacterial suspension. In Fig. 2 the results of antibacterial test on sol–gel and particulate TiO2 layers of different titania loading in 50 cm3 of bacterial suspension can be seen. It is clear that the inhibition effect of the TiO2 layer on E. coli increased as the amount of titania in the layer increased. The advantages of this method and experimental setup are: (i) the effect of UV light irradiation on the inhibition of bacteria in the suspension without TiO2 layer is almost negligible, (ii) the bacterial suspension is stable in dark even when the TiO2 layer is present, (iii) samples of bacteria suspension can be taken during the experiment, (iv) only one sample of TiO2 layer is necessary for the whole experiment (in the glass adhesion test – the ISO standard – one TiO2 sample is necessary for each point of the CFU dependence on the time of irradiation). It seems, that the 50 cm3 method is an ideal test for particulate TiO2 layers with high activity prepared from powder suspensions.

−1

Survival E. coli / log (CFU ml )

3.5 3.0 2.5 2.0

−2

P25 1.0 mg cm −2 P25 0.5 mg cm −2 P25 0.2 mg cm −2 P25 0.1 mg cm just UV Light dark −2 Sol-gel 0.043 mg cm

1.5 1.0 0.5 0.0 0

25

50

75

100

125

150

175

200

Time / min Fig. 2. 50 cm3 method for E. coli antibacterial test. Log scale of surviving bacteria under UV irradiation (1 mW cm−2 ) for particulate titania layers with different amount of P25 as a function of illumination time. The sol–gel layer is also included (250 nm, 0.043 mg cm−2 ).

But use of particulate layers in the practical application (antibacterial glasses and tiles) is not favoured due to the low mechanical stability. TiO2 layers prepared by the sol–gel method have much higher application potential in this field. Sol–gel layers can be applied on various surfaces, such as tiles, glass and metal surfaces and their mechanical stability, compared to powder layers, is much better. On the other hand due to the non-porous structure [39] and much smaller layer thickness (146 nm) than 0.1 mg cm−2 particulate film (thickness 800 nm) the resulting photoactivity of sol gel film measured using Acid Orange 7 as model compound is about 8 times smaller than that for particulate film [38]. Thus we can expect the similar behaviour when comparing the antibacterial properties of particulate and sol gel films. In fact, from Fig. 2 it is clear that the antibacterial activity of the sol–gel layer in the 50 cm3 test is comparable with the antibacterial effect of UV light itself. As a consequence, we had to find and verify different methods for the antibacterial testing of sol–gel layers. An experimental setup with a drop of bacterial suspension (0.2 cm3 ) pipetted onto the coated substrates has been described by Kühn et al. [14]. The volume of the drop in the “drop test” can be smaller, e.g. 0.1 cm3 [16,17], 0.07 cm3 [30] or even 0.01 cm3 [29]. This approach is simple but it has two serious drawbacks. At first, samples with the drop of bacterial suspension were not covered and drying of the drop during irradiation may take place. Secondly, the surface area of TiO2 film in contact with the bacteria drop is not properly defined. Kikuchi et al. [19] solved the drying problem by placing the TiO2 sample with the bacterial drop into a Petri dish with a small amount of water and covering it with a glass lid. However the problem with the definition of surface area remained. The above mentioned drawbacks result in a number of discrepancies as is visible, for example, from comparison of the results on TiO2 layers prepared from Degussa P25. Kühn et al. [14] observed 4 log decrease of E. coli CFU after 1 h of irradiation using a 0.2 cm3 drop (UV light had no effect on the drop of bacterial suspension). Hajková et al. [29] also described a 4 log decrease of E. coli CFU after 1 h of irradiation using a 0.01 cm3 drop, but a 2 log decrease using UV light itself was observed. As a next step, we eliminated the problem of the ill-defined area and drop drying by creating a new antibacterial test method. Our method defines the size of the tested TiO2 sample (25 × 30 mm2 ) which fits well into a small Petri dish (diameter 45 mm). Then we put 3 cm3 of bacterial suspension into the Petri dish to create a thin liquid film above the TiO2 layer and cover the whole system with a glass lid to eliminate evaporation (Fig. 1B). Using this method we decreased the volume of bacterial suspension from 50 to 3 cm3 and also the ratio of the irradiated area to the volume was changed from 1:5 (50 cm3 method) to 1:0.4 (3 cm3 method). Initially, we tried similar initial bacteria concentration as in the 50 cm3 test (2.5 × 104 CFU/ml). After irradiation of the system, we expected a faster decrease of bacteria concentration. However the results showed almost no killing of bacteria (Fig. 3). Secondly, we increased the initial concentration to the range recommended in the ISO

J. Kr´ ysa et al. / Journal of Hazardous Materials 195 (2011) 100–106

3.2. Critical assessment of the ISO standard – adhesion glass method

7 6

−1

Survival E. coli / log (CFU ml )

103

5 4 3 Survival bacteria / %

100

2 1

6

90

3.3×10 CFU ml TiO2 + UV UV

80 70

4

2.5×10 CFU ml TiO2 + UV UV

−1

60 50

0

0 0

−1

20 40 60 80 100 120 140 160 180 Time / min

15

30

45

60

75

90

105 120 135 150 165 180

Time / min Fig. 3. 3 cm3 method for the antibacterial test of E. coli (two initial concentrations). Log scale of surviving bacteria under UV irradiation (0.6 mW cm−2 ) for sol–gel titania layer and pure glass substrate as a function of illumination time. Insert diagram shows the percentage of surviving bacteria.

standard [33] (3.3 × 106 CFU/ml). Again the differences between the effects of UV light itself and irradiated TiO2 layer was not significant. This is possibly due the insufficient contact of bacteria with the TiO2 layer and the existence of “dead volumes of bacteria suspension” with small or no exchange with the volume of bacteria suspension in contact with TiO2 layer. The problem of the dead volume was solved by Sunada et al. [24] who placed a cylindrical frame directly on the TiO2 sol–gel layer and then 1 cm3 of the E. coli suspension was pipetted into it. After 1 h around 50% of the bacteria were killed (only 5% due to UV itself) [24]. If we compare this test (1 cm3 ) with our 3 cm3 test, where we have 35% of killing after 1 h and the same effect of UV (5%), it is clear that, if all the volume of bacterial suspension is in direct contact with the TiO2 layer, the photocatalytic de-activation of microorganisms is faster. Similarly to Sunada et al. [24], Dunlop et al. [40] used a silicone cylinder placed on a TiO2 layer and filled this with a 1 cm3 bacterial suspension of lower concentration (1 × 103 CFU/ml). Even though the experimental setups [24,40] were similar the difference between inhibition efficiency of the TiO2 + UV light and UV light itself is much smaller in the work of Dunlop et al. [40]. In addition to this observed discrepancy the scale of bacteria concentration used may make comparison difficult. Fig. 3 shows that a percentage scale shows a decrease of viable bacteria, but a log scale suggests negligible antibacterial effect. It seems that a log scale is more suitable for confirmation of photocatalytic inhibition effect of TiO2 layers, but for evaluation of the effect of UV itself the percentage scale is more useful.

In the next step, we adapted our experimental setup according to the ISO [33]. However in our laboratory we are not able to fulfil all the recommendations and requirements of the ISO. In Table 1 we show the differences between the ISO and our own glass adhesion test. The differences are in detail discussed in the following three paragraphs. At first, according to the ISO standard, E. coli (G−) is the species of bacterium recommended for the tests (glass adhesion method), but other types of bacteria can be tested, if necessary. In our work we used E. faecalis (G+) as the second test microorganism. The preparation of microorganism suspension according to the ISO is complicated and time consuming (repeated subcultures with one month expiration, many cultivations and dilutions before each experiment). Using the gelatinous pills (CCM) has many advantages: after 24 h the bacterial suspension is ready for the experiment, the concentration of bacteria in the pill is guaranteed, the pill can be stored for 3 years, the purity of bacterium strain is also guaranteed and, finally, it is easy to use. Secondly, according to the ISO standard, nutrient broth must be used for the preparation of the bacterial suspension. However we think that saline solution is better than nutrient broth because it does not contain organic compounds (meat extract and peptone in nutrient broth) which could also be photocatalytically degraded by the TiO2 layers during the test and thus slow down the rate of bacteria inactivation. According to the ISO standard, nutrient agar must be used for bacteria cultivation. From a microbiological point of view, this is not the best choice because of possible contamination from the surrounding environment. For this reason we are using selective media in our laboratory. Finally, according to the ISO standard, the specimen size should be 50 × 50 ± 2 mm2 and the size of adhesive glass should be 40 × 40 ± 2 mm2 . It is also possible to use a different specimen size but the specimen surface must be covered by adhesive glass of dimension in the range from 400 mm2 to 1600 mm2 . Our specimen size was 24 × 30 mm2 and the area covered by adhesive glass was 576 mm2 . The volume of the cell suspension spread on the specimen was different from that mentioned in ISO standard and was adjusted according to the size of the adhesive glass. The amount of bacteria was the same as the concentration recommended in the ISO standard (∼2 × 106 CFU/ml). Fig. 4 shows the results of the adhesive glass test. It is apparent that the difference between the antibacterial effect of the TiO2 layer and UV light itself is much higher than in the case of the 3 cm3 test. The explanation is that in the adhesive glass test, the irradiated surface of TiO2 is in direct contact with bacteria. Comparing the drop test and glass adhesion test the later seems to be more appropriate for TiO2 thin films. In the present adhesive glass test

Table 1 List of significant parameters of ISO 27447:2009(E) method and their comparison with the glass adhesion method used in our laboratory.

Bacterium Bacteria suspension preparation Bacteria cultivation Specimen size Sample size covered by adhesive glass Volume of test bacterial suspension Initial bacteria concentration Exposure time Light source UV light intensity

ISO 27447:2009 (E)

Our laboratory adhesion glass test

Staphylococcus aureus (G+) Escherichia coli (G−) Cultivation in nutrient broth Nutrient agar

Enterococcus faecalis (G+) Escherichia coli (G−) Gelatinous pill m-FC agar (E. coli) Medium Slanetz-Bartley (E. faecalis) 24 × 24 mm2 576 mm2 0.05 cm3 2 × 106 CFU/ml 3h Fluorescent BLB lamp 300–400 nm 0.2–0.6 mW cm−2

50 × 50 ± 2 mm2 400–1600 mm2 0.15 cm3 6.7 × 105 –2.6 × 106 CFU/ml 4–8 h Fluorescent BLB lamp 300–400 nm 0.001–0.25 mW cm−2

104

J. Kr´ ysa et al. / Journal of Hazardous Materials 195 (2011) 100–106

7.0

5.5

−1

6.0 5.5

Survival bacteria / %

5.0 4.5 4.0 3.5

100 90 80 70 60 50 40 30 20 10 0

6

−1

8×10 CFU ml TiO2 + UV UV

5.0 4.5 4.0 100

Survival bacteria / %

−1

Survival E. coli / log (CFU ml )

6.5

Survival E. coli / log (CFU ml )

a

3.5 1

60 40

20 40 60 80 100 120 140 160 180

3.0 30

45

60

75

0

90

30

45

60

105

120

135

150

165

180

5.5 −1

5.0

4.5 100

Survival bacteria / %

Light intensity is an important parameter in the antibacterial tests. Firstly, light intensity is one of the rate determining steps in semiconductor photocatalysis. Secondly, UV light itself (especially of low wavelengths) may inactivate bacteria. In the ISO standard fluorescent black light blue (BLB) lamps are recommended (wavelength 300–400 nm, light intensity 0.001–0.25 mW cm−2 ). However the light sources and their intensities and wavelengths employed in reported antibacterial tests [6,8,10,11,15,21] often vary and, as a consequence, comparison of results is very difficult. For example, Soken et al. [6] describe E. coli disinfection using a AgTiO2 suspension and a UV light intensity of 5.8 mW cm−2 (300 W, 294 nm). Sunada et al. [15] studied the photocatalytic inhibition of E. coli on TiO2 thin films by BLB lamp (15 W, 365 nm, 1.0 mW cm−2 ). Wu et al. [21] used a metal halogen desk lamp to investigate disinfection induced by visible light. The light intensity below 400 nm was less than 0.01 mW cm2 and the visible light intensity was in the range 1.6 mW cm−2 to 0.4 mW cm−2 [21]. Ibanez et al. [10] used an UV-A lamp (maximum at 365 nm) for studying the antibacterial effect of TiO2 (P25) suspension on different Gram(−) microorganisms. Because of the high sensitivity of P. aeruginosa to UV-A, suspensions of these bacteria were exposed to a lower UV-A intensity, i.e. 1.4 mW cm2 . For other microorganisms (E. coli, Salmonella typhimurium, Enterobacter cloacae) a light intensity of 5.5 mW cm−2 was chosen. Benabbou et al. [11] used HPK 125 W light to investigate the disinfection of E. coli in TiO2 suspension. Appling an optical filter they were able to work in UVC, UVB and UVA wavelength regions. In the case of UVA light, the intensity varied from 0.48 mW cm−2 to 3.85 mW cm−2 by virtue of the distance from the light source and the presence of the appropriate grid [11]. It must be emphasized that in all the above mentioned cases, the intensity of the light sources did not fit the interval set by the ISO standard [33]. Fig. 5 shows the results of adhesion glass tests (A – E. coli, B – E. faecalis) using three different light intensities (0.6, 0.4 and 0.2 mW cm−2 ). The lowest value, (0.2 mW cm−2 ) fulfils the ISO standard. It can be seen that the effect of light intensity on E. coli and

90

b Survival E. faecalis / log (CFU ml )

3.3. Influence of light intensity

75

Time / min

Time / min

60% of the bacteria were killed after 20 min and after 2 h almost 99% of the surface was disinfected. On the other hand the drop test (100 ␮l, 106 CFU/ml) on sol–gel layers shows inactivation of only 30% of viable bacteria after 3 h [16]

0.2 mW cm TiO + UV UV

Time / min

15

105 120 135 150 165 180

Fig. 4. Adhesion glass method for the antibacterial test of E. coli. Log scale of surviving bacteria under UV irradiation (0.6 mW cm−2 ) for titania sol–gel layer and pure glass substrate as a function of illumination time. Insert diagram shows the percentage of surviving bacteria.

0.4 mW cm TiO + UV UV

20 40 60 80 100 120 140 160 180

0 0

15

0.6 mW cm TiO + UV UV

20 0

0

Time / min

0

80

4.0

3.5 1

80 60 40

0.6 mW cm TiO + UV UV

20 0

0

0.4 mW cm TiO + UV UV

0.2 mW cm TiO + UV UV

20 40 60 80 100 120 140 160 180

Time / min

0 0

15

30

45

60

75

90

105 120 135 150 165 180

Time / min Fig. 5. Log scale of surviving bacteria under UV irradiation for titania sol–gel layer and pure glass substrate as a function of illumination time – adhesive glass method. (a) E. coli (initial bacteria concentration – 3.8 × 105 CFU/50 ␮l). (b) E. faecalis (initial bacteria concentration – 3.2 × 105 CFU/50 ␮l). Open symbols – UV light itself, full symbols – UV light + TiO2 , light intensity 0.6 mW cm−2 (), 0.4 mW cm−2 (♦) and 0.2 mW cm−2 ().

E. faecalis inactivation is different. In the case of Gram(−) bacterium E. coli, the effect of UV light itself on the bacteria inhibition decreased as the light intensity decreased (Fig. 5A). The percentage of surviving bacteria after 180 min irradiation increased from 38% for the highest light intensity (0.6 mW cm−2 ) to 77% for the lowest light intensity (0.2 mW cm−2 ). After 60 min irradiation the difference was even higher: 40% for 0.6 mW cm−2 and almost 90% for 0.2 mW cm−2 . In the case of Gram(+) bacterium E. faecalis, a decrease in UV light intensity did not have such a definite effect on the bacteria inhibition (Fig. 5B). The percentage of surviving bacteria after 60 min irradiation was around 80% for all UV light intensities. Even after 180 min irradiation the effect of UV light was not as strong as observed in the case of E. coli. The effect of UV light on the amount of surviving E. faecalis has moved from 35% (highest intensity – 0.6 mW cm−2 ) to 46% (lowest intensity – 0.2 mW cm−2 ). According to our experiments 60 min is the minimum irradiation time necessary to distinguish the antibacterial effect of TiO2 from the effect of UV light itself. In the case of E. coli (light intensity 0.2 mW cm−2 ), we observed 10% inhibition by UV light and 60% by TiO2 layer (after 60 min irradiation). When a higher intensity was used (0.6 mW cm−2 ), 60% of the bacteria were killed only by UV, but with a TiO2 layer more than 95% bacteria were inactivated (Fig. 5A). This trend (increasing light intensity) is consistent

J. Kr´ ysa et al. / Journal of Hazardous Materials 195 (2011) 100–106

with recent results of Dunlop et al. [40] who observed, after 40 min irradiation (UVA, 3 mW cm−2 ), 70% inhibition by UV light and 90% inhibition when TiO2 layer was applied. A strong effect of UV light itself was also observed by Foster et al. [41] but in the log scale (2 log decrease after 6 h for 2 mW cm−2 ) while for an order lower light intensity (the ISO) the inhibition due to UV light itself was only 60%. In the case of E. faecalis, for all studied light intensities after 60 min of irradiation 20% of bacteria was inactivated by UV light itself and around 55% was inactivated by using an irradiated TiO2 layer (Fig. 5B). Thus E. faecalis are not as sensitive to UVA light as E. coli. On the other hand their inactivation proceeds with similar rate as that of E. coli. The results are important because there are few other few other data concerning the photocatalytic degradation of E. faecalis on TiO2 . Only Malato et al. [42], in his review, reports that bacterium E. faecalis is generally more difficult to disinfect than E. coli and Mitoraj et al. [43] confirm this experimentally, but for the case of VIS light irradiation. 4. Conclusions Gramnegative E. coli and gram positive E. faecalis were found to be suitable for antibacterial effect evaluation on irradiated TiO2 layers. It was found that the method using 50 cm3 of bacteria suspension is convenient for testing layers with strong antibacterial effect (prepared from powder photocatalysts). A decrease in the bacteria suspension volume to 3 cm3 did not bring the expected result (improvement of the difference between antibacterial effect of irradiated TiO2 and UV light itself). The possible reason is insufficient contact of bacteria with the TiO2 layer and the existence of “dead volumes of bacteria suspension” with small or no exchange with the suspension adjacent to the TiO2 layer. Thus for evaluation of the antibacterial effect of transparent sol gel layers the adhesion glass method based on the ISO standard is the most appropriate. Some parameters stated in ISO 27447:2009(E) can be adapted according to the working conditions used in particular laboratories (sample size, type of microorganism, irradiation time). Furthermore we suggest some improvements: (i) the use of gelatinous pills (CCM) of bacteria leading to simplicity and reproducibility, (ii) the use of saline solution instead of nutrient broth for bacteria suspension preparation, (iii) the application of selective media instead of nutrient agar for bacteria cultivation. Experiments at three UV light intensities (0.2–0.6 mW cm−2 ) confirm the inhibition effect of UV light (even at 365 nm) itself. The lowest value of 0.2 mW cm−2 , fulfilling the requirements of the ISO standard, and irradiation time 60 min was found to be optimal for testing. Acknowledgements The authors acknowledge financial support (project 1M0577) of the Ministry of Education, Youth and Sport of the Czech Republic, the Grant Agency of the Czech Republic (project number 104/08/0435) and the FP7 EU project PILGRIM (No.: 223050). The authors gratefully acknowledge the English correction done by Prof. A.A. Wragg from Exeter University, UK. References [1] Q. Li, S. Mahendra, D.Y. Lyon, L. Brunet, M.V. Liga, D. Li, P.J.J. Alvarez, Antimicrobial nanomaterials for water disinfection and microbial control: potential applications and implications, Water Res. 42 (2008) 4591–4602. [2] M. Yoshinari, Y. Oda, T. Kato, K. Okuda, Influence of surface modifications to titanium on antibacterial activity in vitro, Biomaterials 22 (2001) 2043–2048. [3] A. Paleologou, H. Marakas, Ni.P. Xekoukoulotakis, A. Moya, Y. Vergara, N. Kalogerakis, P. Gikas, D. Mantzavinos, Disinfection of water and waste water by TiO2 photocatalysis, sonolysis and UV-C irradiation, Catal. Today 129 (2007) 136–142.

105

[4] A. Vohra, D.Y. Goswami, D.A. Deshpande, S.S. Block, Enhanced photocatalytic inactivation of bacterial spores on surfaces in air, J. Ind. Microbiol. Biotechnol. 32 (2005) 364–370. [5] J.A.H. Melián, J.M.D. Rodríguez, A.V. Suárez, E.T. Rendón, C.V.d. Campo, J. Arana, ˇ The photocatalytic disinfection of urban waste waters, Chemosphere J.P. Pena, 41 (3) (2000) 323–327. [6] M. Sökmen, F. Candan, Z. Sümerb, Disinfection of E. coli by the Ag-TiO2 /UV system: lipid peroxidation, J. Photochem. Photobiol. A: Chem. 143 (2–3) (2001) 241–244. [7] R. van Grieken, J. Marugán, C. Sordo, P. Martínez, C. Pablos, Photocatalytic inactivation of bacteria in water using suspended and immobilized silver-TiO2 , Appl. Catal. B: Environ. 93 (2009) 112–118. [8] N. Lydakis-Simantiris, D. Riga, E. Katsivela, D. Mantzavinos, N.P. Xekoukoulotakis, Disinfection of spring water and secondary treated municipal waste water by TiO2 photocatalysis, Desalination 250 (2010) 351–355. [9] J.M.C. Robertson, P.K.J. Robertson, L.A. Lawton, A comparison of the effectiveness of TiO2 photocatalysis and UVA photolysis for the destruction of three pathogenic micro-organisms, J. Photochem. Photobiol. A: Chem. 175 (2005) 51–56. ˜ [10] J.A. Ibánez, M.I. Litter, R.A. Pizarro, Photocatalytic bactericidal effect of TiO2 on Enterobacter cloacae: comparative study with other Gram (−) bacteria, J. Photochem. Photobiol. A: Chem. 157 (2003) 81–85. [11] A.K. Benabbou, Z. Derriche, C. Felix, P. Lejeune, C. Guillard, Photocatalytic inactivation of Escherischia coli: effect of concentration of TiO2 and microorganism, nature, and intensity of UV irradiation, Appl. Catal. B: Environ. 76 (2007) 257–263. [12] H.-L. Liu, T.C.-K. Yang, Photocatalytic inactivation of Escherichia coli and Lactobacillus helveticus by ZnO and TiO2 activated with ultraviolet light, Process Biochem. 39 (2003) 475–481. [13] D.M.A. Alrousan, P.S.M. Dunlop, T.A. McMurray, J.A. Byrne, Photocatalytic inactivation of E. coli in surface water using immobilised nanoparticle TiO2 films, Water Res. 43 (2009) 47–54. [14] K.P Kühn, Ir.F. Chaberny, K. Massholder, M. Stickler, V.W. Benz, H.-G. Sonntag, L. Erdinger, Disinfection of surfaces by photocatalytic oxidation with titanium dioxide and UVA light, Chemosphere 53 (2003) 71–77. [15] K. Sunada, T. Watanabe, K. Hashimoto, Studies on photokilling of bacteria on TiO2 thin film, J. Photochem. Photobiol. A: Chem. 156 (2003) 227–233. [16] C.C. Trapalis, P. Keivanidis, G. Kordas, M. Zaharescu, M. Crisan, A. Szatvanyi, M. Gartner, TiO2 (Fe3+ ) nanostructured thin films with antibacterial properties, Thin Solid Films 433 (2003) 186–190. [17] O. Akhavan, R. Azimirad, Photocatalytic property of Fe2 O3 nanograin chains coated by TiO2 nanolayer in visible light irradiation, Appl. Catal. A: Gen. 369 (2009) 77–82. [18] P. Amézaga-Madrid, G.V. Nevárez-Moorillón, E. Orrantia-Borunda, M. Miki-Yoshida, Photoinduced bactericidal activity against Pseudomonas aeruginosa by TiO2 based thin films, FEMS Microbiol. Lett. 211 (2002) 183–188. [19] Y. Kikuchi, K. Sunada, T. Iyoda, K. Hashimoto, A. Fujishima, Photocatalytic bactericidal effect of TiO2 thin films: dynamic view of the active oxygen species responsible for the effect, J. Photochem. Photobiol. A: Chem. 106 (1997) 51–56. [20] X. Zhao, Q. Zhao, J. Yu, B. Liu, Development of multifunctional photoactive selfcleaning glasses, J. Non-Cryst. Solids 354 (2008) 1424–1430. [21] P. Wu, R. Xie, J.A. Imlay, J.K. Shang, Visible-light-induced photocatalytic inactivation of bacteria by composite photocatalysts of palladium oxide and nitrogen-doped titanium oxide, Appl. Catal. B: Environ. 88 (2009) 576–581. [22] J.C. Yu, W. Ho, J. Lin, H. Yip, P.K. Wong, Photocatalytic activity, antibacterial effect and photoinduced hydrophilicity of TiO2 films coated on a stainless steel substrate, Environ. Sci. Technol. 37 (2003) 2296–2301. [23] G. Fu, P.S. Vary, C.-T. Lin, Anatase TiO2 nanocomposites for antimicrobial coatings, J. Phys. Chem. B 109 (2005) 8889–8898. [24] K. Sunada, Y. Kikuchi, K. Hashimoto, A. Fujishima, Bactericidal and detoxification effects of TiO2 thin film photocatalysts, Environ. Sci. Technol. 32 (1998) 726–728. [25] W. Zhang, Y. Chen, S. Yu, S. Chen, Y. Yin, Preparation and antibacterial behavior of Fe3+ -doped nanostructured TiO2 thin films, Thin Solid Films 516 (2008) 4690–4694. [26] L. Caballero, K.A. Whitehead, N.S. Allen, J. Verran, Inactivation of Escherichia coli on immobilized TiO2 using fluorescent light, J. Photochem. Photobiol. A: Chem. 202 (2009) 92–98. [27] N. Baram, D. Starosvetsky, J. Starosvetsky, M. Epshtein, R. Armon, Y. Ein-Eli, Enhanced inactivation of E. coli bacteria using immobilized porous TiO2 photoelectrocatalysis, Electrochim. Acta 54 (2009) 3381–3386. [28] R.v. Grieken, J. Marugán, C. Sordo, C. Pablos, Comparison of the photocatalytic disinfection of E. coli suspensions in slurry, wall and fixed-bed reactors, Catal. Today 144 (2009) 48–54. ˇ J. Krumeich, P. Exnar, A. Kolouch, J. Matouˇsek, P. Koˇcí, [29] P. Hájková, P. Spatenka, Antibacterial effect of metal modified TiO2/PECVD films, Eur. Phys. J. D 54 (2009) 189–193. [30] F.R. Marciano, D.A. Lima-Oliveira, N.S. Da-Silva, A.V. Diniz, E.J. Corat, V.J. TravaAiroldi, Antibacterial activity of DLC films containing TiO2 nanoparticles, J. Colloid Interface Sci. 340 (2009) 87–92. [31] M.T. Madigan, J.M. Martinko, P.V. Dunlap, D.P. Clark, Brock Biology of Microorganisms, 12th edn., Pearson Education Inc., Glenview, IL, USA, 2009, ISBN 978-0321-53615-0. [32] J.P. Nataro, J.B. Kaper, Diarrheagenic Escherichia coli, Clin. Microbiol. Rev. (1998) 142–201.

106

J. Kr´ ysa et al. / Journal of Hazardous Materials 195 (2011) 100–106

[33] ISO 27447:2009(E), Fine ceramics (advanced ceramics, advanced technical ceramics) — test method for antibacterial activity of semiconducting photocatalytic materials, first edition 2009-06-01. ´ A. Mills, Correlation of oxidative and reductive dye bleach[34] J. Zita, J. Krysa, ing on TiO2 photocatalyst films, J. Photochem. Photobiol. A: Chem. 203 (2009) 119–124. ´ Multilayer TiO2 /SiO2 thin sol–gel films: effect of [35] J. Zita, J. Maixner, J. Krysa, calcinations temperature and Na+ diffusion, J. Photochem. Photobiol. A: Chem. 216 (2010) 194–200. [36] ISO 9308-1:2000, Water quality – enumeration of Escherichia coli and coliform bacteria. [37] ISO 7899-2:2000, Water quality – detection and enumeration of intestinal enterococci. ´ ´ Photocatalytic degradation of Acid Orange 7 [38] M. Zlámal, J. Krysa, J. Jirkovsky, on TiO2 films prepared from various powder catalysts, Catal. Lett. 133 (2009) 160–166.

[39] J. Zita, J. Krysa, U. Cernigoj, U. Lavrencic Stangar, J. Jirkovsky, J. Rathousky, Photocatalytic properties of different TiO2 thin films of various porosity and titania loading, Catal. Today 161 (2011) 29–34. [40] P.S.M. Dunlop, C.P. Sheeran, J.A. Byrne, M.A.S. McMahon, M.A. Boyle, K.G. McGuigan, Inactivation of clinically relevant pathogens by photocatalytic coatings, J. Photochem. Photobiol. A: Chem. 216 (2010) 303– 310. [41] H.A. Foster, D.W. Sheel, P. Sheel, P. Evans, S. Varghese, N. Rutschke, H.M. Yates, Antimicrobial activity of titania/silver and titanoa/copper films prepared by CVD, J. Photochem. Photobiol. A: Chem. 216 (2010) 283–289. ˜ M.I. Maldonado, J. Blanco, W. Gernjak, Decon[42] S. Malato, P. Fernández-Ibánnez, tamination and disinfection of water by solar photocatalysis: recent overview and trends, Catal. Today 147 (2009) 1–59. [43] D. Mitoraj, A. Jánczyk, M. Strus, H. Kisch, G. Stochel, P.B. Heczko, W. Macyk, Visible light inactivation of bacteria and fungi by modified titanium dioxide, Photochem. Photobiol. Sci. 6 (2007) 642–648.

Journal of Hazardous Materials 195 (2011) 107–114

Contents lists available at ScienceDirect

Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat

CO2 sequestration by carbonation of steelmaking slags in an autoclave reactor E.-E. Chang a , Shu-Yuan Pan b , Yi-Hung Chen c , Hsiao-Wen Chu b , Chu-Fang Wang d , Pen-Chi Chiang b,∗ a

Department of Biochemistry, Taipei Medical University, Taipei, Taiwan Graduate Institute of Environmental Engineering, National Taiwan University, No. 71 Chou-shan Rd., Taiwan c Department of Chemical Engineering and Biotechnology, National Taipei University of Technology, Taiwan d Biomedical Engineering and Environmental Sciences, National Tsing Hua University, Taiwan b

a r t i c l e

i n f o

Article history: Received 13 May 2011 Received in revised form 3 August 2011 Accepted 4 August 2011 Available online 10 August 2011 Keywords: Accelerated carbonation Alkaline solid waste Calcite Surface coverage model Life cycle assessment

a b s t r a c t Carbon dioxide (CO2 ) sequestration experiments using the accelerated carbonation of three types of steelmaking slags, i.e., ultra-fine (UF) slag, fly-ash (FA) slag, and blended hydraulic slag cement (BHC), were performed in an autoclave reactor. The effects of reaction time, liquid-to-solid ratio (L/S), temperature, CO2 pressure, and initial pH on CO2 sequestration were evaluated. Two different CO2 pressures were chosen: the normal condition (700 psig) and the supercritical condition (1300 psig). The carbonation conversion was determined quantitatively by using thermo-gravimetric analysis (TGA). The major factors that affected the conversion were reaction time (5 min to 12 h) and temperature (40–160 ◦ C). The BHC was found to have the highest carbonation conversion of approximately 68%, corresponding to a capacity of 0.283 kg CO2 /kg BHC, in 12 h at 700 psig and 160 ◦ C. In addition, the carbonation products were confirmed to be mainly in CaCO3, which was determined by using scanning electron microscopy (SEM) and X-ray powder diffraction (XRD) to analyze samples before and after carbonation. Furthermore, reaction kinetics were expressed with a surface coverage model, and the carbon footprint of the developed technology in this investigation was calculated by a life cycle assessment (LCA). © 2011 Elsevier B.V. All rights reserved.

1. Introduction Carbon sequestration is a promising option for reducing carbon dioxide (CO2 ) emissions and alleviating global warming. Both CO2 captured from emission sources and subsequent transport of the captured CO2 to isolated reservoirs are essential for carbon sequestration. Carbon capture is affected by environmental factors, capacity, and cost. Mineral sequestration is a method of carbon capture that accelerates the natural weathering of silicate minerals, allowing them to react with CO2 to form stable products, carbonate minerals, and silica for further usage or disposal [1]. In addition, carbonation is an exothermal reaction; thus, energy consumption and costs may be limited by its inherent properties [1,2]. In all cases, the sequestration chemicals must provide base ions such as monovalent sodium and potassium, or divalent calcium and magnesium ions to neutralize the carbonic acid. Other carbonate-forming elements such as iron carbonates are not practical due to their unique and precious features [3]. In addition to controlling the reaction conditions, choosing suitable mineral feedstocks and properly designing the reactor are crucial to achieving high CO2 sequestration efficiencies.

∗ Corresponding author. Tel.: +886 2 23622510, fax: +886 2 23661642. E-mail address: [email protected] (P.-C. Chiang). 0304-3894/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jhazmat.2011.08.006

One possible feedstock for CO2 sequestration by accelerated carbonation is industrial solid waste, including steelmaking slags, combustion residues, and fly ash, which generally are alkaline and rich in calcium. The use of industrial waste is advantageous because of its low cost and widespread availability in industrial areas [4]. Interest in using industrial alkaline solid wastes as sources of calcium or magnesium oxide for CO2 sequestration has arisen because these materials are readily available, cheap, and usually produced near large-emission sources of CO2 [5]. In this study, carbonation reactions were performed primarily via the reaction of CO2 with raw CaO-based materials, and calcium carbonate (CaCO3 ) was observed to be the predominant carbonation product [6]. The use of this material simultaneously can reduce the amount of waste and neutralize a hazardous material. The objectives of this study were to investigate the carbonation of several steelmaking slags, including ultra-fine (UF) slag, fly-ash (FA) slag, and blended hydraulic slag cement (BHC), in an autoclave reactor. The effects of the operational conditions, including the type of steelmaking slag, reaction time, liquid-to-solid ratio (L/S), temperature, CO2 pressure, and initial pH, on the performance of the carbonation process were evaluated. In addition, reaction kinetics of the carbonation process were tested using a surface coverage model.

108

E.-E. Chang et al. / Journal of Hazardous Materials 195 (2011) 107–114

Fig. 1. Schematic diagram of the experimental set-up for the carbonation of steelmaking slag in an autoclave reactor. 1. CO2 gas cylinder; 2. Circulating bath; 3. Syringe pump; 4. Magnetic stirrer and heater; 5. Reactor (autoclave); 6. Thermo couple; 7. Needle valve; 8. Vent to hood.

2. Materials and methods 2.1. Experiments The aqueous carbonation of UF slag, FA slag, and BHC were conducted in an autoclave reactor that contained distilled water at a designated temperature of 40–160 ◦ C. The UF slag, FA slag, and BHC with a diameter of approximately 1 cm were provided by the CHC

100

Fresh-FA Fresh-UF Fresh-BHC

Weight Percent (%)

95

FA UF

90

Weight loss between 500 - 850 ºC

85

Weight loss due to decomposition of CaCO3

Fresh-UF Fresh-FA Fresh-BHC UF Slag FA Slag BHC

80

2.2. Composition analysis BHC

75 0

200

Resources Corporation (Kaohsiung, Taiwan). All slags were ground and sieved to less than 44 ␮m for all experiments. The BHC contains an intimate and uniform blend of Portland cement and fine granulated blast furnace (BF) slag. The BHC used in this investigation is classified as CEM III/C (∼90% BF slag content) according to EN standards [7]. A schematic diagram demonstrating the carbonation of the steelmaking slag in an autoclave reactor is shown in Fig. 1. CO2 was injected continuously into the reactor at a designated pressure and a constant flow rate. The operational factors, including the reaction time (t), liquidto-solid ratio (L/S), reaction temperature (T), CO2 pressure (P), and initial pH, systematically were varied with the various feedstocks to minimize energy and chemical consumption. After the reaction, the samples of reacted slurry immediately were filtered through a PTFE membrane filter (Millipore, 45-␮m pore size and 47 mm diameter), and then heated in an oven (105 ◦ C) for use. The conversion of the carbonation products was determined quantitatively by thermogravimetric analysis (TGA) and qualitatively by X-ray diffraction (XRD) and scanning-electron microscopy (SEM).

400

600

800

1000

Temperature (ºC) Fig. 2. TGA curves of fresh and carbonate of UF slag, FA slag, and BHC (Carbonation conditions: PCO2 = 650 psig; T = 60 ◦ C; t = 1 h; particle size < 44 ␮m; L/S = 10 mL g−1 ).

Prior to examining the capacity for CO2 capture, the chemical compositions of steelmaking slags were characterized by inductively coupled plasma atomic emission spectroscopy (ICP-AES), after total digestion using aqua regia to dissolve the solid materials in the sample, by the chemistry analysis laboratory in the China Himent Corporation. However, SiO2 was dissolved further by using hydrofluoric acid with increasing temperatures and pressures in a

E.-E. Chang et al. / Journal of Hazardous Materials 195 (2011) 107–114

microwave digestion. The contents of each metal in the extracted solution were measured by the ICP-AES method. Then, the content of metal oxide could be computed using the ICP-AES results. 2.3. TGA The thermal characteristics of the slag before and after carbonation were examined using a thermo-gravimetric analyzer (TGA-51, Shimadzu); this analysis was performed to determine the weight loss using different temperatures for the selected samples. Three weight fractions corresponding to (1) moisture, (2) organic elemental carbon, calcium hydroxide, and MgCO3 , and (3) CaCO3 content were determined mainly at the following temperature ranges: (1) 25–105 ◦ C, (2) 105–500 ◦ C, and (3) 500–850 ◦ C. The weight loss between a temperature range of 500–850 ◦ C is contributed mainly by the decomposition of CaCO3 due to its release of CO2 [6,8,9]. However, it has to be remarked that a continuous weight loss between temperatures of 105 and 1000 ◦ C is due to the dehydration of calcium hydroxide, calcium silicate hydrates, calcium aluminate hydrates, and other minor hydrates [10]. In order to prevent overestimating the CaCO3 content, the weight losses due to the dehydration of hydrates have been modified by a graphical technique and are illustrated in Fig. 2 [10]. Samples were heated linearly in the temperature range of 25–850 ◦ C at a heating rate of 10 ◦ C/min. The TGA weight fraction determined by means of a graphical technique, based on the dry weight, was assumed to be the CaCO3 content, expressed in terms of CO2 (wt%): CO2 (wt%) =

mCaCO3 m105◦ C

(1)

The carbonation conversion (ıCa ) was determined from the total calcium content of the carbonation product, assuming the initial carbonate content was negligible [8,11]: ıCa (%) = [CO2 (wt%)/100 − CO2 (wt%)] × [MWCa /MWCO2 ]/Catotal (2) where ıCa is the carbonation conversion, MWCa and MWCO2 are the molar weights of Ca and CO2 , respectively, in kg/mol, and Catotal is the total Ca content of the fresh sample in kg/kg. 2.4. SEM and XRD SEM (JSM-6500F, JEOL) was used in this study to produce highresolution three-dimensional images of the sample and to study the surface structure of the slag. SEM was useful particularly in identifying CaCO3 formed on the surface of the slag in the carbonation reaction. XRD (X’ Pert Pro, PANalytical) was used to identify and characterize the CaCO3 crystals in the carbonation products. Monochromatic X-rays were used to determine the interplanar spacing of the sample atoms using Cu as the anode material (K␣˚ at 1 wavelength = 1.540598 A˚ and K␣-2 wavelength = 1.544426 A) an angular step of 1◦ held for 1 s with 2 spanning from 20◦ to 70◦ . When the Bragg conditions for constructive interference were obtained, a “reflection” was produced in which the relative peak height was proportional to the number of grains in a preferred orientation. 2.5. Aqueous carbonation Theoretically, the extent of carbonation increases with reaction time. The aqueous carbonation experiments were conducted with reaction times of up to 720 min. The experimental procedures included the following three steps: aqueous CO2 dissolution, Ca leaching, and CaCO3 precipitation. Previous studies by Huijgen et al. [9] had indicated that the influence of the L/S ratio on carbonation was insignificant. Therefore, the L/S ratio was fixed at 10 mL g−1

109

Table 1 Physico-chemical properties of UF slag, FA slag, and BHC (CHC Resources Corporation). Parameters

UF slag

Physical properties True density (g/cm3 ) Mean diameter (␮m) BET surface area (m2 /kg) Total pore area (m2 /g) Porosity

FA slag

BHC

2.89 11.67 148 1.76 0.63

2.78 17.35 237 1.32 0.59

2.94 20.63 115 1.13 0.60

Chemical properties SiO2 (%) Al2 O3 (%) Fe2 O3 (%) CaO (%) MgO (%) S2− (%) SO3 (%)

33.93 14.35 0.35 42.43 6.42 0.24 0.52

34.89 15.75 1.97 38.80 5.59 – 0.52

27.34 8.42 2.71 52.82 4.66 – 1.49

Total (%)

98.24

97.52

97.44

in this study. The conditions of the carbonation experiment were as follows: L/S ratio of 10 mL g−1 , a PCO2 of 700 psig, particle size of less than 44 ␮m, and reaction time of 60 min, unless otherwise specified. To assess the effective initial pH of the water samples for the carbonation conversion, the pH values of the water were prepared from 2 to 12 at a fixed reaction temperature of 100 ◦ C, and a pressure of 700 psig. The initial pH value of solution was adjusted to the designed value using KOH and HNO3 solutions. Then, different steelmaking slags were dispersed intensively in the prepared water at an L/S of 10 mL g−1 . 3. Results and discussion 3.1. Physico-chemical properties of steelmaking slags The physico-chemical properties of the UF slag, FA slag, and BHC feedstocks are presented in Table 1, which shows that the major components of these three steelmaking slags were CaO: 42.43 wt%, 38.80 wt%, and 52.82 wt% for UF slag, FA slag, and BHC, respectively. Minor amounts of SiO2 , Al2 O3 , MgO, Fe2 O3 , S2− , and SO3, which do not contribute to CO2 sequestration because the CO2 -capturing capacity of the slag material is attributed mainly to the CaO components, are also listed in Table 1. A higher adsorption capacity of CO2 on the slags was expected in the carbonation reaction, which was validated by the following experiments. Assuming that all CaO was converted to CaCO3 , the theoretical capacities of the UF slag, FA slag, and BHC were 0.333, 0.305 and 0.415 kg CO2 /kg dry solid, respectively. Fig. 2 shows the weight variation of the fresh and carbonated specimen obtained by using TGA. According to the TGA results, the weight losses of the fresh UF slag, FA slag, and BHC between 500 and 850 ◦ C was insignificant which indicated that the initial hydrate and carbonate contents of the three feedstocks were negligible. In addition, the TGA curves of carbonated specimen indicate that the CO2 release at high temperatures (above 800 ◦ C) can be neglected due to the lack of a peak at 800 ◦ C in the TGA curves. The weight loss of the carbonated BHC was higher significantly than the carbonated UF and FA samples, which can be attributed to the greater CaO content and lower silicon dioxide content of the BHC. Therefore, the BHC conclusively captured a higher amount of CO2 than the other two slags during the carbonation reaction. Fig. 3 shows the SEM images of the fresh (Fig. 3a, c, e) and carbonated (Fig. 3b, d, f) UF slag, FA slag, and BHC, respectively. Comparisons of the SEM images of the feedstocks before and after carbonation showed that cubic particles adhered to the feedstocks

110

E.-E. Chang et al. / Journal of Hazardous Materials 195 (2011) 107–114

Fig. 3. Scanning electron micrographs (SEMs) of (a) fresh and (b) carbonated UF slags; (c) fresh and (d) carbonated FA slags; and (e) fresh and (f) carbonated BHC.

after carbonation. Based on the SEM images and qualitative analysis from XRD, it was determined that the cubic particles were composed of CaCO3 and had diameters ranging from 1 to 2 ␮m, which is similar to the literature [4,11]. The dark cubic particles in the SEM images were found to be CaCO3 which was confirmed by the EDX and XRD analyses. The mineralogical characterizations of fresh and carbonated UF slag, FA slag, and BHC were performed based on the XRD patterns shown in Fig. 4a–c, respectively, which indicated that the main crystal phase of the fresh slags was CaO. In contrast to the XRD results from fresh BHC, CaCO3 was identified as the primary phase in the reaction products. The peaks in the XRD analysis of the carbonated material appeared at 2 values of 23.02◦ , 29.41◦ , 35.97◦ , 43.15◦ , 47.49◦ , 48.50◦ , 57.40◦ , 60.68◦ , and 64.68◦ (in red line), which are indicative of calcium carbonate. These results suggest that the steelmaking slags should be carbonated with CO2 to form CaCO3 in an autoclave reaction.

3.2. Effects of feedstocks and operating factors The effect of reaction time on the conversion ratio of the three feedstocks at 160 ◦ C and 700 psig is shown in Fig. 5, which indicates that the carbonation rate decreased as the reaction time increased. The reaction leveled off after 60 min, indicating that the carbonation reaction had a stationary phase due to the formation of a SiO2 barrier, which strongly blocks the reactive surface sites and inhibits the release of calcium ions from the slag. These effects exhibit a limited conversion of CO2 during the carbonation reaction, which is consistent with the findings suggested by Huijgen et al. [4]. Therefore, the maximum efficiencies in carbonation conversion (ıCa ) after a reaction time of 720 min for the UF slag, FA slag, and BHC were found to be 38.1%, 34.7%, and 68.3%, respectively. The relatively higher conversion of the BHC also could be explained by the chemical compositions of the slags shown in Table 1, indicating that the CaO content of the BHC was (52.82 wt%) and thus higher than that of the UF (42.43 wt%) and FA (38.80 wt%)

E.-E. Chang et al. / Journal of Hazardous Materials 195 (2011) 107–114

a

111

100

700

Intensity

600 500

80

400 300

δCa (%)

200 100 0 20

25

30

35

40

45

50

55

60

65

70

2θ[ °]

Intensity

b

UF slag FA slag BHC

500 450 400 350 300 250 200 150 100 50 0 20

60

40

20

0 0

2

4

6

8

10

12

Reaction Time (hr)

25

30

35

40

45

50

55

60

65

70

Fig. 5. The influence of reaction time on carbonation conversion of steelmaking slags (carbonation conditions: PCO2 = 700 psig; T = 160 ◦ C; particle size 1.5 in the sediments of the Tigris River, suggesting anthropogenic impact on the metal levels in the river. The concentrations of Cr, Cu, Ni and Pb are likely to result in harmful effects on sediment-dwelling organisms which are expected to occur frequently based on the comparison with sediment quality guidelines. PCA/FA and cluster analysis suggest that As, Cd, Co, Cr, Cu, Mn, Ni and Zn are derived from the anthropogenic sources, particularly metallic discharges of the copper mine plant. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Metal contamination in aquatic environments has received huge concern due to its toxicity, abundance and persistence in the environment, and subsequent accumulation in aquatic habitats. Heavy metal residues in contaminated habitats may accumulate in microorganisms, aquatic flora and fauna, which, in turn, may enter into the human food chain and result in health problems [1,2]. Heavy metals discharged into a river system by natural or anthropogenic sources during their transport are distributed between the aqueous phase and bed sediments [3]. Because of adsorption, hydrolysis and co-precipitation only a small portion of free metal ions stay dissolved in water and a large quantity of them get deposited in the sediment [4]. Sediments are ecologically important components of the aquatic habitat and are also a reservoir of contaminants, which play a significant role in maintaining the trophic status of any water body [5]. The measurements of pollutants in the water only are not conclusive due to water discharge fluctuations and low resident time. The same holds true for the suspended material [6]. The study of sediment plays an important role as they have a long residence time. River sediments, therefore, are important sources for the assessment of man-made contamination in rivers. Sediments, not only

act as the carrier of contaminants, but also the potential secondary sources of contaminants in aquatic system [7,8]. Therefore, the analysis of river sediments is a useful method to study the metal pollution in an area [9]. The Tigris River is one of the most important rivers in Turkey. Some reports have been published on the heavy metal levels in sediment samples from the upper regions of the Tigris River [10,11]. In this paper, we report the first comprehensive study on distribution of heavy metals in sediments of the Tigris River that was accomplished through regular monitoring of the river during a period of one year at seven different sites spread over the river stretch of about 500 km. The objectives of this study were (i) to determine the spatial and temporal distributions of heavy metals in surface sediments of the Tigris River, (ii) to define the natural and/or anthropogenic sources of these metals using multivariate statistical techniques, (iii) to explore the degree of heavy metal contamination in the river using contamination indices, (iv) to assess environmental risks of these metals in the study area by comparison with sediment quality guidelines (SQGs). 2. Materials and methods 2.1. Study area

∗ Tel.: +90 412 2266046; fax: +90 412 2266052. E-mail address: [email protected] 0304-3894/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jhazmat.2011.08.051

The Tigris has been an important river throughout history and was one of the main water sources of the ancient Mesopotamian

356

M. Varol / Journal of Hazardous Materials 195 (2011) 355–364

2.2. Sampling sites Fig. 1 shows the locations of the sampling sites. Surface sediment samples were collected from seven sites, namely Maden (Site-1), E˘gil (Site-2), Diyarbakır (Site-3), Bismil (Site-4), Batman (Site-5), Hasankeyf (Site-6) and Cizre (Site-7), along the Tigris River. Uncontaminated sediment samples (USS) were also collected from a mountain stream in the study area for background studies. The brief description of sampling sites selected for this study is recorded in Table 1. 2.3. Sample collection

Fig. 1. Map showing sampling sites on the Tigris River.

Surface sediment samples were collected at monthly intervals between February 2008 and January 2009. The samples collected from each site consisted of 4–5 composite samples. Composite sediments (top 2 cm of surface) were taken by using a self-made sediment core sampler with an inner diameter of 6 cm and length of 60 cm. After sampling, the sediment samples were sealed in clean polyethylene bags, placed in a cooler at 4 ◦ C, and transported to the laboratory immediately for further analysis. 2.4. Chemical analysis

civilizations. The Tigris River originates in the Toros mountains of the Eastern Anatolia region of Turkey and follows a southeastern route to Cizre, where it forms the border between Turkey and Syria for 32 km before entering Iraq. The total length of the river is approximately 1900 km, of which 523 km is within Turkey. It drains a catchment area of about 57,614 km2 [12]. Currently, there are two major dams in operation on the Tigris River in Turkey: the Kralkızı and Dicle. Maximum flows occur from February through April, whereas minimum flows occur from August through October. The annual mean flow of the river in Diyarbakır (upstream) and Cizre (downstream) was calculated to be 28.3 m3 /sn and 211.8 m3 /sn, respectively [13]. The continental climate of the Tigris Basin is a subtropical plateau climate. The annual mean air temperature varied between 14.6 ◦ C (Maden) and 21.8 ◦ C (Cizre) with the highest and the lowest temperature of 35.9 ◦ C and 0 ◦ C, respectively. Annual total precipitation ranged from 294.1 mm Cizre (downstream) to 611.1 mm in Maden (upstream), of which 82% concentrated during the time period of October to April [14].

2.4.1. Reagents and standards All reagents were of analytical grade or of Suprapure quality (Merck, Darmstadt, Germany). Double deionized water (Milli-Q System, Millipore) was used for the preparation of all solutions. The element standard solutions used for calibration were prepared by diluting stock solutions of 1000 mg/l of each element. Stock standard solutions were Merck Certificate AA standard (Merck). All glasswares used were cleaned by soaking in dilute acid for at least 24 h and rinsed abundantly in deionized water before use. 2.4.2. Analysis of sediment samples Sediment samples were air dried; then, stones and plant fragments were removed by passing the dried sample through a 2 mm sieve. The sieved sample was powdered and finally passed through a 500 ␮m sieve and stored in acid washed and deionized water rinsed glass bottles. For heavy metal content determinations, 0.25 g sediment subsamples were digested in teflon vessels with 12 ml HNO3 (65%):HCl (37%) (3:1) mixture in a microwave oven (MARSXpress, CEM) [15]. After microwave digestion, the sample solutions

Table 1 Locations and description of sampling sites along the Tigris River. Sites

Coordinates

Altitude (m)

Description

S-1

38◦ 20’N–39◦ 41’E

860

S-2

38◦ 06’N–40◦ 08’E

616

S-3

37◦ 53’N–40◦ 13’E

576

S-4

37◦ 50’N–40◦ 39’E

538

S-5

37◦ 54 N–41◦ 05’E

540

Site-1 is located about 3 km downstream of copper mine plant in Maden Township. Wastewaters containing heavy metal from plant discharge into the river before this site. Site-2 is located about 2 km downstream of Dicle Dam in E˘gil Township. Agricultural runoff and irrigation return flow are pollution sources at this site. Site-3 is just near On Gözlü Köprü (Ten-Eyed Bridge) in Diyarbakır Province. Some wastewater drains that collect mixed domestic and industrial wastewater empty into the river before this site. Site-4 is situated just near Bismil Bridge. Wastewaters from the Diyarbakır wastewater treatment plant discharge into the river between the site-3 and site-4. In addition, domestic wastewaters from Bismil Township discharge into the river just before it reaches site-4. Agricultural runoff and irrigation return flow are other pollution sources at this site. Site-5 is situated just near Batman Bridge. Agricultural runoff is pollution source at this site. Site-6 is located near Hasankeyf Bridge. Animal manure wastes and municipal wastewater discharges from Hasankeyf Township are pollution source at this site. Site-7 is located just near Cizre Bridge. Wastewater drains from Cizre Township empty into the river directly before this site. Additional pollution sources at this site are the sand pits near the river.

◦

◦

S-6

37 42’N–41 24’E

471

S-7

37◦ 19’N–42◦ 11’E

371

M. Varol / Journal of Hazardous Materials 195 (2011) 355–364

were filtered, adjusted to a suitable volume with double deionized water. The sediment extracts were analyzed for Co, Cr, Cu, Fe, Mn, Ni and Zn by a flame atomic absorption spectrometry (FAAS) equipped with deuterium background correction (AA240FS, Varian). As, Cd and Pb in extracts were measured by using a graphite furnace atomic absorption spectrometry (GFAAS) with Zeeman background correction (AA240Z, Varian).

2.4.3. Quality control The analytical data quality was guaranteed through the implementation of laboratory quality assurance and quality control methods, including the use of standard operating procedures, calibration with standards, analysis of reagent blanks, recovery of spiked samples and analysis of replicates. The accuracy and precision of the analytical procedures were tested by recovery measurements on spiked sediment samples. The sediment samples collected as uncontaminated sediment samples were spiked with metals and digested with the same procedure as the samples. The percentage recoveries of the metals in the spiked samples ranged from 91.4% (Fe) to 105.2% (Pb). The precision of the analytical procedures, expressed as the relative standard deviation (RSD), ranged from 5 to 10%. The precision for the analysis of standard solution was better than 5%. All analyses were carried out in duplicate, and the results were expressed as the mean.

2.5. Assessment of sediment contamination In the interpretation of geochemical data, choice of background values plays an important role. Many authors have used the average shale values or the average crustal abundance data as reference baselines. The best alternative is to compare concentrations between contaminated and mineralogically and texturally comparable, uncontaminated sediments [16–18]. Since there were no data on background concentrations for the investigated Tigris sediment and soils of close areas, the background values in this paper were calculated from the mean concentrations of heavy metals in uncontaminated sediments of the study area. In this study, four different indices were used to assess the degree of heavy metal contamination in sediments of the Tigris River.

2.5.1. Contamination factor (CF) The CF is the ratio obtained by dividing the concentration of each metal in the sediment by baseline or background value (concentration in uncontaminated sediment): CF =

Cheavy metal Cbackground

CF values were interpreted as suggested by Hakanson [19], where: CF < 1 indicates low contamination; 1 < CF < 3 is moderate contamination; 3 < CF < 6 is considerable contamination; and CF > 6 is very high contamination.

2.5.2. Pollution load index (PLI) For the entire sampling site, PLI has been determined as the nth root of the product of the n CF: PLI = (CF1 × CF2 × CF3 × · · · × CFn )1/n This empirical index provides a simple, comparative means for assessing the level of heavy metal pollution. When PLI > 1, it means that a pollution exists; otherwise, if PLI < 1, there is no metal pollution [20].

357

2.5.3. Geoaccumulation index (Igeo) The geoaccumulation index (Igeo) is defined by the following equation: Igeo =

Log2 (Cn ) 1.5(Bn )

where Cn is the concentration of metals examined in sediment samples and Bn is the geochemical background concentration of the metal (n). Factor 1.5 is the background matrix correction factor due to lithospheric effects. The geoaccumulation index consists of seven classes [21]. Class 0 (practically unpolluted): Igeo ≤ 0; Class 1 (unpolluted to moderately polluted): 0 < Igeo < 1; Class 2 (moderately polluted): 1 < Igeo < 2; Class 3 (moderately to heavily polluted): 2 < Igeo < 3; Class 4 (heavily polluted): 3 < Igeo < 4; Class 5 (heavily to extremely polluted): 4 < Igeo < 5; Class 6 (extremely polluted): 5 > Igeo [22]. 2.5.4. Enrichment factor (EF) Enrichment factor (EF) is a useful tool in determining the degree of anthropogenic heavy metal pollution [16]. The EF is computed using the relationship below: EF =

(Metal/Fe) Sample (Metal/Fe) Background

In this study, iron (Fe) was used as the reference element for geochemical normalization because of the following reasons: (1) Fe is associated with fine solid surfaces; (2) its geochemistry is similar to that of many trace metals and (3) its natural concentration tends to be uniform [22]. EF values were interpreted as suggested by Sakan et al. [16], where: EF < 1 indicates no enrichment; 50 is extremely severe enrichment. 2.6. Sediment quality guidelines Sediment quality assessment guidelines (SQGs) are very useful to screen sediment contamination by comparing sediment contaminant concentration with the corresponding quality guideline [23]. These guidelines evaluate the degree to which the sediment-associated chemical status might adversely affect aquatic organisms and are designed to assist in the interpretation of sediment quality. Such SQGs have been used in numerous applications, including designing monitoring programs, interpreting historical data, evaluating the need for detailed sediment quality assessments, assessing the quality of prospective dredged materials, conducting remedial investigations and ecological risk assessments, and developing sediment quality remediation objectives [23]. The consensus-based sediment-quality guidelines (SQGs) were used in this study to assess possible risk arises from the heavy metal contamination in sediments of the study area. The consensus-based SQGs were developed from the published freshwater sediment-quality guidelines that have been derived from a variety of approaches [23]. These synthesized guidelines consist of a threshold effect concentration (TEC) below which adverse effects are not expected to occur and a probable effect concentration (PEC) above which adverse effects are expected to occur more often than not. An apparent advantage of the consensus-based guidelines is that MacDonald et al. [23] evaluated the reliability of the TECs and PECs for assessing sedimentquality conditions by determining their predictive ability that is, the ability of the guidelines to correctly classify field-collected sediments as nontoxic or toxic to one or more aquatic organisms.

358

M. Varol / Journal of Hazardous Materials 195 (2011) 355–364

Table 2 Maximum, minimum, mean and standard deviation values of heavy metals in sediments of all sites studied in the Tigris River. Sites

Maden

E˘gil

Diyarbakır

Bismil

Batman

Hasankeyf

Cizre

Metal concentrations (mg/kg)

Max Min Mean SD Max Min Mean SD Max Min Mean SD Max Min Mean SD Max Min Mean SD Max Min Mean SD Max Min Mean SD

As

Cd

Co

Cr

Cu

Mn

Ni

Pb

Zn

18.0 5.0 8.9 4.0 4.9 2.0 3.3 0.9 6.6 3.5 4.8 1.0 5.2 2.4 3.5 0.8 6.0 2.3 3.6 1.0 3.6 2.2 2.9 0.5 8.5 2.9 5.4 1.8

4.9 1.4 2.4 1.2 2.4 1.4 1.8 0.3 3.0 1.4 1.8 0.4 2.6 1.0 1.6 0.5 1.6 0.7 1.2 0.3 2.5 0.8 1.6 0.5 2.7 1.8 2.2 0.3

389.8 55.6 155.9 107.7 30.6 20.9 25.7 2.9 39.7 23.2 30.3 4.9 25.6 12.3 15.9 3.7 12.2 5.4 9.0 1.8 16.1 5.4 10.0 2.6 19.0 11.1 14.1 2.7

151.7 76.4 119.0 21.8 96.0 56.4 76.4 11.3 163.4 98.1 115.4 16.9 113.2 67.7 83.8 13.6 65.9 35.8 50.5 9.5 90.2 28.4 54.6 15.7 124.4 65.7 93.6 20.2

5075.6 673.1 1941.9 1592.3 131.6 91.6 117.0 12.4 297.2 117.5 189.7 53.4 136.3 50.8 73.9 25.9 36.4 17.2 24.1 5.7 64.2 11.2 28.5 15.5 59.2 27.7 37.3 10.4

1657.0 822.0 1233.3 268.5 752.0 540.5 629.8 63.9 787.5 556.3 663.2 70.7 1228.0 528.7 641.3 189.8 590.8 282.2 420.2 92.2 791.0 329.5 489.7 130.6 982.9 529.8 702.5 124.4

288.0 151.9 216.8 44.9 144.4 113.4 132.0 9.1 174.5 153.3 162.3 5.9 172.8 137.0 149.6 10.1 109.7 79.3 93.9 8.9 125.7 74.0 91.0 15.9 244.7 135.5 173.7 35.9

566.6 144.4 393.9 121.7 358.1 89.1 255.5 104.0 387.7 89.0 250.3 102.7 392.4 185.5 274.3 59.0 299.2 62.3 163.7 95.6 344.6 73.9 221.8 86.9 387.6 93.7 297.3 79.8

2396.6 191.3 530.4 597.9 190.6 134.0 165.2 14.5 247.0 136.0 178.2 27.6 220.8 107.3 146.1 31.7 183.1 84.8 129.6 30.3 189.0 60.1 120.5 30.6 191.2 123.1 152.1 21.0

2.7. Statistical analyses Analysis of variance (ANOVA) was performed to analyze the significant spatial and temporal differences (p < 0.05). Relationships among the considered variables were tested using Pearson’s coefficient with statistical significance set at p < 0.05. Multivariate analysis of the river data set was performed using cluster analysis (CA) and principal component analysis/factor analysis (PCA/FA) techniques. The above statistical analyses were applied to experimental data standardized through z-scale transformation to avoid misclassification due to wide differences in data dimensionality. Kaiser–Meyer–Olkin (KMO) and Bartlett’s sphericity tests were performed to examine the suitability of the data for PCA/FA [24]. KMO is a measure of sampling adequacy that indicates the proportion of variance that is common, i.e., variance that may be caused by underlying factors. A high value (close to 1) generally indicates that principal component/factor analysis may be useful, as was the case in this study, where KMO = 0.76. Bartlett’s test of sphericity indicates whether a correlation matrix is an identity matrix, which would indicate that variables are unrelated. The significance level of 0 in this study (less than

0.05) indicated that there were significant relationships among the variables. Hierarchical agglomerative cluster analysis (CA) was performed on the normalized data set using Ward’s method with squared Euclidean distances as a measure of similarity. Factor analysis (FA) was conducted after principal component analysis (PCA). PCA of the normalized variables (data set) was performed to extract significant principal components (PCs) and to further reduce the contribution of variables with minor significance; these PCs were then subjected to varimax rotation (raw) to generate varifactors (VFs).

3. Results and discussion 3.1. Heavy metals in sediments of the Tigris River The basic statistics for all of the metal parameters measured during the sampling period of one year at seven different sites are summarized in Table 2. During the study period, all heavy metals showed significant spatial variations (ANOVA, p < 0.05). The ranges of metals

Table 3 Heavy metal concentrations reported for previous studies conducted in the Tigris River. Sites

Maden E˘gil Diyarbakır Bismil Maden E˘gil Diyarbakır Bismil Diyarbakır Diyarbakır Bismil

Metal concentrations (mg/kg)

References

Cd

Co

Cu

Mn

Ni

Pb

Zn

– – – – – – – – – nd nd

– – – – 503 118 21 4 32.01 43.13 32.4

3433 1213 904 991 3433 1213 904 991 728.96 137 92.5

– – – – – – – – – 622.9 497.7

– – – – 403 305 50 41 66.35 124.5 99.51

– – – – 102 83 31 24 – nd nd

891 456 405 716 891 456 405 716 369.14 30 42.7

[27] [27] [27] [27] [10] [10] [10] [10] [28] [11] [11]

[16] [29] [30] [5] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] 54–567 24.7–45.5 178–645 8.47–343.47 86.1–708.8 78–2010 42–271 68–5280 11–221 82–3700 404–1920 160–8076 54–4380 21.09–25.66 11–123 3.3–17.3 32–98.5 4.86–156.2 39.3–189 14.7–541.8 11–140 17–13400 3.5–23.2 19–270 522–6880 62–2281 11–681 8.65–38.29 –

17–55 15.4–79.2 52.3–161 4.75–76.08 – 17.5–173.3 19–188 1.6–36 2–112 – – 9–23 6–78 – 490–2316 221–446 – 65.73–834.7 – 442–1655 – – 75–2810 410–6700 936–5240 750–14000

32–162 13.1–38.7 31.7–90.1 3.6–245.33 71.6–420.8 31.1–8088 14–93 22–2700 10–81 18–480 19.5–76.9 51–796 9–1739 6.47–178.61 –

7–23 – – 2.4–88.7 84.4–23.4 26.5–556.5 39–180 11–151 8–274 33–71 – 24–71 29–240.5 28.7–152.73 – – – 1.5–23.4 – – 6.8–42 – – – 13–24

2.08–12.90

0.12–0.55 1.08–3.7 0.09–17.83 1.0–4.3 1.1–32.9 1–11 0.13–12 – 0.1–22 0.95–5.95 0.5–31 0.5–24.8 0.03–0.37 – – – – 8.1–388 1–40 – – 2.8–31 – 21–1543

[3] 32–2200 126–345 – – 207–1660 13–66 22–47 –

–

This study This study Zn

191.3–2396 60.1–247 144.4–566.6 62.3–392.4

Pb Ni

151.9–288 74–244.7 822.0–1657 282.2–1228

Mn Cu

673.1–5075.6 11.2–297.2 76.4–151.7 28.4–163.4

Cr Co

55.6–389.8 5.4–39.7 1.4–4.9 0.7–3 5–18 2–8.5

Cd As

Metal concentrations (mg/kg)

359

Tigris River (site-1), Turkey Tigris River (the rest of sites), Turkey Shing Mun River, Hong Kong Tisza River, Serbia Yes¸ilırmak River, Turkey River Po, Italy Gomti River, India Almendares River, Cuba Danube River, Europa Axios River, Greece Tinto River, Spain Nile River, Egypt South Platte River, USA Tees River, UK Rimac River, Peru Sea Scheldt River, Belgium Luan River, China

The results of contamination factors (CFs) and pollution load index (PLI) are presented in Table 5. The highest CF values for all metals studied were found at site-1 (Maden), which receives

Locations

3.2. Indices of sediment contamination

Table 4 Heavy metal concentrations in sediment samples from the Tigris River and other selected rivers from the literature.

in sediments were: 2.0–18.0 mg/kg for As, 0.7–4.9 mg/kg for Cd, 5.4–389.8 mg/kg for Co, 28.4–163.4 mg/kg for Cr, 11.2–5075.6 mg/kg for Cu, 282.2–1657.0 mg/kg for Mn, 74.0–288.0 mg/kg for Ni, 62.3–566.6 mg/kg for Pb and 60.1–2396.6 mg/kg for Zn. The highest concentrations of heavy metals were found at site-1 (Maden) due to metallic wastewater discharges from copper mine plant in Maden Township. Site-3 (Diyarbakır) which receives untreated domestic and industrial wastewaters from Diyarbakır province, site-4 (Bismil) which receives partially treated domestic wastewater from Diyarbakır wastewater treatment plant, untreated domestic wastewater from Bismil Township and agricultural runoff, and site-7 (Cizre) which receives untreated domestic wastewater from Cizre Township had also high metal concentrations. The lowest mean values of As, Ni and Zn were found at site-6 (Hasankeyf), while the lowest mean values of Cd, Co, Cr, Cu, Fe, Mn and Pb were calculated at site-5 (Batman). In this study, total metal concentrations followed the order of site-1 > site-7 > site-4 > site-3 > site > 2 > site > 6 > site-5. During the study, all metals studied did not show significant temporal differences (ANOVA, p > 0.05). In this study, heavy metal concentrations in assessed sediment samples from the Tigris River were compared with previous studies (Table 3). The mean values of Co, Cu, Ni and Zn except Pb at site-1 (Maden) were lower when compared with an earlier study conducted in 1990 [10] due to reduction of the activity of the copper mine plant. The mean values of Co, Cu, Ni and Zn at site-2 (E˘gil) were found significantly lower than those at the same site reported for the Tigris River owing to the construction of two dams on the river over the last 10 years: Kralkızı and Dicle. It is well known that concentrations of suspended solids and heavy metals in the reservoir water will be decreased significantly due to sediment deposition. The water leaving the reservoir can be clearer, and this could affect the river downstream of the dam. However, the mean values of Co, Ni and Pb except Cu and Zn at site-3 (Diyarbakır) and site-4 (Bismil) were higher than those reported by Gümgüm et al. [10]. In this study, the mean values of Cd, Cu, Mn, Ni, Pb and Zn except Co at site-3 were found higher, while the mean values of Cd, Mn, Ni, Pb and Zn except Co and Cu at site-4 were higher when compared with a previous study conducted in 2000 [11]. The increase in some metal concentrations at site-3 (Diyarbakır) and site-4 (Bismil) may be due to increased anthropogenic activities in the Diyarbakır province which has the largest urban settlement in Tigris Basin. It may have contributed large amounts of heavy metals into the river. Total heavy metal concentrations in the sediment samples from the Tigris River followed the order: Fe > Mn > Cu > Pb > Zn > Ni > Cr > Co > As > Cd. The results were not compatible with previous studies [10,11] conducted in the Tigris River. Karadede-Akin and Ünlü [11] found that Fe was the most abundant in the sediment, followed by Mn, Cu and Co, and the least was Zn, while Cd and Pb were not recorded. Gümgüm et al. [10] reported that the accumulation order of heavy metals in the sediment samples was Cu > Zn > Ni > Co > Pb. Comparison of metal contamination data of the Tigris River with the published data of other rivers (Table 4) reveals that the sediments of site-1 are severely polluted with heavy metals, while sediments of the rest of sites are slightly polluted. The extent of metal pollution in the Tigris River was not much more serious than that in the Tinto River, Danube River and Rimac River, and much worse than the Yes¸ilırmak River, River Po, Luan River, Nile River and Axios River (Table 4).

References

M. Varol / Journal of Hazardous Materials 195 (2011) 355–364

360

M. Varol / Journal of Hazardous Materials 195 (2011) 355–364

Table 5 Metal contamination factors (CFs) and pollution load indices (PLIs) for sediments of all sites studied in the Tigris River. Sites

Contamination factors (CFs)

Site-1 Site-2 Site-3 Site-4 Site-5 Site-6 Site-7 Mean Min Max

PLI

As

Cd

Co

Cr

Cu

Fe

Mn

Ni

Pb

Zn

3.42 1.27 1.85 1.35 1.38 1.12 2.08 1.78 1.12 3.42

2.86 2.14 2.14 1.90 1.43 1.90 2.62 2.14 1.43 2.86

8.66 1.43 1.68 0.88 0.50 0.56 0.78 2.07 0.50 8.66

1.86 1.19 1.80 1.31 0.79 0.85 1.46 1.32 0.79 1.86

34.68 2.09 3.39 1.32 0.43 0.51 0.67 6.16 0.43 34.68

1.11 0.99 1.02 1.08 0.88 0.98 1.10 1.02 0.88 1.11

2.37 1.21 1.28 1.23 0.81 0.94 1.35 1.31 0.81 2.37

2.93 1.78 2.19 2.02 1.27 1.23 2.35 1.97 1.23 2.93

5.79 3.76 3.68 4.03 2.41 3.26 4.37 3.90 2.41 5.79

6.80 2.12 2.28 1.87 1.66 1.54 1.95 2.60 1.54 6.80

a huge amount of metallic discharge from copper mine plant in Maden Township. The CF values for Co, Cu and Zn were >6 in sediments of site-1, which denotes a “very high contamination” by these metals. The CF values for As and Pb in sediments of site1 showed a “considerable contamination”, while the CF values for Cd, Cr, Fe, Mn and Ni indicated a “moderate contamination”. The CF values for metals studied except Pb at other sites showed “moderate contamination”. In this study, Cu had the highest and lowest CF values among the ten metals studied. However, Pb had the highest CF values among the ten metals studied at all sites except site-1. Site-3 (Diyarbakır) which receives municipial and industrial wastewater discharges from Diyarbakır and site7 (Cizre) which receives municipial wastewater discharges from Cizre showed high CF values. Total contamination factors followed the order of site-1 > site-3 > site-7 > site-2 > site-4 > site-6 > site-5. The pollution load index (PLI) ranged from 1.02 to 4.19 (Table 5). According to the mean PLI value (1.88), the Tigris River was moderately polluted. Site-1 had the highest PLI (4.19) within the study area, indicating that the sediments of site-1 were strongly polluted by investigated heavy metals. Other sites where PLI was between 1 and 2 must be classified as moderately polluted. The PLI followed the order of site-1 > site-3 > site-7 > site-2 > site-4 > site-6 > site-5. Table 6 presents Igeo and EF values of the metals studied. The Igeo values of As at sites 2, 4, 5 and 6, Cd at site-5, Co at sites 2, 4, 5, 6 and 7, Cr and Mn at all sites except site-1, Cu at sites 4, 5, 6 and 7, Fe at all sites, and Ni at sites 5 and 6 were less than zero, suggesting that these sites were not polluted by these metals. The Igeo values for Cd, Cr, Mn and Ni were under 1 in the sediments of all sites which usually had “unpolluted to moderately polluted” class. Among ten metals studied, Cu, Co, Zn and Pb had the highest Igeo values, respectively. The highest Igeo values of metals studied were found in the sediments of site-1. The Igeo class of Cu was “extremely polluted” for sediments of site-1. The Igeo class of As and Pb were “moderately polluted” for sediments of site-1, while

4.19 1.67 1.99 1.55 1.02 1.11 1.62 1.88 1.02 4.19

the Igeo class of Co and Zn were “moderately to heavily polluted”. Total Igeo values followed the order of site-1 > site-3 > site-2 > site7 > site-4 > site-6 > site-5. According to Zhang and Liu [25], EF values between 0.05 and 1.5 indicate that the metal is entirely from crustal materials or natural processes, whereas EF values higher than 1.5 suggest that the sources are more likely to be anthropogenic. In this study, the mean EF values for all metals studied except Cr and Mn were >1.5 in the sediments of the Tigris River, suggesting anthropogenic impact on the metal levels in the river. The highest EF values were found at site-1 (Maden) due to metallic wastewater discharges from the copper mine plant in Maden Township. The EF value for Cu in the sediments of site-1 was 31.34, showing “very severe enrichment”, while the EF values for Co, Pb and Zn were between 5 and 10, indicating “moderately severe enrichment”. However, the EF values for As, Cd, Cr, Mn and Ni at site-1 indicated “minor enrichment”. Cu had the highest and lowest EF values among the ten metals studied. Co had the second highest EF value. Pb at all sites except site-1 had the highest EF values among the ten metals studied. The EF values for metals studied in sediments of other sites showed “minor to moderate enrichment”. Total EF values followed the order of site-1 > site-3 > site-2 > site-7 > site-4 > site-6 > site-5. 3.3. Application of sediment quality guidelines It is important to determine whether the concentrations of heavy metals in sediments pose a threat to aquatic life. In this study, heavy metal concentrations in assessed sediment samples were compared with consensus-based TEC and PEC values (Table 7). As, Cu and Zn were lower than the TEC in 96.4%, 28.6% and 16.7% of samples, respectively. Cd, Cr, Cu and Zn were between the TEC and PEC in 95.2%, 71.4%, 46.4% and 79.7% of samples, respectively. Ni and Pb exceeded the PEC in 100% and 83.3% of samples, respectively. Cr exceeded the PEC in 9 of samples at site-1, 6 of samples at site-3, 1 of samples at site-4 and 3 of samples at site-7. Cu exceeded the PEC

Table 6 Geoaccumulation indices (Igeo) and enrichment factors (EF) of heavy metals for sediments of all sites studied in the Tigris River. Sites

Site-1 Site-2 Site-3 Site-4 Site-5 Site-6 Site-7 Mean Min Max

As

Cd

Co

Cr

Cu

Fe

Mn

Ni

Pb

Zn

Igeo

EF

Igeo

EF

Igeo

EF

Igeo

EF

Igeo

EF

Igeo

EF

Igeo

EF

Igeo

EF

Igeo

EF

Igeo

EF

1.19 −0.24 0.30 −0.16 −0.12 −0.43 0.47 0.14 −0.43 1.19

3.09 1.29 1.81 1.25 1.57 1.14 1.90 1.72 1.14 3.09

0.93 0.51 0.51 0.34 −0.07 0.34 0.80 0.48 −0.07 0.93

2.58 2.17 2.10 1.77 1.62 1.95 2.39 2.08 1.62 2.58

2.53 −0.07 0.17 −0.76 −1.58 −1.43 −0.94 −0.30 −1.58 2.53

7.83 1.45 1.65 0.82 0.57 0.57 0.72 1.94 0.57 7.83

0.31 −0.33 0.27 −0.20 −0.93 −0.81 −0.04 −0.25 −0.93 0.31

1.68 1.21 1.77 1.21 0.90 0.87 1.34 1.28 0.87 1.77

4.53 0.48 1.18 −0.18 −1.80 −1.56 −1.17 0.21 −1.80 4.53

31.34 2.12 3.32 1.22 0.49 0.52 0.61 5.66 0.49 31.34

−0.44 −0.60 −0.56 −0.48 −0.77 −0.62 −0.45 −0.56 −0.77 −0.44

1 1 1 1 1 1 1 1 1 1

0.66 −0.31 −0.23 −0.28 −0.89 −0.67 −0.15 −0.27 −0.89 0.66

2.14 1.23 1.25 1.14 0.92 0.96 1.23 1.27 0.92 2.14

0.97 0.25 0.55 0.43 −0.24 −0.29 0.65 0.33 −0.29 0.97

2.65 1.81 2.15 1.87 1.44 1.26 2.14 1.90 1.26 2.65

1.95 1.32 1.30 1.43 0.68 1.12 1.54 1.33 0.68 1.95

5.23 3.81 3.61 3.74 2.74 3.34 3.99 3.78 2.74 5.23

2.18 0.50 0.61 0.32 0.15 0.04 0.38 0.60 0.04 2.18

6.15 2.15 2.24 1.74 1.89 1.58 1.78 2.50 1.58 6.15

M. Varol / Journal of Hazardous Materials 195 (2011) 355–364

361

Table 7 Comparison between sediment quality guidelines (SQGs) with heavy metal concentrations (mg/kg) of all sites studied in the Tigris River.

SQGs Measured values in this study Site-1

Site-2

Site-3

Site-4

Site-5

Site-6

Site-7

Total

TEC PEC Minimum Maximum Average Samples < TEC Samples between TEC and PEC Samples > PEC Samples < TEC Samples between TEC and PEC Samples > PEC Samples < TEC Samples between TEC and PEC Samples > PEC Samples < TEC Samples between TEC and PEC Samples > PEC Samples < TEC Samples between TEC and PEC Samples > PEC Samples < TEC Samples between TEC and PEC Samples > PEC Samples < TEC Samples between TEC and PEC Samples > PEC Samples < TEC Samples between TEC and PEC Samples > PEC

As

Cd

Cr

Cu

Ni

Pb

Zn

9.79 33 2.0 18.0 4.6 9 3 0 12 0 0 12 0 0 12 0 0 12 0 0 12 0 0 12 0 0 81 (96.4%) 3 (3.6%) 0 (0%)

0.99 4.98 0.7 4.9 1.8 0 12 0 0 12 0 0 12 0 0 12 0 2 10 0 2 10 0 0 12 0 4 (4.8%) 80 (95.2%) 0 (0%)

43.4 111 28.4 163.4 84.8 0 3 9 0 12 0 0 6 6 0 11 1 2 10 0 3 9 0 0 9 3 5 (6%) 60 (71.4%) 19 (22.6%)

31.6 149 11.2 5075.6 344.6 0 0 12 0 12 0 0 3 9 0 12 0 11 1 0 9 3 0 4 8 0 24 (28.6%) 39 (46.4%) 21 (25%)

22.7 48.6 74.0 288.0 145.6 0 0 12 0 0 12 0 0 12 0 0 12 0 0 12 0 0 12 0 0 12 0 (0%) 0 (0%) 84 (100%)

35.8 128 62.3 566.6 265.3 0 0 12 0 3 9 0 2 10 0 0 12 0 6 6 0 2 10 0 1 11 0 (0%) 14 (16.7%) 70 (83.3%)

121 459 60.1 2396.6 203.1 0 9 3 0 12 0 0 12 0 3 9 0 5 7 0 6 6 0 0 12 0 14 (16.7%) 67 (79.7%) 3 (3.6%)

in all of samples at site-1 and 9 of samples at site-3. Ni exceeded the PEC in all of samples. Pb exceeded the PEC in all of samples at site-1 and site-4, 9 of samples at site-2, 10 of samples at site-3 and site6, 6 of samples at site-5 and 11 of samples at site-7. Zn exceeded the PEC in 3 of samples at site-1. These results indicate that the concentrations of Cr, Cu, Ni and Pb are likely to result in harmful effects on sediment-dwelling organisms which are expected to occur frequently. An index of toxicity risk, PEC quotients, was also evaluated in this study. PEC quotients were calculated using the methods of MacDonald et al. [23]. Sediment samples are predicted to be not toxic if PEC quotients are site-3 > site-7 > site-4 > site2 > site-6 > site-5. PEC quotients of Cu at site-1, Ni and Pb at all sites exceeded 1.5, suggesting a potential toxicity of these metals in sediments of the river. Conversely, the toxicity risks were much lower for As and Cd at all sites, Cr and Cu at sites 5 and 6 and Zn at all sites except site-1, with PEC quotients 1 that explained about 83.9% of the total variance in the sediment quality data set. The first PC accounting for 54.8% of the total variance was correlated (loading >0.70) with Cd, Co, Cu, Mn and Ni. The second PC accounting for 17.4% of total variance was correlated with Fe. Whereas the third PC accounted for the total variance of 11.7%, it correlated (loading >0.70) with none of the metal parameters. Three VFs were obtained through FA performed on the PCs. The corresponding VFs, variable loadings and the explained variance are presented in Table 9. The loading plots of the first two VFs are presented in Fig. 2. VF coefficients having a correlation greater than 0.70 were considered significant (strong). VF1, which explained 56.5% of the total variance, had strong positive loadings (>0.70) on Cd, Co, Cu and Mn, and a moderate positive loading on Ni. This VF represents anthropogenic sources. In Maden Township (upstream), there is a copper mine plant that discharges

Table 8 PEC quotients of heavy metals for sediments of all sites studied in the Tigris River.

As Cd Cr Cu Ni Pb Zn Mean Min Max

Site-1

Site-2

Site-3

Site-4

Site-5

Site-6

Site-7

0.27 0.48 1.07 13.03 4.46 3.08 1.16 3.36 0.27 13.03

0.10 0.36 0.69 0.78 2.72 2.00 0.36 1.00 0.10 2.72

0.14 0.37 1.04 1.27 3.34 1.96 0.39 1.22 0.14 3.34

0.11 0.33 0.76 0.50 3.08 2.14 0.32 1.03 0.11 3.08

0.11 0.23 0.45 0.16 1.93 1.28 0.28 0.63 0.11 1.93

0.09 0.31 0.49 0.19 1.87 1.73 0.26 0.71 0.09 1.87

0.16 0.44 0.84 0.25 3.57 2.32 0.33 1.13 0.16 3.57

362

M. Varol / Journal of Hazardous Materials 195 (2011) 355–364

Table 9 Loadings of experimental variables (10) on significant principal components for the Tigris River data set* . VF1 As Cd Co Cr Cu Fe Mn Ni Pb Zn Eigenvalue % Total variance Cumulative % variance *

VF2

Apr. Oct.

VF3

−0.096 0.263 0.102 0.258 0.141 0.931 0.304 0.414 0.908 −0.198 1.707 17.065 73.611

0.458 0.726 0.938 0.031 0.945 −0.016 0.732 0.507 0.342 0.561 5.655 56.546 56.546

Feb.

0.792 0.165 0.234 0.911 0.130 0.090 0.533 0.669 0.143 0.547 1.158 11.584 85.196

Nov. Mar. Jul. Jan. May Dec. Jun. Aug. Sep. 0

20

40

60

80

100

120

(Dlink/Dmax)*100

Bold and italic values indicate strong and moderate loadings, respectively. Fig. 4. Dendrogram showing clustering of sampling periods. 1,2 Fe

1,0

Pb

VF2 (17.065%)

0,8 0,6

Ni

0,4

Mn Cd

Cr

Cu Co

0,2 0,0

As

Zn

-0,2 -0,4 -0,1

0,0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

0,9

1,0

VF1 (56.546%) Fig. 2. Loading plots of the first two VFs obtained for the data set.

metallic wastewaters containing high levels of Co, Cu and Ni into the Tigris River [10,26]. VF2, which accounted for 17.0% of the total variance, had strong positive loadings on Fe and Pb. This factor represents lithogenic sources. VF3 (11.5% of total variance) had strong positive loadings on As and Cr, and moderate positive loadings on Mn, Ni and Zn. This VF represents anthropogenic sources. The elements are derived from municipal and industrial wastewaters, and metallic wastewaters of the copper mine plant. 3.4.2. Cluster analysis Cluster analysis (CA) was applied to the river sediment quality data set to group the similar sampling sites (spatial variability). Spatial CA rendered a dendrogram (Fig. 3) where all seven sampling sites on the river were grouped into three statistically signifi-

cant clusters at (Dlink /Dmax ) × 100 < 40. Cluster 1 (Maden) site was located in a high pollution region, which receives metallic wastewater discharges from copper mine plant. Cluster 2 (E˘gil, Diyarbakır, Bismil, Hasankeyf and Cizre) sites were in a moderate pollution region. Cluster 3 (Batman) site was in a region of relatively low pollution. Temporal CA generated a dendrogram (Fig. 4) that grouped the 12 months into two clusters at (Dlink /Dmax ) × 100 < 60, and the difference between the clusters was significant. Cluster 1 included February, April, October, November, March, July, January, May and December roughly corresponding to the wet season in Turkey (October to April). In this study, about 82% annual total precipitation was concentrated in the time period from October to April. Cluster 2 included the remaining months (June, August and September), closely corresponding to the dry season (May to September). However, if 12 months had been empirically divided into spring (March to May), summer (June to August), autumn (September to November) and winter (December to February), or into dry/wet seasons, a mistake in grouping could have been made. In fact, Fig. 4 shows that the temporal patterns in water quality were not purely consistent with the four seasons or the dry/wet season. Similarly, CA was performed to group the analyzed parameters. CA rendered a dendrogram (Fig. 5) where all ten metal parameters were grouped into three statistically significant clusters at (Dlink /Dmax ) × 100 < 85. Cluster 1 includes As and Zn which were identified as contaminants derived from anthropogenic sources (wastewater discharges of copper mine plant). Cluster 2 contains Cd, Mn, Ni, Cr, Co and Cu derived from anthropogenic sources

As Zn

Maden (1)

Cd

Eğil (2)

Mn

Diyarbakır (3)

Ni Cr

Bismil (4)

Co

Cizre (7)

Cu

Hasankeyf (6)

Fe

Batman (5)

Pb

0

20

40

60

80

100

120

(Dlink/Dmax)*100 Fig. 3. Dendrogram showing clustering of sampling sites on the Tigris River.

0

20

40

60

80

100

(Dlink/Dmax)*100 Fig. 5. Dendrogram showing clustering of the analyzed parameters.

120

M. Varol / Journal of Hazardous Materials 195 (2011) 355–364

363

Table 10 Pearson correlation matrix of heavy metals in the Tigris River.

As Cd Co Cr Cu Fe Mn Ni Pb Zn

As

Cd

Co

Cr

Cu

Fe

Mn

Ni

Pb

Zn

1 0.277b 0.404a 0.583a 0.383a −0.013 0.575a 0.473a 0.120 0.681a

1 0.660a 0.331a 0.682a 0.214 0.714a 0.699a 0.466a 0.260b

1 0.317a 0.973a 0.219b 0.835a 0.717a 0.488a 0.393a

1 0.259b 0.267b 0.585a 0.732a 0.398a 0.316a

1 0.213 0.796a 0.673a 0.502a 0.389a

1 0.363a 0.425a 0.853a −0.068

1 0.862a 0.608a 0.528a

1 0.629a 0.342a

1 0.088

1

Bold values represent correlation with significance. a Significance at the 0.01 probability level (2-tailed). b Significance at the 0.05 probability level (2-tailed).

(wastewater discharges of copper mine plant, and industrial and domestic wastewaters). Cluster 3, which contains Fe and Pb, are derived from lithogenic sources.

3.4.3. Correlation matrix In order to establish relationships among metals and determine the common source of metals in the Tigris River, a correlation matrix was calculated for heavy metals in the sedimens. According to the values of Pearson correlation coefficients (Table 10), a significant positive correlation existed among the metals studied. In this study, Fe did not show significant correlation with As, Cd, Cu and Zn, and Pb did not show significant correlation with Zn. Fe was significantly correlated with Pb (r = 0.853, p < 0.01), indicating that the elements were derived from lithogenic sources. The significantly positive correlation of As (r = 0.383, p < 0.01), Cd (r = 0.682, p < 0.01), Co (r = 0.973, p < 0.01), Cr (r = 0.259, p < 0.01), Mn (r = 0.796, p < 0.01), Ni (r = 0.673, p < 0.01), and Zn (r = 0.389, p < 0.01) with Cu showed that the elements were derived from wastewater discharges of copper mine plant and also moving together.

4. Conclusion Different useful tools, methods, guidelines and indices have been employed for evaluation of sediment pollution in the Tigris River, Turkey. The highest concentrations of heavy metals were found at site-1 (Maden) due to metallic wastewater discharges from copper mine plant in Maden Township. Site-3 (Diyarbakır), site-4 (Bismil) and site-7 (Cizre) had also high metal concentrations due to domestic and industrial wastewaters. Total heavy metal concentrations in the sediment samples from the Tigris River followed the order: Fe > Mn > Cu > Pb > Zn > Ni > Cr > Co > As > Cd. The highest values of contamination factor (CF), pollution load index (PLI), geoaccumulation index (Igeo) and enrichment factor (EF) for all metals studied were found at site-1 (Maden), which receives a huge amount of metallic discharge from copper mine plant in Maden Township. Heavy metal concentrations in assessed sediment samples were compared with consensus-based TEC and PEC values. The results have indicated that the concentrations of Cr, Cu, Ni and Pb are likely to result in harmful effects on sediment-dwelling organisms which are expected to occur frequently. Multivariate analysis (PCA/FA, CA) and correlation matrix were used in this study. The PCA/FA applied on the investigated heavy metals identified three varifactors (VFs). VF1 and VF3, which were loaded with As, Cd, Co, Cr, Cu, Mn, Ni and Zn, were related to the anthropogenic sources. The CA classified all the sampling sites into three main groups of spatial similarities. A significant positive correlation is observed among As, Cd, Co, Cr, Cu, Mn, Ni and Zn, indicating that these metals were derived from similar sources and also moving together.

Acknowledgements The author thanks three anonymous reviewers for their valuable comments and constructive suggestions.

References [1] J.A. Cook, S.M. Andrew, M.S. Johnson, Lead, zinc, cadmium and fluoride in small mammals from contaminated grass-land established on fluorspar tailings, Water Air Soil Poll. 51 (1990) 43–54. [2] J. Deniseger, J. Erickson, A. Austin, M. Roch, M.J.R. Clark, The effects of decreasing heavy metal concentrations on the biota of Buttle Lake, Water Res. 24 (1990) 403–416. [3] S.N. Sin, H. Chua, W. Lo, L.M. Ng, Assessment of heavy metal cations in sediments of Shing Mun River, Hong Kong Environ. Int. 26 (2001) 297–301. [4] V.K. Gaur, S.K. Gupta, S.D. Pandey, K. Gopal, V. Misra, Distribution of heavy metals in sediment and water of River Gomti, Environ. Monit. Assess. 102 (2005) 419–433. [5] M. Singh, A.A. Ansari, G. Muller, I.B. Singh, Heavy metals in freshly deposited sediments of Gomti river (a tributary of the Ganga river): effects of human activities, Environ. Geol. 29 (1997) 246–252. [6] U. Förstner, G.T.W. Wittman, Metal Pollution in Aquatic Environment, SpringerVerlag, New York, 1983. [7] J.A. Thornton, A.J. McComb, S.O. Ryding, The role of sediments, in: A.J. McComb (Ed.), Eutrophic Shallow Estuaries and Lagoons, CRC Press, Boca Raton, 1975, pp. 205–223. [8] W. Calmano, W. Ahlf, U. Forstner, Exchange of heavy metals between sediment components and water, in: J.A.C. Broekaert, S¸. Güc¸er, F. Adams (Eds.), Metal Speciation in the Environment, NATO ASI Series, vol. G 23, Springer-Verlag, Berlin, 1990, pp. 503–522. [9] G.E. Batley, Trace Element Speciation: Analytical Methods and Problems, CRC Press, Boca Raton, Florida, 1989. [10] B. Gümgüm, E. Ünlü, Z. Tez, Z. Gülsün, Heavy metal pollution in water, sediment and fish from the Tigris River in Turkey, Chemosphere 29 (1994) 111–116. [11] H. Karadede-Akin, E. Ünlü, Heavy metal concentrations in water, sediment, fish and some benthic organisms from Tigris River, Turkey, Environ. Monit. Assess. 131 (2007) 323–337. [12] E.N. Akbulut, S. Bayarı, A. Akbulut, Y. S¸ahin, Rivers of Turkey, in: K. Tockner, C.T. Robinson, U. Uehlinger (Eds.), Rivers of Europa, Academic Press, New York, 2009, pp. 643–672. [13] Anonymous, Mean Monthly Flow Values of the Tigris River in the Period 2008–2009, General Directorate of Electrical Power Resources Survey and Development Administration, Ankara, 2009. [14] Anonymous, Mean Monthly Values of Temperature and Rainfall in Maden, Ergani, Diyarbakır, Batman and Cizre in the Period 2008–2009, Turkish State Meteorological Service, Ankara, 2009. [15] USEPA, Method 3051a: Microwave Assisted Acid Dissolution of Sediments, Sludges, Soils, and Oils, Revision 1, United States Environmental Protection Agency, Washington, DC, 2007. [16] S.M. Sakan, D.S. Djordjevic, D.D. Manojlovic, P.S. Polic, Assessment of heavy metal pollutants accumulation in the Tisza river sediments, J. Environ. Manage. 90 (2009) 3382–3390. [17] B. Rubio, M.A. Nombela, F. Vilas, Geochemistry of major and trace elements in sediments of the Ria de Vigo (NW Spain): an assessment of metal pollution, Mar. Pollut. Bull. 40 (2000) 968–980. [18] S. Covelli, G. Fontolan, Application of a normalization procedure in determining regional geochemical baselines, Environ. Geol. 30 (1/2) (1997) 34–45. [19] L. Hakanson, Ecological risk index for aquatic pollution control. A sedimentological approach, Water Res. 14 (1980) 975–1001. [20] D.C. Tomlinson, J.G. Wilson, C.R. Harris, D.W. Jeffery, Problems in the assessment of heavy metals levels in estuaries and the formation of a pollution index, Helgol. Wiss. Meeresunters 33 (1980) 566–575.

364

M. Varol / Journal of Hazardous Materials 195 (2011) 355–364

[21] G. Müller, Die Schwermetallbelstung der sedimente des Neckars und seiner Nebenflusse: eine Bestandsaufnahme, Chem. Zeitung 105 (1981) 157–164. [22] M.A.H. Bhuiyan, L. Parvez, M.A. Islam, S.B. Dampare, S. Suzuki, Heavy metal pollution of coal mine-affected agricultural soils in the northern part of Bangladesh, J. Hazard. Mater. 173 (2010) 384–392. [23] D.D. MacDonald, C.G. Ingersoll, T.A. Berger, Development and evaluation of consensus-based sediment quality guidelines for freshwater ecosystems, Arch. Environ. Con. Tox. 39 (2000) 20–31. [24] M. Varol, B. S¸en, Assessment of surface water quality using multivariate statistical techniques: a case study of Behrimaz Stream, Turkey Environ. Monit. Assess. 159 (2009) 543–553. [25] J. Zhang, C.L. Liu, Riverine composition and estuarine geochemistry of particulate metals in China-weathering features, anthropogenic impact and chemical fluxes, Estuar. Coast. Shelf Sci. 54 (2000) 1051–1070. [26] F. Tümen, M. Bildik, M. Baybay, M. Cici, B. Solmaz, Pollution potential of solid wastes from Ergani Copper Plant, Do˘ga Tr. J. Eng. Env. Sci. 16 (1992) 43–55. [27] E. Ünlü, B. Gümgüm, Concentrations of copper and zinc in fish and sediments from the Tigris River in Turkey, Chemosphere 26 (1993) 2055–2061. [28] B. Gümgüm, G. Öztürk, Chemical speciation of heavy metals in the Tigris River sediment, Chem. Spec. Bioavailab. 13 (2001) 25–29. [29] D. Mendil, Ö.F. Ünal, M. Tüzen, M. Soylak, Determination of trace metals in different fish species and sediments from the River Yes¸ilırmak in Tokat, Turkey Food Chem. Toxicol. 48 (2010) 1383–1392. [30] A. Farkas, E. Claudio, L. Vigano, Assessment of the environmental significance of heavy metal pollution in surficial sediments of the River Po, Chemosphere 68 (2007) 761–768.

[31] S. Olivares-Rieumont, D. de la Rosa, L. Lima, D.W. Graham, K. D’ Alessandro, J. Borroto, F. Martinez, J. Sanchez, Assessment of heavy metal levels in Almendares River sediments-Havana City, Cuba Water Res. 39 (2005) 3945–3953. [32] P. Woitke, J. Wellmitz, D. Helm, P. Kube, P. Lepom, P. Litheraty, Analysis and assessment of heavy metal pollution in suspended solids and sediments of the river Danube, Chemosphere 51 (2003) 633–642. [33] A.P. Karageorgis, N.P. Nikolaidis, H. Karamanos, N. Skoulikidis, Water and sediment quality assessment of the Axios River and its coastal environment, Cont. Shelf Res. 23 (2003) 1929–1944. [34] J. Morillo, J. Usero, I. Gracia, Heavy metal fractionation in sediments from the Tinto River (Spain), Int. J. Environ. Anal. Chem. 82 (2002) 245–257. [35] A.E. Rifaat, Major controls of metals’ distribution in sediments off the Nile Delta Egypt, Egypt J. Aquat. Res. 31 (2005) 16–28. [36] J.S. Heiny, C.M. Tate, Concentration distribution comparison of selected trace elements in bed sediment fish tissue in the South Platte River Basin U.S.A. 1992–1993, Arch. Environ. Con. Tox 32 (1997) 246–259. [37] K. Hudson-Edwards, M. Macklin, M. Taylor, Historic metal mining inputs to Tees river sediment, Sci. Total Environ. 194/195 (1997) 437–445. [38] W. Mendez, Contamination of Rimac River Basin Peru, due tomining tailings. TRITA-LWR Master Thesis, Environmental Engineering and Sustainable Infrastructure, The Royal Institute of Technology (KTH), Stockholm, 2005. [39] B. Vandecasteele, B. De Vos, F.M.G. Tack, Temporal-spatial trends in heavy metal contents in sediment-derived soils along the Sea Scheldt river (Belgium), Environ. Pollut. 122 (2003) 7–18. [40] J. Liu, Y. Li, B. Zhang, J. Cao, Z. Cao, J. Domagalski, Ecological risk of heavy metals in sediments of the Luan River source water, Ecotoxicology 18 (2009) 748–758.

Journal of Hazardous Materials 195 (2011) 365–370

Contents lists available at SciVerse ScienceDirect

Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat

Molecular mechanism of kidney injury of mice caused by exposure to titanium dioxide nanoparticles Suxing Gui a,1 , Zengli Zhang b,1 , Lei Zheng a,1 , Yaling Cui a,1 , Xiaorun Liu c,d,1 , Na Li a,e , Xuezi Sang a , Qingqing Sun a , Guodong Gao a , Zhe Cheng a , Jie Cheng a , Ling Wang a , Meng Tang c,d,∗ , Fashui Hong a,∗ a

Medical College, Soochow University, Suzhou 215123, People’s Republic of China Public Health School, Soochow University, Suzhou 215123, People’s Republic of China c Key Laboratory of Environmental Medicine and Engineering, Ministry of Education; School of Public Health, Southeast University, Nanjing 210009, China d Jiangsu key Laboratory for Biomaterials and Devices; Southeast University, Nanjing 210009, China e General Hospital of Jincheng Anthracite Mining Group Co. Ltd., Jincheng 048006, People’s Republic of China b

a r t i c l e

i n f o

Article history: Received 30 June 2011 Received in revised form 17 August 2011 Accepted 17 August 2011 Available online 24 August 2011 Keywords: Titanium dioxide nanoparticules Kidney Inflammatory response Cytokines

a b s t r a c t Numerous studies have demonstrated that damage of kidney of mice can be caused by exposure to titanium dioxide nanoparticles (TiO2 NPs). However, the molecular mechanism of TiO2 NPs-induced nephric injury remains unclear. In this study, the mechanism of nephric injury in mice induced by an intragastric administration of TiO2 NPs was investigated. The results showed that TiO2 NPs were accumulated in the kidney, resulting in nephric inflammation, cell necrosis and dysfunction. Nucleic factor-␬B was activated by TiO2 NPs exposure, promoting the expression levels of tumor necrosis factor-␣, macrophage migration inhibitory factor, interleukin-2, interleukin-4, interleukin-6, interleukin-8, interleukin-10, interleukin18, interleukin-1␤, cross-reaction protein, transforming growth factor-␤, interferon-␥ and CYP1A1, while heat shock protein 70 expression was inhibited. These findings implied that TiO2 NPs-induced nephric injury of mice might be associated with alteration of inflammatory cytokine expression and reduction of detoxification of TiO2 NPs. Crown Copyright © 2011 Published by Elsevier B.V. All rights reserved.

1. Introduction In the development of nanotechnology, nanomaterials are recognized to have potential applications due to their larger surface area to volume ratio, which enhances chemical reactivity and easier penetration into cells. Among the various nanomaterials, customarily titanium dioxide nanoparticles (TiO2 NPs) are regarded as chemical inert, nontoxic and biocompatible material [1–3], they have been widely used in the sunscreen ingredient, pharmaceutical, and paint industries as a colouring material [4–7]. In over ten years, however, TiO2 NPs toxicology has attracted considerable attention owing to their small sizes, large surface per mass and high reactivity. A number of investigations have definitely showed that TiO2 NPs exposure are able to cause injuries in various animal organ types, including lung, liver, spleen, and brain [8–24]. Recently, the toxicity of TiO2 NPs to kidneys has been reported. Scown et al. had found that TiO2 NPs were accumulated in the kidney, but had minimal effects on renal functions in rainbow trout [25]. In contradiction, Wang et al. had observed that TiO2 NPs exposure to mice resulted

in higher blood urea nitrogen and creatinine levels and the renal tubule was filled with proteinic liquids [10]. Chen et al. had also observed renal glomerulus dilatation and proteinic liquids filled renal tubule, but no kidney dysfunction was found with TiO2 NPstreated mice [26]. Furthermore, TiO2 NPs were also suggested to induce nephric inflammation and impair nephric functions, which exerted its toxicity through ROS accumulation [27]. However, the molecular mechanism of TiO2 NPs-induced nephric inflammation remains unclear. Nephric inflammation and dysfunction are due to altered in kidney regardless of the cause of these diseases. Thus TiO2 NPs induced nephric inflammation and dysfunction are able to be monitored through inflammatory cytokine expression levels in kidney. To confirm the above hypothesis, mice were continuously exposed to TiO2 NPs for 90 days by an intragastric administration. The inflammatory cytokine expression in the mouse kidney was determined and the possible mechanism of the TiO2 NPs induced nephric pathogenesis in mice was discussed. 2. Materials and methods 2.1. Chemicals, preparation and characterization

∗ Corresponding authors. Tel.: +86 0512 61117563; fax: +86 0512 65880103. E-mail addresses: [email protected] (M. Tang), Hongfsh [email protected] (F. Hong). 1 Contributed equally to this work.

Nanoparticulated anatase TiO2 was prepared via controlled hydrolysis of titanium tetrabutoxide. The details of the synthesis

0304-3894/$ – see front matter. Crown Copyright © 2011 Published by Elsevier B.V. All rights reserved. doi:10.1016/j.jhazmat.2011.08.055

366

S. Gui et al. / Journal of Hazardous Materials 195 (2011) 365–370

and characterization TiO2 NP were previously described by our previous reports [21,28]. The average particle sizes of powder suspended in 0.5% w/v hydroxypropylmethylcellulose K4M (HPMC, K4M) solvent after 12 h and 24 h incubation ranged from 5 to 6 nm. The mean hydrodynamic diameter of TiO2 NPs in HPMC solvent ranged between 208 and 330 nm (mostly 294 nm), and the zeta potential after 12 h and 24 h incubation was 7.57 mV and 9.28 mV [21]. 2.2. Animal and treatment It has been previously demonstrated by Wang et al. that sensitivity to TiO2 exposure was higher in CD-1 (ICR) female mice than CD-1 (ICR) male mice [1]. Therefore, CD-1 (ICR) female mice were used in this study. 80 CD-1 (ICR) female mice (24 ± 2 g) were purchased from the Animal Center of Soochow University (China). All mice were housed in stainless steel cages in a ventilated animal room. Room temperature of the housing facility was maintained at 24 ± 2 ◦ C with a relative humidity of 60 ± 10% and a 12-h light/dark cycle. Distilled water and sterilized food were available for mice ad libitum. Prior to dosing, the mice were acclimated to this environment for 5 days. All animals were handled in accordance with the guidelines and protocols approved by the Care and Use of Animals Committee of Soochow University (China). All procedures used in animal experiments conformed to the U.S. National Institutes of Health Guide for the Care and Use of Laboratory Animals [29]. The mice were randomly divided into four groups (N = 20), including a control group treated with 0.5% w/v HPMC and three experimental groups treated with 2.5, 5, and 10 mg/kg BW TiO2 NPs, respectively). The mice were weighed, and the TiO2 NP suspensions were administered to the mice by an intragastric administration every day for 90 days. Any symptom or mortality was observed and recorded carefully everyday during the 90 days. After 90 days, all mice were weighed firstly, and then sacrificed after being anesthetized using ether. Blood samples were collected from the eye vein by removing the eyeball quickly. Serum was collected by centrifuging blood at 2500 rpm for 10 min. Kidneys were collected and weighed. 2.3. Coefficient of kidney After weighing the body and kidneys, the coefficient of kidney to body weight was calculated as the ratio of kidney (wet weight, mg) to body weight (g). 2.4. Titanium content analysis Kidneys were removed from the −80 ◦ C and then thawed, and roughly 0.3 g of the kidney was weighed, digested and analyzed for titanium content. Inductively coupled plasma-mass spectrometry (ICP-MS, Thermo Elemental X7, Thermo Electron Company) was used to analyze the titanium concentration in the samples. For the analysis, an Indium concentration of 20 ng/mL was utilized as an internal standard element, and the detection limit of titanium was 0.074 ng/mL. The data were expressed as nanograms per gram fresh tissue. 2.5. Biochemical analysis of kidney functions Kidney functions were determined by uric acid (UA), blood urea nitrogen (BUN), creatinine (Cr), calcium (Ca) and phosphonium (P). All biochemical assays were performed using a clinical automatic chemistry analyzer (Type 7170A, Hitachi, Japan).

2.6. Histopathological examination of kidney For pathological studies, all histopathological tests were performed using standard laboratory procedures [30]. The kidneys were embedded in paraffin blocks, then sliced into 5 ␮m in thickness and placed onto glass slides. After hematoxylin–eosin (HE) staining, the slides were observed and the photos were taken using an optical microscope (Nikon U-III Multi-point Sensor System, USA), and the identity and analysis of the pathology slides were blind to the pathologist. 2.7. Expression assay of inflammatory cytokines The level of mRNA expression of nucleic factor-␬B (NF-␬B), NF-␬B-inhibiting factor (I␬B), tumor necrosis factor-␣ (TNF-␣), macrophage migration inhibitory factor (MIF), interleukin-2 (IL2), interleukin-4 (IL-4), interleukin-6 (IL-6), interleukin-8 (IL-8), interleukin-10 (IL-10), interleukin-18 (IL-18), interleukin-1␤ (IL1␤), cross-reaction protein (CRP), transforming growth factor-␤ (TGF-␤), interferon-␥ (INF-␥), cytochrome p450 1A (CYP1A) and heat shock protein 70 (HSP70) in the mouse kidney was determined using real-time quantitative RT polymerase chain reaction (RT-PCR) [31–33], respectively. Synthesized cDNA was used for the real-time PCR by employing primers that were designed using Primer Express Software according to the software guidelines, and PCR primer sequences are available upon request. To determine NF␬B, I␬B, TNF-␣, MIF, IL-2, IL-4, IL-6, IL-8, IL-10, IL-18, IL-6, IL-1␤, CRP, TGF-␤, INF-␥, Bax, Bcl-2, CYP1A1 and HSP-70 levels in the mouse kidney, an enzyme linked immunosorbent assay (ELISA) was performed using commercial kits that are selective for each respective protein (R&D Systems, USA). Manufacturer’s instruction was followed. The absorbance was measured on a microplate reader at 450 nm (Varioskan Flash, Thermo Electron, Finland), and the concentrations of NF-␬B, I␬B, TNF-␣, MIF, IL-2, IL-4, IL-6, IL-8, IL-10, IL-18, IL-6, IL-1␤, CRP, TGF-␤, INF-␥, CYP1A1 and HSP-70 were calculated from a standard curve for each sample. 2.8. Statistical analysis Statistical analyses were conducted using SPSS 11.7 software. Data were expressed as means ± standard deviation (SD). One-way analysis of variance (ANOVA) was carried out to compare the differences of means among multi-group data. Dunnett’s test was performed when each dataset was compared with the solventcontrol data. Statistical significance for all tests was judged at a probability level of 0.05 (p < 0.05). 3. Results 3.1. Coefficient of kidney and titanium accumulation Significant increases of the coefficients of kidney (p < 0.05 or p < 0.01) were caused by TiO2 NPs exposure for consecutive 90 days (Fig. 1). Furthermore, with increasing TiO2 NPs dose, the obvious accumulation of titanium in the kidney occurred (p < 0.01) (Fig. 2). These results show that the accumulation of titanium in the kidney was associated with the coefficients of kidney of mice. The increase of kidney indices caused by TiO2 NPs exposure may be related to the nephric dysfunction and tissue injury, which are confirmed by the further assays of biochemical parameters and histopathological observation of kidney of mice. 3.2. Biochemical parameters in serum of kidney The changes of biochemical parameters in the blood serum of mice kidney caused by TiO2 NPs exposure are presented in Table 1.

S. Gui et al. / Journal of Hazardous Materials 195 (2011) 365–370

367

Table 1 The changes of biochemical parameters in the blood serum of mice after intragastric administration with TiO2 NPs for 90 days. Indexes

TiO2 NPs (mg/kg, BW) 0

UA (␮mol/L) Cr (␮mol/L) BUN (mmol/L) Ca (mmol/L) P (mmol/L)

222.56 8.81 9.28 2.43 3.28

2.5 ± ± ± ± ±

11.13 0.44 0.46 0.12 0.16

160.21 9.75 8.11 2.48 3.33

5 ± ± ± ± ±

8.01* 0.49* 0.41* 0.12 0.16

110.88 11.68 7.05 2.51 3.46

10 ± ± ± ± ±

5.54** 0.58** 0.35** 0.13 0.17

96.76 13.19 6.32 2.71 3.52

± ± ± ± ±

4.84** 0.66** 0.32** 0.14 0.18

Ranks marked with an asterisk or double asterisks means it is statistically significant different from the control (unexposed mice) at the 5% or 1% confidence level, respectively. Values represent means ± SD, N = 10.

With TiO2 NPs dose increased, the contents of Ca and P of nephric function parameters were not significant compared with the control group (p > 0.05). However, the Cr was increased, and the UA, and BUN were decreased gradually (p < 0.05 or p < 0.01), respectively; demonstrating that long-term exposure to low dose TiO2 NPs impaired nephric functions in mice. 3.3. Histopathological evaluation of kidney Fig. 3 presents the histopathological changes of kidneys in mice treated by TiO2 NPs exposure. In the 2.5 mg/kg BW TiO2 NPs treated group, the nephric tissue is significantly showed to inflammatory cell infiltration and congestion of mesenchyme blood vessel (Fig. 3b). In the 5 mg/kg BW TiO2 NPs treated group, inflammatory cell infiltration, congestion of mesenchyme blood vessel and spotty necrosis of renal tubular epithelial cells were observed (Fig. 3c). Furthermore, a large area of necrosis of renal tubular epithelial cells

was detected in the 10 mg/kg BW TiO2 NPs treated group (Fig. 3d). The findings indicate that the kidney injury was related to a dosedependent manner of TiO2 NPs exposure. 3.4. Cytokine expression To further confirm the role of molecular mechanism in the TiO2 NPs-induced kidney injury, the changes of the inflammationrelated genes or detoxification-related genes and their proteins expression in mice caused by TiO2 NP exposure were detect using real time RT-PCR and ELISA (Tables 2 and 3). The mRNA expression levels of NF-␬B, TNF-␣, MIF, IL-2, IL-4, IL-6, IL-8, IL-10, IL-18, IL1␤, CRP, TGF-␤, INF-␥, and CYP1A1 were increased significantly in the TiO2 NP treated groups (p < 0.05 or 0.01). Interestingly, I␬B and HSP-70 expression levels were decreased significantly compared with control group (p < 0.05 or 0.01). 4. Discussion

Fig. 1. The coefficients of kidney of mice by an intragastric administration with TiO2 NPs for consecutive 90 days. Bars marked with an asterisk or double asterisks means it is significantly different from the control (unexposed mice) at the 5% or 1% confidence level, respectively. Values represent means ± SD, N = 20.

Fig. 2. The contents of titanium in the mouse kidney by an intragastric administration with TiO2 NPs for 90 days. Bars marked with an asterisk or double asterisks means it is statistically significant different from the control (unexposed mice) at the 5% or 1% confidence level, respectively. Values represent means ± SD, N = 5.

In this study, effects of TiO2 NPs on the mouse kidney were studied. After intragastric administrations with 2.5, 5, and 10 mg/kg BW of TiO2 NPs for 90 consecutive days, significant increases of the kidney indices (Fig. 1) and titanium accumulation in mouse kidneys (Fig. 2) were observed, coupled with increase of Cr level, decrease of BUN, UA excretion (Table 1), induced inflammatory response and necrosis of kidneys (Fig. 3). Previous study indicated that abnormal pathological changes of the mouse kidney and the nephric dysfunction were not able to be triggered by intraperitoneal injection with 5 mg/kg BW TiO2 NPs for 14 days, but with 50, 100 and 150 mg/kg BW TiO2 NPs exposure, impairment of kidney functions and severe inflammatory response of kidney were observed [27]. Wang et al. also observed that the 2-week exposure to the 5 g/kg BW TiO2 NPs by a gavage caused nephric dysfunction and tissue damage of mice [10]. In this study, molecular evidences were provided to prove TiO2 NPs induced nephric dysfunction and inflammation of mice by alteration of gene expression levels of the cytokines involved in inflammatory response or detoxification. NF-␬B is known as a critical intracellular mediator of the inflammatory cascade, and it binds to inhibitory proteins (I␬Bs) which prevent NF-␬B from migrating to the nucleus from cytoplasm. When an appropriate inducer existed, I␬Bs are phosphorylated and degraded, allowing nuclear uptaking of NF-␬B and initiating gene transcriptions, including MIF, the proinflammatory cytokines of TNF-␣, IL-1␤, IL6, IL-8, IL-18, CRP, and anti-inflammatory cytokines of IL-2, IL-4, and IL-10 [34]. TGF-␤ is proved to be involved with a dual-role as an anti-inflammatory and a profibrotic cytokine. IFN-␥ and TNF are essential for primary defense against infection [35,36], and mice that lack these two cytokines or their cognate receptors succumb to infection rapidly [37]. In response to TiO2 NPs stimulation, our results suggested that TiO2 NPs exposure for 90 consecutive days could significantly up-regulate mRNA expression levels of several relative inflammatory cytokines genes, including NF-␬B, TNF-␣, MIF, IL-2, IL-4, IL-6, IL-8, IL-10, IL-18, IL-1␤, CRP, TGF-␤, and

368

S. Gui et al. / Journal of Hazardous Materials 195 (2011) 365–370

Table 2 Effect of TiO2 NPs on the amplification of regulating the inflammation, immune and oxidative stress gene mRNA of the mouse kidney by real-time PCR analysis after intragastric administration with TiQ2 NPs for consecutive 90 days. Ratio of gene/actin

TiO2 NPs (mg/kg, BW) 0

NF-B/actin IB/actin TNF-␣/actin MIF/actin IL-2/actin IL-4/actin IL-6/actin IL-8/actin IL-10/actin IL-18/actin IL-1ˇ/actin CRP/actin TGF-ˇ/actin INF-/actin CYP1A1/actin HSP-70/actin

0.30 0.71 0.08 0.21 0.06 0.07 0.09 0.15 0.12 0.28 0.21 0.42 0.26 0.20 0.28 0.41

2.5 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.015 0.036 0.004 0.011 0.003 0.004 0.005 0.008 0.006 0.014 0.011 0.021 0.013 0.010 0.014 0.021

5

0.36 ± 0.0l8 0.58 ± 0.029* 0.11 ± 0.006* 0.32 ± 0.016* 0.09 ± 0.005* 0.09 ± 0.005* 0.13 ± 0.007* 0.19 ± 0.010* 0.17 ± 0.009* 0.31 ± 0.016 0.33 ± 0.017** 0.53 ± 0.027* 0.38 ± 0.019* 0.26 ± 0.013* 0.36 ± 0.018* 0.32 ± 0.016* *

0.57 0.42 0.18 0.49 0.15 0.16 0.18 0.31 0.24 0.46 0.42 0.65 0.53 0.41 0.66 0.25

10 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

**

0.029 0.021** 0.009** 0.025* 0.008** 0.008** 0.009** 0.016** 0.012** 0.023** 0.021** 0.033** 0.027** 0.021** 0.033** 0.013**

0.88 0.32 0.31 0.76 0.23 0.27 0.31 0.51 0.38 0.61 0.58 0.88 0.70 0.58 1.01 0.11

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.044** 0.016** 0.016** 0.038** 0.012** 0.014** 0.016** 0.026** 0.019** 0.031** 0.029** 0.044** 0.035** 0.029** 0.051** 0.006**

Ranks marked with an asterisk or double asterisks means it is statistically significant different from the control (unexposed mice) at the 5% or l% confidence level, respectively. Values represent means ± SD, N = 5.

INF-␥, and decrease I␬B expression. The obvious alterations of these cytokines’ expression indicated the involvement of inflammatory responses in TiO2 NPs-induced kidney toxicity. Studies had showed that TiO2 NPs promoted the expression of inflammatory cytokines in the lung, liver, spleen and brain of rat and mice [9,12,38–41]. Increases of NF-␬B expression in the mouse liver were also detected due to the significant increases of NF-␬B-inducible kinase and I␬B kinase expression and decrease of I␬B expression after treated with TiO2 NPs for 60 days [17]. In this study, significant increase of the CYP1A expression and reduction of HSP70 expression were observed (Tables 2 and 3). CYP1A and HSP70 were selected since they represent different processes that the cells follow to detoxify and/or defend against environmental toxicants [42]. Differences of gene expression of CYP1A and HSP70 were then used to explain the toxic characteristic

signatures of TiO2 NPs. It is well known that CYP1A induction is activated by the aryl hydrocarbon receptor (AHR) pathway, and its protein plays an essential function in the biotransformation and detoxification of endogenous and exogenous compounds. It is a widely accepted environmental biomarker, useful for monitoring the biological effects of several xenobiotic groups, including heavy metals [42]. De Jongh et al. showed that administration of 2,3,7,8tetrachlorodibenzo-p-dioxin (TCDD) to male C57BL/6J mice had caused the increases of both CYP 1A1 and CYP 1A2 hepatic protein levels [43]. In this study the high level expression of this gene and its protein products indicated that TiO2 NPs may cause kidney intoxication in mice. Likewise, higher level expression of HSP70 is often associated with a cellular response to a harmful stress or to adverse life conditions. The reduction of the HSP70 expression in the kidney by exposure to TiO2 NPs indicated a slow biotransformation or

Fig. 3. Histopathological observation of kidney caused by an intragastric administration with TiO2 NPs for consecutive 90 days. (a) Control group (unexposed mice) presents integrated glomerulars and normal kidney tubulars; (b) 2.5 mg/kg TiO2 NPs group presents inflammatory cell infiltration (yellow cycle) and congestion of mesenchyme blood vessel (blue arrow); (c) 5 mg/kg TiO2 NPs group indicates inflammatory cell infiltration (yellow cycle), congestion of mesenchyme blood vessel (blue arrow) and spotty necrosis of renal tubular epithelial cell (green cycle); (d) 10 mg/kg TiO2 NPs group indicates severe necrosis of renal tubular epithelial cell (green cycles). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.) The scale bar presented at the upside of each photomicrograph indicated 100 ␮m.

S. Gui et al. / Journal of Hazardous Materials 195 (2011) 365–370

369

Table 3 Effect of TiO NPs on the inflammatory cytokine protein levels of the mouse kidhey by ELISA analysis after intragastric administration with TiO NPs for consecutive 90 days. Protein expression

TiO2 NPs (mg/kg, BW) 0

NF-␬B (ng/g tissue) I␬B (g tissue) TNF-␣ (ng/g tissue) MIF (ng/g tissue) IL-2 (ng/g tissue) IL-4 (ng/g tissue) IL-6 (ng/g tissue) IL-8 (ng/g tissue) IL-10 (ng/g tissue) IL-18 (ng/g tissue) IL-1␤ (ng/g tissue) CRP (␮g/g tissue) TGF-␤ (ng/g tissue) TGF-␥ (ng/g tissue) CYP1A1 (ng/g tissue) HSP-70 (ng/g tissue)

34.62 18.71 72.83 269 66.45 44.99 6.95 32.93 5.96 6.17 88.94 38.68 21.69 19.50 12.02 11.95

2.5 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

1.73 0.94 3.64 13 3.32 2.25 0.35 1.65 0.30 0.31 4.45 1.93 1.08 0.98 0.60 0.60

38.13 14.26 81.66 582 73.28 49.39 8.23 38.99 7.12 8.29 105.77 53.91 34.71 28.57 21.17 7.14

5 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

1.91 0.71* 4.08* 29** 3.66* 2.47 0.41* 1.95* 0.36* 0.41* 5.29* 2.70* 1.74** 1.43** 1.06** 0.36*

52.95 10.77 171.26 2749 87.39 57.78 9.67 42.98 8.35 12.25 168.81 74.85 51.99 47.92 36.59 5.56

10 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

2.65** 0.54** 8.56** 137** 4.37** 2.89* 0.48** 2.15** 0.42** 0.61** 8.44** 3.74** 2.60** 2.40** 1.83** 0.28**

89.96 7.86 327.79 3129 94.46 71.19 13.99 53.37 10.66 19.99 196.42 95.93 72.48 63.81 52.88 3.78

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

4.50** 0.39** 16.39** 156** 4.72** 3.56** 0.70** 2.67** 0.53** 1.00** 9.82** 4.80** 3.62** 3.19** 2.64** 0.19**

Ranks marked with an asterisk or double asterisks means it is statistically significant different from the control (unexposed mice) at the 5% or l% confidence level, respectively. Values represent means ± SD, N = 5.

detoxification and decreased response to the adverse effects experienced in the kidney [44–47]. About the dose selection in this study, we consulted the report of World Health Organization in 1969. According to the report, LD50 of TiO2 for rats is larger than 12,000 mg/kg BW after oral administration. In the present study, we selected 5, 10, and 50 mg/kg BW TiO2 NPs exposed to mice every day. They were equal to about 0.15–0.7 g TiO2 NPs of 60–70 kg body weight for humans with such exposure, which were relatively safe doses. However, we think, attention should be aroused on the application of TiO2 NPs and their potential long-term exposure effects especially on human beings. In conclusion, the present study shows that mice treated with 2.5, 5 and 10 mg/kg BW TiO2 NPs for 90 consecutive days resulted in significant increases of NF-␬B, TNF-␣, MIF, IL-2, IL-4, IL-6, IL-8, IL-10, IL-18, IL-1␤, CRP, TGF-␤, INF-␥, CYP1A expression and significant decrease of HSP70 expression, leading to the increase of kidney indices, inflammatory responses and cell necrosis in mouse kidney.

Acknowledgements This work was supported by a Project Funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions, the National Natural Science Foundation of China (grant No. 30901218), the Major State Basic Research Development Program of China (973 Program) (grant No. 2006CB705602), National Important Project on Scientific Research of China (grant No. 2011CB933404), National Natural Science Foundation of China (grant No. 30671782, 30972504). References [1] R.W. Tennant, B.H. Margolin, M.D. Shelby, E. Zeiger, J.K. Haseman, J. Spalding, et al., Prediction of chemical carcinogenicity in rodents from in vitro genetic toxicity assay, Science 236 (1987) 933–941. [2] S. Huda, S.K. Smoukov, H. Nakanishi, B. Kowalczyk, K. Bishop, B.A. Grzybowski, Antibacterial nanoparticle monolayers prepared on chemically inert surfaces by cooperative electrostatic adsorption (CELA), Appl. Mater. Interfaces 2 (2010) 1206–1210. [3] J.F. Li, Y.F. Huang, Y. Ding, Z.L. Yang, S.B. Li, X.S. Zhou, et al., Shell-isolated nanoparticle-enhanced raman spectroscopy, Nature 464 (2010) 392–395. [4] M. Hedenborg, Titanium dioxide induced chemiluminescence of human polymorphonuclear leukocytes, Int. Arch. Occup. Environ. Health 61 (1988) 1–6. [5] C. Gelis, S. Girard, A. Mavon, M. Delverdier, N. Paillous, P. Vicendo, Assessment of the skin photoprotective capacities of an organo-mineral broad-spectrum sunblock on two ex vivo skin models, Photoimmunol. Photomed. 19 (2003) 242–253.

[6] K. Kakinoki, K. Yamane, M. Igarashi, M. Yamamoto, R. Teraoka, Y. Matsuda, Evaluation of titanium dioxide as a pharmaceutical excipient for preformulation of a photo-labile drug: effect of physicochemical properties on the photostability of solid-state nisoldipine, Chem. Pharm. Bull. 53 (2005) 811–815. [7] J. Gurr, C. Ass Wang, K.J. Chen, Ultrafine titanium dioxide particle in the absence of photoactivation can induce oxidative damage to human bronchial epithelial cells, Toxicology 213 (2005) 66–73. [8] F. Afaq, P. Abidi, R. Matin, Q. Rahman, Cytotoxicity, pro-oxidant effects and antioxidant depletion in rat lung alveolar macrophages exposed to ultrafine titanium dioxide, J. Appl. Toxicol. 18 (1998) 307–312. [9] H.W. Chen, S.F. Su, C.T. Chien, W.H. Lin, S.L. Yu, C.C. Chou, et al., Titanium dioxide nanoparticles induce emphysema-like lung injury in mice, FASEB J. 20 (2006) 1732–1741. [10] J.X. Wang, G.Q. Zhou, C.Y. Chen, H.W. Yu, T.C. Wang, Y.M. Ma, et al., Acute toxicity and biodistribution of different sized titanium dioxide particles in mice after oral administration, Toxicol. Lett. 168 (2007) 176–185. [11] H.T. Liu, L.L. Ma, J.F. Zhao, J. Liu, J.Y. Yan, J. Ruan, et al., Biochemical toxicity of mice caused by nano-anatase TiO2 particles, Biol. Trace Elem. Res. 129 (2009) 170–180. [12] L.L. Ma, J.F. Zhao, J. Wang, Y.M. Duan, J. Liu, N. Li, et al., The acute liver injury in mice caused by nano-anatase TiO2 , Nanosacle Res. Lett. 4 (2009) 1275–2128. [13] Y.M. Duan, J. Liu, L.L. Ma, N. Li, H.T. Liu, J. Wang, et al., Toxicological characteristics of nanoparticulate anatase titanium dioxide in mice, Biomaterials 31 (2010) 894–899. [14] N. Li, L.L. Ma, J. Wang, J. Liu, Y.M. Duan, H.T. Liu, et al., Interaction between nano-anatase TiO2 and liver DNA from mice in vivo, Nanoscale Res. Lett. 5 (2010) 108–115. [15] Y.L. Cui, X.L. Gong, Y.M. Duan, N. Li, R.P. Hu, H.T. Liu, et al., Hepatocyte apoptosis and its molecular mechanisms in mice caused by titanium dioxide nanoparticles, J. Hazard. Mater. 183 (2010) 874–880. [16] H.T. Liu, L.L. Ma, J. Liu, J.F. Zhao, J.Y. Yan, F.S. Hong, Toxicity of nano-anatase TiO2 to mice: liver injury, oxidative stress, Toxicol. Environ. Chem. 92 (2010) 175–186. [17] Y.L. Cui, H.T. Liu, M. Zhou, Y.M. Duan, N. Li, X.L. Gong, et al., Signaling pathway of inflammatory responses in the mouse liver caused by TiO2 nanoparticules, J. Biomed. Mater. Res. A 96 (2011) 221–229. [18] N. Li, Y.M. Duan, M.M. Hong, L. Zheng, M. Fei, X.Y. Zhao, et al., Spleen injury and apoptotic pathway in mice caused by titanium dioxide nanoparticules, Toxicol. Lett. 195 (2010) 161–168. [19] L.L. Ma, J. Liu, N. Li, J. Wang, Y.M.i Duan, J.Y. Yan, et al., Oxidative stress in the brain of mice caused by translocated nanoparticulate TiO2 delivered to the abdominal cavity, Biomaterials 31 (2010) 99–105. [20] R.P. Hu, X.L. Gong, Y.M. Duan, N. Li, Y. Che, Y.L. Cui, et al., Neurotoxicological effects and the impairment of spatial recognition memory in mice caused by exposure to TiO2 nanoparticles, Biomaterials 31 (2010) 8043–8050. [21] R.P. Hu, L. Zheng, T. Zhang, Y.L. Cui, G.D. Gao, Z. Cheng, et al., Molecular mechanism of hippocampal apoptosis of mice following exposure to titanium dioxide nanoparticles, J. Hazard. Mater. 191 (2011) 32–40. [22] J. Wu, W. Liu, C.B. Xue, S.C. Zhou, F.L. Lan, L. Bi, Toxicity and penetration of TiO2 nanoparticles in hairless mice and porcine skin after subchronic dermal exposure, Toxicol. Lett. 191 (2009) 1–8. [23] J.A. Shin, E.J. Lee, S.M. Seo, H.S. Kim, J.L. Kang, E.M. Park, Nanosized titanium dioxide enhanced inflammatory responses in the septic brain of mouse, Neuroscience 165 (2010) 445–454. [24] Y. Yu, W. Ren, B. Ren, Nanosize titanium dioxide cause neuronal apoptosis: a potential linkage between nanoparticle exposure and neural disorder, Neurol. Res. (2008), doi: 10. 1179/17431 3208X 305391.

370

S. Gui et al. / Journal of Hazardous Materials 195 (2011) 365–370

[25] T.M. Scown, R.Y. van Aerle, B.D. Johnston, S. Cumberland, J.R. Lead, O. Richard, et al., High doses of intravenously administered titanium dioxide nanoparticles accumulate in the kidneys of rainbow trout but with no observable impairment of renal function, Toxicol. Sci. 109 (2009) 372–380. [26] J.Y. Chen, X. Dong, J. Zhao, G.P. Tang, In vivo acute toxicity of titanium dioxide nanoparticles to mice after intraperitoneal injection, J. Appl. Toxicol. 29 (2009) 330–337. [27] J.F. Zhao, J. Wang, S.S. Wang, X.Y. Zhao, J.Y. Yan, J. Ruan, et al., The mechanism of oxidative damage in nephrotoxicity of mice caused by nano-anatase TiO2 , J. Exp. Nanosci. 5 (2010) 447–462. [28] P. Yang, C. Lu, N. Hua, Y. Du, Titanium dioxide nanoparticles codoped with Fe3+ and Eu3+ ions for photocatalysis, Mater. Lett. 57 (2002) 794–801. [29] National Institutes of Health (NIH), Guide for the Care and Use of Laboratory Animals, National Academy Press, Washington D.C., 1996. [30] A.B. Tarabishy, B. Aldabagh, Y. Sun, Y. Imamura, P.K. Mukherjee, J.H. Lass, et al., MyD88 regulation of Fusarium keratitis is dependent on TLR4 and IL-1R1 but not TLR2, J. Immunol. 181 (2008) 593–600. [31] K.J. Livak, T.D. Schmittgen, Analysis of relative gene expression data using realtime quantitative PCR and the 2(-delta delta C(T)) method, Methods 25 (2001) 402–408. [32] W.H. Liu, D.A. Saint, Validation of a quantitative method for real time PCR kinetics, Biochem. Biophys. Res. Commun. 294 (2002) 347–353. [33] L.D. Ke, Z. Chen, A reliability test of standard-based quantitative PCR: exogenous vs. endogenous standards, Mol. Cell. Probes 14 (2000) 127–135. [34] S. Ghosh, M.J. May, E.B. Kopp, NF-␬B and Rel proteins: evolutionarily conserved mediators of immune responses, Annul. Rev. Immunol. 16 (1998) 225–260. [35] N.A. Buchmeier, R.D. Schreiber, Requirement of endogenous interferon-gamma production for resolution of listeria monocytogenes infection, Proc. Natl. Acad. Sci. U.S.A. 82 (1985) 7404–7408. [36] E.A. Havell, Evidence that tumor necrosis factor has an important role in antibacterial resistance, J. Immunol. 143 (1989) 2894–2899.

[37] J.T. Harty, M.J. Bevan, Specific immunity to Listeria monocytogenes in the absence of IFN-␥, Immunity 3 (1995) 109–117. [38] D. Hohr, Y. Steinfartz, R.P. Schins, A.M. Knaapen, G. Martra, B. Fubini, et al., The surface area rather than the surface coating determines the acute inflammatory response after instillation of fine and ultrafine TiO2 in the rat, Int. J. Hyg. Environ. Health 205 (2002) 239–244. [39] J.X. Wang, C.Y. Chen, Y. Liu, F. Jiao, W. Li, F. Lao, et al., Potential neurological lesion after nasal instillation of TiO2 nanoparticles in the anatase and rutile crystal phases, Toxicol. Lett. 183 (2008) 72–80. [40] J.X. Wang, Y. Liu, F. Jiao, F. Lao, W. Li, Y.Q. Gu, et al., Time-dependent translocation and potential impairment on central nervous system by intranasally instilled TiO2 nanoparticles, Toxicology 254 (2008) 82–90. [41] J. Wang, N. Li, L. Zheng, Y. Wang, Y.M. Duan, S.S. Wang, et al., P38-Nrf-2 signaling pathway of oxidative stress in mice caused by nanoparticulate TiO2 , Biol. Trace Elem. Res. 140 (2011) 186–197. [42] N.A. Lewis, T.D. Williams, J.K. Chipman, Functional analysis of a metal response element in the regulatory region of flounder cytochrome P450 1A and implications for environmental monitoring of pollutants, Toxicol. Sci. 92 (2006) 387–393. [43] J. De Jongh, M. DeVito, R. Nieboer, L. Birnbaum, M. Van den Berg, Induction of cytochrome P450 isoenzymes after toxico-kinetic interactions between 2, 3,7,8-tetrachlorodibenzo-p-di-oxin and 2,2 ,4,4 ,5,5 -hexachlorobiphenyl in the liver of the mouse, Fund. Appl. Toxicol. 25 (1995) 264–270. [44] M.P. Mayer, B. Bukau, Hsp70 chaperones: cellular functions and molecular mechanism, Cell. Mol. Life Sci. 62 (2005) 670–684. [45] H.M. Beere, B.B. Wolf, K. Cain, D.D. Mosser, A. Mahboubi, T. Kuwana, et al., Heatshock protein 70 inhibits apoptosis by preventing recruitment of procaspase-9 to the Apaf-1 apoptosome, Nat. Cell Biol. 2 (2000) 469–475. [46] H.M. Li, T. Niki, T. Taira, S.M. Iguchi-Ariga, H. Ariga, Association of DJ-1 with chaperones and enhanced association and colocalization with mitochondrial Hsp70 by oxidative stress, Free Radic. Res. 39 (2005) 1091–1099. [47] H.N. Hong, H.N. Kim, K.S. Park, S.K. Lee, M.B. Gu, Analysis of the effects diclofenac has on Japanese medaka (Oryzias latipes) using real-time PCR, Chemosphere 67 (2007) 2115–2121.

Journal of Hazardous Materials 195 (2011) 371–377

Contents lists available at SciVerse ScienceDirect

Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat

Ozonation kinetics for the degradation of phthalate esters in water and the reduction of toxicity in the process of O3 /H2 O2 Gang Wen a , Jun Ma a,∗ , Zheng-Qian Liu b,∗∗ , Lei Zhao c a b c

State Key Laboratory of Urban Water Resource and Environment, Harbin Institute of Technology, Harbin 150090, People’s Republic of China School of Environmental Science and Engineering, Huazhong University of Science and Technology, Wuhan 430074, People’s Republic of China School of Civil Engineering, Harbin Institute of Technology, Harbin 150090, People’s Republic of China

a r t i c l e

i n f o

Article history: Received 12 May 2011 Received in revised form 11 July 2011 Accepted 18 August 2011 Available online 24 August 2011 Keywords: Phthalate esters Rate constant Ozone (O3 ) Hydroxyl radical (• OH) Hydroxyl radical/ozone ratio (Rct ) Toxicity assessment

a b s t r a c t The oxidation kinetics of four phthalate esters (PAEs) with ozone alone and hydroxyl radical (• OH) were investigated. The toxicity reduction in the process of O3 /H2 O2 was evaluated. The second order rate constants for the reaction of four PAEs with ozone and • OH were determined by direct oxidation method and competition kinetics method in bench-scale experiment, and found to be 0.06–0.1 M−1 s−1 and (3–5) × 109 M−1 s−1 , respectively. The oxidation kinetic rate constant of the selected PAEs (diethyl phthalate, DEP) was confirmed using Song Hua-jiang river water as the background. The results indicated that DEP degradation in this river water was close to the simulated value based on the determined rate constants. The toxicity test performed with bioluminescence test, showed that the toxicity expressed as the inhibition rate changed from 36% to below detection limit in the process of O3 /H2 O2 , which means that catalytic ozonation is an efficient way for DEP degradation and toxicity reduction, but an ineffective method for DEP minimization on the basis of the total organic carbon determination. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Phthalate esters (PAEs) are frequently used as plasticizers for cellulosic and vinyl ester resins to improve their flexibility and softness, and also in ceramic, paper, cosmetic and paint industries [1]. The production of PAEs has reached 3.5 million tons per year [2]. PAEs have been detected in surface and groundwater in ng L−1 –mg L−1 concentration range and associated with birth defects, organ damage, infertility, as well as testicular cancer, and are also known to be among the major endocrine disrupter chemicals (EDCs) [3,4]. Recently, it has been revealed that di-butyl phthalate (DBP) exhibits antagonistic thyroid receptor activity [5]. Drinking water treatment plant is the most important barrier to prevent the organic matters from human being contact. Previous investigation on 13 EDCs removal from traditional waterworks in China has demonstrated that four types of PAEs occurred almost in all samples with concentrations ranging from 20 to 163,760 ng L−1 , which was inefficiently removed during traditional drinking water treatment processes [6]. Advanced treatment processes are required to attenuate the PAEs contamination. Various protocols are explored to enhance PAEs removal during water

∗ Corresponding author. Tel.: +86 451 86282292/86283010; fax: +86 451 82368074. ∗∗ Corresponding author. E-mail addresses: [email protected] (J. Ma), [email protected] (Z.-Q. Liu). 0304-3894/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jhazmat.2011.08.054

treatment processes, including biotransformation [7], adsorption [8] and advanced oxidation processes [9–14]. The biotransformation of phthalates under both aerobic and anaerobic conditions has been investigated, but it is not suitable for applying in drinking water treatment due to less biomass existing in waterworks and requirement of long hydraulic retention time. Adsorption is an efficient way to remove PAEs due to its higher n-octanol/water partition coefficients (Kow ) (Table 1). However, it is only a method to shift the contaminations, but not to minimize them. Advanced oxidation processes would be the most powerful way for PAEs degradation and minimization. Several studies have been conducted for the elimination of PAEs by the processes of TiO2 /UV [9], H2 O2 /UV [10], electro-coagulation [11], ozonation [12] and catalytic ozonation processes [13,14]. Ozone is widely used in drinking water treatment for organic matter decomposition and microbiology disinfection all over the world. Several researchers conducted experiments on PAEs removal in the processes of ozonation and catalytic ozonation [12–14], with more attention to the removal efficiency and the way to improve it. However, the results of these studies illustrated that these processes could not be applied into practical water utilities due to the influence of natural water background. Elovitz and von Gunten [15] developed the hydroxyl radical/ozone ratio (Rct ) concept, which allows the prediction of the transformation of contaminations in natural water background combined with rate constants and oxidant behavior. In fact, ozonation and catalytic ozonation processes always involve in two active species: ozone

372

G. Wen et al. / Journal of Hazardous Materials 195 (2011) 371–377

Table 1 Chemical characteristics of four selected PAEs.

O

O

R

O

R

O Compound

Structure (–R)

CAS number

Molecular weight

Solubility (mg L−1 )

Log Kow

Dimethyl phthalate (DMP) Diethyl phthalate (DEP) Dipropyl phthalate (DPrP) Dibutyl phthalate (DBP)

–CH3 –CH2 CH3 –CH2 CH2 CH3 –CH2 CH2 CH2 CH3

131-11-3 84-66-2 131-16-8 84-74-2

194.2 222.2 250.3 278.4

4200 1100 108 11.2

1.66 2.65 3.27 4.50

and hydroxyl radical (• OH) [16]. The knowledge about second order rate constants of oxidation processes involving in both ozone alone and • OH would provide a powerful tool to optimize the degradation of PAEs. Unfortunately, a few kinetic data are available for ozone alone and • OH with PAEs. David Yao [17] reported the second order rate constant of DMP and DEP for the first time with ozone alone by means of pseudo-first order reaction in excess of ozone. However, there are not any reported rate constants of DBP and DPrP with ozone alone. Haag and Yao [18] estimated the second rate constant of DMP and DEP with • OH using a model method based on structure–activity relationship, but no experimental result has been reported on the second order constant rate of PAEs with • OH. Therefore, detailed kinetic constants of PAEs with ozone alone and • OH are required for further studies. The aim of this study is to determine the second-order rate constants of ozone alone and • OH with four PAEs by bench scale experiments in pure aqueous solution and to validate its applicability through oxidation and simulation of DEP degradation in river water background, where only DEP was selected as a representative due to the similarity of kinetics constant and the structures of 4 types of PAEs. Furthermore, bioluminescence test was performed to evaluate the acute toxicity change in the process of O3 /H2 O2 because of its higher degradation efficiency.

DMP, DEP, DPrP, DBP (Guang Fu Chemical Inc., China, 99.5% purity), p-chloro benzoic acid (pCBA, Sigma–Aldrich Chemicals, USA, 98% purity), high performance liquid chromatography (HPLC) grade methanol (Fisher, American), H2 O2 (30% w/w), ZnO powder (diameter 100 ␮m) and all other chemicals were of analytical grade and were used without further purification. Milli-Q water (Millipore Q Biocel system) was used for sample preparation. Ozone stock solutions were produced by sparging O3 /O2 air into Milli-Q water. The chemical characteristics of four selected PAEs are listed in Table 1.

bench-scale glass reactor. Ozone delivered into the reactor via a medium porosity ceramic was kept constant concentration across the whole experiment process, which was supplied from an ozone generator (DHX-IIB model, Harbin Jiujiu Co.) with the inlet ozone concentration is 0.7 mg min−1 . When the ozone concentration in reactor was reaching a constant value (4–5 mg L−1 ), a small aliquot (10 ml of DMP, DEP, DPrP and 25 ml of DBP) of PAEs stock solution (100 ␮M of DMP, DEP, DPrP and 40 ␮M of DBP) was injected into the reactor, followed by starting the reaction. An aliquot of 0.1 mol L−1 sodium thiosulfate solution was used to quench the reaction after sampling at various intervals. Meanwhile, the ozone concentration in liquid was determined with indigo method [20]. The experiment was carried out in Milli-Q water using tert-butyl alcohol (TBA 10 mM) as the • OH scavenger and was adjusted to pH 2 with perchlorate (1 M). If not stated otherwise, the experiments were controlled at 25 ◦ C by using cooling water. To determine the activation energy for the reaction of ozone with PAEs, the same experiments were also performed at 5, 10, and 15 ◦ C. The experiment was repeated at least three times and the errors given were 95% confidence intervals. Because of rapid reaction of • OH with organic matters, the rate constant of PAEs with • OH is difficult to determine directly, but can be measured by using competition kinetics method [16,18]. Considering the structure of PAEs, the rate constant of • OH with PAEs was constant throughout the pH range evaluated. These experiments were carried out with Milli-Q water at 25 ◦ C and the pH was kept at 10 for ozone decomposition into • OH. The reference compound was pCBA exhibiting a rate constant of k• OH 5 × 109 M−1 s−1 [17]. Under alkaline conditions, the half-life time of ozone is about several seconds, and the dominant reaction is between • OH and organic matters. Therefore, the reaction with molecular ozone can be ignored due to its short half-life time and lower second rate constant [19]. The equal concentrations of the compounds (1 ␮M pCBA and 1 ␮M PAEs) were spiked into the water samples. Thereafter different under-stoichiometric concentration levels of ozone (ranging from 0.1 to 1 ␮M) were added. After ozone injection, the solutions were vigorously stirred. The residual concentrations of target and referenced compound in the flask were analyzed by HPLC.

2.2. Determination of rate constants for the reaction of PAEs with ozone alone and • OH

2.3. Degradation and simulation of DEP decomposition in different ozonation processes in river water

Determination of the second-order rate constant of 4 types of PAEs with ozone was conducted by direct oxidation method under the condition of excessive ozone concentration. A semicontinuous flow reaction model was used to determine the rate constant of PAEs with ozone, which was described by a previous study [19]. Briefly, experiments were performed in a 1 L

Song Hua-jiang river water (water quality parameters are shown in Table 2) was used as the background for PAEs degradation. Due to the similarity of structure and kinetic properties of 4 types of PAEs, DEP was selected as the representative chemical to investigate its degradation in river water. River water was quickly filtered (0.45 ␮m, cellulose acetate) within 8 h after sampled and stored at

2. Experimental 2.1. Materials and reagents

G. Wen et al. / Journal of Hazardous Materials 195 (2011) 371–377

373

Table 2 Water quality characteristics of the filtered Song Hua-jiang river. pH

UV254 (cm−1 )

UV215 (cm−1 )

DOC (mg L−1 )

SUVA (L cm−1 mg−1 )

Alkalinity (mM)

7.9

0.0875

0.5248

3.5

3.36

4.7

4 ◦ C before use. During DEP degradation, river water was buffered to pH 8 by adding 10 mM borate buffer. The low concentration of DEP (1 ␮M) was spiked into river water, and 40 ml cold ozone stock solution (50 mg L−1 ) was injected to start the chemical reaction. Samples were taken at presumed intervals to determine the DEP concentration with HPLC. As described by Elovitz and von Gunten [15], Rct value can be calculated from the extent of the decrease of a probe compound (pCBA) concentration, which reacts fast with • OH but slowly with ozone, and a simultaneous determination of the ozone concentration. Once the Rct value is known, the elimination of a compound (M), which reacts with both oxidants, can be calculated by secondorder kinetics and expressed as a function of Rct , kO3 , kOH , and the  ozone exposure ( [O3 ] dt) according to Eq. (1); ln

 [M]  [M]0



= −(

[O3 ] dt)(kOH Rct + kO3 )

(1)

In order to simulate the degradation of DEP, another experiment was designed as follows. pCBA was added as a probe compound (0.5 ␮M) to determine Rct , H2 O2 and ZnO were used as catalyst in catalytic ozonation experiments, respectively, with a ratio of 0.34 mg of H2 O2 /mg of O3 , and ZnO added with a concentration of 0.1 g L−1 . The other part of this experiment was the same as the DEP degradation described as above. Before DEP and pCBA analysis, the samples were filtered with a filter (0.45 ␮m in pore size, cellulose acetate). 2.4. Toxicity test The acute toxicity of DEP and its degradation intermediate were tested with luminescent bacterium bioassay following the Chinese standard method (GB/T 15441-1995, 1996) [21,22]. It was performed using gram negative luminescent bacteria of the species Vibrio qinghaiensis sp. Nov (Q67). Due to the higher decomposition efficiency of O3 /H2 O2 , which was selected to decrease the toxicity of DEP and its intermediates products, a semi-continuous experiment (volume 1 L) was taken for DEP degradation in Milli-Q water with initial concentration of DEP of 20 ␮M, ozone concentration of 0.7 mg min−1 and H2 O2 concentration of 0.3 mM. Different samples at each interval were taken for toxicity test and TOC determination. The samples chosen for toxicity test were concentrated in Milli-Q water and adjusted to pH 7 before the analysis. Starting from a concentration factor of 300 times, eight double consecutive elution were tested (dilution factor 1:2), and the bioluminescence was then measured with Glomax illumination equipment (Turner Biosystems) [23]. The toxicity variation is expressed as Eq. (2). I(%) =

LB − LS × 100 LB

(2)

where I represents the inhibition of the concentrated sample to luminescent bacteria, LB is the luminescent intensity of blank, and LS is the luminescent intensity of sample. 2.5. Analytical methods The concentration of dissolved ozone in water was determined by the indigo method [20]. The concentration of ozone in gas phase was analyzed by iodometric method [24]. The concentration of PAEs were analyzed using HPLC equipped with an automatic Waters

717 plus autosampler injector and a Waters 1525 binary pump, using a waters symmetry C18 column (4.6 mm × 150 mm, 5 ␮m particle size) and methanol/water (50/50 for DMP, 60/40 for DEP, 70/30 for DPrP, 80/20 for DBP, v/v) as the mobile phase with a rate of 1 mL min−1 . The water sample was detected by a UV detector (Waters 2487 dual absorbance detector) at 230 nm and injected volume was 100 ␮L. The condition for pCBA analysis was as follows: an eluent with a rate of 1.0 mL min−1 consisting of 60/40 (v/v) methanol/water (adjusted to pH 2 with H3 PO4 ), UV wavelength of 240 nm was used. The pH in aqueous solution was measured by pH acidometer (Delta 320, Shanghai Leici Apparatus Fac., China). The TOC was analyzed by a TOC Analyzer (Analytik jena Multi N/C 3100). A Cary 500 UV–Vis spectrophotometer was used to measure UV254 value. The alkalinity of water was analyzed by titration method according to standard method [25]. 3. Results and discussion 3.1. Determination of rate constant for reaction of PAEs with ozone alone Several methods have been reported to determine second-order rate constant between ozone with compounds, including pseudofirst order reaction (ozone in excess or chemicals in excess) [16,17], competition kinetics [17] and direct oxidation method [19]. The second-order rate constant was determined by direct kinetics in semi-continuous batch reactor, and the exact experiment step can be found as above mentioned. Both • OH inhibition and ozone decomposition were considered in this study. The reaction between PAEs with • OH was expelled by using 10 mM TBA as a way for scavenging it, ozone concentration was kept constant during the whole process, so the second-order reaction could be transformed in pseudo-first order reaction. Table 1 shows that PAEs do not dissociate within all the range of pH, indicating that the second rate constant of PAEs is pH-independent. So, the experiment was carried out in pH 2 by using perchlorate for minimizing the reaction of • OH with PAEs. Under the condition of acidic pH and higher concentration of TBA in the ozonation system, molecular ozone reacts with PAEs predominantly, and the influence of • OH can be ignored. So, the rate of PAEs degradation can be written as follows: −

d[PAEs] = kO3 [O3 ][PAEs] dt

(3)

−

d[PAEs] = kO3 [O3 ]dt [PAEs]

(4)

Because of the constant ozone concentration, Eq. (4) can be converted to Eq. (5): ln

[PAEs]t = −kO3 [O3 ]t [PAEs]0

(5)

According to the above Eq. (5), kO3 can be concluded from the plots of PAE degradation versus time. The result of experiment is shown in Fig. 1. Table 3 lists all the results of the present experiment and the previous study. The rate constant of DMP, DEP, DPrP and DBP are 0.072, 0.085, 0.11 and 0.092 M−1 s−1 , respectively. Comparing the rate constants of four PAE compounds, there is no obvious relationship in the oxidation rate constants with the increase of PAE

374

G. Wen et al. / Journal of Hazardous Materials 195 (2011) 371–377 4.5

0.22

4.0

0.20

0.12 0.10

2.0

0.08

1.5

0.06 1.0

0.04

0.5

1.0 0.8 0.6 0.4

0.02

0.0 0

2000

4000

6000

8000

0.2

0.00 10000 12000 14000 16000

Time (s) Fig. 1. Variation of ozone concentration with time and the second-order plots of four types of PAEs degradation by ozone alone. Experimental conditions: [PAEs]0 = 1 ␮M, [TBA] = 10 mM, [O3 ] = 3.85 mg L−1 , pH = 2.0, T = 25 ◦ C.

carbon chain. The rate constants of DMP and DEP are close to those reported in previous study [17]. As can be seen from Table 3, PAEs are difficult to be oxidized by ozone alone. Previous study shows that ozone behaves as an electrophilic species, which reacts only with some electronrich organic moieties, such as phenols, anilines, olefins, and deprotonated-amines [26]. Advanced oxidation process, especially catalytic ozonation, produces more • OH and would be the option for PAEs elimination. The reaction rate constant between PAEs and • OH would be discussed afterwards. In order to figure out the activation energies, four different reaction temperatures were performed. Activation energy can be calculated according to the slope of the Arrhenius plot of ln k against T−1 to fit Eq. (6). ln k = ln A −

Ea RT

(6)

where K is second order rate constant, Ea is reaction activation energy, T is reaction temperature, R is gas constant. DMP, DEP, DPrP and DBP exhibited 66 ± 8, 73 ± 4, 54 ± 10 and 58 ± 9 kJ mol−1 , respectively. They are endothermic reaction, enhanced reaction rate as the temperature increase. The results show no trends between the substitution groups with the activation energies. 3.2. Determination of rate constant for reaction of PAEs with • OH Rate constants for the reaction of 4 types of PAEs with • OH were determined by competition kinetics methods, which is the frequently used. Primary methods for generating • OH include UV/H2 O2 , ␥-radiolysis, O3 /H2 O2 [16], Fenton method, Walling’s method [18] and O3 /OH− [19]. In the present study, competition kinetics with pCBA as reference compound was adopted and the O3 /OH− method was selected for producing • OH. The second-order rate constants of the four types of PAEs compounds (M) with • OH were determined using Eq. (7) by plotting Table 3 Second-order rate constants for the reaction between ozone and 4 types of PAEs. Compound

(T = 25 ◦ C) (M−1 s−1 ) Measured

DMP DEP DPrP DBP

DBP DPP DMP DEP

1.2

0.14

ln(PAEs)0/(PAEs)

2.5

0.16

-ln(C/C0)

-1

O3 (mg L )

3.0

1.4

0.18

O3 DEP DBP DPP DMP

3.5

1.6

0.072 0.085 0.110 0.092

± ± ± ±

0.012 0.021 0.033 0.042

0.0 0.0

Reference

0.99 0.98 0.98 0.98

0.2 ± 0.1 [17] 0.14 ± 0.05 [17] No report No report

0.8

1.2

1.6

ln(pCBA)0/(pCBA) Fig. 2. The second-order plots of the degradation of four types of PAEs by • OH. Experimental conditions: [PAEs]0 = 1 ␮M, [pCBA] = 1 ␮M, [O3 ] = 0.1–1 ␮M, pH = 10.0, T = 25 ◦ C.

the decrease of M versus the reference compound (C), which can be seen from Fig. 2, a typical plot of relative rate experiment for the degradation of two compounds. k•MOH =

ln ([M]0 /[M]∞ ) C k• ln ([C]0 /[C]∞ ) OH

(7)

where k•MOH and k•COH are the rate constant for PAEs and reference compound, respectively. The measured second-order rate constants for the reaction of the PAEs with • OH are summarized in Table 4. The rate constant of DMP, DEP, DPrP and DBP is 2.67 × 109 , 3.98 × 109 , 4.47 × 109 and 4.64 × 109 M−1 s−1 , respectively. The rate constant of DMP is lower than that calculated by Haag, whereas DEP’s rate constant is comparative to Haag’s result [18]. Additionally, it is the first time to report the second order rate constant of DPrP and DBP. As seen from the results in Table 4, the rate constant increases as the carbon chain extends. It is well-known that • OH attacks compounds by abstracting a hydrogen atom (H-abstraction), electron transfer reaction or by addition to an unsaturated bonds (such as C C bond). Seen from PAEs molecular structure, the attack to PAEs mainly depends on Habstraction, and the electron donor capacity of the substitute group follows the sequence: C(CH3 )3 > CH3 CH2 CH2 > CH3 CH2 > CH3 > H, which can explain the phenomenon of the increasing rate constant of PAEs as the increase of carbon chain length. Those second rate constants of PAEs would be further validated to its applicability through PAEs elimination in the processes of ozonation and catalytic ozonation processes and are discussed below. 3.3. Degradation and simulation of DEP decomposition in different ozonation processes in river water In natural water, the reactions of PAEs with ozone and • OH have to be considered together and it is essential to know their exposures Table 4 Second-order rate constants for the reaction between • OH and 4 types of PAEs. Compound

R2

0.4

KOH (T = 25 ◦ C) (M−1 s−1 ) Measured

DMP DEP DPrP DBP

(2.67 (3.98 (4.47 (4.64

± ± ± ±

0.26)E+9 0.21)E+9 0.35)E+9 0.41)E+9

R2

Reference

0.98 1.0 0.99 0.99

4E+9 [18] 4E+9 [18] No report No report

G. Wen et al. / Journal of Hazardous Materials 195 (2011) 371–377

2.0

0.5

375

1.0

1.0

0.8

0.9

O3 alone

0.8

O3/H2O2

O3

-1

-1.0

model of O3 alone

-1.5 -2.0 -2.5

1.2

0.6

-3.0 0

4

8

12

16

20

24

28

32

Time (min)

0.8

0.4

0.4

0.2

0.0

0.0

[pCBA]/ [pCBA]0

O3 (mg L )

pCBA

-0.5

C/C0

1.6

ln ([O3]/ [O3]0)

0.0

model of O3/H2O2 O3/ZnO

0.7

model of O3/ZnO

0.6

0.5 0

4

8

12

16

20

24

28

32

Time (min)

0.4 0

Fig. 3. Variation of ozone concentration with time and the plot of pCBA degradation in the process of ozonation. Experimental condition: [O3 ]0 = 2 mg L−1 , T = 25 ◦ C, [pCBA] = 0.5 ␮M, pH = 8, inset is the decomposition of ozone.

[27]. In natural river water background, a part of matrixes exists as • OH initiator, whereas another part of them serves as • OH scavengers. Rct , can be used to forecast the decomposition of compound in ozonation or catalytic ozonation processes in association with compound’s second-order rate constant of ozone and • OH. 3.3.1. Quantification of Rct of ozonation and catalytic ozonation processes in river water Ozone decomposition in natural water can be divided into an initial and a second phase. During the second phase (>20 s), ozone decomposition follows an apparent first-order rate law, the rate constant in the second phase is 10–100 times smaller than that during the initial phase [28]. As seen from Eq. (8), Rct describes that the ratio of • OH exposure to O3 -exposure, which can be calculated from the decrease in concentration of pCBA and O3 . The • OH exposure can be calculated by means of Eq. (9). The ozone exposure can be calculated from the integral of the ozone concentration versus time. Substitution of Eq. (9) into Eq. (8) gives the result of Rct , shown as Eq. (10). Rct =

 [OH]dt 



[OH] dt = − Rct = −

(8)

[O3 ]dt

ln([pCBA]/[pCBA]0 ) kOH,pCBA

(9)

ln([pCBA]/[pCBA]0 ) kOH,pCBA ·



(10)

[O3 ] dt

The concentration of pCBA (• OH probe) and O3 were detected, and typical results of Rct are shown in Fig. 3. The results show that the decomposition of O3 followed first order rate reaction. The calculated Rct of three oxidation processes are shown in Table 5, indicating that the O3 /H2 O2 process presents the most powerful capability to produce • OH, followed by O3 /ZnO process in generating • OH. As was reported previously, the mechanism of H2 O2 and ZnO for improving organic matter removal is the enhancement of ozone decomposition and conversion of ozone into • OH [29,30]. H2 O2 generates plenty of HO2 − which enhances ozone decompoTable 5 Measured and calculated Rct in different processes. Different process

O3 alone

O3 /H2 O2

O3 /ZnO

Rct (measured) Rct (calculated)

1.20E−08 6.06E−08

1.37E−07 3.19E−07

3.23E−08 6.06E−08

5

10

15

20

25

30

Time (min) Fig. 4. Simulation and degradation of DEP removal during ozonation or catalytic ozonation in river water (dots mean experimental data and lines mean modeled results). Experimental condition: [O3 ]0 = 2 mg L−1 , T = 25 ◦ C, [pCBA] = 0.5 ␮M, [DEP] = 1 ␮M, pH = 8.

sition, meanwhile hydroxyl group on ZnO surface improves the decomposition of ozone and the formation of • OH [31]. Recently, a method for predicting Rct in different natural water background was developed [32], which showed that Rct was dependent on water quality characteristics and could be simulated (R2 = 0.92), using water quality characteristics and experimental conditions (Eq. (11)). log Rct = −10.12 + 2.04 DH2 O2 − 0.325 DOC + 0.747 pH −11.47 UV254 − 0.143 URI

(11)

URI = UV relative index (calculated as the ratio of UV215 over UV254 ); DH2 O2 = peroxide dosage (mg H2 O2 /mg O3 ); UV254 = UV absorbance at 254 nm (cm−1 ); DOC = dissolved organic carbon (mg C L−1 ). In our present study, the model was used to predict the Rct in Song Hua-jiang river water, which can be seen in Table 5. Comparing the experimental results with the simulated results, the model data has the same order as experimental result. However, it cannot differentiate the processes of ozonation alone and O3 /ZnO catalytic ozonation because there is no consideration of the influence of heterogeneous catalyst on ozone decomposition. Therefore, the model is not appropriate for predicting the Rct in heterogeneous catalytic ozonation processes. 3.3.2. Simulation of DEP decomposition in river water Natural water matrixes have important impact on the ozonation process of organic matter, where natural organic matter may promotes or prohibits the radical chain reaction, acting as an initiator or scavenger, while, the alkalinity competes with organic matter for • OH as a scavenger [33]. Batch experiments with river water as background were performed to examine the removal efficiency of DEP in O3 , O3 /H2 O2 , and O3 /ZnO processes. The water quality parameters are given in Table 2. As seen from the experimental results in Fig. 4, the oxidation of DEP was mainly determined in reactions with • OH. The oxidation efficiencies increases with the increase of Rct , namely O3 /H2 O2 process with higher Rct removes DEP much faster than that of O3 alone or O3 /ZnO process.

376

G. Wen et al. / Journal of Hazardous Materials 195 (2011) 371–377 0.5

5

1.0 4

0.8

TOC

3

0.6

DEP

0.4

0.1

0.0 0

5

10

15 Time (min)

20

25

2

0.2

1

0.0

0

-1

0.2

TOC (mg L )

0.3

C/C 0

Inhibitation rate (%)

0.4

Acknowledgements

30

Fig. 5. Toxicity assessment of DEP oxidation and TOC change in the process of O3 /H2 O2 .

Furthermore, it is possible to predict the DEP removal with a function of ozone exposure, Rct , kOH , and kO3 according to Eq. (12). ln

 [M]  t

[M]0



= −(

Rct were conducted. Three different ozonation processes including ozone alone, O3 /H2 O2 and O3 /ZnO were evaluated in Song Huajiang river water background, which showed that the O3 /H2 O2 process had the highest capacity in degrading DEP. The DEP transformation was successfully simulated with the determined second-order rate constant in combination with Rct . The acute toxicity in the process of O3 /H2 O2 was assessed by luminescent bacteria, which showed that O3 /H2 O2 was an efficient way for PAEs degradation and toxicity reduction, but it was a less efficient method for DEP mineralization.

This research was supported by State Key Laboratory of Urban Water Resource and Environment, Harbin Institute of Technology (2010DX10), 863 Hi-tech Research and Development Program of China (Grant No. 2009AA06Z310) and the National Key Special Founding for Water Pollution Control and Management (Grant No. 2008ZX07421-002, 2009ZX07424-005, 2009ZX07424-006) and NSFC (50821002). References

[O3 ] dt)(kOH Rct + kO3 )

(12)

Fig. 4 presents the data for the predicted and measured oxidation of DEP in the three oxidation processes. The results indicated that DEP removal efficiency in river water followed the sequence, O3 /H2 O2 > O3 /ZnO > O3 , which is in accord with the Rct , meaning that O3 /H2 O2 shows the most powerful capability to produce • OH among the three processes. It was calculated with experimental Rct and model Rct for DEP removal with Eq. (12), which showed poor simulation results compared with the degradation data using the calculated Rct (data not shown here). Using the experimental Rct , the simulated results were well in consistence with the experimental data. From the consistence of experimental data with the simulated data, it can be concluded that above determined secondorder rate constants of DEP with O3 and • OH are applicable in natural water. 3.4. Toxicity assessment Fig. 5 shows the variations of toxicity and TOC in the process of O3 /H2 O2 . It is seen from the results that after 30 min oxidation, the acute toxicity expressing inhabitation rate decreases to below the detection limit, namely no acute toxicity was detected. The decrease in acute toxicity of DEP mainly occurs in the first 10 min, which coincides with the decrease of DEP concentration. Therefore, the primary toxicity to luminescent bacteria may come from the DEP parent molecular, while the intermediates contribute a little to the toxicity for Q67. After 30 min oxidation, the TOC decreases from 3.76 to 3.22 mg L−1 , indicating that only 14.4% of TOC is mineralized and the intermediates are accumulated in the oxidation process, which illustrated furthermore that the decrease of toxicity to Q67 was due to DEP parent molecular removal, but not the formation of intermediates. Through continuous oxidation of ozone and • OH, the parent of DEP had been decomposed into nontoxic intermediate. However, the mineralization of DEP requires more powerful oxidation process or in combination with biotransformation. 4. Conclusions The second-order rate constants for the reaction of 4 PAEs with ozone and • OH were determined by direct oxidation method and competition kinetics method by bench-scale experiments. Degradation of DEP in river and simulation decomposition based on

[1] M. Vitali, M. Guidotti, G. Macilenti, C. Cremisini, Phthalate esters in freshwaters as markers of contamination sources – a site study in Italy, Environ. Int. 23 (1997) 337–347. [2] J. Sundell, C.G. Bornehag, Presences of phthalate esters in homes, Epidemiology 17 (2006) S86. [3] C.A. Stales, D.R. Peterson, T.F. Parkerton, W.J. Adams, The environmental fate of phthalate esters: a literature review, Chemosphere 35 (1997) 667–749. [4] C.A. Harris, P. Henttu, M.G. Parker, J.P. Sumpter, The estrogenic activity of phthalate esters in vitro, Environ. Health Perspect. 105 (1997) 80–811. [5] N. Li, D. Wang, Y. Zhou, M. Ma, J. Li, Z. Wang, Dibutyl phthalate contributes to the thyroid receptor antagonistic activity in drinking water processes, Environ. Sci. Technol. 44 (2010) 6863–6868. [6] X.L. Shao, J. Ma, G. Wen, Investigation of endocrine risrupting chemicals in a drinking water work located in Songhua River basin, Environ. Sci. 129 (2008) 2723–2728 (In Chinese). [7] B.V. Chang, C.M. Yang, C.H. Cheng, S.Y. Yuan, Biodegradation of phthalate esters by two bacteria strains, Chemosphere 55 (2004) 533–538. [8] S. Venkata Mohan, S. Shailaja, M. Rama Krishna, P.N. Sarma, Adsorptive removal of phthalate ester (di-ethyl phthalate) from aqueous phase by activated carbon: a kinetic study, J. Hazard. Mater. 146 (2007) 278–282. [9] X.R. Xu, S.X. Li, X.Y. Li, J.D. Gu, F. Chen, X.Z. Li, H.-B. Li, Degradation of n-butyl benzyl phthalate using TiO2 /UV, J. Hazard. Mater. 164 (2009) 527–532. [10] B. Xu, N.Y. Gao, X.F. Sun, S.J. Xia, M. Rui, M.-O. Simonnot, C. Causserand, J.F. Zhao, Photochemical degradation of diethyl phthalate with UV/H2 O2 , J. Hazard. Mater. 139 (2007) 132–139. [11] I. KabdaslI, A. Keles, T. Ölmez-HancI, O. Tünay, I. Arslan-Alaton, Treatment of phthalic acid esters by electrocoagulation with stainless steel electrodes using dimethyl phthalate as a model compound, J. Hazard. Mater. 171 (2009) 932–940. [12] H.Y. Li, J.H. Qu, H.J. Liu, Removal of a type of endocrine disruptors-di-n-butyl phthalate from water by ozonation, J. Environ. Sci. 18 (2006) 845–851. [13] Y.H. Chen, N.C. Shang, D.C. Hsieh, Decomposition of dimethyl phthalate in an aqueous solution by ozonation with high silica zeolites and UV radiation, J. Hazard. Mater. 157 (2008) 260–268. [14] M. Hammad Khan, J.Y. Jung, Ozonation catalyzed by homogeneous and heterogeneous catalysts for degradation of DEHP in aqueous phase, Chemosphere 72 (2008) 690–696. [15] M.S. Elovitz, U. von Gunten, Hydroxyl radical/ozone ratios during ozonation processes I. The Rct concept, Ozone Sci. Eng. 21 (1999) 239–260. [16] U. von Gunten, Ozonation of drinking water: part I. Oxidation kinetics and product formation, Water Res. 37 (2003) 1443–1467. [17] C.C. David Yao, W.R. Haag, Rate constants for direct reactions of ozone with several drinking water contaminants, Water Res. 25 (1991) 761–773. [18] W.R. Haag, C.C.D. Yao, Rate constants for reaction of hydroxyl radicals with several drinking water contaminants, Environ. Sci. Technol. 26 (1992) 1005–1013. [19] J.M. Shen, Z.L. Chen, Z.Z. Xu, X.Y. Li, B.B. Xu, F. Qi, Kinetics and mechanism of degradation of p-chloronitrobenzene in water by ozonation, J. Hazard. Mater. 152 (2008) 1325–1331. [20] H. Bader, J. Hoigné, Determination of ozone in water by the indigo method, Water Res. 15 (1981) 449–456. [21] M. Ma, Z. Tong, Z. Wang, W. Zhu, Acute toxicity bioassay using the freshwater luminescent bacterium; Vibrio qinghaiensis sp. Nov.—Q67, Bull. Environ. Contam. Toxicol. 62 (1999) 247–253. [22] X.W. Zhu, S.S. Liu, H.L. Ge, Y. Liu, Comparison between the short-term and the long-term toxicity of six triazine herbicides on photobacteria Q67, Water Res. 43 (2009) 1731–1739.

G. Wen et al. / Journal of Hazardous Materials 195 (2011) 371–377 [23] S. Monarca, D. Feretti, C. Collivignarelli, L. Guzzella, I. Zerbini, G. Bertanza, R. Pedrazzani, The influence of different disinfectants on mutagenicity and toxicity of urban wastewater, Water Res. 34 (2000) 4261–4269. [24] K. Rakness, G. Gordon, B. Langlais, W. Masschelein, N. Matsumoto, Y. Richard, C.M. Robson, I. Somiya, Guideline for measurement of ozone concentration in the process gas from an ozone generator, Ozone Sci. Eng. 18 (1996) 209–229. [25] A.E. Greenberg, L.S. Clesceri, A.D. Eaton, Standard Methods for the Examination of Water and Wastewater, 20 ed., 1998, Washington, DC, pp. 1325. [26] Y. Lee, U. von Gunten, Oxidative transformation of micropollutants during municipal wastewater treatment: comparison of kinetic aspects of selective (chlorine, chlorine dioxide, ferrateVI, and ozone) and non-selective oxidants (hydroxyl radical), Water Res. 44 (2010) 555–566. [27] M.M Huber, S. Canonica, G.Y. Park, U. von Gunten, Oxidation of pharmaceuticals during ozonation and advanced oxidation processes, Environ. Sci. Technol. 37 (2003) 1016–1024.

377

[28] M.O. Buffle, U. Von Gunten, Phenols and amine induced • OH generation during the initial phase of natural water ozonation, Environ. Sci. Technol. 40 (2006) 3057–3063. [29] E.J. Rosenfeldt, K.G. Linden, S. Canonica, U. von Gunten, Comparison of the efficiency of OH radical formation during ozonation and the advanced oxidation processes O3 /H2 O2 and UV/H2 O2 , Water Res. 40 (2006) 3695–3704. [30] M. Muruganandham, J.J. Wu, Synthesis, characterization and catalytic activity of easily recyclable zinc oxide nanobundles, Appl. Catal. B: Environ. 80 (2008) 32–41. [31] J. Nawrocki, B. Kasprzyk-Hordern, The efficiency and mechanisms of catalytic ozonation, Appl. Catal. B: Environ. 99 (2010) 27–42. [32] S. Vincent, A. Kotbi, B. Barbeau, Predicting hydroxyl radical activity and trace contaminants removal in ozonated water, Ozone Sci. Eng. 32 (2010) 244–251. [33] J. Staehelin, J. Hoigne, Decomposition of ozone in water in the presence of organic solutes acting as promoters and inhibitors of radical chain reactions, Environ. Sci. Technol. 19 (1985) 1206–1213.

Journal of Hazardous Materials 195 (2011) 378–382

Contents lists available at SciVerse ScienceDirect

Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat

Influence of microbial adaption and supplementation of nutrients on the biodegradation of ionic liquids in sewage sludge treatment processes Marta Markiewicz a , Stefan Stolte b , Zofia Lustig a , Justyna Łuczak a , Michał Skup c , Jan Hupka a , Christian Jungnickel a,∗ a

Department of Chemical Technology, Chemical Faculty, Gda´ nsk University of Technology, ul. Narutowicza 11/12, 80-233 Gda´ nsk, Poland Center for Environmental Research and Sustainable Technology, University of Bremen, Leobener Strasse UFT, D-28359 Bremen, Germany c Department of Environmental Analysis, Faculty of Chemistry, University of Gda´ nsk, Sobieskiego 18, 80-233 Gda´ nsk, Poland b

a r t i c l e

i n f o

Article history: Received 8 June 2011 Received in revised form 17 August 2011 Accepted 18 August 2011 Available online 24 August 2011 Keywords: Ionic liquids Biodegradation OECD 301 Adaptation Supplementation

a b s t r a c t As ionic liquids are winning more attention from industry as a replacement of more hazardous chemicals, some of their structures have the potential to become persistent pollutants due to high stability towards abiotic and biotic degradation processes. Therefore it is important to determine the hazard associated with the presence of ILs in the environment, for example biodegradation under real conditions. Standard biodegradation testing procedures generally permit pre-conditioning of inoculum but do not allow for pre-exposition to the test substance. These are usually conducted in a mineral medium which does not provide additional organic nutrients. Though very valuable, as a point of reference, these tests do not fully represent real conditions. In in situ conditions, for example in wastewater treatment plants or natural soils and water bodies, the presence of readily available sources of energy and nutrients as well as the process of adaptation may often alter the fate and metabolic pathways of xenobiotics. Our results have shown that these are the opposing processes influencing the biodegradation rate of ILs in sewage sludge. The results have significant practical implications with respect to the assessment of biodegradability and environmental fate of ILs and other xenobiotics in environmental conditions and their potential remediation options. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Industrial development during the last decades resulted in increased pollution of the environment by xenobiotics. Due to this, the need for understanding the impact of toxic compounds on microbial populations and the catabolic degradation pathways of xenobiotics has arisen. Thus standardized biodegradability and toxicity test were developed to allow for classification of xenobiotics according to the environmental hazard they pose. Bearing in mind the definition of xenobiotics, as man-made chemicals foreign to organisms which inhabit the environment, their biodegradation rate in natural soils and waters is in most cases much lower than that of natural compounds. Nevertheless structural similarities to biomolecules can result in relatively high biodegradation rates if enzymes of low substrate specificity are present. Factors which may influence this rate, among others, include microbial adaptation and availability of additional nutrients [1]. Ionic liquids (ILs) as a non-conventional class of novel solvents are becoming increasingly important owning to a number

∗ Corresponding author. Tel.: +48 58 3472334. E-mail address: [email protected] (C. Jungnickel). 0304-3894/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jhazmat.2011.08.053

of desirable characteristics including negligible volatility, nonflammability, high thermal stability, low melting point, broad liquid range and controlled miscibility with organic compounds or water [2–5]. The negligible volatility limits their impact on air quality, but their release to the environment may affect soil and water. Moreover, some IL structures have the potential to become persistent pollutants due to their high stability towards abiotic and biotic degradation processes. Therefore it is important to determine the hazard associated with the presence of ILs in the environment. Adaptation is defined as a change in the microbial community that leads to an increase in the biodegradation rate, or maximal biodegradable concentration of a given xenobiotic as a result of previous exposure. Examples of such adaptation processes are e.g. rapid degradation of p-nitrophenol by aquatic microorganisms [6] and enhanced degradation rates after elongated exposure of subsurface soil communities to m-cresol, m-aminophenol and aniline [7]. Mechanisms of adaptation usually involve processes such as genetic mutation or horizontal gene transfer, induction of specific enzymes which enhance the degradative capacity of the entire community, and population change such as selective growth of certain strains [8]. All mechanisms may take place simultaneously, or one may dominate and exact prediction of which will occur is not possible [9]. Furthermore it should be noted, that adaptation does

M. Markiewicz et al. / Journal of Hazardous Materials 195 (2011) 378–382

B

1

100

0.9

90

0.8

80

0.7

70

biodegradation [%]

concentration [mM]

A

0.6 0.5 0.4 0.3

60 50 40 30

0.2

20

0.1

10

0 0

5

10

15

20

25

time [days]

379

0

0

5

10

15

20

25

time [days]

), Fig. 1. (A) Biodegradation curves of [OMIM][Cl] as a sole source of carbon (—), with supplementation of glucose (- - -), with supplementation of synthetic feed ( ). (B) Biodegradation (%) normalized for sorption to sewage sludge flocs of [OMIM][Cl] as a sole source of carbon (—), with supplementation of glucose sorption control ( ). (- - -), with supplementation of synthetic feed (

not necessarily occur in every case. Aelion et al. did not observe any adaptation of subsurface microbial communities to chloro- and trichlorobenzene after an eight months adaptation period [7]. Similarly Nyhoim et al. did not note any increase in the biodegradation rate after pre-exposure of activated sewage sludge to aniline and pentachlorophenol [9]. The specific reason for this remains unclear though many theories exist. The most probable reasons include lack of complete enzyme systems within the population, accumulation of toxic degradation products, binding with enzymes causing inactivation or insufficient cell density of inoculum [10]. One of the few papers which discusses the adaptation of soil microorganisms to ionic liquids proposes that the electron-donor ability of the IL effect the biodegradability [11]. A number of research groups have performed biodegradation tests with alkyl substituted imidazolium cations using activated sewage sludge [12–15]. In most cases ILs were used as a sole source of organic carbon and organic nitrogen. This is especially important, because it should be remembered that in wastewater treatment plants or natural environments, other organic substrates are present, which might be preferentially degraded or co-metabolized with the primary contaminant resulting in lower biodegradation rates [16]. Romero et al. [17], discussed the biodegradability of imidazolium ILs in the presence of additional carbon source. It was found that the ILs tested were not biodegradable when D-glucose was available. However, ILs with no additional carbon were also not degraded (2–10%), which is in contrast to other research where, e.g. complete primary biodegradation of 1-methyl-3-octylimidazolium chloride [OMIM][Cl] was shown [12,18]. Results of Romero et al., though very interesting, should be treated with caution due to the very short duration of the test (five days) as well as lack of collaborating results in literature. Standard biodegradation testing procedures generally permit pre-conditioning of inoculum (aeration in the presence of a mineral medium) but do not allow for pre-exposition to the test substance. The purpose of this is to provide repeatable results enabling comparison and standardization of biodegradation rates of different chemicals usually for regulatory purposes. Though very valuable, as a point of reference, these tests do not fully represent real conditions [19]. Therefore, to more accurately predict biodegradation under real conditions it is beneficial to take adaptation into account especially if biodegradation requires induction of specific metabolic pathways, e.g. aromatic ring break-down [6,20]. One of the few works which discuss the adaptation of microorganisms to ILs, conducted by Stolte et al., found a sixfold increase in the biodegradation rate of [OMIM][Cl] over a period of 31 days [12]. Additionally, Docherty et al. observed complete biodegradation of

hexyl-methylimidazolium bromide after extending duration of the test and concluded that though IL could not be classified as readily biodegradable it is not expected to persist in the environment [20]. The aim of this paper is to describe the effect of additional substrates and pre-exposition of bacteria to IL on the rate of biodegradation, and thereby discuss the relevance of including preexposition in standardized tests.

2. Experimental methodology 2.1. Modified OECD 301A DOC Die-Away test – supplementation The ionic liquid used in the test was [OMIM][Cl] provided by Merck KGA, Darmstadt, Germany. The sewage sludge (dry mass 6.5 g L−1 ) was taken from the aeration chamber of the ´ “Gdansk – Wschód” municipal wastewater treatment plant, ´ Poland. Primary degradation was detected by direct Gdansk, determination of the substrate by HPLC – UV. Eight test flasks containing 0.5 L of sewage sludge flocs and mineral medium composed of: 8.5 mg L−1 KH2 PO4 , 21.75 mg L−1 K2 HPO4 , 22.3 mg L−1 Na2 HPO4 ·2H2 O, 1.7 mg L−1 NH4 Cl, 27.5 mg L−1 CaCl2 , 22.5 mg L−1 MgSO4 ·7H2 O and 0.25 mg L−1 FeCl3 dissolved in water were prepared as recommended by OECD procedure [21]. Subsequently, a solution of [OMIM][Cl] was added to yield the concentration of 1 mM and the amount of test solution was made up to 1 L. Each test concentration was conducted in duplicate. Two test flasks were additionally supplemented with glucose and two with synthetic sewage feed (16 g of peptone, 11 g of meat extract, 3 g of urea and 0.7 g NaCl dissolved in 1 L of water). Nutrients were added three times a week, 0.36 g and 2.5 mL, respectively. Also blank samples (without test substance) and chemically sterilized negative controls were prepared. All test vessels were aerated and analytical samples were collected in duplicate at specific time intervals. Mass loss due to evaporation was compensated at every collection interval.

2.2. Modified OECD 301A DOC Die-Away test – adaptation The test vessels were prepared as previously described. Sewage sludge from the same source was used (dry mass 5.5 g L−1 ). Increasing concentrations of [OMIM][Cl] (1 mM, 1.5 mM, 2 mM, 2.5 mM) were added every fortnight. The total time for the adaptation test was two months. All vessels were aerated and analytical samples were collected in duplicate at specific time intervals.

380

M. Markiewicz et al. / Journal of Hazardous Materials 195 (2011) 378–382

conditions were set in accordance to [22] with a capillary voltage of 3500 V, drying gas flow-rate of 5 L min−1 , drying gas temperature at 300 ◦ C and nebulizer at 70 psi.

6

concentraon [mM]

5

3. Results and discussion

4

3.1. Supplementation

3

2

1

0

1

4

7

11 13 17 18 22 25 29 30 32 36 39 44 45 46 50 53 58

me [days] Fig. 2. Biodegradation of [OMIM][Cl] by adapted sewage sludge community (black bars) and abiotic control (grey bars).

2.3. HPLC analysis Analytical samples were centrifuged to remove solids and a supernatant was taken for the HPLC-UV analysis. A Perkin Elmer Series 200 HPLC consisting of a chromatographic interface (Link 600) binary pump, UV/VIS detector, vacuum degasser and Rheodyne injection valve were used. For IL’s cation separation C6-Phenol (Phenomenex) 150 × 4.6 mm column was used in conjunction with detection by UV adsorption at 218 nm. As a mobile phase 27% acetonitrile/water + 0.1% (v/v) trifluoroacetic acid at the flow rate of 0.8 mL min−1 was applied. For preparation of HPLC mobile phase HPLC – grade acetonitrile, Lab – Scan (Dublin, Ireland) and spectrophotometric grade trifluoroacetic acid (Sigma–Aldrich, Germany) were used. 2.4. Metabolites analysis Additional analytical samples were taken from vessels containing live and chemically sterilized inocula at the end of adaptation tests. Samples were centrifuged to remove solids and the supernatant was diluted hundredfold (biotic sample) or thousandfold (abiotic sample) with a 9:1 methanol–water mixture resulting in approximately 5 ␮M concentration of parent compound in all samples. Subsequently samples were analyzed for the parent compound and metabolites by electrospray ionization mass spectrometry equipped with ion trap detector (Brucker-Daltonic GmbH, Germany). Mass spectra for cations were acquired in the positive ion mode in the scan range of m/z+ 50–300. The ESI source

concentration [mM]

2.5 2 1.5 1 0.5 0

0

5

10

15

20

25

30

35

time [days] Fig. 3. Comparison of biodegradation rate of 2 mM [OMIM][Cl] conducted by adapted (solid line) and non-adapted (dashed line) activated sewage sludge community.

The initial, fast decrease in [OMIM][Cl] concentration was due to sorption of [OMIM][Cl] onto flocs of activated sewage sludge (shown in Fig. 1A). The sewage sludge organic matter (especially extracellular polymeric substances) can act as a ‘buffer’ for the IL, initially decreasing the bioavailable concentration and thereby mitigating its toxicity. The supplementation with glucose and synthetic feed increased the time of [OMIM][Cl] primary biodegradation (Fig. 1B). It can be observed that after approximately 14 days the biodegradation of 0.2 mM [OMIM][Cl] remaining after sorption was completed only when no supplements were added. After more than 20 days biodegradation was accomplished in the sample where synthetic feed was added. In the vessel with glucose supplementation complete biodegradation was not observed within the test timeframe. [OMIM][Cl] biodegradation with synthetic feed cannot be explained using the diauxie effect [23], as the feed was added continuously for the duration of the test. For [OMIM][Cl] as a nominal source of organic carbon and organic nitrogen complete primary biodegradation is achieved within 14 days. When synthetic feed is present the biodegradation rate is clearly reduced, even though complete primary degradation is achieved within 23 days. When supplemented with glucose, providing easily available source of organic carbon, [OMIM][Cl] is utilized in less than 20%. Therefore the presence of other nutrients in the sewage or within the environmental media in general can have a strong influence on the biodegradability of ionic liquids. Also compounds which have been classified as “readily biodegradable” might present recalcitrance towards biodegradation under real environmental conditions which has to be taken into account when evaluating their fate in the environment. Generally, the reduced biodegradation rate of [OMIM][Cl] in the presence of glucose is consistent with the research conducted by Lewis et al. [24] where the addition of organic carbon significantly decreased the degradation of p-cresol. The addition of synthetic feed containing organic carbon and nitrogen decreases the rate of biodegradation (relative to non-supplemented tests). Similar results were obtained by Swindoll et al. for p-nitrophenol [25]. On the other hand, Piekarska et al. showed that the addition of other sources of organic carbon and nitrogen increased the efficiency of degradation of diesel oil [26]. This can be explained by the difference in the chemical structure of the primary substrates. Nitrophenol is a pure aromatic compound, biodegradation of which requires the induction of a specific metabolic pathway. Diesel oil is a mixture containing mostly linear or branched hydrocarbons, which are degraded through the ␤-oxidation pathway. The enzymatic systems for this pathway are relatively common in most soil and sewage microorganisms. To summarize, addition of easily available organic carbon and nitrogen sources seems to facilitate biodegradation when the primary pollutant itself is relatively easy to degrade. It can be anticipated that xenobiotics containing structures which are commonly recognized as poorly biodegradable will not be metabolized if microorganisms can obtain carbon from other sources. This hypothesis seems to hold true when the secondary substrate is organic. Inorganic supplements were proven to facilitate biodegradation. The addition of inorganic carbon (NaHCO3 ) and inorganic nitrogen (NH4 Cl) has previously been shown to increase the rate of biodegradation of xenobiotic compounds [9,24]. Utilization of

M. Markiewicz et al. / Journal of Hazardous Materials 195 (2011) 378–382

m/z +

381

intensity biotic sample (0.5 mM OMIM; 1:100 diluted)

abiotic sample (5 mM OMIM; 1:1000 diluted)

195

3*105

3*105

211

1*104