Environmental Issuesand Waste Management Technologies in the Cermaic & Nuclear (Ceramic Transactions Series)

Environmental Issues and Waste Management Technologies in the Ceramic and Nuclear Industries VIII Related titles publ...

Author: S.K. Sundaram

21 downloads 1234 Views 26MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Environmental Issues and Waste Management Technologies in the Ceramic and Nuclear Industries VIII

Related titles published by The American Ceramic Society: Environmental Issues and Waste Management Technologies in the Ceramic and Nuclear Industries VII (Ceramic Transactions,Volume I 32)

Edited by Gary L. Smith, S.K. Sundaram, and Dane R. Spearing

02002, ISBN I-57498- 146-3 Environmental Issues and Waste Management Technologies in the Ceramic and Nuclear Industries VII (Ceramic TransactionsVolume I 19) Edited by Dane R. Spearing, Gary L. Smith, and Robert L. Putnam 0200 I, ISBN 1-57498- I 16- I The Magic of Ceramics

By David W. Richerson

02000, ISBN 1-57498-050-5 Environmental Issues and Waste Management Technologies in the Ceramic and Nuclear Industries V (Ceramic TransactionsVolume I 0 7)

Edited by GregoryT. Chandler and Xiangdong Feng

02000, ISBN 1-57498-097- I Ceramic Innovations in the 20th Century Edited by John B.Wachtman Jr. 0 I 999, ISBN I-57498-093-9 Environmental and Waste Management Technologies in the Ceramic and Nuclear Industries IV (Ceramic TransactionsVolume 93)

Edited by J. C. Marra and G.T. Chandler 0 1999, ISBN I-57498-057-2 Environmental and Waste Management Technologies in the Ceramic and Nuclear Industries 111 (Ceramic TransactionsVolume 8 7) Edited by D. Peeler and J. C. Marra 0 1998, ISBN I-57498-035- I Environmental and Waste Management Technologies in the Ceramic and Nuclear Industries I1 (Ceramic TransactionsVolume 72) EditedV. Jain and D. Peeler 0 1996, ISBN: I-57498-023-8 Environmental and Waste Management Technologies in the Ceramic and Nuclear Industries (Ceramic TransactionsVolume 6 I ) Edited by! Jain and R. Palmer 0 1995, ISBN I-57498-004- I Environmental and Waste Management Issues in the Ceramic Industry I1 (Ceramic TransactionsVolume 45) Edited by D. Bickford, S.Bates,V. Jain, and G. Smith 0 I 994, ISBN 0-944904-79-3 Environmental and Waste Management Issues in the Ceramic Industry I (Ceramic TransactionsVolume 39) Edited by G. B. Mellinger 0 I994, ISBN I-944904-7I-8 For information on ordering titles published by The American Ceramic Society, or to request a publications catalog, please contact our Customer Service Department at 6 14-794-5890 (phone), 6 14-794-5892 (fax), [email protected] (e-mail), or write to Customer Service Department, 735 Ceramic Place, Westerville, OH 4308 I , USA. Visit o u r on-line book catalog at www.ceramics.org.

Environmental Issues and Waste Management Technologies in the Ceramic and Nuclear Industries VIII Proceedings of the Science and ~ e c h ~ ino A~dressing ~ o ~ ~ y Ceramic €nvironmen~alIssues in the Ceramic l n ~ u s and Science and Technology for the Nuclear Industry symposia held at the /04* An~uu/Meeting ofThe American Ceramic Sociely 2002 in St Louis, Missouri A ~ r i28-30, /

€ ~ i by~ e ~ S.K. Sundaram

Pacific Northwest National laboratory

Dane R. Spearing

Los Alamos National Laboratory

john D.Vienna Pacific Northwest National Laboratory

Published by The American Ceramic Society 735 Ceramic Place Westerville, Ohio 4308 I ~.ceramics.or~

Proceedings of the Science and Technology in Addressing Environmental Issues in the Ceramic Industry and Ceramic Science and Technology for the Nuclear Industry symposia held at the I04* Annual Meeting ofThe American Ceramic Societ).:April 28-30,2002 in S t Louis, Missouri

Copyright 2003,The American Ceramic Society. All rights reserved. Statements of fact and opinion are the responsibility of the authors alone and do not imply an opinion on the part of the officers,staff,or members ofThe American Ceramic SocietyThe American Ceramic Society assumes no responsibility for the statements and opinions advanced by the contributors to its publications or by the speakers at its programs. Registered names and trademarks, etc., used in this publication, even without specific indication thereof, are not to be considered unprotected by the law. No part of this book may be reproduced,stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical,photocopying, microfilming, recording, or otherwise, without written permission from the publisher:

Authorization t o photocopy for internal or personal use beyond the limits of Sections I07 and I08 of the US. Copyright Law is granted by the American Ceramic Society, ISSN 1040- I I 22 provided that the appropriate fee is paid directly t o the Copyright Clearance Center; Inc.,222 Rosewood Drive, Danvers, MA 0 I923 USA, www.copyright.com. Prior t o photocopying items for educational classroom use, please contact Copyright Clearance Center; Inc. This consent does not extend t o copying items for general distribution or for advertising or promotional purposes or to republishing items in whole or in part in any work in any format. Please direct republication or special copying permission requests to the Senior Director; Publications,The American Ceramic Society, PO Box 6 136,Westerville,Ohio 43086-6 136, USA.

Cover photo: Scanning electron micrograph o f a crushed glass sample is courtesy ofTAkai, D. Chen, Y Yamamoto,7: Shirakami, K. Urabe, K Kuraoka, and IYazawa, and appears as Figure I in their paper “Sodium Extraction from Waste Glass by Acid Leaching to Obtain Silica Source for Construction Materials,” which begins on page 39.

For information on ordering titles published byThe American Ceramic Society, or t o requert a publications catalog, please call 6 I 4-794-5890. Printed in the United States of America.

4 3 2 1-05 04 03 02 ISSN 1042-1 122 ISBN 1-57498-159-5

Preface

............................................

xi

Recycling of Ceramics and Glasses Industrial Applications for Spent Refractory Materials jJ? Bennett and I(.-S. Kwong

......... 3

Ceramic-Based Magnetic Extractants for Removal of Organics from Water .................................

IS

A. Apblett, S.M. AI-Fadul, and T.M.Trad

Investigation on a Recycling Process ofwaste Colored Glass .............................. D, Chen, H. Masui,TAkai, and TYazawa

Use of Mid-Delaware River Dredge Sediment as a Raw Material in Ceramic Processing ..................... K, Hill and R.A. Haber

1

.23

3I

Sodium Extraction from Waste Glass by Acid Leaching t o Obtain a Silica Source for Construction Materials . . . . . . . . . 39 T. Akai, D. Chen,YYamamoto,%Shirakami, K.Urabe, K. Kuraoka,and IYazawa

Emissions in Glass and Ceramic Industries Analysis of Emissions from Nitrate Containing Glasses S, Luo and L.E,jones

. . . . . . . 49

l Characterizing Particulate Emissions using MicrometernScale X-Ray Fluorescence

................... 59

J.F. Shackelford, FIB, Kelly, S.S. Cliff, M. Jimenez-Cruz,and TA. Cahill

1

Dilatometry and Mass Spectrometry Study of the Decomposition and Sintering of Calcium Carbonate K, Feng and S.J. Lombardo

~

V

.........67

Lead-Free Electronics: Current and Pending Legislation J.M.Schoenung

. . . . . . 75

First Delisting Petition Approval by the US EPA for a Vitrified Mixed Waste ...............................

.83

J.B.Pickett, C.M.Jantzen,and L.C. Martin

Characterization of Defense Nuclear Waste using Hazardous Waste Guidance: lnsights on the Process at Hanford . . . . . . . . . 95 M. Lerchen, L. Huffman,W. Hamel, and K. Wiemers

Effect of TransitionIN on-Transition Metal Modification on the Activity of Ga2O3-AI2O3Catalyst for NOx Reduction by Hydrocarbon under Oxygen-Rich Conditions . . . . . . . . . . . I05 M.H.Zahir; S. Katayama, K.Maeda, and M.Awano

Vitrification Technology and Melter Disassembly COGEMA Experience in Operating and Dismantling HLW Melter .............................

I13

R. Do-Quang, J.L.Desvaux, I? Mougnard,A. Jouan,and C. Ladirat

Conceptual Methods for Disposal of a DWPF Melter and Components .............................

123

M.E. Smith, D.F.Bickford, F.M. Heckendorn, and E.M. Kriikku

Evaluation of Crystallinity Constraint for HLW Glass Processing ...............................

133

I? Hrma,J. Maty65, and D.-S. Kim

Ruthenium - Spine1 Interaction in a Model High-LevelWaste (HLW) Glass ........................

I 4I

TM.Willwater,J.V.Crum, S.M. Goodwin, and S.K. Sundaram

Glass Formulation and Testing Interim Models Developed to Predict Key Hanford Waste Glass Properties using Composition . . . . . . . . . . . . . . . I 5 I J.D.Vienna,D-S. Kim, and I? Hrma

Relationship between Liquidus Temperature and Solubility I? Hrma and J.D.Vienna

vi

. . . . I59

Glass Formulation for INEEL Sodium Bearingwaste J.D.Vienna,D.-S. Kim, and D.K.Peeler

Vitrification of Korean Low-Level Waste

. . . . . . . . 169

. . . . . . . . . . . . . . . . . I77

L.O. Nelson, I? Kong, G.Anderson, K. Choi, C.-W. Kim, and S.-W. Shin

Phase Equilibria7Viscosity7 Durability, and Raman Spectra in the System for Idaho Nuclear Waste Forms . . . . . . . . . . . . . I85 S.V. Raman, B.A. Scholes,A. Erickson, and A.A. Zareba

Measurement of SimulatedWaste Glass Viscosity R.F. Schumacher;TB. Edwards, D.K. Peeler; and A.G. Blum

. . . . . . . . . . I99

Hanford Tank Waste Treatment Hanford Low-LevelWaste Form Performance for Meeting Land Disposal Requirements . . . . . . . . . . . . . . . . . . .2 R.F. Schumacher;C.L. Crawford, N.E. Bibler; D.M. Ferrara, H.D. Smith, G.L. Smith,J.D.Vienna,D.B.Blumenkranz, D.J.Swanberg, I.L. Pegg, and IS. Muller

Leaching Mechanism of Borosilicate Glasses under TCLP Conditions ..............................

2I5

H. Gan and I.L. P e g

Electrochemical Studies of SuIfate-Containing Waste Glass Melts .................................

.225

LVidensky,H. Gan,A.C. Buechele, and I.L. Pegg

Durability Testing and Modeling Modeling High-Level Waste Glass Degradation in Performance Assessment Calculations . . . . . . . . . . . . . . . . . . .235 W.L. Ebert

Waste Glass Corrosion: Some Open Questions I?Hrma,J.D.Vienna,and J.D.Yeager

. . . . . . . . . . . 245

Vapor Phase Hydration of Glasses in H 2 0and D 2 0

. . . . . . . . 253

TR. Schatz,A.C. Buechele, C.F. Mooers, R.Wysoczanski, and I.L. Pegg

vii

Modeling Fluid Chemistry Inside a Waste Package Due t o Waste Form and Waste Package Corrosion . . . . . . . . . . . . . 263 V. Jain and N. Sridhar

Leaching Full-Scale Fractured Glass Blocks Y Minet and N. Godon

. . . . . . . . . . . . . . .275

Development of Sensors for Waste Package Testing and Monitoring in the Long-Term Repository Environments V. lain, S. Brossia, D. Dunn, and L.Yang

Corrosion of Partially Crystallized Glasses I? Hrma, B.J.Riley, and ].D.Vienna

. . . . . 283

. . . . . . . . . . . . . . . 29 I

Ceramic and Alternative Waste Forms Development of Titanate Ceramic Wasteforms and Crystal Chemistry of Incorporated Uranium and Plutonium . . . . . . . . 301 E.R.Vance

Substitution of Zr, Mg,AI, Fe, Mn, CO,and Ni in Zirconolite, CaZrTi20, ..............................

3I3

E.R.Vance,J.V.Hanna, B.A. Hunter, B.D. Beg, D.S. Perera, H. Li, and Z.-M. Zhang

Effects of Sub-Surface Damage Induced by Mechanical Polishing on LeachTesting of Cesium-Bearing Hollandite M.L. Carter; E.R.Vance,DJAttard, and D.R.G.Mitchell

. . . . . 321

Iron Phosphate Glasses for Vitrifying Sodium BearingWaste C.-W. Kim, D. Zhu, D.E. Day, and D. Gombert

. . 329

Phosphate Glasses for Vitrification of Waste with High Sulfur Content.. ...............................

337

Solubility of High-Chrome Nuclear Wastes in Iron Phosphate Glasses ..................................

347

D.-S.Kim, ].D.Vienna, I? Hrma, and N. Cassingham

W. Huang, C.-W. Kim, C.S. Ray, and D.E.Day

Development of a Sampling Method for Qualification of a Ceramic High-LevelWaste Form . . . . . . . . . . . . . . . . . . . 355 TI? O’Holleran and K.J.Bateman

...

Vlll

Microwave Heating for Production of a Glass Bonded Ceramic High-LevelWaste Form ............... , 3 6 3 TF?O’Holleran

Morphology and Composition of Simulant Waste Loaded Polymer Composite - Phase Inversion, Encapsulation, and Durability .......................... H.D.Smith, G.L. Smith, G. Xia, and B.j.j. Zelinski 93NbMAS NMR of Niobium Containing Silicotitanate Exchange Materials ....................... B.R. Cherry, M. Nyman, and TM. Alam

37 I

377

Selective Absorption of Heavy Metals and Radionuclides from Water in a Direct-to-Ceramic Process .............. 385 B.F? Kiran,A , ~Apblett, , and M. Ch~hboun;

ix

This page intentionally left blank

In 2002,The American Ceramic Society hosted several symposia focusing on eight broad topics at its Annual Meeting. Two key symposia, Ceramic Science and Technology for the Nuclear Industry and Science and Technology in Addressing ~nvironmentalissues in the Ceramic ~ndustry, clearly illustrate the delicate balance that exists among the environment, the processes/technologiesthat have been used in glass and ceramic industries, as well as the wastes both nuclear and nonunuclear (hazardous) that have been generated. Ceramics and glasses play a critical role in the nuclear industry. Nuclear fuels and waste forms for low-level and high-level radioactive, mixed, and hazardous wastes are primari~y either ceramic or glass. With increasingly stringent environmental regulations and demands that are placed on our limited natural resources, it is critical to identify and adequately address environmen~lissues in the ceramic and glass industry to ensure longevity and success. In ceramic/glass manufacturing, companies are beginning to focus on green ceramics, performing life cycle analyses, and adopting environmental stewardship to manufacture environmen~llyfriendly products. Effective and responsible environmental stewardship is becoming increasingly more important in the world. The symposia and subsequent proceedings help foster continued scientific understanding, techno~ogicalgrowth, and environmental stewards~ip within the fields of ceramics, glass, and envjronmen~l/nuclearengineering. This proceedings combines key papers that were presented at the above mentioned symposia duringThe American Ceramic Society I 04th Annua~ Meeting & Exposition held April 28-30,2002 in St. Louis, Missouri. This is the fourteenth volume published by The American Ceramic Society in the areas of waste management and environmental issues in relation to ceramics. Previous proceedings on nuclear waste management and environmental issues date from I983 and include Advances in Ceramics volumes 8 and 20 and Ceramic Transactions volumes 9,23,39,45,6 I , 72,87, 93, t 07, I 19, and i32.

xi

The editors thank Robert L. Putnam, Boyd Clark, Linda E. Jones,Jeff Kohli, G. L. Smith and Carol M. Jantzen for their help in organizing the symposia and James I? Bennett and James C. Marra, for their contribution in chairing the sessions and keeping the presentations running smoothly. The editors appreciate the support from John Marra,Vijay Jain, Greg Chandler, and the Nuclear and EnvironmentalTechnology Division and Glass and Optical Materials Division of The American Ceramic Society. The editors acknowledge most importantly, the authors and reviewers. Without them, a high quality proceedings volume and timely publication would not be possible. The editors also appreciate all the support from Kevin G. Ewsuk and Chris Schnitzer. Lastly, the editors thankTeresa Schott of the Pacific Northwest National Laboratory and the book publishing team at The American Ceramic Society: Mary Cassells, Sarah Godby, Greg Geiger, and Mark Mecklenborg. Their support and contributions were instrumental in the publication of this volume. S.K. Sundaram Dane R. Spearing John D.Vienna

xii

Recycling of Ceramics and Glasses


INDUSTRIAL APPLICATIONS FOR SPENT REFRACTORY MATERIALS James P. Bennett, Kyei-Sing Kwong Albany Research Center - USDOE 1450 Queen Ave. SW Albany, OR 97321 Phone: 541-967-5983 FAX: 541-967-5845 E-Mail: Bennett@,alrc.doe.gov ABSTRACT The reusehecycling of spent refractory materials by industry is limited to a few companies because of the lack of economic and legislative driving forces. For most users of refractories in the United States, it is more economical to landfill spent refractory materials than to reuse/recycle the material. Where recycling is successful, applications for the spent refractory materials are primarily as a refractory raw material and as a slag conditioner. Applications for spent refractory materials in steel, aluminum, glass, and brass industries will be evaluated, with emphasis on what cormnon elements make up these programs. INTRODUCTION The reuse of refractory materials after removal from service is not commonly practiced by industry because of economics and a lack of legislated driving forces, with most spent refractory material being land filled after removal. Structured programs for the recyclingheuse of spent refractories do not exist. Interest in recycling has been cyclic, driven in part by legislation, environmental concerns, product stewardship, company environmental policy, high landfill costs, or a lack of landfill space. Marketplace forces have also started to play a role in recycling through companies seeking IS0 14000 certification. Historical recycling of spent refractory materials has been practiced by refractory companies, with off specification material, floor sweepings, old refractory formulations, excess material, grinding, and some spent material often added back to refractory batch formulations. Where used in these formulations, the spent refractory acts as a grog, reducing drying and firing shrinkage, giving volume stability, and reducing energy consumption during firing. In the United To the extent authorized under the laws of the United States of America, all copyright interests in this publication are the property of The American Ceramic Society. Any duplication, reproduction, or republication of this publication or any part thereof, without the express written consent of The American Ceramic Society or fee paid to the Copyright Clearance Center, is prohibited.

Environmental Issues and Waste Management Technologies VIII

3

States, large scale recycling of spent refractory materials is limited to a few industries, such as the glass, where the driving force is environmental regulations due to the use of chrome oxide as a raw material in the refractory. Chrome oxide changes valence from the trivalent to hexivalant state under certain use conditions, causing it to be classified as a hazardous waste. Refractory companies have been reluctant to take spent material back from users for a number of reasons. Top among concerns include the amount and frequency of spent refractory available and concerns over the performance of materials that contain spent refractory. Other concerns center on refractory contamination and its beneficiation costs, shipping distances to a refractory user, andor limited demand for the processed materials. These factors can make beneficiating and recycling material uneconomical when compared to virgin, mined material costs, consistency, and availability. Attempts to recycle in the United States have tended to follow economic cycles, with some companies interested in recycling during good economic times, but focusing on core business activities during a recession. Mergers and consolidation in the refractory industry have not helped recycling programs, with efforts at recycling decreasing or being eliminated as consolidation in the refractory industry has occurred. Established programs with refractory users in the glass industry have survived because of the economic costs associated with hazardous waste disposal. More emphasis has been placed on recycling in Europe because of a lack of landfill space, because the distance between refractory users and producers is generally small, and because regulations have been legislated that impact waste disposal. Some steel producers in Europe, for example, mandate the return of spent refractory material as part of their contract with the refractory producer. In Japan, recyclers have placed emphasis on recyclingfreusing spent refractory materials because of decreasing landfill space, focusing on basic materials as slag conditioners and on high value slide gate valves

’.

Refractory users who want spent refractory reusedrecycled must keep in mind that spent refractory must be utilized in a product for some application, and that there are refractory users who do not want refractories containing spent material because of concerns over decreased service life. Other refractory users do not share the performance concerns with the reuse of recycled materials, but don’t have the storage space for spent material or don’t have refractory producers or recyclers in close proximity to process the spent material economically into a product. Ideally, a refractory user would like a material with the longest life that is the most cost effective to use in an application. Most recycling decisions are based on economics - does it make economic sense to recycle or to landfill spent

4

Environmental Issues and Waste Management TechnologiesVIII

material? Potential applications for spent refractory materials are listed in Table I. Table I - Possible applications for went refractory materials Roofing granules Refractory component Tile body component Insulating powder Raw material for glass Component in cement, aggregate for Highway road aggregatehubbase concrete Ferro alloy (high chrome containing Building component materials) Abrasive Fuel source (SIC, C containing materials) Soil conditioner Filler for bulk items Carbon, silicon source Landscape material Grog in ceramic materials Soil stabilization Sulphur removal in the ladle steels Slag conditioner Waste neutralizatiodtreatment (acids, pathogens) Contamination of refractory materials is an area of great concern in recycling. It can range from process infiltration into the refractory to the mixing of zoned linings on tearout from the furnace. Beneficiation of contaminated material makes up a high portion of refractory recycling costs, underscoring the importance of keeping the material clean on tearout ’. Not only does clean material have more potential applications, it also has lower beneficiation costs. Recycled refractory material uses can range from low to high volume applications with varying material values.

A number of companies recycle spent refractory materials. These companies have established markets or applications for the spent material and take in no more material than they can market. They possess the specialized knowledge and equipment for processing the spent refractory material. Limitations exist on processing the material, which center around contamination and beneficiation costs, shipping distances andor limited demand for the processed materials, factors that can make the recycled material uneconomical when compared to virgin, mined material costs, constancy, and availability. Regardless of who does the beneficiation of the refractory material, flow sheets tend to follow the setup as shown in figure 1. Materials are typically sorted on removal, especially if the refractory lining is zoned or if heavy contamination of the refractory has occurred. This is done by hand at a few recyclers. The spent refractory is next crushed, dried, and screened. Other beneficiation circuits can be added to this flow sheet, but a simple flow sheet is typically used. This type of


5

beneficiation circuit is typically followed by industries such as glass, steel, alwninum, and brass. Common beneficiation flow sheets and applications for beneficiated material at a number of industries in these areas will be discussed. On sitelln place refractory beneficiationlremoval

*

rl Primary crushing

Drying

Figure 1 - General beneficiation process for refractory materials GLASS INDUSTRY RECYCLING Recycling in the textile and fiberglass industry of the United States has been brought about by Federal regulations impacting the disposal of spent refractory materials containing chrome oxide, which in the presence of alkali/alkaline earth at elevated temperatures forms hexivalent chrome, a hazardous material. Regulations governing the amount of hexivalent chrome oxide in a material evolved from the Resource Conservation and Recovery Act of 1976, which was used as the basis for establishing the Toxicity Characteristics Leaching Procedure (TCLP). A leachable limit of 5 mg/L was established for chrome under this act. Because of this legislation, many steel producers have worked to eliminate any chrome containing materials from their plant. Chrome oxide containing refractories have many good properties that cannot be reproduced by other refractories in the glass industry, which has brought about the continued usage of this material in this field. Refractory producers have traditionally placed emphasis on performance, not recycleability. Little thought or effort has been made towards developing easily recycled refractory materials. With the disposal restrictions and costs placed on chrome oxide containing refractories, recycling programs for many chrome oxide-

6


containing products from the glass industry were developed over a period of years. In the U.S., magnesia-chrome spent refractory material from the reheat checkers of glass furnaces are recycled back as a refractory raw material after treatment by a proprietary water leaching process that removes soluble sulfates and chromates '. Fused material from glass furnaces is crushed and added back as a raw material feed for new refractories 5 , however, upper limits have been found to exist on the quantity of material that may be added before the performance of the material is adversely affected below acceptable levels. The performance issues are due to impurities present in the spent refractory.

STEEL INDUSTRY RECYCLING A number of refractory recycling programs have been successfully developed by both integrated and electric arc furnace (EM) steel producers, although these programs are not extensively used. More potential applications for internal reuse of spent refractory materials exist at integrated shops than EAF shops, if the mills are willing andor able to develop these applications. Flow sheets representative of spent refractory reuse as a slag conditioner and as a refractory raw material are shown in figure 2 (a-b). Magnetic material is typically removed from the refractory materials for remelt in the steelmaking furnaces, although this is not shown in both flow sheets. Mill service personnel, who perform other services at the steel mills, typically do magnetic material removal. Reuse of the spent refractory material as slag conditioners is typically limited to basic materials for EAF and basic oxygen furnace applications. Hydration and the speed of hydration limit basic material reuse. It is important to note that when basic materials are reused as slag conditioners, CaO and MgO levels in the sla has a large impact on the slag viscosity, slag foaming, and refractory wear . Additions of basic materials must be made with careful attention paid to the slag chemistry (Si02, CaO, MgO, FeO, and Al2O3), temperature, CaO/(Si02 + A1203) ratio, and the A1203/(Si02 + Al2O3) ratio.

B

Work done at MEFOS in Sweden has indicated spent alumina-silicate materials can be used for s u l k removal from ladle slags. Alumina raises slag viscosity, so one must be careful how much material is added. It is of interest to note that one of the flow sheets in figure 2 shows all spent refractory material being reused as a refractory raw material while the other shows 55 pct being returned to the refractory manufacturer (probably for reuse as a refractory raw material). Also note that some material (5 pct) is disposed of in a landfill in one of the steel mills. No applications could be found for this material because of contamination fi-om the process and because of zoned materials. Other application for materials removed from steel mills depend on material chemistry,


7

but typically involves applications as blast furnace flux,sinter plant additives, sludge thickeners, road and concrete aggregate, and sand blast materials.

AbO,

+-

t

Mill Service

$ . SiO,

MgOlC

*’

Segregated on Removal

Refractory Manufacturer

* Sinter Plant * Ladle Slag Conditioner

2 (b)

Figure 2 (a-b) - Process flow sheets used in the steel industry.

ALUMINUM INDUSTRY RECYCLING Recycling of spent refi-actory materials in the aluminum industry has been limited primarily to carbon bake furnace brick *. Refiactories removed fiom the primary andor secondary melting furnaces are typically monolithic materials and have issues associated with impurities, types of material, size of material removed from furnaces, variable bond strength, and possible anchors in the monolithic. It may be possible to utilize the high alumina materials removed the primary and secondary furnaces to satisfjr the alumina requirements of cement if local raw materials utilized by the cement producer are deficient in alumina. Alumina substitution research was conducted by the Univ. of Missouri at Rolla and

8


indicated this as a possible application '. A flow sheet from a company associated with reuse of carbon bake furnace brick from primary aluminum production is shown in figure 3. Furnace Tearout

(1.4 %)

I

v

Leave intact Reuse

I Storage

I

Segregate *iscard

Clean I Reuse Dense Firebricks

-1 1 I

IFB's

pfFqq Castable

Castable

Figure 3 - Process flow sheet used to recycle spent refractory materials from a carbon bake furnace of an aluminum Droducer.

Note that the process flow sheet shows that spent refractory material removed from the furnace was reused, recycled, or discarded, with over 98 pct of the refractory material being either recycled or reused. The bulk of the recycled refractory was reused in refractory castables and was used to rebuild the carbon bake furnace. The original refractory was approximately 50 pct A1203. Other applications for the dense alumina-silicate refractory were as a decorative aggregate and as a roadway aggregate. This recycling program had the support and commitment of management when enacted. The driving forces were environmental (reduced landfill space) and economic (reduced furnace rebuilding costs). Management endorsed and supported the plan. BRASS INDUSTRY RECYCLING Recycling by the brass and bronze industry is being driven by environmental regulations. Brass oftentimes contains lead, an element that slowly builds up in the brass furnace lining through vapor penetration or through direct contact with the metal and slag l0. The limit of lead in the furnace lining is found to be above the 5 mg/L allowable limit by TCLP testing. A process flow sheet for a brass and


9

bronze producer indicating how the spent refractory is disposedrecycled is shown in figure 4. Material segregation on removal from the furnace

I I I Crushing

Screening

I

-

t

I

Reuse slag conditioner hazardous waste landfill

Figure 4 - Process General beneficiation process for refractory materials

Refractory material that comes from brass and bronze finaces are about 70pct alumina. Because it is considered a hazardous waste and must be disposed of in a hazardous waste landfill, the brass and bronze recyclers have started to crush the refractory and reuse it in the furnace to saturated the slag with alumina, reducing refractory wear through corrosion. COMMON ELEMENTS OF RECYCLING PROGRAMS Companies that recycle have done much work to ensure their programs remain successful. Management typically has appointed an individual to head the program and he has the commitment and backing of management. An evaluation of the types and quantities of refractory materials coming to the plant, areas of usage within the plant, current “disposal” practices, and an assignment of costs associated with the spent refractory handling/disposal (estimate both current and hture costs) should be made. An assessment should also be made as to possible reuse or disposal options for the spent materials and consideration given to production process changes that would result in a refractory waste reduction. This assessment should be made according to the priorities listed in Table I1 l l . Table I1 - Priorities for spent refractory waste reuse/disposal l1 1. Reduction in the process 2. Internal reuse 3. External reuse 4. Treatment 5. Landfill

10


Reductions in the amount of refractory waste generated through an evaluation process can significantly reduce the amount of material to recycle. Technological advances, for example, have played a significant role in reducing both the amount and the frequency of spent refractory material generation. Examples are improved ladle and furnace design, maintenance and repair schedules; better slag and metallurgy control; slag splashing l 2 and improved refractory materials and installation practices. Improved refractory materials and installation practices include better customer service by refractory producers; hot patching; monolithic materials; zoning; gunning; subcontracted refractory installation, tearout and maintenance; the practice of the endless lining concept 13; and refractory wear sensing using laser wear profiling. After considering ways to reduce refractory wastes generated by the process, consideration should be given to internal utilization of the spent refractory material. An example of internal reuse would be the use of basic refractory materials as slag conditioners in electric arc furnaces or the reuse of spent MgO/C materials on the sidewalls of basic oxygen furnaces to freeze slag (slag splashing). Active recyclingheuse programs are necessary to reuse or recycle what refractory wastes are generated and to deal with the changing nature of refractory materials and the process. When initiating any recycling program, it is best to fxst recycle those materials that are the easiest to recycle or which have an established marketlapplication first, leaving a positive recycling experience in the plant. The examples of refractory recycling shown in figures 2-4 are of materials in applications easily accomplished by a refractory user. More difficult materials or materials with limited demand should be evaluated and processed at a later time, or can continue to be disposed of in a landfill if an economic justification or driving force does not exist. A common problem in any recycling program involves the crushing operation where dust can be generated. Mill services in the U.S. have attempted to beneficiate spent refractory material for steel mills but have encountered problems with dust suppression, causing most to discontinue processing the spent refractory material for reusehecycling. Where the basic materials are crushed for reuse as a slag conditioner, care must be taken to process the spent refractory into particle sizes that will dissolve in the slag during the heat (below 2.5 cm in size), yet will not be so fine as to be exhausted into the dust collector, altering the chemistry and lowering the market value of recoverable Zn in the EAF dust.

-

Regardless of how refractory recycling/reuse is approached, the commitment and backing of management is essential to develop and implement a successful


11

program. Management should appoint one individual to head the effort, providing him with strong support. External utilization or processing by companies specializing in recycling spent refractory material should also be considered, an option that would allow the refractory user to concentrate on their core business. The economics of any recycling program should be compared with purchased materials before commitments are made. The last option for handling spent refractory material should be refractory waste treatment andor disposal in a landfill. CONCLUSIONS Spent refractories can be reusedhecycled as a raw material source in a number of products or processes ranging from low to high volume applications and with varying material values. Material cleanliness, beneficiation costs, and material consistency will play a large role in determining these applications. Successful recycling efforts are utilized in a number of industries, including glass, steel, aluminum, and brass. Applications for the spent refractories have typically been internal rather than external. At a minimum, processing of the spent refractory material typically involves segregation into the different types of refractories, than crushing and magnetic separation. The most common applications for refractory reusehecycling are as a refractory raw material or as a slag conditioner. Recyclers are often a viable option for refractory reusehecycling. Barriers to material reuse are oftentimes based more on an unwillingness to change prior practices and on a lack of driving forces than on barriers caused by technical issues. Refractories removed from service are typically landfilled. A number of driving forces to encourage recycling have recently emerged. These include growing concerns over legislation, the environment, and future liability. REFERENCES 1. Takahashi, H., M. Tsuno, and M. Hayaishi, “Recycling of Used Refractories in an Electric Steelmaking Shop,” Journal of the Tech. Assoc. of Refractories, Japan, 20(4), 2000, pp. 249-253. 2. Oxnard, R.T., “Refractory Recycling,” ACerS, 73( 10), Oct. 1994, pp 46-49. 3. U.S. Code of Federal Regulations, Title 40-Protection of the Environment, Part 26 1--Identification and Listing of Hazardous Waste, July 1, 1999. 4. Noga, J., “Refractory Recycling Developments,’’ Ceram. Eng. Sci. Proc., 15(2), 1994, pp 73-77.

12


5. Webber, R.A., “Recycling at Corhart - A 30 Year Success Story,” Ceram. Eng. Sci. Pro., 16(1), 1995, pp 214-215. 6. Kwong, J., and J.P. Bennett, “Achieving MgO Saturated Foamy Slags in the EM,” Proceeding of the 5gthEAF Conference and lgth Process Technology Conf. Proc., Nov. 11-14, 2001, Phoenix AZ,published by the ISS, pp 277285. 7. Viklund-White, C., H.Johansson, and R. Ponkala, “Utilization of Spent Refractories as Slag Formers in Steelmaking,” Proc. of the 6th Int’l Conf. on Molten Slags, Fluzes, and Salts, ed. by S. Seetharaman and D. Sichen, meeting held in Stockholm, Sweden and Helsinki, Finland, June 12-17, 2000, pub. on CD by Division of Metallurgy, KTH, Sweden, paper is 13 pages. 8. Holmes, L., N.S. Schubert, A. Mooney, J. Bennett, and KS. Kwong, “Recycling of Spent Refractory Material from Carbon Baking Fwnaces,” Proc. of the UNITECR 5thBiennial Worldwide Congress held in New Orleans, LA, USA, Nov. 4-7, 1997, pp 477-486.

9. Smith, J.D. and K.D. Peaslee, “Spent Refractory Waste Recycling from NonFerrous Metals Manufacturers in Missouri,” Proc. of the 4th Int’l Symposium on Recycling of Metals an Engineered Materials, edited by D.L Stewart, J.C Daley, and R.L. Stephens, meeting held in Pittsburgh, PA on Oct. 22-25, 2000, published by TMS, pp 1385-1394. 10. Kwong, K.S., J.P. Bennett, K.W. Collins, and A.E. Wynne 111, “The Recycling of a 70 % A1203 Spent Refractory,” Proc. of the UNITECR 5th Biennial Worldwide Congress held in New Orleans, LA, USA, Nov. 4-7, 1997, pp 487-496. 11. Abrino, D.E., “Waste Minimization in Industries Using Refractory Materials,” Proc. of the UNITECR 5th Biennial Worldwide Congress held in New Orleans, LA, USA, Nov. 4-7, 1997, pp.465-471. 12. Goodson, K.M., N. Donagly, and R.O. Russell, “Furnace Refractory Maintenance and Slag Splashing,” Iron and Steelmaker, ISS, 22(6) 1995, pp 3 1-34. 13. Alasarela, E. and W. Eitel, “How Infinite is Endless Linings of Ladles,” Paper in Proc. of the UNITECR, San Paulo, Brazil, Oct 3 1-Nov. 3, 1993, pp 12671278.


13


CERAMIC-BASED MAGNETIC EXTRACTANTS FOR REMOVAL OF ORGANICS FROM WATER Allen Apblett, Solaiman M. Al-Fadul, and Tarek M. Trad Department of Chemistry Oklahoma State University Stillwater, OK 74078-3071 ABSTRACT Magnetic filtration can provide rapid, efficient removal of magnetic materials from a waste stream. However, since most pollutants are non-magnetic, it is necessary to use magnetically-active “extractants” to bind to pollutants and allow their separation by a magnetic filter. Excellent candidates for extractants are particles of magnetic ceramic oxides such as ferrites and magnetite whose surfaces have been derivitized to provide binding sites for toxic metal ions or organic pollutants. Such materials can be used to separate dyes or petrochemicals from water and break oil in water emulsions. INTRODUCTION The application of efficient magnetic filtration to decontamination and waste treatment operations is attractive because it can provide very rapid separation of pollutants from aqueous waste streams. This coupled with the ability to switch the filter on and off electronically (avoiding any need for mechanical contact) allows the minimization of exposure of workers to harmful agents. However, since many environmental contaminants are not magnetic, magnetic filtering aids must be developed that bind the materials and allow their magnetic separation. This problem has been addressed in coal beneficiation by use of magnetic fluids that are composed of magnetic particles, a suspending agent, and a carrier solvent that selectively wets the contaminant particles (oxide minerals in the case of coal) [ 11. Several approaches have been previously developed for separation of oil in water mixtures. The simplest method was to mix an extremely large excess of magnetite (ratio of 40 Fe304 1 oil (fatty acid) by weight) so that the oil absorbed onto the surface of the magnetic powder [2]. Subsequent magnetic filtration reduced the oil from 500 ppm to 2 ppm. The success of the absorption method can be attributed to the oil fEst being emulsified in an ionic form-an approach that is not applicable To the extent authorized under the laws of the United States of America, all copyright interests in this publication are the property of The American Ceramic Society, Any duplication, reproduction, or republication of this publication or any part thereof, without the express written consent of The American Ceramic Society or fee paid to the Copyright Clearance Center, is prohibited.


15

to normal hydrocarbons. When the same approach is applied to a 3000 pprn Bheavy oil in water emulsion, the final concentration of oil was only reduced to 57 ppm [3]. Another method that utilized a magnetic suspension and an acidic coagulant merely achieved a final concentration of 103 ppm [4]. There is also a successful application of the magnetic-filler-in-polymer technology that has been used for oil slick removal. In this case the polymer was polystyrene and the filler was iron oxide. The extractant was sprayed onto an oil slick by a watercraft travelling through the spill [51. The polymer/oil sludge was then collected on rotating magnetic disks. Using this approach, a 99% recovery of spilt oil was achieved with 20 volumes of oil being collected per unit volume of polymer. The objective of the research reported herein is the development of single component systems for use as magnetic filtration aids i.e. magnetic materials that can absorb pollutants and allow their separation from water via magnetic filtration (Figure 1).

Figure 1. Cartoon of Magnetic Extraction Separation of oil/water emulsions might also be feasible using magnetic extraction. The outer shells of the extractants will have a strong affinity for both the oil and the hydrophobic tails of any surfactants that might be present. Therefore, it should be possible to have the surfactant/oil micelles bind strongly to the extractant and be influenced to separate from the water via a magnetic field. This potential alternative to other methods of breaking emulsions such as

16


coalescing filters could provide considerabletimeand money savings to the petroleum industry. EXPERIMENTAL Magnetite was purchased from Strem Chemicals while Sibrid polyimide/siloxane co-polymer was purchased from Geleste. All other reagents were ACS grade and were purchased from Aldrich. All of these chemicals were used as received. Water was purified by reverse osmosis and deionization. Toluene and xylene were HPLC grade and were used without further purification. X-ray powder diffraction (XRD) patterns were recorded on a Bruker AXS D-8 Advance X-ray powder diffractometer using copper K- radiation. Surface areas were measured by nitrogen adsorption isotherms on a Quantachrome Nova 10 instrument. Magnetic filtration was performed using glass pipets packed loosely with steel wool (#OO fine grade) and taped to the side of an electromagnet. The latter magnet was a 24 V extended-reach-electromagnet with a 170 pound pull and dimensions 3"xl"x1.4". The electromagnet was powered by a 30watt direct current power supply. Preparation of PolydimethylsiloxaneMagnetiteComposites Composites of polydimethylsiloxane (PDMS) and magnetite were prepared by cross-linking a PDMS polymer at moderate temperature with the magnetic substrate. PDMS (15.0 g) was stirred with 20.0 g of magnetite to yield a paste that was heated to 280°C for one hour. Afterwards, the material was washed extensively with toluene in a Soxhlet rextractor to remove any unattached PDMS and was then dried in a vacuum oven at room temperature. Two different PDMS oils were used as starting materials: a low viscosity (10 centistokes) material and a moderate viscosity material (1000 centistokes). These yielded strikingly different materials, a homogenous thinly-coated powder in the first case and a rubbery composite in the latter. Both materials are hydrophobic and float on water. Preparation of Octadecylsilsesquioxane-CoatedMagnetite Magnetite powder (30 g) was treated with a solution of octadecyltrimethoxysilane (1.O g) in toluene (25 g). After 12 hours, the derivitized powders were isolated by filtration, washed with toluene, and were dried in vacuo. Preparation of 3-(Ethylenediaminepropyl)silsesquioxane-Coated Magnetite Magnetite powder (10 g) was saturated with water by placing in an enclosed jar containing water for 12 hours. It was then placed in a hybridization tube along with a solution of trimethoxysilylpropylethylenediamine (1.O g) in toluene (80 ml), The mixture was then heated to 90°C while rotating ion a rotisserie in a hybridization oven. After 24 hours, the derivitized powders were isolated by filtration, washed with toluene, and were dried in vacuo.


17

Preparation of SibricUMagnetiteComposite Sibrid (20.0 g of a15 wt% solution in N-methylpyrrolidine) was mixed with 30.0 g of magnetite to yield a sticky paste. The mixture was then placed in an oven for 24 hours at 100°C. Preparation of Magnetic Activated Carbons Magnetic activated carbons were prepared by modifling a literature procedure for activated carbon manufacture. Sawdust (200 g) was saturated with a 10% aqueous solution of sulfiuric acid (50 g) containing iron(I1) sulfate heptahydrate(5 g). The mixture was dried in air to air dry and was then fired to 500°C under a nitrogen atmosphere. ARer cooling to room temperature, the resulting powders were exposed to air at which point they became quite hot due to rapid oxidation of iron(I1) oxide to magnetite. The final result was a magnetite-containing activated carbon. A similar procedure using a nickel sulfate hexahydrate (5 g)/ Fe(S04)*6H20 (10 g) mixture in 10% sulfuric acid (50 ml) yielded a nickelhron impregnated activated carbon. Testing of Magnetic Extractants for Dye Removal from Aqueous Solution Solutions of Congo Red and Bromothymol Blue were prepared with a concentration of 30 ppm. The pH of the latter dye solution was adjusted to 4.0 so that the dye was in its deprotonated, yellow form. Next, 7.0 g of dye solution was mixed with 0.20 g of magnetic extractant and the mixture was shaken for two minutes. The resulting solution was subjected to magnetic filtration and was then analyzed for dye content by UVNisible spectroscopy. The wavelength for the absorption measurements was performed at the maximum visible light absorption for each dye, 497 and 420 nm, respectively. Dye concentrations were calculated using a calibration curve constructed from serial dilutions of the dyes over the range of interest. Testing of Magnetic Extractants for Breaking of an Emulsion A stable emulsion was prepared by diluting a 35:20:45 weight percent paraffin oil/triethanolamine/oleic acid mixture to 1000 ppm in water [6]. This yielded an indefinitely stable white emulsion. A second emulsion was prepared by sonicating 1.0 g of mineral oil and 0.1 f of Brij-35 in 1 L of water. Ten grams of each emulsion was treated with 1.0 g of magnetic extractant by briefly shaking the two materials together in a glass vial for one minute. The mixture was then passed through a magnetic filter. The extent of emulsion removal was then assessed by measuring the solution turbidity using a nephelometer. RESULTS AND DISCUSSION Preparation of Magnetic Extractants A facile method for making magnetically-active activated carbons was developed. These materials were simply prepared by impregnating sawdust with aqueous solutions of iron sulfate or irodnickel sulfate mixtures and then

18


subjecting the treated sawdust to a procedure for preparation of activated carbon [7]. X-ray powder diffraction showed that the resulting materials contained poorly crystalline iron or nickel salts along with traces of calcium sulfate derived from calcium ions naturally present in wood. In the iron-containing activated carbon, the iron-containing phases were amorphous and could not be detected by XRD. The nickelkon derivative displayed broad reflections for y-Fe203 (maghemite) and NiFeS2 (petalite). However, the X-ray diffraction intensity for these phases was weak indicating that the bulk of the metal oxides were dispersed as amorphous small particles. This is beneficial since it prevents the particles from becoming permanently magnetized. Magnetic testing of the powders with a strong electromagnet indicated that the activated carbons were strongly ferromagnetic and no non-magnetic particles were present. Nine of the powders demonstrated any remnant magnetization when the power to the magnet was switched ofc an important property so that the powder will not stick to non-magnetized steel. A recent report indicated that polydimethylsiloxane(PDMS) is a good absorbent for phenanthrene [S] prompted the preparation of composites of PDMS with magnetite. These were synthesized by cross-linking a PDMS polymer at moderate temperature in a paste with the magnetic substrate. This is similar to a procedure reported by Soares et al. for coating alumina, calcium carbonate, and hematite with PDMS [9]. Two different PDMS oils were used as starting materials: a low viscosity (10 centistokes) material and a moderate viscosity material (1000 centistokes). These yielded powders with substantially different thicknesses of PDMS coatings so that the PDMS-1000 product was a rubbery composite while the PDMS-10 product was a loose powder. A different polymer-coated powder was prepared by casting a commercial silioxanehmide co-polymer material from an n-methylpyrollidine solution onto a magnetite powder. The composite thus prepared is stable in water and in nonpolar organic solvents so that they can be used for magnetic filtration of aqueous solutions and then be cleaned for reuse by washing with an organic solvent. The reaction of hydrolyzable organosilicon alkoxides provides another facile method for derivitizing surfaces of metal oxides. In this investigation octadecyl and ethylendiaminegroups were covalently anchored to the surface by treating magnetite powders with octadecyltrimethoxysilaneand N-(trimethoxysilylpropyl)ethylenediamine,respectively. These reagent condensewith surface hydroxyls on the iron oxide surface, leading to a monolayer of pendant octadecyl or ethylendiamine groups grafted to the metal surface via a cross-linked silica layer. Thus, the particle surface becomes coated with a monolayer of polymerized silsequioxanes, (RSiO1.5)~. Testing of Magnetic Extractants The testing of the extractants was performed using 30 ppm aqueous solutions of two dyes, one that was anionic, Congo Red, and one that was neutral,


19

Bromothymol Blue in its yellow, sulfone form (see Figure 2). For the latter dye, the pH of the solution was adjusted to 4.0 to ensure it remained in the sulfone form. Each magnetic extractant was assessed for its ability to separate the dye from water via magnetic extraction. The performance of the extractants varied widely depending on the nature of the binding groups attached to the magnetic core (Table I). The anionic dye, Congo red, was efficiently adsorbed by both the untreated magnetite and the ethylenediamine-derivitizedmagnetite. The surface of magnetite is naturally-positively charged while the amine groups can be protonated by water to generate positive charges and this may account for their enhanced adsorption of Congo red. Non-polar coatings on magnetite led to very poor adsorption of Congo red but, with the exception of the octadecylsilsequioxane derivative, they were moderately successful at removing the neutral dye, Bromothymol blue, from water. These results are encouraging since they suggest that selective extractants for different classes of aqueous contaminants might be developed. Magnetite was a much poorer adsorbant for the Bromothymol blue sulfone then for Congo red as might be expected for the interaction of a positively-charged surface with a neutral and a negatively-charged molecule, respectively. However, the ethylenediamine derivative did a very good job with the neutral dye, possibly due to strong hydrogen bonding between the amine and the sulfone.

S03Na

S03Na

Congo Red

Br

CH(CHd2

Bromothymol Blue (Sulfone Form)

Figure 2. Structures of Dyes Used in this Investigation. The magnetic activated carbons with their relatively high surface areas might have been expected to significantly outperform the other extractants but this was only true for the adsorption of Congo red by the nickel iron derivative. While the activated carbons were useful adsorbants for both dyes, they performed best with the anionic dye suggesting that part of their surfaces is occupied by metal particles. This hypothesis is further supported by the enhanced uptake of dyes by the nickel-containing derivative as compared to the one that contained only iron.

20


TABLE I. Results from Treatment of 7.0 g of Aqueous Solutions of 30 ppm Dye Solutions with 0.20 g of Magnetic Extractants. Concentrations in ppm. Extractant Surface Area (m2/g) [C. Red] [B. Blue] Magnetite Octadecylsilsesquioxane PDMS-10 PDMS-1000 Sibrid 3-(Ethy1enediamine)propy1 Activated Carboaagnetite Activated Carbon/ Nickel Iron

6.5 2.0 4.2 2.1 2.8 6.3 273 265

4.8 25.8 25.8 29.4 25.8 6.3 2.9 0.11

22.3 26.6 9.7 21.0 16.7 6.1 9.7 2.5

Several of the extractants were tested for their ability to break emulsions using magnetic filtration. The idea for this process is to have the extractant bear functional groups that can segregate at the oil water interface so that a magnetic field can sweep the oil particle out of an aqueous mixture. Two types of emulsions were tested, one prepared with a neutral surfactant (Brij 35) and the other an anionic surfactant (oleate). The extractants were briefly mixed with these emulsions and then the mixtures were separated by magnetic filtration. The effectiveness of treatment was determined by use of a nephelometer and the results are displayed in Table 11. Surprisingly, magnetite did a very good job of removing the organics from both types of emulsions and performed best with the neutral surfactant. The derivatized magnetites performed less effectively and this may be a reflection of their tendency to be poorly wetted by water and to float on the surface while magnetite disperses well throughout the solution. The results show that magnetic extractants are capable of breaking emulsions and suggest that optimization of the magnetic extractant could result in complete breaking of an oil in water emulsion. The results also indicate that the best derivative would contain a polar headgroup and a hydrophobic tail so that it can partition effectively at the oil-water interface. The relatively successful breaking of the Brij 35/oil aqueous emulsion using magnetite coated with an imiddimethylsiloxane copolymer (Sibrid) suggests that this approach may be effective since the copolymer consists of alternating blocks of hydrophobic and hydrophillic groups. CONCLUSIONS Magnetic extractants based on magnetite or ferromagnetic activated carbons show promise for separation of organic species from water. Variation of the adsorbant material placed on magnetite particles can be used to change the selectivity of the magnetic extractant. Successful separation of oil from aqueous emulsions is also possible via magnetic filtration using magnetite as an extractant.


21

TABLE 11. Results from Treatment of 15 g of 100 ppm Emulsion with 1.O g of Magnetic Extractants. Results in Nephelometer Units (NTU). Extractant Brii emulsion Oleate emulsion 56.7 Initial emulsion 59.1 1.2 7.8 Magnetite 30.3 17.2 PDMS-1000 20.4 Sibrid 9.1 ACKNOWLEDGEMNT The Integrated Petroleum Environmental Consortium is gratefully acknowledged for supporting this research. The National Science Foundation, Division of Materials Research, is thanked for Award Number 987 1259 that provided funds for the X-ray powder diffractometer used in this investigation. REFERENCES [I] T. A. Sladek, "Coal Beneficiation with Magnetic Fluids" in Industrial Applications of Magnetic Separation Y. A. Liu, Ed. (Institute of Electrical and Electronics Engineers, New York, 1979). [2] W. F. Lorenc, J. A. Hyde, "Oil Removal from Waste Waters" U.S. Patent, 3161511,1974. [3] E. Nagata, H. Iwamoto, M. Kobayashi, "Separation of Oil and Water" Japan Patent, 16111493, 1977. [4] G. S. Pantelyat, V. G. Sleptsov, 3, 18-19 (1998). "Treatment of Wastewaters Containing Lubricants and Detergents by Magnetic Filtration" Vodosnabzh. Sanit. Tekh. 3, 18-19 (1998). [5] B. A. Bolto, D. R. Dixon, R. J. Eldridge, E. A. Swinton, D. E. Weis, Willis, D., H. A. J. Battaerd, P. H. Young, "The Use of Magnetic Polymers in Water Treatment" J. Polymer Sci. Symp. 49,215-225 (1975). [6] S. H. Shin, D. S. Kim, 35, 3040-47 (2001). "Studies on the Interfacial Characterization of O N Emulsion for the Optimization of Its Treatment" Environ. Sci. Technol. 35,3040-3047 (200 1). [7] A. Yehaskel, Activated Carbon Manufacture and Regeneration (Noyes Data Corporation, New Jersey, 1978). [8] J. Poerschmann, T. Gorecki, F.-D. Kopinke, "Sorption of Very Hydrophobic Organic Compounds onto Poly(dimethylsi1oxane) and Dissolved Humic Organic Matter" Environ. Sci. Technol. 34,3824-3830 (2000). [9] R. F. Soares, C. A. P. Leite, W. J. Botter, F. Galembeck, "Inorganic Particle Coating with Poly(dimethy1siloxane)" J. Appl. Polym. Sci. 60,200 1-2006 (1996).

22


INVESTIGATION ON A RECYCLING PROCESS OF WASTE COLORED GLASS Danping Chen and Hirotsugu Masui Conversion and Control by Advanced Chemistry, PRESTO, JST, AIST Kansai, Special Division of Green Life Technology Ecoglass Research Group, 1-8-31 Midorigaoka, Ikeda, Osaka 563-8577, JAPAN

*Tomoko Akai Conversion and Control by Advanced Chemistry, PRESTO, JST and National Institute of Advanced Industrial Science and Technology, AIST Kansai, Special Division of Green Life Technology Ecoglass Research Group, 1-8-31 Midorigaoka, Ikeda, Osaka 563-8577, JAPAN

* *Tetsuo Yazawa National Institute of Advanced Industrial Science and Technology, AIST Kansai, Special Division of Green Life Technology Ecoglass Research Group, 1-8-31 Midorigaoka, Ikeda, Osaka 563-8577, JAPAN ABSTRACT A recycling process for waste colored glasses through phase-separation is newly proposed. The colored soda-lime-silicate glass was leached by an acid solution after being re-melted with B203.It was found that all cations except Si4+ in the soda-lirne-borosilicate glass were leached by the acid. The colored glass was successfully bleached, and highly pure silica powders were obtained using this method. INTRODUCTION A large amount of colored glass waste is produced in high consumption nations. Currently, only a portion of the colored glasses waste is used as raw material to be *Corresponding author **Present address: Himeiji Institute of Technology,2167 Shosha, Himeji, Hyogo, 671-2201,Japan

To the extent authorized under the laws of the United States of America, all copyright interests in this publication are the property of The American Ceramic Society. Any duplication, reproduction, or republication of this publication or any part thereof, without the express written consent of The American Ceramic Society or fee paid to the Copyright Clearance Center, is prohibited.


23

re-melted. Most of the remainder is non-recyclable, and is directly put into landfills, because colored glasses is considered impossible to decolorize [1,2]. The practice of putting this glass into landfills has give rise to environmental, social and economic problems, all of which have been increasing in recent years in many municipalities and countries, especially in Japan, the EU, and Taiwan [l-31. Because of national regulations which reinforce environmental protection measures, there is a strong need to utilize waste glasses. There have been proposals to recycle colored glass waste, including using it as part of the coarse aggregate in cement and concrete [3-61, and the extraction of SiO2 from the glass waste by alkali fusion [7]. Due to the reaction between the alkali in glass and the reactive silica in cement, the use of glass as part of the coarse aggregate in concrete does not work well, because of the strength regression and excessive expansion [5-61. Using an alkali fusion, it is possible to separate Si02 and other contents of the glass waste, but it produces a lot of alkali waste. Therefore, it has been necessary to research new methods for re-utilizing colored glasses waste. We noticed that the transition metal ions always concentrate in the B203-rich phase of the phase separated soda-borosilicate glass [8], as shown in Figurel. Furthermore, as such manufacturing process the Vycor glass, the B203-rich phase can be leached out by the hot acid solution, and the remaining highly porous structure can be sintered together to form an almost pure SiO2 glass. Then, the phase-separating property of the 1 glass may be used to extract the alkali 2 ions and decolorize the colored glass we will propose a Figure 1. Schematic representation of In this recycling process for waste colored phase separation in soda- borosilicate glass glasses based On the Property Of glass l.SiOn-ri& phase; 2, B 2 0 3 - ~ c hphase phase separation.

EXPERIMENTAL The base glass used in the present experiment consisted of two types. One type was colored glass waste, the exact compositions of which were unknown. This glass may have contained the elements of Na, Ca, Al, Si, CryFe, Cu, COand Ti [7]. The composition of the other glass used was 15.2Na20*10.2Ca0*73.2SiO21.3~203*0.1c~03(wt%) (the blue base glass) and 15.2Na20*10.2Ca0*73.2Si02* 1.3&03* CrzO3 (wt%) (the green base glass). The base green or blue glass and the H3B03 were mixed (base glass: B203=100:x,x= 15- 60) in a mortar. In order to treat the colored glass waste and obtain the blocks of glass samples, the waste glass and Si02, H3B03, and Na2C03 as well as Al(OH)3 were mixed in this

24


proportion: glass waste:Si0~:B~0~:Na~0:Al~0~ = 100:50:150:6:6. Next, a 50g batch was melted in a platinum crucible at temperatures of 1400 "C for 4 hours, and the melt was then poured on to a graphite plate. Next, the glass obtained was annealed at 560-72OoC for various amounts of time, ranging from 10 to 80 h. The heat-treated and the non-heat-treated powder or blocks of glass were leached in 0.5- 3.ON HNO3for 24-48h. After the acid leaching, the solutions were filtered and the remaining glass was washed with deionized water. The content of Na in the leaching solutions was chemically analyzed by an atomic absorption spectrophotometer (AA-6800, Shimazu Co.). The contents of B, Ca, Cr, CO and AI in the leaching solutions were chemically analyzed by an inductively coupled plasma emission spectrometer (SPS7800,Seiko Instruments). The compositions of the leached glass were chemically analyzed based on JISR 3101 (the Japanese Industrial Standards for chemical analysis of soda-lime-magnesia -silica glass). The blocks of glass samples before and after acid leaching were polished, and the optical absorption spectra were measured in the 200 to 9OOnm wavelength range with a UV/VIS spectrophotometer(UV-240, Shimadzu Co.). To examine the structural changes with change of glass composition, we measured the "B NMR spectra. All "B NMR spectra were measured on a Chemagnetics CMX-200 spectrometer with v0=64.2504 MHz (B,=4.7 T), H3B03 (1M aq.) used as a frequency reference. The free induction decay was acquired using a single pulse of typically 1 ps with a repetition time of 3 s. The signal was typically accumulated 256 times. RESULTS AND DISCUSSION Table I lists the leaching rate of various elements from the blue base glass re-melted with different amounts of B203 (x). Because of the vaporization of some Na2O and B2O3 during the re-melting process, the leaching rates of Na and B were about 10% lower than their actual leaching rate. However, the leaching rates of Ca, CO and Al were a little higher than their actual leaching rates. The leaching rate of Si was about 1.0%, and did not change with an increased B2O3 content. This result shows that the leaching rate of all of the elements, except for Si, increased with an

B

CO

x

Na

Ca

15

48.6

44.7 47.4 21.9

25

88.6 101.7 82.6

35

88.0 104.7 86.6 105.1 31.5

66.1

Al 4.2 12.1

.

Table Leaching rate (%)of various elements of (green base glass:B,O, =lOO:x) glass after acid treatment x INa Ca B Cr Al

20

53.0

56.4 49.1

30

89.6 101.0 90.5

45

85.8

98.6

3.8

3.9

34.8 22.1

88.1 93.0 90.5


25

Tablem. Analytical composition (wt%) of the (blue base glass:B,O, =100:30) glass after heat and acid treatment Treatmentconditions Si02 Na2O CaO A 1 2 0 3 COO B203 Heat treated at 640°C for 65h 98,0 o.05 0.02 1.7 0.01 0.19 and acid treated in 1N HNO,

increasing B2O3 content. Although this glass was not heat-treated, the elements of Nay Ca, CO and B dissolved in the heat acid as the x=35. Above x=35, the blue glass powder was decolorized to be colorlessness. This implies that the micro-phase separation exists in the quenched transparent glass. However, although the leaching rate of Al increased with the increasing B2O3contentYthe maximum leaching rate achieved was only 31.5%; this indicates that the site of the Al element in the glass is different from the site of Nay Ca and Co. We also analyzed the composition of the blue base glass after it was re-melted with B2O3 and leached, and the result is listed in Table U.The Si02 purity of the glass after the acid treatment reached to 98wt%, and the coloring ions of the cobalt were entirely eliminated. However, as mentioned above, the majority of A l 2 0 3 remained, due to the low amount of B2O3 added to the glass. Table II lists the leaching rates of various elements of the green base glass re-melted with different amounts of B2O3 (x). The leaching behavior of Nay Ca and B was the same as that of blue base glass. However, the leaching rate of Cr was similar to the Al element. Most of the Cr and Al could not be leached out, and the glass had some green or yellow color in it up to x= 45%. Since the leaching experiments showed that the Na, Ca, CO and B in soda-lime-borosilicateglass without heat treatment easily dissolved in the hot acid solution, we deduced that there is a micro-phase separation in the quenched glass, and the cations of NayCa and COare enriched in the surrounding B atoms. However, as for the cations of Cr and Al, they were not dissolved in the hot acid solution until x= 45wt%. This may mainly relate to a change in glass structure with the B2O3 content, because the ratio of B03/B04 in the soda-lime -borosilicate glass structure changed with the 400 200 0 -200-400 ratio of ( Na2O+CaO)/ B203. PPm Figure 2 displays the "B NMR spectra of Figure 2. "B NMR spectra of the the green base glass after the addition of B2O3 green base glass after the addition glass. A narrow sharp signal near 0 ppm of different quantities of B203 corresponds to the tetrahedral boron, BO4, and (green base glass:B203=100:x)

26


two overlapping, broad, split signals between 5 and 20 ppm correspond to the trigonal boron, BO3. It can be seen that the peaks corresponding to the trigonal boron BO3 increase with the amount of B2O3. This suggests that the fraction of tetrahedral boron, BO4 in this glass decreaseswith the addition of B2O3. The phase separation of glass results from the selective bonding between various chemical bonds in the glass structure. In addition, the nature of the chemical bonds, such as the covalent bond or ionic bond, results in the selective bonding. It is known that the nature of the chemical bond in oxide is related to the field strength. The B3' cation in BO3 and B04 as well as Si4' cation in Si04, have different field strength [9]. The order of these field strengths is as follows: B3+(in BO+ Si4' (in Si04)>B3+(inB04) The COions may selectively connect with the chemical bond with a lower cation field strength of in the glasses, such as B-0in B04,Na-0 and Ca-0. Therefore, it may easily concentrate in the B203-rich phase, and easily dissolve in the hot acid solution with boron, as shown in Table I . However, the Cr ions in the glass behave differently. As shown in Table II, the Cr leaching rate of the green base glass re-melted with 30%&03 was 34.8%. When the glass was re-melted under a reducing atmosphere, the Cr leaching rate only decreased to 10%. This result reveals that the low valence of Cr ions is distributed in the SiO2-rich phase, and that the Cr ions dissolved in the hot acid solution may be the high valence of Cr. The Cr6' ion may be similar to the COion; it is distributed in the BzOs-rich phase. As for the A13' and C?' ions in the MO6(M=Al, Cr) octahedron, they may selectively connect with the chemical bond of Si-0, which has a higher cation

""

2 50

I

I

Y

c) Q)

2

M El

40

30

. r (

c 20 0

4

10

- 0 0

20

40

0

60

80

Time (hour) Figure 3. Leaching rate of aluminum in blue base glass with 30% B203 added after heat treatment at 650°C for different times.

L

0

20

40

60

80

T i m e (hour)

Figure4. Leaching rates of chromium aluminum in green base glass with 35% B203 added after heat treatment at 650 "Cfor different times.


27

field strength than that of B-0 in BO4. Then the ions of A13' and C? mainly concentrate in the SiO2-rich phase. However, the fraction of the three-coordinated BO3, with a field strength slightly higher than that of the Si4' cation, increase with the addition of B2O3; some A13+and Cr3' ions in MO6 (M=Al, Cr) octahedron can bond with the chemical bond of B-0 in three-coordinated BO3.ThenYin the case of high B203 content, A13' and C? concentrate in the interface between the SiO2-rich phase and the B203-rich phase as well as the B203-rich phase. These results suggest that the transition metal ions are not always enriched in the B203-rich phase of the phase separated soda-borosilicate glass [8]. The distribution of the ions having multiple valence state in the phase separated glass varied with the glass composition. The discussion above is based on the hypothesis that there is a micro-phase separation in the quenched glass. When the micro-phase separation is developed by heat-treatment, the distribution of various atoms in the phase-separated glass appears to be altered. Figures 3 and 4 show the effect of heat treatment times on the leaching rate of A13' and Cr3' ions. The leaching experiment showed that the change in the leaching rates of the CO, Na and Ca ions, probably concentrated in the B203 phase, were slight. However, with regards to the A13' and C? ions, the leaching rate was greatly changed with different heat treatment times. The Al leaching rate achieved its maximum value when the blue base glass with 30% B2O3 added was heat-treated at 650 "Cfor 40h. Figure 4 shows a similar result for the A13+and Cr3' ions in the green base glass. When the green base glass re-melted with 35% B2O3 was heat-treated at 650°C for 20h, the Al and Cr leaching rates showed their maximum value. The size of the separated phase changed with the temperature and time of the heat treatment, and may influence the Al and Cr leaching rates, especially when the ions of A13' and Cr3' concentrated in the interface between the SiO2-rich phase and BzOs-rich phase. In addition, the glass structure and valence and site of the Cr may undergo some changes during heat treatment, which could also influence the leaching rate. Tomozawa suggested a change in the glass structure with heat treatment temperatures through an anal sis of the immiscibility controversy of borosilicate glass [lO].High- resolution 71B, 29Siand 27AlNMR revealed temperature dependent structural changes in borate, borosilicate and boroaluminate glass [ll].These changes involved the change in

Glass composition lOOWaste green bottlet45B203 *100Greenbase glasst45B203

28

Na 0.91 1.1 (86)

Ca 0.88 0.84 (104)

Cr 1.92 0.74 (98)

Fe 2.53

B 1.25 1.29 (84)

Al 8.3 7.2 (92)

Si 17.1 3.8 (0.9)


the boron coordination numbers and the transformation of AlO6 or AlOs units into A104 units with the lowering of the cooling rate for transforming the liquids into glass. We also measured the NMR spectra of the glass before and after the heat treatment, and the results supported the above conclusion [12]. Since the transformation of AlO6 or AlOs units into A04 units requires a charge compensation [9], this transformation appears difficult in the SiO2-rich phase with low alkali and alkaline earth ions contents. The transformation of AlO6 or AlOs units into AlO4 units implies that the A13' ions leave the SiO2-rich phase and enter the interface. Consequently, the Al ions are easily leached out in the hot acid solution. The size of the separated phase may increases with the amount of time of the heat treatment when heat treatment is beyond 20h. The increasing of the size results in a decrease in the interface of phase separation. This is an important reason for changing the leaching rate when increasing the heat treatment time, as shown in Figure 4. As for the Cr ions, the Cr6+ ions may distribute in the B203-rich phase and the influence of the heat treatment upon the Cr6' ions is slight. However, the changes in the glass structure and the phase separation resulting from the heat treatment may also influence the Cr3' ions, which means that the relationship between the heat treatment time and the leaching rate for the A13+and Cr3' ions are very similar. We treated the actual green bottle glass waste based on the above method and experimental results. The green bottle glass waste was successfully bleached by the phase separation and acid treatment. The results are shown in Table IV. An 100 approximate composition of the green bottle glass can be known in comparison 8o with the leaching rate of the green base glass listed in Table IV. Therefore, the 55 green bottle waste was converted into , colorless, porous, almost pure Si02 glass. , The porous glass obtained after the phase separation and acid leaching treatments was subsequently sintered at llOO°C, and + 20 = became a compact non-porous glass 2.Qlass 2 after possessing properties similar to that of 200 300 400 500 600 700 800 900 silica glass. Figure 5 shows the optical Wavelength (nm) transmission curves of a block of colored glass waste with a thickness of >lmm (l), Figure S. transmission along With the glass after the Phase curves of waste mlored glass Separation and acid leaching treatments with before and after phase separation a lmm thickness (2). and acid leaching treatments.


29

SUMMARY In conclusion, we have demonstrated a new method for recycling colored glass waste through glass phase-separation. Colored soda-lime-silicate glass was re-melted with B2O3, and the soda-lime-borosilicateglass was then heat-treated for phase separating. The non-heat-treated glass and the heat-treated glass were leached with a HNO3 acid solution at 90°C. It was found that all cations except Si4’ in the soda-lime-borosilicate glass could be leached by acid. The powder or blocks of colored glass were successfully bleached, and a silicate glass with a high Si02 purity was obtained. REFERENCES 1. LYasui, “Several Aspects of Glass Recycle”; New Glass, Vol. 16 [2] 9-14 (2001) 2. D. Workman, “Recycling and How to Get the Message Across”. 19th Int. Conger. On Glass. Edinburgh, Scotland, V01.2~12,(2001) 3. N. Su and J.S. Chen, “Engineering Properties of Asphalt Concrete Made with Recycled Glass”, Resources, Conservation and Recycling, Vol. 000 000400 (2002) 4. Y. Shao, T. Lefort, S. Moras and D. Rodriguez, “Studies on Concrete Containing Ground Waste glass”, Cement und Concrete Research , Vol. 30, 91-100 (2000) 5. C.D. Johnson, “ Waste Glass as Coarse Aggregate for Concrete”, J Testing and Evaluation, Vol. 2 [5] 344-350 (1974). 6. K. Asaga, K. Kanai, H.Kuga, S. Hirose and M. Daimon, “Hydration of Portland Cement in the Addition of Waste Bottle Glass Powder”, Inorganic Materials, V01.4~423-430(1997) 7. H. Mori, “ Extraction of Silicon Dioxide from Waste Glasses by Alkali Fusion”, Proc. 19th Int. Conger. On Glass. Edinburgh, Scotland, V01.2, 13-14, (2001) 8. W.Vogel, “Phase Separation in Glass”, J.Non-Cryst. Solid, Vol. 25, 172-215 (1977) 9. H. Scholze, “Glass: Nature, Structure, and Properties”, Springer-Verlag (1990) pp.108-09 and pp.135-138 10.M. Tomozawa, “A Source of the Immiscibility Controversy of Borate and Borosilicate Glass System”,JAm. Ceram.Soc., Vo1.82 [11 206-208 (1999) ll.S.Sen, Z.Xu and J.F. Stebbins, “Temperature Dependent Structural Changes in Borate, Borosilicate and Boroaluminate Liquids: High-resolution llB, *’Si and 27AlNMR’, J.Non-Cryst. Solid, Vol. 226,29-40 (1998) 12. D. Chen, H. Masui, T.Akai and T. Yazawa, to be submitted to J.Non-Cryst. Solid for publication

30


USE OF MID-DELAWARE RIVER DREDGE SEDIMENT AS A RAW MATERIAL IN CERAMIC PROCESSING Kimberly Hill, R. A. Haber, Rutgers University ABSTRACT Traditionally, the millions of tons of sediment dredged from New Jersey's rivers and channels were disposed of in the ocean or at land-based sites. Due to the recent environmental and legislative regulations, ocean disposal is no longer favorable. In addition, capacities of land-based disposal sites are quickly receding. Hence, non-traditional applications and disposal methods must be found for the dredge sediment. This was the first work in which Delaware River dredge sediments were evaluated for ceramic applications. The results show how thesse materials could potentially be introduced into c e d c industries as a new low-cost raw material source. The objective of this work was to characterize dredge material from the mid-Delaware River and to formulate a commercially sound ceramic tile product containing the maximum possible amount of dredge material. A matrix of formulations was prepared with dredge material used alone and also with a low cost local New Jersey clay, other recycled materials and commercially beneficiated materials. The dredge material was a coarse, highly quartz material with little organic content or soluble anions. Characterization results showed that the dredge material needed to be combined with clay and an auxiliary flux to reach the desired water absorption and breaking strength for floor and wall tile applications. I S 0 standards categorize tile by the forming method and water absorption value. Strength for commercial-sized tile must also meet IS0 specifications. The formulation concluded to be most applicable for floor tile consisted of 40% dredge materials, 50% New Jersey clay, 10% limestone. Numerous compositions showed potential for use as wall tile.

BACKGROUND ON DREDGING

Dredging is the practice of excavating material from the bottom of a waterway-rivers, bays, channels and ports-to allow for the safe passage of large vessels. Dredging is necessary to the U. S. economy. One container ship holds the cargo equivalent of 6,000 tractor-trailers or 1500 rail cars.' Thus, ships provide an efficient means of transporting goods. Many waterways through To the extent authorized under the laws of the United States of America, all copyright interests in this publication are the property of The American Ceramic Society. Any duplication, reproduction, or republication of this publication or any part thereof, without the express written consent of The American Ceramic Society or fee paid to the Copyright Clearance Center, is prohibited.


31

which ships travel are naturally only 20 feet deep. However, today’s megaships ride 40-50 feet below the water’s surface. Material must be removed for safe navigation. Traditionally, when material was removed from a waterway, it was placed in ocean disposal sites or in one of two land-based disposal sites. Due to recent legislative and environmental restrictions, disposal in ocean-based sites has become unfavorable. One land-based option was landfills; however, landfills have only limited space for dredge sediment. In addition, tip fees make this type of disposal quite costly. The second option was placing material in confined disposal facilities. These are land sites specifically dedicated to dredge disposal. In New Jersey alone, there are 76 confined disposal facilities (CDFS)? Many of these sites have reached maximum capacity and are no longer operational. Of those that remain open, many are quickly approaching maximum capacity. With a growing number of dredging projects, limited disposal space and increasing environmental regulations for safe material disposal, non-traditional uses and placement of dredge sediment are no longer novel concepts, but are now necessary actions. Non-traditional uses include construction materials, landfill capping, replacement fill and shore protection, to name a few? The objective of this work was to characterize mid-Delaware River dredge sediments with respect to inorganic chemistry, mineralogy, particle characteristics and fired properties. Results of materials characterization were then used to evaluate the applicability of the dredge material for ceramic processes. The goal was to maximize the quantity of dredge used in ceramic bodies, but with minimal beneficiation. Since this was unprecedented work, experimental compositions were designed to explore a wide range of options. The dredge material was used alone, as a blend with other locally available raw materials and as a blend with commercially available raw materials. A secondary objective of this work was to incorporate a locally available New Jersey clay. Experimental compositions for ceramic floor and wall tile were chosen based on properties comparisons with commercially available products. For all compositions, water absorption and breaking stren@h were measured. MID-DELAWARE RIVER DREDGE SEDIMENT By wet sieve and x-ray diffraction analyses, it was found that the midDelaware River dredge sediment was coarse-grained quartzose material. A minor presence of clay minerals was detected in the -200 mesh (74 pm)fiaction of the material, which contributed only 20% of the particle size distribution. The dredge also contained negligible soluble anions, measured by ion chromatography, and less than 1% total organic content, as measured by hydrogen peroxide digestion. The vitrification point of the material was found to be greater than 1500°C. This was determined by firing one-inch diameter dry pressed discs of

32


the material at increasing temperatures fiom 1250°C to 1500"C, exhausting the limits of the kiln in use and the limit of temperatures of interest for using this material in manufacturing. The color of the resulting body was red at lower temperatures and brown at higher temperatures. With such a high vitrification point, it was concluded that employing this material alone in a ceramic body was not feasible. LOCAL, NEW JERSEY CLAY In the mid-l980s, during construction at the Burlington County Solid Waste Facilities Complex in Burlington County, NJ, more than four million tons of clay material were excavated and stockpiled. No commercial uses have been found for large tonnages of this clay; thus, the stockpiles remain onsite at the landfill. Since the New Jersey clay is located in the same county as the midDelaware River dredge material, it would be cost-effective to use these raw materials together in a ceramic body. Estimates from the county place the price of this clay to be $3-8 per ton. X-ray diffraction showed that quartz is the major phase in the local New Jersey clay. Minor phases include illite, smectite, mica and pyrite. Coarse glauconite particles ( > 1 5 0 p ) also appear as a minor phase. The clay contains 10,000 ppm to 16,000 ppm sulfur, depending on the size fraction of the material being analyzed, as determined by Leco carbon and sulfur analysis. Much of the s u l h content appears in mineral form, as pyrite (FeS2). Soluble sulphates measure 30-43 ppdg, according to ion chromatography tests. Due to the pyrite, the clay also contains appreciable amounts of Fe2O3. The decomposition of pyrite causes the material to bloat between 1150°C and 1200°C. With increasing temperatures, the clay fires to orangish red or to brown. The results from materials characterization aided logical development of formulations for ceramic tile, brick and lightweight aggregate applications. Given that the river dredge was a silicious material, it was necessary to combine it with other raw materials to produce a commercially viable product. The New Jersey clay served as one clay component to add to the raw materials matrix. The clay minerals include illite and montmorillonite. The appreciable s u l h and iron contents in the New Jersey clay must be considered when determining firing parameters. Due to the vesicular nature of the New Jersey clay, a white commercial ball clay was used as a replacement in selected compositions. In this case H.C.Spinks', C&C ball clay was chosen To further tailor the frring properties, an auxiliary flux was added to the matrix of raw materials. Both a local New Jersey limestone and Feldspar Corporation's F-4 feldspar were used.


33

CERAMIC TILE Following characterization, the second phase of this project was to formulate compositions including the maximum possible amount of dredge. To evaluate the applicability of the experimental compositions for tile, water absorption was measured for bodies fired to 1150°C and 1200°C. These values were then compared to IS0 standards4for commercial ceramic tile to discriminate between bodies applicable for floor tile or wall tile and bodies not applicable for either. Breaking strength was measured by biaxial flexure. Tile is categorized by the forming method, extruded or pressed, and the water absorption values. Water absorption (E) categories are as follows: E$I 5 million. The delisting portion of the cost savings for the vitrified wasteform, and subsequent disposal as a low-level radioactive waste rather than a mixed waste, saved $1 to 1.5 million (included the petition preparation costs).

CONCLUSIONS The EPA, Region 4 granted the delisting of the vitrified M-Area plating line wastes, and other SRS wastes that were included in the treatment process, on August 2 1,2002. This is the first delisting petition approval for a vitrified mixed waste in the US. During the 45-day comment period allowed in the proposed delisting exclusion, the EPA received neither comments nor requests for public hearings. The authors strongly recommend that a petitioner for mixed waste prepare an “Upfiont Delisting Petition”, if at all feasible. Such a petition needs to include the delisting limits calculated by the DRAS technique, rather than the older EPACML method, as the delisting limits may be stricter for certain constituents. The petition should also clearly determine whether the regulators would use the UTS limits, since these are the strictest limits in some cases. It is imperative to know what delisting limits will apply when designing a waste treatment process and to maximize waste loading in the final wasteform.


93

AUTHOR’S NOTE

This proposed approval by the EPA had been issued when this paper was

submitted for publication at the 2002 St. Louis meeting. Since the final approval was granted in August, prior to publication in the fmal Proceedings, the paper has been updated to reflect the final approval. ACKNOWLEDGMENT The proposed approval of the delisting petition was prepared by Dr. Judy Sophianopoulos, of the US Environmental Protection Agency, Region 4, of the South RCRA Enforcement and Compliance Section, Atlanta, GA 30303. Dr. Sophianopoulos conducted the risk modeling analysis using the DRAS software, and included the modeling results in the delisting approval. Dr. Sophianopoulos 4 prepared the notification of the proposed approval in the Federal Register 5 (March 15,2002), and also the final approval (August 21,2002). Although the delisting petition approval by Dr. Sophianopoulos was based on the information 6 provided in WSRC’s Delisting Petition , the petition would not have been successful Without Dr. Sophianopoulos’ expertise and diligence in preparing the approval. REFERENCES

1. J. B. PICKETT, S . W.NORFORD, J. C. MSJSALL, and D. G. BILLS, “Vitrification and Privatization Success,))Proceedings of the Environmental Issues and Waste Management Technologiesin the Ceramic and Nuclear Industries Vx St. Louis, MO, Mi, 2001, Ceramic Transactions, Vol. 119, page number 219. The American Ceramic Society, Westerville, OH 43018 (2001). 2. EPA (Delisting Section, Office of Solid Waste, U. S . Environmental Protection Agency), Petitions to Delist Hazardous Wastes,A Guidance Manual, Second Edition, EPM530-R-93-007 (March 1993). 3. J. B. PICKETT, “Upfint Delisting Petition for Vitrijied M-Area Plating Line Wastes”, WSRC-TR-96-0244, Westinghouse Savannah River Co., Aiken, SC 29808 (1996). 4. EPA (v. S . Environmental Protection Agency), Hmardous Waste Management System; Identification and Listing of Hazardous Waste: Proposed Exclusion; Proposed Rule, 67 Federal Register 11639, (March 15,2002). 5. EPA (U. S . Environmental Protection Agency), Hazardous Waste Management System; Identijication and Listing of Hazardous Waste: Proposed Exclusion; Final Rule, 67 Federal Register 54124, (August 21,2002). 6. J. B. PICKETT, “Delisting Petition for Vitrified M-Area Plating Line Wastes”, WSRC-TR-96-0244, Rev. 2, Westinghouse Savannah River Co., Aiken, SC 29808 (2000).

94


CHARACTERIZATION OF DEFENSE NUCLEAR WASTE USING HAZARDOUS WASTE GUIDANCE. INSIGHTS ON THE PROCESS AT HANFORD Megan Lerchen Pacific Northwest National Laboratory 902 Battelle Boulevard P.O. Box 999 Richland, WA 99352 William Hamel U. S. Department of Energy, Office of River Protection 2440 Stevens Center PO BOX450, MSIN H6-60 Richland, WA 99352

Lori Huffman U. S. Department of Energy, Ofice of River Protection 2440 Stevens Center PO Box 450, MSIN H6-60 Richland, WA 99352 Karyn Wiemers DMJMH+N 3250 Port of Benton Blvd Richland, WA 99352

ABSTRACT Federal hazardous waste regulations were developed for management of industrial waste. These same regulations are now applicable for much of the nation’s defense nuclear wastes. At the U.S. Department of Energy’s Hanford Site in southeast Washington State, one of the nation’s largest inventories of nuclear waste remains in storage in large underground tanks. The waste’s regulatory designation and its composition and form constrain acceptable treatment and disposal options. Obtaining detailed knowledge of the tank waste composition presents many challenges. Early insights fi-om a performance-based approach to demonstrating achievable quality standards will be discussed in the context of environmental guidance, permitting, and compliance under the hazardous waste regulations. INTRODUCTION The U.S. Department of Energy (DOE) is required to store, treat, and dispose of high-level waste at DOE’S Hanford Site in southeast Washington. Quality data supporting the project’s regulatory and engineering needs must be available. Over the last few years, DOE has made significant progress in defming and putting into effect characterization requirements for currently stored Hanford To the extent authorized under the laws of the United States of America, all copyright interests in this publication are the property of The American Ceramic Society. Any duplication, reproduction, or republication of this publication or any part thereof, without the express written consent of The American Ceramic Society or fee paid to the Copyright Clearance Center, is prohibited.


95

radioactive tank wastes to meet data needs for treatment and final disposition. This effort has relied on a cooperative, teaming approach between DOE, the regulators, and the implementing contractors in order to tailor the characterization effort to obtaining acceptable, usefid data. Hanford Tank Waste Hanford has 53 million gallons of high-level waste, containing 190 million curies of radioactivity, stored in 177 underground tanks. Accumulation of the waste began in 1944 with the inception of the H d o r d defense production mission as part of the Manhattan Project. Current operations consist of waste receipts from activities such as deactivation and decommissioning work, analytical and processing laboratories, ongoing tank waste management operations, and early efforts for tank closure demonstrations. The underground tanks are within ten miles of the Columbia River, the largest river in the Pacific Northwest. Many of the tanks are past their design life, and 67 of the older tanks are known or suspected to have leaked. In addition, the newer tanks are quickly nearing their capacity. The only permanent solution is to treat and immobilize the tank waste into an inert waste form.

Tank Waste Treatment It is planned that the dangerous waste and radioactive constituents in Hanford's high-level tank waste will be initially separated, through pretreatment if necessary, into lower and higher activity fractions followed by fmal treatment to make disposable waste forms. Pretreatment will partition constituents between low-activity and high-level fractions to meet disposal requirements and minimize product volume. All of the higher activity fiaction will be made into durable, disposable glass waste forms through vitrification at the future waste treatment plant. The low activity fraction will likewise be made into durable, disposable glass waste forms or, if applicable requirements are met, solidification by supplemental waste treatment. A significant challenge presented by tank waste is the overall uncertainty in the detailed characterization knowledge as opposed to the bulk waste constituents such as sodium, aluminosilicates, nitrate, and hydroxide. This adds to the difficulty in planning and designing for tank waste treatment facilities because in lieu of certain characterization knowledge, the project has been using bounding or other conservative estimates where needed. RCRA REGULATION OF TANK WASTE Because of past Hanford-specific practices and its location in Washington State, Resource Conservation and Recovery Act (RCRA) requirements and their applicability to Hanford waste differ in some important aspects from other DOE

96


sites. Under the Washington State RCRA program, the tank waste is designated for multiple RCRA waste codes. Each of these codes drives a requirement to use particular treatment technologies andor meet particular numeric performance standards in order to meet the RCRA treatment requirements for disposal. Tank waste characterization data will be used to support upcoming petitions to the regulators for a new treatment variance and delisting. In addition, characterization data are being used to support RCRA permitting for tank waste treatment and may be used to meet other tank waste management data needs related to regulatory compliance. DATA QUALITY OBJECTIVES FOR REGULATED COMPOUNDS In 1998, DOE and the Washington State Department of Ecology (Ecology) agreed upon the jointly prepared Data Quality Objectives (DQO) document referred to as the Regulatory DQO.’ The outcome of the Regulatory DQO was a prioritized list of 173 compounds for analysis that was selected fiom an initial list of nearly 1000 regulated compounds. The selection process involved a systematic review of each compound by a team of tank waste chemistry experts, including representatives from DOE and The evaluation focused on the plausibility of the regulated compounds’ existence in the tank waste matrix and a prioritization based on relative toxicity (Figure l).5 Target U. S. Environmental Protection Agency (EPA) methods (EPA publication SW-846, Test Methods for Evaluating Solid Waste, PhysicalKhemical Methods, EPA 600/4-79-020)or other equivdent environmental methods were identified for characterizing each of these prioritized compounds.6

Figure 1. Logic diagram for analyte selection and prioritization process used in the Regulatory DQO


97

This DQO is focused, in part, on land disposal restrictions and permitting for waste treatment. At the time the DQO was conducted, the permitting and design efforts for the waste treatment plant currently under construction were in their infancy. Therefore, the parties agreed that the DQO implementation would be conducted in a step-wise fashion. This was intended to allow for refining the characterization needs as data are collected and the waste treatment plant permitting and design also move forward. A timeline for Step 1 of the DQO and the baseline schedule for the Waste Treatment Plant are shown in Figure 2. The timeline shows completion of Step 1 in the near future, before commissioning of the waste treatment plant.

Figure 2. Data quality objectives implementation timeline. Note that subsequent to The American Ceramic Society presentation, the milestone for reporting was delayed to first quarter FY2003. RCRA CHARACTERIZATIONMETHODS VALIDATION Methods for nine groups of compounds were investigated for their applicability to the tank waste solid and liquid matrices. The nine groups were metals, anions and organic acids, mercury, ammonia, cyanide, volatiles, semivolatiles, polar volatiles, and polychlorinated biphenylsl pesticides. The status of these analyses as of April 2002 is shown in Table 1.

98


I I

Table 1. Data quality objectives implementation status - April 2002' Analysis

I I

Preparative Method

I I

Analysis Method

I I

status

Liquids: SW-846 3005A Solids: SW-846 3050B ASTM 4503 (modified)

Metals

Reporting in progress

SW-846 9056

I EPA350.32 I

I

I

Ammonia

NIA

Mercury

NIA

Liquids: SW-846 7470A Solids: SW-846 7471A


Cyanide

NIA

SW-846 9012A


Volatiles

Liquids: SW-846 5030B Solids: SW-846 5035

SW-846 8260B

Water and sand testing in progress

Semivolatiles Liquids: SW-846 351OC Solids: SW-846 3550B

SW-846 8270C


Polar VOlatileS SW-846 502 1

SW-846 8260B


SW-846 8081A (Pesticides) SW-846 8082 (PCBs)

Waste testing in progress

Pesticides/ PCBs

Liquids: SW-846 351OC Solids: SW-846 3550B

ASTM- American Society of Testing and Materials EPA-U.S. Environmental ProtectionAgency IC-ion chromatography ICPIAES- inductively coupled plasmdatomic emission spectrometry ICPMS- inductively coupled plasmdmass spectrometry NIA-not applicable PCBs-polychlorinated biphenyls


I

'All analytical methods fiom Test Methods for Evaluating Solid Waste, PhysicaVChemical Methods SW-846, unless otherwise noted. 'Methods for Chemical Analysis of Water and Wastes, EPA 600/4-79-020, March 1983.

The application of EPA's SW-846 guidance document, Test Methods for Evaluating Solid Waste, PhysicaVChernical Methods: to the unique radioactive waste matrix necessitated that a strategy be developed to support analytical method validation. This strategy was developed as part of the Regulatory DQO process. The initial step in this strategy has three main parts: 1. Determining method detection limits (MDLs) and estimated quantitation limits (EQLs) in water and sand to demonstrate the laboratories' ability to pefiorm the method with standard matrices; 2. Determining MDLs and EQLs in tank waste liquids and solids to demonstrate the ability to apply the methods and establish the MDLs and EQLs for liquid and solid radioactive waste matrices; and


99

3. Conducting a holding time and storage condition study to understand the effect of radioactive waste sample management on characterization data measurements. Parts 1 and 2 started with the preparation of an implementing plan* for the Regulatory DQO followed by preparation of detailed laboratory test plans for each method. The results of the analyses were compared with quality control parameters documented in the project’s Quality Assurance Plan. These performance parameters were generally consistent with the SW-846 methods, specific American Society of Testing and Materials (ASTM) methods, or alternative methods approved by Washington State (WAC 173-303-1lO).’ Detection limits were also compared with available (albeit often preliminary) regulatory thresholds and were determined to be generally adequate for decisionmaking purposes. Final results are expected to be published in 2002. The data evaluation for Parts 1 and 2 involved an interactive process between the performing lab, their contractor, DOE, and Ecology. A jointly agreed-to format and protocol for interim data transmittal provided faster access to decisioncritical data (Figure 3) and a fairly thorough verification of the data quality. Decision-makers real-time involvement at key hold points yielded ownership of the results. A strong commitment to the process and the schedule by the contractors, DOE, and Ecology facilitated exceptional communication, minimizing delays.

Figure 3. Data quality objectives data verification and stakeholder review process

100


The third part of the Regulatory DQO Step 1 is the performance of a holding time and storage condition study. This study will evaluate the effects of sample storage upon sample integrity, analyte degradation and loses, and analytical data measurements. While a number of statistical approaches have been evaluated for this study, hture implementation pathways have not been finalized. The DQO Step 1 concludes with a critical decision point. The applicability of methods for RCRA compliance will have been demonstrated for a solid waste matrix and a liquid waste matrix. In parallel permitting, testing of treatment efficiencies, plant design, and alternate treatment and disposal strategies will have advanced. This information will be assimilated to provide a basis for future characterization needs supporting the acquisition of decision-quality data.

’

SUMMARY DISCUSSION With the approaching end of Step 1 of the Regulatory DQO, we are heading into a consciously built-in decision point where the parties to the DQO committed upfront to reexamine whether changes are warranted in Step 2 implementation based on Step 1 results. In this evaluation, it is intended that continued tank waste characterization efforts will be tailored to obtaining effective data based on a better understanding of tank waste characterization capabilities and a current look at what data needs may be met. The validation of preparation and analysis methods for regulated compounds in tank waste matrices was intended to be the first step in providing technically defensible data for continued tank waste storage, preparation of the waste treatment plant permit, preparation of the treatment variance and delisting petitions, and to possibly meet other tank waste management data needs related to regulatory compliance. As the end of Step 1 approaches, deliberate cooperative efforts among regulators, stakeholders, waste management contractors, and the DOE should continue to focus on collecting data that are effective in serving their decision-making needs. Progress toward this end will be best supported by an attitude reflected by Nancy Wentworth, director of Quality staff for the EPA Office of Environmental Information (2002), “Get the right data, Get the data right, and Keep the data right.” ACKNOWLEDGEMENTS The authors would like to acknowledge the contributions of Gertrude Patello and her project staff, Battelle, Pacific Northwest Division; David Blwnenkranz, Bechtel National Inc; Nancy Welliver, DMJM H+N; and Jerry Yokel, Department of Ecology.


101

REFERENCES I KD Wiemers, ME Lerchen, M Miller, and K Meier, “Regulatory Data Quality Objectives Supporting Tank Waste Remediation System. Privatization Project,” USDOE Report PNNL-12040, Rev. 0, Pacific Northwest National Laboratory, Richland, Washington, 1998. * KD Wiemers, RT Hallen, H Babad, LK Jagoda, and K Meier, “A Compilation of Regulated Organic ConstituentsNot Associated with the Hanford Site, Richland Washington,” USDOE Report PNNL-11927, Pacific Northwest National Laboratory, Richland, Washington, 1998. KD Wiemers, P Daling, and K Meier, “Rationale for Selection of Pesticides, Herbicides, and Related Compounds from the Hanford SSTDST Waste Considered for Analysis in Support of the Regulatory DQO (Privatization),” USDOE Report PNNL-12039, Pacific Northwest National Laboratory, Richland, W a s ~ g t o n 1998. , KD Wiemers, H Babad, RT Hallen, LP Jackson, and ME Lerchen, “An Assessment of the Stability and the Potential for In-Situ Synthesis of Regulated Organic Compounds in High Level Radioactive Waste Stored at Hanford, Richland, Was~ngton,”USDOE Report PNNL- 11943, Pacific Northwest National Laboratory, Richland, Washington, 1998. KD Wiemers, ME Lerchen, MS Miller, and NC Welliver, “Logical Selection of Analytes for H d o r d TWRS Privatization Waste Feeds,” 53rdNorthwest Regional Meeting of the American Chemical Society, Richland, W ~ ~ ~ o n , 1998. KD Wiemers, ME Lerchen, and M Miller, “An Approach for the Analysis of Regulatory Analytes in High Level Radioactive Waste Stored at H d o r d , Richland, Washington,” USDOE Report PNNL-11942, Pacific Northwest National Laboratory, Richland, W a s ~ n ~ o1998. n, EPA, “Test Methods for Evaluation Solid Waste PhysicaVChemical Methods,” SW-846,3rd Edition, as amended by Updates I (July, 1992), IIA (August, 19931, IIB (January, 1995), and 111, US. Environmental Protection Agency, Washington, D.C., 1997. GK Patello, TL Almeida, JA Campbell, OT Farmer; EW Hoppe; CZ Soderquiest, RG Swoboda, MW Urie; and JJ Wagner, ‘‘Regulatory DQO Test Plan for Determining Method Detection Limits, Estimated Quantitation Limits, and Quality Assurance Criteria for Specified Analytes,” USDOE Report PNNL13429, Pacific Northwest National Laboratory, Richland, Washington, 2001.

’

102


Chapter 173-303 WAC. “Dangerous Waste Regulations.” Washington Administrative Code, as amended, http://www.ecy.wa.gov/ laws-rules/lawsetc.html. l0 Crumbling, Deana M. et. al “Managing Uncertainty in Environmental Decisions,” Environmental Science and Technology (October 1,2001).


103


EFFECT OF T ~ S I T I O N ~ O N - T ~ S I T IMETAL O N MODI~CATION ON THE ACTIVITY OF GazO3-Al203 CATALYST FOR NOx REDUCTION BY HYDROCARBON UNDER OXYGEN-RICH CONDITIONS

Md. Hasan Zahir, Singo Katayama and Kunihiro Maeda Synergy Ceramics Laboratory, FCRA, Shimo-Shidami, Moriyama-ku, Nag0ya 463-8687, Japan

Masanobu Awano Synergy Materials Research Center, National Institute of Advanced Industrial Science and Technology (AIST), S ~ 0 - ~ oS~ y ~a m a~ - ~ ku, Nagoya 463-8687, Japan

ABSTRACT

The effect of additives on the catalytic performance of Ga203-Al203 has been studied for the selective reduction of NO with ethylene in oxygen-rich atmosphere. Ga203-Al203 with additions of Zn, Ni, Cr, Mn, Fe and La metal oxides were prepared by a co-precipitation method Among the catalysts tested, Zn-Ga203-Al203 with 15% Zn content exhibited the highest activity over a wide range of temperature. The addition of NiO and ZnO enhanced further the activity of Ga203-Al203 in the lower temperature region for the case of Ni doped and higher temperature region (450-600 'C) for the case of Zn-Ga203-Al203catalysts. Zn- Ga~O3-Al203catalysts are active for the re~uctionof NO with both CH4 and CtH4 in the presence of oxygen. The fact that no deactivation behavior was observed during continuous reaction conditions as well as its excellent reproducibility character is remarkable. The high activity and selectivity of nontransition metal doped-Ga203-~203,were att~butedto the spine1 type structure, containing highly dispersed transition and non- transition metal cations in A 1 2 0 3 matrix. INTRODUCTION The selective catalytic reduction (SCR) of NOx by hydrocarbons is under investigation worldwide as the most attractive technique for NO reduction to Nz in exhaust gas streams. In this regard Ga203-Alr03catalyst has the potential for a breakthrough in the catalytic reduction of nitric oxide by hydrocarbon under excess oxygen conditions [I]. The capability to retain high catalytic activity in the presence of considerable excess of oxygen is one of the unique properties of this ~

_

-

_

To the extent a u t h o ~ e under d the laws of the United States of America, a11 c o p ~ g hinterests t in this pub~cationare the property of The American Ceramic Society. Any dup~cation,reproduction, or republicationof this p ~ ~ c a t i or o nany part thereof, without the express written consent of The American Ceramic Society or fee paid to the Copyright ClearanceCenter, is prohibited.

Environ~~ntal Issues and Waste ~ a n a g Tec~nologies ~ ~ ~ n ~VIII

105

catalyst. In addition, Ga203-A1203showed higher tolerance against water, and SO2 than Ga-ZSM-5, which indicates the possibility of designing de-NOx catalysts with high durability by using a nonzeolitic matrix [2]. However, the effective temperature window of the catalyst for NO reduction was relatively narrow and thermal stability is questionable. Some SCR catalysts containing aluminate-like phase were reported to be effective for NO reduction. For example, Hamada et al. [3], Kung et al. [4] studied Co/Al2O3, and showed that as the calcination temperature increased up to 1073 K, supported cobalt species changed from Co304 to C o A l 2 0 4 and the activity and selectivity increased. Okazaki et al. reported the similar results on Ni/Al203, and suggested that active metal species of these catalysts are the corresponding metal aluminates [5]. However, previous studies mentioned that zinc-alumina spinels work well for many applications because of their strong acidbase properties [6]. Although Zn is sometimes regarded as member of the 3d transition metals, ZnO is actually not a transition metal oxide because the 3d orbital of Zn2" is filled [7]. Usage of non-transition metal oxide for NO reduction has effective influence because the high selectivity of Ga203-Al203 was originated fiom non-transition metal (gallium) i.e., non-reducing nature of gallium ion resulting in the low oxidation activity [13. Therefore, we assumed that Zn-Ga2O3-Al203 i.e., structurally isomorphous ZnAl2O4 and ZnGazO4 or ZnAlGa04 system might be interesting for the reduction of NO by hydrocarbon. In this study, we investigated the effect of metal oxide additives namely Zn, Ni, Cr, Mn, Fe and La oxides on Ga203-Al203 with spine1 structure for the selective reduction of NO by CH4 and C2H4. EXPERIMENTAL All metals doped Ga203-&03 catalysts were prepared by co-precipitation method. Metal nitrate was used as metal sources, while ammonium carbonate were used as bases. The loading of metal oxide and Ga203 was fixed at 15 and 30mol %, respectively. Appropriate amounts of starting nitrates were dissolved in distilled water and aqueous ammonium carbonate ((NH4)2C03) solution (2.0 mol/L) was added to the solution to coprecipitate metallic ions. In this procedure, the pH of the mixed solution was kept at ca. 9.0 and then the solution was vigorously stirred for 24 h. The precipitate thus obtained was washed with distilled water three times, followed by drying at 110 OC and calcinations at 800 0 C for 5h in air. The crystalline phases were identified by X-ray diffractometry. Morphology of the product particles was examined using scanning electron microscopy. Particles size and surface area of the powders were also determined

106


by laser diffraction method and Brunauer-Emmett-Teller nitrogen adsorption technique. Catalytic activity of the resulting powder was measured using a fixed bed flow reactor. The sample powders of 200 mg were placed between quartz wool plugs in the reactor. The reactant gas mixture consisted of 0.1 % NO, 1.510%0 2 and 0.2 % C2H4 balanced by helium at a space velocity (SV)of 12,000 h-'. Flow rate of each reactant gas was controlled by a mass flow controller. The total gas flow rate was fixed 50 cm3min-' (WE= 0.24 g s ~ m - ~Decomposition ). of the nitric oxide and total NOx in the reactor effluent was detected using a chemiluminescent NO-NOx gas analyzer. All other reactants were analyzed by on-line gas chromatograph. Before the catalytic reaction, the catalysts were pretreated in air at 800 OC for 5h. NO conversion to NZ and C2H4 conversion to CO2 calculations are based on the following expressions: NO conversion to N2 (%) = {2[N2]/[NO]'"}x 100, C2H4 conversion to CO2 (%) = {(1/2)[CO~]/[C~H4]'"}XlOO where [NOIin or [C2H4Iia are the inlet NO or C2H4 concentration, respectively, and [Nz] or [CO,] are the concentration of N2, or CO2 in the reactor effluent gas. RESULTS AND DISCUSSION Fig.1 shows XRD patterns of ZnO and Ni- GatO3-AlzO3 diffiaction lines due to a spinel phase. Graphs for the X-ray spectra of these powders did not show any excess ZnO, NiO, A1203, or Ga203that was present as unreacted material. On the other hand, as for the catalysts containing La, Mn and Cr, gave the diffraction peaks assigned to each metal oxide additives, indicating the formation of phase pure spinel or y-Al203 might not be possible for this present reaction system. Lines due to ZnO were not detected even on 30 mol.% Zn-Gaz03-Al203, while those due to spinel phase became intense, probably because its composition is close to mzo4 or ZnGazO4 (stoichiometric spinel). It is well known that zn&o4 and ZnGazO4 are structurally isomorphous, where the A13+and Ga3+ cations are interchangeable, and a compound that contains both aluminum and gallium i.e., zinc aluminogallate, can be synthesized [8]. Additive effect of metal oxides: Fig.2 shows the activity of Ni-, Zn-, Mn-, Cr-,Fe-, and La-doped Gaz03N203 for NO reduction by CzH4 in the presence of oxygen. Although Ga203M2O3 catalyzed effectively the NO reduction by CzH1 with high NO conversion more than 98%, the addition of NiO and ZnO enhanced further the activity of Ga203-Al203 in the lower temperature region (350-530 OC) for the case of Ni


107

doped Ga203-Al203 and higher temperature region for the case of Zn-Ga203-Al203 at wide range of temperature 450-600 'C. Usage of non-transit ion metal oxide for NO reduction is important because the high selectivity of Gaz03A1203 was originated fiom non-transit ion metal (gallium) i.e., G a,O,-Al,Os-C r non-reducing nature of gallium ion resulting in the low o~dation activity f2]. Therefore it might assume that two 20 30 40 50 60 70 non-transition metal Dlffroctlon angle, 28 / * ( C u K a ) Fig.1. XRD patterns of various trensitlon metal doped G a 2 0 3 - A k 0 3 catalysts. i O n S (Ga3' and Zn2+) in the inert tetrahedral position enhance the activity at wide range of temperature. That is to say, ZnO on the Gaz03-Al203 exhibit a cooperative effect with Ga3" for the removal of NOx with C2H4, especially when the temperature is between 400 to 600 'C. The ~ a x i NO ~ ~conversion m on Zn- and Ni-Gaz03-Al203, was identical in compare with its non-doped Ga203-AlzO3. This is a remarkable result, because reported results of Inz03-, SnOz-, COO-, CuO- and &-doped Ga203-Al203 shows the maximum NO conversion was much less than that on Gaz03-A1203[2]. This is because h y ~ o c oxidation ~ ~ n by oxygen, which is a side reaction consuming hydrocarbon, proceeds predominantly because of the too high oxidation activity of CuO and Ag. In the case of In203 and SnO2, NO reduction was also decreased, although their propene oxidation activity was not so high. On the other hand for Mn, Cr, and La-modified catalysts, the enhancement in activity was very poor probably because the weak formation ability of y-Al2O3 phase or tiny spine1 structure. The activity of FeO-Ga203-AlzO3 was also low and it can be assume that the presence of large iron oxide particles catalyzing C2H4 oxidation with dioxygen. It is worth noting that on open supports, such as alumina, zirconia and sulfated zirconia, the activity and selectivity of CO,Pd, Ag, Ni, Mn, and Fe are limited by structural

108

E n v ~ o n ~ ~ nIssues t a l and Waste ~ a ~ a g e ~~ee nc ~ t n o ~ oVIII gi~s

effects, especially low dispersion, which renders the catalyst more active for CH4 combustion than for the SCR of NO [2]. RecentIy it has been shown that Ga203/Al203 system exhibits high catalytic activity not only by higher hydrocarbon but also by lower hydrocarbon. In particular, SCR by CH4 is most challenging since CH4 is relatively inert and generally reacts with 0 2 much faster than with NO thermally and over most catalysts [9]. In this study we observed that Zn-Gaz03/A1203 and Ni-Ga203/Alz03 catalysts retained their high catalytic activity even if CH4 concentration becomes 1000 ppm. Effect of 0 2 concentration The capability to retain high catalytic activity in the presence of considerable excess of oxygen is one of the unique properties of this catalyst. At present, the more widespread and well-supported point of view is that the catalytic activity of gallia containing systems for NO reduction in the presence of excess oxygen is due to the presence of gallium coordinately unsaturated cations (cus) on the surface (Lewis acid sites and Gazis)[lO]. In the presence of 1-10%0 2 , the NO conversion did not bend over, achieving almost complete reduction at 500 0 C. In the absence of oxygen, the NO conversion was quite low over Ga203Al2O3 and Zn-Ga203-Al203. The NO reduction to N2 increased up to 98% with increasing oxygen concentration. The results shown that the presence of oxygen is necessary for NO reduction and improves the NO removal activities by accelerating NO oxidation. Lifetime and reproducibility Another remarkable feature of Zn-Gaz03-Alz03 is, as expected, high thermal stability and stable catalytic performance, which is one important criterion for commercial application. The effects of recycling on the activity of the Zn-Ga203-Al203 catalyst are another point worth of discussion. For Zn-Ga203A I 2 0 3 catalysts, a steady-state conversion (96%) was obtained immediately at the


109

beginning of the run at 5OO0C, and the conversion did not decrease even after 24h of continuous reaction. CONCLUSIONS

Transitiodnon-transition metal oxides-aluminogallate, exhibited spinel-type structure, show high activity and thermal stability. Highly effective catalysts for SCR by hydrocarbons (CH4 and C2H4) were obtained without the E M - 5 framework. The activityhtability of Zn-Ga’03-Al203 catalyst remained high under demanding reaction conditions; namely in 10% 0 2 , 1000 ppm NO, 1000 ppm hydrocarbon and high space velocities. No deactivation behavior was observed during continuous reaction conditions as well as its excellent reproducibility character is remarkable. These results clearly demonstrate a strategy to design de-NOx catalysts with high activity and durability by using non-zeolitic matrix. REFERENCES ‘M. Haneda, Y. Kintaichi, T. Mizushima, N. Kakuta, H. Hamada, Applied Catalysis B, 31 81-92 (2001). ’K. Shimizu, A. Satsuma, T. Hattori, Applied Catalysis B, 16 319-326 (1998). 3H. Hamada, Catalysis Today, 22 21-40 (1994). 4JY. Yan, MC. Kung, WMH. Sachtler and HH. Kung, Journal of Catalysis, 172 [l]178-186 (1997). ’N. Okazaki, Y. Katoh, Y.Shiina, A. Tada and M. Iwamoto, Chemistry Letter, 889-890 (1997). 6 R. Roesky, J. Weiguny, H. Bestgen, U. Dingerdissen,Applied Catalysis A, 176 213-220 (1999). 7 T. Tsubota, M. Ohtaki, K. Eguchi and H. Arai, Journal of Materials, Chemistry, 7 [l] 85-90 (1997). 8 SK. Sampath and JF. Cordaro, Journal of the American Ceramic Society, 81 31 649-654 (1998). ‘Jd. Armor, Catalysis Today 29 43-45 (1996). 10 Yu. N. Pushkar, A. Sinitsky, 00. Parenago, AN. Kharlanov, EV. Lunina Applied Surface Science, 167 69 (2000).

110


Vitrification Technology and Melter Disassembly


COGEMA EXPERIENCE IN OPERATING AND DISMANTLINGHLW MELTER

R DO-QUANG (1) ,JL.DESVAllX(2), P. MOUGNARD(2) A. JOUAN (3), C. LADLRAT (3) (1) COGEMA, 2 rue Paul Dautier, BP 4,78141 V & y Cedex, FRANCE (2) COGEMA, 5044 Beaumont La Hague cedex (3) CEA, Valrh6 / Marcoule, BP 171,30207 Bagnols-sur-Ct2z.e Cedex, FRANCE ABSTRACT

The vitrificationof high-level liquid waste produced from nuclear fuel reprocessing has been carried out industrially for over 20 years by COGEMA, with two main objectives : containment of the long lived fission products and reductionof the final volume of waste. Research performed by the CEA (the French Atomic Energy Commission) in the 1950's led to the commissioning of the Marcoule Vitrification Facility (AVM) in 1978. In this plant, vitrified waste is obtained by first evaporating and calcining the nitric acid fed solutioncontainingfission products. The calcine is then fed together with glass fiit into an induction-heated metallic melter. Based on the industrial experience gained in the Marcoule Vitrification Facility, the vitrification process was implemented at an even larger scale in the late 1980's in the R7 and T7 vitrification facilities of the La Hague reprocessing plants.

So far, COGEMA's vitrification facilities have produced more than 11000 high-level glass canisters, representing more than 4500 tons of glass and 4500 million curies immobilized in glass. More than a technical success, in-line vitrificationof HLW produced by o p t i n g reprocessingplants has become a commercial reality. The basic principles leading to the choice of the two-step vitrification process with hot induction metallicmelter are : The separation of the functions, to have simpler and more compact equipment and to limit the size of the melter allowed for complete in-cell assembly and dissasembly with moderate size overhead cranes, master-slave manipulators and remote controlled tools. Easy remote maintenance of the process equipments with an optimization of solid wastes generated during operation The fact that the heating system is outside the metallic melter and thus independent from the melting pot, allow it to be not contaminated by HAL glass, to be not sensitive to the glass melting (no wear, no corrosion, no shorting and easy to start even if the metallic melter is full of glass), to be stop and to be maintain easily .

In parallel consistent and long term R&D programs have enabled continuous improvement of the process. The average melter lifetime now exceeds the design basis value (5000 hours instead of 2000 hours) and less than one week is necessary to stop a vitrification line, to change the melter and to restart the vitrificationoperation. The melting pot replacement is based on preventive maintenance to change equipment or subequipment before the failure in order to minimize the number of component or subcomponent to be change and the level of component's contamination The volume of secondary wastes fiom maintenance operations is thus minimized. Pieces of worn equipment are generally of small size, can be easily splitted for conditioning in accordance with COGEMA solid waste management strategy.



113

1) COGEMA industrial exDerience in Vitritlcation The vitrification of high-level liquid waste produced from nuclear fuel reprocessing has been camed out industrially for over 20 years by COGEMA, with two main objectives : containment of fission products and reduction of the final volume of waste. Research performed by the CEA (the French Atomic Energy Commission) in the 1950's led to the selection of borosilicate glass as the most suitable containment matrix for waste from spent nuclear fbel. This led to the commissioningof the MarcouleVitrification Facility (AVM) in 1978. Based on the industrial experience gained in the Marcoule Vitrification Facility, the A W vitrification process was implemented at larger scale in the late 1980's in the R7 and T7 facilities in order to operate its in line with the UP2 and UP3 reprocessing plants. Both vitrification facilities are equipped with three vitrification lines having each a glass production capacity of 25 kg/h. With two line in operation and one in stand by, each vitrification facility of La Hague reprocessing plants was designed to have a glass throughput of 50 kg/h and to meet the production requirements with sufficient flexibility of operation .

So far, COGEMA's R7 and T7 facilities at la Hague have produced (by the end of 2001) more than 8700 high-level glass canisters (CSD-V), representing more than 3500 tons of glass and 3,6 billion curies immobilized in glass. More than a technical success, in-line vitrification of HLW produced by operating reprocessing plants has become a commercial reality that led, in 1995, to the first return of glass canisters to COGEMA customers. 2)The R7/T7 Process 2.1) The La Hame two steo vitrification orocess

The French Atomic Energy Commission (CEA) began research on the immobilization of HLW in 1957 and led to the choice of a two-step vitrification process implemented first in the Marcoule Vitrification facility (AVM-1978) and extrapolated it in term of capacity and design for La Hague Vitrification Facility (R7-1989 and T7-1992).

114


In the two step process, the nitric acid solution containing the concentrated fission products solution coming from reprocessing operation is adjusted in a stirred vessel to make its chemical composition compatible with specificationsfor the glass product. The solution is then fed to a rotary calciner where it is heated up to 600°C by four distinct zones.

The calcine is heated under air and most of the nitrates are transformed into oxides. Aluminum nitrate is added to the feed prior to calcination in order to avoid sticking in the calciner (melting of NaN03). Sugar is also added to the feed prior to calcination to reduce some of the nitrates and to limit Ruthenium volatility. At the outlet of the calciner, the feed is still in the oxidized state, with significant amounts of nitrates left. The calcine falls directly into the melting pot along with the glass frit which is fed separately. The melting pot is fed continuously but is batch poured. The melting pot is made of base nickel alloys; The glass in the bottom of the melter is maintained heated to a temperature of llOO°C and is fully oxidized. A glass batch is poured into a canister every 8 to 12 hours depending on the composition of the HLW solutions being treated. The canister (CSD-V), which has a volume of 150 liters is filled with two batches of 200 kg each. The maximum activity at the time of vitrification is of 760,000 Ci per canister The maximum contact dose rate of the canister at the time of production can be greater than lO5rad/h Off-gas treatment comprises a hot wet scrubber with tilted baffles, a water vapor condenser, an absorption column, a washing column, a iodine filter and three HEPA filters. The most active gas washing solutions are recycled from the wet scrubber to the calciner. 2.2) Glass Prodact Oaalitv

The R 7 m glass was designed to hold, at the maximum, 18.5 % of radioactive waste oxides (fission products, actinides, noble metals and Zr fines), or equivalently an overall maximum waste loading ratio of 28 %. This limit was in fact set to avoid excessive heating of the glass during interim storage. The glass product has a very high activity content (predominantly 13’Cs, %r) and significant amounts of noble metals (3 wt % max). During the qualification process for the R7 and T7 facilities, waste homogeneity has been demonstrated through grab samples during pouring and destructive examination of canisters. Homogeneity of the product was satisfactory and no undissolved feed was observed. Satisfactory quality of the glass has also been demonstrated through the examination of production samples obtained in both the R7 and T7 facilities. The glasses were homogeneous with no undissolved feed and their characteristicswere in full agreement with the expected values. The residence time of the glass in the melter is in the range of a few hours, which is enough for complete glass elaboration, provided that the temperature is sufficient. The R 7 m formulation is known worldwide to have an outstanding durability, especially in the long term. Normalized releases using a powder test very similar to the 7-day PCT are less than 1/10 of the US acceptability criteria.


115

The basic principles leading to the choice of the two-step vitrification process and the multi lines design of La Hague vitrificationfacilities were : The separation of the vitrification functions, Easy remote maintenance of the process equipment. The separation of the process functions (calcinatiodvitrification) led to simpler and more compact equipment which is always desirable in a highly radioactive environment; Easy remote maintenance of all process equipment allowed for complete in-cell assembly and dissasembly with moderate size overhead cranes, master-slave manipulators and remote controlled tools. From a more general point of view, main process equipment of each vitrification line are located in a separate cell, while pouring and cooling cells are common to the three lines. All of these cells are equipped with cranes, master slaves manipulators and shielded windows for remote maintenance. They are associated with parking cells allowing crane maintenance as well as introduction of new replacement equipment. In the process cells, layout is optimized in order to facilitate access, modifications, and even addition of new equipment. The process equipment considered to be the least reliable are designed to be modular (i.e the calciner), so that their main subcomponentsare relatively compact and easy to replace remotely. The volume of secondary waste generated by maintenance operations is thus minimized. Pieces of worn equipment are generally small in size and can easily be splitted for conditioning in glass type canisters or cemented drum. As a result, maintenance operations are fully integrated into the process design and method of operations, which is of utmost importance to minimize downtime, volume and activity of secondary waste as well as to increase availability for production. 4) The Induction Heated Melter

4.1) General Drincfole The first work on vitrification of radioactive waste began in France in 1957 at the Saclay nuclear center. Techniques developed during this period to produce glass early used an induction-heated metal pot. The melter consisting of copper coil inductors embeded in a concrete structure, is designed to have a very long life time and can be remotely dismantled. The melter surround the melting pot which is the only really consumable item. The obvious advantage of this solution is that the heating system is outside the metallic melter (melting pot) and thus

0

Independent fiom the melting pot Not contaminated by HAL glass Not sensitiveto the glass melting (no wear, no corrosion, no shorting) Easy to start (even if the metallic melter is full or empty), to stop, to maintain or to replace.

Another major advantage of induction heating system is the simplicity to heat by Joule effect a metallic melter by using electric inductors.

116


From PIVER (~O'S), the first industrial-scale prototype unit in the world intended for vitrification of concentrated fission product solutions with induction melter technology (batch direct liquid feeding, 90 kg glass batch pour, 12,5 metric ton of glass for 5,25 million Ci), through AVM (70's) and to R7 and T7 (~O'S), induction heating have been always successllly industrially operated by CEA and COGEMA. 4.2) R7R7 Induction Melter Description

The separation of the calcination function fiom the melting function allows to limit the size melter with regard to the design capacity : 0

The evaporative capacity of the R7R7 calciner is of about 80 Vh. The power delivered to the melting pot is used to melt solid product (calcine and fiit glass) and not to evaporate liquid.

Thus, the melting pot is ovoid (long axis 1 m, short axis 0,35 m, total height 1,40 m, weight around 400 kg), and is made of base nickel alloys.

.M

The melter is supplied with 4000 Hz power by a 200 kVA generator at a voltage adjustable up to 600V through four superposed copper inductors cooled by water and connected to the cooling circuit by flexiblepiping. The generator output voltage allows overall adjustment, and each inductor is supplied to also allow individual power adjustment L

I I

t-

L._L.-.-. L

R7/T7 melter 4.2.1) Glass Donring

The melting pot is equipped with two bottom pouring nozzle, also heated by induction, to fill glass canisters : The first is used for nominal melter operation to periodically fill the canister with 200 kg glass charge. The second ensures a complete emptying of the melting pot at the end of each vitrification campaign or for maintenance intervention, and thus participates to reduce the levels of activity of the secondary waste. 4.2.2) Glass stirring

Initially, the R7 / T7 melters were only equipped with bubble stirring devices which were efficient to produce the glass for nominal values of the noble metal content. However it was demonstrated by CEA that glass viscosity and frit glass - calcine reactivity are influenced by noble metals.


117

As a consequence, COGEMA has undertaken the development of a mechanical stirrer in 1994 and

has deployed first these mechanical stkers on the T7 facility in 1996. The main objectives associated to the deployment of this new technology was to increase the noble metal content in the feed to the maximum specified noble metal content (3%) value and at the same time maintain the throughput capacity of the vitrification lines. The mechanical stirring was developed in addition to the bubbling device.

Melting pot are thus now equipped with both mechanical and bubbling devices; stirring devices have been designed to be compliant with the melting pot life time. A particular R&D program has supported materials selection used for the mechanical stirrer. Mechanical stirring proved to be successful in achieving the objectives since the noble metals content in the glass was increased h m an initial 1.5 wt. % to 3 wt. % (the maximum value) without any pouring problems or any metal accumulation in the bottom of the melter, with respect to the expected glass throughput. In fact, mechanical stirring also participate to the melting pot life time increase by reducing the melter wall temperature dispersion (see 4.2.4). 4.23) Control and monitoring

The instrumentation installed in R7fT7 melting pot enablesmeasurements of 0 0

The level and temperature of the molten glass The temperature of the melting pot’s wall

Since 1995, the level of glass in the R7 and T7 vitrification facility melting pots has been directly measured. There is a perfect correlation between this level measurement and the material balances determined with the crucible feed equipment. Factory calibration of the melter enables conversion of the level measurement into mass of glass It is very important to measure the temperature of molten glass and of the melting pot wall, as the temperature both determines the quality of the glass produced and governs proper operation of the process. Hot crucibles in the COGEMA vitrification facilities are equipped with thermocouples to measure:

0

The temperature of the molten glass, as this ensures that glass production conditions are adequate (a minimum average temperature of the molten glass is required as one of the parameters on which the quality of the glass produced depends), Wall temperatures of the melting pots, as these control heating of the melting pots (1 17OOC maximum under control).

The type of thermocouples used in induction heated crucible facilities at La Hague have been the subject of careful engineering and constant improvement.

A thermocouple characterization and evaluation program including: 0

Long duration endurance tests at constant temperature (1 15OOC) and with numerous temperature cycles (more than 2000), Integration of substantial doses of irradiation,

have been canied out and have validated the type of thermocouples currently used in the La Hague vitrification facility melting pots.

118


A continuous monitoring of the melting pot temperatures and of the electrical induction parameters are used 0 0

To operate the melting pot To follow the melting pot ageing and wear and to prevent failure. 4.2.4) Melter life time

One of the major on-line developments undertaken on the melter has been the work performed to extend the melter's lifetime. At the start of operations in the R7 facility, the lifetime of the oval-shaped metallic melter was less than the design basis value (2000 hours) due to the combined effects of thermal, and mechanical stresses as well as corrosion in the gaseous phase. An important R&D program was launched. Comprehensive studies were performed in order to better understand the electromagnetic, thermal and mechanical behavior of the melting pot at the different stages of operation as well as the corrosion mechanisms at play. In particular, the power transfer fiom the inductor coils to the melting pot wall was analyzed in detail.

These studies helped to determine which species were responsible for corrosion. They also showed that the thermal power released in the melting pot's wall and therefore the temperature gradients in the melting pot could lead to high levels of stress as well as condensation of certain corrosive species. As a result of these studies, the design, material of the melter and the method of controlling the temperatures were modified. These changes led to a sharp increase in the lifetime of the melters. At present, after ten years of operation, the lifetime of the standard melting pot exceeds the design basis wlues by more than a factor of two. The average melting pot lifetime is of about 5000 hours, with an award at 6400 hours corresponding to more than 200 hundred glass canisters produced with a single melter. Today one melting pot by year and by vitrification line is used in R7/M vitrification facilities,that has a great impact on reducing process downtime and secondary wastes. 5) Melter redscement and dismantling

The R7 and T7 are mature vitrification facilities where operation and maintenance principles have been optimized with two main objectives : 0 0

Maximize production availability Minimize volume of secondary wastes

Even if efforts have been made to avoid the need for maintenance (e.g. by improving the reliability of equipment), some equipment, because of their nature or complexity, need to be periodically maintain or replace. A specific effort has been done to facilitate their maintenance and minimize intervention duration.

The R7R7 melter, is thus designed to be small, compact, easy to replace remotely, and as consequence cheap. As a result, melter's maintenance operations are l l l y inkgrated into the process design and method of operations.


119

Remote maintenance is performed incell with cranes and remotely operated tools, using master-slave manipulatorsor servo manipulators. The melting pot replacement is based on preventive maintenance to change equipment before failure in order to minimize. 0

0

The number of componentor subcomponentsto be changed (meltingpot, melters.. .) The level of component’scontamination (limit contact with glass)

The unique experiencegained by COGEMA’svitrificationoperators, and the continuouscontrol of the melting pot operation (temperature, electrical induction parameters, analysis of small drift) allows us to anticipate and to change equipment before fatal wear ; this management is a mix between ‘periodical exchange’and ‘management value’ The main source of preventive intervention are metallic pot wear (corrosion, thermal and mechanical constraints ...) When the decision to change a melting is taken, the melter is emptied fiom its molten glass and the vitrificationline is progressivelystopped. When the melting pot is cold, it is disconnected fiom the calciner, and the melter which is on a mobile cart, is moved back. It is then possible to get the melting pot out of the melter, to transfer it to the dismantling cell (located above the vitrificationcell) and to replace it by a new one.

To have range of magnitude, ten days are routinely necessary to stop and to restart a vitrification line with a new melting pot.

Melting pot transfmji-omthe melter to the dismantling cell 5.2) Melting Dot dismantling Over the whole COGEMA complex, and especially at the L a Hague facilities, effort have been

focused for several year on the minimization and rationalization of the volume of conditioned solid waste coming from maintenance operation (contaminated tools, replaced components...) in standardized waste package. For example the Vitrified residue (CSD-V) and compacted residue (CSDC) have the same external design allowing 0

120

To standardizedhandling operation (same devicesto handle either CSD-C or CSD-V canisters)


0

To simplify transport operation (same shape of transportation cask) To rationalize interim and long term storage if needed (additional possibility of mixed storage CSD-C/CSD-V)

These efforts encompassed all the aspects of waste generation and management, fiom plant and equipment design, to the development of specific repair techniques or to the implementation of rigorous sorting and decontamination of the waste at the source especially for materials in contact with

HLW. Vitrification dismantling cell are equipped with dismantling and decontamination tools, counting cell, and waste preconditioning unit. After dismantling and decontamination operation, pieces of worn equipment are generally of small size and sorted according to their activity level and material nature to be 0

Recycled inside the process, and conditioned in CSD-V type glass canister Send to La Hague waste treatment facility (AD2) to be cemented as technological waste, or, at soon compacted in the new La Hague Compaction facility (ACC) according to their level of activity (LLW / ILW).

For the melting pot, the dismantling operation consists to immobilize the melter in a vice and to cut it in small pieces of 300 x300 rnm by means of a sawing machine. For the melting pot, the dismantling operation consists to immobilize the melter in a vice and to cut it in small pieces of 300 x300 mm by using a sawing machine. Generally several blades (three to five) are necessary to completely dismantle the melting pot in small pieces.

Sawing machine

The mechanical saw alternative movement participate also to separate possible residual cold glass fiom metallic pieces and contribute do downgrade the waste. A complemenm mechanical cleaning of metallic piece could be completed by pneumatic drill, operated by master slave manipulator.

To have range of magnitude, six days of continuous operation are routinely necessary to dismantle a melting pot. Metallic pieces are decontaminated in high temperature nitric acid tank, counted in a measurement cell, sorted and preconditioned in EW basket (650 mm diameter, 1 m height). Three ILW baskets are necessary to precondition a melting pot and associated worn saw’sblades.

To finish the waste conditioning, ILW technological baskets are routinely sent by mean of mobile shielded cask to the A D 2 facility to be cemented and conditioned as technological waste in


121

CBF-C2 residue (1200 liters volume, 4 metric ton max in weight). Generally three 500 Ci CBFC2 are necessary to condition a melting pot and its auxilaries (saw’s blades, current maintenance tools). Soon, a second way for ILW basket conditioning will be to send it (according to their activity) to the new Compaction Facility (ACC) to reduce by factor 4 the final waste volume by using universal standard canisters (CSD-C). 0

The recovered glass coming from dismantling melting pot operation are conditioned in HLW basket (380mm diameter, 1 m height) to be recycled in a specific standard vitrified canister (CSDv).

Melting pots are dismantled in line with the process in La Hague vitrification facilities according to the La Hague waste management strategy, allowing an optimization and standardization of the final waste volume. Because of the little number of melting pots to be dismantle by year, the treatment of melting pot in dismantling cells is a routine and short time operation l l l y integrated in the ordinary vitrification facility life. 6)Conclnsion

COGEMA has been operating industrial HLW vitrification facilities for over twenty years. The feedback fiom hot operations and the long-term R&D programs conducted with the CEA have helped to continuously improve the process in all of its aspects (glass formulation, process, associated technologies, operations and maintenance). The R7 and T7 vitrification facilities operating in-line with COGEMA’s two major 800 ton capacity commercial reprocessing plants have had outstanding records of operation, not only fiom the standpoint of total glass production and plant availability but also with respect to safety, remote incell maintenability, and secondary waste generated, demonstrating the maturity of the French vitrification process. With respect with the COGEMA layout and maintenance concepts, the melter induction technology used in R7/T7 vitrification has been designed to be small, compact, reliable and easy to replace remotely. As a result, melter’s maintenance operations are fully integrated into the process design and method of operations, and leads to optimiz,e maintenance time intervention and volume of secondary waste.

To go a step further in the Induction heating technology development, CEA and COGEMA are

committed in the development of the Cold crucible Melter (CCM) technology. The use of the Cold crucible Melter technology will lead to a virtually unlimited equipment service life and great flexibility in dealing with different types of waste. The high specific power directly transferred to the melt will allow high operating temperatures and high waste loading factor without any impact on the process equipment. And at last, the CCM technology remains always compact, simple and modular taking benefit from COGEMA’s melters operation experience.

References : Major breakfhroughsin high level waste vitrification G. MEHLMAN, R.DO QUANG, A. JOUAN, Waste Management ’99, Tucson

122


CONCEPTUAL METHODS FOR DISPOSAL OF A DWPF MELTER AND COMPONENTS *Michael E. Smith, Dennis F. Bickford, Frank M. Heckendorn, Eric M. Kriikku Westinghouse Savannah River Company Savannah River Technology Center "Building 773-43A Aiken, SC 29808 ABSTRACT The Defense Waste Processing Facility (DWPF) has processed over 1.8 million kilograms of high level waste (HLW) glass since radioactive startup in 1996. The DWPF Melter is the heart of the vitrification process. The current plan is to store failed HLW equipment like the melter in Failed Equipment Storage Vaults. While this storage is acceptable in the short term, technology must be developed for proper long-term storage of these melters and other HLW equipment. Potential methods, including dismantlement sequences, for the disassembly, size reduction, and decontamination of a failed DWPF Melter will be discussed. INTRODUCTION The Defense Waste Processing Facility (DWPF) Melter was started up in 1994. It began processing radioactive feeds in April 1996. To date over 1000 canisters (1/6 of the projected canisters) have been filled with more than 1.8 million kilograms of radioactive glass. During the processing of this high level waste (HLW) at DWPF a need will arise to develop remote andor robotic systems to disassemble contaminated HLW processing equipment. This includes failed melters, vessels, and equipment. The current approach is to store this equipment in Failed Equipment Storage Vaults (FESV). While this storage is acceptable in the short term, technology must be developed to properly dispose of this equipment. This should include dismantlinghize reduction of the equipment, decontamination, disposal of the majority of the material as low level waste (LLW), and disposal of the remaining fraction as HLW materials. The DWPF melter will probably be the most difficult DWPF HLW equipment to disassemble and decontaminate. A single DWPF melter can hold up to 6000 kilograms of HLW glass. If an approach can be developed to dismantle and dispose of the melter, then similar techniques could be used with other HLW processing equipment. The design life of the melter was two years. This was based on a cited corrosion rate of the melter Monofrax K-3 refractories of 5.4 mils per day.' The actual expected corrosion rate should be much less and is based on pilot scale melter work and the operating history of the first DWPF melter (Melter 1). This melter is still operating in the DWPF after seven years. Although the melter design has proven robust, the life of future DWPF melters may not be as long due to the processing of feeds with higher levels of noble metals.* During the DWPF operations time of 25 to 30 years, several melters will most probably fail and be temporarily stored in FESV's. There is no facility at SRS that is currently setup to perform the dismantlement work on these melters. One concept would return failed melters and other large equipment to the DWPF Canyon after waste processing is completed. The equipment would then be disassembled and size reduced in the canyon. The required manipulators, tools, etc. are not currently in place or specified. In addition, there is no existing To the extent authorized under the laws of the United States of America, all copyright interests in this publication are the property of The American Ceramic Society. Any duplication, reproduction, or republication of this publication or any part thereof, without the express written consent of The American Ceramic Society or fee paid to the Copyright Clearance Center, is prohibited.


123

plan on the dismantlement of a DWPF Melter. The need to remotely disassemble and size reduce HLW equipment is a DOE complex wide problem. Making the task more difficult is the fact that the various sites have different melters and facilities available to do the work. This paper identifies potential methods for the disassembly, size reduction, and decontamination of the DWPF Melter. DWPF MELTER DESCRIPTION The DWPF Melter, shown in Figure 1, is a refractory-lined, sealed, stainless steel vessel with a 1.8-meter internal diameter and approximately 2.1 meters of internal height. Figure 2 shows DWPF Melter 1 assembly with the frame and lifting yoke. The 3.8-centimeter thick vessel shell is 304L stainless steel with an outer diameter of 2.56 meters. The shell contains no remotely removable parts. All shell components have at least a two-year design life.

Figure 1 - DWPF Melter Cross-Section Primary glass containment is achieved by use of 30.5-centimeter thick Monofrax K-3 fused cast refractory brick. This refractory is very dense and hard. Korundal XD refractory was used in the vapor space. Zirmul refractory was chosen for the refractory placed underneath the Monofrax K-3. The glass pool is maintained between 1050 - 1170°C. The DWPF Melter electrodes are four uncooled plates fabricated from Inconel 690 of sufficient thickness to last greater than two years. The DWPF Melter vessel shell is penetrated in the vapor space by four pairs of horizontal resistance-type Inconel 690 dome heater tubes. Each tube is 8.3 centimeters in diameter. The DWPF riser heater consists of an Inconel 690 serpentine heater that surrounds the ten-centimeter inside diameter Inconel 690 riser channel. The pour spout heater also consists of an Inconel 690 serpentine heater that surrounds the five-centimeter diameter of the pour spout channel. These strip heaters keep the glass flowing through the riser and pour spout channels at a temperature between 1050-1100°C. The DWPF Melter is composed of various components that are quite bulky and heavy. Table I gives the weights of the major components (total is 67,600 kilogram^).^

124


Figure 2 - DWPF Melter Assembly Table 1 - Weights of DWPF Melter Major Components Component Weight (kilograms) Vessel 17,600 Frame 12,900 Refractory 25,200 Components 9,800 Piping 1,200 Nozzle Material 900 Not only is the DWPF Melter large and heavy, it was not designed to be remotely disassembled. For example, the melter lid is attached to the rest of the melter with 56 bolts that are tightened and then tack welded in place. This lack of design for remote disassembly is also seen in that most of the major melter parts (like the lid) where not designed for remote removal. PREVIOUS MELTER DISMANTLEMENT EXPERIENCE There has been limited remote experience with the disassembly of HLW melters. There have been several pilot-scale or full scale melters that have either been inspected or disassembled after being used in non-radioactive testing. This section discusses these activities. Scale Glass Melter (SGM) Inspection The Scale Glass Melter was a two-thirds scale non-radioactive pilot melter of the DWPF Melter that was operated at SRS between 1985 and 1988. Before the melter was shut down a sequence of “high-risk‘’ tests which thermally cycled the melter were performed. After these tests and additional feed tests, the SGM was drained and shut down, the melter lid removed, and the refractories inspected. The main finding of the inspection of the SGM after all of this thermal cycling was that the K-3 refractory was basically intact with some minimal wear at the melt line. Integrated DWPF Melter System (IDMS) The IDMS was a one-ninth scale, non-radioactive DWPF pilot system. The melter was still operational at the time of its 1995 shutdown, although deposits of noble metals on the melter floor


125

were requiring higher current from the lower electrodes to keep the lower melt pool in the operational temperature range. Findings pertinent to melter disassembly work are as follows4: 0 The melt pool refractory (Monofrax K-3) experienced some thinning and spalling at the sides in the melt zone (area between glass and feed). Overall the K-3 was in excellent condition. 0 The Mullfrax 202 vapor space refractory (similar to Korundal XD) was in excellent shape. The lower electrodes showed wear at their bottom faces. This was due to the channeling of a portion of the lower electrode current flow (due to noble metals) through this area. West Valley Functional and Checkout Testing of Systems (FACTS) Melter Disassembly At the West Valley Demonstration Project (WVDP), non-radioactive vitrification processing operations were conducted from 1984 to 1989 using a slurry-fed ceramic melter similar to the one now in radioactive operation at the WVDP. After the tests were completed, the melter was removed and manually disassembled. The melter was then inspected to determine the impact on the melter internals due to five years of glass processing. The following gives the sequence of steps used to disassemble the melter5: The seal weld between the melter shell and lid were manually plasma arc cut. The lid was then lifted and inverted to inspect the refractory. The refractory was removed from the melter lid by use of jackhammers and pry bars. The process required hammering through the refractory to the fiberboard backing and then working the hammers and pry bars under the refractory and around the nozzles. The size of the refractory waste pieces could not be controlled. 0 The melter cavity was disassembled top down in layers. The vapor space refractory was first removed using the same manual techniques as were used on the lid. Glass penetrations 0.64 to 1.3 centimeters were noted in the refractory cracks below the melt line. The refractories were somewhat difficult to handle even manually. 0 Once the innermost layer of glass contact refractory was removed, the remainder was removed cleanly. Pamela Vitrification Plant Radioactive Melter The Pamela plant in Dessel, Belgium was designed to vitrify high-level liquid reprocessing waste. The plant is somewhat unique in that is was designed to handle the remote dismantling of large HLW equipment such as the melter. Operations were started in 1985. After the initial vitrification program was completed in 1991, two melters and other large scale equipment had to be dismantled so that the plant could be reused for a second vitrification program that was planned to be started in 1999. From October 1991 till March 1994 one melter and three other large pieces of equipment were dismantled and dismantling waste was conditioned. In total 187 drums (200 liters each) containing cemented medium-level waste were transferred to the ap ropriate on-site facilities. Some five tons of low-level dismantling waste were transferred as well.

B

Although the plant had a dismantling cell, it was decided to perform the dismantling operations in the melter cell. Both the melter cell and the dismantling cells are equipped with a heavy-duty manipulator. The melter cell also has a 2-ton overhead crane while the dismantling cell has a 20ton overhead crane. The melter cell has three lead glass windows at ground level with masterslave manipulator pairs at each window. Three more windows are located on the second level. Two of these higher windows have master slave manipulators as well. The design of the melter cell and the dismantling cell took into account the ability to dismantle the melter and process the dismantling waste. This design included having these two cells close to each other and having an adjacent intervention cell that allowed for the maintenance of the cell

126


cranes and heavy-duty manipulators. The dismantling was done by using existing manipulation systems and by adapting traditional tools like grinding discs, impact wrenches, vacuum cleaners, hydraulic jacks, hammer drills and grab tools for remote work. The work was completed within 2.5 years with about 25,000 man-hours. Lessons learned from this task are summarized below. 0 Dismantling techniques should be considered when designing a vitrification plant (access to and visual control in shielded cells must be easy) 0 Manipulators must allow for operation of multiple types of tools Large drums should be used for packaging of waste to reduce size reduction time 0 Personnel doing the work must be experienced remote manipulator operators 0 Grinding discs were better at cutting stainless steel than diamond encrusted saw blades (faster and less expensive) Spare parts used in the dismantlement work should be in stock to minimize time losses RELEVANT DOE D&D ACTIVITIES There have been several DOE funded decontamination and decommissioning (D&D) activities that are relevant to the dismantling of large radioactive equipment. A TFA report from O N 3 summarizes these various activities. These activities include the D&D Chicago-Pile No. 5 (CP-5) research reactor, the size reduction of the Tokamak Fusion Test Reactor Vacuum Vessel, and D&D work at the INEEL South Tank Farm in January 2000. Technologies used included the Dual Arm Work Platform (DAWP) at CP-5 (DAWP developed by ORNL), diamond wire cutting by Bluegrass Concrete Cutting, Inc., and the Modified Brokk Demolition Machine. A jointly written report from SRS and ORNL gives an outline for large-scale system operations and D&D work.7 Finally, an SRS report describing the conversion of a section of one of the separations buildings at SRS from a remote-crane-operated facility into a master-slave-manipulator-operated facility will give some insight into what it might take to convert an existing radioactive facility into a remote dismantlement facility.’ This work includes equipment removal and building decontamination. It also includes the installation of new service and support equipment. CURRENT DWPF MELTER STORAGEhIISPOSALPLAN During the design of the DWPF, special cells and equipment were considered for the dismantlement of failed DWPF Melters but were deemed too expensive. Therefore the current long-term storage/disposalplan for failed DWPF Melters is as follows. Disconnect the melter assembly from the Melt Cell Move the melter assembly to the Remote Equipment Decontamination Cell (REDC) for external decontamination Place the melter assembly into a carbon steel sealed storage box Transfer it via rail car to an underground Failed Equipment Storage Vault (FESV) There are now two existing FESV’s at DWPF. Each FESV is a reinforced concrete structure designed as a Category 1 nuclear facility. One melter assembly can be stored in each vault. Permanent storage of these melters in the FESV’s is probably not a viable option as this would amount to using the Savannah River Site as a HLW repository. Therefore, it is assumed that at the end of the life of the DWPF the failed melters will have to be size reduced and then the parts segregated based on waste classification. This may not have to be done if their radioactive content is low enough to meet incidental waste rules. This may be accomplished by moving the melter assemblies (one at a time) back into the DWPF Canyon to be D&D’d. There is no formal plan at this time as to how to do this D&D work. Because the DWPF was not specifically designed for large KLW equipment D&D, another existing facility or a new facility designed solely for this purpose may instead be used. Finally,


127

there is the option of shipping these failed melters to the HLW repository. The size and weight of the melter assembly along with the potential amount of HLW glass in the melter will probably make this transportation option impossible. Also, the federal HLW repository design has no provision for accepting and storing a package of this size and weight. DWPF MELTER DISMANTLEMENT This section describes the basic tasks that will be required to dismantle the DWPF Melter. The exact details cannot be addressed because the location of the dismantlement has not been determined. In addition, the details of how the melter assembly is disconnected and moved are not discussed here as these will be covered in DWPF procedures. There are basically three different approaches to DWPF Melter dismantlement. The first approach would be to remotely remove enough glass from the melter so that the melter would be considered non-HLW. The second approach would be to dismantle the melter from the outside in. The last approach would be the dismantlement of the melter from the inside out. Each of these options is discussed below. Many of the steps are similar for the various approaches and therefore the tooling required for the tasks are discussed in a later section. Glass Removal Approach The chances of removing enough glass from an intact melter to have the melter disposed of as non-HLW are fairly low. This is because the glass would have to be mined out remotely through the various 10-centimeter diameter top head nozzles. The glass cannot be accessed via the 20centimeter diameter off-gas nozzles because the dome heaters are directly below these ports. It may, however, be conceivable to take off the melter lid and then remove enough of the glass to store the melter as non-HLW. With regards to the maximum radiation levels for the melter to be disposed of as non-HLW, these numbers and rules change with time. Therefore, the specifics are not discussed here as dismantlement work will not occur for probably another 10 to 20 years. This timing could of course be accelerated if multiple melter failures occur due to problems such as the settling of noble metals. The basic steps to perform this dismantlement are as follows. Remove all melter top head components (the components would probably be removed if possible before the melter was shutdown so that residual melt pool glass would not trap these components in the glass) Remove all melter jumpers Cut and grind off the 56 melter lid bolts Remove the melter lid Cut away and remove the melter dome heaters sections on the inside of the melter Breakup/remove the glass fiom inside of the melter and melter lid and place in DWPF canisters or other appropriate disposal vessels Decontaminate the melter assembly and outside of the melter shell as much as possible (may require covering of melter with cover plate) without getting water, etc. inside of melter Determine radiation field of melter lid and main melter, estimate remaining HLW inventory Reinstall the melter lid back onto the melter Place melter assembly in storage box and either return to FESV or bury as non-HLW Place glass removed from melter in DWPF canisters or other appropriate disposal vessels Size reduce top head components and segregate based on activity Ievel (ensure HLW pieces can fit into DWPF canisters or other appropriate disposal vessels) Place HLW top head component pieces into same DWPF canisters or other appropriate disposal vessels that contains the glass removed from the melter and seal canister (decontaminate canister/vessel surface if needed)

I28


The main advantage of this approach is that it is the simplest and least expensive of the three alternatives. Minimal amounts of size reduction of melter materials would be required. The effort to remove the glass from the melter would be about the same as the other two approaches. The main disadvantage is the uncertainty as to whether or not the removal of the glass would then allow for the disposal of the melter as non-HLW. Also, the details of how to reinstall the melter lid back on the melter are uncertain at this time. If repository and waste regulations allow for this approach at the time of disposal, then it is recommended that it be used for the DWPF Melters. Inside-Out Melter Dismantlement Approach This approach is the same basic type of method cited in the West Valley dismantlement report.5 Remove all melter top head components Remove all jumpers Cut and grind off the 56 melter lid bolts Remove the melter lid Removehreakup the melter lid refractories (may be possible to first remove HLW contaminants from surface of these refractories to allow disposal of most of it as non-HLW) Cut away and remove the melter dome heaters sections on the inside of the melter Removehreakup the melter refractories above the glass level (may be possible to first remove HLW contaminants from refractory surfaces to allow disposal of most of it as non-HLW) Breakuph-emoveglass from inside of melter Removehreakup the remaining glass contact refractories (may be possible to first remove HLW contaminants from refractory surfaces to allow disposal of most of it as non-HLW) Remove and decontaminate melter electrodes Cut, remove the riserlpour spout section Cut, decontaminate, and remove the melter shell Cut, decontaminate, and remove the melter frame assembly (melter shell and frame may not need to be cut up or decontaminatedas they may be able to be disposed intact as LLW) Place glass removed from melter in DWPF canisters or other appropriate disposal vessels Size reduce top head components and segregate based on activity level Place HLW top head component pieces into same DWPF canisters or other appropriate disposal vessels that contains the glass removed from the melter This approach allows the melter shell to provide containment for the refractory pieces, therefore heIping to minimize the amount of contamination in the dismantling facility. It also allows for the movement of the melter by the use of the lifting yoke until the melter shell is cut. This approach is similar to the one that was used on the Pamela Melter as previously discussed.6 Outside-In Melter Dismantlement Approach In this approach, the melter shell is removed before the removal of the refractory. This approach is similar to the second disassembly alternative cited in the West Valley melter dismantlement report5 The less contaminated refractory (located at the higher melter elevations) is taken first and then the glass-contaminated refractory is removed. This allows for the bulk of the frame and shell to be disposed of as small pieces of LLW. The basic steps to accomplish this dismantlement are as follows. Perform the first six steps of the Inside-Out Melter Approach previously discussed Cut away, decontaminate, chop up, and remove the top portion of the melter frame assembly Cut, decontaminate, and remove the melter shell (not the bottom portion of the shell)


129

Removehreakup the melter refractories above the glass (may be possible to first remove HLW contaminants from refractory surfaces to allow disposal of most of it as non-HLW) Pryhemovehreakup the melter refractories below the glass (may be possible to first remove HLW contaminants from refractory surfaces to allow disposal of most of it as non-HLW) Breakuph-emoveglass from inside of melter Cut, decontaminate, and remove the remaining bottom portion of the melter shell Cut away, decontaminate, chop up, and remove the remaining bottom portion of the melter frame assembly Place glass removed from melter in DWPF canisters or other appropriate disposal vessels Chop up top head components and segregate based on activity level Place HLW top head component pieces into same DWPF canisters or other appropriate disposal vessels that contains the glass removed from the melter. The easier to reacWless contaminated refractories are removed first. Unfortunately, there is a greater chance of spreading contamination in the dismantling facility as there is no containment provided by the melter shell. In addition, the glass contact refractories will not be supported when being broken apart. This approach could, however, make the handling of the refractories easier. CANDIDATE DISMANTLING TOOLS In considering tools for the disassembly of a melter or other large HLW equipment there are several characteristics of the tools that must be considered. Obviously the tool must be able to remotely perform the task required. Because they will be used in a remote radioactive environment, other criteria should be considered as well. One of these is reliability. The speed at which the tool accomplishes the work along with the ease that is can be done must also be factored. The ease of repair or change out of the tooling is still another factor. Tooling should also be chosen based on whether or not it can be used with the existing manipulators, etc. that are in the dismantlement facility. Another consideration is the amount and type of secondary waste that will be generated by the various tools being considered. Finally, the use of proven technology is a prudent approach when choosing tools for this type of work. The West Valley and Oak Ridge reports have excellent summaries of tooling that could be used.”* The various tasks are given below with tooling that should be considered for each step. Before the melter lid can be removed, the 56 top head bolts that hold the lid to the melter vessel must be cut. This includes the tack welds from the nuts to the bolts and the nuts to the flange. The best choice for this would be a grinder tool with grinder discs. Grinding discs were successfully used during the Pamela Melter dismantlement.6Lifting the melter lid remotely was not considered in the design of the melter. There are four lifting holes located on the top of the melter lid that were used to lift the lid during installation. By using manipulators, chains may be able to be reinstalled at these four points to remove the lid. If not, the lid could be raised by accessing the various nozzles. This technique, however, could cause problems by breaking up the melter lid refractories during movement. A lid lifting jig or fixture could also be fabricated. This suggests that a welder to install lifting and handling tabs may be required. The breakup of the various melter refractories could be accomplished by the use of various tools. These include the following. Hydraulic or mechanically actuated wedge (to move refractory stuck in place) Ram (free falling weighted chisel) Hydraulic or mechanical spreader Jackhammer 0 Diamond wire saw

130


La Bounty Crusher Rotary tool with face grinding wheel (for refractory surface cleaning) Needle gun (for refractory surface cleaning) Abrasive water jet Grapple device (for picking up refractory pieces) Vacuum (to grab smaller pieces) The most likely scenario would be to grind or cut the face of the refractories to remove as much contamination as possible. A vacuum system could be used to collect the waste dusting from the grinding. A jackhammer with the proper force and stroke would then be used to break up the refractory. This was the tool of choice for the Pamela Melter work.6A heavy-use remote arm with “grabbing fingers” and possibly a modified grapple device (attached to an overhead crane) currently used at West Valley would be used to grab and move the broken up refractory pieces. The cutting of the Inconel 690 dome heaters (8.3-centimeter outer diarneted5.7-centimeter inner diameter) will have to be done with the heaters in place. The following tools may be used. Plasma torch Grinding discs CircuIar saw Abrasive water jet Shears Needle gun Grapple device If grinding discs are used, a vacuum system could be used to collect the waste dusting. The top head components are also made out of Incone1690 and have a 7.6-centimeter outer diameter. The grinding disk or a needle gun can be used to remove glass from these components before they are size reduced. Per past experience with the SGM, there should be about 2.5 centimeters of glass in the DWPF Melter if it is drained completely via the drain valve.” A melter that was not drained could have as much as 6000 kilograms of glass. The following tools may be used to remove the glass. Abrasive water jet Vacuum Needlegun 0 Jack hammer Pneumatic chisel Ram (weight chisel) Grapple device The cutting of the Melter Frame Assembly and the Melter Shell could use the following tools. 0 Shears Plasma torch Circular saw Abrasive water jet Grinding discs SUMMARY/RECOMMENDATIONS The current plan is to store failed DWPF Melters in FESV’s located at the DWPF. Because these failed melters may hold up to 6000 kilograms of radioactive HLW glass, they will eventually need to be D&D’d. At this time there is no facility at SRS specifically designed for the D&D of large HLW equipment. It may be possible to take the lid off of failed DWPF Melters and remove enough glass to then classify the melters as non-HLW. This would then allow the melters to be disposed onsite without a full scale D&D effort. If for technical or regulatory reasons this is not possible, two disassembly options are given. With the need for D&D work on HLW equipment recognized by the DOE complex, a DWPF Melter glass removal remote demonstration overseen by SRS is currently planned for FY03 at ORNL’s robotics facility using the Scale Glass Melter. In addition, West Valley is performing D&D tests on HLW vitrification equipment. In closing, the following recommendations are given with regards to this DWPF Melter D&D task.


131

If possible, remove the lid from the failed DWPF melters, remove as much glass as possible, replace the lid, and then store the melters as non-HLW glass. Pick tooling that has been proven to work in other D&D activities. Equipment reliability must be considered as well as ease of operation. Cleaning contaminants from the parts of the melter (refractory, etc) should be done to minimize the amount of HLW to store. The amount of secondary waste that will be produced must be considered when choosing the D&D sequence and tooling involved. If the DWPF Melter cannot be drained and it is still operational (yet deemed to be shut down due to the number of years run or reduced performance), then the flushing of the melter with at least three melter volumes of non-radioactive glass should be considered. Flushing of a drained melter should be considered as well. The design of future DWPF Melters and new HLW vitrification plants such as Hanford and Idaho should consider adding features that will make the D&D of these melters easier: 1. Design a way to remotely disengage and pick up the melter lid. 2. Add several larger nozzles (possibly for the feed tube nozzles) to allow easier glass clean out of the melter (especially noble metals at the melter floor) without removing the lid. 3. Consider making some of the failure mode parts of the melter (for example the riser/pour spout) remotely replaceable. 4. Consider a new melter design that is easily replaced and D&D’d. This design should weigh much less and minimize the amount of glass contact refiactory. Future HLW vitrification sites should consider designing the vitrification facility or a special celllfacility for the remote D&D of large HLW processing equipment. REFERENCES 1. Basic Data Report, Defense Waste Processing Vitrification Facility Sludge Plant, USDOE Report WSRC-RP-92-1186, July 1992. 2. D. F. Bickford, M. E. Smith, “The Behavior and Effects of the Noble Metals in the DWPF Melter System (U)”, USDOE Report WSRC-TR-97-0370, Savannah River Technology Center, November 30,1997. 3. B. S. Richardson, “Melter Glass Removal and Dismantlement”, USDOE Report ORNL/TM2000/324, Oak Ridge National Laboratory, October 2000. 4. C. M. Jantzen, D. P. Lambert, “Inspection and Analysis of the Integrated DWPF Melter System (IDMS) after Seven Years of Operation (U)”, USDOE Report WSRC-RP-96-575, Savannah River Technology Center, February 6, 1997. 5. “Recommended Methods for Decontamination and Decommissioning, Size Reduction, and Disposal of Melter and Components - Evaluation Report”,West Valley Nuclear Services, Co., February 28,2001 (no author or documentnumber cited). 6. P. Luycx, M.Demonie, M. Snoeclat, L. Baeten, (Belgoprocess, Gravenstraat 73, B-2480 Dessel (Belgium)). 1996. “Experience gained with the dismantlingof large components of the Pamela Vitrification Plant”, Proceedings of the International Topical Meeting on Nuclear and Hazardous Waste Management Spectrum ‘96. August 18-23, 1996 at Seattle, WA (U.S.A.), American Nuclear Society, Inc., La Grange Park, IL (U.S.A.). pp. 1717-1727. 7. F. Heckendom and R Kress, “Outline for Large-scale System Operations and D&D Report”, USDOE Report WSRC-TR-2000-00364, Savannah River Technology Center and Oak Ridge National Laboratory, September 2000. 8. R. D. Kelsch, A. J. Lethco, and J. B. Mellon, “Multipurpose Processing Facility to Separate Actinides”, Proceedings of 20* Conference on Remote Systems Technology, 1972.

132


EVALUATION OF CRYSTALLINITY CONSTRAINT FOR HLW GLASS PROCESSWG Pave1 Hrma, Josef Maty69; and Dong-Sang Kim Pacific Northwest National Laboratory P.O. BOX999, MS:K6-24 Richland, WA 99352 ABSTRACT It has been a commonly held assumption that constraining liquidus temperature (7') prevents the accumulation of crystalline phases in the high-level waste (HLW) glass melter because crystals, if they form at all, should dissolve easily in the melt at temperatures above liquidus. This, as the model calculation showed, is not the case in melters with fast circulation flow. If the melt circulates rapidly between cool and hot regions, crystals do not have a sufficient time to dissolve while in the hot zone. As a result, a steady-state size and concentration of crystals is established throughout most of the melter during normal operation. A consequence of this result is that the rate of crystal accumulation in the melter only slightly increases with increasing , 'T but strongly increases with increasing crystal size. For the melter simulated by the model, the TL could be 100°C above the accepted constraint without a serious impact on melter performance. Nucleation agents that keep crystals small abound in most HLWs but are often absent in simulated wastes for experimental melter runs. The weak impact of TL on melter performance is an important finding because without the current TL constraint, the HLW glass volume at H d o r d can significantly decrease. INTRODUCTION One main risk for the continuous operation of high-level waste (HLW) glass melters is the accumulation of solid phases, such as noble metals, spinel, eskolaite, or zirconiumcontaining minerals. To lower this risk, HLW glass is formulated with a constrained liquidus temperature (TA).'An unfortunate consequence of this constraint is to limit the waste loading in the glass, leading to a high volume of waste glass and, in turn,to high capital, production, and disposal costs. Optimizing glass composition to achieve maximum waste loading compatible with glass property constraints' has shown that without the TLconstraint, the waste-glass volume could decrease by 12 to 16%, which for Hanford represents substantial saving. The TAconstraint is based on the assumption that if the estimated glass temperature at the bottom of the melter is higher than TL,spinel or other crystals (except noble metals) would not be present within the melt, and thus the only problem to deal with would be the settling of noble metals. However, the melt temperature is as low as 850°C in the melting Current address: Institute of Inorganic Chemistry, Academy of Sciences of the Czech Republic, V HoleSoviEkhch 41, 18000 Prague 8, Czech Republic.

a



133

zone below the cold cap where spinel or other crystalline phases are likely to exist. It is the commonly held (tacit) assumption that these crystals dissolve as soon as the melt reaches a temperature above the TL. This assumption was checked with the Glass Service-Glass Furnace Model (GSGFM), a mathematical model originally designed for commercial glass making? This model was augmented by adding to it an algorithm for predicting the spinel distribution and accumulationrate in the melter?-5 The algorithm used to augment the model is based on extensive experimental studies6-12conducted fiom 1998 to 2001 to measure the parameters that describe spinel formation and settling in molten glass. This paper summarizes the outcome of the spinel-distributionmodeling and discusses in greater detail the consequences for HLW glass vitrification. EXPERIMENTAL STUDIES Experimental studies described in detail in previously published papers6-12were all conducted with MS-7 glass (Table I), an 11-componet generic HLW glass that formed spinel as the primary crystalline phase. The physical properties of the glass that were needed for mathematical modeling (density, viscosity, and electrical conductivity) were measured as functions of temperature6(Table 11). To measure the density of spinel, spinel crystals were isolated fiom the glass by digesting the glass phase in an acid. The methods used to obtain phase equilibria, the kinetics of nucleation, growth and dissolution of spinel crystals, the settling rate, and the properties of the sludge layer of spinel in MS-7 glass are briefly described below. Oxide A1203 B2O3 Cr2O3 Fe203

Mass Fraction 0.0800 0.0700 0.0030 0.1150

Oxide Li20 MgO MnO Na2O

Mass Fraction 0.0454 0.0060 0.0050 0.1530

Oxide NiO SiO2 Zr02

Mass Fraction 0.0095 0.4531 0.0600

Table 11. Properties of MS-7 Glada) Property Equation or Value Viscosity exp(-12.3 +19723/T) Electrical Conductivio exp(6.97-2914/(T-466)) SpecificHeat Melt Density Spinel Density

Pg

kg-m"

2722.65-0.2077T

kg.mm3

5140

Liquidus Temperature Spinel Equilibrium 0.04334{ l-exp[-5110.7(l/T-l/T')]~ Volume Fraction Mass Transfer Coefficient !Data validity range is 1i.om 850°C to 1200°C;(b)Tis in K

I

134


The equilibrium fraction of spinel in MS-7glass was measured using quantitative Xray diffraction (XRD) as a function of temperature, glass composition (adding to or removing components from the MS-7baseline), and the partial pressure of oxygen (p02)?-~ The model has so far been applied only to data obtained from glass equilibrated with air. Thepoz field will be added to the model in the fbture because it affects both foaming and spinel precipitation in the melter. The main focus of the experimental study was on kinetics, that is, the measurements of the rates of nucleation, crystal growth, and crystal di~solution.6~~ It was discovered early on that spinel nucleation proceeds nearly instantaneou~ly.'~Moreover, the feed melting studies revealed that spinel crystals precipitate already from the primary (nitrate) melt.* These low-chromium crystals dissolved as the conversion progressed, but could also survive as seeds for crystals that form in the glass. The rate of nucleation was measured by counting the number of crystals per unit volume of glass. For the MS-7 baseline glass, the number of nuclei was low at the glassprocessing-temperature range, and strongly increased with decreasing temperature. The number density at the glass-processing temperature interval increased up to 4 orders of magnitude when nucleation agents were added to the glass as minor components. The most effective agents were noble metals! The addition of noble metals resulted in at least a tenfold decrease in the crystal size, and hence at least a hundredfold decrease in the rate of crystal settling. The rate of crystal growth and dissolution was determined by measuring crystal size in thin sections of quenched glass samples that were isothermally heat treated? A special methodology was developed for the settling-rate study to eliminate all possible sources of convection, such as bubbles and surface forces." The measured Stokes coefficient was only slightly different from that for hindered settling" of cubic particles in room-temperature liquids. Finally, the density of spinel sludge was determined both in laboratory crucibles and in a sample taken from a pilot-scale melter.12 For the purpose of mathematical modeling, it was necessary to formulate constitutive equations for spinel-glass-mixture equilibrium and kinetic behavior as h c t i o n s of melt temperature. The rate of spinel crystal growth and dissolution was represented by the Hixson-Crowell equation7in the form

da -=kH(Co -C) dt where a is the crystal size, COis the equilibrium solid-spine1 mass fraction, C is the solidspinel mass fraction, and kH is the mass-transfer coefficient. The mass-transfer coefficient varied with temperature according to the Arrhenius relationship shown in Table 11. MATHEMATICAL MODELING The GS-GFM code was applied to a HLW glass melter (the West Valley type) (Figure 1). The impact of the growing sludge layer on the macroscopic melt flow was omitted for simplicity. The distributions of glass velocities and temperatures in the melter were calculated using a 3-D mathematical model of glass flow, heat transfer, and Joulean heat generation in the melter space coupled through temperature- and compositiondependent properties of the glass. The procedure of control volumes was used. The model


135

involved the kinetics of the growth and dissolution of spinel crystals, the class representation of spinel crystal-size distribution, and the algorithms for the moving and settling of spinel crystals inside the melter.

Figure 1. Stationary temperature and velocity fields in the cross section of a HLW melter Crystals were allowed to grow or dissolve according to Equation (1) and settle following the modified Stokes law. To represent the distribution of crystal sizes in each control volume, 20 classes of crystal sizes were used in calculations, described by their center (the mean size of the i-th size class, ai) and by their equivalent width (wi); thus,

C;=,wi=a,,.

x:=l

The volume fraction of crystals in the control volume was

C= Nv,a,?.The mass balance of the crystalline phase was used for each size class in the form:

where R(Nv,) is the interchange among classes due to size change, si is the volumetric generation of spinel crystals by nucleation, and w, = 0.2O5g(ps - pg)azrj’ is the Stokes velocity of spinel crystals in glass; here g is the gravity acceleration, TJ is the dynamic viscosity of the melt, and p, and pgare the densities of spinel crystals and glass. The thickness of the sludge layer on the melter bottom and the slant melter walls was calculated from the mass balance of spinel crystals in the regions adjacent to the bottom and from the measured concentration of spinel in the sludge. The thin-layer approximation was used. In each control volume above the bottom, the i-th size height, hi, was calculated for the time step At as hi = A t w s i . All crystals within hi fall fast enough to settle on the bottom during At. The settled fraction of crystals in the control volume is si = h i / k where Az is the height of the control volume. The height of the sludge layer is n

h =( k / ~ ) ~ s i N V , a , ? i=l

(3)

where V, is the volume fraction of spinel crystals in the sludge.

136


Monodispersed spinel crystals entered the melt from the cold cap. In the reference case, the input crystals concentration was Cjpj = 110 kg/m3 and their size was aifl= 1 pm. The nucleation density of spinel was ns = exp(13.6224.01823T). The number of spinel crystals moving from one control volume to another in the region with T .c Tt was increased to match ns if the next control volume was cooler than ''7 ,and the current number of particles was lower than n,. New nuclei were generated in the size class of the smallest crystals. For each time step, the model performed the following sequence of operations: 1) Load the temperature and velocity fields from the GS-GFM. 2) Start crystal nucleation. 3) Obtain w,for each size class. 4) Account for growth and dissolution: da dt

[CO

= k(T)

-tN,.,a; i=l

1

(4)

5 ) Redistribute crystals into unified size classes. 6) Calculate crystal settling in control volumes above the bottom and slant walls. 7) Repeat Steps 2 through 6 to reach a stationary state.

RESULTS AND DISCUSSION Figure 1 displays the stationary temperature and velocity fields in the cross section through the melting space calculated by the GS-GFM. The melter produced a rather vigorous circulation flow, assuring a good mixing and homogenizing and a nearly uniform temperature field. The average temperature of the molten glass was 1104°C. The average glass velocity was approximately one order of magnitude higher compared to standard industrial glass fiunaces.

Figure 2. Spine1 concentration distribution


137

The distribution of spinel crystals in the longitudinal and cross section for the calculated reference case is displayed in Figure 2. The fast flow of the glass melt resulted in a nearly uniform distribution of crystals in the melting space? Hence, the melter behaved as a fast (nearly ideal) mixer. The rate of growth of the sludge layer was computed as a function of the position in the melter? The sludge-layer thickness profile at the melter bottom after 30 h of melter performance is shown for the reference case in Figure 3. Note the increased layer thickness towards the melter outflow. The sludge layer grew slowly, only 0.04 &year. Sludge that thin is not expected to interfere with melter operation. Figure 4 presents the result of a parametric study conducted to investigate the influence of the input crystal size and the 'T on the sludge-layer growth. The dependence of the sludge-layer growth rate on the input size of spinel crystals has a power-law form with the exponent of 2.5 (Figure 4, left plot), which is higher than that which the Stokes equation would suggest. When the initial crystal size increased from 1 pm to 5 pm, the sludge-layer Figure 3. The thickness profile of deposited spinel Crystals rate increased from (the sludge layer) at the bottom after 30 h of HLW melter 0.04 to 2 performance (reference case) (for TL = 1078°C). Crystals larger than 10 pm produced a sludge layer several cm thick after one year of melter performance. With continuous operation, > 10-pm crystals would gradually obstruct the melter outflow. The right plot in Figure 4 shows that the sludge-layer thickness mildly grew with increasing . ' T All calculated data with the crystal size interval from 1 to 100 pm and'T interval from 950 to 1200°C can be approximated as

where v h = &/dt is the sludge-layer thickness growth rate, vho = 27.2 d y e a r , a0 is the initial crystal size, a0 = 100 pm, p = 6.30, and q = -2.96~10"I?. As this equation suggests (see also Figure 4), the Stokes exponent is a fimction of 7 ' . This function is approximated as linear in Equation (9,but is slightly nonlinear as the lack of fit of the 10pm line (in Figure 4) reveals. When a0 = 100 pm, vh is independent of . '7 The Stokes exponent increases as the TLdecreases, reaching the value of 2 at TL= 1179°C and 2.5 at TL = 1040°C. These results suggest that within broad ranges of variation, the 'T should not cause technological problems by its impact on spinel accumulation if the melter is a fast

138


mixer. The most important parameter for the sludge-layer growth control is the size of spinel crystals. There are a number of additional applications for this model. The model can potenI00 tially be extended to incorporate the oxi10 dation-reduction equilibrium and the concentration and I sizes (including rates of growth) of gas bubbles. Melter 0.1 idling can also be simulated? The 0.01 model can be used to test different I 10 100 melter designs, the input size of spinel crystals [pm] effect of bubbling, melter operation parameters, and the impact of glass composition on spinel settling. Thus, melter design, melter operation, and glass formulation can be optimized. CONCLUSIONS The mathematical model presented m = enables calculating 0.01 I ; , , . , , , , 950 loo0 1050 I100 1150 I200 the distribution of spinel crystal conT' PCI centration within 7 = 1078°C the melt and the Figure 4. The effects of the initial size of crystals at ' (above) and of 7'~ with input crystal size as the parameter (below) evolution of the sludge-layer thick- on the growth rate of the sludge layer of deposited spinel crystals at the HLW melter bottom ness on the HLW glass melter bottom. It shows that fast-mixing melters keep a nearly uniform concentration and size of crystals during normal operation. Hence, keeping TL> 1050°C may not prevent accumulation of crystals in the melter whereas crystals may not settle in the I


I

139

melter even if TL< 1050°C. The accumulation of solids in the melter is primarily determined by the initial crystal size. Nucleation agents (noble metals) keep crystals small in most HLWs, but are often absent in simulated wastes for experimental melter runs, thus exacerbating the potentially false TLconstraints. Without the seemingly unnecessary current TLconstraint, the HLW glass volume at Hanford can significantly decrease, with correspondingly significant cost savings. This model and the conclusions drawn from modeling should be verifiedhalidated with actual melter operation data. ACKNOWLEDGMENT Pacific Northwest National Laboratory is operated for the U.S. Department of Energy (DOE) by Battelle under Contract DE-AC06-76RLO 1830. This work was funded by the DOE through the Environmental Management Science Program. REFERENCES (‘>D.S.Kim and J.D. Vienna, “Influence of Glass property Restrictions on Hanford HLW Glass Volume,” Ceram. Tram 132, 105-115 (2002). (*)P. Schill, Calculation of Three-Dimensional Steady Flows and Temperature Using Multigrid Method,” Proceedings of the International Congress on Glass, Vol 29, 336343, Leningrad, 1989. (3)J. Matya, J. KlouEek, L. NEmec, and M. Trochta, “Spinel settling in HLW melters,” The gh International Conference Proceedings (ICEM’OI), Bruges, Belgium, 2001. (4)J. MatyhS, Description of the Behavior of Multitude Particles in Non-isothermal Convective Melting Space, PhD. Thesis, Laboratory of Inorganic Materials, Prague, Czech Republic, 2001. Schill and M. Trochta, “Advanced Mathematical Modeling of Special Glass Furnaces,” Proceedings of the 2002 GLASS ODYSSEY, 6‘h EGS Conference, Montpelier, 2002. (@P.Hrma and J. Alton, “Dissolution and Growth of Spinel Crystals in a High-Level Waste Glass,” The gh International Conference Proceedings (ICEM’OI), Bruges, Belgium, 2001. (’)J. Alton, T.J. Plaisted, and P. Hrma, ‘‘Spine1 Nucleation and Growth of Spinel Crystals in a Borosilicate Glass” accepted in . I Non-Cryst. Solids. (‘’P. Izak, P. Hrma, B.W. Arey, and T.J. Plaisted, “Effect of Batch Melting, Temperature History, and Minor Component Addition on Spinel Crystallization in High-Level Waste Glass,” J. Non-Cryst. Solids 289,17-29 (2001). (’)P. Hrma, P. Izak, J.D. Vienna, G.M. Irwin and M-L. Thomas, “Partial Molar Liquidus Temperatures of Multivalent Elements in Multicomponent Borosilicate Glass,”Phys. Chem. Glasses 43 (2) 128-136 (2002).

( O)J. Klouzek, J. Alton, T.J. Plaisted, and P. Hrma, “Crucible Study of Spinel Settling in Hip-Level Waste Glass,” Ceram. Tram. 119,301-308 (2001). (l ’E. Barnea, and J. Mizrahi, “A Generalized Approach to the Fluid Dynamics of Particulate Systems, Part 1,” Chem. Eng. J. 5, 171-189 (1973). (12)M.Jiricka and P. Hrma, “Chemical and Mechanical Properties of Spinel Sludge in High-Level Waste Glass,” Ceramic-Silikaty 46 (1), 1-7 (2002) (13)J.G.Reynolds and P. Hrma, “The Kinetics of Spinel Crystallization from a HighLevel Waste Glass,” Mat. Res. Soc. Symp.Proc. 465,261-268 (1997).

140


RUTHENIUM - SPINEL INTERACTION IN A MODEL HIGH-LEVEL WASTE (HLW) GLASS T. M. Willwater', J. V. Crum, S. M. Goodwin, and S. K. Sundaram Pacific Northwest National Laboratory Richland, WA 99352 ABSTRACT Noble metals (for example ruthenium) act as nucleation sites for the precipitation of spinel (crystalline) phases. The noble metals along with the spinel phases will settle to the bottom of the melter causing local viscosity increase, power fluctuations, and even potentially shorting of electrodes leading to premature melter failure. We studied the partitioning of ruthenium in a model high-level waste glass. Ruthenium oxide was chosen as this was predominantly found in melter tests with feeds containing noble metals at the bottom of the melter. A doping of 10 wt % of ruthenium oxide was selected to simulate somewhat the conditions at the bottom of the melter where noble metals accumulate. The heat-treatment conditions (temperature and duration) were chosen from reported literature, such that large crystals of trevorite (NiFe204) were formed in the glass. The spinel-glass interface was characterized using scanning electron microscopy (SEM) and microprobe characterization. SEM results showed the crystals distributed in the glass matrix. Microprobe measured the ruthenium concentration across and around the spinel-glass interfaces. The results did not show significant partition of ruthenium in the spinel. INTRODUCTION Precipitation and settling of noble metals (for example, rhodium, ruthenium, and palladium) in high-level waste (HLW) glass melts processed in joule-heated melters can lead to operational difficulties. In addition, the noble metals act as nucleation sites for the precipitation and growth of spinel (crystalline) phases, which in turn will settle to the bottom of the melter and cause the viscosity of the melt to increase in that region. Summer Intern,Pre-Service Teacher (PST) program, Pima Community College, AZ To the extent authorized under the laws of the United States of America, all copyright interests in this publication are the property of The American Ceramic Society. Any duplication, reproduction, or republication of this publication or any part thereof, without the express written consent of The American Ceramic Society or fee paid to the Copyright Clearance Center, is prohibited.


141

Spine1 forms during the batch-melting reactions [l] in Fe- and Ni-containing HLW glass batch, dissolves the batch melts, precipitates from the HLW glass again as the melt temperature cools below the liquidus temperature, TL [2-41. Once crystals are formed, they tend to settle down in a melter, which could potentially lead to power fluctuations, current excursions, enhanced electrode corrosion, and even shorting of electrodes causing premature melter failure. It is important to understand the solubility of noble metals as well as their partitioning between the crystal phase formed and the glass. Hrma and coworkers [5-81 have extensively studied the formation, dissolution, and growth of spinel crystals in a model borosilicate glass system (MS-7) (base composition as well as compositions with trace amounts of noble metals). These works have used the Hixson-Corwell equation (based on Fick’s Law) to determine mass-transfer coefficients for dissolution and growth and found that these coefficients were found to fit one Arrehenius function of temperature. Three major melter test campaigns testing noble metals have been completed in the past: 1) PNNL test, 2) German melter test, and 3) Integrated DWPF (Defense Waste Processing Facility) Melter System (IDMS). Noble metals have been included in glass development studies since some of the earliest waste solidification and vitrification work at PNNL [9]. The insolubility of noble metals in glasses was observed at those early stages and was also known from the literature; however, the effect this insolubility could have on melter operation was not known. Early works in 1970s included crucible and laboratory-scale tests. Since then, five major studies, gradient furnace testing (GFT), research scale melter (RSM) testing, engineering scale melter (ESM) testing, modeling, and engineering analysis, were completed at PNNL. German melter tests (1980s and 1990s) showed that the accumulation of noble metals could be greatly decreased by increasing the slope of the melter floor. The IDMS was designed as a pilotscale test facility for the DWPF. Before testing with the IDMS, two short-term noble metals campaigns with a 1/1OOth scale mini-melter revealed a need for extended noble metals testing. Numerous test runs with the IDMS melter addressed the designs of the DWPF feed preparation system, offgas system, and the melter itself. The IDMS engineering-scale melter is prototypic of the DWPF melter, designed with a melt surface area of 0.29 m2 (approximately 1/9th of the DWPF surface area), and a melt volume of 0.20 m3. The IDMS has conducted a total of 16 noble metal-related runs with four different types of wastes sludges containing various amounts of noble metals [lO-121. All these melter tests results clearly indicate that the most commonly found species is high concentrations of Ru02 in the melter and Ru has always been found in association with Ru02 and other noble metals and spinels. Interaction of noble metals and spinel crystal at high noble metal concentrations has not been systematically studied. The objective of the present study is to generate preliminary data addressing this issue.

142


EXPERIMENTAL METHODS AND MATERIALS A model borosilicate glass (MS-7) is chosen as this is the most studied composition used for investigation of kinetics of growth of spinel crystals. Trevorite (NiFezO4) is only dominant crystalline phase formed in this composition that is shown in Table I. Table I. Composition of MS-7 glass Oxide Glass Comp, Source wt% chemical

Na20 NiO Si02

8.00 7.00 0.30 1 1.50 4.50 0.60 0.50 15.30 1 .oo 45.30

Total

100.00

A1203 B203 Cr203 Fe203 Li20 MgO

MnO

zro2

6.00

Total

A1203 H3B03 Cr203 Fe203 Li2C03 MgO

8.00 12.43 0.30 1 1.50 11.13

MnO

0.50 41.59 1 .oo 45.30 6.00 138.35

Na2C03 NiO Si02

zroz

0.60

By following the standard glass melting procedure, a 500g sample of the base MS-7 glass was first prepared then milled in an agate mill for 6 minutes and eventually put in the furnace at 1250" C where it was maintained for one hour. The melt was cooled and the glass was milled again in a tungsten carbide mill for 6 minutes. It was then re-melted again at 1250" C for an hour. The cooled glass was ground into a fine powder with the tungsten carbide mill for 4 minutes. At this point, small batches of 10 grams each were prepared. 10 wt% of Ru02 was now added to each of the samples. The samples were then placed in 1 x 1 x 1 cm3 platinum-gold crucibles. Thee samples were then put heat-treated at 1200°C for 30 minutes and the temperature was then decreased to 800°C. The heat treatment conditions were chosen from reported literature [5]. The conditions corresponded to the highest linear spinel growth rate (21.9 pnh) reported in the base MS-7. The samples were then sectioned and polished for further characterization. Xray diffraction 0 ) was used to confirm the primary crystalline phase was trevorite. Optical microscopy showed significant amount of settling of large trevorite crystals at the bottom of the crucible. The spinel-glass interfaces in the sample near the bottom of the crucible region were characterized using scanning


143

electron microscopy (SEM - JEOL JSM-5900) and microprobe (JEOL JXA- 8600 Superprobe).

RESULTS AND DISCUSSION The largest crystals (about 50 pn) were found near the bottom of where the crucible. Figure 1 shows the secondary electron micrograph of a representative glass-spine1 interface in the sample with 1 0 WWOof Ru02 heat-treated at 800°C for 7 hours. Prominent trevorite crystals are surrounded by Ru/RuO2-rich particulates, confinning glass saturated with excess Ru. Figure 2 shows the microprobe data. The inset shows the back-scattered electron image of the same location shown in Figure 1. The burning marks shown in the image are the points where the Ru concentration was measured. The spots are 1 0 ym apart so they do not interfere with the neighboring spots significantly. The Ru concentration is not measurable from points 1 to 5. Then, it starts to increase steadily. The point 11 is close to a crystal as seen in Figure 2. The increase can not be attributed to the crystal as the points 6 - 1 0 are not on the crystal. Additionally, the Z-contrast shows a spongy white Ru-containing phase in this region.

Figure 1. Secondary Electron Micrograph of the Spine1- Glass Interface (1 0 wt% RuO~,800°C,7 hours)

144


Figure 2. Ruthenium Concentration at the Points Marked on the Back-scattered Electron Micrograph (Inset) of the Region shown in Figure 1 Figure 3 shows the secondary electron micrograph of a representative glassspinel interface in the sample with 10 wt% of Ru02 heat-treated at 800°C for 24 hours. The spinel-glass interfaces features similar to the 7 hours sample. The size of crystals has not changed significantly. Figure 4 shows the microprobe data. The inset shows the back-scattered electron image of the same location shown in Figure 3. The Ru concentration is not measurable from points 1 to 3. Then, it starts to increase through points 4-10 with points 5-7 showing not measurable concentration in the crystal seen in the Figure 2. The point 11 is at the other end of the crystal that is close to the region populated by Ru-rich particdates region. The increase can be attributed to dissolution of Ru into the spinel crystal. A systematic evaluation is proposed to determine distribution of Ru as a function of glass chemistry to establish the mechanism of partition of Ru in spinel crystals.


145

Figure 3. Secondary Electron Micrograph of the Spine1- Glass Interface (10 wt% RuO~,800°C, 24 hours)

Figure 4. Ruthenium Concentration at the Points Marked on the Back-scattered Electron Micrograph (Inset) of the Region shown in Figure 3

146


CONCLUSION Preliminary data indicate increasing Ru concentration in spine1 crytal formed at 800°C for 24 hours in MS-7 with 10 wt% of Ru02. Further data generation and analysis will be needed to establish a partitioning mechanism. ACKNOWLEDGEMENTS We acknowledge United States Department of Energy (DOE) - Office of Science for support to TMW under the Community College Initiative (CCI) Program and Royace Aikin, Science Education Specialist and CCI manager, and his Assistant Dale Johns, for all their support. PNNL is operated by Battelle Memorial Institute for the U.S. Department of Energy under contract DE-AC0676RtO 1830. REFERENCES 1. P. I&, P. Hrma, M. J. Schweiger in Nuclear Site Remediation, Editors: P. G. Eller and W. R. Heneman, ACS Symposium Series, 778, p. 314, American Chemical Society, Washington, DC, 2000. 2. P. Hrma and J. D. Vienna in Proceedings of Waste Management 00, Tucson, AS, 2000, CD-ROM. 3. M. Mika, M. J. Schweiger, P. Hrma in ScientiJc Basis for Nuclear Waste Managementz, Editor: 1. R. Triay, 465, p. 71, Materials Research Society, Warrendale, PA, USA, 1997. 4. P. Hrma, J. D. Vienna, J. V. Crum, G. F. Piepel, M. Mike in Scientijk Basis for Nuclear Waste Management ZXI.1, Editors: R. W. Smith and D. W. Shoesmith, 608, p. 67 1, Materials Research Society, Warrendale, PA, USA, 2000. 5. J. Alton, T. Plaisted, P. Hrma, Dissolution and Growth of Spinel Crystals in a Borosilicate Glass, J. Non-Crystal. Soli&, 311,24-35,2002. 6. J. Klouiek, J. Alton, P. Hrma, T. Plaisted in Ceramic Transactions, Editors: D. R. Spearing, G. L. Smith, and R. L. Putnam, 119, p. 301, American Ceramic Society, Westerville, OH, USA, 2001. 7. T. Plaisted, J. Alton, B. Wilson, P. Hrma in Ceramic Transactions, Editors: D. R. Spearing, G. L,Smith, and R. L. Putnam, 119, p. 291, American Ceramic Society, Westerville, OH, USA, 2001. 8. T. Plaisted, F. MO,C. Young, P. h a in Ceramic Transactions, Editors: D. R. Spearing, G. L. Smith, and R. L. Putnam, 119, p. 3 17, American Ceramic Society, Westerville, OH, USA, 2001. 9. S. IS.Sundaram and J. M. Perez, Noble Metals and Spinel Settling in High Level Waste Glass Melters, PNNL-13347, September 2000.


147

10. N. D. Hutson, J. R. Zamecnik, M. E. Smith, D. H. Miller, and J.A. Ritter in Integrated D WPF Melter System (IDMS) Campaign Report: The First Two Noble Metals Operations (U). WSRC-TR-9 1-400, Defense Waste Processing Technology, Savannah River Laboratory, Aiken, SC, 1991. 11. N. D. Hutson in Integrated D WPF Melter System (IDMS) Campaign Report: Hanford Waste Vitrijkation Plant (HWVP) Process Demonstration (v). WSRC-TR-92-0403, Rev. 1,Westinghouse Savannah River Company, Savannah River Technology Center, Aiken, SC, USA, 1992; N.D. Hutson and M. E. Smith, The Behavior and Effects of the Noble Metals in the DWPF Melter System in Proceedings of the High Level Radioactive Waste Management Conference,American Nuclear Society, La Grange Park, Illinois. 1:541-548, 1992. 12. N. D. Hutson in IDMS Task Summary Report Part 1: The Behavior and Efects of the Noble Metals in the DWPF Melter System. WSRC-TR-93-0458, Savannah River Technology Center, Aiken, SC, USA, 1993.

148


Glass Formulation and Testing


INTERIM MODELS DEVELOPED TO PREDICT KEY HANFORD WASTE GLASS PROPERTIES USING COMPOSITION John D. Vienna, Dong-Sang Kim, and Pave1 Hrma Pacific Northwest National Laboratory, Richland, WA 99352 ABSTRACT Over the past several years the amount of waste glass property data available in the open literature has increased markedly. We have compiled the data from over 2000 glass compositions, evaluated the data for consistency, and fit glass property models to portions of this database [13. The properties modeled include normalized releases of boron (rg), sodium (r& and lithium (rLi) from glass exposed to the product consistency test (PCT) [2], liquidus temperature (TL)of glasses in the spine1 and zircon primary phase field, viscosity (q) at 1150°C (q1150) and as a function of temperature (q& and molar volume (V). These models were compared to some of the previously available models and were found to predict the properties of glasses not used in model fitting better and covered broader glass composition regions than the previous ones. This paper summarizes the data collected and the models that resulted from this effort. INTRODUCTION Efforts are being made to increase the efficiency and decrease the cost of vitrifying radioactive waste stored in tanks at U.S. Department of Energy(D0E) waste sites. The compositions of acceptable and processable higklevel waste (HLW) and low-activity waste (LAW) glasses need to be optimized to minimize the wastexorm volume and, hence, save cost. Glass composition and associated properties from glasses tested at Pacific Northwest National Laboratory, West Valley Demonstration Project, Savannah River Technology Center, Vitreous State Laboratory at Catholic University of America, Idaho National Engineering and Environmental Laboratory, and several other institutions were reviewed and compiled into a single database. This database, although not comprehensive, represents a large fraction of data on waste-glass compositions and properties that were available at this tine. The compositions of glasses in this database were converted to mole fiactions of oxides (and elements in the case of TLmodels) using standard techniques. The compositions were screened for applicability to immobilization of Hanford HLW andor LAW. These data were then used to fit the composition parameters or coefficients in glass property models. The models were validated using subsets of the data not used in their development and the validation results and composition ranges of validity were compared to a number of previously reported glass property models including those reported by Hrma et al. [3,4] along with others. Due to space limitations, the



151

composition region of validity for these models were not included in the paper but can be found in the detailed report [13. RESULTS Product Consistency Test Response Normalized ByLi, and Na releases in the PCT are calculated with the formula: N

h [ r j (glrn2>]= C r j i x i i=l

where,j is the element released (B, Li, and Na), i is oxide component, N is number of N

components, xi is the i-th component mole fraction where

xi = 1,and rji are the

i=l

coefficients listed in Table 1. The basis for this model form has been frequently published (see, for example [S]). Also listed in the table are the summary statistics of the model fits including R2 (the fraction of variation of ln[q] accounted for by the model), R2,aj (adjusted for the number of coefficients), R 2 d (the R2that would be calculated for each glass if it were removed from the model, the model fit to remaining glasses), s (root mean square error), number of glasses, and minimum, maximum, and mean of response. Generally, the R2 values are lower than those from models reported earlier. However, these models were better able to predict PCT responses of glasses not used to fit the model than any of the previous models compared. It is not surprisingthat the R2values are low, as, ln[ri] is not generally linear with composition except over small composition regions. Table 1. Coefficients for PCT Response

* For those components not listed and those listed with “---” as a coefficient, the ‘cothers”coefficient should be used.

152


Viscosity The qllso is calculated according to the equation:

where, hi is the i-th component coefficient listed in Table 2. This, linear, model form has been shown to be highly successful in modeling of q with composition. As can be seen by the R2,97% of the variation in ln[qllso]is accounted for by this simple linear approximation. The generally accepted relationship between q and absolute temperature (T) is given by: ln[q] = C+D/(T-To). However, over narrow ranges of T, ln[q] is nearly linear with 1IT (e.g., ln[rlj = A+B/T) and the coefficients describing these temperature effects are known to vary linearly with composition. Since a majority of the qT data is over a suficiently narrow temperature range to be easily approximated with the linear relationship and that relationship contains only two parameters that must be fit to composition, q~is represented by Equation (3) in this study, where Ai and Bi are the coefficients listed in Table 2. This model describes nearly 98% of the variation in data.

Table 2. Coefficients for Viscosity

* LN203 is the combined lanthanides and yttrium oxides.

I'

I

Liquidus Temperature The TL models were developed based on the work of Vienna et al. [q. As a thermodynamic quantity, TL can be related to the state functions according to:


153

-AG~ TL =RlnK

(4)

where A d is the free energy of formation of the crystalline phase from the melt, and K is the reaction constant. In simple systems, such as crystallization of X Y from solution, K is taken as the inverse of the product of X and Y concentrations in solution. However, for crystallization of crystals that are generally solid solutions from multicomponent waste glass melts, the value of K is a significantly more complicated function of composition which includes activities of components in the crystal solid solution and their activities in the melt. As A d i s also a function of composition, the quantity on the right hand side of Equation (4) has been empirically fit to composition. Previous studies [3,4,6] have shown this factor to be linear with composition, having coefficients (3;) similar to ci,hi,Aj, and Bj, discussed above. A TL model which accounts for the effect of component concentrations in the melt on the activities of spinel components in the melt using ion potential (Pi)was developed [q. Using these relationships, models for TL of waste glasses in the spinel ([Fe,Mn,Ni][Fe,Cr,Mn]~O~) and zircon (ZrSiO,) primary phase fields were fit to the appropriate subsets of experimental data. The 7"'s of melts in the spinel primary phase field are calculated with the formula:

where, i represents the electropositive-element components, Pi is the ratio of the i-th component valence to crystal radius reported by Shannon [7J, and Z,tjony @ion, tcov, and Ocov are coefficients reported in Table 3. In this model, components are broken into three groups represented by the three terms in Equation (5). The first group includes the major spinel components minus Fe, the second group includes the alkali and alkaline-earth components, and the last group includes all components not in the first two groups. This model fits the data very well with roughly 90% of the variation in TL explained by the model. Although the previously published model [qshowed slightly better summary statistics, the advantage of this model is a better estimate of Mn effects on TLand a significantly broader composition region of model applicability. Likewise this model was fitted to TL data in the zircon primary phase field according to:

where the only component in the first term is Zr. The coefficients and summary statistics for the TL model in the zircon primary phase field are also listed in Table 3. This model also fits the data well, explaining roughly 87% of the variation in TLdata. Insufficient data was available to fit TL - composition models in other primary phase fields of interest to waste immobilization.

154


Table 3. Coefficients for TL

Molar Volume The database includes density (p) data on glasses at room temperature. In an ideal mixture, the volume of the mixture is given by the sum of partial volumes of the mixture constituents. Clearly glass is not an ideal mixture, however, a model based on volume is more likely to be linear than one based on p. Therefore, the molar volume ( y) is fit to data according to:

v = Cv,x, N

i=l

(7)

where yl. is the partial molar volume of the i-th component in glass. The yl. values are listed in Table 4. Density (p)is then calculated according to:

where Mi is the molecular mass of the i-th component. The experimental data was suEcient to estimate 6 for 18 glass components. However, it is also possible to estimate these volumes using standard ionic radii. Through the use of Shannon’s crystal radii [7], the of all 56 components found in the database were estimated according to:

zi + br; vi = aro3 2


(9)

155

where vi is the apparent yl. per cation in the component, ro is the radius of oxygen, ri is the cation radius, Zj is the cation valence, and a and b are empirically fit parameters. Table 4 lists the resulting yl. values for the 56 component model along with summary statistics of both models. Table 4. Partial Molar Volumes (cm3) . 56-Comp. Vi 1 Component I 18-Comp. V, I 56-Comp. yi 46.149 I 120.000 30.048 18.866 15.214 7.526

ItLi20 Na20

NiO

SiOz

SrO Ti02 zro2 ZnO BeO Bi203

19.943 19.834 12.668 25.316 17.611 17.964 27.081 15.069

CdO

coo

cszo 40.000 156.000 122.250 1 43.OOO ' 47.000 l 28.000 128.800 52.000 3 1.700 i 35.000

R'dj

V'

Rz p' R'adi

pw

365 0.949 0.946 0.921 0.918

365 0.946 0.937 0.9 17 0.902

* - The R' and RLadjvalues were calculated

on both the molar volume (V) and density (p)

bases.

~

CONCLUSIONS A series of models were developed and validated for use in predicting key waste glass properties as functions of composition. These models included models forre, rN,, rLi, q1150, q ~ V/p, , and''7 in the spine1and zircon primary phase fields. The fit statistics of

156


these models suggest that they are roughly as good as previously published models, however, these models cover broader composition regions and were able to better estimate data not used in model fitting. We recommend that these models be applied (within their appropriate composition regions of validity) for the purpose of rough property estimation over relatively broad composition regions. For more precise property estimation in relatively small composition regions, new models should be fitted to data specifically developed in those composition regions. ACKNOWLEDGEMENTS The authors are grateful to Carol Jantzen (SRTC) and Ian Pegg (CUA) for supplying their data for inclusion in the database; Scott Cooley (PNNL), Steve Lambert (NHC), David Peeler (SRTC), Greg Piepel (PNNL), and Jacob Reynolds (WGI-WTP) for careful review of this work and helpful comments; and Bill Holtzscheiter (SRTC) and Ken Gasper (CHG) for programmatic guidance and support. This work was funded by the DOE Office of Science and Technology under the Tanks Focus Area Immobilization Program. Pacific Northwest National Laboratory is operated for the U. S. Department of Energy by Battelle Memorial Institute under Contract DE-AC06-76RLO 1830. REFERENCES [11 JD Vienna, DS Kim, and P Hrma, Database and Interim Glass Property Models for Hanford HL W and LA W Glasses, PNNL- 14060, Pacific Northwest National Laboratory, Richland, WA (2002). [2] ASTM International, Standard Test Methodsfor Determining Chemical Durabiliq of Nuclear, Hazardous, and Mixed Waste Glasses and Multiphase Glass Ceramics: The Product Consistency Test (PClJ,ASTM C 1285-02, West Conshohoken, PA (2002). [3] P Hrma, GF Piepel, JD Vienna, SK Cooley, DS Kim, €URussell, Database and Interim Glass Property Modelsfor Hanford HL W Glasses, PNNL 13573, Pacific Northwest National Laboratory, Richland, WA (200 1). [4] P Hrma, GF Piepel, MJ Schweiger, DE Smith, DS Kim, PE Redgate, JD Vienna, CA LoPresti, DB Simpson, DK Peeler, and MH Langowski, Property/Composition Relationshipsfor Hanford High-Level Waste Glasses Melting at I 15OoC,PNL- 10359, Pacific Northwest Laboratory, Richland, WA (1 994). [5] CM Jantzen, ‘‘ThermodynamicApproach to Glass Corrosion,” in Corrosion of Glass, Ceramics, and Ceramic Superconductors,eds., DE Clark and BK Zoitos, Noyes Publications, Park Ridge, NJ (1992). [6] JD Vienna, P Hrma, JV Crum, and M Mika, “Liquidus Temperature Composition Model for Multi-Component Glasses in the Fe, CryNi, and Mn Spine1 Primary Phase Field,” . I Non-Cryst. Sol.,292: 1-24 (200 1). [7] RD Shannon, “Revised Effective Ionic Radii and Systematic Study of Interatomic Distances in Halides and Chalcogenides,”Acta Cryst. A32:75 1-767 (1976).


157


RELATIONSHIP BETWEEN LIQUDUS TEMPERATURE AND SOLUBILITY Pave1 Hrma and John D.Vienna Pacific Northwest National Laboratory P.O. BOX999, MS: K6-24 Richland, WA 99352 ABSTRACT The literature on high-level waste glass crystallization uses three basic ways of organizing data: 1) solubilities of sparsely soluble glass components are plotted as functions of temperature; 2) liquidus temperature (T') of glass is expressed as a b c t i o n of glass composition; and 3) fractions of crystalline phases at equilibrium with glass are measured as a function of temperature. To make the results mathematically tractable, the response functions are constructed by fitting simple mathematical expressions to data. The relationship between solubility-based and T'-based formulae is discussed. INTRODUCTION It is common in materials science that material properties, such as viscosity or heat conductivity, are represented as functions of thermodynamic state variables, i.e., temperature (9, pressure, and composition. These functional relationships are called response functions. A special class of properties, such as liquidus temperature (TL), describes the state of the material at equilibrium. The purpose of this contribution is to review basic concepts used to characterize the phase behavior of high-level waste (HLW) glasses, including solubility limits and solubility products, and to discuss their relationship to .'7 SOLUBILITY LIMIT The solubility limit of an oxide (component A) in a HLW glass melt at a given temperature is commonly determined by adding component A to the glass until a solid phase appears at equilibrium. However, caution is needed when using this term. If the solid phase that forms on adding A to the mixture is a compound AB of A with component Bywe should more correctly speak about the solubility limit of AB, not A. For example, when Cr2O3 concentration is systematically increased in a HLW glass containing NiO and FezO3, the solid phase that first appears can be eskolaite (Cr2O3) or spinel, a solid solution of chromite (FeCr204) and nichromite (NiCr204) with magnetite (Fe304). When the primary phase is eskolaite, the solubility limit is that of Cr2O3 in that particular glass. If the primary phase is spinel, we can still talk about Cr203 solubility, but the solubility limit is that of spinel. Chromium, which exists in HLW glass in two dominant valences, as Cr(II1) and Cr(VI), brings an additional level of complexity to the phase behavior of HLW glass. As To the extent authorized under the laws of the United States of America, all copyright interests in this publication are the property of The American Ceramic Society. Any duplication, reproduction, or republication of this publication or any part thereof, without the express written consent of The American Ceramic Society or fee paid to the Copyright Clearance Center, is prohibited.


159

Figure 1 shows, in a high-basicity HLW glass with 1.50 mass% Fe203 and 0.15 mass% NiO, the primary phase at temperatures above 1000°C is eskolaite.' Below 1000°C, a liquid chromate precipitates as a separate phase. Spinel forms only at temperatures below 850°C. In the eskolaite region, the Cr203 solubility limit is lowest at 1200°C (2.3 mass%) and increases at temperatures higher or lower than 1200°C. Around lOOO"C, the Cr2O3 fraction dissolved in this glass could be as high as 3.5 mass %. It is advantageous to express 1500 glass composition in terms of ,400 associate species. The concept of glass structure composed of ~ 1 3 ~ associate species, usually identical g12W 2 to crystalline phases, was proved useful in understanding phase k 1 1 ~ behavio? and has been ,!?,m experimentally evidenced by spectroscopic studies? For example, postulating the existence 8oo of dissolved nichromite, we can 0 0.5 1 cr203tAisaCtion &ass%f5 3.5 write Figure 1. Equilibrium phase diagram' for the Cr2O3 in a HLW glass with low content of NiO Cr203(2)+NiO(Z) * NiCr204(Z) (1) and Fe203 where the symbol ( I ) indicates a species dissolved in glass. Precipitation of solid nichromite can be described as a reaction

70-

8 a c1

cu

60 ;,

4-

O 3oIt

s .2 20-

9

160

10;,

*Fe "i '0 o

-

I

~

A oNi+S+ik l ~

precipitate from glass are seen in Figure 2. Spinelforming oxides also participate in a number of other associate species, including acmite (NaFeSi206) that forms


a solid solution, or a segregated liquid. Accepting such hard limits for HLW glass formulation leads to unnecessary low waste loadings. LIQUIDUS TEMPERATURE The TL versus composition function has generally a tractable form within a single primary phase field.' As any mixture property, TLcan be expressed in the form of partial properties, i.e., 6 N

T, = zTLixi

(3)

i=l

where TLj is the i* component partial molar TLand xi is the i* component mole fraction. The T'is are generally functions of composition. Fortunately, the NiO MgO ranges of concentrations of individual components in HLW glasses are usually sufficiently narrow to allow us to approximate Q 1050. TLjS as constants. Thus, the TL hypersurface within each primary phase field is approximated as a flat 950 t hyperplane. An example of the 4.w -0,OZ 0 am 0-04 4ptW=-l3 multiple slopes of such a hyperplane is shown in Figure 3 that displays the effect of a number of glass F i p e 3 . SpinelprimaryphaseTLaSafunction of addition (&j) of oxides to a baseline glass7 components added to or removed fiom MS-7 glass7 (containing, in mass%, 0.3 Cr203, 11.5 Fe2O3, and 0.95 NiO). t

I

i

1

SOLUBILITY PRODUCT Jantzen* suggested estimating TL using free energies of formation of selected crystalline phases. Plodinec' proposed a solubility product model for TL when the primary phase was either an associate species, such as nichromite, or an end-component, such as Gd2O3. Expressing the equilibrium constants of the corresponding reactions, such as (1) and (2), and approximating activities as concentrations, one obtains the relationship xCr,&xNio= A exp(-B / T)

(4)

where A and B are constants for a given glass. At T = TL, xCrzq and xNi0are equal to Crz03 and NiO fractions in the original glass. To compare Equation (4) with (3), we linearize Equation (4) by resolving exp(-B/TL) around a baseline temperature, TB. On neglecting small, higher-order terms (assuming that TL- TBor = 70 mass % waste loading.

1.35

1.32

1.87

1.85

2.69

2.71

2.75

2.77

2.84

2.84

2.86

2.89

2.92

2.97

496 105

495

495 108

498 109

500 113

478 125

463 126

455 129

461 131

464 135

106

1

0

RESULTS AND DISCUSSION The vitrification of this simulated high chrome waste was achieved by simply adding a source of P2O5 to the waste. No other addition was needed. All the wasteforms containing I 65 mass % waste were glassy. Those containing 2 70 mass % waste contained a small amounts ( 4 . 5 mass % ) of crystalline Cr2O3, see Fig. 3. On the basis that the IP65W composition contains a total of 2.6 mass % Cr203, but is free of crystalline Cr2O3, it is concluded that the solubility limit of Cr2O3 in these iron phosphate melts ( for 2 h melting ) is about 2.6 mass %. This is at least 2.6

352


times larger than the amount of Cr2O3 that can be dissolved in borosilicate melts, (solubility of Cr203 in AABS glass is c 1 mass % 16]) . A quantitative estimate for the amount of crystalline C1.203 present in the samples containing > 65% waste was made by comparing the intensity (cps) of the

Fig. 4 Content of Cri03 crystal (b) calculated fiom the calibration curve in (a) of several as-made iron phosphate wasteforms. The amount of Cr203present in the batch for the respective wasteforms is also shown (b).

most intense XRD peak (28 = 24.5") of Cr2O3 with a previously determined calibration curve. This calibration curve was determined fiom XRD measurements of samples prepared by mixing known amounts (1 to 10 mass %) of crystalline Cr203 with powdered of IP65W glass, which was shown to be amorphous by XRD. The concentration of crystalline Cr2O3 in the mixtures was then plotted as a function of the relative intensity of the XRD peak at 24.5" for the mixture compared to that for pure IP65W glass (L /Ig) to obtain the calibration curve shown in Fig. 4 (a). The concentration of crystalline Cr2O3 detennined by this method in the melts containing = or > 60% waste is shown in Fig. 4 (b). The Cr2O3 content in the as-made batch is also shown for comparison. Clearly, the amount of crystalline Cr2O3 in the samples increased with increasing waste loading above 65%, but was only about 1.3% for the IP8OW sample. As shown in Fig. 1, the dissolution rate in DIW at 90 OC for all the iron phosphate wasteforms containing 55 to 80 mass % waste is upto 50 times smaller than that for soda-lime-silica window glass, even though, those wasteforms containing 70 to 80% waste contain near1 20 mass % Na2O. The IP65W wastefkom, had the lowest dissolution rate (5.9 x 10-IT g/cm2/min)or highest chemical durability in water at 90 "C. The presence of a small amount (up to 1.5 mass%) of crystalline Cr203 in these glasses does not appear to adversely affect their chemical durability to any measurable degree. The excellent chemical durability indicated by the DR measurements (Fig. 1) for these iron phosphate glasses is confinned by the VHT (Fig. 2) and PCT (Table 3)


353

results. As shown in Fig. 2, the LP70W wasteform, which contains about 0.8 mass % crystalline Cr2O3, does not show any visible or detectable corrosion layer on its surface after 7 days in DIW at 200 “C.The average total normalized mass release for IP70W (Table 3) fiom PCT is only 1.33 g/m2.For the IP75W sample, which contains about 1.2 mass% crystalline Cr2O3, a thin corrosion layer only 11 pm thick was observed on the surface after the VHT (Fig. 2). Based on this corrosion layer, the corrosion rate for this glass was calculated to be only 3.3 g/m2/day. Corrosion rates of 140 to 196 g/m2/day have been reported [71 for the LAW-33 and LD6-5412 borosilicate glasses. The average total normalized mass release (1.86 g/m2) for the IP75W sample is a little higher than that of IP70W (Table 3), but is still quite low. CONCLUSIONS The present results clearly show that the high chrome (up to 4.5 mass %) waste at Hanford can be vitrified by simply adding about 30 mass % phosphate to the waste and melting the mixture at 1250 OCfor 2 h. The simulated blend of the three high chrome wastes at Hanford used in the present study are about 10 mass % of the total waste at Hanford. The solubilitylimit of Cr203 in these iron phosphate melts is about 2.6 mass%, compared to < lmass % in common borosilicate glasses. Iron phosphate wastefonns having waste loading of 55 to 75% of the high chrome HLW have an exceptionallyhigh chemical durability (as determined by VHT and PCT). ACKNOWLEDGMENT This work was supported by Department of Energy (DOE) under EMSP grant DOE DE-FG07-96ER45618. REFERECES ‘High- Level Waste Melter Study Report, PNNL-13582, July 2001. 2R.A. Kirkbride, “Tank farm contactor operation and Utilization Plan (TWRSOUT)”, HNF-SD-WM-SP-012, Rev.2, CH2M Hill Hanford Group, Inc., Richland Washington, 2000. 3X. Yu, “Properties and Structure of Sodium-Iron Phosphate Glasses,” Journal of Non-Crystalline Solids, 215 21-3 1 (1997) 4PNNL Technical Document, “Vapor Hydration Test Procedure, GDL-VHT” ’ASTM Standard Test Method for Determining Chemical Durability of Nuclear, Hazardous ,and Mixed Waste Glasses: The Product Consistency Test, C 1285-97 6X.Feng et al, “Glass Optimization for Vitrification of Hanford site Low-Level Tank Waste”, PNNL-10918, Pacific Northwest Laboratory, Richland, WA, 1996 7A. Jiricka et al, “The Effect of Experimental Conditions and Evaluation. Techniques on the Alteration of Activity Glasses by Vapor Hydration,” Journal of Non-Crystalline Solids,292 25-43 (2001).

354


DEVELOPMENT OF A SAMPLING METHOD FOR QUALIFICATION OF A CERAMIC HIGH-LEVEL WASTE FORM T. P. O'Holleran Argonne National Laboratory - West P. 0. Box 2528 Idaho Falls, ID 83403-2528

K. J. Bateman Argonne National Laboratory - West P. 0. Box 2528 Idaho Falls, ID 83403-2528

ABSTRACT A ceramic waste form has been developed to immobilize the salt waste stream from electrometallurgicaltreatment of spent nuclear fuel. The ceramic waste form was originally prepared in a hot isostatic press (HIP). Small HIP capsules called witness tubes were used to obtain representative samples of material for process monitoring, waste form qualification, and archiving. Since installation of a fullscale HIP in existing facilities proved impractical, a new fabrication process was developed. This process fabricates waste forms inside a stainless steel container using a conventional h a c e . Progress in developing a new method of obtaining representative samples is reported. INTRODUCTION Electrometallurgica1 treatment of spent nuclear fuel produces two waste streams: metal fiom cladding hulls and salt from electrorefining. A ceramic waste form has been developed to immobilize the salt waste. A hot isostatic press (HIP) was originally used to prepare the ceramic waste form. Small, easily fabricated HIP capsules called witness tubes were shown to be a practical way to obtain representative samples of cerarnic waste form material for process monitoring, waste form qualification, and archiving.' However, the HIP was found to be impractical for production of full-scale waste forms. A ''pressureless consolidation" process was developed to replace the HIP. This process uses a conventional furnace to fabricate waste forms inside a stainless steel container that becomes part of the waste form. A new method of obtaining representative



355

samples for process monitoring, waste form qualification, and archiving must be developed and integrated into the full-scale production process. The objective of this work is to develop and qualiQ a standardized method for obtaining samples for product consistency testing during production of the pressureless consolidated ceramic waste form. The effort is divided into two phases. During Phase I, candidate sampling methods will be developed using small "laboratoryffscale waste forms to investigate materials interaction issues and develop the methodology. The primary goal of Phase I is to specify the sampling method to be used on full-scale waste forms. A secondary goal is to identiQ a back-up methodology to reduce technical risk. All experiments during Phase I will be performed using non-radioactive materials. During Phase 11, the sampling methodology developed in Phase I will be tested with full-scale production equipment. This testing will be performed in conjunction with process development so that the candidate methodology will emerge from Phase I1 completely integrated with the production process. Phase I1 testing will involve extensive sampling in order to develop the data base necessary to establish the statistical relation between the properties of product consistency samples and the production waste form material. Currently, we are engaged in Phase I of this effort. This paper reports progress to date, and outlines future plans. TECHNICAL APPROACH The Waste Acceptance System Requirements Document (WASRD)2 requires that the Product Consistency Test (PCT)3, process knowledge, or a combination of the two be used to demonstrate that waste forms meet specifications during production. This approach has been adopted for the ceramic waste form produced during electrometallurgical treatment of metallic sodium bonded spent nuclear fuel as described in the Waste Form Compliance Plan4 While waste form qualification during production will rely heavily on process knowledge, some sampling and testing will be conducted on a statistical basis. The samples required are not large (20 - 40 g) however, the size and weight of the production scale waste form (about 0.5 m in diameter and 1 m tall and weighing up to 450 Kg) makes sampling problematic. As the process is currently laid out, there is no room in the hot cell for equipment large enough to obtain samples of waste form material by conventional methods such as cutting or core drilling. Sampling activities could take place at various times during the production process. Basically, the points during the process where sampling activities could occur can be defined based upon whether the waste form material is at processing

356


temperature (hot) or has cooled to ambient temperature (cold). Furthermore, sampling can be defined as a two step process: probing the waste form to isolate a small amount of material, and physically retrieving the material. For example, the Defense Waste Processing Facility obtains samples of high-level waste glass by inserting a small cup into the molten glass pour stream (probing), then removing the cup after collecting enough molten glass for analysis (retrieving).' This sampling activity would be described as hot probing and hot retrieving. On the other hand, the West Valley Demonstration Project obtained glass samples after the glass had cooled by reaching into the canister with remote manipulator (probing) and removing a shard of glass (retrieving)? This sampling activity would be described as cold probing and cold retrieving. Using these concepts, four types of potential sampling activities were defined for the ceramic waste form production process as shown in Table I. Table I. Potential sampling activities defined in terms of the waste form temperature at each of the sampling steps, along with some waste form material properties that could be of interest Probe Retrieve Material Properties of Interest Hot Hot Viscosity, Rheology Hot Cold Viscosity, Rheology, Chemical (interactions), Mechanical Cold Hot Viscosity, Rheology Cold Cold Chemical (interactions), Mechanical EXPERIMENTAL To test sampling methods, experimental waste forms were produced from a 3/1 (by weight) ratio of salt-occluded zeolite A to borosilicate glass frit. The salt was a eutectic mixture of LiCl and KC1, containing simulated (non-radioactive) fission product salts. This mixture of powders was placed into a 500 ml stainless steel beaker. A stainless steel weight slightly smaller in diameter than the inside of the beaker and 4.5 cm thick was placed on top of the powder charge to provide some pressure to assist in consolidation. The beaker was placed into a pot furnace, heated to 915' C, and held for six hours. Three of the potential sampling activities described in Table I have been tested so far. A hot probe - hot retrieve sampling method was devised based on a method for obtaining soil sample^.^ In this method, the hot probe step involves removing the steel weight and inserting a thin-walled stainless steel tube into the waste form material at the end of the heat cycle while still at maximum processing


357

temperature. The steel tube had an outside diameter of 1.9 cm, an inside diameter of 1.7 cm, with a 60"taper on the outside of one end. The outside of the tube was coated with boron nitride mold release. The hot retrieve step is removing the stainless steel tube containing a hot sample of material. Two variations of hot probe - cold retrieve methods were attempted. The first was simply a variation of the hot probe - hot retrieve method described above, where the stainless steel tube was not removed until after the waste form had cooled. In the second method, a cavity was drilled into the underside of the stainless steel weight. This was to allow material to flow into the cavity during the heat cycle (the hot probe step). The resulting protrusion of waste form material was to be mechanically removed after the waste form had cooled and the weight was removed (the cold retrieve step). For the cold probe - cold retrieve method, a hole 1.6 cm in diameter was drilled completely through the stainless steel weight and coated with boron nitride mold release. After the stainless steel beaker was filled with starting material and the modified weight placed on top, the hole was filled about half way with additional starting material (the cold probe step). A steel rod the same length as the thickness of the weight and slightly smaller in diameter than the hole, coated with boron nitride mold release, was then inserted into the hole. The purpose of the rod was to apply the same pressure to the potential sample as was applied to the bulk of the waste form material. When the waste form had cooled after the heat cycle, the resulting protrusion was to be mechanically removed (the cold retrieve step).

RESULTS AND DISCUSSION The hot probe - hot retrieve sampling method using a stainless steel tube failed to produce a sample. The tube only penetrated the waste form material about a centimeter, and only with great difficulty. The tube was easily removed, but no waste form material remained in the tube. This method was abandoned after several attempts. The hot probe - cold retrieve method using the stainless steel tube also failed to produce a sample. After cooling, the tube was adhering to the bulk waste form, but it was easily broken free. However, no waste form material was retained in the tube. A curious ring structure remained in the annular depression in the waste form left by the tube. Scanning electron microscopy with energy dispersive X-ray spectroscopy revealed that this ring structure consisted essentially of a layer of oxidized stainless steel containing remnants of the boron nitride mold release (see Figure 1). Iron and chromium were also found to have diffused about ten to

358


twenty microns into the waste form. These results indicate that the boron nitride mold release was ineffective in preventing the steel fiom adhering to the waste form. Separation occurred by mechanical failure within the oxide layer that had formed on the surface of the stainless steel.

Figure 1. Back scattered electron image of the outer portion of the ring structure left behind when the stainless steel tube was removed from the waste form, showing the boron nitride and oxidized stainless steel layers. The second hot probe - cold retrieve method using the modified weight failed to produce a useful sample. Only a small amount of material penetrated into the cavity during the heat cycle. This material broke fiee and remained in the cavity when the weight was removed, but was easily dislodged from the cavity. The waste form material that was retrieved was visibly more porous than the rest of the waste form, suggesting poor consolidation .from lack of pressure in the immediate vicinity. The viscosity of the mixture of molten glass and salt occluded zeolite (or sodalite after the phase transition) at processing temperature is apparently too high to allow the amount of flow needed for this method. The cold probe - cold retrieve method succeeded in producing a sample. When the modified weight was removed from the waste form, the protruding waste form material broke off fiom the bulk waste form, and was retained in the cavity. The sample was removed in one piece by tapping the steel rod to drive the sample out of the cavity.


359

In order to be useful for waste form qualification, the sample material must be representative of the bulk waste form material. Properties need not be identical, as long as differences are consistent and preferably small. Measurements of the heights of the weight and the steel rod were made before and after heat treatment to compare the consolidation of the bulk waste form and the sample. If a “consolidation factor” is defined as the ratio of the green height to the fired height, then the bulk waste form achieved a consolidation factor of 1.83, compared to a consolidation factor of 1.77 for the sample. This suggests that the sample material did not achieve quite the fired density of the bulk waste form. Density measured by helium pycnometry (which does not measure open cell porosity) confirmed that the sample material was slightly less dense than the bulk waste form material. The bulk material had a density of 2.25 g/cm3, while the sample had a density of 2.17 g/cm3. Both the consolidation factor and the density of the sample material are only 3% less than the corresponding values measured for the bulk waste form material, which is very near the uncertainty in the measurement and therefore considered acceptable. These slight differences may be attributable to friction between the steel rod and the cavity walls that reduces the effective pressure applied to the sample material by the steel rod. If so, the small differences between the sample material and the bulk waste form material could presumably be eliminated by simply lengthening the steel rod. X-ray powder diffkaction was also performed to compare the phase composition of the sample to the phase composition of the bulk material. The results showed sodalite as the primary crystalline phase, with halite and nepheline as minor phases for both materials. This is the expected phase composition of the ceramic waste form. Most importantly, the phase compositions of the sample and bulk materials are virtually identical as shown in Figure 2.

360


10

20

30 40 Degrees 2 theta

50

60

Figure 2. X-ray powder diffraction patterns of (top) bulk waste form material, and (bottom) sample material. CONCLUSIONS A sampling method has been demonstrated that can be used to obtain waste form qualification samples of ceramic waste form during production. The method is fairly simple, requires no large scale equipment, and should have little impact on the overall process. Samples obtained by this method axe representative of the bulk material as determined by density and phase composition. Thus data obtained from such samples will be acceptable for waste form qualification and process verification. FUTURE PLANS Experiments with the cold probe - cold retrieve sampling method will continue in order to generate additional material for testing and characterization. The Product Consistency Test3will be used to compare leach behavior of test material and bulk material. The use of different materials for the steel weight will also be investigated. Phase I1 activities will begin when full scale production equipment becomes available for testing.


361

ACKNOWLEDGEMENTS Argonne National Laboratory is operated for the U. S. Department of Energy by the University of Chicago. This work was supported by the Department of Energy, Nuclear Energy Research and Development Program, under contract no. W-31-109-ENG-38. The authors wish to thank Mr. T. DiSanto, Mr. E. A. Reseigh, and Dr. S. M. Frank for their assistance.

REFERENCES ‘T. P. O’Holleran, S. G . Johnson, and K. J. Bateman, “Ceramic Waste Form Qualification Using Results fiom Witness Tubes,” Radioactive Waste Management and Environmental Restoration, to be published. 2U.S. Department of Energy, “Waste Acceptance System Requirements Document,” DOERW-035 1, Revision 03, DOC ID: E00000000-00S11-170800001 REV 03 (1999). 3American Society for Testing and Materials, “Test Methods for Determining Chemical Durability of Nuclear Waste Glasses: The Product Consistency Test (PCT),” C1285-97, Annual Book of ASTM Standards, 12.01 (1998). 4t’Waste Form Compliance Plan for the Waste Forms Jiom Electrometallurgical Treatment of Spent Nuclear Fuel,” Argonne National Laboratory - West Document No. F0000-0031-ES, REV. 00 (1999). 5N.E. Bibler, J. W. Ray, T. R. Fellinger, 0. B. Hodoh, R. S. Beck, and 0. G. Lien, “Characterizationof the Radioactive Glass Currently Being Produced by the D WPF at Savannah River Site,” Waste Management ’98 Proceedings (1998). 6V. A. DesCamp and C. L. McMahon, “VitrificationFacility at the West Valley Demonstration Project,” Topical Report DoElNE144139-77 (1996). 7American Society for Testing and Materials, “Standard Practice for ThinWalled Tube Sampling of Soils for Geotechnical Purposes,” D 1587-00, Annual Book of ASTM Standards, 04.08 (2001).

362


MICROWAVE HEATING FOR PRODUCTION OF A GLASS BONDED CERAMIC HIGH-LEVEL WASTE FORM T. P. O'Holleran Argonne National Laboratory - West P. 0. Box 2528 Idaho Falls, ID 83403-2528 ABSTRACT Argonne National Laboratory has developed a ceramic waste form to immobilize the salt waste fiom electrometallurgical treatment of spent nuclear fuel. The process is being scaled up to produce bodies of 100 Kg or greater. With conventional heating, heat transfer through the starting powder mixture necessitates long process times. Coupling of 2.45 GHz radiation to the starting powders has been demonstrated. The radiation couples most strongly to the salt occluded zeolite powder. The results of these experiments suggest that this ceramic waste form could be produced using microwave heating alone, or by using microwave heating to augment conventional heating. INTRODUCTION During much of the ceramic waste form development effort, a hot isostatic press (HIP) was used to consolidate the powder starting materials. The HIP applies heat and pressure to melt the glass binder and consolidate the powder into a dense solid body. This fabrication route was necessary when the desired end product was a glass-bonded zeolite. Processing temperatures had to be kept relatively low to avoid transforming the zeolite to sodalite, with concurrent release of excess salt. The low processing temperature required the use of pressure to achieve densification. With the selection of glass-bonded sodalite as the frnal waste form (which necessitated a reduction in salt loading), higher processing temperatures could be used for consolidation. A t . the higher processing temperatures used to fabricate the glass-bonded sodalite waste form, densification could be achieved without the application of pressure. The current baseline



363

process for fabricating the glass-bonded sodalite waste form achieves densification using heat only, and is called "pressureless consolidation." As the size of the waste form is scaled up from laboratory scale (on the order of a few centimeters in diameter) to full scale (about 1/2 meter in diameter), heating material in the center of the powder charge becomes more difficult. Thermal conductivity through the loose powder is relatively low, so as the size of the waste form increases, processing time must also be increased to fully densify the material. In the HIP process, this problem was partially alleviated by the application of pressure, since pressure drives densification, and thermal conductivity increases with density. However, since the baseline process for fabricating the glass-bonded sodalite waste form relies solely on temperature to achieve densification, processing times for the full-scale waste form may become excessive. Commercially available microwave ovens operating at 2.45 GHz and at power levels fiom 450 to 850 W have been used to heat zeolites and other aluminosilicates.1s2One advantage of microwave heating is that heat is evolved within the load as microwave energy penetrating the material is absorbed. This results in rapid heating of the load. At 600 W, complete melting of a 10 g, 2.5 cm diameter pellet of Linde 4A was achieved in less than 2 min.' It has been proposed that the initial heating of zeolite 4A (below about 400" C) depends on the degree of hydration, and that dehydrated zeolite could be difficult to heat with microwave radiation alone? The zeolite material used to fabricate the glassbonded sodalite waste form contains essentially no water (< 0.5 wt. %), but does contain approximately 2.5 molecules of occluded chloride salt per pseudo unit cell. The starting material for the glass-bonded sodalite waste form also contains 25 wt. % borosilicate glass. The microwave heating behavior of these materials, alone or in combination, has never been reported in the open literature. The objectives of this work were therefore to determine whether microwave energy would couple sufficiently with the starting material for the glass-bonded sodalite waste form to cause heating, and, if so, to determine if microwave heating could be applied to a production process. EXPERIMENTAL In order to heat materials to high temperatures in a conventional microwave oven, thermal energy generated within the load must not be allowed to escape freely into the microwave cavity. For these experiments, an insulating chamber with internal dimensions 5 cm by 5 cm by 7.5 cm high was constructed from 2.5 cm thick Zircar@ECO-1200B refractory insulating board. When inserted, the 50

364


ml high purity alumina crucibles used for these experiments nearly fill this

chamber. Experiments were performed in two phases. In phase 1, the objective was simply to determine if glass-bonded sodalite starting materials would couple to a microwave field efficiently enough to achieve high temperatures. An uninstrumented commercial microwave oven (CEM model MDS8 1D) with a nominal power output of 850 W at 2.45 GHz was used to qualitatively evaluate the coupling efficiency of the glass-bonded sodalite waste form starting materials In these experiments, separately and as the standard starting mixture. incandescent light escaping through joints in the insulating chamber served as an indicator that the load had reached high temperature. The elapsed time from application of microwave power to observation of incandescent light was used as a relative measure of coupling efficiency. The objective of phase 2 experiments was to quantify the thermal response of the glass-bonded sodalite waste form starting material to a microwave field to allow assessment of potential production applications. For these experiments, a commercial Magic Chef model MCD990B with a nominal power output of 900 W at 2.45 GHz was modified to accept a metal sheathed, ungrounded type K thermocouple. A small hole was drilled through the roof of the oven to allow insertion of the thermocouple, and a corresponding hole was drilled through the roof of the insulating chamber so that the thermocouple could be inserted into the center of the load. The materials used in these experiments were a dehydrated, salt occluded zeolite 4A from UOP (Des Plaines IL),and a borosilicate glass frit from Pemco Corp. (Baltimore MD). Both materials were in powder form, with a nominal particle size of -60+200 mesh. The composition of the glass is given in Table I.

~~

Table I: Composition (as oxides) of the glass frit used to make the glass-bonded sodalite waste form Compound Weight Percent Si02 66.5 B203 19.1 A1203 6.8 Na20 7.1 K?O 0.5 ~~

Salt occluded zeolite 4A was prepared by first drying zeolite 4A at 550" C under vacuum, then loading the dried zeolite with simulated (non-radioactive)


365

electrorefiner salt (8.33/1 zeolite to salt mass ratio) at 500" C in a heated V-mixer. The composition of the salt is shown in Table 11. Table 11: Composition of the salt used to make the salt occluded zeolite used in these experiments Wt. % Salt Salt Wt. % LiCl-KC1 eutectic 69.7 BaC12 1.02 14.9 LaC13 1.22 NaCl KBr 2.3 X 10'2 CeCI3 2.33 RbCl 0.33 PrC13 1.15 src1* 1.01 NdC13 3.89 YC13 0.70 PmCI3 0.11 KI 0.15 SmCI3 0.69 CSCl 2.50 EuC13 4.71 X 10-2 RESULTS AND DISCUSSION Phase 1 Experiments In the first phase 1 experiment, 14.8 g of salt occluded zeolite 4A was loaded into the crucible, filling it about half way. Power was switched on, and incandescence was observed after 225 s. Power was immediately switched off. After cooling, the crucible was removed and examined. Most of the powder appeared unaffected, but a region in the center was cracked and seemed to have begun to sinter. There was a hollow space below this region, at the bottom of which was about a 1 cm piece of material that had melted. The first experiment was repeated with 14.6 g of glass f i t as the load. After three consecutive 10 min runs, the oven was opened and the lid of the insulated enclosure removed to observe the load. The glass powder was quite warm, so another pre-programmed run,this time for 30 min was initiated. Incandescence was observed 503 s into that run. After the crucible had cooled, visual observation showed that approximately half the glass had melted. The same experiment was repeated with 14.8 g of a 311 mixture (by weight) of salt occluded zeolite 4A to glass frit. Incandescence was observed 239 s into the run. After cooling, visual examination revealed that a small portion of the charge had consolidated into a spheroid about 1.5 cm in diameter by 1 cm thick. The results of these experiments are summarized in Table 111. The results of the phase 1 experiments show that both the salt occluded zeolite 4A and the borosilicate glass frit used to make the glass-bonded sodalite waste form can be heated to high temperatures in a microwave field. However, the

366


Table 111. Results of phase 1 microwave heating experiments at a nominal microwave power of 850 W Material Mass of load (g) Time to Incandescence (s) Salt Occluded Zeolite 4A 14.8 225 Borosilicate Glass Frit 14.6 >1800 3/1 (weight) mixture of Salt 14.8 239 Occluded Zeolite to Glass salt occluded zeolite clearly couples more efficiently to the microwave field than the glass. As can be seen from Table 111, the thermal response of the 3/1 (by weight) mixture of salt occluded zeolite to glass used to make the glass-bonded sodalite waste form closely resembles the thermal response of the pure zeolite. This implies that at least initially the salt occluded zeolite component is performing the energy conversion function that heats the entire mixture. The rapid onset of incandescence in both materials is typical of the phenomenon known as thermal runaway, that has been widely reported in the literature?J94This phenomenon can seriously limit the use of microwave heating for production applications. This is especially true for the glass-bonded sodalite waste form, because melting causes the radionuclide-bearing salt to phase separate into halide inclusions that are readily soluble in water. So, while the phase 1 experiments showed that glass-bonded sodalite waste form starting materials can be heated to high temperatures using microwave radiation, the question of whether microwave heating could be used in waste form production remained unanswered. Phase 2 Experiments Some way to control or avoid thermal runaway is necessary to use microwave heating in the glass-bonded sodalite waste form production process. The phase 2 experiments addressed this problem by first quantiQing the thermal response of the starting mixture to identify the onset of thermal runaway, then testing the microwave duty cycle (power setting) as a means of controlling the temperature of the load. A type K thermocouple was inserted approximately into the center of 22.68 g of the salt occluded zeolite 4Nglass baseline mixture for m h g the glass-bonded sodalite waste form. The first experiment was run logging temperatures every 10 s. After 70 s, the temperature jumped from about 400" C to nearly 1200' C, indicating the onset of thermal runaway (see Figure 1). Power was immediately shut off, and the load allowed to cool. While cooling, several power settings above


367

and below 7 were tested. Below 7, the load continued to cool, and above 7 temperature increased to thermal runaway. run - power setting 10 -Re-start

-Initial

-

2

1400 1200 loo0

8

800 600 400

B#

F

- power setting 10

200

0 0

20

40 60 80 Elapsed Time (sec)

100

120

Figure 1: Plot of temperature vs. time for the two full power thermal runaway events encountered during the phase 2 experiments. The load was allowed to cooled to about 100' C, whereupon heating at full power was re-initiated. Thermal runaway was again encountered, although at a slightly longer elapsed time (see Figure 1). Power was switched off, and when the load cooled to about 600' C, heating was re-initiated at a power setting of 7 (70% duty cycle, about 20 s on and 10 s off). At this setting, the temperature quickly rose to about 950' C, then began to oscillate with the duty cycle. Monitoring the temperature for about three minutes indicated that the load was approaching dynamic equilibrium, with a mean temperature around 900' C (see Figure 2). This heating schedule was continued for ten minutes, whereupon the programmed run was automatically terminated. Heating was immediately re-initiated at a power setting of 7 for a programmed time of 30 min, but the oven shut down automatically in response to an overtemperature protection device after 657 s. The experiment was terminated at that point. After cooling, a spheroid approximately 1 cm in diameter was found loosely attached to the thermocouple. The remainder of the powder was apparently unaffected. The solid piece was easily dislodged from the thermocouple, and sectioned for analysis. It appeared to consist of two distinct layers; a fiiable, poorly consolidated layer on the outside, and a well consolidated core on the inside. Optical and scanning electron microscopy revealed that the inner core

368


consisted of a multi-phase outer layer and an amorphous-looking inner core (see Figure 3).

!

1200

1

1000 -

400

G

--

-

~-

200

-

0

0

50

100

150

200

250

Elapsed Time (sec)

F i m e 2: Plot of temperature vs. Time at 70% duty cycle showing approach to dynamic equilibrium at a temperature of about 900" C.

Figure 3: Optical micrograph showing the multiple layers of the solid body formed in the phase 2 heating experiments. Tick marks on the scale at the bottom are one millimeter apart.


369

Except for excess porosity near the inner glassy core, the outer portion of the well-consolidated material resembles normal glass-bonded sodalite waste form material. X-ray powder diffiaction confirmed the phase composition as resembling the conventionally prepared waste form, except with a bit more nepheline (in this case a thermal decomposition by-product of sodalite). The outer, poorly consolidated layer showed less halite (a by-product of glass/sodalite interactions) and a lower amorphous content than the inner portion. These results are consistent with the visual observations evident in figure 3. CONCLUSIONS Glass-bonded sodalite waste form starting materials, particularly the salt occluded zeolite, effectively couple to microwave radiation resulting in heating. While thermal runaway resulting in undesirable melting is possible, simple duty cycle power modulation appears to give sufficient temperature control to make microwave heating for waste form production feasible. Further testing is required to determine how best to apply microwave heating. For example, microwave heating could be used exclusively for producing glass-bonded sodalite waste forms, or it could be used as a boost in conjunction with conventional heating to accelerate heating of the central portion of full-scale waste forms, thereby reducing processing times. Future experiments are planned to address this question. ACKNOWLEDGEMENTS Argonne National Laboratory is operated for the U. S. Department of Energy by the University of Chicago. This work was supported by the Department of Energy, Nuclear Energy Research and Development Program, under contract no. W-3 1-109-ENG-38. The author wishes to thank Ms.M. L. Adamic and Mr. J. R. Krsul for their assistance. REFERENCES S. Komarneni and R. Roy, “Anomalous Microwave Melting of Zeolites,” Materials Letters, 4 [2] 107-1 10 (1986). T. Ohgushi, K. Ishimaru, and S Komarneni, ”Nepheline and Carnegieite Ceramics from A-Type Zeolites by Microwave Heating,” Journal of the American Ceramic Society,84 [2] 321-327 (2001). T. Ohgushi, S. Komarneni, and A. S. Bhalla, “Mechanism of Microwave Heating of Zeolite A,” Journal of Porous Materials, 8 23-35 (2001). 4B. I. Whittington and N. B. Milestone, “The Microwave Heating of Zeolites,” Zeolites, 12 815-8 18 (1992).

’

370


MORPHOLOGY AND COMPOSITION OF SIMULANT WASTE LOADED POLYMER COMPOSITE-PHASE WERSION, ENCAPSULATION, AND DURABILITY Harry D. Smith, Gary L. Smith, Guanguang (Gordon) Xia Pacific Northwest National Laboratory, 902 Battelle Blvd, Richland, Washington, 99352’ Brim J.J. Zelinski Department of Materials Science and Engineering, University of Arizona, Tucson, Arizona, 85721

ABSTRACT Because of their good physical and chemical durability, relatively high salt loading capacity, and low leachability, sol-gel-derived, organic-inorganic hybrid materials (polycerams) show promise as media that could be used to stabilize high salt wastes. Use of this technique has been hindered by the need for highly volatile and flammable organic solvents in the fabrication process. In an effort to overcome this hinderance, we carried out initial development of an alternative production approach based on an aqueous emulsion technology and a “phase inversion” phenomenon that results in encapsulation of the waste form. Our major interests focused on understanding the phenomena and optimizing fabrication methods to produce a final waste form with excellent waste stabilization characteristics. Scanning electron microscopy was used to obtain the microstructures of the waste forms for understanding the migration, distribution, and encapsulation of the salt in the waste forms. The leaching rate of the salt from a waste form was quantified by means of conductivity measurement. INTRODUCTION Over the past 50 years, large amounts of mixed low-level wastes have been generated at U. S . Department of Energy (DOE) sites and other related industries. Salt-containing wastes are always troublesome for treatment due to the high solubility of salts in water and the possible involvement of a broad range of chemical species. Polymexeramic hybrids (polycerams) have been demonstrated to be promising candidates for encapsulating salt wastes (Smith, et al., 1999) in comparison to the other developed technologies such as vitrification and grout. The need to use organic solvents with high volatility and flammability in the fabrication of these polymer and polyceramic materials has offset their advantages. Developing alternative approaches that employ aqueous emulsion systems in waste

Pacific Northwest National Laboratory is operated for the U. S.Department of Energy by Battelle under Contract DE-AC06-76RL01830. To the extent authorized under the laws of the United States of America, all copyright interests in this publication are the property of The American Ceramic Society. Any duplication, reproduction, or republication of this publication or any part thereof, without the express written consent of The American Ceramic Society or fee paid to the Copyright Clearance Center, is prohibited.


371

form fabrication process to avoid using organic solvents were initiated (Liang, et al., 2002) based on discussions with the University of Arizona. For the initial study, efforts were focused on the identification of a good model system for fabricating durable waste forms using aqueous-based systems. This approach is based on the water/oil like phase inversion concept. For our study, a model aqueous emulsion mixture of polystyrene butadiene and epoxy resin was used. The discontinuous emulsified droplets suspended water may first congeal into a continuous, waste-encapsulating phase and then form a tough and durable waste monolith during a curing process. The model waste form itself possesses tough mechanical strength at moderate salt loading with low leachability, and apparently good chemical durability. However, little has been known about the occurrences and outcomes resulted from the “phase inversion” during the waste form fabrication processes. The objectives of this study were to understand the “phase inversion” concept and related processes, which are believed to be crucial to the development of the final waste forms, silica-incorporated polycerams that will meet land disposal requirements. Scanning electron microscopy was used to obtain detailed information about the microstructure of the waste forms for understanding the migration, distribution, and encapsulation of the salt in the waste forms. The leaching rate of the salt from a waste forms was quantified by means of conductivity measurement. MATERIALS AND EXPERIMENTS A commercially available aqueous emulsion, polystyrene-butadiene (PSB) latex (Styronal ND 656, BASF), and epoxy resin (Epo-Kwick Resin, Buehler) were used as the ingredients for fabricating our model polymeric composites. With the aid of a surfactant (sorbitan monooleate, Aldrich), the PSB latex and epoxy resin were emulsified by vigorously stirring. Waste salt surrogate, sodium nitrate, along with a crosslinking agent diethylenetriamine (DETA, Aldrich) was then mixed with the emulsion thoroughly. After the mixture was cured at 80°C in a glass baker for about two days, a robust waste form was produced. The process is depicted in Figure 1. After curing, waste monolith was rinsed with de-ionized water to remove the salt crust which was observed to form on the free surfaces. This salt was collected and quantitatively determined in order to obtain a precise inventory of the salt associated with the sample. The microstructures of broken or cut surfaces of cured samples were characterized using a JEOL 5900 LV Scanning Electron Microscope (SEM) with a built-in EverhardtThronley secondary electron detector and a Robinson series VI scintillation-based backscattered electron detector (BSE). The local salt distribution of unleached and leached samples was analyzed by energy dispersive spectroscopy (EDS). The salt leaching behavior was determined by measuring the conductivity of the leached solution as a function of time. Typically, a sample of 2-5 mm in thickness was sectioned fkom the monolith and immersed in a known volume of dsionized water for salt leaching tests. The conductivity of the solution vs. time was measured with a conductivity meter and recorded. The amount of salt leached out from the sample at a given time therefore was determined.

RESULTS Before curing, the aqueous mixture of polymer precursors and salt solution was a milk-like emulsion. During the system was heated at 80°C in an oven, polymerization was observed while some water droplets were found on the wall of the beaker. The final

372


Polystyrene-butadiene

Epoxy resin and surfactant

NaNOs/Water

1

Mixed and stirred for 30 min

[

1

Cured at 8OoCfor 2 days

1

Figure 1. A Schematic Flowsheet of Aqueous-based Fabrication Process for Producing Polymeric Waste Forms. waste form was a tough solid with some salt residuals on the monolith surfaces. Figure 2 shows photo images that were taken during different fabrication stages.

Figure 2. Photo Tmages of the Fabrication of Waste Form During Curing Process. A) Emulsified mixture of aqueous polymer precursors and salt solution prior to curing. B) and C) Polymerization taking place during curing. D) Final waste form.


313

On a macroscopic scale, most of samples prepared appear to be homogeneous. However, SEM examination revealed that these samples are inhomogeneous at the microscopic scale. In general, as the waste form cures, the salt was distributed throughout the interior with the highest concentration near the free surfaces. The salt that actually migrated to a free surface formed a salt crust there, as shown in Figure 3. In the matrix of the waste form, the salt particles exist in different forms, such as large or small particles dispersed in the polymer matrix and pockets which encased some well crystallized salt particles, as indicated in Figure 4. The salt Figure 3. Backscattered SEM Image of Near retention percentage (S) vs. time for the half a waste Form Showing disk of a waste form (containing 22 wt% the SurfaceNaN03) is shown in Figure 5.

DISCUSSION AND CONCLUSION

Phase Inversion and Salt Distribution

One of the important goals of this study was to evaluate the extent of the “phase inversion” and document its relationship to the leaching rate of the surrogate salt waste (sodium nitrate). Figure 6 gives a schematic representation of phase inversion process. When emulsified epoxy and PSI3 were thoroughly mixed, the resulting emulsion was stable for several days without visible phase separation even with 30wt% salt present in the aqueous phase. Upon water removal from the aqueous mixture during the curing process, the emulsion is believed to transform from the oil-in-water (O/W) type into a water-in-oil (W/O) type. The phase inversion results in the encapsulation of salt particles within the polymeric matrix. Figure 4 may provide us with the evidence of the occurrence of the phase inversion. The salt crystals are observed to be completely entrained in the matrix or encapsulated in voids.

Figure 4. Backscattered SEM Images of the Interior Potion of a Typical Waste Form. On the left is a cut surface of the waste form. Salt particles are embedded in the matrix. On the right is a fracture surface of the same waste form that showing salt crystals trapped in a void.

374


The phase inversion process is considered to be the key step controlling the microstructure of the waste form, which in turn controls the salt distribution, leachability, strength, and chemical durability. The occurrence of the phase inversion arises fiom the attainment of a critical ratio of organic component to water as water is removed during the curing process. Once the phase inversion and polymerization occur, the further

80

= 4

30 20 10

0 0

50

100

150

200

250

300

350

Time (h)

Figure 5. Salt Retention (wt’Y0) vs. Leaching Time for the Sample (22wt% Salt). About 60 wt% of salt leached out from the sample at a very short leaching period, which corresponded to surface salt on the leaching sample. Only a small amount of salt (-2 wt%) was retained in the sample after about 300 hours. evaporation of water from the waste form surface results in the partial migration of the salt from the interior to the free surfaces of the monolith. Note that during the curing process, the waste salt is precipitating out of the aqueous phase. These precipitates will automatically be entrained in the matrix, while any droplets of waste salt solution that are trapped will form spherical voids with salt crystals in them. The water from those voids probably diffuses through the polymer matrix until it reaches open porosity and evaporates. In fact, the salt crystals were always found on the surfaces of the final products. As seen in Figure 4, the microstructure of the central portion of the sample consists of sac-like structures that may be filled with salt. These structuresare imbedded in the matrix phase, appear to completely encapsulate that salt, and are resistant to leaching.

Salt Retention and Leaching As seen in the Figure 5, the salt retention curve of the sample (22 wt% salt) shows different leaching behaviors. At the beginning of the leaching test, the salt retention drops rapidly, which corresponds to the salt crust or interior salt exposed by the sectioning process. This salt can be washed off very quickly and the quantity of the salt can be calculated from the diffusion curve by the extrapolation of the long time leaching behavior to zero time. The leaching rate of the salt fiom the interior of the sample was


375

Figure 6. A Schematic Representation of Phase Inversion Process During Curing for the Fabrication of an Aqueowbased Polymeric Waste Form. On the left, emulsified polymer precursors (filled circles) suspended in water forms the oilin-water ( O N ) type emulsion. Upon water removal from the aqueous system during curing, the emulsion is expected to transfbrm into a water-in-oil (W/O) type (right). The phase inversion results in the encapsulation of salt particles (empty circles on the right) within the polymeric matrix. relatively slow, as shown in the salt retention curve at the later leaching stage. The fact that almost all the salt in the waste form eventually leached out after over 300 hours leaching test indicates open porosity still exists in the polymeric matrix. The formation of the open porosity is believed due to the water evaporation and salt migration towards the surface of the waste form during the curing process. Obviously, the closing of the open porosity by appropriate methods will help to reduce the salt leachability. Post treatment methods, such as reheating or hot-pressing waste forms may enhance the capability of resistance to salt leaching, ACKNOWLEDGEMENTS The authors thank the Laboratory Directed Research and development (LDRD) program supported by Pacific Northwest National Laboratory (PNNL) We also would like to thank Dr. Willam Kuhn and Mr. Jim Buelt for their advises and Dr. Liang Liang for his early work on this project. REFERENCES 1. G a y L. Smith and Brian J.J. Zelinski, Stabilize High Salt Content Waste Using SolGo1 Process, Innovative Technology -DOE/EM-0473, OST Reference #2036, Mixed Waste Focus Area, Prepared for U.S. Department of Energy, Office of Environmental Management, Ofice of Science and Technology, September 1999.

2. Liang, L., Smith, H., Russell, R., Smith, G. & Zelinski, B. J. J. Aqueous Based Polymeric Materials for Waste Form Applications. In G.L. Smith, S.K. Sundaram, and D.R. Spearing (Eds.), Proceedings of the International Symposium on Environmental Issues and Waste Management Technologies in the Ceramic and Nuclear Industries VII, Westerville, Ohio: Ceramic Transactions, The American Ceramic Society (2002).

376


93NbMAS NMR OF NIOBIUM CONTAINING SILICOTITANATE EXCHANGE MATERIALS Brian R. Cherry', May Nyman2,and Todd M. Alam' 'Department of Organic Materials and 2Departmentof Geochemistry, Sandia National Laboratories, Albuquerque, NM 87 185 USA

ABSTRACT Crystalline silicotitanate (CST), HNa3Ti&i2014*4H20, is a highly selective Cs ion exchanger, making it an attractive material for removal of 13'Cs from nuclear waste solutions. The Cs selectivity can be improved further by replacing a fraction of the framework titanium with niobium to form NbCST. High-speed 93Nb magic angle spinning (MAS) nuclear magnetic resonance (NMR) spectroscopy was utilized to characterize framework changes as a function of Cs loading in a series of Nb-CST materials. Based on these 93NbMAS NMR studies it is argued that the niobium octahedra present in Nb-CST have near uniform Nb-0 bond lengths and are slightly distorted from cubic symmetry. INTRODUCTION Technologies that selectively remove radioactive Cs or Sr from nuclear defense wastes are of great interest to the U.S. Department of Energy (DOE) in that the radionuclides 137Csand "Sr are responsible for the majority of the radioactivity in these waste solutions. The nuclear wastes stored at Hanford, the Savannah River Site (SRS), Oak Ridge National Laboratory (ORNL), West Valley, and Idaho National Engineering and Environmental Laboratory (INEEL) present challenges to present Cs removal technologies because these wastes contain very high concentrations of dissolved salts and may be extremely basic (Hanford, SRS) or acidic (INEEL). Further, proposed Cs removal technologies must be able to withstand high radioactive doses without diminished performance. In the early 199O's, Dosch, Anthony, and Gu at Sandia National Laboratories [1, 21 discovered a new silicotitanate inorganic ion exchanger that selectively sorbs 50 ppm Cs from solutions containing -5 M sodium salts. This material, To the extent authorized under the laws of the United States of America, all copyright interests in this publication are the property of The American Ceramic Society. Any duplication, reproduction, or republication of this publication or any part thereof, without the express written consent of The American Ceramic Society or fee paid to the Copyright Clearance Center, is prohibited.


377

called Crystalline Silicotitanate or CST, shows excellent Cs removal capabilities in both highly acidic and highly basic solutions, and is stable at these extreme pHs, as well as in extreme radioactive environments. Further, substitution of 25% of the framework Ti with Nb provides an approximate fourteen-fold increase in the selectivity for Cs, as measured by the distribution coefficient E,ml/g][3] for Cs over Na.[1, 21 In a cooperative agreement with Sandia, Universal Oil Products (UOP LLC) developed the Nb-substituted CST as a product known as IE910, and a granular form known as UOP IONSIV IE91 lTM. Until August 2000, the addition of Nb to the CST framework was a trade secret, protected by a U.S. patent assigned to Sandia National Laboratories.[4] Therefore, the mechanism for increased Cs selectivity with addition of Nb to the CST framework has never been properly investigated. We are currently using a variety of solid-state N M R techniques to determine structural and compositional changes that result from substitution of N b into the CST framework, and how this affects the Cs selectivity of these sorbents. These investigations are being carried out on a series of CST and Nb-CST materials with variable Cs-loadings.

Figure 1: Powder X-ray diffraction spectra of the CST and Nb-CST materials.

A

10

20

30

40

50

60

2-theta

378


Figure 2: CST/Nb-CST framework structure, Ti@b)O6 octahedra (gray), SiO4 tetrahedra (black), framework Na (spheres).

CST Structure and the Nb-site: The 25% Nb-substituted CST has a composition of approximately HNa2Ti3NbSi2014.4H20. The structure is essentially the same as that reported for the Nb-free CST, ~a3Ti4Si2014*4H20.[5]Powder X-ray diffraction spectra comparing the CST and Nb-CST reveal essentially no major crystallographic differences between the two samples (Figure 1). Work in progress on structural investigations on the same series of CST, Nb-CST and Cs-exchanged CST and Nb-CST show that differences in the framework structure between CST and NbCST are minimal.[6] For the sake of discussion, a view of the CST/Nb-CST framework is shown in Figure 2. The TiO6 octahedra (gray clusters) are arranged in cubane clusters of four edge-sharing octahedra. These clusters are cornerlinked to each other in the z-direction through two octahedra per cluster. Overall, the Ti06 octahedra make up double zig-zag chains along the z-direction. These chains are linked in the x- and y- directions by corner-sharing with SiO4 tetrahedra (black clusters). These Si04tetrahedra form chains in the z-direction by alternating with edge-sharing framework sodium sites (spheres). This structural arrangement gives rise to TiO6 octahedra in which all the oxygen atoms are bridging (i.e. no terminal oxygen atoms), and thus very regular Ti-0 bond lengths. The Ti-0 bond lengths reported for Nb-free CST range from 1.89 - 2.07 A. The Nb disordered over 25% of these framework octahedral sites may be either regular like the Ti06 sites, or they may be distorted, as framework NbO6

-


379

octahedra often exhibit axial distortion with one short axial bond (-1.8 A), one long axial bond (- 2.2-2.4 A) and four regular equatorial bonds (-1.9 - 2.1 A).[7, 81 The purpose of this paper is to characterize the niobium coordination environment with 93NbMAS NMR as a function of Cs loading.

EXPERIMENTAL (1) Material Preparation Synthesis of Nb-Substituted C'stalline Silicotitanate (Nb-CST): Titanium isopropoxide (TIPT, 3.43 g, 12 mmol), tetraethylorthosilicate (TEOS, 3.33 g, 16 mmol) and Nb205 (0.54 g, 4 mmol Nb) were added to 50 ml aqueous NaOH (6.6 g, 165 mmol) solution in a 100 ml Teflon liner to an autoclave Parr reactor. The mixture was stirred for 0.5 hr, and then placed in a 200 "C oven for three days. The resulting product, a white microcrystalline powder was collected by filtration (yield 2.2 g of HNa2Ti3NbSi201404H20;86 % yield based on Ti). A small amount of crystalline byproduct was inevitably formed with the major Nb-CST product. Before analyses of the sample, the byproduct was removed by a two step treatment: 1) the Nb-CST with the byproduct was first exposed to a 1 M aqueous HCl wash for three hours at room temperature, and 2) the Nb-CST with byproduct was exposed to a 1 M NaOH wash for three hours at 40 "C. The first step amorphizes the byproduct, and the second step dissolves the resulting amorphous byproduct.

-

Ion Exchange: A series of Cs-exchanged Nb-CST materials were prepared by ion exchange. The maximum amount of Na in Nb-CST that can be readily exchanged for Cs was approximately 25%. For each ion exchange, 3 g of Nb-CST was combined with 50 ml aqueous CsCl solution, containing the appropriate amount of CsCl to obtain a Cs-exchange Nb-CST sample with 3.8, 6.4, 9.0 and 9.6 wt % cesium. The Nb-CST samples were shook with the CsCl solutions at room temperature for 12 hours, and the Cs-exchanged samples were collected by filtration. Inductively Couple Plasma Mass Spectroscopy (ICP MS) was used for compositional analysis of these Cs-exchanged Nb-CST materials. Powder X-ray hffraction was used to examine phase identification, purity and crystallinity, and thermogravimetric analysis (TGA) was used to determine water and OH content. (2) NMR Anal sis The static z3Nb N M R spectrum of the Cs fiee Nb-CST was obtained on a Bruker Avance6OO at 146.72 MHz with a 4mm broadband probe, a Hahn-Echo sequence, a 25 ps inter-pulse delay, and 120K scan averages. All 93Nb MAS N M R spectra were obtained on a Bruker Avance600 at 146.72 MHz with a 2.5 mm broadband probe. Direct polarization MAS spectra were obtained with sample spinning speeds between 31-33 kHz, using a n/12 93Nbpulse ( d 2 = 3 ps),

380


high power TPPM 'H decoupling, 500 ms recycle delay, 1 MHz SW (681 1 ppm) and 128K scan signal averaging. Spectra were processed using linear rediction of the first 8-12 time domain points to reduce acoustic baseline roll. 93Nb spectra were referenced to the secondary standard NbCls in wet acetonitrile, with the sharp resonance, due to [NbC16]-,assigned to 6 = 0.0 ppm.

RESULTS AND DISCUSSION The 93NbStatic NMR spectrum of the Nb-CST material prior to Cs exchange is shown in Figure 3a. The static spectrum results from a complex mixture of quadrupole (CQ) and chemical shift anisotropy (CSA) interactions, with a CQ on the order of 20 MHz. The relationship between the CQand the CSA is the subject of ongoing investigations. Under high speed MAS conditions, o,= 33 kHz, the 93Nbline width is dramatically reduced, (full width half maximum, FWHM 74,000 Hz to 15,000 Hz) Figure 3a. The high rate of spinning speed is necessary to narrow the resonance plus separate the fjrst spinning side bands from the isotropic resonance. The 93NbMAS N M R spectra for the Nb-CST materials as a function of Cs loading are shown in Figure 3b; the frequency shifts and line widths are listed in Table I. The MAS NMR spectra of all the Nb-CST samples show a single, nearly symmetric resonance. Within experimental error, there is no variation in the 93Nbfrequency shift with increasing Cs loading. The line widths span a range from 13000 to 15600 Hz. It should be noted that resonances with large quadrupolar couplings (CQ > 60 MHz) may not be observable using these standard MAS techniques at the given field strength. Interestingly the overall signal intensity does not vary greatly with increasing Cs loading, supporting the argument that the addition of Cs does not produce a new unobservable resonance, with the corresponding loss of the original peak. To date, there have been only a limited number of solid state 93NbMAS investigations.[9-11] The Nb in these Nb-CST materials have an observed 93NbNMR frequency shift (Table I) that is downfield from the 6 = -900 through -1 100 ppm shifts reported for alkali niobate perovskites, lead niobate pyrochlorates, lead magnesium niobate (PMN) and a range of PMNAead titanate (PT) solid solutions (all octahedral bonding configurations, ranging from cubic symmetry to distorted rhombic).[9, 111 The downfield shift of the Nb-CST is consistent with the de-shielding nature of the octahedra participating in the cubane cluster. The niobium sites present in the PMN and PMNPT solid solutions are all corner sharing.[9, 111 The observed line widths for the Nb-CST (Table I) are in the same range observed for slightly distorted octahedral in the PMNPT solids, corresponding to a CQw 17 MHz. In those PMNPT materials, this CQwas assigned to axial (tetragonal) Nb.The line width for the Nb-CST is also much larger than the narrow 2000 Hz resonance observed for the very symmetric cubic Nb site reported in the PMNPT materials.

-


-

381

Figure 3: a) 93Nb Static and MAS NMR spectra of the 0% Cs Nb-CST material. b) 93NbMAS NMR spectra of the Nb-CST materials as a function of Cs loading.

A

a.

b.

A

i\

I

"

'

I

-500

"

"

I

-1000

.

ppm

Table I. 93NbMAS NMR characterization of cesium exchanged Nb-CST materials. wt % C S ,Nb-CST

&so

(ppm)"

FWHM (Hz)

0

-722

15160

3.8

-726

13725

6.4

-726

13155

9.0 9.6

-725 -726

13559 15598

a. Apparent frequency shift, second order quadrupolar effects not determined in these single field strength studies. Estimated error f 5 ppm.

382


In addition to the PMN and PMNPT solid solutions, two other 93NbNMR studies on model compounds helped place bounds on the symmetry of the niobium octahedra present in the Nb-CST materials. The 93Nb MAS NMR line widths in Nb-CST (Table I) are very similar to that found in Sn2Nb207, where the niobium octahedral is characterized by a single Nb-0 bond length of 1.98 A and two 0-Nb-0 bond angles, 90.59' and 89.42".[10] In contrast, the niobium octahedral present in SnNb2O6 (Foordite) is significantly distorted with the Nb-0 bond distances ranging between 1.85 to 2.16 A. The corresponding static 93Nb NMR spectra of Foordite shows a very broad second-ordered broadened yuadrupolar powder pattern with a CQ of -38 MHz.[lO] Based on the observed Nb MAS N M R frequency shifts and line widths observed for the exchanged NbCST (Table I) we conclude that the niobium octahedra present in Nb-CST have near uniform Nb-0 bond lengths and are slightly distorted from cubic symmetry, like the TiO6 sites found in the Nb free CST materials, where the Ti-0 bond length range from 1.89 - 2.07 A. In the Nb-CST materials a small upfield shift of the 93Nbresonance is also observed with the initial Cs exchange. Further Cs uptake has no effect on the observed 93Nbfrequency shift. This observation suggests that as the initial Cs is incorporated into the Nb-CST, the framework adjusts slightly, allowing the exchanged Cs to occupy an optimal binding site in the center of the tunnel. Once this small change in the structural framework has occurred, additional Cs exchange has no further impact on the framework structure. Evidence for this change in the framework structure as Cs is initially loaded has also been observed in the 29Si MAS NMR data.[12] These 93Nb MAS N M R investigations have provided insight into the structural environment of Nb-CST materials, and have demonstrated that no major variation-in the Nb-0 octahedral symmetry occur with incorporation of Cs. This NMR data, along with the ongoing crystallography investigations, suggest the addition of N b to the CST framework does not affect the Cs selectivity by direct interaction of Cs with the framework Nb. We are continuing to investigate the mechanism responsible for the improved Cs selectivity through 23Na,29Si,'H and 133CsNMR experiments. ACKNOWLEDGMENTS Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy under Contract DE-AC04-94AL85000. This work was supported by the D.O.E. (Office of Science) Environmental Management Science Program, project #8 1949.


383

REFERENCES 1.

2.

3. 4. 5. 6. 7.

8.

9.

10. 11.

12.

384

Dosch, R.G. and R.G. Anthony, Hydrous Crystalline Sitico-Titanates: New Materialsfor Removal of Radiocesiumfrom Concentrated Salt Solutions with pH's in the 1-14 Range, A Topical Report, 1995, Sandia National Laboratories: Albuquerque, N.M. Gu, D., TAM-5, A Hydrous Crystallline Silicotitatnatefor Removal of Cesiumform Dilute Aqueous Waste, in Kinetics, Catalysis and Reaction Engineering Laboratory, Department of Chemical Engineering. 1995, Texas A&M University: College Station, TX. Zheng, Z., C.V. Philip, R.G. Anthony, J.L. Krumhansl, D.E. Trudell, and J.E. Miller, Ion Exchange of Group I Metals by Hydrous Crystalline Silicotitanates. Ind. Eng. Chem. Res., 1996.35: p. 4246-4256. Anthony, R.G., R.G. Dosch, and C.V. Phillip, U. S. Patent # 6,1lU,378, Sandia National Laboratories: U.S.A. Poojary, D., R. Cahill, and A. Clearfield, Synthesis, Crystal Structure, and Ion-Exchange Properties of a Novel Porous Titanosilicate. Chem. Mater., 1994.6(12): p. 2364-2368. Tripathi, A. and A. Clearfield,personal communication with M Nyman. 2002: Texas A & M University. Jehng, J.M. and I.E. Wachs, Structural Chemistry and Raman-Spectra of Niobium Oxides. Chem. Mater., 1991.3(1): p. 100-107. Nyman, M., T.M. Nenoff, A. Tripathi, J. Parise, and R.S. Maxwell., Sandia Octahedral Molecular Sieves (SOMS): Structural and Property Efects of Charge-Balancing the MV-Substituted (M = Ti, Zr) Niobate Framework. J. Am. Chem. Soc., 2002.124(8): p. 1704-1713. Fitzgerald, J.J., S. Prasad, J. Huang, and J.S. Shore, Solid-state 93NbNMR and 93NbNutation Studies of Polycrystalline Pb(Mgl/3Nb2/3)03and (1x)Pb(Mg1&b2,3)03/xPbTi03 Solid-Solution Relaxor Ferroelectrics. J. Am. Chem. Soc., 2000.122: p. 2556-2566. Cruz, L.P., J.-M. Savariault, J. Rocha, J.-C. Jumas, and J.D.P.d. Jesus, Synthesis and Characterization of Tin Niobates. J. Solid State Chem., 2001. 156: p. 349-354. Prasad, S., P. Zhao, J. Huang, J.J. Fitzgerald, and J.S. Shore, Niobium-93 M Q M S NMR Spectroscopy Study of Alkali and Lead Niobates. Solid State Nuclear Magnetic Resonance, 2001. 19: p. 45-62. Cherry, B.R., M. Nyman, and T.M. Alam, In preparation.


SELECTIVE ABSORPTION OF HEAVY METALS AND RADIONUCLIDES FROM WATER IN A DIRECT-TO-CERAMIC PROCESS B.P. Gran, Allen W. Apblett, and Mohamed Chehbouni Department of Chemistry Oklahoma State University Stillwater, OK, 74078. ABSTRACT The ability of molybdenum hydrogen bronze, HMo206to absorb heavy metals and radionuclides from water was investigated. It was found that it could remove substantial amounts of metal ions from water and was selective for those that are chemically-soft or have large radii. The products from uranium, thorium, and neodymium uptake were discovered to be layered metal molybdates while that from lead was wulfenite, PbMo04. In the light of this result, the application molybdenum trioxide for lead adsorption was investigated and it was found to perform similarly to the hydrogen bronze. INTRODUCTION The use of reactive barriers to prevent the spread of pollutants in aquifers is a promising technology that can greatly curtail any environmental endangerment. Furthermore, the reagents used for construction of reactive barriers are generally also amenable to application in pump and treat operations or for treatment of wastewaters. In 1989, the use of granular iron was proposed for in situ remediation of groundwater containing chlorinated organic contaminants. Since that time, the technology has been adopted at numerous sites and has been applied to remediation of other types of organic compounds, inorganic species, and radionuclides [ 1-41.



385

Molybdenum hydrogen bronze, HMo206(also called molybdenum blue) is a promising reagent for environmental remediation that has a number of unique properties which suggests it could perform better than other reductants for treatment of contaminated waters and the construction of permeable reactive containment barriers to prevent spread of pollutants within an aquifer. For example, when reductions of inorganic or organic pollutants are performed in a column-type reactor, the color change fiom royal blue to white would greatly facilitate monitoring of the column’s remaining reductive capacity. Unlike other reductants that can be employed in the presence of water and oxygen (such as iron), molybdenum blue has an open layered structure (Figure 1) that allows the entire reductive capacity to be used and enhances the rate of reaction by providing a tremendously enhanced area for the reaction to take place. Since both reduced and oxidized forms of the oxide materials have layered structures through which reactants and products can intercalate, passivation due to build up of oxidized product on the surface does not occur. This is in significant contrast to iron that can form a crust of rust that arrests further reaction of iron particles with contaminant species. Finally, molybdenum blue is easily recycled after use in redox reactions since regeneration only requires treatment with hot isopropanol in the presence of a trace of HCl or with zinc/HCl. In fact, the regeneration process with isopropanol only produces acetone as a by-product and, in actual industrial production, the acetone could be captured and sold as a commodity chemical.

Figure 1. Structure of Molybdenum Blue Molybdenum blue has been demonstrated to be a useful reagent for dechlorination of halocarbons such as carbon tetrachloride[5,6]. In such reactions, the bronze acts as a source of hydride so that, for example, CCl, is reduced to chloroform, CHC13. This may be somewhat surprising since the protons present

386


in the bronzes are attached to oxygen atoms to yield hydroxides that bridge between two molybdenum centers. Nevertheless, the combination of the hydroxide and a molybdenum(V) center provides a water and air-stable source of hydride. The question posed in this investigation is whether or not molybdenum blue can also reduce metal ions and remove them from aqueous solution. Alternatively, the possibility exists that the hydrogen ions can be exchanged with metals allowing their uptake. If so, the metals would be readily released by oxidation of the bronze to Moo3providing a "switchable" ion exchanger amenable to highly concentrating metal ions in a similar fashion to the materials developed by Dorhout et al. [7]. EXPERIMENTAL All reagents were commercial products (ACS Reagent grade or higher) and were used without further purification. Thermogravimetric studies were performed using 10-20 mg samples on a Seiko ExStar 6200 TGA/DTA instrument under a 50 ml/min flow of dry air. The temperature was ramped from 25 to 600°C at a rate of 5"C/min. Bulk pyrolyses at various temperatures were performed in ambient air in a digitally-controlledmuffle furnace using ca. 2 g samples, a ramp of lO"C/min and a hold time of 4 hr .X-ray powder diffraction (XRD) patterns were recorded on a Bruker AXS D-8 Advance X-ray powder diffractometer using copper K,radiation. Crystalline phases were identified using a searcWmatch program and the PDF-2 database of the International Centre for Diffraction Data [8]. Preparation of Molybdenum Blue A round bottom flask was charged with 30.00 g of Moo3, 300 ml of n-butanol and 5 ml of concentrated HC1. The mixture was refluxed for 12 hours at which time it had turned a very dark blue color. At this point, the reaction mixture was cooled to room temperature and was filtered through a fine sintered-glass filter funnel. The dark blue solid was washed with n-butanol and was then dried in a vacuum oven at room temperature. The yield was 28.30 g (94%). The XRD pattern of the product corresponded to that of Mo205(OH)(ICDD #14-0041). Measurement of the Uptake of Metals by Molybdenum Blue Molybdenum blue was tested for the ability to remove Pb2+,Th4', U O F and Nd3+from aqueous solution. HMo206(1.0 g) was reacted with 100 ml of individual approximately 0.lM solutions of Pb2+,Th4', U O F and Nd3+.In all cases, nitrate salts were used with the exception of uranyl where both a nitrate and an acetate salt were tested. After stirring magnetically for a sufficiently long time for complete reaction, as indicated by complete disappearance of the blue color,


387

the mixtures were separated by filtration through a 20 ym nylon membrane filter. The solid products were washed copiously with distilled water and then were dried in a vacuum desiccator. They were subsequently characterized by infrared spectroscopy, thermal gravimetric analysis, and X-ray powder diffraction. The uranium and neodymium concentrations in the treated solutions were analyzed using UVNisible spectroscopy (h=415 nm and 521 nm, respectively). Lead was determined gravimetrically as lead chromate [9]. Quantitation of thorium was performed colorimetrically using the blue complex (h=575 nm) formed between thoriwn and carminic acid [101. Selectivity Determination The selectivity of molybdenum blue for actinides was tested by competition experiments with calcium. Thus, the reactions between uranyl nitrate and molybdenum blue were repeated in the presence of 0.5, 1.0, and 5.0 molar equivalents of calcium nitrate per mole of uranyl ion and the uptake of uranium was determined by UVNisible spectroscopy. Reaction of Lead Nitrate with Moo3 Pb(NO& (3.98 g, 12 mmol) was dissolved in 100 ml of water and the resulting solution was stirred with Moo3 (1.44 g, 10 mmol) for 72 hours. The resulting white solid was isolated by filtration was washed with water and dried in a vacuum oven at room temperature for 12 hours. The yield was 1.73 g, corresponding to an uptake of 0.27 g of Pb. XRD analysis showed the solid to be a mixture of PbMo04 (ICDD # 44-1486) and unreacted Moo3. RESULTS AND DISCUSSION Molybdenum blue was tested for its ability to remove Th4+ (as a model for plutonium(IV)}, UO? (of interest in its own right and as a model for PuO?), and Nd3' (as a surrogate for the later transuranics , radioactive lanthanides and Pu3+) from aqueous solution. Also, the uptake of lead as a model heavy metal was also investigated. The experiments that were performed were designed to determine the capacity of the blue reagents for the various metals and attempt to identify the mechanism of metal uptake. Molybdenum blue was reacted with an aqueous solution of each of the metals listed above. The stoichiometry was adjusted so that there was at least a one-fold excess of contaminant metal ions {on the basis of one molar equivalent of metal ion per M(V) site in HMo206}. The experimental conditions and results for the molybdenum blue/metal ion reactions are listed in Table I while the results of the analyses and binding capacity calculations are given in Table 11. The results show that molybdenum

388


blue has a remarkable capacity for absorption of actinides and heavy metals. Molybdenum blue absorbed 122% by weight of uranium, 37% by weight of thorium, 61.6% by weight of neodymium, and 110% by weight of lead. The substitution of acetate ions for nitrate ions has a small, negative effect on the uptake of uranium. These extremely high capacities bode well for the eventual application of these materials in environmental remediation. The uptake of the metals in terms of milliequivalentsper gram of molybdenum blue were 4.27 for neodymium, 5.14 for uranium, to 5.29 for lead. Thus, the moles of metal that can be absorbed by molybdenum blue varies with the metal used. Within the group of doubly-charged metal ions, the moles of metal absorbed are almost equivalent. In the case of the latter metals, the uptake of metals may be expressed as 1.5 moles per mole of HMo206 and is therefore larger in magnitude than the number of Mo(V) centers. This result indicates that the molybdenum(V1) centers in molybdenum blue also play a role in metal binding. The uptake of neodymium was 1.24 moles per mole of molybdenum blue. Table I. Experimental Conditions for Metal Uptake Experiments Weight of Molybdenum Blue (g)

Weight of Solid Product (g)

Color of Solid Product

Uranium acetate

1.04

2.15

Yellow

Uranium nitrate

1.05

2.32

Yellow

Thorium

1.oo

1.40

White

Neodymium

1.10

1.48

Grey

Lead

1.04

2.74

White

Metal

Clearly, the results indicate that the uptake of the metals is not a simple ionexchange reaction since the uptake exceeds the number of exchangeable ions. The color changes observed during the uptake of the metals indicate a redox reaction in which the molybdenum(V) is oxidized to molybdenum (VI) but the final colors also demonstrate that the contaminant metals are not reduced so that the responsibility of a redox process for metal uptake may be ruled out. Presumably, the oxidation of the molybdenum blue is due to reaction with atmospheric oxygen.


389

A major concern for the application of molybdenum blue in the field is its selectivity for actinides and heavy metals as opposed to benign cations normally found in natural waters. Therefore, the selectivity of molybdenum blue for uranyl ion over calcium ions was determined. The results are displayed in Table 111 and demonstrate that molybdenum blue is highly selective for uranium. Even a fivefold higher concentration of calcium ions over uranyl ions had little effect on the absorption of uranium. Curiously, the minor effect that calcium does exert on uptake is greatest when it is below equimolar amounts, least when it is present in the same concentration as uranium, and increases thereafter. Table 11. Results of Metal Uptake Experiments Final Concentration

Uptake (mmol)

Metal Capacity (m0Vg)

Metal Capacity (weight %)

0.053 M

4.7

4.5

108%

0.046 M

5.4

5.1

122%

Thorium

0.084 M

1.6

1.6

37.0%

Neodymium

0.047 M

5.3

4.8

69.5%

Lead

0.045 M

5.5

5.3

110%

Uranium acetate Uranium nitrate

Table 111. Results of Competition Experiments Calcium: Uranium Ratio

Final Uranium Concentration

Weight Percent of Uranium Absorbed ~~

0: 1

0.046 M

122%

1:2

0.055 M

100%

1:l

0.049 M

121%

1:5

0.052 M

111%

390


Infrared spectral analysis of the various solid products and molybdenum blue was performed in order to gain a better understanding of the nature of the products and to perhaps shed some light on the mechanism of metal uptake. The positions of the molybdenum-oxygenstretches are given in Table IV. Molybdenum blue was found to have a characteristic absorption at 857 cm-' which is different from the bands observed in molybdenum trioxide. In all cases, except for neodymium, this band, attributable to Mo(V)-0 stretching vibrations has disappeared. The neodymium compound is unusual because the molybdenum centers appear to be freely rotating in the solid so that there is rotational structure to the infrared absorptions making it difficult to assign the positions of the vibrations. Nevertheless, the data in Table IV demonstrate that the solid products from reaction of molybdenum blue with uranium, uraniumlcalcium mixtures, and neodymium all contain network polymers based on Moo6 octahedra, as demonstrated by multiple MO-0stretches. By contrast, the lead product had a single strong MO-0absorption at 786 cm-' attributable to a tetrahedral Moo4 center. Table IV. Metal-Oxygen Stretching Frequencies Observed in the Infrared Spectra Nd+ Pb + U+ U+Ca+ Moo3 MoBlue MOBlue MOBlue MOBlue MOBlue

998

999 (w)

998

884

970*

980"

972"

892

90 1

913

857

866 840*

849* 786

569

572

551

533

502

572 498

* U-0 stretches of the uranyl ion In addition to the M-0 stretching bands, the infrared spectra of the solids also contain bands attributable to a small amount of anions that are also absorbed from aqueous solution. The solids from reaction of molybdenum with uranyl, lead,

EnvironmentalIssues and Waste Management TechnologiesVIII

391

and neodymium nitrate all display an infrared absorption at 1384 cm-'. Notably, this band is due to ionic nitrate and not nitrate covalently-bound to the contaminant metals [111. The product from uranyl acetate has weak bands at 1506 and 1436 cm-' that are due to acetate ions - again the positions of these bands do not correspond to acetate bound to uranium (15 14 and 1480 cm-') that was determined from the infrared spectrum of the starting material. The uptake of the anions indicates that when the metals are bound, the charge is not entirely compensated by the negative charge of the molybdate framework. Nevertheless, the absorptions for the extraneous anions are weak indicating a low degree of incorporation into the solid products. This conclusion was supported by the fact that the ceramic yields derived from heating the solids to 600°C in a thermal gravimetric analyzer were quite high (Table V). Indeed, the majority of the weight losses occur between room temperature and 200°C and can be attributed to dehydration and dehydroxylation reactions. Table V. Results of TGA Experiments Metal Salt

Temperature Range of Weight Loss

Ceramic Yield

U acetate

25-436°C

88.2%

U Nitrate

25-469°C

90.2%

Thorium Nitrate

25-502°C

94.9%

Nd Nitrate

25-209"C

96.5%

Pb Nitrate

215-495°C

99.2%

X-ray powder diffraction analysis of the product from lead uptake by HMo206revealed that it consisted mainly of PbMo04 (wulfenite, ICDD # 441486) plus a small amount of molybdite (Moo3, ICDD # 05-0508). The other metals, however, formed crystalline phases that did not match normal molybdate salts. However when heated to 600°C these unidentified phases were converted to a small amount of Moo3 and U02(Mo04),Th(MoO&, or Nd2(Mo04)3 , depending on the metal. These results in combination with the infrared spectral data suggest that the metal ions intercalate between the layers of HMo206and react to give what is likely to be phases that consist of negatively-charged slabs of Moo6 octahedra with the contaminant ions residing between the layers. In the

392


case of lead, however, the interaction between HMo206and Pb” is so strong that the molybdenum oxide layers are destroyed to yield a normal ortho-molybdate salt. Since the uptake of metals does not rely on ion exchange or redox chemistry there is an implication that the reduction of Moo3 to HMo206is not necessary for metal uptake since the parent oxide also consists of layers of Moo6 octahedra. Reaction of aqueous lead nitrate with Moo3 demonstrated that the trioxide could absorb lead ions. In a 72 hour reaction at room temperature, 1.44 g of Moo3was found to absorb 0.27 g of lead from a 1.2 M lead nitrate solution. XRD analysis showed that the reaction was not complete so that unreacted Moo3 was present along with the expected product, PbMo04. CONCLUSION In conclusion, it has been demonstrated that molybdenum blue has an extremely high capacity for absorption of contaminant metals. Considerable information has been collected concerning the mechanism of metal absorption and the results obtained so far suggest intercalation of the metal ions between the layers of HM0206 followed by reaction to yield solids in which the metal ions tare trapped as counterions to the freshly-generated molybdate sites. These reactions are highly selective for heavy metals and suggest considerable promise for application in environmental remediation and as reactive barriers for the prevention of the spread f contaminant plumes. ACKNOWLEDGEMENT Support for this research by Oklahoma State University’s Environmental Center is gratefully acknowledged. The National Science Foundation, Division of Materials Research, is thanked for Award Number 9871259 that provided funds for the X-ray powder diffractometer used in this investigation. REFERENCES 1

D. R. Burris, R. M. Allenking, V. S. Manoranjan, T. J. Campbell, G. A. Loraine, and B. L. Deng, “Chlorinated Ethene Reduction by Cast-Iron: Sorption and Mass Transfer’’J. Environ. Eng., 124, 1012-1019, 1998. L. Charlet, E. Liger, and P. Gerasimo, “Decontamination of TCE-Rich and URich Waters by Granular Iron: Role of Sorbed Fe(I1)”J. Environ. Eng., 125,2530, 1998.


393

T. L. Johnson, W. Fish, Y.A. Gorby, and P. G. Tratnyek, “Degradation of Carbon-Tetrachloride by Iron Metal Complexation Effects on the Oxide Surface” J. Contaminant Hydrology, 29,379-398,1998. S. F. Ohannesin and R. W. Gillham, “Long-Term Performance of an In-Situ Iron Wall for Remediation of VOCS” Ground Water, 36, 164-170, 1998.

A.W. Apblett, L.D. Byers, and L.E. Reinhardt, “Dechlorination of Chlorocarbons by Molybdates and Vanadates” in Preprints of Papers Presented at the 213th ACS National Meeting, 37,300-302, 1997. Allen W. Apblett, B.P. Gran, and Katie Oden “ReductiveDechlorination of Chloromethanes Using Tungsten and Molybdenum Hydrogen Bronzes or Sodium Hypophosphite”in Chlorinated Solvents and DNAPLS; Reactive Permeable Barriers and Other Innovations, (ACS Book Series, Washington, DC, 2002), 154164. P.K. Dorhout and S.H. Strauss, “The Design, Synthesis, and Characterization of Redox-Recyclable Materials for Efficient Extraction of Heavy Element Ions from Aqueous Waste Streams”, ACS Symposium Series, 727,53-68, 1999. “Powder Diffraction File (PDF-2)” (International Centre for Diffraction Data, Newtown Square, PA). 9

A. I. Vogel, G. H. Jeffery, J. Bassett, J. Mendham, and R. C. Denney, Vogel’s Textbook of QuantitativeAnalysis, (Longman Scientific and Technical: Burnt Mill, Harlow Essex, UK, 1989), pp. 458-459. l0 F. D.

Snell, C. T. Snell, and C.A. Snell, “Thorim by Carminic Acid” in ColorirnetricMethods of Analysis, Vol. IIA, (D. Van Nostrand Co.: Princeton, N.J., 1959), pp. 518-519. l 1 K. Nakamoto, Infiared

and Raman Spectra of Inorganic and Coordination Compounds, 4th ed. (John Wiley & Sons:, New York, 1986).

394


KEYWORD AND AUTHOR INDEX Absorption, 385 Actinides, 301, 314 Akai, T., 23, 39 Alam, T.M., 377 Al-Fadul, S.M., 15 Anderson, G., 177 Apblett, A., 15, 385 Attard, D.J., 321 Awano, M., 105 Barrium hollandite, 23 1 Bateman, K.J., 355 Begg, B.D., 313 Bennett, J.P., 3 Bibler, N.E., 209 Bickford, D.F., 123 Blum, A.G., 199 Blumenkranz, D.B., 209 Borosilicate glass, 169, 209, 215 Brossia, S., 283 Buechele, A.C., 225,253 Cahill, T.A., 59 Calcium carbonate, 67 Carter, M.L., 321 Cassingham, N., 337 Catalyst, 105 Cement, 39 Ceramic waste form, 355, 363 Cesium, 231,377 Chehbouni, M., 385 Chen, D., 23,39 Cherry, B.R., 377 Choi, K., 177 Chromium oxide, 347 Cliff, S.S., 59 COGEMA, 113 Construction material, 39

Corrosion, waste glass, 245, 291 Corrosion, waste package, 263 Crawford, C.L., 209 Crum, J.V., 141 Crystalline silicotitanate, 377 Crystallinity, constraint, 133 Day, D.E., 329,347 Decomposition, 67 Defense Waste Processing Facility (DWPF), 123 Delisting, 83 Desvaux, J.L., 113 Dilatometer, 67 Disposal, 123 Dissolution, 235 DOE, 95 Do-Quang, R., 113 Dredge sediment, 31 Dunn, D., 283 Durability, 185 Ebert, W.L., 235 Edwards, T.B., 199 Electronic applications, 74 Emissions, 49, 67, 105 Emissions, particulate, 59 EPA, 83 Erickson, A., 185 Extractants, magnetic, 15 Feng, K., 67 Ferrara, D.M., 209 Fission products, 3 13 Fluid chemistry, 263 Fracture, glass, 275 Fuel, nuclear, 113


395

Gan, H., 215,225 Glass block, 275 Glass bonded, 363 Glass crystallization, 159 Glass fracture, 275 Glass melter, 123, 133, 141 Glass viscosity, 199 Glass waste, 39 Glass, borosilicate, 169, 209, 2 15 Glass, colored, recycling, 15 Glass, partially crystallized, 29 1 Glass, phase separation, 15 Glass, simulated waste, 199 Glass, iron phosphate, 329, 347 Glass, nitrate containing, 49 Glass, phosphate, 337 Glass, sulfate containing, 225 Godon, N., 275 Gombert, D., 329 Goodwin, S.M., 141 Haber, R.A., 31 Hamel, W., 95 Hanford, 95, 151, 209,225,253, 347 Hanna, J.V., 3 13 Health, 74 Heavy metals, 385 Heckendorn, EM., 123 High-level waste, 123, 133 High-salt waste, 37 1 High-chrome, 347 High-level waste (HLW), 95, 141, 159, 185, 235, 263, 275, 283, 291, 301, 337,347,355,363 High-level waste, liquid, 113 Hill, K., 31 Hollandite, 23 1 Hot isostatic pressing (HIP), 263, 355, 363 Hrma, P., 133, 151, 159, 245, 291, 337 Huang, W., 347 Huffman, L., 95

396

Hunter, B.A., 313 Hydration, 253 Idaho, 185 INEEL, 169, 177, 329 Iron phosphate glass, 329, 347 Jain, V., 263,283 Jantzen, C.M., 83 Jimenez-Cruz, M., 59 Jones, L.E., 49 Jouan, A., 113 Katayama, S., 105 Kelly, P.B., 59 Kim, C.-W., 177,329,347 E m ,D.-S., 133, 151, 169, 337 Kiran, B.P., 385 Kong, P., 177 Kriikku, E.M., 123 Kuraoka, K., 39 Kwong, K.-S., 3 Ladirat, C., 113 Land disposal, 209 Leach testing, 321 Leaching, 39,215,275 Lead, 75 Legislation, 74 Lerchen, M., 95 Li, H., 313 Lombardo, S.J., 67 Low activity waste (LAW), 209, 225, 253 Low-level waste, 177 Luo, s.,49 Maeda, K., 105 Magnetic extractants, 15 Martin, L.C., 83 Mass spectrometer, 67 Masui, H., 23


MatyA-, J., 133 Melter, 123, 133, 141 Microwave heating, 363 Micro-XRF, 59 Mid-Delaware River, 31 Minet, Y., 275 Mitchell, D.R.G., 321 Mixed waste, 185 Modeling, 133,235, 263 Models, 151 Molybdenum hydrogen bronze, 385 Monitoring, waste package, 283 Mooers, C.F., 253 Mougnard, P., 113 Muller, I.S., 209

Refractories, 3 Regulations, 95 Repository, 235,263,283 Reuse, 3 Riley, B.J., 291 Ruthenium oxide, 141

Salt wastes, 371 Sampling, 355 Savannah River Site (SRS), 83 Schatz, T.R., 253 Schoenung, J.M., 75 Scholes, B.A., 185 Schumacher, R.F., 199, 209 Sediment, dredge, 3 1 Selective catalytic reduction (SCR), 105 Nelson, L.O., 177 Sensors, 283 Niobium, 377 Shackelford, J.F., 59 Nitrates, 49 Shin, S.-W., 177 NOx, 49, 105 Nuclear fuel, spent, 113,355,263,363 Shirakami, T., 39 Silica, 39 Nyman, M., 377 Silicotitanate, 377 Simulation, 199 O’Holleran, T.P., 355, 263 Sintering, 67 Organics, 15 Sludge, wastewater treatment, 83 Smith, G.L., 209, 371 I?articulate emissions, 59 Smith, H.D., 209,371 1?eeler, D.K., 169,199 Smith, M.E., 123 1?egg, I.L., 209, 215, 225, 253 Sodium bearing waste, 169, 329 I?erera, D.S., 313 Sodium extraction, 39 1?hase equilibria, 185 Solubility, 159 1?hase separation, glass, 15 Sridhar, N., 263 1?ickett, J.B., 83 Strontium, 377 1?lutonium, 301 Sulfate, 225 1?olycerams, 371 Sulfur, 225,337 1?olymer composite, 37 1 Sundaram, S.K., 141 Swanberg, D.J., 209 Radionuclides, 385 Synroc, 231,313 Raman, S.V., 185 Raw material, 31 TCLP, 215 Ray, C.S., 347 Test, 199, 291, 283 Recvclinrz, 3, 15


397

Tile, 31 Titanate ceramics, 23 1,301 Toxicity, 215 Trad, T.M., 15 Urabe, K., 39 Uranium, 301

Zahir, M.H., 105 Zareba, A.A., 185 Zelinski, B.J.J., 371 Zhang, Z., 313 Zhu, D., 329 Zirconolite, 3 13

Vance, E.R., 301, 313, 321 Vapor phase hydration, 253 Vidensky, I., 225 Vienna, J.D., 151, 159, 169, 209, 245, 291,337 Viscometer, 199 Viscosity, 185 Vitrification, 113, 169, 177, 225, 329, 337 Vitrified mixed waste, 83 Waste glass, 39, 151, 159, 225, 235 Waste glass, corrosion, 245 Waste package, 263,283 Waste, high-level, 95, 141, 159, 185, 235, 263, 275, 283, 291, 301, 337, 347,355,363 Waste, sodium bearing, 329 Wastewater treatment sludge, 83 Water, 15, 385 Wiemers, K., 95 Willwater, T.M., 141 World Trade Center, 59 Wysoczanski, R., 253 Xia, G., 371 X-ray fluorescence, 59 Yamamoto, Y., 39 Yang, L., 283 Yazawa, T., 23, 39 Yeager, J.D., 245 Yucca Mountain, 235, 263, 283

398


Environmental Issuesand Waste Management Technologies in the Cermaic & Nuclear (Ceramic Transactions Series)

Ceramic Nanomaterials and Nanotechnology: Ceramic Transactions (Ceramic Transactions Series)

Advances and Applications in Electroceramics: Ceramic Transactions (Ceramic Transactions Series)

Advances in Sintering: Ceramic Transactions (Ceramic Transactions Series, Volume 209)

Ceramic Materials and Components for Energy and Environmental Applications: Ceramic Transactions Volume 210 (Ceramic Transactions Series)

Competitive energy management & environmental technologies

Nuclear Decommissioning Waste Management and Environmental Site Remediation

Advances in Electroceramic Materials: Ceramic Transactions, Volume 204 (Ceramic Transactions Series)

Advances in Biomedical and Biomimetic Materials: Ceramic Transactions,206 (Ceramic Transactions Series)

Advances in Porcelain Enamel Technology: Ceramic Transactions Volume 211 (Ceramic Transactions Series)

Advances in Materials Science for Environmental and Nuclear Technology II: Ceramic Transactions

Materials for Electrochemical Energy Conversion and Storage (Ceramic Transactions, Vol. 127) (Ceramic Transactions Series)

Materials Processing and Texture: Ceramic Transactions, Volume 200 (Ceramic Transactions Series)

Processing of Nanoparticle Structures and Composites: Ceramic Transactions (Ceramic Transactions Series)

Novel Processing of Ceramics and Composites: Ceramic Transactions (Ceramic Transactions Series, Vol. 195)

Advanced Nuclear Fuel Cycles And Radioactive Waste Management: Nuclear Development

Environmental Education and Solid Waste Management

Environmental education and solid waste management

Waste Management

Advances in Environmental Accounting and Management, Volume 2 (Advances in Environmental Accounting & Management) (Advances in Environmental Accounting & Management)

The Reporter’s Handbook on Nuclear Materials, Energy, and Waste Management

Environmental Technologies

Applied Ecology & Environmental Management

Radioactive Waste in Perspective: Nuclear Development

Olive Processing Waste Management

Basic Hazardous Waste Management

Electronic Waste Management

Environmental Issuesand Waste Management Technologies in the Cermaic & Nuclear (Ceramic Transactions Series)

Ceramic Nanomaterials and Nanotechnology: Ceramic Transactions (Ceramic Transactions Series)

Advances and Applications in Electroceramics: Ceramic Transactions (Ceramic Transactions Series)

Advances in Sintering: Ceramic Transactions (Ceramic Transactions Series, Volume 209)

Ceramic Materials and Components for Energy and Environmental Applications: Ceramic Transactions Volume 210 (Ceramic Transactions Series)

Competitive energy management & environmental technologies

Nuclear Decommissioning Waste Management and Environmental Site Remediation

Advances in Electroceramic Materials: Ceramic Transactions, Volume 204 (Ceramic Transactions Series)

Advances in Biomedical and Biomimetic Materials: Ceramic Transactions,206 (Ceramic Transactions Series)

Advances in Porcelain Enamel Technology: Ceramic Transactions Volume 211 (Ceramic Transactions Series)

Advances in Materials Science for Environmental and Nuclear Technology II: Ceramic Transactions

Materials for Electrochemical Energy Conversion and Storage (Ceramic Transactions, Vol. 127) (Ceramic Transactions Series)

Materials Processing and Texture: Ceramic Transactions, Volume 200 (Ceramic Transactions Series)

Processing of Nanoparticle Structures and Composites: Ceramic Transactions (Ceramic Transactions Series)

Novel Processing of Ceramics and Composites: Ceramic Transactions (Ceramic Transactions Series, Vol. 195)

Advanced Nuclear Fuel Cycles And Radioactive Waste Management: Nuclear Development

Environmental Education and Solid Waste Management

Environmental education and solid waste management

Waste Management

Advances in Environmental Accounting and Management, Volume 2 (Advances in Environmental Accounting & Management) (Advances in Environmental Accounting & Management)

The Reporter’s Handbook on Nuclear Materials, Energy, and Waste Management

Environmental Technologies

Applied Ecology & Environmental Management

Radioactive Waste in Perspective: Nuclear Development

Olive Processing Waste Management

Basic Hazardous Waste Management

Electronic Waste Management

Recommend Documents