The Roots of Organic Development (Industrial Chemistry Library)

Industrial Chemistry Library, Volume 8 The Roots of Organic Development Industrial Chemistry Library Advisory Editor...

Author: J.-R. Desmurs | S. Ratton

57 downloads 1178 Views 16MB 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

Industrial Chemistry Library, Volume 8

The Roots of Organic Development

Industrial Chemistry Library Advisory Editor: S.T. Sie, Faculty of Chemical Technology and Materials Science Delft University of Technology, Delft, The Netherlands

Volume 1

Progress in C1 Chemistry in Japan (Edited by the Research Association for C 1 Chemistry)

Volume 2

Calcium Magnesium Acetate. An Emerging Bulk Chemical for Environmental Applications (Edited by D.L. Wise, Y.A. Levendis and M. Metghalchi)

Volume 3

Advances in Organobromine Chemistry I (Edited by J.-R. Desmurs and B. Gdrard)

Volume 4

Technology of Corn Wet Milling and Associated Processes (by P.H. Blanchard)

Volume 5

Lithium Batteries. New Materials, Developments and Perspectives (Edited by G. Pistoia)

Volume 6

Industrial Chemicals. Their Characteristics and Development (by G. Again)

Volume 7

Advances in Organobromine Chemistry II (Edited by J.-R. Desmurs, B. Gdrard and M.J. Goldstein)

Volume 8

The Roots of Organic Development (Edited by J.-R. Desmurs and S. Ratton)

Industrial Chemistry Library, Volume 8

The Roots of Organic Development Edited by Jean-Roger

Desmurs

Rh6ne Poulenc Industrialisation, CRIT/Carrikres, 85 Avenue des Frbres Perret, 69192 Saint-Fons Cedex, France Serge Ratton

Rh6ne Poulenc Organic Intermediates Enterprise, 25 Quai Paul Doumer, 92408 Courbevoie Cedex, France

1996 ELSEVIER Amsterdam

~

Lausanne

~

New York --

Oxford --

Shannon

~

Tokyo

ELSEVlER SCIENCE B.V. Sara Burgerhartstraat 25 P.O. Box 211, 1000 AE Amsterdam, The Netherlands

ISBN: 0-444-82434-0 9 1996 Elsevier Science B.V. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written permission of the publisher, Elsevier Science B.V., Copyright & Permissions Department, P.O. Box 521, 1000 AM Amsterdam, The Netherlands. Special regulations for readers in the USA - This publication has been registered with the Copyright Clearance Center Inc. (CCC), Salem, Massachusetts. Information can be obtained from the CCC about conditions under which photocopies of parts of this publication may be made in the USA. All other copyright questions, including photocopying outside of the USA, should be referred to the copyright owner, Elsevier Science B.V., unless otherwise specified. No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein.

This book is printed on acid-free paper. Printed in The Netherlands

FOREWORD It is our belief within RHONE-POULENC that the key to building long term customer relationship in our industry is superior technology backed up by outstanding service. Benefits of superior technology in Organic Chemistry are multiple : lower cost raw materials, shorter synthesis routes, improved yields, selectivity and kinetics, resulting in better productivity. Higher transformation rates of less hazardous materials leads to healthier, cleaner operations with reduced waste disposal issues. Last but not least, process safety is continually upgraded as more intimate knowledge of chemical reactions and other unit operations is achieved. For our worldwide customers such technical progress creates multifaceted value: reliability, shorter response time, more competitive economics, improved quality leading to faster registration, and safer and more environmentally responsible operations. Furthermore, it enables us to extend the use of this expertise to the adaptation of decisive physical or chemical properties of molecules to provide our customers with desired use properties. Making our overall skills available to customers to solve their problems is indeed the basis of our ,, Chimie Nouvelle ,, approach. In this spirit we expect and look forward to provide, along with our Organic Intermediates technology, whatever services are required to make our joint success complete : efficient pilot facilities, advanced analytical equipment with expert staff, toxicology and eco toxicology support, environmental services, formulation capabilities, ... we do this throughout the world. This is the way we aim to become your preferred partner in organic chemistry, to gain your confidence and be able to participate early in your most important projects. May this book demonstrate to all our existing or potential partners our commitment to top level organic chemistry. We are proud of the achievements and expertise of our teams. May our partners keep challenging them to build leaderships together.

Bertrand LOUVET Rh6ne-Poulenc Chemical Sector Executive vice-President

Ted ZIEMANN President of Rh6ne-Poulenc Organic Intermediates Enterprise

This Page Intentionally Left Blank

PREFACE It seems to us, symbolic, important, and above all promising for the furore, that the year in which Rh6ne-Poulenc holds its centenary celebrations also sees the publication of a scientific review, gathering together organic chemistry research carried out in common by groups from universities and other large organisations, and with Rh6ne-Poulenc research workers. The development of an industrial group, especially one such as Rh6ne-Poulenc, is directly linked to the possibility of innovation. For this, it is necessary to rely, not just on the groups own resources and strengths, but also on the research and discoveries made by external research bodies. The General Management of the Group, as far back as 1974, was aware of this need to have a closer association with upstream research, and so signed the first contract with the CNRS (National Research Centre). This was only the first step, although an important one, and several years passed before Rh6ne-Poulenc opened its research doors to the outside world. From 1981 until the present day, with the support and constant incitement of the Group managers, a network of collaborators has been set up, at first in France, and then abroad. This has required, from everyone involved, efforts regarding mutual understanding, always within a climate of trust. The first organised meetings were RP-CNRS symposia based on themes, and focusing on problems directly related to the Groups chemical interests (homogeneous catalysis, chemical reactivity, regioselectivity...) during which our research workers and those of the CNRS exchanged information and results and initiated future collaborations. Today these symposia have been replaced by ,, Journ6es RP-CNRS ,, where several themes are examined over a two day period, using a format akin to a seminar. It was then decided to modify these [CoIIIc12(CF3CO)] + [(CoIIIc12(CF3CO2)]

2 [COIIC12] + (CF3C0)20

(16)

It is also clear that these Co nI species can either oxidize the aromatic or react by electrophilic attack (refs. 34,35). The aromatic dimer is formed as a result of this monoelectronic oxidation : [CoIIIc12(O2CCF3)] + PhOCH 3 PhOCH3~

+ PhOCH 3

> [COIIC12] + (PhOCH3) "+, CH3CO 2-

> dimer H + H +

dimer H + [CoIII(o2CCF3)]----~ dimer + CFaCO2H + Con

45

(17) (18) (19)

The real mechanistic alternative is indeed whether the C-C bond is formed by a classic electrophilic attack (eqn. 20) or by the interaction of [ColII(c1)2(COCH3)] with the radical cation of the aromatic (eqn. 21). PhOMe + [ComC12(COCF3)] PhOMe"+ + [CoIIIc12(COCF3)]

> p-CF3COPhOCH3 + [CoIII(c1)] q-HC1

(20)

> p-CF3CPhOCH3 + [ColICI2] -k- H +

(21)

The latter possibility has been suggested by Iqbal in the case of the Co II catalyzed

acetylation

of

anisole

by acetylchloride

(ref.

31).

However,

no

experimemal support of the mechanism has been provided so far. Oxidation and electrophilic attack can be competitive pathways or oxidation can be a common pathway for both dimerization and trifluoroacetylation.

CONCLUSION We have reported here the catalyzed trifluoroacetylation of methoxyaromatics by TFA for the first time. No solvent is used in the reaction (coordinating solvents inhibit the reaction). In the case of anisole, variations of the experimental conditions allow to selectively lead to either paradimerization or trifluoroacetylation. Given the inconvenience of the production of large quantities of aluminium waste in the A1C13 induced process and the delicate handling of trifluoroacetic chloride due to its boiling point o f - 27~

the finding reported here should prove very practical for

the trifluoroacetylation of methoxyaromatics on the industrial scale.

References .

2. 3. 4 5 6 7 8 9 10. 11. 12.

J.H. Simons, E.O. Ramler, J. Am. Chem. Soc., 65,389, (1943) J.H. Simons, W.T. Black, R.F. Clark, J. Am. Chem. Soc., 75, 5621, (1953) D.D. Tanner, A. Kharrat, J. Am. Chem. Soc., 110, 2968, (1988) P.J. Wagner, M.J. Thomas, E. Puchalski, J. Am. Chem. Soc., 108, 7739, (1986) O. Ichitani, S. Yanagida, S. Takamuku, C.J. Pac J. Org. Chem., 52, 2790, (1987) A.G. Anderson, R.J. Anderson, J. Org. Chem., 27, 3578, (1962) W.H. Pirkle, D.L. Sikkenga, M.S. Pavlin, J. Org. Chem., 42,384, (1977) R.K. Mackie, S. Mhatre, J.M. Tedder, J. Fluorine Chem., 10, 437, (1977) S. Clementi, F. Genel, G. Marino, J. Chem. Soc. Chem. Commun., 498, (1967) V.G. Gluckhovtsev, Y.V. II'In, A.V. Ignatenko, L.Y. Brezhnev, Isz. Akad. Nauk., SSSR, Ser. Khim. (Engl. Transl.), 2631, (1988) W.D. Cooper, J. Org. Chem. 23, 1382, (1958) M. Hojo, R. Masuda, E. Okada, Tet. Lett. 28, 6199, (1987) 46

13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35.

T.R. Forbus, J.C. Martin, J. Org. Chem. 44, 313, (1979) R. Stewart, K.C. Teo, Can. J. Chem. 58, 2491, (1980) P.J. Wagner, R.J. Truman, A.E. Puchalski, R. Wake, J. Am. Chem. Soc., 108, 7727, (1986) X. Creary, J. Org. Chem., 52, 5026, (1987) G. Friour, G. Cahiez, J.F. Normant, Synthesis, 37, (1984) S. Sibille, V. Ratovelomanana, J. P6richon, J. Chem. Soc. Chem. Commun., 283, (1992) D. Naumann, M. Finke, H. Lange, W. Dukat, W. Tyrra, J. Fluorine Chem., 56, 215, (1992) X. Creary, J. Org. Chem., 52, 5026, (1987) T. Keumi, M. Shimada, M. Takahashi, H. Kitajima, Tet. Lett. 1990, 783. M. Hiro, K. Arata Chem. Lett., 325, (1978) K. Nomita, Y. Sugaya, S. Sasa, M. Miwa, Bull. Soc. Chim. Japn., 53, (1981) T. Yamaguchi, A. Mitoh, K. Tanabe, Chem. Lett., 1229, (1982) F. Effenberger, G. Epple, Angew. Chem. Int. Ed. Engl., 11,300, (1972) F. Effenberger, F. Steegmiller, Chem. Ber., 121, 117, (1988) T. Mukaiyama, H. Nagaoka, M. Ohshima, M. Murakani, Chem. Lett. 165, (1988) T. Mukaiyama, T. Ohno, T. Nishimura, S.J. Han, S. Kobayashi, Chem. Lett., 1059, (1991) T. Mukaiyama, K. Suzuki, S.J. Han, S. Kobayashi, Chem. Lett., 435, (1992) A. Kawada, S. Mitamura, S. Kobayashi, Synlett, 545, (1994) J. Iqbal, M.A. Khan, N.K. Nayyar, Tet. Lett., 5179, (1991) J.-P. Begu6, D. Bonnet-Delpon, Tetrahedron Report N ~ 290, Tetrahedron, 47, 3207, (1991) G. Harvey, G. M/ider, Collect. Czech. Chem. Commun., 57,862, (1992) J.K. Kochi in "Metal-Catalyzed Oxidation of Organic Compounds" Academic Press, New York, pp. 120-133, (1981) D. Astruc in "Electron Transfer and Radical Processes in Transition-Metal Chemistry" VCH, Chapter 7, New York, (1995)

47

CATALYSIS BY RARE EARTH PHOSPHATE II 9 SELECTIVE O - M E T H Y L A T I O N OF PHENOLS BY M E T H A N O L IN VAPOR PHASE

LAURENT GILBERT a) MARCELLE JANIN a) ANNE-MARIE LE GOVIC b) PASCALE POMMIER b) AND ALAIN AUBRY b) a) Rh6ne-Poulenc Industrialisation, Centre de Recherche, d'Ing6nierie et de Technologie, 85 Avenue des Fr~res Perret, B.P. 62, 69192 Saint-Fons, France b) Rh6ne-Poulenc Recherches, Centre de Recherches d'Aubervilliers, 52 rue de la Haie Coq, 93308 Aubervilliers Cedex, France

ABSTRACT Vapor phase catalytic alkylation of phenols with methanol was carried out on various phosphates as catalysts. The best activity and selectivity was observed on boron, rare-earth and niobium phosphate. With boron phosphate, the reaction is very selective for O-alkylation even at high temperature. On this catalyst omethoxy-phenol is selectively obtained from 1-2-dihydroxybenzene. With rareearth phosphate calcinated at 400~ and with niobium phosphate, O-alkylation selectivity decreases with an increase of reaction temperature. For rare-earth phosphates it is possible to improve the selectivity by calcination at higher temperature or by a wetness impregnation of cesium hydrogenophosphate. An explanation of these results is proposed.

INTRODUCTION Alkylaryl ethers have distinctive, pleasant odors flavors which make then valuable for the perfume and flavor industries. They are also valuable intermediates for agrochemicals, pharmaceuticals, food preservatives and antioxidants (ref. 1). Among those products, 2-methoxyphenol 1 (guaiacol) is especially interesting as it is used as intermediate for the synthesis of important food additives like vanilline.

48

The most versatile method of preparation is the williamson ether synthesis (ref. 2) particularly by reaction of sodium phenate with halogen derivatives of hydrocarbons. To obtain a higher reactivity, this reactant can be substituted by a dialkyl sulfate or dialkylcarbonate but those alkylating agents scarcely contribute to a more economic process. The preparation of methylarylether by reaction between phenols and methanol was the object of many studies as such a process present main advantages on an economic point of view due to the lower cost of methanol compare to methylchloride and an environmental advantage due to the limitation, and even the total removal of saline effluents. However, to be an attractive one, such a process needs to be highly selective in O-alkylated products with a strict limitation of Calkylated by-products. Numerous studies concerning O-alkylation of phenol were reported (refs. 3-9). Described catalysts belong to all the catalyst families 9oxides (ref. 3) ; phosphates (refs. 4, 5) ; metallosilicates (ref. 6) ; aluminophosphates (ref. 7) ; ion exchange resin (ref. 8). On the other hand, the selective mono-O-alkylation of diphenols was little reported and mainly in patent literature (ref. 9). The main studies deal with synthesis of guaiacol by methylation of 1-2-dihydroxybenzene 2 (catechol) (eqn. 1) catalyzed by boronphosphate eventually doped or supported (ref. 9). The main difficulties of this reaction consists in physical instability of the catalyst which is eluted in the reaction stream conducting to the formation of methylborate as a byproduct which has to be separated. It is then needed to add some new catalyst continuously. We have studied the O-alkylation of catechol in guaiacol by solid-gas catalysis. In this reaction, 1,2-dimethoxybenzene 3 (veratrole) can be produced as a coproduct and also some C-alkylated derivatives as by-products. We have chosen to mainly study phosphates derivatives and to realize a large screening of catalysts. In the preoccupation to screen all the periodic classification of the elements, we have tested numerous metallic phosphates that we have arbitrarily classified as a function of the cation valence. As far as it was possible, we have limited the synthetic routes of catalysts to one or two to be able to compare their intrinsic activity.

~~]

OCH 3

OH

.,~ OH

catalyst +

+

CH3OH vapor phase

2

1

49

H20

(1)

EXPERIMENTAL Catalysts preparation : the various metallic phosphates were prepared by wet synthesis except for niobium oxide phosphate which was obtained from CBmm. According to the solubility of the metallic precursor, two synthetic routes were used: - precipitation (for a totally soluble precursor) - dissolution, reprecipitation (for a less soluble precursor) For precipitation, the synthetic process consists of addition of orthophosphate anions to an aqueous solution of the metallic salt. The precipitated metallic orthophosphate was then recovered by filtration, then washed, dried at l l0~ and then calcinated at various temperature. By this method CaHPO4, FePO4, Ce(HPO4)2, Zr(HPO4)2 were obtained. The dissolution, reprecipitation process consists of preparing a metallic salt suspension of known solubility and in adding to it an aqueous orthophosphate anion solution in order to displace the following equilibrium: M x Ly

+

z Hz-n PO4 n+

-"

"-

Mx H(z-n)PO4

+

Y L t-

with nz = ty

n from 1 to 3 L = CO32-, OH-, C204 z-

The metallic orthophosphate is then recovered using the same workup as reported for the precipitation method. Phosphates of rare earth metals were prepared by this method using rare earth carbonates as precursors, boron phosphate starting from orthoboric acid, and yttrium phosphate from yttrium oxyde. Cesium phosphate synthesis is a little special due to its high solubility in water and cesium phosphate crystals were obtained by water evaporation from an aqueous solution of cesium hydrogenophosphate resulting from the neutralization of an aqueous solution of cesium hydroxyde by phosphoric acid. Crystalline structure (DRX) and specific surface area (BET method) were systematically determined on the dried and calcinated catalysts at 400~ Rare earth phosphates were more precisely characterized and the results are reported in the following paper (ref. I0). Catalytic tests 9The reactions were performed in a vapor phase tubular quartz reactor packed with the catalyst (stationary bed) and heated in a shell oven under nitrogen at the test temperature. During the catalyst's reactions screening were performed between 250 and 390~ using 2.5 ml of catalyst. After thermal 50

equilibria has been reached, nitrogen was introduced via volumetric flow meter (1 l/h). A catechol solution (2.25 mmol/g) in methanol (molar ratio CH3OH / C 6 H 6 0 2 = 10) was introduced with the aid of a syringe at a flow of about 5.5 g per hour. The reaction products were collected and were analysed by liquid chromatography.

RESULTS AND DISCUSSION Catalytic activity of metallic phosphate

Table 1 summarize the catalytic performances of metallic orthophosphates at 270~ The main following conclusions can be then extracted : - calcium phosphate has no activity

- generally the other metallic phosphate are selective for the O-alkylated products. Cesium hydrogenophosphate presents the lowest selectivity but as for zirconium and cerium IV hydrogenophosphate the measurement of selectivity is rather imprecised due to a low activity. Moreover with cesium hydrogenophosphate another difficulty is linked to its high physical instability and so a rapid decrease of the conversion, - in those conditions cerium III and samarium phosphate are the only catalysts giving a significant (# 5 %) yield of veratrole, - the best activities are obtained for boron, rare earth (lanthanum, samarium, cerium) phosphates and niobiumoxyphosphate. All those conclusions are simplified in Figure 1 which show that in terms of productivity, trivalent and pentavalem metallic phosphates are the most active catalysts for this reaction. On the other hand, if we reported the activity as a function of specific surface area (Figure 2) we observed a very high activity of cesium hydrogenophosphate, all the other active catalysts showing a comparable activity.

Table 1. Activity of metallic phosphates

Phosphate

S BET(m2/g)

Cs2HPO4

o rn

5~

|

i

BPO4

37

99

99

58

99

99

76

98

100

60

98

98

77

98

98

82

64

64

60

92

97

79

89

100

95

24

28

72

96

99

88

88

100

92

6C

70

62

99

99

63

92

92

64

68

68

....

LaPO4 (ex carbonate)

CePO4 SmP04 (ex carbonate) NbOPO4

55

Cesium hydrogenophosphate presents a remarkable catalytic activity since its specific surface area is very low. However, this solid cannot be used as a catalyst due to its physically instability under the reaction conditions. We, therefore, focussed our attention in trying to support this catalyst, supports being chosen among the one active in the reaction : BPO4, REPO4 (RE = Sm, La) and NbOPO4. The different catalysts are prepared by wetness impregnation. Initially, we have fixed arbitrarily the cesium hydrogenophosphate content to 10 % in weight. This value is superior to the one needed to generate a monolayer and we can therefore observe, the intrinsic activity of the supported cesium hydrogenophosphate. Table 4. Catalytic activity of Cs2HPO4 impregnated metallic phosphate Entry

Catalyst

Temperature

Conversion

(~ 1

LaPO4

300 '

98

330

l'

82

.....

64

360

i

88

57

0 '

0 0

300

52

98

0

(Cs 10 % p/p)

330

86

88

10

360

97

64

26 12

S m P O 4,

4

Cs2HPO4

(Cs 10 % p/p)

5

77

Selectivity Veratrole (%)

LaPO4, CszHPO4

SmPO

4

Selectivity Guaiacol(%)

(%)

BPO4

300

88

88

330

ii

92

60

360

'

94

" '

35

10 '

5

300

71

100

0

330

90

90

10

360

98

75

24

300

58

99

0

330

76

98

2

360

83

96

3

BPO4, Cs2HPO4

300

18

100

(Cs 10 % p/p)

330

30

100

360

32

100

0

300

52

92

0

330

63

68

0 i

7

8

NbOPO4

NbOPO4,Cs2HPO

|

i

|

0

0

360

64

38

0

300

51

98

2

330

59

86

360

61

81

4

(Cs 10 % p/p)

i

56

Boron phosphate, as previously seen, effectively catalyzed the O-alkylation of catechol to guaiacol (Table 4, entry 5). At different temperature, the reaction leads selectively to guaiacol. The consecutive reaction of veratrole formation being limited even at high conversion. The impregnation of this catalyst by cesium hydrogeno-phosphate gives rise to an heterogeneous product as it was demonstrated by the RX, STEM analysis. The cesium is concentrated in amorphous zones without any interaction with the boron phosphate crystallites. The obtained catalyst presents a lower activity but still an excellent selectivity in guaiacol (Table 4, entry 6). Niobium phosphate exhibits a lower selectivity in Oalkylated products, in particular at high temperature (360~ (Table 4, entry 7). The impregnation of this catalyst by cesium hydrogenophosphate leads to an uniform cesium distribution. The activity of niobium phosphate is not greatly enhanced while the selectivity is increased (Table 4, entry 8). However, cesium hydrogenophosphate interacts mildly with niobium phosphate since it is eluted under the reaction conditions. In the case of rare earth phosphates (Table 4, entries 1-4), the impregnation by cesium hydrogeno-phosphate does not modify noticeably the activity. However the selectivity in O-alkylated products is greatly enhanced. At high conversion one can also observed the formation of veratrole in significative amount by consecutive O-alkylation of guaiacol. Those catalysts having a good activity as well as selectivity were further characterized. The results of this studies are presented in the following article (ref. 10). The influence of the cesium hydrogenophosphate content was examined in the case of lanthanum phosphate. Figure 1 presents the selectivity observed at 330~ for a 80 % conversion. 100 8O ~,

60

~

4o

~

2o 0

I

0

3

6

9

12

Content of Cs2HP4 (% p/p)

15

Fig. 1. Influence of cesium hydrogenophosphate content

5?

This curve shows that an excellent selectivity in O-alkylated products is obtained starting at about 5 % of Cs2HPO4. This value corresponds to the cesium hydrogeno-phosphate quantity which is necessary to obtain a monolayer. The increase in the cesium hydrogenophosphate content does not lead to a lower activity. The study of O-methylation of phenol as well as 1,4-dihydroxybenzene in anisole and 4-methoxyphenol respectively and the condensation of catechol with ethylene glycol demonstrates that the use of L a P O 4 , Cs2HPO4 as a catalyst is a powerfull methodology to selectively access numerous alkylarylethers. Results are presented in Table 5. Table 5. O-alkylation of phenols catalyzed by LaPO 4, Cs2HPO 4

Reaction OH

OCH3

Catalyst

Reaction conditions

Results

LaPO 4 Cs2HPO 4

MeOH/PhOH = 10

Conversion = 53 %

0 = 360~

(8 % p/p)

tc# ls. OH

OMe

LaPO 4 Cs2HPO 4 (8 % p/p)

OH

MeOH/H20/H Q = 16/7/1

Selectivity = 90 % o-cresol

10 %

Conversion = 25 % Selectivity = 94 %

0 = 330~

OH

tc# ls.

OH

HO

LaPO 4 Cs2HPO 4

Ethylene

Conversion = 100 %

(8 % p/p)

Glycol/PC = 10w1

Selectivity = 98 %

0 = 330~ tc# ls.

DISCUSSION The sum of results published in the literature on phenol alkylation using methanol are not clear and one cannot easily conclude to the relation between the acidity and the basicity of the catalyst and the selectivity in O or C alkylated products. However it seems that O-alkylation products can be obtained by the use of acidic catalysts (ref. 11). An increase of the acidity of oxide type catalysis (ref. 3) or mixed aluminium phosphate-alumina (ref. 12) gives rise to an increase of selectivity in O-alkylated products. However for strongly acidic catalysts Calkylated products, which are the more thermodynamically stable, can be obtained either by isomerization or by reaction between phenol and methylarylether. 58

On mildly acidic catalysts, C-alkylated products can also be obtained by competitive reaction on residual basic sites. This latter mechanism is generally evidenced by examining the regioselectivity of the reaction, a mechanism involving a basic site leading to the ortho isomer via a surface phenolate. When strongly basic catalysts are used, C-alkylated products are mainly formed. O and C-alkylation mechanisms necessitate the cooperation between acid and basic site in order to activate at the same time phenol and methanol. With this results in mind, we believe that it is now possible to explain the observed datas. The catalysts having shown the best activity are acidic catalysts. The niobium phosphate has a high activity (ref. 13) that may originate from the low selectivity in O-alkylated products at high reaction temperature. Rare earth phosphates calcinated at 700~ (ref. 10) and boron phosphate which have a medium acidity correlate well with a good selectivity in O-alkylation. In the case of rare earth phosphate calcinated at 400~ the decrease in selectivity while increasing the reaction temperature is due to the presence of basic sites on the catalyst. On the other hand, the good to excellent selectivity in O-alkylated products observed with cesium hydrogenophosphate alone or supported is difficult to explain. Indeed this catalyst is exclusively basic. The absence of C-alkylated products by the reaction between guaiacol and methanol tends to suggest that the guaiacolate intermediate at the surface has a different behaviour on lanthanum phosphate as well as lanthanum phosphate doped with cesium hydrogenophosphate.

59

R

OH

~

B~(~)~OCH3

OH I

~OCH3

R

o•R

H\o/CH3 [ ~ R

path A

A~

H(E) MeOI--I

H|

R

"-

R R

pathBr_

Ho

~

CH3~o/H e

!e

MeOI-I = I A ~ _ ~

@

H20

%oc.3

OCH3 H20

,

R

Scheme 1. Proposed mechanism for the basic catalysis alkylation of phenol

In addition, cesium hydrogenophosphate has an interaction with the rare earth phosphate as it was demonstrated by further characterization. The mechanism in basic catalysis goes via a phenolate anion that can be chemosorbed either by an acidic site (path A) or on a neighbouring Lewis acid site (path B). Methanol will be activated by an acid site and therefore will be able to react in O or C alkylation. The C alkylation will be favoured as the acidic site chemosorbing the phenolate will be harder. For example, the excellent selectivity in O-alkylated products observed on cesium hydrogenophosphate is due to the softness of the cesium ion,

60

(A e = Cs 9 in path B), while path A seem to be favoured (A 9 = La3e) on LaPO4 calcinated at 400~

CONCLUSION In this study, we have demonstrated that boron,

niobium and rare earth

phosphates are excellent catalysts for the selective O-alkylation of pyrocatechine in guaiacol and veratrole.

The reaction is conducted in the vapor phase using

methanol as the alkylating agent.

In the case of rare earth phosphates

the

calcination temperature has a very important effect on the selectivity of the reaction. This phenomenon,

linked to the synthetic procedure,

is due to the

residual basicity on the rare earth phosphate calcinated at 400~

Wetness

impregnation of rare earth phosphate by cesium hydrogenophosphate give rise to very active as well as selective catalysts.

References

1. Kirk Othmer Encyclopedy, 4th Edition, John Wiley, Chap. 9, p. 860, New York, (1995). 2. J. March in "Advanced organic chemistry", 4th Edition, John Wiley, pp. 386-387, (1992). 3. T. Kotanigawa, M. Yamamoto, K. Shimokawa, Y. Yoshida, Bull. Chem. Soc. Jpn. 47, 950, (1974). A.B. Mossman, US 4.611.084 (25/11/1985), (to AMOCO Corp). D. Farcassu, US 4.487.976, (30/08/1982), (to EXXON US). 4. F. Nozaki, I. Kinuira, Bull. Chem. Soc. Jpn, 50, (3), 614, (1977). 5. P. Pierantozzi, A.F. Nordguist, Appl. Catal., 2!, (2), 263, (1986). F.M. Bantista, J.M. Campelo, A. Garcia, D. Luna, J.M. Marinas, A. Romero, J.A. Navio, M. Macias, Appl. Catal. A, 99, (2), 161, (1991). M. Marczewski, G. Perot, M. Guisnet et al. (Eds), in "Studies in surface science and catalysis", p. 273, Elsevier (Science Publishers BV ?), Amsterdam, (1988). E. Fischer, O. Skiner, G. Wih, Wiss. Z. Univ. Rostock, Naturwiss, Reihe, 39, (7), 67, (1990). V. Durgakumari, S. Narayanan, L. Guczi, Catal. Letters, 5, 377, (1990). G.A. Olah, J. Kaspi, J. Org. Chem., 43, 16, (1978). Y. Shioni, Y. Nakamura, T. Manabe, S. Furusaki, M. Matsuda, M. Saito, EP 509927, (16/04/1992), (to Ube Industries). S. Furusaki, M. Matsuda, M. Saito, Y. Shiomi, (03/04/1991), (to Ube Industries). S.P. Bhatnagar, A. Prakash, S.C. Misra, M.S. Raiker, IN 158895, (18/11/1983), (to Reckitt and Colman). 10. A.M. Le Govic, A. Aubry, L. Gilbert, P. Pommier, M. Janin, following paper in this issue. 11. F.M. Bautista, J.M. Campelo, A. Garcia, D. Luna, J.M. Marinas, A. Romero, Applied Catalysis A: , 99, 161, (1993). 2. E. Santa Cesaria, D. Grasso, Applied Catalysis, 64, 83, (1990). 3. A. Florentino, P. Cartraud, M. Magnoux, M. Guisnet, Applied Catalysis, 89, 143, (1992). R.L. Martins, W.J. Schitine, F.R. Castro, Catalysis today, 5, 483, (1989). .

CATALYSIS BY RARE E A R T H P H O S P H A T E lII. C H A R A C T E R I S A T I O N OF SAMARIUM P H O S P H A T E AND SAMARIUM P H O S P H A T E - C E S I [ ~ H Y D R O G E N O P H O S P H A T E AS KEY CATALYSTS FOR O - A L K Y L A T I O N OF P H E N O L S

ANNE-MARIE LE GOVIC a~, PASCALE POMMIER a~, ALAIN LAURENT GILBERT b~AND MARCELLE JANIN b~

AUBRY a~,

a) Rh6ne-Poulenc Recherches, Centre de Recherches d'Aubervilliers, 52 rue de la Haie Coq. 93308 Aubervilliers Cedex, France b)Rh6ne-Poulenc Industrialisation, Centre de Recherche, d'Ing6nierie et de Technologie, 85 Avenue des Fr~res Perret, BP 62, 69192 Saint-Fons Cedex, France

SUMMARY Samarium phosphates, impregnated or not by cesium hydrogenophosphate, selective catalysts for O-alkylation of phenols, have been characterised by various techniques. This study has shown that : the cesium salt added by wetness impregnation (10 % w/w) has a sintering effect on its calcination. The examination of structural and textural datas shows that the cesium does not enter the crystalline network. The cesium salt is uniformly distributed on the crystalline surface and the special morphology of samarium phosphate makes the cesium retained in the porosity of the solid. - Samarium phosphate has an intrinsic acidic activity which can only be observed on products calcinated at a temperature of 700~ and which therefore possess a -

monoclinic structure. Samarium phosphate calcinated at lower temperatures, with an hexagonal structure has acido-basic characteristics highly dependant on the synthetic route use : - a totally basic activity is observed for samarium phosphate being neutralized with ammonia after precipitation.

62

products which have not been treated according to the previous step present an acidic activity -

- the addition of cesium by wetness impregnation on the dry catalyst produces a totally basic behavior.

INTRODUCTION Even if the use of rare-earth phosphates in heterogeneous catalysis for fine chemicals has been reported from more that 20 years, those catalysts were little characterized. Recently, doping those catalysts with cesium salts has greatly improved the activity, selectivity of the transformation as well as the life time of the catalysts (refs. 1,2). Particularly, a synergism between cesium hydrogenophosphate and samarium phosphate has been observed for the O-alkylation of dihydroxy-benzene (ref. 2). We described in this paper some characterizations of this solid doped or not, that may allow to explain catalytic results.

EXPERIMENTAL Samarium phosphate was prepared by wet synthesis starting from samarium carbonate (Sm2(CO3)3, originated RP). Precipitation of phosphate by phosphoric acid is conducted at 80~ by addition of a samarium carbonate suspension in a vessel containing phosphoric acid. After the end of the addition, the solid could be treated by ammonia at pH - 9. Cesium hydrogenophosphate is introduced by wetness impregnation of the dried (110~ solid. Transmission microscopy is realized using a Philips CM30 apparatus at 300 KV. DRX spectra were realized on a diffractometer Philips 1700 by scanning between 5 to 70 ~ at l~ Porous distribution is determined by mercury intrusion after elimination of gas over night at 200~ in an oven (Autopore II 9220 V3.01). Acidobasic properties characteristics of solids were estimated by studying the reactivity of 2-methyl 3-butyn-2-ol (MBOH) (ref. 3).

RESULTS AND DISCUSSION I n f l u e n c e of th e c e s i u m salt on t h e r m a l stability

We have compared the specific surface area thermal evolution of samarium phosphate just dried or samarium phosphate impregnated by cesium hydrogeno63

phosphate (10 % w/w) between 400~ and 800~ Results are reported in Figure 1 and Table 1 and show a sintering effect of the cesium salt.

Table 1. Specific surface area of SmPO4, SmPO4, Cs2HPO4 as a function of calcination temperature SmP04 (m2/g)

SmPO4, Cs2HPO 4 (m2/g)

dried

124

91

400

107

84

500

97

67

700

39

10

Temperature (~

140 ;pecific surface area 9 (m2/g) 120 100 80

--0--

60

SmP04

SmP04, Cs2HP04

I

40 20

200

,

|

.

,

,

|

300

400

500

600

700

800

Temperature (~

Fig. 1. Specific surface area of SmPO4" SmPO4, Cs2HPO 4 as a function of calcination temperature

64

~ T I V E

I~IOll

vl

+ tnil'mltofl.

I I -i

0I~4ETER

9e x t ~ m l l ~

IHii-i t-

.....

__

_..,

.

.

.

0.:'

O.l

..... iillili I1 i'~"il~ i.~ ,

-,

......

. i

l!t~'--,r ..... i ~1"t-I-i~;--!

-.' .... tbt4.-t-,~-/~if.bt-l-i-i .... ! _i /!lit 1/i!Iilili ]

9

l

i iO

,, !i

:i

roLL._!.......~--

100

~ --

.

-

I-

( licromtt~s

~ ii'i _L~_L.., O.Ol

O.

i

OIiUtr[l~

~io

•

.

i!ll ! /IIIIILL~__$___!

.4

'..

i

)

Porous repartition of SmPO4

"3

CUHUt.ATIVE IHTRU6IOH + tntJ~ltOn

vl

OIAHETEH

m ixt.r~ston

' ,.I_4~H_~L y i

, i l t I i.~. ~il]ill

ili]i

i ! t

1.,11, i I

I [tNitl! i [ i,li!!t, ,l!I tt ,~,~i[!

I !l!i]li I-! -~--~l!Ii-[~ii i iii!i!'il !, !it~i!t!i!l .

O.II

0.1

il

II

I

~

'

t0

ti

.. .tiltlii

1!1:! ~tli~li liilli

'

!t:~

ilil .

t

i!i[!li

.~!.!!!!!, ~!i!i!l i .. o. !

1

oz.uq~n~.

Fig. 3.

t

LIiLLLl~

l

l!il i !/

111,~,~ I :

Porous repartition of SmPO4, Cs 2 HPO4

65

( eicroutar~

)

0.01

,

For the same thermal treatment (fixed duration and temperature), specific surface area of SmPO4, Cs2HPO4 are systematically weaker of about 20 m2/g (or even 30 m2/g) than those of SmPO4. On an other hand, we have checked that the ammonia treatment has no effect on specific surface area. Those results have been maked up, in the case of solid calcinated at 500~ by porosity measuremem and by electronic microscopy analysis. 9 Poro$imetry

9 samples expand a porous volume of 0,69 cm3/g with a

microporous volume of 0,20 cm3/g. Introduction of cesium lower the porous volume without any change in the porous repartition (Table 2, Figs. 2 and 3).

Table 2. Porous repartition of SmPO, and SmPO4, Cs2HPO4 Porous volume Medimn pore Medium pore between 0,1 rn diameter (m) diameter (A) and 37 A

Catalyst

Total porous voltt3Ine (cm /g)

Porous volume between 30 ~,d 0,1 m (cm/g)

SmPO4

0,69

0,45

0,20

15

65

SmPO4, Cs2HPO4

0,56

0,40

0,16

12

7o

Electronic microscoov The sintering effect of the cesium salt has been made visible by transmission electronic microscopy. Comparison of electronic microscopic stereotypes of product calcinated at 500~ (6 h.) without (Fig. 4) or with Cs2HPO4 (Fig. 5) leads to the following remarks 9 SmPO4 is formed of agglomerated polydispersed small stick of size between 10 -

and 100 nm. In the case of SmPO4, the periphery of those sticks is well defined (frame -

bones). - In the case of SmPO4, Cs2HPO4, the periphery of the sticks is badly defined and they are linked by amorphous zone of molten aspect enriched in cesium, as shown on cartography analysis (Fig. 6). The STEM-EDS cartography analysis does not show if the cesium is uniformly widespread on each cristallite or if it creates a solid solution in the cristalline structure of the phosphate. It is the reason why a DRX structural analysis was realized.

66

Samarium phosphate precipitates as an hexagonal phase and shows a phase transition between 600 and 700~

to form a more closed monoclinic phase. Those

results are in good agreement with schneider's published datas (ref. 4) reviewing crystalline structure studies about rare earth orthophosphates. Rare earth orthophosphates can be subdivided into several families according to their crystalline structures and the polymorphic modification as a function of the temperature. The first family regroups light rare earth (so called ceric phosphates) including the following elements : La, Ce, Pr, Nd, Sm, Eu. Those phosphates are dimorphiques. Indeed, they precipitate at low temperature under hexagonal phase and evoluate at higher temperature to the thermodynamically stable phase, the monoclinic one, isomorphic to monasite CePO4. The phase transition temperature is accompanied by an exothermic phenomena linked to the cation ionic radius and is higher as the cation radii is lower (ref. 5). 9 DRX DRX studies of SmPO4 and SmPO4, Cs2PO4 calcinated at various temperatures (between 200 and 800~ show that cesium has no visible effect on crystalline structure of products : the crystalline phase transition (from hexagonal to monoclinic) occurs between 600 and 700~ independently on the presence of cesium (Table 3) - the mesh parameters are similar for SmPO4 and SmPO4, CszHPO4 (Table 4) DRX shows no formation of cesium pyrophosphate which is usually obtained as -

early as 300 ~ The complete analysis of the crystalline structure by DRX and EXAFS of impregnated structure shows that the cesium does not enter the crystalline network in SmPO4. The comparison of those results and the electronic microscopy analysis leads to the conclusion that the cesium is uniformly distributed on crystallite surface and that the excess of cesium is retained in the porosity of the solid, probably as amorphous cesium phosphate.

67

Fig 5. Electronic microscopy of

SmPO4,CsHPO 4 68

Fig. 6-2.

STEM - EDS cartography of Sm PO4, Cs2HPO 4 | localisation of P 69

Fig. 6-3.

STEM - EDS cartography of Sm PO4, Cs2HPO4 | localisation of Cs

ACIDO-BASIC PROPERTIES Characterization of SmPO4 We have examined the influence of surface chemistry of SmPO4 on its acidobasic properties. Characterization by reactivity of MBOH was realized for product treated at pH = 9 with an ammoniacal solution or not. The MBOH test permits to determine without any doubt the acido-basic characteristic of surface site. We have reported in Table 5, methylbutynol conversion at the 12th pulse and the acidic (A), basic activities (B) and activity due to acid base pairs (B) obtained for each samarium phosphate. The evolution of acido properties - conversion of MBOH and selectivity in the various products formed as a function of preparation methods and calcination temperatures are reported Figures 7 and 8. 9 The acidic selectivity is the sum of selectivity in 3-methyl 3-buten-l-yne (MBYNE) and in prenal which are formed on acidic sites. 9 The basic selectivity is the selectivity in acetone or acetylene which are formed on basic sites.

70

9 The acid base pairs selectivity is the sum of selectivity in 3-hydroxy 3-methyl butan-2-one,

in 3-methyl 3-buten 2-one and methylisopropylketone which are

formed supposedly to be on acid-base pairs.

Table 5. Acidobasicity of samarium phosphate determined by reaction of MBOH Calcination

Conversion of

temperature

MBOH (%)

Catalyst

A

B

AB

dried

11

94

1

5

500~

14

97

1

2

700~

30

98

1

1

dried

27

2

98

-

treated by an

500~

24

7

93

-

ammoniacal solution

700~

91

8

1

SmPO4

dried

0,7

_

_

_

Cs2HPO4 (10 % p/p) ,,

500~

0,5

SmPO4,

dried

32

100

Cs2HPO4 (10 % p/p)

500~

39

100

S m P O

S m P O

4

4

(treated by an ammoniacal solution) If we compare products calcinated at 500~

SmPO4 without treatment has a

totally acidic behaviour while the sample treated at pH = 9 as a totally basic behaviour. The basic behaviour observed for SmPO4 treated at pH = 9 indicates the presence of residual anions coming from the ammoniacal neutralization step. When calcinated at 700~

both products present comparable behaviour with an

higher acidity for the phosphate not treated. At 700~

we find the intrinsic acidic

activity of samarium phosphate. SmPO4, Cs2HPO4 was also characterized by the MBOH test. Results reported in Table 5 show that the presence of the cesium salts exalt the surface basicity. The addition of the cesium salts induces a totally basic like behaviour of this catalyst. The observed difference in activity should be interpreted with some caution due to the high basic activity of cesium oxide, the presence of which, even in small quantities, can not be excluded.

Conversion of MBOH (%)

40-, 3O

A w

--43- SmPO4 without treatment

,.,

20-

SmPO4 treated at pH = 9

100~ [ i

0

100

200

300

400

500

600

700

Calcination temperature of SmPO4 (~ A

4030 J

- - u - SmPO4 without treatment 20~, SmPO4 treated at pH = 9

L

10~ ~

,

v

0' 200

100

300

400

500

600

700

Calcination temperature of SmPO4 (~

A 9activity = mmol MBOH transformed per surface unit and per hour Fig. 7. Acido-basic properties of samarium phosphate

SmP04 Selectivity (%) 100 80

60

--{3----o- % MBYNE (Acidity) ~

I,

40

~

2o

~ 100

% Prenal (Acidity)

200

300

400

500

600

700

Calcination temperature of SmP04 (~

72

SmPO4 treated at pH = 9 Selectivity (%) 100 80 -i L

60

- - o - % Acetylene (basicity) i

40 ~

- - I - % MBYNE (acidity)

% Acetone (basicity) r

20O~ _20100

:

% Prenal (acidity)

v

200

300

400

500

600

700

Calcination temperature of SmPO 4 (~

Fig. 8. Acido-basic properties of samarium phosphate 9selectivity on each catalyst

CONCLUSION This study leads to the following conclusions. Cesium salt added by wetness impregnation (10 % w/w) has a sintering effect on the calcination of samarium phosphate. The examination of structural and textural data shows that the cesium does not enter the crystalline network. The cesium salt is uniformly distributed on crystallites surface and the special morphology of samarium phosphate makes the cesium retained in the porosity of the solid. Samarium phosphate has an intrinsic acidic activity which can only be observed on products calcinated at 700~ and therefore with a monoclinic structure. Samarium phosphates calcinated at a lower temperature, with an hexagonal structure has acido-basic characteristics highly dependant on the synthetic route used - a totally basic activity is observed for samarium phosphate being neutralized with ammoniac after preparation - products which have not been treated according to the previous step present an acidic activity the addition of cesium by wetness impregnation on the wet product gives it a -

totally basic activity.

73

References 1. P.J. Tirel, C. Doussain, L. Gilbert, M. Gubelmann, H. Pernot, J.M. Popa, Studies in surface science and catalysis, 78,693, (1983) 2. L. Gilbert, M. Janin, A.M. Le Govic, P. Pommier, A. Aubry, Preceeding paper in this issue 3. H. Lauron-Pernod, F. Luck, J.M. Popa, Applied Catalysis, 78,213, (1991) 4. L. Niinist6, M. Leskelii in "Handbook on the Physics and Chemistry of rare earth" F.A. Gschneider, J.R. Eyring, L. Eyring Eds., Vol. 9, Chapter 59, p. 91. 5. R. Kijkowna, Nieorg. Zwiazki Fosforowe, 7,239, (1976)

74

SELECTIVE FUNCTIONALISATION ORGANOSILICON INTERMEDIATES

OF

FLUOROAROMATICS

VIA

BERNARD BENNETAU a), PIERRE BABIN b) AND JACQUES DUNOGUES a) a) Laboratoire de Chimie organique et organom6tallique (URA 35 CNRS), Universit6 Bordeaux I, 351, Cours de la Lib6ration, 33405 Talence, France b) Laboratoire de Pharmacie chimique, Universit6 Bordeaux II, Place de la Victoire, 33000 Bordeaux, France

INTRODUCTION The importance of fluorinated organic compounds is demonstrated by the increase in the number of published novel compounds prepared during the last two decades. This fact reflects the interest of scientists, both academic and industrial, in utilizing fluorine to modify the physical and chemical properties of organic compounds. The introduction of fluorine increases thermal and oxidative stability, lipophilicity and also closely mimics hydrogen in particular from a sterical point of view. These properties range from the high stability of fluorinated polymers to the enhanced properties of agrochemicals and phamaceuticals. For instance, organofluorine compounds have been used as lubricants, refrigerants, fire extinguisher agents, inhalation anesthesics and surfactants. Otherwise, many fluoroaromatics find wide use in biomedical applications (ref. 1), agrochemicals and pharmaceuticals (ref. 2) because their efficacy is improved by the presence of fluorine (lower dosage, lower toxicity and increased selectivity). The regioselective functionalisation of fluoroaromatics or the selective introduction of fluorine into aromatic molecules under mild conditions are of great importance to the chemical industry and are a considerable challenge to organic chemists. So, the regio- and stereospecific requirements have created needs for developing special reagents and procedures; many strategies to introduce fluorine into a fluorinated aromatic ring or to introduce substituents into fluoroaromatics have been developed.

75

The aim of the present account is to provide comprehensive, if not exhaustive, highligths of the selective functionalisation of fluoroaromatics, and in a few cases, electrophilic fluorination of arylsilanes. E L E C T R O P H I L I C CLEAVAGE REACTIONS OF CARBON SILICON BONDS IN FLUORO-ARYLSILANES WITH OR WITHOUT FRIEDELCRAFTS CATALYSTS Eaborn et al. (ref. 3) have shown that the displacement of the trimethylsilyl moiety in aryl(trimethyl)silanes by electrophiles was analogous to that of hydrogen in electrophilic aromatic substitution :

iMe3

RO

iMe3 E+ -..

Nu"~

R

slow

~

R

0E

+ Me3Si~Nu

fast

Scheme 1.

Moreover, the efficacy of the ipso factors for a series of electrophilic desilylation processes gave rise to the expectation that aryl(trimethyl)silanes might be successfully employed for synthetic purposes; this was confirmed by many works reviewed in 1993 (ref. 4). With fluoroaromatics, the ipso effect of the trimethylsilyl group was involved for regiospecific electrophilic substitutions. For example, the increasing sophistication of nuclear medicine techniques has presented challenges to the synthetic chemist involved in the preparation of imaging agents labeled with radionuclides. In order to compare their utility as substrates for regiospecific aromatic halogenation, some para-substituted aryltrimethylsilicon,-germanium, and -tin compounds were treated with no-carrier-added (n.c.a.) 77Br and 131I (ref. 5). Results are summarized in Scheme 2 :

76

Radiochemical yield (%) F

F

MMe3

.A.

77Br

131I

Si

18

Ge > Si. On the other hand, only few examples of direct electrophilic fluorination of aromatics have been reported since the ability of fluorine to behave as an electrophile

is not

easily

achieved;

however,

radiofluorination

of

aromatic

compounds has been described but, in the reactions which have been reported, it is apparent that strong electron-donating groups are required on the aromatic ring when mild fluorinating reagents are used. Nevertheless, without activating groups, an alternative route to 18F-labeled radiopharmaceuticals, using 18F-labeled reagents has

been

proposed,

involving

arylsilanes

(ref. 6).

The

scope

of

this

fluorodemetalation reaction as well as the influence of the metal displaced and of aromatic substituents has been studied (ref. 7). The reaction is given and the yields mentioned in Scheme 3. Chemical yield (%) Y

Y i CF3

MMe3

18F

i) [18F]-F2 or [18F]-CH3CO2F M = Si, Ge, Sn Y = F, CF3 Scheme 3.

77

Sn

Ge

Si

74

56

30.5

35

10.5

2.5

As previously noted (ref. 5), yields are lower when arylgermanium or -silicon derivatives are used. However, for activated aromatic systems, it was pointed out that aromatic halodesilylation proceeds under convenient conditions and arylsilanes being less expensive than their germanium and tin analogues and much less toxic than the corresponding aryltins. Another example of electrophilic fluorination of fluoro(trialkyl)silanes by acid-catalyzed metal-metal exchange is given below (ref. 8): X

X

X

ArF (%) 43

CF3 SiMe3

180~ Temperature

The temperature influences both the reactivity of the phenate and the selectivity, as shown in Table 3.

122

Table 3. The influence of temperature 9carboxylation of potassium phenate Selectivities (%)

T (~

Yield (%)

,,,

140 190 210 250

A.S. 59 29 6

pHBA 41 70 94 > 98

39 43 48 > 48

Conditions 9 Bulk carboxylation Duration 9salification 1 h30mn Carboxylation 4h P (CO2) = 1 bar These values show clearly the thermal restructuring of bipotassium salicylate in parahydroxybenzoic acid, as shown in figure 3. C O 2 pressure The minimum required pressure for the carboxylation of phenate corresponds to the dissociation pressure of the complex formed between the phenol, the alkaline metal and the CO2 molecule, and to the applied temperature (ref. 5). Between 120~

and 160~

the dissociation pressure of the complex (sodium phenate CO2)

is around 4 bars. At a given temperature, an increase in CO2 pressure, above this minimum value has no effect on the system selectivity, but it can favour the reaction kinetics. The role of water

The presence of traces of water leads to a reduction in yields (Table 4). Table 4. The influence of residual water in the phenol / phenate mixture on the carboxylation yield H20 content (ppm)

Yield loss (%)

5O0 5 000

1.25 12.5

In fact, any molecule of water present, breaks down the phenate into a molecule of alkaline hydroxide, which then consumes some CO2, to regenerate water (eqns. 13-15).

123

OM

OH

2 ~

+2

2 MOH

+

CO 2

H20

,~.

+ 2 MOH

"~

(13)

M2CO3 + H20

OM

(14)

OH

+H20+CO22

+ M2CO 3

(15)

The water can also form in situ by 9 2 MOH

+ CO 2

~,

"~

M2CO3

+

H20

(16)

Ueno (ref. 5) draws attention to the fact that the production of water can also be due to a parasite etherification process (eqns. 17, 18). OH

~

O - - ~

+

H20

(17)

OH + CO 2 +

2 H20

(18)

II o

The influence of the system The carboxylation reaction of phenol requires total anhydricity. This condition can be fulfilled : - either in solid phase : good selectivity will be achieved is the gas / solid exchanges are efficient. These conditions requires the use of special technology.

124

- o r in suspension in an inert solvent (biphenyl ether, kerosene, di-or terphenyl, etc.). - or in a phenol phase system.

SYNTHESIS OF OTHER HYDROXYAROMATIC COMPOUNDS The influence of substitute core groups

The reactivity of the substrate is influenced by the steric space requirement and the core substitute groups. - An alkyl group (either ortho or para) favours the ortho-carboxylation of the phenol group. - A donor group (NH2, OCH3, OH, X, etc.) enables good yields to be achieved ( > 80%). An acceptor group (NO 2, CN, COOR, etc.) will inhibit the reaction. The differences in reactivity and orientation are summarised in Table 4.

-

125

Table 5. The influence of substituents (ref. 6) Substituents

Majority product

Substrate

HO--@R

Conditions

H O - ~ R HOOC

Yield

Ref.

t ~ = 160- 220~ P = 4 0 - 100 b dur6e = 4 - 8 h

25 h 85 %

(7) (8) (9) (10) (11) (12)(13)

t ~ = 125- 175~ P = 100b dur6e=4-8h

70~83 %

(7) (8) (9) (12) (13) (14) (15) (16)

t ~ = 200~ P = 40b dur6e = 6 h

37 %

(13)

t ~ = 200~ P = 40b dur6e = 6 h

3O %

(7) (13) (17) (18)

t ~ = 210oC P = 35 b dur6e = 40 h

17 %

(18) (19) (20)

t ~ = 250~ P= 30b dur6e = 5 h

90 %

(7)

t ~ = 180~ P=63b dur6e = 5 h

84 %

(7) (21) (22)

t~ =90-225~ P = 8 - 100b dur6e = 4 - 43 h

5-90 %

(7) (13) (23) (24) (25) (26)

t ~ = 210~ P =40b dur6e = 4 h

0-19%

HOOC Alkyl (CH3,

_ ~ OH

C2H5"" ")

R

R R

R

R

R'

R

HOOC

R'

HOOC HO-~R' R

Phenyl

HO@~R'

R

.o-@~ HOOC

HOOC Donor (NH2, OMe, OH, X...)

Ho- G HOOC ~

Acceptor (NO2, CN, COOR...) HOOC

126

(13)

CONCLUSIONS The carboxylation reaction of phenol by CO2 is well known and industrially developed using various technologies. Chemically speaking the key parameters are shown in Table 6. Selectivity

Key parameters Cation P (CO2) T~

ortho

para

Na

K

equilibrium shift 130 + 50

210 + 30

a

Three main technology types can be used : - fluidised bed, ,, LIST ,, reactor, -

in a dispersion system,

- bulk.

References

1. ,, Liquid crystal polymers ,,, SRI, report N~ 86 C. 2. J. March in ,, Advanced Organic Chemistry ,,, 3ieme Edition, John Wiley, p. 491-492, NewYork, (1985). 3. R. Schmitt, J. Prakt. Chem., 397, (1885). 4. A.J. Rostron, A.M. Spivey, J. Chem. Soc., 39, (1964). 5. Ueno Ryuzo, Masada Yoshiyasu, EP 254 596, (1986). 6. A.S. Lindsey, H. Jesrey, Chem. Rev., 583, (1957). 7. O. Bame, G.F. Adamson, J. Org. Chem., 19, 510, (1954). 8. F. Beilstein, A. Kuhlberg, Ann., 156, 206, (1870). 9. B.I.O.S. Final report N ~ 664, His Majesty's Stationery Office, London. 10. D. Cameron, H. Jeskey, J. Org. Chem., 15,233, (1950). 11. R. Ihle, J. Prakt. Chem., 2 (14), 443, (1876). 12. P. Spika, Gazz. Chim. Ital., 8, 421, (1878). 13. F. Wessely, K. Benedikt, Monatsh, 81, 1071, (1950). 14. C. Brunner, Ann., 351,320, (1907). 15. A. Engelhardt, J. Russian. Phys. Chem. Soc., 1, 220, (1869). 16. M. Filiti, Gazz. Chim. Ital., 16, 126, (1886). 17. R.C. Fuson, J. Corse, J. Amer. Chem. Soc., 63, 2645, (1941). 18. L. Palfray, Bull. Soc. Chim., 956, (1948). 19. J.A. Jesurun, Ber., 19, 1414, (1886). 20. H. Kolbe, E. Lautemann, Ann., 115,201, (1860). 21. Heyden, Fabrik, German patent 61125, (1891) ; Frdl 3,828. 22. H. Schwazz, Ber., 13, 1643, (1880). 23. M. Calvin, US 2 493 654, (1950) ; CA. : 44, 2559, (1950).

127

24. L.N. Ferguson, R.R. Holmes, J. Amer. Chem. Soc., 72, 5315, (1950). 25. V.M. Rodionov, Bull. Acad. Sci. U.R.S.S., Classe Sci. Chim., 3 (421), (1940) ; C.A. : 35, 5101, (1941). 26. L. Varnholt, J. Prakt. Chem., 2 (36), 19, (1897).

128

ACCESS TO POLYCHLOROPHENOLS : CHEMISTRY OF INTERMEDIATES

JEAN-ROGER DESMURS a), SERGE RATTON b), RENE JACQUEROT a), JEAN DANANCHE a), BERNARD BESSON a) AND JEAN-CLAUDE LEBLANC ~ a) Rh6ne-Poulenc Industrialisation, Centre de Recherche, d'Ing6nierie et de Technologie, 85 Avenue des Fr6res Perret, B.P. 62, 69192 Saint-Fons Cedex, France. b) Interm6diaires Organiques, 25 quai Paul Doumer, 92408 Courbevoie Cedex, France. c) Rh6ne-Poulenc Chimie, B.P. 17, 38800 Le Pont-de-Claix, France.

INTRODUCTION U.S. production between 1977-1980, shown in Table 1, illustrates the industrial importance of these aromatic derivatives, used either as synthesis intermediates, or active components for some agrochemical compounds or pharmaceuticals specialities.

129

Table 1. Production levels and uses of chlorophenols (ref. 1) Product

US production t/year

Uses

a few thousands tons

intermediates

23000

intermediates

19000

fungicide used in wood protection formulations

OH ~ C 1

OH CI

C1 OH C1

C1

C1

The agrochemicals sector provides the main outlets for chlorophenols, as shown in Table 2 which shows a few well-know products.

130

Table 2. Shows agrochemical and pharmaceuticals specialities prepared from chlorophenols. Structure

Trade-mark

C1 24 D Acid (herbicide)

I~~-O--CH2COOH C1 ~ H3 C1----~,,),/x----O--CH--COO H

Dichloroprop (herbicide)

C1 /-.~ ~ CH3 C1 - - - - ~ . ) / ~ O--~~)//x--- O-- CH-- COOH3

Hoelon (herbicide)

C1 Bifenox (herbicide)

CI-~O~~N~--NO2 COOCH3 CI c3.7

C1-----{(

) )---O--CH-~--CH2--N

\~_;(

-

~C1

_

,c_N.~N

O//

Sportak (fungicide)

\--/

C1 // O~CH2~CH2~NH--NH~CxNH2

\ -(X

Guanochlor (antihypertensor)

C1 CI

C1

NH--

Diclofenac (analgesic)

CH2CO2H

131

Generally a (trichlorophenols,

quality problem is created when heavy chlorophenols tetrachlorophenols and pentachlorophenols) are processed.

Impurities such as polychlorophenoxyphenols, polychlorodibenzodioxins, polychlorodibenzofurans and polychlorodihydroxybiphenyls, are quite often found in some industrial products. Table 3 below shows the influence of the structure of the series of dibenzodioxins on the DL 50 (refs. 2, 3) (number and position of the chlorine atoms). Table 3. Shows LD 50 by mouth of various polychlorodibenzodioxins. DCDD

TCDD HCDD (2,3,7,8 isomer) (mixed isomers)

OCDD

r j 0.114 mg/kg

mouse

0.022 mg/kg

J

rat

> 1000 to 2000 mg/kg

guinea-pig {

J

0.045 mg/kg

> 4000 mg/kg

100 mg/kg > 1000 to 2000 mg/kg

0.0006 mg/kg 0.0021 mg/kg

J rabbit

I

0.115 mg/kg (0.275 mg/kg) - through percutaneous route > 0.252 mg/kg through intraperitoneal route

2,3,7,8-Tetrachlorodibenzodioxin 1 has a toxicity clearly above that of all the other compounds of this series. Thus it is imperative that industrial products are free of this impurity or its precursors.

132

Literature on this subject tells that chlorophenols cause these unwanted byproducts to form (refs. 4, 5, 6) when exposed to thermal, photochemical or basic conditions (eqn. 1). OH + C1

Cim

~

C

H

CI~ + 2HCI (1) ~

"O"

But information provided in the literature is not sufficient to explain how these by-products, observed during experiments carried out on chlorination reactions, are formed. Consequently our objective was to study chemical phenomena occurring during the phenol chlorination process in order to understand how by-products were formed with the aim of reducing the amount formed.

HYPOTHESIS It seems that the creation of polychlorophenoxyphenols, polychlorodibenzofurans or polychlorodibenzodioxins is closely related to the varying degrees of chlorination. During the mono and dichlorination of phenol, no parasite chemistry appears. But if we introduce a third and especially a fourth or fifth chlorine atom within the aromatic core then a problem is created. This is problably linked to the high activating quality of the OH group which tends to orientate an electrophilic substitution - in an ortho or para positioning (Fig. 1).

"x

cl

OH

cl

Cl

OH

cl

Fig. 1. Activated position towards an electrophilic attack of chlorine When the carbon has already been substituted by an atom, ipso attack leads to some polychlorinated

gem-dichlorocyclohexadienones (Fig.

133

2).

0

0 C1

0

CI~.~C1

C I ~ C I

C1

C1

0

0 ,C1

C

C1

C1

~1

"1

Fig. 2. Polychlorinatedgem-dichlorocyclohexadienones generated by chlorination of 2,4 or 2,6dichlorophenols or 2,4,6-trichlorophenol We assumed for this process that gem-dichlorocyclohexadienones were actual intermediates acting in the formation of chlorophenols (eqn. 2) containing one or two atoms of chlorine in meta position relative to the OH group. 0

C1

OH

.C1

.~

C I ~ C 1 (2)

isomerisation y

"C1

C1 but they could also be the precursors of polychlorophenoxyphenols (eqn. 3)

0 C1

OH CI+CI

CI

0 .9 . C I ~ C I condensation "- ~CU. ~"

C1 + HCI

OH 9 _~ C I - . ~ C 1 (3)

C1

y o

CI

CI

CI

Cl

134

"C1

or polychlorodibenzodioxins (eqn.3)

OH I

?

+ HCI

d

+clQ

I

CI

or polychlorodibenzofurans (eqn.5)

135

1) Red 2) -HC1

(4)

and polychlorohydroxydiphenyls (eqn. 6) O C1

OH ~~/C1

CI

OH 9

C

+

l

~

~

+ HC1 C1

(6)

k~.,.OH C1

To verify these hypotheses, we have successive by performed the following actions : - synthesising polychlorinated gem-dichlorocyclohexadienones ; - developing a method of analysis enabling characterization and quantitative analysis of the products ; - d e m o n s t r a t i n g that polychlorinated gem-dichlorocyclohexadienones existed within chlorination reaction masses, and establishing how the formation of gemdichlorocyclohexadienones and that of polychloro by-products were correlated ; - examining the reactivity of polychlorinated gem-dichlorocyclohexadienones.

SYNTHESIS

OF

POLYCHLORINATED

GEM-DICHLOROCYCLOHEXA-

DIENONES

Polychlorinated gem-dichlorocyclohexadienones can be obtained through the action of t-butyl hypochlorite (ref. 7), hypochlorous acid (ref. 8) or chlorine in acetic acid (refs. 9, 10) with a chloro phenol. OH C1

C1

O 1 I C I ~ C 1

O CI

C1 (7)

ch c1 + tBuOC1

ch c1 + tBuOH

n=0.1 or2 The methods described by P. SVEC (ref. 6) enabled us to prepare 2,4,4,6-tetrachlorocyclohexa-2,5-dien- 1-one 2, 2,3,4,4,6-pentachlorocychlohexa

136

2,5-dien- 1-one 3, 2,3,4,4,5,6-hexachlorocyclohexa-2,5-dien- 1-one 2,3,4,5,6,6-hexachlorocyclohexa-2,5-dien- 1-one 5. 0

0

C1

C1

C1

0

CI

C1

and

0

C I ~ C 1

CI.~~C1

CI"Ol./~.cI'CI~..

C1

C1 C1

2

ANALYSIS

3

OF

4

POLYCHLORINATED

5


DIENONES

As our aim was to bring to light the polychlorinated gem dichloro cyclohexadienones, within chlorination reaction masses, one of the most important parts of our study was to develop an efficient method of analysis capable of detecting traces amounts of these intermediates. The only items of interest to be found in the literature are those provided by P. SVEC and coll. fiefs. 6, 11) using gas phase chromatography in c.c.m.. P. SVEC and V. KUBELKA (ref. 11) observed that the polychlorinated gemdichlorocyclohexadienones were broken down in gas phase chromatography. This result was confirmed in the laboratory by work performed on different gas phase chromatography columns which showed almost quantitative reversion of the polychlorinated gem-dichlorocyclohexadienones back to the initial state of chlorophenol (eqn. 8). For this reason GPC cannot be used. 0 C1

OH C1

G P C - 2 0 0 to 250 ~

C1

CI

(8) C1

Some polychlorinated gem-dichlorocyclohexadienones were separated by P. SVEC (6) using thin-layer chromatography on a silica plate with eluants of hexane, cyclohexane, and benzene type. Using hexane as an eluant it was possible to transpose this separation of elements to HPLC with a silica column (Fig. 3). 137

Column : HIBAR MERCK Si 60 (5 u) ; L = 12.5 cm Hexane eluant

t N

m

h

\D

Y)

N

*I /, -

-. ?'

rr.

r-

..-.

Fig. 3. Chromatogram of the separation of polychlorinated gem-dichlorocyclohexadienonesand gem-dichlorocyclohexenones with silica column 138

These separation conditions proved difficult to handle, especially for chlorination masses, since the large quantities of chlorophenols introduced caused the base line to shift rapidly. This resulted in any quantitative analysis being impossible. These circumstances obliged us to research into other conditions using HPLC. We observed that the polychlorinated gem-dichlorocyclohexadienones were eluted (Fig. 4) when placed under the conditions that we used to quantitatively analyse the chlorophenols, but unfortunately the retention periods were interfering with those of the chlorophenols (Fig. 5).

139

,.j],

"el.,...,

7, *'2. i

-... *. :21

r. r, !

-

u-.

--.

~.~ I--

i.-.,

~4

,2.-, L.L.I

,-,-, L , J

J

~

i.."~-

r'_L~_

."

L"-: ".'-'

C':,CI

I

(..-_" o- -o

,2,-_. ;'%

,--,

~-" k - J L.~ 9 .,:-P.-..4

,., "2". :-~

_.,

Fig. 4. 2,4,4,6-Tetrachlorocyclohexa-2,5-diene-l-one on coiunm R.P. 18 e l u a n t methanol- acetate buffer pH 4.1 (78.22) 140

,.j],

"el.,...,

7, *'2. i

-... *. :21

r. r, !

-

u-.

--.

~.~ I--

i.-.,

~4

,2.-, L.L.I

,-,-, L , J

J

~

i.."~-

r'_L~_

."

L"-: ".'-'

C':,CI

I

(..-_" o- -o

,2,-_. ;'%

,--,

~-" k - J L.~ 9 .,:-P.-..4

,., "2". :-~

_.,

Fig. 4. 2,4,4,6-Tetrachlorocyclohexa-2,5-diene-l-one on coiunm R.P. 18 e l u a n t methanol- acetate buffer pH 4.1 (78.22) 140

Because similar retention periods are used for some chlorophenols and polychlorinated gem-dichlorocyclohexadienones a problem of peak times was created and it was necessary to use a double detection method enabling differentiation of both products. In order to obtain the best possible sensitivity and specificity for the polychlorinated gem-dichlorocyclohexadienones, we opted for electrochemical detection based on reduction of gem-dichlorocyclohexadienones at an imposed potential o f - 0 volt (Fig. 6). ,.C1 O"

0 C1

C1

C1

00 C1

C1 ~

C1

C1|

+ 2 e-

~~.

C1

~

+ C1| C1

C1|

The chlorophenols,

(9)

(10) cyclohexadienones,

benzoquinones

and chlorophenoxy-

phenols are analysed using UV detection. Simultaneous electrochemical detection enables specific analysis of electro active compounds in reduction. This analytical technique proved to be very efficient as only one injection provided us with all the required information on the composition of the reaction masses.

142

8.67 9.95 11.48 14.14-

2 , 3 , 6 - T r i c h l o r o p-benzoquinone Chloranil 2,4,4,6-Tetrachlorocyclohexa2,5-dien-l-one 2,4,6-Trichlorophenol

15.9218.33 19.67 -

2,3,4,4,6-Pentachlorocyclo hexa-2,5-dien-l-one 2,3,4,6-Tetrachlorophenol Pentachlorophenol

O

E O

~i!j;I ~ r. 9o

~ ~

.--.

,i

"~ .

-

.-, .

.

.

:,

"

.

.

7.

.

" 9 r.~, ..'~ ..,-"

'i , -

!

"i.

Fig. 6. Benzoquinones, cyciohexadienones and phenols mixture in UV and electrochemical detection 143

The limit of detection of polychlorinated gem-dichlorocyclohexadienones by HPLC using equipment fitted with a double detection system, UV and electrochemical, is approximately 0.01% in a synthetic chlorophenol mixture.

D E T E C T I O N OF POLYCHLORINATED GEM-DICHLOROCYCLOHEXADIENONES WITHIN THE CHLORINATION MASS Since the analytical method we used enabled us to detect down to 0 . 0 1 % of polychlorinated gem-dichlorocyclohexadienone in a chlorophenol mixture, we were able to detect this compound in a chlorination reaction mass. This confirmed some of the assumptions we had made when we started work on the subject. Detection of the polychlorinated gem-dichlorocyclohexadienones was performed by means of numerous tests and only one example of the chlorination of 2,4,6-trichlorophenol in the presence of A1C13 (ref. 12) is described below : OH C1

OH CI + C12

MC13

~ C1

i1

(11)

1

C1

C1

2,4,6-Trichlorophenol and A1C13 are heated to a temperature of 100~ Then chlorine is introduced at the rate of 5 1/hour. Samples of the product are taken at the following times : corresponding to each of the respective, introduced quantities of chlorine : 164.6 mM (equivalent 0.5) 329.2 mM (equivalent 1) 658.4 mM (equivalent 1.5) 987.6 mM (equivalent 2) The formation of 2,4,4,6-tetrachlorocyclohexa-2,5-dien-l-one is observed from the very first sample. Formation of dichlorobenzoquinone, trichlorobenzoquinone, tetrachloro-phenol and polychlorinated phenoxyphenols is also observed (Fig. 7).

144

0

c 0 .L(

a

m N C

m 0

I !

I

,

Fig. 7. Chromatogram of the chlorination of 2,4,6-trichloropheno1 after introduction of 0.1 of chlorine equivalent 145

During the 2,4,6-trichlorophenol chlorination process (Fig. 8), the concentration of 2,4,4,6-tetrachlorocyclohexa-2,5-dien-l-one remains more or less constant while a marked increase of other impurities is noted. From these results, the following observations can be made : -Chlorophenoxyphenols and other polychlorinated impurities were released when the polychlorinated

gem-dichlorocyclohexadienones were

generated in the reaction

mixture. This fact confirms our hypothesis that polychlorinated

gem-dichlorocyclo-

hexadienones are probably the cause of the parasite chemistry. - The

content

of

polychlorinated

gem-dichlorocyclohexadienones

remains

approximately constant afterwards, and this fact supports our assumption that intermediates for the reaction with one or two chlorine atoms in meta position could be used. Whilst these initial results showed us that our assumption was correct, we still had to demonstrate that the polychlorinated gem-dichlorocyclohexadienones, placed under the conditions of the chlorination process, were capable of producing the different products observed during the experiment. An in-depth study conducted on the reactivity of the polychlorinated

gem-dichlorocyclohexadienones confirmed

these facts.

146

During the 2,4,6-trichlorophenol chlorination process (Fig. 8), the concentration of 2,4,4,6-tetrachlorocyclohexa-2,5-dien-l-one remains more or less constant while a marked increase of other impurities is noted. From these results, the following observations can be made : -Chlorophenoxyphenols and other polychlorinated impurities were released when the polychlorinated

gem-dichlorocyclohexadienones were

generated in the reaction

mixture. This fact confirms our hypothesis that polychlorinated

gem-dichlorocyclo-

hexadienones are probably the cause of the parasite chemistry. - The

content

of

polychlorinated


remains

approximately constant afterwards, and this fact supports our assumption that intermediates for the reaction with one or two chlorine atoms in meta position could be used. Whilst these initial results showed us that our assumption was correct, we still had to demonstrate that the polychlorinated gem-dichlorocyclohexadienones, placed under the conditions of the chlorination process, were capable of producing the different products observed during the experiment. An in-depth study conducted on the reactivity of the polychlorinated

gem-dichlorocyclohexadienones confirmed

these facts.

146

REACTIVITY

OF


POLYCHLORINATED

DIENONES

Since 2,2,4,6-tetrachlorocyclohexa-3,5-dien-l-one changes very rapidly, when in chlorophenol medium, to give 2,4,4,6-tetrachlorocyclohexa 2,5-dien 1-one (eqn. 12), we deliberately restricted the scope of our study to only deal with this last compound. This was because we wanted to avoid analysis of mixtures for which the interpretation would have been very difficult. 0 Cl~....~C1 ~ T ~ "C1

0 70~

C1

,Cl (12)

"~ trichlorophenol

C1 Thus we examined in turn : - the intrinsic thermal stability of polychlorinated gem-dichlorocyclohexadienones, -the thermal stability of polychlorinated gem-dichlorocyclohexadienones within chlorophenols, - the behaviour of polychlorinated gem-dichlorocyclohexadienones with acids, -the action of water and chlorine on polychlorinated gem-dichlorocyclohexadienones.

T H E R M A L STABILITY OF POLYCHLORINATED GEM-DICHLOROCYCLOHEXADIENONES Literature does not provide any information on this subject. Before studying the reactivity itself, it seemed important to us to know the range of temperatures which could be used, by examining the basic thermal stability of these compounds (Table 4). Since a chlorination process takes 8 hours on average, we opted to use this time period to perform all trials necessary to conduct the study.

148

Table 4. Percentage of polychlorinated gem-dichlorocyclohexadienonestransformed after heating for 8 hours at different temperatures. Temperatures Cyclohexadienones

Melting point

70~

125 ~

180~

O

C1y

[~,CI

C1/

122 ~

0 %

0 %

100 %

112 ~

0%

0%

100

106 ~

0 %

0 %

100 %

~C1

C1

,C1

%

xCl 0 C1

Cl

When heated throughout, polychlorinated gem-dichlorocyclohexadienones remain stable up to a temperature of 150~ and do not produce the impurities observed during the chlorination process. At a temperature of 180~ the formation of polychlorodibenzodioxins is observed with several break-down products.

149

Table 5. Amount of dioxins in ppm formed after heating polychlorinated gem-dichlorocyclohexadienones for 8 hours at a 180~ 0

0

i

ci

Dioxins

CI~

i Cl

0 CI

C I ~ ~C1

Cl~

c1

C1/

C1

C

1,3,6,8-Tetrachlorodibenzodioxin

10

not detected

not detected

2,3,6,8-Tetrachlorodibenzodioxin

not detected

not detected

not detected

Pentachlorodibenzodioxins

16

1

3

Hexachlorodibenzodioxins

8

40

< 3

Heptachlorodibenzodioxins

50

1000

< 3

Octachlorodibenzodioxin

140

260

4200

THERMAL

STABILITY

OF

POLYCHLORINATED

GEM-DICHLORO-

CYCLOHEXADIENIONES IN CHLOROPHENOLS Since polychlorinated

gem-dichlorocyclohexadienones possibly

formed during

the process of chlorination are in contact with chlorophenols, we determined their stability in the presence of chlorophenols. This experiment was performed with mixtures of 1 mM of polychlorinated

gem-dichlorocyclohexadienonesand

10 m M of

chlorophenols heated for 8 hours at different temperatures. Obtained results indicate the strong influence of the type of phenol and polychlorinated

gem-dichlorocyclohexadienones on

the changes in reaction mixture.

This is the reason why we will distinguish : a) the process related to phenols containing chlorine atoms with a 2.4.6 position from that b) of phenols containing at least one hydrogen with a 2.4.6 position.

Process performed with chlorinated atoms with a 2.4.6 position In the presence of 2,4,6-trichlorophenol, 2,3,4,6-tetrachlorophenol or pentachlorophenol no significant transformation of the cyclohexadienone 2 occurs when performed at temperatures up to 150~

(refer to tables 4, 5, 6).

150

Table 6. Percentage of 2,4,4,6-tetrachlorocyclohexa-2,5-dien-l-one 2 transformed after heating for 8 hours at different temperatures. Melting point of phenol

Phenols

temperature 9 125~

70~

180~

OH

c' 7c 65~

2%

114~

4%

173~

4%

C! OH CI~,~ CV' ' ~

C1 I~ CI OH

CI.~

C1

4%

100 %

Cl~'~~C1 CI

At 180~ 2,3,4,6-tetrachlorocyclohexa-2,5-dien-l-one 2 in isomerizes to tetrachlorophenol 6 with a 75 % yield. 0 C1

pentachlorophenol

OH ,C1

180o

C1

,C1 (13)

Pentachlorophenol

C1 C1

2

6__

151

With pentachlorocyclohexadienone 3, we observe the same stability.

C1

C1

C1

Table 7. Percentage of 2,3,4,4,6-pentachlorocyclohexa-2,5-dien-l-one heating for 8 hours at different temperatures.

3 transformed after

Temperatures Phenols

Phenol melting point

70 ~ C

65 ~

5 %

114 ~

3%

173 ~

2%

125 ~ C

180 ~ C

2%

99 %

OH Cl~/~]./CI

C1 OH CI.~

/CI

Cl OH CI~

CI

c1

When the process is performed at 180 ~

with pentachlorophenol, an exchange

reaction can be observed with 2,3,4,4,6-pentachlorocyclohexa-2,5-dien-l-one

3

leading to the formation of tetrachlorophenol and 2,3,4,4,5,6-hexachlorocyclohexa2,5-dien- 1-one _4.

152

Parallel to this reaction, an isomerization process of 2,3,4,4,6-pentachlorocyclohexa-2,5-dien-l-one 3 in pentachlorophenol _7 takes place. Hexacyclohexadienone 4 i s even stable up to 180~

as indicated in the Table 8.

O C I ~ C 1

CIcl cl'c1 Table 8. Percentage of 2,3,4,4,5,6-hexachlorocyclohexa-2,5-dien-l-one _4 transformed after heating for 8 hours at different temperatures. Temperatures Phenol

Phenol melting point

70 ~ C

65 ~

2 %

114 ~

2%

173 ~

1%

125 ~ C

180 ~ C

2 %

2%

OH C1

C1

CI OH

C1~ CI

OH

civic CI.~

C

Cl

153

At first sight, polychlorinated

gem-dichlorocyclohexadienones seem

to have

thermal stability in the presence of trihalogenated phenols in 2.4.6 position. But detailed analysis of the reaction mass - now possible thanks to the development of the analytical method described above - shows that in all the cases a 3 to 15 % formation of polychloro phenoxyphenols occurs. After heating for 8 hours at 70~ the 2,4,4,6-tetrachlorocyclohexa-2,5-dien-l-one

2

(lmM)

in 2,4,6-tetrachloro-

phenol (10 mM), the following products : 2,4-dichloro 6-(2,4,6-tetrachlorophenoxy) phenol 8 and 2,6-dichloro 4-(2,4,6-trichlorophenoxy) phenol 9 (Fig. 9) are detected within the reaction mass.

~

"

1JV

~

'

~

.

.

: :.,,,: ~~,.~. ... . ...-.. ----t

.

0 CI~CI

. o

g ~

- ' "~"': " . . . . . . . ....

_ .... ".--

J

"

_ -

J

CI~

"C! OH

CI ~ C I C1

CI CI"~Oc_ ~ CIHO - o~

~

~

CI

~

CI

CI

Fig. 9. Chromatogram of a mixture of tetrachloroclohexadienone 2 and 2,4,6-trichlorophenol after being processed for 8 hours at a temperature of 70~ 154

C,o C,

OH

oH

C1

C1

C1 CI

0

C1

C1

C1

C1

9

8

The formation of phenoxyphenols 8 and 9 at 70~

which was only observed

with the mixture of tetrachlorocyclohexadienone, 2,4,6-trichloro-phenol (Table 4) seemed

to

be

a

transformation

of

the

2,4,6-trichlorophenol

catalysed

by

tetrachlorocyclohexadienone. Table 9. Formation of phenoxyphenols after a heating for 8 hours, on different mixtures of products.

Reactants initial composition in %

OH CI~ 0 C1 0/ 7C1 C1CI~C1CI CI

CI

C1

c,

O

0%

not detected

not detected

2%

9.6 %

6.4 %

not detected

not detected

100 % O

OH

CICI~C1CI +CI~ ~CI CI 10,5 %

89,5 %

OH

7.c, 100 %

155

Phenoxyphenols 8 and 9 are the result of an attack by SN2 or SN' 2 of 2,4,6-trichlorophenol on cyclohexadienone 2 with intermediate formation of polychlorophenoxycyclohexadienones 10 and 11.

0 (SN'2 C1

OH

I~C1

+

C1

C1~ N/C ]--~~'~SN 2

2_

Cl

C1

0 CI~ /u,,. ~C1

(16)

0 C1 C1

.C1

+

]~

' 0 ~ C1

C1

C1

C1

C1 1_!

As a second step the polychlorophenoxycyclohexadienones 10 and 11 oxidise 2,4,6-trichlorophenol back to tetrachlorocyclohexadienone 2 and polychlorophenoxyphenols 8 and 9.

156

cl*cl

+

A

c1

11

+

“0c1 c1

C1

Ci

3

c1 8

This mechanism shows the catalytic role of polychlorinated dichlorocyclohexadienone 2 explained in a different way. (Fig. 10)

157

gem-

OH

0 CI

/CI

O Cl

Cl

9

Cl

/

C1 C1

"~

- HCl

CC1~IO~ C1

C1

Fig. 10. Catalytic process of the formation of polychlorophenoxyphenols At temperatures up to 150~ for 7 hours, chlorophenols on their own do not produce any chlorophenoxyphenols. At the same temperature and time as above, polychlorinated gem-dichlorocyclohexadienones do not produce any chlorophenoxyphenols and consequently are stable products. A mixture of 2,4,6-trichlorochlorophenol and polychlorinated gem-dichlorocyclohexadienones produces some polychlorophenoxyphenols. A balance of this reaction indicates that the polychlorophenoxyphenols formed are the result of the condensation of the two chlorophenol molecules. Polychlorinated gem-dichlorocyclohexadienones have a catalyst function. This very important result shows the essential part played by the polychlorinated


in the parasite chemistry of the chlorination of

phenols.

Process performed with chlorophenols containing at least one hydrogen atom in a 2,4,6 position Apart from their greater reactivity, the behaviour of these chlorophenols with polychlorinated gem-dichlorocyclohexadienones is close to that of trihalogenated chlorophenols with a 2.4.6 position.

158

In this case too, the development of the mixture is a function of both the phenol (number and position of the chlorine atoms) and the polychlorinated gem-dichlorocyclohexadienone. Table 10. Percentage of transformed cyclohexadienones after heating 1 mM cyclohexadienones in 10 mM of phenol at 70~ for 8 hours.

of

% of transformed cyclohexadienones

Phenol

--OH

CI

C1

O CI~C1

O C I ~ C I

C1CI~-CI

C1CI~-cfCI

100 %

100 %

53 %

100%

15 %

7%

100 %

100%

49 %

10%

10%

6%

70 %

18%

7%

C1

C I @ O H C1

C1 CI

The reactivity of polychlorinated gem-dichlorocyclohexadienones decreases as the number of chlorine atoms increases. Most of the by-products formed during these reactions belong to the polychloro phenoxyphenols category. This is shown in the following example with the 2,4-dichlorophenol.

159

0

OH

C1

C1

CI

C1 C1 12 % yield

2

12

C1

C1

HO

c1

C1

+ C1

27 % yield C1

13 C1

C1

+ C1

6 % yield 9

el

C1

HO

C1

C1 8 % yield

8

C1

C1

C1 + C1

OH

26 % yield

C1

The

formation of the

condensation

of

SN 2 and

structures 12 SN' 2 of

and 13 is easily explained by

2,6-dichlorophenol

on

the

tetrachlorocyclo-

hexadienone 2, via phenoxycyclohexadienones 14 and 15 (eqns. 19, 20). 2,4,6-Trichlorophenol is the result of an intramolecular transfer of the chlorine according to the following equations "

160

.CI

0

OH

V

OH

" 0 ~ c1

~

Cl

C1

1

Cl

Cl

Cl

1__4

1__2

0

OH

CI

+

CI

OH

1

C1

OH C1

CI

+ C1

C

C1

0

C

13

0 C1

OH C1

CI

[ Cl OH

1

C1 ~

2

(20)

(21)

2

el

C1

The 2,4,6-trichlorophenol thus formed produces polychloro phenoxy phenols 8 and 9 through the process described in the previous section. Equations 19, 20 and 21 show the capacity of 2,4-dichlorophenol to be chlorinated by the ipso intermediates which results in consumption of polychlorinated gem-dichlorocyclohexadienones. This explanation easily demonstrates that phenols with a smaller amount chlorine appear to have the highest reactivity. As soon as they reach 70~

phenols with low chlorine content (phenol,

monochlorophenols, dichlorophenols) produce polychloro phenoxyphenols in the presence of polychlorinated gem-dichlorocyclohexadienones. In this case there is some consumption of polychlorinated

gem-dichlorocyclohexadienones.

161

BEHAVIOUR OF P O L Y C H L O R I N A T E D GEM-DICHLOROCYCLOHEXAD I E N O N E S W I T H ACIDS Chlorination of 2,4,6-trichlorophenol to tetrachlorophenol or pentaclorophenol is usually performed with an acid (refs. 12, 13). For this reason it was important to observe

the

reactivity

of

polychlorinated


in

chlorophenols in the presence of acids. To do this we studied the behaviour of a mixture containing the following components at 70~ - 1 mM of a polychlorinated

:

gem-dichlorocyclohexadienone

- 10 mM of a phenol - 2 mM of acid As the type of phenol used is mainly responsible for the formation of products, we will distinguish - as we did for the thermal stability - trihalogenated phenols in 2.4.6 position, and phenols containing at least one hydrogen atom in position 2.4.6.

Process performed with phenols containing chlorine atoms with a 2.4.6. position In the presence of either a Lewis acid or a strong Bronsted acid, the main reaction observed is the isomerization of polychlorinated

gem-dichlorocyclohexa-

dienone to chlorophenols (eqns. 22, 23). 2,4,6-Trichlorophenol acts as a solvent in this process.

0 C1

OH C1

AIC13(2.5 mM) 70 ~ - 8 h ~ C I ~ / ~ C 1 2,4,6-Trichlorophenol (10 mM)

Yield = - 90 %

(22)

Yield =-~ 80 %

(23)

CI

2

Cl

OH C I ~ C 1

C%l.> ~ C 1 U" " 3

70~

8hr OH

C1

C1 CI

C1

1.68 mM intermolecular (10 mM) C1 The reaction starts off with chlorine transfer to pentachlorocyclo-hexadiene 3 with 2,4,6-trichlorophenol to produce some 2,3,4,6-tetrachlorophenol and tetrachlorocyclohexadienone 2 (eqn. 25).

164

O

c, c, +Cl CI~cI.~C

1

OH

OH

c,

c, c, c,+ CU

C1

0

.Cl (25)

T

C1

3

2

OH

A1Cl3

C1

,C1

2

CI CI

During the following step the cyclohexadienone 2 isomerizes to tetrachlororophenol in an intramolecular way. This last point was perfectly demonstrated by heating a mixture of tetrachlorocyclohexadienone 2 (1 mM) and 2,3,4,6-tetrachlorophenol (10 mM) in the presence of A1C13, at a temperature of 70 ~ for 8 hours. 0

OH

CI~CI+

CI~,/C1

(21/ "(21 2

CU

OH A1C13 C1

T C1

CI

absence of pentachlorophenol

(26)

C1 C1

As opposed to this, the transformation process of the same cyclohexadienone 3 (1 mM) under the same conditions with trifluoromethanesulphonic acid produces a mixture of 2,3,4,6-tetrachlorophenol and pentachlorophenol, by intermolecular and intramolecular processes. OH CI~CI

CF3SO3H(2.5 mM)

ClcI,,/~-Cl

70~ - 8 h OH CI~CI

CI

jCl?l~

OH [~Cl (27)

2

[

CI

CI~

CI

CI CI 0.5 mM 0.67 mM INTERMOLECULAR INTRAMOLECULAR

(10 mM)

CI 165

Results obtained can be explained by the following reaction mechanisms 9

C!

OH -- cICI~~CI CI

0......H|

0

CIcI~C! CI - H+

OH ~:~CI

CICI~CI CI

~MOLECULAR (28)

x ~~ CI Cl

~,,,,,,,H

OH CI~/CI+ H

CI

CIl ~ C i

OH CI

CU y CI

_-

C1/ y CI

Fig. 11. Intramolecularand intermolecular mechanisms To conclude, strong acids and Lewis acids transform polychlorinated gemdichlorocyclohexadienones in chlorophenols. Polychlorinated gem-dichlorocyclohexadienones are true intermediates in the formation of chlorophenols that have one chlorine atom in meta position of the OH. According to the reaction system used (nature of the acid, of the polychlorinated gem-dichlorocyclohexadienone, and of the polychlorophenol) either an intramolecular migration of the chlorine (isomerization) or a intermolecular transfer occurs. - I n addition to the transformation of polychlorinated gem-dichlorocyclohexadienones to polychlorophenols, formation of polychlorophenoxyphenols - in quantities varying according to the acid used in the process - also takes place. Process performed with chlorophenols containing at least one hydrogen atom in a 2.4.6. position When these

chlorophenols

are

used

to

run

the

process,

hardly

any

transformation of polychlorinated gem-dichlorocyclohexadienones in chlorophenols takes place.

166

O

OH

C1

C1

Acid

CI~

C1 (28)

X Chlorophenol containing at least 2.4.6

c1

C1

one hydrogen element in

The products formed are mostly polychlorophenoxyphenols and polychloro dihydroxybiphenyls, which are other families of by-products found in the chlorination masses. Polychloro dihydroxybiphenyls are mostly found with phenol and o'-chlorophenol (eqns. 30, 31). 0 Cl~/Cl

OH

AICI3(2.5 mM)

+ I~~..;.Cl

or

CF3SO3H (2.5 mM)

OH CI,,.~C1 (29)

70 ~ - 8 h

C1~

"CI

10 mM

1 mM

"~ C1

"C1

Yield

=2%

CI + CI

~

O

HO

CI Yield

OH

O

A1CI3(2.5 mM)

H C1 = -~60 %

C1

C1

OF

CF3SO3H(2.5 raM)

cl Icl c

1

+

70~ - 8 h

+C I ~ ~ ~ - O H /

HO

10 mM

1 mM

\ Yield

C1 = - 40 %

(30) 17

3 Preparation and characterization of the biphenyls was performed using the following technique.

167

CI

c1

c1

The formation of polychlorodihydroxybiphenylsis the result of a nucleophilic attack on the protonated form of the cyclohexadienone (Fig. 12).

168

o CI~

.H O"

o~ C1 -.. H+ ,._ C I. ' ~ ~ ,L, _ u C _I

C1

C1

OH

OH

OH

1 OH CI

.C1

C1

/~"~OH

Fig. 12. Mechanism of formation of polychlorodihydroxybiphenyls When the process is performed with 2,6-dichloro and 2,4-dichlorophenol, polychlorophenoxyphenols formed as well as polychlorodihydroxybiphenyls (eqn. 31).

169

cQcy$

1 rnM

NCI3 70"(2.5mM) - 8h

*

c1

CI 10 rnM

0.94 rnM

OH

I

c1 0.04 mM

0.18 rnM

Cl

c1

0.45 mM

CI

The formation of biphenyl 20 can be explained by the oxydation of 2,4-dichloro phenol by 2,4,4,6-tetrachlorocyclohexa-2,5-dien-l-one 2 (eqn. 32).

170

0 C1

OH C1

2

OH

CI +

~

2

+ 2 HC1

C1

2_

C1

CI

2

C1

(32)

C1

OH

r/"~

C1

CI

+

c1

20

Fig. 13. Mechanism of formation of biphenyls 20 Polychlorinated gem-dichlorocyclohexadienonesin the presence of a strong acid do not change into chlorophenols when processed in chlorophenols containing at least one hydrogen atom in ortho or para position. The main products formed are : polychlorophenoxyphenols and polychlorodihydroxybiphenyls.

REACTIVITY WITH WATER P. SVEC (ref. 6) demonstrated that polychlorinated


dienones hydrolysed in halogenated benzoquinones. O C1

O c1

C1

c1 (33)

+ H20

Chlorinated benzoquinones are also a family of by-products produced during the chlorination of heavy chlorophenols (trichlorophenols, tetrachlorophenols, pentachlorophenols) (Fig. 13). We observed that the quantity of benzoquinones formed is linked to the presence of water contained in the reactants.

171

0 C1

OH C1

2

OH

CI +

~

2

+ 2 HC1

C1

2_

C1

CI

2

C1

(32)

C1

OH

r/"~

C1

CI

+

c1

20

Fig. 13. Mechanism of formation of biphenyls 20 Polychlorinated gem-dichlorocyclohexadienonesin the presence of a strong acid do not change into chlorophenols when processed in chlorophenols containing at least one hydrogen atom in ortho or para position. The main products formed are : polychlorophenoxyphenols and polychlorodihydroxybiphenyls.

REACTIVITY WITH WATER P. SVEC (ref. 6) demonstrated that polychlorinated


dienones hydrolysed in halogenated benzoquinones. O C1

O c1

C1

c1 (33)

+ H20

Chlorinated benzoquinones are also a family of by-products produced during the chlorination of heavy chlorophenols (trichlorophenols, tetrachlorophenols, pentachlorophenols) (Fig. 13). We observed that the quantity of benzoquinones formed is linked to the presence of water contained in the reactants.

171

REACTIVITY WITH CHLORINE

Chlorine reacts with polychlorinated gem-dichlorocyclohexadienones to produce polychlorinated cyclohexenones (ref.6) (eqns. 34, 35). O

O + C1

~

(34)

C C1 l i ~

Clcl.> pyridinic N > CN > CF3 > COX (X = Hal, OR) > CHO-~ COR > C1n These electron-withdrawing substituents must be located in ortho or para position to the leaving group and, sometimes, their efficiency can be enhanced by an halogen in meta position. Some comparative examples are given below (ref. 4) :

NO2 X

NO2

C1

~

X

F

DMF Table 1. Influence of substituents X

Conditions

ArF (%)

NO2

140-150~ / 0.5 h

77

CF3

160~ / 3.5 h

76

COzMe

155~ / 4 h

67

245

The leaving ability of the group to be displaced is as follows : R3N + > NO2 > C1 > Br

Displacement of ammonium moieties is of huge interest for the rapid synthesis, under very mild conditions, of 18F-labelled radio-pharmaceuticals useful for medical imaging. Following Clark's work and some others in the 80's (refs. 2,3) increasing interest is paid to fluorodenitration but this method is limited by the availability of 1,2 or 1,4 dinitrobenzenes. Thus, on an industrial scale, exchange of F for C1 is much preferred (ref. 5). The kinetics of chlorine displacement is strongly dependent on the position of the activating substituent : 4-chloronitrobenzene reacts faster than 2-chloronitrobenzene with potassium fluoride whereas, in 2,4-dichloronitrobenzene, chlorine in the 2-position is exchanged more rapidly than chlorine in the 4-position (refs. 4 - 8). These observations are consistent with an addition-elimination process involving an anionic adduct (Meisenheimer's complex) which has been observed by 1H and 19F NMR (ref. 9) or UV spectroscopy (ref. 3) :

E W G ~ C 1

-..

+F |

"-

EWG

-F|

~

C1 \

/

F

4--

-CI@

~

EWG

+ C1 @

EWG = electron-withdrawing group Scheme 2. Addition-elimination process for aromatic halogen exchange.

When located in 2- or 4-position to the chlorine atom, the nitro group stabilizes the Meisenheimer's adduct both by inductive and mesomeric effects, the latter one being predominant. To maximize this effect, the nitro group must be coplanar with the aromatic ring. This is not the case when a bulky chlorine atom is presem in ortho position : the nitro group is then twisted out the plane and 2-chloronitrobenzene is less reactive than 4-chloronitrobenzene. In 2,4-dichloronitrobenzene, mesomeric activation of 2- and 4- positions is affected to the same extend and inductive effect becomes predominant. Nevertheless, this effect decreases very fast with the distance so that the ortho position is more activated than the para one. Inductive activation by halogens can also explain the higher reactivity of 3,4-dichloronitrobenzene compared to 4-chloronitrobenzene. The same phenomenon has been observed with 2-chloropyridines (ref. 10) : 246

X2

X1 C1

Me2S02

Table 2. Inductive effects in the "Halex" reaction

X1

X2

Conditions

Fluoropyridine (%)

H

200- 210~

C1

201~

h

65

194~

h

70

C1

h

49.5

Concerning the effect of the magnitude of twisting in chloronitrobenzenes, we observed that 2,6-dichloronitrobenzene, in which the nitro group is more twisted than in 2,4-dichloronitrobenzene, does not react under conditions where the latter isomer is quite completely transformed (sulfolane / 180~ If the

Meisenheimer's

adduct

is

stabilized

by

/ 11 h). the

electron-withdrawing

substituent, it is also destabilized by +I~ interactions between the negative charge and the p-electrons of the two halogens. Thus, the formation of this adduct can be considered as an equilibrated process since +I~ effect is more pronounced with fluorine than with chlorine, which forms longer bonds than fluorine. On the other hand, as fluorine is more electronegative than chlorine, fluoroaromatics should be more electrophilic than chloroaromatics, and the second step in Scheme 2 could be also considered as an equilibrium. It will be seen that, under some conditions, chloroaromatics can be generated from fluoroaromatics and chloride anions. Nevertheless, reactions depicted in Scheme 2 are usually shifted to the right because C-F bonds are stronger than C-C1 ones (542 kJ/mol vs 339 kJ/mol) (ref. 11). The first " H a l e x " experiments have been carried out with neat chloroaromatics at high temperatures (400 - 500~

but the introduction of dipolar aprotic solvents

in the late 50's brought a dramatic improvement for the use of this process on a large scale under realistic conditions (0 ~ _< 2 0 0 - 250~

(ref. 4). It can be noticed

that protic solvents, which decrease the nucleophilicity of the fluoride anion by strong hydrogen-bonding, are less adapted than aprotic ones. Commonly used dipolar aprotic solvents are : dimethylsulfoxide (DMSO), tetramethylenesulfone (or 247

sulfolane), N,N-dimethylformamide (DMF), N,N-dimethylacetamide (DMAc), Nmethylpyrolidinone (NMP) or benzonitrile. For fluorodenitration, the following order of efficiency has been reported : DMSO > tetramethylenesulfoxide > DMAc > NMP > sulfolane (ref. 12). For other "Halex" reactions, DMSO remains the best solvent (provided that the reaction temperature is lower than 150~ but some changes can be observed in the order of efficiency for other solvents. It can be noticed that, under rather drastic conditions, some by-products can result from the solvent, for instance N,N-dimethylanilines from DMF (refs. 13,14) or thioanisoles from DMSO (ref. 15). Thus, because of its low cost, its thermal stability (up to 250~ and its high boiling poim (Eb760 -- 285~ sulfolane is often preferred. Potassium fluoride is the cheapest source of fluoride and is thus widely used on large scale. However, it is only slightly soluble in aprotic solvents and large difficulties arise from this fact both on a process point of view and on a fundamental point of view (concerning the elucidation of the mechanism). Thus, the "Halex" reaction has been also studied with organosoluble fluorides.

F L U O R I N A T I O N WITH ORGANOSOLUBLE FLUORIDES Tetraalkylammonium fluorides, commonly available, are soluble in a wide range of solvents. However, they are very hygroscopic (ref. 16) and several hydrates are known for Me4NF.nH20 (n 1,2,3,4), Et4NF.nH20 (n = 1,2,2.75,3,5), (n-C3H7)4NF.nH20 (n = 2,3,6), (n-C4H9)4NF.3H20 or BnMeaNF.H2 O (Bn = benzyl) (ref. 16). Among them, Me4NF.4H2 O, Et4NF.2H20, Bu4NF.3H20 and BnMe3NF.H20 are commercialy available. Even under high vacuum, complete dehydration fails at room temperature and Hoffman degradation occurs, under heating at 60~ when hydrogen is present in [3-position to the nitrogen centre (refs. 17,18) : | R3N_CH2~CH2~

R,

Fe

A

| R4N--CH2~CH2~R' Fe+ HF

R3N + R ' - C H - - C H 2

+ HF

| R4N--CH2~CH2~R'

HF2e

Azeotropic dehydration with benzene also fails (refs. 17, 18).

248

Table 3.

Dehydration of R4NF, nH20 according to (ref. 17) and (ref. 18)

R4NF, nH20

20~

Treatment

State

Residual water (Karl-Fischer)

Observed species (19F NMR) (see below)

none

solid

n=3

F-, nH20

solid

n = 2.5

F-, nH20

solid

n=2

F-, nH20 -at-HF2-(traces)

mbar/13h

lyophilization Bu4NF, 3H20

Me4NF, 4H20

20~

sieves

oil

F-, nH20 + HF2- (40-60%)

60~

mbar/20 h

oil

n =0.6

60~

mbar/27 h

oil

n = 0.2

HF 2-

45~

mbar/22 h

solid

n=2

F-, nI-I20

solid

n=2

F-, nH20

solid

n = 1

F-, H20

azeotropic (C6H6) BnMe3NF, H20

azeotropic (C6H6)

.

.

.

.

HF 2-

Hydrated tetraalkylammonium fluorides can be, nevertheless, used in " H a l e x " reactions but water, the nucleophilicity of which is enhanced by hydrogen-bonding with F-, competes

with the fluoride and delivers phenols and diaryl ethers as by-

products. A typical situation is shown below for 3,4-dichloronitrobenzene (ref. 17).

249

C1 O2N~-~-C1 1

C1 R4NF, nH20 Solvent/120~ (F-/Solv. = 1/25 mol/mol)

CI

O 2 N @ F

+ O 2 N @

2

OH

~

C1

Table 4.

Reaction of 3,4-dichloronitrobenzene with soluble hydrated tetraalkylammonium fluorides according to (ref. 17).

RaNF, nH20

BuaNF, 3H20

Solvent

Conv. 1 (%)

Yield 2 (%)

Sulfolane

60

42

20

4.29

56 43

20

3.11

20

3.91

DMSO DMF

Yield (3 + 4) Initial rate of (%) formation for 2 102.Vo (min-1)

Me4NF, 4H20

Sulfolane

60

48

12

4.85

Et4NF, 2H20

Sulfolane

55

45

10

4.17

BnMe3NF, H20

Sulfolane

52

47

5

2.41

Table 4 shows that : - The initial rate of formation of aryl fluoride, but also the yield of oxygenated by-products, increase with the water content of the reagent, -

The final yield of aryl fluoride increases with the solvent polarity,

-

The initial rate of formation of aryl fluoride increases when the solvent polarity

decreases. The effect of the water content in the ammonium fluoride is illustrated on Figure 1 (from (refs. 17,18))

250

NO 2

NO 2

1 BnMe3NF, n H20 CI

Sulfolane/120~ / 2 h F

100 80 "5 6O - - I - - Yield 2

4020 0

|

0

1

m

m w

|

I

2

3

4

n H20

Fig. 1.

Influence of water on the reaction of 3,4-dichloronitrobenzene with BnMe3NF, nH20 (refs. 17,18)

Thus, such extended side-reactions constitute a severe drawback for the use of hydrated alkylammonium fluorides in aromatic fluorination. This problem can be +

-

avoided by using other solvates of onium fluorides, especially species like RaM F , nHF and, for instance, BuaN+HF2 - (n = 1), BunN+H2F3 - (n = 2), Bu4N+H3F4 (n = 3), Bu4P+HF2 -, PhnP+HF2 - or Ph4As+HF2 -. These compounds are easily prepared from R4M § nH20 or R4M § by phase-transfer techniques (refs. 17,19 - 21) or ion-exchange on resins (refs. 17, 20, 22 - 27). However, their ability to displace aromatic chlorine is dramatically dependent on the degree of solvation of the fluoride, as shown on Figure 2 (refs. 17,18) :

251

NO2

NO2 Bu4NF, n HF (1 eq.) Sulfolane / 120~ 92 h. (HF2-/solv. = 1/25)

~]~C1 C1

80 60 O

E

40 2O

0

v

!

|

1

2

3

4

n (F-, n HF) Fig. 2. Influence of the degree of solvation in Bu4NF, nHF on the fluorinating power (refs. 17,18).

In practice, only monohydrogenofluorides are efficient for "Halex" reactions. Moreover, two equivalents of Bu4NHF 2 are needed to obtain quite quantitatively 2_ from 1 but, under the above conditions, the reaction is completely chemoselective (Fig. 3) (refs. 17,18).

252

NO2

NO2 Bu4NHF2

,._,._

Sulfolane 120~ 92 h

CI

C1

C1

100 80 6O 9

40 20 "V"

|

0

1

!

2

3

Mol.ratio Bu4NHF 2 / 1 Fig. 3. Influence of the excess of Bu4NHF2 on the fluorination yield

This p h e n o m e n o n can be explained by the f o r m a t i o n of n o n - r e a c t i v e highersolvated f l u o r i d e s w h e n the reaction p r o c e e d s 9

ArC1 + Q + HF2-

~-

Q + HF2- + HF ArC1 + Q + H2F3-

A r F + Q +CI + H F Q+H2F 3-

/N/ ~ ,.._

ArF

ArF + Q + C1- + Q + H2F3-

ArC1 + 2 Q+HF2

253

Fluorination yields are less sensitive to solvent effects with Bu4NHF 2 than with BuaNF, 3 H20 (Tables 4 and 5) but the same tendency is observed. It can be noticed that this effect can be correlated to the variation of 19F NMR chemical shifts for solvated F- in different solvents (see below). Thus, the measurement of 19F chemical shifts could be a fruitful guide to choose a convenient solvent for fluorination. C1 O2N

~ ~ _ _

CI

R4M +HF2-(2 eq.)

~ O2N

Solvent (HF2- / solv. = 1/25) 120~ / 2 h

_1

Table 5.

C1 F 2

Fluorination of 3,4-dichloronitrobenzene with onium hydrogenofluorides according to (ref. 17). Solvent

Yield 2 (%)

102.Vo (min1)

DMSO

92

7,34

DMF

91

8,08

Sulfolane

88

8,10

Bu4P+HF2-

Sulfolane

68

Ph4P+HF2

Sulfolane

44

Ph4As+HF2-

Sulfolane

60

RaM+HF2-

Bu4N+HF2

Table 5 clearly shows that ammonium hydrogenofluorides

are far better

fluorating agents than other onium hydrogenofluorides 9 BuNHF 2 > Bu4PHF 2 > Ph4AsHF 2 > Ph4PHF 2 As previously mentionned, the possibility of a reverse reaction, that is the formation of chloroaromatics from fluoroaromatics, must be investigated, especially with onium halides since, in this case, no solubility problem can disturb the eventual equilibria. Table 6 indicates that tetrabutyl ammonium or phosphonium chlorides, when

dried,

react to a very limited extend with fluoroaromatics,

even in

concentrated medium with very activated substrates like 2,4-dinitrochlorobenzene (ref. 17). 254

However, water enhances dramatically this " R e t r o - H a l e x " reaction but only when performed on very activated substrates (ref. 17). An explanation could be found in the fact that the fluoride anion can be far more strongly solvated than the chloride anion : hydration could thus be an effective driving force for the "Retro-Halex" reaction.

X

X + BunM+CI, xH20 Solvent ~ O 2 N - ~ C 1 120~ / 2 h (1 eq.)

O2N-~F

X + O2N~OH

Table 6. Displacement of aromatic fluorines by onium chlorides. X

NO2

C1

M

x(H20)

Solvent*

Conv. ArF (%)

Yield ArC1 (%)

Yield ArOH (%)

N

1

Sulfolane

52

44

8

DMF

80

68

12

N

0

DMF

8

P

0

DMF

13

N

1

Sulfolane

* CI- / Solv. = 1/2 (mol/mol)

As already noticed, the anionic fluorinated species are cominuously changing as the halogen exchange proceeds in homogeneous media and this reaction is sensitive to the nature of the solvent. In order to have a better knowledge of the process at every moment and to quantify the solvent-solute interactions, 19F N M R spectra of different solvated fluorides have been recorded in several solvents (refs. 17,18) (Table 7).

255

T a b l e 7.

C h e m i c a l shifts (19F N M R ) f o r s o l v a t e d f l u o r i d e s in d i f f e r e n t s o l v e n t s ( f r o m (ref. 17))

19F N M R ; 8 ( p p m vs. CHCl3)a'b); JHF ( H z ) o r Av ( H z ) |

Solvent

Bu4NF, 3 H20

Bu4NHF2

Bu4NH2F 3 d)

Bu4NH3F4 d)

MeaNF, 2 H20

MeaNF, 4 H20

EtaNF, 2 H20

2~C-VHg_

NMe-T, H20

HCONMe 2 DMSO

-93.2

- 131.7

- 152.0

- 104.0

- 143.8

- 160.2

- 99.7

-

108.6

-

101.9

(Av=12) DMSO-d6

- 150.1

- 161.8

(d,J=119) CH3COCH3

- 109.4

- 151.3

-

167.6

HCONHMe

- 113.2

- 144.3

-

163.6

CD2C12

- 118.7

- 1 5 3 . 8 c)

- 170.4

EEl 4

- 113.2

- 144.8

- 164.9

HMPT

- 114.1

- 150.4

- 168.2

MeCN

- 114.1

- 149.9

- 165.4

- 170.6

-

118.1

-

111.7

(Av= 10) -114.4

DMF

-149.4

- 166.3

-

112.6

(Av=10) - 116.1

DMAc

- 150.6

- 166.4

-

114.6

(Av=31) PhNO 2

-

116.5

- 153.0

- 164.5

-

119.4

(Av=32) PhCN

- 116.5

- 151.3

- 167.7

DMPU

- 116.5

-155.1

- 169.3

DMEU

-

118.1

- 155.8

- 168.8

- 120.4

- 155.9

- 168.4

Sulfolane

-114.3

-120.7 (zXv=40)

MeOH

- 147.4

C1CH2CH2C1

- 160.1

- 172.5

- 157.4

- 167.6

HMPT

= hexamethylphosphoramide

DMEU

= N,N-dimethylethyleneurea

;DMPU

a)

Singlet if no other indication provided

b)

F- / s o l v . = 1 / 5 0 ( m o l / m o l )

c)

becomes a doublet at- 56~

d)

large singlets

= N,N-dimethylpropyleneurea

(8 = - 1 4 2 . 5 p p m ; JHr = 120 H z )

256

;

- 97.4

The chemical shifts have been correlated satisfactorily with the solvent parameters AN (acceptor number (ref. 28)), DN (donor number (ref. 29)) and (dielectric constant) for a set of nine solvents (acetone, acetonitrile, DMF, DMAc, nitrobenzene, sulfolane, HMPT, benzonitrile, methanol) (ref. 17) (Fig. 4). The predominant weight of AN indicates clearly the basic character of solvated fluorides which, however, is strongly modulated by HF-solvation and can be quantified in that way. Thus, the correlation between the chemical shift and the reactivity of soluble fluoride anions could, in principle, allow to predict their fluorination efficiency in any solvent.

FLUORINATION WITH ALKALINE FLUORIDES Because of their price and availability, alkaline fluorides are the most attractive anionic fluorinating agents. Caesium fluoride is the most reactive of them but, because of its price, is only devoted to the preparation of products with very high added values. At the other bottom, lithium and sodium fluorides are completely unreactive. Thus, potassium fluoride, which presents the best ratio between cost and reactivity, is the most popular reagent to perform the "Halex" reaction on a large scale. In fact, no other inorganic fluorides than alkaline ones have been claimed for this technique. Taking into account that all inorganic fluorides and chlorides are sparingly soluble in aprotic solvents and that solubility could be a significant parameter in the "Halex" process, this point can be understood when looking at the lattice energy of solid inorganic fluorides and chlorides. Indeed, a reaction in which one reagent and one product are both in the solid state, must be favoured if the lattice energy of the product is larger than the lattice energy of the reagent. In fact, the lattice energies of fluorides are always larger than those of chlorides. This gap is partly balanced by a higher solvation energy for fluorides and a larger energy for C ~ F bonds (452 kJ/mol) than for C----C1 bonds (339 kJ/mol) (ref. 11). Nevertheless, minimizing the difference between lattice energies of chlorides and fluorides must favour the process. This difference lies between -196 and -83 kJ/mol for alkaline salts, between- 146 and-280 kJ/mol for alkaline earth salts and between-272 a n d - 1 8 8 kJ/mol for transition metal salts (ref. 31). Thus, alkaline fluorides appear to be the less unadapted reagents for substitutive aromatic fluorination. Some of their thermodynamic data are reported on Table 8.

257

BU4NF.3H20

Bu4NHF2

Y( I)=89.6+ I.16 AN+O. 14 DN+0.2 c

Y(2)= 144.3 + 0.29 AN -0.04 DN + O.14 r

170

150

165

140 E ca

'~ 160 130

Dr.l~

6 120

155

Sulfolanea

I

d

G

eL!

t.)

II0 I00

150 145

"

100

!

"

110

i

120

9

i

130

9

!

9

140

Y experimental (ppm)

140,

//" I

i

150

140

150

CI,BOH

9

l

160

Y expeMmental (ppm)

9

, _

170

Bu4 N H2. F3 Y(3)= 16 I.I +0.2 AN +0.09 DN +0.04 c

175

E e,t

170

4D U

>" 165

160 160

9

i

165

9

i

170

9

175

Y expeMmental (ppm) Fig. 4. Correlations between ]9F NMR chemical shifts of solvated fluorides and. solvent parameters (NEMROD program (ref. 30)).

258

ArC1 + MF

~

ArF + MCI

Table 11. Thermodynamic data for alkaline fluorides and chlorides

AEL (MC1-MF) a) AG~176

b) AH(ArC1---~ArF)c)

EL('MF)a)

EL(MC1)a)

(kJ/mol)

(kJ/mol)

(kJ/mol)

(kJ/mol)

(kJ/mol)

Li

1030

834

-196

194

+83

Na

910

769

-141

152

+28

K

808

701

-107

122

-6

Rb

774

680

-94

114

-19

Cs

744

657

-87

102

-26

Fr

715

632

-83

M

a) b) c)

-30

E L = lattice energy; according to (ref. 31) and (ref. 32) according to (ref. 33) according to (ref. 11)

Table 11 can only provide general tendencies since it cannot explain the fact that caesium fluoride is not much more soluble than potassium fluoride and, even, can be less soluble, as reported from our o w n measurements in Table 12.

259

Table 12. Apparent solubilities of caesium and potassium fluoride (electrochemical analysis of the supernatant solution after stirring for 1 h and decantation for 10 min at 0~

Solvent

[KF] (ppm) 40~

lO0~

[CsF] (ppm) 150~

40~

IO0~

150~

2,4-C12C6H4-NO2

23

65

Benzonitrile

42

105

25

24

10

DMAc

18

290

40

28

155

165

DMF

25

70

110

20

120

150

N,N-Diethyl acetamide

85

225

170

50

50

70

Acetonitrile

50

Sulfolane

44

490/190"

200

Sulfolane + 1% H20

70

100

240

Sulfolane + 5 % H20

30

130

130

NMP

81/20"

635/160"

360/140"

41

215

90

DMSO

20

70/80*

40

35

75

220

DMPU

135/20"

335/100"

140/70"

610

190

240

95

of phenylchloroformate

High yields are achievable by using an excess of anhydrous hydrogen fluoride as shown in Table I. The increase of this ratio to a value of 75 does not greatly improve the activity and selectivity.

303

Table 1 " Fluorodecarboxylation of phenylchloroformate-Influence of the ratio HF/C6HsOCOC1

A1203, 300~ OCOC1 + HF

Entry

Without

catalyst,

W

F +

HF/C6H5OCOC1

Yield (%)

1

18

76

2

56

94

3

74

97

CO 2

we do not obtain at this temperature

+

HC1

any fluorinated

derivatives (Table 2, emry 1). Several catalysts (Table 1) are suitable for the fluorodecarboxylation

of

phenylchloroformate

to

fluorobenzene;

such

as

oxyfluoride (AI, entry 2; Cr, entry 4; Zr, entry 5; Ti, entry 10) or fluorides (A1, entry 3; La, entry 6; Ce entry 7; Mg entry 9). The best activity and selectivity is observed using as catalyst aluminium fluoride or oxyfluoride (entries 2 and 3).

Table 2 9 Fluorodecarboxylation of phenylchloroformate - Influence of catalyst nature Entry

Catalyst

Temperature (~

Yield PhF (%)

1

without catalyst

400

2

A1203

300

> 95

3

A1F3

300

> 95

4

Cr203

300

70

5

ZrO2

300

46

6

LaF3

300

40

7

CeF3

300

40

8

ZnO

300

14

MgF2

400

14

TiO2

220

54

10

304

< 1

In contrast to the decarboxylation of arylfluoroformate, anhydrous hydrogen fluoride

vapor

phase

fluorodecarboxylation

of

alkylchloroformate

can

be

successfully applied to many substrats (Table 3). Cresylchloroformates react in a similar way as phenyl chloroformate (entry 1 to 3). The use of more electrodonating substituents like methoxy is also successfull eg for the synthesis of 4-fluoro anisole (entry 4). But, the obtained yield is modest due to demethylation of 4-fluoro anisole or of the starting material leading to tars. For electron withdrawing groups, reactivity is lower but fluoro-chloro, fluorobromo and difluorobenzene are prepared in high yield (entries 6 to 8). Table 3.

Fluorodecarboxylation of arylchloroformate on aluminium oxyfluoride

O--

OCOC1 + HF

~ Vapour phase

R

F + HC1 + CO2 R

Emry

Substituent (R)

1

2-CHs

300

> 95

2

3-CH3

300

> 95

Temperature (~

Yield (ArF) (%)

3

4-CH3

300

> 95

4

4-CH30

400

24

5

4-F

300

< 1

6

4-F

400

> 95

7

4-C1

300

85

8

4-Br

400

70

9

4-NO2

300

Traces

10

1-Ph

350

24

11

1-naphthyl

300

10

12

1-naphthyl

400

> 95

Tars are obtained starting from 4-nitrophenylchloroformate (entry 9); this substrate is mostly unstable under the chosen reaction conditions. 2-Phenyl phenyl chloroformate 3 reacts in a different manner leading to a mixture of

1-fluorobiphenyl 4

(yield 24

%,

entry

intramolecular acylation (yield 50 %) (Scheme 2).

305

10) and of 5 obtained by an

F~ O

Fluorodecarboxylation

II

_4

HF. Al203 ~.. / r 300~ Vapour phase

3

v Internal Friedel & Crafts 5 Scheme 2. Fluorodecarboxylation of 2-phenylphenylchoroformate

Finaly 1-fluoronaphthalene can be obtained with a very high yield from c~naphthylchloroformate (entry 12). The influence of reaction temperature was examined in the case of the decarboxylation of meta-tolylchloroformate 6 in 3-fluorotoluene _7. Results are given in Table 4 and Figure 1. To proceed, the reaction needs a minimum temperature. At low temperature (250~ obtained in rather significant quantities.

Table 4.

Fluorodecarboxylation of 3-tolylchloroformate - Influence of the temperature

~ / CH 3

Entry

entry 1), cresol and cresylcarbonate are

Temperature

(~

OCOC1 + HF

~

F

/ CH 3

6

+

CO2 + HC1

7

Yield (%)

Yield (%)

Yield (%)

Yield (%)

3-fluorotoluene

3-chlorotoluene

m-cresol

carbonate

1

250

78

0,6

10,5

10,5

2

300

92

2

6

0

3

350

90

6

3

1

306

Yield 8, 9, 10

Yield 7 m

100

- 30

90

-E}-- Yield 7 - ~ - Yield 8 Yield 9 Yield 10

20

80 70

10

60 0

50 250

20O

300

350

400 Temperature (~

Fig. 1

Influence of the temperature on the selectivity of fluorodecarboxylation of 3-

tolylchloroformate

m-Cresol 8 is probably obtained by decomposition of chloro (or fluoro) formate by anhydrous hydrogen fluoride following "

/~OCX CH 3

+ HF II O X = C1, F

~

~

H + FCOX

O

/ CH 3

8

Cresylcarbonate 9 result from an acylation of cresyl haloformate by ArOCO + cation 9

~--~OCX ,,

/ CH3

O X=C1, F

/ CH3

+ ~r~OCQ)

XO

,,

CH3

O 9

O

\

~

(~XQ

/ CH3

C II"X O

~ CH3

+ X2CO

CH3

An increase of temperature favours fluorodecarboxylation and limits this side reactions, the main coproduct becomes then 3-chlorotoluene 10 which results from a intramolecular decarboxylation. 307

~OCC1

""

II

/ CH3

~ , ~

O

/ CH3

The synthesis of fluorinated aromatics by starting from the phenol derivative has been derivatives (Table V) like diphenylcarbonate (entry 2), ot,c~,ct-trifluoroanisole (entry 5). selectively

C1

+ CO 2

10

the elimination of a small molecule successfully generalized to various (entry 1), phenylchlorothioformate The reaction proceeds also very

starting from phenylfluoroformate.

For each substrats, an acidic

activation (Lewis or Bronsted) is necessary to produce an activated intermediate, probably the cation ArOCO § or an analogue, by the liberation of phenol or hydrogen fluoride. Table V 9 Influence of the leaving group nature X ~ O ~ Y

HFvaporphaseA1203 ~-

Entry

F + COX + HY

Temperature (~

Yield (ArF) (%)

1

O

OPh

300

60

2

S

C1

300

91

3

O

F

300

> 95

4

C12

CI

400

23

5

F2

F

400

50

R E A C T I O N A L MECHANISM For the decarboxylation of arylfluoroformates, Christe and Pavlath suggest an internal nucleophilic substitution mechanism (Scheme 3). 0 OC --F

R

8+

8-

~

O

R

Scheme 3. Internal nucleophilic substitution mechanism 308

~ F

+

CO 2

This mechanism was preferred to a mechanism via the cation ArOCO § due to the high stability of the acylium cation. This cation does not decarboxylate under the usual conditions for the decarboxylation of arylchloroformate and alkylfluoroformate. Decarboxylation of arylfluoroformiate on 7-alumina (ref. 4d) or fluorodecarboxylation of arylchloroformate under anhydrous hydrogen fluoride did not seem to proceed via the same mechanism. Indeed, the high activation energy of the internal nucleophilic substitution explains why such high temperatures such as 700-800~ are needed. At lower temperatures, the reaction must proceed via a nucleophilic displacement of carbon dioxide in the cation ArOCO § To obtain a very high selectivity in fluoro aromatics, it is necessary to have available nucleophilic fluoride ; the lower selectivity obtained in decarboxylation of phenyl fluoroformate using aluminium fluoride as catalyst can be explained assuming that the fluoride is not nucleophilic enough to permit the reaction. This is in accordance with the necessity of a large excess of anhydrous hydrogen fluoride for the fluorodecarboxylation of arylchloroformate on aluminium fluoride. So, this methodology permits the preparation of aryl fluoroformate with various susbtituents. The use of a large excess of anhydrous hydrogen fluoride compensates for the lack of reactivity of some arylchloroformates.

OCX " 0

+

HF

~

CQ

XHF

R )~"J/

-------~

F

+CO2

+ HX

R

X = C1, F

Scheme 4. Bimolecular nucleophilic substitution mechanism

Also the transformation of phenol to fluorobenzene using a mixture of hydrogen fluoride and antimony pentafluoride has been previously described (ref. 7), we have shown that under our conditions phenol is not reactive. Fluorobenzene is thus produced mainly from phenylchloroformate or phenylfluoroformate. Scheme 5 shows the reaction routes for the main products :

309

OH

+HF/ "COF2 ' + HF

-

CO2 C1

~

~OF

~ -HC1 + -CO 2 " ~ ~

_

CO2

G Scheme 5. Fluorodecarboxylation of arylchloroformate 9product's filiation

The

reaction of substituted arylfluoroformate

regiochemistry

shows

a high retention of

in accordance with the proposed mechanism.

In the case of

fluorotoluene no isomerisation is observed under the reaction conditions.

310

C A T A L Y S T LIFE TIME The catalyst life time was examined in the case of 3-fluorotoluene preparation by fluorodecarboxylation of 3-methylphenylchloroformate decarboxylation. We observed a rapid deactivation of the catalyst after 5 hours of continuous running. This deactivation is linked to the formation of non volatile derivatives. After 5 hours the amount of carbon on the catalysis is 4 %. The catalyst activity can be recovered by calcination under air for 3 hr at 450~ In this way, several reaction - reactivation cycles have been realised without significant modification of the catalytic activity (Scheme 6). CH 3

CH 3

I

HF vapour phase

f"/"'N"l

A1203, 300~ OCOC1

Yield (%) 100

80 60 40 20

0

10

20

30

40

50

60

70

80

Duration on stream (h.)

Scheme 6. Fluorodecarboxylation of 3-methylphenylchloroformate 9catalyst life time

311

CONCLUSION The fluorodecarboxylation of arylfluoroformates in the vapor phase, under anhydrous hydrogen fluoride has been successfully realised by using catalysts such as aluminium fluoride or aluminium or chromium oxyfluoride. This transformation is quite versatile, the main limitation being the stability of substrates or products under the reaction conditions. The catalyst used, deactivates rapidly by coking but can be reactivated by a simple oxidative treatment. This new access to fluorinated aromatics derivatives appears to be an attractive industrial alternative to the diazotisation of anilines.

References

1.

R.D. Chambers in "Fluorine in Organic Chemistry", Wiley-Interscience, New York, (1973). 2. G. Balz, G. Schiemann, Ber., 60, 1186, (1927). H. Susdritsky in "Adv. in Fluorine Chem.", vol. 4, M. Stacey, J.C. Tatlow, A.G. Sharpe, Eds., Butherwerthesm, London, (1965) ; Organofluorine chemistry : principles and commercial applications, Banks, Smart, Tatlow Eds, Plenum Press, chap. 9, (1994). 3. R.L. Ferm, C.A. Van Der Werf, J. Am. Chem. Soc., 72, 4809, (1950). 4. a) K.O. Christe, A.E. Parlath, J. Org. Chem., 30, 3170, (1965). b) K.O. Christe, A.E. Parlath, J. Org. Chem., 30, 4104, (1965). c) K.O. Christe, A.E. Parlath, J. Org. Chem., 31,559, (1966). 5. D.P. Ashton, T.A. Ryan, B.R. Webster, B.A. Wolfmdale, J. Fluorine Chem., 27, 263, (1986), EP 118241, (1983) (to ICI). 6. N. Isvashenko US 3.499.942, (1966). I. Hisamoto, C. Maeda, M. Nishiwaki EP 57443 (1981), (to Daikin Kogyo) J.F. Bieron, D.Y. Tang, US 4792 618, (1984), (to Occidental Chem. Corp.). M. Tojo, S. Fukuoka, J 63054332 and J 63088146, (1986), (to Asahi Chem. Ind.). F.J. Weignet, US 4754 084, (1987), (to Du Pont de Nemours). 1) L. Gilbert, H. Garcia, B. Langlois 2) L. Gilbert, H. Garcia, C. Rochin 3) L. Gilbert, B. Langlois, FR 2647106, FR 2647107, FR 2647111, (1989), (to Rh6ne-Poulenc). 7. J.I. Darragh GB 1582427 (1976), (to ICI). 8. A. Werckmann, GE 857 350, (1943). 9. H. Erlingafeld, Angew Chem., 72, 836, (1960). 10. S. Nakanishi, J. Am. Chem. Soc., 77, 3099, (1955). 11. P. Beak, R.J. Trancik, D.A. Simpson, J. Am. Chem. Soc., 91, 5073, (1969). 12. M. Janin, B. Langlois, L. Gilbert, M.C. Perrod, Private communication

312

MILD T R I F L U O R O M E T H Y L A T I O N OF ORGANIC COMPOUNDS

CLAUDE WAKSELMAN AND MARC TORDEUX SIRCOB-CNRS, Equipe Fluor, B~timent Lavoisier, Universit6 de Versailles, 45 avenue des Etats-Unis, 78 000 Versailles, France

INTRODUCTION Numerous organic molecules bearing a trifluoromethyl group have found industrial applications as pharmaceutical or agrochemical products (ref. 1). They are classicaly prepared by the use of aggressive or toxic reagents. However, the understanding of the reactivity of a small fluorinated molecule has recently allowed the proposal of new trifluoromethylation ways in much milder conditions.

PROPERTIES OF FLUORINATED SUBSTITUENTS Owing to the extreme electronegativity of fluorine (4.0 on the Pauling scale), fluorinated groups behave inductively as electron-withdrawing substituents. The trifluoromethyl group shows an electronegativity (3.5) higher than that of the chlorine atom (3.0) (ref. 2). The inductive effect is balanced by the mesomeric one in trifluoromethoxy and trifluorothiomethoxy groups as for an halogen substiment (ref. 3). On the other hand, fluorination can exert an influence on the lipophilicity of organic molecules, particularly at positions adjacent to atoms or groups having electrons. Hansch constants derived from octanol/water partition coefficients of substituted benzenes (ref. 4) are summarized in Table 1.

313

Table 1. Hydrophobic constants of various substimems

Substituent

OCH 3

F

CH 3

SCH 3

C1

CF 3

OCF 3

SCF 3

n

-002

0.14

0.56

0.61

0.71

0.88

1.04

1.44

. ,

*

Hansch constants derived from octanol/ water partition coefficients of subtituted benzenes (ref. 4).

These data are more appropriate to the nonspecific equilibrium binding of the various compounds to tissues in general than to their genuine kinetics of absorption and distribution in living systems (ref. 5). They can explain the interest for the introduction of the trifluomethylated substituents CF3, OCF3, SCF3 (considered to induce a higher lipophilicity than chlorine) in pharmaceutical and agrochemical products.

CLASSICAL PREPARATIONS The industrial route employed for the elaboration of the trifluoromethyl group is based on an halogen exchange in hydrogen fluoride (Fig. 1) (refs. 1,2). Only very stable molecules can survive in such drastic conditions.

ArCC13

HF

~

ArCF 3

ArYCC13

HF

~

ArYCF3

(Y = O, S)

Fig. 1. Halogen exchange reactions

Numerous alternative preparations have been described (ref. 2). They often make use of toxic (CF3SC1) or fragile (CF3Cu, CF3SCu) reagents. Radical trifluoromethylation can also occur starting from the expensive trifluoromethyl iodide.

R E A C T I V I T Y OF T R I F L U O R O M E T H Y L HALIDES Trifluoromethyl halides CF3X (X = I, Br) are known to be resistam to nucleophilic attack on the carbon atom (refs. 6, 7). This behaviour is explained by the unusual polarisation of the C - X bond (Fig. 2) and also by steric effects and lone-pair repulsion forces associated with fluorine substituents. 314

5- S + CF3mX

( X = B r , I)

Fig 2: Polarization of the CF3-X bond

Pioneering studies showed that attack by strong nucleophiles (alcoholates...) on the larger halogen X can occur, as shown by the formation of fluoroform in protic medium (Fig. 3) (ref. 8). solvent Nu- + CF3X

~

NuX +

CF3

~

CF3H

Fig. 3. Halogenophilic attack

However, since the mid 1970s, some reactions of trifluoromethyl iodide with soft nucleophilic reagents, enamines (ref. 9), or thiolates (ref. 10), have been interpreted as single electron transfer (SET) processes (Fig 4, X = I) (refs. 9, 10,

11). Nu- + CF3X

N u " + [CF3X] 9

[CF3X] 9

X" + CF3"

Nu

[NuCF3] 9

o

+ CF 3 9

[NuCF3] . " + CF3X

NuCF3 + [CF3X] "

Fig 4. SET process with a charged nucleophile.

Trifluoromethyl bromide, produced as a fire extinguishing agent (refs .12, 13), is much cheaper than the corresponding iodide. Unfortunatly, its reactivity is much lower too. At that time, this bromide was considered as rather inert.

315

T R I F L U O R O M E T H Y L BROMIDE REACTIONS

1. First substitution reaction by thiolates Thiolates are powerful nucleophilic reagents. However, we observed no reaction when trifluoromethyl bromide is bubbled through a potassium thiophenoxide solution in DMF at room temperature. This failure was in agreement with the inertness reputation of this halide. Assuming that a mechanism involving radical anions (Fig. 4, X = Br) could occur, a minimal concentration of the halide should be necessary to maintain the chain process. In order to increase the amount of trifluoromethyl bromide in solution, we performed the reaction under pressure. Indeed, condensation occurred at room temperature in a glass apparatus under 23 bars (Fig. 5) (refs. 16,17). Inhibition of this condensation by nitrobenzene was clearly observed, in agreement with the SET process (Fig 4, X = Br). A similar trifluoromethylation of thiols by trifluoromethyl bromide in liquid ammonia under UV irradiation has also been described (ref. 18).

ArSK + CF3Br

ArSCF3 + KBr

Fig 5. Trifluoromethylation of potassium thiophenoxide.

2. Reaction with metals and carbonyl compounds We remarked that the first step of the radical-anion chain mechanism (Fig. 4) can be considered as a reduction of the halide by the nucleophile. Consequently, we tried to use well known reductants such as zinc. However, no reaction occurred when the halide is placed in the presence of zinc in various solvents. By analogy with the thiophenoxide condensation, we attempted the transformation in DMF under slight pressure. Consumption of the reagents was only observed when electrophilic substrates, such as carbonyl compounds, are present since the beginning of the reaction. These Barbier like condensations started more easily in pyridine than in DMF (ref. 19). Moderate yields were obtained with aldehydes as substrates (Fig. 6).

2-4 bars / CF3X RCHO + Zn

RCH(OH)CF3 lutidine

Fig 6. Barbier condensation with aldehydes. 316

The reaction was more difficult with ketones. In the case of acetone, no addition product was even observed. Curiously enough, the presence of this simple ketone initiated the formation of trifluoromethyl zinc derivatives (Fig. 7).

CF3Br + Zn

2-4 bars

CF3ZaaBr + (CF3)2Zn

pyridine acetone

Fig 7. Formation of trifluoromethylzinc derivatives.

When benzaldehyde was subsequently introduced into such a medium, no condensation

product

was

detected,

showing

that

these

strongly

solvated

organometallics are poorly reactive. In the case of an a-keto ester, the addition of the trifluoromethyl group occurred as expected to the keto group (Fig. 8).

2-4 bars RCOCOOEt + Zn + CF3Br

R

pyridine

CF3 !

COOEt

I OH

Fig. 8. Addition to ketoesters.

Simple esters did not lead to addition products. However, the Barbier procedure was effective, even at atmospheric pressure, when the ester was activated by an electron-withdrawing group; ethyl trifluoropyruvate and hexafluoroacetone were respectively obtained from diethyl oxalate and ethyl trifluoroacetate (Fig. 9). 2-4 bars EtO2CCO2Et + Zn + CF3Br

CF3CO2Et + Zn + CF3Br

~ EtO2CCOCF3 pyridine id.

~ CF3COCF3

Fig. 9. Reaction with activated esters.

317

Some acid anhydrides were also trifluoromethylated (Fig. 10). O ~

O

O 3-4 bars

+ Zn + CF3Br

pyridine

O Fig. 10. Trifluoromethylation of phtalic anhydride It is well known that iminium salts show a reactivity comparable to that of carbonyl compounds towards organometallics. Consequently, we tried a similar Barbier condensation with the Eschenmoser's salt, but without success (Fig 11).

3-4 bars CH2zN+Me2 + Zn + CF3Br

CF3CH2NMe2 pyridine

Fig. 11. Failure of the Barbier condensation with the Eschenmoser's salt This failure can shed some light on the nature of the intermediate involved in these Barbier condensations. In contrast to the case of carbonyl compounds, iminium ions do not present a partial negative charge able to coordinate with the metal. On the contrary, carbonyl adsorption leaves the possibility for these compounds to play the part of a ligand around the nascent organometallic, formed at the zinc surface, and to react in the coordination sphere. Following this interpretation, the nascent organozinc intermediate is not completly surrounded with pyridine. It can be more reactive than the strongly solvated organometallics detected in the pyridine-acetone medium (vide supra).

3. Chemical synthesis of triflic acid Barbier conditions were also employed thiocyanates (Fig. 12) (ref. 20).

RSCN + Zn + CF3Br

2-5 bars

RSCF3

pyridine Fig 12. Trifluoromethylation of thiocyanates

318

for

the

trifluoromethylation

of

Benzyltrifluoromethylsulfide, formed from benzylthiocyanate, can lead to triflyl chloride by oxidative chlorination under slight pressure (Fig. 13). Cleavage of the carbon-sulfur bond was easy in this example because the benzyl group is able to stabilize a positive charge. El2, 4 bars C6H5CH2SCF3

C6HsCH2C1 + CF3SOaC1

H20, 5~ Fig. 13. Formation of triflyl chloride by oxidative chlorination of benzyltrifluoromethylsulfide

Owing to a limited yield obtained in the preparation of

benzyltrifluoro-

methylsulfide another route to triflic acid was prefered " the direct Barbier condensation with sulfur dioxide (Fig. 14) (ref. 21). 3-4 bars SO2 + Zn + CF3Br

CF3SO2ZaaBr DMF

Fig. 14. Formation of zinc bromide trifluoromethanesulfinate In contrast to the carbonyl condensations, where no reaction occurred between the substrate and the metal, an initial attack on zinc by sulfur dioxide in DMF was actually observed. Then, introduction of trifluoromethyl bromide under slight pressure led to the formation of zinc triflinate. Homologous zinc sulfinates have been obtained from the much more reactive long-chain perfluoroall~l iodides when these halides were introduced at atmospheric pressure, before or after sulfur dioxide, in a suspension of zinc-copper couple in DMSO or DMF (ref. 22). This condensation was imerpreted as occurring at the metallic surface and was tought to involve an adsorbed organozinc intermediate (ref. 23). In order to check this hypothesis, we performed the following experiment : the supernatant liquid formed in the reaction of sulfur dioxide with zinc in DMF was transfered to a second flask containing perfluorohexyl iodide. Formation of the corresponding sulfinate was detected by NMR. Consequently, a reaction had occurred in this solution. Moreover, inhibition of sulfinate formation was noticed when nitrobenzene was mixed with the perfluoroalkyl iodide. These results can be interpreted by a single electron transfer process (Fig. 15) (refs. 19, 21) involving an intermediate trifluoromethyl radical.

319

~--

Zn + 802

Zn+ +

502 "-(

SO2 +

Br" +

~

1/2-O2SSO2) j,

SO2-- +

CF3Br

SO2-- +

CF3-

.~

CF3-

CF3SO2"

Fig. 15. SET prosess for the triflinate formation

This condensation with sulfur dioxide is rather peculiar. To the difference with carbonyl electrophiles, sulfur dioxide is more easily reduced than trifluoromethyl bromide. As already pointed out, initial consumption of zinc by this anhydride was obvious, producing the sulfoxylate radical anion which is known to be in equilibrium with the dithionite anion (Fig. 15). Incidentally, this salt mixed with sodium bicarbonate in aqueous acetonitrile was used for the transformation of liquid perhalogenoalkyl halides into their corresponding sulfinates (ref. 24). We have been able to transform the gaseous and poorly reactive trifluoromethyl bromide into sodium trifluoromethanesulfinate. However, the reaction conditions (Fig. 16) (ref. 25) were modified because no transformation occurred in the medium employed for the sulfinato-dehalogenation of the liquid halides. 13 bars

Na2S204

+ CF3Br +

Na2HPO4

-- CF3SO2Na DMF-H20

Fig. 16. Formation of sodium trifluoromethanesulfinate from sodium dithionite

The triflinate salt was also obtained with sodium hydroxymethanesulfinate (Rongalite) in the presence of sodium metabisulfite in anhydrous DMF (Fig. 17) (ref. 25). 3-5 bars NaO2SCH2OH

+ CF3Br +

Na2S205

~

CF3SO2Na

DMF Fig. 17. Formation of sodium trifluoromethanesulfinate from sodium hydroxymethanesulfinate

The triflinate salt can be transformed to its corresponding triflate derivative by oxidation with hydrogen peroxide, then to triflic acid by action of sulfuric acid (Fig. 18).

320

30 %

CF3SO2Na Fig. 18.

H20 2 H2SO4 ~ CF3SO3Na ~ CF3SO3H

T r a n s f o r m a t i o n o f s o d i u m triflinate to triflic acid

4. Trifluoromethylation of aromatic compounds We have interpreted the formation of zinc triflinate by a SET process (Fig. 15). In order to test for the presence of the electrophilic trifluoromethyl radical in this reaction, we have added anilines to the medium. Indeed, alkylation at electron-rich positions of the ring was observed (Fig. 19) (ref. 26). NH2

NH2

~/

Zn + SO2 + CF3Br, 3-5 bars Na2S203, DMF

NH2

/CF3 +

CF3

Fig. 19. Trifluoromethylation of aniline

In this experiment, a decimolar quantity of zinc and sulfur dioxide was used. In order to explain such a catalytic effect, we have considered that the sulfur dioxide which is formed in the medium (Fig. 15) can be reduced back to its radical anion by an intermediate cyclohexadienyl radical. This step could induce a chain formation of the trifluoromethyl radical (Fig. 20) (refs. 26, 27). X X X

~

CF3

cv3Br

X CF3

~

SO2 . .

CF3Br 9

~

I-i+

$02

~

"

"

Fig. 20. Radical trifluoromethylation of aromatic compounds

321

CF3Br

"~CF3

Other electron-rich aromatic compounds can be employed as substrates. Pyrroles were trifluoromethylated regioselectively at the 2-position (ref. 27). Recently, the system trifluoromethyl iodide-zinc-sulfur dioxide in DMF at low temperature was used for the trifluoromethylation of aminonaphtalenes and aminoquinolines (ref. 28). Computational results support the mechanism in which the electrophilic trifluoromethyl radical intertact with the aromatic ring at the sites with the greatest electron density of the HOMO orbitals. Similar trifluoromethylation reactions of anilines, phenols or pyrroles were performed in the presence of sodium dithionite (ref. 27).

5. Trifluoromethylation of disulfides Owing to the susceptibility of the weak sulfur-sulfur bond in disulfides to free radical attack, the trifluoromethylation of these compounds was attempted. Indeed, trifluoromethyl sulfides were obtained (Fig. 21) (ref 29).

SO2 9precursor RSSR

+

RSCF3

CF3Br Na2HPO4, DMF-H20

Fig. 21. Trifluoromethylation of disulfides

Experimems were performed with various sulfoxylate radical anion precursors: sodium dithionite, sodium hydroxymethanesulfinate or a mixture of sulfur dioxide with a reductant, such as zinc or sodium formate (refs. 29, 30).In contradistinction with the trifluoromethylation of aromatic compounds (Figs. 19, 20), a stoiechiometric amount of the sulfoxylate radical anion precursor was necessary. In the disulfide case, there is no intermediate able to reduce back the sulfur dioxide which is formed in the medium (Fig. 22). o

SO2"CF3

+ +

2 RS.

CF3Br RSSR

~

CF3" + X

+

~

RSCF3

RS.

~

RSSR

+

SO 2

Fig. 22. Mechanism of the disulfide trifluoromethylation

322

CONCLUSION For many years now, the reactivity of trifluoromethyl bromide has been underestimated. During the past decade the major breakthrough in this area has been the realisation that trifluoromethylation of organic compounds with this halide can be induced by mild reductants such as thiolates, zinc or sulfoxylate radical anion. Nowadays, a great variety of fluorinated products are available by these new methods

:

sodium

triflinate

trifluoromethyl-containing

and

aromatic

triflic

acid,

compounds,

trifluoromethylated ethyl

alcohols,

trifluoropyruvate,

trifluoromethylsulfides ....

AKNOWLEDGEMENTS Cooperation between Rh6ne-Poulenc and our CNRS team began with Dr. C. Kaziz and was kept on with Drs. J-C. Lanet and S. Ratton. We thank all of them for their interest in this research programm. Part of the work has been developed at the " Centre de Recherche, d'Ing6nierie et de Technologie des Carri6res " in Lyon. We thank Drs. B. Langlois, J-L. Clavel, R. Nantermet and the other members of the Rh6ne-Poulenc laboratory for this close cooperation. Similarly, we aknowledge the active participation of Dr. C. Francese during the preparation of her thesis in our laboratory.

References

1. R. E. Banks, B. E. Smart , J. C. Tatlow in " Organofluorine Chemistry: Principle and Commercial Applications ", Plenum Press, New York (1994). 2. M.A. McClinton, D. A. McClinton, Tetrahedron, 48, 6555, (1992). 3. L. M. Yagupolskii, A. Ya. Ilchenko, N. V. Kondratenko, Russian Chem. Rev., 43, 32, (1974). 4. C. Hansch, A. Leo in " Substituent Constants for Correlation Analysis in Chemistry and Biology ", Wiley, New York, (1979). 5. P.N. Edwards in ref. 1, p.530. 6. R.D. Chambers in " Fluorine in Organic Chemistry ", Wiley, New York, (1973). 7. C. Wakselman, A. Lantz, in ref. 1, p. 178. 8. J. Banus, H. J. Emeleus, R. N. Haszeldine, J. Chem. Soc., 60, (1951). 9. D. Cantacuzene, C. Wakselman, R. Dorme, J. Chem. Soc., Perkin Trans. 1, 1365, (1977). 10. V. N. Boiko, G. M. Schchupak, L. M. Yagupolskii, Zhur. Org. Khim., 1057, (1977). 11. C. Wakseman, J. Fluorine Chem., 59, 367, (1992). 12. Until now the main application of trifluoromethyl bromide was its use as Halon 1301 in aeronautic ( ref. 13).Unfortunatly, Halons are implicated in the depletion of stratospheric ozone. The participants of the 1992 Montreal Protocol Meeting in Copenhagen agreed to phase out Halon production by the year 1994, except for some essential fire-fighting uses. Research on alternative agents have been initiated in order to find new products with low or zero Ozone Depletion Potential. However, the numerous candidates examined so far present 323

13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29.

massive problems: toxicity (CF2I),harmful production of hydrogen fluoride on discharge into the hot flame (CF3CHFCF3, CF3CHF2, CF3H), atmospheric life time higher than a millenary (CF3CF2CF 3, CF3CF2CF2CF 3) (refs. 14, 15). Another question is related to the quantity of material needed to put out a fire owing to the lower efficiency of the alternative agents. This issue is crucial for aircrafts where weight carried is a critical factor. These difficulties explain why Halon 1301 is still in use. However, introduction of possible substitutes should leave important stocks of this classical extinguishing agent. Their use for trifluoromethylation reactions should be more useful than a simple destruction. C. Wakselman, A. Lantz, in ref. 1, p.185. R. E. Banks, J. Fluorine Chem., 67, 193, (1994). M. Freemantle, Chem. Eng. News, 29 , (september 19 issue, 1994); idem, 25 (january 30 issue, 1995). C. Wakselman, M. Tordeux, J. Chem. Soc., Chem. Comm., 793, (1984). C. Wakselman, M. Tordeux, J. Org. Chem., 50, 4047 (1985). N. V. Ignatev, V. N. Boiko, L. M. Yagupolskii, Zh. Org. Khim., 21,653, (1985). M. Tordeux, C. Francese, C. Wakselman, J. Chem. Soc., Perkin Trans. 1, 1951, (1990). M. Tordeux, C. Francese, C. Wakselman, J. Fluorine Chem., 43, 27, (1989). C. Wakselman, M. Tordeux, Bull. Soc. China., 868, (1986). H. Blancou, P. Moreau, A. Commeyras, J. Chem. Soc., Chem. Comm., 885, (1976). A. Commeyras, Ann. Chim. (Paris), 9, 673, (1984). W-Y. Huang, J. Fluorine Chem., 32, 179, (1986). M. Tordeux, B. Langlois, C. Wakselman, J. Org. Chem., 54, 2452, (1989). C. Wakselman, M. Tordeux, J. Chem. Soc., Chem. Comm., 1701, (1987). M. Tordeux, B. Langlois, C. Wakselman, J. Chem. Soc., Perkin Trans. 1, 2293, (1990). L. Strekowski, M. Hojjat, S. E. Patterson, A. S. Kiselyov, J. Heterocycl. Chem., 1413, (1994). C. Wakselman, M. Tordeux, J-L. Clavel, B. Langlois, J. Chem. Soc., Chem. Comm., 993, (1991).

324

FORMYLATION

OF

AROMATIC

COMPOUNDS

IN

SUPERACIDIC

MEDIUM

LAURENT SAINT-JALMES, CHRISTOPHE ROCHIN, ROBERT JANIN AND MARCEL MOREL Rh6ne-Poulenc Industrialisation, Centre de Recherche, d'Ing6nierie et de Technologie, 85 Avenue des Fr~res Perret, B.P. 62, 69192 Saint-Fons Cedex, France.

INTRODUCTION Chemists were always interested in introducing in one step a formylgroup into aromatic rings to obtain aldehydes 9 O II ArH ~ Ar~C--H Several methods to obtain aromatic aldehydes are well-known. For example the reaction between aromatics and : -disubstituted formamides in presence of phosphorus oxychloride : Vilsmeier-

Haack reaction (ref. 1), - chloroform in basic medium : Reimer-Tiemann reaction (ref. 2), - zinc cyanide in acidic medium (for example HCI) : Gattermann rea ction (ref. 3). Carbon monoxide and formic derivatives can also be used to make formylations of aromatic compounds (ref. 4) in acidic or superacidic medium. Superacidic medium are characterized as medium whose acidity is superior to those of H2SO 4 (100 %) (ref. 5). In these medium the acidity cannot be measured by pH. The ,, Hammet-Deyrup acidity function ,,, Ho (ref. 6), which shows the facility of protonation of a weak base by a superacid, allows to class the superacids (ref. 7) (Fig. 1).

325

H2SO4 100 %

10

11 12

13

tl I I I I L A

2,4

14

A

15

4~

16

17

18

19

20

1

I

I

I

A

A

Illlllllll A

21 22

I

I

23

I

24

25

I ~---HO A

A HF

HF/BF3

HF/SbF5 ( 1 0 0 / 1)

CF3CO2H

FSO3H/SbF5 (1 / 1)

CF3SO3H FSO3H CFsSOsH/SbF5

(90 / 10) Fig. 1. Hammet-Deyrupacidity scale Formylation

using carbon

monoxide

The first example of electrophilic formylation was reported by Gattermann and Koch in 1897 (ref. 8). They found that carbon monoxide and HC1 can react with alkylbenzenes in presence of A1C13 and cuprous chloride to give aldehydes (Fig. 2).

ArH + CO + HC1

mc~3 CuC1

0 II

Ar~C--H

Fig. 2. Gattermann-Koch reaction From this first example, others acidic systems have been studied to try to generalize this type of formylation and to study selectivities. Formylation of aromatic rings with carbon monoxide requires the use of superacidics to activate carbon monoxide by protonation and to protonate the formed aldehyde which is the thermodynamic driving force of the reaction. O II Even if the formyl cation H--Ce was never observed, the mechanism of formylation by CO / superacids seems to be an electrophilic one.

326

Highly basic aromatic rings can be protonated in superacid medium to give or-complex. In this case, the formylation rates of these basic compounds decrease, which is consistent with an electrophilic mechanism. The use of HF-BF 3 mixture as catalyst for aromatic formylation has been reported

mostly

in

the

patent

literature

(refs. 10, 11).

Formylations

of

alkylbenzenes can be obtained with good yields (Fig. 3).

+ CO

+

20 bars

HF + BF3 10 eq.

~

_ 25oc

1,2 eq.

(ref. 10)

2

79 % yield Fig. 3. Formylation of alkylbenzenes with CO / HF/BF 3 In the cases of phenols,

formylations by C O / H F / B F

3 system require

pressures of carbon monoxide of about 50 to 100 bars to obtain good yields (refs. 12, 13)(Fig. 4).

OH

OH + HF + BF3 50 eq.

CO 40 bars 40~ 1h

2 eq.

~

(ref. 13)

3_ 4

Fig. 4. Formylation of phenol with CO / HF / Others

catalysts

such

as

80%

BF 3

HF-SbF 5

(ref. 14),

HF-SbF 7

(refs. 15, 16),

HSO3F-SbF5 (ref. 17) have also been used to make formylation of alkylbenzenes with carbon monoxide. Triflic acid, alone (ref. 18) or in presence of Lewis acid (ref. 19), catalyses formylation of alkylbenzenes by CO. Even if CO is much more soluble in CF3SO3H than in H2SO 4, pressures of CO superior to 100 bars are required to have good yields in benzaldehydes if Lewis acids are not present (Fig. 5). 327

CH3

0

CH3 CO (1200 psi) + CF3SO3H 6 eqt

3,5 h- RT~

-~

79%

~H3

CH3 CO (1 atm) + CF3SO3H + SbF5 3, 5 h, 0~

:--

59%

Fig. 5. Formylation with CO / triflic acid

Formylation using formic derivatives Very few Friedel-Crafts methods equivalent to the acylation with acyl halides or anhydrides are available to make formylation, undoubtly because halides and anhydrides of formic acids are very unstable. Formyl chloride has been shown to be stable at less than-60~

(ref. 20).

Formic anhydride, which can be obtained from formic acid by deshydratation, is unstable above -40~ Formic anhydride gives formates with phenols but fails to formylate aromatic rings (ref. 21). Mixed anhydrides of formic acid with acids, such as acetic formic anhydride, are stable but give acetylation of aromatic compounds with evolution of carbon monoxide (ref. 22). Formyl fluoride is the more stable halide of formic acid. It can be prepared from formic acid with potassium fluoride and benzoyl chloride in 16 % yield (ref. 23), other fluorinating agents are also used such as KHF2 (ref. 22) (Yield : 35,4 %). Reaction of sodium formate with benzoyl fluoride give also formyl fluoride (ref. 24) with 36 % yield. Formyl fluoride in presence of boron trifluoride forms a complex which has been used in aromatic formylations of alkylbenzenes (ref. 22, Table 1).

328

Table 1. Formylation with HCOF / Aromatic

BF 3

Product

Benzene _7 Toluene Xylene Naphtalene 9

Yield, %

Benzaldehyde 8 Tolualdehyde Dimethylbenzaldehyde Naphtaldehyde 10

56 75 78 20 - ot isomer 67 - [3 isomer

Methyl formate, a stable and commercialy available compound, is a C-1 building block (ref. 25) and so can be a theoretical formylating agent. We have found that formylation of a large type of aromatic substrates can be obtained with fair to good yields using methylformate in the presence of HF-BF3 complex (ref. 26). The

aromatic

compound

and

methylformate

are

diluted

in

anhydrous

fluorhydric acid and a pressure of boron trifluoride is added to the mixture. Amounts

of boron

trifluoride

can be changed to increase

yields.

Table 2

summarizes the obtained results. Benzene

can

be

formylated

very

easily.

Fluorobenzene

gives

fluorobenzaldehyde with a total selectivity in para isomer. On the contrary chloro and bromobenzene are transformed in poor yield, and chloro and bromo toluenes 17 are obtained next to halobenzaldehydes 16, showing that methyl formate can act as an alkylating agent. Table 2. Formylation with H C O 2 C H 3 / HF /

Substrates

a)

HF

BF3 pressure at 0 ~

T~

BF 3

time

conversion %

Yields %

Major products CHO

40 eq.

10 bar

50 ~

6h

72

90 %

11 9

12

F 10 bars

60~

6h

95 %

2,5 bars

40 ~

4h

55 %

~ )

81%

40 eq. 13

51% CHO 9

14

. . .

329

Substrates a)

HF

BF3 pressure at 0°C

To

time

anversion %

Yields %

Major products CI

I

c1 I

40 eq.

10 bars

60°C

6h

15 %

+

9

--

40 eq.

?

40 eq.

50°C

5h

80 %

50 %

10 bars

50°C

5h

85 %

53 %

10 bars

30°C

5h

20 %

22

13 %

OH

5 bars

50°C

5h

90 %

(yo"

rcHo I

CHO

-

I

40 eq.

10 bars

50°C

4h

25 %

22

40 %

21

23

OCH;

0

II

5 bars

OH

&"'

bCH3

-0

15 %

OCHj

I

24

11 %

I

--

@ \

40 eq.

2.5 bars

25

40°C

4h

69 %

- --

a) in every case 1,3 eq. HC0,Me

330

CHO

49 %

Two by-products 27, 28, (Fig. 6) are obtained during formylation of diphenylether 18 next to diphenylaldehyde. The formation of 27 and 28 can be explained by hydride transfer (Fig. 6), a type of transfer in acidic medium already mentionned in literature (ref. 27).

C o-o

(y~

18

19

CHO

C< I

+

C

OH-.~.H

27

0

28

Fig. 6. Formation of by-products during formylation of diphenylether Formylation of guaiacol 20 gives not only vanilline 21 but also aldehyde 29 and surprisingly acetophenone 30 (Fig. 7).

~ OH

OH

OMe OMe

OMe

OH OMe

OMe

20 C yield

3 40 %

30 %

18 %

21

29

30

Fig. 7. Formylation of guaiacol with HCO2Me / HF / BF3 system In the case of anisole 22, acetophenone 31 is also observed (Fig. 8).

331

OCH3

OCH3

(~CH3

OCH3 CHO

v-

+

~

+

22

CH3 11% 24

15% 23

5 % 31

yield

Fig. 8. Formylation of anisole with HCO2Me / HF / BF3 system Phenol is not formylated with H C O 2 M e / H F / B F 3. On the contrary phenyl formate 32 gives by Fries rearrangment p-hydroxybenzaldehyde 33 (Fig. 9). Phenol does not give phenyl formate 32 with methyl formate in the presence of HF / BF 3 (Fig. 9). OH

HCO2Me, HF, BF3

/

X

~CHO o II O--C--H

HCO2Me, HF, BF3

X o II

O--C--H

OH I-IF/BF 3 33

32

CHO

Fig. 9. Formylation of phenol with HCOzMe / HF / BF3 1,3,5-Trimethoxybenzene and 1,4-dimethoxybenzene do not give formylation in our conditions, but demethylations of these compounds are noted. During the reaction, formation or carbon monoxide is noted, showing that methyl formate is, in part, decomposed in the reaction conditions. Two mechanisms can explained the reaction of formylation with methyl formate in HF / BF 3 medium.

332

In route A (Fig. 10), methyl formate is decomposed in CO and methanol. Carbon monoxide is protonated to give formylation of an aromatic ring. In route B (Fig. 10), methyl formate is protonated without decomposition and can be attacked by an aromatic ring to give formylation. 0 II H--C--O--CH 3

|

-H II H--C--O--CH3

xx,,,

Route A ~ , , -'/

RouteBH O" I

H--C--O--CH3 (9

CO + CH3OH + H+

O

R

OH

[

O

+]

H OCH3

~~

CHO

/( CH3OH

R R

Fig. 10. Mechanism of formylation with HCOzMe / HF / BF3 system Acetylated products which are formed during formylation of guaiacol or anisol are difficult to explain. We can suppose that methanol which is the by-product in the reaction of formylation with methyl formate is, in part, deshydrated in our superacid conditions to give the carbocation CH3 + (Fig. 11). This carbocation can be attacked by CO in a nucleophilic way, like in Koch-Haft formylation of aliphatic

333

O compounds, to give the acyl cation CH3

, which is attacked by aromatic rings

to give acetophenones.

CH3~OH

H§

+ H20

|

O CH3~ICI | R

0 R

CH3~ O Fig. 11. Supposed mechanism of acetylation 9a side reaction during formylation Why this acetylated compounds are only formed starting from guaiacol and anisol is not explained. In conclusion, formylation of aromatic rings can be obtained in HF-BF3 medium using methyl formate, a stable and cheap material. High pressures are not required to make these formylations.

Different

functionalized aromatic rings are formylated in our conditions with fair to good yields. Optimizations of the yields obtained will be studied in the future.

References .

2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

C. Jutz, Adv. Org. Chem., 9, part 1,225, (1976). H. Wynberg, E.W. Meijer, Org. React., 28, 1, (1982). W.E. Truce, Org. React., 9, 37, (1957). G.A. Olah, L. Ohannessian, M. Arvanaghi, Chem. Rev., 4, 671, (1987). R.J. Gillepsie, T.E. Peel, Adv. Phys. Org. Chem., 9, 1, (1972). L.P. Hammet, A.J. Deyrup, J. Am. Chem. Soc., 54, 2721, (1932). G.A. Olah, G.K.S. Prakash, J. Sommer in "Superacids", J. Wiley & Sons, (1985). L. Gattermann, J.A. Koch, Chem. Ber., 30, 1622, (1897). M. Tanaka, M. Fujiwara, H. Ando, J. Org. Chem., 60, 2106, (1995). S. Fujiyama, B.E. Patent 887021, (1981), (to Mitsubishi Gas Chem). S. Fujiyama, T. Kasahara, Hydrocarbon Process, 57, 147, (1978). 334

12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.

K. Kudo, N. Sugita, H. Teranishi, Y. Takezaki, Sekign Gakkaishi, 11,690, (1968). L. Weisse, R. Neunteufel, H. Strutz, EP 599148, (1992), (to Hoechst). J.M. Oelderik, A. Kwantes, GB 1 128 966, (1968), (to Shell). M. Tanaka, Y. Souma, J. Chem. Soc. Chem. Comm., 1551, (1991). M. Tanaka, M. Fujiwara, H. Ando, Y. Souma, J. Org. Chem., 58, 3213, (1993). M. Tanaka, J. Iyoda, Y. Souma, J. Org. Chem., 57, 2677, (1992). B.L. Booth, T.A. E1-Fekk~, G.F.M. Moori, J. Chem. Soc., Perkin I, 181, (1980). G.A. Olah, K. Laali, O. Faroqs, J. Org. Chem., 50, 1483, (1985). H.A. Staab, A.P. Datta, Ang. Chem. Int. Ed. Engl., 3, 132, (1964). G.A. Olah, Y.D. Vankar, M. Arvanaghi, J. Sommer, Ang. Chem. Int. Ed. Engl., 18, 614, (1979). G.A. Olah, S.J. Kuhn, J. Am. Chem. Soc., 82, 2380, (1960). A.N. Nesmejanov, E.J. Kahn, Chem. Ber., 67,370, (1934). A.I. Mashentseo, J. Gen. Chem. USSR, 1_.6_,203, (1946). J.S. Lee, J.C. Kim, Y.G. Kim, Applied Catalyst, 57, 1, (1990). C. Rochin, M. Crochemore, S. Ratton, EP 300861, (1988), (to Rh6ne-Poulenc Chimie). R.M. Roberts, J. Org. Chem., 52, 1591, (1987).

335

H I G H S E L E C T I V I T I E S IN HYDROGENATION OF H A L O G E N O N I T R O B E N Z E N E S ON Pd, Pt OR RANEY NICKEL AS CATALYSTS

GEORGES FERRERO

CORDIER,

Rh6ne-Poulenc

JEAN-MICHEL

Industrialisation,

Centre

GROSSELIN

de

AND

Recherche,

ROSE-MARIE

d'Ing6nierie

et

de

Technologie, 24 Avenue Jean Jaur~s, 69153 D6cines-Charpieu Cedex, France.

INTRODUCTION z

X NO2

y NH 2

+ 2 H20 X, Y, Z (CI, Br, F, CF3)

Aromatic haloamines have a wide range of applications in the production of pharmaceuticals and agrochemical substances. The main route to these haloanilines involves reduction from the corresponding nitro compounds either with group VIII metal catalysts and hydrogen or with iron and hydrochloric acid. Most of the time the reduction of halonitroaromatic to the corresponding amine is accompanied by simultaneous dehalogenation which lowers the yield and results in the formation of corrosive halogen acids. At least 2 families of haloanilines can be distinguished. One of these families is included molecules in which the fluorine atom is contained within a CF 3 group bonded to the benzene core. This CF 3 group is insensitive to hydrogenation under operating conditions that are commonly used to reduce the NO2 group in the nitrated precursor. The other family includes molecules containing at least one fluorine atom that is directly bonded to the aromatic core. Sometimes this aromatic core also contains chlorine atoms. These

fluorine and chlorine atoms can be removed under conventional

hydrogenation conditions for the NO2 group, by reaction with hydrogen. 336

Such hydrodehalogenation is often unwanted, but it is sometimes necessary, especially in terms of the chlorine atom when we want to obtain haloanilines from halogenonitrobenzenes. Such hydrodechlorination can occur at the same time as the hydrogenation of the NO2 group with the same catalyst. It can also occur after the hydrogenation of the NO2 group, possibly with a different catalyst.

HYDROGENATION CATALYSTS The literature is filled with various processes and catalyst compositions and systems for these transformations. Promoted platinum and sulfided platinum are the most selective group VIII metal catalysts but depending on reaction conditions and the nature of the halogenonitrobenzene, some undesirable halo-azo and azoxy compounds are left in the product (refs. 3, 11). Catalysts suppliers as Johnson-Mattey (ref. 12), A.G. Degussa (ref. 13) or Heraeus provide special platinum on carbon catalysts at industrial scale. Depending the reaction conditions they claim for these catalysts very high selectivities in halogenoanilines. One other well known catalyst for these hydrogenation is the Raney Nickel catalyst. This catalyst is a very attractive industrial catalyst because of its low cost compared to platinum. However it is also well known to cause extensive dehalogenation during the hydrogenation of halonitroaromatics (ref. 14). Numerous methods have been employed to overcome this problem including modification of the nickel catalyst by adding base (ref. 15) phosphorus (ref. 16), amines (refs. 17, 18) and various sulfided additives (refs. 19, 20). Since the catalytic systems in the literature were not suitable for our application, we tried to develop precious metal or Raney Nickel based catalysts which offer high performance (selectivity, activity and stability) for the hydrogenation of various. In the particular case of Raney Nickel, Rh6ne-Poulenc Chimie has patemed in 1990 Raney Nickel catalysts promoted by iodides (ref. 21) or thiourea (ref. 22).

Some Industrial Hydrogenation Processes One of the most common hydrogenation processes is the batch process in liquide phase with a slurry catalyst ; in a stirred tank reactor (STR) or in a loop reactor as Buss Loop for example.

337

This process decreases hydrodehalogenation especially on the chlorine atom which is one of the most removal. It increases heavy products formed by coupling reactions on intermediaries of hydrogenation and favours by-products coming from nucleophilic substitution on the aromatic fluorine atom. One other process is to proceed by nitro-compound injection at the same rate it is consumed. So there is no accumulation of intermediaries in solution. This process is also carried out in liquid phase with slurry catalyst in STR or loop reactors. It favours hydrodehalogenation and absolutely needs an adapted and selective catalyst.

It decreases heavy products

and especially suppresses

HF

formation coming from nucleophilic substitution on the aromatic fluorine atom.

Some results at Laboratory and Industrial scale The results given as examples in this paper have been obtained by the way of injection pressure as indicated previously. We

selected

five of the current

halogenonitrobenzenes

which

give very

representative pictures in selectivities and activities regarding the type of catalyst involved in hydrogenation. Table 1 gives the results obtained with 0.5 % Pd / A1203 and Table 2 with 0.4 g of 0.5 % Pt / A120 3. Table 1. Hydrogenation of halogenonitrobenzenes over 0.5 % Pd / A1203

Nitro compounds

1 ~mp. ~

4-CNB

353

Injection Yield % rate mole h-1 g-~ chlorides (catalyst) 1.2

4-FNB

373

1.25

3-C-4-FNB

333

0.55

1.05

13.5

3,5-DCNB

353

0.95

9.5

= = = =

0.12

Aniline 90.12 %

7

4-Fluoroaniline " 13 % 4-Chloroaniline 95 % Aniline" 6 % 4-Chloroaniline 911% 3-Chloroaniline 2.5 % 3-Chloroaniline 7.5 % Aniline 910 %

Aniline" 13 %

19

353

CNB DCNB FNB C-FNB

Other by-products

13

3,4-DCNB

Yield % in wanted halogeno aniline

Yield % fluorides

chloronitrobenzene dichloronitrobenzene fluoronitrobenzene chlorofluoronitrobenzene 338

,,,

-~ 85 99,8 ~75 i

85 85

Table 2. Hydrogenation of halogenonitrobenzenes over 0.5 % Pt / A1203 Injection Yield rate -1 % mole h -1 g C1(catalyst)

% F-

Other by-products

1.9

-

0.05

Aniline: 1.9 %

= 98

-

< 0.05

-

Aniline : traces

< 99,9

0.20

0.5

0

0.087

4-Fluoraniline : 0 . 5 %

= 99,4

353

0.26

0.75

-

0.12

4-Chloroaniline : 0.7 %

= 99.0

373

0.32

1.8

-

0.032


= 98

Temp. (~

4-CNB

373

0.37

4-FNB

373

0.35

3-C 4-FNB

353

3,4-DCNB

Yield

Yield % in wanted halogeno aniline

Yield % Azo + Azoxy

Nitro compounds

Aniline : 0 . 2 % 3,5-DCNB

353

0.22

0.4

0.08


= 99.5

375

0.31

0.5

0.035


= 99.5

Hydrogenation on Raney Nickel W i t h R a n e y N i c k e l catalyst we p e r f o r m e d the h y d r o g e n a t i o n s first w i t h o u t and t h e n with selectivity p r o m o t e r s .

Hydrogenation on Raney Nickel without promoters At

preselected

injection

rates

of

the

halogenonitrobenzene

as

described

p r e v i o u s l y we c o m p a r e d the activity and selectivity for e a c h h a l o g e n o n i t r o b e n z e n e at 353 ~

T h e r e s u l t s w e r e as follows.

9 R E A C T I O N R A T E ( m o l e h -1 g-1 Ni) 3,5-DCNB

...........

3,4-DCNB 3-C-4-FNB

t .......... 0.18

t ............... 0.23

4-CNB

t ..........................

t ........

0.4

* TOTAL HYDRODEHALOGENATION 4-FNB 3,5-DCNB 3,4-DCNB

4-FNB

"-

0.7 Y I E L D (%) 3-C-4-FNB

4-CNB

-t ........... t .......... t ............. t .......................... 3 8 10 14

339

t ........ 21

H y d r o g e n a t i o n on Raney Nickel with promoters A calculated and optimized quantity of selectivity p r o m o t e r was added to the catalyst and solvent mixture before the halogenonitrobenzene was injected to obtain the best c o m p r o m i s e b e t w e e n reaction rate and selectivity because all the p r o m o t e r s tested partially deactivated the catalyst. Table 3 shows the comparative results obtained with 3,4-dichloronitrobenzene in the p r e s e n c e o f the previously k n o w n promoters (Na2S, thiophene, cyanoguanidine) and our n e w one, thiourea. Table 3. Effect of promoters on the hydrogenation of 3,4-dichloronitrobenzene Promoter

(P)

Catalyst

Yield % CI

g/1

Injection rate mole h -1 g-1 (catalyst)

(P)

g/1

Yield % mzo + Azoxy

Temp.

Pres.

~

bar

Yield % in wanted halogenoaniline

Cyanoguanidine

3.0

20

0.18

0.7

< 0.001

353

20

~ 99.3

Thiophene

0.3

20

0.18

3.5

< 0.001

353

20

~ 96.5

Na2S- 9 H20 Thiourea

3.3 1.0

20 20

0.18 0.18

0.4 0.4

< 0.001 < 0.001

353 353

20 20

= 99.5 ~ 99.5

??????????? as p r o m o t e r in halogenonitrobenzene hydrogenation. Table 4. Effect of thiourea on the hydrogenation of halogenonitrobenzenes over Raney Nickel Nitro compound

4-CNB 4-FNB 3-C-4-FNB 3,4-DCNB 3,5-DCNB

T

P

Thiourea Catalyst

~

bar

g/1

g/1

Injection rate mole h -1 g-1 (catalyst)

353 373 353 353 353

20 20 20 20 20

1.0 1.13 3.30 1.0 1.0

20 20 20 20 20

0.28 0.45 0.13 0.20 0.16

Yield % C1or F-

Yield Yield % % Azo + in wanted Azoxy halogenoaniline

_ 95 %). Pure carbinol can also be synthesized simply by performing complete hydration of the fuchsone - carbinol mixture in the presence of H20 / AcOEt/ H3PO4 with extraction and recrystallisation (PF = 180~

358

NMR purity = 100 %).

Fuchsone, as used throughout this document, refers to the mixture of fuchsone and carbinol. Optimization Table 5 below shows balances of all the experiments conducted to perform optimizations of the synthesis of fuchsone. Table 5 Reactive agents E n t r y . equiv, amounts Ph2CO THIOL 18

1

T~ of reaction

0.2

110

Yields h

Yields

TT RR RT Ph.CO Fuchsone Fuchsone

4

99.2

84.6

85.3

h

TT RR Ph.CO Fuchsone

.

.

.

RT Fuchsone

.

19

1

0.2

80

4

96.4

75.8

78.6

5

97.6

79.6

20

1

0.05

80

4

80

75

93.8

7

88.7

86

97

21

1

0.05

110

4

91

86.3

94.8

6

94.3

93.3

98.9

22

1

0

110

4

47

46.3

98.5

6

58

55.3

95.3

1

0.05

140

4

95

88

92.6

.

.

24

1.5

0.05

110

5

76.9

76

98.8

7

80.7

78.4

97.2

25

1

0.005

110

4

65.7

65

98.9

7

75

72.7

96.9

1

0.2

110

4

99

83.7

84.5

.

27

0.5

0

110

4

56.6

55.3

97.7

7

74

70.7

28

1

0

140

5

84.8

73

86.1

7

89.7

77.8

86.7

29

1

0

110

4

45.2

43.7

96.6

7

60.5

57.8

95.5

1

0

110

4

43.6

41.3

94.8

7

59

58.7

99.4 98.9

23

26

i

i

i

I

30*

.

.

81.6

.

.

. 95.5

2

0

110

4

34

33.5

98.5

7

44

43.5

32

2

0.005

110

4

41

40.5

98.8

7

50

50

100

33

2

0.05

110

4

59.5

58

97.5

7

66.5

65

97.7

34

2

0.1

110

4

67

64.5

96.3

7

74.5

67.5

90.6

35

3

0

110

4

22.3

22.3

100

7

30

30

100

31

i

Phenol = 10 equivalent amounts - CHaSO3H = 8 equivalent amounts T T = c o n v e r s i o n - R R = yields - R T - s e l e c t i v i t i e s * Recycling o f methasulfonic acid (trial 29)

Most of the time it appears that on one hand the selectivity of fuchsone decreases as the temperature rises, and on the other hand that the thiol concentration can be greatly reduced whilst retaining a marked kinetic effect and high selectivity (entries 20-21). But, strong acid can never be a catalyst in the real sense of the term, as observed

throughout

the

preceeding

experiments. 359

We

will

see

that

it is

stoichiometrically consumed to form a stable salt with the fuchsone, and it will become a catalyst only after the hydrolysing action of this salt releases fuchsone and acid (which can then be recycled). Synthesis mechanism for fuchsone " All the above experiments confirm the following process 9 Ar2C=O + R--SO3H --..

Ar2C+__OH,-O3S_R _..

"- Ar2C+OH + O3S--R (3)

2 'S OH OH

Ar2C+--OH + ArOH O3S--R

"O3S--R OH Ar2C

OH + R--SO3H

Ar2C~OH,-O3S--R

Ar2C

~ ....... ~

OH Ar2C

OH (4) + RSO3H

w-- A r 2 C ~ O H ,

A r 2 C @ O

OH +nH20 I ~ OH,O3S--R .,t-~ Ar2C

"O3S--R + H20 (5)

+ R--SO3H

(6)

OH + R--SO3H (7)

A r 2 C ~ O

Carbocation which results from protonation of benzophenone (eqn. 3) is paraselectively added to phenol (this paraselectivity results from the steric hindrance created by the two nuclei attached to the same carbon atom) to give carbinol (reaction 4). Carbinol under the action of a strong acid R-SO3H dehydrates immediately to produce a salt (eqn. 5) which is the stable form of fuchsone in this medium. The formation of this salt leads to the consumption of one mole of strong acid per mole of fuchsone formed (the acid is therefore not catalytic). 360

We are mistaken when we imagine, at first, that the use of a weaker acid would give a protonation of benzophenone (eqn. 3) without dehydrating the formed carbinol (eqn. 5). In reality this is impossible since the benzophenone is less basic (donor number DNN 17) (ref. 11) than carbinol (DNN 38). So protonation of benzophenone needs a strong enough acid causing the dehydratation of carbinol and the formation of a salt. Hydrolysing the salt (eqn. 7) is necessary to isolate the fuchsone (mixed with carbinol). The regenerated acid CH3SO3H remains in the aqueous phase, and the fuchsone is extracted by an appropriate organic solvent (e.g. : isopropyl ether). Trials conducted on the synthesis of the fuchsone at high temperatures, aiming to shift equilibrium (6) to the right, and to give back the catalytic acid, resulted in failure. This was confirmed by the I.R. study on the salt at variable temperature (1" 240~ in which there is no modification of spectra. The formation of the main by-product 6 is explained by the addition of phenol to the cationic part of the dissociated salt (eqn. 8) :

Ar2C~OH,

O3SCH3

--- A r 2 C L - ~ O H + CH3SO3 +A

H

OH (8)

__(~OH

Ar2C - .-

Ir 6_

OH A clear demonstration of the existence of the three equilibria was made : it was brought to light that this biphenol within a strong acid turns completely into fuchsone salt, and that fuchsone in a phenol system with low acidity leads quantitatively to the biphenol. The influence of thiols results from a rapid addition of this product to the protonated benzophenone Ar2C + - O H (A) to form an hemithioketal which immediately reacts to give a new electrophilic species (B) :

361

/ [ Ar2C+OH ] + A~SH

-..

OH

"~ ar2C

~.,

"-

[ Ar2C+---S--R ] + H20

\ S - - R + H+

(A)

03)

Fig. 7. Addition of thiols to the protonated benzophenone

This species fixes on the phenol and gives a sulphide (C) which - through proton action- releases the cation (D), and regenerates the thiol. SR

I

Ar2C+---S~ R

1

~

ArzC

OH

A r 2 C ~ O H

+ R--SH

(13)

_H§ (C)

(D)

The kinetic effect is explained by the fact that since (B) is more electrophilic than (A), the limiting step of the overall process (addition of A on phenol) is replaced by a quicker reaction (addition of B to the phenol).

Synthesis trials conducted in liquid phase and with heterogenous catalysis Processing of the heterogenous acid catalysts (sulphonic resin, zeolites, clays...) is conducted the same way as it is for homogeneous catalysis: the active acid parts are progressively blocked as 'fuchsone salts'. Regeneration of these parts takes place through hydrolysis which releases the formed fuchsone. Heterogeneous catalysis will have to work first in absorption (the salt is formed) and then in desorption (washing with H20 + organic solvent). Between each of the cycle the washing water is eliminated through drying. It was observed that the efficiency of acid parts of sulphonic resins (NAFION or BAYER K 2431 : 0.8 to 5 meq. H+/g resins) always remains low (1-1.5 mole of fuchosne/10 equivalent amounts H § and that the use of prohibitive quantities of resins is necessary. The same procedure applies to resins having very large pores (BAYER K 1221, LEWATIT 4 % DVB). With mineral solid acids, efficiency is lower than in the previous examples (zeolites, clays, metallic oxides ..... ).

362

Synthesis trials conducted in gaseous phase on mineral solid acid catalysts 9 A few trials were conducted on various solid acids (SiO 2 - A1203 - Nb205, HZSM-5, ...) " Phenol/benzophenone/Argon = 2 / 1 Hold up time = 1 to 4 sec.

/ 1 in moles 0 = 350~

WHSV = 2 - 3.5 h -1

(WHSV = weight hour space velocity Fuchsone is an instable product under these conditions which explains why it is always absent in the process, although benzophenone is always partly transformed. The process cannot be improved by reducing the hold up time.

R E A C T I O N O F F U C H S O N E W I T H H202 We saw before that H202 reacts stoichiometrically with fuchsone leading selectively to HQ and benzophenone (reaction 2). How does this reaction work ? .Co-catalysed hydroxylation of phenol by fuchsone. Influence of [Fuch~oneJ on RR .(diphenols) and on HQ/PC ratio. Several hydroxylation experiments of phenol (by H202/HC104) were conducted starting with increasing quantities of fuchsone. All the results are shown in Table 6 and Diagram 3. Table 6

Trials

BP/36 BP/37 BP/38 BP/39 BP/40 BP/41 BP/42 BP/43 BP/44

Conversion Selectivity Selectivity Selectivity TT RT RT RT PC/HQ (H202) (HQ) (PC) (DP)

(Fuchsone / (HC104 / (H202 H202)o HzO2)o phenol)o % mol % mol % mol 4.65 10.3 14.5 19.9 21.8. 33 57 77.2 101

0.95 0.85 0.95 0.80 0.88 0.75 0.89 0.87 0.80

5.35 5.0 5.2 5.25 4.4 5.3 4.65 4.7 5.1

1 1 1 1 0,5 0,5 0,5 0,5 0,5

99 100 100 100 100 100 100 100 100

DP " Diphenols 363

39.5 42.5 44.5 48 47.5 54 68.5 81 93.5

41.5 40 36 34.5 34.5 29 18 9 2

81 82.5 80.5 82.5 82 83 86.5 90 94.5

1.05 0.94 0.80 0.72 0.73 0.54 0.26 0.11 0.02

PC/HQ

Selectivity% RT (PC +HQ)

1,5

95

1

90

0,5

85

0

80 100

0

25

50

75

PC/HQ (2) RT (PC + HQ) / H202 (1)

ratio (Fuchsone / H202)o % MOL. Diagram 3. By increasing the initial ratio (fuchsone/H202) both the selectivity in HQ and the diphenol yield are also increased. These results show to what extent cost savings could be made by using a phenol process in parallel with a fuchsone plant functioning with a (H202/FUCHSONE) molar ratio = 1 and producing hydroquinone with much greater yields of phenol and H20 2 ( - - 9 5 % ) than those achieved (80 %/H202 and 88 %/phenol) using phenol hydroxylation with H202/HC104 (with or without benzophenone). Remark : when co-cocatalysis is performed using fuchsone, a modulation of the PC/HQ ratio can be obtained with a reduced circulation of benzophenone, this ratio being very much higher than when co-catalysis of benzophenone occurs on its own (ref. 7) (best result for benzophenone co-catalysis : (HQ/PC) - 1 (Diagram 4). Ratio PC/HQ

1,5 J......

0,5

0

I 10

i 20

i 30

i 40

Ratio (Benzophenone / H202) 0 % moI.

Diagram 4.

364

50

BENZOPHENONE I FUCHSONE

It is thought that within the hydroxylation process in which fuchsone is the cocatalyst, this product will first react with H202. It was demonstrated that fuchsone is approximatively 200 times more reactive than phenol, quantitatively producing HQ and benzophenone. Hydroxylation continues further in the presence of the formed benzophenone. Total selectivity of HQ is due to the ex-fuchsone HQ on one hand, and on the other hand to the co-catalysis of hydroquinone. The above indicates that fuchsone is a mixture of fuchsone and carbinol, and that these two compounds lead to the same compounds in the presence of H202. But it is to be noted that carbinol is 5 times less reactive than fuchsone (i.e. approx. 40 times more reactive than phenol). Important remark : Proper separation of the fuchsone from its synthesis medium (CH3SO3 H) is a key element ensuring the adequate oxidation of this product by H202 as the yields percentage will not exceed 30%, with a HQ/PC r a t i o - 0,5 in the presence of CH3SO3H. This demonstrates that the fuchsone salt does not react with H202 and that non-identified secondary reactions occur. Bv-oroduct of the oxidation of fuchsone 4-Hydroxy benzophenone (Yield = 1.5%) is the only by-product identified. Figure 9 shows its formation mechanism 9 ,,

A

,,OH O" H202~ Ar2CI~ O H

Ar2C~O

(1)~ +H + - H20

Ar2C+O~OH

H20

Ar2C--O + HQ

(14)

0+

4-'~0~ ---OH

_-

O

+ ArOH

Fig. 9. The main reaction is (1) as the nucleophilic group p-HO-Ar migrates more easily than the Ar group which is less nucleophilic.

365

FUCHSONE-SELECTIVE PROCESS IN HYDROQUINONE We thought it would be better to let the fuchsone dissolve in the appropriate solvent (isopropyl ether) until its transformation to HQ by H202 occurs, rather than isolating fuchsone in solid state and facing the various difficulties of this technique.

Synthesis of fuchsone Phenol / benzophenone / CH3SO3H = 10 / 1 / 8 in moles ; 0 = 110~ ; 7 h. The duration of the reaction can be be shortened by adding thiols (such as water-soluble HS (CH2)2SO3H which can be recycled using the same procedure as that for CH3SOaH ). Processing of the reaction mass Isopropyl ether and water are added to the above reaction mass (hydrolysis of the salt to fuchsone/carbinol with acid release) : the aqueous phase contains 97.5 % of introduced CH3SO3H with a small quantity of phenol and trace amounts of fuchsone. The organic phase contains almost the entire quantity of phenol which did not react, fuchsone (corresponding to a yield of 58.5 %/benzophenone) and the residual benzophenone (conversion = 59.8 %). Oxidation of fuchsone by H 2 0 2 : A trace amount of HC104 is added to the organic phase, and then the stoichiometric quantity of H202 is added dropwise. Temperature is maintained at 40~ for one hour. HQ with a yield of 100 % / fuchsone is obtained, together with benzophenone with 100 % recovery rate. Then this organic phase is sent towards the distillation section of the diphenols unit. Recycling of CI-I3803I-I : The aqueous phase containing CH3SO3H is first dehydrated at atmospheric pressure, then at reduced pressure. This action ends with a temperature of 176~ in the boiler at 50 tors). CH3SO3H recovery rate is approx. 97 %. Then this acid is recycled to the fuchsone synthesis process without any loss of either activity or selectivity.

366

CONCLUSION This fuchsone route allows increase, as required of the HQ selectivity of a RHONE-POULENC type diphenols unit, while yields versus H202 and phenol are also increased. It is based on the principle of working with two independent hydroxylation processes (phenol and fuchsone) before joining the two fluxes to continue the process. New finding~ are : the paraselectivity of the condensation of benzophenone on the phenol and high HQ selectivity, of oxidation of fuchsone by H202_ (and high benzophenone selectivity which becomes a catalytic compound). This process solves the problem raised by the flexibility of the HQ/PC ratio required by a unit producing HQ and PC at the same time.

References 1. 2. 3. 4. 5.

6.

a) b) a) b) a) b) c) a) b)

7. 8. 9. 10. 11.

a) b)

W.H. Sheard and co-workers, Ind. Eng. Chem. 44, 1730, (1952). M. Dorn and co-workers, EP 368.292 (09/11/1988), (to Peroxide Chemie). E. Nowak and co-workers, U.S. 4.463.198 (23/08/1982), (to Goodyear). M. Taramasso and co-workers, BP 2.024.790 (22/06/1978), (to Snamprogetti S.p.A.). A. Esposito and co-workers, FP 2.523.575 (19/03/1982), (Anic). G. Bellussi and co-workers, EP 200-260 (23/04/1985), (to Enichem Sintesi). M. Marinelli and wo-workers, FP 2.657.346 (19/01/1990), (to Enichem Synthesis) ; A. Thangaras, P. Ratnasamy and A. Kumar, J. Catal., 131,294 (1991). Y. Ben Taarit, C. Naccache, J. Mol. Catal., 68, 45, (1991). F. Bourdin, M. Costantini, M. Jouffret, G. Lartigau, FP 2.071.464 (30/12/1969), (to Rh6ne-Poulenc). J. Varagnat, Incl. Eng. Chem. Prod. Res. Dev, 15 (3), 212, (1976). M. Costantini, M. Jouffret, EP 480.800 (04/10/1991), (to Rh6ne-Poulenc). M. Costantini, D. Laucher, EP 558.376 (01/09/1993), (to Rh6ne-Poulenc). H. Burton, G.W.H. Cheesman, J. Chem. Soc., 1955, 3089. W.T. Lewis and co-workers, J. Am. Chem. S0c., 101, 5717, (1979). M. Costantini, D. Michelet, D. Manaud, EP 606.182 and 606.183 (08/01/1993), (to Rh6ne-Poulenc). Y. Marcus in ,, Ion Solvation ,,, John Wiley and Sons Ed., (1985).

367

THE MECHANISMS OF NITRATION OF PHENOL

PASCAL METIVIER AND THIERRY SCHLAMA Rh6ne-Poulenc Industrialisation, Centre de Recherche, d'Ing6nierie et de Technologie, 85 Avenue des Fr6res Perret, B.P. 62, 69192 Saint-Fons Cedex, France

INTRODUCTION Nitration of phenol is an old reaction that has been described for the first time in 1875 (ref. 1). Nitrophenols are of great interest for the industry since they can be used as precursors for dyes, pharmaceuticals (e.g. acetominophen), and agrochemicals (e.g. parathion, phosalone). Documents dating from 1898 (ref. 2) can be found in the archives of the Rh6ne-Poulenc company indicating the early industrial interest toward this reaction. To our knowledge, Rh6ne-Poulenc is the only company carrying out this reaction on a large industrial basis. This reaction has the reputation to be messy (ref. 3), and despite important studies, the different mechanisms involded in this reaction have not been completely solved. Two mechanisms are currently considered to be active in this reaction. The first one which is well established and has been fully studied in aqueous media involves two steps, nitrosation followed by and oxidation (Scheme 1). Nitrosation is an electrophilic aromatic substitution involving the nitrosonium ion and is mainly para selective (refs. 4, 5), and the oxidation is due to the nitrogen dioxide in equilibrium with nitrous acid and nitric acid. Two mechanisms involving nitrogen dioxide have been proposed for the oxidation step (ref. 6). This total paraselectivity can only be obtain in a two step procedure where nitrosation is performed in the absence of nitric acid followed by oxydation to paranitrophenol by addition of nitric acid to the mixture.

368

OH

OH

+

HNO 2

~N O"

OH

3

+

H20

+

HNO 2

+

H20

OH

I +

HNO 3

0-~ N

0~. N+ O-

OH

OH

+

HNO 3

r

..N+ 0

O-

Scheme 1. 9The nitrosation / oxydation pathway, a para selective route to nitrophenol

The reaction of phenol with nitric acid catalysed by nitrous acid conducts to mixture of para and ortho nitro phenols, with a varying ratio and low yield in aqueous media, and an unvariable 55/45 ortho/para ratio when the reaction is carried out in biphasic media. The biphasic procedure, named nitrous acid catalysed nitration of phenol in the literature has been first pointed out by Kagan and coworkers (ref. 7) and has been since, subject to many studies. J.H. Ridd and coworkers, have shown throught CIDNP effects (Chemical Induced Dynamic Nuclear Polarisation) in the nitration reaction that the mechanism involves the ArO'NO2 radical pair (ref. 8). M.J. Thompson and P.J. Zeegers (ref. 9), have correlated the ortho/para ratio of the nitration of various phenols with the unpaired electron spin density of phenoxy radicals using semi-empirical calculations and shown that they are in very good agreement with the experiments. They have proposed a mechanism in which the phenoxy radical is the key intermediate. In this mechanism the first step is an electron transfer with NO + as the transfer agent, followed by a deprotonation step leading to the phenoxy radical which than reacts with nitrogen dioxide to give the nitrophenols (Scheme 2). This mechanism, also 369

enables to explain the formation of the major side product which is benzoquinone. The phenoxy radical can be over oxidated throught another electron transfer step, leading to the phenoxonium cation which reacts with water to give hydroquinone and is then oxidised to benzoquinone.

HNO 2 +

H+

..,,

"-

H20

OH

+

NO +

OH

+

NO +

#

+

+

OH

O.

[~

~ ~

O,

+ H+ OH

+

NO

NO 2

+ 2 HNO 3

2 NO 2 +

H20

OH

...,

"-

3 NO 2 +

_..,

~

HNO 2 +

~

H20 HNO 3

OH

I [ ~

NO

+

HNO 3

~N/~_

+

H20

Scheme 2. 9Nitrous acid catalysed nitration of phenol, mechanism proposed in literature

Both mechanisms (nitrosation / oxydation and nitrous acid catalysed) involve an initial step with NO + as the reacting specy. In water, it would react as an electrophilic reagent and in organic media as an electron transfer acceptor. Results obtained in our laboratory are not in good agreement with this explanation, and we decided to try to identify more precisely the role of each active specy in this reaction. 370

RESULTS Experiments were carried out with different oxydation states of nitrogen to try to identify the species that are active in the different mechanisms. The approach chosen was to react phenol with NIII, NIV and NV in water and in organic media and to observe the different products that are formed. This basic procedure which seems simple needs to be undertaken caustiously. Nitrophenols are very pourly soluble in water (2 g/1 for orthonitrophenol at 20~

and tends to demix from

water giving two liquid phases. Since this reaction is considered to be messy, one must be sure that the carbon balance is correct so to avoid misinterpretation. All the reactions were carried out in relative dilute state ( 1 % ) and reaction medias are quenched with aqueous cold water containing sulfamic acid to destroy nitrous oxydes, and titration of products is done throught HPLC analysis. Nitrosation : In aqueous media, as described in literature (refs. 4, 5) the main reaction product is paranitrogophenol.

Side products are paranitrophenol and

orthonitrophenol (Table 1). Table 1. 9 Nitrosation of phenol in aqueous media with NaNO2 (2 equivalents) / H2SO4 system n2so4 (%) weight

42

Temperature (~

42

70

70

23

-4

25

Conversion

(%)

41.9

63.8

24.1

41.3

. . . . .

Paranitrosophenol

(%)

35.2

44.4

19.9

23.7

YIELD

Paranitrophenol

( %)

1.7

0.8

2.8

5.5

HPLC

Orthonitrophenol

(%)

0.6

0.3

0.6

0.2

98.6

88.4

titration

At 23~

Benzoqu inone

( %)

2,4-Dinitrophenol

(%)

2,6-Dinitrophenol

(%)

Carbon balance

(%)

0.5

96.1

91.7

paranitrophenol becomes as important as paranitrosophenol, whereas

orthonitrophenol remains very low. This can be explained by the dismutation equilibrium of NIII in water to give NIV and NII (ref. 10). This NIV specy can then oxidise the paranitrosophenol to give paranitrophenol. The small quantity of orthonitrophenol

formed

indicates clearly that the nitrosation mechanism

preponderant.

371

is

In organic media (Table 2) results are more surprising in that, that according to the proposed ,, nitrous acid catalysed ,, mechanism, one would not have predicted paranitrosophenol

to

be

the

main

product.

Particularly,

in

toluene

or

dichloromethane where biphasic nitration of phenol takes place very rapidly, no reaction is observed.

Table 2. 9 Nitrosation of phenol in acetonitrile media with NOBF4 CH3CN

Temperature (~ Conversion

(%)

50

3

-1

72

83

56.4

65

37.2

1.5

3.2

Paranitrosophenol (%)

44.3

YIELD

Paranitrophenol

(%)

1.56

HPLC

Orthonitrophenol (%)

0.8

titration

Benzoquinone

0.6

(%) 0.5

2,4-Dinitrophenol (%) 2,6-Dinitrophenol (%) Carbon balance

The

carbon

balance

84

75

(%)

is not very

good

when

83.8

reaction

takes

place,

so

interpretation must be cautious. But even if this carbon loss reveals an electron transfer from phenol to NO + , this reaction is slow and does not explains results obtained in ,, nitrous acid catalysed ,, nitration of phenol. Reaction of phenol with dinitrogen tetroxide (N204) : In contrast to nitrosation in organic media, reaction with N204 in organic media is fast and leeds to the caracteristic products of ,, nitrous acid catalysed ,, nitration (Table 3). A striking exemple is reaction in toluene where nitrosation does not take place whereas reaction occurs with N204 leading to nitrophenols, benzoquinone and no trace of paranitrosophenol.

Furthermore the ortho/para ratio is roughly invariable and

corresponds exactly to the expected ratio for ,, nitrous acid catalysed ,~ nitration. Dinitrophenol and especially 2,4-dinitrophenol is formed in small quantities during this

reaction.

We

have

looked

at the

compared

reactivity

of

ortho

and

paranitrophenol and show that paranitrophenol reacts more quickly with N204 than orthonitrophenol to give dinitrophenol.

Nitration of orthonitrophenol

occurs leads to a ratio of 65/35 in favour of 2,4-dinitrophenol. 372

when it

From these

experiments we can conclude that in organic medium, nitrogen tetroxide is the reagent involved in the first step of the nitration leading to the phenoxy radical.

Table 3.

9 Reaction of phenol using N204 in organic solvents Toluene ,,Temperature

(~

N204

Conversion

(%)

p aranitrosopheno! . (%)

CH2C12

CH3CN

Stflfolane

AcOEt

0

24

24

0

24

0

24

24

0

24

24

0.5

0.5

0.6

0.5

0.5

0.5

0.5

0.5

1.1

1.1

0.5

90

57.8

56.8

64.8

73.5

96.5

55.8

43.6

100

100

81.7

/

/

/

/

/

/

/

/

/

/

/

..Paranitrophenol

'% )

34.3

11.6

11.0

21.5

10.8

40.6

18.3

15.1

40.3

30.9

29.9

Orthonitrophenol

( %)

40.9

23.5

25.3

24.6

28.4

47.2

19.6

18.8

47.0

35.6

34.5

Benzoquinone

( %)

4.2

1.9

1.7

1.3

1.8

2.1

1.3

1.0

1.9

1.9

1.7

.2,4-Dinitrophenol

(%)

0.8

7.7

8.5

1.5

15.7

0.8

5.6

/

2.0

14

3.2

2,6- Dintrophenol

( %)

/

/

/

/

/

/

2.6

/

1.0

6.3

1.0

Carbon balance

(%)

90

86.5

89.4

84.6

83.3

94.4

91.7

91.6

92.0

88.7

88.5

54.0

54.8

54.8

51.6

51.8

53.7

50.8

55.5

53.2

48.3

51.8

(ONP+2,6-DNP) / 2; nitrophenols

Reaction w i t h .

N204 ~

aqueous media and DMSO : As described in literature,

introduction of N204 in an aqueous media results in an immediate dismutation to nitrous acid and nitric acid (ref. 10). As a result of this dismutation results of this type of reaction leads to the same results than with nitrous acid in diluted sulfuric acid, i.e. nitrosation is the predominent mechanism (Table 4), with no trace of benzoquinone. In dimethyl sulfoxide (DMSO), the same type of result is obtained. This means that D M S O behaves the same way than water toward N204. This is confirmed through literature results, N204 is known to racemise chiral sulfoxide (ref. 11), and that its action on an O18 maked sulfoxide leads to an oxygene exchange between the two molecules (ref. 12). The explanation of this reactivity passes throught the dismutation of N 204 in sulfoxides as described in Scheme 3.

373

Table 4. - Reaction o f . N204. in water and in DMSO with phenol

WATER

DMSO

Temperature

(~

24

24

24

N204

(eq.) (%)

0.5

0.5

0.5

38.5

26.7

42.1

Paranitrosophenol

(%)

26.8

23.8

14.3

Paranitrophenol

(%)

0.9

1.1

8.3

Orthonitrophenol

(%)

1.5

1.4

1.9

Benzoquinone

(%)

/

/

/

2,4-Dinitrophenol

(%)

2,6-Dinitrophenol

(%)

Carbon balance

(%)

Conversion

6.4

o,.

o

O

CH3-

N. "n-

~ S-:'-O"

'

~'

"~

C H~ S--O

CH3

O

CH 3

O

89.0

99.5

90

O II

N-. oN~O

O

9

N'-----O

""-

* S---O

CH 3

'N--O

I_ O

O~N*IO H3

/S*---Or, )

O~N'~OI_ ~

OH 3

N--O

O

CH3 S'-"-O/ CH 3

s'---oCH3 I

I

O~

"N II O

O~N-tO O~\ 4-

O ae

N*-"'-N" O O 9

I /

S---O

C,H3

Scheme 3. " Dismutation of N204 with DMSO 374

O-~ N

II O

Reaction of the nitronium ion in organic media : reaction of phenol with NO2 + in organic media leads to the same type of results than with N204 (Table 5). The formation of benzoquinone is systematically observed and the ortho/para ratio is again invariable.

A small quantity of paranitrosophenol

is observed which

corresponds to nitrosonium contained in the nitronium product used as starting material (-- 4 % for NO2SbF6).

Table 5. 9 Reaction of NO2+ in organic media CH3CN

CH2C12

Temperature

(~

24

24

24

Conversion

(%)

28.1

64.8

63.3

Paranitro sophenol

( %)

3.3

1.2

0.3

YIELD

Paranitrophenol

( %)

9.2

20

18.7

HPLC

Orthonitrophenol

( %)

12.0

24.0

21.8

(%)

1.85

2.6

0.3

TITRATION Benzoquinone 2,4-Dinitrophenol

( %)

4.4

2,6-Dinitrophenol

(%)

0.2

Carbon balance

(%)

Nitration agent (ONP + 2,6-DNP) / E nitrophenols (%)

99.0

93.1

83.0

NO2SbF6

NOzBF4

NO2SbF6

56,5 %

54,5 %

49 %

Reaction of $ulfonitric medium with ohenol " Reactions in sulfonitric media are carried out using either sulfamic acid or urea as nitrous acid scavenger. Results obtained with different sulfuric acidities are given in Table 6. With diluted sulfuric acid and a nitrous acid scavenger, no reaction takes place, indicating that nitric acid by itself is not an active specy. With 80 % sulfuric acid, where the nitronium ion begins to be significative (ref. 13), the results are the same than in organic media. With 70 % sulfuric acid, a non expected result is obtain, in that, the ortho/para ratio moves up to 65/35. In that case the reaction is more orthoselective than ever. With sulfuric acid concentration over 80 %, the preponderant reaction is sulphonation of phenol. With intermediate type sulfuric acids (30-60 %), after a varying induction time where no reaction takes place, the media turns suddenly to tars and results are not interpretable.

375

Table 6. : Reaction of phenol with various sulfonitric medium Nitrous acid scavenger

no

urea

urea

urea

urea

urea

H2S04 (%)

80

80

80

70

70

20

Temperature (~

24

24

0

0

24

24

Conversion (%)

100

100

78.2

35.7

83.5

2.8

Paranitrosophenol

0.1

/

/

0.4

0.3

/

YIELD

Paranitrophenol

32.8

35.1

35.1

11.8

27.3

/

HPLC

Orthonitrophenol

42.2

45.9

42.4

22.1

51.9

/

Benzoquinone

0.1

/

/

/

/

/

2,4-Dinitrophenol

5.5

2.8

/

/

/

/

2,6-Dinitrophenol

2.4

1.2

/

/

/

/

Carbon balance (%)

83.1

85.0

99.3

98.7

96.0

97.2

(ONP+2,6-DNP) / Z nitrophenols

53.8

55.4

54.7

65.0

65.0

/

TITRATION

GENERAL DISCUSSION From our results, it appears clearly that the first step in nitrous acid catalysed nitration of phenol is not a monoelectronic transfer from phenol to the nitrosonium specy, but rather a reaction with N204 leading to the formation of the phenoxy radical, nitrous acid and NO2. The mechanism that we proposed for this reaction of phenol with N204 passes throught the intermediate formation of phenylnitrate, which then decomposes homolytically to form the phenoxy radical and NO2 (Scheme 4). Semi-empirical calculation (MNDO, PM3) on the homolytic scission of phenylnitrate shows that the Enthalpy of reaction to give the phenoxy radical and NO2 is -1,7 kcal/mol and though should be spontanneous (ref. 14). The N204 reaction with phenol that we propose here is formally the same then the reaction of water (ref. 17) with N204 leeding to dismutation into nitrosonium nitrate, also the same than the desmutation reaction with DMSO (ref. 15) and aliphatic alcools. This mechanism corresponds also partially to the one proposed by R.G. Coombes (ref. 18) for the reaction of 2,4,6-trialkylphenol with nitrogen dioxide in organic solution.

376

O

II

H~

H~oc/N~,.O.

o-\

/

o-

N~---N *

+ N O 2-

O

Same reaction as described for aliphatic alcohols

O

II

H ~ O (." N~,.O.

0 / N~"~0 -

ROH ~

l

+ N O 2-

+

HN02

RONO 2

t 1

J

0

II Semi empirical calculations

I

&H = - 1,7 kcal/mol.

+ NO 2

i + NO 2

>

H

OH

~ t

[ ~

+

NO 2

45%

NO 2

55 %

Scheme 4. 9 Proposed mechanism for nitrous acid catalysed nitration of phenol

With the nitronium ion results indicate that the same intermediate phenylnitrate is formed, which then follows the same path to give 55/45 ortho/para mixture of nitrophenols (Scheme 5). It is interesting to note that in organic medium the N204 nitration of phenol is much faster than with the nitronium ion. This mechanism involving

the

initial

formation

of

the

phenylnitrate

which

decomposes

homolitycally to the phenoxyradical and NO2" has already been proposed by J.H. Ridd in the case of nitration of paranitrophenol with the nitronium ion (ref. 16).

H~O NO2 ~"

OH +

NO2 +

~

O"

NO 2

~

O ~

.

+ NO 2

.

OH ~

Scheme 5. 9 Proposed mechanism for the nitration of phenol with the nitronium ion

377

NO 2

The case of nitration with at 70 % sulfonitric mixture seems particular. In this zone of acidity, the main specy is neither the nitronium ion neither nitric acid but protonated nitric acid HzNO3 + (ref. 13). In this case one can invoque a cyclic transition state to explain the ortho selectivity that is observed.

CONCLUSION According to our results, three mechanisms can be effective in the nitration of phenol. The first one which has been well described is the nitrosation oxydation pathway,

which is paraselective and involves paranitrosophenol

as the key

intermediate. The two other mechanisms involve the same key intermediate : the phenoxy radical which combines with NO2 to give a 55/45 ortho/para nitration mixture. This intermediate can be formed either via fast reaction with dinitrogen tetroxide (N204), or slow reaction with the nitrosonium ion. The results we obtain suggest that the first step is the formation of the phenylnitrate intermediate, which undergoes an homolitic breakage of the oxygen-nitrogen bond leading to the phenoxy radical and nitrogen dioxide. In the case,

OH

I H20

N204

O

0

ONO 2

O

O.

~

0 OH

+

NO2

Scheme 6. 9 The ,, nitrous acid ,~ catalysed nitration of phenol - overall proposed mechanism of the biphasic procedure, the formation of N204 results from the well know reaction of nitric acid with nitrous acid. This N204 is then extracted to the organic media where fast reaction with phenol takes place as depicted in Scheme 6.

378

References

1. K6rner, Gazz. Chim. Ital., 4, 440, (1875). 2. L. Benda, Internal report, (April 1898), (to Soci6t6 Chimique des Usines du Rh6ne). 3. T. Mc Cullough, K. Kubena, J. Chem. Educ, 67,801, (1990). 4. C.A. Bunton, E.D. Hugues, C.K. Ingold, D.I.H. Jacobs, M.H. Jones, G.J. Minkof, R.I. Reed, J. Chem. Soc., 2628, (1950). 5. B.C. Challins, J.H. Higgins, A.J. Lawson, J. Chem. Soc., Perkin Tram. II, 1831, (1972). B.C. Challins, J. Chem. Soc. 03), 1971, 770 ; B.C. Challins, J.H. Higgins, J. Chem. Soc., Perkin Trans. II, 1597, (1973). 6. Y. Ogata, H. Tezuka, J. Org. Chem., 1968, 33, 3179 ; G.V. Bazanova, A.A. Stotskii, J. Org. Chem. USSR, 1427, (1981). 7. M. Ouertany, P. Girarg, H.B. Kagan, Tetrahedron Lett., 23, 4315, (1982). D. Gaude, R. Le Goaller, J.L. Pierre, Synth. Comm., !6, 63, (1986) ; M.J. Thompson, P.J. Zeegers, Tetrahedron Lett., 29, 2471, (1988). 8. A.H. Clemens, J.H. Ridd, J.P.B. Sandall, J. Chem. Soc., Perkin trans. II, 1667, (1984) ; M. Ali, J.H. Ridd, J.P.B. Sandall, S. Trevellick, J. Chem. Soc., Chem. Commun., 1168, (1987); J.H. Ridd, S. Trevellick, J.P.B. Sandall, J. Chem. Soc., Perkin trans. II, 573, (1992). 9. M.J. Thompson, P.J. Zeegers, Tetrahedron, 45, 191, (1989) ; M.J. Thompson, P.J. Zeegers, Tetrahedron, 46, 2661, (1990). 10. J.W. Mellor in ,, A comprehensive treatise on inorganic and theoretical chemistry ,,, Volume 8, Logman Editor, pp. 454-469 for nitrous acid, pp. 529-549 for nitrogen tetroxide. 11. S. Oae, N. Kunieda, W. Tagaki, Chem. & Ind., 1790, (1965). 12. C.R. Johnson, Jr D. Mc Cams, J. Am. Chem. Soc., 86, 2935, (1964) ; C. Lagercrantz, Acta Chem. Scand., 23, 3259, (1969). 13. D.S. Ross, K.F. Kuhlman, R. Malhotra, J. Am. Chem. Soc., 105, 4299, (1983) ; G.F. Scheats, A.N. Stachan, Can. J. Chem., 56, 1280-1283, (1978) ; R.B. Moodie, K. Schofield, P.G. Taylor, J. Chem. Soc., Perkin trans. II, 133, (1979). 14. Phenylnitrate is not described in literature. 15. C.R. Johnson, J.R.D. Mc Cant, J. Am. Chem. Soc., 86, 2935, (1964) ; C. Lagercrantz, Acta Chem. Scand., 23, 3259, (1969). 16. J.H. Ridd, H.A. Clemens, J.P.B. Sandall, J. Chem. Soc., Perkin trans. II, 1667-1672, (1984). 17. T.A. Turney, G.A. Wright, Chem. Rev., 59, 497, (1959). 18. R.J. Coombes, A.W. Diggle, S.P. Kempsel, Tetrahedron Lett., 34, 8557, (1993).

379

OXIDATION OF ALKYLPHENOLS TO HYDROXYBENZALDEHYDES

ERIC FACHE, DOMINIQUE LAUCHER, MICHEL COSTANTINI, MONIQUE BECLERE AND GILLES PERRIN-JANET Rh6ne-Poulenc Industrialisation, Centre de Recherche, d'Ing6nierie et de Technologie des Carri6res, 85, Avenue des Fr6res Perret, BP 62, 69192 Saint-Fons Cedex, France.

4-methylphenol is oxidized into p-hydroxybenzaldehyde by oxygen or air, in water / acetic acid media, in the presence of catalysts like Pd / C or Pd-Sn / C (66 % selectivity for total conversion). The former catalysts are not able to yield ohydroxybenzaldehyde from 2-methylphenol. However, good performances are reach with a Pd-Pt / C catalyst (60 % selectivity for total conversion )

INTRODUCTION Among hydroxybenzaldehydes, the o- and p- hydroxy isomers are the most important for commercial applications in agricultural, flavor and fragance, pharmaceutical or polymer fields (ref. 1). The two main processes for the manufacture of hydroxybenzaldehydes are both based on phenol. The most widely used process is the saligenin process. Hydroxybenzyl alcohols (o- and p- isomers) are produced from base - catalyzed reaction of formaldehyde with phenol (ref. 2). Air oxidation of these alcohols over a suitable catalyst (based on palladium or preferentially on platinum) produces hydroxybenzaldehydes (ref. 3). The Reimer Tiemann process allows the coproduction of o- and p- hydroxybenzaldehydes (ref. 4). Treatment of phenol with aqueous chloroform and sodium hydroxide leads to benzal chlorides which are rapidly hydrolyzed by alkaline medium to aldehydes. The previous processes need two chemical steps and produce salt effluents. More recently, the direct formation of hydroxybenzaldehydes by the oxidation of the corresponding alkylphenols was reported. However, the oxidation of 4- and 2methylphenol respectively into p- and o-hydroxybenzaldehydes remains difficult, leading very often to heavies. For instance, the catalytic systems used for the 38O

oxidation of substituted cresols (Table 1) are no efficient and / or no selective in the case of o- and p-cresols (ref. 6).

Table 1 : Usual catalytic systems for the oxidations of substituted cresols or derivatives

Catalysis type

Catalysts

Substrates

References

Basic

tBuOK (Stoichiometric reaction)

2,6-tert-butyl-4-methylphenol

5a

Homogeneous

Ce(OAc)3

2,6-tert-butyl-4-methylphenol 2,6-dimethoxy-4-mrthylphenol

5b

Homogeneous

Co(OAc)2 + Mn(OAc)2

3,4,5-trimethoxytoluene

5c

Homogeneous

CuC12 + amines or oximes

2,4,6-trimethylphenol

5d

Heterogeneous

Pd/C

2,4,6 trimethylphenol

5e

However, with cobalt or iron catalysts, good results are obtained when the phenol group is protected, either under its acetate form after reaction with acetic anhydride in an acetic acid medium (ref. 7), or as phenate when the oxidation takes place in a basic medium (at least three equivalents of base versus cresol) (ref. 8). These oxidations suffer from great drawbacks : in acetic medium, the reaction leads to poor selectivities at high conversions (acid formation) and needs one more step to recover the aldehyde under the phenolic form. The main limitation of the oxidation in basic media is the large coproduction of salt. Moreover, the oxidation of o-cresol appears more difficult than the p-cresol one. In this paper, we wish to report efficient methods to oxidize 4-methylphenol and 2methylphenol into the corresponding aldehydes which avoid the previous drawbacks.

OXIDATION

OF

4-METHYLPHENOL

IN

ACETIC

MEDIA

IN

THE

PRESENCE OF PALLADIUM-BASED CATALYSTS Oxidations of methylaromatic compounds, without phenolic group, in the presence of palladium based catalysts are well documented (ref. 9). Toluene (ref. 9), o-, m-, p-xylene (refs. 9c-e), mesitylene (ref. 9c), hexamethylbenzene (ref. 9c), o-methylanisole (ref. 9e) and p-methylanisole (ref. 9d) are among the main substrates which have been studied. The solvent of choice for the reaction is acetic acid and the main product is the corresponding benzylic acetate. Aldehyde 381

selectivity is low. According to our knowledge, the oxidation of 4-methylphenol in acetic acid medium has been reported only one time with a catalyst based on palladium, bismuth and chromium or manganese or silicium (ref. 10). Under these conditions, the main product is 4-hydroxybenzylacetate. As expected, we found that the oxidation of 4-methylphenol in acetic acid medium in the presence of Pd-Sn / C * catalyst leads to 4-hydroxybenzylic acetate with a good selectivity (Table 2, entry 2). The study of the reactionnal intermediates in such a medium shows the difficulty to oxidize the acetate into 4-hydroxybenzaldehyde under these conditions (Table 2, entry 5) while the esterification of alcohol by acetic acid is complete (Table 2, entry 3).

OXIDATION OF 4-METHYLPHENOL IN WATER / ACETIC MEDIA IN THE PRESENCE OF PALLADIUM-BASED CATALYSTS In the case of Pd-Sn / C catalysts, the addition of water to the acetic medium allows the shift of the oxidation selectivity from the acetate towards 4-hydroxybenzaldehyde (Table 2, entry 1). Moreover, in this new medium, 4-hydroxybenzylic alcohol is converted into 4-hydroxybenzaldehyde with 61% selectivity (Table 2, entry 4). The best selectivities in aldehyde are reached with media containing at least 50% in volume of water (Table 2, entries 1, 2, 6-7). At lower water concentration, 4-hydroxybenzylacetate is the main product of the reaction (entry 2). The presence of water is necessary to allow the equilibrium between 4-hydroxybenzylacetate, which is not oxidized, and 4-hydroxybenzylalcohol which is easily converted into 4-hydroxybenzaldehyde (Scheme 1).

OH~CH2OAc

~ ~

,IOH

NNN~

OH

OH

CHO

COOH

H20

O H ~ C H 2 O H / ~

Scheme 1 : Oxidation of 4-methylphenol in the presence of Pd-based catalysts in water-acetic acid media. 382

T a b l e 2. 9Oxidation o f cresols with palladium-based catalysts in acetic acid / water m e d i a

Ea'ms'

Substrate

AcOH/ H20

Catalyst

At (h)

ml/ml

Nature

mmol Pd

Conversion

Selectivities

%

1

2

3

%

1

p-cresol

25 / 25

Pd-Sn/C

0.25

2

87

2

9

64

9

2

p-cresol

50 / 0

Pd-Sn/C

0.25

0.25

84

5

70

10

3

3

4-hydroxybenzylalcohol 50 / 0

Pd-Sn/C

0.25

0.2

100

4

4-hydroxybenzylalcohol 20 / 30

Pd-Sn/C

0.25

1

100

-

-

61

7

5

4-hydroxybenzylacetate 50 / 0

Pd-Sn/C

0.25

2

35

-

-

47

11

100

6

p-cresol

40 / 10

Pd-Sn/C

0.25

2

93

3

27

50

6

7

p-cresol

10 / 40

Pd-Sn/C

0.25

2

76

1

7

60

8

8

p-cresol

25 / 25

Pd/C

0.15

2.5

55

8

20

61

5

9

p-cresol

25 / 25

Pd/C

0.15

21

99.5

e

e

66

31

10

o-cresol

25 / 25

Pd/C

0.15

4

95

-

-

-

4

11

o-cresol

25 / 25

Pd-Pt/C**

0.27

4

70

-

-

14

-

12

o-cresol

25 / 25

Pd-Pt/C

0.27

4

100

-

-

60

+BiO(NO3)

Rt/Bi=2.45

Substrate " 10 m m o l " K O A c " 10 m m o l

100~

9o x y g e n - 5 1/h.

! " 2 or 4 - h y d r o x y b e n z y l a l c o h o l 92 2 or 4 - h y d r o x y b e n z y l a c e t a t e 9 3 " 2 or 4 - h y d r o x y b e n z a l d e h y d e 94 - 2 or 4 - h y d r o x y b e n z o i c acid

383

CATALYSTS USED TO OXIDIZE 4-METHYLPHENOL IN WATERACETIC MEDIA Various Pd-based catalysts used in acetic acid / water media allow the oxidation of 4-methylphenol into 4-hydroxybenzaldehyde with good selectivity. The best results are obtained with two kinds of catalysts : Pd-Sn / C and prereduced Pd/C* (Table 2, entries 1 and 8). With total conversion of cresol, 4-hydroxybenzaldehyde can be obtained with an average selectivity of 65 %. 4-hydroxybenzoic acid is the main by-product (Table 1, entry 9). The key of the oxidation seems to be the oxidation degree of palladium which has to be as low as possible (reduction of Pd(OAc)2 by Sn (II) derivatives in the case of Pd-Sn catalysts, or reduction, for instance by hydrogen, in the case of Pd / C catalysts). Hence, in the transformation of 4-hydroxybenzylalcohol into 4hydroxybenzaldehyde, palladium behaves as a dehydrogenation catalyst and has to be at a low oxidation degree. Moreover, it is also well known that low oxidation state palladium is implicated in benzylic oxidations, via the cleavage of the benzylic C-H bond, while more electrophilic palladium with high oxidation degree favors the attack and functionnalization of the aromatic ring (ref. 11).

MAIN SIDE-REACTIONS IN THE CASE OF 4-METHYLPHENOL OXIDATION IN WATER / ACETIC MEDIA IN THE PRESENCE OF PALLADIUM-BASED CATALYSTS When initial concentration of 4-methylphenol is increased, the selectivity in aldehyde and more generally in the corresponding alcohol, acetate, aldehyde and acid decreases (Table 3). The loss of selectivity in concentrated media, is mainly due to polyethers like :

C H 3 ~ O(---CH2~ O)O- - CnH 2 ~ Formation of these compounds, not or poorly oxidable under the reactionnal conditions, is due to the nucleophilic reaction of 4-methylphenol on 4-hydroxybenzylalcohol or on the corresponding acetate. With these results, we can complete the reactionnal scheme (Scheme 2).

384

Table 3- Oxidation of p-cresol in water-acetic acid (1/1) media by palladium-based catalysts Influence of the concentration of p-cresol on the selectivity of the oxidation

[p-cresol] M

Conversion %

Selectivities 3

%

!+2+3+4

0.2

99.5*

66

97

2.0

80**

35

50

KOAc 910 mmol; Pd/C (3 % Pd, Pd 0.15 mmol), AcOH / H20 =25 / 25 ml; 100~ oxygen 95 1 / h., * 21 h and p-cresol 10 mmol, ** 26 h, p-cresol 110 mmol and KOAc 100 mmol.

1" 4-hydroxybenzylalcohol" 2- 4-hydroxybenzylacetate; 3 4-hydroxybenzaldehyde; 4_ :4-hydroxybenzoic acid.

~ Cresol

OH

1 Polyethers

/

HO--@CH2ObI

CHO

Scheme 2 9 Oxidation of 4-methylphenol in water-acetic acid media in the presence of palladium based catalysts 9

H o w e v e r , the formation of previous by-products can be strongly decreased by continuous injection of cresol in the medium in order to keep a low instantaneous concentration of cresol in the medium. M o r e o v e r addition of bismuth allows a significant increase on aldehyde selectivity (Table 4).

385

Table 4 " Oxidation of p-cresol in water-acetic acid (1/1) media by palladium based-catalysts. Continuous injection of cresol.

BiONO3/Pd

Conversion %

Selectivities % 3

1+2+3+4

0

96.5

50

70

0.3

93.8

65

82

KOAc " 28 mmol; Pd/C (3% Pd, Pd 0.15 mmol), AcOH / H20 = 12.5 / 12.5 ml- 100~ oxygen" 5 1 / h, 20 h initial p-cresol 4 mmol, injection of p-cresol 41 mmol (12 h). 1" 4-hydroxybenzylalcohol; 2 94-hydroxybenzylacetate" 3 94-hydroxybenzaldehyde; _4:4-hydroxybenzoic acid.

RECYCLING

O F C A T A L Y S T S ; E X A M P L E O F T H E O X I D A T I O N O F 4-

METHYLPHENOL The industrial reality of a catalyst is only achieved if it can be easily recovered and if it keeps its activity and selectivity. These conditions are nearly gathered only with Pd-Bi / C catalysts (Figs 1-2). Finally, it has been found that promotion of palladium by bismuth not only increases the selectivity in aldehyde but also limits the deactivation of the catalysts. Similar results have been published in the past decade on the partial oxidation of alcohols with similar catalytic systems (ref. 13). Various interpretations on the role of bismuth have been suggested : among them, resistance of Pd/C against overoxidation and surface orientation of the reactant suppressing the formation and strong adsorption of poisoning intermediates are also problably the main reasons of the improved performances in the oxidation of p-cresol.

386

100 90 80

l"

70 60

Conversion with Pd/C I ,[! Conversion with Pd/C + Bi

TT% 50

40 30 20 10 0

1

0

1

,

,

,

2

3

4

Recycle

Fig. 1 9 Influence on the activity with the recycling of catalysts Pd / C and Pd / C + Bi (conditions described in Table 4)

80 B

70

[l

60 50

I II Selectivity with Pd/C I m Selectivity with Pd/C + Bi

4O RT % 30 2O 10 0

1

2

3

4

Recycle

Fig.2 9 Influence on the selectivity with the recycling of catalysts Pd / C and Pd / C +Bi (conditions described in Table 4)

O X I D A T I O N OF O T H E R C O M P O U N D S IN W A T E R / A C E T I C ACID M E D I A IN T H E P R E S E N C E OF Pd B A S E D C A T A L Y S T S . C A S E OF 2 - M E T H Y L P H E N O L (ref. 14) Of course, previous catalytic

systems

allow

the

oxidation

of

other

methylaromafic compounds into aldehydes, especially compounds which are not phenolic

and

(p-methylanisole,

which

bear

electrodonnating

3,4-(methylenedioxy)toluene

groups

on

the

aromatic

for instance). However,

ring

different

catalytic methods already exist for the oxidation of these kinds of substituted cresols (Table 1). So, the new systems would be really interesting only if they allow the oxidation of substrates, which are very difficult to oxidize by classical methods. 387

Among these substrates, we can find not only 4-methylphenol but also 2-methylphenol. Unfortunatly, oxidation of 2-methylphenol with the previous catalysts (Pd-Sn/C and Pd/C) only gives small amounts of 2-hydroxybenzoic acid and heavies (table 2, entry 10). These heavies are polyethers probably obtained by reaction of o-cresol itself with 2-hydroxybenzylacetate or 2-hydroxybenzylalcohol. Apparently, palladium catalysts activate the benzylic C-H bond of o-cresol, but the oxidation of the intermediates seems less rapid than side reactions. On the other hand, we have check that platinum catalysts, which are known to be excellent catalysts for the oxidation of 2-hydroxybenzylalcohol into 2-hydroxybenzaldehyde in basic aqueous medium (ref. 3), is unable to activate efficiently the benzylic C-H bond of cresols. We synthesized bimetallic catalysts, Pd-Pt / C** , with the hope that palladium would activate benzylic C-H bond and platinum would accelerate the oxidation of intermediate alcohols. Effectively, this new catalyst allows to recover 2-hydroxybenzaldehyde with 14 % selectivity at 70% conversion (Table 2, entries 11-12). Addition of bismuth salts are known to improve the aldehyde yield in the saligenin process. With such additives, the selectivity of the aldehyde can reached 60% for a total cresol conversion. Of course Pd-Pt / C can also oxidize 4-methylphenol but it does not bring significant improvement compared to initial catalysts.

CONCLUSION So, we have discovered new and original catalytic conditions which allow an easy transformation of alkylphenols into the corresponding hydroxybenzaldehydes. Hence, 4-methylphenol is oxidized into p-hydroxybenzaldehyde by oxygen or air, in water / acetic acid media, in the presence of catalysts like Pd/C or Pd-Sn/C (66 % selectivity for total conversion). The former catalysts are not able to yield o-hydroxybenzaldehyde from 2-methylphenol. However, good performances are reach with a Pd-Pt/C catalyst (60 % selectivity for total conversion).

References

1. 2. 3. 4.

In ,, Encyclopedia of chemical Technology ~,, Kirk-Other, third edition, 13, pp.70, John Wiley (New-York), (1981). K.C. Eapen and L. M. Yeddanapalli, Makromol. Chem., 1968, 119, 4. J. Le Ludec, Ger Often 2,612,844, (1976), (to Rh6ne-Poulenc SA). H. Wynberg, Chem. Rev., 60, 169, (1969).

388

5.

a) b)

c) d)

e) 6. 7. 8.

9.

a) b) c) d)

e) 10. 11. 12. 13. 14.

A. Nishinaga, T. Itahara, T. Shimizu, T. Matsuura, J. Am. Chem. Soc., (1978), 100 (6), 1820. T. Yuschikuni, J. Mol. Catal., 1992, 72, 29 ; N. Kitajima, S. Sunaga, Y. Moro-Oka, T. Yoshikuni, M. Akada, Y. Tomotaki, M. Taniguchi, Bull. Chem. Soc. Jpn., 61, 1035 (1988). N. Kitajima, S. Sunaga, Y. Moro-Oka, T. Yoshikuni, M. Akada, Y.Tomotaki, M. Taniguchi, Bull. Chem. Soc. Jpn., 61,967, (1988). M. Shimuzu, Y. Watanabe, H. Orita, T. Hayakawa, K. Takehira, Bull. Chem. Soc. Jpn., 66, 251 (1993). M. Shimuzu, Y. Watanabe, H. Orita, T. Hayakawa, K. Takehira, Tet. Lea., 32 (18), 2053 (1991). K. Takehira, M. Shimuzu, Y. Watanabe, H. Orita, T. Hayakawa, Tet. Lett., 31 (18), 2607, (1990). U.S. 4, 915,875 (04/11/1986), (to Dow Chemical). S . N . Sharma, S. B. Chandalia, J. Chem. Tech. Biotechnol., 49, 141, (1990), and references therein. JP 63154644, (1986) (to Mitsui Petrochemical), JP 62242644 A, (1986), (to Mitsui Petrochemical). J. Dakka, D. A. Sheldon, NL. 9200968-A (1992) (to DSM NV) ; JP 2172940, 2172941 and 2172942 (1988) (to Sumitomo) ; T. A. Andrew, M. Needham, US. Pat. 4,453,016 (1982) and U.S. Pat. 4, 471, 140 (1984) (to Dow Chemical) ; K. Freimund RShrscheid, U.S. Pat. 4, 748, 278 (31/05/1988) (to Hoechst), EP 323290-A (1987) (to Air Liquide) ; A. Schnatterer, H. Flege, US. 4929766 (1989) and US.Pat. 5130493 (1991), (to Bayer), A. Nishinaga, T. Itahara and T Matsuura, Angew. Chem. Internat. Edit., 14 (5), 356, (1975). S.K. Tanielyan, R. Augustine, J. Mol. Catal., 87, 311, (1994). E. Benazzi, H. Mimoun, C. J. Cameron, J. Catal., 140, 311, (1993). E. Benazzi, C.J. Cameron and H. Mimoun, J. Mol. Catal., 69, 299 (1991). D.R. Bryant, J. E. McKeon, B. C. Ream, J. Org. Chem., 33 (11), 4125 (1968). A.B. Goel, Inorg. Chim. Acta, 129, L31, (1987). A. B. Goel, Inorg. Chim. Acta, 121, L l l , (1986). A. B. Goel, Inorg. Chim. Acta, 90, (1984), L15. C. H. Bushweller, Tet. Lett., 58, 6123 (1968). D.R. Bryant, J. E. McKeon, B. C. Ream, Tet. Lett., 30, 3371, (1968) Matsuda, Teruo and Shirafuji, Tamio, JP 7879832 (1976), (to Sumitomo Chemical Co, Ltd). J.E. Lyons, Catalysis Today,1988, 3, 245. J.F. Lepage, in ,< Catalyse de contact ,,, Tecnnip Editions, 1978. T. Mallat, Z. Bodnar, P. Hug and A. Baiker, J. Catal., 153, 131, (1995) and references therein. E. Fache, M. Costantini, D. Laucher, FR 9207950, (29/06/92) and FR 9303488, (26/03/93) (to Rh6ne-Poulenc).

Pd-Sn / C catalyst is obtained by adding to a solution of 2.44 g palladium (II) acetate (10 mmol, Johnson-Matthey), 9.82 g of potassium acetate (100 mmol) in 400 ml acetic acid firstly 20 g of Ceca 3S charcoal (optionally treated with concentrated nitric acid according to known procedure (ref. 9b)) and then 16.1 g tin (II) 2-ethylhexanoate (39,75 mmol, Strem). The suspension is vigourously stirred and heated at 100~ for 4 hours. After cooling, the catalyst is recovered by filtration, washed with acetic acid and water and dried 5 hours under reduced pressure (50 mbar) at 50~ (Pd : 4,3 %; Sn : 2%).

389

Pd / C catalysts are synthezised according to usual methods (ref. 12). For instance, catalyst can be prepared by adding to a solution of 0.32 g palladium (II) acetate (1.4 mmol, JohnsonMatthey) and 0.98 g of potassium acetate (10 mmol) in 100 ml acetic acid 5 g of ceca 3S charcoal (optionally treated with concentrated nitric acid according to known procedure9b). The suspension is introduced in a stainless steel autoclave, heated at 100~ under 20 bar of hydrogen for 15 hours. After cooling, the catalyst is recovered by filtration, carefully washed with water and dried 5 hours under reduced pressure (50 mbar) at 50~ (Pd : 3.5%). Immediatly before the oxidation, the catalyst is reduced by hydrogen (200~ hydrogen : 1 l/h, 2 hours) ** The Pd-Pt / C catalyst is prepared according to the procedure describe for Pd / C catalyst, by mixing with the palladium salt, a platinum (II) or (IV) salt (hexachloroplatinic acid for instance) in the ratio (Pd: 2.85 %; Pt : 0.25 %).

390

LARGE P O R E TI-BETA ZEOLITE WITH VERY LOW ALUM]NIUM CONTENT 9AN ACTIVE AND SELECTIVE CATALYST FOR OXIDATIONS USING HYDROGEN PEROXIDE

MIGUEL A. CAMBLOR a) MICHEL COSTANTINI b) AVELINO CORMA a) PATRICIA ESTEVE a) LAURENT GILBERT b) AGUSTIN MARTINEZ a) AND SUSANA VALENCIA

a)

a) Instituto de Tecnologia Quimica (CSIC-UPV), Avda. Los Naranjos s/n, 46071 Valencia, Spain. b) Rh6ne-Poulenc Industrialisation, Centre de Recherche, d'Ing6nierie et de Technologie des Carri~res, 85 Avenue des Fr~res Perret, B.P. 62, 69192 SaintFons Cedex, France.

ABSTRACT The new large pore Ti-Beta zeolite has been synthesized in a wide range of chemical compositions and its activity and selectivity in the epoxidation of olefins and the hydroxylation of phenol has been tested. Several new synthetic procedures have been developed yielding innovative materials with chemical compositions out of the range previously known for zeolite Beta and with a predesigned composition profile in the crystallites. Catalysts with a reduced A1 content of up to 0.1 A1 atoms per unit cell of 64 tetrahedra and below with all the A1 confined into the very inner core of the crystallites show a good activity in the epoxidation of n-hexane with an enhanced selectivity to the epoxide. Optimization of the catalyst and of the reaction conditions for the selective hydroxylation of phenol can yield a valuable catalyst for this industrially important reaction. INTRODUCTION Zeolites are microporous crystalline solids which find a wide variety of industrial applications in the fields of ion exchange and separation, purification and catalytic transformation of organic compounds. As heterogeneous catalysts, most of their uses have been as acid catalysts where the combination of high acidity, high specific surface area and the shape selectivity derived from the size and shape of their 391

microporous channel systems made zeolites outstanding materials with no competitors in catalytic cracking and other petroleum and petrochemical processes. This has been the field of zeolites for over 30 years. However, recent advances in zeolite science are spreading the interest of zeolites as catalyst for a number of applications other than acid catalysis, including base and oxidation catalysis and photochemistry. An increased importance of zeolite catalysis in commodities and fine chemicals production can be thus envisaged. The interest of zeolites as oxidation catalysts begun with the synthesis of TS-1 (titaniumsilicalite-1) (ref. 1) and the subsequent reports on its catalytic performance using hydrogen peroxide in the presence of water. TS-1 is an active and selective catalyst in the epoxidation of olefins, the hydroxylation of aromatics, the ammoxidation of cyclohexanone (with NH 3 and H202), the oxidation of alcohols to ketones (ref. 2) and even the oxidation of alkanes to alcohols and ketones (refs. 3, 4). It has the outstanding property of being highly active in the presence of water, in contrast to other heterogeneous catalysts, even if they have the same overall composition (like the TiO2/SiO2 catalyst of Shell) (ref. 5). This can be an effect of having Ti species confined into the hydrophobic microporous channels, or of having Ti in a special environment or coordination, and makes Ti-zeolites an important subject of study from both the academic and the industrial points of view. TS-1 has a medium pore channel (-5.5,4,) which imposes severe geometrical restrictions to the size of the organic substrates to be oxidized (ref. 6), and also restricts the use of oxidating agents to H202. To overcome this limitations Ti-zeolites with larger pores were desirable. Along these lines of thought we reported the first synthesis of a large pore zeolite, Ti-Beta, with a tridimensional system of large pore channels (7-6.5A) (ref. 7). It was shown that this material while having lower intrinsic activity than TS-1 for the oxidation of organic substrates small enough to have no restrictions to enter the TS-1 pores, is more active than TS-1 for carrying out the oxidation of larger compounds (ref. 8). Additionally, Ti-Beta is active in oxidation reactions using tertbutyl hydroperoxide (ref. 9). Up to now, all the Ti-Beta samples reported contain A1 in framework positions. This implies that in activated samples, besides the Ti Redox sites, acid sites associated to framework A1 will also be present. The presence of the acid sites or the associated A1 may negatively affect the oxidation activity, but it certainly can catalyze other reactions such as formation of diols from epoxides, and undesired polymerization reactions. It is therefore of clear interest to prepare A1 free Ti-Beta zeolite by direct synthesis. Here we report a synthesis strategy which has allowed us to produce Ti-Beta samples with a much higher Si/A1 ratio than any one reported up to now, together with predesigned zeolite crystals containing very low A1 content all of it located in the 392

inner core of the crystallites, while leaving Ti in the outershell. This was expected to influence the catalytic activity and selectivity of Ti-Beta, and the results will be presented here. EXPERIMENTAL

Synthesis The synthesis mixtures were prepared using tetraethylammonium hydroxide (40 % aqueous solution, Alfa or 35 % aqueous solution, Aldrich) with a very low alkali

cations

content

(Na < 2ppm,

K < 0.Sppm),

deionized

water,

tetraethylorthotitanate (TEOT) or tetrabutylorthotitanate (TBOT) as a source of Ti and amorphous silica (Aerosil 200, Degussa) or tetraethylorthosilicate (TEOS, Aldrich) as the source of silica. Depending on the synthesis method a source of aluminum can be directly added to the synthesis mixture (metal A1, aluminum halide, etc.) or incorporated in the seeds of aluminosilicate zeolite Beta. Four synthetic procedures were developed and Table 1 summarizes the typical ranges of chemical composition of the initial mixture and the typical results of the syntheses. The methods are denoted according to the nature of the silica source and, in one case, the use of seeds. The preparation of the reaction mixtures was as follows" Amorphous silica method - TEAOH was diluted in a certain amount of water and the TEOT and Aerosil were added sequentially under stirring. Finally, a solution of aluminum nitrate in TEAOH and water was also added. (ref. 10) TEOS method- TEOS is hydrolized in an aqueous solution of TEAOH with stirring, then TEOT is also hydrolized. Finally, a solution of aluminum nitrate in TEAOH and water is added and the mixture is left, while stirring, until all the ethanol formed in the hydrolisis is evaporated. (ref. 11) TEOS/seeds method - TEOS is hydrolized in an aqueous solution of TEAOH under stirring, then TEOT is also hydrolized and the mixture is left, while stirring, until complete evaporation of the ethanol produced. If desired, H202 can be added either before or after TEOT addition. Then seeds are added to the clear solution formed and the mixture is kept under stirring to get an homogeneous mixture. Typically, the amount of seeds is around 2.5-3 g of zeolite Beta crystals per 100 g of SiO2 in the reaction mixture. No A1 solution is added. (ref. 12) TiO2/SiO 2 cogel method - A TiO2/SiO 2 coprecipitate is used as the source of Si and Ti and this is wetness impregnated with a solution containing A1 and TEAOH. The TiO2/SiO 2 cogel is prepared by first hydrolizing TEOS in a mildly acidic solution (HC1). Then a solution of TBOT in isopropanol is added under stirring. The pH of the resulting clear solution is then rised to 6.0 by addition of a small amount of a base (typically, tetraethylammonium hydroxide or tetrapropylammonium hydroxide). This 393

causes the precipitation of the TiO2/SiO 2 cogel, which is then dried to 110~ The dried cogel is then wetness impregnated with a solution containing TEAOH and a source of A1. (ref. 13) The synthesis mixtures were crystallized by heating at the crystallization temperature (usually 135-145~ in PTFE lined stainless steel 60 ml autoclaves. During crystallization the autoclaves are rotated at 60 rpm. After quenching with tap water, the solids are recovered by centrifugation and washed with distillate water until pH < 9. Then the solids were dried at 100~ 580~

for several hours and calcined at

to remove the tetraethylammonium cations occluded into the zeolitic channels.

Characterization Phase purity of the zeolites was determined by conventional X Ray powder diffraction (XRD) methods using a Philips 1060 diffractometer with a graphite monochromator and a variable divergence slit operating in the constant area mode. Cu K~ radiation (k = 1.541A) was used. The crystallinity was determined by measuring the total area under the main diffracted peak (2 0 -- 22.5 o) and comparing it with that of a highly crystalline aluminosilicate zeolite Beta. The chemical composition was determined by atomic absorption spectroscopy using a Spectra A-10 Plus Varian spectrometer. The absence of Ti oxides out of the zeolite framework was checked by diffuse reflectance Uv/Visible spectroscopy (Shimadzu UV-210PC spectrophotometer, reference BaSO4). Catalytic tests 1-hexene oxidation was carried out at 50~ in a round bottom glass flask equipped with a condenser and a magnetic stirrer. In a typical run 33 mmol of 1hexene, 23.6 g of methanol and 0.8 g of H202 aqueous solution (35 wt %) are mixed in the flask and heated to the reaction temperature under vigorous agitation. Then 0.2 g of catalyst is added to the reaction mixture (time zero). The kinetics of the reaction was followed by taking aliquots at five reaction times (between 0.5 and 5 hours). The products were analyzed by gas chromatography using a capillary column (5 % methylphenylsilicone, 25 m length) and a FID detector. For the H202/1hexene ratio used the maximum conversion of 1-hexene would be 25 %. The catalytic oxidation of phenol with hydrogen peroxide was performed in a round-bottom flask equipped with a condenser and a magnetic stirrer. In a typical reaction, given amounts of catalyst, phenol and solvent were mixed in the flask and heated to 80~ under vigorous agitation. The reaction was started by dropwise addition of 35 wt % aqueous hydrogen peroxide solution using a perfusion pump (addition time --1 min.). The reaction was stopped at 3 h by cooling the flask to 394

room temperature and then, the mixture was filtered to remove the catalyst. When water was used as solvent, methanol was added before filtering to homogenize the mixture. Products were analized by HPLC in a Waters Associates 440 apparatus equipped with a UV detector (254 nm) using a 100 RP-18 column (5 ~tm, 125 mm.). The amount of unreacted hydrogen peroxide was determined in both cases by iodometric titration. RESULTS AND DISCUSSION

Design of the synthetic procedure Four methods have been developed for synthesizing highly crystalline zeolite TiBeta. The aim was to prepare a catalyst with as low Al content as possible, in view of the detrimental effect of the acid character and the hydrophilic nature of the material in the catalytic oxidations with H202. Two of these methods afford Ti-beta with Si/A1 ratios more than 50 % larger than those previously claimed for zeolite Beta (5-100). However, the A1 content was still high. Then we designed a synthesis procedure which, while producing Ti-Beta with very high Si/A1 ratios, yields a material with extreme chemical zoning: an inner core of aluminosilicate zeolite is covered by a titanosilicate outer shell. The following illustrates how this synthesis procedure was designed. The

first

methods

reported

for

the

synthesis

of Ti-Beta

involved

the

crystallization, without the aid of seeds, of synthesis mixtures prepared using one of two different sources of Si (amorphous silica (7) and TEOS (11)) in the absence of alkali cations. Both methods gave similar results (Table 1) except that the TEOS method produced, for the same crystallization time, temperature and overall chemical composition in the starting mixture, higher Si/A1 ratios in the final zeolite. That just means that TEOS is a more reactive silica source than Aerosil. However, the upper limit for the Si/A1 ratio of the zeolite was the same in both cases (--150) (14). The isomorphous substitution of Si by Ti was evidenced by XRD, IR, XANES and EXAFS, (ref. 15) and the catalysts were found active in the oxidation of alkanes to alcohols and ketones (ref. 7) and in the oxidation of olefins (ref. 11) using hydrogen peroxide or tert-bu~l hydroperoxide (ref. 9). However, when H202 was used as the oxidant selectivity to the epoxide was found to be low due to the presence of acid sites that catalyzed the addition of the solvent (usually methanol) to the oxirane ring (see below). Using tert-butyl hydroperoxide as the oxidant afforded -- 100% selectivity to the epoxide with a somewhat lower oxidation rate. (ref. 9)

395

Interestingly, we observed a competition between Ti 4+ and A13+ in the crystallization of zeolite Ti-Beta (ref. 14). This competition was reflected in the following facts : -the higher the A1 content in the initial mixture, the lower the Ti content in the zeolite.

-the higher the Ti content in the initial mixture, the lower the A1 content in the zeolite. -

the Si/A1 (or (Si + Ti)/A1) ratios of Ti-Beta largely exceed the upper limit previously

found for zeolite Beta synthesized from gels containing alkali cations in the absence of Ti (Si/A1 ---40) (ref. 16). - t h e surface chemical analysis (XPS) indicated that, for Ti-Beta samples with Si/A1 ratios higher than 100, the outer shell of the crystallites contained no A1 at all, contrarily to what is found in the absence of Ti. To explain this competition we hypothesized that Ti could play a role similar to that of A1 in the crystallization of zeolite Beta, that is, the creation of negative charges in the framework and thus its stabilization by interaction with the TEA § templating cations. This hypothesis was also supported by the fact that the amount of TEA § cations decomposing at T > 6 2 0 K in air (as determined by thermal analysis) was dependent on the total amount of A1 + Ti, rather than only on A1 (ref. 14). This hypothesis required the ability of zeolitic Ti to change its coordination number, something which obtained substantial support from XANES and EXAFS measurements (ref. 15). As a result from this early work we thought that zeolite Beta crystals can grow without the incorporation of a trivalent element (A1, Ga, B, Fe,...) provided that Ti is incorporated into the framework. This was actually supported by the fact that, as mentioned above, Ti-Beta samples with Si/A1 ratios higher than 100 have no A1 in its outer shell, which means that in the last steps of its crystallization these samples grow without A1 incorporation. Unfortunately, we were unable to synthesize Ti-Beta in the absence of aluminium (or other T3+), the conclusion being that a trivalente element is necessary for zeolite Beta to nucleate. Obviously, crystallization of pure titaniosilicate zeolite Beta without A1 could then be possible if the nucleation problem was solved. We have done this by seeding with highly active zeolite Beta seeds comprised of very small zeolite Beta crystals (typically --0.05 mm and below) showing good stability in the synthesis media (TEOS/Seed method) (ref. 12). In this way it is possible to synthesize highly crystalline zeolite Ti-Beta with Si/A1 ratios well above those obtained by other synthesis procedures, for example Si/A1 ratios about 1000. Additionally, as shown by XPS, the crystals obtained by this procedure consist of an inner core of 396

aluminosilicate composition (which are basically the seeding crystals) covered by an outer shell of titanosilicate composition and essentially free of A1. The Ti-Beta outer shell can account for up to about 97.5 % of the mass of crystals. Fig. 1 schematically illustrates the chemical zoning in these "second generation" Ti-Beta materials.

Fig. 1. Schematic representation of chemical zoning in Ti-Beta catalysts prepared by the TEOS/Seed method

Finally, we have developed still another synthetic procedure (TiO2/SiO 2 cogel method), aimed to afford very high yields of Ti-Beta zeolite using a small amount of tetraethylammonium hydroxide (ref. 13). The method consists in the crystallization of a TiO2/SiO 2 cogel after wetness impregnation with a solution containing A13+ and TEAOH. This method gives good results for intermediate Si/A1 ratios but it doesn't allow the synthesis of Ti-Beta with A1 contents as low as those obtained with the TEOS/Seed method (Table 1).

397

Table 1. Methods for the synthesis of zeolite Ti-Betaa Typical Ti-Beta products

Typical gel compositions Method

SiO2/A1203 H20/SiO2

Zeolite Betaa

30-800

10-15

Amorphous silica

50-800

10-15

TEOS/seeds TiO2/SiO2 cogel

c. d.

20400

Si/A1

%TiO2 Yieldd

0.3-0.5

1140

0

0-10

0.5

50-150

1-6

-- 10

1-6

--- 10

, . .

TEOS

a. b.

SiO2frio 2 TEAOH/SiO2

50-800

10-15

20400

0.5

50-150

400-2000 b

10-15

20-1000

0.5

100-1(130b 0.3-6 c

15-30

30-120

0.4

50-300

15-30

50-800

5

1-6

The synthesis of zeolite Beta with no Ti is included for comparison (16) All the A1 is incorporated to the synthesis mixture in the aluminosilicate zeolite Beta seeds, and is confined to the inner core of the final product. Ti is incorporated to the outer shell of the crystallites. g of zeolite per 100g of initial mixture. To further illustrate the differences between the methods, Fig. 2 and 3 show the

yield and the Si/A1 ratio, respectively, of the zeolite as a function of time of crystallization for synthesis runned by the four methods. Obviously, it is not possible to compare synthesis with the same chemical compositions because of the differences of the methods. Accordingly, in Fig. 2 and 3 we compare synthesis runs that gave, for every method, high Si/A1 ratios and high zeolite yields. In this figures it is seen how our

"classical" syntheses of Ti-Beta zeolite (amorphous silica and TEOS

methods)

have

been

surpassed

by the

new,

previously

unpublished

methods

(TEOS/seed and cogel methods), if the Si/A1 ratio of the zeolite and its yield are compared. Furthermore, the TEOS/seed method is the most versatile one in terms of varying the chemical composition of the zeolite, and thus we have prepared materials wich are essentially pure silica (A1 plus Ti contents below 0.2 atoms per unit cell of 64 tetrahedra).

398

g zeolite / 100 g gel 25

20

J

S

15

10

0

0

5

10

15

Time (days) Fig. 2. Variation of the yield of Ti-Beta zeolite as a function of crystallization time at 135~ Gels prepared by the amorphous silica (o), TEOS (+), TEOS/Seed (*) and cogel ( I ) methods

Si / A1 in zeolite 1.000

800

600

400

200 _____.-.+ O

L

0

5

10

15

Time (days) Fig. 3. Variation of Si/A1 ratio of Ti-Beta zeolite as a function of crystallization time at 135~ Gels prepared by the amorphous silica ( ) , TEOS (+), TEOS/Seed (*) and cogel (n) methods. 399

C A T A L Y T I C TESTS

Epoxidation of 1-hexene Table 2 lists results obtained in the oxidation of 1-hexene with H20 2 using representative Ti-Beta zeolites prepared by the new TEOS/seed and cogel methods. The influence of the A1 content of Ti-Beta on the selectivity to the epoxide is clearly seen in Table 2. For the same level of 1-hexene conversion, the lower the A1 content in the catalyst the higher the selectivity to the epoxide. This is a consequence of the presence of strong acid sites due to the presence of A104- units in the framework. These acid sites act as catalysts for the opening of the oxirane ring by addition of either water produced in the decomposition of H20 2 (to give the glycol product) or methanol (to give the methyl glycol ether). The relationship between the selectivity and the A1 contem for a given conversion is not linear, being the enhancemem in selectivity as the A1 content increases more remarkable, when the lower the A1 content is. Table 2. Influence of the A1 content of the Ti-Beta on the selectivity to epoxide during the oxidation of 1-hexene at -- 4 % hexene conversion Si / A1 ratio

epoxide selectivity (%)

43 300 470 550 700

3 43 61 73 85

The epoxide selectivity problem can be completely solved by exchanging the zeolitic H § by Na § The Na § form of the catalyst show a --100% selectivity to the epoxide, while only a minor decrease in the activity is found (Table 3). However, it should be considered the leaching of Na § with time.

400

Catalyst

Type

SilAl

%Ti02

L'ogel

43

4.2

Cogel

300

4.7

React ion Time (h) 0.5 I 2 3.5 5 0.5 1

TEOSIS

470

3.3

2 3.75 5 0.5 1

TEOS/S

TEOSIS

550

550 (Na')'

3.0

3.0

2 3.75 5 0.5 1 2 3.83 5 0.5 1

TEOSIS

700

6.3

2 3.83 5 0.5 1.2 2 3.5 5

Product selectivity (mol %)

1-c6

conver. (mol%) 2.32

i:::

9.13 10.97 3.68 6.25 9.47 13.72 15.66 1.84 3.25 5.35 8.10 10.22 2.00

i::;

1 1.20

12.12 1.10 2.41 5.03 11.03 12.93 7.02 12.17 16.37 19.93 19.68

1 I II I

I

I

I

1 1

I

1 I

I

I

Epoxide

Glycol

MGE

8.09 3.38 1.36 0.71 0.46 45.94 25.87 15.28 9.69 7.63 87.73 66.49 46.11 32.96 27.30 98.72 75.82 52.00 33.38 31.74 100.00 100.00 100.00 99.27 98.47 56.11 37.73 27.77 19. I7 8.59

0 0 0 0 0 0 0 0 0 1.47 0 0 0 0 3.15 0 0 0 0.35 2.09 0 0 0 0 0 0 0.9 4.40 5.66 7.13

91.91 96.62 98.64 99.29 99.54 54.06 74.13 84.72 90.31 90.90 12.27 33.51 53.89 67.04 69.55 1.28 24.18 48.00 66.27 66.17 0 0 0 0.73 1.53 43.89 61.37 67.83 75.17 84.28

H202 (niol %)

1 I I

Conv.

Selec.

10.94 23.51 36.79 50.95 58.09 17.80

80.95 63.09 67.85 68.35 72.02 78.61 77.76

I

I

75.93 7.76

78.42 90.94 80.38

I

I I

I I

52.83 8.23

74.03 92.12

30.34

79.49

59.66 4.72

76.88 88.92 81.16

I 1

59.66 35.90

82.91 84.39

72.5 I

97.40

94.10

90.22

a) Nat form of Ti-Beta zeolite, obtained by contacting the zeolite with a Na acetate solution under reflux conditions.

TON (moVmol of Ti 7.5 12.5 21.1 29.4 35.3 10.6 18.0 27.3 39.5 45. I 7.8 13.7 22.6 34.3 43.2 9.1 16.4 28.9 50.7 54.9 5.0 10.9 22.8 50.0 58.6 15.I 26.1 35.2 42.8 42.3

In addition to the effect on the product selectivity, it is seen in Table 3 that, for a given synthetic method, the activity (1-hexene and H202 conversion) as well as the selectivity of H202 increases as the A1 content of the zeolite decreases. These results show that the AI content of zeolite Ti-Beta is one of the most important factors in determining its activity and selectivity in oxidation reactions, and the benefits that the new methods of synthesizing Ti-Beta with low A1 content can provide.

Hydroxylation of phenol For this reaction a high selectivity to diphenols with a high para-selectivity is desired. It appears that both parameters are related, so generally the higher the selectivity to diphenols the lower the catechol/hydroquinone ratio. We have found with Ti-Beta catalysts that the synthesis procedure is very important in determining this relationship. Thus, as it is shown in Fig. 4 (hydroxylation using acetone as a solvent), with catalysts synthesized by the TEOS/seed procedure it is possible to obtain a much higher selectivity to diphenols for a given catechol/hydroquinone ratio.

Diphenols selectivity (mol/100rnol H202) 70

60

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

'

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

a 9

40 .

.

.

.

w

.

.

.

.

.

.

.

.

.

.

.

30

20

1

1,1

1,2

1,3

1,4

1,5

1,6

1,7

1,8

1,9

2

2,1

2,2

CTOL/HO.

Fig. 4.

Selectivity to diphenols (in mol % relative to H202) vs catechol to hydroquinone ratio in the hydroxylation of phenol with H202 using several Ti-Beta catalysts. Phenol / acetone / U20 = 50 / 65 / 6.5 (mol) : Ti-Beta / phenol = 4.5 g/mol ; H202/phenol = 5 % (mol). Reaction temperature 80~ 3h Reaction time. Ti-Beta materials prepared by TEOS / Seed (=), cogel (11), aerosil (+) and TEOS (*) methods. 402

In Table 4 we presem the results of catalytic tests in several conditions using a TiBeta prepared by the cogel method. There it can be seen that water is a good solvent from the point of view of activity and selectivity and that it is possible to obtain about equimolecular amounts of catechol and hydroquinone while keeping the selectivity to diphenols above 50 %. Table 4. Hydroxylation of phenol using a Ti-Beta catalysta Results referred to H202 (molar)

Reaction conditions H20 Solvem (g) (g) 0.23 Acetone, '7.55 1.55 t-BOH, 7.4 t-BOH,7.4 9.4 9.4 18.8 18.8 18.8 CH3CN18.8 9.0 Dioxane, 1.6 -

T (~ 80 80 80 94 80 80 80 80 80 80

Time (h) 3 3 3 3 3 3 1b 3c 3 3

Conversion ( %) 94 95.5 98 100 98 100 98.5 100 94 97

HQ

CTOL

17.5 14 17 24.5 25 24.5 27 26.5 17.5 23

23 13 16 24.5 22 24.5 28.5 25 28.5 21

Total CTOL/HQ diphenols 40.5 1.3 27 0.95 33 0.92 49 1.0 47 0.88 49 1.0 55.5 1.05 51.5 0.94 46 1.6 44 0.91

a) 9.4g phenol; H202/phenol:5%mol 9 0.45g catalyst, Ti-Beta (cogel method), Si/A1 73, 3.4%TIO2. b) 0.9g catalyst c) 1.35g catalyst

CONCLUSIONS It has been presented here that there is not a unique Ti-Beta material, but the characteristics and catalytic performance strongly depend on chemical composition and synthesis procedure. Then, new synthesis procedures which allow to prepare samples with much lower A1 content than any one reported before have been developed. Moreover, by using highly reactive and stable seeds, crystals of Ti-Beta zeolite have been produced, which have an inner core of aluminosilicate composition, covered by an outer shell of Titanosilicate which accounts for about 98 % of the mass. These synthesis methods have lead to samples which present an improved catalytic behaviour for reactions such as olefin oxidation and phenol hydroxylation using H202 as oxidant.

403

References 1. M. Taramasso, G. Perego, B. Notari, U.S. 4 410 501, (1983). 2. U. Romano, A. Esposito, F. Maspero, C. Neri, M.G. Clerici, Stud. Surf. Sci. Catal., 55, 33, (1990). 3. T. Tatsumi, M. Nakamura, S. Negishi, H. Tominaga, J. Chem. Soc., Chem. Commun., 476, (1990). 4. Huybrechts, D.R.C., L. De Bruycker, P.A. Jacobs, Nature, 345,240, (1990). 5. R.A. Sheldon, J. Mol. Catal., 7, 107, (1980). 6. T. Tatsumi, M. Nakamura, S. Negishi, H. Tominaga, J. Chem. Soc., Chem. Commun., 476, (1990). 7. M.A. Camblor, A. Corma, A. Martinez, J. P6rez-Pariente, J. Chem. Soc., Chem. Commun., 589, (1992). 8. A. Corma, M.A. Camblor, P. Esteve, A. Martinez, J. P6rez-Pariente, J. Catal., 145, 151, (1994). 9. A. Corma, P. Esteve, A. Martinez, S. Valencia, J. Catal., 152, 18, (1995). 10. M.A. Camblor, A. Corma, J. P6rez-Pariente, Sp. Pat. 2,037,596, (1993). 11. M.A. Camblor, A. Corma, A. Martinez, J. P6rez-Pariente, S. Valencia, Stud. Surf. Sci. Catal., 82, 531, (1994). 12. M.A. Camblor, A. Corma, M. Costantini, L. Gilbert, J. P6rez-Pariente, S. Valencia, FR Pat. 95/01824, (to Rh6ne-Poulenc), (17/02/95). 13. M.A. Camblor, A. Corma, M. Costantini, L. Gilbert, J. P6rez-Pariente, S. Valencia, FR Pat 95/01823, (to Rh6ne-Poulenc), (17/02/95). 14. M.A. Camblor, A. Corma, J. P6rez-Pariente, Zeolites, 13, 82, (1993). 15. T. Blasco, M.A. Camblor, A. Corma, J. P6rez-Pariente, J. Am. Chem. Soc., 115, 11806, (1993). 16. M.A. Camblor, A. Mifsud, J. P6rez-Pariente, Zeolites, 11,792, (1991).

404

PEPTIDE SYNTHESIS BY SAPPHO TECHNOLOGY

JEAN-MARIE BERNARD, KAMEL BOUZID, JEAN-PIERRE CASATI, MARIE GALVEZ, CHRISTIAN GERVAIS, PIERRE MEILLAND, VIRGINIE PEVERE, MARIE-FRANCE VANDEWALLE, JEAN-PAUL BADEY AND JEAN-MARIE ENDERLIN

Rh6ne Poulenc Industrialisation, Centre de Recherche, d'Ing6nierie et de Technologie, 85 Avenue des fr~res Perret, B.P. 62, 69192 Saint Fons Cedex, France.

INTRODUCTION Peptides are molecules very active at low concentration. They are used in pharmaceutical, agrochemical and nutritional areas. The synthesis of these compounds (ref. 1) is very dependant on the sequence and the quantity required. To synthesize short peptides, at high volume and low cost, chemist prefer to use the N carboxyanhydrides (NCA) of aminoacids, protected if necessary on the side chain reactive functional groups. This is the case for intermediates of angiotensing converting enzyme peptides such as L-alanyl-L-proline and Ne-(TFA)-L-lysyl-Lproline. This method is however very difficult to use repetitively for long peptides, because it gives quantities of oligomers (more than 2 - 3 %) at each coupling step and the purity of the final peptides is poor. This is the reason why peptide chemists, to decrease the problems of purification prefer for long peptides to use protecting groups (tert-butyloxycarbonyl (t-Boc), benzyloxycarbonyl (Z), fluorenylmethyloxycarbonyl (FMOC) .... ) and classical reagents such as T.B.T.U. (O-1H-benzotriazol-l-yl)-l,l,3,3-tetramethyl uronium tetrafluoroborate), B.O.P.(benzotriazol-l-yl-oxy-tris (dimethylamino) phosphonium hexafluorophosphate and so on in polar solvents such as N,N-dimethylformamide or N-methylpyrrolidone. But this solvents are not compatible with the acidic deprotection reagents such as trifluoroacetic acid and 405

necessite many solvent exchange or precipitation at each coupling or deprotection step. That is the reasons these methods are long and expensive. Recently, we have searched new, for economical ways to synthesize long peptides. We have now developed new methodology to produce peptides at low cost, continuously and automatically without precipitation and isolation at different steps of the synthesis. The name of this process is SAPPHO (in french : Synth~se Automatis6e de Peptides en Phase HOmog~ne). Different peptides or peptide fragments have been synthesize by this technology. We present here the synthesis of fragments of the salmon calcitonin.

THE SAPPHO TECHNOLOGY SAPPHO is a powerful new innovative technology for the automated large scale synthesis of peptides in solution phase, which can lead to significant cost saving of up to 50 % compared to the traditional technologies (in homogeneous phase or Merrifield phase). SAPPHO technology which combines a modular approach to synthesis with an innovative solubilisation system guarantees 9 the simplicity and reproducibility of the synthetic process the direct synthesis of the final product in high purity on line quality control at each stage. The patented solubilisation system enables the same solvent to be maintained throughout each coupling cycle, thus avoiding any need for interim purification, precipitation or difficult solvent exchange. The solubilisation system comprises 9 An organic solvent, non miscible with water (ref. 2). - A phenolic additive to enhance solubility (ref. 2), which gives no side reaction with the peptides all along the coupling cycle. The additive gives hydrogen bonds with the amide groups of the peptide and decreases the interactions between the molecules of peptide.The major consequences are a better solubilisation of the peptides in the medium which permits synthesis at high molar concentration. - A lipophilic, non polymeric, aromatic carboxyl protecting group protector (ref. 3), having good solubility in the organic solvent, stable throughout the synthesis. This protecting group is easily cleaved at the end of the synthesis by classical reactions like saponification, hydrogenolysis or other technologies. The choice of this protecting group and the cleavage method is dependent of peptide sequence and functional alpha amino and side chains protecting groups. -

-

-

-

406

Each successive amino acid is coupled with the growing peptidic chain through the repetition of a simple 4 step cycle: coupling, extraction, drying, and deprotection of the alpha nitrogen protecting group. In the SAPPHO process, the N-tert-butyloxycarbonyl (N-t-BOC), the Nallyloxycarbonyl (N-ALLOC), and the N,N-diallyl can be used for the protection of the alpha amino function. The functional side chains can be protected by different classical orthogonal protecting groups. Different coupling reagents such as N-hydroxysuccinimide activated esters of amino acids, T.B.T.U. ((O-1H-benzotriazol-l-yl)-l,l,3,3-tetramethyl-uronium tetrafluoroborate), or N protected N carboxyanhydrides of amino acids (UNCA), and so on can be used in the SAPPHO technology. The excess of reagents (coupling reagents and N protected amino acids) is very low (10 to 20 %, compared with the excess used (400 to 600 %) in solid phase synthesis (Merrifield synthesis)). After the coupling step, the excess reagents are transformed into hydrophilic species eliminated with the co-products during aqueous washing. The N protecting group is cleaved by the appropriate patented method. (ref. 4). For the cleavage of the N-ALLOC protecting group, the drying step can be avoided. Quality control at each step of the cycle enables high yields of 97.5 % per cycle (more than 99.5 % for the coupling and deprotection steps) to be achieved and simplifies the final purification. The SAPPHO process also permits peptide fragment coupling. This technology has been successfully applied to the synthesis of various peptides (Leucine Enkephaline, Luteinising Hormone Releasing Hormone (L.H.R.H.), calcitonin fragments...).

CALCITONIN FRAGMENTS SYNTHESIS Salmon calcitonin is a calcium regulated hormone which inhibits the bone resorption of calcium ions. It is a polypeptide of 32 amino acids. The Salmon calcitonin is currently manufactured by solid or liquid phase synthesis. Two protected fragments of the salmon calcitonin (1 to 10 and 25 to 32) have been synthesized by the SAPPHO process. All the aminoacids have the L configuration.

407

P r o t e c t e d (1-10) salmon calcitonin

Boc~Cys~Ser(O- BzI)- Asn-- Leu--Ser (O-Bzl)--Thr (O-BzI)--Cys-- Val~ Leu--Gly--COOH !

I

S

S

I

t

P r o t e c t e d (25 - 32) s a l m o n calcitonin

HC1, Thr (O-Bzl)-Asn-Thr(OBzl)-Gly-Ser(OBzl)-Gly-Thr(O-Bzl)-Pro-O-GPC The structure of the GPC group (carboxylic protective group) is 9

O

SYNTHESIS OF P R O T E C T E D (25 - 32) SALMON CALCITONIN The procedure of Gisin (ref. 5) has been used for esterification of L proline.The cesium salt of the N-tBoc L-Proline reacts with the (3-phenoxybenzyl) 4 chloromethylbenzoate (CI-GPC), in N,N-dimethylformamide (DMF). (Fig. 1)

Cs §

o -o

CI

9

-o CI-GPC

80~ + CSCI

Fig. 1. Synthesis of the lipophilic ester of N-tBoc L-proline

Then, after deprotection of the N-tBoc group with dry HC1 gas, the protected N tBoc aminoacids are successively introduced, at room temperature, on the chlorhydrate of the lipophilic L-Proline ester, by using T.B.T.U. as coupling reagent and diisopropylethylamine (Fig. 2.). A 15 % excess of T.B.T.U and N tBoc amino acids is used for the coupling steps. 408

Thr

Asn Thr

Gly Ser

Gly Thr

Pro Boc-Pro-OGPC

Boc-Thr(OBzl) -Pro-OGPC Boc-Gly -Thr(OBzl) -Pro-OGPC Boc-Ser(OBzl)-Gly -Thr(OBzl) -Pro-OGPC Boc-Gly-Ser(OBzl)-Gly-Thr(OBzl) -Pro-OGPC Boc-Thr(OBzl)-Gly-Ser(OBzl)-Gly-Thr(OBzl) -Pro-OGPC Boc-Asn-Thr(OBzl)-Gly-Ser(OBzl)-Gly-Thr(OBzl) -Pro-OGPC Boc-Thr(OBzl)Asn-Thr(OBzl)-Gly-Ser(OBzl)-Gly-Thr(OBzl) -Pro-OGPC coupling reagent: TBTU / Room temperature//deprotection reagent: dry HC1 gas//SAPPHO process

Fig.2. Scheme of synthesis of protected (25-32) Salmon Calcitonin

The molar concentration of the peptide in the solvent ranges between 0.25 M/L for the dipeptide Boc Thr (O-Bzl) - Pro - GPC and 0.1 M/L for the octapeptide Boc Thr (O-Bzl) - A s n - T h r (O Bzl) - Gly - Ser (O-Bzl) - Gly -Thr (O-Bzl) - Pro OGPC. The co-products (hydroxybenzotriazol, N,N,N',N'- tetramethylurea and the excess of N tBoc amino acids) are eliminated by aqueous extractions. The N tBoc group is cleaved with dry HC1 gas which gives volatile co-products (CO2, isobutene and tButyl chloride) eliminated by distillation. Yields of each coupling and deprotection steps are always more than 99.5 %. The reaction times of coupling and deprotection are always, respectively, less than 6 hours and less than 2 hours. 100 g of the protected (25 - 32) salmon calcitonin is isolated by precipitation with diisopropyl ether (71.5 %). The high performance liquid chromatography (HPLC) profile of the crude product is given in Figure 3.

409

--\

\ ql I . 7~. ' ;4.,06 --'-~_

.~1

:," . : : j

,

~,

....... "~"

" ~ ,~: . $ 6

I ~

ClATG ~

.'3 31X Z : ~

Fig.3. HPLC Profile of protected (25 - 32) Salmon Calcitonin

S Y N T H E S I S O F P R O T E C T E D (1 - 10) S A L M O N C A L C I T O N I N The synthesis of this fragment has been synthesized, using procedure. (Fig. 4.) Cys

Ser

Asn

Leu

Ser

Thr

Cys

Val

Leu

the

same

Gly

Boc-Gly-OGPC Boc-Leu--- GIy-OGPC Boc-Val . . . . Leu --- Gly OGPC Boc-Cys(S-Acm)-Val . . . . Leu--- GIy-OGPC Boc-Thr(OBzl)-Cys(S-Acm)-Val . . . . L e u - - - Gly-OGPC Boc-Ser(OBzl)-Thr(OBzl)-Cys(S-Acm)-Val . . . . Leu --- G ly-OGPC Boc-Leu . . . . Ser(OBzl)-Thr(OBzl)-Cys(S-Acm)-Val . . . . Leu --- GIy-OGPC Boc-Asn---Leu . . . . Ser(OBzl)-Thr(OBzl)-Cys(S-Acm)-Val . . . . Leu --- GIy-OGPC Boc-Ser(OBzl)-Asn---Leu . . . . Ser(OBzl)-Thr(OBzl)-Cys(S-Acm)-Val . . . . Leu --- Gly-OGPC Boc-Cys(S-Trt)-Ser(OBzl)-Asn---Leu . . . . Ser(OBzl)-Thr(OBzl)-Cys(S-Acm)-Val . . . . Leu--- GIy-OGPC coupling reagent: T B T U / R o o m t e m p e r a t u r e / / d e p r o t e c t i o n

reagent: dry HCI g a s / / S A P P H O

Fig. 4. Synthesis of the protected (1-10) Salmon Calcitonin 410

process

We have observed that T.B.T.U. gives a little dehydration of the side chain of asparagine. This side reaction has been confirmed by synthesis of an authentic cyanoalanyl peptide and HPLC analysis. Optimisation conditions have been fc,und to decrease the level of this side reaction : low temperature, minimisation of the quantity of diisopropylethylamine used during the coupling step and use of hydroxysuccinimide ester of asparagine as the activated aminoacid. The cyano alanyl peptide can be eliminated by preparative HPLC at the end of the synthesis. The analysis data (Mass spectra and HPLC profile) of the crude protected (1-10) salmon calcitonin are given in Figure 5 and Figure 6. FAB+ Magnet EpM:243 File T e x t : K B - 1 5 9 7 - D M F / N B A 100%

Bpi:22003712

AutoSpecEQ

24311 .......

90.

TIC:99330400

Flags:NORM

50.00 ............

lo0.o0 .........

i 2i2E 0E7 8E7

80.

5E7

70

307.08

60

3E7 1

5o~ 40

91.00

6E6

30_

364.14

0

165.08

"'-STxo....

4

t , 14q~.15 661.26 l,.lt,i~.lt/ 4160.1~ 552,-2~08 27 [

20. 10

IE7

IIIdtll, ddJI, Ldt.lll .... k k . . ~ . L i ~ , , l

26o

-~--S~o . . . . .

4;0 ....

i6o ....

x100.00

100~

,.~, ~ ' ~ , _ l J . . . . . .

6;0 ....

759.37

90,.4

J,. _.i~,m,.i,_L . . . . .

7~o . . . .

.do ....

tx o . . . .

950 . . . .

;0 ~

4E6

~ 2~.6 0E0

./=

2 2E7

2 0E7

9o_~ 8o~

1 8E7

1702 75

1 5E7

70-

1 3E7

60

1 IE7

50

_8

40

8E6

6 6E6

30

1425.62 i

lO.~

i

ii00

1200

1300

1400

.....

1579.74

1500

1600

[

]~..t "

i

~

'

.

4 4E6 2 2E6 ,L_. I

1700

...........

1800

Boc__ ?ys--Ser-- Asn- Leu--Ser-- Thr--?ys-- Val-- Leu--Gly OH SFrt

OBzl OBzl S-Acre

L.. 91 (100 %) (35 %) 243

[MH]+ = 1679 - B ~ [MNa]+ = 1701

S-Trt

1579 -Acm"-- 1508 ~

1425

Fig.5. Mass spectra of the crude protected (1-10) salmon calcitonin 411

_t

1900

~ __.~j_ 0 0 E O

2000

M/Z

"

E" Z

s:.

~, 2

. "

Fig. 6. H.P.L.C. profile of the crude protected (1-10) salmon calcitonin

The protected peptide is isolated, before disulfide bridge formation, by precipitation from its N-methylpyrrolidone solution (95.5 % yield) with water. After saponification in DMF, the carboxylic protected (1 - 10) salmon calcitonin fragment is isolated by precipitation with acidic aqueous medium. The precipitate is washed with acetonitrile to give a white powder (80.5 %). The scheme of the the disulfide bridge synthesis is shown in Figure 7. Different protecting groups (S-Trityl, S-Acetamidomethyl...) have been introduced on the cysteine side chains to optimise the reaction conditions of disulfide bridge formation. The conditions developed by Kamber and his group have been used to make the disulfide bridge (ref. 5). The best results have been obtained when we use S trityl on the L-cysteine and S-acetamidomethyl on the 7 cysteine in N,N-dimethylformamide in the presence of an excess of iodine (4 equivalents). The excess of iodine is eliminated with ascorbic acid. The final peptide is isolated by precipitation with water.

412

Boc--CysmSer(O-Bzl) - Asn~ Leu--Ser (O-Bzl)- Thr (OBzI)--Cys-- Val-- Leu~Gly--O nGPC I

I

SwTrt

SmAcm

R.T.

1

NaOH

Boc~Cys--Ser(O- Bzl)- Asn~ LeuwSer (O- Bzl)- Thr(OBzl)--Cysm Val-- Leu--GIy--COOH I

I

S--Trt

S--Acm DMF I R.T.

I2

Boc--Cys--Ser(O-BzI)-AsnmLeumSer (O-BzI)-Thr (OBzt)--Cys--ValmLeu--Gly--COOH I

I

S

S

1

I

Fig. 7. Synthesis of protected ( S - S) (1-10) salmon calcitonin

The protected (S - S) (1-10) salmon calcitonin is purified by preparative chromatography

on

silica

as

stationary

phase

with

a

solvent

mixture

(dichoromethane / methanol / acetic acid (93 / 7 / 2 v / v / v)) as eluent phase. After precipitation with water the pure product is analysed by HPLC

on

Lichrospher 100 RP 18 (125 x 4 mm) 5 micron as stationary phase and with a mobile phase (methanol / water / N,N-dimethylformamide / trifluoracetic acid 70 / 30 / 5 / 0.4 v / v / v / v). The D M F is introduced in the eluent phase to solubilise the protected ( S - S) (1-10) salmon calcitonin. The H P L C profile of the pure protected (S - S) (1-10) salmon calcitonin is shown in Figure 8.

413

,

|ll

.

.

.

.

II

,

II

".

9~ - 1 4 oeoo

~

"Jo,~

JPE

-1~.

:-: _

t i

,

.

.

.

.

.

.

|

.

7

oam ~111~

9

/

i

Fig. 8. HPLC profile of the pure protected (S - S) (1-10) salmon calcitonin

The protected (S - S) (1-10) salmon calcitonin has been successfully coupled on a (11-32) protected fragment grafted on a polystyrene resin. After final HF deprotection, the salmon calcitonin has been obtained with a better yield than stepwise synthesis on a polystyrene resin. A gain of 50 % of final salmon calcitonin is obtained using this procedure.

CONCLUSIONS We have demonstrated that the SAPPHO process is a new and powerful method of synthesizing peptides, at a low cost, with very good yield and purity. It is the first automated peptide synthesis technology which can be used to synthesize peptide or peptide fragments from 3 to 15 aminoacids.

414

References 1. a Methoden der Organischen Chemie / Houben Weyl 15 / 1 and 2 Published Georg Thieme Verlag Stuttgart, ( ) b Peptides / Gross Meienhoffer N~ c Principles of Peptide Synthesis - M. Bodansky Ed. Springer- Verlag 2, ( ) M.F. Maurice, M. Galvez, EP 0432022 (02/10/1989), (to Rh6ne-Poulenc Chimie). 3. J.M. Bernard, K. Bouzid, C. Gervais, EP 0421848, (02/10/1989), (to Rh6ne-Poulenc Chimie). 4. V. P6v~re, EP 0537089 (11/10/1991) (to Rh6ne-Poulenc Chimie), J.M. Bernard, E. Blart, J.P. Genet, M. Savignac, EP 0566459 (15/04/1992 ~ (to Rh6nePoulenc Chimie) J.M. Bernard, E. Blart, J.P. Genet, S. Lemaire-Audoire, M. Savignac, French Applications, N ~ 930423; N ~ 9304232 and N ~ 9304233, (09/04/1993), (to Rh6nePoulenc). 5. B.F. Gisin, Helv. Chim. Acta, 56,1476, (1973) 6.' B. Kamber and coll Helv. Chim. Acta. :51, 2061, (1968); Idem 53,556, (1970); Idem 54, 398, (1971)

415

A NEW AND PRACTICAL REMOVAL OF ALLYL AND ALLYLOXYCARBONYL GROUPS PROMOTED BY WATER-SOLUBLE Pd(0) CATALYSTS

SANDRINE LEMAIRE-AUDOIP~ a), MONIQUE SAVIGNAC GENET a) AND JEAN-MARIE BERNARD b)

a), JEAN-PIERRE

a) Laboratoire de Synth~se Organique associ6 au CNRS URA 1381, Ecole Nationale Sup6rieure de Chimie de Paris, 11 rue Pierre et Marie Curie, 75231 Paris Cedex 05, France b) Rh6ne-Poulenc Industrialisation, Centre de Recherche, d'Ing6nierie et de Technologie, 85 Avenue des Fr~res Perret, B.P. 62, 69192 Saint-Fons Cedex, France.

INTRODUCTION Among the usual protecting groups for amino, hydroxyl and carboxylic functions, the allyloxycarbonyl (Alloc) and the allyl moieties were largely developed during the last twenty years, since a new methodology using n-allyl palladium complexes was introduced for their cleavage (ref. 1). O II R~ z / C ~ o . ' " ' - . ~ Z=N,O

Pd(O)Ln

C02 NuH

\_ ....

I |

+ RZ |

_-

R--ZH

+

Pd(0)Ln

-,} ~ R _ Z Z " - . . ~ Pd(0)Ln Scheme 1. 416

In literature, various conditions involving different allyl scavengers such as formic acid (ref. 2), morpholine (ref. 3), tributyltinhydride (ref. 4), or potassium 2-ethylhexenoate (ref. 5) in anhydrous medium have been reported. However, these systems suffer some limitations, especially for the deprotection of secondary amines which leads to the competitive undesired reaction of N-allylation (Scheme 1, path (2)). Although recent progress has been made using silylated amines as allyl scavengers (ref. 6), a simple and unexpensive method for the cleavage of allyl carbamates derived from secondary amines would be of great interest. In addition, one of the greatest drawbacks of homogeneous metal catalysis is the separation of the reaction product from the active catalyst, which often requires costly and toxic procedures. A solution to this problem consists in anchoring the catalyst on an organic or inorganic polymer (ref. 7) insoluble in the reaction medium. Another elegant alternative consists in using water soluble ligands which once complexed to the metal make the catalyst poorly soluble in organic media. These systems combine the advantages of homogeneous and heterogeneous catalysis : easy separation of the product from the catalyst, high reactivity and high selectivity. At presem, sulfonated phosphines, e.g. TPPMS (ref. 8) and TPPTS (ref. 9), constitute the most widely used class of water soluble ligands. They found various industrial applications in the field of hydrogenation (ref. 10), hydroformylation (ref. 11), reduction of saturated and unsaturated aldehydes (ref. 12) and coupling reactions (ref. 13). In our continuing imerest in the area of palladium promoted reactions (ref. 14), we have developed a water soluble catalyst prepared in situ from Pd(OAc)2 and the water soluble ligand TPPTS. This system demonstrated high activity for various cross-coupling reactions in aqueous medium (ref. 15). We have also found that this catalyst allowed smooth and selective removal of allyl and Alloc groups in the presence of diethylamine as allyl scavenger, in homogeneous (CH3CN/H20) and biphasic (CaH7CN/H20) media (ref. 16). This palladium promoted deprotections proceed with high to quantitative yields, and the use of a two-phase system allows the reaction to occur with remarkable selectivity, in particular for the cleavage of allylcarbamates derived from secondary amines. Moreover, we developed the first efficient conditions for the chemoselective removal of allylcarbamates in the presence of substituted allyl carboxylates ; in the same way, allyloxycarbonates can be cleaved without affecting dimethylallylcarbamates in the same molecule (ref. 17).

417

RESULTS AND DISCUSSION

Deprotection of alcohols and carboxylic acids The cleavage of allyloxycarbonates was investigated in the presence of molar 2 % of Pd(0) catalyst and diethylamine as nucleophile ; our results are summarized in Table 1. Using 5 eq. of HNEt 2, primary alcohols such as (R)-citronellol were deprotected in homogeneous medium (CH3CN/H20) with excellent yield (emry 1). When HNEt 2 was replaced by formic acid, the yield and the rate of the reaction considerably decreased. The deprotection of allyloxycarbonates derived from secondary alcohols 2 was also performed under the above conditions (Pd(OAc)2 / TPPTS (1/2) tool 2 % ; HNEt2) to recover the parent molecules with good yield (entry 2). Moreover, our catalytic system allowed the selective and quantitative cleavage of the allyloxycarbonyl moiety from 6-O-Alloc-methyl-2,3-dibenzyl-ot-Lglucopyranoside 3 without affecting the other protecting groups present on the molecule (entry 3). The reaction was then carried out in a biphasic medium (C3HTCN/H20), and 1-menthol could be quantitatively deprotected within 30 minutes. Taking advantage of a two-phase system, we were able to recycle the water-soluble catalyst up to 10 times with no loss of efficiency, which is a major asset from an industrial view point (see scheme given in ref. 18). In addition, the use of HNEt2 as an unexpensive allyl scavenger is very attractive since both the allylated by-product and the excess of nucleophile are simply removed by evaporation, affording after extraction very clean crude products. The above procedure was also applied to the cleavage of allylic esters. In the presence of tool 2 % of Pd(OAc)2/TPPTS and 5 eq. of HNEt 2 in homogeneous medium, phenylacetic acid was rapidly deprotected, in quantitative yield (entry 5). In the same way, the allyl group was smoothly removed from the carboxylic acid of the base sensitive cephalosporine 6 with 93 % yield (entry 6).

418

Table 1. Deprotection of Alcohols and Carboxylic Acids using Pd(OAc)2/TPPTS catalyst F_my

!

Substrate

HNEt2! Solvent ' Cat (eq.) (mol %)

Product

[ Time l Yield i

I

o/A lloc

I

OH

CH3CN / H20

2.2

CH3CN / H20

2.2

CH3CN / H:O

~

5

94 a)

10

80

20

99

30

100

5

100

HO. ~ . . . . O7OCH3

Alloc O ~ . . . . O7OCH3

HO...... "// ......OCH2Ph OCH2Ph 3

---OH

(min.) i (%)

/

,

2.2 !

OAlloc

HO...... T ......OCH2Ph OCH2Ph

C3H7CN/ H20

2.2

CH3CN / H20

5

.-.

OH

4

D

5

_fA lj~,,,,x,~O~

5

~y

OH

a) with HCOOH as nucleophile 9t = 2.5 h" yld = 51%

Deprotection of primary and secondary amines Then, we have investigated the cleavage of allylcarbamates (Table 2). The reaction was first conducted on primary amines in homogeneous medium. Under treatment with mol 2 % of Pd(0) catalyst and 2.2 eq. of nucleophile N-Allocbenzylamine _7 was quantitatively cleaved to recover the parent molecule within 10 minutes (entry 1). However, when N-allyloxycarbonyl-N-methyl benzylamine 8 was allowed to react under the same conditions, the undesired reaction of N-allylation

419

I

occurred preferentially to give a (30/70) mixture of the free amine and the N-allylated side product 9 (entry 2). This competition between the nucleophile and the flee secondary amine for the capture of the rt-allyl palladium complex intermediate is explained by the mechanism of deprotection (Scheme 2). L.

"Pd~

o

L

|

I

II

R-.z~C-.o/'--.~

~

~'~..~,

CO~

Z=N,O L\pd~L |

I

Pd(O)Ln

A

b

R--ZH + Et2N / ' ' ' ~

Et2NH

Scheme 2. In the first step, oxidative addition of the zerovalem palladium species on the allyl moiety of the protected substrate leads to a rt-allyl complex, followed by decarboxylation of the carbamate. Then, intermediate A is trapped by the nucleophile (diethylamine in the present case) resulting in the deprotected product with regeneration of the palladium (0) species (path a). Nevertheless, when RZ- has a strong nucleophilic character, and this is the case of secondary amide bases, it also acts as an allyl trapping agent to give the undesired N-allylated side-product (path b). Anticipating that the reaction may be more selective in a biphasic system, the deprotection was carried out in C3H7CN/H20 (6/1) medium, with mol 5 % of Pd(0) catalyst ; under these conditions, the free secondary amine was quantitatively recovered without any undesired N-allylated product (entry 3). The use of a twophase system thus offers an interesting alternative for the efficient removal of 420

allylcarbamates derived from secondary amines, avoiding the competitive N-allylation. It is reasonable to think that in such a biphasic medium there is almost no contact between the catalyst present in the aqueous phase and the deprotected substrate liberated in the organic layer, resulting in an enhanced selectivity toward path a. Other protected secondary amines such as (1R, 2S)-N-allyloxycarbonylephedrine 10 and N-allyloxycarbonyl-L-proline !1 reacted equally well upon treatment with 5 fold excess of HNEt 2 (entries 4 and 5).

Table 2. Deprotectionof Primary and Secondary Amines using Pd(OAc)2/TPPTS catalyst F_my

Substrate

HNEt2 Solvent Cat (mol%)[ (eq.)

Product

Time Yield (mill.)

(%)

10

100

i

1

~/~All~ N

2"2 CH3 H20 CN/

0"5

(" l ~

"NH2

7 2

~

/~I AII~ Me

2

_8

2

CH3CN / H20

100 30 %

70 %

4

OH p h / ~I ! Me

5

C3H7CN/ H20

5

CH3CN / H20

100 %

0%

100

OH

_

ph.d./Me

Me/N~Albc 10

~N~"~COOH I Alloc 1_!1

2

15

100

15

100

Me/N~H

2.2

CH3CN / H20

'"'N Z C O O H i H

421

This efficient and unexpensive methodology thus allows the removal of allyl and allyloxycarbonyl groups from various substrates and the particularly mild conditions are compatible with polyfunctionalized molecules. Moreover, both Pd(O) catalyst and N-allyl diethylamine by-product are easily separated from the free alcohols, amines and carboxylic acids which are recovered in almost pure form.

Chemoselective removal of allylic protecting groups At this stage of our study, we have compared the rate of deprotection of several phenyl acetic allyl esters under the above homogeneous or biphasic aqueous conditions. We found that in (CH3CN/H20) medium the dimethylallyl group is cleaved at a lower rate than the cinnamyl group in the presence of 2 to 2.5 mol % of palladium (0). Under the same conditions, the allyl moiety is instantaneously removed. By comparison, in a biphasic system (C3HTCN/H20), the cinnamyl and the dimethylallyl groups remain imact in the presence of mol 5 % of Pd(0) water soluble catalyst, even after 3 days at room temperature ; whereas the allylic ester is still cleaved to give phenyl acetic acid in excellent yield. Based on these results we investigated the selective cleavage of an allylcarbamate in the presence of a dimethylallyl carboxylate in the same molecule (ref. 19).

0

O...~~/R1

Pd(OAc)2fI~PTS(12) mol 1% CH3CN/H20 HNEt2 5eq.

N 0~0~..,

mol 3 to 5 %

N I H

~

Pd(OAc)2flTPTS (1:2)

O~O..~~RI

o ..ou

CH3CN/H20 HNEt2 5 eq. I H

422

As shown in table 3, the allyloxycarbamate of isonipecotic acid 12 was selectively and quantitatively cleaved under homogeneous conditions, in the presence of 1% of Pd(0), without affecting the dimethylallyl carboxylate (entry 1). The resulting monodeprotected product 13 was then deprotected using a higher amount of catalyst (mol 5 %). The same scheme of selective deprotections was achieved on a base sensitive cephalosporin 14 (entry 2) ; with 2.5 % of water soluble catalyst the Alloc moiety was selectively removed to give the dimethylallyl carboxylate 15 within 30 minutes, and then the carboxylic acid was quantitatively recovered using 5 % of Pd(0).

423

Table 3. Selective Cleavage of Allyloxycarbamates in the presence of Substituted Allyl Carboxylates Entry

Substrate

Product

Time (min.)

Yield c) (%)

Product

Time (min.)

Yield d~

10

100

60

100

45

100

40

86

(%)

H 1

20

O....~O,,,,,N,~

96

H

H 13

12

.~.N/O...~NH

-S

O O ~ N ~/~CH3

H2N~ S " - . ] O ~ N/N......~CH3 30

100 a)

!_5

14

O

O

-(

H

~__

H

o

O H

H

4

-=-\Ph

18

99

17

16

o

30

19

a) Pd(0) 9mol 2.5 % 9b) solvent 9butyronitrile-water " c) crude product 9d) isolated yield

We also applied these conditions of selective deprotection on amino acids ; it was possible to cleave the N-allyloxycarbamate of the L-proline derivative 16 without affecting the carboxylic acid protected by the dimethylallyl moiety (entry 3). Nevertheless, when the dimethylallyl group was replaced by the cinnamyl group, the use of 1% of Pd(0) in homogeneous medium led to complete removal of the allyloxycarbonyl group with a certain amount of the deprotected carboxylic acid (ref. 20). In order to overcome this difficulty it was necessary to operate under biphasic conditions, in the presence of 1 % of catalyst, giving the expected cinnamyl-L-prolinate 17 in quantitative yield (entry 4). Then, the selective cleavage of aUyloxycarbonates in the presence of dimethylallylcarbamates was performed with high efficiency (Table 4). 0

H 0 0. ~. . ~ . . . ~ . . I

Pd(OAc)2/TPPTS(1:2) mol 5 %

)

C3H7CN/H20 HNEt2 5 eq.

N

Pd(OAc)2/TPPTS (1:2) mol 5 %

CH3CN/H20 HNEt2 5 eq.

;

OHI (C)

N

OH I N

A first attempt to cleave selectively the allyloxycarbonate from (1R, 2S)-(-)ephedrine doubly protected 20, under homogeneous conditions, using 1% of Pd(0), led to total deprotection of the amino function together with partial removal of the dimethylallyloxycarbonyl group. Taking advantage of a biphasic medium, the reaction was thus conducted in a butyronitrile-water system with 5 % of Pd(0) ; under these conditions, the allyloxycarbonyl group was smoothly removed from oxygen without affecting the dimethylallylcarbamate. In a second step, the amine could be deprotected using an homogeneous medium, with acetonitrile as cosolvent, to recover the parent molecule within 15 minutes, with 100 % yield

425

(entry 1). An other example on 1-(2-O-Allyloxycarbonylethyl)-N-dimethylallyloxycarbonyl piperazine 22 gave similar results, and thus confirmed the selective cleavage of an allyloxycarbonate in the presence of N-dimethylallylcarbamate in the same molecule (entry 2).

426

Table 4. Selective Cleavage of Allyloxycarbonates in the presence of Dimethylallyl carbamates Substate

Entry

Time (min.)

Product

Yield c~ (%)

Product

Time (rain.)

Yield dr (%)

15

100

O OH OH

ph.~,,,.,~CH3

ph,,."k-,,,,/CH3

p h ~ . . , . ~ CH3

CIt3/N

o ~ O ' ~

~

20

100 CH3

O

llq

21 20

U ) r o.~o~ O

J

H

I

iS

r OH

OH 23

22 a)

Pd(0) 9mol 5 % / butyronitrile-water / crude product 9b) Pd(0) 9mol 5 % / acetonitrile-water / isolated yield

\H

CONCLUSION In summary, we have developed a smooth and efficient methodology for the cleavage

of allyloxycarbonates,

allylcarbamates

and

allyl

carboxylates

using

Pd(OAc) 2 / TPPTS catalyst in aqueous medium. The free parent molecules are easily

isolated

from the reaction

mixture

by

simple

aqueous

work-up

and

extraction; they are generally pure enough to be used in another step without any further purification. Moreover, the use of a two-phase system (C3HvCN/H20) affords a valuable solution for the deprotection of secondary amines which are obtained without any N-allylated side product. In addition, in a biphasic medium the recycling of the active catalyst is particularly attractive from an industrial view point. Finally, chemoselective deprotection of bifunctional molecules containing differently substituted allylic groups was performed with high efficiency. Various applications of this technique are under investigation in our laboratory, especially in the field of peptide synthesis.

References

1.

J.W. Greene, P.G.M. Wut in ,~ Protective group in organic synthesis >,, Ed. John Wiley, New-York (1991). 2. a) I. Minami, Y. Ohashi, I. Shimizu, J. Tsuji, Tetrahedron Lett., 26, 2449, (1985). b) Y. Hayakawa, S. Wakabayashi, H. Kato, R. Noyori, J. Am. Chem. Soc., 11.2, 1691, (1990). 3. a) H. Kunz, H. Waldmann, Angew. Chem. Int., Ed. Engl., 23, 71, (1984). b) H. Kunz, H. Waldmann, U. Klinkhammer, Helv. Chim. Acta, 71, 1868, (1988). c) H. Kunz, C. Unverzagt, Angew. Chem. Int. Ed. Engl., 23,436 (1984). 4. a) F. Guib6, Y. Saint M'Leux, Tetrahedron Lett., 22, 3591, (1981). b) F. Guib6, O. Dangles, G. Balavoine, A. Loffet, Tetrahedron Lett., 30, 2641, (1989). c) O. Dangles, F. Guib6, G. Balavoine, S. Lavielle, A. Marquet, J. Org. Chem., 52, 4984, (1987). d) P. Boullanger, G. Descotes, Tetrahedron Lett., 27, 2599, (1986). 5. P.D. Jeffrey, S.W. McCombie, J. Org. Chem., 47, 587, (1982). 6. A. Mermouk, F. Guib~, A. Loffet, Tetrahedron Lett., 33,477, (1992). 7. P.W. Wang, M.A. Fox, J. Org. Chem., 59, 5358, (1994). 8. S. Ahrland, J. Chatt, N.R. Davies, A.A. William, J. Chem. Soc., 276, (1958). TPPMS = Triphenylphosphinomonosulfonate sodium salt. 9. E.G. Kuntz, US Patent 4 248 802 (1981), (to Rh6ne-Poulenc Industries) ; D. Sinou, Bull. Soc. Chim. Fr. (3), 480, (1987). TPPTS = Triphenylphosphinotrisulfonate sodium salt. 10. a) Y. Dror, J. Manassen, J. Mol. Catal., 2, 219-222, (1977). b) A.F. Borwski, D.J. Cole-Hamilton, G. Wilkinson, Nouv. J. Chim., 2, 137, (1978). c) F. Joo, Z. Toth, M.T. Beck, Inorg. Chim. Acta, 25, L61, (1977). d) C. Larpent, R. Dabard, H. Patin, Tetrahedron Lett., 28, 2507, (1987). C. Larpent, H. Patin, J. Mol. Cat., 61, 65, (1990). 11. a) W.A. Hermann, J. Kellner, H. Riepl, J. Organomet. Chem., 3_8_9_,103, (1990). b) P. Escoffre, A. Thorez, P. Kalck, J. Chem. Soc., Chem. Commun., 146, (1987).

428

12. a) E. Fache, F. Senocq, C. Santini, J.M. Basset, J. Chem. Soc. Chem. Commun., 1776, (1990). b) A. B6nyei, F. Joo, J. Mol. Catal., 58, 151, (1990). c) J.M. Grosselin, C. Mercier, G. Allmang, F. Grass, Organometallics, 10, 2126, (1991). 13. N.A. Bumagin, P.G. More, L.P. Beletskaya, J. Organomet. Chem., 371,397, (1989). 14. a) D. Ferroud, J.M. Gaudin, J.P. Gen6t, Tetrahedron Lea., 27, 845, (1986). b) J.P. Gen6t, J.M. Gaudin, Tetrahedron, 43, 5315, (1987). c) J.P. Gen6t, S. Jug6, S. Achi, S. Mallart, J. Ruiz-Mont6s, G. Levif, Tetrahedron, 44, 5263, (1988). d) J.P. Gen6t, S. Grisoni, Tetrahedron Lea., 29, 4543, (1988). e) J.P. Gen6t, J. Uziel, S. Jug6, Tetrahedron Lett., 29, 4559, (1988). f) J.P. Gen6t, M. Port, A.M. Touzin, S. Roland, S. Thorimbert, S. Tanier, Tetrahedron Lett., 33, 77, (1992). g) J.P. Gen6t, N. Kardos, Tetrahedron : Asymmetry, 5, 1525, (1994). 15. a) J.P. Gen6t, E. Blart, M. Savignac, Synlett, 715, (1992). b) E. Blart, J.P. Gen6t, M. Sail, M. Savignac, D. Sinou, Tetrahedron, 50, 505, (1994). 16. a) J.P. Gen6t, E. Blart, M. Savignac, J.M. Paris, Tetrahedron Lett., 34, 4189, (1993). b) J.P. Gen6t, E. Blart, M. Savignac, S. Lemeune, S. Lemaire-Audoire, J.M. Paris, J.M. Bernard, Tetrahedron, 50, 497, (1994). S. Lemaire-Audoire, M. Savignac, E. Blart, G. Pourcelot, J.P. Gen6t, J.M. Bernard, 17. Tetrahedron Lett., 35, 8783, (1994). The catalytic system can be recycled up to 10 times as presented in the following scheme 18. (the procedure is applied on N-methyl N-allyloxycarbonyl benzylamine), without loss of efficiency. After completion of the reaction, the first schlenck tube containing the free amine in the organic layer and the catalyst in the aqueous layer is linked, by a siphon tube, to another schlenck tube containing the protected amine dissolved in butyronitrile. The aqueous layer with the active catalyst is transferred under argon pressure into the second tube, over the fresh solution of N-allyloxycarbonyl-N-methyl benzylamine. aqueous layer catalytic system Argon --I~

II.,

I,II

Argon

~TC,.. Ph~N,H I

Me ~'~~mEt2 + HNEt2 C3HTCN V = 3ml Catalyst + H20 V -- 0.5 mi

T

)(

I ///'// ~__~

)(

429

Ph~N ~ ' ~ "] ~" l~le ~ ~HNEt,. (2.2 eq) ~C3HTCN V = 3ml

19.

20.

The doubly protected substrates are readily prepared by addition of allyl chloroformate on the amino function, followed by esterification of the carboxylic acid with the appropriate substituted allylic bromide in the presence of DBU. When the substrate was treated with mol. 1 % Pd(O) under homogeneous conditions, the cinnamyl carboxylate was partially cleaved, and the reaction led to a mixture of the selectively N-deprotected prolinate with the fully deprotected amino acid in a (65 : 35) ratio.

430

SAFETY OF C H L O R I N A T I O N REACTIONS

JEAN-LOUIS GUSTIN AND ALEXANDRE FINES Rh6ne-Poulenc, Centre de Recherche, d'Ing6nierie et de Technologie, 24 Avenue Jean Jaur~s - 69151 D6cines - France

ABSTRACT Chlorination reactions are part of various processes in the chemical industry, to manufacture heavy chemicals, specialty chemicals, pesticides and pharmaceuticals, in inorganic and organic chemistry. They are a valuable tool in organic synthesis. The hazard of processing chlorine involves : Gas phase explosion ; Runaway reaction or thermal explosion in the condensed phase. Gas phase explosion hazard with chlorine as an oxidizer is present in gas phase chlorination processes as well as in chlorinations in the condensed phase. Gas phase chlorination processes are mostly continuous processes operating in the flammable area. Gas phase explosion hazard is related to burner malfunctions. Where chlorination is made by chlorine injection in the liquid phase, gas phase explosion hazard is related to chlorine evolution in the vapour phase, giving a flammable mixture with the solvent or reaction mixture vapour. Here hazard assessment is achieved by comparing the gas phase composition with the flammable area of the gaseous mixtures. Auto-ignition is also considered because the autoignition temperature of gaseous mixtures containing chlorine is close to the ambient temperature. The relevant flammability data is obtained in a specially designed 20 litre sphere. The main features of this explosion vessel include : Hastelloy C 276 walls, central ignition with spark, hot wire or pyrotechnic ignition source, 200 bar pressure resistance, ambient to 300~ initial temperature, easily opened for frequent cleaning. This apparatus allows precise determination of the flammability limits, autoignition temperature, explosion overpressure, rate of pressure rise and flame -

-

speed. A review of flammability data in chlorine is given. 431

Runaway reaction hazard in chlorination reactions is related to a series of dangerous process situations or process deviations such as : Delay in reaction initiation -

-

Reaction mixture instability

- Production of unstable species like chloramines, nitrogen trichloride, chloro nitroso compounds. -

Demixion or segregation of unstable species in case of chlorination made in

aqueous solution, because the chlorinated compounds are less soluble in water than the initial reactant. A full review in runaway reaction hazard in chlorination reactions is given with examples from the literature and from the laboratory.

INTRODUCTION Quite similar to oxygen, chlorine is used as an oxidizer in a wide range of chemical processes where it is reacted with organic and inorganic compounds to produce chlorinated products or intermediates. A wide range of useful products are obtained such as bleach, metallic chlorides, reactive monomers to manufacture plastics, heat exchange fluids, chlorinated solvents and intermediates in organic synthesis to produce specialty chemicals, pesticides and pharmaceuticals. Chlorine is involved in a wide range of process situations including gas phase reactions in a burner or on a catalyst, solid/gas reactions in a fluid bed, gas/liquid reactions in a packed column, gas/liquid reactions by injecting chlorine in a liquid phase in a semi-batch process or in a continuous process. The reaction of chlorine takes place without catalyst, in the presence of a catalyst or in photochemical reactions. Compared to oxygen, chlorine is a more reactive gas because it is processed as a pure gas whereas oxygen is mostly reacted using air. More problems would occur with oxygen if the use of pure oxygen was widespread in the chemical industry. Compared to pure oxygen, chlorine is even more reactive. The self-ignition temperature of gaseous mixtures of organic vapours with chlorine is much lower than that of their mixtures with oxygen. Natural light can split the chlorine molecule to produce reactive chlorine radicals. Many reactions of chlorine take place near the ambient temperature. The combustion of iron in chlorine can be initiated at temperatures slighly above 100~ Chlorine is toxic to man and animals. Many chlorinated compounds are also toxic.

432

For all the above reasons, the chemical processes where chlorine is involved are submitted to careful safety studies where the specific chemical properties of chlorine are considered

T H E R M A L E X P L O S I O N HAZARD IN THE CONDENSED PHASE Chlorine is a strong oxidizer. Mixtures of chlorine and organic fuels may have a high energy content and are unstable. The thermal instability of condensed phases containing chlorine can appear in various process conditions " 9 When chlorine is injected in a liquid reaction mixture, the chlorination reaction may not start immediately allowing chlorine to accumulate in the reaction mixture. The reaction may start suddenly when a large concentration of chlorine is present in the reaction mixture and give a severe runaway reaction producing a large quantity of insoluble HC1. An example of such an induction period in chlorination is mentionned in the literature for the chlorination of ketones in methanol (ref. 1). To avoid this type of incident, the reaction onset should be checked before allowing a large concentration of unreacted chlorine to be

-

dissolved in the liquid phase. When chlorine is reacted with an organic fuel in a liquid reaction mixture, highly unstable substitution products may be obtained. This process situation is dangerous in two cases : if a high concentration of unstable chlorination product is obtained in the condensed phase if a chlorinated liquid phase separates from the bulk liquid phase segregation".

"by

The latter situation is frequent in the chlorination of aqueous solutions of organic reactants because the chlorinated products are less soluble in water than the initial reactants. Examples of this dangerous process situation are the synthesis of alcohol hypochlorites by injecting chlorine in an alcaline aqueous solution of alcohol. Traugott Sandmeyer described the synthesis of methyl and ethyl hypochlorites (refs. 2, 3) and suffered severe injuries. Roland Fort and Leon Denivelle (ref. 4) described the synthesis and properties of a series of other alcohol hypochlorites obtained following Sandmeyer's Method. The oxidation of organic compounds containing nitrogen in their formula (amines, amides, cyanides) using chlorine, gives unstable chloramines. The very unstable nitrogen trichloride is finally obtained. NC13 is only slightly soluble in 433

water and can separate from aqueous solutions giving a very sensitive dense oil. Liquid NC13 can detonate. NC13 can be obtained by chlorination of aqueous solutions containing ammonium ions. NC13 compound was first obtained by Pierre Louis Dulong (1785 1838) by chlorination of ammonium chloride solutions. Dulong was seriously injured by several explosions of liquid NC13. (ref. 5) At least one accident is known where liquid NC13 could separate in a wastewater treatment where bleach was used to oxidize cyanide ions. After an agitation failure, the actuation of a bottom valve triggered the detonation. Direct chlorination would lead to the same dangerous situation. The chlorination of organic compounds with a N - O bound will leave this chemical bound unaffected. The chlorination of oximes will give chloro oximes or chloronitroso compounds (refs 6, 7) which can demix from aqueous solutions giving an unstable dense oil. This ends in process situations similar to NC13 formation and demixion. 9 Accumulation of unstable chlorinated compounds in the bulk liquid phase. The accumulation of unstable chlorinated products in the bulk liquid phase is most likely when a solvent is used where this product is soluble. The most common example is nevertheless the accumulation of NC13 produced by electrolysis of KC1 or NaC1 salt containing ammonium ions, in a NC13 removal process using extraction in carbon tetrachloride. If NC13 is not continuously thermally decomposed, high NC13 concentrations in the CC14 solution are obtained with a potential runaway decomposition hazard. Such an incident is known in the literature (ref. 8). Note that the decomposition kinetics of NC13 in CC14 solutions is strongly influenced by the wall material. Recommendations When chlorine is reacted with organic reactants, specially if nitogen comaining compounds or ammonium ions are present, the possible formation of unstable chlorinated compounds should be considered. Any segregation of a separate phase from the bulk liquid is potentially dangerous and should be investigated carefully. The demixion of an unstable liquid phase may induce a high vapour pressure of the unstable product in the gas phase because the gas phase is in equilibrium with the separated unstable liquid. This problem should be considered. If no segregation occurs, the process situation is safer, however it is necessary to check for low concentration of unstable chlorinated compounds (NC13, alcohol hypochlorites, others...) in the bulk liquid phase. 434

GAS PHASE EXPLOSION HAZARD IN CI-~O INATION REACTIONS Gas phase explosion hazard is present when chlorine is mixed with a fuel in the gas phase. The fuel may be hydrogen, a solvent or organic vapour, ammonia, etc. When chlorine is reacted with a fuel in a burner, as in the manufacture of HC1 from Hydrogen and chlorine or in the manufacture of chlorinated solvents from hydrocarbons and chlorine, most incidents occur when the burner is set on-stream, either by lighting the burner with a pilot flame or by preheating the gas and the burner. Of course the gas mixture is in the flammable range and explosions occur due to maloperation. When chlorine is reacted with a fuel on a catalyst bed, maloperation will result in catalyst burn- out and/or gas phase explosion before or after the catalyst. Here the determination of the fuel gas flammable limits in chlorine are of interest if the feed gas is not in the flammable range in normal process conditions. When chlorine is injected or bubbled in a liquid phase containing a reactant and/or a solvent, chlorine evolution in the gas phase may produce a flammable mixture with the reactant, product, solvent or reaction mixture vapour. Here inertizing is difficult as in other oxidation processes because the oxidizer is bubbled through the liquid reaction mixture. As far as possible, it is recommended to keep the gas phase composition outside the flammable range. Various methods are used : 1) Lowering the fuel vapour pressure below the Lower Flammability Limit in chlorine by lowering the process temperature. 2) Raising the fuel vapour pressure above the Upper Flammability Limit in chlorine by raising the process temperature. 3) Inertizing the gas phase by flushing the reactor gas phase with an inert gas such as Nitrogen, CO2, HC1. To keep the reactor gas phase below the L.F.L. in chlorine (method 1) is the safer method where only proper temperature control is necessary. To keep the reactor gas phase above the U.F.L. in chlorine (method 2) may not be quite safe. On start-up the temperature must be set to the process normal value ensuring enough fuel vapour pressure before chlorine injection. If a condensor is used where the fuel vapour pressure is depleted, the gas flow composition may enter the flammable range. Glass condensors are better not used or protected from light. Inertizing (method 3) is a difficult technique when the chlorine flow evolving from the liquid reaction mixture may change. 435

If chlorine does not evolve in the gas phase in normal process conditions, an inert gas flush in the reactor gas phase is recommended (see below). If a chlorine flow evolves from the liquid reaction mixture unreacted, enough inert gas flush must be provided in the reactor gas phase to lower the chlorine concentration below the minimum oxidizer concentration (MOC) of the fuel flammable range. If HC1 is released in the gas phase, this gas contributes to the reactor gas phase blanketing. However one should take into account rapid changes in the process conditions, if the wanted chlorination reaction stops due to catalyst depletion or reactant consumption. More unreacted chlorine can be released in the gas phase, the HC1 production can disappear. Therefore monitoring of the gas phase chlorine concentration using a chlorine analyser is recommended.

SELF-IGNITION, DEFLAGRATION AND DETONATION IN THE GAS PHASE Self-ignitions of gaseous mixtures containing chlorine and a fuel, near the ambient temperature, are known. Self-ignitions can turn into severe deflagrations or detonations. Self-ignition occurs in mixtures with a composition both in the flammable range and outside the previously determined flammable range. This phenomenon can be explained as follows : - The self-ignition temperature of gaseous mixtures is not a clear-cut limit. It is best represented by an induction period versus temperature relation. Self-ignition will be observed at lower temperature if a longer induction period is allowed. - Near the self-ignition temperature, the flammable area is enlarged to a wide range of equivalence ratios. - When long induction periods are necessary, weak ignition sources can initiate the explosion, such as light, wall effects, tar deposits, catalyst deposits on the wall, NCI3 decomposition flame (refs. 5, 9). A combination of these influences may explain the above mentioned self-ignition phenomenon outside the flammable range. As an example, the self-ignition of gas phase mixtures of dioxane and chlorine was thoroughly investigated by F.

Battin-Leclerc (refs. 10, 11). Dioxane

is

sometimes mentioned as a solvent for chlorination processes (ref. 12) whereas selfignitions of dioxane + chlorine mixtures is easily obtained near the ambient temperature.

436

The flammable limits of dioxane + chlorine mixtures were determined in a 4.6 litre explosion vessel together with the explosion overpressures and maximum rates of pressure rise (ref. 12). The explosion overpressures obtained are of the same order of magnitude of that of explosion of gaseous fuel in air but half of the expected thermodynamic explosion overpressures in chlorine. The gas phase detonation of gaseous mixtures of dioxane and chlorine was successfully investigated in shock tubes by A. Elaissi (refs. 13, 14). This mixture was shown to be very sensitive to detonation compared to mixtures of fuel in air or oxygen. The full investigation of this example, chosen for convenience, shows that mixtures of organic fuels with chlorine can exhibit self-ignition followed by deflagration and detonation thus explaining violent explosions observed in the past.

EXPERIMENTAL SET-UP TO STUDY EXPLOSION LIMITS, EXPLOSION CHARACTERISTICS AND SELF-IGNITION OF GASEOUS MIXTURES A new explosion vessel, a 20 litre sphere, was built to investigate gas phase explosions with special attention for experiments using chlorine as an oxidizer (Fig. 1).

Fig. 1. 20 litres explosion vessel

437

This new facility allows the measuremem of : -

The flammability limits of gaseous mixtures using various ignition sources : single spark, fusing wire, chemical ignitors.

-

The explosion characteristics i.e. explosion overpressure and maximum rate of pressure rise.

- The laminar burning velocity deduced from the pressure-time history of the explosion. (ref. 15) The pressure is recorded at a rate of 20.000 Points/s. - The self-ignition temperature and induction period of gaseous mixtures, down to a few minutes. Sampling is possible to check for gas phase reaction. - Flash points in chlorine The main features of this explosion vessel are : - Hastelloy C276 walls to lower wall effects i.e. to prevent the reaction of chlorine with fuel before ignition, catalysed by stainless steel. - The vessel is made of two half-spheres connected through a flange assembly kept tight by clamps. The upper half-sphere is fixed, the lower half-sphere is movable, using a pneumatic jack, to allow quick opening of the vessel for frequent cleaning. Combustion in chlorine produces soot deposits on the walls, which may promote or prevent ignition of subsequent mixtures. Cleaning after each positive test is necessary to obtain reliable flammable limits in chlorine. The vessel design pressure is 200 bar, thus allowing initial pressure of 10 to 20 bar according to the expected explosion pressure. The vessel temperature can be set between ambient temperature and 300~ Mixing is ensured before ignition using a propeller mixer. Central ignition is made using spark, hot wire or a pyrotechnic ignition source.

REVIEW

OF

FLAMMABILITY

DATA

OF

GASEOUS

MIXTURES

CONTAINING C H L O R I N E Flammability

limits

A review of flammability limits of gaseous mixtures containing chlorine was first given by Mal'tseva, Roslovskii and Frolov (refs. 16, 17). The experimental set-up used to obtain these data was a double-wall vertical glass cylinder, 80 mm in diameter and 120 mm high. The experiment initial temperature was set by thermostating the vessel. The fuel was introduced after evacuation, and allowed to 438

vaporize. Then chlorine was admitted in the explosion vessel. Mixing was only by molecular diffusion (no stirring). A 10 min waiting time was observed before ignition by a spark. Our comment on this method is that the wall material is correct, mixing is poor or not effective and the waiting time before ignition is too long and may have allowed the mixture to react before ignition. The results are summarized in Table 1 for the reader convenience.

439

Table 1. Flammability limits of gaseous fuels in chlorine. Data of Mal'tseva (refs. 16, 17) Fuel

Temperature (~

LFL % vol

UFL % vol

Hydrocarbons |

i

CH4

20- 22

5,6

63,0

C2 H6

i20 - 22

,4,95

, 55,4

I

C3 H8

'20 - 22

!4,30

50,0

C4 H 10

20 - 22

3,31

49,5

i

C5 H12

i

|

120 - 22

2,42

43

i

Alcohols CH3 OH

70

13,8

73,5

C2 H5 OH

83

5,06

64,1

C3 H7 OH

102

3,03

51,5

C4 H9 OH

120

2,53

48,8

C5 Hll OH

143

il,98

37,6

i 105

' 27,62

82,0

122

15,83

56,0

I

Carboxylic acids H COO H |

CH3 COO H ,

,

i

,i

C2 H5 COO H

145

'9,33

50,8

C3 H7 COO H

170

7,81

49,8

C4 H9 COO H

190

5,84

48,8

CH3 C1

20

10,2

56,0

116,7 i

52,9

Chloro-alkanes CH2 C12

'50

CH C13

not combustible

C2 H5 C1

20

|

8,98

49,2 |

1-2 C2 H4 C12 !

100 J

i C2 H3 C13

] !

16,4

136,8

,

not combustible

C3 H7 C1

60

6,88

41,8

C3 H6 C12

100

9,95

35,0

C3 H5 C13

not combustible

C4 H9 C1

100

5,42

44,5

NB compositions are in percents by volume.

440

Dokter (ref. 18) and Medard (ref. 19) published some more data collected in Table 2, together with interesting discussions. Table 2. Flammability limits of gaseous fuels in chlorine. Data published by Dokter (ref. 18) and Medard (ref. 19). Fuel

Temperature

(~ H2

cn4 00 .. 200

CH4 CH4 CH3 C1 C~. H6 H2 CH4

20

cn4 cn4

200

UFL % vol 89 63 66

LFL % vol 3,5 5,51 3,6 0,6 10,2 4,95

C2 H6 C3 H8 CH3 C1 CH2 C12

(18) (18) (18) (18) (18) (18) (19) (19) (19) (19) (19) (19) (19) (19)

63 58,8 86 70

5,6 3,6 0,6 6,1 5 10 16

100

Ref

58 " 40 63 53

Further flammability data obtained either using our 4.6 litre stainless steel cylinder (ref. 12) (C) or our 20 litre Hastelloy C sphere,(S) are given in Table 3. Table 3. Flammability limits of gaseous fuels in chlorine Fuel

Temperature

CH3 C1 C3 H8 C3 H6 C12 MTBE 1-4 Dioxane CH3 c o o H Acetone Chlorobenzene Toluene 2 chloro toluene a chloro toluene c~ dichloro toluene c~ trichloro toluene 2 Fluoro toluene

25 70* 200** 60 80 120 60 130 160 150 160 160 160 100

(~

* Po = 1,7 Atmosphere abs. ** Po = 1,3 Atmosphere abs.

LFL % vol 7 2 4,5 2 2,5 5 4,5 7,5 3,5 5 4 6 9 4

UFL % vol 65 60 33 41 36 ...... 60 43,5 50 45

37

(C) = 4.6 litre cylinder (S) = 20 litre sphere 441

Apparatus (C) (S) (S) (C) (C) (c) (C) (S) (S) (C) (S) (S) (S) (C)"

Miscellaneous data can be found in the literature, like the flammability limits

of

benzene in chlorine (ref. 20) 9L F L = 8 % vol, UFL = 52 % vol. The experimental data is given under atmospheric initial pressure, unless otherwise specified.

Self-ignition temperatures Data on self-ignition temperature of gaseous mixtures of fuel and chlorine are given by Mal'tseva (ref. 17), Dokter (ref. 18) and others. A collection of data is given in Table 4. Table 4. Auto-ignition temperature of gaseous fuels in chlorine Fuel

AIT in chlorine Author (~

CH4

318

(17)

C2 H6

280

(17)

dimethyl ether

ambient

(17)

C 1-C3 carboxylic acids

300 - 320

(17)

C4-C7 carboxylic acids

230 - 190

(17)

C2-C4 carboxylic anhydrides

290- 215

(17)

C3-C5 ketones

325 - 205

(17)

C1-C8 alcohols

1225 - 210

(17)

C2-C7 aldehydes

,110 - 160

(17)

H2

207

(18)

CH3 C1

215

(18)

CH2 C12

, 262

(18)

C2 H6

205

(18)

C3 H6

150- 100

(18)

1,2 C3 H6 C12

180

(18)

Dioxane (0,26 ATA)

100

(10 - 11)

chloro benzene

> 165~

20 litre sphere

i

C3 H8 (1,7 ATA)

!

165 ~

20 litre sphere

1

|

I

C 3 H6 (1,7 ATA)

160~

. 20 litre sphere

442

CONCLUSION Owing to the importance of chemical reactions involving free chlorine in the chemical industry, the collection of experiences and experimental data is of great interest. This should contribute more to process safety than information on less dangerous chemicals or processes. It is surprising that only limited effort or support is devoted to collect safety data on chlorination reactions. The literature on the safety of chlorination reactions is very limited compared to the literature on oxidation reactions using oxygen. The authors hope that their contribution will promote further experimental work in this field. The new 20 litre explosion vessel, specially designed to study the flammability of gaseous mixtures containing chlorine as an oxidizer will allow the obtention of reliable data at a reasonnable cost, for a wide range of initial conditions.

References

1. 2. 3. 4. 5. 6. 7. 8. 9.

10. 11.

12.

13. 14. 15. 16. 17.

R.R. Gallucci, R. Going, J. org. chem., 46, 2532, (1981). T. Sandmeyer, Ber. XIX, 857, (1886). T. Sandmeyer, Ber. XVIII, 1767, (1885). R. Fort, L. Denivelle, C.R. Acad. Sci., 234, 1109, (1954). F. Baillou, Thesis dissertation, 27 Septembre 1990, Universit6 d'Orl6ans, France. Piloty, Steinbock, Ber. 35, 3113, (1902). W. Steinkopf, W. Mieg, J. Herold, Ber. 53, 1148 (1920). "Nitrogen Trichloride, a collection of reports and papers", The Chlorine Institute, Report n~ Ed.2, New York, (1975). F.Baillou, R. Lisbet,G. Dupr6, C. Paillard, J.L. Gustin, "Gas phase explosion of nitrogen trichloride : Application to the safety of chlorine plants and chlorination processes", 7th International Symposium on Loss prevention and Safety promotion in the Process Industries. Taormina, Italy, May 1992, Paper n ~ 106. F. Battin-Leclerc, Thesis dissertation, 15 Jan. 1991, INPL-ENSIC, Nancy, France, (1991). F. Battin-Leclerc, P.M. Marquaire, G.M. Come, F. Baronnet, J.L. Gustin, "Auto-ignition of gas phase mixtures of 1,4 Dioxane and chlorine", 7th International Symposium on Loss prevention and Safety promotion in the Process Industries, Taormina, Italy, May 1992, Paper n ~ J.L. Gustin, "Gas-phase explosions of mixtures of organic compounds with chlorine", 6th International symposium Loss prevention and Safety Promotion in the Process Industries, Oslo, Norway, June 19- 22, Paper n ~ 91, (1989). Abdelkrim Elaissi, "Propri6t6s explosives des m61anges 1,4 Dioxanne + chlore en phase vapeur", Thesis dissertation, University of Orleans, France, 14 March 1994. A.Elaissi, G. Dupr6, C. Paillard, paper presented at the 8th International Symposium Loss prevention and Safety promotion in the Process Industries, Antwerpen, (1995). D. Bradley, A. Mitcheson, Combustion and Flame, 26, 201-217, (1976). A.S. Mal'tseva, Yu. E Frolov, V.L. Sushchinskiy, The Soviet Chemical Industry, 1, 23-25, (1971). A.S. Mal'tseva, A.T. Rozlovskii, Yu. E. Frolov, Zhurnal Vses. Khim. Ob-va im. Mendeleeva, 19, 5,522-551, (1974). 443

18. T. Dokter, J. Hazardous Materials 10, 73-87, (1985). 19. L. Medard, Les explosifs occasionnels, i, pp. 172-173, Lavoisier Ed., Paris, (1987) 20. G. Calingaert, W. Burt, I.E.C., 43 (10), 1341, (1951).

444

SODIUM AMIDE IN ORGANIC SYNTHESIS

JEAN-MARIE POIRIER URA n ~ 464 du CNRS, UFR Sciences, Universit6 de Rouen et IRCOF, F-76821 Mont Saint Aignan Cedex, France.

Sodium amide, NaNH2 (mp 210~ is slightly soluble in liquid ammonia, about 1 mole per litre at -33~ NaNH2 is a powerful basic reagent and very useful in organic synthesis. This compound acts essentially as a deprotonating reagent but in some cases as a nucleophilic reagent. The acidic acetylenic proton of alkynes can be easily removed by treatment with sodium amide in anhydrous liquid ammonia and the resulting anion reacted with various electrophilic reagems (refs. 1-4). In the same manner, the anion of indole (ref. 5) is methylated leading to the 1-methylindole in high yield. Diphenyl methane is metallated by NaNH2 and alkylated in 90-95 % yields (ref. 6). The disodiosalts of 13-diketones have been prepared and alkylated (refs. 7-8) or acylated (ref. 9) (Fig. 1). These salts can be reacted with chloro- and bromoacid salts leading to dioxocarboxylic acids (ref. 10). The acid salts must be prepared beforehand because of the rapid reaction of an excess of NaNH2 with the halogen of the acid. Treated with 2 equivalents of NaNH2 in liquid ammonia, unsymmetrical 13-diketones lead to disodiosalts in which two alkylation sites are possible (refs. 1113). When R 1 = H , R 2 = Me (Fig. 1) the methylation is very selective on the a site (a " b = 89 911). Increasing the steric hindrance on the b site (R 1, R 2 - - Me) yields almost exclusively methylation on the a site (a : b = 99 : 1) except when the anion is stabilized by resonance (R 1 = H, R 2 = Ph), in this case the methylation takes place on the b site. With unsymmetrical f3-diketones the following rule of selectivity has been proposed (refs. 9, 12) : phenylacetyl > acetyl > propionyl > isobutyryl. This order is valid whatever the alkyl halide used and the authors suggest that it is also valid for acylation and carbonatation.

445

Sodium enolates of ketones have been prepared by reaction of these ketones with NaNH2. For example, the alkylation of the sodium enolate of cyclohexanone by allylbromide (Fig. 2) leads to 2-allylcyclohexanone accompanied by a little of the dialkylated product (ref. 14). Dimethyl ethynyl carbinol was obtained by reaction of the enolate of acetone (prepared by reaction of solid NaNH2 in ether) with acetylene. Although a prepared ketone enolate is used, this reaction can also be considered as an aldolisation reaction of the acetylide with acetone (ref. 15). Hauser and coll. react sodioenolates of ketones prepared in ether (ref. 16) with acid chlorides (Fig. 2). O

O

O O .-,-- / ~ J ~ ~ O

1) NaNH2

2) X(CH2)nCO2M 3) H +

OH

X =C1, Br n = 1-3,5,6,10 M =Li, Na 0 0 / / ~ / R 1

NaNH2~

R2 O

~

site b

O 1) NaNH2 2) CO2 3) H +

~

R

--O

R2

site a

O

R . ~ ~

--O

-0 -0 ~ R 1

O ~

OH O

R = Ph, Me, Ph(CH2)2, n-Bu, PhCH=CH, Ph2C=CH, H, OEt Fig. 1. Reaction of disodiosalts of diketones.

446

~

ONa //-,.,....,,Br

O NaNH2 ,.._ NH3 liq.

EhO

ONa

O

NaNH2

1) C2H2

NH3 liq.

2) H +

O RI~.,/R

2

NaNH2 0~ Et20

O R

~...I

O

ONa R I ~

O

R3COCI R2

R I ~ R 3 R2

R

C1

NaNH2 NH3

N

)
5,000 mg/kg DL 50 (rat) : 3700 g/kg This document aims to review of the use of anisole as a solvent.

481

ORGANIC SYNTHESIS

Grignard reaction Anisole, as a solvent, has been widely used in the Grignard reaction. R.N. Lewis and J.R. Wright (ref. 3) report that anisole, which is a slightly basic solvent, gives a quick reaction between Grignard's reagent and acetone.

+

Anisole

~MgCI

0

OH

Other authors also describe different synthesis reactions 9 via an organomagnesium compound in anisole (ref. 4 ) HO --N

/~ /

HO MeMgI Anisole

/ /

t=2h. T= 85~

or in ether/anisole mixtures (ref. 5).

oi< ~

MeMgI Ether Anisole

OH

1 -0,33 h

Organolithium condensation Besides lithiation of anisole, anionisations by butyllithium of more acidic compounds can be obtained (ref. 6).

482

BuLi (1,6 M) CuC12

S

~

~

Hexane/ Et20 / Anisole - 1 0 . 100 ~

(2 eq.)

Yield = 90 %

Halide exchange Anisole has been used to perform the synthesis of fluorinated compounds through halide exchange (ref.7).

~

" RMnH + A" 490

RM n H is the polymer chain containing n monomer units and AH is a chain transfer agent or an inhibitor. The active radical site is transfered to another molecule.

Termination step by radical combination

RM n"

+

RMm ~

--~

RMn+mR

by disproportionation

RM n.

+

RM m"

--~

RMnH

+ Pm

Pm is a polymer chain. These two reactions lead to radical destruction.

General concepts of inhibition Inhibitors are substances that stop radical polymerization. Those have quite a rapid action on initiator and propagator radicals. They transformed them either into a non radical form or into a radical with low reactivity in propagation reaction. They will block the radical polymerization over a so-called induction period that will vary according to its concentration in the reaction environment and to the experimental conditions (temperature, catalyst, etc.). Beyond this induction period, polymerization will occur at the same rate as it would if the inhibitor was not present. (ref. 8). The reaction scheme for ideal inhibition is set out below X --> 2 R" R"

+

Z

-->

inactive product (rate constant kz)

This reaction competes with the chain-growth process :

R" + M --> RM" (rate constant kp) On this basis, it is then possible to determine an inhibition constam. This constant requires an exact kinetical analysis of each separate process, which has not been carried out in most cases. Generally only the ratio kz/kp is obtained. Some values (refs. 9 - 11) are shown in Table 1.

491

Table 1. Inhibition constants at 50~ Inhibitor Nitrobenzene

Monomer Styrene Methyl acrylate

z = kz/kp 0.326 0.00464

1,4-Dinitrobenzene

Vinyl acetate Styrene

11.2 13.52

1,3,5-Trinitrobenzene

Vinyl acetate (45~ Styrene

68.5 64.2

Methyl acrylate

0.204

p-Benzoquinone

Vinyl acrylate Styrene

404 518

Acrylonitrile

0.91

Methyl methacrylate Styrene

5.7 2040

Chloranile Copper (II) chloride

Methyl methacrylate (44~

0.26

Styrene

11000

Acrylonitrile (60~

100

Diphenylpicrylhydrazine (DPPH) Hydroquinone

Methyl methacrylate (60~ Methyl methacrylate (44~ Vinyl acetate

1027 2000 0.7

1,2,3-Trihydroxybenzene 2,4,6-Trimethylphenol Oxygen

Vinyl acetate Vinyl acetate Styrene

5 0.5 14600

Methyl methacrylate

33000

Phenol

Methyl acrylate

0.0002

p-Nitrophenol

Vinyl acetate Methyl acrylate

0.012 0.0649

References 9, 10 and 11.

It can be noticed that the inhibition constant kz/kp initially varies considerably as a function of the reactivity and the polarity of the chain growth radical. It is therefore difficult to extrapolate the efficiency of an inhibitor for a given monomer such as styrene to other monomers, for instance of acrylic type. Furthermore, as this table shows, oxygen is one of the strongest known free radical inhibitors : it acts on alkyl radicals to form peroxy radicals. However, these peroxy radicals can also graft onto the monomer to form an oxidized polymer. The 492

fact is that oxygen can have a dual role : one as an inhibitor (blocking the alkyl radicals in the reaction medium) and the other as polymerization initiator. To illustrate this,

we can try to diagramatically represent an inhibition

mechanism that is close to reality (Scheme 1).

X

~- [R~

02

~

ROO*

M

~

RO2M 9

02

RO2MOO ZH

@

z

RM

Scheme 1. Inhibition mechanism of free-radical polymerization

In fact the inhibition mechanism is much more complex than showed above since the propagating radicals can be alkyl radicals and/or peroxy radicals.

THE RANGE OF M O L E C U L A R INHIBITORS Radical polymerization inhibitors are therefore molecules that are able of reacting with the radicals present in the monomer (either alkyl radicals or peroxy radicals) to give very low reactive radicals which will stop the chain growth. It should be noted that the formation of these products (inert in terms of the polyaddition reaction) can result from several basic and consecutive reactions. This is why inhibition mechanism can sometimes be rather complex. Generally speaking, the chemical structures of these free radical inhibitors can be classified into two main family (ref. 12) : -

acceptor type radical inhibitors,

-

donor type radical inhibitors.

Acceptor type radical inhibitors Acceptor radical inhibitors will be capable of oxidizing the alkyl radicals R" by accepting an hydrogen atom or even an electron by means of an addition

493

mechanism. They are more reactive in an oxygen deficient environments. The main chemical families are presented below.

Quinones Quinones are the most extensively studied inhibitors of radical polymerization and they represent an important inhibitor class. They have complex behaviour and numerous inhibition mechanisms. Even if some details are not yet fully understood, it is now accepted that the main mode of reaction is an addition of the propagating radical to the oxygen of the quinone, as in the following reaction : Rn" + O

~

O

~

R n O @ O *

This aryloxyl radical may terminate a second chain as follows R n O ~ O

~

+ Rm~

.~

RnO~ORm

According to F.Tudos's investigations (ref. 13) on styrene polymerization, quinones can be divided into three groups : Benzoquinone and its non-halogen-substituted derivatives (1,4-Benzoquinone (BQ), 2-Me-BQ, diMe-BQ, triMe-BQ, tetraMe-BQ, MeiPr-BQ, diiPr-BQ, Bu-BQ etc...) - Halogen-substituted derivatives of benzoquinone (C1-BQ, Chloranil, etc...) Quinones with condensed ring systems (Naphtoquinone, Anthraquinone, Phenanthrenequinone etc...) -

-

Aromatic nitro compounds This class includes compounds such as : nitrobenzene, (di)nitrobenzenes, dinitrophenols, (di)nitrocresols or even (di)nitro(alkyl)phenols or cresols. Generally, these compounds are more effective with electron-rich monomers (vinyl acetate or styrene) but have very little effect on methyl methacrylate or methyl acrylate. Here again there are several inhibition mechanisms (ref. 14).

494

Nitroso compounds This inhibitor class can be broken down into two sub-groups: -

aromatic nitroso compounds type O N - ~ ~ ) ~

with R = H, Me, Et, OH,

R

or Nitrosonaphtol, 2-methyl

OMe, C1, p-NH2, NHCH3, NMe2, N(Alkyl), NPh2 4-nitrosophenol, Me-nitrosophenol

- N-nitroso compounds (N-nitrosodiphenylamine, N-nitrosophenylhydroxylamine, Cupferron...)

Metal salts A number of metal salts are well known to be radical inhibitors under certain conditions. In fact besides the rather pro-degrading nature of metal reactions 3+ 2+ + [ Fe + RH ~ Fe + H + R ~ ], metals salts may performed stabilization reactions depending on the redox potential of the system in question. Metal salts may also have some radical-trapping type stabilizing reactions. This is often the case for copper ions:

Cu

2+

+ R" ~

Cu

+

+ R

+

9

For example propagating radicals derived from acrylonitrile could be terminated by Fe(III), Mo(III), Ti(III), Cr(II)...

Other inhibitors Besides these main products, other inhibitors can also be mentioned such as : - dibenzofulvene derivatives - aromatic azo compounds - phenylacetylene - pyridinic derivatives etc...

Donor type radical inhibitors Donor radical inhibitors tend to reduce peroxy radicals by giving an hydrogen atom or an electron by transfer. They act most favourably in oxygen-rich environments. The main chemical families are presented below.

Phenolic compounds Phenolic compounds (phenol, hydroquinone, hydroquinone monomethylether, methylhydroquinone,

tertbutylhydroquinone, 495

catechol,

tertbutylcatechol

etc...)

represent the major class of polymerization inhibitors for vinyl monomers. Phenols inhibition mechanism can be represented as follows : ROO ~

+ HO ~

X

~

ROOH +

"O

X

Phenolic inhibitors in this case exhibit a so called oxygen synergy. This reaction is much more rapid than the transfer reaction on an alkyl radical. Moreover, the slightest trace of oxygen will very rapidly form a peroxy radical from an alkyl radical. Aromatic amines

Alkyl diphenylamines, alkyl p-phenylenediamines, phenothiazine are the main amines which are used.

~

.oo-

~

N H - R '

ROOH +

R

,

The same oxygen synergy as with phenolic inhibitors is exhibited with diphenyl aromatic compounds and phenylene diamine compounds (ref. 15). However, phenothiazine is a rather different case. Indeed it is well known that phenothiazine directly reduces alkyl radicals and not peroxy radicals (ref. 16), and that it works better in anaerobic environments (ref. 17). A lky lh y drox y lamin e s RO0~

+

\ N-OH

/

~

ROOH

\ N --0"

+

/

This type of compound (diethylhydroxylamine, dibutylhydroxylamine etc...) can form both a donor and an acceptor inhibitor system since the nitroxylated radicals that are so-formed are excellent acceptor inhibitors. They may therefore have either function according to the operating conditions in which they are used. Metals salts

It is again the redox potential of the environment that will determine whether certain metal salts act as donor inhibitors, as is the case for copper ions e.g. Cu

+

+

H

+

+ ROO ~

--)

Cu

2+

496

+

ROOH

THE USE OF P H E N O L I C INDUSTRY

INHIBITORS

IN THE VINYL M O N O M E R

The importance of phenolic inhibitors in the vinyl monomer industry. The role of phenolic compounds as stabilizers and antioxidants has been studied very extensively in polymers and copolymers (refs. 18, 19). Many papers are devoted to this problem. Studies have been made on the optimization of phenolic structure based on hydroquinone (ref. 20) or catechol (ref. 21) as an antioxidant in polypropylene for example. Others have dealt with the influence of the polarity or stearic effect for different phenolic compounds or substituted phenols on the kinetics of antioxidation reactions - for example in polyvinyl acetate (ref. 22). Lastly, many papers have discussed on kinetic effects. However, it is rather surprising that this type of product has not represented a major role as an inhibitor of the radical polymerization of vinyl monomers. No indepth theoretical research has been published, as it is the case of antioxidants, specifically dealing with the inhibition of monomers. The sparse information that is available is limited to a few papers and reviews. This is indeed less surprising since the problem of stabilization of vinyl monomers (particularly styrene, butadiene, (meth)acrylates, etc...) during manufacture and storage is as important as the commercial stabilization of polymers. The fact is that all vinyl monomers tend to polymerise very easily, especially at high temperatures. During their industrial production, a purification operation is generally needed to obtain products of the required purity. Since distillation is the most commonly used process to perform this purification; inhibition of the polymerization of vinyl monomers during distillation (sometimes carried out at high temperature) is of great importance. If polymerization takes place during this distillation step, the formation of high molecular weight compounds of different structures may occur and cause deposits on the internal surface of processing equipment, often rendering it inoperative. This undesirable effect results in a loss of monomer, limits plants efficiency, increases security risks and lowers product quality. Inhibitors used in these processes belong to various classes of compounds. The principal selection criteria of such "process-inhibitors" are the radical trapping efficiency and the ease of separation from the final product. The consumption of the phenolic inhibitors used during this process step is either on their own or mixed with other inhibitors from different families and represents around 50 % of the total consumption of process inhibitors. 497

On the other hand, the stabilization of monomers for storage in the factory or during road, rail or sea transportation, is a great problem in terms of monomer quality, and especially in terms of safety. Indeed, for all vinyl monomers, the polymerization reaction is exothermic and autocatalytic once initiated. It is therefore very difficult to have storage equipment that can withstand such pressures (ref. 23). The main selection criteria for such "package-inhibitors" are their effectiveness at ambient temperature, their colourlessness and their ability to be removed before polymerization or to be overcome by the use of moderate amounts of initiator under the customer's polymerization conditions. At the present time, phenolic inhibitors are greatly preferred by all the major production companies in the vinyl monomer industry. Therefore the consumption of phenolic inhibitors represents over 95 % of the consumption of package inhibitors. As mentionned in the introduction, inhibitors which are preferably use during manufacturing may not be suitable under stockage conditions. In any case the choice of phenolic products is clearly manifest. In fact, in the vinyl monomer industry, over 50 % of the polymerization inhibitors (process and package) are phenolic compounds.

Synergistic effect with oxygen. The particularity of the phenolic inhibitors used in industry, is their synergistic effect with oxygen. These inhibitors require a minimum amount of oxygen in order to be as effective as possible. This effect has been described in literature for the stabilization by hydroquinone or paramethoxyphenol of acrylic acid (refs. 24, 25), methacrylic acid (ref. 17), methyl methacrylate (ref. 26), vinyl acetate (refs. 27, 28) and styrene (refs. 29, 30). The effect occurs due to the extremely rapid reaction of the first alkyl radical with oxygen [R" + 02 ~ either termination reaction

ROO'], several orders of magnitude faster than [R ~ + ZH --> RH + Z ~ or propagation

[R" + M - - ~ RM'] e.g.. for methacrylic acid kl = 20,000 kp. The resulting peroxy radical reacts very rapidly with the phenolic inhibitor in a termination step. Industrially, the presence of oxygen in the production lines arises from the contact between hydrocarbons and air, or gases containing oxygen and in systems operating at reduced pressure as a consequence of leaks in the equipement. Therefore inhibitors which work in aerobic conditions are the most common. Generally for storage and transportation of such monomer, contact with air or a 50/50 nitrogen/air mixture is maintained. 498

Specifications

for

radical

polymerization

inhibitors

There are a certain number of factors to consider when selecting an inhibitor to avoid unwanted polymerization. The main key factors to take into account are : -

the operating environment (distillation column, reboiler, condenser, tank, drum,

etc) -

the nature of chemical compounds in the operating system (liquid, gas, pH,

reactivity of these products with inhibitors) -

temperature and pressure

-

presence of oxygen

-

ecotoxicity with respect to recycling or elimination of process effluents. Consequently the characteristics of a process inhibitor and a package inhibitor

are not the same. The main factors in selecting a process-inhibitor and a packageinhibitor are summarized in the Table 2.

Table 2. Characteristics of process inhibitor and package inhibitor Inhibitor Process-inhibitor Concentration range 20 ppm to 1% - trapping efficiency Characteristics ease of separation from the final product - solubility in crude monomer, raw material and coproducts - thermal stability - volatility degradability - toxicity ease of handling - price

-

-

to -

Package inhibitor 2 ppm to 2000 ppm trapping effectiveness color formation solubility ease of handling ease of removal or ability override toxicity

Selection of an inhibitor must therefore be made carefully while keeping in mind these characteristics. Several phenolic compounds may have the right profile in each inhibitor category and not effecting the quality and cost of monomer production. Amongst other things, this explains the success of phenolic inhibitors in the vinyl m o n o m e r industry.

499

Examples of phenolic inhibitors for few vinylic monomers

Styrenic m o n o m e r s Styrenic monomers are the aromatic vinyl monomers such as styrene, vinyltoluene, a-methylstyrene and divinylbenzene (DVB). Styrene is one of the most important monomers produced by the chemical industry today. It is separated from the products of dehydrogenation of ethylbenzene by a multi-stage rectification. The process inhibitors for styrene are mainly nitroaromatic compounds or aromatic amines and phenolic compounds such as ptertbutylcatechol or 2,6 ditertbutylhydroxytoluene mixed with other radical acceptor inhibitor. Regarding stabilization for storage and transportation of these monomers, the best known inhibitor is paratertbutylcatechol (p-TBC) which is widely used by every styrenic monomer producer. It imparts no color, but does require a minimum of 15 - 20 ppm oxygen. It is easy to remove prior to polymerization by alkaline washing, by distillation or by passing through an activated alumina column. The amount of p-TBC varies between 12 and 50 ppm for styrene, vinyltoluene and amethylstyrene and could rise to about 1000 ppm for DVB. An illustration of the effectiveness of p-TBC is shown in Figure 1.

500

25-

J

20

J

15

J

10

J

J

J

J J

J _

0 ppm

J

ppm m

_ 0,2.' TIME (hours)

Figure 1. Styrene polymerization at IO0~

Influence of concentration of p-TBC

Butadiene The most important source of butadiene world-wide is C4 fractions obtained as a by-product from the cracking of naphta and gas oil for ethylene. Butadiene a colourless flammable gas, may polymerize in three ways briefly described as : to a liquid dimer which appear in small quantities during storage. However, the major portion tends to remain dissolved in the liquid. - to a dark rubber-like heavy polymer usually formed in the liquid phase. This type can cause trouble by plugging lines and equipment. to a crystalline-like white polymer called "popcorn" usually formed in the vapour phase. It causes severe trouble by plugging up refining still columns, valves, pumps, pipes. -

-

Different process-inhibitors may be used during manufacturing depending on each producer. The most widely used phenolic process-inhibitor is the p-TBC with a mixture of other non phenolic process-inhibitors, p-TBC is the best inhibitor for retarding the "popcorn seed" formation.

501

On the other hand, p-TBC is generally used in the final butadiene product between 70 and 150 ppm. The 2,6-ditert-butylhydroxytoluene (BHT) could also be used as an alternative.

(Meth)acrylic acid and (Meth)acrylates Acryclic acid is obtained by the catalytic oxidation of propylene and acrylates (methyl acrylate, ethyl acrylate, butyl acrylate, 2-ethylhexyl acrylate...) by alcohol esterification of the acid. The preparation of methacrylic acid involves the acidic hydrolysis of acetone cyanohydrin and methyl methacrylate is obtained by a similar process involving the methanolysis of acetone cyanohydrin. During their manufacture, these very reactive monomers are found in very diverse chemical systems with great acidity (H2SO 4 may be a catalyst) or basic pH (caustic soda may be a neutralizer). Process inhibitors are also often mixed with inhibitors from different chemical families that are capable of being effective in each system that is encountered. Industrially, inhibitors as diverse as benzoquinone, cupferron, manganese or cerium salts, hydroxylamine, copper alkyldithiocarbamate, hydroquinone, phenothiazine, etc. may also be used. The current worldwide trend is to use a mixture of phenothiazine and hydroquinone as a process inhibitor system. For package inhibitors, manufacturers use hydroquinone (HQ), hydroquinonemethylether or paramethoxyphenol (HQMME or PMP), 2,4-dimethyl 6-tertbutyl phenol and BHT as an alternative. However the most common is the hydroquinonemonomethylether due to its colourlessness and its efficiency (ref. 31). As an example, the concentration of PMP in the products usually varies between 10 and 300 ppm for methyl acrylate and methacrylic acid respectively. In summary, Table 3 presents a certain number of phenolic inhibitors that are used either during the process or storage of the main monomers including vinyl monomers such as acrylonitrile, vinyl acetate or acrolein.

502

Table 3. E x a m p l e of few phenolic inhibitors in vinyl monomer

Inhibitor

Formula

HQ

+ + OH

OH

OH

OCH3

acrylic acid

*

methyl acrylate

*

ethyl acrylate

*

butyl acrylate

*

2-ethylhexyl acrylate

*

methacrylic acid

*

methyl methacrylate

*

ethyl methacrylate

*

2-ethylhexyl methacrylate

*

acrylonitrile

*

vinyl acetate

PMP

TBC

OH

BHT

2,4-diMe 6tBu Phenol

OH

OH

CH3

CH3

*

vinylidene chloride acroleine

*

styrene c~-methylstyrene

vinyl toluene divinylbenzene butadiene isoprene chloroprene

CONCLUSION The inhibition of radical polymerization is important in the chemical industry to prevent unwanted polymerization of vinyl monomers during processing, storage and transportation. Having discussed some of the principles of free-radical polymerization, a brief overview of possible inhibition mechanisms is presented. A short summary of the major classes of radical inhibitors gives an idea of the wealth of choice of chemical compounds available for use in this application. Owing to specifications for industrial inhibitors and the attractive intrinsic properties of 503

phenolic compounds, these inhibitors are the most common class used in this industrial application world-wide. In fact phenolic inhibitors are either used on their own or mixed with other inhibitors from different chemical families. In virtually all vinyl monomer synthesis processes they are either as a process or as a package inhibitor. Finally, a brief overview of some phenolic products used for this application is presented for a few vinyl monomers.

References

1. G.Odian in "La Polym6risation Principes et applications", Polytechnica, Paris, (1994). 2. G.C.Eastmond in "Comprehensive chemical kinetics", Bamford, Tipper, Eds Elsevier, Vol 14A, pp127- 152, Amsterdam, (1976). 3. P.J. Flory in "Principles of polymer chemistry", Comell Univ. Press, Ithaca, New York (1953). 4. G.E.Ham in "Vinyl polymerization", Ed Ham, Dekker, New-York, (1967). 5. J.C.Bevington in "Comprehensive Polymer Science" Allen, Bevington, Eds Pergamon Press, Vol 3, Ch.6, pp 65, (1989). 6. J.C. Bevington in "Radical polymerization", Academic Press, (1961). 7. Encyclopedia of Polymer Science and Engineering, Wiley, Vol 13, New York, (1988). 8. J.C.Bevington, N.A.Ghanem, H.W.Melville, J.Chem.Soc., 2822, (1955). 9. Encyclopedia of polymer science and technology, Wiley, Vol 7, pp 658, New York, (1969). 10. J.Brandrup, E.H.Immergut in "Polymer Handbook", Interscience publishers, (1975). 11. G.C.Eastmond in "Comprehensive chemical kinetics", Bamford, Tipper, Eds Elsevier, Vol 14A, pp153, Amsterdam, (1976). 12. F.Tudos, Z.Fodor, M.Iring in "Oxidation inhibition in organic materials", Eds Pospisil, Klemchuk, Vol II, p. 219, CRC, (1990). 13. F.Tudos, T.Foldes-Berezsnich, Prog. Polym.Sci., 14, 717-761, (1989) 14. G.C.Eastmond in "Comprehensive chemical kinetics" Eds Bamford, Tipper, Eds Elsevier, Vol 14 A, Chapter 2, p. 104, Amsterdam, (1976). 15. R.E.Winkler and E.B.Nauman, J.Polym.Sci., Polym. Chem.Ed, 26, 2853, (1988) 16. L.B.Levy, J. Polym. Sci., Polym. Chem. Ed., 23, 1505, (1985). 17. A.Nicolson Plant/Operations Progress, 10 (3), 171, (1991). 18. Atmospheric oxidation and antioxydants, G.Scott Eds Elsevier (1993). 19. Oxidation inhibition in organic materials, J. Pospisil, P. P. Klemchuk Eds, CRC Press (1990). 20. J.Pospisil, L.Kotulak, L.Taimr in "Stabilization of polymers and stabilizer processes advances in chemistry series 85, R. F. Gould Ed., American Chemical Society Publications, Chapter 14, (1968). 21. J.Pospisil, L.Kotulak, L.Taimr in "Stabilization of polymers and stabilizer processes advances in chemistry series 85, R. F. Gould Ed., American Chemical Society Publications, Chapter 13, (1968). 22. M.Simonyl, F.Tudos, J.Pospisil, Eur.Polym.J.,3, 101, (1967). 23. J.L.Gustin, P.Vandermarliere presented at the " 7th International Symposium on Loss Prevention and Safety Promotion in the Process Industries", Taormina, Italy 4 - 8 May 1992. 24. L.B.Levy Plato/Operations Progress, 6 (4), 188-189, (1987). 25. J.J.Kurland, J. Polym. Sci., Polym. Chem. Ed., 18, 1139, (1980). 26. R.G.Caldwell, J.L.Ihrig, J. Am. Chem.Soc., 84, 2878, (1961). 27. L.B.Levy, Process Safety Progress, 12 (1), 47, (1993). 5o4

28. 29. 30. 31.

L.B.Levy, L.Hinojosa, Journal of Applied Polymer Science, 45,537, (1992). G.L.Batch, C.W.Macosko, Thermochimica Acta, 166, 185- 198 (1990). A.A.Miller, F.R.Mayo, J.Am.Chem.Soc., 78, 1017, (1956). L.S.Kirch, J.A.Kargol, J.W.Magee, W.S.Stuper Plant/Operations Progress, 7 (4), 270-274, (1988).

505

T R A C I N G B A C K T H E O R I G I N O F V A N I L L I N BY S N I F - N M R

GERARD J. MARTIN Universit6 de Nantes - Laboratoire de R6sonance Magn6tique Nucl6aire et R6activit6 Chimique -URA CNRS/472 - France

INTRODUCTION A chemically pure compound, such as 99.99 % vanillin, is in fact a complex mixture of isotopomers, themselves a combination of isotopes of the elements concerned (C, H and O), and distributed according to their isotopic abundances. Probability rules predict 884736 isotopomers of vanillin molecule, C8H80 3. At the

natural abundance of the stable isotopes of C, H and O (Table 1) the light isotopomer of vanillin, 12C 81H816r'-'3, has an occurrence probability of 0.90685, whereas the heavy isotopomer, 13C82H81803, has only a 6.4478 E-55 chance of occurring in nature. In other words, we would have to produce a mass of vanillin much larger than the mass of the Sun to be able to observe its heavy isotopomer !

Table 1.

Isotopic abundances (in ppm) of the natural isotopes of H, C and O. The boxes indicates the nuclei which have a non-zero magnetic moment and may be studied by NMR. The brakets indicates radioactive nuclei.

1

1H 2 1

H

OXYGEN

CARBON

HYDROGEN

999844.26

155.74

10-10

12 6C

988887.67

16

O

997625.81

17 8~

373.00

8

13 6C

11112.33

[1% 6 ]

10-6

506

18 8~

2001.9

From a practical point of view, NMR spectroscopy is able to observe 13 isotopomers of vanillin containing one heavy isotope (8 for 13C and 5 for 2H, Fig.l a, l b), and in extreme signal-to-noise conditions, may detect the double natural labels 2H-13C. On the other hand, Isotope Ratio Mass Spectrometry (IRMS) and Liquid Scintillation Counting (LSC) or Tandem Mass Spectrometer Ion Accelerator (TMSIA) lead to overall contents in 2H, 13C, 180 and 3H, 14C respectively. This isotopic fingerprint is a very powerful means for assigning the origin of vanillin and can also provide useful indications on the transformation processes undergone by the raw materials and precursors of vanillin, since the isotopic abundances of a given chemical species do not follow random distributions bqt obey well defined physical, chemical or biochemical laws.

So I yen'l"

I

2

3

Ref.

4

, , I , , , , i , , W , l , l , w | ' ' ' W l W W ' ~ l t0.0 9.0 8.0 7.0 6.0 PPN

5.0

~

'

l

w

' ' 4.0

5

~

l

w

' ' 3.0

'

21

.0

'

9

Fig. la. Fig. 1.2H (Fig. la) and ]3C (Fig. lb) NMR spectra of vanillin sample prepared from guaiacol, recorded at 11.4 T 507

So I v e n t

i

DD~

9 "

2

'

9

I

180

9 "'

'

',

i

160

"

,

3

'~

4

9

~

t 0

,-

5

.,

6

i..

i

120

i..

7

~

T.

i

t00

,

~

,

i

80

9

""

r

9

Fig. lb.

The thermodynamic and kinetic isotope effects which are always associated with a transformation (i.e. distillation, chemical reaction...), induce varying degrees of isotopic fractionation of the product with respect its substrate. However, these differences in isotopic composition may be very small and accurate and precise quantitation methods are required if NMR Spectroscopy or Mass Spectrometry are to be used for isotopic analysis. The following points will be presented successively to illustrate the ability of isotopic analysis to trace back the origin of vanillin : i - NMR and isotopic methodologies. ii - Authentication of the natural or synthetic status of vanillin : the analytical aspect. iii - Explanation of the isotopic fractionation observed 9the mechanistic aspect.

508

]

60

9

M E T H O D O L O G Y O F I S O T O P I C ANALYSIS BY N M R S P E C T R O S C O P Y Individual resonances in {1H} decoupled 2H, 13C... spectra correspond to well defined isotopomers and their abundances are directly proportional to the signal intensities. In the case of the 2H-spectrum of vanillin, five signals are observed at 11.4 T

H~

1 "C~ O

3

4

6

08CH3 OH since even at this high field, the H 3 and H7 resonances are deceptively equivalent. The chemical shifts are presented in Table 2. Table 2. Chemical shifts of vanillin in ppm/TMS Site

1

2

3

4

5

6

7

8

82H/TMS

9.8

/

7.2

7

8.1(OH))

/

7.2

4

5X3C/TMS

153.2

91.7

88.5

77.1

114.6

110

72.1

17.6

For curve fitting procedures, it is more conveniem to assign isotopomers in decreasing chemical shift order which give respectively 9 for 2H Isotopomer :

1

2

3

4

5

Chemical shift : 1 for 13C

5

3,7

4

8

Isopotomer

1

2

3

4

5

6

7

8

Chemical shift

1

5

6

2

3

4

7

8

The assignment of 2H and 13C signals to the isotopomers of vanillin presents no difficulties.

On the other hand a precise and accurate determination of line

intensities is a considerable challenge. Precision depends mainly on the signal-tonoise ratio (S/N) achievable in a reasonable period of time which in turn is a function of the acquisition time, AQ, the delay between pulses, PD, and the number 509

of scans, NS. In the case of the 2H spectrum of vanillin, a value of 3.4s for AQ easily fulfills the condition of complete relaxation of the nuclei after a 90 ~ pulse width (PW) and a minimum of 2400 scans are required to obtain a (S/N) value higher than 200 for most of the isotopomers examined at 11.4T. To limit random errors which could be associated to short term instabilities of the spectrometer, a given spectrum is repeated three times. Using an automatic sample changer, three vanillin samples can therefore be studied in 24 hours. The sensitivity requirements of 13C spectroscopy are somewhat less severe than for 2H-NMR. Usually, S/N values higher than 200 are obtained for 13C isotopomers in 2 hours at l l.4T for a mixture of vanillin and relaxation reagents. (AQ=6.8s N S = 128 and P W = 9 0 ~ (ref. 1). The precision of the intensity determination, expressed in terms of repeatability (mean standard deviation of NE replications of NS scans) is of the order of 0.5 % to 2 % in relative values depending on the nature of the isotopomer and for a given treatment of the FID's. Analytical methods used for quality control purposes must also be accurate. Accuracy is defined as the closeness of agreement between the true value and the mean of a series of replicates. Reference materials certified by international organizations are available for judging the accuracy of an analytical procedure. As far as NMR spectroscopy is concerned, systematic errors which are frequently related to the lack of accuracy arise mainly during the acquisition step and in the treatment of the FID's. In order to minimize bias in signal acquisition, oversampling and very stable decoupling procedures are recommended, and, to obtain very reproducible intensity values, an operator independem curve fitting procedure should be implemented in the analytical sequence. The Interliss algorithm developed in Names (ref. 2) iterates in the complex plane from the raw frequency spectrum obtained after Fourier transformation without any phasis nor base line correction. The reproducibility of the FID treatment is thus optimized and small differences in the line intensities of different samples can be clearly shown. The 2H isotope ratios of vanillin are usually obtained by internal referencing from the signal intensity according to equation (1).

(D / H) i =

PwsMvmwsSi (D / H)ws

(1)

PiVMwsmvSwstv

where m, M, Pi and S are respectively the weights of products used and their molecular weights, the number of equivalent hydrogen in position i and the signal

510

intensity of the vanillin sample V and of the working standard WS. (D/H)ws is the isotope ratio of the reference used as the working standard and tv the weight/weight parity of vanillin. The site-specific isotope ratios (D/H)i are related to the overall isotope ratio (D/H) of the whole vanillin sample, obtained by IRMS after combustion of the product, by equation (2). n

Z Pi(D / H)i (D/H)i= i n ZPi

(2)

i

where the summation is carried out on the n isotopomers observable at the field considered. (D/H)i may therefore also be computed from IRMS data, (D/H), if for any reason, the NMR spectra of vanillin are run without addition of the working standard ws 9 n

fi

(D / H) i = Z. P i l l1 (D/H)

(3)

1

fi is the observed molar ratio of isotopomer i computed from the ZH-NMR spectrum

fi-

Si n ZSi

(4)

i

Equations 1 to 4 apply for all elements considered and the isotope ratio R, which are related to the isotopic abundance A R -

A A-1

(5)

are usually reported with respect to the international references V. Smow (2H and 180) and PDB (13C). The isotope ratios of these standards (Rref) were determined by rigourous international studies (Table 1) and are also used to define a relative scale of isotopic contents 5, expressed in per mil (%0):

511

5i =

((Ri))

Rref - 1 * 1000

(6)

Equations 3 and 4 are recommended to determine the Site-specific Isotope Ratios of carbon, since the existence of differential residual Nuclear Overhauser Enhancements in vanillin samples doped with Cr(AcAc)3 may induce more or less significant bias. To conclude this section, it should be recalled that data treatment is an important part of the analytical approach. The results of a SNIF-NMR experiment constitute a matrix of data where the variables are the isotope ratios or signal intensities of the different isotopomers and the individuals are the NE observations for a given sample. In this sense, SNIF-NMR is a second-order procedure (ref. 3) and multi-variate analysis is the appropriate method for evaluating the results if a linear behaviour of the variables may be assumed. When the whole set of vanillin samples from different origins is considered, we are faced with the necessity of choosing a classification procedure and defining belonging rules. A first approach which assumes a Gaussian distribution of the measurements is based on the computation of the Mahalanobis distances (ref. 4) from the means and the variances of the different groups considered. d2(M) = (~--g) C -1 (x-g)

(7)

where C is the variance-covariance matrix of the n isotope ratios measured for the n authentic vanillins of the group, ~t the vector mean of the group and x the coordinates of the unknown. It is recommended to associate to d(M) a probability P which will delineate the contour of the group as a function of Z2(et;n). In fact, the Mahalonobis distance gives a default classification since it does not take into account the random errors of measurement and the non Gaussian contribution in the distribution of the data. A better way to determine wether an unknown belongs to a reference group and to compute any mixture composition is to carry out a Monte-Carlo simulation (ref. 5). An overall belonging probability P is computed in the form of the summation of the different Gaussian distributions associated to each of the different measurements which constitute the reference group R. However it happens frequently that an unknown does not belongs frankly to a group of well defined origin but is in fact a mixture (A) of several origins (m) : 512

(8)

A = a 1 + a 2 ........ a m

The measurement function R M (1,n) determined on the unknown is a linear combination of the n isotope ratios Rij of the m origins and A is readily computed from the least square solution

A = ( R R ) -1 R -1

RM

(9)

A multi-variate probability may also be computed if we include the dispersion (variance) of the reference group by computing the Mahalanobis distances between the unknown and well defined mixtures of vanillins from different origins. It is obvious that the same limits as those described for the classification step apply to this procedure and a Monte-Carlo simulation should refine the determination of the mixture composition. The only drawback of the Monte-Carlo approach is the computation time since in the case of four isotope ratios measured for three groups of vanillins, a 1% precision requires 10Ell tests, i.e. one week of computation for a 60 MHz-Pentium based micro-computer (ref. 5) ! Even if the Mahalanobis distance gives less precise results, it is preferable to the Monte-Carlo simulation for most routine analyses.

A U T H E N T I C A T I O N OF V A N I L L I N S F R O M D I F F E R E N T O R I G I N S : THE

ANALYTICAL APPROACH The vanillin market is in fact shared by two products : the natural vanillin contained in vanilla beans, an extremely highly priced (~ 4000 $/kg) and scarce product, and the synthetic (from phenols) or hemisynthetic (from lignin) vanillins which are produced in large quantities and are relatively cheap (40 $/kg). According to the main food industry regulations in the European Union and the USA, a flavour is said to be natural (Table 3) only if the raw material comes from the living pool and if physical or biochemical processes only have been involved during the different manufacturing steps.

513

Table 3.

Classification of the status of a food additive (i.e. aroma) as a function of the nature of the raw material and process used in its manufacture

RAW MATERIAL Living

PRODUCT CLASSIFICATION Natural

Natural

Artificial

(named botanical)

(oiotechnological

(hemisynthetic)

product) Fossil

Artificial

Artificial

Artificial

(Biotechnological

(synthetic)

product) Process

Biotechnical

Physical

Chemical

Obviously, such a situation is likely to encourage cupidity and in the food industry fairly substantial quantities of vanillin ex-lignin have been mislabelled as natural. Various compositional criteria have been developed in an attempt to combat such fraud but give the considerable price difference between both qualities of vanillin, adjusting the chemical composition of adulterated samples to resemble that of the natural product was not a great price to pay. As far as stable isotope analysis is concerned,

13C-IRMS would have been a choice method to detect such

adulteration, since, as shown in Table 4, there is a very significant difference between natural 613C ---20.5 %o and hemisynthetic vanillin (613C -- -28 %o).

514

Table 4.

SNIF-NMR determination of the isotope ratios of vanillin isopotomers 1

2

3

4

5

Total

VN

130

(a)

160

200

130

145

VL

110

(a)

130

175

110

130

VG

350

(a)

140

150

2H Isotopomer .,

ppm/V.SMOW

,

X3CIsotopomer VN %o/PDB

%o/V.SMOW

8

Total

-45

-10

-50

-20 -28

VL

-55

-20

-55

-50

-20

-85

-30

1

2

3

Total

VN VL

,

160

2 to 7

VG 180 Isopotomer (b) I

120

1

2O (c)

(c)

L vc (a)

(c)

16 18

Due to fast chemical exchange with lable hydrogens of the medium this site is irrelevant for authentivation of vanillins

(b)

SMRI determination

(c)

Not yet available

Unfortunately, the natural product is heavier than its synthetic counterpart and it is relatively easy and not prohibitively expensive to add a small quantity of

13C

enriched vanillin to the synthetic product. From this point of view, vanillin exlignin is preferred to vanillin ex-guaiacol since the quantity of labelled material required is lower. Applying isotopic dilution equations, it is apparent that less than 13 one hundred milligrams of 99% C-labelled vanillin are required to transform l kg vanillin ex-lignin into lkg natural vanillin ! As early as the 1980's, it was possible to purchase 0.1g of such enriched compounds for 300 to 600 $, depending on whether the carbonyl, methoxy or ring was labelled. This rendered invalid the ~13Ccriterion for authenticating vanillins and lead to the investigation of deuterium NMR as a potential method to discriminate between synthetic and natural products (ref. 6). As shown in Table 4, the deuterium finger print of natural vanillin exhibits large differences to those of its synthetic analogues. The Mahalanobis distance dM between the gravity centres of natural and guaiacol vanillin is twenty times larger than between those of natural and lignin vanillin. However, even in this less favourable case, dM is much greater than the 99% probability contour of the natural vanillins defined by Z2(0.99 ; 4). 515

It is even possible to characterize the addition of 5 % lignin vanillin in the natural product at a 95 % confidence level. In fact the existence domain of a given group defined by 4 isotope ratios is a hyper-ellipse and it is more convenient to represent the populations of the three main groups of vanillin in a plane constructed from the two canonical functions D~ and D2. Figure 2 illustrate the ability of 2H-NMR to discriminate between the vanillin groups. Since three groups are involved, the whole variance is distributed in only two orthogonal directions. The main axe D1 is closely related to the deuterium content of the formyl group, whereas the second function D2 makes a clear distinction between natural and lignin vanillin. The classification is 100% complete at a 99.9 % confidence level.

516

b...,.

to

o

I0.00

26.70%

o 8.00

The Roots of Organic Development (Industrial Chemistry Library)

Industrial Organic Chemistry

Handbook of Industrial Chemistry. Organic Chemicals

Handbook of Industrial Chemistry: Organic Chemicals

The Organic Chemistry of Sugars

Fundamentals of Organic Chemistry

Organic Chemistry

Organic chemistry

Survey of Industrial Chemistry

Organic Chemistry I (Organic Chemistry Fundamentals) (SparkCharts)

Organic chemistry

Organic chemistry

Dictionary of Organic Chemistry

Organic Chemistry

Organic Chemistry

Organic Chemistry

Organic Chemistry

Organic Chemistry

The Organic Chemistry of Sugars

The Organic Chemistry of Sugars

Organic Chemistry II (Organic Chemistry Reactions) (SparkCharts)

The industrial development of Poland

Survey of Industrial Chemistry

Organic Chemistry of Explosives

Organic chemistry of coal

Organic Chemistry

Organic Chemistry

Organic Chemistry

Organic Chemistry

Organic Chemistry

Organic chemistry

The Roots of Organic Development (Industrial Chemistry Library)

Industrial Organic Chemistry

Handbook of Industrial Chemistry. Organic Chemicals

Handbook of Industrial Chemistry: Organic Chemicals

The Organic Chemistry of Sugars

Fundamentals of Organic Chemistry

Organic Chemistry

Organic chemistry

Survey of Industrial Chemistry

Organic Chemistry I (Organic Chemistry Fundamentals) (SparkCharts)

Organic chemistry

Organic chemistry

Dictionary of Organic Chemistry

Organic Chemistry

Organic Chemistry

Organic Chemistry

Organic Chemistry

Organic Chemistry

The Organic Chemistry of Sugars

The Organic Chemistry of Sugars

Organic Chemistry II (Organic Chemistry Reactions) (SparkCharts)

The industrial development of Poland

Survey of Industrial Chemistry

Organic Chemistry of Explosives

Organic chemistry of coal

Organic Chemistry

Organic Chemistry

Organic Chemistry

Organic Chemistry

Organic Chemistry

Organic chemistry

Recommend Documents