摘要:

Chemical Society Reviews Vol 10 No 2 1981 Page MELDOLA MEDAL LECTURE The Relationship between Metal Carbonyl Clusters and Supported Metal Catalysts By J. Evans 159 Photochemistry and Photocyclization of Aryl Halides By J. Grimshaw and A. P.de Silva 181 Singlet Molecular Oxygen By A. A. Gorman and M. A. J. Rodgers 205 CHEMICAL ASPECTS OF TRACE CONSTITUENTS OF THE DIET I Control and Surveillance of Trace Constituents -Is There a Need? By D. G. Lindsay 233 II Sources of, and Analytical Advances in, Trace Inorganic Constituents in Food By C. J. Pickford 245 III Advances in the Analysis of Trace Organic Constituents of the Diet, with Particular Reference to Mass Spectrometry By J. Gilbert and R. Self 255 IV Nutritional Chemistry of Inorganic Trace Constituents in the Diet By J.K.Chesters 270 V Sources and Biogenesis of Trace Organic Constituents of the Diet By S. A. Slorach 280 The Royal Society of Chemistry London Chemical Society Reviews EDITORIAL BOARD Professor K. W. Bagnall, B.Sc., Ph.D., D.Sc., C.Chem., F.R.S.C. (Chairman) Professor K. R. Jennings, M.A., D.Phil., C.Chem., F.R.S.C. Professor G. W. Kirby, M.A., Ph.D., Sc.D., F.R.S.E., C.Chem., F.R.S.C. Professor B. L. Shaw, B.Sc., Ph.D., F.R.S. Chemical Society Reviews appears quarterly and comprises approximately 20 articles (ca. 500 pp) per annum. It is intended that each review article shall be of interest to chemists in general, and not merely to those with a specialist interest in the subject under review.The articles range over the whole of chemistry and its interfaces with other disciplines. Although the majority of articles are intended to be specially commissioned, the Society is always prepared to consider offers of articles for publication. In such cases a short synopsis, rather than the completed article, should be sub-mitted to The Managing Editor, Books and Reviews Section, The Royal Society of Chemistry, Burlington House, Piccadilly, London, W1V OBN. Members of the Royal Society of Chemistry may subscribe to Chemical Society Reviews at €10.50 per annum; they should place their orders on their Annual Subscription renewal forms in the usual way. 1981 Annual subscription price, U.K. E31.00, Rest of World €33.00, U.S.A.$78.00 including air speeded delivery. Application to mail at second class postage rate is pending at Jamaica, N.Y. 11431. Change of address and orders with payment in advance to The Royal Society of Chemistry, The Distribution Centre, Blackhorse Road, Letchworth, Herts SG6 1HN England. Air Freight and mailing in the U.S. by Publications Expediting Inc., 200 Meacham Avenue, Elmont, New York 11003. All other despatches outside the U.K. by Bulk Airmail, and Accelerated Surface Post outside Europe. Note to subscribers. Regrettably publication of the four issues has still not reverted to the usual quarterly dates. The cause of this is a persisting shortage of articles (the production problems of recent years have been largely overcome) but the setting-up of an Editorial Board should result in an increase in the commissioning of reviews. 0 Copyright reserved by The Royal Society of Chemistry 1981 ISSN 0306-0012 Published by The Royal Society of Chemistry, Burlington House, London, WlV OBN Printed in England by Eyre & Spottiswoode Ltd, Thanet Press, Margate.

ISSN:0306-0012    

DOI:10.1039/CS98110FP003

出版商:RSC    

年代:1981

数据来源: RSC

ISSN:0306-0012    

DOI:10.1039/CS98110FX005

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

Chemical Society Reviews Vol 10 No 2 1981 Page MELDOLA MEDAL LECTURE The Relationship between Metal Carbonyl Clusters and Supported Metal Catalysts By J. Evans 159 Photochemistry and Photocyclization of Aryl Halides By J. Grimshaw and A. P. de Silva 181 Singlet Molecular Oxygen By A. A. Gormtnn and M. A. J. Rodgers 205 CHEMICAL ASPECTS OF TRACE CONSTITUENTS OF THE DET E Control and Surveillance of Trace Constituents -Is There a Need? By D. G. Lindsay 233 I1 Sources of, and Analytical Advances in, Trace Inorganic Constituents in Food By C. J. Pickford 245 I11 Advances in the Analysis of Trace Organic Constituentsof the Diet, with Particular Reference to Mass Spectrometry By J. Gilbert and R. Self 255 IV Nutritional Chemistry of Inorganic Trace Constituents in the Diet By J. K. Chesters 270 V Sources and Biogenesis of Trace Organic Constituents of the Diet By S. A. Slorach 280 The Royal Society of Chemistry London

ISSN:0306-0012    

DOI:10.1039/CS98110BX007

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

MELDOLA MEDAL LECTURE* The Relationship between Metal Carbonyl Clusters and Supported Metal Catalysts By J. Evans DEPARTMENT OF CHEMISTRY, THE UNIVERSITY, SOUTHAMPTON, SO9 5NH 1 Introduction An important class of heterogeneous catalysts consists of small metal particles (with average diameters of 1 to -20 nm) dispersed on a high surface area N oxide support.1 These materials are heterogeneous in many senses since both the metal particle sizes and oxide supports are non-uniform and this makes it virtually impossible to describe their chemistry on a molecular level. For example, a wide variety of potential chemisorption sites exist and it is difficult, on the basis of any spectroscopic or chemical probe, to ascertain which ones are occupied by a substrate.In the past few years, studies of the chemistry of single-crystal metal surfaces have markedly increased the understanding of chemisorption pro- cesse.x2 However, even in these idealized cases, identifying adsorption modes can be difficult. Transition-metal carbonyl clusters are an ever extending class of complexes, which can be isolated and their structures determined by the conventional methods of X-ray diffraction, vibrational spectroscopy, and nuclear magnetic resonance.3 Presently, complexes with up to 38 metal atoms are known.4 There is an appealing parallel between cluster complexes and metal surface chemistry, which has been reviewed at length.5 In this review we shall discuss two con- tributions that cluster chemistry can make to surface studies, viz, as structural models and catalyst sources.2 Clusters as Structural Models A. Metal Polyhedra.-A summaryof the metal skeletons adopted for tri- to *Based on the lecture given on 25th September, 1980, at the Autumn Meeting, University College, Cardiff. J. R. Anderson, ‘Structure of Metallic Catalysts’, Academic, New York, 1975. * e.g., J. C. Buchholz and G. A. Somorjai, Acc. Chem. Res., 1976, 9, 333; G. A. Somorjai, ibid., p. 248. B. F. G. Johnson, ed., ‘Transition Metal Clusters’, Wiley, New York, 1980. P. Chini, J. Organomet. Chem., 1980, 200, 37. E. L. Muetterties, T. N. Rodin, E. Band, C. F. Brucker, and W. R. Pretzer, Chem. Rev., 1979, 79, 91; E. L. Muetterties, Zsr. J, Chem., 1980, 20, 84. Relationship bet ween Metal Carbonyl Clusters and Supported Metal Catalysts hexa-nuclear metal carbonyl clusters is given in Table 1.In each case, the most commonly found geometry is listed first and is in accord with chemical intuition. Table 1 Geometries of tri- to hexa-nuclear metal skeletons in metal carbonyl clusters No of metals Metal polyhedron Triangular Linear Bent 4 Tetrahedron Butterfly Planar, D2h Triangular + 1 terminal 5 Trigonal bipyramid Square pyramid Edge bridged tetrahedron Vertex linked triangles (planar) (skewed) Triangle + 2 terminal 6 Octahedron Trigonal prism Capped trigonal pyramid Capped square pyramid a P. R. Raithby in ref. 3, p. 5; b G. Longoni, M. Manassero, and M. Sansoni, J. Am. Chem. Soc., 1980, 102, 3242; C D.H. Farrar, B. F. G. Johnson, J. Lewis, J. N. Nicholls, P. R. Raithby, and M. J. Rosales, J. Chem. SOC.,Chem. Commun., 1981,273;d J. R. Shapley, G. A. Pearson, M. Tachikawa, G. E. Smidt, M. R. Churchill, and F. J. Hollander, J. Am. Chem. SOC.,1977,99, 8064. However the geometry of a metal skeleton in a particular complex is dependent upon both electronic and steric effects. Two electron counting procedures are commonly applied to these systems, viz, the effective atomic number rule, assuming each edge of the skeleton to be equivalent to a two-electron bond, and the skeletal electron counting rules.6 Neither of these systems is reliable, although the latter scheme is fairly effective when the majority of the skeletal vertices are isolobal with the BH units for which these rules were formulated, e.g., M(C0)3.7 For example, the reduction of OS~(CO)~~ to its dianion is accompanied by a skeletal change from a capped trigonal bipyramid to an octahedron.8 These structures would be anticipated for six and seven skeletal pairs, respectively.On the other hand, protonation of [Fe4(C0)13l2- to give [HFe4(C0)13]-, which involves no change in the total number of electrons, also causes a change in metal * K. Wade, Adv. Inorg. Chem. Radiochem., 1976, 18, 1.'M. Elian and R. Hoffman, Inorg. Chem., 1976, 15, 1148. 8C. R. Eady, B. F. G. Johnson, and J. Lewis, J. Chem. SOC.,Chem. Commun., 1976, 302; M. McPartlin, C. R. Eady, B. F. G. Johnson, and J. Lewis, ibid.,p. 883. Evans geometry, from a tetrahedron to a b~tterfly.~ This is probably due to the steric problem of accommodating 14 ligands around an iron tetrahedron, and is possible for the larger osmium system.lO This is not a simple situation to apply to supported small metal particles.For each nuclearity there is a choice of geometries that is influenced by the ligands and so the structures of supported metal particles may well alter under catalytic reaction conditions. Metal particles of this size range are very difficult to study and estimations of particle size are made on the basis of chemisorption and elec- tron microscopy measurements. Even then, reports of particle shape are rare. Electron micrographs of a sample of 1 % rhodium on silica gel were thought to show individual metal atoms (2.7 8,diameter of 2.69 A in the metal).ll A penta-gon of rhodium atoms 3.75 A apart (too distant for metal-metal bonding) was thought to be discernible. There has been a more recent attempt to observe the structure of supported rhodium,l2 but in this case the particles were much larger (-20 A). The structural classes of the larger metal carbonyl clusters are shown in Table 2.[Pt38(CO),44l2- represents the largest yet identified and the diameter of the metal skeleton in that complex is 11.6 A.4 This is comparable with the average Table 2 Structural classes of larger metal carbonyl clusters (> 9 metal atoms) (a) Bulk metal types: hcp : [Rh13(C0)24H5-nln-,a [Pt26(co)32]2-,” [N~I~(CO)~IH~-~I~- a ccp : [osloc(co)24l2-,~ [Ptas(CO)-4412- a bcc: [Rh14(C0)25]4- a distorted: ABCB sequence: [Rhz~(CO)37]~- a bcc/close packed :[Rh14(CO)26]~-,~[Rh15(C0)27]3-,a [Rh22(C0)35HzI5-(b) Non-bulk types: Vertex linked octahedra, i.e., [Rh12(C0)3ol2-a Stacked triangles: [Pt3(CO)6]n2- (n = 1-6)a Stacked quadrilaterals : [Rh17(C0)~$3)2]~-,~[R~~~(CO)Z~(C)~I~-,~ [RhioE(CO)2zl3- (E = P,AS)^ Stacked pentagons: [Pt~(CO)z21~-,~[Rh15(C0)28(C)2]-a Separated metals : [PdsFes(C0)24H]3-a Irregular :Rh~,(Cz)(C0)25& a Ref.4; P. F. Jackson, B. F. G. Johnson, J. Lewis, M. McPartlin, and W. J. H. Nelson, J. Chem. SOC.,Chem. Commun., 1980, 224; C J. L. Vidal, Znorg. Chem., 1981, 20, 243; J. L. Vidal, W. E. Walker, and R. C. Schoening, ibid., p.238;d V. G. Albano, P. Chini, S. Martin-engo, M. Sansoni, and D. Strumolo, J. Chem.SOC.,Dalton Trans., 1978,459; C J. L. Vidal, R. C. Schoening, and J. M. Troup, Znorg. Chem., 1981,20,227 M. Manassero, M. Sansoni, and G. Longoni, J. Chem. SOC.,Chem. Commun., 1976, 919. 10 P. A. Dawson, B. F. G. Johnson, J. Lewis, D. A. Kaner, and P. R. Raithby, J. Chem. SOC., Chem. Commun., 1980, 961. l1 E. B. Prestridge and D. J. C. Yates, Nature (London), 1971,234, 345. B. Tesche, H. Knozinger, and B. C. Gates, J. Catal., 1980, 64, 235. Relationship bet ween Metal Carbonyl Clusters and Supported Metal Catalysts diameters reported for many highly dispersed supported metal catalysts. Thus on the basis of size, this class of complex overlaps with the smaller supported metal particles, and they may represent a similar situation to the latter when exposed to carbon monoxide.The variety of structural forms is striking. Several of these are fragments of bulk metal structures, but there is also an extensive series of other skeletal forms. These alternative structures are particularly favoured by inter- stitial heteroatoms, but carbonylate anions also adopt non-bulk forms. The five-fold symmetry13 of [Ptig(C0)22]4- is reminiscent of some of the favoured forms of small arrays of atoms predicted from Leonard Jones potential calcula- tions.14 A centred icosahedron rather than a centred cubo-octahedron (the ccp fragment) was predicted for a 13-atom assembly. The preferred 19-atom structure is compared to the skeleton of the platinum complex in Figure 1 and the dif- ference is only in the relative orientations of the three pentagons.Interestingly, the diffraction interference patterns obtained for amorphous metal-containing materials, e.g., NiP and PdSi, are close to those calculated for a centred icosa- hedr0n.1~ Figure 1 Metal polyhedra of (a) [Pt,,(CO)22]4-and (b) calculated for a 19-atom array B. Ligand Sites.-Polynuclear metal complexes offer the types of ligand sites that are also likely to be available to a substrate on a metal surface. Carbonyl complexes have been used for many years as a data bank for assigning vibra- tional data recorded on chemisorbed CO, and the frequencies of the v(C0) bands used to differentiate between terminal (‘atop’) and bridging (either ~2 or ~3) sites.The chemisorption of carbon monoxide to alumina-supported rhodium provides lS D. M. Waschecheck, E. J. Wucherer, L. F. Dahl, A. Ceriotti, G. Longoni, M. Manassero, M. Sansoni, and P. Chini, J. Am. Chem. Suc., 1979, 101, 6110. l4 M. R. Hoare and P. Pal, J. Cryst. Growth, 1972, 17, 77; Adv. Phys., 1971, 20, 161. l6 C. L. Briant, Discuss Faraday Soc., 1976, 61, 25. Evans a good example of this approach. Three types of site have been identified by infrared spectroscopy.16 Two of these are probably Rh-CO (vco 2070 cm-l) and p2-Rh-CO (VCO 1870 cm-l) units on rhodium particles, and these frequencies correlate well with those observed for CO (surface coverage of Q) on an exposed close-packed rhodium plane (1 11) by electron energy loss spectroscopy (EELS).17 Two other bands occur at 2101 and 2031 cm-l, and although there has been a suggestion that these are due to dicarbonyl sites on the edges of two-dimensional rafts,l* 13C n.m.r.evidence indicates that these are from isolated RhI(C0)z metal centres.19 Recently, new carbonyl bonding modes have been identified that should also be considered in surface studies (Figure 2). The dinuclear complex (1) contains both a semibridging (VCO 1700 cm-I) and anbonded carbonyl group (vco 1560 1-0 /"\,C. I ,/' 'C \ Figure 2 Examples of n-bonded carbonyl ligands l6 J. T. Yates, jun., T. M. Duncan, and R. W. Vaughan, J. Chem. Phys., 1979, 70, 1219. l7 L. H. Dubois and G. A. Somorjai, Surf.Sci., 1980, 91, 514. D. J. C. Yates, L. L. Murrell, and E. B. Prestridge, J. Catal., 1979, 57, 4. l@T.M. Duncan, J. T. Yates, jun., and R. W. Vaughan, J. Chem. Phys., 1980,73,975. 163 Relationship bet ween Metal Carbonyl Clusters and Supported Metal Catalysts cm-1).20 The latter are comparatively rare, but there are two other variations reported. In [HFe4(C0)13]- (2), even in the absence of them bond, the carbon atom would be bridging and again the oxygen atom is bonded to one metaLg The carbon-oxygen stretching frequency of this group seems in doubt. The lowest frequency band is reported at 1723 cm-l in a mull but 100 cm-l higher in solution.21 The niobium trimer (3) contains am-bonded carbonyl group that itself bridges two metal atoms.22 This CO bond is relatively long (1.30 A) and the carbonyl stretching frequency is very low (1330 cm-1).These bonding modes may be more difficult to detect in surface studies for two reasons. First, since the carbon-oxygen axis is not perpendicular to the metal surface, the component of the dipole-moment change normal to the metal, and therefore allowed by the surface selection rule, will be reduced. Secondly, the n-bonded groups have fairly low stretching frequencies that may be obscured by some supports. Interestingly though, photoelectron spectroscopy data on CO on a carbonized osmium surface has been interpreted in terms of a weak C-0 bond, nearly parallel with the surface,23 and also EELS spectra of CO in Ni(ll0) at low coverages exhibit a band at 1565 ~rn-l.~~ This may be due to am-bonded carbonyl and this particular surface contains potential sites similar to those shown in Figure 1.The observation of av-bonded cyanide ligand in (NEt4) [(CSH~)ZMOZ(CO)~(CN)],acting in a similar manner to the m-carbonyl in (l), suggests this bonding may be available for a variety of n acid ligand~.~5 Metal carbonyl clusters can stabilize a wide variety of organic moieties26 and the binding modes established for two-carbon ligands are given in Figure 3. They can be used to provide reference spectra for chemisorbed hydrocarbons, but as yet detailed vibrational analyses of these species are relatively rare. EELS studies of acetylene and/or ethylene on Ni(l1 l),2712* Pt(l1 1),29 and Rh(l1 l)30 have been reported. Spectra of acetylene were observed on all three surfaces and are pre- sented in Table 3.There are substantial differences in these spectra. This is particularly evident in the C-H stretching region. All show a marked lowering in these frequencies as compared with v1 of free acetylene (3374 ~m-l),~l indicating a change towards an alkene or alkane type of carbon atom. The four bonding modes shown in Figure 3b are all available to acetylene on these surfaces and it may be that the three surfaces do not use a common chemisorption site. There is reasonable agreement between the data on platinum with the analysed spectrum *OA. A. Pasynskii, Yu. V. Skripkin, J. L. Eremenko, V. T. Kalinnikov, C. G. Aleksandrov, V.G. Andrianov, and Yu. T. Struchkov, J. Organomet. Chem., 1979, 165, 49. *l K. Farmery, M. Kilher, R. Greatrex, and N. N. Greenwood, J. Chem. SOC.A, 1969, 2339. W. A. Herrmann, M. L. Ziegler, K. Weidenhammer and H. Biersack, Angew. Chem., Znt. Ed. Engl., 1979, 18, 960. aa Y. Fukuda and J. W. Rabalais, Chem. Phys. Lett., 1980, 76, 47. 24 J. C. Bertolini and B. Tardy, Surf. Sci., 1981, 102, 131. M. D. Curtis, H. R. Han, and W. M. Butler, Znorg. Chem., 1980, 19, 2096. ao A. J. Deeming, ref. 3, p. 391. J. E. Demuth and H. Ibach, Surf. Sci., 1979, 85, 365. J. E. Demuth and H. Ibach, Surf. Sci., 1978, 78, L238. 2s H. Ibach and S. Lehwald, J. Vac. Sci. Technol., 1978, 15,407. L. H. Dubois, D. G. Castner, and G. A. Somorjai, J. Chem. Phys., 1980,72, 5234.31 Y. Iwashita, F. Tamura, and A. Nakamura, Znorg. Chem., 1969, 8, 79. Evans H M- M CH2 II M CH3 I /?M-M CH3 I CH3I Figure 3 Co-ordination modes of C, hydrocarbon ligands in metal complexes: (a) ethylene,(b) acetylene, md (c)other fragments Relationship bet ween Metal Carbonyl Clusters and Supported Metal Catalysts Table 3 Vibrational data of chemisorbed acetylene (frequencies in cm-l) On Ni(1 1 1)27 On Rh(11 1)30 On Pt(11 1)29 C~H~CO~(CO)~~~Assignment 3085 3010 2910 1400 2984 3086,l,,> VCH 1220 1310 1403 vcc 1065 870 887 706 985 770 8941 768 1 8CH 510 510 323 340 60515511 VCM of C2H2CO2(CO)6 (4),31 suggesting that form of edge bridging site is used on the 5d metal.A face bridging site is provided by Os3(CO)lo(CzHz) (5)32 and this complex has markedly lower YCH frequencies (2989 and 2939 and are in (4) (5) better agreement with the data on the nickel surface. EELS spectra of associatively adsorbed ethylene on Pt(111)29-34and Ni(111)34 have been observed at low temperatures. The only model complexes for which vibrational data are available are thewbound systems shown in Figure 3a. Even so there can be significant variations in the spectra. For example, the three i.r. allowed VCH bands for ethylene under C2, symmetry are at 3079, 3013, and 2988 cm-l in Zeises salt, K[P~C~~(CZH~)].H~O,~~but when complexed to an iron(0) centre, which will probably provide more back donation into the ethylene T* orbital, these frequen- cies are lowered to 3020,2970, and 2930 cm-1 [in Fe(C0)4(C~H4)1.~~ Allowing for these variations and the requirement that the observed bands should be polarized perpendicular to the metal surface, then the majority of the bands observed on these two metals fit chemisorbed molecular ethylene.Interestingly though a low frequency VCH was observed at 2690 cm-l on the nickel surface, suggesting an additional binding mode. There are no exact model complexes for the ‘di-a’ sa A. J. Deeming, S. Hasso, and M. Underhill, J. Chem. SOC.,Dalton Trans., 1975, 1614. s8 J. Evans and G. S. McNulty, unpublished results. J. E. Demuth, H. Ibach, and S. Lehwald, Phys. Rev. Lett., 1978, 40,1044. c6 J. Hiraishi, Spectrochim. Acta, 1969, 25, 749.D, C. Andrews and G. Davison, J. Organomet. Chem., 1972, 35, 161. Evans bonded ethylene (6) proposed for the major site on silica-supported nickel and platinum and exhibit VCH bands in the region of 2780-2880 ~m-1,3~but a trans rotamer of p(CH2CH2)[Re(CO)& has been recently reported.38 CH3 I / \-M-M The vibrational data obtained for ethylene on Rh(l11) from -270 to 420 K were not in accord with simple adsorption (Table 4).30 A new, partially dehydro- genated, species is formed that is also formed from acetylene and hydrogen and Table 4 Comparison of the EELS data for the stable species observed for C2H4 on Pt and Rh(ll1) with i.r. and Raman spectra of model ethylidyne complexes (frequencies in cm-1) ~t(111)29 Rh(11 l)30 3050w 3000w 2933 291 6 2920w 2900m 2892 2892 1420sh sh 1425 1438 1350s 1350s 1359 1367 1 130s 1 130s 1171 1147 9OOW 880m 1008 1037 435w 450m 536,393 appears to be analogous to the stable species observed for CZH4 on Pt(l1 l).29 The known geometries of ethynyl (CgH), vinylidene (CCH2), vinyl (CHCH2), ethylidyne (CCH3), and ethylidene (CHCH3), moieties in metal complexes are shown in Figure 3c.Initially the EELS data were rationalized in terms of an ethylidene species, using CH3CHC12 as the reference spectrum.29 However, a subsequent interpretation of low energy electron diffraction (LEED) intensities favoured face bridging ethylidyne species with the C-C axis within 15" of normal to the metal surface.39 This led to a reappraisal of the EELS data in terms of the new model, now using cH3CB1-3~0 and CH3CCo3(C0)9 (7)41 as references.A ST B. A. Morrow and N. Sheppard, Proc. R. Soc. London, Ser. A, 1969,311, 391. 38 B. Olgemoller and W. Beck, Chem. Ber., 1981, 114, 867. L. Kesmodel, L. H. Dubois, and G. A. Somorjai, J. Chem. Phys., 1979, 70, 2180. 40T.R. Stengle and R. C. Cooper, J. Mol. Spectrosc., 1970, 34, 33. W. T. Dent, L. A. Duncanson, R. G. Guy, H. W. B. Reed, and B. L. Shaw, Proc. Chem. SOC.London, 1961, 169. Relationship bet ween Metal Carbonyl Clusters and Supported Metal Catalysts major problem in this comparison was that the strong band at 1130 cm-l cor- related with a CH3 rocking vibration, which should be polarized parallel to the metal surface and therefore be very weak; this left the relatively weak band at 900 cm-l as the v(C-C).On this basis, Demuth proposed a vinyl species using Sn2(CH=CH2)6 as his reference.42 The geometry of this vinyl is not one found in cluster chemistry and is certainly quite unlike that of the tin complex. However, a normal co-ordinate analysis on the CH3CCo3 unit of (7) has shown that the original assignment of these two vibrations should be inverted and this gives a good fit with the surface species.43 It must be stressed that the model complexes should be as close as possible to surface species. In this case, the frequencies of the two vibrations in question, VC-c and PCH, in the CH3CX3 species, are highly dependent upon the nature of the halide40 and differ again from those observed on the cluster analogues.However, within a set of three ethylidyne complexes, (7), H3Ru3(CO)sCMe, and H30s3(CO)sCMe, these variations are much less and so frequencies can be transferred to metal surfaces with more confidence. As shown in Table 4 the strong surface bonds can all correlate with the expected positions of the surface allowed C-CH3 vibrations. The three weak features at N 3050, N 1420, and N 900 cm-l, however, remain problematical. They may be due to an uncharacterized second-surface species. While molecular acetylene and ethylene were observable on Ni( 1 1 1) at 300 and 140 K, respectively, more vigorous conditions give rise to new EELS spectra. At room temperature, ethylene is apparently dissociated to acetylene.44 High surface coverages of acetylene at 300 K have been variously described as N causing trimerization to give benzene45 and cleavage of the carbon-carbon bond;28 very different spectra were reported indicating that the surface reactions are very sensitive to the conditions employed.At 450 K, the vibrational spectrum is quite simple (2980m, 1300vw, 790 s cm-l). The two stronger bands were assigned to the v(CH) and 8(CH) vibrations of a tilted bridging methine (CH) group. These modes have been observed at 3041 and 850 cm-l, respectively, in CO~(CO)SCH,~~ giving support to the C-C cleavage process. From the discussion above, it is evident that organometallic complexes, including clusters, are a rich source of appropriate models for surface structures.One additional feature, not developed here but which has been reviewed re- ~ently,~~is that these complexes exhibit many forms of non-rigidity. These include migration of CO and alkyne ligands over clusters, hydrides within the metal framework, and also rapid motions of the metal skeleton itself. It J. E. Demuth, Surf. Sci., 1980, 93, L82. P. Skinner, M. W. Howard, I. A. Oxton, S. F. A. Kettle, D. B. Powell, and N.Sheppard J. Chem. Soc., Faraday Trans. 2, submitted for publication. I4J. C. Bertolini and J. Rousseau, Surf. Sci., 1979, 83, 531. 45 J. C. Bertolini, J. Massardier, and G. Dalmai-Jmelik, J. Chem. Soc., Faraday Trans. I, 1978, 74, 1720. M. W. Howard, S. F. A. Kettle, I. A.Oxton, D. B. Powell, N. Sheppard, and P. Skinner, J. Chem. SOC.,Faraday Trans. 2, 1981, 77, 397.''J. Evans, Adv. Organomet. Chem., 1977,16,319; E. Band and E. L. Muetterties, Chem. Rev., 1978, 78, 639; R. E. Benfield and B. F. G. Johnson, ref. 3, p. 471. Evans 3 Clusters as Catalyst Sources A. New Heterogeneous Catalysts.-A conventional preparation of a supported metal catalyst involves the introduction of a metal salt or co-ordination complex and subsequent reduction, normally by hydrogen at high temperatures. A wide variation of particle size distributions can be achieved by varying metal loadings and reaction conditions. This arises from competing kinetic processes, including the rate of reduction of the metal salt and the rate of metal migration and aggregation.Metal carbonyl clusters offer an alternative method of catalyst preparation since they contain preformed metal particles at, or near, a zero oxidation state. So interaction with an oxide support and thermal dissociation of some or all of the CO ligands could in principle give rise to a material that has a single metal particle size, and, by virtue of its unsaturation, be catalytically active. This more specific synthetic route would have dual advantages. First, since observations will be on a single metal species they can give more definite results and, secondly, these systems could be more specific in their activity. The simplest preparative procedure is to interact a metal carbonyl with the oxide, dry the material, and then to heat it in vacuo.This follows through the sequency of initial physisorption, chemisorption, and further carbonyl dissocia- tion. Methods of the initial interaction vary. One commonly used procedure is evaporation of the solvent from a suspension of a high surface area oxide in a solution of the carbonyl cluster in an organic solvent. While some physisorption probably occurs, this method will in many cases cause most of the complex to be in the form of crystallites. For example, the v(C0) i.r. bands of OSs(C0)rz on silicagel agree well with those of the carbonyl as a nujol mull.48 However, the i.r. spectra of Os3(CO)12 in this region differ markedly between the solid and solution, primarily due to intermolecular coupling in the crystal lattice.So this result indicates that crystallites of Oss(C0)lzare formed on silica gel. Evaporation of the solution must be avoided if simple physisorption of separated complex molecules is to be achieved. Chemisorption could involve a variety of chemical reactions (Figure 4): (i) substitution of a carbonyl by a surface hydroxyl group, (ii) oxidative addition of the hydroxyl group, (iii) nucleophilic attack on the carbonyl carbon atom, and (iv) a Lewis acid interaction with a carbonyl oxygen atom (this is more likely for bridging groups). There is no compelling evidence for the first process at present. However, chemisorption of Oss(CO)12 on silica involves loss of two carbonyl ligands and oxidative addition of a surface silanol group to afford (8).48*49This species was characterized by comparing its v(C0) i.r.bands with those of a close analogue, HOss(CO)ro(OSiPhs), and also by Extended X-Ray Absorption Fine Structure (EXAFS) measurements. This technique, which can give considerable structural information from amorphous as well as crystalline materials, is an A. K. Smith, B. Besson, J. M. Basset, R. Psaro, A. Fusi, and R. Ugo, J. Organomet. Chem., 1980,192, C31. IoB. Besson, B. Morawek, A. K. Smith, J. M. Basset, R. Psaro, A. Fusi, and R. Ugo,J. Chem. SOC.,Chem. Commun., 1980, 569. Relationship bet ween Metal Carbonyl Clusters and Supported Metal Catalysts (i) (CO)m-lMn-CO + HO--f ___3 (CO),--1Mn-O~I H 0 (iii) (CO)m-iMn--CO + HO --# (C0),-1Mfl--C-" 9 I IH E Figure 4 Possible chemisorption processes for a carbonyl cluster on an oxide surface: (i)substitution, (ii) oxidative addition, (iii) nucleophilic attack on the carbon atom, and (iv) Lewis acid interaction with the oxygen atom exciting prospect in this field.Similar species have been identified for Os3(CO)12 and Ru3(CO)12 on several o~ides.**~~~ However, interaction of Fe3(C0)12 and Fe(CO)5 with magnesia and 7-alumina gives rise to [HFe3(CO)11]- ionically bound to the surfaces.51 This anion is probably produced via nucleophilic attack by a surface hydroxyl group, and subsequent elimination of C02, which probably forms an adsorbed carbonate. Finally, evidence for a Lewis acid binding has been obtained for Fe3(CO)12 on a dehydrated HY type of ~eolite.5~This is generally found by a lowering of the vco frequency.Here this interaction is weak and may involve hydroxyl protons in the supercage. Interaction of MeMn(C0)s with alumina produces a YCO band at 1510 cm-l, and the i.r. evidence indicates that CO insertion had occurred rapidly to yield (9).53 J. Evans and B. P. Gracey, J. Chem. SOC.,Chem. Commun., 1980, 852. K1 F. Hugues, A. K. Smith, Y. Ben Taarit, J. M. Basset, D. Commereuc, and Y. Chauvin, J. Chem. SOC.,Chem. Commun., 1980, 68. s8 D. Ballivet-Tkatchenko and G. Coudurier, Inorg. Chem., 1979, 18, 558. I8F. Correa, R. Nakamura, R. E. Stimson, R. L. Burwell, jun., and D. R. Schriver, J. Am. Chem. Soc., 1980,102, 5112. Evans CH3 Nuclearity is not always maintained however.In the presence of oxygen, Rh4(C0)12 produces Rh1(C0)2 sites on silica.54 The vco bands in the infrared region indicated coupling between these metal centres, and the structure of this species has been assigned as The pyrolysis stages after initial chemi- oc co \ OC\Rh sorption also appear to cause nuclearity changes. In general, these pyrolysis products exhibit rather broad vco bands in the infrared region and definite characterization is extremely difficult. However, it has been noted that the same species is formed on y-A1203 from three different ruthenium clusters [Ru3(CO)12, H4Ru4(C0)12, and RUsC(C0)17] after heating above lo00C,56 and a similar phenomenon was reported for osmium clusters at 200 0C.48 In both cases these were assigned as having MI1(C0)2 or 3 centres.So the ideal situation of obtaining a controlled catalyst with a particular well established metal nuclearity is hard to attain. Nevertheless, as was shown by some of the earliest work in this area, partially decarbonylated species that are catalytically active can be generated.57 Tracer studies were used to identify the stoicheiometry of a butene isomerization, hydroisomerization, and hydrogena- tion catalyst formed from Ru~(C0)12 and silica to be Ru3(C0),5. The activity of this material did differ from that of ruthenium metal. The supported metals produced by complete thermal decomposition of metal carbonyls can differ from those prepared conventionally. For example, thermal activation of iron carbonyls 64 A.Theolier, A. K. Smith, M. Leconte, J. M. Basset, G. M. Zanderighi, R. Psaro, and R. Ugo, J. Organomet. Chem., 1980, 191, 415. 66 R. Whyman, ref. 3, p. 545. E.* V. L. Kuznetsov, A. T. Bell, and Y. I. Yermakov, J. Cutul., 1980, 65, 374. 67 J. Robertson and G. Webb, Proc. R. SOC.London, Ser. A, 1974, 341, 383. Relationship bet ween Metal Carbonyl Clusters and Supported Metal Catalysts on alumina under CO and H2 give a small metal particle size of less than 20 A.58 So this is a valuable new method of preparing supported metal catalysts. In general, low-oxidation-state metal carbonyls have relatively little affinity for oxygen-donating ligands. So functionalization of the oxide support with a soft donor ligand might help to maintain cluster nuclearity over a wider range of conditions.A second advantage is a much more substantial set of model com- plexes for reference spectra. A versatile anchoring method is to introduce a hydrolysable silyl group into the ligand, which can be used to react with surface hydroxyl groups (Figure 5a).59960 Phosphines have been most commonly used and the results of some reactions are shown in Figure 5b. The success of a particular anchoring procedure is highly dependent upon the chemistry of the precursor. In some cases, notably Rh4(C0)d8 and Rhs(CO)16,6l the cluster breaks down to give mononuclear Rh* dicarbonyl sites. However, the frequencies of the vco bands differ from those observed on unfunctionalized silica gel, indicating these rhodium centres are bound to a phosphine.A more common problem is that of a multiplicity of substitution products, illustrated by R~3(C0)12.6~ In general these can be identified by comparing the i.r. spectra with those of close analogues, but this can be difficult in practice because the i.r. bands tend to be broader on oxide surfaces than are those observed for model complexes in hydrocarbon solutions, and so overlap more. Jn some situations, the rates of stepwise carbonyl substitutions are sufficiently different that a single substitution product can be obtained, as is the case for RUsC(C0)1763 and substituted iridium ~arbonyl.~~In these circumstances, the agreement of the i.r. spectra of the car- bony1 stretching vibrations between anchored and model clusters, e.g., RU~C(CO)~~(PP~~C~H~SIL)and Ru&(CO)lsPPhzEt, is very close, i.e., within-2 cm-1.This increases the confidence in structural assignments and also indi- cates virtually no transmission of electronic effects from the support through the ligand chain to the cluster. This is in marked contrast to observations on clusters directly bonded to oxides. For example, the frequencies of the carbonyl vibra- tions of HOs3(CO)lo(OSiPh3) and Os3(CO)12 on silica gel (8) differ by up to -10 cm-1. Electronic effects are transmitted through the cluster indicating that silica gel is a more electron-withdrawing ligand than is OSiPhs, as would be expected. Substitution of a labile acetonitrile ligand and addition to an un- saturated cluster, as in (v) and (vi), allowed isolation of anchored osmium clusters from reactions under ambient conditions.65-67 68 D.Commereuc, Y.Chauvin, F. Hugues, J. M. Basset, and D. Oliver, J. Chem. SOC.,Chem. Commun., 1980, 154. gB K. G. Allum, R. D. Hancock, I. V. Howell, S. McKenzie, R. C. Pitkethly, and P. J. Robin-son, J. Organomet. Chem., 1975, 87, 203. R. L. Burwell. Chem. Technol.. 1974.4. 370. 61 H. Knozinger; E. W. Thornton, and hi. Wolf, J. Chem. SOC.,Faraday Trans. 1, 1979, 75, 1888. 62 S. C. Brown and J. Evans, unpublished results. 6a S. C. Brown, J. Evans, and M. Webster, J. Chem. Soc., Dalton Trans., to be published. 64 T. Catrillo, H. Knozinger, J. Leito, and M. Wolf, Znorg. Chim. Acta., 1980, 44, L239. 6s S.C. Brown and J. Evans, J. Chem. Soc., Chem. Commun., 1978, 1063. 66 R. Pierantozzi, K. J. McQuade, B. C. Gates, M. Wolf, H. Knozinger, and W. Ruhmann, J. Am. Chem. Soc., 1979,101,5436. 67 B. C. Gates and J. Leito, Chem. Technol., 1980, 195 and 248. Evans ,-) (4 X -OH +X&(CH&-L l>Si(CH&&OI 1-OH (X3Si -L) e.g., SIL -L ~i~-Rh(C0)2(PPhz S1L)XN Ru~(CO)II(PP~~-SIL) +Rus(CO)io(PPhz -SIL)2 RUsC(CO)i6(PPha -SIL) Tr4(CO)lo(PPha)(PPhz SIL)N Oss(CO)ii(PPhz N SIL) HzOss(CO)io(PPhz -SIL) J(vii) HzOss(CO)s(PPheN SIL) (i) Rh6(co)1661;(ii) Ru~(CO)~~'~;(iii) RU&(coh,e3 ;(iv) Ir,(CO)l,(PPh,) ";(v) Os8(CO)11-(MeCN)66;(vi) H20s3(CO)l,; (vii) Figure 5 (a) Generalized functionalization of oxides, (b) some reactions of phosphinatedsilica gel The chemistry of clusters anchored to phosphine functionalized oxides has been restricted to investigations of their thermal stability and some catalytic studies.H2Os3(CO)g(PPhz -SIL) and HAUOS~(CO)IO(PP~~-SIL) have been reported tocatalyse but-I-ene isomerization.66 The former material and Os3(CO)il- Relationship bet ween Metal Carbonyl Clusters and Supported Metal Catalysts (PPh2- SIL) also catalyse ethylene hydrogenation, albeit slowly.68969 Indeed the anchored complexes have lower turnover rates than their dissolved analogues under similar conditions. This may partly be due simply to having a third phase present, but i.r. studies demonstrated that the appended catalysts were being deactivated in two ways.One parallels the solution chemistry in which the ethylene is metallated and provides the ethylidyne ligand in (11). The second is specific to the oxide surface and appears to yield osmium(1r) carbonyl sites bound to a phosphine ligand. More active olefin hydrogenation catalysts are provided by RusC(C0)17~~ on functionalized silicas, but in each and rhodium carb~nyls~~ case aggregation to the metal ensued; 12-15 A rhodium particles were identified in the latter study. It is interesting that the amine functionalization hindered aggregation more. So although simple pendant ligands may retard the drawbacks of direct inter- action between oxides and cluster complexes, viz, cluster breakdown and aggregation to metals, they are not prevented.Attempts to overcome this problem are currently under way using bridging and polydentate anchoring ligands. The first such species reported was (lZ), which incorporates a face bridging acetylide.66 More recently, ruthenium and osmium trinuclear complexes OXIDE S -OXIDE II (13) M = Ru,Os (14) S. C. Brown and J. Evans, J. Mol. Catal., 1981, 11, 143. INS. C. Brown and J. Evans, J. Organomet. Chem., 1980, 194, C53. E. W. Thornton, H. Knozinger, B. Tesche, J. J. Rafalko, and B. C. Gates, J. Catal., 1980, 62, 117. Evans have been anchored on pendant thiolates (13)50971 and tricobalt carbon complexes (14) have been irnmobili~ed;~~ these procedures were also generalized to other oxides. The osmium derivatives of (13) are particularly stable and synthetic chemistry has been performed on them.72 Oxidation of one carbonyl group by Me3NO in the presence of acetonitrile has afforded HOs3(CO)g(MeCN)(S -OXIDE) on silica and alumina.The acetonitrile ligand can be readily substituted and thus forms an incipient vacant site, giving the complex more potential for catalytic activity. Interestingly, HOs3(CO)9(MeCN)(SPrn) catalyses pentene isomerization under conditions in which HOs3(CO)1o(SPrn) is inactive. It is apparent that the ideal of a durable active heterogeneous catalyst with a single cluster species is not trivial to attain. Direct interaction and anchoring with monodentate ligands do not seem appropriate. However, use of bridging and polydentate ligands, such as HC(PPh&, which co-ordinates over a face of the rhodium tetrahedron in Rh4(C0)9[HC(PPh3)3],73 may well increase the resistance to fragmentation sufficiently to achieve this goal.B. Synthesis Gas Chemistry.-The impetus for this work is to provide processes for petrochemical synthesis from a synthesis gas, which could be coal-based.74 Presently, most synthesis gas is derived from methane and other petroleum products and is mainly used for conversion to hydrogen via the water gas shift reaction, for ammonia manufacture, and also methanol manufa~ture.~4*~5 The viability of converting a particular process to a coal base is dependent upon the economic and political climate in different countries. It has been suggested that one of the first processes to convert in the USA would be the manufacture of ethylene-derived oxygenates (ethylene glycol) and this could be attractive in 6-14 years.74 However, in 1978, outside Eastern Europe, there were 18 Lurgi gasifiers operating for the production of synthetic natural gas (SNG: 2632% C02, 17-21 % CO, 36-43 % H2, and 8-13 % CH4).76 Thirteen of these were in Sasolburg (South Africa) and a further 40 are planned there by 1981, increasing the capacity by ca.42 Mm3 per day. The majority of the synthesis gas produced was used to manufacture 1.5 Mton of gasoline per year,77 using two processes with iron catalysts. About 6000 tons of coal per day were used for the gasifiers with an additional 5400 tons of coal per day used in the direct production of electricity and steam for the synthesis project.The extended capacity will consume ca. 40 OOO tons of coal per day. The scale of such operations causes problems of site,74978 but even so a pilot plant for coal liquefaction has been operating in ?lT.Catrillo, H. Knozinger, and M. Wolf, Znorg. Chim. Acta., 1980, 45, L235. J. Evans and B. P. Gracey, unpublished results. 7sA.A. Arduani, A. A. Bahsoun, J. A. Osborn, and C. Voekler, Angew. Chem., Znt. Ed. Engl., 1980, 19, 1024. J. E. Johnson, Chem. Technol., 1981, 44. 75 C. M. Bartish and G. M. Drissel, Kirk-Othmer Encycl. Chem. Technol., 3rd Ed., 1978, 4, 772. 76 J. Huebler, and J. C. Janka, Kirk-Othmer Encycl. Chem. Technol., 3rd Ed., 1980, 11, 410. 77 C. D. Kalfadelis and H. Shaw, Kirk-Othmer Encycl.Chem. Technol., 3rd Ed., 1980, 11, 447. J. Elkington, The Guardian, 26th March, 1981. 175 Relationsh@ between Metal Carbonyl Clusters and Supported Metal Catalysts West Germany and another is planned for North Wales to go on stream in 1984.78 The chemical interest, however, is in the design of specific catalysts. Efficient catalysts, based on zinc-copper-alumina, already operate for methanol manu- facture.75 Indeed, dehydration of methanol using 'shape selective' zeolites, e.g. ZSM-5,79 is an alternative route to synthetic gasoline.80 Mobil's process gives a relatively narrow molecular-weight range with less than 3 % methane and ending by Cii.N On the basis of two surveys of metal carbonyl catalyst precursors, one below 200 atm and 200 "C81and another at 2000 atm and 230 "C,g2 it appears that Co, Ru, Rh, Ir, and Pt complexes form the most active catalysts for synthesis gas conversion.Activity appears to be strongly solvent dependent and we shall discuss results obtained on the first three elements, which seem to have most potential at this stage. Reports on three studies in which C02(CO)8 was used as a precursor have been p~blished;*l-~~different reaction conditions were used in each study. In all cases, liquid products predominated. At 200 "C, under an initial pressure of 130 atm (CO:H2 = 1 :1) in glyme or diglyme, then the major product from synthesis gas was ethanol (80%).8l Under a higher pressure (300 atm) and in benzene, heptane, and p-dioxan, the major products were methanol and methyl f~rmate.~~ Rates were solvent dependent (p-dioxan > benzene > heptane) and in the ether solvent ethanol and propanol were also detected and shown to be secondary products from methanol.At 230°C and ZOO0 atm, rates were faster in toluene than in N-methylpyrrolidone (NMP), with ethylene glycol (25 %) as an additional product to methanol (31 %) and methyl formate (35 %).g2 These conditions approach those presented 30 years ago for the synthesis of ethylene glycol, glycerol, and their esters and ethers.83 Co(0Ac)z was a typical catalyst precursor at pressures of 3000 atrn.S4 Samples obtained after the pressure was released N indicate that HCo(C0)4 was the principal (or only) cobalt-containing species in solution in the 300 atm and 2000 atm studies.The cobalt chemistry evident from the recovered solutions from the 200 atm reactions is complex.81 Diglyme solutions that have operated between 120 and 180"C are red-brown and exhibit i.r. bands that were assigned to Coe(CO)s, [(diglyme)nCo] [Co(CO)4]3, and an unusual polynuclear complex (15). However, after operation at 200 "C, these solutions are yellow and exhibit no v(C0) i.r. bands attributable to metal car- bonyls. 79 M. S. Spencer and T. V. Whittam in 'Catalysis', ed. C. Kemball and D. A. Dowden (Specia- list Periodical Reports), The Chemical Society, London, 1980, Vol. 3, p. 189. S. L. Meisel, J. P. McCullough, C. H. Lechthaler, and P. B. Weisz, Chem. Technol., 1976, 86.R. J. Daroda, J. R. Blackborow, and G. Wilkinson, J. Chem. SOC.,Chem. Commun., 1980, 1098. 8a W. Keim, M. Berger, and J. Schlupp, J. Catal., 1980, 61, 359. 83 J. W. Rathke and H. M. Feder, J. Am. Chem. SOC.,1980, 102, 3625. 84 W. G. Gresham, B. P. 655 237, 1951 (Chem. Abstr., 1952, 46, 7 115h); B.P. 665 698, 1952, U.S.P.2 623 906, 1952 (Chem. Abstr., 1952, 46, 11 232f). Evans Methanol and methyl formate have been, identified as major reaction products from Rus(C0)~initiated catalyses under a variety of conditions, although again, actual product distributions were sensitive to solvent and conditions.81782185 Little gas uptake was noted at 200"C, under 300 atm in glyme solvents, but catalysis was evident in 2-methoxyethanol under 100 atm of synthesis gas at 170 oC.81No metal carbonyls were observed in the recovered solutions; an alkox- ide gel was present.Apparent Fischer-Tropsch catalysis by RUQ(CO)IZ under 100 atm at 300 "C in heptane has been shown to be a heterogeneous reaction that N is only apparent after the carbonyl starts to decompose to ruthenium After -1 day, all the carbonyl is decomposed. Studies in THF solution indicated that the fate of the ruthenium is strongly pressure dependent.85 Hydrogenation of carbon monoxide proceeds at 1300 atm and 225-275 "C to give > 99 % methanol and methyl formate. A variety of catalyst precursors, e.g., Ru~(CO)~Z, [RU~C(CO)I~]~-,and Ru(acac)3, gave similar results and the i.r. spectra of samples extracted from working solutions show only the ruthenium to be in the form of Ru(C0)5.Under milder conditions (80°C and 135 atm of synthesis gas) the pentacarbonyl and Ru3(C0)12 are present in approximately equal concentrations, but increasing the temperature without markedly raising the gas pressure pro- motes decomposition to the metal. It is interesting that under conditions of mixed homogeneous and heterogeneous catalysis, methanol formation decreased with decreasing Ru(C0)5 concentration. Ru(C0)5 has also been detected after, and during, synthesis gas conversion, in under -340 atm of CO and Hz and N 200 "C in acetic acid solution.87 Again a variety of precursors gave equivalent results. Methyl acetate and ethylene glycol diacetate were the major products. The rate of production of the latter is strongly related to the concentration of carboxylic acids.Partial pressure dependences of H2 and CO favoured involve- ment of HzRu(C0)4. There is obvious difficulty in ascribing these catalytic activities to particular species. In none of these cases do clusters appear to be the active species. In situ i.r. monitoring of the ruthenium system gives good evidence for the involvement of mononuclear carbonyl complexes. This method has also been employed on rhodium-catalysed systems, where there is evidence for cluster catalysis. A variety of rhodium carbonyls including Rh(CO)z(acac) can be used as precursors for the J. S. Bradley, J. Am. Chem. SOC.,1979, 101, 7419. sE M. J. Doyle, A. J. Kouwenhoven, C. A. Schapp, and B.van Oort, J. Organomet. Chem., 1979, 174, C55. a7 B. D. Dombek, J. Am. Chem. SOC.,1980, 102,6855. 177 RelationshQ between Metal Carbonyl Clusters and Supparted Metal Catalysts catalytic synthesis of ethylene glycol and other polyhydric alcohols.88 In situ i.r. studies have been reported on these systems at pressures of up to lo00 atm, somewhat lower than some of the optimum catalytic c0nditions.8~ In the presence of amines and/or alkali-metal carboxylates, solutions of Rh(CO)z(acac) in tetraglyme or sulfolane, when treated with lo00 atm of CO/Hz (1 :1) at tem- peratures below 180 "C, give i.r. bands indicative of [Rh5(C0)15]- (16)90 and 1-[Rh(CO)4]-; other clusters, e.g., [Rh12(C0)30]2-, behaved similarly. At higher temperatures (-250 "C) more approaching optimum catalytic conditions for polyol formation, higher nuclearity cluster complexes (Rh13-15) are evident.Three clusters containing encapsulated atoms, [Rh~(co)l~c]~-, [Rh9P(C0)21l2-, and [Rh17(Sz)z(C0)3~]~-, show increasing stability and are active catalysts under conditions in which they are the only detectable species. Although they are less active than Rh(CO)z(acac)-based systems (by a factor of 2-10), they represent the most probable examples of catalysis by clusters. Studies of this type represent one of the most effective ways of investigating these systems, when linked to conventional mechanistic methods. A very recent report has demonstrated the applicability of 13C n.m.r. to studying high-pressure reactions, potentially a very powerful te~hnique.~~ Intermolecular exchange even under loo0 bar of CO/H2 (2.1 :1) was sufficiently slow to allow clear observation of the spectra due to the dissolved complexes.These clearly showed the conversion of [Rh1z(C0)30l2- to [Rh5(C0)15]- at room temperature. At low temperatures (N -50 "C)this trans- formation is slow. R. L. Pruett and W. E. Walker, U.S.P. 3 833 634, 1974 (Chem. Abstr., 1973, 79, 78 081). 89 J. L. Vidal and W. E. Walker, Inorg. Chem., 1980, 19, 896. A. Fumagilli, T. F. Koetzle, F. Takusagawa, P. Chini, S. Martinengo, and B. T. Heaton, J. Am. Chem. SOC.,1980, 102, 1740. B. T. Heaton, J. Jonas, T. Eguchi, and G. A. Hoffman, J. Chem. Soc., Chem. Commun., 1981, 331. Evans Synthesis gas conversion has also been effected by materials formed by the deposition of polynuclear rhodium carbonyls on a variety of oxides, followed by pyrolysis in vacuo or under hydrogen.g2 Although these catalysts are not well characterized, they show interesting facets.First, these reactions were carried out under relatively mild conditions, 205-225 "Cand 65 Torr of CO/H2 (-1:2). Secondly, the product distributions were support dependent. While methane was predominant on acidic supports, e.g., Si02, ethanol was the major product on more neutral oxides, e.g., La203 and TiOz, and methanol predominated on basic supports, e.g., MgO. Thirdly, on La203, the percentage of ethanol formed varied within the range of 29-61 % for different complexes under similar conditions {(CsHs)zRhz(C0)3< Rh4(CO)i2 > Rhs(C0)is > (NEt4)3[Rh7(CO)isl > [NBu~]~ [Rh13(C0)23H3J }.This ethanol selectivity was substantially greater than a conventional heterogeneous rhodium catalyst (17 % ethanol). Pyrolysis of Fe3(CO)12 on oxides such as MgO and A1203 also produces Fischer-Tropsch catalysts with unusual sele~tivities.5~~~3 At 176 "C, after 1 % CO conversion, propene was the major product (32%), with methane (26 %) also being formed in reasonable quantities. The catalyst has very small metal particles initially (< 20 8)but these increase to 200-500 8 by the end of the run. The selectivity for propene drops concomitantly, indicating a unique process for the smaller particles. Exposure of a freshly prepared catalyst to ethylene also yields propene as the major product.This suggests that methylene groups are formed on the iron surface. Hydrogenation can then afford methane, and further reaction with ethylene to form a metallocyclobutane is a possible route towards propene. Finally, highly active methanation catalysts have been prepared by supporting Mo(CO)~ and W(CO)6 on alumina.94 Initially, these materials appear to contain adsorbed 'su b-carbonyl' mononuclear species [M(CO)s or M(C0)3] and possibly the activity may be due to these centres. 4 Conclusions Surface science and cluster chemistry are both rapidly developing fields, and so conclusions about their relationships may change in the next few years. However, it seems that clusters have a valid contribution to make as models for metal surfaces.This approach has its limitations ;a structure or mechanism determined on a homogeneous system is in no way binding to a heterogeneous one. In the author's view the most incisive contribution cluster chemistry can make is to form a bank of spectroscopic data on well defined systems. Work on clusters as catalyst sources is in its early stages, particularly in heterogeneous systems. Nevertheless some results, even though they are not yet well understood, seem promising. Recent spectroscopic developments offer exciting extensions to the situation five or so years ago. These include Fourier transform infrared (with greatly increased sensitivity, particularly in the far i.r.), high-resolution solid- 9g M.Ichikawa, J. Chem. SOC.,Chem. Commun., 1978, 566; Bull. Chem. SOC.Jpn., 1978, 51, 2268 and 2273. 93 F. Hugues, B. Besson, and J. M. Basset, J. Chem. SOC.,Chem. Commun., 1980, 719. A. Brenner and D. A. HUCUI,J. Am. Chem. SOC.,1980,102,2482. 179 Relationship bet ween Metal Carbonyl Clusters and Supported Metal Catalysts state n.m.r. (using magic angle sample spinning and multiple-pulse sequences), and EXAFS. These methods offer the promise of greatly improving the degree of characterization of heterogeneous systems, and used in combination can increase this near to the level currently expected for solution studies. I wish to thank the Society of Maccabaeans and the RSC for the award of the Meldola Medal. This reflected the stimulating environments in which I have worked at Cambridge, Princeton, and Southampton and I am grateful to all my colleagues in these departments.

ISSN:0306-0012    

DOI:10.1039/CS9811000159

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

Photochemistry and Photocyclization of Aryl Halides By J. Grimshaw DEPARTMENT OF CHEMISTRY, QUEEN’S UNIVERSITY, BELFAST BT9 SAG and A. P. de Silva DEPARTMENT OF CHEMISTRY, UNIVERSITY OF COLOMBO, COLOMBO, SRI LANKA 1 Introduction The vast range of photochemical reactions that haloarenes undergo with other species can broadly be divided into a few classes. Photochemically induced homolysis reactions of the carbon-halogen bond have been known since around 1960, and bond reactivity follows the order expected from carbon-halogen bond energy considerations. Iodobenzenel was investigated first, and then later bromobenzene2 was shown to undergo equivalent reactions, and chlorobenzene to react with a low quantum yield. Thus, whereas iodo- and bromo-arenes have received attention as potential sources of aryl radicals, chloro-compounds have received little attention.Here we consider the broad range of photohomolysis reactions, some of which are of value in synthesis, while others are important as possible routes by which polutant chloro- and bromo-arenes can be degraded3 under environmental conditions. Inevitably, to give a coherent account of photochemistry and photocyclization, the discussion must include reactions which cover a broader range of mechanisms than simple homolysis. Fluoroarenes undergo a variety of photoinduced electrocyclic reactions where the carbon-fluorene bond remains inta~t.~ Such reactions are not relevant to this review. Nucleophilic photosubstitution reactions occur on haloarenes ;some activating substituent such as nitro is usually necessary. A variety of nucleophiles including alkoxy, hydroxy, amino, and thiocyanate can be used and fluoride or chloride are the most satisfactory leaving groups, although other leaving groups as well as halides are possible.Such reactions have been known since around 1956. The bromo- and iodo-arenes show competition between nucleophilic substitution and J. McD. Blair, D. Bryce-Smith, and B. W. Pengilly, J. Chem. SOC.,1959, 3174; J. McD. Blair and D. Bryce-Smith, J. Chem. SOC.,1965, 1788; W. Wolf and N. Kharasch, J. Org. Chem., 1961, 26, 283; ibid., 1960, 30, 2493. * T. Matsuura and K. Omura, Bull. Chem. SOC.Jpn., 1966, 39, 944; G. R. Lappin and J. S. Zanucci, Tetrahedron Lett., 1969, 5085.N. J. Bunce, Y. Kumar, and B. G. Brownlee, Chemosphere, 1978, 7, 155. * R. D. Chambers, J. R. Maslakiewicz, and K. C. Srivastava, J. Chem. SOC.,Perkin Trans. 1, 1975, 1130; M. G. Barlow, D. E. Brown, and R. N. Haszeldine, ibid., 1978, 363; B. Sket and M. Zupan, J. Am. Chem. SOC.,1977, 99, 3504; J. Libman, Z. Ludner, B. Louris, and V. Yakhot, J. Chem. Res. (S), 1978, 472; (M), 1978, 5557. Photochemistry and Photocyclization of A ryl Halides bond homolysis. Nucleophilic photosubstitution reactions have received adequate discussion elsewhere5 and have largely been excluded from this review. A class of reactions that proceed according to the example in Scheme 1 was discovered in 1970. The reaction is initiated by photochemically induced electron- transfer to give the substrate radical-anion in catalytic amounts. The reaction can also be promoted electrochemically.These radical-anion substitution reactions have been reviewed by Bunnett.6 PhBr + electron source ,&> [PhBr]*-+ residue [PhBr]*-jPh* + Br-Ph* +.-> MeCOeH2 [PhCH2COMe]--+,-> PhBr [PhCH&OMe]*-PhCHzCOMe + [PhBr]*-Overall: h>PhBr + MeCOCH2 ,iq.NHS> PhCHzCOMe + Br-Scheme 1 Bryce-Smith and his colleagues showed that the photolysis of iodobenzene gives rise to phenyl radicals as reactive intermediates by comparing the relative yields of products from reaction in isopropylbenzene and in neat iodobenzene with the relative yields from decomposition of dibenzoyl peroxide, a known phenyl radical source, in these solvents.1 Kharasch and his school showed that a reaction of preparative value results when conditions are arranged under which aryl radicals, generated by photohomolysis from an iodoarene, can interact with a second arene ring.The photolysis of iodobenzene in benzene solution to yield diphenyl is one example of a general intermolecular reacti0n.l The equivalent intramolecular process has been used to effect a large number of ring-closure reactions, in yields which vary from small to acceptable, and constitutes an important step in the synthesis of many alkaloids. A selection of syntheses is given in Scheme 2 to serve as an introduction to the mechanistic discussion in the following section. Aryl radicals intermediate in some of these reactions can be formed by other routes, for example by reductive decomposition of diazonium salts in the Pschorr reaction, and the relative merits of the various related processes have been discussed recently.' r, J.Cornelisse, G. P. de Gunst, and E. Havinga, Adv. Phys. Org. Chem., 1975, 11, 225; J. Cornelisse and E. Havinga, Chem. Rev., 1975, 75, 353; E. Havinga and J. Cornelisse, Pure Appl. Chem., 1976, 47, 1; J. Cornelisse, G. Lodder, and E. Havinga, Rev. Chem. Intermed., 1979, 2, 231. J. F. Bunnett, Acc. Chem. Res., 1978, 11, 413, M. Sainsbury, Tetrahedron, 1980, 36, 3327. Grimshaw and de Silva ... 111 (4, x = C1) 44% Reagents: i, Aq. MeOH, HCl, NaHSO, if R = Me or H, hv;ii, C6H6,Na,S20, if R = Ac, hv iii, H,O, NaOH, NaBH,, hv; iv, hv, CsHs, ButOH, KOBu' Scheme 2 It has been suggested that some photocyclization reactions of haloarenes which possess a stilbene system [e.g.(4)] proceed by an electrocyclic intermediate, followed by elimination to reform an aromatic system.8 In the other reactions shown in Scheme 2, photohomolysis of the aryl-halogen bond was considered to be the first step, and so either iodo- or bromo-substrates were used. The benzene ring upon which substitution occurs can be either without phenolic substituents,g * M. P. Cava, M. J. Mitchell, S. C. Havlicek, A. Lindert, and R. J. Spangler, J. Org. Chem., 1970, 35, 175. See also M. P. Cava, P. Stern, and K. Wakisaka, Tetrahedron, 1973, 29, 2245 ;M. P.Cava and S. S. Libsch, J. Org. Chem., 1974,39,577 ;L. Cleaver, S. Nimgirawath, E. Ritchie, and W. C. Taylor, Aust. J. Chem., 1976, 29, 2003. S. M. Kupchan, J. L. Moniot, R. M. Kanojia, and J. B. O'Brien, J. Org. Chem., 1971, 36,2413; J. L. Neumeyer, K. H. Oh, K. K. Weinhardt, and B. R. Neustadt, ibid., 1969,34, 3786. Photochemistry and Photocyclization of Aryl Halides as in (l), or can carry a phenoxide substituent,lO as in (2) when this substituent exerts a pronounced ortho, para directing effect, allowing the formation of spiro-compounds. Spirodieneones formed in this way are themselves subject to photodegradation, so that better yields are obtained by in situ reduction of the first formed carbonyl compound to an alcohol, (3), with sodium borohydride.Cyclization of substrates with a tertiary amino-function (1, R = alkyl) is best carried out in acidic solution where the amino-function is protonated, while secondary amines are best acylated to give (1, R = acyl) and photolysed in benzene. Where aryl radicals are intermediate in a ring-closure reaction, they can be diverted by hydrogen abstraction from the solvent, or some other source, to give overall replacement of halogen by hydrogen. Often the yield of cyclized product is low. The early work in this field of haloarene photochemistry has been reviewed.ll-l3 2 Reaction Mechanisms There have been a number of mechanistic studies on the photochemical reactions which lead to replacement of halogen by hydrogen and the results are relevant to the more useful preparative aspects of aryl halide photochemistry.Only a few such studies have been made of the preparatively useful ring-closure reaction. A. Energy Transfer and Homolytic Bond Cleavage.-The U.V. spectrum of iodobenzene shows evidence for a na* excited state, involving the iodine group, which is dissociative in nature. This transition is of low intensity and partly overlapped by an intense vv* transition. Simple homolytic fission in iodo- arenes may arise from population of nu* or ua* states. This mechanism is not available in the ordinary solution photochemistry of bromo- and chloro-com- pounds, which do not possess a low energy u* orbital, however, homolysis of cholorobenzene via a triplet aa* state is suggested for the short wavelength vapour phase phot01ysis.l~ Radiationless decay of the vv* state to a vibration- ally excited ground-state can also lead to homolysis of the weakest bond in the system before the vibrational energy is dispersed as heat.15 For chloro-, bromo-, and iodo-arenes the weakest bond is usually, but not necessarily, the carbon- halogen bond.These two processes for energy transfer to the a-bond lead ultimately to radical species which may recombine before one radical is able to break from the solvent cage (see Scheme 3). Further reactions of the solvent- separated aryl radical lead to the final products. In the chemistry of thermally-generated free radicals, cage recombination can be demonstrated because the rate of the forward reaction is diminished in lo Z.Horii, Y. Nakashita, and C. Iwata, Tetrahedron Lett., 1971, 1167. l1 R. K. Sharma and N. Kharasch, Angew. Chem., Int. Ed. Engl., 1968, 7, 36. l8 P. G. Sammes in ‘The Chemistry of the Carbon-Halogen Bond’, ed. S. Patai, Wiley, New York, 1972, p. 747. lS T. Kametani and K. Fukumoto, Acc. Chem. Res., 1972,5,212. l4 G.Porter, ‘Reactivity of the Photoexcited Molecule’, Interscience, London,1967, p. 104. l5 G. Porter and B. Ward, Proc. R. SOC. London, Ser. A, 1965, 287,457. Grimshaw and de Silva A solvents of high viscosity.16 In the context of a photodissociative process, cage recombination shows itself by the influence of solvent viscosity on the quantum yield for decomposition of the starting material. The quantum yield for the photoreduction of bromobenzene has been measured in mixtures of ethanol and ethane-l,2-diol and shows a decrease with increasing solvent viscosity.17 For the photocyclization of (6) and the related bromo- and iodo-compounds, the quantum yield showed a smooth variation with solvent viscosity in a range of solvents including cyclohexane, methanol, and ethanel,2-dio1.18 For pre- parative work with reactions which follow this mechanism, less viscous solvents will allow shorter reaction times because of this greater quantum yield for bond homolysis.Simple energy considerations indicate which aryl halides can be expected to undergo the simple homolysis reaction, since the energy of the excited state must be greater than the bond energy.Substitution by the halogen changes the excited-state energies of the parent aromatic system very little, so we can compare the energy data given in Table 1 in order to justify the quantum yield data given in Table 2 for some reactions that have been studied in detail. Table 1 Triplet energies compared to carbon-halogen bond energies in arene compounds Arene ET/W rno1-l Haloarene Dc-xlkJ mo1-l Benzene 352 PhCl 397 Biphenyl 275 PhBr 334 Naphthalene 255 PhI 268 Data from: S. L. Murov, Handbook of Photochemistry, M. Dekker, New York, 1973; K. W. Egger and A. T. Cocks in 'Chemistry of the Carbon-Halogen Bond', Part 2 ed. S. Patai. Part 2, Wiley, 1973 Flash excitation of fluoro- and chloro-benzene in solution shows emission spectra from triplet states.Bromo- and iodo-benzene show rapid flash photo- lysis and a triplet intermediate cannot be detected.19 Chloro- and bromo- substituted naphthalenes and biphenyls show slow photolysis via the triplet state even though from bond energy considerations cleavage of the carbon- lS D. Gegiou, K. A. Muszkat, and E. Fischer, J. Am. Chem. Soc., 1968,90, 12; D. Booth and R. M. Noyes, ibid., 1960, 82, 1868; W. A. Pryor and K. Smith, ibid., 1970, 92, 5403. l7 J. Szychlinski and L. Litwin, Rocznik Chem., 1963, 37, 671. J. Grimshaw and A. P. de Silva, Can. J. Chem., 1980, 58, 1880. I. Loeff, H. Lutz, and L. Lindquist, Isr. J. Chem., 1970, 8, 141. Table 2 Quantum yields (a)for the decomposition of haloarenes Haloarene Solvent @ Haloarene Solvent @ 5.;3:1-Chloronaphthalenea MeOH 0.005 4-Chlorobiphenylb iso-octane 0-0006 %1-Bromonaphthalenea MeOH 0.170 2-Br omobiphenyl c cylohexane 0.046 b 2-Chloro biphenylb iso-octane 0.39 2-Bromobiphenylc cyclohexane 0.004 5 3-Chlorobiphenylb iso-octane 0401 4-BromobiphenylI: c ycl ohexane 0.015 %Q aL. 0.RUZO,N. J. Bunce, and S. Safe, Can. J. Chem., 1975,53,688; N. J. Bunce, P. Pilon, L. 0. Ruzo, and 0.J. Sturch, J. Urg. Chern., 1976, 41, 3023. bN. J. Bunce, Y. Kumar, L. Ravanal, and S. Safe, J. Chem. SOC.,Perkin Trans. 2, 1978, 880; CN. J. Bunce, S. Safe, and L. 0.RUZO,J. ,$Chem. SOC.,Perkin Trans. 1, 1975, 1607. Grimshaw and de Silva halogen bond is unlikely since the reaction would be so endothermic.A mechanism of photoinduced intermolecular electron-transfer (see section B) has been proposed to account for the observed reactivity. The triplet energy of biphenyl is raised by 2-substitution as a result of the benzene rings being forced out of plane, and this accounts for the higher reactivity of 2-halobiphenyls. Cyclization of the chlorophenylpyrazole (6) is also not expected from simple energy considerations, since for 1, 3, 5-triphenylpyrazole, values of ES = 351 and ET = 272 kJ mol-l have been found. The reaction can be effected either directly through the singlet state or by triplet sensitization, and either way the quantum yield is of the same order for chloro-, bromo-, or iodo-analogues. Solvent effects indicate the direct singlet reaction to be a radical process and, surpris- ingly, the reaction leads in all cases to quantitative cyclization with no replace- ment of halogen by hydrogen, as a side reaction.A transition state (7) leading to an intermediate involving n-complexation by the adjacent benzene ring of radical species which develop during bond homolysis has been proposed to account for these observations.18 Complexation can be expected to lower the energy require- ment for bond homolysis. A related effect attributed to complexation of radicals by aw-cloud has been noted in general free radical chemistry.20 B. Electron Transfer and Radical-anion Decomposition.-Photochemically induced electron transfer between dialkylanilines and aryl halides is well documented.21 Irradiation by light (h = 313 nm) of a mixture of dimethylaniline and a halobenzene leads to exciplex formation between the excited state of the amine and the halobenzene followed by electron transfer from the amine.The fluorescence of dimethylaniline is quenched by this process and decomposition of the halobenzene radical-anion then yields a phenyl radical and halide ion according to Scheme 4. Aliphatic amines also promote bond cleavage in aryl halides by a similar mechanism where an electron is transferred from the nitrogen lone pair to the excited state of the aryl halide.22 Such photoinduced electron-transfer could ao T. W. Koenig and J. C. Martin, J. Org. Chem., 1964,29,1520; D. B. Denney, R. L. EIlsworth, and D. Z. Denney, J.Am. Chem. SOC.,1964, 86, 1116. p1 T. Latowski, 2. Naturforsch., Teil A, 1968, 23, 1127; C. Pac, T. Tosa, and H. Sakuri, Bull. Chem. SOC.Jpn., 1972, 45, 1169; M. Grodowski and T. Latowski, Tetrahedron, 1974, 30, 767. pa J. Nasieliski and A. Kirsch-Demesmaeker, Tetrahedron, 1973, 29, 3153; M. Ohashi, K. Tsujimoto, and K. Seki, J. Chem. SOC.,Chem. Commun., 1973, 384. Photochemistry and Photocyclization of Aryl Halides c1 NMez Scheme 4 clearly give rise to unwanted side-reactions in the photochemical ring-closure step for the synthesis of a number of alkaloids and the earlier workers discovered empirically that best yields for cyclization are obtained when a free amino-group is protected either by protonation or by conversion to an N-acyl derivative, Added triethylamine also promotes the photochemical decomposition of chlorinated biphenyls,23 par~-terphenyls,2~ and bromobiphenyls25 by electron transfer to the excited state.Related reactions may be in part responsible for the photodegradation of chlorinated aromatic compounds in the environment. A related photochemically induced electron-transfer process has been used to explain carbon-halogen bond cleavage in the triplet state of chloronaphtha- lenes2s and chloro- and bromo-biphenyls25 where the triplet energy is less than that required for bond cleavage (Table 1). Electron transfer is proposed between the excited state and some other aromatic species in solution, either the halo- arene or photochemically produced arene, to give the haloarene radical-anion which undergoes rapid bond cleavage.The rate of carbon-halogen bond cleavage in some radical-anions is known to be slow, which appears to be the case for a variation of the electron-transfer mechanism shown by the photolysis of 9,lO-dichloroanthracene in the presence of 2,5-dimethyl hexa-2,4-diene.27 Irradiation in acetonitrile gives rise to a singlet exciplex which relaxes to a pair of radical-ions. That the radical-anion of 9,lO-dichloroanthracene is protonated by traces of water and then loses a chlorine atom to give 9-chloroanthracene is demonstrable since, in the presence of 0.06M deuterium oxide, deuterium is incorporated into the product. Deuter- ium is not incorporated into the product after photolysis in [2Hs]acetonitrile as N.J. Bunce, Y.Kumar, L. Ravanal, and S. Safe, J. Chem. SOC.,Perkin Trans. 2, 1978, 880. %* B. Chittim, S. Safe, N. J. Bunce, L. 0. RUZO, K. Olie, and 0. Hutzinger, Can. J. Chem., 1978,56, 1253. s6 N. J. Bunce, S. Safe, and L. 0. RUZO,J. Chem. Soc., Perkin Trans. 1, 1975, 1607. L. 0. Ruzo, N. J. Bunce, and S. Safe, Can. J. Chem., 1975,53,688. W. K. Smothers, K. S. Schanze, and J. Saltiel, J. Am. Chem. SOC.,1979, 101, 1895. Grimhaw and de Silva indicating that, in this case, the radical-anion does undergo significant carbon- chlorine bond cleavage to give a a-radical. C. Electron Reorganization to an Electrocyclic Intermediate and Subsequent HX Loss.-Compounds such as (4) bear a formal resemblance to stilbene and so can be expected to show cis, trans photoconversion and to form the intermediate (5) on irradiation, subsequently losing hydrogen chloride. This cyclization elimina- tion mechanism was proposed to account for the ready reaction of the chloro- arene (4, X = C1).8 Cyclization onto carbon-6’ must also be expected but, if we assume that formation of the electrocyclic intermediate is reversible, then rapid irreversible loss of HCI will direct the overall process to cyclization onto the halogen-bearing carbon-2’.The value of halogen substituents in directing the electrocyclization step is, however, not clear and a number of examples can be cited where a halogen substituent has been used to block the oxidative photo- c ycliza t ion. Thus, 2-c hl orostilbene undergoes oxidative p ho t ocycliza t ion to 1-chlorophenanthrene in 57 % yield28 and the diolefin (8) undergoes a double cyclization without loss of bromine.29 Bromine has been used as a blocking group in other helicene synthe~es.~~ Cyclization of (9) to 9-nitrophenanthrene is thought to proceed by carbon- iodine bond homoly~is~~ and subsequent intramolecular arylation, rather than through an electrocyclic intermediate, because attempted oxidative ring-closure of unsubstituted a-nitrostilbene fails.Some other examples of cyclization of C. S. Wood and F. B. Mallory, J. Org. Chem., 1964, 29, 3373. OD C. F. Wilcox, P. M. Lahti, J. R. Rocca, M. B. Halpen, and J. Meinwald, Tetrahedron Lett., 1978, 1893; R. H. Martin, C. Eyndels, and N.Defay, Tetrahedron Lett., 1972, 2731. S. M. Kupchan and H. C. Wormser, J. Org. Chem., 1965,30, 3792. Photochemistry and Photocyclization of Aryl Halides NO2 30 % ortho-iodo compounds [e.g, (lO)]31 have been shown to proceed by first photo- chemical reduction of the carbon-iodine bond (via bond homolysis) and then photo-oxidative ring closure of the stilbene in the presence of iodine formed from hydrogen iodide eliminated in the first step. It appears, therefore, that in simple stilbenoid systems an ortho-halogen substituent tends to adopt a rotational conformation so as to avoid the other phenyl ring in the excited state of the cis-isomer, and this reduces the probability for cyclization at its root. In the case of iodo-substituents homolysis of the weak carbon-iodine bond is usually the preferred reaction to electrocyclization and oxidation.The successful dehydrohalogenative cyclization of (4) reflects the opposite conformational preferences of this crowded system and is not a typical example. 3 Reactions; a Survey Most of the known photochemistry of aryl halides can be viewed formally as resulting from substitution processes occurring at the carbon-halogen bond, with displacement of the halogen by an incoming group. From this starting point we can classify these photoreactions according to the reactive atom of the incoming €TOUP. A. Reactions with Enolate-carbanionic Centres.-An important photoreaction of enolates with aryl halides in liquid ammonia has been reviewed by Bunnett.6 Analogous intramolecular reactions should have considerable potential in synthesis, and a few of these, such as the synthesis of (1 l),a2 have been described. The reaction proceeds by a chain mechanism, where the initiating step is *l G.Deluca, G. Martelli, P. Spagnolo, and M. Tiecco, J. Chem. SOC.(C), 1970,2504; R. M. Letcher and K. M. Wong, J. Chem. SOC.,Perkin Trans. 1, 1977, 178. 39 M. F. Semmelhack, R. D. Stauffer, and T.D. Rogerson, Tetrahedron Lett., 1973, 4519; M. F. Semmelhack and T. M. Barger, J. Org. Chem., 1977,42, 1481. Grimshaw and de Silva (11, 99%) transfer of an electron from some source to the aryl halide. One of the two species involved is photoexcited and the energy required for electron transfer is offset by a return from the excited state to the ground state.The aryl halide anion-radical then undergoes anion exchange with the enolate. The product anion-radical transfers its electron to a second aryl halide group so that a chain of anion exchange reactions can occur from one initiating electron transfer. B. Reactions with Alkene Carbon Centres.-The photocyclization of enamides has been a fruitful route for the synthesis of six- and seven-membered nitrogen heterocycles. Two distinct routes are followed. Cyclizations typified33 by the reaction of (12, X = C1 or Br) proceed through an electrocyclic intermediate and then loss of HX to yield the product. In these reactions other leaving groups besides halogen can be used. Other examples,34 such as the cyclization of (13), probably occur by a photochemically induced electron-transfer mechanism between the aryl halide, and triethylamine added as a scavenger for the hydrogen halide.Me0 Me0 (12) X = F, C1, Br, OAc, OMe, or NO2 85% when X = F 50% when X = C1 46 % (1 3) An example of a reaction which may involve homolytic bond cleavage followed by intermolecular radical addition to an olefin is shown by the pyridine com- 33 G. R. Lenz, J. Org. Chem., 1974,39,2839; T. Kametani, T. Sugai, Y. Shoji, T. Honda, F. Satoh and K. Fukumoto, J. Chem. SOC.,Perkin Trans. 1, 1977, 1151. 34 I. Tse and V. Snieckus, J. Chem. SOC.,Chem. Commun., 1976, 505; H. Iida,T. Takarai and C. Kibayashi, J. Org. Chem., 1978, 43, 975.Photochemistry and Photocyclization of Aryl Halides pound (14).35 The reaction mixture, however, shows no products from the radical polymerization of ethylene. Similarly, irradiation of an equivolume mix- ture of chlorobenzene and cyclopentene gives a mixture of cis-and trans-l-chloro-2-phenylcyclopentane.35 F F ’ + CH2=CHz hv no solvent 77% (14) C. Reactions with Aromatic Carbon Centres.-The large number of reactions belonging to this class is divided into two categories. Intermolecular Reactions. Reactions in this category mostly proceed by the homolysis mechanism and are typified by the conversion of 4iodobiphenyl to p-terphenyl on irradiation in benzene solution. Kharasch’s observations on the poor reactivity of 4-chloro- and 4-bromo-biphenyl promoted the generalization that photoreactivity decreased sharply on passing to the lower ha1ides.l This observation is not, however, general because both bond-dissociation energy and excited-state energy are determining factors for reactivity.Simple halobenzenes with higher energy excited states relative to biphenyls can arylate benzene whether the halogen group is iodo,l bromo,2 or chlor0.3~ In cases where both chlorine and iodine are attached to an arene, e.g. (15), then the carbon-iodine bond is preferentially cleaved.37 Interestingly, 2-chlorobiphenyl is able to take part in photo-arylation reactions,36 whereas the Cisomer is not, presumably because the excited-state energy levels in 2-substituted biphenyls are higher,s*v39 so that bond homolysis becomes possible.The photolysis of halogenophenols (16) in aqueous alkaline medium provides another example where the reactivity is comparable for all three halides, though the product distribution differs in the three cases.4o Formation of coupled products in this reaction seems to have inspired several alkaloid syntheses which are described below. Intramolecular Reactions. Photocyclization reactions have been used extensively 36 M. G. Barlow, R. N. Haszeldine, and J. R. Langridge, J. Chem. SOC.,Chem. Commun., 1979,608; D. Bryce-Smith, W. M. Dadson, and A. Gilbert, J. Chem. SOC.,Chem. Commun., 1980, 112. 36 G. E. Robinson and J. M. Vernon, J. Chem. SOC.,(C), 1971, 3363. 37 J. Bratt, B. Iddon, A. G.Mack, H. Suschitsky, J. A. Taylor, and B. J. Wakefield, J. Chem. SOC.Perkin Trans. 1, 1980, 648. 38 C. M. O’Donnell, K. F. Harbaugh, R. P. Fisher, and J. D. Wine-Fordner, Anal. Chem., 1973, 45, 609. P. J. Wagner, J. Am. Chem. SOC.,1967, 89, 2820; P. J. Wagner and B. J. Scheve, ibid., 1977,99,2888. 40 K. Omura and T. Matsuura, Tetrahedron, 1971, 27, 3101. Grimhaw and ak Silva C1 c1 (16) X = C1, Br, or I (17) +(17, X = H) + (16, X = H) + (16, X = OH) in synthesis and a number of these reactions were used to illustrate the previous discussion of reaction mechanisms. The photosynthesis of isoquinoline alkaloids was reviewed in 1972.13 The activating chromophore in these reactions is usually a benzene ring so that bromo-compounds such as (18), in addition to iodo-compounds, undergo satisfactory reaction by a bond homolysis process.l3* 41~4~A complication when using secondary amines is the parallel formation of berberines by addition of photo-liberated formaldehyde to the amino-f~nction.~~ Thus (18) gives rise to (19) in addition to the expected product.Formaldehyde hv, aq. acidBr t NaHSO, + OH OMe OH OH is liberated from substrates which contain methylenedioxy- or methoxy-groups and, in small amounts, from substrates which contain neither of these groups.42 The influence of excited-state energy on the rate of bond cleavage can be seen in some examples of cyclization and in lack of reactivity reported in heterocyclic series. Thus, the pyridine compounds (20) cyclize satisfactorily43 on irradiation, whereas the more conjugated quinoline (21) gives no cyclization owing to its 41 S.Rajeswari, H. Suguna, and B. R. Pai, Indian J. Chem., Sect. B, 1977, 15, 592; B. R. Pai, S. Natarajan, H. Suguna, and G. Manikumar, ibid., 1977, 15, 1042. 4a T. R. Govindachari, K. Nagarajan, S. Rajeswari, H. Suguna, and B. P. Pai, Helv. Chim. Acta, 1977, 60, 2138. 43 A. Fozard and C. K. Bradsher, J. Org. Chem., 1967, 32,2966; C. K. Bradsher and C. F. Voigt, ibid., 1971, 36, 1603. 193 Photochemistry and Photocydization of Aryl Halides X p,Y -t (20) X = C1 or Br,Y = H 75 % when (21) X = C1 or Br X = H, Y = C1 or Br x= H,Y = Br low excited-state energy and to the poor n-donor ability of the pyridinium ring, which reduces the possibility for assisted homolysis.44 Commonly encountered side reactions are of two types.Firstly, photoinduced homolytic bond cleavage occurs at the weakest bond in a structure so that unexpected reactions can be observed alongside the expected cyclization. Thus, reaction of (22) leads to a mixture of products where compound (23) is thought 0 hV N ____+ HCb CH HOk&d, O4 / to result from a photo-Fries type migration of the o-iodobenzoyl group to the 3-position in competition with carbon-iodine bond h0molysis.~5 Secondly, ring photosyntheses involving aryl a-radical intermediates are successful only if performed in a solvent of low hydrogen-donor ability, otherwise replacement of halogen by hydrogen occurs to a significant extent. Benzene or t-butanol are often used as solvent, and also aqueous acid for amine substrates or aqueous alkali for phenol substrates.The majority of the reported photocyclization reactions involve formation of six-membered rings. There are a few examples of five-membered ring-forming reactions and some rings larger than six-membered have been successfully constructed by this photochemical route in spite of the inevitable competition with simple replacement of halogen by hydrogen. Examples are found of the 44 D. E. Portlock, M. J. Kane, J. A. Bristol, and R. E. Lyle, J. Org. Chem., 1973, 38, 2351. W. Carruthers and N. Evans, J. Chem. SOC.,Perkin Trans. 1, 1974, 1523. 194 Grimshaw and de Silva formation of a seven-membered ring4s from (24), an eight-membered ring47 from (25) and a nine-membered ringM from (26).35% 1 IH CH2-N-CH2 hv0: H F2 aq.HCI ’ Ph (25) 57% CHzCH2 \NH \ hw, aq.Me0 c=0 -+ MeOH, NaOH Me0 OMe A wide range of alkaloids have been synthesized by cyclization of substrates where the unhalogenated aryl moiety possesses a phenolic group and the reaction is carried out in aqueous alkaline solution. The first reports of this type appeared in 1971.49 In these cases the substrates possess a basic nitrogen centre and so the phenoxide ion must effectively compete with the nitrogen lone pair for association with the excited aryl halide moiety in giving rise to cyclization by the electron- transfer mechanism. This cyclization shows a high degree of regiospecificity for arylation either ortho-or para-to the phenoxide group.The alternative view that 46 N. E. Brightwell and G. W. Griffin, J. Chem. SOC.,Chem. Commun., 1973,37. 47 P. W. Jeffs and J. F. Hansen, J. Am. Chem. SOC.,1967,89,2798; P. W. Jeffs, J. F. Hansen, and G. A. Brine, J. Org. Chem., 1975,40,2883. 48 K. Ito and H. Tanaka, Chem. Pharm. Bull. Jpn., 1974,22,2108. 49 T. Kametani, H. Nemoto, T. Nakano, S. Shibuya, and K. Fukumoto, Chem.Ind. (London). 1971, 788; R. J. Spangler and D. C. Boop, Tetrahedron Lett., 1971,4851. Phottrchemistry and Photocyclization of Aryl HalzHes these reactions follow the energy transfer, homolytic bond cleavage mechanism dismisses the possibility of electron transfer from the amine lone pair which is known to occur in competition to simple homolysis.The directing effect of the phenoxide group has been exploited for the synthesis of a large number of spiro-systems. Reducing conditions obtained by addition of sodium iodide have been used5O to prevent the accumulation of bromine which would attack the substrate (27). Addition of sodium borohydride which reduces kv, H,O NaOH,NaI ’ OH OMe (27) photosensitive cyclized products in situ has been made, for example in reaction of (2), so as to reduce the likelihood of photodegradation.8 However, the known photoreduction of aryl halides by sodium borohydride merits caution in its use. Even though the yields of cyclization products are rarely good, the simplicity of this reaction has been reason enough for its continued use.The yields improve when fully aromatic isoquinolines such as (28) are used as s~bstrates.5~95~ OH RZ Phenanthridine systems are another class of heterocycles whose synthesis via aryl halide photochemistry has been examined in some detail. The reported photosynthesis of phenanthridone by irradiation of o-iodobenzanilide probably T. Kametani, T. Kohno, R. Charubala, and K. Fukumoto, Tetrahedron, 1972, 28, 3227. s1 T. Kametani, R. Nitadori, H. Terasawa, K. Takahashi, M. Ihara, and K. Fukumoto, Tetrahedron, 1977, 33, 1069. s* S. M. Kupchan and P. F. O’Brien, J. Chem. Soc., Chem. Commun., 1973, 915. Grimshaw and de Silva proceeds by formation of benzanilide which is known53 to be further converted to phenanthridone in the presence of an oxidant.Since this early work, benzanilides have been demonstrated to show restricted rotation about the peptide bond and to adopt a configuration with trans-phenyl groups, so that cyclization cannot be expected for radicals formed by carbon-halogen bond homolysis of o-iodobenzanilides. Later work with o-iodo-N-methylbenzanilides, which adopt a cis-diary1 configuration, resulted in higher yields of the phenan- thridone from bond homolysis followed by radical substitution, together with dimeric compounds resulting from cyclization to give a five membered ring.54 This approach of cyclizing N-alkyl-o-halogenobenzanilideshas led to several successful applications including the synthesis of substituted phenanthridones55 such as (29) and the synthesis of benzophenanthridones45~~~from o-bromo-N- met h y1benznap hthamides .o-Chlorobenzanilides can be made to undergo efficient photocyclization even though they adopt the trans-coniiguration in s0lution.5~ Because the excited state energy of benzanilide is less than the carbon-chlorine bond energy, the trans- form of (30) is photochemically stable and undergoes only the trans, cis inter-& (29) 22% hv cyclohexane 0 (30) B. S. Thyagarajan, N. Kharasch, H. B. Lewis, and W. Wolf, J. Chem. SOC.,Chem. Commun., 1967, 614. 64 D. H. Hey, G. H. Jones, and M. J. Perkins, J. Chem. SOC.,Perkin Trans. 1, 1972, 1150. 65 A. Mondon and K. Krohn, Chem. Ber., 1972, 105, 3726; R.K.-Y. Zee-Cheng, S.-J. Yan, and C. C. Cheng, J. Med. Chem., 1978,21, 199. m W. J. Begley and J. Grimshaw, J. Chem. SOC.,Perkin Trans. 1, 1977, 2324. J. Grimshaw and A. P. de Silva, J. Chem. SOC.,Chem. Commun., 1980, 302. Photochemistry and Photocyclization of Aryl Halides conversion. The cis-configuration has the assisted homolysis mechanism available to it and undergoes efficient cyclization. Thus, efficient photocyclization from these flexible molecules is possible for chloro-compounds but not for bromo- or iodo-compounds, since for these the trans-configuration undergoes carbon- halogen bond homolysis. o-Bromobenznaphthamides have also proved use- ful substrates for cyclization,5* and a similar mechanism may operate here where the energy considerations make bromo-compounds the most satisfactory substrates. The photoinduced electron-transfer between halobenzenes and aromatic amines has been applied to intramolecular examples.Reaction of (31) affords phenanthridine,59 the expected dihydro-product dehydrogenates under the reaction conditions, and a dimer that may arise by photoalkylation*O of phenanthridine by dihydro phenanthridine under the reaction conditions Compounds which have a fixed cis-diary1 configuration show an increased yield of cyclized product, often with the virtual exclusion of radical reactions with the solvent. For the synthesis of phenanthridine-type compounds this configuration has been achieved by incorporating the carbon-nitrogen bond into a heterocyclic system,’* into a cyclic system (32) containing a hydrogen bond,e1 or into a co-ordination complex (33).62 Compound (32) showed a h”3MeCN>HIO, NaOH q+ [Jy2 -(31) X = C1 or Br 14 % 28% with X = C112% 9% withX = Br Ph N’ x Ph (32) X = Br or I (33) X = C1 or Br 68 S. V. Kessar, 0. Singh, and P. Balakrishnan, Tetrahedron Lett., 1974, 2269; I. Ninomiya,J. Naito, and H. Ishii, Heterocycles, 1975, 307. 59 K. Mizuno, C. Pac, and H. Sakurai, Bull. Chem. SOC.Jpn., 1973, 46, 3316. (O F. R. Stermitz, R. Pua-Seiber, and D. E. Nicodem, J. Org. Chem., 1968, 33, 1136. 61 J. Grimshaw and A. P. de Silva, J. Chem. SOC.,Chem. Commun., 1980, 301. S. Prabhakar, A. M. Lobo, and M. R. Tavares, J. Chem. SOC.,Chem. Commun., 1978,884. Grimhaw and de Silva wavelength-dependent cyclization quantum yield demonstrating that, in this case, carbon-halogen bond fission occurs from an upper excited state and that it can compete with the normally fast internal conversion to a lower, and in this case inactive, excited state.There are a number of examples of photocyclization of polychloro-compounds to indicate that, in general, in the benzene and pyridine series, ortho-halogeno- substituents are more photolabile than meta- or para-substituents. Thus, (34) is converted to a dichlorobenzofurane3 and the chloropyridine (35) is ~yclized3~ without loss of further chloro-substituents. c1usc1’ hhexanev, c1 c1fcDclCI (35) 58 % This field of photocyclization is very large and some relevant mechanistic studies have already been discussed, yet in general the field suffers from a lack of any but the most rudimentary quantitative information. An exception is the photoreaction of (36) where the proportion of cyclized product, (37), is large + (36, X = H) + (36, X = Ph) (36, X = Cl, Br, or I) (37) O3 A.Norstrom, K. Anderson, and C. Rappe, Chemosphere, 1976, 5, 21 199 Photochemistry and Photocyclization of Aryl Halides from the chlorocompound but negligible from the iodocompound. 64 These facts were originally explained in terms of thermodynamic versus kinetic control determined by the oxidizing ability of the halogen atom. According to the assisted homolysis model, such halogen-atom dependent product distributions can arise by the higher electron affinity of chlorine causing a strongly complexed transition-state, which leads to cyclization.In the case of the iodo-compound the tendency will be to proceed by simple unassisted homolysis, especially where a high strain energy is associated with complexation. D. Reactions with Hydrogen Centres.-Reductive dehalogenation is a ubiquitous phenomenon in aryl halide photochemistry and is the chief reason for the low yields attained in photocyclizations conducted in organic solvents. In extreme cases cyclization is totally inhibited45165 by the direct route, but secondary electrocyclic reactions of the dehalogenated product may allow oxidative ring-closure. Even during photoarylation reactions carried out in neat aromatic solvents, photoreduction takes place in varying degree~~J~*3~ because hydrogen atoms are transferred to aryl radicals from hydrogen iodide and species of the type (38).Traces of oxygen can assist the desired reaction by scavenging species (38), but excess oxygen also scavenges aryl radicals.66 Reductive dehalogenation has been turned to preparative advantage in a few cases. Irradiation of pentachloropyridine in dioxan gives 2,3,4,6-tetrachloro- pyridine.37 Positions on an aromatic ring can be blocked by halogenation and then unblocked by a photochemical reaction, as appropriate, during a syn-thetic sequence.ll Irradiation of the aryl halide in [2Helacetone leads to specific replacement of the halide by deuterium,ll although nonphotochemical methods are usually more practical for purposes of specific deuterium labelling.Environmental photochemists have studied photodehalogenation to assess the role of solar degradation in the detoxification of halogen-containing pollutants. Points of possible interest in general preparative work which arise are the following. Ortlzo-chloro- and ortho-bromo-substituentsshow enhanced photoreactivity, with respect to their positional isomers, in the biphen~1,~3925~67 64 W. A. Henderson and A. Zweig, J. Am. Chem. SOC., 1967, 89, 6778; W. A. Henderson, R. Lopresti, and A. Zweig, ibid., 1969, 91, 6049. 65 J. Stumpe, A. Mehlhorn, and K. Schwetlick, J. Photochem., 1978, 8, 1. 86 N. Kharasch and R. K. Sharma, J. Chem. SOC., Chem.Commun., 1966, 106. 67 T. Nishiwaki, M. Usui, K. Ada, and M. Hida, Bull. Chem. SOC. Jpn., 1979, 52, 821. Grimhaw and de Silva ter~henyl,~~~68 and diphenyl ether631 69 series. Some benzene derivatives show this effect, the most notable exceptions being substituted phenols and anilines where the rneta-chloro-isomers are the most reactive.70 The enhanced lability of ortho-substituents can be explained because, firstly in the biphenyl series ortho-substitution is known to raise the excited-state energy, and secondly elimination of a crowded substituent causes relief of strain. Increased photodecomposition rates of aryl halides of low intrinsic reactivity have been attained by adding aliphatic amine~~~923925 or aqueous alcoholic alkali,67$71and these effects could be important in pollution control.Exciplexes are demonstrable between aryl halides and aliphatic amines and the dehalo- genation rate increases with solvent polarity, implying electron transfer as the key step. Quantum yields considerably larger than unity have been observed for dehalogenation in the presence of additives, and are attrib~ted~~9~' to a chain reaction propagated by carbonyl radical-anions originating from the alchohol solvent. A chain propagation by borane radical-anions is reponsible for the high photoreactivity of halobenzenes in the presence of sodium borohy- dride.72 E. Reactions with Halogen Centres.-Halogen exchange is a little studied aspect of aryl halide photochemistry. Some of the early examples of iodine exchange do not involve photoexcitation of the aromatic system, but rather proceed by (visible) light-induced fission of molecular chlorine or bromine, followed by radical attack on the i0doarene.~3 Genuine cases of participation by the photo- excited states of aryl halides are found during the photolysis of the iodoarene in halogenated aliphatic solvents, such as carbon tetrachloride and bromo- forms4974 Arene radicals generated by homolysis of the carbon-iodine bond attack the halogenated solvent.In a reaction of some utility, molecular iodine containing a radioactive isotope can be exchanged with aryl iodides under the influence of ultraviolet, but not visible, radiation. This, and also the intra- molecular reaction75 of (39) are understandable in terms of simple homolysis of the carbon-iodine bond followed by further reaction of the aryl radical on iodine. F.Reactions with other Heteroatom Centres.-Aryl halides form only a small part of the large number of aromatic substrates which undergo nucleophilic photosubstitution, and this reaction has been reviewed.5 Thus alkoxy-, hydroxy-, amino-, and thiocyanato-groups undergo photosubstitution on a variety of B. Chittim and S. Safe, Chemosphere, 1977, 6, 269. On A. Norstrom, K. Anderson, and C. Rappe, Chemosphere, 1977, 6, 241. 70 H. Parlar, P. G. W. Steven, R. Baumann, and F. Korte, 2. Naturforsch., Teil B, 1979, 34, 113. 71 C. Parkanyi and Y. J. Lee, Tetrahedron Lett., 1974, 1115. 72 J. A. Barltrop and D.Bradley, J. Am. Chem. Soc., 1973, 95, 5085. 73 W. Voegtli, H. Muhr, and P. Lauger, Helv. Chim. Acra, 1957, 37, 4170; B. Miller and C. Walling, J. Am. Chem. Soc., 1957, 79,4187; J. T. Echols, V. T. C. Huang, C. S. Parrish, J. E. Rose, and B. Milligan, ibid., 1967, 89, 4081. F. Kienzle and E. C. Taylor, J. Org. Chem., 1970, 35, 528. " T. Sato, S. Shimada, and K. Hata, Bull. Chem. SOC.Jpn., 1971,44, 2484. Photochemistry and Photocyclizatian of Aryl Halides (39) 21 % halogenated benzene, naphthalene, and other aromatic substrates. The halogen of choice for these reactions is either fluorine or chlorine, so as to avoid excessive competition from carbon-halogen bond homolysis, and subsequent radical processes. Radical-producing photochemical processes can also lead to heteroatom substitution, as in the formation of phenols during the photolysis of aryl iodides in the presence of oxygen.66 The photoconversion of 2-iodo-2’-methylthio-biophenyl to dibenzothiophene76 is probably a radical process.Similarly, aryl iodides and bromides can be sulphenylated with disulphides,77 or converted to phenyl phosphonates with trialkylpho~phites.~~ COzEt hw---+ cisdorm A (40) x = c1 Y = C1 or OMe Y = CI or OMe (42) 69 % ’% J. A. Kampmeier and T. R. Evans, J, Am. Chem. SOC.,1966,88,4096. 77 T. Fujisawa and H. Ohta, Bull. Chem. SOC.Jpn., 1976, 49, 2341.’* J. B. Plumb and C. E. Griffin, Tetrahedron Lett., 1966, 5049. Grimshaw and de Silva A few intramolecular heteroatom-centre reactions have been reported, and there appears to be considerable scope for further exploration.Photocyclization of (40)is interesting in that 2,6-disubstitution appears necessary for ring closure. This may be due to the 2-substituted compound adopting a conformation where the halogen atom and ethoxycarbonyl group are remote. Low-temperature studies support an electrocyclic reaction followed by elimination of RCl to yield (41), which can cyclize to give the product.79 No information is available to decide if the cyclizationgO of (42) is a nucleophilic photosubstitution or a radical reaction. 4 Conclusions Photohomolysis of haloarenes has proved to be a useful source of aryl radicals in synthesis and the intramolecular cyclization of the radicals so formed onto an adjacent arene group is a useful step for the synthesis of a large number of alkaloids.However, the yields of cyclized product have often been low and the route is, in many cases, considered attractive only because of its simplicity. Quantitative information on these and related reactions has accumulated to indicate a range of mechanisms for carbon-halogen bond cleavage. It is now possible to select reaction conditions so as to achieve better yields, at least for some cyclizations. We hope that this review will assist in defining areas where photocyclization can be successfully used. 7B R. Arad-Yellin, B. S. Green, and K. A. Muszkat, J. Chem. SOC.,Chem. Commun., 1976, 14. R. Paramasivan, R. Palaniappan, and V. T. Ramakrishnan,J. Chem. SOC.,Chem. Commun., 1979, 260.

ISSN:0306-0012    

DOI:10.1039/CS9811000181

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

Singlet Molecular Oxygen By A. A. Gorman CHEMISTRY DEPARTMENT, UNIVERSITY OF MANCHESTER MANCHESTER M13 9PL U.K. and M. A. J. Rodgers CENTER FOR FAST KINETIC RESEARCH UNIVERSITY OF TEXAS AT AUSTIN AUSTIN, TEXAS 787 12, U.S.A. 1 Introduction In 1928 an article in Nature appeared which presented an interpretation of the atmospheric oxygen absorption bands and established the electronic levels of the oxygen mo1ecule.l Atmospheric spectroscopy allowed measurements of the energy differences between the ground state (3c,-)and the low-lying ld, and 1c,+states.2 With a single notable exception the ld ,state remained within the province of spectroscopy until the early 1960’s, at which time it attracted the attention of solution-phase chemists and biologists; a phenomenal literature on the subject has since arisen, by far the major part since 1965.Singlet oxygen is now believed to be involved in many dye-sensitized photo-oxidations of organic systems and in the quenching of ,electronically excited molecules. It has been invoked as an intermediate in phenomena such as chemiluminescence, photo- dynamic action, photocarcinogenicity, and decomposition of 02-rich compounds. In this review we attempt to present a glimpse of the main features of the accu- mulated research and to indicate where the future lies. Many excellent re~iews3-l~ have appeared over the years which discuss the several aspects of the physics, chemistry, and biology of this intriguing metastable entity. Of the sixteen electrons in the 02 molecule, twelve have valence implications : R.S. Mulliken, Nuture,1928, 122, 505. a W. H. J. Childs and R. Mecke, Z. Physik, 1931, 68, 344. C. S. Foote, Acc. Chem. Res., 1968, 1, 104. K. Gollnick, Adv. Photochem., 1968, 6, 1. J. W. Hastings and T. Wilson, Photophysiology, 1970, 5, 49. R. P. Wayne, Adv. Photochem., 1969, 7, 311.’M. Kasha and A. U. Khan, Ann. N. Y. Acad. Sci., 1970, 171, 5. D. R. Kearns, Chem. Rev., 1971, 71, 395. W. Adam, Chem. Z., 1975,99, 142. lo‘Singlet Molecular Oxygen’, ed. A. P. Schaap, Dowden, Hutchinson and Ross, Strouds-burg, Pennsylvania, 1976. l1 B. Stevens, Acc. Chem. Res., 1973, 6, 90. la C. S. Foote, in ‘Free Radicals in Biology’, Volume 11, ed. W. A. Pryor, Academic Press, New York, 1976.l3 K. Gollnick, in ‘Radiation Research : Biomedical, Chemical and Physical Perspectives’, ed. 0. F. Nygaard, H. I. Adler and W. Sinclair, Academic Press, New York, 1975. l4 ‘Singlet Oxygen’, ed. H. H. Wasserman and R. W. Murray,Academic Press, New York, 1979. 205 Singlet Molecular Oxygen the inner four (KK) probably retain essentially atomic character. According to MO theory the remaining twelve occupy the lowest available bonding and anti- bonding orbitals. The 20, (bonding) and 2au (antibonding) pair are each doubly occupied and contribute only weak interactions. The 2pz atomic Orbitals from each atom give rise to a 30, bonding pair which is doubly occupied, gives 02 its bond, and defines the molecular axis. The doubly degenerate Inu bonding pairs are both doubly occupied, leaving two electrons for the 17, antibonding pair. Strong interactions are contributed by the 3ag, and 17rU and lrg electrons; thus 02has one a-bond, two n-bonds, and one 7-antibond, i.e.two electon-pair bonds, one a and one n. The distribution of the two electrons in the ITg orbitals deter- mines the overall symmetry, angular momentum, and spin characteristics of 02. The combination of two indistinguishable electrons with two orbitals results in six electronic sub-states having different electron distributions, energies, and magnetic properties. There are three degenerate 3Zg-states, two equal-energy ld states, and a unique lC,+state. The ld I and lC,+ states are 0.98 eV and 1.63 eV above the sz,-ground state respectively. The three states have almost identical binding energies and dissociate at a common energy limit (ca.5eV). The transitions ld ,t3zg-and lzu+4-3Z8-are highly forbidden and corres- pond to very weak absorption coefficients and long upper-state lifetimes at zero pressure. A full and lucid description of the electronic structure of oxygen has appeared re~ent1y.l~ The upsurge in interest in the chemistry of 02(ld,) stemmed from studies of the orange-red luminescence emitted from mixed hydrogen peroxide-sodium hypochlorite solutions. Originally observed in 192716 it was rediscovered in 1960,17 shown to consist of two red bands, and incorrectly assigned to a solvent- shifted 1CQ+-+3cs-transition in 1963.IS Comparison of the spectra of emissions from the H202-OCl- system and from radiofrequency discharges in 02 gas led19 to assignment of the chemiluminescence peaks to transitions from an 04 species. It is now accepted that the weak bands at 634 and 703 nm giving rise to the glow arise from the simultaneous transitions2* of equations (1) and (2), the energy difference corresponding to the vibrational spacing in ground- 02('dS)+ O2('dg)-+ 02(3Zg-) + 0,(34-) + hv (634 nm) (1) 02('4)+ 02('dg)-+ 02PZg-) + 02(3 + hv (703 nm) (2)Zg-) state 02.Simultaneous transitions can occur as a consequence of collisions between excited molecules and can lead to 'pooling' of the excitation energy into a single photon. Such processes are only usually observed when the forbiddenness of the single-molecule transition is so large that sufficient collisions between excited states can occur. The formation of the molecular pair state (Id,) (Id,) of overall singlet multipicity is able to correlate with the singlet component of l5 M.Kasha and D. E. Brabham, Ref. 14, Chapter 1. l6 L. Mallet, C.R. Hebd. Seances Acad. Sci.,1927, 185, 352. l7 H. H. Seliger, Anal. Biochem., 1960, 1, 60. l* A. U. Khan and M. Kasha, J. Chem. Phys., 1963, 39, 2105. 19 S. J. Arnold, E, A. Ogryzlo, and H. Witzke, J. Chem. Phys., 1964, 40, 1769. ao A. U. Khan and M. Kasha, J. Am. Chem. SOC.,1970,92,3293. Gorman and Rodgers the (3 cg-)(3 ground-state pair which has singlet, triplet, and quintuplet 2,) contributions. Energy pooling via simultaneous transitions has been observed21 for all the possible ld,, Z,+, 32,-combinations leading to photons of wave- lengths as short as 380 nm.Only recently have singlet oxygen luminescences been observed as a result of photosensitization in the liquid phase. The long radiative lifetime and the efficient deactivation processes which occur in liquids lead to very low quantum yields of luminescence. Nevertheless, high-intensity light sources, sensitive photodetectors, and noise-discrimination techniques have led to detection of the ldg-+3~,-transition at 1268 nm in chlorocarbon and fluorocarbon solvents where lifetimes approach 1 OV3 second^.^^-^^ A Russian groupz3 has reported time resolution of the 1do+3cg-emission (1270 nm) in cc14 solution.During the early 1960’s the attention of organic chemists was drawn towards singlet oxygen. In a series of contiguous communications in 1964 it was shown that organic substrates were oxidized by (i) the peroxide-hypochlorite reacti0n,~5 (ii) dye-photosensitization,26 and (iii) oxygen submitted to electrodeless di~charge,~’in such a way as to allow the conclusion that Oz(ld,) was the common reactive species. Eventually it became established that many dye- sensitized photo-oxidations proceed via an 02(ld ,) intermediate3 and it trans- pired that the mechanistic conclusions made in the early 1930’~2812~ and largely ignored until the mid-1960’s concerning the intervention of excited 02 molecules in dye photo-oxidations were revived and substantiated.At this time considerable evidence exists that Oz(ld,) is a reactive inter- mediate in photo-oxidation reactions. 02(1 c,+)is, however, more elusive in condensed phase. It has been estimated30 that in solution lC,+is relaxed to Idgin ca. 10-9 s by collisional quenching interactions and although it is well characterized as a gas-phase intermediate31 its solution properties are uncertain. The remainder of this review concerns Id, only. An excellent reprint volume which collects together the benchmark papers in this subject has been published.1° 2 Methods of Generation Three methods of production of O~(1d ,) have been important for performing quantitative measurements of its reactivity : peroxide decomposition, high- frequency discharge, and energy transfer.In addition singlet oxygen formation has been invoked in a miscellany of chemical and biological systems.32 21 K. Furukawa, E. W. Gray, and E. A. Ogryzlo, Ann. N. Y. Acad. Sci., 1970,171, 175. 22 A. A. Kraznovsky, Photochem. Photobiof., 1979, 29, 29. 2s I. M. Byteva and G. R. Gurinovitch, J. Lumin, 1979, 21, 17. 24 A. U. Khan and M. Kasha, Proc. Natl. Acad. Sci. USA, 1979, 76, 6047. 25 C. S. Foote and S. Wexler, J. Am. Chem. SOC., 1964, 86, 3879. 2a C. S. Foote and S. Wexler, J. Am. Chem. Soc., 1964, 86, 3880. E. J. Corey and W. C. Taylor, J. Am. Chem. SOC.,1964,86, 3881. H. Kautsky and H. de Bruijn, Naturwissenschqften, 1931, 19, 1043. 29 H. Kautsky, H. de Bruijn, R. Neuwirth, and W. Baumeister, Chem. Ber., 1933, 66, 1588.30 S. J. Arnold, M. Kubo, and E. A. Ogryzlo, Adv. Chem. Ser., 1968, No. 77, p. 133. s1 E. A. Ogryzlo, in Ref. 14, Chapter 2. s2 R. W. Murray, in Ref. 14, Chapter 3. Singlet Molecular Oxygen A. Peroxide Decomposition.-The H202-OCl-system in water has already been mentioned. The reagent can be generated either by adding aqueous OC1- solution to an alkaline solution of H2Oz or by bubbling Cl2 gas into alkaline hydrogen peroxide. The reaction proceeds according to equation (3). Spectro-scopic evidence shows that both ld, and L',+ states are generated and in OCl-+ Ha02 + C1-+ HSO + 0% (3) methanolic solutions the efficiency of ld , production approaches 80z.32 Using l8O-enriched H202 or OC1- it was demonstrated that the 02 is derived wholly from H202 and not from OC1- or water.33 A variation on this method uses bromine and alkaline H202.34 In aqueous alkaline solution organic peracids decompose with evolution of oxygen, some of which was proposed to be in the ld, ~tate.3~More recent work35 casts doubt on whether Id, production is a significant channel.The proposal,36 based on trapping experiments, that the oxidation (by Ce4+ has been confirmed by spectro- ,) 04ldof s-butylhydroperoxide gives rise to ions) scopic measurement^.^^ A discussion of whether various transition-metal- oxygen complexes (e.g. potassium perchromate) release 02(ld ,) on decompo- sition has recently been published.32 Endoperoxides of aromatic compounds formed by the interaction of the substrate with singlet oxygen have been shown to re-form O~(1d g) when heated.Substrates such as rubrene, 9,10-diphenylanthracene,38 and 2,5-diphenyl-furan39 have been shown to behave in this way. A similar scheme has been proposed for 1,3-dipheny liso benzo furan. 40 Organic phosphites interact with ozone to yield adducts which are stable at low temperatures but break down on warming to phosphate and molecular oxygen. Multiplicity considerations and experimental observations led to the conclusion that singlet molecular oxygen was a product of this decomp~sition.~~ A detailed account of phosphite-ozone reaction is contained in a recent singlet oxygen review volume.32 The question as to whether the superoxide ion dismutes to yield 02(ld g) has been the subject of much discussion.The proposition was initially made based on a work-up of products from the decomposition of KO2 in DMSO containing 2,5-dimethylfuran.42 Luminescence observations of the decomposing KO2- DMSO system, however, showed only a weak when small amounts of 33 A. E. Cahill and H. Taube, J. Am. Chem. Soc., 1952,74, 2312. 34 E. McKeown and W. A. Waters, J. Chem. Soc. (B),1966, 1040. R. F. Boyer, C. T. Lindstrom, B. Darby, and M. Hylarides, Tetrahedron Lett., 1975,4111. 36 J. A. Howard and K. U. Ingold, J. Am. Chem. SOC., 1968,90, 1056. 37 M. Nakano, K. Takayama, Y. Shimizu, Y.Tsuji, H. Inaba, and T. Migita, J. Am. Chem. Soc., 1976, 98, 1974. 38 H. H. Wasserman, J. R. Scheffer, and J. L. Coopei', J. Am. Chem.Soc., 1972, 94, 4991. 39 A. M. Trozzolo and S. R. Fahrenholtz, Ann. N.Y. Acad. Sci., 1970, 171, 61. 40 A. Singh, N.R. McIntyre, and G. W. Koroll, Photochem. Photobiol., 1978, 28, 595. 41 R. W. Murray and M. Kaplan, J. Am. Chem. SOC., 1968, 90, 537, 4!4 A. U. Khan, Science, 1970, 168, 476. 43 R. Nilsson and D. R. Kearns, J. Phys. Chem., 1974,78, 1681. Gorman and Rodgers water were present and no oxidation of singlet-oxygen-reactive substrates was found. On the other hand, electrolytically produced 020-converted 1,3-diphenyl- isobenzofuran into dibenzoylben~ene.~~ The production of 02’-in the xanthine- xanthine oxidase system45 is accompanied by luminescence which has been proposed as arising from 02(1dg).46 Other work on this enzyme system led to the suggestion that hydroxyl radicals produced in a biological cycle similar to the Haber-Willstatter reaction, equation (4), can oxidize 02’-to 02(ldg).47 The lack of inactivation of the enzymes lysozyme and ribonuclease under 02*-+ H,O, + 0,+ OH-+ *OH (4) y-radiolysis conditions48 where HO2-(pH 3.1) or 029’ (pH 7.6) are the only radical species present, strongly suggests that Oz(ld,,) is not formed from HO2* or 02*-.This is supported by recent observations of the decomposition of tetramethylammonium superoxide in DMSO in the presence of cholesterol supported on microbe ad^.^^ From measurements of the yields of cholesterol oxidation products (specific for singlet oxygenso) it was concluded that no more than 0.2% of 02(1dg)was produced even when corrected for quenching by 02.’.B. Electrical Discharge.-ElectricaI discharges have been used as a means of generating metastable molecules in the gas phase since the turn of the century. The development of high-power radio and microwave generators has allowed elimination of electrodes in the gas stream since such radiation can penetrate glass walls. The most commonly used generators are microwave (2450 MHz), the power from which can be localized and directed via waveguides. Gaseous oxygen excited in such a way yields 02(ld ,,) and 02(l cg+),which are detected by their light emission. A system like this can be used for measuring the rate parameters for reaction of excited oxygen states with gas-phase substrates, or in a preparative manner by exposing the gas stream to a stirred condensed- phase reactor containing substrates.31 C. Energy Transfer.-Singlet States.Oxygen has long been known to quench the fluorescence of many organic molecules with diffusion-controlled rate constants, but the exact mechanism of this quenching remains controversial. The interaction of an organic molecule (1M) in its first excited state (1M*) with molecular oxygen (3 zg-)produces a collision complex of triplet multi- plicity, designated by 3(M-02)*. The decay of S(M-02)* to 1M and OZ(~C~-) is regarded as unlikely, owing to the large amount of electronic energy to be dissipated by the complex and the small Franck-Condon factor.51 Decay of O4 E. A. Mayeda and A. J. Bard, J. Am.Chem. SOC., 1974, 96, 4023. I6I. Fridovich, J. Biol. Chern., 1970, 245, 4053. J. Stauff and H. Wolf, 2. Naturforsch., Teil B, 1964, 19, 87. O7 R. M. Ameson, Arch. Biochem. Biophys., 1970, 136, 352. 4* G. E. Barlow, R. H. Bisby, and R. B. Cundall, Radiat. Phys. Chem., 1979, 13, 73. OS C. S. Foote, F. C. Shook, and R. A. Abakerli, J. Am. Chem. SOC., 1980,102,2504. so M. J. Kulig and L. L. Smith, J. Org. Chem., 1973, 38, 3639. s1 A. U. Khan and M. Kasha, Ann. N. Y. Acad. Sci., 1970, 171, 5. 209 Singlet Molecular Oxygen 3(M-02)* to 3M* and 02(3&-) is considered to be more probable. The dis- sociation of 3(M-02)* to 3M* and Oz(ld,) is a spin-allowed process, and debate has surrounded the ability of this process to complete with the overall deactiva- tion to yield 3M* and ground-state oxygen. In order for Oz(ld,) to be formed in the energy-transfer step, the lM*-+3M* gap of the organic molecule must exceed 94.5 kJ mol-l, the energy of ld,.Since rate constants for oxygen quenching of 1M* emission have been shown to be diffusion controlled regardless of the size of this energy gap for aromatic species, it has been concluded that energy transfer to 02is not necessary for efficient quenching.52 In studies of self-sensitized photoperoxidations, it was found that the products of 1M* quenching were 3M* and 02(3&-) rather than 3M* and 02(ldp) for species M with high intersystem crossing yields.53 However, subsequent studies54055 involving the strongly fluorescent hydrocarbon rubrene suggest that O2(ldg) may indeed be produced as a consequence of oxygen quenching of 1M*.The experimental result for rubrene (which has also been observed for several other aromatics with high @F~~)was that singlet oxygen was produced with a quantum yield in excess of unity; this phenomenon can be accounted for by processes (5) and (6). lM*+ 01(3Zu-)-+ 3M* + Ol('dU) (5) 3M* + 0,(3Zu-)-+ 'M + O4(ldg) Therefore, it appears likely that certain aromatic singlet states may sensitize singlet oxygen formation, particularly if the fluorescence quantum yield is high. Another reports' has provided evidence that singlet excimers of pyrene and naphthalene sensitize the formation of singlet oxygen. Triplet States. Tripet states are also quenched by molecular oxygen, but rate constants for this process have values that are often an order of magnitude below those for quenching of singlet states.A factor of one-ninth can be explained on the basis of spin statistics.8 A collision complex of the triplet aromatic (3M*) with ground-state oxygen can be formed with singlet, triplet, or quintet multiplicity, i.e. with only a one in nine probability of singlet formation. Only the singlet complex can yield singlet oxygen directly; the triplet complex dissociates to give ground-state products and the quintet can only return to 3M* and 02(3cg-). The observed limiting value of one-ninth of the diffusion- controlled rate constant thus suggests that quenching occurs only via the singlet complex. Several cases have been reported of quenching rate constants exceeding one-ninth of the diffusion-controlled rate constant,58 particularly for amines. These results have been explained in terms of an intermediate charge-transfer 6a C.S.Parmenter and J. D. Rau, J. Chern. Phys., 1969, 51, 2242. 63 B. Stevens and B. E. Algar, J. Phys. Chern., 1968, 72, 3468. 64 B. Stevens and J. A. Ors, J. Phys. Chern., 1976, 80,2164. 65 K. C. Wu and A. M. Trozzolo, J. Phys. Chern., 1979, 83,2823. 66 K. C. Wu and A. M. Trozzolo, J Phys. Chern., 1979, 83, 3180. 67 D. M. Shold, J. Photochem., 1978, 8, 39. 6* A. Garner and F. Wilkinson, Chern. Phys. Lett., 1977, 45, 432. 210 Gorman and Rodgers complex with an energy E(lM*) < ECT< JT(~M*).This mechanism allows for enhanced oxygen quenching rates by the processes of intersystem crossing between singlet and triplet charge-transfer complexes and/or deactivation of the triplet collision complex to ground-state products via a CT state.Since this scheme results in deactivation of 3M* without production of O2(ldg), the quantum yield of singlet oxygen will be reduced from unity (as observed for aromatic hydro~arbons5~) to as little as one-quarter if statistical proportions of triplet complex decay to ground-state products. Measurements of the fraction of sensitizer triplet states quenched by O2pzg-)that lead to 02(ld g) in benzene solution have shown a strong dependence on sensitizer.60 Excited-state sensitizers for singlet oxygen generation via energy transfer can be produced by light sources (continuous or pulsed) or by electron beams.The former directly excite the sensitizer states; the latter populate solvent excited states which can rapidly transfer their energy to sensitizer molecules.60 For quantitative studies of the kinetic properties of O2(ldg) in its reactions with added substrates the photosensitization technique has been the most widely and profitably used. 3 Determination of Singlet Oxygen Lifetime and Reactivity Parameters As described in Section 2 the most common means of production of 02(ldg) involves transfer of electronic energy from a sensitizer triplet state. This method has been used for virtually all kinetic and much product analysis work, the H202-OCI-system having frequently been employed as supporting evidence for the intermediacy of 02(ld Q).The sequence of photochemical events leading to the formation of this species and the channels for its subsequent decay are summarized in equations (7)-(14) for a sensitizer S,a substrate A, which reacts exclusively to give an adduct AO2, and a second quenching species Q. The determination of the rate constants for the last three steps involving O~(ld,), for a variety of media and a host of reactants, and the significance of such rate constants with respect to mechanism, have been the subject of a vast body of research which has appeared over the past fifteen years. s hv (7) @T (8) IS* -3S* (9) ks .3s* kS (10) 02(1dg)+ A kA + AO, (13) 02(lAs) + Q kg + loss of 02(1d@) (14) B.E. Algar and B. Stevens, J. Phys. Chem., 1970, 74, 3029. O0 A. A. Gorman, G. Lovering, and M. A. J. Rodgers, J. Am. Chem. SOC.,1978, 100,4527. 211 Singlet Molecular Oxygen A. Steady State Techniques.-In the absence of a quencher Q the quantum yield for production of A02 is given by equation (15)wheredh is the triplet yield on excitation of S. Experiments are most commonly carried out at constant oxygen concentration by bubbling and thus equation (16) will hold, where Cis a constant. QA0a-l = c(1+ kd/kA[A]) (1 6) Slope/intercept ratios for plots of Q~0g-lvs. [A]-l yield the ratios kd/kA, the so called /%values. In thepresenceofQ,equation(14), therelationshipofequation(17) will hold and values of kd/kA and k,/kA are readily abstracted from slope/ intercept treatments for varying [Q].6l If two substrates are known to react with O2(ldg) to give adducts, direct measurement of competitive rates of product formation, or less preferably loss of substrate, will yield relative rate constants for reaction.62-64 Thus, steady-state techniques only allow determination of rate constants for reaction of a species with Oz(ld g) relative to the rate constant for natural decay of the latter under the conditions of the particular experiment.Nevertheless a whole body of pioneering work based on the above approach made a significant advance towards an understanding of the factors which influence the reactivity of 02(ldu) towards organic/biological substrates.A distinct disadvantage concerned the fact that conclusions concerning solvent effects on reactivity were jeopardized by lack of knowledge of the variation of Oz(ld,) lifetime, i.e. of kd [equation (12)], with solvent. Solvent effects were therefore generally studied on the basis of product distributions rather than of rate constants. Thus the determination of 02(ld u) lifetimes using time-resolved techniques represented a major advance in 02(ldu) research. B. Time Resolved Techniques.-Early work had indicated that the O~(ld,) lifetime in common organic solvents was longer than 10ps65 and, it was generally assumed, relatively insensitive to solvent variation. This was shown not to be the case by experiments in which pulsed laser excitation of sensitizer molecules allowed rapid production of 02(ld8) via channels (7)-(11).The subsequent decay of this species could be witnessed by use of a suitable monitor. Diphenylisobenzofuran [DPBF (l)] absorbs intensely at 415 nm and reacts rapidly with 02(ld g) according to equation (18)to give a colourless intermediate (2) which collapses to pr0ducts.6~9 The time-resolved bleaching of DPBF 61 C. S. Foote, in Ref. 14, Chapter 5. 62 K. R. Kopecky and K. J. Reich, Can J. Chem., 1965, 43, 2265. 63 T. Wilson, J. Am. Chem. SOC.,1966, 88, 2898. 64 R. Higgins, C. S. Foote, and H. Cheng, Adv. Chem. Ser., 1968, No. 77, p. 102. 65 C. S. Foote and R. W. Denny, J. Am. Chem. SOC.,1968,90, 6233. 68 M. P. Stevens, F. Nahavandi, and F. Razmara, Tetrahedron Lett., 1973, 301.67 R. W. Murray and D. P. Higley, J. Am. Chem. SOC.,1974, 96, 3330. 21 2 Gorman and Rodgers (18) has been monitored at 415 nm after production of O2(ldg) via ruby-laser excitation of Methylene Blue as ~ensitizer.~*-~l It can be easily shown that, for a solution in which the initial 02(ld g) concentration is small compared with that of DPBF and of any quencher Q, the rate of loss of DPBF is given by equation (19) and a plot of In(& -Om) against time will be linear with a slope k' given by equation (20) where Dt and D, are respectively optical densities of DPBF at k' = kd + kA [DPBF] + kq[Q] (20) time t and at completion of the quenching event. The rate parameters kd, kA, and kq can be evaluated from plots of k' against (i) DPBF concentration (in the absence of quencher) for kd and kA and (ii) quencher concentration (at fixed DPBF concentration) for kg.It was shown71 that the O~(ld,)decay was inde- pendent of the sensitizer employed and this was confirmed in a later time- resolved study in which the pulse radiolysis technique was used to produce the sensitizer triplet states.60 It is generally accepted that carotenoid pigments play a role in the protection of biological systems and that their mode of action may well be Oz(ldg) quenching.65~72 The laser flash technique has been used to show that p-carotene quenches Oz(1d g) via collisional energy transfer [equation (21)J and that sub-sequent decay of triplet /%carotene proceeds via channels (22) and (23).73At sufficiently high carotene concentrations the carotene triplet formation is rate O1(ldg) + F-car kt o,(sz~-)+ 3p-caf* 3fl-car* + 02p&-) -P-car + Op(a&-) (21) (22) 38-car* -+8-car (23) limiting and therefore the rate constant for decay of 3P-car*, k', follows the decay of 02(ld 9) according to equation (24) in the presence of added quencher , k' = kd + kt [B-car] + kq [QI (24) P.B. Merkel and D. R. Kearns, Chem. Phys. Lett., 1971, 12, 120. as P. B. Merkel and D. R Kearns, J. Am. Chem. SOC.,1972,94, 1029. 70 P. B. Merkel and D. R. Kearns, J. Am. Chem. SOC.,1972,94, 7244. 71 D. R. Adams and F. Wilkinson, J. Chem. SOC.,Faraday Trans. 2, 1972, 68,586. C. S. Foote, Y. C. Chang and R. W. Denny, J. Am. Chem.SOC.,1970,92,5216. 73 A. Farmilo and F. Wilkinson, Photochem. Photobiol., 1973, 18, 447. 213 Singlet Molecular Oxygen Q.73t74 Extraction of the rate constants kd, ks, and kq is as for DPBF bleaching. It turns out that /%carotene is a diffusion-controlled quencher of Oz(ld g), kt = 1.3 x 1O1O 1 mol-l s-1,75 and this places its triplet energy at least slightly below that of Oz(ld,) (94.5 W mol-l; cf. ref. 76). The same has since been shown to be the case for a wide range of carotenoid substances.77.78 A pulse technique for O~(ld g) production using direct excitation with a Nd-YAG laser has been developed.79 Rate constants determined, usually for Freon-113 as solvent, are usually significantly lower than other reported values. The above work allowed determination of the absolute rate constants for interaction of Oz(ld g) with any molecule of interest which was soluble in organic solvents. In addition, determination of the Oz(ld 8) lifetime, 74, for specific solvents allowed all previously determined relative rate constants derived from /%values or competitive experiments to be placed on an absolute basis. C.Differentiation Between Physical and Chemical Quenching.-The quenching rate constants determined as described above may represent either a process which removes quencher (chemical quenching), one which leads to oxygen and quencher ground states (physical quenching), or a combination of both.The separation of the respective contributions generally involves experiments in which the amount of quencher actually lost is compared with the extent of removal of a substance known to quench Oz(ld,) exclusively via chemical reaction. The nature of particular quenching processes will be discussed in Section 4.D. Solvent Effects on ?-,.-The lifetime of Oz(ld,) is now known for a whole range of solvents.80 Values vary over a wide range from a few microseconds for water to a millisecond in perfluorinated hydrocarbons. In addition a significant deuterium isotope effect is observed, 74 usually, although not always, being appreciably longer in the deuteriated medium. It has been shown70 that the intensity of the solvent i.r. absorption near 7880, 6280, and 4700 cm-1, resonant with the O’-+O, 0’-+1 , and 0’+2 bands for the ld Z,-transition, correlates well with the rate constant for natural decay of Oa(ld,).In effect the electronic energy of the excited oxygen molecule is converted into solvent vibrational energy. In the 7880 and 6280 cm-l region the i.r. overtone bands of common solvents correspond to C-H and 0-H vibrations. Thus C-and 0-deuteriation normally lowers absorption intensity in this region, resulting in larger values of 7,. Of particular importance was the fact that, in agreement with theory, 74 74 A. Garner and F. Wilkinson,.in Ref. 75, p. 48. ‘Singlet Oxygen Reactions with Organic Compounds and Polymers’, ed. B. Ranby and J. F. Rabek, Wiley, N.Y., 1978. ‘16 W. G. Herkstroeter, J. Am. Chem. SOC.,1975, 97, 4161. ’I7 F. Wilkinson and W. T. Ho, Spectrosc.Lett., 1978, 11, 455. 78 M. A. J. Rodgers and A L. Bates, Photochem. Photobiol., 1980, 31, 533. 7B I. B. C. Matheson, B. C. Lee, B. S. Yamanashi, and M. L. Wolbarsht, J. Am. Chem. SOC., 1974,96, 3343. D. R. Kearns, in Ref. 14, Chapter 4. 214 Gorman and Rodgers was found to be ten times larger in D2O than in H20.81 It has recently been sh0wns~-~3that ‘74 in D2O is in fact 53 If: 3 ps, longer than that originally determined (20 ps) and larger than that in H2O (4 ps) by a factor of 13, in agreement with an earlier ratio based on steady-state-derived /3-values.84 Nevertheless, the D20/H20 lifetime difference previously observed81 has led to a powerful means of identification of 02(ld ,) in photobiological processes (Section 5A).Owing to expanding interest in the biological aspects of 02(ld g) reactivity the emphasis has shifted towards experimentation in aqueous systems and aqueous systems containing hydrophobic regions. A new water-soluble monitor for 02(ld 8), 9,lO-anthracenedipropionicacid has been introduceds5 which allows 74 and reactivity measurements for water-soluble quencher~.~~ Several papers have appeared concerning the reactivity of 02(ld ,) in aqueous micellar systems.S6- s9~s2Generally these systems are DzO based because of the more favourable ‘74 value and in such situations 02(ld,) behaves kinetically as if in pure DzO. Lipid-soluble substrates within the micelles exhibit rate constants for reaction with 02(ld8) that are typical of such reactions in organic media.E. Solvent Effects on Reactivity of 0z(ldg) with Organic Substrates.-In general rate constants for quenching of 02(ld,) by organic substrates are insensitive to solvent and this has been a key factor in mechanistic discussions. Individual effects or the lack of them will be commented upon in the following section. 4 Mechanism of Reaction of 02(ld g) with Organic Substrates It has been mentioned that the quenching of 02(ld 8) may be physical or chemical in nature. Two physical mechanisms, energy-transfer and charge-transfer quenching, have been well characterized and are discussed in the sections on carotenes, amines, and phenols. For many compounds with a r-system which reacts chemically with Oz(ld g) physical quenching is probably unimportant although in the majority of cases this is not established.A. Carotenoids and Other Energy-transfer Quenchers.-As already noted, carotenoids are particularly efficient quenchers of 02(ld 8). In several cases it has been clearly demonstrated that the mechanism involves essentially diffusion- controlled energy transfer as witnessed by the time resolved monitoring of carotenoid triplet formati~n.~~.~~ Clearly such a mechanism operates because 81 P. B. Merkel, R. Nilsson, and D. R. Kearns, J. Am. Chem. Soc., 1972, 94, 1030. 8z B. A. Lindig and M. A. J. Rodgers, J. Phys. Chem , 1979, 83, 1683. B. A. Lindig, M. A. J. Rodgers, and A. P. Schaap, J. Am. Chem. SOC.,1980, 102,5590. C. S.Foote, in Ref. 75, p. 135. A. P.Schaap, A. L. Thayer, K. A. Zaklika, and P. C. Valenti, J. Am. Chem. SOC.,1979, 101,4016. A. A. Gorman, G. Lovering, and M. A. J. Rodgers, Photochem. Photobiol., 1976, 23,299. A. A.Gorman and M. A. J. Rodgers, Chem. Phys. Lett., 1978, 55, 52. I. B. C. Matheson, A. D. King, and J. Lee, Chem. Phys. Lett., 1978, 55, 55. I. B. C. Matheson and R. Massoudi, J. Am. Chem. SOC.,1980, 102, 1942. 215 Singlet Motecular Oxygen these extended conjugated n-systems exhibit triplet energies close to or below that of Oz(ld g). Other compounds with extended chromophores, e.g. (3),69 76990 and metal complexes73~certain dye~,~O$ 91.92 are diffusion-controlled quenchers but direct evidence for energy transfer has not been obtained. Ph Ph B. Amin=.-Species that quench 02(1d @) via a charge-transfer process generally react chemically as well, to an extent depending on individual structure and reaction conditions.Amines typify such behaviour and the primary process is envisaged as formation of a complex between the electron-donating quencher and the electron-deficient oxygen species as shown in equation (25).93-97 The fact that quenching rate constants correlate with amine ionization potential~~~~~5 supports such a mechanism. The subsequent partitioning of the triplet complex pQ+ Oda4-1 Q + Oa('dc) +[Q. ..OJ1+ (25) IQ.*0213\ oa* determines the ratio of physical to chemical quenching. Formation of oxidation products requires an abstractable hydrogen a to nitrogen, N-methyl groups being particularly susceptible.98-100 Otherwise only physical quenching is observed.This is the case for diazabicyclo-octane [DABCO (4)], chemical W. F. Smith, W. G. Herkstroeter, and K. L. Eddy, J. Am. Chem. SOC.,1975, 97, 2764. 91 F. Wilkinson, in Ref. 75, p. 27. 92 D. J. Carlsson, T. Suprunchuk, and D. M. Wiles, Can. J. Chem., 1974, 52, 3728. 93 C. Ouannes and T. Wilson, J. Am. Chem. SOC.,1968,90, 6258. s4 K. Furukawa and E. A. Ogryzlo, J. Photochem., 1972, 1, 163. 95 R. H. Young and R. L. Martin, J. Am. Chem. SOC.,1972,94, 5183. n6 R. H. Young and D. R. Brewer, in Ref. 75, p. 36. 97 E. A. Ogryzlo, in Ref. 75, p. 17. J E. Lindner, H. J Kuhn, and K. Gollnick, Tetrahedron Lett. 1972, 1705. 99 M. H. Fisch, J. C. Gramain, and J. A. Olesen, Chem.Commun., 1971, 663. looD. Bellus, H. Lind, and J. F. Wyatt, Chem. Commun., 1972, 1199. 216 Gorman and Rodgers reaction presumably being forbidden on steric grounds. Inhibition of reaction DABCO is commonly used as a means of identifying O2(ldg) intermediacy. C. Reaction with Simple Olefins: The Ene Reaction and Dioxetan Formation.- The Ene Reaction. The process whereby olefins possessing an allylic hydrogen react with O~(ld,) to form allylic hydroperoxides [equation (26)] has been a subject of major interest and controversy for 15 year~.3~~~*J~l Despite an increas- ing rate of appearance of papers, both experimental and theoretical in nature, no definitive mechanistic conclusions have been drawn. Early e~periments10~-~0~ demonstrated that the double bond migrates, thus eliminating the simple H-abstraction process of equation (27).At the present moment discussion centres around four of the mechanisms outlined in equation (28), which involve either a concerted six-centre transition state typical of the classical ene reaction (a), a perepoxide or closely related intermediate (b and c) or a biradical inter- mediate (d). General points concerning the reactivity of simple olefins towards 041d g) are as follows. Quenching rate constants, which correspond to chemical reaction only, are relatively low, 103-107 1 mol-l s-l lo1 compared with lO6-lO9 1 mol-1 s-1 for amines.61 Some correlation between rate constant logarithms and ionization potentials,l05 together with data from Hammet plots,l06 indicates the electrophilic character of Oz(ld g) in these reactions. Solvent effects are minimal.Activation energies for reactions in solution have only been reported in one paper,lo7 which contains discrepancies concerning rate constant data. Nevertheless it is clear that activation energies are very low (0-21 kJ mol-l) and 8-17 kJ mol-1 lower than for corresponding gas-phase reactions.31 Thus, in solution these processes exhibit low enthalpies of activation and highly negative entropies of activation. This, coupled with a high overall exotherm- lol K. Gollnick and H. J. Kuhn, in Ref. 14, Chapter 8. loaC. S. Foote, R.Wexler, and W. Ando, Tetrahedron Lett., 1965, 4111. lo3 K. Gollnick and G. 0. Schenck, Pure Appl. Chem., 1964, 9, 507. lo4A.Nickon and J. F. Bagli, J. Am. Chem. SOC.,1961, 83, 1498. lo5 D. R. Kearns, J. Am. Chem. SOC.,1969, 91, 6554. lo8C. S. Foote and R W. Denny, J. Am. Chem. SOC.,1971,93, 5162. lo’ E. Koch, Tetrahedron, 1968, 24, 6295. 217 Singlet Molecular Oxygen i \ 0-li“”/<1 X Y 9XY%Y X>=(Y c -+ (28)X Y 9-0-0-074 x Y +tXY 1 icity,lOl points to a highly ordered rate-determining transition state with little change in substrate structure. The fact that isotope effects are either very small or absentlo8-ll2 is therefore not surprising. It would be impossible to cover the vast amount of discussion and argument presented over the years on the basis of stereochemical and electronic effects.It suffices to say that experimenta- lo*K. R. Kopecky and J. H. Vander Sande, Can. J. Chem., 1972, 50,4034. logA. Nickon, V. T. Chaung, P. J. L. Daniels, R. W. Denny, J. 3. DiGiogio, J. Tsunetsugu,H. G. Vilhuber, and E. Werstiuk, J. Am. Chem. SOC.,1972, 94, 5517. 110 L. M. Stephenson, D. E. McClure, and P. K. Sysak, J. Am. Chem. SOC.,1973, 95, 7888. 111 M. B. Grdina, M. Orfanopoulos,andL. M. Stephenson,J. Am. Chem. SOC.,1979,101,3111. 118 M. Orfanopoulos and L. M. Stephenson, J. Am. Chem. SOC.,1980,102, 1417. 218 Gorman and Rodgers tion has not allowed definitive mechanistic conclusions to be drawn and the reader is referred to a comprehensive compilation of data.101 The advent of theoretical calculations in this area has in no way decreased mechanistic controversy.Orbital correlation diagrams,8 HOMO-LUMO analysis,l13 and both CND0/2 CP4 and MIND0/3115 calculations favour initial perepoxide formation. In contrast an ab inirio/thermochemical approach116 is claimed to rule out the perepoxide on energetic grounds and favour the biradical intermediate of equation (28d) with varying zwitterionic character dependent on solvent and substituents. Whereas others3v8 have ruled out such an intermediate on the basis of the absence of Markovnikov directing effects it was reasoned that the least-substituted biradicals of the type shown in equation (28d) will be the most stable.116 However, thermochemical and kinetic arguments against irreversible biradical formation,l17 which is necessary to explain the lack of olefin cis-trans isomerization, were not countered and it was emphasized that it was impossible to distinguish between the biradical and concerted mechan- isms.Of particular recent interest is work on 2-methoxynorbornene (5) in which products incorporating the solvent, methanol, clearly appear to provide evidence in favour of a zwitterionic intermediate, perepoxide or open form as shown in equation (29).l18 The substrate cannot, in this case, participate in allylic hydroperoxide formation for steric reasons and the authors' claim that this result supports cal~ulations~~~~ 115 in favour of perepoxide intermediacy in general does not seem justified. 113 S.Inagaki, H.Fujimoto, and K. Fukui, Chem. Lett., 1976, 749.11* S. Inagaki and K. Fukui, J. Am. Chem. SOC.,1975,97, 7480. 115 M. J. S. Dewar and W. Thiel, J. Am. Chem. SOC.,1975, 97, 3978. 116 L. B. Harding and W. A. Goddard, J. Am. Chem. SOC.,1980, 102, 439. 11' F.A.Litt and A. Nickon, Adv. Chem. Ser., 1968, No. 77, p. 118. 118 C. W. Jefford and C. G. Rimbault, J. Am. Chem. Soc., 1978,100, 6437. 219 Singlet Molecular Oxygen Finally it has been suggested119 that the reaction of 0z(ldg) with olefins involves irreversible formation of a w-or charge-transfer complex, the geometry of which, and therefore the stereochemical aspects of subsequent decay processes, is governed by frontier orbital interactions. However, whether the low rate constants for reaction of O@d g) with olefins, compared with other substrates, are compatible with irreversible complex formation is a matter of conjecture.Dioxetan Formation. Olefins in which the double bond possesses an electron- donating heteroatom, generally N, 0, or s, react with O2(ldg) with overall cleavage of the double bond according to equation (30) and it is known that reaction proceeds via a dioxetan, the exact origin of which is as much a matter of controversy as that of the allylic hydroperoxide. Activated substrates undergoing dioxetan formation quench Oz(ld g) with rate constants somewhat higher gener- ally than for the simple olefins although no clear-cut correlation between rate constants and ionization potentials exists.120 As for the ene process, activation energies are very 10w107 and solvent effects are minima1.121J22 The two processes are clearly closely related and indeed may compete where an allylic hydrogen is a~ailable.~~3J~4Discussion concerning the mechanism of dioxetan formation centres around the intermediates of equation (28), which could be common to both ene and dioxetan reactions, and the discrete concerted four-centre process of equation (28e).A concerted ,2s + ,2s cycloaddition is forbidden but would become allowed when the olefin HOMO is higher in energy than the O~(ld,) LUM0.8J25 However, both the ,28 + ,2s addition, which experiment 123,126 shows must be antarafacial with respect to oxygen as shown in equation (28e), and perepoxide formation are In contrast to previously mentioned calculations on simple olefins114J15 the MIND0/3 treatment of vinylamine and 2,3-dihydropyranll5 reactivity towards 02(ld ff)indicates rate-limiting formation of an open zwitterion as in equation (28c).This may rearrange to dioxetan or perepoxide, the latter of which may participate in the ene reaction, if this is sterically possible. The ab initiolthermochemical approach,lls which suggests the 119 L.M.Stephenson, Tetrahedron Lett., 1980, 21,1005. lao G.R. Faler, Ph.D. Dissertation, Wayne State University, Detroit, Michigan, 1977. 121 A. P.Schaap, Ph.D. Dissertation, Harvard University, Cambridge, Massachusetts, 1970. lza C. S. Foote, A. A. Dzakpasu, and J. W.-P. Lin, Tetrahedron Lett., 1975, 1247. la3 P. D. Bartlett and A. P. Schaap, J. Am. Chem. SOC.,1970, 92,3223.124 A. P. Schaap and P. D. Bartlett, J Am Chem. Sot. 1970, 92,6055. la5 C.S.Foote, Pure Appf. Chem., 1971, 27,635. lZ6 G.Rio and J. Berthelot, Bull. Sac. Chirn. Fr., 1971, 3555. Gorman and Rodgers biradical as an intermediate common to both ene and dioxetan reactions as in equation (28d), has already been discussed. Several dioxetans have now been isolated127-131 using low-temperature techniques. All decompose with the emission of light at elevated temperatures according to equation (30). This supports the original proposal that breakdown of a dioxetan in a concerted symmetry-controlled process should give one ground- state carbonyl fragment and one electronically excited singlet state.132 However, considerable work on such processes has now shown that by far the major excited product is the carbonyl triplet state.133 A proposal 134 that the forbidden spin inversion is an integral part of the concerted cleavage has received support from MIND0/3 calculations135 which suggest that, for the parent dioxetan, this process is favoured over that involving an intermediate biradicaloid, which is in turn better than the standard cycloreversion to give singlet states, one electronically excited.In contrast the activation parameters for dioxetan decompositions are in excellent agreement with thermochemical calculations based on the biradical mechanism.136-139 Intersystem crossing would then take place within the biradical as shown in equation (3 1).Additional support for this mechanism comes from CND0/2l409141 and CNDO/S142 calculations.The preparation, properties, and reactivities of dioxetans have been comprehensively re~iewed.1~3 D. 1,4-Addition to m-systems: Endoperoxide Formation.-The earliest report 12' S. Mazur and C. S. Foote, J. Am. Chem. SOC.,1970, 92, 3225. lZ8 A. P. Schaap, Tetrahedron Lett., 1971, 1757. lasT. Wilson, D. E. Golan, M. S. Harris, and A. L. Baumstark, J. Am. Chem. SOC.,1976.98, 1086. 130 K. A. Zaklika, P. A. Burns, and A. P. Schaap, J. Am. Chem. SOC.,1978, 100, 318. 131 K.A. Zaklika, A. L. Thayer, and A. P. Schaap, J. Am. Chem. SOC., 1978, 100,4916. 132 F. McCapra, Chem. Commun., 1968, 155. 133 A. P. Schaap and K. A. Zaklika, in Ref. 14, Chapter 4. 134 N. J. Turro and P. Lechtken, J.Am. Chem. Soc., 1973, 95, 264. 135 M. J. S. Dewar and S. Kirschner, J. Am. Chem. SOC.,1974, 96, 7578. 136 H. E. O'Neal and W. H. Richardson, J. Am. Chem. SOC.,1970, 92, 6.553. 13' W. H. Richardson, M. €3. Yelvington, and H. E. O'Neal, J. Am. Gem. SOC.,1972, 94, 1619. 138 W.H.Richardson, F. C. Montgomery, M. B. Yelvington, and H. E. O'Neal, J. Am. Chem. SOC.,1974, 96, 7525. 139 K. R. Kopecky, J. E. Filby, C. Mumford, P. A. Lockwood, and J.-Y. Ding, Can.J. Chem., 1975, 53, 1103. 140 E.M.Eveleth and G. Felar, Chem. Phys. Lett., 1973, 22, 499. 141 G.Barnett, Can. J. Chem., 1974, 52, 3837. 142 D. R. Roberts, Chem. Commim., 1974, 683. 143 P. D. Bartlett and M. E. Landis, in Ref. 14, Chapter 7. 221 3 Singlet Molecular Oxygen of photo-~xidationl~~was the result of 1,4-addition of 04ld g) to a conjugated 77-system, that of naphthacene as shown in equation (32), although the structure of the endoperoxide was not established until much later.145 The first such product actually isolated was that formed by dye-sensitized addition of oxygen to the plant hormone ergosterol according to equation (33),146f147 and sub- sequent work extended these results to simple 1,3-dienes.145 It is now well documented that such reactions typically take place between 02(ld8) and conjugated 77-systems in general, acyclic, cyclic, aromatic, heteroaromatic, etc. The initially formed endoperoxides, either spontaneously or at elevated tempera- tures, may undergo fragmentation back to reactants as indicated by equation (32) or rearrange to products formed by homolytic fission of the 0-0 bond.Such products are often epoxides or further rearrangement products thereof. Fragmentation to give back Oz(ld g) is most typical of polyaromatic substrates, in particular those with phenyl groups attached to the 1- and 4-positions of the diene unit. For instance, the thermal decomposition of 9,lO-diphenylanthracene 9.10-peroxide in benzene, has been utilized as a high-yield chemical source of 02(ldg).148 Transfer of the latter species from 2,5-diphenylfuran endoperoxide has also been dem~nstrated.~~ The endoperoxides were previously ac~epted~J~~ as the products of a sym- metry-allowed cycloaddition process of the Diels-Alder type, equation (34a).The latter reaction is itself now the subject of considerable contro\ersy,150 said to proceed via a highly unsymmetrical biradicaloid species according to 144 M. Fritzche, C.R. Hebd. Seances Acad. Sci.,1867, 64, 1035. 145 K. Gollnick and G. 0.Schenck, in ‘1,4-Cycloaddition Reactions’, ed. J. Hamer, Academic Press, New York, 1967. 146 A. Windaus and J. Brunken, Liebigs Ann. Chem., 1928, 460, 225. 14’ W. Bergman and M. J. McLean, Chem. Rev., 1941, 28, 367. lP8 H. H. Wasserman and D. L. Larsen, J. Chem. SOC.,Chem. Commun., 1972, 253. log A. U. Kahn and D. R. Kearns, Adv. Chem. Ser., 1968, No. 77, p. 143. l50 P. Cararnella, K. N. Houk, and H. C. Domelsmith, J. Am. Chem. Sac., 1977, 99, 451 1. Gorman and Rodgers MIND0/3151 and via a synchronous reaction according to two a6 iititio cal-~ulations.~5~J5~Recent MIND0/3 data154 suggest that the reaction of butadiene with Oz(1d ,) yields the endoperoxide via a perepoxide, equation (34b).It has recently been shown155 that the reactions of Oz(ld,) with a series of furans in toluene exhibit AH+.values of zero and dS* values of -80 to -122 J K-1 mol-1. The highly negative entropies of activation are typical of Diels- Alder reactions156 and this fact, together with the absence of significant solvent effects,l57 appears to favour the concerted process. However, as for the ene- and dioxetan-forming reactions one cannot exclude perepoxide formation via an early transition state. g + OZ('AS) E.Phenols.-In contrast to the polyaromatics, phenols quench Oz(ld,) via a combination of physical and chemical processes.It has been shown that for a series of 2,4,6-trisubstituted phenols the quenching rate constant logarithms are a linear function of the phenol half-wave oxidation potentials15s and that both phenols and the corresponding ethers fit the same plot.159 This together with the direct observation of a phenoxy-radical led to the mechanistic proposal sum- marized in equation (39.158 The partitioning of the phenoxy radical between steps a and b determines the make up of the overall quenching. Although this mechanism appears of general applicability it has been suggested that in the case of highly activated phenols 1,4-~ycloaddition to give the endoperoxide is the primary step as in equation (36).160 The fastest phenolic quencher of Oz(ld ,) is a-tocopherol [Vitamin E (6)], which plays an important biological role in the 151 M.J. S. Dewar, A. C. Griffin, and S. Kirschner, J. Am. Chem. SOC.,1974, 96, 6225. lS8 R. E. Townsend, G. Ramunni, G. Segal, W. J. Hehre, and L. Salem, J. Am. Chem. SOC., 1976, 98,2190. 15s L. A. Burke, G. Leroy, and M. Sana, Theor. Chim. Acta, 1975, 40, 313. 154 M. J. S. Dewar and W. Thiel, J. Am. Chem. Soc., 1977, 99, 2338. lS5 A. A. Gorman, G. Lovering, and M. A. J. Rodgers, J. Am. Chem. SOC.,1979,101, 3050. 156 A. Wasserman, 'Diels-Alder Reactions', Elsevier, Amsterdam, 1965. 15' R. H. Young, K. Wehrly, and R. L. Martin, J. Am. Chem. SOC.,1971, 93, 5774. 158 M. J. Thomas and C.S. Foote, Photochem. Photobiol., 1978, 27,683. 159 I. Saito, M. Imuta, and T. Matsuura, Tetrahedron, 1972, 28, 5307. 180 I. Saito, N. Yoshimura, T. Arai, A. Nishinaga, and T. Matsuura, Tetrahedron, 1972, 28, 5131. 223 Singlet Molecular Oxygen R R -Qo; OMe protection of tissue and lipids from oxidation damage. Physical quenching predominates over chemical reaction by factors of 13.5,161 and 120163 in methanol, benzene, and pyridine respectively. The products of chemical reaction on proflavin-sensitized oxidation in methanol are summarized in equation (37).'64 F. Heterocycles.-The importance of O@d ,J with respect to biological damage has stimulated research into the reactivity of this species with heterocycles of every type.The majority of work has been based on product analysis and generally the extent of physical as opposed to chemical quenching is unknown. It would appear, however, that the simpler heterocycles are predominantly chemical quenchers and this has been shown to be true for DPBF (l)mJs5 and some furans. 155366 In the majority of cases the products of chemical reactions may be rationalized in terms of initial formation of an endoperoxide (Section 4D), otherwise of a dioxetan or corresponding open zwitterion (Section 4C). Isolated products are usually the result of rearrangement involving cleavage of the 0-0 bond, often with participation of an alcoholic solvent. Products may be isolated which correspond stoicheiometrically to two molecules of substrate and one molecule of oxygen.These possibly result from either the presence of water lel C. S. Foote, T.-Y. Ching, and G. G. Geller, Photochem. Photobiol., 1974, 20, 511. leP B. Stevens, R. D. Small, and S. R. Perez, Photochem. Photobiol., 1974, 20, 515. leaS. R. Fahrenholz, F. H. Doleiden, A. M. Trozzolo, and A, A. Lamola, Photochem. Photobiol., 1974, 20, 505. l'' G. W. Grams, E. Eskins, and G. E. Inglett, J. Am. Chem. SOC.,1972,94, 886. le6 p. B. Merkel and D. R. Kearns, J. Am. Chem. SOC.,1975, 97,462. I*' Y. Usui and K. Kamogawa, Photochem. Photobiol., 1974,19,245. Gormun and Rodgers -+ 02pdg) HO 0 37)J/ in the solvent, effectively replacing -0-C-OOH by -O-C-OH,167 by reaction of the initial adduct or some rearranged form thereof with a second substrate molecule, or by reaction of two molecules of adduct with concomitant loss of 02(ld g).67 Such overall behaviour is typified by aryl-substituted furans, several endoperoxides of which have been isolated and their reactivity exam- ined.3QJss-170 Equation (38) exemplifies the product dependence on reaction ccmditions.16s~171 Much of the interest in the reactivity of pyrroles towards Oz(1d g) stems from the observation that neonatal jaundice may be treated by exposure of the infant to visible light.172 It is thought that such phototherapy depends on the ability of 04ld g) to degrade the pigment bilirubin (7) to water-soluble products which may be excreted.173 The products of sensitized photo-oxidation of bilirubin in methanol are biliverdin (8) and the fragmentation products (9)-(12).174-176 g), aboutBoth bilirubin and biliverdin are efficient quenchers of O~(ld ld7 C.S.Foote, M.T. Wuesthoff, S. Wexler, I. G. Burstain, R. Denny, G. 0. Schenck, and K. J. Schulte-Elte, Tetrahedron, 1967, 23,2583. 168 R. E. Lutz, W. J. Welstead, R. G. Bass, and J. I. Dak, J. Org. Chem., 1962, 27, 11 1 1. C. Dufraise, G. Rio, and A. Ranjon, C.R. Hebd. Seances Acad. Sci. Ser. C, 1967,264,516. 170 C. Dufraise and S. Ecary, C.R. Hebd. Seances Acad. Sci., 1946, 223,735. 171 H.H.Wasserman and A. Liberles, J. Am. Chem. SOC., 1960, 82,2086. 17a R. J. Cremer, P. W. Perriman, and D. H. Richards, Lancet, 1958, 1, 1094. lT3 R.Schmid, N. Engl. J. Med., 1971, 285,520. 17* D.A.Lightner, D. C. Crandall, S. Gertler, and G. B. Quistad, FEBS Lett., 1973, 30,309. 175 A.F. McDonagh, Biochem. Biophys. Res. Commun., 1971, 44, 1306. 170 R. Bonnet and J. C. M. Stewart, J. Chem. SOC.,Perkin Trans. I, 1975, 224. Singlet Molecular Oxygen Ph Me000 Pha:h OMe Ph Ph Ph Ph an order of magnitude down on diffusion control. Whereas chemical reaction makes a ten percent contribution in the case of bilirubin it is much less significant for biliverdin.177J78 The latter is a product of bilirubin oxidation but does not lie on the principal degradative pathway. Compounds (9) and (10) result from cleavage of dioxetans formed by 02(ld g) addition to the double bonds exocyclic to rings A and D whereas (11) and (12) are the consequences of endoperoxide formation at rings B and c respectively.The latter reaction is strictly analo- gous to the dealkylation process observed during similar treatment of a simple tetra-substituted pyrrole as shown in equation (39).17Q The reactivity of indoles towards 02(ld ,J is essentially characteristic of electron-rich olefins (Section 4C), the principal products at room temperature being those resulting from dioxetan fragmentation as in equation (ma). Much recent work,180-182 in which alcohol and amine trapping products have been formed at low temperature, has been claimed to establish the primary formation of the open zwitterionic species of equation (40b) and therefore to support previously mentioned MIND0/3 calculation^^^^ (Section 4C).However, in C. S. Foote and T.-Y. Ching, J. Am. Chem. SOC.,1975, 97, 6209. 178 B. Stevens and R. D. Small, Photochem. Photobiol., 1976, 23, 33. ITS D. A. Lightner and G. B. Quistad, Angew. Chem. Int. Ed. Eng., 1972, 11, 215. 180 1. Saito, M. Imuta, Y. Takahashi, S. Matsugo, and T. Matsuura, J. Am. Chem. SOC., 1977, 99, 2005. ln1 I. Saito, S. Matsugo, and T. Matsuura, J. Am. Chem. SOC., 1979, 101, 7332. la*M. Nakagawa, S. Kato, S. Kataoka, and T. Hino, J. Am. Chem. SOC.,1979, 101, 3136. Gorman and Rodgers H 02C CO2H HO2C CO2H (7) via dioxetanesI \ via endoperoxides\ HO2C CO2H OHC + + Singlet Molecular Oxygen Me P IMe H /b the authors' opinion no evidence presented eliminates a situation in which the primary step is concerted formation of a dioxetan, the retrocyclization of which no longer competes with solvent-induced processes at low temperature.The n.m.r. characterization of dioxetan (13), formed by photo-oxidation of 2-t-butyl-l,3-dimethylindoleat -78 "C in CFC13l83 certainly does not argue against such a possibility. It has recently been sh0wnl5~ that the reactions of a series of indoles with 0z(ldy) in toluene exhibit AH* values of zero and dS* values in the range -97 to -143 J K-l mol-l, certainly in agreement with the Me Me highly ordered transition-state requirement of concerted dioxetan formation. The examples referred to in this section have been chosen to typify Oz(ld,) reactivity towards heterocycles. A recent excellent review184 provides a much more detailed survey.5 Biochemical and Biological Significance of Singlet Oxygen Oxidation processes are of paramount importance in cellular systems. Many aspects of metabolic change rely on oxidation steps for their proper functions; other oxidation events induced by exogenous or adventitious initiators can be severely damaging to organisms. With the experimental realization that Oz(ld p) Ins I. Saito, S. Matsugo, and T. Matsuura, J. Am. Chem. SOC.,1979, 101, 4757. ln4 H. H. Wasserman and B. H. Lipshutz, in Ref. 14, Chapter 9. Gorman and Rodgers has a significant lifetime and is capable of undergoing a number of oxidative reactions with the chemical components of biological assemblies, many observers of biological change have attempted to rationalize their data in terms of singlet oxygen mechanisms.It is possible to categorize the effects under two major headings : (1) Oxidations induced by the presence of light, a sensitizer (extrinsic or intrinsic), and oxygen-Photodynamic Effects. (2) Naturally occurring metabolic processes that proceed via oxidation without assistance from radiant energy. These will be considered in order. A. Photodynamic Effects.-An enormous volume of published research is concerned with these effects both at the biochemical and biological level. Excellent recent reviews are a~ailable.l~J~~-~~~ Requirements are the presence of a photosensitizer, light, and oxygen, all of which are, of course, important for the photogeneration of singlet oxygen. These three requirements are not a sufficient criterion of singlet oxygen intervention since the excited sensitizer molecule may react directly with the substrate (atom abstraction, electron transfer, etc.) subsequent to which 02 (ground state) can react with the modified substrate to produce long-term change.Such mechanisms are referred to as Type I; mechan-isms involving 02(ld8) as a demonstrated intermediate are Type 11. Additional criteria have been developed for demonstrating Type I1 behaviour. These usually are (a) enhancement of the effect in a D2O-based system and (6)prevention of the effect in the presence of recognized singlet oxygen quenchers, azide ions, DABCO etc. (see Sections 3D and 4B).Photo-oxidations have been studied at every level from small molecules in aqueous solutions to the sensitization of multicellular animals to damage or death. A comprehensive reviewla6 covers most recorded work an a variety of systems and rather than attempt to cover the multitude of published data hereit is better to concentrate on a few examples that have been especially well documented. Erythropoietic protoporphyria (EPP) is a disorder which is brought on in susceptible patients by exposure to sunlight and is characterized by swelling, erythema, and lesions. The photosensitivity of EPP patients can be reduced by oral administration of P-carotene.la8 The red blood cells of EPP patients contain large amounts of free protoporphyrin [a visible-light chromophore which efficiently sensitizes Oz(1d g)].These cells are haemolysed upon irradia- tionl*gJgO with visible light via photo-oxidation of membrane components. lS6 N. I. Krinsky, in ‘The Survival of Vegetative Organisms’, ed. T. G. R. Gray and J. R. Postgate, Cambridge University Press, London, 1976, pp. 209-239. lS6 J. D. Spikes, in ‘The Science of Photobiology’, ed. K. C. Smith, Plenum Press, New York, 1977, Chapter 4. J. D. Spikes and R. Straight, in ‘Oxygen and Oxy-Radicals in Chemistry and Biology’, ed. M. A. J. Rodgers and E. L. Powers, Academic Press, New York, 1981. M. M. Mathews-Roth, M. A. Pathak, T. B. Fitzpatrick, L. C. Harber, and E. H. Kass, N. Engl. J. Med., 1970, 282, 1231. L. C. Harber, A. S. Fleischer, and R.L. Baer, J. Am. Med. A~soc.,1964, 189, 191. lBoB. D. Goldstein and L. C. Harber, J. Clin. Invesf., 1972, 51, 892. Singlet Molecular Oxygen Incubation of EPP red cells with Vitamin E tocopherol)^^^ and derivativeslgO affords protection from haemolysis. Further, it has been shown192 that 3~-hydroxy-5a-hydroperoxy-d%holestene, the product of 04ld 8) attack on cholester01,~~is produced on irradiation of aqueous suspensions of normal red blood cell ghosts incorporating protoporphyrin. It has been concluded that photo-oxidation of cholesterol, initiated by 02(ld g), leads eventually to break- down of erythrocyte membranes in EPP patients. These conclusions are sub- stantiated by observations that membrane damage to egg-lecithin liposomes has been shown to arise from singlet oxygen processes.193 Similar processes appear to be involved in damage to membrane lipids in certain viruses.194 Inactivation and genetic changes in Saccharomyces cervisiae have been shown195Jgs to be photosensitized by xanthene, thiazene, and acridine dyes. These effects, which have been attributed to 02(ldg) initiation are D2O-enhanced and quenched by azide.l95,196 At the biochemical level there is clear evidence that some amino-acids, part- icularly histidine, methionine, tyrosine, and tryptophan, are photo-oxidized via a singlet oxygen mechanism.12J97 Leading on from this the deactivation of the enzymes alcohol dehydrogenase and trypsin under photodynamic conditions is enhanced in D2O and quenched by NaN3.lg8 Again the fairly clear conclusion can be reached that 04ld 8) is involved.Similarly the photodynamic oxidation of tryptophan residues in several enzymes (lysozyme, papain) has been attributed to singlet oxygen reaction.199-201 In all biological photo-oxidations examined so far in which 02(ld Q) has been invoked as a precursor to lethal or sub-lethal damage, nowhere has its inter- vention been directly demonstrated. All the evidence relies on the correlation of oxidation products with those of established singlet oxygen reactions, on the effects of D2O and small molecule quenchers, and on the identification of specific reaction products from added reactive substrates. It is to be hoped that continued research effort in biological and model systems will close the gap and offset the need for the long extrapolation from chemical properties to biological effect.B. Metabolic Events where 02(ld 8) is Invoked.-The microbicidal activity of phagocytes and the action of some oxidase enzymes have been examined for any possible role played by singlet oxygen. A thorough review of these aspects lB1G. D. Ludwig, D. Bilheimer, and L. Iverson, Clin. Res, 1967, 15, 284. lB1A. A. Lamola, T. Yamane, and A. M. Trozzolo, Science, 1973, 179, 1131. leaS. M. Anderson, N. I. Krinsky, M. J. Stone, and D. C. Clagett, Photochem. Photobiol., 1974, 20, 65. lB4W. Snipes, G. Keller, J. Woog, T. Vickroy, R. Deering, and A. Keith, Photochem. Photobiol., 1979, 29, 785. le6 T. Ito and K. Kobayashi, Photochem.Photobiol., 1977, 26, 581. lB6T. Ito, Photochem. Photobiol., 1977, 25, 47. lB7L. I. Grossweiner and A. G. Kepka, Photochem Photobiol., 1972, 16, 305. lo*R. Nilsson and D. R. Kearns, Photochem. Photobiol., 1973, 17, 65. leBA. G. Kepka and L. I. Grossweiner, Photochem. Photobiol., 1973, 18,49. looG. Jori, G. Galiazzo, and 0. BUSO,Arch. Biochem. Biophys., 1973, 158, 116. G. Jori, M. Folin, G. Gennari, G. Galiazzo, and 0. BUSO,Photochem. Photobiol., 1974, 19, 419. Gorman and Rodgers was published recently.202 The microbicidal action of polymorphonuclear leukocytes depends on (i) recognition of the invading organism, (ii) enclosure into a membrane-bound vesicle (phagosome), and (iii) fusion with enzyme- containing granules.During the enclosure and fusion steps 02.-and H202 are released and it has been speculated that Oz(ld g) is also involved. This latter was specifically examined203 by incubating human polymorphonuclear leukocytes with two strains of the coccus Sarcina Zeutea, one containing carotenoid and the other a mutant without carotenoid. The wild strain showed no significant killing, whereas the mutant was rapidly killed under the experimental conditions. Parallel experiments with Toluidine Blue as photosensitizer showed the same type of effect, i.e. protection by carotenoid. It was inferred that protection against bacteriocidal action parallels that against photodynamic killing, which is consistent with the proposal that singlet oxygen may be a microbicidal agent in leukocytes.Suggestions that oxidase-substrate systems generate 02(ld g) stem from observations of chemiluminescence during the enzyme-substrate reaction.46 Further experimental work during the past decade has failed47p204-206 to show conclusively that singlet oxygen is formed. The most that can be stated is that 02(ldg) production does not occupy a major role in oxidase activity. Many aspects of oxidase and other enzyme biochemistry are presented in a 1979 review.202 N. 1. Krinsky, in Ref. 14, Chapter 12. '03 N. I. Krinsky, Science, 1974, 186, 363. z04 T. C. Pederson and S. D. Aust, Biochem. Biophys. Res. Commun., 1973, 52, 1071. *06 E. W. Kellog and 1. Fridovich, J. Biol. Chem., 1975, 250, 8812. Io6 K.-L. Fong,P. B. McCoy, .I.L. Poyer, H. P. Misra, and B. B. Keele, Chem. Biol. Interact., 1976, 15, 77. 231

ISSN:0306-0012    

DOI:10.1039/CS9811000205

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

CHEMICAL ASPECTS OF TRACE CONSTITUENTS OF THEDIET* I Control and Surveillance of Trace Constituents -Is There a Need? By D. G. Lindsay FOOD SCIENCE DIVISION, MINISTRY OF AGRICULTURE, FISHERIES AND FOOD, GREAT WESTMINSTER HOUSE, LONDON SWlP 2AE 1 Introduction Although the incidence and levels of certain pesticide residues in food in the United Kingdom had been investigated in surveys carried out prior to 1970, it was not until this time that a systematic programme of examining food for residues of potentially toxic substances was developed. The need for a national programme of food surveillance was realized after the discovery that methyl- mercury compounds were present in a range of species of fish. Because methyl- mercury had been shown to be the causative agent in the aetiology of Minamata disease, it was necessary to obtain data as quickly as possible to assess the possible health significance to fish consumers.Since this time, support for a nationally co-ordinated programme of food monitoring and surveillance has continued to grow as more and more use has been made of the data that have been generated. From a comparison of the incidence of chronic disease amongst populations and of the changes observed in the patterns of disease which occur when these populations resettle in other parts of the world, there is good epidemiological evidence that the diet is one of the important factors which predispose individuals to chronic disease, particularly in the incidence of cancer.’ Whether the factors responsible for these effects are major or minor components of the diet is not known.If scientific progress is to be made to determine what in the diet are the compositional factors which may be the cause of disease, basic information on the chemical composition of food will be necessary to test any hypothesis. One of the most important objectives of a food surveillance programme is to attempt to obtain such data and to observe whether or not there are significant trends in the intakes of individual components of the diet amongst selected populations. These data may be used to undertake risk assessments, which in turn are the basis for deciding whether or not there should be controls on the amounts of the toxin present in food, and to decide on the most appropriate methods of control.The enormous advances in the analysis of trace substances in food which have *These papers were originally presented at a symposium on Chemical Aspects of Trace Constituents of the Diet’ organized by The Royal Society of Chemistry Food Chemistry Group and held at the Scientific Societies’ Lecture Theatre, Savile Row, London W1, on 15th October 1980. E. L. Wynder and G. B. Gori, J. Nut. Cancer Znst., 1977, 58, 825. 233 Chemical Aspects of Trace Constituents of the Diet. Part I occurred within the last decade have been in the application of improved g.c.- m.s., h.p.l.c., and a.a. techniques which have enabled individual compounds to be measured at the parts per 106-109 level.Much of the impetus for the development of these techniques came from the need to determine the presence -or absence -of chemicals which have been approved for the production or the manufacture of food, such as pesticides or therapeutic substances used in livestock production. The analytical techniques required for regulatory purposes have been developed to the point at which analytical sensitivity is such that very low levels of toxic substances can be detected in food -substances which are an inherent component of that food or which are adventitious contaminants of food and have not been subjected to any form of risk assessment procedure as have the majority of the chemicals which are deliberately added to food. 2 Risk Assessment We must start on the basis that no chemical is ever completely safe, but there are ways in which many of them may be used safely.This is the fundamental from which risk assessment commences, and in this connection no one has yet defined risk. Inevitably risk assessments for humans depend heavily upon interpretations of the results of administration of chemicals to animals, usually by incorporation in their diets, which are themselves inevitably man-made. Immediately, an unknown is introduced into the assessment since direct com- parisons between human effects and those observed in laboratory animals are rare. Nevertheless, for the last 25 years assessments of chemicals intended for human consumption as part of the total diet (such as food additives) have been made by utilizing a hypothetical acceptable daily intake (ADI) for the chemical, calculated from the level in the diet of a test animal demonstrated experimentally to induce no observable adverse effect during its life time, by the application of an arbitrary safety factor.It is pertinent to remark here also that no one has yet defined ‘observable adverse effect’, although experience would lead one to suppose that ‘statistically significant pathological variation from a parallel control group of animals not exposed to the chemical’ is the intended criterion. The safety factor used to divide into the maximum no-effect level also varies according to circum- stances, and in this area there are as many opinions as there are experts! Tradi- tionally the accepted figure has been 100, but this is sometimes increased when the quality of the data is not of the required standard or when the nature of the toxic effects at the higher dose levels gives cause for concern.The safety factor may be less than 100 (sometimes as low as 10) if the substance is a normal constituent of the human diet or a normal metabolite of the human body or where a substantial proportion of the available data is derived from experiments in man. Despite all the disadvantages and qualifications which one can apply to the ADI, and recognizing that modern toxicological thinking dictates a move away from ‘rigid protocol’, the concept may be said to have served us well for those areas of risk assessment related to the deliberate addition of chemicals to food to achieve necessary technological functions.It also enables comparisons, albeit Lindray crude ones, to be made between actual human intake (which can be derived from surveillance programmes) and the socially acceptable yardstick of safety represented by the ADI. Chemical substances which unintentionally become part of an individual’s food intake as ‘contaminants’ of food are considerably harder to assess in terms of human risk. Their quantities are frequently minute, their distribution in food is invariably random, their nature and chemical identity may well be unknown until the emergence of an unacceptable biological effect, and their origins in the food chain are frequently obscure. If one adds to this plethora of ‘unknowns’ the fact that many of these xenobiotics may be classified as of ‘natural’ origin, the presumption must be made that man has been exposed to them over countless generations. So what effects are we concerned with in any risk assessment of the uninten- tional food contaminants? In the present state of the toxicological art it is safe to assume that the battery of biological tests to which chemicals deliberately intended to be present in food must be subjected before they are used will eliminate any possibility of the induction of non-reversible effects, such as an increase in tumour incidence or effects resulting from the storage of the compound in body tissues.Contaminants of food either adventitious or naturally present do not have the privileges of precursive biological testing, except in the case of pesticide formulations and certain substances used in animal husbandry and therapeutics.This makes a judgement on the potential hazard difficult to make. As the amounts of such substances are usually very small, it is unlikely that their presence in food will lead to manifestations of acute toxicity. We are left therefore with the possibility that they might induce non-reversible effects, such as a carcinogenic response or mutagenicity, or that they might, in the long term, accumulate in body tissues. In the case of proximate electrophilic carcinogens which alkylate DNA bases (e.g. aflatoxin, vinyl chloride) and which act at the cellular level to produce heritable changes leading to malignant tumour forma- tion, it has not yet proved possible to demonstrate a threshold dose below which no effect occurs.The effect of decreasing the dosage is simply to increase the time to tumour formation. Attempts have been made to apply mathematical procedures to calculate the risks of exposure to low doses of carcinogens as, for example, the Mantel-Bryan procedure,2 which calculates a virtually safe dose, the one-hit model (and its modification by Weibull),3 which assumes that the dose-response curve is linear in the low-dose region, and the Armitage-Doll multi-stage modeL4 However, the choice of mathematical procedures for low-dose extrapolations has no firm biological basis and must, to some extent, be arbitrary.There is also no reason to suppose that, should such mathematical techniques be applied, an ‘acceptable’ or virtually safe dose for some carcinogens would result that would allow a N. Mantel and W. R. Bryan, J. Nat. Cancer Inst., 1961, 27, 455. D. G. Hoel, D. W. Gaylor, R. L. Kirshstein, V. Sattiotti, and M. A. Schneiderman, J. Toxicol. Environ. Health, 1975, 1, 133. P. Armitage and R. Doll, Proc. 4th Berkeley Symposium on Mathematical Statistics and Probability, ed. Lecam and Neyman, 1961, Vol. 4(4), p. 19. Chemical Aspects of' Trace Constituents of the Diet. Part I maximum level of a carcinogen in food to be above the limit of detection of present-day analytical techniques. 3 Policy for the Control of Adventitious Toxic Substances in Food Precisely because it is difficult to determine the risk associated with a chronic toxin which is present adventitiously in food, any policy which attempts to regulate the risk should not be so inflexible that no account is taken of this uncertainty or of the fact that the removal of a proportion of food from the market-place may achieve very little in reducing overall exposure in the long term.This is because the substance is rarely present in food with a frequency and at levels that can be represented by a normal distribution. Where the pattern of residues is represented by a highly skewed distribution, as is usual, the removal of a small proportion of the most highly contaminated food, even if this could be achieved by vigorous enforcement, will do little to reduce the long-term exposure.Such a policy might be more effective where there are sections of the com- munity exposed to higher than average levels of the substance in their diets, but rarely do nationally enforced controls need to be considered to deal with such situations. Should action be considered necessary, or prudent, it is far better to reduce exposure by dealing with the problem at source wherever this is practicable, consistent with available technology and economic cost. Such a policy has more often than not been adopted in the United Kingdom to limit the contamination of food and explains why there are not many regulations made under the Food and Drugs Act which limit the maximum amount of a toxic substance that is permissible in a food.Instead food surveillance can be used to provide data for risk assessment, to identify the source of the residue, and to investigate practical ways of minimizing exposure without the recourse to national legislation. Food surveillance programmes carried out since 1970 have defined the extent and degree to which a range of toxic substances are present in the United Kingdom food supply. The results of some of this work, together with the approaches which have been adopted to deal with the situation which has been revealed, are well illustrated with reference to some of the work which has been undertaken on the contamination of food by mycotoxins, toxic substances in food present as a result of manufacturing practices, and by certain industrial chemicals.4 Mycotoxins in Food Although there have been many reports in past history of the occurrence of acute ill-health, notably ergotism, resulting from the ingestion of mouldy food, it was not until the 1960s that there was an explosive interest in investigating the metabolites of fungi which naturally contaminate food. The impetus for this interest was the discovery that the aflatoxins, metabolites of a strain of the ubiquitous soil-borne fungus Aspergillus flaws, were highly carcinogenic. Since that time more than 100 compounds which could be classified as mycotoxins Lindsay have been reported in the literature, although to date few of these compounds have been shown to be present in food.Epidemiological studies indicate an association between the intake of afla-toxins and the incidence of primary liver-cell cancer. This is supported by data for at least eight species of experimental animals5 Consequently the levels and incidence of these metabolites in retail food supplies have been investigated in some detail. So far this work has established that the aflatoxins are detected primarily in nuts and nut products and in milk and dairy products. Table 1 summarizes the data which were obtained for peanuts imported into the United Kingdom between 1977 and 1978. Although these peanuts were intended for human consumption, extensive quality-control procedures and processing in the factories will have reduced consumer exposure considerably.Table 1 AfIatoxin levels in imported peanuts (1977-1978) Country of Number of’ Number ojsamples containing average levels of origin sumplesbanalysed aflatoxinn in stated rangelpg kg-l __-_ _-__ 0-5 5-30 30-100 100-300 >300 Brazil 2 0 0 2 0 0 Egypt 2 1 0 1 0 0 Gambia 5 0 3 1 0 1 India 35 18 4 8 4 1 Malawi 53 27 17 7 2 0 S. Africa 6 6 0 0 0 0 U.S.A. 56 43 9 3 1 0 a In most cases aflatoxin B, was the major component. b 20 kg samples taken from each 20 ton batch. Low levels of aflatoxin MI, a bovine metabolite of one of the most frequently encountered aflatoxins, which has also been demonstrated to be carcinogenic, were detected in a proportion of the milk and dairy products tested.The results obtained are summarized in Table 2. The likely origin of this contamination is Table 2 Range of aflatoxin MI levels in milk and dairy products (1978-1979) Samples Number of samples Runge of levels joundl examined pg kg-l Producer-retailer milk 278 < 0.03-0.52 Milk powder 203 < 0.104.80 U.K. cheese 223 < 0.10--0.40 Imported cheese 143 < 0.104.50 Whey powder 88 < 0.10--0.60 Environmental Health Criteria 1 1 : Mycotoxins, World Health Organisation, Geneva, 1979. Chemical Aspects of Trace Constituents of the Diet. Part I through the utilization of groundnut meal as a component of the compound rations fed to dairy cattle. The United Kingdom Feedingstuffs Regulations limit the maximum allowable level of aflatoxin B1 in ‘complementary’ feedingstuffs to 20 pg kg-l.In general the levels of aflatoxin MI in milk are 300-fold lower than the levels of the feed,6 which should limit the maximum concentration of aflatoxin in milk to about 0.1 pg 1-l. At one time this was the lower limit of detection of the method, but improvements in the anaIytical methodology presently give a lower limit of detection of 0.03 pg 1-1 for liquid milk and 0.1 Fg I-1 for milk products.’ The surveillance programme has also looked into the contamination of certain commodities by other mycotoxins. Ochratoxin A was detected in maize products in the range < 5-200 pg kg-1, in soya and soya products in the range < 50-500 pg kg-l, and in cocoa beans in the range < 100-500 pg kg-l.Ochratoxin A has been detected in animal products for human consumption in both Denmarks and S~eden,~ and in a limited survey carried out in the United Kingdom ochra- toxin A has been detected in pig meat and kidneys. A national survey of the incidence of ochratoxin A in barley and pig meat is under way at the present time. Because of the carcinogenic potency of the aflatoxins it is prudent to keep exposure to a minimum. The practical ways in which this can be achieved are being actively investigated. As far as nuts for human consumption are concerned, developments in quality control have been adopted by the industry, including the careful choice of raw materials through improved sampling and analytical techniques.More fundamental work, such as the development of aflatoxin-resistant strains of nuts, improved harvesting, and chemical preservation tech- niques, is being investigated on a world-wide basis. At the present time, however, it is impossible to ensure that all edible nuts are free of aflatoxin. The policy of attempting to minimize the problem requires a close working relationship between Industry and Government which we are fortunate in having. In investigating ways of keeping the contamination of milk to a minimum, options for control include the possibility of decontaminating oil-seed meal by a process such as ammoniation, avoidance of the use of oil-seed meal as a straight feed or in dairy rations, and the development of methods of sampling to detect contaminated batches.The feasibility of these and other options must be evaluated with regard to cost and practicability. As far as the contamination of food by ochratoxin A is concerned, there are still insufficient data on the incidence and levels of residues in the diet and of factors which contribute to the contamination, as well as insufficient toxicological data, to consider options for control. Nonetheless, because of the potential risks * J. V. Rodricks and L. Stoloff in ‘Mycotoxins in Human and Animal Health’, ed. J. V. Rodricks, C. W. Hesseltine, and M. A. Mehlman, Pathotox Publishers Inc., Illinois, 1977, p. 67.’D. S. P. Patterson, E. M. Glancy, and B. A. Roberts, Food Cosmet. Toxicol., 1978, 16,49. P. Krogh, Nord.Veterinaerrned., 1977, 29, 402. L. Rutqvist, N. E. Bjorklund, K. Hult, and S. Gatenbeck, Zentralbl. Veterinaermed., Reihe A, 1977,24,402. Lindsay associated with the presence of mycotoxins in food it is possible to investigate methods which could reduce, or eliminate, this potential, e.g. through the use of non-toxigenic strains of fungi in the manufacture of those foods, such as cheeses and fermented sausage, where mould ripening is a fundamental part of the production process. 5 Contamination of Food by Manufacturing Processes A particularly striking example of the use of surveillance to identify and resolve avoidable exposure to toxic substances has been the recent detection of nitro- sodimethylamine (NDMA) in beer as a result of the direct-fired kilning of malt.In 1978 it was reported that traces of NDMA were being found in beers produced in Germany, the Netherlands, and the U.S.A., and this contaminant was sub- sequently detected in beer in the United Kingdom. Levels were low, typically around 2 pg kg-l. In view of the potent carcinogenicity of NDMA and the amount of beer consumed in the United Kingdom (240 pints per head of popula- tion in 1979) the matter was viewed with some concern. During malt production the germinated barley is dried and heated to a moder- ately high temperature. In the United Kingdom the drying process is based on blowing hot air through the germinated barley over a period of 24-48 hours; in the majority of United Kingdom kilns the hot air is obtained by taking a suitable fuel, burning it, and diluting the combustion gases with an appropriate amount of air.In the early 1970s commercial supplies of natural gas became available and this was seen as offering significant advantages to the malting industries as a fuel. The cost was very attractive and as a low sulphur fuel it was environmentally ‘clean’ and offered the prospect of introducing energy recovery systems in the form of copper heat-exchangers in the exhaust air; with more conventional fuels these were not economically feasible because of corrosion problems arising from the sulphur oxides in the exhaust gases. With these advantages in mind, gas-fired kilns were introduced widely in the United Kingdom during the middle and late 1970s.The mechanism of NDMA production appears to involve a reaction sequence in which nitrogen oxides (NOx) in the combustion gases react with an amine in the ‘green’ malt and NDMA is released later in the ‘kilning’ cycle. The reaction is inhibited by sulphur dioxide, and it seems that quite fortuitously anthracite and ‘heavy’ oil contained sufficient sulphur to prevent NDMA formation when they were used as fuels in the past. But the change-over to natural gas also removed the S02;NDMA formation was no longer inhibited and in the worst cases levels of some hundreds of p.p.b. of NDMA were formed in malt. Happily, as a result of intensive collaboration between Government, maltsters, the Gas Board, and the burner manufacturers the problem has rapidly been brought under control.Within two years of the first recognition that a problem might possibly exist, the nature of the problem has been identified and measures have been devised which have dramatically reduced NDMA levels. New gas burners have been designed; these have lower flame temperatures and give low NOx concentrations in the combustion gases. These burners either alone or in Cheinical Aspects of Trace Constituents of the Diet. Part I combination with burning sulphur (or injecting SOZ)have brought NDMA levels down well below 10 pg kg-l in most cases; in kilns which previously produced malts with very high NDMA content it is common to find that the levels of NDMA have been reduced by factors of 30-100.There is also a real prospect that the identification of the role of NOx in combustion gases in determining NDMA levels in malt and beer will lead to the development of a novel catalytic burner with intrinsically low NOx characteristics which will find applications in processes extending far beyond the malt-kilning application which prompted its development. 6 Environmental Contaminants in Food With the exception of the gaseous pollutants, the exposure of the general popula- tion to industrial chemicals is primarily through food consumption. There are many well documented accounts of where food has been so heavily contaminated by the industrial discharge of chemicals, or through careless controls over the use of certain chemicals, that food consumption has led to chronic disease. The Minimata tragedy in Japan was first recognized in the 1950s and traced to the contamination of fish by methylmercury.10 Contamination of rice oil by poly- chlorinated biphenyls (PCBs) and polychlorinated dibenzofurans led to the outbreak of Yusho disease in Japan in 1968,*l and Itai-Itai disease has also been attributed in part to the contamination of rice through cadmium pollution.12 Less severe effects were observed in Turkey in 1955 when hundreds of people consumed grain treated with hexachlorobenzene and developed ~0rphyria.l~ More recently residents of Michigan were found to have been exposed to polybrominated biphenyls (PBBs) as a result of the accidental use of this fire- retardant agent in animal feed.Although effects were observed on lymphocyte production and function,l4 the short- and long-term implications of these differences are not known. 7 Methylmercury Levels in Fish Extensive monitoring of fish caught from United Kingdom coastal waters has been undertaken as part of the food surveillance programme since 1970. These data have demonstrated that average methylmercury levels in fish caught in the north-east Irish Sea were raised to about 0.3 mg kg-1 compared with levels of about 0.15 mg kg-1 in other areas. Before any assessment of the risks associated with the ingestion of this chronic toxin could be made, it was necessary to obtain information about the likely exposure of consumers to these raised levels since it is only this information which can be compared with the extensive toxicological data existing in both experimental animals and humans on the toxic effects of lo ‘Minamata Disease’, ed.M. Katsuna, Kumamoto University, Japan, 1968. l1 S. Katsuki, Fukuoka Acta Med., 1969, 60, 403. la ‘Cadmium in the Environment’, ed. L. Friberg, M. Piscator, and G. Nordberg, CRC Press, Cleveland, Ohio, U.S.A., 1971, p. 111. C. Cam and G. Nigogosyan, J. Am. Med. Assoc., 1963, 183, 88. l4 J. G. Bekesi, J. F. Holland, H. A. Anderson, A. S. Fishbein, W. Rom, M. S. Wolff, and I. .I. Selikoff, Science, 1978, 199, 1207. Linhay methylmercury compounds. This in turn requires detailed information about the food consumption habits of the exposed population, and these data are as important a part of the food surveillance programme as is information about the actual levels of a toxic substance present in the diet.Dietary intake studies have been organized in which consumers of fish taken from the north-east Irish Sea were compared with a control population. The study group was chosen from amongst the highest consumers of fish in each population, and the corresponding levels of mercury in whole blood and hair were measured. Consumers were found in both communities whose intake exceeded the W.H.0.-recommended tolerable weekly intake of methylmercury of 200 pg. Table 3 compares the data which were obtained. In comparison with the mercury levels in blood and hair assessed by a W.H.O. Task Forcels as likely to result in a slight increase in the earliest signs of methylmercury poisoning, namely parasthesia, the consumer with the highest intake of mercury was shown to be protected by at least an 8-fold factor of safety from the levels in blood and at least a 4-fold factor of safety from the levels in hair.16 Table 3 Intake of mercury and corresponding levels in whole blood and hair mwngst U.K.fishing communities North-east Control area Irish Sea Intake of mercury/pg week-1 range 5-443 4-560 (70 kg)-l mean 130a 94 Mercury in whole blood/pg range 0.04-2.58 0.04-1.21 (100 ml)-l mean 0.50a 0.35 Mercury levels in hair/pg g-l range 0.1-1 1.03 0.27-5.8 mean 2.03a 1.28 a Significant at 95 % level.Although it was judged that this increased exposure to methylmercury amongst north-east Irish Sea communities was not a health risk, these findings have resulted in a review of the level of discharges into the north-east Irish Sea with the ultimate aim of ensuring that fish caught in the north-east Irish Sea do not have significantly raised methylmercury levels in comparison with other areas of the United Kingdom.8 Polychlorinated Biphenyls in Food These compounds are some of the most persistent contaminants that are known. The fact that they are stable to acidic and basic hydrolysis and to heat and that they have a high dielectric constant led to their use in such applications as dielec- l6 Environmental Health Criteria 1 : Mercury, World Health Organisation, Geneva, 1976.l6 J. Haxton, D. G. Lindsay, J. S. Hislop, L. Salmon, E. J. Dixon, W. H. Evans, J. Reid, C. J. Hewitt, and D. S. Jeffries, Environ. Res., 1979, 18, 351. Chemical Aspects of Trace Constituents of the Diet. Part I trics and heat-exchange fluids. Although they have been used since the 1930s they were only detected in the environment in 1966 when developments in the analysis of the organochlorine pesticides revealed unexplained peaks. Their persistence has resulted in the contamination of food principally by three routes: (a)the direct contamination of food or animal feedstuffs by an industrial accident, (b) migration of PCBs from packaging into food, and (c) absorption from the environment by fish. Studies on the incidence and levels of contamination in fish by PCBs have shown detectable levels of PCBs in a large proportion of the fish 1anded.l’ Toxicological evaluation of the PCBs have shown that there is a direct rela- tionship between PCB exposure in laboratory animals and an increasing inci- dence of various sub-chronic and chronic toxic effects including adverse repro- ductive effects, tumour production, and possible carcinogenicity.l* The U.S.Food and Drug Administration have concluded that there are insufficient toxicological data with which to make an assessment of risk. Since it is not possible to derive a ‘no-effect’ level of intake it is impossible to set an acceptable level of intake. Because of the generally low levels of P CBs that are found in food and because there is no identifiabIe point source of discharge of PCBs in the United King- dom, it is not practicable to consider imposing controls over the levels of PCBs in food other than by controls over the manufacture, use, and disposal of the PCBs.As a consequence of the health and environmental effects of the PCBs their manufacture has ceased altogether in the U.S. and the United Kingdom, and their use is restricted solely to closed systems such as transformers and capacitors. However, it is expected that the levels of PCBs in food will not decline appreciably for many years owing to their continued release from old equipment in land-fills or because they are currently in use in equipment which was manufactured some years ago. 9 Miscellaneous Industrial Chemical Residues in Food Table 4 summarizes some of the classes of chemical compounds that have been detected in freshwater fish in U.S.monitoring programmes.19 These classes cover a Table 4 Classes of chemical compounds detected in freshwater fish in U.S.A.lg Aromatic amines Brominated aromatics Chlorinated aliphatics Chlorinated benzenes Chlorinated benzotrifluorides Chlorinated non-aromatic cyclics Chlorinated toluenes Triaryl phosphates l7 Ministry of Agriculture, Fisheries and Food, unpublished data. ‘Polychlorinated Biphenyls (PCBs) :Reduction of Tolerances’, US Food and Drug Adminis- tration, Fed. Register, Washington DC, 1979, Vol. 44, p. 38 330. P. Lombardo, Ann. N. Y. Acad. Sci., 1979, 320, 673.Lindsay very wide range of individual compounds, some of which may fall into the category of ‘recognized contaminants’. However, there are likely to be industrial compounds not yet recognized which are contaminants of food or potential contaminants of food. 10 Future Developments in Surveillance In part the problems of food contamination that have occurred have arisen as the result of the lack of any system of control over the manufacture, use, and disposal of industrial chemicals, and they have been the impetus for the introduction of registration schemes for the production of new and existing chemicals in a number of countries. The application of such schemes will no doubt be highly effective in minimizing food contamination problems, but they in turn will require the development of appropriate monitoring programmes, particularly as no registra- tion scheme can avoid the occurrence of accidents.All the monitoring work which has been carried out so far for environmental contaminants has concentrated on establishing the extent to which chemicals known to be stable in the environment are present in food. In order to avoid having to take arbitrary judgements on what is an acceptable level of intake of a chronic toxin, it would be better to consider the development of a system of monitoring which was able to detect at the very earliest stage changes in the pattern of residue distribution in samples of food which are sentinels of environ- mental contamination, e.g. fresh fish, milk, or meat or other foods which are not normally highly processed The analysis of suitable extracts of such foods by high-resolution capillary g.c.coupled to m.s. and computer storage facility would enable a ‘fingerprint’ of the output of the chromatographic system to be stored in terms of both retention time and peak intensity. The stored data could be compared with the output of the analysis of a number of similar samples taken from the same or different locations and any changes in the fingerprint observed. Significant alterations could be identified and the source controlled before the problem developed into a major issue. Although such an approach could detect compounds which are detectable by g.c. methods, this may well overlook classes of compounds which are not amenable to g.c.analysis. Those techniques which will have a great impact on surveillance programmes include developments in m.s. instrumentation such as pulsed positive- and negative-ion detection techniques, developments in selective detectors for g.c. and particularly 1.c. systems, and developments in computer system software capable of handling massive volumes of data of the type obtained from monitoring programmes. There is no doubt that had such a surveillance system been in operation widespread pollutants such as PCBs could have been detected long before they became ubiquitous. So far this approach has mainly been limited to the detection of unknown contaminants in water,20 but a number of laboratories in the U.S.are loJ. A. McFall, W. Y.Huang, and J. L. Laseter, Bull. Environ. Contam. Toxicol., 1979, 22, 80. Chemical Aspects of Trace Constituents of the Diet. Part I evaluating the possibility of applying this approach to food monitoring. The application of such techniques should enable us to develop from a situation where we are wise after an event into one where we become wise before an event, thereby minimizing the economic costs of environmental and industrial constraints. 11 Food Surveillance-Is There a Need? The results of some of the recent work of the food surveillance programme, which have been reviewed in this paper, serve to illustrate the wide range of different topics that have been covered so far and have come to light through improvement in analytical methodology.This systematic collection of data on the levels of toxic substances in food has: (a) enabled assessment to be made of the health significance associated with the consumption of a particular chemical residue in the diet, (6) highlighted localized problems of food contamination, (c) shown the effects of changes in the composition of food through changes in technology both in agriculture and food manufacture, (d) provided information on whether or not the restrictions placed on the use of toxic substances in food production are being met in practice, (e) allowed assessment to be made of the need to introduce new legislation or to revise existing legislation, (f)indicated whether non-tariff barriers to international trade have been set up unreasonably, and (g) provided information about the incidence of potentially toxic substances in the diet, which in turn has been the stimulus for further R and D work.Little of this activity has resulted in legislative control over toxic substances in food since, where appropriate, modifications in manufacturing practice have been adopted. This has been a direct consequence of the response of United Kingdom industry. Because of the enormous complexity of people’s diets, of knowing what individuals eat over a period of time, of finding suitable control populations, and of bringing about a change in exposure through the diet, epidemiological methods attempting to study the possible factors in the diet that are responsible for disease will be fraught with limitations. It is clear that developments in the analytical chemistry of food will outpace our knowledge of the potential health significance of low levels of trace non-nutritive constituents of food. Nonetheless, a continued search for such constituents is justified in order that the highest standards of food quality are maintained and, when appropriate, that further studies may be made of the relationship of diet to health.

ISSN:0306-0012    

DOI:10.1039/CS9811000233

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

It Sources of, and Analytical Advances in, Trace Inorganic Constituents in Food By C. J. Pickford ENVIRONMENTAL AND MEDICAL SCIENCES DIVISION, AERE, HARWELL, NR. DIDCOT, OXON. 1 Introduction This review seeks to highlight some of the significant advances that have occurred in recent years concerning trace inorganic constituents in foodstuffs. Some specific sources of trace-element contamination have also been considered in some detail. Because of the very great current interest in those elements, the toxic metallic elements have been dealt with in greater detail than the non-toxic and non-metallic elements. In 1977 Crosbyl reviewed the determination of metals in foods, giving an excellent general picture of the subject up to that time. Biennial reviews are published in Anal.Chem. and are a valuable source of applications data. A. Essential Trace Elements.-Underwood2 has classified trace elements into three groups: (a) those essential for higher animals, (6) those possibly essential, and (c) the non-essential elements. The definition of ‘essentiality’ has been dealt with in some detail by Cotzias,3 and the elements that are generally included in this group are Cr, Co, Cu, F, Fe, I, Mn, Mo, Ni, Se, Si, Sn,V, and Zn. Ti, Pb, As, and Cd4 are under investigation, with other elements, as being possibly essential. B. Toxic Trace Elements.-Other elements have no known function in higher organisms but are always present without apparent effect. Some elements, how- ever, are known to be harmful, and their presence may cause toxic, carcinogenic, or other undesirable effects. Many essential elements are toxic at higher concen- tration levels.The list of these toxic or carcinogenic elements includes4 As, Ba, Be, Bi, Cd, Co, Cr, Cu, Fe, Hg, Li, Mn, Ni, Pb, Sb, Se, Sn, Te, TI, V, and Zn. The specificity of many essential- and toxic-element effects is such that it is important to have an accurate knowledge of the distribution of many of these elements at all parts of the food chain, starting from environmental levels, either natural or polluted, through foodstuffs of animal or vegetable origin, to man himself. N. T. Crosby, Analyst, 1977, 102,225. E. J. Underwood, ‘Trace Elements in Human and Animal Nutrition’, Academic Press, New York, 1977. G.C. Cotzias, ‘Trace Substances in Environmental Health’, ed. D. D. Hemphill, University of Missouri, Columbia, 1967, p. 5. G. H. Morrison, CRC Crit. Rev. Anal. Chem., 1979, 8, 287. Chemical Aspects of Trace Constituents of the Diet. Part I1 C. The Origins of Environmental Levels of Trace Elements.-Ultimately, the distribution of trace elements in our environment, and hence food chain, may be traced back to the average composition of the Earth’s crust. Data for some elements of interest are given in Table 1.5 Concentrations range from high levels of elements such as Si of nearly 30% to levels of 0.002 pg g-1 and lower for Te and rarer elements. Table 1 Distribution of some typical essential and toxic trace elements in the Earth’s crusta Element Average concentration in earth’s crustlpg 8-l Ba 250 Cr 200 V 150 Zn 132 Ni 80 cu 70 Sn 40 Pb 16 Be 6 As 5 Sb 1 T1 0.6 Hg 0.3 Bi 0.2 Cd 0.13 Se 0.09 Te 0.002 a Data from ref.5. Mobilization of trace elements may occur by a number of routes, such as mining of mineral aggregates, volcanic activity, etc., for volatile elements, and leaching processes, to give the raised levels we refer to as ‘environmental pollu- tion’. D. Typical Levels Found in Food.-These may vary from the c 0.001 pg 8-1 level for elements such as V to levels of up to 300 pg 8-1 for essential elements, such as Zn, and some other elements present as a result of contamination, such as Sn. There are many sources of data (e.g.ref. 6) for trace-element levels in food; those quoted were gathered from member states of the European Economic Community by the Health and Safety Directorate of the CEC. Typical data are given in Table 2.6 B. Mason, ‘Principles of Geochemistry’, John Wiley and Sons, 1959. a J. Bouquiaux and M. Langevin, EUR 5730C,Health & Safety Directorate, Luxembourg, 1977. Pickford Table 2 Typical levelstmg kg-1 ojtrace inorganic constituents in fooda As 0.001 -+ 0.1 Cd 0.003 -+ 0.3 -+Cr 0.01 3.0 cu 0.1 -+ 30 Hg 0.003 -+ 0.1 Pb 0.03 -+ 1.0 Sn 0.03 -+ 10 (up to > 300) Zn 1.0 -+ 100 a Data from ref. 6. E. Atypical Levels in Food.-As well as the typical levels quoted, some food products may sometimes contain high levels of certain trace elements.Attention is focused upon this, of course, when the element is toxic, as in the case of Hg in marine fish, up to 3.1 pg g-l,’ Cd in brown crab meat, 10 pg g-l,* and Pb in acid canned foods such as fruit, 10 pg g-l.9 It is clear that there are two dif- ferent problems involved : (a) contamination of food by processes such as canning and (b) concentration of environmental levels of trace elements by various organisms. In the latter case high environmental levels as a result of man’s activities will be concentrated even further. A specific example of this is the concentration of methylmercury in shell fish, from water contaminated by acetaldehyde plant operationlo in the Minamata Bay in Japan. 2Sources of Inorganic Constituents A.Major Sources of Trace Inorganic Constituents.-The major sources of trace inorganic constituents in food can be grouped together to common origins in many cases. Sources for some typical elements are given in Table 3, from data cofitained in reference 6. B. Lead from Automotive Exhaust.-The use of lead additives in petrol to raise the octane rating has been a source of much debate in recent years, and a number of studies have pointed to the potential damage caused to the environment. Most of the concern has arisen over lead levels in air and dusr, but in view of the raised lead levels found in soils and plants growing near to major roadsll and the subsequent passage to man via the food chain this should also be of concern. Food is normally considered to be the major source of lead to industrially A.V. Holden, ‘Mercury Contamination in Man and His Environment’, IAEA Technical Report No. 137, Vienna, 1972. * B. Otto, ‘The Levels of Selected Trace Elements in Food’, May 1975, document prepared by the Commission of the European Communities. J. Bouquiaux, Report of working group, July 1974, C.E.C. document V/F/1966/74e. lo L. T. Kurland, S. N. Faro, and H. Sielder, World Neurol., 1960, 1, 370. l1 H. L. Cannon and J. M. Bowles, Science, 1962, 137, 765. Chemical Aspects of Trace Constituents of the Diet. Part II Table 3 General sources of inorganic constituents in fooda Element Source As Combustion, pesticides, veterinary products Cd Zinc smelters, battery industry, pigments, solders Cr Plating, pigments, tanning, dyeing cu Industrial use, pesticides Hg Mining, volcanic activity, chloralkali, electrical industry, pesticides and fungicides, dentistry Ni Alloys, plating Pb Natural, combustion, paints, industry Sn Canning a Data from ref.6. unexposed individuals, with an average intake of 200-300 pg day-l l2so that any increase in this amount isviewed with some trepidation .One particular study13 has sought to establish the pathways of lead from petrol to man by using an isotopically abnormal form of lead in tetraethyl-lead added to petrol in the Turin area. By analysing air, water, particulate, food, and blood samples an attempt is being made to correlate blood lead levels with petrol consumption over a fixed period of time.C. Trace-element Contamination from Sewage Sludge.-There is some concern that the repeated application, at excessive rates to the soil, of sewage sludge containing trace elements may eventually cause injurious or toxic effects to either crops or animals. Sewage sludge may contain high levels of Cu (0.8%), Ni (0.33%),Zn (up to 4.9%), Pb (up to 0.3%), and Cd (up to 0.15%), expressed as percentage dry weight, from 42 sewage sludges from England and Wales.14 In the U.S.A. levels of up to 0.34% have been reported15 for Cd; such high levels of trace elements are traceable to industrial sources operating within the zone covered by the sewage facilities. Non-industrial areas have much lower levels in genera1,16 although substantial levels of Zn are still rep0rted.1~ Much of the concern about the health consequences of this problem has centred on Cd owing to its high toxicity and la K.R. Mahaffey in ‘The Biogeochemistry of Lead in the Environment’, Elsevier/North- Holland Biomedical Press, 1978. l3 S. Facchetti, International Conference on ‘Management and Control of Heavy Metals in the Environment’, London, Sept. 1979. l4 M. L. Berrow and J. Webber, J. Sci. Food Agric., 1972, 23, 93. l6 D. R. Zenz, B. T. Lynam, C. Len-Hing, R. R. Rimkus, and T. D. Hinesley, Metropolitan Sanitary District of Greater Chicago, Dept. of Research and Development, Report No. 74-20, 1975. Is ‘Survey of Sewage Sludge Composition’, Department of the Environment, 1978, un- published.l7 D. Purves, paper presented at ‘Management and Control of Heavy Metals in the Environ- ment’, London, Sept. 1979. Pickford ability to replace Zn in the body, with dangerous consequences. Page et aL1*have reported data for pea, wheat, and corn uptake from sludge-amended soils as a function of Cd content of the sludge, pH, and plant species. These data show that more Cd is mobilized at a lower pH and that the highest Cd concentration was found in the leaves, rather than the seeds or grain, of the plant. However, the concentration of Cd in peas grown at a lower pH (4.4-5.2) was practically linearly related to Cd added from the sewage sludge. Davis and Cokerls have reviewed the agricultural effects of Cd in sewage sludge used as fertilizer.They consider the difficulty of assessing the extractable part of the Cd present to assess the portion which is taken up by plants. The guidelines used at present to control Cd levels in sewage sludge used as fertilizer are cautious to allow for the limited knowledge currently available about uptake rates, etc. European legislation allows from 10to 167 g ha-l yr-l to be added to agricultural land, whereas the U.K. suggests a soil increase of up to 2.3 p.p.m. to be achieved over 30 years or more. 3 Analytical Methods A. Sample Preparation.-Before a sample can be analysed, some form of pretreatment is almost invariably required. Samples are generally dried to constant weight and then either analysed directly or wet or dry ashed and the ash or a solution of it is analysed by a solution technique. A full appraisal of preparation methods for food can be found elsewhere.' B.Atomic Absorption Spectroscopy.-The breakdown made by Crosbyl of published analytical applications showed that in the years up to 1977 a.a.s. had been by far the most popular technique used for trace inorganic constituents in food. An examination of recent literature confirms this trend. The techniques of hydride generation and electrothermal a.a.s. are becoming particularly popular as a result of improved instrumentation. The hydride generation a.a.s. technique has been used for the analysis of As, Se, Sb, Te, and certain other elements. Evans, Jackson, and De1lar2O have recently described a comprehensive procedure for determining Sb, As, and Sn in foodstuffs using this technique combined with a specific wet oxidation step.Fiorino et aLZ1have used a similar hydride generation technique coupled with an argon-entrained hydrogen Aame for As, Se, Sb, and Te in food. Their results for As and Se are compared with data obtained by i.n.a.a. (instrumental neutron activation analysis), a.a.s., fluorimetry, and spectrophotometry and show excellent agreement. A similar study for As and Se has been made by Ihnat and Miller.22 l8A. L. Page, A. C. Chang, and F. T. Bingham, as in ref. 17. lBR. D. Davis and E. G. Coker, as in ref. 17. *O W. H. Evans, F. J. Jackson, and D. Dellar, Analyst, 1979, 104, 16. 81 J. A. Fiorino, J. W.Jones, and S. G. Capar, Anal. Chem., 1976, 48, 120. 89 M. Ihnat and H. J. Miller, J. Assoc. OH.Anal. Chem., 1977,60,1414. Chemical Aspects of Trace Constituents of the Diet. Part 11 There is no doubt that, despite interferences that may arise during the hydride- forming step from transition metals and other species, the hydride-generation a.a.s. technique is becoming very widely used and is one of the most convenient techniques for these elements. Electrothermal a.a.s. is now very extensively used for the determination of certain trace metals in acid digests of food material. The very high sensitivity attained by this technique permits direct determination of trace elements such as Pb in food digests at normal environmental level^,^^^^^ although many workers prefer to use an extraction process to separate heavy metals before electrothermal a.a.s.determination (e.g. ref. 25). In addition to improving sensitivity even further, a solvent-extraction step has the added advantage of reducing the matrix and background effects which are often associated with this technique. Verymuchwork isstill done usingflame a.a.s. combined with an extraction tech- nique to improve detection limits where necessary. Allenby et aZ.26have published a routine screening procedure for Cd, Cu, Pb, and Zn in food for levels down to 0.02 pg g-l, using such a procedure with an optional DDC/heptan-3-one extrac- tion step. Using similar procedures, many workers claim that it is unnecessary to use electrothermal rather than flame a.a.s.since the extraction step increases the sensitivity by a sufficiently large factor. C.Inductively Coupled Plasma Optical Emission Spectrometry.-This technique (i.c.p.o.e.s.), originally pioneered by Greenfield27 in the U.K. and by Fassel and co-workers in the U.S.A.,28 is rapidly becoming one of the most widely used multi-element techniques for trace metals. Although the capital cost of the equipment is high, a large number of elements may be determined simultaneously in a sample, so that the unit cost per element determination may be much lower than with a.a.s. The technique has been extensively reviewed by Barnes.Z9 Detection limits from i.c.p.0.e.s. are similar to those from flame a.a.s. for many elements, although much better for refractory elements such as B, Zr, etc.There has been only a limited investigation of interference effects so far (ref. 29 and refs. therein), but it would appear that chemical effects such as the phosphate/calcium effect seen in a.a.s. are reduced or absent, but that spectral interference, especially when direct-reading spectrometers are used with wide slit widths, occurs as in other forms of emission spectrometry. Irons et al.30 have compared the technique with X-ray fluorescence spectro- metry (x.r.f.) amongst several laboratories, using digests of NBS standard ref- erence materials such as Bovine Liver and Orchard Leaves for the comparison. They conclude that i.c.p.0.e.s. is the more sensitive technique for most elements a3 D.Harbach and H. Diehl, 2.Anal. Chem., 1978, 290, 145. a4 Z. Grobenski, M. Melcher, and B. Weltz, 2. Anal. Chem., 1978, 290, 144. 86 R. W. Dabeka, 25th Canadian Spectroscopy Symposium, Sept. 1978, Quebec, Canada. z6 P. Allenby, J. W. Robertson, and F. C. Shenton, J. Assoc. Public Anal., 1977, 15, 61. S. Greenfield, I. L. W. Jones, and C. T. Berry, Analyst, 1964, 89, 713. R. H. Wendt and V. A. Fassel, Anal. Chem., 1965, 37, 920. a* R. M. Barnes, CBC Crit. Rev. Anal. Chem., 1978, 7, 203. *O R. D. Irons, E. A. Schenk, and R. D. Giauque, Clin. Chem., 1976,22,20218. Pickford except Ni and Pb. 1.c.p.o.e.s. requires a liquid sample, x.r.f. a solid pellet; this may be advantageous or not, depending on the sample type. Accuracy and pre- cision were found to be similar for the two techniques, although the much simpler data evaluation required for i.c.p.0.e.s.meant that even a very small minicomputer was sufficient. A.a.s. and i.c.p.0.e.s. have been compared by Munter et a1.31 for a number of food materials. They presented a large amount of data obtained with an i.c.p.0.e.s. system for standard reference materials and other food products, previously analysed by 15-23 other laboratories by a.a.s. They conclude that the inductively coupled plasma is a near-ideal multi-element source, although they point out that lower detection limits are needed for the elements Cr, Cd, Ni, and Pb in order to reach normal levels of these elements. D. X-Ray Fluorescence Spectrometry.-This technique is much used for environ- mental monitoring (e.g.air filters) but has not found a wide application for the analysis of food, perhaps owing to the high cost of the equipment and the fact that samples and standards must be closely matched to obtain meaningful data. Giaque32 describes the use of the technique for a number of elements in biological materials and compares the technique in another publicati0n3~ to i.c.p.0.e.s. Proton-induced x.r.f. (p.i.x.e.) has been used by Campbell33 to determine trace elements in wines and complex biological matrices. Pelletization of freeze-dried solids has been applied to milk ~amples3~ and other types of foods35 and appears to provide a convenient means of sampling in many cases. E.Neutron Activation Analysis.-This multi-element technique is more sensitive than x.r.f. or i.c.p.0.e.s.for many elements and is essentially blank free, although it does require access to irradiation facilities. Generally, sensitivities are lower if no chemical separation is made (i.n.a.a.), and the presence of large amounts of Na, etc., may require chemical separation to be carried out, at a much greater cost. Schele~z3~determined 25 elements in human diet by this route, and with DiehP7 he described the application of the technique to the determination of up to 28 elements in food and similar materials, with detection limits in the range 10-6-10-11 g for a 0.15 g sample. They report matrix effects for chromium when it was measured in potato flour. 31 R.C. Munter, R.A.Grande, and P. C. Ahn, ICP Information Newsletter, 1979, 5, 368. R.D. Giaque, F. S. Goulding, J. M. Jaklevic, and R. H. Pehl, Anal. Chern., 1973,45, 671. J. L. Campbell, B. H. Orr, A. W. Herman, L. A. McNelles, J. A. Thompson, and W. D. Cook,Anal. Chem., 1975, 47, 1542. A. L. Langhorst, N. A. Bonner, F. Bazau, and D. R. McIntyre, U.S. Energy Res.Dev. Adm. Rep., 1976. 3s R.T. King, J. Sci. Food Agric., 1977, 28, 631. 36 R.Schelenz, J. Radioanal. Chern., 1977, 37, 539. 37 J. F. Diehl and R. Schelenz, Lebensrn.- Wiss. Technol., 1975, 8, 154. Chemical Aspects of Trace Constituents of the Diet. Part II Other determinations using this technique involve extensive chemical separa- tion and are better described in that context. F.Other Analytical Techniques.-The summary by Crosbyl of the citation of analytical methods used for food shows that, in addition to the techniques mentioned, flame photometry, spectrophotometry, titration, and electrochemical methods are the only others to be used to any extent. The use of these techniques is restricted to mainly routine applications using standard published methodolo- gies. G. Separation and Preconcentration Techniques.-There are many circumstances where the sensitivity or selectivity of a method is such that some form of separa- tive or preconcentration procedure must be used. This is particularly true of the toxic trace elements such as Pb, Cd, and As, and many publications (e.g. refs. 25 and 26) deal with the a.a.s.determination of these elements in food after chelation and solvent extraction. Such a separation does, of course, raise the sensitivity of flameless a.a.s. even further, permitting interference-free determination at low environmental levels. With the increasing availability and use of multi-element instrumental methods, there is a particular need for preconcentration methods that separate elements en bloc from interfering matrix elements. Thus Menke38 used ion exchange to preconcentrate trace elements from food, prior to x.r.f. determination. Van der Sl00t39 has studied the use of activated charcoal as a collecting medium for trace metals, either in simple acid solution or chelated with ammonium pyrrolidine dithiocarbamate, the final determination being made by i.n.a.a.Chelating resins have also been used for preconcentration of trace elements from various matrices before determination by ~.r.f.~O or i.c.p.o.e.s.*l Similar materials have been used42 for the trace determination of metals in foods and other materials by a.a.s. The concentration of trace impurities plays an important role in neutron activation analysis, either before the activation step43 or afterwards, although in this case the recovery can be incomplete since inactive carriers are generally added to improve, and assess, the yield obtained. Although chemical separation or preconcentration techniques may lengthen an analysis somewhat, this may be at least partially offset if a group of elements is simultaneously separated and determined. 38 H.Menke, Fresenius’Z. Anal. Chem., 1977, 286, 31. 39 H. A. van der Sloot, ECN-1, Petten, Netherlands, 1976. 40 D. E. Leyden, G. H. Luttrell, W. K. Nouidez, and D. B. Werho, Anal. Chem., 1976, 48, 67. R. M. Barnes and J. S. Genna, Anal. Chem., 1979, 51, 1065. 4x D. S. Hackett and S. Siggia, ‘Environmental Analysis’, Academic Press, New York, 1977, p. 253. 43 M. H. Yang, P. Y. Chen, C. L. Tseng, S. J. Yeh, and P. S. Weng, J. Radioanal. Chem., 1977, 37, 801. Pickford 4 FutureTrends A. Speciation of Trace Elements.-The importance of the exact form of a trace element with regard to its toxicity, and hence the maximum allowable concentra- tion that may be present in food, has been recognized for some time.The methyl form of mercury is known to be the more toxic f0rm,4~ compared to inorganic mercury, and specific analytical techniques are capable of distinguishing between the two f0rms.~5 In the case of mercury, conversion of inorganic mercury to methylmercury has been reported in and explains the build-up of organic mercury in fish products such as tuna, which was recently the source of much publicity. Similarly, the toxicity of chromium depends on the valence state of the element. This is reflected in the different air levels allowed for CrIII and CrVI. Considered to be of sufficient importance, the topic of metal speciation merited a session at the recent conference ‘Management and Control of Heavy Metals in the Environment’ in London, September 1979.Florence and Batle~~~ have reviewed the possible approaches for chemical speciation based on chemical modelling procedures or analytical measurements following separation. A number of other publications at the above conference dealt with metal speciation in aquatic systems, and there seems no doubt that future work on foodstuffs will be directed towards determination of the form as well as the concentrations of metal species. There are, of course, considerable difficulties in extension of this topic to food. Most meat, vegetable, efc. products, unlike water, require considerable chemical pretreatment before analysis, with the attendant risks of changing the form of the metal present. B. Synergistic Effects.-Synergistic effects already play a role in trace-element toxicology, e.g.the role of Se in Hg toxicity has been studied,44 the uptake of Pb is enhanced from Ca- or Fe-deficient diets,48 and in agriculture the inter-relation- ship of Mo and Cu is well known.49 It seems reasonable to suppose that this topic will also assume greater importance in the trace-element chemistry of food. It is easy to envisage the legislative nightmare that could result. In the study of synergistic effects, multi-element analytical techniques such as i.c.p.o.e.s., x.r.f., or i.n.a.a. are clearly advantageous, and the present widespread use of computers for complex data handling renders the task of establishing inter-element correlations rather easier. C. New Elements.-As more understanding is obtained about the effects of trace E.J. Underwood, ‘Trace Elements in Human and Animal Nutrition’, Academic Press, New York, 1971, and refs. therein. 46 L. Magos, Analyst, 1971, 96, 847. 46 S. Jensen and A. Jernelow, Nature, 1969, 223, 753. 47 T. M. Florence and G.E. Batley, CRC Crit. Rev. Anal. Chem., 1980, 9, 219. K. R. Mahaffey, T. A. Banks, C. L. Stone, S. Capar, J. Compton, and M. Glubik in ‘Proc. Tnt. Conf. Heavy Metals Environ.’, Toronto, Canada, 1977. 4s A. T. Dick, Aust. J. Agric. Res., 1954, 5, 511. Chemical Aspects of Trace Constituents of the Diet. Part 11 elements on man, toxic effects of ‘new’ trace elements are discovered. Thallium, long known as a toxic element in factory situations but only recently being considered in any detail, is now becoming of interest to environmentalists (see, for example, ref. 50) and, no doubt, other elements will create similar interest in the future. Naturally the interest in new elements reflects to a certain extent the ‘state of the art’ of analytical techniques. Thallium is, as it happens, a rather difficult element to determine at trace levels by most of the techniques available for instrumental analysis, so that until recently determination of this very toxic metal at environmental levels was a difficult analytical task. S.Tooze, Ecologist, 1980, 5, 163.

ISSN:0306-0012    

DOI:10.1039/CS9811000245

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

111 Advances in the Analysis of Trace Organic Constituents of the Diet, with Particular Reference to Mass Spectrometry By J. Gilbert MINISTRY OF AGRICULTURE, FISHERIES AND FOOD, FOOD SCIENCE DIVISION, HALDIN HOUSE, QUEEN STREET, NORWICH NR2 4SX and R. Self AGRICULTURAL RESEARCH COUNCIL, FOOD RESEARCH INSTITUTE, COLNEY LANE, NORWICH NR4 7UA 1 Introduction What is meant by a trace organic constituent of a food? Each generation of food chemists would give a different answer. For us in 1980 it is appropriate to define ‘trace’ as meaning constituents present at or below the parts per billion (p.p.b.) level, but because we are primarily concerned with instrumental techniques it is more precise to talk about submicrogram levels of detection. The adequacy of such levels of sensitivity will depend upon the nature of the problem.1 For example, it may only be necessary to establish that the compound of interest is above or below a specified limit; on the other hand, it may be the minimum possible detectable level that is sought, for monitoring known potently toxic or carcinogenic constituents.From an analytical point of view, achieving the latter presents the greatest challenge, and it is the wide range of different methods available for trace organic analysis that is in need of careful assessment to ascer- tain which ones are suitable for a particular job. Usually there are four stages in the estimation of a trace organic compound, sampling, extraction, separation, and detection, and it is important to be able to identify initially the most likely sources of error.Whilst the importance of samp- ling from biological materials should not be overlooked, it is not normally a serious methodological problem in food analysis, provided that the proper procedures have been implemented. The extraction of the compound of interest from this sample, however, can involve several processes which are difficult to perform accurately because of the physical and compositional complexities of food matrices. Relatively little progress has been made of late in this area, which contrasts sharply with the innovations in the separation and detection stages. The purpose of this review is to consider the assay as a whole, from sampling through to detection, and to draw upon some specific examples from the litera- ture to illustrate critically both the advantages and shortcomings of alternative techniques that might be considered for a particular trace analysis problem.Throughout, mass spectrometry has been used as a linking theme, not because it 1 W. Horwitz and J. W. Howard in ‘Trace Organic Analysis: A New Frontier in Analytical Chemistry’, NBS Spec. Publ. 519, ed. H. S. Hertz and S. N. Chesler, Washington D.C., 1979, p. 231. 255 Chemical Aspects of’Trace Constituents of the Diet. Part 111 can in any way be described as a ‘widely used’ or ‘freely available’ technique but because technically mass spectrometry is unquestionably in the forefront as regards recent advances in trace analysis and because as an analytical tool it is unsurpassed both for versatility of application and in terms of specificity and sensitivity.2 Sample Preparation The sample submitted to the analytical laboratory may not be representative of the batch of material from which it was taken. Furthermore, even if it were when it was collected, its temporary storage may have seriously affected the level of any trace organic constituent. The losses caused by adsorption onto container sur- faces, evaporation, thermal decomposition, and exposure to light can often be great enough to invalidate the results of the analysis2 The maximum sample size that can be used for analysis is often determined by the extraction/concentration procedures that follow. It is not always possible to take samples large enough to be truly representative of the bulk material because unacceptable errors would be introduced at the concentration stage if large volumes of solvent have to be used to ensure complete extraction.It is at the extraction stage that an internal standard can be added to help compensate for these losses and for any further losses that may be attributed to adsorption later on in the assay.3 The conventional practice of ‘spiking’ with the standard compound at the extraction stage to determine the recovery has always been viewed with caution because it cannot be proved that spiking truly simulates the physical condition of the trace compound in the biological matrix. But no matter how problematical these procedures are, it is essential to persevere with the quantification of this unsatisfactory aspect of the assay.Solvent extraction is seldom specific for the analyte, and often trace impurities are introduced from the solvent itself. Depending on the specificity of the detection method employed, some degree of clean-up will be required. The most popular methods are adsorption chromatography, gel permeation, and ion ex- change, which can cope with the large volume of solvent, but after an evaporation step (unfortunately prone to large errors) other conventional chromatographic techniques, e.g. thin-layer chromatography, have been used. Extra care is required when internal standards are present, to ensure that they pass through the chroma- tography stage with the compound of interest.The cleaned-up extract is then evaporated to low volume or dryness, depending on the next stage. Even with an internal standard and after deactivation of the glassware with silylating agents, this is still a vulnerable procedure. There have been few improvements in the isolation and concentration method- ology since the pioneering work in flavour chemistry during the 1960s,*but the H. S. Hertz, W. E. May, S. A. Wise, and S. N. Cheder, Anal. Chem., 1978, 50, 428A. B. J. Millard, Quantitative Mass Spectrometry in Life Sciences 11, Proc. 2nd Int. Symp., Ghent, 13-16 June 1978, ed. A. P. de Leenheer, R. R. Roncucci, and C. van Peteghem, Elsevier S.P. Co., Amsterdam, 1978, p. 83. * J. M. H. Bemelmans, ‘Progress in Flavour Research’, ed.D. G. Land and H. E. Nursten, Applied Science Publishers, London, 1979, p. 79. Gilbert and Self use of organophilic resins, viz XAD-2 and Tenax, is worthy of note.215 The improvement of microchemical techniques for trace organic analysis must be the foundation work urgently required to increase the reliability of this stage of the assay. It may be necessary to alter the molecular structure of the compound of interest in order to provide enhanced volatility for further chromatography or to incor- porate a feature which will increase the specificity to detection. The most popular derivatives are those that do both, e.g. pentafluorophenyldimethylsilylether.697 There is a wide choice of derivatives and derivatization methods,* but only those capable of working efficiently on the microscale will be acceptable.It may be concluded that detection should be as sensitive and selective as possible so that the number and extent of these preliminary stages can be kept to a minimum. 3 Identification and Confirmation of Identity Compound identification is an obvious prerequisite of any quantitative assay. Mass spectrometry and combined g.c.-m.s. are often the only techniques with sufficient sensitivity for trace organic analysis. Modern rapid-scanning mass spectrometers with on-line computerized data acquisition and processing facilitate the collection of several thousand spectra during the g.c.-m.s. analysis. These can be tested for information content (mass resolved chromatography), the redundant spectra discarded, and the remainder searched, in the forward or reversed mode, using a library collection.Reversed searching using small special- ized collections is particularly useful for the semi-quantitative screening of trace compounds. Alternatively, the mass-spectral data accumulated during g.c.-m.s. runs can be presented as a chromatographic profile by using a particular ion or any selection of ions in the spectrum (mass chromatography) in order to indicate the presence of particular compounds in complex mixtures submitted for rapid testing. In conjunction with the g.c. retention data, a known compound can be recognized from the partial mass spectrum. Typically 10-100 ng samples are required for interpretation.4 Chromatography with Gkneral-purpose Detectors Typical extracts submitted for analysis require further purification to ensure that the compound of interest is free from substances which could cause chemical interference with its quantitative estimation. General-purpose detectors are invaluable in dealing with mixtures of unknown volatile compounds, and the flame-ionization detector (FID) for gas chromatography has been extensively used in food analysis. However, if a general-purpose detector is to be used, it is W. Averill, Chromarogr. News].,1978, 6, 4. A. J. Francis, E. D. Morgan, and C. F. Poole, J. Chromatugr., 1978, 161, 111. A. J. Francis, E. D. Morgan, and C. F. Poole, Org. Mass Spectrom., 1978,13, 671.K. Blau and G. King, ‘Handbook of Derivatives for Chromatography’, Heyden and Son Ltd., London, 1978. Chemical Aspects of Trace Constituents of the Diet. Part III essential to optimize the selectivity of the chromatographic separation. There have been several developments in this area (see below). A. The Improved Manufacture of Glass Capillary Columns.-The high adsorption levels previously associated with glass support-coated and stainless-steel wall- coated open tubular columns have been greatly reduced by improved surface deactivation techniques applied to the latest glass and fused-silica column~.~J~ It is now possible to detect low picogram quantities of, e.g., 2,4-dimethyl aniline and 2,6-dimethylphenol (acid/base test compounds) with a FID.I1 A particularly effective use of high-efficiency capillary column analysis is in the separation of amino-acid enantiomers with specially prepared chiral stationary phases.12 The unnatural enantiometers can then be used as internal standards for quantitative studies.B. New Developments in Capillary Column Injector Design.-The previous inherent disadvantage of capillary columns -the low capacity for injected volumes of solvent -can now be overcome by new injectors utilizing the ‘solvent effe~t’l3.1~(injections of up to 5 pl of solvent), meaning that a wide range of compounds of interest can be accommodated. C. The Use of Mixed Stationary Phases.-A development with potential applica- tion to trace quantitative analysis is the manufacture of columns specially designed to separate a particular component from the rest of the mixture.If two different monophasic columns cannot separate the compound of interest from interfering constituents, it should be possible to calculate the percentage composition of a mixture of the phases and manufacture a column that will produce the desired effect.15 The introduction of liquid chromatography (1.c.) to quantitative analysis has extended the range of compounds to include those which are thermally unstable or too involatile for g.c. analysis even when derivatized. Recent advances in this area have been confined largely to the diversification of its application, but the development of microbore columns shows considerable promise.16 5 Chromatography with ‘Specific’ Detectors Given unlimited time and resources, no doubt chromatography alone could provide the specificity required for use with general-purpose detectors, but in practice it is expedient to use detectors having increased specificity in conjunction with readily available chromatographic methods to provide combinations with * G.Schornburg, H. Husrnann, and H. Behlau, Chromatographia, 1980, 13, 321. lo K. Grob and G. Grob, J. High Res., 1980, 3, 197. R.Self, unpublished observations. W. A. Konig and G. A. Nicholson, Anal. Chem., 1975, 47, 951. la K. Grob and K. Grob, jun., J. High Res. Chromatogr. Chromatogr. Commun., 1978, 1, 57. l4 K. Grob and K. Grob, jun., J. Chromatogr., 1978, 151, 3 11. l6 R.J.Laub, J. H. Purnell, D. M. Summers, and P. S. Williams,J. Chromatogr., 1978,155, 1. laT. Takeuchi and D. Ishii, J. Chromatogr., 1980, 190, 150. Gilbert and Self appropriate selectivity. The enormous variation in the properties of candidate organic compounds precludes any logical plan to simplify the present diversifica- tion in methodology. At this time it is only possible to make superficial com- parisons (see below). A. Comparison of Gas Chromatography Detectors.-A large number of g.c. detectors are available, of which those shown in Figure 1 are by no means exhaustive (additional examples being photoionization and ultrasonic and i.r.-u.v. spectroscopic detectors). The detection limits shown are very much general guidelines and in actual analyses the smallest detectable amount may be much more or much less, but they serve to give an impression of the relative merits and in particular indicate the position held by mass spectrometry when used for ion monitoring.The recently introduced negative-ion chemical ionization (n.i.c.i.) for quantitative estimation of electron-capturing compounds, or chemi- cally prepared electron-capturing derivatives, offers unsurpassed sensitivity and considerable selectivity. The most selective detector is the thermal-energy analyser (TEA)17 which was specially built for nitrosamine analysis.l* B. Comparison of Liquid Chromatography Detectors.-In Figure 2 the attainable limits of detection for I.c. are compared. There is no equivalent in 1.c.to the universal and sensitive FID used in g.c. Refractive index and nephelometric detectors are really too insensitive for trace analysis. U.V. detectors are the most widely used for compounds with a chromophore, utility being extended by monitoring at 190-200 nm with the proper selection of a non-absorbing mobile phase. With fluorescence, one is beginning to approach comparable limits of detection to those of g.c., but this technique is necessarily limited to the few compounds with native fluorescence, e.g. aflatoxins, polyaromatics, or those compounds which can be derivatized to incorporate a fluorophore. It is the recent advances in 1iquid chromatograph y-mass spectrometry (1 .c.-m .s.) coup1 inglg which indicate the possible future direction of detection in 1.c.The technique is in its infancy and many different approaches are being made. Moving belts were the first commercial system, but orifice types are making progress now. A direct coupling can be made if the 1.c. solvent, after vaporization, can be re-used as the reagent gas for chemical ionization mass spectrometry (c.i.m.s.). Other innova- tions include the use of a microbore column16 to reduce the quantity of solvent to a level which can be handled by conventional g.c.-m.s. interfaces, again provided that c.i.m.s. is employed. For quantitative analyses c.i.m.s. is not an impediment and, by analogy with g.c. capillary columns, microbore techniques will increase the selectivity (and sensitivity) of the assay. By using the mass spectrometer, it means that for the first time a specific, sensitive, and yet general- purpose detector is available, enabling previously intractable problems in food analysis to be tackled.I'D. H. Fine, R. Rufeh, D. Lieb, and D. P. Rounbehler, Anal. Chem., 1975, 47, 1188. l8T. A. Gough, K. S. Webb, and R. F. Eaton, J. Chromatogr., 1977, 137, 293. lS P. J. Arpino and G. Guichon, Anal. Chem., 1979,51, 683A. FID ECD * AFlD (nitrogen)4 I 3iFlD (phosphorus) Y ms.(s.isn.)f I I ms.(s.im -n.i .c.i. c g Fy(sulp FPD (phosphorys) < TEA (nitrc < MPED Hall electrolytic conduct ivit v t I I I I 1 Figure 1 Comparison of limits of detection and linear response for gas chromatographic detectors [FID = flame-ionization detector, ECD = electron-capture detector, A FlD = alkaliflame-ionization detector, m.s.(s.i.m.) = mass spectrometry (selected-ion monitoring), n,i.c.i.= negative-ion chemical ionization, FPD =flame photometric detector, TEA = thermal-energy analyser, MPED = microwave plasma-emission detector; the dashed right-hand end of an arrow indicates that the upper rimit is uncertain] Gilbert and Self Chemical Aspects of Trace Constituents of the Diet. Part III 6 Chromatography with the Mass Spectrometer as Detector Some of the detectors mentioned above allow a choice to be made, e.g. among the different wavelengths, at which a compound may absorb radiation, so that selectivity can be maximized. This type of flexibility is extended by using the mass spectrometer as detector.The multitude of ionic species created from the sample and its chemical background (interference) are separated into discrete ion beams which, at low resolution, can contain mixtures of both isomeric and isobaric species but at optimum resolution contain only a mixture of isomeric ions. The species considered to offer the greatest specificity for the compound of interest are selected for detection -selected-ion monitoring (s.i.m.). There are two types of s.i.m.: (a) single-ion monitoring, carried out with the mass spectrometer in the static mode, affords maximum sensitivity at optimum resolution but excludes the use of both isotopically labelled internal standards and multicomponent analyses, where two or more characteristic ions have to be detected, and (b) multiple-ion monitoring (mim.) with the mass spectrometer in the switched mode, which, although free from the above exclusions, can only be used with limited resolution and sensitivity.S.i.m. methods in general are susceptible to instrument instability (drift). The method of displaying on the oscilloscope the regions of the spectrum con- taining the selected ion (selected mass scanning) prevents drift errors and also gives a visual warning of any chemical interference from ions of closely similar mass. This has been the method of choice for some nitrosamine and mycotoxin assays. It is generally accepted that single-ion monitoring will give higher sensitivity than m.i.m. at the same resolution, but the trade-off between sensitivity and resolution20 and between precision under m.i.m.and high-resolution single-ion monitoring must be carefully ‘balanced’ against the requirements of every indi- vidual assay. Faced with the problems of whether to use ion ratios or peak areas (heights), labelled or unlabelled internal standards, low-resolution g.c. with high-resolution mass spectrometry or vice versa, the analyst has resorted to trial and error methods, and only recently have attempts been made to rationalize the design of an assay from first principles. Methods of ionization other than electron impact (e.i.) may provide both extra selectivity and increased sensitivity in certain cases, and there have been important recent advances in these areas (see below).A. Chemical Ionization.21s22-Chemical ionization mass spectra are charac-terized by fewer fragment ions and intense even-electron pseudo-molecular ions. *O B. J. Millard, ‘Quantitative Mass Spectrometry’, Heyden and Son Ltd., London, 1978, p. 135. 11 K. R. Jennings, ‘Gas Phase Ion Chemistry’, Academic Press, 1979, Vol. 2, p. 123. Ia R. E. Mather and J. F. J. Todd, Int. J. Mass. Spectrom. Ion Phys., 1979, 30, 1. Gilbert and Self This simplification of the spectrum means that m.i.m. of simple mixtures is then possible by direct-probe m.s. The enhancement of the molecular ion may also provide increased sensitivity, and an intensification of the ions of higher mass increases the specificity by reducing interference from chemical background ions which are usually of lower mass. The use of g.c.-c.i.m.s.is increasing for both rapid screening and accurate assays, especially for those compounds which are subject to thermal and e.i. induced decomposition. B. Negative-ion Chemical I~nization.~~-Two approaches to n.i.c.i. are possible : (a) for compounds of high electron affinity the use of conventional reagent gases (e.g. nitrogen, methane, or argon) to form electron attachment spectra or (6) for compounds of low electron affinity the use of reagent gases which themselves form negatively charged ions (e.g. Freons, methylene chloride, nitrous oxide, or even sulphur hexafluoride), whose role is anion-substrate formation through charge exchange or cluster f0rmation.2~ By judicious choice of reagent gases ion-molecule reactions can be manipulated to yield structural information or to provide high-sensitivity detection by s.i.m.The increased sensitivity attainable with n.i.c.i. compared with positive-ion c.i. is due to the greater total ion current achieved for one specific ion (often 1000 times more negative ions than positive ones2*) and through the much reduced and simplified background. For the future, the application of n.i.c.i. in trace food analysis will most likely be to problems where confirmation or quantification of electron-capturing compounds is required. In particular one thinks of pesticide analysis25 where the complexity of other interfering compounds of environmental origin can cause problems of detection by conventional electron-capture detector, but it remains to be seen whether because of the increased specificity of n.i.c.i.-m.s.reduced clean-up procedures could be adopted. 7 The Mass Spectrometer as Its Own Chromatograph Even under optimized g.c.-m.s. conditions, s.i.m. suffers from chemical inter- ference introduced by residual sample impurities, chromatographic bleed, and the increasing problem of environmental pollution, and these problems will increase as lower detection levels are needed. Metastable-ion mass spectrometry offers a partial solution. Methods have been developed, e.g. mass-analysed ion kinetic-energy spectrometry26 and various forms of linked-scanning,27 which detect metastable ions free from normal ion interference, i.e.free from ‘chemical’ noise. The acronym m.s.-m.s. has been coined to suggest two distinct mass spectral separations in tandem.28 Therefore it is possible to work with relatively crude extracts, largely eliminating the time-consuming and error-prone sample a3 D. F. Hunt and F. W. Crow, Anal. Chem., 1978, 50, 1781. ar H. Brandenberger, Recent Dev. Mass Spectrom. Biochem., 1979, 2, 227. aaD.W. Kuehl, M. J. Whitaker, and R. C. Dougherty, Anal. Chem., 1980, 52, 935. B6 M. H. Bozorgzadeh, R. P. Morgan, and J. H. Beynon, Analyst (London),1978, 103,613. 87 A. P. Bruins, K. R. Jennings, and S. Evans, Int. J. Mass Spectrom. Zon Phys., 1978, 26, 395. F. W. McLafferty, Acc. Chem. Res., 1980, 13, 33. Chemical Aspects of Trace Constituents of the Diet.Part III clean-up and pre-chromatographic stages.29 A1though metastable peaks are of relatively low intensity compared to normal ions, the low background facilitates maximum signal amplification, and already picogram-level analyses have been described.30 High selectivity has been demonstrated with crude extracts,31 but the possibility remains that identical precursor ions could originate from more than one compound in the mixture. However, the successful differentiation of isomeric species generates considerable optimism for the widespread application of the method. Work is just beginning on the optimization of mass-spectral conditions for the production of intense metastable ions. Collision chambers32 and longer field free regions will help this process, as will the new high-perfor- mance instruments that are being introduced.33 The present sensitivity of these techniques for quantitative analysis is described as similar to that obtained using high-resolution mass spectrometry (h.r.m.s.) for elemental analy~is.3~ The most likely application in food science will be for rapid assays of halogenated contaminants, possibly in conjunction with n.i.c.i.35 The general application will be to the estimation of unusual compounds in bland matrices.8 Immunological Assays as an Alternative to Mass Spectrometry Many of the methods described so far would be considered to be too expensive for routine quality-control laboratories, and cheaper methods are sought.Immunological assays have been claimed to be more cost-effective, but do they provide true values ? Because g.c.-m.s.-s.i.m. methods in clinical chemistry are further advanced than in other fields and have entered the stage of being de- scribed as ‘definitive’ methods, a number of comparisons have been made with radioimmunoassays (r.i.a.).36-39 The present situation would seem to be that for r.i.a. the preparation of the antiserum is difficult and most antisera demonstrate some degree of cross-reactivity, lowering the ~pecificity.3~ The added incon- venience of having to synthesize the radiolabelled analogue and calculate the compensation factor for its use contrasts with an assay methodology which has been described as both simple39 and convenienL37 The sample preparation procedure is also described as simple but should be qualified by ‘but extensive’.Whereas it is clear that extensive clean-up is advisable for r.i.a.,38 more work is required to determine the optimum level of preparation of samples for s.i.m. as R. W. Kondrat and R. G. Cooks, Science, 1978, 199, 978. 30 S. J. Gaskell and D. S. Millington, Biomed. Mass Spectrum., 1978, 5, 557. 31 T. L. Kruger, R. G. Cooks, J. L. McLaughlin, and R. L. Ranieri, J. Org. Chem., 1977,42, 4161. 39 R. S. Stradling, K. R. Jennings, and S. Evans, Org. Mass Spectrum., 1978, 13, 429. 33 F. W. McLafferty, P. J. Todd, D. C. McGilvery, and M. A. Baldwin, J. Am. Chem. Sac., 1980, 102, 3360. 34 S. J. Gaskell, R. W. Finney, and M.E. Harper, Biomed. Mass Spectrum., 1979, 6, 113. 36 R. W. Kondrat, G. A. McClusky, and R. G. Cooks, Anal. Chem., 1978,50, 1222. m S.Baba, Y.Shinohara, and Y. Kasuya, J. Chromatogr., 1979, 162, 529. 37 K. H. Powers and M. H. Ebert, Biomed. Mass Spectrum., 1979, 6, 187. 38 K. Fotherby, J. Steroid Biochem., 1979, 10, 121. 39 V. W. Winkler, J. M. Strong, and R. A. Finley, Steroids, 1977, 29, 739. Gilbert and Self It is likely that h.r.s.i.m. in particular will require much less pre-purification and therefore may not be any slower when the total analysis time is considered. Not- withstanding this, most commentators agree that s.i.m. is superior to ria. in specificity and is viewed as the most reliable confirmatory method. 9 Selected Examples of Applications in Trace Analysis A.Polychlorinated dibenzodioxins (PCDDs).-The PCDDs are a group of highly toxic compounds found as stable contaminants in commercial herbicides and which have additionally, through industrial accidents, been released to the environment. The possible contamination of soil and water through both routes gives rise to concern about the appearance of these compounds in the food chain. There are 75 isomers of PCDDs containing one to eight chlorine atoms (and 22 tetrachloro-positional isomers), which vary greatly in their acute toxicity and biological a~tivity.~O$~l Hence a methodological requirement is the need to be able to distinguish between these similar compounds, and furthermore the extreme toxicity of 2,3,7,8-TCDDnecessitates monitoring for it at p.p.t.(10-12 g g-1) levels in foods. An added problem is the frequent presence in biological samples of much larger amounts of halogenated aromatics (PCBs, PBBs, and DDE), possibly at a 106-fold excess over the TCDD and having only fractional dif- ferences in molecular weights from the compounds of interest. Analytical procedures therefore have necessarily involved extensive clean-up to remove potential interference and/or sophisticated detection to achieve both high sensi- tivity and the desired specificity. Early work on TCDDs involved analysis of the commercial herbicides for the individual isomers, and techniques used have included capillary column g.c.- m.s.42 for identification and g.c.-rn.s.-~.i.m.~3-~5 at m/z = 320 for quantification (m/z = 322, although a more intense ion in 2,3,7,8-TCDDwas not used in order to avoid PCB interference).For the analysis of TCDD in Vietnamese fish46 a high-resolution signal-averaging technique has been described which enhances the signal-to-noise levels of the dioxin peak. The sum of 60 scans presented as a high- resolution mass spectrum showed a dramatic increase in sensitivity and signal quality. The detail of the clean-up for dioxins has been scrutinized, and neutral clean-up procedures have been described for milk4' and seafood.4* However, for beef, liver, and milk samples,49 using solvent extraction and column chroma- 40 E. J. McConnell, J. Moore, J. Haseman, and M. Harris, Toxicol.Appl. Pharmacol., 1978, 44, 335. 41 A. Poland and E. Glover, Mol. Pharmacol., 1977, 13, 924. 42 H.-R. Buser, J. Chromatogr., 1975, 114, 95. 43 H.-R. Buser and H. P. Bosshardt, J. Chromatogr., 1974, 90, 71. 44 H.-R. Buser and C. Rappe, Chemosphere, 1978, 7, 199. 45 W. W. Blaser, R. A. Bredeweg, L. A. Shadoff, and R. H. Stehl, Anal. Chem., 1976,48,984. 46 R. Baughman and M. Meselson, Environ. Health Perspect. (Exp. Issue), 1973, 5, 27. 47 P. W. O'Keefe, M. S. Meselson, and R. W. Baughman, J. Assoc. Ofl.Anal. Chem., 1978,61, 621. 48 K. Fukuhara, M. Takeda, M. Uchiyama, and H. Tanabe, Eisei Kagaku, 197521, 318. 49 L. A. Shadoff and R. A. Hummel, Biomed. Mass Spectrum., 1978, 5, 7. Chemical Aspects of Trace Constituents of the Diet. Part III tography clean-up (1OOO:l concentrations for a 10 g sample) of the TCDD and selected mass scanning at m/z 320 and 322, a sensitivity of 10 p.p.t.was achieved,49 and criteria for positive results were taken as (a) correct g.c. retention time and (6) correct ratio 320/322. If an incorrect ratio was observed at low resolution, then the assay was repeated first at medium and then at higher resolution if required. More recently it has been shown50 that, with a clean-up procedure designed specifically to handle samples with high lipid content and excess amounts of chlorinated aromatics (alkaline sample digestion and extraction, sulphuric acid column chromatography, and reversed-phase h.p.1.c. final clean-up), it was possible to attain the desired specificity and sensitivity with packed-column g.c.and with m.i.m. of the molecular-ion cluster of TCDD.50 TCDDs present an example of one of the most demanding assays to be encountered at the present time, one where sophisticated m.s. detection systems (high-resolution s.i.m. and/or low-resolution m.i.m.) are an essential prerequisite, but additionally multi-stage clean-up procedures and great attention to detail are required in order to produce accurate and reliable data. The application of n.i.c.i.-s.i.m. illustrates the increased sensitivity of the technique over e.i.m.s. Of 63 liver assays, 49 were shown to contain measurable amounts of TCDD by n.i.c.i., while only 24 could be detected by e.i.m.s. under similar experimental conditions.51 Still further improvement was predicted for the use of capillary columns.B. Analysis for Nitrosamines in Foods.-Volatile nitrosamines found originally in cured meats52 but more recently in alcoholic beverages53 are generally extrac- ted by an initial distillation, chemically separated from extraneous constituents by a somewhat lengthy clean-up procedure, and then finally concentrated in an organic solvent by perhaps a 1000-fold over the initial concentrations. Analysis is normally by g.c. (packed columns most frequently used) with detection by either specific nitrogen g.c. (Coulson or Hall), chemiluminescence (TEA), or mass spectrometry. Nitrogen-specific Detectors. For nitrosamines, nitrogen-specific detectors are now relegated to the role of preliminary screening of samples, for example in a large-scale survey where it is necessary to minimize the use of m.s.instrument time. For primary measurements it is now generally accepted that nitrogen detectors are not sufficiently specific for nitrosamines, and this can be clearly illustrated from the results of a survey of 154 food samples where about half the positive results for N-nitrosodimethylamine (NDMA), using the Coulson detector, were shown to be false by m.s.54 Hence, the two recognized detectors for nitrosamines are now the TEA and mass spectrometer. 50 L. L. Lamparski, T. J. Nestrick, and R. H. Stehl, Anal. Chem., 1979, 51, 1453. 51 J. R. Hass and M. D. Friesen, Ann. N.Y. Acad. Sci., 1979, 320, 28. 52 T. A. Gough, Analyst (London), 1978, 103, 785.63 B. Spiegelhalder, G. Eisenbrand, and R. Preussman, Food Cosmet. Toxicol., 1979, 17, 29. 54 K. Goodhead and T. A. Gough, Food Cosmet. Toxicol., 1975,13, 307. Gilbert and Self Chemihminescence Detectors. TEAS, in contrast, have considerable selectivity for nitrosamines and were designed exclusively for this particular determination. The TEA is an expensive detector compared to conventional g.c. detectors but is only one fifth of the cost of h.r.m.s. Operator skills are not needed at the same level as for m.s., but the TEA is not a versatile detector -its application is limited exclusively to nitrosamines. The selectivity of the TEA is based on three factors: (i) choice of suitable pyrolysis conditions for N-NO bond cleavage, (ii) selective cold trapping of NO free from potential interfering species, and (iii) NO excitation and chemilumi- nescent emission.Sensitivity is of the order of 50 pg of NDMA injected.55 There have been very few erroneous results; one case was reported of a false positive non-nitrosamine chemiluminescent species,S6 and a second case55 (one out of 98 samples analysed) gave a high NDMA with a TEA but could not be confirmed by m.s. Many comparisons for nitrosamine assays have been made both between use of a TEA and m.s.55$57958 and among the various options available for deter- mining NDMA by s.i.m.58 The low molecular weight (74) of NDMA causes particular problems for s.i.m., and by low-resolution m.i.m., even when, for example, three ions of m/z 30, 42, and 74 are monitored, commonly occurring fragment ions in co-eluted components in food extracts can interfere, causing signal suppression or enhancement compared with the pure standard.For h.r.s.i.m. the molecular ion at 74.0480 for NDMA is present at high relative abundance, but 29SiMe, is a potential interferent (arising both from antifoam tablets and certain types of silicone rubber ~epta),~~*~O and the compound from which this fragment is derived is eluted at a retention time close to that of NDMA. A resolution of greater than 7000 is required for the separation of this and other ions of the same nominal ma~s.~5 The degree of refinement in m.s. methodology which has been achieved for nitrosamine assays is well illustrated by the fact that even the mode of operation of s.i.m.to be employed (peak matching compared with precise ion monitoring) has been strongly argued.58 The current situation appears to be that the high cost of the TEA as a single- purpose detector can be justified, and many are in routine operation for surveillance of nitrosamines in foods and the environment. The clean-up require- ments are less rigorous for the TEA than for m.s., and its specificity of detection has been clearly proven in the field. Nevertheless, when high levels of nitrosamines or unexpected positives are found by TEA, the usual procedure would be to repeat the assay by m.s.(s.i.m.) -not, as normally would be the case, for con- firmation (which presumes m.s.to be superior) but as a cross-check using a detection method based on a different physical principle. 55 K. S. Webb, T. A. Gough, A. Carrick, and D. Hazelby, Anal. Chem., 1979, 51, 989. 56 T. A. Gough, K. S. Webb, and M. F. McPhail, Food Cosmet. Toxicol., 1977, 15, 437. 57 D. H. Fine, D. P. Rounbehler, and N. P. Sen, J. Agric. Food Chem., 1976, 24, 980. T. A. Gough, K. S. Webb, M. A. Pringuer, and B. J. Wood, J. Agric. Food Chem., 1977,25, 663. aeT.A. Gough and K. S. Webb, J. Chromatogr., 1973, 79, 57. 6o C. J. Dooley, A. E. Wasserman, and S. Osman,J. Food Sci., 1973, 38, 1096. Chemical Aspects of Trace Constituents of the Diet. Part 111 C. Mycotoxins.-Microfungal attack on food plants during growth and sub- sequent storage gives rise to contamination with mycotoxins (fungal metabolites), a diverse group of compounds, many of which are highly toxic.These com- pounds are unique in the wide range of techniques which are currently being employed for their analysis, ranging from thin-layer chromatography (t.1.c.) to h.r.s.i.m. T.1.c. (and more recently h.p.t.l.c.)61p62 is perhaps the most commonly used method for routine monitoring of aflatoxins63.64 where, although there is the disadvantage of a lack of quantitative precision (coefficients of variation range from 15 to 70 %),65 the native fluorescence of certain aflatoxins allows a detection limit of 0.5 nge6 (under long-wavelength U.V. light). The method is inexpensive and is suitable for the screening of large numbers of samples. H.p.1.c.has tended to replace t.1.c. for mycotoxin analysi~67~~~ because the associated techniques of silica-gel packed ceWg and laser-fluorescence detection70171 offer improved sensitivity and precision, but the rather low selectivity makes extensive sample clean-up compulsory. R.i.a. techniques are said to offer the advantage over t.1.c. and h.p.1.c. of needing less sample clean-up (ideal for routine survey or quality-control work). Using a simple solvent extraction, aflatoxin BI has been determined in foods with good reproducibility (14-16 7372973 and a sensitivity of the order of 1 p.p.b.73 Most mycotoxins are too involatile for g.c., but g.c. and g.c.-m.s. have found application for the appropriate derivatives of trichothecenes74~75 and pat~lin'~ in foods.M.s. has found extensive application for mycotoxin analysis and has been reviewed elsewhere.77 However, the point to emphasize regarding m.s. (and the theme of this review) is the possibility of carrying out analyses for mycotoxins by m.s. directly on foods or on crude extracts. Some published has shown that by direct insertion probe m.s., where the limitation to sensitivity was the sample capacity of the probe (100 pg), it was possible to detect by h.r.s.i.m. 61 K. Y. Lee, C. F. Poole, and A. Zlatkis, Anal. Chem., 1980, 52, 837. 6e P. A. Biondi, L. Gavazzi, G. Ferrari, G. Maffeo, and C. Secchi, J. High Res. Chromatogr. Chromatogr. Commun., 1980, 3,92. 6s 0.L. Shotwell and M. L. Goulden, J. Assoc. Off.Anal.Chem., 1977, 60, 83. 64 A. E. Waltking, J. Assoc. Off.Anal. Chem., 1970, 53, 104. 65 S. Nesheim, 'Trace Organic Analysis: A New Frontier in Analytical Chemistry', NBS Spec. Pub]. 519, ed. H. S. Hertz and S. N. Cheder, Washington, 1979, p. 355. 66 T. R. Romer, J. Assoc. Off.Anal. Chem., 1975, 58, 500. 67 G. M. Ware and C. W. Thorpe, J. Assoc. Off.Anal. Chem., 1978, 61, 1058. D. C. Hunt, A. T. Bourdon, P. J. Wild, and N. T. Crosby, J. Sci. Food. Agric., 1978, 29, 234. 6sT.Panalaks and P. M. Scott, J. Assoc. Off.Anal. Chem., 1977, 60, 583. 70 G. J. Diebold, N. Karny, R. N. Zare, and L. M. Seitz, J. Assoc. Of.Anal. Chem., 1979, 62, 564. 71 G. J. Diebold, N. Karny, and R. N. Zare, Anal. Chem., 1979, 51, 67. 7B W. 0. Harder and F.S. Chu, Abstr. Annu. Meet. Am. SOC. Microbiol., 1979, 79, 204. 73 P. S. Sun and F. S. Chu, J. Food Safety, 1977, 1, 67. 74 R. M. Eppley, J,Assoc. Of.Anal. Chem., 1979, 56, 824. 75 Cs. Szathmary, J. Galacz, L. Vida, and G. Alexander, J. Chromatogr., 1980, 191, 327. 76 J. D. Rosen and S. R. Pareles, J. Agric. Food Chem., 1974, 22, 1024. 77 R. Self, Biomed. Mass Spectrom., 1979, 6, 361. W. H. Haddon, M. S. Mastri, G. Randall, R. H. Elsken, and B. J. Meneghelli, J. Assoc. Off. Anal. Chem., 1977, 60,107. Gilbert and Self (resolution = 5000) 0.03 ng of aflatoxin B1. Surprisingly the limit of detection was better in extracts than standard solutions owing to irreversible bonding effects on glass surfaces. Finally, 1.c.-m.s. is a technique with enormous potential for application to the field of mycotoxins.9 Concluding Remarks In this review we have attempted to make an appraisal of the role of m.s. in trace organic analysis by comparing its sensitivity and specificity with that attainable by other techniques and to examine some of the recent advances in instrumenta-tion. We have used some examples from areas of food chemistry where m.s. has been proved to be useful at subnanogram levels, e.g. in the analysis of aflatoxins, dioxins, and nitrosamines. For the future it is clear that n.i.c.i. will be widely applied for the ultra-sensitive detection of trace electron-capturing components, and examples are already emerging from the literature. L.c.-m.s. coupling is very much in its infancy, and not until more of the technical difficulties have been resolved will its full potential in food chemistry be realized. However, once established, it will not only enable identification of the less volatile components separated from mixtures, but it will open new frontiers in areas of trace analysis where 1.c. has been inapplicable through lack of a suitable detector. Further ahead still is m.s.-ms., an interesting academic tool so far, but potentially one which could lead the way to high specificity in trace quantitative measurements made directly on foods.

ISSN:0306-0012    

DOI:10.1039/CS9811000255

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

IV Nutritional Chemistry of Inorganic Trace Constituents in the Diet By J. K. Chesters ROWETT RESEARCH INSTITUTE, BUCKSBURN, ABERDEEN Although improvements in techniques have greatly increased the precision with which the inorganic trace constituents of a diet can be determined,l nutritional studies have clearly shown that analytical values for individual elements are not sufficient to allow one to predict whether an animal would be able to obtain adequate amounts of these elements from the diet.2J Similarly, it is often difficult to predict the deleterious effects of toxic trace elements merely by determining their dietary concentrations.4 These difficulties are now known to result from interactions between the elements, from differences in their chemical forms, and from influences exerted by other dietary constituents. A well known example of the influence of the chemical form of an element is the difference between the biological potency of Co as an inorganic salt and as a component of vitamin B12.Only as the latter compound will Co fulfil its essential role for most animals, and therefore estimates of total dietary Co intake are insufficient. The only exception to this is for ruminant animals where the micro- organisms of the rumen are able to convert Co to vitamin B12 and supply the animal’s needs. Problems arise even in this situation, however, because a variable proportion of the dietary Co is converted into analogues of vitamin BIB which are inactive in the host animal.5 A somewhat similar situation seems to be emerging with regard to Cr nutrition of most animals including man.Current research suggests that Cr is required in the form of a specific organic complex, glucose tolerance factor, for it to be fully active in potentiating the functions of insulin.6 The precise structure of this complex is unknown7 but appears to represent a variable fraction of the total Cr content of a diet. Analyses of the latter therefore C. J. Pickford, Chem. SOC.Rev., 1981, 2, 245. I. Bremner and N. T. Davies, Rep. Rowett Znst., 1973, 29, 126. N. T. Davies, Proc. Nutr. SOC., 1974, 33, 293. N. T. Davies, Proc. Nutr. SOC., 1979, 38, 121. E. J. Underwood, ‘Trace Elements in Human and Animal Nutrition’, 4th edition, Academic Press, London, 1977, p. 146.R. A. Anderson, M. M. Polansky, J. H. Brantner, and E. E. Roginsk in ‘Trace Element Metabolism in Animals’, ed. M. Kirchgessner, Arbeitskreis fur Tierernahrungsforschung, Weihenstephan, West Germany, 1978, p. 269. W. Mertz, R. A. Anderson, W. R. Wolf, and E. E. Roginski in ‘Trace Element Metabolism in Animals’, ed. M. Kirchgessner, Arbeitskreis fur Tierernahrungsforschung, Weihenste-phan, West Germany, 1978, p. 272. 270 Chesters provide inadequate information to assess the adequacy of the diet with regard to Cr.8,gt10 The above examples of a nutritional requirement for a specific pre-formed complex appear to be exceptional. However, most of the trace metals tend to form complexes with organic constituents of the diet, and these frequently render the metals insoluble and not available for absorption.Entry of these complexes into the acid medium of the stomach often results in dissociation of the dietary complexes. However, subsequent neutralization of the digesta during passage down the small intestine results in recombination of the metals with dietary ingredients and digestive juices leading to the formation of a whole new range of complexes.11J2 In sheep, for example, the low pH of the stomach resulted in the expected release of Zn and Mn into the soluble fraction of the digesta, but the solubility of Cu actually decreased.12 It is possible that this resulted from the formation of insoluble cupric sulphide when both the cupric and sulphide ions were liberated from previously soluble complexes by the acid of the abomasum.In man a mucopolysaccharide secreted by the stomach binds Fe liberated from dietary complexes by the fall in pH and appears to retain the Fe in a soluble and available form during subsequent neutralization of the dige~ta.1~ The formation of a complex salt during neutralization of the digesta also appears to underlie the profound effects of dietary phytic acid [myo-inositol 1,2,3,4,5,6-hexakis(dihydrogenphosphate)] on the availability of Zn, Cu, and Mn from diets.l* Diets based on maize and soya-bean products, which are rich in phytic acid, have in the past caused parakeratosis, symptomatic of Zn defi- ciency, in pigs.15 Delayed sexual maturation in man in the Middle East has also been associated with Zn deficiency caused by a staple diet of unleavened bread rich in phytic a~id.1691~ At neutral to alkaline pH in conditions similar to those in the gut, phytic acid forms insoluble salts containing predominantly Ca but also Zn, Cu, and Mn,l8 These salts appear to pass relatively unchanged through the absorptive regions of the gut and thus render the associated trace metals un- available.For any given Ca concentration the extent of precipitation of Zn depends on its molar ratio to the phytic acid, and the ratio of phytate to Zn has also been shown to control the availability to rats of Zn from diets with relatively 8k.M. Hambidge, J. Hum. Nutr., 1978, 32, 99. R. A. Anderson, .I.H. Brantner, and M. M. Polansky, J.Agric. Food Chem., 1978,26, 1219. lo C.T.Gurson, Adv. Nutr. Res., 1977, 1, 23. l1 I. Bremner and A. H. Knight, Br. J. Nutr., 1970, 24, 279. l2I. Bremner, Br. J. Nutr., 1970, 24, 760. l3 A. Jacobs and P. M. Miles, Br. Med. J., 1969, 4, 778. l4N.T. Davies and R. Nightingale, Br. J. Nutr., 1975, 34, 243. l6 P. K. Lewis, W. G. Hoekstra, R. H. Grummer, and P. H. Phillips, J. Anim. Sci., 1956, 15, 741. l6 J. A. Halsted, H. A. Ronaghy, P. Abadi, M. Haghshenass, G. H. Amirhakemi, R. M. Barakat, and J. G. Reinhold, Am. J. Med., 1972, 53, 277. l7 J. G. Reinhold, A. Parsa, N. Karimian, J. W. Hammuk, and F. Ismail-Beigi, J. Nutr., 1974, 104, 976. N. T. Davies and S. E. Olpin, Br. J. Nutr., 1979, 41, 590. 271 Chemical Aspects of Trace Constituents in the Diet.Part IV Table 1 Efect of increasing concentrations of phytic acid on the solubility oj'Zn at pH 6.3 in vitro and on the growth of ratsa Phytate: Zn Zn remainingb Average daily weight gain& ratio soluble at pH 6.3 6 g Ca kg-l/g 12 g Ca kg-l/g in vitro/% 0 I 5.76 k 0.27d 5.78 & 0.48d 10 2.08 5.01 & 0.15 6.34 k 0.26 15 1.85 5.81 k 0.37 5.60 k 0.23 20 1.63 5.13 k 0.16 4.11 k 0.20 30 1.40 4.94 k 0.18 3.16 f 0.23 Data from ref. 18. b The solutions contained 2.4 g Ca I-' and 7.2 mg Zn 1-l. c The diets contained 6 or 12 g Ca kg-I and 18.5 mg Zn kg-l. d Values are means & S.E. high Ca contents.l8Jg However, in assessing the availability of Zn from diets containing phytic acid, the Ca content of the ration is at least equally important.20 Thus parakeratosis in pigs was severely aggravated by an increase in the Ca content of their ration from 8 g kg-1 to 12 g kg-1 15 and, in rats, phytate:Zn ratios that severely impaired growth on diets containing 12 g Ca kg-I failed to influence it when the dietary Ca concentration was only 6 g kg-1.18 Furthermore a recent study has shown that even moderately high phytate:Zn ratios can be tolerated when the total Zn present is substantially above the minimum required in the absence of phytate, and the Ca intake is not excessive.20 These findings together suggest that equilibria are established within the gut between soluble Ca ions and phytic acid which result in the precipitation of a portion of the latter as the insoluble Ca salt.The availability of the trace metals present depends on the extent to which these divalent metals substitute for Ca in the precipitated p h yt a tes . Interest in the potential influence of dietary phytates on the trace-metal status of man has increased recently with the trend towards substitution of meat by products based on processed soya-bean protein. Until recently meat has been a major dietary source of readily available trace metals, whereas the highly processed, textured vegetable proteins tend to have relatively low concentrations of Zn and are rich in phytic acid.21 Furthermore animal studies of phytic acid metabolism have shown that the enzyme responsible for partial degradation of phytic acid in the gut, phytase, appears to be Zn dependent.22 Any reduction in availability of Zn caused by phytic acid in the diet may therefore reduce the activity of phytase in the gut, increase the proportion of the dietary phytate, which remains undegraded, and thus aggravate the reduction in Zn availability.Nevertheless the importance of phytic acid in determining man's trace-element status may be less than animal studies have suggested because of the relatively lS D. Oberleas, Pruc. West. Hemisph. Nutr. Cungr., 1975, IV, 156. *O E. R. Morris and R. Ellis, J. Nutr., 1980, 110,1037. 81 N. T. Davies and H. Reid, Br. J. Nutr., 1979, 41, 579. *p N. T. Davies and A. Flett, Br. J. Nutr., 1978, 39, 307. Chesters low Ca intakes of man23 and the general diversity of ingredients within single meals.Furthermore whole-meal bread, which can be a major source of phytic acid in human diets, has been shown to have a higher Zn content than white bread.23a Therefore, although the fractional absorption of dietary Zn was less from diets containing whole-meal bread, the amount of Zn absorbed was actually greater than when white bread was eaten. In addition to phytic acid the fibre components of plant products have been implicated in binding the essential trace elements. Studies with rats showed a 32 %decrease in the fractional absorption of Zn when 6 %cellulose was added to a semi-synthetic diet,24 and 14.2 g cellulose/d added to the diet of a group of adol-escent boys increased by over 30 % the faecal excretion of Zn and CU.~~ Further-more, when diets supplying close to the recommended daily allowances of Zn and Cu were offered to men, they remained in positive balance for these elements if the diets were low in fibre but showed a net loss of Zn and Cu when the diets were rich in fibre from fruit and vegetables.26 Similarly supplementation of human diets with hemicellulose markedly reduced the fractional retention of Zn.27 Bacterial cell walls have also been found to bind trace element@ and may be partially responsible for an observed increase in trace-element requirements in conventionally reared animals relative to their germ-free equivalent^.^^ In the walls of Baciffussubtifis the metals were bound to carboxyl groups,3O but a wider investigation of the nature and degree of binding to bacterial cell walls in general will be needed before their potential influence on trace-metal availability can be assessed.Although the formation of complexes between trace inorganic constituents and the organic components of diets frequently reduces the availability of the inor- ganic components, these complexes can also enhance availability. Mention has already been made of the situations where these complexes remain intact during absorption and function as vitamins within the animal, but other complexes appear to be important only during the actual processes of absorption. Although most biological forms of Fe appear to be equally available to animals, Fe present in the diet as part of the haem molecule is preferentially absorbed.31 Typical values for the fractional absorption of Fe have been reported as 20-25% of haem Fe and only 3-8% of non-haem Fe.Calculations of the adequacy of Fe as ‘Household Food Consumption’, Ministry of Agriculture and Fisheries, London, 1980. B. Sandstrom, B. Arvidsson, A. Cederblad, and E. Bjorn-Rasmussen, Am, J. Clin. Nutr., 1980, 33, 739.** W. M. Becker and W. G. Hoekstra in ‘Intestinal Absorption of Metal I,ons, Trace Elements and Radionuclides’, ed. S. C. Skoryna and D. Waldron-Edward, Pergamon Press, Oxford and New York, 1971, p. 229. Is L. M. Drews, C. Kies, and H. M. Fox, Am. J. Clin. Nutr., 1979, 32, 1893. Ie J. L. Kelsay, R. A. Jacob, and E. S. Prather, Am. J. Clin. Nutr., 1979, 32, 2307.x7 C. Kies, H. M. Fox, and D. Beshgetor, Cereal Chem., 1979, 56, 133. C. F. Mills, Soil Sci., 1958, 85, 100. J. C. Smith, E. G. McDaniel, L. D. McBean, F. S. Doft, and J. A. Halsted, J. Nutr., 1972, 102, 711. so R. J. Doyle, T. H. Matthews, and U. N. Streips, J. Bacteriol., 1980, 143, 471. 31 E. R. Monsen, L. Hallberg, M. Layrisse, D. M. Hegsted, J. D. Cook, W. Mertz, and C. A. Finch, Am. J. Clin. Nutr., 1978, 31, 134. Chemical Aspects of Trace Constituents in the Diet. Part IV intake of man based on dietary analyses must therefore take into account the proportion of the element present as haem. However, Fe absorbed in this form appears to be liberated from the porphyrin moiety within the mucosa and thereafter enters the same metabolic pool in plasma, as does the dietary non- haem Fe.32 Absorption of non-haem Fe occurs preferentially in the ferrous form and is markedly enhanced by ascorbic acid in the diet.33 The latter appears to act both as a reducing agent and as a chelator of Fe within the gut.A wide range of other low molecular weight substances capable of forming chelates with Fe also facili- tate absorption of non-haem Fe. Among these is citric acid, which has also been suggested to be involved in the absorption of Zn from milk.5s34 Milk diets have long been recognized to result in relatively high fractional absorptions of many essential and toxic trace elements.35~~~ In many instances, however, the animals receiving these diets have been young, and the effects of milk on absorption have been confounded by those of age, relatively high frac- tional absorptions of the trace metals being characteristic of the neonates of many species.35936 However, the studies by Kostia136 have shown that, even with Table 2 Influence of age and milk diet on trace-metal absorptiona AbsorptionJZ of oral dose Element Sucklings Weaned animals Milk diet Standard diet Pb 52 23 0.4 Cd 26 7 0.5 Mn 40 6 0.05 a Data from ref.36. previously weaned rats, a return to a liquid milk diet increased the retention of oral doses of Pb, Cd, and Mn. This effect is largely dependent on milk being the sole nutrient. When fed with solid diets its effects have been much less dramatic, and in some instances milk has even reduced the uptake of trace metal~.3~ Furthermore, the availability of trace elements from milk may vary with the species of origin of the milk.Thus there is a rare inherited disease of man, acro- dermatitis enteropathica, associated with an inability to absorb adequate amounts of Zn from normal diets. Generally the sufferers are able to obtain adequate "L. R. Weintraub, M. B. Weinstein, H.-J. Huser, and S. Rafal, J. Clin. Invest., 1968, 47, 531. 33 J. D. Cook and E. R. Monsen, Am. J. Clin. Nutr., 1977, 30, 235. 34 B. Lonnerdal, A. G. Stanislowski, and L. S. Hurley, J. Znorg. Biochem., 1980, 12, 71. ssC. F. Mills and N. T. Davies in 'Development of Mammalian Absorptive Processes', C.I.B.A. Symposium No. 70, Excerpta Medica, Amsterdam and Oxford, 1979, p.247. 36 K. Kostial, D. Kello, S. Jugo, I. Rabar, and T. Maljkovic, Environ. Health Perspec?., 1978, 25, 81. 37 J. Quarterman and E. Morrison, Proc. Nutr. Sac., in press. Chesters amounts of Zn while receiving human milk but develop the symptoms of the disease if transferred to cow milk or solid diets.38 This does not result from simple differences in Zn content since human milk generally contains a lower concentra- tion of Zn than bovine milk and most normal diets. Early comparisons of human and bovine milk showed that despite the former’s lower total Zn content it had a higher concentration of Zn associated with the low molecular weight fraction. This led to the hypothesis that the high availability of Zn from human milk was caused by the presence of a low molecular weight chelator which facilitated abs~rption.~~One group has suggested that citric acid is the critical compound,34 yet bovine milk contains higher concentrations of citric acid than human milk,40 and rat milk, which is essentially devoid of citric acid, gave the highest availability of Zn when this was tested with rat~.~1 Picolinic acid, a metabolite of tryptophan, has also been suggested,42 but Cousins40 considered that none of these compounds has sufficiently high binding constants for Zn to be significantly better chelators of it than many of the amino-acids present.His view was that the higher protein concentration of bovine milk results in most of the Zn being bound to high molecular weight components and that the Zn in low molecular weight form in human milk is not bound to any unique component.However, overshadowing the above controversy is the lack of convincing evidence that the differences in availability actually depend on the nature and proportions of the Zn in low molecular weight complexes. Furthermore a recent study has shown that lacto- ferrin, a protein present in much higher concentration in human milk than in bovine, also binds Zn43 The presence of species-specific proteins which facilitate trace-metal absorption from milk cannot be ruled out. The influence of milk on trace-metal absorption clearly needs further investigation. It may even be found that much of the effect of milk on absorption is caused by its high digestibility resulting in relatively little faecal residues and therefore little competition for trace metals between non-digested solids and the animals’ transport systems.Trace-metal availability is influenced by compounds secreted into the gut as well as those derived from the diet, and the influence of gastric mucopoly- saccharides on Fe absorption has already been mentioned. The absorption of Pb depends markedly on bile components and was reduced ten-fold when bile flow was diverted from the gut of rats.44 There have also been reports of low molecular weight ligands able to bind trace metals in the lumen of the gut and in the cytosol of mucosal cells.45~46 These ligands have been variously characterized, but present evidence suggests that they may be artefacts derived from a metal binding pro- 30 V. M.de Kaloustian, s. s. Musallan, s.A. Sanjad, A. Murib, w. D. Hammad, and Z. H. Idriss, Am. J. Dis. Child., 1976, 130, 421. L. S. Hurley, J. R. Duncan, M. V. Sloan, and C. D. Eckhert, Proc. Nut. Acad. Sci. U.S.A., 1977,74, 3547. 40 R. J. Cousins and K.T. Smith, Am. J. Clin. Nutr., 1980, 33, 1083. 41 P. B. Johnson and G. W. Evans, Am. J. Clin. Nutr., 1978, 31,416. IaG. W. Evans and P. E. Johnson, Fed. Proc., Fed. Am. SOC.Exp. Biol., 1979,38, 703. 43 E. W. Ainscough, A. W. Brodie, and J. E. Plowman, Am. J. Clin. Nutr., 1980, 33, 1314. 44 J. Quarterman, J. N. Morrison, and W. R. Humphries, Proc. Nutr. Soc., 1977, 36, 103A. 46 J. R. Duncan and L. S. Hurley, Am. J. Physiol., 1978, 235, E556.I6B. R. Schricker and R. M. Forbes, Nutr. Rep. Znt., 1978,18, 159. Chemical Aspects of Trace Constituents in the Diet. Part IV tein, metallothionein, which has been partially degraded by digestive enzymes during the preparation of extract^.^'^^^ As well as containing naturally occurring binding agents diets have frequently been supplemented with chelators or chelated trace metals with the intention either of increasing the availability of dietary trace minerals or of supplying available forms of the trace metal~.*~-51 The addition of EDTA to phytate-rich diets was found to increase the availability of Zn to turkeys,52 and in subsequent studies with both turkeys and chicks several chelating agents were investi-gated5O~~~(see the Figure).The diets fed to the turkeys contained phytate-rich 2 0 wk I *2-I 4 STABILITY CONSTANTS FOR Zn Figure Relation of stability constant for Zn of several chelating agents to their growth- promoting efects in turkey poults when added to a Zn-deficient diet (Reproduced with permission from J. Nutr., 1964, 82,249) soya-bean protein and only 26 mg Zn kg-1. They were inadequate in available Zn as shown by a major response in growth to Zn supplementation. Adding to the basal diet chelators with log stability constants for Zn from 13 to 17 resulted in growth responses similar to those obtained with Zn. Chelators of lower or higher affinity for Zn produced lesser growth responses. It was suggested that below the critical range of affinities the chelators were unable to sequester Zn from the dietary ingredients while above the range Zn removed from the diet by the chelator was so firmly bound that it was unavailable to the animal and was 47 R.J. Cousins, Nutr. Rev.,1979, 37, 97. R. J. Cousins, K. T. Smith, M. L. Failla, and L. A. Markowitz, Life Sci., 1978, 23, 1819. 40 C. L. Keen, B. Lonnerdal, M. V. Sloan, and L. S. Hurley, J. Nutr., 1980, 110, 897. so F. H. Nielsen, R. L. Sunde, and W. G. Hoekstra, J. Nutr., 1966, 89, 35. slN. W. Solomons, R. A. Jacob, 0. Pineda, and F. E. Viteri, J. Nutr., 1979, 109, 1519. s1 F. H. Kratzer, J. B. Allred, P. N. Davis, B. J. Marshall, and P. Vohra, J. Nutr., 1959, 68, 313. b8 P. Vohra and F. H. Kratzer, J. Nutr., 1964, 82, 249.Chesters excreted along with the chelator. Within the critical range Zn removed from dietary ingredients by the added chelator was still sufficiently labile to be taken up by the animals' tissues and be used in its essential metabolic roles. The organic components of a diet can also influence the utilization of trace inorganic constituents even after they have been absorbed. Thus it has been known for many years that a range of substances, the goitrogens, will impair iodine metabolism. Research has shown that they can be divided into two groups according to their mode of action. Those of the cyanogenetic type inhibit uptake of I by the thyroid gland and are thought to act by giving rise to thiocyanate or isothiocyanate ions which are of similar ionic size to the iodide ion and compete with it for the iodide transport system of the gIand.S4 The other main group of goitrogens, the thio-oxazolidones, prevent iodination of tyrosine residues within the gland, presumably by inhibiting iodide peroxidase, and their effects are not generally reversible by increasing the I content of the diet.A wide range of mutual interactions between the trace metals also influence their availability. Hill and Matrone55 first advanced the hypothesis that ions having similar electronic structure might compete with each other in metabolic pathways. High concentrations of Zn in the diets of chicks reduced growth rate and survival, and these effects were found to be reversible on increasing the Cu content of the diet.55 Similar interactions have also been observed when Cd was added to the diet.56 Subsequent studies have confirmed many of the postu- lates of Hill and Matrone and have shown that interactions can occur between Cu, Zn, and Cd and between Pb and Ca and Mo and W.In many instances the underlying mechanisms have still to be clarified, but the induction of Cu defi-ciency by high dietary concentrations of Zn or Cd is better understood. This appears to originate from the induction by Zn or Cd of a protein, metallothionein, in the gut muc0sa.5~ Cysteine residues account for almost 30% of the amino- acids in metallothionein, which acts as storage protein for several types of diva- lent cation. Even when induced by one cation such as Zn2+ or Cd2f the protein is still capable of binding other metal ions such as Cu2+ whose passage through the mucosal cells towards the bloodstream is thus interrupted.Since the normal functioning of the mucosal epithelium results in sloughing of the cells as they reach the tip of the villus, any metals still bound to metallothionein are returned to the lumen of the gut. High dietary concentrations of Zn and Cd therefore interfere with the absorption of Cu. Another likely source of interaction results from competition of analogous ions for transport systems. It seems likely that such a mechanism underlies the antagonistic effects of the s0g2-,M004~-,and wo42-ions.58 Interactions between trace metals occur not only during absorption but also within the body.Thus high dietary concentrations of Ca reduced the fraction of 54 N. T. Davies, Proc. Nutr. Soc., 1979, 38, 121. 16 C. H. Hill and G. Matrone, Fed. Proc., Fed. Am. Soc. Exp. Biol., 1970, 29, 1474. 66 I. Bremner and J. K. Campbell, Environ. Health Perspect., 1978, 25, 125. &' A. C. Hall, B. W. Young, and I. Bremner, J. Znorg. Biuchem., 1979, 11, 57. 68 C. J. Cardin and J. Mason, Biochim. Biophys. Acta, 1976, 455, 937. Chemical Aspects of Trace Constituents in the Diet. Part IV Table 3 Eflect of dietary Zn concentration on the absorption of an oral dose of 64Cua Dietaryb Zn concentrationlm Fraction of dose of g Wu absorbed1 % 64Cuon metallo-thionein in mucosal Concentration of Zn on metallo- kg-1 cpm mg-l thionein in mucosal PS g-l 30 43 10 0.9 150 39 11 1.6 450 43 17 3.4 900 27 63 24.4 a Data from ref.57. b All diets contained 3 mg Cu kg-l. Pb absorbed by rats but also reduced the rate of release of Pb from the body once Pb was removed from the diet.59pgo The interaction during absorption probably resulted from competition for transport proteins since low dietary Ca intakes stimulated Pb absorption and vitamin D affected Pb uptake in a manner similar to its effect on Ca transport.61 However, the influence of Ca on the retention of Pb, which had been previously absorbed, was less readily explicable since both low and high intakes of Ca increased the retention of Pb relative to diets con- taining normal amounts of Ca.59.60 Once trace inorganic constituents have been absorbed they may be temporarily stored in organs such as the liver prior to being utilized for their essential func- tions or being excreted.During each of these processes their fate is often in- fluenced by interactions with other essential or toxic elements. For example, high dietary Cd concentrations tend to increase liver Zn storage as a result of induction of metallothionein synthesis in the liver,62 and the toxic effects of Cd on the testis can be reduced by increasing the Zn intake of the animal.63 These and many other aspects of research into the metabolism of trace elements merit closer examination by chemists from all branches of the science. The history of attempts to understand the influence of milk on trace-metal absorption can be used to illustrate this.In order to identify factors which influence absorp- tion, milk must be separated into its components. The first approach has been to fractionate it on the basis of particle size, but attempts to use gel exclusion chro- matography or molecular filters have been hindered by the inherent ion-exchange properties of these materials. Although slight and often stated to be insignifi- cant, metal binding by these systems is often excessive for satisfactory separations of the trace metals present in biological samples. Chemical modifications devised to reduce these binding affinities would be invaluable. When separated on the basis of particle size, milk contains complex micelles J.Quarterman and J. N. Morrison, Br. J. Nutr., 1975, 34, 351. 'O J. Quarterman, J. N. Morrison, and W. R. Humphries, Environ. Res., 1978, 17, 60. I1C. M. Smith, H. F. DeLuca, Y.Tanaka, and K. R. Mahaffey, J. Nutr., 1978, 108, 843. IaM. D. Stonard and M. Webb, Chem.-Biol. Interact., 1976, 15, 349. J. Parizek, J. Endocrinol., 1957, 15, 56. Chesters accounting for much of its protein and lipid. A substantial proportion of the metals present is also in these micelle~.~~ A better understanding of their structure and physical chemistry and of the processes governing partition of metals between aqueous, lipid, and insoluble phases would aid assessment of their importance in modifying metal absorption. A recurrent theme in studies of the trace metals has been their binding to the other components present in biological systems, for example the low molecular weight constituents of milk.Attempts to assess the relevance of this binding have been hindered by the lack of mathematical methods for assessing the complex interactions of binding species. Even where such methods are being developed their usefulness has often been limited by lack of suitable data. There is still a great need for estimates of metal-binding affinities which have been determined under physiologically relevant conditions of pH, ionic strength, and temperature. Methods are also needed for including in simulations of binding equilibria the high molecular weight, multi-ligand compounds such as proteins, which play such an important role in biological situations.Finally, although analytical methods for the trace elements have improved immensely, there is still a need for ultra-micro methods of analysis. At present, metal concentrations generally have to be determined by methods which tend to average out variations in concentration within a tissue or cell. In many instances these structures must be disrupted and fractions pooled to obtain adequate quantities for analysis. The consequent disruption of natural compartmental barriers provides ample opportunity for redistribution of the metals in vitro and causes uncertainty as to the relevance of the results to the situation in vivo. Methods such as electron-probe microscopy and laser-probe atomic absorption or fluorescence applied to frozen-dried tissue preparations would, if developed to a sufficient degree of sensitivity, greatly aid our understanding of the biological location and role within animals of the trace inorganic constituents of diets. 64 J. E. Piletz and R.E. Ganschow, Am. J. Clin. Nutr., 1979, 32,275.

ISSN:0306-0012    

DOI:10.1039/CS9811000270

出版商:RSC    

年代:1981

数据来源: RSC