摘要:

CHEMICALSOCIETY Volume 26 Pages 73-1 46 April 1997 ISSN 0306-0012REVIEWS Issue 2 CSRVBR 26(2) 73-1 46 Peptide nucleic acid. A DNA mimic with a pseudopeptide backbone Peter E. Nielsen and Gerald Haaima 73-78 Trends in isothermal microcalorimetry Ingemar Wadso 79-86 Modern valence bond theory J. Gerratt, D. L. Cooper, P. B. Karadakov and M. Raimondi 87-100 Developments in metalorganic precursors for semiconductor ....................... 1I @Q 4 growth from the vapour phase Anthony C. Jones 101-1 10 Modern tanning chemistry Anthony D. Covington 11 1-126 Carbon+arbon bond-forming reactions mediated by cerium(1v) reagents Vijay Nair, Jessy Mathew and Jaya Prabhakaran 127-132 Lead, glass and the environment Michael J. Hynes and Bo Jonson 133-146 Articles that will appear in forthcoming issues Functionalised conducting polymers S.Higgins Semiconductor micromachining David Schiffrin Surface science aspects in semiconductor electrochemistry H. J. Lewerenz Reactions of complex metalloproteins studied by protein film voltammetry Fraser A. Armstrong, Hendrik A. Heering and Judy Hirst Trends in organic electrosynthesis James H. P. Utley Modern aspects of battery design John Owen Electrochemistry for a cleaner environment Daniel Simonsson Electrochromic materials Roger J. Mortimer Microdialysis sampling coupled on-line to microseparation techniques Malonne I. Davies and Craig E. Lunte Microwave chemistry Saskia A. Galema Intramolecular vibrational energy redistribution Dean Boyall and Katharine E. Reid New mass spectrometric methods for the study of noncovalent associations of biopolymers Richard D. Smith, James E. Bruce, Qinyan Wu and Q. Paula Lei Some aspects of organic pigments Zhimin Hao and Abul Iqbal MELDOLA MEDAL: Understanding the properties of urea and thiourea inclusion compounds Kenneth D. M. Harris Speciation of trace metals in the environment Steve J. Hill Chemical Society Reviews, 1997

ISSN:0306-0012    

DOI:10.1039/CS99726FP003

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Chemical Society Reviews Editorial Board Jean-Pierre Sauvage (CNRS, Strasbourg) [Chair] Vicenzo Balzani (Bologna) Ed C. Constable (Basel) Chris Elschenbroich (Marburg) Tim C. Gallagher (Bristol) Editorial Office Martin Sugden (Managing Editor) David Bradley; Peter Whittington (Production) Debbie Halls (Editorial Secretary) http://c hem i stry .rsc. org/rsc tel: +44 (0)1223 420066 Chemical Society Reviews publishes concise, succinct and lightly referenced articles that provide an introductory overview to topics of current interest in chemistry. The articles appeal to the general research chemist as well as to the expert in the field and provide an essential starting point for further reading. Advanced undergraduates, postgraduates and experienced re- searchers should all benefit from reading Chemical Society Reviews.Chemical Society Reviews (ISSN 0306-0012) is published bimonthly by the Royal Society of Chemistry, Thomas Graham House, Science Park, Cambridge, UK CB4 4WF. 1997 subscription rate: &I30(USA $234). Customers in Canada will be charged the sterling price plus a surcharge to cover GST. Individuals can subscribe for &45 (USA $80) providing their institutional library takes a full price subscription. All orders accompanied by payment should be sent directly to The Royal Society of Chemistry, Turpin Distribution Services Ltd, Blackhorse Road, Letchworth, UK SG6 lHN, (NB Turpin Distribution Services Ltd., distributors, is wholly owned by the Royal Society of Chemistry.) Payment should be by cheque in pounds sterling payable on a UK clearing bank or in US dollars payable on a US clearing bank.Second class postage is paid at Zdenek Herman (Prague) Horst Kunz (Mainz) John P. Maier (Basel) D. Mike P. Mingos (Imperial) Jeremy K. M. Sanders (Cambridge) Royal Society of Chemistry Thomas Graham House Science Park Cambridge UK CB4 4WF csr@ rsc.org fax: +44(0) 1223 420247 The Editorial Board commissions articles that encourage international, interdisciplinary dialogues in chemical research. The Board welcomes any suggestions for new articles. A guide for authors and synopsis form can be found in the first issue of this year’s volume or on the RSC’s World-Wide Web home page (URL above).Alternatively, they can be requested from the Managing Editor, in paper or electronic form (postal and e- mail address above). Jamaica NY 1141-9998. Airfreight and mailing in the USA by Publications Expediting Services Inc., 200 Meacham Avenue, Elmont, NY 11003 and at additional mailing offices. US Postmaster: send address changes to Chemical Society Review, c/o Publication Expediting Services Inc., 200 Meacham Ave- nue, Elmont NY 11003. All dispatches outside UK by bulk airmail within Europe and Accelerated Surface Post outside Europe. 0 The Royal Society of Chemistry, 1997 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, recording, or otherwise, without the prior permission of the publishers. Typeset and printed in Great Britain by Black Bear Press Limited.

ISSN:0306-0012    

DOI:10.1039/CS99726FX005

出版商:RSC    

年代:1997

数据来源: RSC

ISSN:0306-0012    

DOI:10.1039/CS99726BX007

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Peptide nucleic acid (PNA). A DNA mimic with a pseudopeptide backbone Peter E. Nielsena and Gerald Haairnab a Center ,for Biomolecular Recognition, Department .for Biochemistry and Genetics, Biochemistry Laboratory B, The Panum Institute, Blegdamsvej 3c, DK-2200, N Copenhagen, Denmark Fax: +45 31396042; Phone: +45 35327762; E-mail: penpanum@biobase.dk h Department of Organic Chemistry, The H. C. grsted Institute, Universitetsparken 5, DK-2100 (JCopenhagen, Denmark PNA (peptide nucleic acid) is a DNA mimic with a pseudopeptide backbone composed of N-(2-aminoethyl)gly- cine units with the nucleobases attached to the glycine nitrogen via carbonyl methylene linkers. PNA was first described in 1991 and has since then attracted broad attention within the fields of bioorganic chemistry, medici- nal chemistry, physical chemistry and molecular biology due to its chemical and physical properties, in particular with regard to efficient and sequence specific binding to both single stranded RNA and DNA as well as to double stranded DNA.The present review discusses the structural features that provide the DNA mimicking properties of PNA and gives an overview of structural backbone modifications of PNA. 1 Introduction Self-recognition by nucleic acids is one of the fundamental processes of life and also one of the most straightforward principles of molecular recognition in complex systems. Only four basal recognition units exist, the nucleobases adenine (A), cytosine (C), guanine (G) and thymine (T) [uracil (U) in RNA], that recognize each other two by two forming A-T and G-C base pairs by simple hydrogen bonding between com-plementary hydrogen bonding acceptor and donor sites of the nucleobases (Fig.1). Thus, it is not surprising that this simple four building block system has been a paradigm and inspiration for chemists attempting to and succeeding in making self- organizing systems (e.g.ref. 1). Also, a large number of close as well as more distant analogues of DNA itself by modifications of the sugar phosphodiester backbone have also been made mainly with the aim of developing efficient antisense drugs ~~~ ~~~~ ~~~~~~~ Peter E. Nielsen was born in Copenhagen, Denmark, in 1951. He rzceived his PhD in Chemistry in 1980 from the University of Copenhagen and has, apart from a postdoctoral stay at UC Berke- ley, worked at the Department of Biochemistry, The Panum In- stitute, University of Copenha-gen, where he is presently a professor and director of the Center for Biomolecular Recog- nition.His research interest is molecular recognition in gen- eral and DNA recognition in particular. He has developed the uranyl photofootprinting tech-nique and is a co-inventor of PNA. Peter E. Nielsen HI Guanine Adenine Fig. 1 Nucleobase pair recognition by Watson-Crick hydrogen bonding (Fig. 2).2 Peptide nucleic acid (PNA) represents a much more dramatic deviation from the natural DNA structure since the entire phosphodiester backbone has been replaced by a pseudopeptide backbone (Fig.3). Thus, from a chemical point of view PNA is a hybrid between an oligonucleotide (the nucleobases) and a ‘protein’ (the backbone) and consequently shows properties from both ‘worlds’. Gerald Haaima hails from the ‘Pride of the South’, Otago, New Zealand. He obtained his BSc with first class honours in 1985 and followed this up with a PhD (1 988) under the supervision of Rex Weavers. After a year with Gary Molander (Boulder, USA) and two years with Lew N. Mander (ANU, Australia) he re- located to Denmark and started work with the late Ole Buchardt and Peter Nielsen on Peptide Nucleic Acids. His research in- terests include nucleic acid rec- ognition and the development oj novel molecular recognition systems. Gerald Haaima Chemical Society Reviews, 1997 73 proteinI w mRNAI nucleus cvtovlasm I Fig.2 Principle in 'antisense' and 'antigene' strategies. In the antisense strategy, an oligonucleotide (analogue) binds to the mRNA by Watson- Crick hybridization and thereby inhibits the translation to the protein product. This can occur either by inducing RNAse H degradation of the mRNA or by physical blocking of the ribosomes. An antigene agent binds to the gene itself, the double stranded DNA either by triplex formation (oligonucleotides) or duplex invasion (PNA) and thereby inhibits the transcription of the gene to mRNA. PNA was originally designed as a reagent to sequence specifically target double stranded DNA as a mimic of triplex forming oligonucleotides which bind as a third strand in the major groove of a DNA double helix via T-A-T and C+.G-C Hoogsteen base pairing (Fig.4).3 However, the PNA backbone turned out to be a much better substitute for the normal sugar phosphate backbone than anticipated: and therefore much effort has been devoted to exploring the general DNA mimicking properties of PNA as well as its potential as an antisense and antigene drug, including being a sequence specific ligand for binding to double stranded DNA. The chemistry,5 physical chemistry6 as well as molecular biology/drug aspects of PNA7 have been presented quite extensively in the recent reviews. The present paper will focus on the 'structure-activity ' relationships of peptide-like back- bones in terms of DNA mimics.Briefly, PNA is composed of a backbone built up from aminoethylglycine units (a reduced dipeptide backbone) in which the nucleobase is attached to the central amine via an OH G4 ?0-y-0-ia1 -I,yG _.---:::-c Qo=p-o-/--A 4i, <AT----&H Fig. 3 Chemical structures of DNA and PNA. A, C, G and T designate the nucleobases adenine, cytosine, guanine or thymine. 74 Chemical Society Reviews, 1997 I T A H H ,', c I Watson-Crick G Fig. 4 T-A-T and C+.G-C triplets involving Hoogsteen and Watson-Crick base pairings. Note, that N3 of cytosine needs to be protonated in order to donate a hydrogen bond to the N7 of guanine. acetyl linker. This particular arrangement of atoms resembles the '6 + 3' number of bonds arrangement found for DNA.A simple geometry dissection of the DNA backbone reveals that six bonds separate each nucleobase unit, while the distance between the backbone and the nucleobase is three bonds. There have been a number of other approaches to the synthesis of PNA monomers*-10 since the first reports on PNA,"-l3 all of which essentially disconnect the molecule about the central amide bond presenting a synthesis of suitably protected nucleobase acetic acid and a protected backbone. The chemistry for these is well established in the literature and does not present new synthetic problems. More de novo approaches have been reported in which the monomer is built up from simple units during the solid phase oligomerization. This approach removes the need for monomer synthesis but its utility in producing high quality product has not been demonstrated.IO As with peptide synthesis, PNA monomers come in two major varieties, Boc and Fmoc each of which present their own possibilities and limitations.This has been reviewed recently,5 and by way of example a set of Boc-monomers are shown in Fig. 5. 2 PNA hybridization PNA oligomers bind strongly and with high sequence discrimi- nation to complementary oligomers of DNA, RNA or another PNA, and in general the hybrid thermal stabilities (T,) for identical sequences follow the order: PNA-PNA > PNA-RNA > PNA-DNA ( > RNA-DNA > DNA-DNA).4.14 Fur-thermore, the stabilities of PNA hybrids are, in contrast to hybrids between two anionic oligomers like DNA-DNA or RNA-DNA, fairly independent of ionic strength because of the neutral PNA backbone.lS It is also noteworthy that PNA hybrids can be formed both in the antiparallel (amino-terminal of PNA facing the 3'-end of the oligonucleotide) and the parallel configuration even though the antiparallel complexes have the higher stability (in general a AT, of 1-2 "C per base pair between antiparallel and parallel complexes are observed).4 Homopyrimidine PNAs distinguish themselves by forming PNA-NA-PNA triplexes of extremely high thermal stabil- ity.8716 Again, the charge neutral backbone of PNA can account for at least part of the triplex stabilization, but an X-ray crystal structure of a PNA2-DNA triplex shows specific interactions (hydrogen bonding) between each amide N-H of the backbone 0 Thymine Cytosine 8 COOH B0CN4-cmH BOCN€f-&V Adenine Guanine Fig.5 PNA monomers used for oligomerization using the Boc (tert-butoxycarbonyl) strategy of the Hoogsteen PNA strand and a phosphate oxygen of the DNA backbonel7 thereby also contributing to the stability. The high stability of PNA2-DNA triplexes also helps explain why hompyrimidine PNAs when targeted to double stranded DNA prefer not to form traditional PNA-DNA2 triplexes, but instead invade the DNA double helix forming an internal PNA2-DNA triplex (having conventional Hoogsteen and Watson-Crick nucleobase interactions j in a strand displace- ment c0mplexl8,~9 (Fig.6). This novel binding mode has opened a new avenue for the attempts to develop sequence specific dsDNA binding ligands, e.g. as gene therapeutic agents I,Ii Fig. 6 Schematic representation of a PNA-triplex strand displacement complex involving a PNA-DNA-PNA triplex via Watson-Crick and Hoogsteen hydrogen bonding (PNA is shown in heavier type than DNA) (antigene strategyj20.21 or as biomolecular tools in genome analyses.22 3 Biological aspects of PNA PNA has many of the properties of a promising antisense or antigene drug, such as stable and highly sequence specific binding to the complementary mRNA or dsDNA gene target, high biological and chemical stability.23 The easy synthetic accessibility and not least synthetic flexibility of PNA should also allow further optimization of the structure, especially with regard to bioavailability and pharmacokinetic properties.Thus it is not surprising that the drug aspects together with the utility as a molecular biology tool of PNA technology is being actively pursued, and the results so far are very encouraging7 4 PNA structure Four high resolution structures of PNA complexes are available at present. Two structures, a PNA-RNA24 and a PNA-DNA25 duplex, were determined by NMR methods, while a PNA2- DNA triplex17 and a PNA-PNA duplex26 were solved by X-ray crystallography. It is quite clear from these structures that the PNA oligomer is to some extent able to structurally adapt to the oligonucleotide complement, but it is equally clear that the PNA has a preferred structure of its own termed the 'P-f~m'.~~,~~ This is, of course, most obvious fromJhe structure of the pure PNA duplex which is a very wide (28 A diameter) helix with an accordingly large base pair helical displacement and a very large pitch (I 8 bp) [Fig.7(aj]. A canonical B-form helix which is typical for DNA duplexes has a diameter of ca. 20 8, and a pitch of ten base pairs per turn. The base pairs are perpendicular to the helix axis and stack through the centre of the helix. A canonical A-form Felix, typical of RNA duplexes, also has a diameter of ca. 20 A but a pitch of 11 base pairs per turn, and the base pairs are tilted ca. 20" relative to the helix axis.Furthermore, the base pairs are displaced away from the helix leaving a central 'tunnel' in the helix, analogous to that seen in the P-form. It is also noteworthy that the base pairs in the P-form are practically perpendicular to the helix axis (B-like) with only minor variations in slide, tilt and propeller twist angles between individual base pairs, and with an interbase-pair stacking overlap that is remarkably close to that found in canonical A-form RNA helices (Fig. 7). It is likewise notable that the Chemical Society Reviews, 1997 75 B-form P A-fom P-form Fig. 7 (a) Structure of a PNA duplex from X-ray crystallography. The structure was determined from a self-complementary hexamer26 but the full turn (18 bp) of the helix has been modelled from these data.Only the right- handed form is shown, but the unit cell contains both a right-handed and a left-handed form. (h)Base pair overlaps in A-, B- and P-form helices. Two consecutive A-T base pairs are shown as viewed from the end of the helix for a canonical B-form, a canonical A-form or for the crystal structure of the PNA-PNA duplex P-form. backbone structure found in the PNA duplex26 is almost identical to that seen in both the Watson-Crick and the Hoogsteen strand of the PNA2-DNA triplex17 and that the basic features, such as carbonyl orientations are also in common with the two PNA-oligonucleotide d~plexes.6,~~?~5 One conclusion to be drawn from the above described structural data is that despite the ability of PNA to efficiently bind and recognize DNA or RNA, the conformation adopted by PNA in these hybrid complexes is not optimal, because the P-form helix preferred by PNA is distinct in terms of important parameters such as helical width and pitch from both the B-form preferred by DNA and the A-form of RNA.Thus, one should be able to obtain a better peptide nucleic acid DNA mimic if one could construct a backbone that in its lowest energy state would adopt a B-form (or A-form) helix. However, the compactness and simplicity of the PNA structure pose severe restrictions as to which modification can be implemented and still result in chemically reasonable structures. Some of these possibilities are discussed below.5 PNA backbone modifications Since the first reports on PNA, a large number of PNA backbone derivatives have been described and investigated (Tables 1 and 2) in order to explore the ‘structural space’ in which this type of PNA mimic operates, as well as in an effort to obtain a molecular understanding of the chemical and structural parameters that determine the DNA mimicking properties of a (peptide) nucleic acid analogue. Thereby we should hopefully also gain a better understanding of the DNA (and RNA) molecules themselves. The results so far have shown (Tables 1 and 2) that only certain alterations of-or deviations from-the original ami- noethyl-glycine backbone are ‘allowed’ without severe penal- ties in DNA/RNA-PNA hybrid stability.As an overall conclusion at this stage one cannot touch the basic structure of the PNA backbone (l),e.g. by extending either of the ‘linkers’ [ethyl -+propyl (2), glycine -+(3-alanine (3) or acetyl -+ pro-pionyl (4)],27 reducing the nucleobase linker amide (5)2* or even reverse the amide linkage within the backbone (6).29 However, much freedom seems to be in placing (functional) substituents on the backbone as exemplified by exchanging the non-functional glycine for other natural amino acids30 (10-18, Table 2), although the type of substituent and also the stereochemistry at the now created chiral centre have different effects on the PNA hybridization properties. Even cyclic substituents, as exemplified by the ‘cyclohexyl substitution’ at the amino ethyl linker (8, 9) is possible provided the ‘right’ stereoisomer is chosen (Lagriffoule, Nielsen et al., in prep.). Naturally, the PNA analogues described to date have far from exhausted the imagination of chemists and more will un-doubtedly be investigated, now that ‘pure peptide chemistry’ has been introduced successfully in the ‘oligonucleotide analogue’ field.More bold chemists may even do away with both sugar-phosphate and peptide backbones and come up with totally novel constructions. 6 Why is PNA a good DNA mimic? One might ask: what features of the original PNA structure are responsible and required for its DNA mimicking properties and also what improvements might be possible? However, prior to engaging in this discussion, it is informative to make some thermodynamic considerations.Hybridization of com-plementary oligomers whether these being DNA, RNA or RNA is characterized by a large enthalpy gain and a significant entropy ~oss.~,~~The decrease in entropy upon hybrid forma- tion, naturally, is due to the formation of a highly ordered and fairly rigid duplex structure from two rather flexible and much less ordered single strands. (It should be kept in mind that an entropy gain which cannot compensate for the above loss is also accompanying hybrid formation due to release of ordered water molecules around the hydrophobic nucleobases). Therefore constraining the single stranded PNA (or other oligomer) in a conformation identical to or close to that found in the hybrid should greatly reduce the entropy loss and therefore increase the free energy upon binding.Thus restricting backbone flexibility, e.g.by introducing cyclic structures is an obvious strategy in the quest for oligomers of improved hybridization potency. This principle has been met with some success using bicyclic DNA analogues for triplex f~rmation,~~ and the idea was also the rationale for making the cyclohexyl derivatives 7 and 8 of PNA (Lagriffoule, Nielsen et al.) (Table 1). Disappointingly, neither the (SS)-nor the (RR)-isomer conferred improved hybridization although the (SS)-isomer had no serious adverse effect. Very interestingly, however, a thermodynamic analysis revealed that when compared to a normal PNA, a 10-mer PNA containing three (SS)-cyclohexyl units only showed an entropy loss (-AS) of 280 J mol-1 K-1 (versus 375 J mol-’ K-I for the normal PNA), whereas the enthalpy gain (-AH) was 127 kJ mol-l (versus 153 kJ mol-l). Therefore, one may conclude that the 76 Chemical Society Reviews, 1997 Table 1Effects on thermal stability (AT,,,/OC) per modified unit for structurally modified PNA T-monomers when incorporated into the oligomer sequence H-GTA GAT CAC T-NH2 AT, DNA/ AT,,, R" Entry Structure Backboneflinker "C "C Ref.Base 1 Eth ylgl ycine 0 0 H Base 2 Prop ylgl ycine -8.0 -6.5 27 3 Ethyl-6-alanine -10 -7.5 27 4 Propionyl linker -20 -16 27 Ethyl linker -22 -18 285 \OAN%NA, H Bqse 6 Retro-inverso -6.5 nd 29 7 (S,S)-Cyclohexyl -0.7 -0.5 Submitted 8 (Rp)-Cyclohexyl -8 -7 Submitted BaseJA0NH L-Omithine -14 -8 34,359 H structural constraints most likely has had the desired effect of Therefore apart from having the proper 'intra-backbone' producing a more ordered single strand state, but, unfortunately, distances (6 + 3), we believe that the constrained flexibility not in the optimal conformation(s) for DNA (or RNA) imposed by the two amide functions in the PNA backbone is hybridization.crucial. However, the poor DNA mimicking properties of the Although thermodynamic data are not yet available for other 'retro-inverso' PNA (6) which is a true isomer of the original PNA derivatives, it is obvious that in addition to any adverse aminoethylglycine PNA obeying both the '6 + 3-rule' as well as steric or structural constraints imposed by the changes of having the same number of constraining amide bonds, shows derivatives 1-5,these will all result in more flexible molecules.that more subtle factors such as dipole4ipole interactions and This is especially illustrative for derivative 5, in which all changes in hydration patterns that we do not fully understand distances are retained, but one of the amide bonds has been also make significant contributions. On the other hand, using reduced. the 'PNA system'4ue to the synthetic accessibility of a wide Chemical Society Reviews, 1997 77 variety of analogues--could help us further unravel the general principles for structure-activity relationships at the molecular level as well as improve our ability to translate chemical structures into three-dimensional structures.Table 2 Effects on thermal stability per monomer (ATmPC) for the PNA sequence H-GTA GAT CAC T-NH2a incorporating three chiral monomers as compared to an unmodified PNA30 Entry R Chirality ATm DNAPC ATm RNA/”C 10 CH3 L -1.8 nd 11 CH3 D -0.7 nd 12 Bus L -2.6 -3.0 13 CH2OH L -1.0 -1.0 14 CH20H D -0.6 -1.0 15 CH2C02H L -3.3 nd 16 CH2CH2C02H D -2.3 nd 17 (CH2)4NH2 L -1.0 -1.3 18 (CH2)4NH2 D +1.0 0 7 Outlook The results obtained with PNA have bearing on many areas of chemistry and biology ranging from basal molecular recogni- tion, self-assembly and chiral induction aspects14 over mole- cular biology tools and gene therapeutic drugs to our under- standing of the structure and function of Nature’s genetic material, DNA, and its possible prebiotic predecessors and origin.32 Even some novel materials have their origin in PNA.33 Therefore, PNA should not be viewed only as a DNA mimic, but as a structural and self-recognizing system in its own right, and we foresee that the properties of PNA and related compounds will prove of increasing interest and utility in both the traditional ‘oligonucleotide field’ as well as in other areas of science-including ones which at this stage are not imagined.8 Acknowledgements This work was supported by the Danish National Research Foundation. 9 References E. A. Wintner, M. M. Conn and J. Rebek, Jr., Acc.Chem. Res., 1994,27, 198. A. D. Mesmaeker, K.-H. Altmann, A. Waldner and S. Wendebom, Curr. Biol., 1995, 5, 343. P. E. Nielsen, M. Egholm, R. H. Berg and 0.Buchardt, Science, 1991, 254, 1497. M. Egholm, 0. Buchardt, L. Christensen, C. Behrens, S. M. Freier, D. A. Driver, R. H. Berg, S. K. Kim, B. NordCn and P. E. Nielsen, Nature, 1993, 365, 556. B. Hyrup and P. E. Nielsen, Bioorg. Biomed. Chem., 1996, 4, 5. 6 M. Eriksson and P. E. Nielsen, Quart. Rev. Biophys., in press. 7 H. J. Larsen and P. E. Nielsen, in Anulysis of antisense and related compounds, ed. A. S. Cohen and D. L. Smisek, CRC Press, in press. 8 S. A. Thomson, J. A. Josey, R. Cadilla, M. D. Gaul, F. C. Hassman, M. J. Luzzio, A. J. Pipe, K. L. Reed, D. J. Ricca, R. W. Wiethe and S.A. Noble, Tetrahedron, 1995, 51, 6179. 9 D. W. Will, G. Breipohl, D. Langner, J. Knolle and E. Uhlman, Tetrahedron Lett., 1995, 51, 12069. 10 S. Lutz and R. N. Zuckermann, Bioorg. Med. Chem. Lett., 1995, 5, 1159. 11 M. Egholm, 0. Buchardt, P. E. Nielsen and R. H. Berg, J. Am. Chem. SOC., 1992, 114, 1895. 12 K. L. Dueholm, M. Egholm, C. Behrens, L. Christensen, H. F. Hansen, T. Vulpius, K. Petersen, R. H. Berg, P. E. Nielsen and 0. Buchardt, J. Org. Chem., 1994, 59, 5767. 13 L. Christensen, R. Fitzpatrick, B. Gildea, K. H. Petersen, H. F. Hansen, T. Koch, M. Egholm, 0. Buchardt, P. E. Nielsen, J. Coull and R. H. Berg, J. Peptide Sci., 1995, 3, 175. 14 P. Wittung, P. E. Nielsen, 0. Buchardt, M. Egholm and B. NordCn, Nature, 1994,368,561.15 S. Tomac, M. Sarkar, T. Ratilainen, P. Wittung, P. E. Nielsen, B. NordCn and A. Graslund, J. Am. Chem. SOC.,1996,118,5544. 16 S. K. Kim, P. E. Nielsen, M. Egholm, 0. Buchardt, R. H. Berg and B. NordCn, J. Am. Chem. SOC.,1993, 115,6477. 17 L. Betts, J. A. Josey, J. M. Veal and S. R. Jordan, Science, 1995,270, 1838. 18 D. Y. Chemy, B. P. Belotserkovskii, M. D. Frank-Kamenetskii, M. Egholm, 0. Buchardt, R. H. Berg and P. E. Nielsen, Proc. Natl. Acad. Sci. USA, 1993,90, 1667. 19 P. E. Nielsen, M. Egholm and 0. Buchardt, J. Mol. Recog., 1994, 7, 165. 20 J. C. Hanvey, N. C. Peffer, J. E. Bisi, S. A. Thomson, R. Cadilla, J. A. Josey, D. J. Ricca, C. F. Hassman, M. A. Bonham, K. G. Au, S. G. Carter, D. A. Bruckenstein, A. L. Boyd, S.A. Noble and L. E. Babiss, Science, 1992, 258, 1481. 21 P. E. Nielsen, M. Egholm and 0. Buchardt, Gene, 1994,149, 139. 22 A. G. Veselkov, V. V. Demidov, P. E. Nielsen and M. Frank- Kamenetskii, Nucl. Acids Res., 24, 2483. 23 V. Demidov, V. Potaman, M. D. Frank-Kamenetskii, 0. Buchardt, M. Egholm and P. E. Nielsen, Biochem. Pharmacol., 1994,48, 1309. 24 S. C. Brown, S. A. Thomson, J. M. Veal and D. G. Davis, Science, 1994, 265, 777. 25 M. Eriksson and P. E. Nielsen, Nature Struct. Bzol., 1996, 3,410. 26 H. Rasmussen, J. S. Kastrup, J. N. Nielsen, J. M. Nielsen and P. E. Nielsen, Nature Struct. Biol., 1997, 4, 98, in press. 27 B. Hyrup, M. Egholm, P. E. Nielsen, P. Wittung, P. B. Nordkn and 0.Buchardt, J. Am. Chem. SOC., 1994,116,7964. 28 B. Hyrup, M. Egholm, 0. Buchardt and P. E. Nielsen, Bioorg. Med. Chem. Lett., 1996, 6, 1083. 29 A. H. Krotz, 0.Buchardt and P. E. Nielsen, Tetrahedron Lett., 1995,36, 6941. 30 G. Haaima, A. Lohse, 0. Buchardt and P. E. Nielsen, Angew. Chem., 1996,35, 1939. 31 R. Jones, S. Swaminathan, J. F. Milligan, S. Wadwani, B. C. Froehler and M. D. Matteucci, J. Am. Chem. Soc., 1993,115,9816. 32 C. Bohler, P. E. Nielsen and L. E. Orgel, Nature, 1995,376, 578. 33 R. H. Berg, S. Hvilsted and P. S. Ramanujam, Nature, 1996, 383, 505. 34 K. H, Petersen and P. E. Nielsen, Bzoorg. Med. Chem. Lett., 1996, 6, 793. 35 E. Lioy and H. Kessler, Liebig Ann., 1996, 201. Received, 7th November I996 Accepted, 17th December I996 78 Chemical Society Reviews, 1997

ISSN:0306-0012    

DOI:10.1039/CS9972600073

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Trends in isothermal microcalorimetry 1::‘d) ,l,e,._.---_-__-. A Ingemar Wadso Thermochemistry, Chemical Center, Lund University, PO Box 124, S-22 I00 Lund, Sweden Isothermal microcalorimeters are of increasing importance in thermodynamics and as general ‘process monitors’. Recent developments in instrumentation and in experi- mental methods have been significant and several easy-to- use instruments are now commercially available. Important application areas include investigations of solute-solvent interactions and ligand binding processes, sorption pro- cesses, living cellular systems and the assessment of stabil-ities of technical products. The combination of isothermal calorimetry with different specific analytical techniques seems to be particularly promising.1 Introduction All calorimeters are thermodynamic instruments but some are also used in kinetics or as analytical tools. Differential (temperature) scanning calorimetry (DSC) has for a long time been one of the most important techniques in thermal analysis and more recently ‘isothermal microcalorimeters’ are gaining an increasing importance as analytical instruments, in particular in some applied areas. The term ‘isothermal microcalorimeter’ is not well defined, but is now commonly used for calorimeters designed for work in the microwatt range conducted under (essentially) isothermal conditions. ‘Nanocalorimeters’, the name of which usually indicates a detection limit approaching one nanowatt, are here included in the group of ‘microcalorimeters’.Recent develop- ments in isothermal microcalorimetry have been substantial and several easy-to-use instruments are now commercially availa- ble. Some of them have the character of modular systems, which allow several general and specialised measurement functions. When complex processes are characterised by calorimetric measurements, for example in technical products or in living materials, it may not be possible to express the results in terms of thermodynamic or kinetic quantities referring to well-defined reaction steps. In such cases, isothermal microcalorimeters have found important but so far limited use as general ‘process monitors’. Ingemar Wadso is Professor Emeritus in Thermochemistry at the Chemical Center, Lund University, where he has spent most of his career.His thermochemical research has centred around in-strumental developments, and investigations of solute-water interactions and of living cel- lular systems. He has been ac- tive in IUPAC. He received the Arrhenius plaquette, I970 (Swedish Society for Chemists), The HufSman Memorial Award, I979 (Calorimetry Conference), The Award for Applied Chem- ical Thermodynamics, 1984 (Swiss Society for Thermal Anal- ysis) and the Lavoisier Medal, I990 (International Society for Biological Calorimetry). He was awarded the MD honoris causa by Lund University in 1992. This review will discuss some properties and current uses of isothermal microcalorimeters.A special focus will be on microcalorimeters used as process monitors in applied areas and on developments of instrument assemblies where specific analytical measurements are conducted in parallel with the calorimetric experiments. 2 Some measurement principles and design characteristics From the point of view of heat measurement principles it is common to divide calorimeters into three main groups: adiabatic, heat conduction and power compensation cal-orimeters. 2.1 Adiabatic calorimeters In an ideal adiabatic calorimeter there is no heat exchange between the calorimetric vessel and its surroundings. Adiabatic conditions are usually obtained by placing an ‘adiabatic shield’ between the vessel and the surroundings. During a measurement the temperature difference between the vessel and the shield is kept at zero.The heat quantity which is evolved or absorbed during an experiment with an ideal adiabatic calorimeter is equal to the product between the temperature change and the heat capacity of the calorimetric vessel, including its content. Semi-adiabatic calorimeters, often called isoperibol calorime- ters, are more commonly used than the (close to) ideal adiabatic instruments. When semi-adiabatic instruments are used in accurate measurements it is necessary to apply corrections for the heat exchange between the vessel and the surroundings. 2.2 Heat conduction calorimeters In a heat conduction calorimeter heat released (or absorbed) in the reaction vessel is allowed to flow to (or from) a surrounding heat sink, usually an aluminum block.Normally, a thermopile positioned between the sample container and the heat sink is used as a sensor for the heat flow. Its driving force, i.e. the temperature difference between the vessel and the heat sink, will give rise to an electrical potential, U, over the thermo- pile. Provided that the temperature is uniform in the vessel and in the heat sink, the thermal power (the rate of heat production) released in the vessel, is given by the Tian equation [eqn. (I)]: P = E (U +.tdU/dt) (1) where P = dq/dt is the thermal power, E is the calibration constant, IC the time constant of the instrument and dU/dt the time derivative of the thermopile potential. For a steady-state process eqn.(1) is reduced to eqn. (2). P=Eu (2) For any process the heat quantity released in the vessel is given by the potential time integral eqn. (3). q = EjUdt (3) (The initial and final thermopile potentials are assumed to be the same). Heat conduction microcalorimeters are usually equipped with semi-conducting thermopiles, often called ‘thermocouple Chemical Society Reviews, 1997 79 plates' or 'Peltier effect plates' as sensors for the heat flow. They have a relatively large thermal conductance and the temperature difference between the microcalorimetric vessel and the heat sink is small, typically in the order of one mK.1 Such instruments can therefore normally be considered as isothermal calorimeters. For processes which are slow on a timescale given by the time constant (typically, the order of a few minutes) the simple eqn. (2) may describe the rate of heat evolution, and thus the kinetics of the process, with an adequate precision.However, in order to obtain accurate rate values for fast processes it is necessary to apply Tian's eqn. (l), or one of its more advanced forms.' With modern commercial microcalorimeters such 'dynamic correc- tions' can be made automatically. However, reported in literature it is common to find thermal power-time curves derived using the simple eqn. (2), in cases where the time constant term in eqn. (1) is clearly significant. For more detailed discussions of properties of heat conduction microcalorimeters, see ref.1 and 3. Most isothermal microcalorimeters in current use are of the heat conduction type, for example the microcalorimetric systems marketed by CSC (earlier Hart Scientific) (USA), Setaram (France) and Thermometric (Sweden). In contrast, Microcal's (USA) titration microcalorimeter uses the same principle as an adiabatic shield DSC. The temperature is allowed to increase, very slowly, during an experiment and the heat evolution from a reaction is balanced by a corresponding change in the heating rate (power compensation, see below). 2.3 Power compensation calorimeters In a power compensation calorimeter the thermal power from an exothermic process is balanced by a cooling power, in microcalorimetry normally by use of Peltier effect cooling.For endothermic processes, compensation can be achieved by reversing the Peltier effect current or by use of an electrical heater. 2.4 Some design features Isothermal inicrocalorimeters form a heterogenous group of instruments and many different designs have been described. Regardless of the calorimetric principle used, most microcal- orimeters are designed as twin or differential instruments. The 'reaction vessel', which is used for the investigated process, and the 'reference vessel', which is charged with an inert material, should preferably be nearly identical, in particular with respect to heat capacity and thermal conductance. In some cases microcalorimetric reaction vessels are taken out from the calorimeter at cleaning and charging operations ('insertion vessels'), alternatively vessels are permanently mounted in the heat sensitive zone of the calorimeter. Measurements can be conducted as batch experiments, with or without agitation (stirring) of the content, or as continuous or stopped flow experiments. Ongoing processes, for example in living materials, and slow degradation and relaxation processes in materials and products of technical importance are often measured using simple sealed ampoules as reaction vessels.Injection techniques are usually employed when liquid or gaseous reactants are used to initiate a batch process. Such methods are particularly well suited to automation, for example in titration experiments. Several microcalorimetric designs have been reported where processes are initiated by bringing reagents together in a flow mixing vessel or in a stirred perfusion vessel.The mixing of a reagent with a heterogenous system (e.g. a suspension of solid particles, which tend to sediment) can be difficult to achieve without causing large heat effects. In such cases rotating or rocking calorimeters using bi-compartment vessels can be the best choice. Electrodes and other analytical sensors can be positioned in the reaction vessels and light can be introduced by use of light guides. Designs and properties of different kinds of isothermal microcalorimeters have been reviewed.24 80 Chemical Society Reviews, 1997 2.5 Direct and indirect determination of enthalpy changes Recent developments in isothermal microcalorimetry have opened several areas for direct calorimetric measurements where earlier indirect methods had to be used. Enthalpy values for well defined chemical processes can be derived from values for equlibrium constants determined as function of temperature ['van't Hoff enthalpies', Table 1, eqn.(4)]. However, such Table 1 Some basic thermodynamic relationships d In K AHo -(van't Hoff equation) (4)dT RT2 AG" = -RT 1nK (5) AG" = AH" -TAS" (6) AC," = dAH"/dT (7) AtransHm = AsolHm(2)-Aso~H~(l) (8) AsolvH = AsolH -AvapHo (9) Cp,2m= AsolCpm+ Cp* (10) Symbols: R = the gas constant (8.314 J K-1 mol-1). K = equilibrium constant. T = temperature (in kelvin). AGO, AH", AS" and AC," = standard changes in Gibbs energy, enthalpy, entropy and heat capacity, respectively.Cp,2m= partial molar heat capacity at infinite dilution. C,* = heat capacity for a pure compound. Subscripts: p, trans, sol, solv, vap = constant pressure, transfer, (dis)solu- tion solvation and vaporization, respectively. values usually have a low accuracy and changes in heat capacity, AC,", derived from van't Hoff enthalpies are in most cases only marginally useful.5 Table 2 gives a summary of expected statistical uncertainties for AH " and AC," values derived from equilibrium constants of different precision determined over different temperature ranges. It is seen that in order to obtain precise values for AH", and in particular for AC," very high precision is required in determination of equilibrium constants.Further, many determinations over a wide temperature range must be made-conditions which are rarely met for van? Hoff values reported in literature. Table 2 Propagation of errors in log K in calculation of van? Hoff AHo and AC," values at 25 "C for a 1 :1 binding reaction (5SD). From King6 f SD in fSD in 5 SD in log K Experimental temperaturesPC AHo/kJ mol-1 AC,"I J K-1 mol-1 0.02 20, 22, 24, 26, 4 2800 28,30 0.001 5, 10, 15, 20, 0.04 6 25, 30, 35, 40, 45,50 Energy-material balances of living systems are often esti- mated by 'indirect calorimetry',7 meaning that values for heat production are derived from analytical values for substances consumed and produced during metabolism (usually only the respiratory gases).Microcalorimetric techniques are now available by which thermal power values can be determined accurately for small samples of living cells and tissues,*J under well defined physiological conditions. 2.6 Process monitoring The very broad application range for non-specific methods like calorimetry can be attractive both in thermodynamic measure- ments and in analytical work. As practically all processes are accompanied by heat effects, calorimetry is particularly well suited to the discovery of unexpected or unknown processes in samples of any aggregation state. Further, in contrast to spectroscopic methods, calorimetry does not require optically clear objects. In particular when heat conduction calorimeters are used, the experiments can be conducted over long periods of time-weeks or longer.These properties can make isothermal microcalorimeters ideal as monitors for slow and complex processes, not the least for solids where chemical and physical processes can be difficult to record continously without interfering with the processes. However, the lack of specificity in heat measurements will also lead to serious limitations for such methods, cf Section 5. It is important to keep in mind that a calorimetric experiment will lead to thermodynamic data, even if the instrument is employed only for analytical purposes. Derived thermodynamic values can sometimes be compared with values estimated from results of chemical analyses combined with thermodynamic data from compilations.It is therefore often important also to be concerned about the accuracy of the calorimetric results, when the instrument is used as a 'process monitor' 3 Important applications Isothermal microcalorimetry is used in a wide range of applications and a comprehensive review cannot be given here. In this section some comments are made on current activities in a few important areas. 3.1 Ligand binding and aggregation processes in solution One of the main applications for isothermal microcalorimetry is the investigation of non-covalent binding processes by means of titration techniques, sometimes referred to as ITC (isothermal titration calorimetry). In such experiments the titrant solution is normally injected stepwise into a stirred reaction vessel, volume typically ca.1 ml. The interpretation of calorimetric titration experiments is based on the assumption that the heat quantities accompanying the injections, corrected for dilution effects, are proportional to the amount of titrant reacted. If the concentration equilibrium constant, K,, is very high (for example, K, > 1 X lo8 for a 1 :1 binding reaction) there will be nearly zero concentration of free titrant following each injection step, until the equivalence point has been reached. Such experiments will lead to information about the stoichiometry of the process and will lead to a value for the molar enthalpy change, but K, values cannot be derived.For processes with moderately high K, values a significant fraction of the titrant will not be consumed and this fraction will increase as the injections continue. The fractions of nonreacted titrant depend on the equilibrium constant for the process and a certain binding model must be predicted in order to derive values for AH" and K,. Results of the calorimetric measure- ments are fit to the assumed binding model using K, and AH" as fitting parameters5 Commercial titration microcalorimeters are now delivered with computer programs for such fitting procedures. For a correct binding model small and random least-squares residuals are obtained as a result of the minimization calcula- tion. Such results will support, but will not prove the correctness of a certain model or combination of models. In particular for processes which appear to be more complex than a 1:1 binding reaction it is desirable that the stoichiometry is also supported by results of specific chemical analyses.Values for the standard Gibbs energy, AG ", and corresponding entropy change, AS O, are calculated from eqns. (5) and (6), respectively. From experiments conducted at different temperatures a value for the heat capacity change, AC,", can be derived, eqn. (7). When a 1 : 1 complex is too strong to allow the determination of K, it is sometimes possible to employ a 'displacement' titration technique in which the binding reaction is divided into two steps, where each has a K, value which is sufficiently small to be e~aluated.~ The time constant for a heat conduction calorimeter is larger than for comparable adiabatic and power compensation calo- rimeters, often 2-3 min when 1 ml vessels are used.' It will then take ca.25 min until the heat released in a fast reaction has been conducted to the heat sink and the measurement time for a binding experiment with 15 titration steps will thus be > 6 h. However, 'dynamic correction' techniques1 can decrease the time required for a titration experiment to the same level as found with adiabatic and power compensation instruments, without any loss of accuracy. As an example, Fig. 1 shows a record from a protein ligand binding experiment conducted with a heat conduction microcalorimeter used with a dynamic correction technique.15 30 45 t / min Fig. 1 Stepwise titration of ribonuclease A (ca. 60 pmol) by 2'-cytidine mo- nophosphate using a heat conduction titration microcalorimeter (stainless steel reaction vessel, volume 2 ml, T = 210 s). A 'dynamic correction' technique was employed and the time between injections was reduced to 4 min. Two curves are shown: the one with low and rounded peaks is the experimental curve, the curve with sharp peaks is the corrected curve. (Courtesy of Thermometric AB). Specific binding reactions between biopolymers and low molecular mass compounds are now investigated in many laboratories by use of titration microcalorimetry. Other equilib- rium reactions studied include aggregation between bio-polymers and interactions between proteins and membrane receptors.In the pharmaceutical industry the binding of drugs and related compounds to biopolymers is currently developing as an important part of techniques used in 'rational drug design'.* Specific binding processes involving macrocyclic compounds, in particular cyclodextrins,5 have been much studied. Investigations of micelle formation by detergents, phospholipids and other amphiphilic molecules is another field where titration microcalorimetry is frequently used. These latter experiments are normally conducted by stepwise injection of amphiphile solutions, at concentrations higher than their critical micelle concentration (cmc), to the reaction vessel which is initially charged with pure solvent.9 The heat quantities measured will thus refer to the deaggregation process until the solute in the reaction vessel has reached a concentration above its cmc.Values for cmc and (after correction for dilution effects) the enthalpy of micelle formation can thus be obtained. Similarly, the thermodynamic properties at pairwise inter- actions can also be derived.I0 3.2 Dissolution and mixing processes Results from calorimetric measurements of enthalpy of dissolu- tion of pure substances (gases, liquids and solids) are essential for our understanding of the thermodynamics of processes in solution.3.11 For aqueous solution systems, in particular, values for their temperature derivatives, i.e. corresponding heat capacity values, are of major importance.Detailed studies of solute-solvent interactions can be made by determination of enthalpies of dissolution, AsolH,preferably Chemical Society Reviews, 1997 81 at different temperatures leading to corresponding heat capacity values. In microcalorimetric dissolution experiments the con- centration of the solutes are often low enough to regard the solutions as infinitely dilute. The difference between dissolu- tion enthalpies for a compound in different solvents will give the enthalpy of transfer for the compound between the solvents, eqn. (8). Similarly, the difference between AsolHmand the enthalpy of vapourization A,,&O (the ideal gas phase value), will give the enthalpy change for the transfer of the compound between gas phase and the solution, often called the enthalpy of solvation, AsolvHm, eqn.(9). The functions A,,,,H and AsolvH and corresponding ACPm values reflect changes in solute-solvent interactions, which are free from contributions from interactions between the solute molecules in solution or in their pure form. The same applies to the partial molar heat capacity of solutes at infinite dilution, Cp,2=,which can be derived from eqn. (10). Development work in dissolution microcalorimetry3~~ has* resulted in several instruments for dissolution of slightly and easily soluble compounds (gases, liquids and solids) into water and other solvents. Results have been reported for many slightly soluble compounds in water, for example, the rare gases, low molecular mass gaseous and liquid hydrocarbons, several other liquid hydrophobic molecules and a few slightly soluble solid compounds.The development of microcalorimetric dissolution techniques has been important, in particular for investigations of biochemical model systems and the hydrophobic effect. However, at present there is little fundamental work conducted in the field of dissolution microcalorimetry, presumably due to the lack of commercial instruments. In addition to its use in studies of solute-solvent interactions, it is likely that dissolution microcalorimetry will become important for the characterization of the state of solids and liquids. For example, dissolution microcalorimetry has been used to characterize solid materials with respect to different polymorphic forms,12 an area very important to the phar- maceutical industry.Enthalpies of mixing of organic liquids, of theoretical and practical importance, are often determined by use of flow microcalorimeters~~4thus avoiding a gas phase in the calorime- tric vessel, which may cause evaporation or condensation effects. In addition, flow calorimetric measurements can easily be automated. 3.3 Sorption processes Enthalpy measurements of sorption (adsorption or absorption) of solutes on small solid particles and on fibres can often be conducted using titration microcalorimeters. The solid material is then (more or less) suspended in the liquid and can be titrated by the solution.For larger pieces of solid material it may be appropriate to use, for example, a rotating sample holder to bring the material into intimate contact with the solution. One area of current practical importance in this field is the thermodynamic characterization of the binding of amphiphile molecules (detergents) to mineral particles, in connection with oil recovery techniquies. The sorption of vapour (especially water vapour) by materials like foodstuff, fibres, pharmaceuticals and building materials is of significant practical importance and vapour sorption equilib- rium curves (sorption isotherms), are often determined in industrial laboratories. It was recently demonstrated in two pharmaceutical laboratories1*,13 that new and valuable informa- tion can be obtained from very simple sorption experiments conducted by isothermal microcalorimetry.An open tube containing a saturated salt solution is placed in the micro- calorimetric vessel which is charged with the sample, Fig. 2. The atmosphere above the salt solution has a well defined relative humidity and vapour will gradually be adsorbed by the sample until equilibrium is reached. The calorimeter will thus continuously measure the sum of an endothermic vapourization process which is nearly balanced by the exothermic vapour sorption on the sample. In some cases the sorption process will initiate other physical or chemical changes in the material, which may be correlated with properties of technical im- portance.15 Using a more sophisticated technique16 a flow of carrier gas, with a predetermined and variable concentration of vapour, is allowed to pass through the sample container of a calorimetric vessel.Such measurements will lead to well-defined sorption enthalpies. Fig. 2 Measurement of vapour sorption using a simple ‘microhygros- tat’.’3J4 A microcalorimetric vessel (c)is charged with a solid sample (b) and an open tube with a saturated salt solution (a)is inserted in the vessel immediately before the experiment is started. It was recently shownl7 that the sorption enthalpy and the sorption isotherm can be determined simultaneously by use of a double twin microcalorimeter, Fig. 3. The upper calorimetric vessel serves as a vapourization chamber for a vapour forming liquid and the lower vessel, which is charged with the sample, is the sorption chamber.Vapour will diffuse to the sorption vessel and the enthalpy change for the sorption process is determined as a function of time. From the rate determined for the vapourization process it is possible to calculate the sorption isothenn. (Values for A,,&, the dimension of the connecting tube [Fig. 3(e)]and the diffusion coefficient of the vapour must be known. It is assumed that equilibrium conditions prevail in the sorption vessel.) It is judged that the technically very important field of vapour sorption microcalorimetry will continue to develop over the next few years. 3.4 Vapourization and sublimation processes There is a strong need for enthalpy of sublimation data for substances with very low vapour pressure, for example in connection with investigations of biothermodynamic model systems.However, very little development work and few measurements have been reported during the last few decades in vapourization/sublimation microcalorimetry. More advanced microcalorimetric techniques are much needed in this field. 3.5 Curing and degradation processes Microcalorimeters have for a long time been used in develop- ment work and as control instruments in the cement industry. Cement hydration, polymerization processes and other indus- trial curing processes are accompanied by the release of large quantities of heat and sensitive calorimeters are rarely needed in such work.But microcalorimeters, especially of the heat conduction type, are usually well suited for measurements of processes releasing thermal powers several orders of magnitude larger then the detection limit for the instrument. For observa- tion of rates of heat production during post curing processes and during physical and chemical aging of the products, micro- calorimeters are needed. The estimation of the ‘shelf-life’ of a chemical product, in some cases including hazard evaluation, is often a critically important property which is carefully evaluated in product development work and in quality control. Stability tests by various physical and chemical methods should preferably be conducted at a temperature close to the normal storage 82 Chemical Society Reviews, 1997 Fig.3 Simplified picture of a vapourization-sorption microcalorimeter for simultaneous determination of sorption isotherms and enthalpy of sorp-tion.17 (a) Vapour forming liquid, (b) solid sample, (c) vapourization chamber, (d)thermopile, (e) connection tube, (f> sorption chamber. temperature, but such assessments will usually require very long observation periods-months or years. It is therefore common to perform accelerated tests by performing the measurements at different elevated temperatures, for example at 150-200 OC, and estimate a value for the rate of degradation at the storage temperature by use of Arrhenius' relationship. Such accelerated tests are useful but may lead to erroneous conclu- sions as the nature of a complex degradation process can change significantly over a large temperature range.With respect to thermally induced run-away reactions, calorimetry is obviously the most fundamental measurement method. However, in some branches of the chemical industry isothermal microcalorimetry is also used more generally in the study and control of slow degradation processes, including those caused by incompatibility between different materials. Measurements are then conducted directly at the storage temperature of the products or at slightly increased tem-peratures, for example at 40-70 OC, from which reasonably safe extrapolations can be made to ambient temperatures. Several microcalorimeters primarily designed for kinetic investigations of relatively fast processes (halflife < 1 s) have been reported, but they have not reached any wider use.Calorimeters are inherently more suitable for accurate kinetic investigations of slow processes, for example many curing and degradation processes. Willson et a1.18 have developed a method which permits a direct analysis of calorimetric results in terms of reaction order, rate constant and enthalpy of reaction, with no prior assumptions concerning reaction mechanism etc. In case the reaction mechanism changes during the course of the experiment, parameters for each contributing process can be evaluated. The amount of heat evolved from a chemical process depends on the (specific or molar) enthalpy change, which will lead to very different detection limits for different processes.For example, oxidation of organic material might be accompanied by a heat evolution of 600 kJ mol-l of 02 whereas enthalpy changes for an ester hydrolysis are often close to zero. Table 3 shows the limits of detectability estimated for an isothermal microcalorimeter of the heat conduction type used in measure- ments of some hypothetical degradation processes. The detec- tion limit for the instrument is taken as 0.1 pW, which may not seem as an impressive value by today's standards. However, the useful detection limit in this type of experiment normally depends on the value of the irreproducibility of the instrument baseline (zero thermal power), which is determined in separate experiments where the reaction vessel is chargedhecharged with an inert material.The variability in that value is typically significantly larger than, for example, the detection limit for deviations of the thermopile signal during a baseline experi- ment. It is values of this latter kind which are normally reported in papers describing properties of microcalorimeters and in printed materials from manufacturers. Table 3 Limits of detectability in the study of some hypothetical decomposition processes using a heat conduction microcalorimeter. The instrument detection limit is assumed to be 0.1 pW and the amount of sample 3 g Detection limit, expressed as % Molecular mass of substance AH/ kJ mol-' decomposition per year ~~ 100 10 1 50 0.2 500 0.02 10000 10 100 50 20 500 2 Many isothermal microcalorimeters presently employed in industrial laboratories are used for stability measurements, mainly in the explosives industry, at military arsenals and in the pharmaceutical industry.Compatibility experiments are im- portant and are easy to perform. Two components of a product are brought into contact and the thermal power-time curve for the mixed material is measured. The result is compared with the theoretical curve (perfect compatibility) constructed from results of separate measurements of the two components. 3.6 Living materials A large part of the method work conducted in isothermal microcalorimetry has been concerned with instruments and working procedures for measurements of living cellular sys- tems.Much of that work has been directed towards use in applied areas: clinical analysis, pharmacology, ecology, bio- technology and agriculture. Results of many 'demonstration experiments' and of some fundamental investigations in cell physiology have been reported but, so far, no application area of real practical importance has been established. 19 Isothermal microcalorimetry is currently used in the study of micro- organisms, human and animal cellular systems, small animals and on material from plants; representative examples of reviews and reports from work in these areas have been published in Special Issues of Thermochim. Acta.2s22 3.6.1 Microorganisms In calorimetric experiments conducted on microbial systems it is common that comparatively large thermal powers are released and the detectability limit of the instruments is rarely a problem.A more important problem is often that cultures may turn anoxic due to limitations in the supply of oxygen.',* Flow calorimetric techniques are often used. The suspension of microorganisms can then be pumped from a conventional fermenter where the properties of the culture are accurately controlled. Yeast cells are heavy which will often cause problems as they may sediment in calorimetric vessels and in flow lines. Microcalorimetric investigations on microbial systems have been conducted at the fundamental level and in several applied areas. For example, the effect of drugs has been studied extensively, in order to explore the possible practical use of the technique in areas such as clinical analysis and pharmacology.The results of method investigations on microbial systems in Chemical Society Reviews, 1997 83 soil, sludge and waste water systems are judged to be important in the characterization of ecological systems. But the identifica- tion of microorganisms by use of their thermal power-time curves, which at one stage appeared promising, does not now seem to be reali~tic.~3 3.6.2 Mammalian cells and tissues Much progress has been made during the past decade in the development of microcalorimetric techniques for the investi- gation of mammalian cells and tissues. In particular, many measurements have been made on the main fractions of human blood cells from healthy subjects and from patients.Investiga- tions have also been made on, for example, macrophages, adipocytes and fat tissue, hepatocytes, muscle fibre bundles, sperm cells, cancer cells and many types of cultured tissue cells. Most of these studies have been part of exploratory studies aimed at method developments in clinical analysis. Investiga- tions at the thermodynamic and physiological level have also been conducted.24 The thermal power (P) of many types of human cells, obtained from groups of healthy individuals and measured under standardized conditions, has been shown to have characteristic average values. As expected, the biological variation is much larger then the precision of the measurements and such standard P values show a large spread.Standard P values for several groups of patients are significantly different from the corresponding values for the healthy groups. Such differences are of fundamental medical interest, but standard P values for the patient groups typically have an even larger spread than the values from the healthy groups and AP values will only have a limited diagnostic value. It seems more realistic that isothermal microcalorimetry may become of practical importance as a prognostic tool in recording changes in P values following medical treatment or the progress of a disease. As an example, see the extensive studies by Monti et al.25 on non- Hodgkin’s lymphoma (NHL) patients.A large fraction of the studies on mammalian cells and tissues have been conducted using simple static reaction vessels. Such experiments are easy to perform but for most cell types sedimentation will take place, or at least the cell concentration will be different in different parts of the vessel. This may lead to a poorly defined concentration of medium constituents, including the concentration of oxygen and the pH. Further, such vessels do not allow addition of reagents (e.g. drugs) during a measurement. Suitable injection (titration) vessels allowing gentle but efficient stirring and injections of reagents,2J are now available. Such vessels can also be fitted with different holders for tissue pieces making them well exposed to the medium.Alternatively, such experiments can be conducted with similar vessels operated as perfusion vessels through which fresh medium is continuously perfused. The vessels are also suitable for measurements of cells attached to a film or to micro- carriers.2.3 Stirred injection vessels have been equipped with electrodes for measurement of 02 concentration and pH.26 Flow microcalorimetric techniques are of less importance in measure- ments of mammalian cells than in work on microorganisms. 3.6.3 Small animals Several microcalorimetric investigations have been conducted on small aquatic animals, often in connection with studies of anoxia. Insects can often easily be accomodated in simple microcalorimetric vessels but few investigations have been reported.Animals such as frogs and lizards have been investigated in large (100 ml) microcalorimetric vessels. 3.6.4 Materials from plants Very few microcalorimetric studies on living material from plants were reported until about ten years ago. Most work within this area has been conducted on tissue pieces (leaves, stems and roots) and has often concerned the effect of stress factors such as high or low temperature, excessive salt concentration and various pollutants. Correlations have been investigated between thermal power produced by tissues from trees and their biomass production, an area which may develop to one of practical importance? Few studies on plant cells have been reported. Surprisingly, photo-microcalorimetric measurements con-ducted on material from plants are very rare.4 A note about systematic errors and standards It was pointed out earlier that practically all processes are accompanied by heat effects. That property can be useful when calorimeters are used as general analytical tools but it will also make calorimeters uniquely sensitive to systematic errors. As may be expected, such errors tend to be more difficult to control in microcalorimetric experiments than in cases where larger thermal powers or heat quantities are measured. Modern isothermal microcalorimeters are usually very easy to work with, but errors from effects such as evaporation, condensation, adsorption and incomplete mixing may easily ruin the result. In work with living cellular sytems there are several additional important sources of errors of which the experimentalist should be aware.28 The calibration constant for an isothermal microcalorimeter is normally determined by release of electrical power, which can easily be measured with a higher accuracy than needed in any microcalorimetric experiment.However, the heat flow pattern from the electrical heater can be significantly different from that released by an investigated process. For flow-through vessels systematic errors in the electrical calibration value can be particularly serious.3.28 When the aim is to conduct accurate microcalorimetric measurements it is important to test the performance of the instrument-and of its user-by means of test reactions.3.28 5 Increasing the specificity It was pointed out earlier that the non-specific nature of heat measurements can make it difficult to interpret results from measurements of complex reaction systems in sufficient detail.This is believed to be the main reason why isothermal microcalorimeters have not been much used as process monitors, except in a few niches of the chemical industry and by some academic research groups. The thermal power-time curve for a process represents a complete record of the process, but it may not be possible to interpret the record at the molecular level. Calorimetric records would be more informative if they were supported by results of specific analyses, if possible performed at the molecular level.In flow calorimetric experi- ments it is easy to take out samples for analysis from the calorimetric flow line. In batch experiments it is possible to extract liquid or gaseous samples from the reaction vessel without causing uncontrolled disturbances, even if the experi- ments are conducted on the microwatt level. However, analyses should preferably be performed continuously and in parallel with the calorimetric measurements. In a few cases such combined measurements have been explored, without receiving much attention. 5.1 Analytical sensors in the reaction vessel Several types of analytical sensors can be positioned in microcalorimetric vessels. Different electrodes and probes for measurement of spectral changes seem to offer interesting possibilities.5.1.I Electrodes It has been shown that miniaturized electrodes can be positioned in a microcalorimetric vessel and be used in parallel with the heat mesurements without causing any significant distur-bances.26 Fig. 4shows results from an experiment with cultured tissue cells in a reaction vessel equipped with pH and 02 electrodes. Following an equilibration period, a small sample of cells was injected at the time indicated by the arrow. The 84 Chemical Society Reviews, 1997 I I I II I I 0 10 20 30 tlh 10 20 30 tlh Fig. 4 Results from measurement of metabolic heat from T-lymphoma cells using a heat conduction microcalorimeter fitted with a stirred perfusion vesseL26 Miniaturised electrodes for determination of pH and oxygen concentration were positioned in the vessel which was without gas phase.(A) Results from parallel measurements of thermal power, P, (-), pH (--.--a) and oxygen concentration (---) are shown. The perfusion of the medium was stopped at time zero and cells were injected at the time indicated by the arrow. (B)The total heat production divided into an aerobic part (shaded) and an anaerobic part (hatched), cf. the text. calorimetric curve increased momentarily after which a gradual increase in the thermal power due to aerobic growth is observed. The signal from the oxygen electrode showed a fast decrease, reaching zero after ca. 5 h. The pH also decreased. The changes in the metabolic pattern during the anaerobic phase are reflected by the changes in the thermal power and the pH.Clearly, the thermal power curve is better understood through the support from the two electrode curves. Further, the rate of oxygen consumption, together with corresponding release of thermal power, is essential for interpretation of the energetics of the metabolic process. 5.1.2 Measurements of spectral changes Light can easily be introduced into a microcalorimetric vessel, for example by use of a rod or a fibre bundle made from quartz, glass or plastic. Such light guides have been used in several macro- and micro-calorimeters for the study of photochemical processes.3 By means of light guides it is also possible to use a microcalorimetric vessel simultaneously for heat measurements and as a light absorption cell, Fig.5. Several years ago Schaarschmidt and Lamprecht29 employed quartz rods as light guides in a microcalorimetric vessel used for the simultaneous measurement of thermal power and optical density of a cell suspension. Using a slightly different arrangement this group monitored oscillating reactions (for example the Belusov- Zhabotinskii reaction), calorimetrically in parallel with the changes of the absorption spectrum.30 5.2 Continues analysis conducted outside the reaction vessel With flow or perfusion calorimeters it is easy to connect analytical instruments on-line with the reaction vessel; tech- niques which have been used in several cases. Similarly, gases (d1 Fig. 5 Principle arrangement of a microcalorimetric titrationbight absorp- tion vessel. An electrode is also positioned in the vessel.(a)Lamp and other optical equipment, (b) light guide, (c), titration syringe, (d)micro-calorimetric vessel, (e) electrode, (f) light path in the medium, (g) spectrometer. may diffuse from a calorimetric reaction vessel to analytical instruments outside the calorimeter. A few examples will be given. Gnaiger and coworkers have made extensive use of perfusion microcalorimeters in studies of small aquatic animals, under different conditions of oxygen supply. Using oxygen electrodes positioned in the flow line before and after the calorimetric vessel, the oxygen consumption and its average concentration in the vessel could be measured continuously during calorimetric experiments.31 Criddle et al.32 have used an assembly consisting of two microcalorimetric vessels and a pressure sensor to analyse the energetics of plant tissue in the dark.The three instruments form a closed system connected by gas tight tubes, Fig. 6. One of the microcalorimetric vessels, A, is the reaction vessel for the plant tissue. Calorimetric vessel B contains a solution of NaOH, and serves as a trap for the CO2 formed. From the thermal power measured in vessel B the rate of release of CO2 can be determined. From these results, combined with values from the pressure measurements and the gas volume of the closed instrument system, changes in the oxygen concentration can be derived.Fig. 6 Schematic picture of a microcalorimetric assembly used for the simultaneous determination of thermal power and production/consumption of C02 and 02,in experiments with plant tissue and in the dark. (Adapted from Criddle et a1.32). The metabolic heat is measured by one calorimeter, (a),only the reaction vessel is indicated, CO2 from respiration of the tissue will diffuse to the second microcalorimeter (b),where it is absorbed by NaOH solution. The two calorimetric vessels and a pressure meter (P) are connected by gas tight tubes to form a closed system. Chemical Society Reviews,1997 85 Fig 7 shows a simplified picture of a wet gas perfusionassembly for studies of living tissues, in particular plant tissues 33 The reaction vessels of two twin heat conduction microcalorimeters (A, B) are connected by gas-tight tubing and are positioned in the same thermostatted bath (Only the reaction vessels are indicated) Vessel A contains the biological tissue and vessel B is charged with NaOH solution Perfusing gas is prehumified in a separate vessel (a) and the final water vapour concentration is adjusted by passing the gas through humidifiers (c, parts of the reaction vessel of calorimeter A) The gas leaving the reaction vessel passes vessel B which serves as an on-line CO2 analyser, cf Criddle et a1 32 Gaseous reagents from a syringe (b)can be supplied to the incoming gas-flow and samples can be taken out for continues or stepwise analyses through a septum (f> Dynamic corrections are carried out to compensate for the inertia of the calonmeters and for the slow flow rate of the gas (ca 50 ml h-1) Electrodes and light guides have now been added to the reaction vessel of calorimeter A (which is designed as a modular system) 34 fA Y t ’-A B Fig.7 Schematic picture of a wet gas microcalonmetric perfusion system for simultaneous determination of metabolic heat and C02 production by living tissues 33 (a) Prehumidifier, (b) inlet for gaseous reagents, (c) humidifiers, (4reaction vessel, (e)calorimetnc vessel used as trap for CO2, (#)gas sampling septum 6 Conclusions Some commercially available isothermal microcalonmeters can now be used at a level of sensitivity which approaches the nanowatt level and it has become realistic to work with compounds that are much more expensive, less soluble, have a higher molecular mass etc ,compared to the situation of a few years ago Further, it is judged there will soon be increased possibilities for such work conducted at higher temperatures and pressures However, careful method studies should be conducted before such goals will be realized, remembenng how vulnerable microcalonmetric measurements are to systematic errors Further developments of chemical test processes and an increased use of such processes are needed Much development in isothermal microcalorimetry has been initiated by needs in biochemistry and cell biology In some cases, in particular regarding the thermodynamic character-ization of ligand binding reactions involving biopolymers, and of model systems, the development work has been very fruitful Instrument and method development work directed towards applications in cell biology have had a significant impact on the progress of modern isothermal microcalorimetry, but has not yet been of much importance for the intended target areas In the late 1960s,following the pionering work of Cal~et~~in France and Benzinge1-3~in the USA, the development of modem isothermal microcalonmeters began to accelerate At that time 86 Chemical Society Reviews, 1997 and dunng the following decade there were great expectations concerning the practical use of such instruments as general analytical tools From our relatively short perspective those expectations were not realistic However, based on the concept of combining isothermal microcalorimetry with specific analyt-ical instruments and by a wealth of information from 30 years of method work, a new optimism can now be sensed 7 References 1 P Backman, M Bastos, D Hallen, P Lonnbro and I Wadso J Bzochem Biophys Methods, 1994, 28, 85 2 I Wadso, in Thermal and Energetic Studies of Cellular Systems, ed A M James, 1987, Wnght, Bnstol, p 34 3 I Wadso, in Solution Calorimetry, ed K N Marsh and P A G O’Hare, 1994, Blackwell, Oxford, p 267 4 J B Ott and C J Wormald, in Solution Calorimetry, ed K N Marsh and P A G O’Hare, 1994, Blackwell, Oxford, p 161 5 M Stodeman and I Wadso, Pure Appl Chem , 1995,67, 1059 6 E J King, Aczd-Base Equlibrza, 1965, Pergamon Press, Oxford, p 184 7 E H Battley, Thermochem Acta, 1995,250, 337 8 P R Connelly, R A Aldape, F J Bruzzese, S P Fitzgibbon, J Matthew, M A Fleming, S Itoh, D J Livingston and M A Vaviva, Proc Natl Acad Sci , (USA),1994,91, 1964 9 I Johansson and G Olofsson, J Chem Soc Faraday Trans 1, 1988, 84,551 10 L G Soldi, Y Marcus, M J Blandamer and P M Curtis, J Soh Chem , 1995,24,201 11 S J Gill, J Chem Thermodynamics, 1988,20, 1361 12 G Salvetti, E Tognoni, E Tomban and G P Johan, Thermochim Acta, 1996,285,243 13 M Angberg, C Nystrom and S Castensson, Znt J Pharm , 1991, 81, 153 14 L -E Bnggner, G Buckton, K Bystrom and P D’Arcy, Int J Pharm , 1994, 105, 125 15 G Buckton, Thermochzm Acta, 1995,248, 117 16 A Bakn, Application Note 22021, Thermornetnc AB, Jarfalla, Sweden, 1993 17 I Wadso and L Wadso, Thermochim Acta, 1996,271, 179 18 R J Willson, A E Beezer, J C Mitchell and W Loh, J Phys Chem , 1995, 99, 7108 19 I Wadso, Thermochim Acta, 1995,2691270, 337 20 Special Issue of Thermochzm Acta, ed R B Kemp, 1990, 172 21 Special Issue of Thermochzm Acta, ed I Lamprecht, W Hemminger and G W H Hohne, 1991,193 22 Special Issue of Thermochim Acta, ed R B Kemp and B Schaarsch-midt, 1995,250 23 R D Newell, in Biological Microcalorimetry, ed A E Beezer, 1980, Academic, London, p 163 24 R B Kemp, Thermochim Acta, 1993,219, 17 25 M Monti, L Brandt, J Lkomi-Kumm and H Olsson, Eur J Haematol , 1990,40,250 26 P Backman and I Wadso, J Biochem Biophys Meth, 1991, 23, 283 27 L D Hansen,M S Hopkin,D K Taylor,T S Anekonda,D R Rank, R W Breidenbach and R S Cnddle, in ref 22, p 215 28 I Wadso, Thermochim Acta, 1993,219, 1 29 B Schaarschmidt and I Lamprecht, Expenenta, 1973, 29, 505 30 B Schaarschmidt, I Lamprecht, T Plessner and S C Muller, Thermo chim Acta, 1986, 105, 205 31 J E Doeller, D W Kraus, E Gnaiger and J M Shick, in ref 20, p 171 32 R S Cnddle, A J Fontana, D R Rank, D Page, L D Hansen and R W Breidenbach, Anal Biochem, 1991,194,413 33 P Backman, R W Breidenbach, P Johansson and I Wadso Thermochim Acta, 1995,251, 323 34 P Johansson and I Wadso, unpublished 35 E Calvet and H Prat, Recent Progress in Microcalorimetry, Pergamon, London 1963 36 T H Benzinger and C Kitzinger, in Temperature-Jts Measurement and Control in Science and Industry, ed ,J D Hardy, Reinhold, New York, 1963, vol 3 Received, 18th November 1996 Accepted, 6th December 1996

ISSN:0306-0012    

DOI:10.1039/CS9972600079

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Modern valence bond theory J. Gerratt,a D. L. Cooper,b P. B. Karadakovc and M. Raimondid a School of Chemistry, University of Bristol, Cantock's Close, Bristol, UK BS8 ITS Department of Chemistry, University of Liverpool, PO Box 147, Liverpool, UK L69 3BX c' Chemistry Department, University of Surrey, Guildford, UK GU2 5XH d Dipartimento di Chimica Fisica ed Elettrochimica and Centro del CNR per lo studio delle relazioni tra struttura e la reattivitci chimica, Universita di Milano, Via Golgi 19, 20133 Milano, Italy The spin-coupled (SC) theory of molecular electronic structure is introduced as the modern development of classical valence bond (VB) theory. Various important aspects of the SC wave function are described. Attention is particularly focused on the construction and properties of different sets of N-electron spin functions in different spin bases, such as the Kotani, Rumer and Serber.Applications of the SC description to a range of different kinds of chemical problems are presented, beginning with simple examples: the H2 and CH4 molecules. This is followed by the description offered by the SC wave function of more complex situations such as the insertion reaction of H2 into CH2(lA1), the phenomenon of hypervalence as displayed by molecules such as diazomethane, CH2N2, SF6 and XeF2. The SC Joe was born in 1938 and obtained his first degree in chemistry from Hertford College, Oxford in 1961 and his PhD in 1966 with Professor I. M. Mills in Reading. He was a post-doctoral fellow with Professor W.N. Lipscomb at Harvard, where he laid the foundations of spin-coupled theory, culminating in a major paper in Advances in Atomic and Molecular Physics in 1971. He has been at Bristol since 1968, where is is Reader in Theoretical Chemistry. Mario was born in Gravedona (Como), near Milan (Italy) in 1939. He obtained his first degree from the University of Milan. He was a post-doctoral fellow with Professor M. Karplus at Harvard. He collaborated with Professor M. Simonetta. In 1976, he met Joe at CECAM, at Orsay, where they first began their collaboration. He was appointed to a Chair in Physical Chemistry in 1994. David was born in Leeds in 1957. He obtained hisfirst degree from Brasenose College, Oxford in 1979 and his DPhil in 1981 with description of the ground and excited states of benzene is briefly surveyed.This is followed by the SC description of antiaromatic systems such as C4H4 and related molecules. 1 Introduction The description of the behaviour of electrons in molecules involves the application of quantum mechanics to very complex systems. Our ultimate objective is not simply to confirm theoretically what we already know from experiment. This merely assures us that quantum mechanics is correct. What we seek is much more: we seek insight into the behaviour of the electrons in a molecule, an explanation of the formation of Graham Richards, also in Oxford. He was a Smithsonian Fellow at Harvard College Observatory from 1981, where he met Joe and joined forces with him and Mario.He returned to Oxford in 1984 as a Royal Society lecturer, before taking up an appoint- ment in Liverpool where he is now Reader in Physical Chemistry. Peter was born in Sofia, Bulgaria in 1959. He graduated in Chemistry from the University of Sofia in 1981 and obtained his PhD there in 1983. From 1981 to 1986 he worked as a research assistant and as an assistant professor at the Faculty of Chemistry of Sofia University. In 1986, he moved to the Bulgarian Academy of Sciences, where he was a research jellow until 1994. In 1990-1 995 he was a visiting research associate with Joe Gerratt at the University of Bristol. In 1995, Peter moved to the University of Surrey where he is now a lecturer in Physical Chemistry.Chemical Society Reviews, 1997 87 chemical bonds, the characteristics of such bonds, their strength, type, how they form and how they break. However, one should not lose sight of the fact that a description, if it is to carry any conviction, must also provide reliable numerical results which, in addition, must be capable of refinement if one so wishes. The solutions to these problems are not simple, nor even unique. Quite certainly, no answers of much value are obtainable by straightforward application of the considerable computing power now available to find some kind of numerical solution to the Schrodinger equation. Instead the problem remains with ourselves and as a result-unsurprisingly-several different approaches have been tried.Over the last forty or more years, the most fruitful approach has seemed to be the molecular orbital (MO) or self-consistent field (SCF) approach, in spite of the fact that the MO wave function does not describe correctly even the most basic of chemical processes: the breaking of a chemical bond. Never- theless, many developments have flowed from the MO approach, some of them of great conceptual importance, such as the rules governing the conservation of orbital symmetry in pericyclic reactions,’ others of a technical nature which allow us to progress to more complex wave functions in which electronic correlation effects missing from the SCF approach (including those which are responsible for providing a correct description of bond breaking) are taken into account.This last is the method of configuration interaction (CI method), which has now reached a stage where ca. 108 or 109 configurations can be handled by various computer codes. Sophisticated extensions of the SCF method, such as the multiconfigurational SCF (MCSCF) approach and the ‘complete active space SCF’ (CASSCF) approach, have been developed and these are embodied in highly efficient computer codes, such as GAUSSIAN96, GAMESS, MOLPRO, MOLCAS and other packages which are widely available. While these techniques have benefited from several genera- tions of development work by many talented research workers to produce codes that must surely be close to optimal for scalar, vector- and even parallel-processing machines, the effect of the large numbers of configurations, which are inevitably involved, seriously affects our vital chemical and physical insight into the problem.More recently, density functional theory (DFT) a technique that has been in use for many years by the solid state physics community (see e.g. ref. 7), has caught the attention of many quantum chemists.8 A great deal of development work has been carried out in recent times, as is obvious to anyone who attends quantum chemistry conferences, both nationally and interna- tionally. DFT clearly has a number of advantages as compared to the ab initio techniques based upon MO theory mentioned above, but questions concerning the foundations of DFT, particularly the origin of the all-important exchange correlation potential, remain and have indeed become more urgent.Concurrent with the introduction of MO theory and its variants, is the theory of Heitler and London, or valence bond (HL or VB) theory. In fact, it was Heitler and London who first showed convincingly that the explanation of the strength of covalent bonding lay with quantum theory.2 Just as important, was the clarity of the description offered by this approach. In particular, the HL theory identifies the ‘exchange effect’ as the fundamental phenomenon responsible for those properties which we associate with a covalent chemical bond: its capability of holding together two electrically neutral atoms, valency itself, the saturation of valency and the idea of the directonality of chemical bonds; concepts which lie at the very heart of chemistry.On the basis of these ideas, Heitler and his students were able to produce a compelling explanation, at least at a qualitative and even at a semi-quantitative level, of many, if not most, aspects of chemical b~nding.~?~ Heisenberg further showed that this very same approach is crucial to the understanding of the many different forms of magnetism. To this day, the Heisenberg theory remains the only explanation of this central phenomenon of the physics of condensed matter. It would therefore seem natural for VB theory to have received most attention and development effort. For a short time, it did so. However, the origin of the exchange effect lies in the overlap between the wave functions of the participating atoms.This overlap, or non-orthogonality, between the relevant atomic wave functions has been the source of serious technical difficulties in the wide application of the Heitler-London approach. Such problems remained until new algorithms implemented on modern workstations with large memory, extensive disk storage and high speed I/O, effectively overcame them. An important extension to the HL theory was the introduction of ‘ionic structures’ into the wave function, i.e. the introduction of chemical structures in which the distribution of the electrons is such that two or more of the participating atoms bear formal positive and negative charges. Nevertheless, the introduction of ionic structures gives rise to severe problems, not least from the interpretational point of view.Even in the simplest case of H2, in order to obtain reasonable quantitative accuracy, it is crucial to add to the original Heitler-London (covalent) wave function eqn. (1.1): yc .-{@lsA(rl)@lSB(r2)+@lSB(rl)@lSA(r2)} C(a1C32 -B1a2>, (1.1) ionic structures of the form eqn (1.2): yl -{ @lSA(rl)@lSA(r2)+ @lSB(rl)@lSB(r2)1 C(a182 -P1a21, (1.2) giving as the total wave function for the H2 molecule a linear combination of wave functions (1.1) and (1.2): (1.3). Y,,, = C’Y, + C*Y,. (1.3) Here @lsA (rl) and @IsB (r2) denote 1s-like orbitals for electrons 1 or 2, centred on hydrogen atoms A or B. Coefficient C2 is small but by no means negligible, C2/C1 = 0.25 for H2 near its equilibrium geometry. We are thus invited to view the H2 molecule, which as far as every chemist is concerned, is quintessentially covalent, as a resonance mixture between a covalent contribution, represented by wave function (l.l), and an ionic part, represented by wave function (1.2), a physical picture which flies in the face of one’s every chemical instinct.For larger molecules, many more ionic structures can be formed. From a chemical perspective, most of them undoub- tedly appear rather unlikely if not extraordinary. Nevertheless, this mode of description is still widely used in a number of contemporary texts in inorganic and organic chemistry (see e.g. ref. 5). But as the number of valence electrons increases, the possible number and type of ionic structures grows to such an extent as to obscure the original clarity of the VB description.However, in organic chemistry, there are some situations in which ionic structures play an altogether more positive role. For example, resonance between covalent and ionic structures provides a direct explanation of the ortho-/para-or meta-directing properties of different substituents of a benzene ring under electrophilic attack by various substituents. There is no doubt that still today, organic chemists, at least in the privacy of their laboratories, find that this explanation is the simplest and most satisfying. However, in a remarkably under-valued paper, Coulson and Fischer,6 using the H2 molecule as a simple example, showed that the ionic structures express nothing more than the deformation of the atomic orbitals that occurs when they participate in chemical bonds.They showed that wave function (1.3) can be rewritten as eqn. (1.4), qtot = @A(rl)@B(r2)-k @B(rl)@A(r2)1 fl(al(32 -Bla2>, (1.4) 88 Chemical Society Reviews, 1997 which is precisely of the form of the original HL function, but where the orbitals @A and @*,instead of simply being atomic, are now (in unnormalized form): eqns. (I .5) and (1.6) @A = @isA + Wisg, (15) and @B = $isg + h+is.A (1.6) the mixing parameter h being the same as coefficient C2 in function (1.3).1- Note that orbitals @A and @B overlap: eqn.(1.7). (@A I @B) = AAB (1.7) Thus we see that in cases such as this, the occurrence of ionic structures does in fact do no more than to allow the original atomic orbitals to delocalize somewhat into the neighbouring atoms as the molecule forms, which is of course perfectly reasonable. As the internuclear distance R increases, h -+ 0 and the orbitals revert to pure atomic form. Wave function (1.3) or (1.4) contains ‘left-right’ correlation which is necessary for a correct description of the dissociation of the Hz molecule. It yields 85% of the observed value of D,, the binding energy of H2, compared to 77% for the Hartree- Fock or molecular orbital wave function. This is not the only form of electron correlation in H2. In particular, angular correlation about the internuclear axis is missing.But the type of ‘non-dynamic’ correlation present in wave function (1.4) ensures that molecular dissociation, however complex, is always correctly described. Generally speaking, the deformations of the atomic orbitals in a molecule are large or small, depending upon such factors as the type of chemical linkage (single, double, triple, aromatic or anti-aromatic), the disparity in the electronegativity of the atoms concerned and the bond length: in all cases, as the distance between the atoms becomes large, i.e. as the bond breaks, the deformation of the atomic orbitals decreases to zero, usually occurring quite suddenly at a critical internuclear distance, and the isolated atom form is regained.The passage from wave function (1.3) to (1.4) gives us a new perspective on the role of the ionic structures and suggests an entirely novel direction for constructing electronic wave functions for molecules, to which we now turn. 2 Spin-coupled wave functions Generalization of the foregoing leads us to propose the following wave function for a molecular system: eqn (2.1) YSC = VSM = d{ -.* v:,$1@2 .*.@N@$@:M~ (2-1) which is known as a spin-coupled (SC) wave function. It incorporates a number of features which do not arise in MO-based wave functions and these are described below. In the following Section, the construction of spin functions @:,M will be briefly discussed and after this we shall be ready for a description of the physical interpretation of the spin- coupled wave function.Function (2.1) describes a system with a total number of electrons N,: eqn. (2.1 .i). N, = 2nc +N. (2.1 .i) Of these, 2nc electrons are ‘inactive’ or ‘core’ electrons, described by n, doubly occupied orbitals vl,v2, ..., vn,.They are not considered to take part in the chemical process under study. In addition we have N ‘active’ or ‘valence’ electrons, which are the objects of our investigation. They are described by N distinct, singly occupied orbitals @I, $2, . . . @p, . . . @N. These orbitals are non-orthogonal, i.e. they overlap: eqn. (2.l.ii). (+p I @v) = Apv (2.1 .ii) 7 Coefficient C, is equal to (1 + h2). They are determined in the familiar way as linear combinations of basis functions (approximate atomic orbitals) X,, chosen beforehand and sited on all the atomic nuclei in the molecule. Thus: eqn.(2.2) m p=l where rn is the total number of basis functions. The coefficients cppare determined by minimising the total energy of the system =E, as we shall see. Note that ACLp 1, i.e. the orbitals are normalised. The core or inactive orbitals vrare similarly determined as linear combinations of basis functions, eqn. (2.3), m p=l but with the added proviso that they are not only normalised, but are also orthogonal to one another: eqn. (2.4). (2.4) This property of the core orbitals simplifies many of the subsequent formulas considerably and may always be imposed without changing the form of the total wave function (2.1).Note that in addition there is a further simplification: The core orbitals vlmay always be taken to be orthogonal to the active orbitals @p, again without changing the form of our assumed total wave function (2.1): eqn. (2.5). (@p 1 v,)= 0 (p, = 1,2, ...,N; i = 1,2, ...,n,) (2.5) These properties of the orbitals enable us to write the total energy in a compact form with a clear physical meaning, as we shall see. @;:We now turn to the functions and @!M which also appear in the total wave function (2.1) and play an important role in the theory. These are many-electron spin functions. The function 0i;lcdescribes the coupling of the spins of the 2nc electrons in the core. It has the simple form eqn. (2.5.i) 0:; = G(a1fi2-fila2) G(a3fi4-B3a4)X ...(2.5.i) * * -x ma2nc -1C32nc -P2nc -1a2nc) showing that the electron spins form n, pairs, each pair having a net spin of zero.Whenever there are orbitals that are doubly occupied, this spin function, known as the perfectly paired spin function, is the only one permitted by the Pauli principle. Function @KM is different. It is an N-electron spin function for the N active electrons. The subscripts indicate that the net spin of these electrons is S with z-component M. A characteristic feature of the spin-coupled approach now appears. Since the N valence orbitals are singly occupied, there are several distinct ways of coupling the individual spins of the electrons to each other in order to form the required overall resultant spin S.This number is denoted byf; and is given by the simple formula (see ref. 9): eqn. (2.6). (2s + 1)N! f? = ($/+S+l)!(;N-S)! ‘ More will be said about the important topic of spin functions in the next Section. Thus spin function OFMoccurring in eqn. (2.1) has the form of a linear combination of all the linearly independent spin functions, @gM,k, = 1,2, .....,f?:eqn. (2.7). k (2.7) k=l Chemical Society Reviews, 1997 89 The coefficients CSk are known as spin-coupling coefficients and are also determined by minimising the total energy. Their physical significance will be described shortly. The use of sets of spin functions as in (2.7) is one of the main features of the spin-coupled approach. Finally the operator SQ stands for the antisymmetrising operator.It ensures that the entire function following in brackets { ...} in eqn. (2.1) is antisymmetric, i.e. obeys the Pauli principle; that is when any pair of space and spin coordinates in the total wave function YsMare transposed, YsMchanges sign. Wave function (2.1) incorporates a number of parameters that are to be optimised by minimising the total energy. It is useful at this stage to summarise these: There-are the coefficients cpp in eqn. (2.2) for the spin- coupled orbitals. Since the orbital index p ranges over the values 1 to N and the index for the basis functions, p, from 1 to rn, there are Nrn such coefficients. They are not all independent, since it should be recalled that each orbital $p is normalised, a condition which, in effect, fixes one coefficient per orbital.Similar considerations apply to the coefficients clp(2.3) for the core orbitals, though in this case, the constraints of normalisation and orthogonality, eqn. (2.4), reduce the number of independent coefficients c,<considerably. In addition there are the fl spin-coupling coefficients CSk, appearing in eqn. (2.7). However, we usually require that the total spin function be normalised: eqn. (2.7.i) (@s,m I @S,M) = 1, (2.7 .i) in which the angular brackets in this case (... I ...) denote integration over all spin coordinates. If the individual spin functions @gM,kare orthonormal (see Section 3): eqn. (2.8), then the spin coupling coefficients must satisfy eqn.(2.8.i), rNpCSk = 1 (2.8.i) k=l which shows that there are onlyfl -1 independent parameters CSk-For two electrons, the spin-coupled wave function (2.1) reduces to wave function (1.4). Thus for N = 2, n, = 0 (there are no core orbitals), for total spin S = 0, there is only a single spin function = l), eqn. (2.8.ii), @&= Ma1P2 -P1a2) (2.8. ii) and (2.1) reduces to eqn. (2.9), yo0 = SQ{~102@&1 = {$1(r1)02(r2) + +1(r2)$2(r1) } Wa1Pz -ha217 (2.9) which is the same as wave function (1.4). We now turn to the construction of the spin functions @&4. k' 3 Construction and properties of spin functions There is a very large number of ways of constructing the N-electron spin functions @FM,kand in this section we provide an elementary survey of those which are the most common and have proved themselves to be the most useful in actual applications.Mathematical details are omitted. They may be found in e.g. ref. 9, 10 or 20. The full physical significance of the different bases of spin functions and of transformations between them will become clear in the course of studying the various applications in this review. The most common method of constructing spin functions is simply by using the rules for coupling of angular momenta in quantum mechanics. Let us denote the spin function for electron 90 Chemical Society Reviews, 1997 i as 0,.This represents a state of the electron with spin s = + (which is of course the same for all electrons) and z-component rn, = q.The value of 0,may be +; or -;.In the former case, it is usual to denote the spin function as a,and in the latter as PI. We begin from electron 1. It has a spin of 4.We then couple a second electron to it. According to the rules for coupling angular momentum, the possible values of the resultant spin of the two-electron system, S2, are 4+ f = 1 or ;-+ = 0. To this we now add a third electron. If S2 = 0, then the value of the spin for the three-electron system, S3, must be f. However if S2 = 1, then the value of S3 may be 1 + + = +,or 1 -f = +.Thus for a three-electron system we see that there are two distinct ways of forming a spin function with net spin S = 4,distinguished by the value of S2. In the first case S2 = 1 and in the second, S2 = 0.In contrast there is only one way to form a three-electron function with net spin S = +. As the number of electrons increases, the possibilities multiply. If the net spin for N -1 electrons is SN- 1, then the spin for N electrons can be SN + 4 or S, -f. The process of constructing N-electron spin functions in this way can be conveniently represented by the so-called branching diagram in which the total number of electrons N is plotted against the net spin S. This is shown below in Fig. 1. Each path through the 7124r 5123l 2 312 1 1 /2 0 1 2 3 4 5 6 7 s 9ioiii2 Fig. 1 diagram, whereby one moves from the point (N = 1, S = f) to the right, represents a possible spin function.At each point (N,S)on the diagram, the number of different paths which may be followed, starting from N = 1, S = f is shown in a circle. This number is just the same asfl [eqn. (2.6)] and it may be observed that this is equal to the sum of the two numbers shown at the points for N -1 which are directly connected to the chosen point, f: = fy:; +f:I;, (3.1) a relationship which may be verified directly from eqn. (2.6). It is clear that a particular path on the branching diagram is defined by a series of intermediate spins S2, S3, . . ., and hence the index k which specifies a particular spin function can be represented as eqn. (3.2). kE(S2S3.....SN- 1). (3.2) In this, it is unnecessary to specify S1, since it is always equal to f and similarly it is not necessary to give SN, as this is just the total spin, S.Spin functions constructed in this way are known as Kotani- Yamanouchi or simply Kotani spin functions after those who introduced them.9910 They are orthonormal [see eqn. (2.8)]. Another basis of spin functions which has proved itself of great value in chemistry is that due to Rumerll and was much used in classical VB theory. It is specially suited to describing chemical bonds and hence it is applied almost exclusively to systems (or subsystems) with zero net spin. The functions in this basis are constructed by considering all distinct pairs of electrons i, j and coupling the associated spins oi,oj to singlets eqn. (3.2.i). (i -j) = *(aipj -piaj). (3.2.i) But simply pairing up the spins of all electrons i, j in all possible ways will produce an over-complete set of spin functions. This can be seen most easily in the case of four electrons.For a net spin S = 0, eqn. (2.6) tells us there are two linearly independent spin functions, but following the elementary Rumer prescrip- tion, we would obtain three, the spins paired up as (1-2)(3-4), or (14)(2-3), or (1-3)(24). However, there is a simple graphical technique, devised by Rumer, to eliminate the redundant spin functions, leaving just the required number, eqn. (3.2.ii). N!N--gV!($N+l)! * (3.2.ii) We place N numbered dots in sequence around a circle, and join them up in pairs such that the joining lines do not cross. Thus for four electrons we have: and this leads to two spin functions only (3.2.iii) = fi(alP2 -Pla2)fl(a3P4 -P3a4), (3.2.iii) and (3.2.i~) @:,0;2 = G(alP4 -P4al)fl<a2P3-P3a2) (3.2.iV) as required.A general Rumer function is constructed by pairing the spin of electrons p and q, Y and s, t and u, etc. to singlets. We label the resulting spin function by: eqn. (3.2.v) k-(p-q,r-s,t-u ,...). (3.2.v) The Rumer basis of spin functions has found extensive use in organic chemistry in the description of the mechanisms of a great variety of organic reactions such as aromatic electrophilic and nucleophilic attack and a host of addition reactions. Even after forty or more years of development of molecular orbital methods, this mode of description obstinately remains a major part of theoretical organic chemistry.The Rumer basis has proved to be particularly convenient in describing aromatic systems, where the well-known Kekul6 structures play a fundamental role. The two KekulC structures for benzene are illustrated as (1) and (2),together with the three corresponding structures for naphthalene (3)-(5].$ An im- $ In the case of benzene, three more spin functions-those corresponding to the so-called 'Dewar' or para-bond structures-are needed to complete the set of five spin functions for six JC electrons with net spin S = 0.In the case of naphthalene with ten JC electrons,fAO = 42, so that in addition to the three KekulC structures shown, there are no less than 39 further possible spin pairings-a fact which most organic chemists do not wish to know.As will be seen, an important result of spin-coupled theory is that the role of the unwanted extra 39 structures may be considered negligible when the orbitals are optimized. portant result which arises out of our extensive use of different bases of spin functions, is the great utility of the little-known Serber basis.12 This set of spin functions is constructed by considering pairs of electrons (1,2), (3,4), ....., (N -1, N> in a similar manner to that of Rumer. The pairs of spins are then coupled to form either a singlet (S = 0) or triplet (S = 1) spin, which are subsequently coupled successively together to form the final spin. A particular function in this basis is identified by the quantum numbers, eqn.(3.2.vi), k = ((...((s12, s34)S4; sS6)S6; . . .)S (3.2.vi) in which s12, s34, ..., etc. is equal to 0 or 1, depending on whether electrons 1 and 2, or 3 and 4 form a singlet or triplet. S4, Sg, .... is the net spin for four, six, etc., electrons. The Serber function eqn. (3.2.vii), ((...(0,0)0;0)0.....)O (3.2.vi i) in which the electron pairs 1 and 2, 3 and 4, ...,N -1 and N form singlets is identical to the Rumer spin function eqn. (3.2.viii), (1-2,3-4 ,....., (N-l)-N) (3.2.viii) and to the last Kotani spin function eqn. (3.2.i~). k = (OiOi.. . . . .$) (3.2. ix) The Serber spin functions are particularly useful in displaying the spatial symmetry properties of spin-coupled wave functions, when it is obvious that those spin functions which do not lead to the required overall symmetry of the total wave function have zero spin-coupling coefficients.This has turned out to be of great utility in cases where the introduction of electron correlation leads to unexpected additional symmetries of, for example, the o electrons in a planar x system. Examples of this will be presented in due course. From a more general point of view, it is often physically and chemically meaningful to divide the electrons into groups and it is very convenient if this division is reflected in the mode of construction of the spin functions. Consider a system in which we wish to focus attention upon a group consisting of N1 electrons and another containing N2 electrons, where N1 + N2 = N, the total number of active electrons.For example, this division might refer to N1 electrons in o orbitals and N2 electrons in a n system, or reflect the fact that in one particular mode of dissociation, the molecule forms two fragments consisting of N1 and N2 electrons. For such purposes it is useful to form a set of spin functions of the type eqn. (3.2.x), Ml,4 in which the index k of the total spin function is characterised by eqn. (3.2.xi), k = (SJ2klk2) (3.2.xi) and kl, k2 describe the 'internal couplings' of the component spin functions and @!&,;,. The symbol (SM [ S1S2M1M2) stands for the vector coupling coefficient which couples together the two angular momenta S1 and S2. As an example of this, a spin-coupled calculation was carried out of the dissociation of the HCN molecule in its ground state, involving the dissociation of the C-N triple bond.'? eqn.(3.2.xii). HCN(X12+)+CH(42-) +N(4S). (3.2.xii) The 14-electron molecule of HCN thus decomposes into two fragments, each consisting of seven electrons. A very conve- nient basis of spin functions is one in which the seven electrons of the CH radical are coupled to a definite spin S1 (equal to t in the combined product, CH + N, of lowest total energy) and the N atom is in its ground state, also with a spin S2 of +.Since the Chemical Society Reviews, 1997 91 total spin S = 0 in this example, the two sub-spins S1,S2 must be equal. The other possible value of S1, S2, which may be of interest, (4, i),would for large C-N distances correspond to the dissociation of the molecule into the ground (2n)state of CH and the excited 2P state of N, which corresponds overall to an excited state of the products.§ Using this ‘707’ basis at C-N distances close to equilibrium, tells us how much this excited state contributes to the total ground state of HCN.Of course for equilibrium internuclear distances, HCN is better described in terms of electron-pair bonds, i.e. in the Serber or Rumer basis of spin functions. This brings us to the problem of transforming from one set of spin functions to another. In practical terms the problem is as follows: A calculation has been carried out using, e.g. the Kotani basis (which may be the most convenient) and we have obtained a set of spin-coupling coefficients CSk (k = I, ... ,fi)corresponding to this basis. What are the values of the spin-coupling coefficients c’sk in some other basis? This is not merely a question of theoretical interest, since it is intimately connected to important physical and chemical information which a spin-coupled calculation can yield about a given system. Certainly such a transformation is always possible and it remains to develop an efficient algorithm for carrying it out. This has been accomplished by means of the code SPINS (see ref. 15), which also runs on a personal computed and which transforms a set of spin-coupling coefficients between the Kotani, Rurner and Serber bases of spin functions.It is also possible to combine any of these transformations with a reordering of the active orbitals in a spin- coupled wave function in any manner. It is, hopefully, clear from the foregoing that there are many possible ways of coupling the individual electron spins to form a given resultant S and the choice is mainly dictated by the actual system under consideration and the process (e.g.reaction or dissociation, etc.) under study. An appropriate choice of basis of spin functions sheds much light upon the behaviour of the wave function, in a very compact and physically meaningful manner. Unsuspected symmetries are often exposed and, combined with the shapes of the orbitals, a great deal of physics and chemistry of the system is revealed.Such information often suggests, for example, energetically the most favourable reaction path, or, in the case of a molecule in an electronically degenerate state, frequently suggests the most likely Jahn- Teller distortion of the molecule. 4 The physical interpretation of the spin-coupled wave function Having discussed the various distinctive features of the spin- coupled wave function, we are in a position to describe its physical and chemical significance and to assess its general quality and reliability in relation to other available ab initio quantum chemical approaches. The spin-coupled approach, as we have seen, describes a system with N active electrons, by N distinct, singly occupied, non-orthogonal orbitals, the spins of which are coupled together in all allowable ways to form the required overall resultant S.The single occupancy allows the electrons to avoid one another, thus incorporating a significant amount of correlation between them. The overlap between the orbitals, on the other hand, allows for quantum interference effects which are crucial for a good description of bonding. Perhaps the most characteristic feature of the wave function is the linear combination of many spin functions. The pairing of the electrons in all possible ways, together with the optimization of the shapes of the orbitals, introduces further correlation effects. These two attributes always allow for a correct description of molecular dissociation, however complicated. As the interatomic distances increase, the orbitals regain their pure atomic shape, and the mode of coupling of the spins also reflects the separation of the parts of the molecule.This aspect of the behaviour of electronic wave functions is perhaps the most important in chemistry, since the making and breaking of chemical bonds, or their rearrangement, constitutes the very essence of chemistry at the molecular level. The spin-coupled wave function thus incorporates a con- siderable amount of chemically significant electron correlation in a compact and highly visual form. One phenomenon which made its appearance early on in this work is the fact that chemical bonds do not appear to break gradually, but on the contrary, are little affected by increasing bond distances until a critical value is reached, when large changes are observed to occur in the wave function over a very short internuclear distance: ca.0.5~0,when the orbitals rapidly lose the deformation characteristic of bond formation and at the same time, the spin coupling coefficients also vary rapidly. This occurs at an interatomic distance of ca. 4.5~0, which remains surprisingly constant for many different species and processes. Indeed, one might assert that in a chemical reaction, A+B +C +D, to a good approximation, nothing occurs between the reagents A and B until they approach to within 4.5~0of one another, when the reaction occurs with surprising suddenness. Hence a fairly universal total reactive cross-section might be estimated to be 4n (4.5a0)2 -70 X 10-20 -7 X 10-19 m*.As we have seen, the spin-coupled wave function is very flexible, but it is important to understand the limits of this flexibility: we have the freedom to order the orbitals in any way we please, we may choose the set of spin functions in a wide variety of ways. Some of this freedom can be well utilized in order to reduce the amount of computational effort necessary in determining the wave function (see below). However, a concomitant danger in this apparent freedom is the opportunity for deriving formalisms which may appear different but do not introduce anything new into the theory. It is also important to appreciate clearly the distinction between SC theory and the older or ‘classical’ valence bond theory.In classical VB theory, the orbitals are taken to be predetermined, either as simple atomic orbitals or hybrids of atomic orbitals. These hybrids, moreover, are fixed, for example either as sp, sp2 or sp3, etc. type orbitals. In SC theory, in contrast, no such preconceptions are imposed. The orbitals are optimized as linear combinations of basis functions (usually approximate AOs) much as in MO-based approaches. However, in common with classical VB theory, the spin coupled orbitals in general overlap with one another (except, of course, in the case of orbitals of different symmetry), or, since the SC orbitals are often localized, by virtue of the physical separation between them. Generally speaking, no constraints, apart from normal- ization, are applied to the SC orbitals and as a result they may be as localized or as delocalized as the situation demands.Bearing in mind that the SC orbitals are always singly occupied, this last means that their shapes are determined by whatever produces the optimum balance between the greatest extent of avoidance of the electrons in different orbitals and quantum interference effects, which arise from the overlap between orbitals. In practice, we have found that this invariably means that the SC orbitals turn out to be localized and indeed often resemble atomic or hybrid atomic orbitals, or semi-localized, 8 States with (S,,S,) equal to (3,of interest and are almost certainly repulsive as far as the interaction of CH meaning that the SC orbitals spread over two or, at most three centres.We have found a greater degree of delocalization than this to be rare. :)and ($ $) are too high in energy to be and N are concerned. 1 Copies of which may be obtained by contacting P.Karadakov@sur- One should also bear in mind that, in contrast to classical VB rey.ac.uk or from http://rs2.ch.liv.ac.uk/dlc/SPINS.html descriptions, the SC orbitals remain singly occupied, the 92 Chemical Society Reviews, 1997 optimization of the orbitals, as explained above, obviating the need to introduce ionic structures for covalent molecules. The determination of the SC wave function, the orbitals and values of the spin-coupling coefficients, is attained by the usual procedures in quantum mechanics.According to the variational principle, we must calculate the total energy E and minimize it with respect to these same parameters, cpp,c,~and CSk. This is done by calculating the first and second derivatives of E with respect to cpp, cIp and CSk. Much of the algebraic detail is set out in the original paper20 and here we do no more than quote one or two equations. The total energy corresponding to the spin- coupled wave function is given by eqn. (4.1). = ppl4(opIhlou)+; ~~(p~,i.)(o.B..Is,oiolA p u=l Ir. u 1r=l ) (4.1) Here, X stands for the usual non-relativistic electronic Ha- miltonian, containing the operators for the kinetic energy of the electrons and all the Coulomb interactions between the electrons and the nuclei making up the molecule.The normalisation integral (Ys, I YsM) is written on the right hand side of (4.1) as A and the usual one- and two-electron integrals as (@pI h I @v) and (@p@v I g I @h@t).The D(p I Y) and D(pv IAt)are elements of the one- and two- electron density matrices. This, together with the normalization integral A is where all the effects of the non-orthogonality between the orbitals occurs. If inactive orbitals are present, it is sufficient to modify the one-electron operator h which occurs in (& I h I &)./IThe spin-coupled wave function as represented by eqn. (2.1) is, of course, based on a single spatial configuration. There are several situations which are not covered by this form of wave function.In particular, degenerate electronic states, frequently accompanied by a Jahn-Teller distortion of the nuclear framework, pose interesting problems (as indeed such states also do in the MO framework). The further systematic refinement of the spin-coupled wave function and the treatment of excited states also present areas not covered by wave function (2.1). Excited states in particular have proven to be a difficulty for the traditional valence bond approach, but are fully accounted for in the spin-coupled approach. This matter is mentioned briefly in Section 8. The spin-coupled wave function does not-cannot-incor- porate all the different types of electron correlation effects. An obvious extension is the development of a multiconfiguration spin-coupled wave function and steps in this direction were already attempted with the earliest applications of spin coupled theory.17.18 Since the spin-coupled wave function incorporates as much ‘radial’ correlation as is possible within the framework of a single-configuration form, the main improvements to the wave function must stem from the inclusion of doubly excited configurations in which two occupied orbitals are replaced by orbitals which differ from them, in some sense, by symmetry. In the case of the very simple diatomic molecules H2, LiH and Liz, (see ref. 17 and 18 above), whose occupied orbitals are all of 0 symmetry, it is fairly clear that the main improvement is due to I( This is achieved by replacing h by the operator F,, which is the Fock operator for the core.c2hl + c(F, = 2J, -Kl), /=I / =I where J, and K, are the usual Coulomb and exchange operators of Hartree- Fock theory, but here are constructed from the inactive orbitals only. replacement of the two 0orbitals involved in the bond by two orbitals of n symmetry, n+n-, if complex orbitals are used, or by nXdx+ JT~’~,if real orbitals are employed. In the remaining parts of this Review, we present a selected series of applications of spin-coupled theory to different parts of chemistry. 5 Simple examples The most elementary example of course is just the H2molecule. As described by eqns. (1.4) or (2.9), there are two orbitals, @A and @B which overlap. These are displayed in Fig.2 below. On the right of Fig. 2, contour plots of the two orbitals are shown, while on the left, orbital @A is shown as a three-dimensional shape with the intemuclear axis superimposed. As the inter- nuclear distance R increases, the deformation of each orbital @* and @B decreases to 0, leaving just a pure hydrogen 1 s orbital on each atom. The potential curve for H2 given by the spin-coupled wave function is compared to the results from a number of other wave functions in Fig. 3. We see that the spin-coupled result remains close to that for a full configuration-interaction (full-CI) wave function** for all values of R and then it becomes identical with it as R +CO. Also shown on this diagram is the potential curve given by the original Heitler-London wave function, eqn.(1. l), and that given by MO theory (the ‘self-consistent field’ (SCF) function). It can be seen that the SCF wave function does not describe dissociation correctly. The potential curve of the lowest triplet state of H2, which is repulsive, is also shown. According to eqn. (2.1), the wave function for this state (N = 2, S = 1) is given by eqn. (5.1). ** A full-CI wave function is the most general variational wave function that can be constructed from a given basis set, either in the MO framework or that of VB (in which case it is called a full-VB wave function). The two wave functions (full-CI and full-VB) are entirely equivalent. Chemical Society Reviews, 1997 93 -0.2 rlbohr Fig. 3 which for A4 = 0, can be obtained simply from that for the singlet state by reversing the signs in eqns.(1.1) or (1.4). A somewhat more complicated example is provided by the methane molecule CH4. In this case we have eight valence electrons and two core electrons, so that N = 8, n, = 1. The ground state is a singlet, S = 0, so that now there is a total of fourteen spin functions. The SC wave function is determined by minimizing the total energy as described in Section 2. Using a very large basis set,?? the energy of the spin-coupled wave function is ca. 0.065 Hartrees lower than that of a SCF function calculated with the same basis, which is a fairly substantial difference. This indicates that a significant amount of the effects of electron correlation is incorporated in the SC wave function.On examining the wave function, we find that the eight SC valence orbitals form four sets of two orbitals each. The orbital pairs are symmetrically equivalent and are permuted amongst themselves by operations of the tetrahedral group, Td. Within each pair, one of the orbitals is very largely localized on one of the H atoms and clearly resembles a H(1s) orbital with small deformations, almost exclusively onto the C atom. The second orbital of the pair is localized mainly on the C atom, but with some small degree of delocalization, almost exclusively onto the H atom of the pair. This is shown in Fig. 4. The deformation of the C-based orbital can be seen quite clearly. Thus, without the imposition of any preconceptions on our part, the SC wave function for methane produces what are clearly four, somewhat deformed, sp3 hybrid orbitals and also four deformed H(1s) orbitals.It goes without saying that this corresponds closely to our usual concept of four-valent carbon with bonds to four H atoms. In addition, however, we must consider the spin functions of which there are fourteen. This example illustrates very clearly some of the features of different spin functions. Table 1 shows them, using the Serber basis, with the orbitals ordered as ($1,$2) (c$~,$~),($5,$6), ($7,&J. The notation of column 2 has been explained in Section 3. Of these, spin function 14 corresponds to the perfectly paired case and this has a weight of 89.4%.This of course is what we ~~ ~ ~ tt 'Triple-zeta valence plus polarization' (TZVP). 94 Chemical Society Reviews, 1997 imagine to be the case in a simple molecule such as methane. All the other spin functions represent different possible ways of coupling the eight valence electrons. It can be seen that a number of them are zero by symmetry and that six others are exactly equal (as long as the molecule has tetrahedral sym- metry). These represent the case when there are two triplet bonds, the two triplets giving a zero net spin. The weight of such an unlikely spin arrangement is naturally rather small: 1.7%. With four symmetrically equivalent bonds, there are six such unique pairs of triplet-coupled bonds. Together with the perfectly paired spin function, this constitutes the entire spin function, for 89.4 + 6 X 1.7 = 100%. However, once the H atoms move away from their equilibrium position in CH4, spin- coupling patterns other than the perfectly paired case assume a much more important role.Table 1Spin coupling coefficients for CH4 (Serber basis) Spin-Coupling pattern Coefficient Weight 1 0.0487 0.0023 2 -0.1298 0.0168 3 0.0 0.0 4 -0.1298 0.0168 5 0.0 0.0 6 0.0 0.0 7 0.0 0.0 8 0.0542 0.0029 9 -0.1298 0.0168 10 -0.1298 0.0168 11 0.0 0.0 12 -0.1298 0.0168 13 -0.1298 0.0168 14 0.9453 0.8936 If we transform these spin-coupling coefficients to the Rumer basis, we find that the perfectly paired spin function now contributes ca. 80% to the total spin function.The remainder is made up almost entirely from structures containing a single H-H pairing and only two C-H bonds. There are four such structures which, at the equilibrium geometry of CH4, are all symmetrically equivalent. The occurrence of this type of spin pairing with significant contributions (ca. 20%), provides us with some insight into the topic of the following section: the insertion of H2 into CHZ(lA1). 6 The CHZ(lA1) + H2 insertion reaction Following on from this discussion of methane, it is interesting to study a related process: that of the insertion of the H2 molecule into the carbene radical CH2 in its low-lying 'A1 excited state to form CH4, eqn. (6.1). H2 +CHz('A1) +CH4.(6.1) This reaction is obviously more complex than the simple bond- breaking process, eqn. (6.2) CH4 +CH3(2A'I) + H (6.2) and involves a fairly complex rearrangement of bonds. Clearly the spin-coupling of the electrons on the right hand side of eqn. (6.1)is completely different to that of the left hand side. In order to understand the pathway for this reaction, it is necessary to look briefly at the SC description of CH2 in this excited state.24 This is shown in Fig. 5. ; H H H II I Fig. 5 Orbitals (a) $1; (b)$2; (c) $3; (4$4; (e)$5; (f) $6 Orbitals are plotted in the plane of the molecule. We clearly see two C-H bonds ($1, $2) and ($3, +4). Orbitals +5 and $6 are plotted in a plane perpendicular to the molecular plane and bisecting the H-C-H angle.If the two C-H bonds are regarded as pointing into the corners of a tetrahedron, then $5 and @6 point approximately in the direction of the two remaining corners of the same tetrahedron. Thus, according to the SC description, the 'A1 state of CH2 closely resembles methane itself from which two H atoms have been plucked. This is quite different from the simple MO description of this state, according to which the two non-bonding electrons of CH2 are considered to occupy a single lone pair orbital of a1 (0) symmetry, stemming from the C atom and pointing along the C2 axis away from the H atoms. The following discussion focuses upon the utility of the SC description. It would appear that the most obvious approach of an H2 molecule to CH2( 'Al) is a simple symmetric path where the axis of the incoming H2 is perpendicular to the C2 axis of the molecule and in a plane perpendicular to that of CH2, bisecting the H-C-H angle.However, this implies the simultaneous breaking of the H2 bond and the formation of two new C-H bonds. According to SC calculations, this path encounters a very high-energy barrier and thus is very unfavourable. But on optimising the various parameters which define the H2 + CH2 system (see ref. 23), we find the minimum energy path is as shown schematically in Fig. 6. The H2 molecule HI-H2 approaches along the line of the lobe of one of the singly occupied orbitals of CH2, $5, say. Then as Fig. 6 Schematic drawing of the pathway for the insertion of H2 into CHZ(I'41) the H1-C distance decreases to ca.3ao, the H1-H2 bond begins to break and a new HI-C bond begins to form-as shown clearly by the sharp variation in the coefficients of the two most important spin-couplings. Simultaneously, the second incoming H atom, H2, now swings right around so as to interact with the other singly occupied orbital of CH2, $6, so forming the second new C-H bond. There is no barrier along this path and the value of the exothermicity of the reaction, calculated from the energies of the SC wave functions for reactants and products is 474.5 kJ mol-1, compared to an experimental value of 490 kJ mol-1. With the benefit of hindsight, this reaction path is entirely reasonable: one new C-H bond starts to form, followed swiftly by a second. Energetically, this is obviously more favourable than the simultaneous formation of two new C-H bonds.The overall process is nonetheless synchronous. Although one bond starts to form first, this process is not completed before the second bond begins to form. However, it would be hard to predict this highly asymmetric reaction path without knowing the form of the SC orbitals for CH2( lA1). If one were to reverse the motions along this path and study the disintegration of CH4 by, say, the pumping of much energy into appropriate vibrational modes, then this picture predicts that as the first H atom starts to depart from CH4, it takes a second H atom with it and, moreover, that the nascent H2 molecule will possess a great deal of angular momentum, i.e.the H2 will be in a highly excited rotational state. There is indeed experimental evidence for this. 7 Hypervalence The impossibility of drawing satisfactory Lewis structures for a whole range of molecules, such as NO2, HCNO, N20, 03,is well-known. These are often referred to as 1,3-dipolar mole- cules. In order to bring their electronic structure and apparent valency in line with other, less unusual systems, many chemists fall back upon the concept of 'resonance'. The resonance structures that are commonly drawn for NO2 are well-known to all chemists. For diazomethane (CHZNZ), which is planar, it is common to indicate resonance between the structures shown below, which indicate that diazomethane has some diradical character.Although it is unstable and tends to dissociate explosively into N2 and an excited state of CH2, there is otherwise nothing in its vibrational or rotational spectrum to indicate that this molecule possesses any kind of unpaired electron character. Chemical Society Reviews, 1997 95 Spin-coupled calculations on diazomethane and a number of other 1,3-dipolar molecules listed above have been carried out at various levels, the simplest example to consider being just the four n electrons of CH2N2. According to the SC model, they are accommodated in four distinct orbitals, shown in Fig. 7. The I II contours are plotted in a plane parallel to the molecular plane and la0 above it.It can be seen that orbital $1 is largely a x orbital centred on the C atom of CH2N2, but distorted somewhat towards the neighbouring N atom, N,. Similarly, $2 is centred on N, and is deformed towards $l. However, there is a second xorbital $3 centred on N, but distorted towards the terminal N atom, N, and finally there is $4, which is a JI orbital on N, but distorted towards $3. The spin-coupling coefficients show that, in spite of the very large overlap between the two orbitals centred on N,, $2 and $3 (A23 = 0.785, using a very large basis set of Gaussian-type orbitals$$), the electrons occupying these orbitals are not paired with each other, but instead there is an almost perfect pairing of the electron spins in the two orbital pairs (@I, $2) and ($3, $4).This corresponds to the chemical structure: H\ C=NEN H’ in which the central N atom takes part in five electron-pair bonds. The large overlap A23 between $2 and $3 should not be forgotten. We tend to think that if a pair of electrons forms one bond, then it cannot contribute to another bond, i.e. that pairs of orbitals forming a bond are orthogonal to all other pairs. Here, we see that this is not so: in spite of the clear formation of five fully fledged electron-pair bonds, the overlaps between them are such that the average number of electrons around N, (the Mulliken population) is very close to 7: i.e. the central N atom is almost exactly neutral. The same is true of N,. However, the C atom is slightly negative and the two H atoms are correspondngly slightly positive.This provides the H atoms with a distinct acidic character, all of which is fully in accord with the known chemistry of CH2N2. $$ A ‘triple zeta plus valence-shell polarization’ (TZVP) basis. 96 Chemical Society Reviews, 1997 Similar considerations hold for other 1,3-dipoles. The N20 molecule is best presented as N=N=O in which the central nitrogen atom again participates in five electron-pair bonds. The obseryed bond lengths of NzO (RN -N = 1.128,RN -0 = 1.184 A) are certainly reminiscent of an N-N triple bond and a ‘normal’ N=O double bond. Thus, from the perspective of SC theory, the problem presented by these molecules resolves itself in a remarkably simple fashion.Contrary to what we were all taught as undergraduates, the nitrogen atom does indeed form five covalent linkages and the availability or otherwise of d orbitals has nothing to do with this state of affairs. This usually shows itself in terms of multiple bonds, such as in CH2N2, N20, F3N0, etc., rather than as five single bonds, simply because the small size of the N atom normally precludes the presence of five nearest neighbours. This is not so for the phosphorus atom. The apparent difference in valency between the first and second rows of the periodic table is therefore a consequence of the size of the atoms and is not primarily due to the availability or otherwise of 3d orbitals for bonding. This conclusion is reinforced by spin-coupled and other studies on SF6 and PF5.The nature of the bonding in these molecules presents a direct challenge to conventional views on valency. According to the SC calculations, the results§§ for SF6, for example, show that the 12 valence electrons are accommo- dated in twelve well-localized orbitals which form six highly polar S-F bonds. Furthermore, the nature of the orbital pairs which form these bonds is very little affected by whether or not d orbitals on the S atom are included in the basis set. Two of the resulting orbitals, forming one of the S-F bonds, are shown in Fig. 8. Once more it should be emphasized that these orbitals are the straightforward outcome of an ah initio calculation, with no constraints imposed as to the final form of the orbitals, nor the type of pairing of the electron spins.It remains to rationalize this result in simpler terms. It is worth recalling that explanations of the bonding in SF6 which are based upon the supposed d2 sp3 hybridization of the orbitals on the S-atom cannot explain properly why SH6, for example, which is unknown, should not be just as stable as SF6 itself. The polarity of the S-F bonds is clearly an essential part of the answer. This can be simply visualized as follows. Consider one of the bond pairs $2) of SF6, illustrated above. Orbital $1 consists of a combination of a 2p orbital on one of the F atoms and an sp hybrid Xs(sp) on the central S atom, where Xs(sp) is of the form in eqn. (7.1). XS(SP) -%3s) S(3P) (7.1) so that eqn.(7.2) holds. $1 = WP) fhXs(sp> (7.2) (see the upper orbital in Fig. 9). The other member of the pair, $2, is an almost pure F(2p) orbital [eqn. (7.3)]. $2 = WP) (7.3) (see the lower orbital in Fig. 9). Hence the bond pair ($1, $2) is of the form eqn. (7.3.i). ($1, $2) = { (F(2p) + hXs(sp)), WP) I = (mp), F(2p)) fh(Xs(sp), WP)) (7.3.i) In other words the S-F bond has significant ionic character, but with sufficient covalency to provide directionality. $3 The calculations on the twelve valence electrons included all 132 spin functions, as specified by eqn. (2.6). However, at least at the equilibrium geometry, only the perfectly paired spin function plays any significant role. F F F F F F Fig.8 It is important to note that we may form six sp hybrids of the form (7.l), according to whether we choose the x-,y-, or z-axis. These hybrids however are linearly dependent, since we started from just four orbitals, S(3s), S(3px), S(3pY), S(3pJ and arrived at the six functions (7.1). This is resolved by the incorporation of F(2p) character. The polarity of the S-F bonds in SF6 is therefore a necessary part of the SC description. Similar considerations apply to PFs and the original paper (ref. 25) should be consulted for further details. It remains to add that almost precisely the same description goes through for XeF2, Fig. 9. There are two bonding pairs of orbitals, each one of which is very similar to the pair (Ql,Q2) described above for SF6, leading to very polar Xe-F bonds.It is clear from the results presented in this section, that the time has come from the much-loved octet rule to be superseded. Presented with sufficient energetic incentives, almost all valence electrons can take part in bonding. We need retain only an 8-electron rule, similar to the 18-electron rule of transition metal chemistry. Polar bonds which shift density away from the central atom appear to be favoured, particularly if the formal number of bonds is very high. Hence differences in electro- negativity and the size of the central atom can be useful first guides to the possible existence of a particular hypervalent species. Xe Fig. 9 8 Aromaticity and antiaromaticity The concepts of aromaticity and antiaromaticity lie at the very heart of organic chemistry.The first useful description of benzene was due to Kekul6 who drew the structures (1)-(2) in Section 3 (see also the footnote) and whose ideas of resonance between the different C-C bonds were later justified and clarified on the basis of quantum theory by Pauling in terms of different spin-pairings of the electrons in C(2p,) orbitals, i.e.in terms of resonance between the so-called Kekul6 and Dewar (or para-bond) structures. Molecular orbital theory, however, gives an entirely different type of description: that of n orbitals delocalized around the benzene ring. The associated MO energy level diagram is shown below, with the appropriate labels for the point group of the molecule, D6h(Fig.10). a2, c6 Fig. 10 Accordingly, the electron configuration of the ground state is (a:" etg). By generalizing this diagram, a simple but very useful Chemical Society Reviews, 1997 97 rule was obtained by Hiickel which predicts that if n is the number of carbon atoms in a ring system, those molecules with 4n + 2 C atoms will be aromatic, while those with 4n C atoms are predicted to be antiaromatic. Over the last forty or so years this has become the accepted description of benzene, while that in terms of Kekulk and para-bond structures has become somewhat less common. However, in 1986 a spin-coupled calculation was carried out on the n electrons of benzene.26 As emphasised previously in section 4, no constraints or preconceptions were imposed upon the form of the orbitals, nor upon the type of coupling between the spins.The result showed six n orbitals, each of which is localized around one of the C atoms constituting the ring. One of them is depicted in Fig. 1 1 in which the contours of the orbital H Fig. 11 are drawn in a plane parallel to the molecular plane and lao above it. It can be seen that although this orbital is clearly localized, there are obvious deformations towards the neigh- bouring C atoms on each side. The remaining spin-coupled orbitals are obtained from the one shown by successive rotations of 2x/6 about the principal symmetry axis of the molecule. The spin-coupling coefficients in the Rumer basis have the values shown in Table 2.These numbers change only very slightly with different basis sets for the orbitals. Spin functions 1 and 4 correspond to the two Kekulk structures and we see that they each make a contribution of ca. 40.5% to the total wave function. The remaining three spin functions, 2,3 and 5 are the Dewar or para-bond functions and each of them contributes ca. 6.4% to the total. Table 2 Spin-coupling pattern Coefficient Weight 1 (1-2,3-4,5-6) 0.5 1638 0.4046 2 (1-4, 2-3, 6-5) -0.09461 0.0636 3 (2-5, 3-4, 6-1) -0.09461 0.0636 5 4 (1-2, 6-3,54) (2-3,4-5, 6-1) -0.09461 0.5 1638 0.0636 0.4046 The most significant feature of this result is that the energy improvement (energy lowering) obtained by the spin-coupled wave function over that of the MO wave function (for a given basis set) is no less than ca.92% of the maximum attainable improvement using a wave function of whatever type, MO or VB, constructed from six electrons and six orbitals. In other words, a fully correlated wave function for the x electrons of benzene approximates closely to the spin-coupled wave func- tion. There is thus very much more to the Kekulk description of benzene than was hitherto realized. This calculation has since been repeated many times with basis sets of varying size and with complete optimization of valence and all inactive orbitals. The results vary very little. Spin-coupled calculations have subsequently also been carried out on many aromatic systems, such as heterocyclic five- and six-membered rings, on naphthalene and on azulene.For naphthalene and azulene, with ten n electrons, the orbitals obtained are very similar to those of benzene, with the exception of the two orbitals localized at each of the C atoms which bridge the two rings. These orbitals display a three-way deformation, towards each of the three adjacent carbon atoms. In addition formula (2.6) shows that for a ten-electron system there are 42 possible spin functions which should be taken into account. But since the spin-coupled orbitals are fully optimized, it turns out that the only spin functions which play any significant role in these molecules are those corresponding to the Kekulk struc- tures [in the case of naphthalene, structures (3),(4) and (5)in the diagram in Section 31 and that the contribution of the other 39 structures may be neglected.The MO description also predicts a number of excited states of benzene. Thus, a single excitation of an electron from an occupied MO (a2u or el,) to one of the unoccupied MOs shown on the diagram above (e.g.to the e2" or bZg orbitals) gives rise to a number of valence excited states. In addition, there is also a large number of Rydberg states with energies below that of the first ionization potential. A constant aim of theoretical studies is to determine these excited states, preferably without going beyond the o/n approximation. Certainly for the ground state, to abandon the o/n separation would be to ignore a vast body of chemical experience, but the situation may be different in excited states.It turns out that the excited states of benzene (with energies less than the first ionization potential) fall into three classes: covalent, ionic and Rydberg. An example of a covalent state is the first singlet excited state, lBzu, lying at an energy of 4.90 eV above the ground state. It may be represented to an excellent approximation (see ref. 27) simply as eqn. (7.3.ii) Y('BZu) = K1-K2 (7.3 .ii) i.e. as the negative combination of the two Kekulk structures of the ground state. Covalent states are in general fairly easily described within the o/x approximation. On the other hand, ionic states require linear combinations of structures of the type: in which two x orbitals occupy one C-atom site, while on the neighbouring C atom, there are none.Ionic states of benzene are much harder to describe within the o/xframework. Physically, it is obvious that this is due, in part, to the existence of positive and negative charges in the n-electron distribution, which causes a static polarization of the 0core, so that the core differs from that of the covalent states. In addition, there are further dynamic o-x interactions which fall outside the o/xapproxima-tion. In any case, a reliable description of these effects requires extensive basis sets which include diffuse atomic orbitals. Lastly, there are the Rydberg states. These are characterized by one orbital which is very diffuse and extends a significant distance from the molecule.Given a basis set which includes such diffuse atomic orbitals (even to the extent of centring them at the midpoint of the molecule), such states are not too difficult to describe well. The spin-coupled description of the excited states of benzene thus leads to an important and useful classification: the valence states are covalent or ionic, the latter being significantly harder to describe than the covalent states, and Rydberg states, which differ physically from the valence states, but otherwise are not difficult to determine accurately. In the MO description, all the valence states arise from one or two singly excited reference configurations, such as (gue:g e2J or (a$ue:g b2& and it is not at all clear from this why the various valence excited states should be so physically distinct.The simplest antiaromatic system is cyclobutadiene, C4H4. A similar diagram as for benzene for the energies of the molecular orbitals of C4H4, assuming a square-planar geometry, gives: 98 Chemical Society Reviews, 1997 / C4 from which it follows that the electron configuration of the n electrons in the ground state is (a;, ei). However, the eg MO is doubly degenerate and according to Hund’s rules, the two electrons with this energy will distribute themselves one in each MO and with spins aligned parallel. MO theory thus unambigu- ously predicts that the ground state of square-planar cyclobuta- diene should be a triplet (3Al.J.This is not so. It is now well-established that the ground electronic state of cyclobutadiene is a singlet. In the square- planar geometry (D4h symmetry), the state has the symmetry lBlg, with the 3A2g state lying at an energy of ca. 0.43 eV above it. Furthermore, the square-plane of the ground state is unstable, the molecule preferring to distort to a rectangular geometry with two short C-C double bondsnn and two longer C-C single bonds. However in the 3A2, state, the square plane geometry is stable. From the spin-coupled wave function for the ground state, one obtains four orbitals of n symmetry without imposing any predetermined form on these orbitals, nor on the type of spin coupling. Two of the resulting orbitals are shown in Fig.12. HI Fig. 12Orbitals (a) $,; (h)$2 They are plotted in a plane parallel to the molecular plane but la0 above it. It can be seen that and $2 are centred about the two horizontal C atoms, Cl and Cz. Orbitals & and $4 are the same as $, and $2, but are rotated by x/2 about the main symmetry axis and instead are centred about C3 and C4 in the diagram, where C3 lies vertically above C4. We note that orbital $2 possesses an extra nodal plane passing through C3-C4, compared to $1 (similarly $4 and $3). Even more remarkable are the spin couplings, for these show that orbitals 42)are almost exactly coupled to a triplet and the same holds for ($3, $4), the two triplets coupled to give an overall value of the spin S for the four electrons of zero.l((( Analysis of how orbitals $,-$4 behave under the operations of D4h, shows that the total wave function for the ground state of C4H4 has the correct lB1, symmetry.Furthermore, the square planar geometry is not stable and the molecule distorts to a rectangle. In the course of this distortion, orbitals rapidly become localized over the four atoms C1-C4 and clearly show the formation of two C-C double bonds which are shorter than the remaining two C-C single bonds. 17 A ‘second order Jahn-Teller instability’. I[[(The phrase ‘almost exactly’ is important here (spin coupling coefficients 0.999865 and 0.0166468), for If (GI, $2) and ($3, $4) were each exactly coupled to triplets (with spin coupling coefficients 1.0 and O.O), then the overall wave function would remain unchanged by the replacement of The four orbitals of the triplet state are remarkably similar to those of the singlet ground state.The spin pairing is also very similar, orbital pairs ($1, $2) and ($3, Q4) each forming a triplet. These two triplets, however, are now coupled to form an overall triplet, as required for this state. This is found to have an energy 0.410 eV higher than that of the ground state, which compares well with the experimental value of -0.43 eV (see above). It thus appears that antiaromatic character is connected to the formation of a triplet spin from a pair of electrons in two distinct orbitals, such as and $2 above. We refer to such a combination of orbitals as an anti-pair.In order to place the concept of anti-pairs found for cyclobutadiene within a wider context, several related systems were studied, one of which is 2,4-dimethylenecyclobutane-1,3-diyl (DMCBD), shown below: &b This molecule has six electrons in six orbitals of n symmetry (it is an isomer of benzene) and can be regarded as being derived from C4H4 by removing two H atoms from cyclobutadiene and substituting them with methylene groups. From this, one would predict that only one of the anti-pairs found in cyclobutadiene would remain. This is indeed the case. The ground state of DMCBD is a triplet. Orbitals $2) and ($3, 44) are, as indicated in the diagram above, four highly localized C(2n) orbitals and (together with the appropriate 0 orbitals) form normal C-C double bonds.The two remaining orbitals of the C4 ring, denoted by a and b,form an anti-pair very similar to one of those in cyclobutadiene itself. Even more remarkable is the bismethylenebiscyclobutyl-idene molecule (BBB), shown below. This is similar to DMCBD, but with an extra C4 unit, plus methylene group, added. b d Orbitals ($5, +J,(c$~,$8) and (&,, form the n componentsof fairly conventional C-C double bonds. However, while the terminal orbitals Q7 and $9 are deformed towards their respective partners, $8 and $10, the other orbitals centred on Cg, Cg, C8 and Clo are deformed in three directions, due to the presence of three C-atom neighbours. BBB turns out to have anti-pairs in both C4 rings, i.e.orbitals (a,b) and (c,d). That is, each ring has associated with it a net electron spin of S = 1.The spins of the two C4 units, however, are aligned antiparallel with each other, giving the ground state of BBB a net spin of zero. This is therefore very much akin to an antiferromagnetic system, except that the spins stemming from each C4 unit each have the value of unity. It is not too hard to imagine an organic polymer consisting of an infinite number of such C4 units and displaying this kind of antiferromagnetic behaviour. According to the Hiickel4n rule, cyclooctatetraene (CSHS) is the next member of the ‘antiaromatic’ series after cyclobuta- diene. Consequently, one would expect that the SC picture of bonding in this molecule would, in some way, remind one of that observed for C4H4.However, SC calculations recently carried out at the lowest-energy tub-shaped 02d geometry of CsHs, as well as at two idealized geometries: a D8h regular octagon and a D4h octagon with alternating carbon-carbon bond lengths show something different;29 see Fig. 13. The eight active orbitals at the D2d and D4hgeometries form four identical, largely independent olefinic carbon-carbon JC (or Chemical Society Reviews, 1997 99 Dt3h D4h D2d D2d (view fiom above) (view from side) Fig. 13 at the tub-shaped geometry, almost n;) bonds. Resonance is insignificant (perfect pairing within the bonds represents by far the most important spin function) and the conclusion is that, at these two geometries which include the one experimentally observed, cyclooctatetraene is definitely non-aromatic.Anti- aromaticity is restricted to the idealized regular octagonal structure. However, the nature of the SC wavefunction at this geometry is different from that for cyclobutadiene. The eight equivalent SC orbitals are localized and very similar to those in benzene; there are no antipairs. The key to the low stability and higher reactivity of the molecule-the two main characteristic features of antiaromatic systems-is in the spin-coupling pattern: in the Serber spin basis, spin functions involving triplet pairs are responsible for 81% of the spin function, with 75% contributed by a spin function made up of triplet pairs only, (((1,112;1) 1; 1).The comparison between the SC descriptions of cyclobuta- diene, benzene and cyclooctatetraene clearly indicates that the reason for the lower stability and higher reactivity of anti- aromatic systems is due to a simultaneous unfavourable coupling of the spins of all valence orbitals to triplet pairs, which discourages bonding interactions and suggests diradical character.9 Summary and conclusions In this short survey we have attempted to describe a range of different chemical systems to which spin-coupled theory has been applied and, hopefully, demonstrated the clarity and freshness of the chemical insights that the theory offers. The whole style of description used in this article differs radically from that traditionally employed by the more orthodox methods of quantum chemistry. Inevitably, a different choice of topics could have been made, so that those covered fall far short of the many applications so far of spin-coupled theory.Among the topics that have been quite arbitrarily excluded is the description of degenerate states and the study of Jahn-Teller distortions, the application of spin- coupled theory to electron-deficient compounds such as the boranes, a more detailed account of virtual orbitals and their use in refining the ground state wave function and the determination of excited states of molecules, intermolecular forces, electro- cyclic addition reactions, N-S heterocyclic ring systems and charge-transfer collisions in plasmas. On the other hand, there still remain many technical developments and improvements that could be incorporated into the spin-coupled codes.Since these mostly do not involve any fundamentally new theory, but straightforward extensions of known methods (e.g. gradients, see ref. 28), such develop- ments have taken second priority to the wide application of spin-coupled theory to many different types of chemical systems. It has long been considered that the use of non-orthogonal orbitals would lead to a formalism of immense complexity, which in turn would require computing resources that would make such an approach hopelessly inefficient. In fact, we see that the opposite is true: the formalism leads to a description of molecules and chemical systems that is extremely compact and highly visual, and hence long expansions of the wave function in terms of different configurations, which obscures all our vital insight, are avoided.This, finally, is the major success of spin-coupled theory. 10 References I R. B. Woodward and R. Hoffmann, Angew. Chem., Int. Ed. Engl., 1969, 8, 781. 2 W. Heitler and F. London, 2.Physik, 1927, 44, 455. 3 W.Heitler, Marx ffandh. Radiologie, 1934, 11, 485. 4 G. Nordheim-Poschl, Ann. Physik., 1936, 26, 258. 5 N. N. Greenwood and A. Eamshaw, Chemistry of the Elements, Pergamon Press, Oxford 1984. 6 C. A. Coulson and 1. H.-Fischer, Phil. Mag., 1949, 40, 386. 7 J. C. Slater, Quantum Theory of Molecules and Solids. Vol. IV, McGraw-Hill Book Co., New York, 1974. 8 R. G. Parr and W. Yang, Density Functional Theory for Atoms and Molecules, 1989. 9 M. Kotani, A. Amemiya, E. Ishiguro and T. Kimura, Tables of Molecular Integrals, 2nd edn., Maruzen, Tokyo 1963. 10 R. Pauncz, Spin Eigenfunctions, Plenum Press, New York, 1979. 11 G. Rumer, Gottingen. Nachr., 1932, 337. 12 R. Serber, Phys. Rev., 1934,45,461; J. Chem. Phys., 1934,2, 697. 13 M. Sironi, D. L. Cooper, J. Gerratt and M. Raimondi, J. Mol. Struct. (Theochem.),1991,229,279. 14 J. C. Manley and J. Gerratt, Computer Phys. Commun., 1984,31, 75. 15 P. B. Karadakov, J. Gerratt, D. L. Cooper and M. Raimondi, Theoretica Chim. Act., 1995, 90, 51. 16 G. A. Gallup, R. L. Vance, J. R. Collins and J. M. Norbeck, Adv. Quantum Chem., 1982,16, 229. 17 S. Wilson and J. Gerratt, Mol. Phys., 1975, 30, 777. 18 N. C. Pyper and J. Gerratt, Proc. Roy. Soc., 1977, A355, 407. 19 J. Gerratt and M. Raimondi, Proc. Roy. SOC., 1980, A371, 525. 20 J. Gerratt, Adv. Atomic Mol. Phys., 1971, 7,141. 21 P. A. Hyams, J. Gerratt, D. L. Cooper and M. Raimondi, J. Chem. Phys., 1994,100,4408; 1994,100,4417. 22 P. B. Karadakov, J. Gerratt, D. L. Cooper and M. Raimondi, J. Chem. Phys., 1992, 97, 7637. 23 M. Sironi, D. L. Cooper, J. Gerratt and M. Raimondi, J. Am. Chem. SOC., 1990,112,5054. 24 M. Sironi, D. L. Cooper, J. Gerratt and M. Raimondi, J. Chem. Soc., Faraday Trans 2, 1987,83,1651;S. C. Wright, D. L. Cooper, J. Gerratt and M. Raimondi, J. Chem. Soc. Perkin Trans. 2, 1990, 369. 25 D. L. Cooper, T. P. Cunningham, J. Gerratt, P. B. Karadakov and M. Raimondi, J. Am. Chem. SOC., 1994, 116,4414. 26 D. L. Cooper, J. Gerratt and M. Raimondi, Nature, 1986, 323, 699. 27 E. C. da Silva, J. Gerratt, D. L. Cooper and M. Raimondi, J. Chem. Phys., 1994, 101, 3866. 28 J. Gerratt and I. M. Mills, J. Chem. Phys., 1968, 49, 1719. 29 P. B. Karadakov, J. Gerratt, D. L. Cooper and M. Raimondi, J. Phys. Chem., 1995, 99, 10186. Received, 2nd October I996 Accepted, I7th December 1996 100 Chemical Society Reviews, 1997

ISSN:0306-0012    

DOI:10.1039/CS9972600087

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

....................... Developments in metalorganic precursors for semiconductor growth from the vapour phase Anthony C. Jones Epichem Limited, Power Road, Bromborough, Wirral, Merseyside, UK L62 3QF Inorgtech Limited, 25 James Carter Road, Mildenhall, SufSolk, UK IP28 7DE Volatile metalorganic compounds are being increasingly used for the deposition of compound semiconductors from the vapour phase by metalorganic vapour phase epitaxy (MOVPE) or chemical beam epitaxy (CBE). Developments in precursor chemistry, such as improved synthesis and purification techniques and the use of alternative precursors, have been central to the progress of MOVPE and CBE. In this paper some of these recent precursor developments are reviewed and the current thinking on gas-phase and surface mechanisms occurring in MOVPE and CBE is discussed. 1 Introduction Compound semiconductors, based on combinations of elements from Groups 111 and V and 11 and VI,? have had a significant impact on our everyday lives.Materials such as gallium arsenide (GaAs), aluminium gallium arsenide (AlGaAs), gal- lium nitride (GaN), indium phosphide (InP) and zinc selenide (ZnSe) have a wide variety of applications in satellite TV receivers, optical fibre communications, compact disc players, bar-code readers and full colour advertising displays. The devices used in these applications require single crystal films of semiconductor material and simple device structures with layer thickness > 10 pm (e.g. light emitting diodes) have been traditionally deposited from mixtures of the molten elements on an appropriate substrate, a process known as liquid phase epitaxy (LPE). However, although LPE can produce high Tony Jones trained as un organometallic chemist at the University qf Manchester, where he received both his BSc (1976) and PhD (I 979).From I9804983 he undertook post-doctoral research at the Uniiwsity of Litserpool into the development of novel synthetic routes to main-group organometallic compounds for use in MOCVD. He MUS a founder menzher of Epichem Ltd., established in I983 spcc-$cally to manufacture chemicals for the semiconductor industry. He is currently Principal Scientific Consultant to Epichem and Chief Scientist at Inorgtech Ltd. He also holds Visiting Professorships at the University of Salford and Impe- rial College of Science, Technol- ogy and Medicine London.In 1996, he received the Michael A. Lunn Outstanding Contributor Award for research on precur- sors for indium phosphide and related materials. His current research interests include the development of precursors for the CVD of compound senzi-conductors, metals, metal oxides and metal nitrides. ~~ ~ ~~ j-Non-IUPAC nomenclature (i.e.111-VJI-VI rather than 13-15, 12-16) has been universally adopted by the semiconductor industry and is therefore used throughout this review. purity semiconductor films, growth of very thin films ( < 0.1 pm) is difficult and there are problems of doping and of interface control during the growth of complex multilayer 111-V and 11-VI devices.Therefore, a great deal of effort has gone into the development of vapour phase deposition techniques, more suitable for multilayer growth. These include molecular beam epitaxy, using vapour phase mixtures of the pure metals; and metalorganic vapour phase epitaxy (MOVPE) and chemical beam epitaxy (CBE) which use vapour phase mixtures of metalorganic compounds. The application of metalorganic chemistry has played a vital role in the recent development of MOVPE and CBE semi- conductor technology. In this review the use of metalorganic and related compounds in MOVPE and CBE is discussed, with emphasis on chemical interactions occurring in the gas phase and at the semiconductor substrate surface.2 Metalorganic vapour phase deposition techniques The use of metalorganic compounds to deposit semiconductor materials was first described in detail by Manasevit.' He showed that metal alkyls of the Group 111 A elements (MR,; M = Al, Ga, In; R = methyl, ethyl) and the Group I1 B elements (MR,; M = Zn, Cd; R = methyl, ethyl) could be used as volatile precursors for growing 111-V and 11-VI semiconduc- tors. For example, solid GaAs can be deposited by the pyrolysis of trimethylgallium vapour at 600-800 "C in the presence of arsine (AsH3). Manasevit named the technique metalorganic chemical vapour deposition (MOCVD) to describe the vapour transport of the metal as a metalorganic compound. In simple form the deposition of GaAs and ZnSe by MOCVD can be summarised in Scheme 1.Scheme 1 These semiconductor alloys are generally deposited on single crystal wafer substrates so that each layer assumes the same crystalline orientation as the substrate. This process is called epitaxy and the specific process involving metalorganic pre- cursors is called metalorganic vapour phase epitaxy (MOVPE). MOVPE is an extremely versatile technique which can be used to grow a large range of mixed 111-V and 11-VI semiconductor materials, (see Table 1). The solid composition of these alloys is controlled simply by varying the concentration of the precursors in the vapour phase. MOVPE is suitable for growing large areas of highly uniform semiconductor layers and allows precise control of layer thickness, purity and doping concentra- tion.Consequently, it is fast becoming the standard technique for producing advanced 111-V devices, which often have sophisticated multilayer structures. The MOVPE process involves a series of gas phase and surface reactions, Although actually a complex process, it can be divided into several steps: (a)evaporation and transport of Chemical Society Reviews, 1997 101 precursors, (b)pyrolysis of precursors leading to the semicon- ductor material and (c) removal of the remaining fragments of the decomposition reactions from the reactor zone. The basic processes which underlie MOVPE or MOCVD processes are shown in Fig. 1. A significant feature of MOVPE is the presence of a layer of hot gas immediately above the substrate.This is termed the boundary layer, and gas phase reactions occurring in this layer play a significant role in MOVPE deposition processes, [see Fig. 2(a)]. The CBE process,2 shown schematically in Fig. 2(b)is a high vacuum process in which Group I11 metalorganic precursors are pyrolysed on a hot substrate (450-600 “C) in the presence of Group V atoms. (e.g. As2 derived from precracked AsH3). Under the high vacuum conditions of CBE, the boundary layer is absent and chemical interactions are limited to the substrate surface, which should in principle lead to more uniform semiconductor layers with sharper interfaces. Another ad- vantage of CBE is that the use of high vacuum allows the in situ monitoring of layer growth by surface science techniques such as modulated beam mass spectroscopy, reflection high energy electron diffraction, Auger electron spectroscopy etc.Developments in metalorganic precursor chemistry have been central to the progress of MOVPE and CBE. For instance, improvements in synthesis and purification techniques (e.g. adduct purification, see Fig. 3), coupled with the use of increasingly sensitive analytical techniques, have allowed the routine production of ultra-high purity metalorganics (total metal impurities < 1 ppm), necessary for the growth of device- quality semiconductor layers. These advances have been Table 1 MOVPE of 111-V and 11-VI compound semiconductors Semi-conductor Reactants Typical growth tempPC GaAs AlGaAs Me3Ga, AsH3 Me3A1, Me3Ga, AsH3 600-750 600-800 111-v InGaAs InGaAlP GaSb InP Me&, PH3 Me31n, Me3Ga, AsH3 Me&, Me3Ga, Me3A1, PH3 Me3Ga, Me3Sb 650 650 750 520 11-VI ZnSe ZnSSe CdSe CdTe CdHgTe Me2Zn, H2Se MezZn, H2S, H2Se MezCd, H2Se Me2Cd, Pr12Te Me2Cd, Hg, Pr12Te 250-350 250-350 25&350 350-400 350-400 Main Gas Flow -_______f I___t Gas Phase Reaction reviewed elsewhere3 and will not be discussed here.Another significant chemical contribution to the field of semiconductor growth has been the development of metalorganic precursors, specifically designed for the MOVPE or CBE process, which have improved gas-phase and surface decomposition character- istics, leading to higher quality semiconductor films.These precursor developments form the basis of this review. 3 Precursors for the growth of 111-V semiconductors Both MOVPE and CBE have been widely utilised for the growth of 111-V materials. Broadly speaking, there are two approaches to growing these materials; the conventional approach in which separate Group I11 and V precursors are used (see Table 1); and the use of ‘single-source’ precursor molecules in which both elements of the compound semicon- ductor are combined. Each of these approaches has inherent advantages and disadvantages, but it is worth noting that many of the original precursors introduced by Manasevit are still favoured today. The single-source precursor approach has yet to make a significant impact on the growth of 111-V electronic materials, although this method may find applications in the future, where low temperature growth is important.3.1 Growth of 111-V semiconductorsby MOVPE 3.1.I Conventional precursors In the development of the MOVPE process, the metalorganic precursors traditionally employed have been those which are readily available commercially, and which have convenient vapour pressures. These include the volatile Group I11 trialkyls, trimethylgallium (Me3Ga), trimethylaluminium (Me3A1) and trimethylindium (Me31n), in combination with the Group V hydride gases arsine (AsH3) and phosphine (PH3). To some extent, this choice was fortuitous since the 111-V layers grown contained remarkably low levels of carbon contamination, considering that carbon-containing metalorganics were em-ployed. This can be attributed to the large quantity of ‘active’ atomic hydrogen produced by the pyrolysis of AsH3 or PH3, which allows the clean removal of carbon-containing fragments from the growth surface.For GaAs grown from Me3Ga and AsH3, a mechanism has been proposed4 for carbon removal which involves the adsorption of [Ga-Me] species (produced by Me3Ga pyrolysis) on the substrate surface, (see Scheme 2). Ga( CH3)x +ASH, + Ga(CH3)x-+ASH, -1 +CH4 ? Scheme 2 e Desorption ofVolatile Transport to Surface Desorption Surface Reaction Products ofPrecursor 1 Surface Diffusion Step Growth Adsorption of I( /Film Precursor Nucleation and Island Growth Fig. 1Scheme to show the transport and reaction processes underlying MOCVD 102 Chemical Society Reviews, 1997 The vast majority of [CH3] radicals interact with ‘ASH,’ species and are removed as the stable methane molecule.However, a small proportion become more strongly adsorbed and subse- quently decompose to incorporate carbon in electrically active form (i.e. as a p-dopant) at an arsenic surface site. On the basis of infrared data under UHV conditions, it has been proposed5 that carbon incorporation proceeds via the dehydrogenation of adsorbed methyl radicals to give strongly bound carbene-like species [ =CH2], see Fig. 4. Further dehydrogenation of these species leads to carbon inclusion in the GaAs films. The concentration of carbon incorporated in GaAs grown from Me3Ga/AsH3 is extremely low (ca.0.001 atomic ppm) and the layers are high purity, typically demonstrating high electron mobilities (p > 100000 cm2 V-1 s-1 at 77 K) and low n-type residual carrier concentrations (n ca.1014 cm-3). Similarly, high purity InP, essentially free from carbon impurities, can readily be grown from Me31n and PH3. Using Me31n, purified (a)MOVPE by adduct formation (see Fig. 3) from trace metals like Si or Zn which could act as dopants, InP layers with electron mobilities as high as 300 000 cm2 V-I s-l at 77 K have been grown.6 This represents some of the highest purity 111-V semiconductor material grown by any technique and is suitable for all device applications. Therefore, with regard to the purity of GaAs and InP layers, there is little requirement for alternative precursors.However, safety and environmental considerations have stimulated re- search into safer liquid replacements for the toxic Group V hydride gases,’ and similar reasons have prompted chemists to investigate single source precursors to 111-V materials. In other areas of 111-V MOVPE, most notably in the growth of AlGaAs, the Group I11 antimonides, and Group I11 nitrides, the use of conventional precursors has sometimes proved problematic and alternative precursors have been sought. Some of these developments are highlighted below. 3 Partly-pyrolysed alkyls in the gas stream ..............................................*.......,Stagnant boundary (J, layer Surface reactions Substrate (b)CBE Group 111 Metal Organic Beam with AS2 3 I Pyrolysis at Surface Substrate Fig.2 Schematic representation of growth processes: (a)MOVPE. (h)CBE. Chemical Society Reviews, 1997 103 3.1.2 Alternative Group V sources AsH3 and PH3 are extremely toxic gases, stored in high pressure cylinders so that there is a severe risk of toxic release during transport and use in MOVPE. Much research7 has therefore been aimed at developing safer liquid alternatives which disperse more slowly in the atmosphere in the case of accidental release. Trialkylarsine compounds, such as trimethylarsine (Me3As) and triethylarsine (Et3As) are not useful precursors, as they lead to heavily carbon contaminated GaAs layers due to the absence of active 'ASH,' species necessary for carbon removal (cf.Scheme 2). Therefore, arsenic precursors which contain one or more hydrogen atoms such as the ethylarsines (Et,AsH, EtAsH2) or tert-butylarsine (Bu'AsH2) must be used. ButAsH2 is by far the most successful liquid arsenic source to date with a convenient vapour pressure (107.2 mbar$ at 10 "C) and pyrolysing at a lower temperature than AsH3; 50% pyrolysed at 3 I mbar = 1.101325 X lo2 Pa. 425 "C compared with 575 "C for AsH3. The more efficient pyrolysis of ButAsH2 relative to AsH3 allows the growth of GaAs at lower V/III ratios and the increased concentration of active 'ASH,' species on the substrate reduces carbon con- tamination.It is likely* that the pyrolysis of ButAsH2 proceeds by homolytic fission of the arsenic-carbon bond to give [C4H9-] and [*AsH2]. Subsequent radical disproportionation, recombi- nation and exchange reactions then lead to the formation of C4H8, C4HIO and to a lesser extent, CgHlg.8 In an alternative mechanism,g two competing decomposition pathways were identified. The dominant route was proposed to be an intramolecular hydrogen transfer in ButAsH2 leading to the elimination of C4HI0 and formation of [ASH] species. However, at substrate temperatures > 350 "C this is accompanied by a minor decomposition route involving the p-hydride elimination of C4Hs and the formation of AsH3. In addition to safety considerations, there are good technol- ogical reasons for seeking a replacement for PH3.The high thermal stability of PH3 (only 50% decomposed at 700 "C) in the presence of less thermally stable AsH3 leads to problems of I Starting Materials I May contain impurities eg Si, Zn, Sn Synthesis of Organometallic 1 IR3M (Lewisacid) May contain R&i, RzZn, R4Sn or other impurities and/or hydrocarbons addition of Lewis base eg DIPHOS; adduct formation R3M.L Removal of volatile non-adducting impurities under reduced pressure Mild thermal dissociation IPurified R3M isolated by distillation 1 The involatile ligand, and any more strongly ladducting species are left in the reaction vessel I Fig. 3 Principles of adduct purification *CH3 1 / Possible route to carbon incorporation ""\ / HI / -H H\ /H ___)H\ /H -/cH3 + H-c -H Ga I /c\ ii M: As or Ga Fig.4 Mechanism proposed for the decomposition of Me3Ga on a GaAs surface. After ref. 5. 104 Chemical Society Reviews, 1997 composition control in the growth of quaternary alloys such as InGaAsP The trialkylphosphines (eg Me#, Et3P) are not useful, as they are more thermally stable than PH3 and would in any case lead to increased carbon contamination The most successful alternative phosphorus source to date is tei t- butylphosphine (ButPH2)7 which is a liquid with a convenient vapour pressure ( 184 9 mbar at 10 "C) suitable for a wide range of MOVPE applications ButPH2 (LCso > 1000 ppm) also has a much lower intrinsic toxicity than PH3 (LCso = 11-50 ppm) and pyrolyses at a significantly lower temperature (50% pyroly sed at 450 "C), probably by homolytic phosphorus- carbon bond fission The reduced thermal stability of ButPH2 relative to PH3 allows the growth of InP at lower V/III ratios and leads to big improvements in the uniformity of InGaAsP As well as these advantages, ButPH2 has a more favourable gas phase chemistry than other RPH2 precursors, which prereact with Me&, even at room temperature to liberate methane and deposit a white solid [probably (MeInPR),, p~lymer],~~) (see Fig 5) Uniquely, ButPH2 undergoes little or no prereaction and this effect may be due to the large steric hindrance of the bulky teit butyl group, which inhibits the formation of gas-phase adducts such as [Me31nPH2But], likely precursors to the polymeric (MeInPR),, deposit I I I I I 0 100 200 300 400 500 TI"C Fig.5Plot of methane evolution as a function of substrate temperature for a range of orgdnophosphine precursors in the presence of Me31n (0tert butylphosphine 0cyclohexylphosphine A cyanoethylphosphine 0 benzylphosphine W cyclopentylphosphine) After ref I0 3 I3 Sinqle-souice alteinatites A wide variety of 111-V semiconducting materials including GaAs, InAs, InP, GaP and AlAs have been deposited from single source precursors I 1 These compounds potentially ex-hibit d number of advantages over conventional MOCVD precursors (I) Air and moisture stability (21) Reduced toxicity (AsH3 and PH3 are eliminated from the growth process) (111) Prereaction is limited, there is only a single source in the supply stream (11) Low temperature growth is often possible Despite these potential advantages, single source precursors have yet to be commercially developed, mainly due to the following disadvantages (I) Low volatility, which makes them difficult to use in conventional MOCVD equipment (11) Control of stoichiometry can be difficult, especially in the growth of ternary and quaternary alloys (A1,Gal -,, As, In,Gal -,, As,, PI -, etc ) (111) Polynuclear decomposition fragments may have a low surface mobility, inhibiting epitaxial growth The most widely investigated 111-V single source precursors have the dimeric form shown in Fig 6(a),although a wide range of other types have also been used12 such as the monomeric GaAs precursor [see Fig 6(h)]and the dimeric GaP single source molecule, [see Fig 6(0] But But But But \$ \/ But I Ga I\At-P\ ,P-Ar' Ga I But (c) Fig.6 Single source precursors for GaAs and GaP deposition Polycrystalline GaAs can be deposited by vacuum MOCVD using the homologous precursors [R2Ga(AsBuf2)]2 (R = Me, Et, Bun) The methyl-based precursor was found to be the most thermally stable and was not completely decomposed at 350 "C in contrast to the butyl derivative which was fully decomposed at that temperature The methyl-based precursor also gave GaAs films which were heavily carbon contaminated (1018-1 0'9 cm-3), due to the decomposition of surface methyl radicals (cf Fig 4) An important feature of the dimeric precursors [R2Ga(AsBut2)I2 is that they appear to retain the required 1 1 stoichiometry during pyrolysis This contrasts with the trimeric compounds [Me2GaAsMe2]? and [Me2GaAsPr2]? which liber- ate unwanted diarsines on pyrolysis at relatively low tem- peratures (ca 150 "C), presumably vra the fission of [Ga-As] bonds In view of the large investment in research effort and equipment development using conventional multi-source 111-V precursors, single-source alternatives are unlikely to have a significant commercial impact unless they show a distinct technological advantage One possible area of future application is in the low temperature growth of Group I11 nitrides, discussed in section 3 1 6 3 I 4 Pi ec ui soi s foi AlGaAs gi ovijth Carbon is a major acceptor impurity in AlGaAs epitaxial layers grown using Me3A1, Me3Ga and AsH3 13 This may be due to the increased strength of the aluminium-carbon bond (Al-CH3, 65 kcal mol-1) (1 cal = 4 184 J) compared with gallium-carbon (Ga-CH3, 59 kcal mol-1) which leads to an increased concentration of methyl radicals on the growth surface The minority pathway of methyl radical decomposition (cf Fig 4) becomes more significant and carbon incorporation (up to 10 ppm) in the AlGaAs layers occurs, ira the formation of surface [A1= CH2] carbene-like species Therefore, in order to lower carbon contamination, it is necessary to use A1 precursors which do not contain the [Al- Me] group Triethylaluminium (Et3Al) is a dimeric molecule with a very low vapour pressure (0 053 mbar at 27 "C) making it impractical to use in MOVPE However, the aluminium- hydride adducts trimethylamine alane, A1H3(NMe3), and dime- thylethylamine alane, A1H3(NMe2Et) are monomeric, with suitable vapour pressures, ([a 2 6 mbar at room temp ) for MOVPE High purity AlGaAs epitaxial layers, essentially free from carbon contamination, have been grown using AIH3(NMe3)14 or AIH3(NMe2Et)ls with Et3Ga and AsH3 Although carbon-containing ethyl radicals are present on the growth surface, they desorb readily by the (3-hydride elimina- tion of ethene from the growth surface, (see Fig 7) Another advantage of AIH3(NMe?) precursors is that they do not form volatile oxygen-containing impurities (in contrast to Me3AI which can oxidise to form volatile (MeO)A1Me2 species), which should allow the fabrication of higher efficiency AlGaAs optoelectronic devices Chemical Society Reviews, 1997 105 Substrate Fig.7 The desorption of ethyl radicals from an AlGaAs surface by the b-hydride elimination of ethene Disadvantages associated with the use of A1H3(NR3) pre- cursors in MOVPE are their low thermal stability (they decompose to deposit A1 at ca. 170 "C) and a marked prereaction occurs with Et3Ga, leading to the formation of highly unstable alkylgallanes (e.g. Et,GaH, EtGaH,) resulting in poor AlGaAs layer uniformity.16 The prereaction problem can be avoided by the use of tert-butylaluminium (But3A1).The steric hindrance of the tert-butyl groups renders But3Al monomeric with an adequate vapour pressure for MOVPE (ca. 0.7 mbar at room temp.) and high purity AlGaAs epilayers, with no detectable carbon, have been grown using But3A1, Et3Ga and ASH?.17 Pyrolysis studies indicated that surface decomposition of But3Al occurs at ca. 277 "C, with the evolution of isobutene (see Fig. 8). This shows that the tri-tert-butyl radical is readily *I 80 -70 -?' 60-2e9 50-ctl x e0 40-Q, 30-20 -10- ". I I I I 250 350 450 550 650 TIK Fig. 8 Surface decomposition of But3AI monitored by mass spectrometry. eliminated from the substrate surface via fl-hydride elimination (see Fig.9), making But3A1 suitable for the growth of carbon- free AlGaAs, as well as high purity A1 films by MOCVD.18 Fig. 9 fl-Hydride elimination from a coordinated re?t-butyl radical to give isobutene and surface-bound hydrogen 3.1.5 Precursors for Group III antimonides The MOCVD growth of GaSb, AlGaSb and InSb is complicated by their low melting points (e.g. 525 "C for InSb compared with 1240 "C for GaAs) and the very low stability of stibine, SbH3, which decomposes at room temp. Consequently, a range of 106 Chemical Society Reviews, 1997 trialkylstibine precursors, SbR3 (where R = Me, Et, Pr', vinyl, alkyl) have been investigated.'g The lower alkyls have rela- tively high thermal stabilities, incompatible with the low melting points of the antimonide alloys (e.g.Me3Sb is only 50% pyrolysed at 525 "C). However, triisopropylstibine (Pri3Sb) is a promising low temperature source (pyrolysing some 200 "C lower than Me3Sb). Unfortunately, although InSb can be grown at temperatures as low as 300 "C, from Pri3Sb and Me31n20 the low vapour pressure of Pr13Sb (0.5 mbar at 25 "C) makes it difficult to use in conventional MOVPE. The mixed alkyl source ButSbMe2 has a higher vapour pressure (ca.4.7 mbar at 25 "C), but unfortunately the Me31n/ButSbMe2 combination leads to poor morphology InSb epilayers with very low growth rates.21 Tris-dimethylaminostibine, Sb(NMe2)3, is a more promising Sb source pyrolysing at low temperature (50% decomposed at 340 "C), primarily by antimony-nitrogen bond homolysis.21 The new Sb(NMe2)3 source has been used with Me31n to grow InSb at 275-425 "C with much higher growth efficiency than other SbR3 precursors.It was suggested22 that Sb(NMe& may accelerate the decomposition of Me31n by a free-radical mechanism and that carbon contamination is reduced by the reaction of [.NMe2] radicals with [Me.] radicals to form the stable volatile molecule NMe3, although a more complex mechanism may be in operation (see later, Fig. 14). There is also a requirement for alternative Group 111 precursors. For instance, it has traditionally been very difficult to grow high quality AlGaSb using the conventional precursor Me3A1. The strong aluminium-carbon bond, and the absence of active hydrogen species required to remove methyl radicals from the growth surface (cf.Scheme 2) leads to heavy carbon contamination (as high as ca.lo1*CM-~in Alo ZGa, 8Sb). These hole concentrations are so high that it is impossible to n-dope the AlGaSb layers for device applications. (e.g. diode lasers). This problem has been solved by the use of But3Al in combination with Et3Ga and either Et3Sb and Me3Sb,23 which led to over one order of magnitude reduction in carbon contamination in AlGaSb compared with layers grown using Me3AI. It has also proved possible to n-dope AlGaSb epilayers grown using But3A1, opening up the possibility of producing GaSb/AlGaSb optoelectronic devices. 3.1.6 Precursorsfor Group III nitride growth The MOCVD of AlN, GaN and InN has traditionally been carried out using mixtures of Me3A1, Me3Ga or Me3In with ammonia (NH3).24 However, the high thermal stability of NH3 (only 15% pyrolysed at 950 "C) necessitates the use of high substrate temperatures (typically > 1000 "C) and high V/III ratios (e.g.2000 : 1) are needed to inhibit nitrogen desorption. This seriously limits the choice of substrate material available and the inefficient use of toxic NH3 gas requires the installation of expensive gas-scrubbing systems. Therefore, much research effort has been directed towards the development of alternative precursors which will allow the growth of Group 111nitrides at lower temperatures and reduced V/III ratios. One approach has been to use nitrogen precursors which are less stable than NH3, such as hydrazine, N2H4 (decomposes at ca.400-450 "C) and dimethylhydrazine, (MeNH)2. In combination with either Me3Al or Me3Ga these allow the growth of AlN and GaN at temperatures down to 500 "C and with low V/III ratios < 20 : 1. However, N2H4 and (MeNH)2 are both toxic and highly unstable,which has restricted their widespread use in MOCVD. 3.1.7 Single-source precursors An alternative approach to lowering the growth temperature of A1N and GaN is to use single source precursor molecules, which already contain a direct Al-N or Ga-N bond.25 A1N films, with no detectable carbon (by AES) have been grown by low pressure MOCVD at 400-800 "C from the trimeric precursor [Me2A1NH2]3,*6 (see Fig.10). The relatively low levels of carbon in the A1N films can be attributed to the removal of W Fig. 10 Molecular structure of the single source AlN precursor, (Me2AlNHzh [CH3.] radicals by the NH2 groups bonded to the A1 centre. Low temperature A1N growth (400-800 "C) has also been achieved from the parent adduct [Me3A1NH3].25 It is probable that [Me3A1NH3] pyrolyses in the hot zone of the MOCVD reactor to form [Me2A1NH2I3 species in situ prior to layer growth. In contrast, [Me3GaNH3] and [Me2GaNH2I3 pyrolyse with cleav- age of the relatively weak gallium-nitrogen bond to deposit Ga metal rather than GaN.27 A large range of other single-source precursors have been used for A1N or GaN growth. These have been extensively reviewed elsewhere28 and include; [Me2A1NHRI2 (R = Prl, But), [M(NR&] (M = Al, Ga; R = Me, Et), [Et2M(N3)] (M = Al, Ga) and [(Me2N)2Ga(N3)]2.However, although these precursors allow the growth of AlN or GaN at low/moderate temperatures (400-800 "C) they generally have only very low vapour pressures ( << 1 mbar) at room temp., which necessitates the heating of source and reactor lines and the use of high vacuum MOCVD equipment. Also, organo- metallic azides [e.g.Et*M(N3)] are of unknown stability and are possibly hazardous. It has been proposed29 that the 'amine- stabilised' organometallic azides, shown in Fig. 11, are more stable than conventional azides. These have been successfully used for the growth of AlN, GaN and InN at < 600 "C,29 although the vapour pressure of these precursors is still very low, requiring the use of high vacuum MOCVD equipment.N3 1 v 3 Fig. 11 Monomeric 'amine-stabilised' Group I11 organometallic azides used for the low temperature growth of GaN, InN and AIN. After ref. 29. In order to combine the advantages of low temperature growth associated with single-source precursors with the benefits of volatile source materials, A1N has been grown from mixtures of Me3A1 or But3Al with the primary alkylamines RNH2 (R = Prl, But)? It is likely that [Me2A1NHR]2 species (see Fig. 12) are formed in the gas phase, allowing A1N growth at temperatures between 500 and 700 "C. The films, however, contained high levels of carbon (4.7-17.0 at %), arising from the decomposition of the organic radical in RNH2.2s R--H Fig.12 Likely elimination product formed in the gas phase during AIN growth from mixtures of Me3Al and RNH2 (R = Prl, But) 3.2 Growth of 111-V semiconductors by CBE CBE is a high vacuum technique and, to prevent elemental desorption during semiconductor deposition, growth temper- atures are generally lower (450-550 "C) than those used in MOCVD.2 In order to grow good morphology GaAs and AlGaAs epitaxial layers at these low temperatures, the AsH3 precursor must be precracked (at ca. 900 "C) outside the growth chamber to give As2 species (see Fig. 3). However, this leads to a shortage of As-H species, which are needed to remove carbon-bearing alkyl fragments from the growth surface (cf Scheme 2).Therefore, the use of the 'conventional' precursors Me3Ga and Me3A1 results in severe carbon contamination in GaAs and AlGaAs grown by CBE (see mechanism in Fig. 4). This problem is only partly solved by using AlH3(NMe3) or AIH,(NMe*Et) with Et3Ga, indicating that an alternative Ga source is required. Although gallane adducts such as GaH3(NMe3) seem to be ideally suited to CBE, these are highly unstable, decomposing in a matter of days at room temperature to deposit Ga metal. A more promising approach is to use R3Ga compounds containing sterically hindered alkyl groups which can eliminate more readily than ethyl radicals from the growth surface by P-hydride elimination. These include triisopropylgallium (Pr'3Ga), triiso- butylgallium (Bul3Ga) and tri-tel-t-butylgallium (But3Ga).PrI3Ga has proved to be a highly successful new Ga source, leading to GaAs and AlGaAs layers with carbon concentrations approximately one order of magnitude less than the equivalent layers grown from Et3Ga.30 Similarly, at low substrate tem- peratures (400 "C), Bu13Ga leads to reduced carbon contamina- tion in GaAs and AlGaAs. However, at higher substrate temperatures (550 "C) both Bu13Ga and Et3Ga lead to similar carbon levels due to the elimination of the (3-methyl group from an isobutyl group. This results in surface methyl groups and increased carbon deposition (see Fig. 13). Surprisingly, But3Ga is not suitable for GaAs growth by CBE, giving growth rates which were over one order of magnitude less than those given by PrIsGa at similar source pressure^.^^ This can be attributed to steric hindrance from the bulky and rigid But group which prevents effective chemisorption of But3Ga, greatly reducing growth efficiency.Fig. 13 The pyrolysis (ca. 550 "C) of an isobutyl radical in CBE The use of the alternative As source, tris(dimethy1aminoar- sine), As(NM~~)~ also leads to a significant reduction in carbon contamination in GaAs and AlGaAs grown by CBE, even when methyl-based precursors are used.32 As(NMe& decomposes at low temperature (ca. 300 "C) so that it can be used without precracking. The surface decomposition of As(NMe2)s [Fig. 14(a)] involves (3-hydride elimination from [N-CHJ] which produces N-methylene imine and surface hydrogen.?2 This allows the removal of surface methyl radicals (generated by the pyrolysis of Me3Ga) as the stable CH4 molecule, thereby reducing carbon contamination in the GaAs and AlGaAs layers [see Fig.14(h)]. In addition to carbon contamination, oxygen contamination is a particular problem in AlGaAs layers grown at the low temperatures associated with CBE. This can severely reduce the efficiency of AIGaAs-based optoelectronic devices. This oxygen contamination can be correlated directly with traces of diethyl ether (<0.1%) in the RJGa source, deriving from the metalorganic synthesis procedure. This usually involves alkylation of a gallium trihalide GaX3 by a Grignard reagent RMgX or an alkyllithium compound RLi, dissolved in an ether solvent. Therefore, recent research3? has been aimed at eliminating Et2O from the synthesis route and the metalorganic adducts Prl3GaNR3 (NR3 = NEt,, NMe2Et) have been prepared from the reaction between GaX3 and Pr'MgBr in trialkylamine (NR3) solvent.The use of Pr13Ga-triethylamine adduct prepared in this manner, together with A1H3(NMe2Et) (also prepared by an ether-free route) resulted in improved purity AlGaAs with oxygen levels below the SIMS detection limit (ca. 4 X 1OI6 cm-3).33 Despite the advances made in the growth of GaAs and AlGaAs by CBE, there are still significant problems associated with the growth of complex indium-containing alloys such as InGaAs and InGaAsP. This is due to the desorption of R2Ga species, enhanced by the presence of In on the growth surface, which leads to strongly temperature-dependent growth rates and to severe uniformity problems.Unfortunately, all the GaR3 and GaHJ(NR3) precursors investigated to date suffer from this effect, so that alternative solutions are required, such as the precise control of temperature across the substrate wafer. 4 Precursors for the growth of 11-VI semiconductors The wide band gap Zn- and Cd-based chalcogenides have been grown both by conventional MOVPE and the single-source precursor approach. These 11-VI materials form defects and interdiffuse at temperatures above ca. 500 "C, so that low temperature growth is a major consideration. The introduction CH N ,CH3 ctN ,CH3 of nitrogen, as an active p-type dopant into ZnSe is another important issue.4.1 Conventional MOVPE The growth of ZnSe/ZnS and CdSe/CdS has proved difficult using conventional precursors such as MelZn or Me2Cd as they prereact in the gas phase with the Group VI hydride (H2Se or H2S) to deposit 11-VI material at the MOVPE reactor inlet. A possible mechanism for the prereaction during ZnSe growth is shown in Fig. lS(a). For ZnSe and ZnS growth, this problem has been largely solved by the use of adducts such as Me,ZnNEt,, which allow the growth of ZnSe at low to moderate tem- peratures (ca. 250-350 "C) without prereaction. The introduc- tion of an amine (e.,?. NEt,, pyridine) separately into the gas phase also eliminates the prereaction between MelZn and Group VI hydrides, and it is likely that amines stabilise intermediate elimination products such as [MeZnSeH] [Fig. 15(h)],inhibiting prereaction, and allowing the transport of Zn species to the growth zone.? This mechanism is strongly supported by the recent isolation and characterisation of the stable pyridine adduct of [M~Z~SBU~],~~ (see Fig.16). An effective solution has not yet been found for the elimination of prereaction during CdSe or CdS growth. Me2Cd is a weaker acceptor than Me2Zn, so that nitrogen- or sulfur-donors added to the gas phase will not form strong enough complexes with the Cd centre to inhibit prereaction. Despite the improvements in ZnSe and ZnS purity arising from the use of the Me2ZnNEt3 adduct, these materials have yet to be commercially exploited, largely due to the difficulty in obtaining p/n junctions, required for optoelectronic devices.Although ZnSe can be easily n-doped using alkyl halides (e.g. BuI), reliable p-doping has not proved possible. Nitrogen is the only viable p-dopant for ZnSe, forming a stable acceptor state on a Se site; therefore, much research has gone into methods of incorporating nitrogen in an electrically active form.J5 For successful nitrogen doping, growth temperatures must be kept low (ideally <350 "C) so that the p-type conductivity of the ZnSe layers is not counteracted by the generation of intrinsic defects. Although H2Se does allow low temperature growth, high VI/II ratios are needed for good layer morphology, making the substitution of nitrogen on a selenium site very difficult, and hydrogen deriving from the pyrolysis of H2Se is likely to deactivate any incorporated nitrogen.Thus R2Se compounds (R = Me, Et, Prl, But) have been investigated,35 with Bu'2Se proving the most suitable for low temperature growth. Un- fortunately, extensive research into a range of potential nitrogen precursors [e.g. EtJN, RNH2, EtN3, MeTSiN?, (allyl)3N] has H Fig. 14 Decomposition mechanism of As(NMe2)i. (a) 0-hydride elimination to give surface hydrogen. (h) Removal of surface methyl radicals. After ref. 33. 108 Chemical Society Reviews, 1997 H ,C"3 Q s'-'\" H/SB--Zn (7Fig. 15 ((0 A possible mechanism for the prereaction between Me2% and H2Se.(h)Possible mechanisms for the inhibition of prereaction by nitrogen donors such as pyridine. failed to demonstrate sufficiently high levels of electrically active nitrogen in ZnSe. The single-source precursor Zn[N(SiMe3)2]23h has recently shown more promise, with low temperature photoluminescence spectroscopy indicating the incorporation of significant levels of nitrogen in the ZnSe.35 However, the layers remained n-type probably due to deactiva- tion of electrically active nitrogen by hydrogen incorporated in the crystal lattice. 4.2 Single-source precursors The problems in achieving p-type doping in ZnSe coupled with the recent successes in producing commercial blue optoelec- tronic devices from Group I11 nitrides24 may force materials chemists to explore other applications for Zn-and Cd-chalcogenides, such as their use in solar cells and flat panel luminescent displays. Low temperature growth on heat-sensi- tive substrates (e.,?.polymers) may be required, for which single-source precursors may be particularly suitable. This area has been comprehensively reviewed37 and will only be discussed briefly here. Early studies showed that CdS films could be deposited from precursors such as Cd(S2PMe2)2 whilst recent research has shown that polycrystalline Zn- and Cd-chalcogenide films can be deposited from M(ER)2 compounds. [M = Zn, Cd; E = Se, S; R = 2,4,6-But3C6H2 or Si(SiMe3),]; the presence of bulky R groups inhibits aggregation of M(ER)2 units and increases precursor volatility.The other main line of research has involved dithio-or diseleno-carbamates such as [M(E2C-Fig. 16 X-Ray crystal structure of [MeZnSBut(CSHSN)12 NEt2)2]2 (M = Zn, Cd; E = Se, S). However, metal films rather than metal chalcogenide films are sometimes deposited. In contrast, the alkyl complexes [RM(E2CNEt2)]2 lead only to the deposition of the desired metal chalcogenide films.38 Once again however, it must be noted that these polynuclear compounds have very low vapour pressures and MOCVD must be carried out in high vacuum (10-2-10-6 mbar). This may restrict the widespread application of these precursors. Me Me\ \ /Me Me Me Fig. 17 Typical structure of a mixed alkyl/selenocarbamate of cadmium or zinc 5 Conclusions The organometallic chemist has undoubtedly played a major role in the development of methods for depositing semicon- ductors from the vapour phase.As well as the important developments of improved precursor purity and consistency, alternative precursors specifically designed for use in MOVPE and CBE processes have led to rapid progress, especially in the deposition of AlGaAs, Group 111-antimonides and -nitrides and ZnSe/ZnS. Significant contributions have also been made to process safety and environmental awareness by the develop- ment of ButAsH2 and ButPH2 as safer liquid replacements for the highly toxic gases AsH3 and PH3. Nevertheless, a number of problems still remain such as the need for an effective p-type dopant for ZnSe and related alloys. Also, oxygen contamination remains a problem in high Al-content alloys (i.e. those containing > 50% Al) and methods of analysing, and then removing, ppm levels of oxygen impurity need to be devised.These problems will provide continuing challenges for organo- metallic chemists and materials technologists in the future. Chemical Society Reviews, 1997 109 6 Acknowledgements I would like to acknowledge the significant contribution of my colleagues at Epichem Limited and throughout the field of compound semiconductor MOCVD. Particular mention must be made of fruitful long term collaborations with Professor P O’Brien (Imperial College of Science, Technology and Medi- cine, London, UK) and Dr J.S. Roberts (University of Sheffield, UK) 7 References 1 (a)H M Manasevit,Appl Phys Lett, 1968,12,156,(h)H M Manase- vit, US Patent, 4, 368, 098C (Published Jan 1983) 2 W T Tsang, J Ciystal Giowrh, 1990, 105, 1 3 A C Jones, Cheni Bi , 1995,31, 389 4 T F Kuech and E Veuhoff, J Cijstal Giouth, 1984, 68, 148 5 K F Jensen, D I Fotiadis and T J Mountziaris, J Ciystal GioMth, 1991, 107, 1 6 E J Thrush, C G Cureton, J M Trigg, J P Stagg and B R Butler, Chemtionics, 1987, 2, 62 7 G B Stringfellow, J Ci vstal GIow rh, 1990, 105, 260 8 P W Lee, T R Omstead, D R McKennd and K F Jensen, J Ciystal Giouth, 1988, 93, 134 9 C A Larsen, N I Buchan, S H Li and G B Stringfellow, J Ciystal Giouth, 1988,93, 15 10 H H Abdul-Ridha, J E Bateman, G H Fan, M Pemble and I M Povey, J Electiochem SOL , 1994, 141, 1886 11 A H Cowley dnd R A Jones, Anqew Chem Int Edii Engl , 1989,28, 1208 12 A H Cowley and R A Jones, Poljhedior?, 1994,8, 1149 13 T F Kuech, E Veuhoff, T S Kuan, V Deline and R Potemski, I Ciystal Gio~th, 1986, 77, 257 14 (a)A C Jones and S A Rushworth, J Ciystal Gi OMrh, 1990,106,253, (h)W S Hobs0n.T D Hams,C R AbernathyandS J Pearton,AppI Phys Lett, 1991, 58, 77 15 S M Olsthoom, F A J M Driessen, L J Giling, D M Frigo and C J Smit, Appl Phjs Lett, 1992, 60, 82 16 B L Pitts, D T Emerson and J R Shealy, Appl Phyr Lett, 1993,62, 1821 17 C A Wang, S Sahm, K F Jensen and A C Jones, J Electioii Mate) , 1996,25, 77 1 18 A C Jones, J Auld, S A Rushworth dnd G W Critchlow, J Cijsral Gio~th, 1994, 135, 285 19 S H Li, C A Larsen, G B Stringfellow and R W Gedridge, J Election Mateiials, 1991, 20, 457 20 C H Chen, Z M Fdng, G B Stringfellow and R W Gedridge, Appl Phys Lett , 1991,58, 2532 21 C H Chen, K T Huang, D L Drobeck and G B Stringfellow, J Ciystal GioMth, 1992, 124, 142 22 J Shin, A Verma, G B Stringfellow and R W Gedridge, J Ciystal Gi OM th, 1994, 143 15 23 C A Wang, M C Finn, S Salim, K F Jensen and A C Jones, Appl Pln~Lett, 1995, 67, 1384 24 S Nakamura, Ad\ Muter , 1996, 8, 689 25 A C Jones, C R Whitehouse and J S Roberts, Chem Vup Deposi-tion, 1995, 1,65 26 Z P Jiang and L V Interrante, Cheni Mutei , 1990, 2, 439 27 S A Rushworth, J R Brown, D J Houlton, A C Jones, V Roberts, J S Roberts and G W Critchlow, Ad\ Mutei Opt Elec , 1996, 6, 119 28 D Neumeyer and J G Ekerdt, Cheiii Matci , 1996, 8, 9 29 A Miehr and R A Fircher, Pupei PDSP 5 Ahsti Book of 8th Int Conf On MOVPE, 9-13 June, 1996, Cardiff, UK 30 P A Lane, T Martin, C R Whitehouse, R W Freer, M R Houlton, P D J Calcott, D Lee, A D Pitt, A C Jones and S A Rushworth, Mutei Sci En?, 1993, B17, 15 31 A C Jones P A Lane, T Martin, R W Freer, P D J Calcott, M R Houlton and C R Whitehouse, J Cijstal Gimth, 1992, 124, 81 32 D A Bohling, C R Abernathy and K F Jensen, J Ciystal CIoMth, 1994,136, I I8 33 R W Freer, T J Whitaker, T Martin, P D J Calcott, M Houlton, D Lee, A C Jones and S A Rushworth, Adi Matei , 1995, 7,478 34 M A Mdlik, M Motevalli, J R Walsh, P O’Brien and A C Jones, I Mutei Cheni 1995, 5, 731 35 W Taudt, S Lampe, F Saurlander, J Sollner, H Hamadeh, M Heuken, A C Jones, S A Rushworth, P O’Brien and M A Malik, J Ci ystal GIOMth, 1996, 169, 243 36 W S Rees Jr, D M Green, T J Anderson, E Bretschneider, B Pathangey, C Park dnd J Kim, J Eleitiuii Mater , 1992, 21, 361 37 M Bochmann, Chem \‘up Deposition, 1996, 2, 85 38 M B Hursthouse M A Malik, M Motevalli dnd P O’Brien, J Matei Chem , 1992, 2, 949 Reteitled, 19th Octohei 1996 Ac cepred, 31d Decenzhei I996 110 Chemical Society Reviews, 1997

ISSN:0306-0012    

DOI:10.1039/CS9972600101

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Northampton, UK NN2 7AL The range of chemistries used for tanning leather is reviewed; traditional methods of tanning are explained and the newer processes are described. The areas of tanning include: vegetable tanning with plant polyphenols, mineral tanning with metal salts, in particular chromium(m), oil and aldehyde tannages, synthetic tanning agents and organic tannages based on natural polyphenols or synthetic organic oligomers. The fundamental nature of the tanning reaction and the origin of hydrothermal stability are discussed. 1 Introduction The conversion of animal hides and skins into useful artefacts may be man’s oldest technology. Untreated skins have limited value, because when wet they are susceptible to bacterial attack and so they putrefy, but if they are dried they become inflexible and useless for purposes such as clothing.Those effects are eliminated by tanning: tanning is defined as a process by which putrescible biological material is converted into a stable material which is resistant to microbial attack and has enhanced resistance to wet and dry heat. Skin is vulnerable to heat, undergoing shrinking in water at ca. 60 “C or above, an effect observed in scalding with hot water. This transition is sometimes referred to as melting, when the fibrous structure becomes rubber-like; at higher temperature the skin becomes more amorphous, familiar as gelatine, and further heat degrada- tion results in animal glue. Tanning raises the temperatures at which these changes in structure are initiated.The main protein in skin is collagen. Other proteins are present in native skin, such as keratin, albumins and globulins, but these are removed during the early stages of the leather- making process, together with other non-collagenous compo- Professor Tony Colington comes ji-om a background of a PhD in physical organic- chemistry with Professoi- R. P. Bell, then post- doctorul 1-eseai-ch in solute-solvent interactions niith Professor A. K. Coiington (no I-elation). In 1976, he M!ent to Mioi-kfor the British Leather Manufacturer-s’ Researdi Association, later BLC The Leather- Technology Centi-e of Northampton, ultimately becoming Head of the Rawstock and Tanning Dept. 1n 1995, he joined the British School of Leather Technology at Nene College, North- ampton; in 1996, he was avturded a Personal Chair in Leather Science.He is the cur-rent President of the Inter-national Union of Leather. Technologists and Chemists So-cieties (IULTCS), the technical organisation qf the world’s leather industries, covering moi-e than 30 national societies. IULTCS celebrates its centen- uiy Kith a congress in London in September 1997-address E-muil enquiries to: tonv.c.oiing- ton@ nene.ac-.uk nents, such as glycosaminoglycans, especially hyaluronic acid, proteoglycans, especially dermatan sulfate, and triglyceride fats. Removal of these non-collagenous proteins is necessary to produce soft leather. Otherwise, the fibre structure may be cemented together during drying. Removal of glycosaminogly- cans and particularly proteoglycans is necessary to allow the fibre structure to split apart; this ‘opening up’ effect allows penetration of tanning and lubricating agents, to produce the range of organoleptic properties or handle or feel required by the consumer.The fibrous structure of hide (from big animals) or skin (from small animals) is illustrated in Fig. l(a) and (b); in the Figures, the hair has been removed by dissolving it in a solution of sodium sulfide buffered with calcium hydroxide and then the structure has been opened up by prolonging that treatment at pH 12.5, this was followed by the action of proteolytic enzymes (called bating) at pH 9 to break down albumins and globulins, then pickling to pH 3 with sulfuric acid in brine prepared the hide for tanning with chromium(rr1) salts.The hierarchy of skin structure is well defined and is illustrated in Fig. 2(a) and (b): fibres are made of fibril bundles, the fibrils are the lowest level of structure that is visible in intact collagen and they are characterised by a repeating banding pattern which can be emphasised by staining with heavy metals. Disrupting the fibrillar structure by swelling it in acid reveals another level of structure, shown in Fig. 3, but its nature is not known. To date, at least 12 collagens have been identified, each performing a different function in skin or other animal tissue. I Collagen is characterised by its glycine content, one glycine at every third residue (-gly-X-Y-), its uniquely high proline content, often next to glycine in the sequence (-gly-pro-Y-) and its unique hydroxyproline content, usually next to proline in the sequence: -gly-pro-hypro-gly-.’ The presence of proline in the sequence causes the chain to twist, forming a left-handed helix.The presence of glycine at every third residue allows three a helices to twist together in a right-handed triple helix, with the glycine methylene groups situated in the centre of the structure. The presence of the hydroxyproline provides a powerful stabilising effect by hydrogen bonding. The structure is illustrated in Fig. 4. In the case of type 1 collagen, the major collagen in skin, the triple helix consists of two al(1) and one a2(1) chains, each 1052 residues long, distinguished only by minor differences in amino acid sequences. Collagen is a polar protein; type 1 collagen contains 4.4% aspartic acid residues, H2N.CH(CH2)(COOH)COOH, 7.2% glutamic residues H2NCH(CH2CH2COOH)COOH and 2.8% 1ysine residues H2N-CH(CH2NH2COOH.The pattern of charge distribution is repeated every 234 residues, although not necessarily with the same amino acid sequence; this is called the D period. The structure is further illustrated in Fig. 5(a) and (b), which are computer generated images of the model, based on the known sequences and water content, taken from the work of Brown and her coworkers at the US Department of Agriculture Eastern Regional Research Center, Pennsylvania.3 The triple helices are packed together longitudinally in a ‘quarter stagger’ arrangement, shown in Fig.6. In this way, the Chemical Society Reviews, 1997 111 Fig. 1 Photomicrographs of tanned bovine hide (a) Light photomicrograph (X40) of cross section i grain or corium major containing hair follicles ii grain-corium junction iii flesh layer iv residual muscle (meat) and subcutaneous fat (b) Scanning electron photomicrograph (x40) of corium pattern of charge is matched and gives rise to the characteristic banding pattern of stained collagen Also shown in Fig 6 is that at each end of the helical region of the procollagen molecule, there is a non-helical or telopeptide region, the natural covalent crosslinks are situated between adjacent helical and telopeptide sites Whilst there is some evidence that packing in the third dimension involves subunits of five triple helices, the current view is that packing is continuous, a distorted hexagonal closest packing The structure of collagen has recently been rigorously reviewed by Kadler 2 Tanning The chemical nature of collagen allows it to react with a variety of agents, often resulting in its conversion to leather, of the changes in appearance and properties that are a consequence of tanning, one of the more important is the increase in hydrothermal stability This can be measured by observing the point at which a specimen shrinks, when it is held in water, heated at a rate of 2 "C per minute, this is the conventionally measured shrinkage temperature, T, It is necessary to specify the conditions, because shrinking is a kinetic process and, as 112 Chemical Society Reviews, 1997 such, can be treated thermodynamically This is illustrated in Table 1, which shows the relationship between shrinkage temperature and free energy of activation at 60 "C for the shrinking transition and the enthalpy of the endothermic reaction, the specimens represented in the Table are raw collagen in the form of kangaroo tail tendon or sheepskin flesh layer, tanned with either basic formato aluminium(m) sulfate or basic chromium(rI1) sulfate The relationship between T, and AHendo indicates that breakdown of the tanning interaction is not the cause of shrinking, indeed, 27Al NMR studies6 demonstrate that even the weak, hydrolysable aluminium tannage is not reversed during shrinking The reaction which is manifested as heat shrinking is a breakdown of the hydrogen bonding in collagen or leather, Table 1 The thermodynamics of collagen shrinking in wet heat at 60 "C Raw 60 103 104 Ali" 73 112 49 Cr"' 107 139 52 Fig 2 Elements of skin fibre structure (a) A fibre composed of fibril bundles emerging from the ice surface in a cry0 SEM photomicrograph mag X 850 (b) Fibril bundles in cross section SEM photomicrograph mdg x 10000 that is, regardless of the tanning process, the shrinking reaction is the same The hydrogen bonding is illustrated in Fig 7, taken from the work of Ramachandran,7 it is this structure component that breaks down during shrinking This begs the question where does hydrothermal stability come from? That is, if the tanning process only modifies the shrinkage temperature, without changing the shrinking mecha- nism, what causes the T, to rise? The answer may have something to do with the size of the cooperating unit in the shrinking process the larger the unit, the slower are the kinetics, the higher is the shrinkage temperature 6 From measurements of rate of shrinking and entropy of activation, it was found that the size of the cooperating unit in raw collagen, with shrinkage temperature 60 OC, is 25 residues, but aluminium and chromium tanned collagens with shrinkage temperatures 73 and 107 "C have cooperating units containing 7 1 and 206 amino acid residues respectively The nature of these crosslinks or cooperating units is not clear, crosslink may function through the natural covalent crosslinks and additional hydrogen bonding structure elements in collagen, supplemented or modified by the tanning effects of hydrogen bonding or covalent crosslinking at polar groups on the amino acid sidechains or through multiple interactions at the peptide link itself Fig 3 Skin fibril structure (a) Fibrils have a characteristic banding pattern when stained with heavy metals SEM photomicrograph mag x 100000 (b) Sub fibrillar structure is revealed by acid swelling Thus, the tanning reaction may be highly complex on the molecular level, to the extent that there is no clear model of its effect on hydrothermal stability (1) T, appears to be related to the size of the artificial crosslink, which may be related to the size of the unit involved in the shrinking reaction, ([I) The effect of the tannage may be related in part to the chemistry, that is to the particular reaction sites within the protein that are involved, (zzz) High T, is achieved either by the precise effects of chromium(rI1) or by controlled multiple interaction between tanning species and collagen (see below) Current tanning technology is dominated by chromium(m), it was introduced about 130 years ago and by the turn of the century had begun to replace the traditional tannages, which were based on plant polyphenols (so-called vegetable tannins) Today, La 90% of the world's leather production is chrome tanned, the remainder is tanned with vegetable tannins, mostly for leathergoods or shoe soles Typically, chrome tanning alone is insufficient to provide the aesthetic requirements for the wide range of leather types produced by the industry, so the main or Chemical Society Reviews, 1997 113 Fig. 4 The tropocollagen molecule.Three left-handed helices are twisted together, to form a right-handed triple helix prime tannage is complemented with other tannages, which are applied after chrome. The full range of options available to the modern tanner will constitute the body of this review. 2.1 Vegetable tanning Many plant materials contain polyphenols which can be used in tanning.To be effective, the molecular mass must be 500-3000; lower molecular mass fractions in the tannin are referred to as non tans and higher molecular mass species are gums. Tanning products may be powdered plant parts or aqueous extracts of those parts; the properties they confer to the leather are as varied as the many sources from which they are obtained. Tannins are classified as follows: hydrolysable or pyrogallol tannins, sub- classified as gallotannins (examples are Chinese gallotannin or tannic acid, sumac, tara) or ellagitannins (examples are myrobalan, chestnut, oak), and condensed or catechol tannins (examples are mimosa, quebracho, gambier). Hydrolysable tannins are sugar derivatives, based on glucose, but may be larger polysaccharides; plant extracts may contain tetrasaccharides as less useful gums. Gallotannins are charac- terised by glucose esterified by gallic acid; esterification may occur directly with the glucose ring or as depside esterification of bound gallic acid (1).8 Ellagitannins have sugar cores, esterified not only with gallic acid, but also with ellagic acid (2) and chebulic acid (3).Examples of structures of hydrolysable tannins are chebulinic acid (4) from myrobalan, chebulagic acid (5) and vescalagin (6) and castalagin (7) of chestnut. The traditional way to tan with vegetable tannins is in pits, where the slow penetration of large reactive molecules can take place over a prolonged period of time; when leather quality was overseen by the Guilds, hide had to stay in the pits for 'a year and a day', a form of quality assurance. Today, that tanning period has been reduced to a few weeks.A feature of pit tanning with hydrolysable tannins is that they deposit 'bloom'; natural fermentation breaks down the tannin into sugar acids and precipitates components such as ellagic acid. The bloom has a 114 Chemical Society Reviews, 1997 G 1 G-G OG G' Chinese gallotannin (tamic add) 1 HO \ bH II0 HOOCI "9 PH Elhgicaad 2 Chebulicadd 3 filling effect within the fibre structure, useful because sole leather is sold by weight, and the organic acid salts provide a buffer against the detanning effects of sulfur oxides and nitrogen oxides in the atmosphere.This latter reaction is known as 'red rot', familiar as the cracking and disintegration of bookbinding leathers, a phenomenon once investigated by Michael Faraday. Hydrolysable tannins typically raise the shrinkage temperature of collagen to 75-80 "C. Condensed tannins are based on the flavonoid ring system, shown in Fig. 8. The A ring usually contains phenolic hydroxy groups and the presence of the C ring makes both rings reactive to forming carbon-carbon bonds; the B ring does not exhibit the same reactivity, it often contains the catechol group, hence the P Hd ChebuEdc acid 4 myrobalantannin H*COO-C& I Chebulagicacid 5 OH R' = OH, R2= H, w?scalagm6 R'= H, ~2=OH, castalagin 7 alternative name for this group of compounds.The condensed tannins are illustrated in the generalised structure (8)9 and the monomeric units of mimosa (9) and quebracho (10) tannins. The condensed tannins do not undergo hydrolysis, instead they may deposit a precipitate, an aggregate of polyphenol molecules, called 'reds' or phlobaphenes. Unlike the hydroly- sable tannins, which are relatively lightfast, the condensed tannins redden markedly upon exposure to light; this is understandable in terms of their linked ring structure and ability to undergo oxidative crosslinking. Condensed tannins typically raise the shrinkage temperature of collagen to 80-85 "C. R R = H, procyanidin Condensed tannins 8 R = OH, prodelphinidtn P mimosa 9 quebracho 10 Vegetable tannins react with collagen primarily via hydrogen bonding, as indicated in the model presented in Fig.9. This type of interaction is inferred from studies of reaction with polyamides. Also, it is known that polyphenols fix to amino sidechains by electrostatic salt links with carboxylate or hydrogen bonding with carboxylic acid groups (depending on pH). It is known that condensed tannins have an additional mechanism for reaction, because they are more resistant to removal by hydrogen bond breakers. For example, treating leather tanned with myrobalan (hydrolysable tannin) with 8 M urea removed 80% of the bound tannin, reducing the shrinkage temperature by 10-20 OC, but mimosa (condensed tannin) tanned leather treated under the same conditions lost only 50% tannin and the shrinkage temperature fell by 4-5 "C.It has been suggested that this additional interaction is covalent reaction between the protein and aromatic carbon in the tannin molecules via quinoid structures. Note that quinone itself can tan protein effectively, raising the shrinkage temperature to 90-95 "C; toxicity considerations rule the reaction out com- mercially. Modem vegetable tanning, especially for sole leather, is still conducted in pits and the procedure retains some of the traditional elements. For example, the high affinity of poly- phenols for protein means that the tannage must start in weak liquors and the hides are progressively moved through a series of pits containing increasing concentrations of the vegetable Chemical Society Reviews, 1997 115 Fig.5 Computer models of collagen. (Reproduced by kind permission of Dr Ellie Brown, USDA, Philadelphia). (a) A 40 nm segment of the microfibril (4 triple helices). Four of the five helices are coloured red: the fifth triple helix has its individual helices coloured green, blue and white. (b) The microfibril model shown with the residues coloured according to their sidechain properties: hydrophobic is green, polar is cyan, positively charged is purple, negatively charged is red. There are regions of higher and lower hydrophobicity and regions of higher and lower charge density. tannins; depleted liquors are strengthened by topping up with stronger liquors and in this way the hides go one way through the system and the liquors go in the opposite direction, the countercurrent method of pit tanning.The use of extracts, rather than the plant material itself, allows highly concentrated solutions to be employed and by warming the pits, ‘hot pitting’, the whole process takes only a few weeks. Vegetable tanning can also be conducted in rotating drums, but there is a need to make the hide less reactive, to allow the tannins to penetrate into the thick hide; the oldest method is to pretan with synthetic analogues, syntans (see below), and one of the newer methods is to precondition the hide with polyphosphate, which has a weak tanning action. In this way, vegetable tanning can be shortened to a few days. The study of plant polyphenols is an active field, not only because of application to tanning technology, but also (inter-alia) because it is a fruitful area for a wide range of products, either isolated compounds or chemically modified polyphenols: that range includes adhesives, inhibitors to fungal, bacterial and viral growth and antitumour 116 Chemical Society Reviews, 1997 Fig. 6 Staggered packing of triple helices Fig.7 Hydrogen bonding in collagen AC B Fig. 8 The flavonoid ring system, on which condensed tannins are based 6+ II-N--c-cl-t--I I k k k k3 k k4 Fig. 9 Model of hydrogen bonding between plant polyphenols and col1agen 2.2 Mineral tanning A review across the Periodic Table of the tanning effects of simple inorganic compounds reveals that many elements are capable of being used to make leather 12 But, if the practical criteria of effectiveness, availability, toxicity and cost are applied, the number of useful options is much reduced In all cases, the benchmark for comparison is tannage with chrom- ium(rI1) T, > 100 "C is easily achieved, it is readily available, with large reserves in Southern Africa, it is relatively cheap and has minimum health hazards or environmental impact From the Periodic Table Groups 1, 2, 6, 7, 8 do not tan Groups 3, 4, 5 elements of the first period do not tan The remainder have weak tanning powers, but only elements of the second period are of practical interest, of those, aluminium(r~~) has the best effect, silicates and polyphosphates have auxiliary functions in tanning Transition elements only titanium, zirconium, chromium and iron have practical possibilities Lanthanides individual elements or mixtures have moderate tanning properties So, in the whole Periodic Table there are only five elements plus the lanthanides that might find application in tanning, but only four of those play a significant role in the modem industry, summarised in Table 2 Table 2 Comparison of minerdl tanning agents Extent of complex formation (more $tars = better reaction) ('rill All11 TIIV ZrlV Oxy complexes ** **** *** * Carboxy complexes **** ** ** %** Amino complexes ** * ** ** Equivalent mass of oxide 38 25 40 61 Maximum T,PC > 120 90 96 97 The chemistries of these tanning agents are presented below Note, iron tannage was used in Germany when chromium was not available, but its effect is not unlike AlIIi, to date, tannages employing mixtures of lanthanides have exhibited similar performance to TitV 2 2 I Chionzrum(lrr) salts There is a fortuitous coincidence of reactivities in chrome tanning The reaction occurs at ionised carboxy groups, aspartic and glutamic acid sidechain carboxys have pK, values 3 8 and 4 2 respectively, providing a reaction range at pH 2-6 Chromium(1rr) forms basic salts in the range pH 2-5, although in practice the useful range is pH 2 7-4 2, where the basicity ranges from 33 to 67% Note, Schorlemmer basicity is defined by the number of hydroxy groups associated with the metal ion, relative to the maximum number allowed by the valency In that useful range, the number of chromium atoms in the molecular ion increases from 2-3 to > 3 and the availability of ionised collagen carboxys increases from 6 to 47% of the total number Hide is prepared for tanning by pickling with sulfuric acid in a solution of sodium chloride, the neutral electrolyte is necessary to prevent osmotic swelling of the collagen Chrome tanning is usually initiated at pH 2 5-3 0, using 33% basic chromium(II1) sulfate in the form of spray dried powder, obtained from sulfur dioxide reduced chromic acid During the course of the tanning process, the pH is raised to 3 54 0, causing the number of reaction sites on collagen to increase and the chrome species to increase in size Starting the process under conditions of low reactivity of both collagen and chrome favours fast penetration of chrome into the substrate, but slow reaction, increasing the pH increases the reactivity of both components of the reaction, resulting in reduced penetration rate To obtain a continuing balance between reaction rate and penetration rate IS part of the tanner's art, this is not simple, especially if the skin is thick and it is not uncommon for hedvier hides to be more than 1 cm thick in places The changes that occur in chrome species are set out in Fig 10 The formation of olated species was first suggested by Bjerrum, to explain the hysteresis tn delayed back titration of basified chromium(rI1) and the compound containing both hydroxy and sulfato bridges is well characterised l4 Further change, into the oxolated (0x0 bridged) species is postulated to happen during ageing after tanning L L 2+ 0\ /O 0/No L hexaquo-p-drhydroxy -p-sulfato dchromiun(111) oxolation11 L __I octaquo -p-0x0 -p-sulfato dchromim(I II) Fig.10 Change? in chromium(ir1) species with pH The presence of coordinated sulfate is necessary for the efficient reaction of chromium(m), l5 selective tannage with the isolated dichromium species (11) is 11 "C lower if sulfate is not present and 15 "C lower for tannage with the trimeric chromium species (12) The coordination of ligands to chromium(Irr), to modify the properties of the salt, is routinely exploited as 'masking' Monodentate ligands, especially formate, may be applied at different ligand to metal ratios, depending upon the degree of Chemical Society Reviews, 1997 117 r H~Q H20 L dimericdKorniun(III)ll trimeric ctromiun (III) 12 effect desired.Reduction of cationic charge and statistical reduction in the number of reaction sites on the molecular ion make the species less reactive to collagen, hence enhancing penetration rate. In addition, such masking can increase the pH at which the salt precipitates; in these circumstances the final pH of the tannage may be elevated beyond that of unmasked tannage, thereby enhancing the reactivity of the collagen. In this way, the reaction rate can be accelerated, but without the same effect of increasing the size of the chrome species.Masking with bidentate ligands, which are capable of crosslinking chromium ions, causes a big increase in size, resulting in a statistically higher number of reaction sites per molecular ion; dicarboxylates containing two or more methylene groups perform this function [shorter chain molecules preferentially chelate chromium(111)1, but phthalate is the salt of choice. Whichever masking salts are used, they are usually added to the tanning bath and the masking reaction proceeds at the same rate as the tanning reaction because the reactions are identical, the formation of carboxy complexes with chromium(Ir1). * The origin of the high hydrothermal stability of chrom- ium(m) tanned leather is interesting because it is unique among solo tannages.By selectively deactivating collagen carboxy groups by esterification and amino groups by acylation, Sykesl7 showed that hydrothermal stability was controlled by reaction between chromium(1Ir) and the carboxy groups, although the removal of these reaction sites did not result in zero chrome fixation. Hence, chrome fixation can occur in three ways: (i) Covalent reaction between one chromium ion and one carboxy, this is unipoint fixation. (ii) Covalent crosslinking between one chromium ion and at least two carboxys, this is multipoint fixation. (iii) Hydrogen bonding between chromium species and the protein, especially along the polypeptide backbone.It has always been assumed that the reaction which determines the high hydrothermal stability is multipoint fixation; unipoint fixation probably provides little hydrothermal stability (see Section 2.3 on oil tanning) because there are many examples of hydrogen bonding in tanning, including vegetable tanning, which confer only moderate hydrothermal stability. The effect of chromium(Ir1) on crosslinking can be calculated as follows. T, > 100 "C is achieved with a chrome content of approx. 2.5% Cr on dry leather mass. Since complexes are binuclear or bigger, 2.5% Cr = 0.5 g atom kg-1 = 0.25 mol molecular ion kg-I. 1 kg of dry collagen contains 1 mol of carboxy groups, therefore, only one quarter of the carboxys react with chrome.But, it has been estimated18 that only 10%of the bound chrome is involved in crosslinking. Therefore, only 1/40 of the carboxy groups are involved in crosslinks with chrome. Collagen contains 11.6% acidic residues (asp + glu) and, since there are 1052 residues per chain, this is equivalent to 120 residues per chain = 360 residues per triple helix. Therefore, the number of residues reacted in crosslinks = 9. Or, the number of crosslinks per triple helix = 5. The size of the cooperating unit in chrome tanned leather is calculated to be 206 residues, so that suggests the effect of a chromium crosslink extends ca. 50 residues either side of each end of the crosslink, which is almost two complete twists of the triple helix. 118 Chemical Society Reviews, 1997 The options for crosslinking are threefold, illustrated in Fig.11:19 they are intra single helix, intra triple helix and inter triple helix. From the known amino acid sequences of the a(1)1 and a2( 1) chains of type 1 collagen, the relative numbers of the crosslink types can be calculated; it is assumed that aspartic acid and glutamic acid sidechain carboxy groups are equally reactive and that a reaction can occur between two sidechains no more than three residues apart. The results are presented in Table 3; two sets of results are given, for native collagen and for completely deamidated collagen. Note, hydrolysis of amide groups on the sidechains of asparagine and glutamine does occur during the alkali treatment of hide, typically to the extent of ca.50% converted to carboxy groups. I Intra-(triple) helix cross-link Fig. 11 Model of crosslink types in metal tanned collagen. (Reproduced by permission of J. Soc Leather Techno/. Chem.). Table 3 Calculated distribution of crosslinks possible in one triple helix of type 1 bovine collagen Intra single Intra triple Inter triple helix helix helix Native collagen: aspasp 5 15 6 aspglu 15 7 15 glu-glu 13 34 18 Total, all crosslinks 33 56 39 Fraction 0.26 0.44 0.30 Deamidated collagen: aspasp 22 24 7 aspglu 38 10 26 g1u-gl u 34 52 33 Total, all crosslinks 94 86 66 Fraction 0.30 0.35 0.27 The assumption of equal reactivity may not be justified, since model studies indicate that poly glutamic acid is less reactive to A1111 than poly aspartic acid, presumably entropy controlled.20 It is interesting to speculate that, if the most likely stabilising crosslinks are between triple helices, then 50% deamidated collagen could form 52 crosslinks, but if only 10% of bound chrome is involved in crosslinking, perhaps only 10% of possible crosslinks are formed; this is in agreement with the calculation presented above.Experiments on native and vari- ously deamidated collagens do indicate a difference in tanning effects, but it is not clear whether this is due to a difference in the availability of reaction sites or to a difference in the distribution of potential crosslinking sites. Chrome fixation is accelerated by elevated temperature and pH; the higher the chrome content of the leather, the higher is the shrinkage temperature.But, the industrial requirement is to obtain high shrinkage from the minimum amount of chrome used. Studies have shown that temperature and pH do not have equivalent effects, demonstrated in Fig. 12, in which tanning effectiveness is measured by the rise in shrinkage temperature per unit of bound chrome.21It can be seen that under tanning condition of constant temperature and pH, rise in shrinkage temperature is controlled by pH. Furthermore, tanning effec-tiveness is better when low basicity chrome salts are basified during tanning, than if moderate or high basicity salts are employed at the beginning of tanning; this may reflect the crosslinking reactivity of species polymerising in situ, com-pared to starting the tanning reaction with large, olated 30.1 25 OC, pH 4.0 25 2 25 OC, pH 3.5 3 25 OC, pH 3.0 h 0P 20 . 4 35 OC, pH 3.5 0 8 \0 ov u) 15 a,c a,>.-+ 0a, % 2 P a,0, C 10 5- OJ 0 1 2 3 4 5 Chrome in leather (YOCr,O,) Fig. 12 Tanning effectiveness as a function of tanning conditions. (Reproduced by permission of J. Am. Leather Chem. Assoc.). Closer examination of the results for tanning at 25 "C and pH 3.5 reveals that there is a maximum effectiveness of tanning, see Fig. 13. Optimum tanning occurs at an offer of 0.7% Cr203, although this corresponds here to a shrinkage temperature of only 77 "C.Here, 4% Cr203 in the leather corresponds to a shrinkage temperature of 107 "C; this chrome content is the 20 , OJ 0123456789 Chrome in leather (YOCr203) Fig. 13 Tanning effectiveness as a function of chrome fixation industry standard to provide a degree of guarantee that the leather will withstand boiling water for at least 2 min. Chrome tanned leather is highly versatile, largely due to the low level of tanning agent needed to achieve the desired stability. This means that the variety of retanning materials which might be applied to the part processed leather can produce a wide range of final products; indeed, from any one chrome tanned cattle hide, it is possible (at a pinch) to produce a sole leather or combat boot upper leather or softee shoe upper leather or upholstery leather or garment leather, all as full grain or suede leathers.The part processed chrome tanned leather is called 'wet blue', because it is wet and it is blue. The colour comes from the protein carboxy complexes of chromium(II1); the colour also depends on whatever masking salts are used, typically it is bright pale blue, but it can range from pale green to purple. This does introduce some difficulties regarding dyeing, but it has not restricted the colour range of fashion leathers to any great extent, with the possible exception of pure pastel shades. 2.2.2 Alurniniurn(III) salts The use of potash alum in leathermaking is almost as old as leathermaking itself; it is known that the Egyptians used it 4000 years ago, because written recipes survive.Throughout tanning history, alum was often used in conjunction with vegetable tannins (see below); for example, in medieval times, Cordovan leather (from Cordoba in Spain, hence the name cordwainer, meaning shoemaker) was in widespread use in Europe, made by vegetable tanning then dyeing with cochineal. Used by itself, alum (solution pH 2) interacts only weakly with collagen, scarcely raising the shrinkage temperature and having little leathering effect. However, in a mixture of water, salt, flour (to mask the aluminium ion and fill the fibre structure) and egg yolk (the lecithin content is an effective lubricant), skin can be turned into a soft, white, leathery product, traditionally used in the past for gloving.But, even in this case, the shrinkage temperature is not raised (hence, it is possible to discriminate between leathering and tanning) and the aluminium salt can be washed out of the leather if it gets wet; for these reasons, this process is called 'tawing', to distinguish it from tanning. The reaction sites for aluminium(Ir1) are the collagen carboxys, but unlike chromium(II1) to which it bears a superficial resemblance in a tanning context, A1111 does not form defined basic species nor does it form stable covalent com-plexes with carboxy groups; that interaction is predominantly electrovalent, accounting for the ease of hydrolysis. The reaction can be optimised for tanning by modifying aluminium sulfate with masking salts, such as formate or citrate, and basifying the tannage to pH 4, close to the precipitation point.(There is a rule of thumb in tanning technology, that any metal salt has its greatest tanning effect just before it precipitates). In this way, reversibility of tannage is minimised and shrinkage temperatures as high as 90 "C can be achieved. Basic A1111 chlorides are also well known in leathermaking and several commercial tanning formulations are available. As solo tanning agents, they are slightly superior to salts based on the sulfate. However, the leathering effect of aluminium(II1) is inadequate, producing firm leather, which may dry translucent due to the fibre structure resticking. Therefore, as tanning agents, alumi-nium(rr1) salts have limited value.Where aluminium salts are useful is their ability to accelerate chrome tanning; this is demonstrated in Fig. 14.22. Following a pretreatment with aluminium(m), the incoming chrome dis-places it, shown in Fig. 15. It is known that the rate of exchange of solvate ligands is 106times faster at A1111 than at CP, so it is possible to postulate that AP can react quickly with other ligands, including collagen carboxys, in a loose association that is entropy favoured or not disfavoured. This allows orientation of the carboxy bearing sidechains, from lowest energy con-formation to conformation for reaction with metal salts, including crosslinking. When chromium(II1) enters the system, Chemical Society Reviews, 1997 119 0.15 ,103 -C: b 5 0.5 0: 110 OC 0.25 -110 OC 105 "c 100 OC 0 I I I I I 2.0 2.25 2.5 2.75 3.0 Cr,O, offer (Yo) Fig.14 The influence of aluminiurn(II1) and chromium(rr1) offers on the shrinkage temperature of rinsed bovine wet blue. (Reproduced by permission of J. Soc. Leather Technol. Chem.). .-C I 3 1 2 3 4 5 Time in chrome tan (h) Fig. 15 The displacement of bound A1111 from hide by CrrIr at pH 3.840; mineral offers were 0.25% A1203 then 2.0% Cr203. (Reproduced by permission of J. SOC.Leather Technol. Chem.). the activation barrier is lower, because the reactants are effectively shifted along the reaction coordinate, so the reaction is accelerated. Since aluminium(rI1) is present only at low concentration and plays a facilitating role, this process appears to be catalytic.A similar effect has been observed for pretreatment of collagen with ethanolamine.23 The mechanism is more specu- lative, but paper chemistry interactions between ethanolamine and the collagen carboxys can be proposed (13), operating in an analogous way to that argued for aluminium(1Ir). 13 Aluminium salts are not much used in the tanning industry because of regulatory pressure and the perception of toxicity; the latter influence is fuelled by the (continuing, but misinfor- med) association with Alzheimer's disease and accidental poisoning, such as the recent experience at Camelford, UK. 120 Chemical Society Reviews, 1997 2.2.3 Titanium(1v) salts In tanning terms, the chemistry of titanium(1v) salts lies somewhere between AllI1 and CrlI1.Empirically, the chemistry is dominated by the titanyl ion Ti02+, but the species are chains of titanium ions bridged by hydroxy and sulfato ligands,24 like CrlI1. However, the coordinating power is weak with respect to carboxy complexation, so the interaction is more electrostatic than covalent. The traditional use for titanium(1v) in tanning was in the form of potassium titanyl oxalate, to retan vegetable tanned leather for hatbanding, a product for which demand has reduced in the latter half of the 20th century. Titanium solo tanning is only moderately effective, because large quantities are required to achieve the highest shrinkage temperatures, > 95 OC, but this causes the leather to be overfilled, although remaining soft.In addition, high hydro- thermal stability is only achieved when the collagen is pretreated with phthalate, possibly resulting in interaction of the following type: Collagen-NH~+----0~C~C~H~~C0~H----S0~[Ti(OH)2],S0~-Also, Russian work has shown that the tannage is best basified with a mixture of hexamethylenetetramine and sodium sulfite; the mechanism is not understood. An advantage of tanning with titanium(rv) is that it is a colourless tannage and therefore makes white leather. Hence, it has found application in tanning sheepskins with the wool on; this is a problem area, firstly because the use of chromium(I1r) produces a discolouration by reaction with the partially degraded keratin at the weathered wool tips and secondly because the reaction must be conducted using high solution to skin ratios, to avoid tangling (felting) the wool.However, one of the problems of using TiIV salts is their tendency to hydrolyse and precipitate in dilute solution. A comparison between the properties of A1111 and TiIV is in Table 4. Table 4 Tanning effect Weak Weak Stability to hydrolysis Good Poor Filling effect Poor Good Optimum tanning pH 4 4 It was argued that a mixture of salts might produce a tanning complex of better value than the individual salts. It was found that mixtures of the metal sulfates could be stabilised against hydrolysis at pH 4 by complexing (masking) with gluconate [HOCH2(CHOH)4C02-] and in this way the mixed salt could be used to tan white sheepskin rugs; in more concentrated tanning solutions, shrinkage temperatures as high as 95 "C can be achieved.It is known that A1111 and TiIV salts can form mixed complexes in which the ions are bridged by hydroxy and sulfato ligands, but they still interact with collagen in a primarily electrostatic manner, just like the individual ions. 2.2.4 Zirconium(1v) salts The development of zirconium tannage is relatively re~ent,~5 but it soon gained industrial acceptance. From its position in the Periodic Table, ZrIV might be expected to display similar tanning properties to Ti1". A comparison of those properties is set out in Table 2.It can be seen that the tanning power exceeds that of AP, but in no way matches CP. Whilst the tanning effects of Zrlv and TiTV are similar, the chemistries of their salts are different. Zirconium(1v) salts are characterised by eight- coordination and high affinity for oxygen, resulting in a tetrameric core structure; the basic unit of structure is four Zr ions at the comers of a square, linked by diol bridges, above and below the plane of the square, (14). By hydrolysis or basification, the tetrameric units can polymerise, by forming more diol or sulfato bridges. In this 14 way, zirconium species may be cationic, neutral or anionic and large ions can form. So, tanning may involve all the polar sidechains of collagen, those bearing carboxy, amino or hydroxy groups.Hydrogen bonding via the hydroxy groups in the ZrIV species is an important feature of the tanning reaction; together with the filling effect by the big molecules, the overall tanning effect is somewhat similar to tanning with plant polyphenols, hence zirconium tanning has been referred to as the inorganic equivalent of vegetable tanning. Zirconium(1v) is not often used as a solo tannage, partly because of its indifferent effectiveness at raising the shrinkage temperature and partly because the acidity of the salts and the vulnerability to hydrolysis mean they must be applied at high concentration and at pH < 1, therefore running the risk of osmotic swelling in the hide; its main use is for retanning, to fill and firm (tighten) the grain or to make better suede.2.3 Oil tanning The familiar wash leather (chamois or ‘chammy’) is tanned with unsaturated oi1;26 the preferred agent is cod liver oil. Useful oils contain fatty acids, either free or as glycerides, which are polyunsaturated; the degree of unsaturation is critical, because if there is too little unsaturation the oil will not oxidise readily and therefore function only as a lubricant, if there is too much unsaturation the oil will crosslink itself and harden with oxidation, like linseed oil. In this tannage, sheepskins are processed in the normal way to the pickled state, when they are then swollen by the osmotic effect in water, so they can be split more easily.Usually the skins used for chamois have inferior or damaged grain surfaces and this portion of the skin is discarded. The flesh splits are then treated with the cod oil; a traditional method of forcing the oil into the wet pelt was to use ‘fulling stocks’ or ‘kickers’, in which the oil was literally hammered into the skin by wooden mallets, nowadays rotating drums are sufficient. Blowing warm air into the vessel serves two functions: the skins are dried a little, to aid oil penetration and autoxidation of the oil is initiated, which is the basis of the process. The actual nature of the tannage is not known, except for the following observations: (i) The unsaturation decreases. (ii) Peroxy derivatives are formed. (iii) Hydroxy function appears.(iv) Acrolein, CH,=CHCHO, is produced. It is thought that the tannage may be due in part to an aldehyde reaction (see below) and to polymerisation of the oil; the presence of the latter effect could account for the difference between the characteristics of oil and aldehyde tanned leathers. The situation is further complicated by the observation that oil tanning hardly raises the shrinkage temperature of collagen; so this is a leathering process rather than a tanning process, based on the accepted criteria of tanning. Oil tanned leather exhibits the interesting Ewald effect: if the leather is heat shrunk in water at 70 OC, but immediately placed in cold water, it rapidly relaxes to regain ca. 90% of its original area and this is repeatable.(Normally, heat shrinking is an irreversible phenomenon). Furthermore, if the wet leather is held under tension whilst it is being heat shrunk, the dried leather remains soft and flexible, unlike other leathers which may come hard and brittle. This process, known as ‘tucking’, is used to mould leather to a desired shape, but keeping its feel. Note, heat moulding can be applied to vegetable tanned leather, when it is known as cuir bouille; here, the collagen is partially gelatinised, causing the fibre structure to be glued together when it is dried and thereby producing a hard, inflexible, non-porous material, which can be used for a variety of purposes, including drinking vessels. The most remarkable feature of oil tanned leather is its hydrophilicity, surprising considering its tanning process.A well tanned chamois leather is expected to take up at least 600% water on its dry mass and to be hand wrung to 180%. Also, this must be repeatable after drying. In use, no grease must be exuded to cause smearing. A synthetic version of oil tanning is to use a sulfonyl chloride, which reacts predominantly with the amino groups on colla-gen: Collagen-NH2 + C16H&02C1 + Collagen-NH-S02.C 16H33+ HCI Clearly this is not a crosslinking reaction, so it is not surprising that the shrinkage temperature is not raised by this tannage; but there is a powerful leathering effect and the product exhibits similar properties to oil tanned leather. It was pointed out that acrolein is produced during the oil tanning reaction; indeed, acrolein itself can be used to make a leather similar to oil tanned.But, although it is not used itself for this purpose (for toxicity reasons), it is used indirectly, as a component of wood smoke. The traditional method of preserv- ing hides and skins used by the plains dwellers, such as the North American Indians and the Mongols, is to use brains tanning. In this process, the animal brain is partly cooked in water, so it can be mashed into a paste, which can be worked into the pelt. The leathering effect turns the skin into a soft, open structured leather, buckskin, largely due to the lubricating power of the phospholipids of the brain. The Sioux Indians have a saying: ‘every animal has enough brains to tan its own hide’.The leathering effect is serviceable, as long as the pelt is not rewetted, because then it will harden on drying due to the fibres resticking. To make the leather resistant to wetting, the solution is to smoke it over a wood fire; the multiplicity of free radical and other reactions do not adversely affect the handling qualities and they are made permanent. 2.4 Aldehyde tanning 2.4 .I Formaldehyde tanning The archetypal aldehyde tannage is with formaldehyde, proba- bly most familiar in preserving biological specimens or in embalming. Reaction occurs primarily at amino groups: Collagen-NH2 + HCHO + Collagen-NH-CH20H The N-hydroxymethyl group is highly reactive and crosslink- ing can occur at a second amino group: Collagen-NH-CH20H + H2N-Collagen +, Collagen-NH-CH2-HN-Collagen In this way, the shrinkage temperature can typically be raised to 80-85 “C.However, the crosslinking is relatively inefficient, probably because the formaldehyde species are not monomeric. Amongst the species formed in solution is paraformaldehyde, HOCH,(CHOH),CHO. The presence of polyhydroxy species and their reaction with skin produces a white, spongy, hydrophilic leather, although the absorptive property of oil tanned leather is not matched. The health and safety implications associated with formal- dehyde mean that its use as a tanning agent is effectively banned. The only remaining common functions are for fixing casein, which is fused to the grain surface in glazing operations, or to impose permanent straightening to the wool of rugskins or Chemical Society Reviews, 1997 121 sheepskin clothing leathers, by reacting with the keratin under conditions of wet heat and tension.2.4.2 Glutaraldehyde tanning Of the many mono-and multi-functional aldehydes which might be used for tanning (and all can be made to work), only glutaraldehyde and its derivatives have found commercial acceptance, with the possible exception of the more expensive starch dialdehyde. The reactions of glutaraldehyde are set out in Fig. 16. The crosslinking options are wider than for simple aldehydes, but the result is the same, a shrinkage temperature of 85 "C maximum. In the same way that formaldehyde is not a simple species in solution, glutaraldehyde is polymerised, shown in Fig.17. The terminal hydroxy groups of the polymer are active and capable of reacting with amino groups. The polymer itself can interact with the collagen peptide links by hydrogen bonding via the alicyclic oxygens and so the leather is given its spongy, hydrophilic character. Collagen-N=CH (CY)3CH=N-Collagen T T I Collagen-NH-CH-NH-CollagenI (CY)3-CHO Fig. 16 The tanning reactions of glutaraldehyde Fig. 17 The reactivity of glutaraldehyde Tanning with glutaraldehyde itself confers a yellow-orange colour to the leather, which is undesirable. Several attempts have been made to modify the chemistry, to prevent colour development, including making the monobisulfite addition compound or hemiacetals, but none has been totally successful.Glutaraldehyde is coming under scrutiny with regard to health and safety implications, so it too may have to be phased out of the tanners' options. 2.4.3 Oxazolidine tanning An alternative to aldehyde tanning, but which retains the essential reactions, is to use oxazolidines, developed less than 20 years ago.27 These compounds are alicyclic derivatives of an amino alcohol and formaldehyde (19, (16); under hydrolytic conditions, the rings can open, to form an N-hydroxymethyl compound (17), which can react with one or more amino sites, in an effective though acridly odiferous tannage. CH3I cH3-c-cH,II HN 0\/w2 4,4-dimethyl-l,3-oxazolidine 15 RII HOYC CyOH 5-(ethyl or hydroxyethy1)-1- 1,3-dihydroxy-2-(ethyl or aza -3,7-dioxabicyclo[3.3.0] hydroxyethy1)-2-N,N bis octane 16 (hydroxymethyl) propane 17 An important effect of these agents is their influence on the chrome tanning reaction, promoting the fixation; it is not known whether this is a function of the chemistry analogous to ethanolamine (see above).2.4.4 Active hydroxy compounds The reactivity of the carbon in N-hydroxymethyl compounds is not the only example of an active hydroxy. Similar reactivity has been observed in the following compounds: [P(CHzOH>,]+ (S042- or Cl-) (HOCH2)3CN02 te traki s-h y drox yme thy1 tris-hydroxymethyl phosphonium sulfate or chloride nitromethane The phosphonium salts are available as bactericides (be- cause they tan the bacteria), but they are not available in sufficient quantities to be industrial tanning agents (see below).So far, the nitromethane derivative is only a chemical curiosity. 2.5 Syntans The term syntan means synthetic tanning agent. This class of tanning agents was introduced early this century, with the purpose of aiding vegetable tanning, although the range of reactivities currently available means that they may serve several different functions. They are classified into three types, according to their primary properties. 2.5.1 Auxiliary syntans These compounds are frequently based on naphthalene and are synthesised by the 'Nerodol' method, i.e. the base material is sulfonated to a high degree and then may be polymerised by formaldehyde, illustrated in Fig.18; the products are usually relatively simple chemical compounds. The presence of the sulfonate groups means that these compounds can interact strongly with the amino sidechains of collagen at pH < 6: Collagen-NH3+----03S-Syntan In this way, reaction sites for vegetable tannins can be blocked (see Section 2. l), promoting penetration through the hide cross section. At the same time, they serve to solubilise the aggregated phlobaphenes of condensed tannins, thereby reduc- ing reaction with the hide surfaces. Similarly, they can disperse acid dyes (most commonly used in leathermaking) and reduce the reactivity of the leather to dyeing, producing more level colouring.A further function of these simple reagents is to act as non- swelling acids for pickling to low pH, to avoid osmotic 122 Chemical Society Reviews, 1997 material is polymerised with formaldehyde and then the product may be partially sulfonated, as illustrated in Fig. 19. F"&yo3, Fig. 18 The Nerodol synthesis of syntans swelling. The auxiliary syntans are characterised by their low tanning power. Some may have tanning properties, by virtue of their phenolic hydroxy content, from the base material. The tanning power is a function of the number of phenolic hydroxy groups in the molecule, and the degree of swelling they produce in pickling is inversely proportional to the rise in shrinkage temperature conferred by the 'non-swelling acid', demonstrated in Table 5.From Table 5, there is no clear dividing line between non swelling acids and auxiliary syntans or the next class of syntans, combination syntans. 2.5.2 Combination or retanning syntan These syntans are usually based on simple phenolic compounds, they are synthesised by the 'Novolac' method, i.e. the base Table 5 Examples of non-swelling acids/auxiliary syntans (Ts = shrinkage temperature) Acid pH T,/"C Swelling (%) None 6.3 66 100 bS04 1.2 43 235 1.8 49 198 1.8 58 193 *CH2-s03H 2.1 65 62 5S03H S03H Fig. 19 The Novolac synthesis of syntans The products are more complex than the auxiliary syntans, having higher molecular masses, and may be crosslinked in two dimensions.Their enhanced tanning functionality means that they can confer hydrothermal stability and their larger molec- ular size means that they can have a filling effect. (Note, the fibre structure of a hide or skin varies over its area, being markedly looser in the belly region; part of the tanner's art is to make the non-uniform raw material into a uniform leather and this is accomplished in part by filling up the interstices of the fibre weave). Because they are relatively small polymers, with conse-quently weak tanning power, these syntans work best as retanning agents; they are applied after main chrome tannage, to modify the handling properties of the leather. 2 S.3 Replacement syntans By increasing the tanning power of syntans, the agents may be classified as replacement syntans, by which it is meant that they could replace vegetable tannins.These syntans can be used for solo tanning, because their properties of tanning are comparable with plant polyphenols. Again, there is no clear distinction between the retanning syntans and the replacement syntans, the difference lies in the degree of the effects. Base materials for syntans can range from the simple to the relatively complicated (18)-(20). In addition, the bridging groups may be more diverse, including dimethyl methylene, ether, urea. They rely less on sulfonate groups for their reactivity, but synthesis by the Novolac method may incorporate some sulfonic acid func- tionality. OH 1.7 62 112m2cDs03HHO2C 2.3 61 101m03H 1.8 60 114HO3S /@3rO3" 2.1 68 63 1.7 63 87 0 2-naphthol resorcinol bis (4-hydroxyphenyl) sulfone 18 19 20 The replacement syntans vary in their effects on leather, but can produce properties similar to vegetable tannins, including raising the shrinkage temperature to 80-85 "C.They are still used to prepare hide to receive vegetable tannins, though they can be used in their own right, to make leather that is more lightfast than vegetable tanned leather; a common use is for making white leather. 2.6 Organic tanning In the modem leather industry, the preferred method of tanning is to use chromium(rI1) salts, because they are versatile, Chemical Society Reviews, 1997 123 effective, easy to apply and have low environmental impact.Despite the evidence to the contrary, the latter aspect of chromium(m) is not wholly accepted by regulatory authorities (with the exception of the USA) and consent limits for discharge of chromium are increasingly stringent; discharge limits of < 5ppm Cr (< M) are typical. Consequently, although there is a commitment to continuing to use chrome, the industry is constantly searching for a viable alternative. The ideal tannage to rival chrome tanning should incorporate the following features: (i) High hydrothermal stability, T, > 105 "C. (ii) No metal salts. (iii) White or pale coloured leather. (iv) Lightfast. (v) Low environmental impact. (vi) Comparable cost. The most difficult criterion to achieve is that specifying shrinkage temperature, which has hitherto been impossible to achieve with organic compounds alone.However, recent developments have indicated that the target is achievable. 2.6.1 Semi metal tanning The only established organic tannage capable of producing leather with high hydrothermal stability is that in which the collagen is first tanned with vegetable tannin, then retanned with a metal salt, preferably aluminium(m); semi alum leathers made with condensed tannins typically have shrinkage tem- peratures ca. 90 "C (mimosa is exceptionally higher, possessing pyrogallol groups), whilst semi alum leathers made with hydrolysable tannins have shrinkage temperatures of 115-120°C. There is a correlation between shrinkage temperature and the presence of the pyrogallol group.This effect can be seen even by treating collagen with catechol or pyrogallol themselves, then retanning with aluminium(Ir1); resulting shrinkage tem- peratures are 71 and 98 "C respectively and the same pattern is obtained for more complex polyphenols. The synergistic interaction between the polyphenol and the aluminium(III) may arise from one of the following options: Collagen-Al-veg-Al-Collagen Collagen-veg-Al-veg-Collagen Collagen-veg-Al-Collagen It is known that applying the aluminium salt before the vegetable tannin produces only moderate shrinkage temperat- ure, characteristic of aluminium alone. Therefore, the first and third options are unlikely.The most probable mechanism is for the aluminium(m) to crosslink the vegetable tannin.28 In effect, the crosslinking polyphenol on collagen is itself crosslinked, to form a matrix within the collagen matrix, to stabilise the collagen by a multiplicity of connected hydrogen bonds in the new macromolecule (21). 21 Because of the importance of the presence of pyrogallol groups, semi metal tannages are confined to the hydrolysable tannins and the condensed tannin mimosa. Many metals can perform this function, depending on the affinity for phenolic hydroxy, but aluminium(II1) is the best. It is probable that this tannage is thousands of years old, because of the availability of native potash alum, which must have been used together with vegetable tannins, if only for the purposes of mordanting for dyeing.2.6.2 High stability tannages based on natural polyphenols The alternative to reaction at the phenolic hydroxy groups is to exploit the reactivity of the A and C rings of the condensed tannins; this is illustrated in Fig. 20. Aldehydes other than formaldehyde can perform this crosslinking function, although at much slower rates; some of those reactions are important for making for example wood glues. This type of chemistry has been used in tanning, but the perceived benefit was to reduce the leachability of vegetable tannins from leather. PH OH OH Fig. 20 Recently,29 it has been shown that specific reactions are potentiall; commercially useful, demonsirated in Table 6, in which crosslinkers are applied to vegetable tanned leather (East Indian buffalo calf, tanned with an unknown mixture of locally avail able plant extracts) .Table 6 Shrinkage temperatures of EI buffalo calf retanned with 10% crosslinker on dry weight Shrinkage temperaturePC Reaction temperaturePC 20 40 50 Control, no crosslinker 84 Phosphonium salta 94 93 93 Gluaraldhyde derivative 90 91 92 Oxazolidineb 100 105 110 a See Section 2.4.4. h See Section 2.4.3. The reason for using elevated temperature is because the usual control parameter of pH is not allowed; aldehyde or other active hydroxy tannage is accelerated by pH > 7, but vegetable tannins are stripped from collagen at pH > 6, alternatively, vegetable tannins are firmly fixed at pH < 4, but aldehyde tannage is very slow at pH < 6.Hence, the reaction can only be driven by heat. The most effective and most temperature dependent reaction is crosslinking with oxazolidine. This 124 Chemical Society Reviews, 1997 2 h.4 Synthetic organic tanning The advantage of tannages based on plant polyphenols is that the reagents are obtained from a renewable resource, but the primary disadvantage is the presence of the natural components of the extract that do not contribute to the chemistry of the process, which means that the controlled. Clearly, one option would be to isolate the preferred species and that is a additional cost. A better option might precisely the required properties of the organic system and to synthesise the reagents.Recently, a major breakthrough was claimed;*' it is based on the use of melamine-formaldehyde polymers, shown in Fig. 22. It was found that oligomers could be further reacted in situ on collagen with additional crosslinker, to raise the shrinkage temperature to previously unachievable chemistry is analogous to the crosslinking of vegetable tannins, insofar as the pH requirements for the components are mutually exclusive and the reaction must be temperature. Not all crosslinkers will work equally well in this reaction; for example collagen tanned with 10%melamine resin (Granofin MH ex Hoechst) then reacted overnight at 50 "C with glyoxal or oxazolidine had a shrinkage temperature of 106"C, but crosslinking with the organic phosphonium salt produced 112 "C.Not all melamine resins work, indeed most of the available commercial products yield only moderate shrinkage temperat- ures, even in the presence of the preferred crosslinker; this is a tanning option that is well known. However, this new tannage depends on a property of melamine relatively small particles and it is in this form that they usually react with 8 7 6 5 +-sumac t-myrobatan4 +vatonla 3 2 1 9 0 1 -1c-r" u -2 -3 -4 -5 -6 -7 -8 -9 -10 0 12 3 4 5 8 7 8 91011121314 Oxazolidine offer (Yo) Fig. 21 (a), (b) Change in shrinkage temperature showing the interaction between tannage and retannage above the additive effect: hydrolysable tannin tannage (20% offer), then oxazolidine retannage.(c) Change in shrinkage temperature showing the interaction between mimosa tannage and oxazolidine retannage above the additive effect. reaction has been further studied,'O to determine the magnitude collagen. The hitherto unobserved effect of high temperature of the synergistic effect, shown in Fig. 21(a-c). processing is to break those aggregates, so that the polymer will This is the first time it has been demonstrated that it is react in the form of the oligomers. This is demonstrated in possible to achieve a commercially exploitable organic tannage, Fig. 23, in which the effect of reaction temperature on polymer conferring high hydrothermal stability to the leather, which can particle size and consequent shrinkage temperature clearly then withstand the rigorous requirements of modern shoe- exhibits a preferred particle size range of 50-60 nm.making. The major shortcoming of this leather is its hydro- Furthermore, the particle size effect outweighs even the philicity, due to the presence of the vegetable tannin. However, melamine :formaldehyde ratio, since the relationship between by applying fluorochemicals, it has been shown that adequate shrinkage temperature and particle size is the same for water and soil resistance for shoes can be obtained. melamine to formaldehyde ratios of 5 or 7, typically used in commercial products. I NHI Fig. 22 Melamine-formaldehyde polymers 120 I I 115 110 0< 105 I2 100 95 1 I II 40 50 60 70 Tanning temp./"C Fig.23 The effects of reaction temperature and particle size (nm) on shrinkage temperature 18 17 18 15 14 *mimosa 13 +gambier -0-quebrrcho 12 0 reaction is not precisely possibility for the future, but at an be to target more high values. The driven by elevated resins to aggregate; oligomers clump together to form larger 20 (b)18 I1 L* 4 108 '.,.-----.-. 9 -1 I 8 7 6 5 0 12 3 4 5 8 7 8 91011121314 0 12 3 4 5 8 7 8 91011121314 Oxazolidine offer (Yo) Oxazolidine offer (Yo) Chemical Society Reviews, 1997 125 These new tannages satisfy the criteria of hydrothermal stability, metal free, white and lightfast. An additional benefit is the speed of the reaction, shown in Fig.24; maximum shrinkage temperature is reached in 3 h, compared with the 15 h typically required for chrome tanning. The only problem in leather terms is that the leather is weaker than a chrome tanned counterpart. This is due primarily to the filling effect of the resin and the loss of strength is proportional to the amount of resin used. It is probable that the tannage has not been optimised with regard to applying a single or major resin species, so by targeting the reaction more precisely, by minimising the quantity of resin required to obtain the desired shrinkage temperature, the strength of the leather will be maximised. 115 I I 110 105 100 95 90 85 ! I I I I I I 0 1 2 3 overnight tlh at 50 "C Fig.24 The rate of tanning with THPS and melamine-formaldehyde resins This could be the basis of the tanning chemistry of the 21st century. The remaining questions are: what is the level of the environmental impact of the change in tanning methods and what is the cost in comparison to chrome tanning? The first question cannot be answered until a full environmental impact study is undertaken, but it is known that melamine-formal-dehyde polymers are non-toxic and the crosslinkers are safe if handled properly. Cost will depend on the scale of adoption; the basic cost is going to be higher than for chromium(II1) salts, but savings may be made through reduced quantities of dyestuffs and lubricants needed for the new leathers.3 Discussion Contrary to popular perception, the tanning industry is eco-logically sound. It has a reputation for producing noxious waste and smells, but it should be remembered that the raw material for leather is a byproduct of the meat industry; if the tanners did not turn the hides and skins into leather, the abattoirs would be faced with a disposal problem that would far outweigh the alleged pollution by tanneries. In recent years, tanners in many countries have cleaned up their act, to the extent that they can operate successfully in modern city centres, unlike their historical predecessors, who were frequently compelled to work away from habitation. It is perhaps an indication that tanning is becoming more environmentally acceptable, that tanners world-wide no longer automatically occupy the lowest level of the social totem pole.Tanners have a good record of improving the quality of their waste products. The traditional method of tanning in closed drums means that processes were developed which would progress satisfactorily without constant monitoring or needing frequent changes in the chemical conditions. The consequence is that these traditional processes tend to be less environmen-tally acceptable. They are gradually being phased out, replaced by cleaner technologies, incorporating more precise process control, more recycling of chemicals and greater reliance on biochemical rather than chemical agents. It might be thought anachronistic, not only to continue to make leather but also to develop the chemistry of leathermak-ing, when so many synthetic materials are available, with such a range of excellent properties.Reconsider that thought in the light of these questions. What is the material of choice for your shoes or handbags? How many shoe upper materials allow wet air to pass freely from one side, whilst completely preventing liquid water passing through from the other side? How many fabrics will abrade, as you skid along the ground having fallen off your motorcycle, without fusing with your soft tissues? How many materials can resist the effects of splashing with molten metals (e.g. sodium, steel), without melting or trans-mitting the heat to the wearer of protective clothing? How many materials are used both for fire resistant upholstery and as fire blankets? How many materials retain their integrity and performance in the temperature range -100 to + 100 "C? Many materials can exhibit some of those properties, but only leather has them all! 4 References 1 A.J. Bailey, J. SOC.Leather Technol. Chem., 1992, 76(4), 111. 2 E. Heidemann, J. SOC.Leather Technol. Chem., 1982, 66(2), 21. 3 E. M. Brown, J. M. Chen and G. King, Protein Eng., 1996, 9(1), 43. 4 K. Kadler, Protein Profile, 1994, 1(5), 519. 5 C. E. Weir, J. Am. Leather Chem. Assoc., 1949, 44(3), 108. 6 A. D. Covington, R. A. Hancock and I. A. Ioannidis, J. SOC.Leather Technol. Chem., 1989, 73(1), 1. 7 G. N. Ramachandran, J. Am. Leather Chem. Assoc., 1968,63(3), 160. 8 E. Haslam, Leather, April, 1993. 9 E. Haslam, J. SOC.Leather Technol. Chem., 1988,72(2),45. 10 E. Haslam and Y. Cai, Nut. Prod. Rep., 1994, 11,41. 11 Plant Polyphenols, Basic Life Sciences, vol. 59, ed. R.W. Hemingway and P. E. Laks, Plenum Press, 1992. 12 H. P. Chakravorty and H. E. Nursten, J. SOC.Leather Trades Chem., 1958,42(1),2. 13 A. D. Covington, J.Am. Leather Chem. Assoc., 1987, 82(1), 1. 14 H. M. N. H. Irving, J. SOC.Leather Technol. Chem., 1974, 58(3), 51. 15 T. Gotsis, L. Spiccia and K. C. Montgomery, J. SOC.Leather Technol. Chem., 1992, 76(6), 195. 16 A. E. Russell and S. G. Shuttleworth, J. SOC. Leather Trades Chem., 1965, 49(6), 221. 17 R. L. Sykes, J.Am. Leather Chem. Assoc., 1956,51(5),235. 18 K. H. Gustavson, J.Am. Leather Chem. Assoc., 1953,48(9),559. 19 A. D. Covington, J. SOC.Leather Technol. Chem., 1986,70(2),33. 20 A. D. Covington, R. A. Hancock and I. A. Ioannidis, unpublished results. 21 A. D. Covington, J. Am. Leather Chem. Assoc., 1991, 86(10), 376. 22 A. D. Covington, J. SOC.Leather Technol. Chem., 1986, 70(2), 33. 23 W. Prentiss and I. V. Prasad, J.Am. Leather Chem.Assoc., 1981,76(10), 395. 24 J. N. Chatterjea, Leather Scz., 1983, 30(10), 291. 25 A. L. Hock, J. SOC.Leather Technol. Chem., 1975,59(6), 181. 26 J. H. Sharphouse, J. SOC.Leather Technol. Chem., 1985,69(2),29. 27 S. Dasgupta, J. SOC.Leather Technol. Chem., 1977, 61(5), 97. 28 J. H. Hernandez and W. E. Kallenberger, J. Am. Leather Chem. Assoc., 1984, 79(5), 182. 29 A. D. Covington and S. Ma, UK Patent 2,287,953,June 1996. 30 A. D. Covington and B. Shi, Proc. 3rd Asian Int. Conf of Leather Science and Tech., Japan, Sept. 1996. Received, 3rd June I996 Accepted, 29th November 1996 126 Chemical Society Reviews, 1997

ISSN:0306-0012    

DOI:10.1039/CS9972600111

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Carbon-carbon bond forming reactions mediated by cerium(IV) reagents Vijay Nair,* Jessy Mathew and Jaya Prabhakaran Organic Chemistry Division, Regional Research Laboratory (CSIR), Trivandrum, 695 019, India Organic synthesis using carbon centred radicals generated by one electron oxidants is of current interest. Although M~(OAC)~has received the most attention, it appears that cerium(rv) ammonium nitrate (CAN) would be a very useful reagent for the generation of radicals. The recent applica- tions of CAN in the construction of C-C bonds are highlighted in this review. CAN has been shown to mediate the facile oxidative addition of 1,3-dicarbonyl compounds to activated and unactivated alkenes leading to dihydrofurans and furanones. In most cases the reactions occur under milder conditions and the yields are superior when com- pared to Mn(0Ach mediated reactions.The results available thus far suggest that CAN offers vast potential in radical mediated organic synthesis. The fundamental task in the synthesis of organic compounds, simple or complex, is the construction of carbon-carbon bonds. Among the various synthetic methods involving polar, radical or pericyclic reactions available for C-C bond formation the free radical reactions have received the least attention until recently, largely due to the erroneous notion that they lack selectivity and are uncontrollable. The past decade, however, has witnessed a dramatic resurgence of interest in the use of radical methodology' which is attributable in large measure to Vijay Nair was born and raised in Konni, Kerala State, India.He received MSc and PhD degrees from Banaras Hindu University (with Professor R. H. Sahasrabudhey) and a second PhD degree from the University ofBritish Columbia under the direction ofJames P. Kutney. Subsequently he did post-doctoral work with Josef Fried at the University of Chicago, Peter Yates at the University of Toronto and Gilbert Stork at Columbia University. He joined the Medical Research Division (Lederle Laboratories) of American Cyanamid Company as Senior Research Chemist in 1974 and became Principal Research Chemist in 1987. In 1981 he received the Outstanding Scientist Award ofAmerican Cyanamid Company. In 1990 he returned the conceptualization and demonstration by Stork2 that the controlled formation as well as the addition of vinyl radicals to alkenes offers a unique and powerful method for complex carbocyclic constructions.The insightful investigations of Julia,3 Beckwith: Ingolds and others leading to a clear understanding of the structure and reactivity of carbon centred radicals and the innovative synthetic applications1 by Giese, Curran and Pattenden6 have contributed significantly to the acceptance of radical methodology. Today there is widespread appreciation of the potential offered by radical processes especially in the synthesis of structurally fascinating and biologically important natural products. Various procedures involving chemical,' electrochemical7 and photochemical8 methods are known for the generation of radicals.In particular, oxidative methods1,9,10 mediated by metal salts with adjacent stable oxidation states like those of Mn"', Co", CdI, Fe"1 and VV have received considerable attention. The major difference between oxidative methods and the traditional approaches is the dual role the metal oxidants play in these reactions. i.e., the initial one electron oxidation of the carbonyl compound will generate the electrophilic carbon radical which is not easily oxidised but can add efficiently to electron rich alkenes. The resulting adduct radical is more susceptible to oxidation and the products are often derived from the intermediate cation by inter- or intra-molecular capture of nucleophiles or by loss of a proton to form an alkene.to India to join the Regional Research Laboratory (CSIR) in Trivandrum as its Deputy Director. His research interests include cycloadditions, radical mediated C-C bond forming reactions and auxiliary directed chiral synthesis. Jessy Mathew has MSc and MPhil degrees with distinction from Mahatma Gandhi University. She completed her PhD (1995) under the supervision of Vijay Nair on C-C bond forming reactions mediated by CAN and has a number ofpublications in this area. Currently she is a post-doctoral Fellow with Professor Michael Kahn at Molecumetics in Bellevue, Wash- ington, USA. Jaya Prabhakaran is a Gold Medalist (MSc) of Kerala University. She completed her PhD (1996) under the guidance of Vijay Nair on asymmetric synthe- sis using chiral auxiliaries derived from bile acids and D-ghCOSe and has several publications in this area.Currently she is a Research Associate in Dr Nair's group. Despite the fact that Mn(OAc)3 occupies a unique position among the one electron oxidants and has served as a key reagent in the synthesis of a number of important natural products,9JO the procedural problems associated with the use of this reagent often limit its application. Naturally, there has been consider- able interest in developing newer reagents and methods for generating radicals. The pioneering work of Heiba and Dessaul and some subsequent investigations have shown that cerium(1v) salts can be used for the generation of radicals. This review focuses on recent advances in the application of cerium(rv) ammonium nitrate (CAN) in the construction of C-C bonds highlighting our own investigations in this area.2 Oxidative addition of carbonyl compounds to alkenes and arenes The first example of the CeIV mediated generation of carbon centred radicals involves electrophilic carbon radical] 1 CHX2 generated by Ce(OAc)4 from CHzXz and its addition to alkenes to give products 3, 5 or 6 according to the general reaction sequence given below. CH2X2 + Ce" oCHX2 + Ce"' + H' 1 X$H-CH2-CH-R X2CH-CH2-CH2-R XzCH-CH= CH-R X2CH-CH2-CHY-R 5 6 X = COOR, COR etc., and Y is any nucleophilic species present in the reaction mixture Scheme 1 The success of the reaction rests on the selective oxidation of radical 2 by the metal ion.The initially formed a-keto radical 1 is not easily oxidised due to the electron withdrawing character of the carbonyl group whereas the secondary alkyl radical 2 undergoes rapid oxidation leading to the product 5 or 6. When an aryl ketone such as acetophenone was used, in addition to the three expected products 7,8 and 9, a new product a-tetralone 10 was obtained (total yield, 64%) presumably by the addition of the intermediate radical to the benzene ring [eqn. (l)]. R=alkyl 7 8 (11 +w OAc R 9 10 49%, yield based on Mn 3+ whereR= C2H5 As an extension of this reaction, lactones were synthesized from alkenes and carboxylic acids12 [eqn.(2)]. 128 Chemical Society Reviews, 1997 Aromatic nitromethylation13 and acetonylation14 have been reported to give substitution products as shown in eqns. (3) and (4). R=H, CH3 R = H,55%, R = CH3,100% o = 56.5, m = 19.1,~= 24.4 13 R= CH37 OCH3 7 C1, F R = CH3,5%, R = OCH3,74% R= Cl, 25%, R= F, 29% It may be pointed out that except for lactone synthesis12 and the nitromethylation of aromatics'3 in which CAN was used as the oxidising species, all the earlier experiments involved the use of the unstable Ce(OAc)4 in acetic acid medium. An important advance in this area was made by Baciocchi and Ruzzi~oni's.~6 who demonstrated that CAN-mediated oxidative addition reactions proceed well in solvents such as methanol and acetonitrile, thus leading to wider acceptance of CAN.The malonylation of aromatic compounds15 mediated by CAN in methanol is an example [eqn. (5)]. COOMe CAN ArH -Al-(cooMe+ (COOMe COOMe (5) l4 Ar=ph,53% o = 50.8, m = 21.4,~= 27.8 3 Oxidative additions to dienes There are limited reports16 on reactions in which 1,3-butadiene acts as a radical trap. The CAN mediated oxidative addition of carbonyl compounds such as acetone, butan-2-one and ethyl acetoacetate to butadiene resulted in a mixture of 1,2-and 1,4-adducts whereas the addition of dimethyl malonate fur- nished substituted cyclopropanes [eqns. (6) and (7)]. 15(30%) 16 (38%) Recently trimethyl silyl enol ether17 was added to 1,3-butadiene in presence of CAN and the resulting mixture of nitrates subjected to nucleophilic attack with sodium dimethyl malonate in presence of Pdo; alkylated adducts 22 and 23 were obtained (Scheme 2).COOMe .COOMeg -+COOMe + ogoc<COOMe COOMe(COOMe ON02 17 18 (17+18)=97% (7) MeOOC' 19 R'\z2--R1& +R1& R2 ON02 R2 ON02 @R2= -1 20 21 20 + 21 N4P40)*R'Jpq +d& R2 Nu R2 Nu 22 23 e~.(22+23)~'=Et,R~ =H, 43% molarratio =98:2 Scheme 2 4 Oxidative additions to activated alkenes Activated alkenes, such as enol ethers, enol acetates and enol silyl ethers take part in the oxidative addition reactions smoothly. Synthesis of 4-ketoaldehyde dimethyl acetals and 3-acyl furans by the addition of carbonyl and 1,3-dicarbonyl compounds respectively to enol acetates l8 are examples.+ @OAc &OM. re& OMe (8).. R=alkyl 24 R=CH3,70% ~RZ=CH~ R3=H ., 25d=H, R2 =OEt R', R2=CH3,56%R'= H, R2 =OEt,9% The cross coupling of 1,2-disubstituted silyl enol ethers with other enol ethers has been exploited in the synthesis of 1,4-dicarbonyl compounds. l9 Additions of silyldienol ethers to silyl enol ethers,20 silyloxy cyclopropanes to 1 ,3-butadiene21 and 1,3-dicarbonyl com- pounds to allyl trimethyl silanes22 are some other contributions in this area. The cation radicals generated from allyl phenyl sulfides and enamines undergo addition to silyl enol ethers23 leading ultimately to the unsaturated ketone 27 and the 1,4-dicarbonyl compound 28 [eqns. (1 1) and (12)]. Similarly, the reaction of a-nitroalkyl radicals,24 generated by the oxidation of nitronate anions, to silyl enol ethers proceeds well to afford P-nitroketones.eg, R =Ph, 79% 'IBDMS CAN, But O (12)28 (63%) 5 Oxidative additions to cyclic and acyclic unactivated alkenes It is evident from the foregoing discussion that most of the CAN mediated oxidative additions of carbonyl compounds reported earlier involved activated alkenes. There has been no systematic investigation of CeIV mediated generation of radicals and their additions to cyclic and acyclic alkenes particularly unactivated ones. We have therefore undertaken a detailed study of CeIV mediated oxidative addition of active methylene compounds to such systems.25 The dicarbonyl compounds chosen for the study include dimedone, acetylacetone, ethyl acetoacetate and dime- thy1 malonate (Fig.1); the alkenes used are shown in Fig. 2. 0boaa aa,, (COOMeCOOMe Fig. 1 Fig. 2 CAN mediated addition of all the dicarbonyl compounds except dimethyl malonate occurred smoothly and rapidly furnishing dihydrofurans in good yields.25 A typical example is shown in eqn. (13). Similar reaction of 1,3-dicarbonyl com- pounds with alkenes in the presence of cobalt(I1) acetate and oxygen26 has been reported to yield dihydrofurans in moderate to good yields. 29 (98%) A comparative study of these reactions*5J7 vs. those mediated by Mn(OAc)3 has revealed that CAN is a superior reagent in terms of the mildness of the procedure, experimental simplicity and higher yields of products.As an illustration, the formation of 29 in 98% yield by CAN mediated reaction in methanol at 5 "C may be contrasted with its formation in 41% yield by the Mn(OAc)3 mediated procedure at the reflux temperature of acetic acid. Unlike dimedone and acetylacetone, dimethyl malonate adds to most of the alkenes to provide lactones.27 For example the reaction of dimethyl malonate with cinnamyl methyl ether in aqueous methanol furnished 30 [eqn. (14)]. Chemical Society Reviews, 1997 129 MM$ v 30 (58%) The formation of lactone 30 may be explained as follows. The addition of dimethyl malonyl radical to cinnamyl methyl ether followed by oxidation of the resulting radical with CAN would produce the cation 31.Since the reaction is done in aqueous methanol, 31 would be easily converted to 32 which would undergo lactonization to render 30. Alternatively, 34 formed from 31 via the stabilized cation 33 can lose methanol to afford 30 (Scheme 3). 31 32 I f 33 34 Scheme 3 As expected, extension of these reactions to exocyclic alkenes (Fig. 3) provided spirodihydrofurans28 [eqn. (1 5)]. Fig. 3 i% 35 (77%) 6 Oxidative additions to styrenes An unprecedented and mechanistically fascinating reaction was encountered between dimethyl malonate and ~tyrene.~9 When dimethyl malonate was reacted with styrene in presence of CAN, we obtained the ketone 36 and the lactone 37 as the major products [eqn. (16)].Small amounts of 38 and 39 were also isolated. Interestingly, this contrasts with an earlier report3O on the exclusive formation of 38 and 39 in this reaction. 0 COOMe (cooMe 0"CANNY%+ PCOOMe COOMe 36 (42%) 39 X =-0Me (5%) A tentative mechanism along the following lines may be suggested for the formation of these products. The benzylic 130 Chemical Society Reviews, 1997 radical 40 formed by the addition of malonyl radical to styrene gets quenched by molecular oxygen to furnish the peroxyl radical 41 which will be easily converted to the hydroperoxide 42. The oxidative cleavage of 42 would lead to the ketone 36. Support for this rationale is obtained from the fact that when the experiment was carried out under argon atmosphere only products 37, 38 and 39 were isolated. The fragmentation of the radical anion 43, derived from 40 by ligand transfer or reaction with nitrate, followed by protonation of the resulting alkoxide to the carbinol and lactonization of the latter would furnish 37. The mechanism for the formation of nitrate 38 can involve either the oxidation of 43 or the trapping of the benzylic cation 46 by N03-.Similarly 46 can be trapped by methanol to afford 39 (Scheme 4). An analogous reaction was observed with dimethyl malonate and a variety of ring substituted styrenes. COOMe 40 ,.H 0.O' COOMeoh",,,. A 41 42 9' -YJ-00 COOMe 0-COOMe 43 44 H OOMe 37 -&'kOOMe 45 ONO;738 L39 46 MeOH Scheme 4 While attempting to add dimethyl malonate to 4-methoxy styrene, we observed the exclusive formation31 of dimeric products 47 and 48. QCH3 moCH3 H3COgy-= H3CO 47 (40%) + 48(34%) The formation of 47 and 48 may be explained as follows.Methoxy styrene gets oxidised by CAN to the radical cation 49 which then adds to another molecule of methoxy styrene leading to the intermediate 50. Methanol addition to 50, oxidation of the resulting radical 51 to the cation 52 and 49 53 Scheme 5 subsequent trapping of the latter by methanol would furnish 47. Dimerization of radical cation 49 followed by the trapping of the dication by methanol 53 is also a valid possibility (Scheme 5).Alternatively, radical 51 may be trapped by molecular oxygen to give the peroxyl radical 54 which will be easily converted to the hydroperoxide 55.The latter on oxidative fragmentation would provide the ketone 48 (Scheme 6). c 55 Scheme 6 7 Intramolecular and tandem cyclizations Intramolecular cyclization of radicals generated by CeIV reagents has received only limited attention vs. Mn(OAc)3. The cyclization of 1-benzyl-2,6-disubstituted-4-piperidone-3-car-boxylic acid methyl ester with Ce(SO& to 56 in low yields is an example.32 The oxidative cyclization of silyl enol ethers of aryl ketones has been reported to provide tricyclic ketones.33 57 (70%) A report on the oxidative cyclization of dimethyl-4-pentenyl malonate has also appeared.34 The oxidation of substituted diethyl a-benzyl malonate in the presence of alkenes produced highly functionalized naphthalenes.35 R= H,*l RI = y CH~ R2 = CN, OAc, aryl etc 58 (58-89%) R3= H, CN, COOMe etc 8 C-C bond forming reactions mediated by CeIV vs.other reagents It can be discerned from the foregoing discussion that CerV reagents will prove to be of great value in organic synthesis. Comparative studies of CeIV with other species, especially, MnIII would be particularly worthwhile. With few excep- tions, 1,26no systematic investigations involving such compari- sons have been carried out. It has been reported that the order of relative reactivity of the oxidation of secondary alkyl radicals by MnIrI, CeIV and CulI is of the order of 1 :12 :350.1' In the nitromethylation of aromatic compounds with nitromethane, CeIV salts were found to be the most promising.In the synthesis of isomeric a-nitroxylenes from xylenes and a-nitrotoluene from toluene, CeIV gave nearly 100% yield. Lower yields are observed when CeIV salts are used together with CuT1,13 and only with CorI1 acetate was any nitromethylated product observed. It is also known that CerV reagents gave higher yields and are superior in their control of regiochemistry in alkan- 2-one coupling. They can be applied efficiently in inter- molecular reactions, which is not the case with Mn"1 promoted coupling.' The available data clearly indicate that in inter- Chemical Society Reviews, 1997 131 molecular C-C bond forming reactions CeIV mediated proc- esses are more efficient than those mediated by Mn"1 9 Conclusion The credentials of CAN as a useful reagent for C-C bond formation have been recognized Most of the work using CAN has involved intermolecular reactions, it however is reasonable to assume that CAN will prove its usefulness in the intra- molecular situation as well There are indications that investiga- tions in this direction will be rewarding From a future perspective, investigations aimed at unravelling the mechanistic details of CeIV mediated reactions and efforts to make them catalytic will be very worthwhile The solubility of CAN in solvents such as methanol and acetonitnle, the mild reaction conditions and the relatively low toxicity of cenum vis transition metals are all factors that will help CAN find a prominent place in the repertory of synthetic organic che- mistry 10 References 1 For reviews see (a) B Giese in Radicals in Organic Synthesis Formation of Carbon-Carbon Bonds, Pergamon, Oxford, 1986, D H R Barton, Aldrzchm Acta, 1990, 23, 3, (b) D P Curran, in Comprehensive Organic Synthesis, ed B M Trost and I Fleming, Pergamon, New York, 1991,vol 4, p 715, (c)A L J Beckwith, Chem SOC Rev, 1993, 143, (6) P I Dalko, Tetrahedron, 1995,51,7579 2 G Stork and P M Sher, J Am Chem SOC, 1983, 105, 6765, and references therein 3 M Julia, Pure Appl Chem , 1974, 40,553 4 A L J Beckwith, Tetrahedron, 1981, 37, 3073, and references therein 5 C Chatgilialoglu, K U Ingold and J C Scaiano, J Am Chem SOC, 1981, 103, 7739, and references therein 6 G Pattenden, A J Smithies, D Tapolczy and D S Walter, J Chem Soc Perkin Trans 1, 1996, 7 and references therein 7 L Becking and H J Schafer, Tetrahedron Lett, 1988, 29, 280 8 D 0 Cowan and R L Drisko, in Elements of Photochemistry, Plenum, New York, 1976 9 W J De Klein, in Organic Synthesis by Oxidation with Metal Compounds, ed W J Mijs and C R H de Jonge, Plenum, New York, 1986,p 261 10 B B Snider, Chem Rev, 1996, 96, 339 and references cited therein 11 E I Heiba and R M Dessau, J Am Chem SOC , 1972,94,2888 12 E I Heiba,R M Dessau andP G Rodewald, J Am Chem SOC,1974, 96,7977 13 M E Kurz and P Ngoviwatchai, J Org Chem , 1981,46,4672 14 M E Kurz, V Baru and P N Nguyen, J Org Chem, 1984, 49, 1603 15 E Baciocchi, R Ruzziconi and D D Aira, Tetrahedron Lett, 1986,27, 2763 16 (a)E Baciocchi and R Ruzziconi, J Org Chem , 1986,51, 1645, (b) E Baciocchi and R Ruzziconi, Gazz Chim Ira1 , 1986, 116, 671 17 A B Paolobelli, P Ceccherelli, F Pizzo and R Ruzziconi, J Org Chem , 1995,60,4954 18 E Baciocchi and R Ruzziconi, Synth Commun , 1988,18, 1841 19 E Baciocchi, A Casu and R Ruzziconi, Synlett, 1990, 679 20 A B Paolobelli, D Latini and R Ruzziconi, Tetrahedron Lett, 1993, 34,721 21 A B Paolobelli, F Gioacchini and R Ruzziconi, Tetrahedron Lett, 1993,34, 6333 22 J R Hwu, C N Chen and S S Shiao, J Org Chem , 1995,60, 856 23 K Narasaka, T Okauchi, K Tanaka and M Murakami, Chem Lett, 1992,2099 24 N Arai and K Narasaka, Chem Lett, 1995,987 25 V Nair and J Mathew, J Chem SOC Perkin Trans I, 1995, 187 26 J Iqbal, B Bhatia and N K Nayyar, Tetrahedron, 1991, 47, 6457 27 V Nair, J Mathew and K V Radhaknshnan, J Chem SOC Perkin Trans 1, 1996, 1487 28 V Nair, J Mathew and S Alexander, Synth Commun, 1995, 25, 398 1 29 V Nair and J Mathew, J Chem SOC Perkin Trans 1, 1995, 1881 30 E Baciocchi, B Giese, H Farshchi and R Ruzziconi, J Org Chem , 1990,55,5688 31 V Nair and J Mathew, unpublished results 32 R Haller, R Kohlmorgen and W Hansel, Tetrahedron Lett, 1973,15, 1205 33 B B Snider and T Kwon, J Org Chem , 1992,57, 2399 34 E Baciocchi, R Ruzziconi and A B Paolobelh, Tetrahedron,1992,48, 4617 35 A Citteno, R Sebastiano and M C Carvayal, J Org Chem , 1991,56, 5335 Received, 1st April 1996 Accepted, 28th November I996 132 Chemical Society Reviews, 1997

ISSN:0306-0012    

DOI:10.1039/CS9972600127

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Lead, glass and the environment Michael J. Hynesa and Bo Jonsonb a Chemistry Department, University College, Galway, Ireland h Glafo, Box 3093, S-350 33, Vaxjo, Sweden The history of lead usage spans some 5000 years. Lead has found considerably wider usage than its natural abundance might suggest. The role of lead in glass manufacture and in particular lead crystal is outlined. Problems involved in replacing lead in glass are discussed. While lead is one of the oldest industrial poisons, it is only in the latter part of this century that it has been elevated to the status of a premier environmental concern. The consequences of this have been acute. The roles of international bodies such as the Organisation for Economic Co-operation and Development (OECD), the World Health Organisation (WHO) and the International Program on Chemical Safety (IPCS) in reducing exposure to lead are reviewed.1 Lead Lead consists of four stable isotopes, 204Pb, 206Pb, 207Pb and 208Pb. The latter three arise from decay of the radioactive series 238U, 235U and 232Th respectively. Unlike gold for example, lead rarely occurs in the elemental state, rather it occurs as galena (PbS), carbonate (PbC03) and anglesite (PbS04). The use of lead began in the neolithic era, around 3500 BC so that the history of lead use spans some 5000 years. Lead is a most ubiquitous metal in that it has found many uses. Since it occurs in highly concentrated ores from which it is readily extracted, the availability of lead is far greater than its natural abundance might suggest.Its environmental significance derives from both its utility and availability. The desire for silver, an important economic indicator in early times, was the principal driving force for lead production. For each ton of silver produced, approximately 400 tons of lead metal were first smelted from ores. The silver was separated from the lead by a process known as cupellation, perhaps the oldest metallurgical process. The metals were melted in a cupel or crucible at ca. 1100 “C.Air was blown onto the fused metals and this oxidised lead to litharge (PbO) which floated on the Michael J. Hynes is a lecturer in inorganic chemistry at University College Galway. He obtained his BSc and PhD degrees in Chemistry at Galway following which he carried out postdoctoral -reseach with Rabh G.Pearson at Northwestern University in Illinois. Thus began his interest in the kinetics and mechanisms of inorganic reactions. He also spent some time with the Eigen group in Gottingen as a Senior Ciba-Geigy Fellow. His current research interests include the kinetics and mechanisms of in-organic reactions, mineral sup- plementation, risk assessment of chemical substances and the chemistry of lead crystal. Michael J. Hynes surface of the melt and was mechanically removed. Silver, which is much more difficult to oxidise, was not oxidised under these conditions, and could be recovered in a relatively high state of purity. Apart from its use in cupellation, the uses of lead range from the production of ornaments to the construction of the hanging gardens in Babylon some 4000 years ago.It was also used extensively in enamels in pottery glazes, a procedure now largely discontinued in many countries, and in cosmetics which whiten the face. Organolead additives were also widely used to improve the octane ratings of gasoline. The use of lead to improve the quality of glass probably originated in Egypt as evidenced by the wall paintings found at Tel-el-Emara showing the various stages of glassmaking. Lead is also widely used in lead-acid batteries, in piping, solder, cable sheathing, radiation shielding and ammunition while ‘white lead’, Pb3(C03)2(OH)2 was very extensively used until relatively recently as a major component of white indoor paint.Fig. 1 sketches lead use patterns over the past 5000 years while Fig. 2 shows the distribution of lead use by percentage category of use in 1990.’ 2 Glass2-10 2.1 Definition of glass ‘Glass’ is a term frequently used by most people. Few of them will however have reflected on the definition of the material or its typical properties. Glass is often described as ‘a liquid’ or ‘undercooled liquid’. A more stringent way of defining the material is the following: ‘glass is a melt that has solidified without being subject to cry stallisation’. The latter definition can be regarded as sufficient for most industrial applications. A glass can, however be produced or synthesised by routes other than melting solid substances and quenching the melt.Condensation of gaseous components or transforming liquids via a gel state to a glass are alternative ways of glass making. To exclude the production route from the Bo Jonson studied inorganic chemistry at the University of Lund where he obtained both his BSc and PhD degrees. He joined Glafo in 1988 and since 1990 has been research manager there. Glafo is a non-profit membership or- ganization which provides R&D and technical support for Scan-dinavian glass producers as well as crystal producers in the UK, Ireland and France. Bq Jonson’s own research at Glafo is concen-trated on subjects related to the crystal industry. These include glass chemistry, lead free batch formulations, colour origin and I‘improved glass quality.Bo Jonson Chemical Society Reviews, 1997 133 Gasoline______.Additives Roman Empire i Fall ofi Roman Empire f Industrial.---RevoIution 1500 1700 700 2501 300 1100 1750 BC 0 AD Fig. 1 Use of lead during the past 5000 years (Reproduced with permission of the International Council on Metals and the Environment, Ottawa, Canada) Gasoline additives (2%) Shot 8 ammunition (3%) Miscellaneous (4%) Alloys (4%) Cable shealthing (5%) Rolled and extruded products (9%) Pigments and other components (10%).. Batteries (63%) 0 15 30 45 60 75 Percentage Fig. 2 Distribution of lead use by percentage category of use in 1990 definition of glass, it is worthwhile inspecting the pattern of how the specific volume correlates with temperature for glass and crystalline materials, Fig.3. From Fig. 3 it is obvious that crystalline materials and glasses have different characteristics. The crystalline material shows a very abrupt change in volume on passing from the liquid to the solid state. The glass, on the other hand, shows a less pronounced change in volume until the temperature is lowered to the region where the material solidifies. This temperature region is called the ‘transformation region’. In order to define the transformation temperature, it is necessary to define the annealing rate, since this influences its absolute value. This transformation point phenomenon can be utilised to produce a more stringent definition of glass: ‘glass is a material which exhibits the phenomenon of a transformation region’.134 Chemical Society Reviews, 1997 /Liquid Meltingpoint -Temperature Fig. 3 Specific volume versus temperature for glass and crystalline materials 2.2 Structural aspects of glass The classic theory of glass structure dates from the 1930s and was formulated by Zachariasen. He described glass as a network and postulated both coordination and bonding criteria for the components in the glass (Zachariasen’s Rules). In the opening sentence of his classic 1932 paper entitled ‘The Atomic Arrangement in Glass’ Zachariasen states, ‘It must be frankly admitted that we know practically nothing about the atomic arrangement in glasses’.While the ideas and work of Zachar- iasen are still valid today, modern X-ray and spectroscopic techniques in combination with kinetic and thermodynamic considerations have contributed significantly to our current knowledge of glass structure in terms of bonding distances, atom distances and coordination polyhedra. For the purpose of this review we may think of the glass structure as a continuous random network possessing medium range order rather than a disordered amorphous material. For glass where silica is the dominant component, a model using the concept of a depolymerised silica structure can be useful. Two principal types of Si-0 bonds can be found: ‘bonding’ i.e. those bonding to additional Si atoms and ‘non- bonding’ i.e.those bonding to other components of the material.Adding additional components to silica, i.e. making a silicate glass, increases the degree of silica depolymerisation and consequently the average number of non-bonding Si-0 per Si. The latter number can be used for a description of the glass in terms of the expected structural units. The understanding and theories of glass structure are however probably best developed for vitreous silica or quartz glass. An increase of the number of components complicates the description. A general model for glass structure in relation to its physical properties is lacking. This is probably due to the complexity of modelling a non-crystalline structure built from up to 15 components.3 Chemical constitution of glass 3.1 Silicate glasses A wide variety of substances can be transformed to or form glasses. Examples include organic compounds, elements, oxides, halogenides, chalcogenides. From a technological point of view, the expression ‘glass’ in most cases refers to a vitreous material based on silicates and produced by use of quartz sand (Si02) as the dominant ingredient. Silicate glasses are described with their chemical components formulated as oxides, dis- regarding (or lacking knowledge of) the type of compounds actually formed in the glassy matrix. The chemical composition is usually given as mass% or mol% oxides. The development of special optical glasses has led to research and development on a variety of non-silicate glasses.The description of those, is however outside the scope of this review. The ingredients of a glass batch? may be conveniently discussed in terms of their structural and technological functions. With this in mind, the following sections will discuss the chemical composition of glass in terms of network formers, fluxes, stabilisers, refining agents and colouring/decolouring elements. 3.2 Network or glass formers In silicate glasses, the network is comprised of silica which is probably present in the form of anionic irregular ring structures. The quantity of silica in a silicate glass is normally between 50 to 75 mass%. Lower amounts (up to 5 mass%) of the glass forming oxides B2O3 and P205 may also be present in a silicate glass.3.3 Fluxes The melting point of quartz sand is approximately 1730 “C .To facilitate the melting, or more strictly, to lower the viscosity of the melt, Si-0-Si bond breaking compounds are introduced. This is achieved using a variety of elements termed fluxes; however the most commonly used are the alkali metals. The quantity of fluxes present ranges from 10to 20 mass%. For cost reasons, lithium, sodium and potassium carbonates are the only ingredients normally used in a glass batch. Viscosity is a critically important factor in forming hollow lead crystal items such as stemware using either hand-blown or moulding operations, hence the need to carefully control it. In order to form a piece of hollow stemware such as a wine glass in a handblown operation, a ‘gather’ of glass is taken from the furnace on the end of a hollow steel tube called a blowing pipe.The craftsman blows through the blowing pipe to form a bubble within the gather of glass. The body of each piece is blown in a water cooled mould which determines its shape. The moulds are usually made of graphite and lined with cork. A jacket of steam between the mould and the glass gives it its smooth surface. Too low a viscosity will result in the gather of glass falling off the blowing pipe while too high a viscosity will result in it being difficult or impossible to blow and shape. 3.4 Stabilisers or modifiers A two component alkali silicate glass is also known as ‘water glass’ and is water soluble.The introduction of di-, tri-, tetra- or penta-valent elements renders the glass durable towards attack by water. The introduction of additional components, i.e. modifiers, affects the chemical and physical properties of the glass. In practice, systematic variation of the modifier content is employed when reformulating the chemical composition in order to achieve the desired properties. From 10 to 35 mass% of modifiers may be present. Non-colouring modifying elements for silicate glasses may be classified in terms of the oxide groups: Divalent: MgO, CaO, SrO, BaO, ZnO and PbO Trivalent: A1203, Bi2O3, Y2O3, La203 Tetravalent: Ti02, Zr02, Sn02 Pentavalent: Ta2O5, Nb205 I Batch: the mix of all necessary raw materials that is charged to the melting furnace. 3.5 Refining agents The glass batch contains compounds in which binary anions, typically carbonates, are present.Melting of the batch, at temperatures up to 1500 “C, causes decomposition and carbon dioxide evolution. Most of the gas will be transported away by the flue gases in the furnace, however some remains, together with air and batch humidity, as gaseous inclusions in the melt. The removal of the gas bubbles is called refining and operates by two mechanisms. One is pure physical bubble-rise according to Stoke’s law, the other is chemical dissolution of the remaining gas in the melt. Both mechanisms are facilitated by the addition to the batch of compounds referred to as refining agents; these result in gas evolution or sublimes.They also affect the redox chemistry of the melt. Three of the most commonly used combinations are: Redox active: As203 and/or Sb2O3 or Ce02 in combina- tion with an oxidising agent, most fre- quently sodium or potassium nitrate. Sulfate: sodium or calcium, often in combination with coke Salts: various chlorides or fluorides, nowadays rarely used due to their adverse environ- mental impact. 3.6 Colouring/decolouring agents Colour in glass can be generated by four basic mechanisms. (a)Colouration by addition of transition metal ions Introduction of first series transition metals into a silicate glass causes coloration due to coordination of oxygen donor atoms to the metal ion. The resultant colours are often analogous to those found in aqueous solutions of the corresponding ion.Chrom- ium(II1) gives rise to an intense green colour and is used in the manufacture of green bottles. Iron(m) produces a rather weak yellow-green colour, which is shifted towards a more intense blue-green colour upon reduction to iron@). For some transition metal ions, the chemical composition and the resulting structure of the base glass influences the colour obtained. (b)Coloration by non-metallic elements Selenium and sulfur may be utilised to colour glasses, either on their own or in combination with other elements. Sulfur can cause yellow-brown colours, selenium yellow or pink depend- ing on the composition of the host glass. Sulfur and selenium may be used to modify the coordination environment around a transition metal colouring element in a silicate glass.They co- ordinate to them and thus give rise to different ligand fields. Iron-sulfur chromophores result in amber or brown glass while cadmium-selenium/sulfur chromophores produce red. (c) Colouration by colloidal particles The coinage metals, copper, silver and gold can be dissolved in a silicate glass and heat treated under rather strict conditions to form colloidal metallic (probably oxides in the case of copper) particles of nanometre size. These give rise to yellows or reds. (d)Decolouring In common with all industrial processes, glass manufacturing is influenced by process costs. The raw materials used for most glass making are not reagent grade chemicals as used in laboratories, but rather minerals or bulk chemicals.As a result green/yellow colouring contaminants such as Fe2O3 and Cr2O3 will be introduced into the glass. When producing a clear colourless glass, the colours produced by the contaminants are compensated for by addition of small amounts of elements which generate the complementary colours, thus giving a grey or slight pink/violet/blue tint to the glass. Chemical Society Reviews, 1997 135 4 Examples of typical lead-containing silicate glasses 4.1 Optical glass Optical glasses, designed to cover a wide and strictly defined range of refractive indexes and dispersions, were traditionally composed of silicate glasses. Lead was introduced to increase refractive index and dispersion (variation in refractive index with wavelength).Optical glasses may have a lead content which can vary from just a few mass% to more than 50 mass% PbO. Table 1 shows the composition of optical glass F8 having a refractive index (nd) of 1.5955 and dispersion (v) of 39.18. 4.2 Sealing glass Sealing glasses are used to join metals to glass in a variety of products. The demands for matching thermal expansion and good adhesion are high. The insulating properties of glasses are advantageous when used in electronics. Lead is also used in some of these glasses. A typical composition is shown in Table 1. 4.3 Cathode ray tube glass The cathode ray tubes used in televisions and monitors utilise three different types of glass in the screen, cone and neck respectively. While lead has been used in all of the glasses, the practice nowadays appears to be that only the neck which contains the glass-metal joint is made from a lead-containing glass.A typical composition for a cathode ray neck tube glass is shown in Table 1. 4.4 Crystal Crystal or lead crystal is the term used to describe a glass composition used for hand or machine based production of more decorated glass items. According to an EU directive” a glass traded within the EU must contain more than 24% PbO to be called ‘crystal’. The directive also states minimum values of refractive index nD (1.545) and density (2.9 g ~m-~).These minima are however the minimum values expected from a glass having such a lead content.Lead gives the glass properties favourable for this type of production, as will be discussed later. Typical compositions are shown in Table 1. 5 Physical and chemical properties of melts and glasses-special emphasis on the impact of lead 5.1 The relation between composition and properties An alteration in the ratio of the components in a silicate glass will result in changes in the physical and chemical properties. These are however, the sum of the influences of all components on the glass structure, so that not only one, but many properties will be affected upon altering the chemical composition. This is schematically described in Fig. 4. Fig. 4 implies that for instance increasing the amount of Na20 in order to reduce the viscosity also results in increased thermal expansion (a)and Table 1 Chemical composition of glasses Chemical composition (mass%) Optical glass F8 (nd: 1.5955, v: 39, 18) Sealing Glass Si02 50.2 57.1 A1203 - 1.5 B203 0.4 - Na20 3.8 4.9 K20 5.6 7.0 PbO 39.7 29.5 ZnO - - As203 0.3 - Sb203 - - reduced durability. Because of the effect on many properties of one component, all glass compositions used in manufacturing can be regarded as compromises, with some critical properties at optimum and the remainder at acceptable levels.Tendency to devlbficaQon t CaO k II Increased densty, refractwe index, dispersion \ /lncreaseased chemical durabilrty 1 \J lncreaised viscosty Decreased chemical durability Decreaseda Increased a Fig. 4 Properties-composition paradox in glasses.The length of the vector indicates how large the effect of the component is (a= thermal expansion coefficient) As previously stated, there is no available structural model for predicting the relationship between composition, structure and properties. The methodology applied for calculating the relationships between properties and composition is based on experimental data. Data on known glass compositions and measured properties over a well-defined range of compositions are treated by mathematical/statistical means. From the analy- sis, a model is designed to describe and calculate the relationships between the chemical composition and properties of interest.5.2 Viscosity The viscous flow-temperature dependence of the melt is probably the most crucial technological parameter in glass production. It determines the thermal input requirements, furnace design and materials and sets requirements for forming equipment and conditions. The viscosity of a glass (7)ranges over several orders of magnitude during the melting process. Viscosity is frequently described by the so-called Vogel- Fulcher-Tamman (VFT) eqn. (l), log ‘1= A +B/(T-To) (1) where 7 is viscosity (dPa); T is temperature and A, B, To are constants. The viscosities of some silicate glasses are plotted using the VFT equation in Fig. 5. It is apparent that the lead glass has a smooth curve; this results in a long temperature range (or time) for the forming process.Such glasses are termed ‘soft’ or ‘long’ Cathode ray tube Lead crystal Lead crystal) (neck glass) (> 30% PbO) (> 24% PbO) 56.2 55.2 59.5 1.7 -0.6 0.8 4.2 0.4 1.9 8.4 11.7 11.0 29.3 31.8 24.5 --1.5 -0.3 0.4 0.3 -0.4 136 Chemical Society Reviews, 1997 glasses. Another characteristic of the lead glass is a low transformation region; thus lead glasses are suitable for low temperature glazing purposes. 5.3 Crystallisation Extensive cry stallisation or devitrification of all glasses can be induced by thermal treatment. The kinetics and phases sepa- rated are largely influenced by the chemical composition and presence of nucleating agents.Lead-containing glasses are found to be stable towards crystallisation and they exhibit low liquidus temperatures. Crystallisation is often found to be initiated at the surface. This is likely due to lead evaporation from the surface resulting in a change in the chemical composition tending towards alkali silicates which have different crystallisation characteristics. 5.4 Surface tension Although surface tension is also a parameter of prime importance in glass technology, its impact on processes related to glass production is not fully understood. Surface tension is largely dependent on the overall chemical composition of the melt. The Na20/K20 ratio has a significant influence on the surface tension. The introduction of large quantities of lead results in a large decrease in surface tension.For example, at 1300 "C the surface tension for a lead-containing glass can be expected to be ca. 200-230 mN m-1 or 20% lower than in lead free silicate glasses having similar viscosities and applica- tions. 5.5 Thermal properties Specific heat and thermal conductivity, properties most relevant at higher temperatures, are less sensitive to variation in either temperature or chemical composition. Lead has little influence on these parameters, however high-lead glasses have been reported to have thermal conductivities at the lower end of the narrow range of variation. Thermal expansion is to a large extent influenced by the overall chemical composition. Most components, apart from SiOz and B203 cause an increase in the thermal expansion coefficient.The effect of lead is rather Fig. 5 Viscosity-temperature curves for silicate glasses. A: soda lime glass, B: lead crystal glass, C: unleaded crystal. Chemical Society Reviews, 1997 137 modest, with alkalis having the most pronounced effect on this property. 5.6 Electrical properties The electrical properties of a glass melt (solid glass is insulating) are of technical interest when using electrically heated furnaces. Conductivity in a glass melt arises from the mobility of fluxing and stabilising ions. Alkali metal ions cause the greatest increase in conductivity while lead causes a reduction. 5.7 Mechanical properties The factors influencing the mechanical properties of glass are difficult to quantify.For instance, the tensile strength is not only dependent on the chemical composition, but also on sample size and surface conditions of the test specimen, test procedure, thermal history etc. Nevertheless, some properties of interest can be briefly discussed. The modulus of elasticity is strongly dependent on the chemical composition. Lead reduces the value of this property significantly to ca. 60 Gpa; this is some 15-20% lower than for lead-free glasses. Surface hardness is also dependent on composition. Lead has a similar effect on hardness as on the modulus of elasticity in that the hardness is significantly reduced on introduction of lead. This greatly facilitates mechanical treatments such as cutting, grinding and polishing. However, the softer lead glasses will also be more readily scratched by uncontrolled mechanical forces.5.8 Optical properties The optical properties most frequently discussed in relation to glass are the refractive index (nSpectralline), main dispersion (nF -nc) and reciprocal dispersive power. The dispersive property is commonly expressed as the Abbe number (v) [eqn. (2)] where C (red hydrogen line 656.3 nm) , D (yellow sodium line 589.3 nm) and F (blue hydrogen line 486.1 nm) represent the relevant Fraunhofer lines of the spectrum of the Sun. v = (nD -I)/(nF -nc) (2) A typical lead crystal containing 24% PbO will have an Abbe number of ca. 48. Both refractive index and dispersion are strongly dependent on the chemical composition and are affected by alterations to it.The influence of a modifying element on those properties does not necessarily have the same effect, i.e. a substitute increasing the refractive index may not necessarily increase the dispersion (decrease in the Abbe number). Ti02 is probably the oxide contributing most to an increase in both refractive index and dispersion in silicate glasses. The effect of PbO is similar to that of Ti02. 5.9 Colouring properties The origin of colour in silicate glasses has been briefly discussed above. No primary effect on colour is expected from the introduction of lead in a silicate glass, although yellowish colouration has been reported for high lead glasses.The latter effect is probably due to contaminants originating in either the raw materials or the crucible. Lead glasses show strong absorption in the ultraviolet spectral region. This can result in secondary effects on the colour originating from ionic or elemental colour centres as discussed above. Another secondary effect can be ascribed to the structure of lead glasses and the fact that several lead glasses contain little or no sodium or lithium but rather potassium as fluxing component. For instance, NiO colours lead-potassium glasses purple but results in a brownish colour in lead-sodium host glasses. Selenium colours lead glasses yellow, but gives rise to a pink colour in non-lead glasses. 5.10 Density Density is also an additive property of the glass components.Lead is one of dominant components in increasing the density. 138 Chemical Society Reviews, 1997 30% PbO glasses have densities of ca. 3 g ~m-~. Increasing the lead content to 60% might double this value. 5.11 Chemical durability The durability of glasses is a very complex subject. A variety of standardised test methods have been described covering a wide range of media and ambient conditions, the thermal history, storage conditions and treatments of the glass item also influencing the result. In general, the higher the silica content in a silicate glass, the more stable towards attack it is. Addition of alkali metals results in a significant deterioration in durability. With regard to the effects of modifying elements, it is difficult to find one single general pattern.One investigation reports increased durability when substituting Na20 in the system Na20-SiO2 with equimolar amounts of CaO, MgO, PbO and ZnO. In this case, ZnO resulted in the greatest improvement in durability, however, this may not necessarily be so for other glass compositions or indeed for other test methods. 6 Advantages and disadvantages of lead in glass In a summary of the above section on the properties of glasses relevant to technology and manufacturing, several benefits of lead can be identified. Lead makes the glass soft, both in the melt and as a solid glass being processed. It gives the glass high values of desirable physical properties such as density and refractive index, properties used for classification of crystal glasses.It promotes a low deformation temperature, useful for low temperature glazing procedures. Some technological drawbacks can also be identified. The soft glass is easily damaged during handling thus increasing the number of repairs necessary. The long viscosity curve may result in lower productivity at the hot end$ working process. Lead is volatile and both the work place and the environment are exposed to lead oxide vapours during manufacture of lead glasses. The most serious drawback of lead is of course its well documented adverse impact on the environment and health. This may eventually lead to its eventual replacement as is occurring in gasoline. 7 Substitution of lead in silicate glasses It is difficult to find one single element which promotes in an alkali silicate glass all the properties achieved by the introduc- tion of lead.It imparts high refractive index and high dispersion without colouring the glass. It provides economic melting temperatures and a long working temperature range suitable for the traditional hand-working methods, high density, a hardness suitable for mechanical cutting together with a chemistry suitable for acid polishing following the cutting process. Finally, all these properties are provided at an economic cost. From a chemical and structural point of view, the neighbouring element bismuth, present in the glass as Bi203, gives similar properties to lead.However, due to cost and availability, bismuth is a less than ideal substitute. The total quantity of bismuth currently mined would only substitute approximately 10% of the lead currently used by the crystal industry. The cost of the raw material is some ten times higher than lead oxide, resulting in unrealistic process costs. From the point of view of the crystal industry, two principal ways of substituting lead in crystal can be identified. One is the use of rather high amounts of BaO together with minor additions of other modifiers. This results in glasses having high densities and refractive indices and melts having viscosities suitable for hand-working. However, the widespread use of barium as a lead substitute has been questioned, primarily due to $ Hot end this refers to procedures applied when the glass is in the melt stage, i.e.at temperatures higher than the transformation range.Forming the melt to the desired shape is the dominant activity. the toxicity of soluble barium compounds. The other route towards unleaded formulations has been the use of combina- tions of the modifiers mentioned above. For cost reasons, a suitable combination of CaO, MgO, SrO and ZnO together with smaller additions of other modifiers, results in glasses having properties similar to those of lead crystal. In general, the unleaded formulations will be expected to show a somewhat shorter hot-end working range, however, this might even be beneficial for productivity. The major changes will be required at the cold ends working, e.g.cutting, grinding and polishing. The cold-end processes currently in use have been developed over many decades and have been optimised for soft lead glasses having well known chemical and physical surface properties. The use of harder glasses having different surface properties will dramatically change the requirements for these processes. 8 Lead as an environmental issue Lead is one of the oldest industrial poisons. Contamination by industrial lead has occurred everywhere on the earth. Estimated natural emissions of lead aerosols to the Earth’s atmosphere are estimated to be only 1/100 those of industrial lead emissions.12 This point is amplified by the knowledge that the skeletal lead concentrations in typical Americans in 1980 were elevated some 500-fold above the natural concentrations measured in the bones of Peruvians who lived in a relatively unpolluted environment some 1800 years ago.The Agency for Toxic Substances and Disease Registry (ATSDR) ranks lead as its prime concern13 while the EU’s concern regarding lead toxicity is highlighted by the fact that it is one of the few elements for which not only are specific regulations in place to protect workers from exposure to it,14 there is also a Directive on biological screening of the population for lead.Is Lead is now one of the premier environmental concerns and pervasive lead poisoning is the principal environmental health issue for American children. How did this come about? Why is lead such an issue compared to other toxic elements such as arsenic, cadmium and mercury? Mushakl6 has discussed the criteria for ranking an element as an environmental health issue.These are shown in Table 2. It will become apparent that lead meets most or all of these criteria. Table 2 Criteria for ranking lead exposure/lead poisoning as an environ- mental health issue 1 Economic and sociopolitical (a) Economic and historic centrality (h) Dominance of economic/social over health issues (c) Limited scope of decision making (4 Control of research and information 2 Scientific and public health (a) Indestructible-accumulates in environment and body, multimedia contaminant (h) Toxicity in numerous organs/systems with few protective body barriers (c) Toxicity with low or no threshold in huge numbers of the most vulnerable (4 Effects persist in target organs 3 Societal risk assessment (a) Adverse health effects across total spectrum of detection, preclinical and clinical (h) Compelling evidence of widespread toxicity is present (c) Requires role for both environmental control/preventive medicine and traditional medical interventions (4 Requires use of societal-level (macro risk) and individual (micro level) components of cost-benefit analyses The high rank of lead as an environmental contaminant is readily visualised when the position of lead as a global 9 Cold end this refers to procedures applied to a cooled glass, commonly at room temperature.economic commodity is examined. Since the introduction of cupellation, the total world production of lead is estimated by Flegal and Smith17 to be ca. 300 million metric tons. Total world consumption up to 1989 is estimated to be 275 million tons. Mercury also has a long history in that its earliest use occurred some 3500 years ago18 and by the first century AD, the Roman population was consuming 4.5 metric tons per annum. Total production of mercury since it was first used is estimated at 780000 metric tons. Cadmium is a relative newcomer to the industrial scene and while significant quantities were produced during the production of zinc, most of it was discarded and it is only in ca. 1910 that commercial production began.Total world production to date is estimated at ca. 1.4 million metric tons. Arsenic is produced as a by-product during copper production and its commercial production dates to the beginning of the nineteenth century, a total production span of ca. 190 years. Total world production is estimated at ca. 78 million metric tons. From the foregoing, the premier role of lead as a potential environmental contaminant can be readily appreciated. It has been used for longer than the other elements and significantly larger quantities of it have been produced and used. 9 History of lead toxicology Castellin019 provides a detailed history of the incidence and recognition of the adverse effects of lead. The first documented case of an illness which can be definitely ascribed to lead poisoning is ascribed to Hippocrates (fifth to sixth century BC) who described a severe colic in a worker involved in extracting the metal.Marcus Vitruvius Pollio (first century BC) who lived in the time of Caesar and Augustus noted the sickly appearance of people who were in contact with lead and Pliny also noted similar effects. Lead was widely used by the Romans in plumbing, with tin as an alloy for making kitchen utensils, as a sweetener in cooking, a food additive and for the preparation of alcoholic drinks. The main source of lead poisoning for the Roman aristocracy was ‘grape syrup’. This was made from unfermented grape syrup and was boiled down to a third of its original volume in lead-lined bronze cauldrons.Despite these early reports of the adverse effects of lead, the public health history of lead is surprisingly short and is virtually non-existent for some 98% of the time in which lead has been used. Of the 300 million tons of lead produced, approximately 50% is estimated to have been dispersed to the various environmental compartments.20 It is really only since the 1970s that the pre-eminent role of lead as an environmental toxicant has been recognised. There are many reasons for this. However, the three main reasons are l(h), (c) and (4as outlined in Table 2. The dominance of socioeconomic issues over health issues is nothing new and many examples exist. Decision-making regarding the use of lead as well as the control of research and information which might highlight the adverse effects of lead exposure was largely vested in the producers and users of lead. One of the pioneers in the fight against lead exposure was Alice Hamilton. Her biography*I and the review by Lippmann22 provide a detailed insight into the difficulties she encountered in her fight against lead.She carried out a systematic study of lead poisoning among industrial workers in Illinois in 19 10 and the results encouraged her to embark on a career in occupational medicine. Tetraethyl lead has occupied a pre-eminent position as a source of environmental lead. This is largely due to its volatile nature and its widespread and rapid dissemination. The history of leaded gasoline began in Ohio in 1923 when gasoline containing lead additives was first used and since then the motor car has played a central role in the dissemination of lead.Since organic lead compounds were first added to gasoline in 1923, background lead levels have increased world-wide, including the Greenland ice cap. Hamilton first expressed her anxiety Chemical Society Reviews, 1997 139 about the use of tetraethyl lead as early as 1925 and it is now patently obvious that failure to take account of her warnings have had tragic consequences from the point of view of public health. On May 20, 1925 the US Surgeon General convened a conference to assess the tetraethyl lead situation. The industry situation was summarised as follows: (a) leaded gasoline was essential to the industrial progress of America; (b) any innovation entails certain risks; (c)the major reason that deaths and illnesses occurred at plants was that the men who worked there were careless and did not follow instructions.Dr Yandell Henderson, a physiologist at Yale University presented severe criticism of the use of tetraethyl lead at that meeting.23 He spoke about the probability that leaded gasoline would be in universal use in cars. Arising from the conference, a special committee was established to conduct an investigation of leaded gasoline. The committee concluded some seven months later that ‘in its opinion there are at present no good grounds for prohibiting the use of ethyl gasoline .. . provided that its distribution and use are controlled by proper regulations.’ The committee saw its report as an interim one and recognising its limitations, they concluded that further studies by the Government were essential. In view of what has come to pass it is unfortunate that these investigations did not take place for more than four decades. The tetraalkyl lead compounds added to gasoline function by decomposing at engine temperatures into atomic lead and alkyl free radicals. These atoms of lead must be removed from the engine, otherwise build-up of lead deposits would eventually destroy the engine. In order to achieve this, small quantities of ethylene dibromide and dichloride are also added so that lead is emitted as PbC12, PbBr2 and PbBrC1.These are eventually converted to PbO. The lead oxide exists as particulate matter in the atmosphere for a considerable period so that it is widely dispersed. Arising from this process, a significant percentage of ingested lead has its origin in leaded gasoline. Ca. 1% of the organolead content of petrol is emitted into the atmosphere via the exhaust unchanged and the elegant studies of Harrison and his coworkers on aerosols24 provide details of the sources, speciation and transport modes of alkyllead compounds in the environment. The requirement to equip cars sold in the US and Europe with catalytic converters has accelerated the conversion to the use of unleaded gasoline. This has had a dramatic effect on both ambient lead levels in air and blood lead levels in these countries as is illustrated in Figs.6 and 7.25 Although only a small percentage of world lead production ends up in gasoline (Fig. 2), it represents a very efficient method of distributing lead 1805’160b n 2 140 0 c 00 120 r ai .r-100 8a CJ) 80 .-c U a, 60f 2 40 8 x 20 a,J in the environment. Leaded petrol is however still used in many countries. 10 Distribution of lead in the environment Fig. 8 outlines the most significant contributors to current body burdens of lead.25.26 It is apparent that human intake can occur by a variety of pathways. These include inhalation of airborne particles, water, food, wines and other alcoholic drinks.It may also be ingested as a result of leaching of lead from cooking and drinking utensils. Drinking water represents a significant source of lead for many people and in recognition of this, many Governments impose maximum concentrations of lead in drinking water. The current World Health Organization recommendation and EU limit is 50 pg 1-l. The concentration of lead in drinking water depends on the lead content of the water source which in turn depends on physicochemical properties such as pH and hardness. It also depends on the nature of the plumbing. Lead in food can arise from the inherent concentrations of lead present, but in certain circumstances, significant increases can arise from the containers in which the food is stored.In general, meat and other animal products make a relatively small contribution to total human lead intake due to fact that lead is not preferentially deposited in edible portions of animals. The lead content of vegetables is highly variable and depends greatly on the lead content of the soil. In the past, canned foods contributed significantly to lead intake.12 This was due to the lead solder used in manufacturing the cans. Acidic foods and foods stored in acid media (vinegar) enhanced lead leaching/migration and significantly increased the lead content of the food. In New Zealand for example, canned food makes up 8.38% of the diet but is responsible for 34.4% of the lead i11take.~7 Replacement of soldered cans by welded cans is eliminating this problem.Significant quantities of lead may be ingested by workers in industries using large quantities of lead-containing materials. This is evidenced by the relatively high blood lead levels (PbB) found in workers in these industries. In recognition of this, international organisations such as OSHA have recommended a permissible exposure limit of 50 pg m-3 in the working place air. The current EU limit for lead in ambient air is 2 pg m-3 and the WHO assume that each 1 pg m-3 of lead in ambient air contributes ca. 1.9 pg dl-1 to a child’s blood and ca. 1.6 pg dl-* to an adult’s. A current WHO draft guideline recommend a fourfold reduction in the ambient air concentration to 0.5 pg m-3. 01 I I I I I I I I I I0 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 Calendar year Fig.6 Comparison of lead used in gasoline and ambient lead concentrations, 1975-1983.25 140 Chemical Society Reviews, 1997 v)s 110 I I 1 I -17 +-.-0 4 Lead used in gasoline (A)+-L -16E 100 -b m rnz -15 2=-90 --u) .-0 -14 a,Pb," 80--0a c. t -130 070-0 .-u) -12 1 Q,(dBlood lead levels (0) rn TJ 60--11 5 s a3 50--10 Q,-(d 1 I I I A9-40P 1976 1977 1978 1979 1980 Fig. 8 Pathways of lead from the environment to and within man25.26 Why is there such a concern about lead? The main reason is represent safety margins of 10-1000 depending on the quantity that the safety margin between unavoidable ingestion from food and quality of the toxicity data available.In the case of and water and levels which constitute a significant risk to health lead, such margins do not exist. The FAO/WHO Provisional is relatively small. In the case of chemicals for which Tolerable Weekly Intake of dietary lead is 3 mg for an adult. occupational exposure levels exist, these levels frequently The average person in Great Britain is estimated to have a Chemical Society Reviews, 1997 141 dietary intake of up to 0.42 mg per week, excluding the contribution from drinking water.28 This same report shows that the intake from drinking water can vary greatly from 2.5 mg per week in areas such as Glasgow and Ayr, areas with a plumbosolvent water supply and much lead piping to 0.19 mg per week in Birmingham, an area where lead levels in water are not unduly elevated.11 Fate of ingested lead Once ingested, lead is absorbed from the gastrointestinal tract or excreted. The fraction absorbed varies with the nutritional state of the victim but it is generally accepted that adults absorb on average 10-15% of the quantity ingested while children can absorb up to 50%. Absorption is greatly enhanced by fasting and both iron, zinc and calcium deficiency. Typically, men in the 60-70 age bracket will have accumulated ca. 200 mg of a lead body burden while women will have somewhat lower body burdens. This represents an accumulation of an average of some 9 pg per day. The ionic radius of Pb2+ (132 pm) is similar to that of Ca2+ (106 pm) and Sr*+ (127 pm) and in the body it has a very similar chemistry.Once in the blood, lead is distributed primarily among three compartments (a) blood, (b) soft tissues e.g. kidney, bone marrow, liver and brain and (c) mineralising tissue e.g. bones and teeth. Since alkaline earth elements are concentrated by bone surfaces and mineralising tissues, it is not surprising that they contain ca. 95% of the total body burden of lead in adults. Lead substitutes for calcium in bone and is anaemia arising from inhibition of haem synthesis, chronic encephalopathy, cognitive impairment, sleeplessness, headaches, aggressive behaviour, convulsions, disruption of the motor system and renal effects. Fig. 9 shows the lowest observable adverse effect levels (LOAEL) of inorganic lead in both children and adults.29 13 Neurologic effects of lead-effects on cognitive development in children By far the most sensitive target of lead poisoning is the central nervous system.In children, neurologic deficits have been documented at exposure levels once thought to be safe. Indeed, as recently as 1969, a blood lead level of 60 pg dl-1 was thought to be safe. In 1970, the US Surgeon General reduced the toxicity level to 40 pg dl- l. In response to research which suggested that cognitive deficiency resulted from exposure at this level, the Centers for Disease Control reduced the intervention level to 30 pg dl-1 in 1978, to 25 pg dl-l in 1985 and finally to 10 pg dl-l in 1991. Lead toxicity in children is a very emotive issue.The landmark paper in this area was by Herbert Needleman and published in the New England Journal of Medicine of 29th March 1979.3O He concluded that children having high dentine lead levels scored significantly less well on the Wechsler Intelligence Scale for Children than children having low dentine lead levels. This is perhaps one of the most controversial scientific papers ever published and was subjected to much criticism, in particular by Claire Ernhart, a psychologist at Case Western Reserve University. Such was the controversy that the incorporated into the hydroxyapatite crystal, C~,O(PO~)~(OH)~. EPA convened a special panel to examine both Needleman’s This occurs both during bone formation and remodelling arising from bone turnover.Most of the lead ingested is excreted in the urine with smaller quantities lost in faeces, sweat, hair and nails. Due to bone turnover in the body, the original surface deposits of lead are redistributed. 99% of lead in the blood is associated with erythrocytes while the remaining 1% is in the plasma where it is available for transport to the tissues. The lead in mineralising tissues is distributed between two compartments, a labile pool which readily exchanges lead with blood, and an inert pool which poses a special risk in that it may be mobilised in times of stress, thus providing a significant endogenous source of lead which can significantly increase the level of lead in the blood. In adults, it is estimated that the half life of lead in blood is ca.25 d, in soft tissue ca. 40 d and in the inert pool in the bone, ca. 25 y. This highlights the risks associated with chronic lead exposure and underlines the difficulties of dealing with lead poisoning. 12 Health effects of lead-saturnismus A wide variety of adverse health effects arise from exposure to lead. In earlier times lead was associated with the planet Saturn, hence the expression ‘saturnismus’ is frequently applied to lead poisoning. Castellino’9 expands on several theories regarding the origin of this analogy. Since lead is the oldest metal, it was associated with Saturn, who was considered one of the oldest gods on Mount Olympus. Others ascribe the association to the fact that lead absorbs other metals and has thus been represented by the symbol of the god Saturn who devours children! A graphic account of the onset of saturnism is provided in an article by Prendergast and he reported a 9.4% fatality rate in adults with a 6.4% fatality rate in females. Acute exposure to lead leads to (a)gastrointestinal cholic and (b)encephalopathy in children.However, acute exposure to lead is thankfully now a rarity and chronic effects are more common. Lead primarily affects the peripheral and central nervous systems, the blood cells and metabolism of vitamin D and calcium. There are many chronic effects arising from lead exposure, however the following are some of the more common: and Ernhart’s work. They found several deficiencies in Needleman’s ~ork.3~ Ironically, they also found fault with Ernhart’s work.The controversy still rage~.~~-3~ Taking all the data p~blished,~6 it can be safely concluded that measurable effects on cognitive development are evident in children having blood lead levels > 10 pg dl-l. Postnatal as opposed to prenatal exposure would appear to have a greater influence on cognitive development. Nutritional deficiencies of essential metals can increase the effects of lead exposure by enhancing the absorption and toxicity of dietary lead. The metals having the greatest influence are calcium, iron and zinc. 14 Monitoring of lead exposure Two aspects of lead are of interest. In the first the actual lead concentrations are measured in order to provide a measure of the dose index.The dose index most frequently used is the blood lead concentration (PbB). This is usually expressed in terms of pg dl- or pmol l-1 (1 pmol 1-= 20.7 pg dl-l). The former is the index used by most health authorities in setting the maximum allowable body burden of lead and it is the index most frequently quoted in studies of lead exposure. The PbB represents the equilibrium existing between lead intake, reten- tion, body stores and excretion rate. However, other indices can also be used and it is often asserted that lead concentration in teeth and bone provide a better indication of the body lead burden. Urinary excretion of lead following administration of EDTA is also used. Considerable interest has been expressed in indices of the biochemical effects of lead absorption.Ideally, these should highlight effects arising from relatively low levels of exposure before the usual clinical symptoms are manifested. Perhaps the most widely studied biological indicator of lead exposure has been the measurement of erythrocyte zinc protoporphyrin (ZPP). Lead interferes with the synthesis of haemoglobin by interfering with several steps in the haem pathway. The incorporation of Fe2+ into the molecule of protoporphyrin IX represents the final step of haem synthesis. The enzyme ferrochelatase which catalyses the insertion of iron into 142 Chemical Society Reviews, 1997 Lead concentration Children in blood Adults1 1 (pg Pb dl-') 150 Death --+ Encephalopathy -3 Naphropathy -W 100 --Encephalopathy Frank anaemia Frank anaemia + Colic - Decreased longevity 50 4-Hemoglobin synthesis Hemoglobin synthesis -40 Peripheral neuropathic Infertility (men) Nephropathy Vitamin D metabolism ? -W 30 Systolic blood presure (men) ? Hearing acuity 4 +--Erythrocyte protoporphyrin (men) ? Nerve conduction velocity 44 20 Erythrocyte protoporphyrin ? Vitamin D metabolism -11-Developmental toxicity - C-Erythrocyte protoporphyrin (women) ? 10 +--Hypertension ? Growth & Fig.9 Effects of inorganic lead in children and adults-lowest observable adverse effect levels (ATSDR-toxicological profile of lead, 1989). ? ,increase;& ,decrease protoporphyrin IX is very sensitive to lead.Exposure to lead inhibits the enzyme activity of ferrochelatase and thus reduces the availability of Fe2+. When Fe2+ is not available, the enzyme binds Zn2+ instead and thus the concentration of ZPP in blood is elevated. It has been demonstrated that measurement of ZPP is not sufficiently sensitive at low levels of lead poisoning to be acceptable as a reliable indicator of lead exposure. However, the ZPP level in blood declines more slowly than blood lead so that ZPP levels can be an indicator of intoxication arising from lead resorbed from bone. 15 Chelation therapy Chelation therapy is a therapeutic procedure in which a chelating agent is administered to a patient in order to selectively remove a toxic metal ion such as lead. Such chelating agents can act in two ways.They can increase excretion of the metal, thus reducing the body burden or alternatively they may limit absorption of the metal from the gastrointestinal tract. Ideally, the chelating agent should have a high affinity for the metal ion to be removed and have a low affinity for essential metal ions such as Zn2+, Ca2+ and Fe2+. Several chelating agents have been used in the treatment of lead poisoning. Examples include calcium disodium ethylene- diaminetetraacetate, 2,3-dimercaptopropanol (British Anti Le- wisite, BAL), 3-mercapto-~-valine (D-penicillamine) and meso-2,3-dimercaptosuccinic acid (succimer, chemet). Although CaNa2EDTA has been used for over fifty years, relatively little is known regarding its efficacy.33 Succimer was the first orally administered drug to be approved by the FDA for treatment of children with blood lead levels of > 45 pg dl-1.Clinical experience with all of these drugs is relatively limited and emphasis should be placed on removing the victim from the source of lead rather than on chelation therapy. 16 Migration (leaching) of lead from lead crystal 16.1 Background In January 1991 Graziano and Blum published a paper37 in which they demonstrated that significant concentrations of lead leached or migrated from crystal decanters and glasses containing port. Despite the fact that earlier papers38739 reported similar findings, the Graziano and Blum paper achieved wide publicity and articles appeared in newspapers and magazines world-wide questioning the safety of lead crystal stemware. This caused significant anxiety in the lead crystal industry.In response, the International Crystal Federation (ICF), an um- brella organisation of European and US crystal manufacturers which was set up in April 1991, undertook an urgent investigation of the problem. International Standards IS0 7086/1 and 7086/2 regulate the release of lead and cadmium from glassware and ceramicware in contact with food. The test consists of filling the container with 4% (v/v) acetic acid and leaving it to standing at 22 "C for 24 h. The maximum Chemical Society Reviews, 1997 143 permissible release of lead from small hollow-ware (< 600 ml) is 5 mg 1-l and from large hollow-ware (> 600 ml) 2.5 mg 1-1.The ICF have adopted voluntary standards for lead migration which are significantly lower than those demanded by the IS0 standard. These are 1.5 mg 1-1 for small hollow-ware and 0.75 mg I-' for large hollow-ware. In addition, they recommend that spirits should not be stored in decanters for extended periods unless the decanters have been treated to eliminate lead migration. Adherence to these guidelines will not result in any undue exposure to lead having its origin in lead crystal. Wines themselves may contain lead concentrations in the range 10-500 pg 1-1. Typical 24% PbO crystal currently manu-factured can be expected to leach ca. 10 pgPb during three 1 h contacts of 150 ml of wine per contact (a total of 450 ml). Most of the leaching occurs during the first two contacts so that successive contacts will not significantly increase the lead intake.These conditions represent an extreme case in view of the quantity ingested and the contact time. However, such an intake does not present any undue risk to the average adult, particulary in view of the fact that lead absorption during meals in less than 10%. The leaching of lead from lead crystal glass is a diffusion controlled ionic exchange process and follows Fick's law. This is demonstrated by eqn. (3), where t is the time. The slope of the plot, a, is largely controlled by the composition of the glass. [Pb] = a*+b (3) The ordinate value, b, represents the initial 'spike' released and is influenced by acid polishing! and surface treatments.16.2 Reduction of lead migration Various approaches have been used to reduce or eliminate lead leaching. These include: (a) The use of non-lead glass liners in decanters. (b) Alterations to the lead glass composition. These include slight increases in the silica content, reduction of the alkali oxide (K20, Na20) content, optimising the Na20/K20 ratio and addition of trace quantities of other oxides such as MgO, ZnO. (c) Use of various surface treatments. These include acid polishing, ammonium sulfate fuming and the use of coatings such as sol-gels and various polymers. Acid polishing of lead crystal items in sulfuric acid-hydroflu- oric acid baths has been shown to significantly reduce lead migration by reducing the concentration of lead in the surface layer of the glass.The use of ammonium sulfate fuming, a process pioneered many years ago by Waterford Crystal, is a remarkably simple procedure that significantly reduces lead migration. A small quantity of ammonium sulfate (0.1-0.2 g) is placed in the glassware before it is placed in the lehr for annealing. Best results are obtained when the temperature is in the range 470-490 "C. The procedure is particularly suited for decanters, however they must be sealed before being acid polished as acid contact diminishes the impact of the ammonium sulfate treatment. The chemistry involved is believed to be as follows: 100-360 "C Decomposition of ammonium sulfate (NH4)2S04 + t NH3 + (NH4)HS04 > 360 "C Vapourisation of ammonium sulfate and reaction with crystal (NH4)HS04+2 Si-0-M + 2Si-OH + M2SO4 + t NH3 M = Na, K fi Following mechanical cutting of the crystal, the surface of the cuts is opaque.Acid polishing in mixtures of sulfuric and hydrofluoric acid renders the crystal transparent once more. 144 Chemical Societv Reviews. 1997 (NH4)HS04 + 2Si-0-M = 4Si-OH + PbSO4+ NH3 Condensation of Si-OH 4 Si-OH +2 Si-0-Si +2H20 This treatment reduces lead migration significantly and has good long-term behaviour. Typical effects results in reduction from 600 ppb to less than 100 ppb using the IS0 test. 17 Lead-what is the future? Lead is an important element and still has many industrial uses.However, there are undoubtedly significant risks associated with excessive exposure to lead either in the home or in the workplace. In recognition of this the Organisation for Economic Cooperation and Development (OECD) signed a ministerial declaration on risk reduction for lead at their meeting in Paris on 20th February 1996. This was the conclusion of a long debate within the OECD on the need to place mandatory restrictions on lead products. The main points in the OECD declaration are: 0 Phase out of the use of lead in gasoline except where required for essential or specialised uses where there are no practical alternatives. 0 Elimination of exposure of children to lead arising from products intended for use by children.0 Eliminate exposure to lead from food packaging. 0 Phase out the use of lead in paint and rust-proofing agents except in cases of essential or specialised uses for which there are no practical alternatives. 0 Restrict exposure to lead from the leaching/migration of lead from ceramic ware and crystalware used for food and beverages. 0 Restrict the use of lead shot in wetlands. 0 Promote the use of alternatives to lead weights used for fishing in shallow water. 0 Reduce lead levels in drinking water. 0 Reduce the levels of lead in occupational environments. 0 Limit lead in air emissions. 0 Establish strategies to abate significant exposures to lead arising from the historic use of lead-containing materials in buildings. These measures are not inconsistent with the views of Simms40 who proposed a scientific basis for regulating lead contamination and there is no reason to believe that their implementation will not result in a significant reduction in exposure to lead.This will be manifested by decreases in the average blood lead levels in the general population and a reduction in the numbers at risk to elevated blood lead levels. Fig. 10 summarises the findings of the third National Health and Nutrition Examination Study (NHANES 111) carried out to assess the health and nutritional status of the US population. The previous study, NHANES 11, carried out from 1976-1980 reported an average blood lead level of 12.8 yg dl-1. The initial results of NHANES 111, which covers the period from 1988-1 99 1 has found that the average blood lead level in the US population is now 2.8 pg dl-1.The decline is significant across all age groups. The largest decline was in the 6-19 years age group with large reductions also found in the most sensitive age group, 1-5 years where the average blood lead level decreased from 14.9 to 3.6 pg dl-l. It is estimated that approximately 1.7 million US children currently have blood lead levels greater than 10 yg dl-1, the intervention level set by the Centers for Disease Control. This is considerably less than the number claimed by environmental pressure groups. Surely the most significant finding of this study is that these striking reductions in blood lead levels have been achieved despite an increase in lead consumption from 4.8 to 5.26 kg per person over the same period.This clearly demonstrates that lead can be used safely in modem society. These findings are supported by recent data from Australia. In 1993 the National Health and Medical Research Council 12.8~~3dl-' JO 1980 1990 LO Lead use and exposure trends for US population (NHMRC, recommended that the blood lead levels of all Australian children be held below 10 pg dl-1 and set a goal to have 90% of children in the 1-4 years age group below this level by 1998. A nation-wide study was carried out in early 1995 and the results have been recently published (Lead in Australian Children). The results show that the NHMRC goal has already been surpassed.Of the 1575 blood lead samples collected, 92.7% (1460) were less than the NHMRC target level of 10 pg dl-1. This survey in common with most previous surveys found that social factors such as lower income and education, cleanliness of the home, peeling of interior paint and year of home construction were associated with high blood lead levels. The implementation of OECD recommendations should provide further impetus to a general reduction in blood lead levels. In addition, further reductions in the permissible concentrations of lead in ambient air will be implemented. In the EU, this is currently 2 pg m--3. However, the WHO working group on inorganic air pollutants has recently recommended a fourfold reduction to 0.5 pg m-3 and the UK Department of the Environment has recently proposed that such a limit be implemented. Bearing in mind that the upper limit for lead of nonanthropogenic origin in blood is 3 pg dl-1, such a proposal would ensure that children would have levels of less than 10 pg dl-*, the level at which the WHO considers the onset of adverse effects to begin.Since traditional medical approaches such as chelation therapy are neither appropriate nor logistically feasible for the treatment of people who have been exposed to low levels of lead, the sole option for preventing low-level lead effects in millions of young children is to preclude further exposure. Similar arguments apply to adults. This will have important implications in the workplace.Although significant advances have been made in reducing blood lead levels, the growing awareness of the potential of chemicals such as lead to have neurotoxic effects and to induce impairment of the central nervous system may in the near future result in these effects challenging cancer as a prime focus of toxicity research on chemicals. The International Program on Chemical Safety (IPCS), a joint venture of the United Nations Environmental Program, the International Labour Organisation and the World Health Organization have finally released their position document on lead.26 This document addresses most of the major health impacts of lead and the lead industry accepts that it presents a thorough and balanced view.It outlines a number of areas for future research as follows: To define the health significance of biochemical changes associated with exposure to lead, with particular attention to alterations associated with blood lead concentrations of the order of 15 pg dl-1 or less. To define the bioavailability of lead from various sources and to establish the relationship between exposure (source and speciation) and body burden. To define the influence of host-related factors (particularly nutrition) affecting absorption and distribution of lead. To intensify kinetic studies of lead in order to provide an improved database for extrapolation between species. To elucidate mechanisms of accumulation and mobilisation of lead from bone with particular attention to the influence of pregnancy and ageing on the kinetics.To investigate the pharmacokinetics of lead in pregnancy in relation to transfer of lead to the developing embyro and foetus and factors that mitigate such transfer. To determine the effects of pre- and post-natal exposure to lead. To improve the defining of paternally mediated effects of lead exposure on the reproductive process and outcomes. It is apparent that these recommendations address many of the acknowledged deficits highlighted in this review. It is worthwhile noting that the International Lead Zinc Research Organization (ILZRO) has already initiated major research programmes in a number of these areas.41 18 References 1 OECD, Risk reduction monograph No.I: Lead background and national experience with reducing risk, OCDE/GD(93)67, Organisation for Economic Co-operation and Development, Paris, 1993. 2 The properties of optical glass, ed. H. Bach and N. Neuroth, Springer Verlag, Berlin, Heidleberg, 1995. 3 D. R. Bamford, Colour generation and control in glass, Elsevier, Amsterdam, 1977. 4 J. Hlavac, The technology of glass and ceramics, Elsevier, Amsterdam, 1983. 5 H. Rawson, Properties and applications of glass, Elsevier, Amsterdam, 1980. 6 H. Scholze, Glass: Nature, structure and properties, Springer Verlag, New York, 1991. 7 M. B. Volf, Chemical approach to glass, Elsevier, Amsterdam, 1984. 8 M. B. Volf, Technical approach to glass, Elsevier, Amsterdam, 1990. 9 W.A. Weyl, Coloured glasses, Society of Glass Technology, Sheffield, 1990. 10 J. Zarzycki, Glasses and the vitreous state, Cambridge University Press, Cambridge, 199 1. 11 69/493/EEC, OJ, 1969, L326, 599. 12 D. M. Settle and C. C. Patterson, Science, 1980, 207, 1167. 13 ATSDR, Top 20 Hazardous Substances. ATSDRIEPA Priority list for 1995, 1996. 14 82/605/EEC, OJ, 1982, L247, 12. 15 77/312/EEC, OJ, 1977, L105, 10. 16 P. Mushak, Environ. Res., 1992, 59, 281. 17 A. R. Flegal and D. R. Smith, Environ. Res., 1992, 58, 125. 18 E. Farber, The Evolution of Chemistry, Roland, New York, 1952. 19 N. Castellino, P. Castellino and N. Sannolo, Inorganic lead exposure: metabolism and intoxication, Lewis Publishers, Boca Raton, 1995. 20 NAS, National Academy of Sciences: Lead in the Human Environment, National Research Council, Washington, DC, 1980. 2 I A. Hamilton, Exploring the dangerous trades, Little-Brown, Boston, 1943. 22 M. Lippmann, Environ. Res., 1990, 51, 1. 23 D. Rosner and G. Markowitz, Am. J. Public Health, 1985, 75, 344. 24 M. Radojevic and R. M. Harrison, Sci. Total Environ., 1987, 59, 157. 25 EPA, EPA-60018-83-028 US. EPA, Environmental Criteria and Assessment Office, Research Triangle Park, NC, June 1986., 1986. 26 IPCS, Environmental Health Criteria I65: Inorganic Lead, WHO International Program on Chemical Safety, Geneva, 1996. 27 L. Pickston and J. Drysdale, Food. Technol. NZ, 1986, 21, 16. Chemical Society Reviews, 1997 145 28 MAFF, Ministry of Agriculture, Fisheries and Food. Lead in food: Progress report. The twenty-seventh report of the Steering Group on food surveillance. The working party on inorganic contaminants in food. Third supplementary report on lead, HMSO, 1989. 29 ATSDR, Case studies in environmental medicine: lead toxicity, Agency for Toxic Substances Disease Registry, Atlanta, GA., 1990. 30 H. L. Needleman, C. Gunnoe, A. Leviton, R. Reed, H. Peresie, C. Maher and P. Barrett, N. Engl. J. Med., 1979, 300, 689. 31 E. Marshall, Science, 1983, 222, 906. 32 J. Palca, Science, 1991, 253, 842. 33 H. L. Needleman and R. J. Jackson, Pediatrics, 1992,89, 678. 34 C. B. Emhart, Arch. ofEnviron. Health, 1994, 49, 77. 35 H. L. Needleman, J. A. Riess, M. J. Tobin, G. E. Biesecker and J. B. Greenhouse, J. Am Med. Assoc., 1996, 276, 363. 36 S. Tong, P. Baghurst, A. McMichael, M. Sawyer and J. Mudge, Brit. Med. J., 1996, 312, 1569. 37 J. H. Graziano and C. Blum, The Lancet, 1991,337, 141. 38 H. Scholze and R. Sauer, Glastech. Ber., 1974, 47, 149. 39 E. A. DeLeacy, Med. J. Aust., 1987, 147, 622. 40 D. L. Simms, Sci. Total Environ., 1986, 58, 209. 41 ILZRO, Environmental Update, 1996, 6, 2. Received, 12th August I996 Accepted, 19th December 1996 146 Chemical Society Reviews, 1997

ISSN:0306-0012    

DOI:10.1039/CS9972600133

出版商:RSC    

年代:1997

数据来源: RSC