摘要:

QUARTERLY REVIEWS ACETYLEMC COMPOUNDS AS NATURAL PRODUCTS By JOHN D. Bu’Loc~ Ph.D. (UNIVERSITY OF MANCHESTER) IN the past ten to twenty years the study of natural products has amply demonstrated the versatility of biochemical systems by bringing to light an enormous diversity of chemical structures not only in biologically functional compounds (coenzymes hormones etc.) but also in that wider group of “ secondary metabolites ” compounds whose metabolic role if they have any remains obscure. I n the long list of instances which could be compiled there are many new compounds without close parallels or with only a few near relatives but there are also a few new categories of natural products in which common features recur in numerous examples from a variety of sources. Compounds containing carbon-carbon triple bonds appear to form one such new category and the study of the naturally derived acetyl- enes has reached a stage a t which attempts to review the field may be profitable.Though the list of types of natural acetylenes is almost certainly incomplete it is sufficiently long to warrant some provisional ordering and an effort to relate that ordering to the general background of metabolic studies. This account of the natural compounds is therefore drawn up in a way which it is hoped will call attention to the wider aspects of the problem. It may well be that the study of natural acetylenic compounds will play an important part in linking studies of primary and secondary metabolism for whilst they are quite clearly exotic substances special features in the general pattern of Nature they seein to be constructed on rather simple plans unlike e.g.the alkaloids or sapogenins. In this sense they are the simplest of t.he more complicated natural products. Historical The first record of the natural occurrence of what are now known to be acetylenic derivatives appears to be the observation by Bretz and Elieson (1826) of the ready crystallisation of the essential oil of Artemisia vulgaris. Carthaus (1907-1910) isolated a compound C,,H,,O from Artemisia sp. but the constitution of A . vulgaris oil was not thoroughly investigated until 1 Bretz and Elieson cited in Gildmeister and Hoffmann “ Die atherischen Ole ” Carthaus Jaarb. Dep. Landb. in Ned.-Indie Batavia 1907 66 ; 1910 65 ; cf. also Schimmel Miltitz 1931 Vol. 111 p. 1018. ref. 1. A A 37 1 372 QUARTERLY REVIEWS the work of Stavholt and Sorensen 3 in 1950.Meanwhile credit for the first characterisation of a natural acetylene goes to Arnaud who isolated tariric acid (I) in 1892 and was able to determine its structure ; it was not CH,*[CH2]10*C=C.[CH,]4*C02H (I) synthesised until 1952.5 At about the same time Hebert described unstable fatty acids from isano oil the composition of which is still not fully understood. Semmler and Ascher had earlier isolated " Carlina oxide " from CarZinu acauZis for which they proposed an allenic structure on the basis of its molecular refractivity.' The correct acetylenic structure (11) was established (and synthesised) in 1935 by Pfau et al. whose paper provides an interesting example of the early use in structure studies of from vibrational spectra.The first natural product to be characterised as a polyacetylene with more than one C-C unit) was the lachnophyllum ester (111) isolated by Willja,ms Goljmov and Smirnov in 1935 and the recent increase in data (i.e. importance of the natural acetylenes arises primarily from work in Norway CH,CH,*CH,*CrC*C~CCH=CH-CO,Me (111) by the Sorensens and their collaborators on this and other constituents of Cornposit= beginning in 1941 and resumed after the war. Professor Sorensen made a useful summary l o of some aspects of this work in 1953. Parallel with this development there was rapid progress in the pure chemistry of acetylenic compounds particularly relevant in the present connection being the work of Jones Whiting and their eo-workers on the synthesis and properties of polyacetylenic compounds 11 and the similar work of Bohlmann.l2 At first this work assisted the study of na'tural polyacetylenes mainly by providing comparative data on spectroscopic properties ; as more powerful synthetic methods were developed it became possible to use the synthesis of possible structures as a direct method of establishing the identity of a natural product. A recent development made possible by the full use of modern methods has been the study of polyacetylenic compounds in fungi. Antibiotics later shown to contain triple bonds and allene units were described by Anchel Stavholt and Sorensen Acta Chem. Xcand. 1950 4 1567. Arnaud Compt. rend. 1892 114 79 ; 1896 122 1000 ; 1902 134 473 547 842. Hebert Bull. SOC. china. France 1896 15 935 941. 5 Lumb and Smith Chem.and Ind. 1952 358 ; J. 1952 5032. 7 Semmler Chem. Ztg. 1889 13 1158 ; Ber. 1906 39 726 ; Semmler arid Ascher 8 Pfau Pictet Plattner and Susz Helw. Chim. Acta 1935 18 935 ; cf. also Gilman Ber. 1909 42 2355. van Ess and Burtner J. Amer. Chem. SOC. 1933 55 3461. Willjems Smirnov and Goljmov Zhur. obshchei Khim. 1935 5 1195. lo Sorensen Chem and Ind. 1953 240. l1 Bowden Heilbron Jones and Sargent J. 1947 1579 and subsequent papers. la Bohlmann Chem. Bey. 1951 84 545 and subsequent papers. BU’LOCK NATURAL ACETYLENES 373 and her co-workers l3 in 1950 and in 1952 Celmer and Solomons 1* elucidated the remarkable structure of the antibiotic mycomycin (IV) the first naturally occurring optically active allene to be fully characterised. The appropriate fungi promise to be the most useful material for the study of the interesting problems of the biogenesis of the natural acetylenes.H-CEXXEC *CH=C -CH-CH=CH *CH-CH *CH2.C02H (IV) Previous reviews of this subject include summaries by Sorensen lo and Anchel l5 of work on the Composits and Fungi respectively and an article by Bohlmann l6 which is especially useful for its summary of the synthetic methods applied in this field. The latter have also been described more generally by Jones and others.17 Distribution An acetylenic component was detected 18 in a fraction b.p. 95- 105O/12 mm. from niethylated butter-fat acids by the characteristic Raman bands (ca. 2040 and 2230 cm.-l) ; apart from this the occurrence of acetylene derivatives in animal products has not been authenticated and the natural acetylenes are derived from higher plants and fungi.Those from plants fall into two main groups one comprising some acetylenic C18 acids occurring as glycerides in the seed-fat of certain angiosperms and the other the acetylenic compounds found in essential oils particularly of certain groups Types of natural acetylenes Source Fungi. . . Compositae and Urnbellifere Seed oils . . Chain length 8 9 10 11 12 13 10 12 13 14 17 18 Conjugated carbon -carbon unsaturation * [ CEX],* t 0 *CH=CH*[CE32] Functional groups -CH,*OH -CHO -CO ,H -CO ,Me,-CO*NH,,-CN )CH.OH >S (thiophen) -CO-O- (y-lactono) -CO,Me -CO-NHBd )CH.OH )C=O )O (furan) -C,H -CO,H (as glycerides) )CH*OH -CH,*OH -CH,.OBc of Composita. The acetylenes of microbiological origin are produced mainly by Basidiomycetes the so-called higher fungi with the exception of mycomycin which is produced by an Actinomycete.A summary of the l3 Anchel Polatnick and Kavanagh Arch. Biochem. 1950 25 208 ; Kavanagh Hervey and Robbins Proc. Nut. Acad. Sci. 1950 36 1 102. 14Celmer and Solomons J . Amer. Chem. Soc. 1952 74 1870 2245. 16Anchel Trans. New York Acad. Sci. U.S.A. 1954 16 337 Is Bohlmann Angew. Chem. 1955 67 389. l7 Jones J. 1950 754 ; Raphael “ Acetylenic Compounds in Organic Synthesis ” Hutterworths London 1955. 18 Yvernault Oleagineux 1946 1 189 ; Chem. Abs. 1947 41 3697. 374 QUARTERLY REVIEWS salient features of the natural acetylenes so far encountered is given in the Table on the previous page. General Properties Stability.-The ease with which acetylenic compounds of the types found in Nature can be handled varies within rather wide limits.A few are quite stable t o heat and may be melted or steam-distilled without decomposition and many can be recrystallised without special precautions. On the other hand most polyacetylenes are apt t o polymerise particularly on exposure to light. The effect is sometimes spectacular for example crystals of isonemotinic acid (p. 386) become bright pink after a few seconds’ exposure. This photosensitivity is shown only in condensed phases and is therefore ascribed to photocatalysed cross-linking of the carbon chains of adjacent molecules. Such a reaction would give three-dimensional networks contain- ing various conjugated chromophores the exact course being determined by the crystal structure of the monomer. This accords with the fact that where the acetylenic carbon atoms are particularly exposed as in the group H*[C=C]; photosensitivity is most marked whereas in compounds with more bulky substituents it may be quite absent.19 Susceptibility to atmo- spheric oxidation appears to be less marked than with polyenes but may be considerable when a number of double bonds is also present.Extreme pH values are in general to be avoided in work with polyacetylenes since they may bring about hydration or ionotropic rearrangements ; some allenic compounds undergo prototropic rearrangement even a t pH 7. Spectra. 20-The spectroscopic properties of the polyacetylenes are very characteristic and most useful in constitutional studies. In general the ultraviolet absorption spectra of chromophores in which triple bonds pre- dominate show band series with a characteristic fine structure (vibration) spacing of about 2000 cm.-l a frequency comparable to that of the C-C stretching vibration in the ground state.These bands are in two groups the wavelengths and relative intensities of which characterise the chromo- pliore almost completely though such “ isomeric ” chromophores as *CH=CH*[C=C],*CH=CH* and *[CH=CH],*[C-C],* may not be iminedi- ately distinguishable. The individual bands are usually very sharp and iheir intensities may be as high as lo5 or more so that polyacetylenic com- pounds can often be detected and to some extent characterised in crude extracts.* Thus in Fig. 1 the spectrum of a crude extract from a fungus shows clearly the presence of a polyacetylenic compound of which the whole sample contained less than 5pg. ; the absorption spectrum of the pure compound is given for comparison.Bohlmann Chem. Ber. 1953 86 657. 20 Cf. especially Jones Whiting Armitage Cook and Entwhistle Nature 1951 168 900 ; Bohlmann Chem. Ber. 1953 86 63 ; also Bohlmann 16 l9 and Jones et aZ.ll * However i t should be noted that for purely acetylenic chromophores the longer- wavelength series of bands is of very low intensity and easily masked even in relatively pure preparations. Where there are more than three triple bonds a shorter-wavelength band series of high intensity is easily observed under normal conditions but in other cases this feature has led t o confusion as in the case of “ biformin ” (p. 386). BU’LOCK NATURAL ACETYLENES 375 The double bonds of allene units are spectroscopically independent of each other ; when each is involved in further conjugation as in mycomycin the resultant absorption spectrum is approximately the sum of those of the two independent chromophores.14 The infrared spectra of these compounds are also of importance in struct,ural studies an early use of Raman spectras in this connection having already been noted ; the genera’lly simple linear structures of most natural acetylenes are reflected in the relative simplicity of their spectra.In absorption bands due to the triple bond itself may be rather inconspicuous ; frequently multiple they occur between 2000 and 2200 cm.-l but are of variable intensity (the same vibration causes strong bands in the Ra,man I 300 350 Wavelength (mp ) F I G . 1 Use of ultraviolet absorption spectra for the detection of polyacetylenes.lJpper curve absorption spectrum of crude ethanol extract from slope culture of a fungus. Lower curve absorption spectrum of the acetylenic component after puri$cation. (Reviewer’s unpublished work.) spectrum l*). Allene groups are associated with a sharp peak a t about 1950 cm.-l which is the most reliable test for the presence of such a group ; free ethynyl groups are readily detected by the rC-H stretching band at about 3300 cm.-l and infrared spectra are also useful in assigning con- figurations to double bonds. More generally they may be used to char- acterise the functional groups of the molecule whether or not these form part of a chromophore system deduced from ultraviolet absorption data. Moreover infrared spectra offer a most valuable method for comparing a synthetic and a natural product in a field where melting-point determina- tions are frequently impracticable.As an example of the use of spectro- scopic measurements for structure-determination in this field the case of the antibiotic nemotin (p. 386) may be cited. With the added information that nemotin gives undecanoic acid on hydrogenation the ultraviolet and 376 QUARTERLY REVIEWS infrared spectra are sufficient to establish the structure. The former indicates the presence of an ene-diyne group as the only conjugated system the latter fixes the remainder of the structure as shown in Fig. 2. Methods of Investigation The C, acids from seed oils have generally been isolated and st,udied by the usual techniques of fat chemistry though recent advances have involved the use of more refined methods better suited to this class of compound.In some of the studies of essential oils and in a recent study of volatile constituents of a fungus compounds were isolated by steam- distillation but the most generally used methods have involved solvent- extraction. Subsequent working-up has sometimes employed simple or fractional crystallisation but chromatographic methods have been widely used for example in the separation of a number of polyacetylenes from Polyporus anthracophilus.21 The use by Lythgoe and his associates of morin-alumina columns on which light-absorbing compounds appear as dark bands under ultraviolet light may be noted.22 For dealing with more unstable compounds special methods may be required. Celmer and Solomons described methods for low-temperature recrystallisation in an inert atmosphere 1b whereas the Reviewer and his co-workers prefer techniques such as solvent-extraction and counter-current distribution which minimise the handling of solid products.23 2 1 Bu’Lock Jones and Turner Chem.and Ind. 1955 686 and unpublished work. 2 2 Anet Lythgoe Silk and Trippett J. 1953 309 ; Brockmann and Volpers Chem. 23 Bu’Lock Jones and Leeming J. 1955 4270. Ber. 1947 80 77. BU’LOCK NATURAL ACETYLENES 377 Provided that the compound can be obtained pure in solution character- isation need not necessarily require its isolation as a solid. Spectroscopic and other data can be obtained by transfer to suitable solvents; electro- metric titrations and the study of addition or rearrangement reactions can if necessary be carried out similarly as can hydrogenation to a more stable product.When the compound is sufficiently stable to be manipulated in crystalline form more precise data can of course be obtained. The structures of hydrogenation products are readily established by conven- tional means ; in considering evidence so obtained it should be remembered that some allylic alcohols and esters are susceptible to hydrogenolysis under quite mild conditions of hydrogenation. Classical methods of degrada- tion are of little use with compounds in which the unsaturation extends over almost the entire molecule though oxidation was used to establish the structures of some of the seed-oil acids and hydrative breakdown of the conjugated chain with alkali was used with the lachnophyllum ester.9 The triple bond reacts rather slowly with ozone and per acid^,^^ so that double bonds may be oxidised selectively as in one determination of the structure of ximenynic (santalbic) acid.25 I n the same way since ene-ynes give maleic anhydride adducts only under forcing conditions the Diels-Alder reaction can be used to detect trans-trans-diene groups as in isomycomycin 26 a t the same time providing data on the residual chromophore of the adduct.In many cases the study of ultraviolet and infrared spectra and the identification of the’ hydrogenation product have been sufficient to justify the assignment of one or a few possible structural formulae decision between these being made by synthesis of the relevant compounds. As examples may be quoted syntheses by Bohlmann et al. of a series of related ketones in order to determine the structure of a compound from Artemisia vulgaris 27 and the assignment z8 of the correct structures to agrocybin and diatretyne I.Acetylenic Acids from Seed Oils The acetylenic acids from seed oils form a clearly defined group in that all those so far encountered have the well-known C, straight chain of stearic acid; on the other hand not all the members of this group are as well characterised as might be wished. The simplest is tariric acid (I) the structure of which was unambiguously established a t an early date.4 Glycerides of this acid make up about 95% of the seed fat of Picramnia sow and also occur in P. camboita P. carpinterg and P. lindeniana but not apparently in other genera of the group (Simarubaceae) to which Picramnia belongs. 29 Another acid with a species-specific distribution is ximenynic acid (V) 24Bohlmann and Sinn Chem.Ber. 1955 88 1869. 25Gunstone and McGee Chem. and Ind. 1954 1112. 26Celmer and Solomons J. Amer. Chem. SOC. 1952 74 3838. Bohlmann Mannhardt and Viehe Chem. Ber. 1965 88 361. 28 Bu’Lock Jones Mansfield Thompson and Whiting Chem. and Ind. 1954 990. 2g Steger and van Loon Rec. Trav. chim. 1933 52 593 ; Grimme Chem. Rev. Fett.- Harx-Ind. 1910 17 158; 1912 19 51. 378 QUARTERLY REVIEWS octadec- 1 l-en-9-ynoic acid.30 The seed oils of several South African Ximenia species contain up to 25% of the glycerides of this acid. The genus belongs to the order Santalales and ximenynic acid appears to be character- istic of XantaZum sp. The so-called santalbic acid from 8. album,31 is in CH3fCH2]5*CH=CH*C~C*[CH,]7*C02H (V) fact ximenynic acid 32 (the name santalbic acid is therefore redundant) and the same acid constitutes over 40% of the seed-fat acids of X.acuminatu,s and S. murrayana. The seed-oil of a species belonging to another genus of the Santalales Ongueka gore Engler (“ isano ” “ boleko ”) presents a more complex picture. Isano oil was first studied by Hebert,6 who in 1896 isolated from it an un- saturated photosensitive fatty acid isanic acid which was not fully char- acterised. Later investigations by Steger and van Loon 33 and Castille 34 were complicated by the isolation of artefacts the failure to obtain homo- geneous products and the use of two different names isanic and erythrogenic acid for the main constituent of the fatty acid mixture. In the Reviewer’s opinion the name isanic acid should be retained to refer to the acid for which structure (VI) 33 34 was finally established by degradation by Seher 35 and by synthesis by Black and W e e d ~ n .~ ~ Castille’s “ erythrogenic acid ” con- sisted mainly of the acid (VI) but its ultraviolet absorption spectrum shows CH2=CHfCH,]4*C~C*CrC*[CH,] ,*CO,H (VI) it to have contained about 10-15y0 of a conjugated ene-di~ne.~’ The oil also contains hydroxystearic acid derivative^,^^ and Seher 35 ascribes to one such component (not isolated) the ene-diyne structure (VII) This structure bears no obvious relation to that of isanic acid and since there is also CH *[CHZ],*CH=CH*C~C C r C CH CH (OH) *[CH,] ,*CO,H (VII) evidence for the existence of both 8-hydroxyisanic acid 38 a,nd a non- hydroxylated ene-diyne acid 39 in isano oil it is apparent that the detailed constitution of this oil remains unsettled.30 Lighthelm and Schwartz J . Amer. Chern. SOC. 1950 72 1868 ; Lighthelm and 31 Madhuraneth and Manjunath J . I n d i a n Chena. Soc. 1938 15 389. 3 2 Hatt and Szumer Chern. and I n d . 1954 962 ; Grigor MacInnes McLean and Hogg ibid. p. 1112; J. 1955 1069; also ref. 25. 33 Steger and van Loon Pette u. Seifen 1937 44 243 ; Rec. Trav. chirn. 1940 59 1156; 1941 60 107. 3 4 Castille Annalen 1939 543 104 ; Bull. Acad. TOY. Med. Belg. 1941 6 152. 3s Seher Annalen 1954 589 222. 36Bla~k and Weedon J. 1953 1785. 38 Steger and van Loon Fette u. Seifen 1937 44 243 ; Rec. Trav. chim. 1941 60 107 ; Kaufmann Balter and Herminghaus Fette u. Seifen 1951 53 537 ; Riley J. 1951 1346. Meade personal communication.von Holdt J. 1952 1088 5039. 37 Cf. Jones et al. ref. 20. BU’LOCK NATURAL ACETYLENES 379 Acetylenic acids thus occur in the seed oils of the order Santalales and of the genus Picramnia. Although the simpler chromophores of tariric and ximenynic acid would not be readily detected spectroscopically it seems likely that the application of the spectroscopic methods used in the examination of e.g. the Compositz essential oils would reveal a wider distribution of more highly unsaturated acetylenic acids. The potentialities of the use of Raman spectra in which even isolated triple bonds give rise to strong and characteristic bands should be noted ; the method has been used by Yvernault l8 to detect acetylenic constituents in butter-fat and in dry-distilled castor oil. Compositae essential oils The best-known class of natural acetylenes is that comprising compounds from the essential oils of many Composits and some Umbelliferae ; they occur in many familiar wild and cultivated plants.This is at present the largest class and is of interest) not only because of the chemistry of individual compounds but also because of the relations between them. Thus the distribution of certain compounds accords well with the taxonomy of the Compositz and may be used in some cases to supplement the botanical classification. lo 40 Moreover the compounds are such that correlations can be made between their structure so that some plausible biogenetic hypotheses can be put forward. The occurrence of several of the Compositz polyacetylenes has been summarised by Sorensen lo with whom the subject is particularly associated.In the present Review the compounds will be considered in order of their chain length (cf. Table p. 373). The compounds are in general neutral (esters ethers and hydrocasbons) and the Norwegian school have in fact isolated many of them from steani-distilled oils. The amide anacyclin is a more polar compound and belongs to a different category. C, Series.-The most widely-occurring natural polyacetylenes are un- saturated esters and alcohols derived from n-decane. These are especially characteristic of Compositae though some of them also occur in fungi (p. 385). The first to be discovered was the lachnophyllum ester shown by degradation to be the &-isomer of structure (III).g This was first isolated from Lachno- phyllum gossypinum Bg.but is widely distributed in Composits particularly in various sections of the genus Eriger~n.~ The isomeric ccp-dihydromatri- caria ester (VIII) has also been reported from several C~rnpositae.~~ How- ever bhis ester was characterised as the derived acid after alkaline hydrolysis CH,*CH,.CH,.CrC~C~C.CH=CH.CO,Me (111) CH,CH=CH.C~C*CrCCH,.CH,.C@,Me (VIII) 40 Tronvold Nestvold Holme Sorensen and Sorensen Actn Chem Scund. 1953 41 Ba,alsrud Holme Nestvold Pliva Sorensen and Sorensen ibid. 1952 6 883. 7 1375. 380 QUARTERLY REVIEWS and other esters giving (VIII) on alkali treatment may also have been present. Similarly in BeZEis perennis the trans-isomer of the ester (111) occurs together with an ene-diyne of unknown structure which after saponi- fication ultimately gives the conjugated diyne-ene-acid corresponding to (111) ; 42 again a more complex possibly allenic compound appears to have been present in the actual plant extracts.The alcohol (IX) matricarianol occurs (principally as the acetate of the trans-trans-isomer) in some Com- positae.*os 43 CH,*CH=CH*C~C *C=C.CH=CKCH,*OH (IX) The most characteristic polyacetylene of the C1 series from Compositae is the matricaria ester (X) the cis-cis-isomer having been found in nearly fifty species particularly of the genus Erigeron (Tripleurospermum) .4Os 43 This isomer was first isolated from Matricaria inodora (scentless mayweed) and its structure established by degradative methods. 44 The assignment of the configuration originally rested upon photoisomerisation to the 2-trans- 8-cis-isomer which was ~ynthesised.~~ This isomer too has been isolated from Matricaria inodora and other C o m p o s i t ~ .~ ~ The more common 2-cis- 8-cis-isomer has now been synthesised as also has the cis-lachnophyllum ester ; 46 since there is some confusion in the literature the melting points CH,*CH=CH*C~C*C=C*CH=CH*CO,Me (X) 2-cis-8-cis- m.p. 37" (natural and synthetic) 2-trans-S-cis- m.p. 2" (natural and synthetic) 2-trans-S-trans- m.p. 61" [natural (fungi) and synthetic] of the known isomers of (X) are given here. The ultraviolet absorption spectra of all the matricaria esters are somewhat anomalous in that their acetylenic fine structure is not sharp. The most highly unsaturated decane derivatives so far encountered in Compositz are the dehydromatricaria esters (XI) about which there has also been some confusion.The first to be isolated from Artemisia wuZgaris,47 has m.p. 112" and is the cis-isomer as was subsequently shown by Sorensen's group 489 49 and by Bohlmann and Mannhardt ; this too has recently been synthesised by Bell Jones and Whiting.46 The trans-isomer of (XI) which melts a t 105" when pure was synthesised 48 before its isolation from Matricaria CH3*C~C*C~C.CrC*CH=CH*CO,Me (XI) As well as the fully characterised polyacetylenes of the C, series there 42Holme and Sorensen Acta Chem. Scand. 1954 8 280. 431dem ibid. p. 34. 44 Sorensen and Stene Annalen 1941 549 SO. 45 Bruun Christensen Haug Stene and Sorensen Acta Chem. Scand. 1951 5 1244. 4G Bell Jones and Whiting Chern. and I n d . 1956 548. 47 Stavholt and Sorensen Acta Chem. Scand. 1950 4 1567. 48 Christensen and Sorensen ibid.1952 6 602. 49 Sorensen Bruun Holme and Sorensen ibid. 1954 8 26. 50 Bohlmann and Mannhardt Chem. Ber. 1955 88 429. BU'LOCK NATURAL ACETYLENES 381 also exist in Composita? several apparently related compounds lacking the characteristic ultraviolet absorption spectra of polyacetylenes. Besides the compounds mentioned above as possible precursors of (111) and (VIII) there are compounds such as the " composit-cumulen I " (broad absorption maximum a t ca. 3500 8 ) from Matricaria in odor^,*^ 51 for which a cumulene structure was proposed but now seems unlikely. Also associated with decane derivatives (in Artemisia vulgaris) are a C, ketone (see p. 382) and a dienetriyne hydrocarbon of unknown chain The hydrocarbon appears to be related to certain acetylenic hydrocarbons of unknown struc- ture which occur in Centaurea sp.together with related p o l y e n e ~ . ~ ~ Other Compounds from Comp0sitae.-A hydrocarbon capillene from the essential oil * of Artemisia capillaris Thunb. was assigned by Harada 53 the structure (XII) on the basis of hydrogenation to n-hexylbenzene the presence of acetylenic bands in the infrared spectrum and oxidative degrada- t)ions from which phenylacetic and acetic acid were isolated but no acetalde- hyde. The physical constants of capillene (d20 0.9735 nio 1.5698) agree closely with those reported for agropyrene (d20 0.9744 n;O 1.5695) one of the few acetylene derivatives from plants which are not Compositze. Agro- Ph*CH2*CH=CH*C+3CH (XII) pyrene was isolated by Treibs from a 22-year-old sample of essential oil from couch grass (Triticum yepens Agropyrum repens) and was also assigned structure (XII).53 However a synthesis of the trans-isomer of (XII) gave a product with properties (Pe5 0.9450 ng6'5 1.5510) differing from those recorded for agropyrene and ~apillene.~4 The recorded properties of the natural products whilst insufficient to characterise them as being the cis- form of (XII) rule out all phenylhexatriene structures and make the alterna- t ive formulation Ph*CH,*CzC*CH=CH*CH appear unlikely.Whatever its structure agropyrene would seem to be related to a group of C, compounds some of which contain also the phenyl group. The best authenticated of these are the compounds (XIII) (which occurs as the acetate in Carlina vulgaris),55 (XIV) (from Coreopsis sp. as the acetate) 5 6 amd (11) the " Carlina oxide " (from Carlina acaulis).These three isomeric compounds are thus closely related both chemically and botanically. The CH,-CH.CH=CHC_C *C~GC*CGC *CH=CH*CH,*OH trans - Ph * C r C .C_C CH=CH *CH ,.OH (XIII) (XIV) 51 Sorensen and Stavholt Actn Chem. Scand. 1950 4 1080. 5 2 Lofgren ibid. 1949 3 82 ; Hellstrom and Lofgren ibid. 1952 6 1024 ; Sorensen 53 Harada J. Chm. SOC. Japan 1954 75 727 (Chem. Abs. 1955 49 10235) ; 54 Cymerman-Craig David and Lake J. 1954 1874. 5 5 SBrensen and Sorensen Actu Chem. Scand. 1954 8 1763. 66 Idem ibid. p. 1741. * This oil also contains the C, ketone l-phenylhexa-2 4-diyn-l-one (Imai J. and Stavholt ibid. 1950 4 1575. Treibs Chem. Ber. 1947 80 97. Yharm. SOC. Japan 1956 76 397; Chem. Abs. 1956 50 10340). 382 QUARTERLY REVIEWS acetates of (XIII) and (XIV) were characterised spectroscopically and by hydrogenation; 5 5 7 56 that of (XIV) and of the all-trans isomer of (XIII) have been synthesised.57 Similarly " Carlina oxide " was eventually char- acterised by its Raman spectrum hydrogenation and synthesis ; the earlier workers 7 preferred an allenic structure to the acetplenic structure (11) partly because of the molecular refractivity and partly because they con- sidered the acetylenic link unlikely to exist in a natural product. The biogenetic implications of the occurrence of this group of compounds and of other arylacetylene derivatives are considered below (p. 393). There also exist in Cornpositz a series of C, hydrocarbons which may perhaps be regarded as related to the alcohols (XIII) and (XIV). These may be represented by the formuIz (XV)-(XX); all are very unstable and occur only in small amounts so that their characterisation has been (XV) CH,-CH-CH=CH~CH-CH~C~C~C~C~CH=CH*CH (XVI) CH,=CH-CH=CH.CzC.CrC.C~C*CH=CH.CH (XVII) CH2-CH*CtC.C~C.C~C*C~C*CH=CH*CH (XVIII) CH,=CH*C_~C*CEGC.C~C*CEIC.CEEC*CH (XIX) P h * C r C -C_C -CH=CH*CH (XX) Ph*C=C.CrC-C=CCH extremely difficult." 58 The structures of (XVIII) one of three highly unsaturated hydrocarbons from HeZipterum sp.58 and of (XVII) have been confirmed by synthesis,59 and an isomer of (XV) has also been synthesised.60 The hydrocarbons (XV) and (XVII) occur together in Coreopsis sp56 as do (XVI) and (XIX) ; 61 (XX) was also isolated from Coreopsis S P . ~ ~ 62 The C1 ketone from Artemisia ~uZgaris,~7 for which the formula (XXI) was eventually established by synthesis 27 a t present remains without close parallel.CH3*C~C~C~C*C~C~CH=CH*CH,*CH,~C0*CH2*CH (XXI) Acetylenic structures have been established for the toxic principles of two species of Umbelliferx? Cicuta virosa (hemlock water drop- wort) and GEnanthe crocata (cowbane water hemlock). Earlier work on these compounds which have some importance as stock poisons is sum- mnrised by Lythgoe and his c o - ~ o r l ~ e r s ~ ~ who succeeded in resolving the rather complex mixture of compounds in each source and showed that the UmbeZZiferz. 57 Bruun Skattebol and Sorensen Acta Chem. Scand. 1954 8 p. 1757 ; Bohlmenn and Inhoffen Chem. Ber. 1966 89 21. 58 Sorensen Holme Borlaud a,nd Sorensen Acta Chem. Scand. 1054 8 1769. 59 Jones Skattebol and Whiting personal communication ; cf.Jones Thompson 6o Bohlmann and Mannhardt Chem. Ber. 1955 88 1330. 6 1 Sorensen and Sorensen personal communication ; also Tids. Kjemi 1955 15 6 2 The author is especially grateful t o Professor Sorenseri for advance information 6 3 Anet Lythgoe Silk and Trippett Chem. and Iizd. 1952 757 ; J. 1953 309. and Whiting Acta Chem. Scand. 1954 8 1944. 120 ; cf. Bohlmann Chem. Ber. 1955 88 1755. regarding this group of C 1 3 hydroca,rbons. BU'LOCK NATURAL ACETYLENES 383 toxic compounds were glycols accompanied in each case by varying amounts of related compounds. The compounds from ananthe crocata were assigned structures (XXI1)-(XXIII) ; those from Cicuta virosa (XXIVa and b) are clearly related but have the unsaturation differently arranged. The structures of the (Enanthe compounds and of cicutoxin ( X X I V a ) have been confirmed by synthesis.64 HO*CH2CH=CH.C~C*C~C.CH=CK.CH~CH=CH*CH2*CH2R (XXIIU) R = CH(OH).Prn.(XXIIb) R = Bun. CH,*CH=CH~C~C*C~C-CH=CH*CH-CH-CH,~CH,~COPrn HO ~CH,~CEI,CH,*CEEC.CE~C~CH=CH~CH=CH.CH=CHR (XXTVa) It = CH(OH).Prn. (XXIVb) R = Bun. There have been isolated from a variety of plants of the Compositz and Rutaceze a series of N-isobutylamides of unsaturated acids several of which show notable insecticidal activity. Some earlier confusion about the structures of certain of these amides has now been cleared up largely by Crombie and his collaborator^,^^^ 66 and certain common structural features can now be discerned. The compounds are all derived from even-numbered fatty acids (C,, CI2 CI4 and C18) and their structures are consistent with a derivation from C (acetate) units.One member of this group in other respects typical contains two triple bonds ; this is anacyclin for which structure (XXV) has been established.GG The (XXIII) Insecticidal amides. CH,*[CH,],.C~C.C~C~[CH,],.CH=CH.CH-CH.CO.NHHui (XXV) insecticidal isobutylamides have been isolated by solvent extraction from the plant materials ; it is not yet known whet,her related compounds (such as esters of the corresponding acids) occur in t,he same or related species. Micro-organisms Distribution.-Versatility of secoiidary metabolism is characteristic of micro-organisms generally and in the recent extensive research in this field i t is scarcely surprising that a number of acetylene derivatives should have been encountered ; however their distribution seems to be somewhat restricted.The selection of material for study has been guided mainly by the search for antibiotics and hence is not truly representative ; nevertheless some groups of micro-organisms have been quite intensively studied (e.g. Streptomycetes) without any acetylenic metabolites having been en- countered. Acetylene derivatives have been reported only once outside the group of Fungi proper vix. in a species of the fungus-like Actinomyce- tales Nocardia acidophihs. In the Fungi the known producers of acetylenic compounds are all of the class Basidiornycetes (" higher fungi ") sub-class 64Hill Lythgoe Merrish and Trippett J. 1055 1770; Bohlmann and Viehe 65E.g. Crombie J. 1955 995. 6 6 Idem ibid. p. 999 Chena. Ber. 1055 88 1245. 384 QUARTERLY REVIEWS Homobasidiomycetes order Agaricales and all belong to only two families of the Agaricales vix.Agaricaceze (including Agrocybe Clitocybe Coprinus Drosophila and Marasmius spp.) and Polyporaceae (including Dzdalea Fistulina Polyporus and Poria spp.). The extent to which this distribution reflects any fundamental classification of micro-organisms can only be decided by further studies both chemical and taxonomic. In a few cases polyacetylenes of the same or related structures have been found in more or less closely related species of Fungi. Thus Agrocybe dura produces agrocybin (XXVI) 67 whilst Clitocybe diatreta in a related genus produces the two compounds (XXVII) and (XXVIII).6* The structures of the amides (XXVI) and (XXVII) were Types of Compound.-C series. (XXVI) HO *CH,*C_C *C_C *CrC*CO *NH (XXVII) HO,C*CH=CH*C~C *C_C C O *NH (XXVIII) HO2C*CH=CH*C~C*C~C*CN finally established by synthesis the arrangement of the unsaturation in the latter having been previously ascertained by measuring the pK of the carboxyl gr0up.6~ Though these three compounds are not dissimilar a fourth C compound from fungi bears no apparent relation to them.This is the acetylenic thiophenaldehyde junipal (XXIX) which together with anisaldehyde is responsible for the characteristic odour of DadaZea juniperina cultures. The nitrile (XXVIII) has also been synthesised.'O. ___ (XXIX) CH,*CzsC-II )-CHO 'S Junipal which occurs together with a (?)vinylogous compound and p-anis- aldehyde was isolated by steam-distillation and characterised spectro- scopically and by a two-step oxidation to thiophen-2 5-dicarboxylic acid.7 The aldehyde group and sulphur atom are so far unique amongst natural acetylenes; junipal is in fa,& the only simple natural thiophen known.The compounds (XXV1)-( XXVIII) were originally studied because of the antibiotic activity of the metabolic liquors in which they occur and the aldehyde (XXIX) because of its odour. Another group of fungal polyacetylenes was first encountered in a spectroscopic " screening ) ' of wood-rotting fungi. The most interesting results so far have been obtained with Polyporus anthracophilus an Australian species from cultures of which at least twelve acetylenes have been isolated and ~haracterised.~~ These may be represented by formula (XXX)-(XXXV) ; the acids occur free as the methyl est,ers or as esters with the alcohol groups of (XXXI) and (XXXII ; R = Me).Clo series. 137 Kavanagh Hervey and Robbins Proc. Nut. Acud. Sci. U.S.A. 1950 36 102. 68 Anchel (a) J . Amer. Chem. XOG. 1952 74 1588 ; 1953 75 4621 ; ( b ) Science 69 Bu'Lock Jones Mansfield Thompson and Whiting Chem. and I n d . 1954 990. 70 Jones and Whiting personal communication. 7l Birkinshaw and Chaplen Biochem. J. 1955 60 255. 7 2 Bu'Lock Jones and Turner Chem. und I n d . 1955 686 and unpublished work. 1955 121 607. BU’LOCK NATURAL ACETYLENES 385 Mixtures of these compounds were separated by chromatography on alumina and the structures all of which have been confirmed by comparison with synthetic materials were assigned largely on the evidence of spectro- scopic data and the identification of hydrogenation products.In all the (XXX) HO,CCH=CH*CEXX~H=CH*CO ,H (XXXI) CH,*CH=CHC~C*CrC*CH~CH*CH,*OH (XXXII) HO*CH2*CH,*CH,*CrC*C~C*CH==CHCOzR (XXXIII) RO,C*CH2*CH,-C~C*C~C~CH===CH*CO ,R (XXXIV) (XXXV) RO,C*CH=CH*C~C*CE~C*CH=CH*CO,R compounds the double bonds have trans-configurations ; this the occurrence of a C compound (XXX) and the presence in several of oxygen functions a t both ends of the molecule all distinguish the P. anthracophilus group from the otherwise similar group of C, compounds from Compositz. The alcohol (XXXI) (trans-irans-matricarianol) in fact occurs in both (p. 380). The compounds (XXX1)-(XXXV) are also produced by at least one other fungus Polyporus j%rno~us.~~ The first acetylenic antibiotics to be described were as it happened not simple polyacetylenes but compounds containing a still more remarkable group the optically active allene unit ; the fulfilment of van’t Hoff’s predictions by a micro-organism would have pleased Pasteur.Allenic formula? had been put forward incorrectly for a number of natural products including Carlina oxide but the first natural allene to be identified as such was the antibiotic mycomycin from Nocardia acido~hilus.~~ For this the structure (IV) was established by Celmer and S o l o r n ~ n s ~ ~ making full use of the newly established spectroscopic data; elucidation of the structure was assisted by the observation that the very labile mycomycin is transformed by alkali into a more tractable compound isomycomyciii with changes in ultraviolet and infrared spectra and complete loss of optical activity. isoMycomycin forms a maleic anhydride adduct and its structure HC=C *CEC*CH=C=CH*CH=CH CH=CH *CH C 0 2H CH *CH=CH.C=C.CrC *CH=CH *CO ,R Allenes.(IV) (XXXVI) CH3*C~C*C~C*C~C*C.H=CH*CH=CH*CH2.COzH (XXXVI) has been confirmed by synthesis.75 Mycomycin itself with a free ethynyl group is extremely unstable and a t room temperature undergoes almost complete resinification in a very short time. Mycomycin is a C, compound; somewhat similar compounds with shorter carbon chains are produced by three Polyporales Poria corticola P. tenuis and an un-named fungus B.841.13 These produce a mixture of acetylenic allenes which when resolved by counter-current distribution methods afforded the C, acid nemotinic acid (XXXVII) its lactone nernotin Johnson and Burdon J. Bact. 1947 54 281. 7 4 Celmer and Solomons J . Anaer. Chern.SOC. 1952 74 1870 2245 3838; 1953 76Bohlmann Chem. Ber. 1954 87 712. 75 1372 3430. 386 QUARTERLY REVIEWS ( XXXVIII),76 and smaller amounts of the homologues odyssic acid (XXXIX) and odyssin (XL).77 The C1 compounds are as labile as mycomycin but (XXXVII) HC~CC~C.CH=C=C€I-CH(OH)CH,*CH,*CO,H I (XXXVIII) HC~C*C=(:.CH-C=CH*CH.C!H,.CH,~CO*b (XXXIX) CH3*C~C*C~C.CH=C=CH~CH(OH)CH,CH,*CO2I'I (XL) CH3.C~C.C~C*CH=C-CH*CH*CH,*CH,*C0 *O the C1 compounds are a little more stable and some derivatives obtained in crystalline form. The alkali-catalysed isomerisations of were these allenes differ from that of mycomycin nemotinic acid being converted rather slowly into the triyne (XLI) (isonemotinic acid) whereas its lactone isomerises very readily to the enetriyne (XLII) (nemotin A).78 Odyssic acid and odyssin behave similarly and the structure (XIJII) assigned to odyssin A has been confirmed by synthe~is.'~ The allenes (XXXVI1)-(XL) combine optically active allene units with a second asymmetric centre which retains its activity in for example isonemotinic acid (XLI).(XLI) HCrC*C~C*C~C.CH,~CH(OH).CH~~CH,*CO,H (XLII) HC~C.C~C.C~C.CH=CH.CH,.CH,*CO,H (XLIII) CH3.C~C*CrC-C~C.CH=CH*CH,*CH,C0,H Two compounds which shows changes of ultraviolet absorption spectrum on alkali treatment reminiscent of the nemotin-nemotin A reaction have been detected amongst the metabolites of Drosophila substrata,80 and these too may be allene derivatives. (The partial characterisation 81 of fucoxan- thin a carotenoid from marine algze as an allene though not strictly relevant may perhaps be noted a t this point.) Othev compounds.Several fungi produce polyacetylenes of as yet unknown structure. Besides metabolites of two species of Coprinus and one of Marasmius,82 some compounds from Polyporus biformis may be especially noted. One of these may be a C, glycol ; a second originally described as a C glycol with a diene-diyne chromophore is probably nona- 4 6 8-triyne-1 2-diol which would show only weak absorption in most of the ultraviolet region together with about 5% of a diene-di~ne.8~ In several fungi including P. biformis and Agrocybe dura the amount of solvent-extractable material in the culture medium may be increased by 76 Bu'Lock Jones and Leeming J. 1955 4270. 77 Idem unpublished work. 78 Bu'Lock Jones Leeming and Thompson J. 1956 3767. 79 Jones Skattebol and Whiting personal communication.Anchel Arch. Biochem. Biophys. 1953 43 127. Sorensen personal communication ; Tids. Kjemi 1955 15 129 ; Torto and Weedon Chem. and Ind. 1955 1219. 8 2 Doery Gardner Burton and Abraham Antibiotics and Chemotherapy 1951 1 409 ; Anchel personal communication ; Benz Scandinavian Chem. Meeting VIII June 14-17th 1953. s3 Robbins Kavanagh and Hervey PTOC. N a t . Acad. Sci. U.S.A. 1947 33 176 ; Anchel and Cohen J. BioE. Chern. 1954 208 319 ; Anchel personal communication. BU’LOCK NATURAL ACETYLENES 387 boiling ; 6 7 9 8 3 since the known polyacetylenes in each case contain hydroxyl groups it would seem likely that boiling liberates some of the material from easily hydrolysed water-soluble conjugates of unknown nature. Biological Activity It will have been apparent from preceding sections that many of the natural sources of acetylene derivatives were first studied because of their physiological activity.However the activity of the source has not always been traced to the acetylenic constituents and in certain cases where active acetylenes have been isolated other closely related acetylenes have also been found and proved to be inactive. Thus there is no reason to expect any general kind of physiological activity in acetylenic compounds. For example the poisonous Umbelliferae (Enanthe crocata and Cicuta virosa have very similar pharmacological action causing violent convulsions ; both plants are important stock-poisons. The toxic effects are due to two different C, glycols (XXIIa) and (XXIVa) ; these toxins are accompanied in the plants by the corresponding monoalcohols (XXIIb) and (XXIVb) but these closely related compounds are virtually non-t0xic.6~ A similar situation exists in regard to the antibiotic activity of some fungi.Thus that of Clitocybe diatreta is due not to the amide (XXVII) but solely to the corresponding nitrile (XXVIII).68b The patterns of anti- biotic activity of the allenes nemotin (XXXVIII) and mycomycin (IV) are completely changed when these compounds are isomerised to nemotin A (XLII) and isomycomycin (XXXVI).13y 74 The activities of some poly- acetylene antibiotics have been summarised by Anchel ; l5 it is there noted that several are unusually active against mycobacteria and some also show marked antifungal activity. None of the antibiotics so far studied has any clinical usefulness ; some are far too unstable and several are highly toxic to animals.However it is not impossible that some more useful member of the series will be dis- covered either as a natural metabolite or as a synthetic material. For most of the polyacetylenes from Compositae no physiological data have been recorded (cf. ref. 9). The N-isobutylamide anacyclin (XXV) belongs to a class of compounds some of which show strong insecticidal and/or sialogogue activity and others of which are inactive. The factors determining these physiological activities are not yet clear but would appear to be a t least partly stereochemical. Anacyclin itself is only slightly toxic to houseflies but half-hydrogenation of the diyne group gives a product with high activity.66 Bios y nt het ic Aspects The problem of the biosynthesis of acetylene compounds is twofold being concerned ( a ) with the biological formation of triple bonds and ( b ) with the biosynthesis of particular compounds and their relationship to more conventional metabolites.There has as yet been very little in the way of a direct attack upon either problem but some pointers to probable solutions can be found in the evidence a t present available. B B 388 QUARTERLY REVIEWS Origin of Triple Bonds.-Several lines of evidence suggest that the origin of triple bonds should be considered as part of a wider problem that of the biological origin of carbon-carbon multiple bonds in general ; this problem is unsolved even for the simplest important example oleic acid. Several of the natural acetylenes belong to well-defined classes of natural products the majority of which possess no triple but only double bonds.Thus the acetylenic acids from seed oils are typical in all respects save in the presence of triple bonds of the far bigger category of unsaturated C, acids from similar sources Equally the amide anacyclin is typical of the group of natural N-isobutylamides. I n the same way fucoxanthin appears to differ from more conventional carotenoids only in the presence of an allene group and it may well be that some of the fungus polyacetylenes are similarly related to members of a growing class of microbiological products with unbranched polyene chains ; unfortunately structures of only a few of these are knowas* It is important not to overestimate the thermodynamic instability of acetylenes. Thus to consider the simplest case the formation of acetylene from ethylene by oxidation with atmospheric oxygen would be only some 10 kcal.mole-l less exothermic than the corresponding oxidation of ethane C$,(g) + 302(g) + C&14(g) + H20(1) + 32.5 k d * C2H4(g) + +02(g) + CJ&(g) + H20(1) + 23.0 kcal. Differences of the same order-modified somewhat by entropy factors and changes in first- and second-order conjugation-would be expected between the heats of various pairs of reactions in which double and triple bonds were formed by oxidation. Thus for enzymic oxidations it might be con- cluded that the aerobic dehydrogenation of ethylenes to acetylenes would be thermodynamically feasible and sufficiently exothermic to allow the intermediate participation of various redox systems if required e.g.-CH=CH- X -c=c- XXH,X; Such direct dehydrogenations are known to operate in at least two important biosyntheses of double bonds the succinoxidase reaction H02C*CH2*CH2*C02H + trans-HO,C*CH=CH-CO,H and the various acyl- coenzyme A ap-dehydrogenations e.g. with butyryl- coenzyme A CH3*CH2*CH2*CO*SA + trans-CH,-CH=CH*CO*SA (Coenzyme A = A-SH) Both enzymes play a fundamental part in synthesis and breakdown of fats ; 85 they are metalloflavoprotein systems with similar electron-acceptor 84 Cf. OEoshnik Vining Mebane and Taber Science 1955 121 147. For meful summaries see Lynen Nature 1954 174 962 and Ann. Rev. Biochem. 1955 24 653. BU’LOCK NATURAL ACETYLENES 389 The intimate mechanism of Similar considerations apply equally to what may be called the “de- Enzymic reactions of this requirements,86 and both give trans-ethylenes.their action is obscure. hydrative ” routes to unsaturated compounds. kind leading to double bonds are known e.g. “ crotonase ” 87 ~~cwM-CH~.CH=CH*CO*SA + CH,.CH( OH)*CH,*CO*SA + CH,=CHCH,*CO*SA I n such dehydrations the heat of reaction is generally rather low ; once more the corresponding reaction dehydration of an enol by way of a suitable enol derivative to a triple bond appears to be thermodynamically feasible ; the reverse reaction is of course well known in vitro and has been shown to occur biologically (cf. p. 394). -COCHZ- $ -C=CH- + -C=C- I ox Consideration of the origin of double-bond unsaturation thus constitutes a relevant analogy for the problem of triple-bond biosynthesis. Unfortun- ately the known enzymic reactions in which double bonds are formed are all rather special and throw little light on the general case that of multiple bonds more or less insulated from other functional groups.Of this general case oleic or tariric acid may be taken as typical. Three routes to unsatur- ated fatty acids appear to warrant consideration. (i) “ Residual ” unsaturation. On this hypothesis unsaturated and saturated fatty acids arise by diverging routes from the same precursors the unsaturation in the former being “ left behind ” as it were in the process of biosynthesis. The routes branch a t a P-hydroxyacylcoenzyme A stage where alternative modes of dehydration are possible (cf. crotonase above). Formation of a saturated fatty acid would involve dehydration to an ap-unsaturated acyl derivative reduction and further repeated condensa- tion with coenzyme A.The alternative dehydration product a /Iy-unsatur- ated compound by reacting directly with further acetylcoenzyme A would give rise to an unsaturated fatty acid R*CH,*CH(OH) *CH,*CO*SA + R*CH=CH.CHz*CO.SA R.CH=CH.[CH,~,,+t,.CO,H Ac*SA etc. Ac-SA ctc. I R*CH,.CH-CH*CO*SA -+ R*[CET&*CO.SA + R.[CH,],n+t,COzH This hypothesis outlined briefly by Lyneqs5 has the advantage of explaining acceptably the patterns of unsaturation most often found in the natural fatty acids (e.g. R*CH==CH*[CII,] 2n+l *CO,H) and in this respect is superior g6 Mahler J . Biol. Chern. 1954 206 13 ; Kearncy and Singer Biochim. Biophys. 87 Wakil and Mahler J . Bid. Chem. 1954 807 125 ; Stern Raw and Del Campillo Acta 1955 17 596. Fed. Proc. 1954 13 304.390 QUARTERLY REVIEWS to the schemes of Hilditch Variations leading to the less common acids with e.g. conjugated double bonds may be formulated readily and the simpler acetylenic acids might be formed similarly by way of allenic intermediates. The relative amount of unsaturated acids formed would be controlled by the availability of reduced coenzyme rather than by the position of equilibrium of the " crotonase " reaction. Experiments in which abnormal (deuter- ated or odd-numbered 91) fatty acids supplied to various organisms were re-isolated as for example A9,10-derivatives suggest that dehydrogenation (or hydroxylation followed by dehydration) can occur at specific points on the hydrocarbon chain or Robinson.89 (ii) Direct dehydrogenation. If such reactions in fact occur there is no a priori reason why similar reactions leading to acetylenic acids should not take place.A third explanation requires the known c$-dehydrogenation step to be followed by controlled double-bond migration (iii) Indirect deh,ydrogeizntion. CH,.[CH2],+,.CH=CH.CO,H + CH3*[CH2]m*CH=CH.[CH2]n'CO2H The fact that till these schemes suitably amended and also the hypothesis of enol dehydration (p. 389) could explain the biosynthesis of the acetylenic seed-oil acids is of wider significance since as will appear from the following section other natural acetylenes probably arise in an essentially similar manner i.e. as offshoots of more normal routes of fat synthesis. Origin of the Carbon Skeleton.-In the first place it may be observed that all the known natural acetylenes are derived from unbranched chains of carbon atoms.Formally this generalisation applies even to those com- pounds containing heterocyclic or aromatic rings and in fact suggests that these may arise from open-chain precursors (cf. p. 393). Such a situation contrasts sharply with the best known group of natural polyenes the branched-chain carotenoids but is parallel to that in two groups already mentioned the natural fatty acids and the unbranched polyenes from micro-organisms. It is only natural to attempt to explain the biosynthesis of all such compounds in similar terms particularly since the biosynthesis of fatty acids is comparatively well ~nderstood.~~ The role of the two-carbon unit of acetyl coenzyme A in the biosynthesis of saturated fatty acids is of course firmly established and schemes extending this to the common unsaturated acids have already been outlined with the observation that given an acceptable hypothesis for the origin of triple bonds the schemes can be extended to cover the examples of the C, 88 Hilditch OZCngineuz 1955 10 83 ; " The Chemical Constitution of Natural Fats " Robinson " The Structural Relationships of Natural Products " Oxford Univ.Chapman and Hall London 1956 p. 463 et seq. Press 1955 p. 6. 90 Schoenheimer and Rittenberg J. BioZ. Chem. 1936 113 505. 91 Appel Bohm Keil and Schiller 2. physioZ. Chern. 1947 282 220. BU’LOCK NATURAL ACETYLENES 39 1 acetylenic fatty acids. However the natural acetylenes do not in general display that marked preference for chains of even numbers of carbon atoms which is characteristic of the fatty acids proper.Chains of 8 9 10 11 12 13 14 17 and 18 carbon atoms have all been found. In particular cases there may be special explanations for the occurrence of an odd-numbered chain; a carbon atom may have been gained by C-methylation or lost by decarboxylation. The latter explanation seems rather likely on u p i o r i grounds for compounds lacking a terminal oxygen function. However even in the series of acids from natural fats the occurrence of odd-numbered carbon chains is not unknown 9 2 and it is unlikely that their biosynthesis proceeds by routes essentially different from that established for the even- numbered fatty acids. For example during the building-up of the chain further condensation of an even-numbered acid might take place with a four-carbon unit (such as succinic or oxaloacetic acid) instead of with acetic acid and be followed by decarboxylation of the /3-keto-acid giving an acid with an odd-numbered carbon chain R*CO,H + CHz-{~OH~AZ&H -+ RCOCH- I I CO,H COzH CH,.CO ,H RCO*CH,-{ C O.CO,H In addition to the role of acetylcoenzyme A in primary metabolic pro- cesses biosynthesis from “ acetate ” units has been demonstrated for a variety of complex metabolites and has been inferred for others from their pattern of o~ygenation.~~ The latter approach is not readily applicable to polyacetylenic compounds ; moreover acetate is not the only “ building- unit ” to merit consideration.The transfer of hydroxyacetaldehyde units is important in plant biosynthesis and succinylcoenzyme A is important as a precursor e.g. of the pyrrole pigments.Faced with such considerations it seems advisable to limit further speculation to cases where some experi- mental data are available. At present such evidence has only been obtained for certain fungal acetylenes. C, and C, Allenes.-A biosynthetic scheme for nemotinic acid must account for the following facts the precursors must be intermediates in the “ normal ” glucose metabolism of the fungus ; 94 labelled carbon supplied as aceta,t,e is efficiently converted into polyncetylenes ; the formation of polyacetylenes is stimulated by the addition of acetic succinic or rnalonic acid ; g5 in addition to the C, compound nemotinic acid the fungus also produces a C1 compound odyssic acid.96 The scheme must also account for the formation of the allene groups which in nemotinic and odyssic acid are unstable.All these requirements can be met by a scheme such as that on page 392 which assumes the biological equivalence of -C0*CH2- and 9 2 Cf. e.g. Shorland Gerson and Hansen Biochem. J. 1955 59 350; 61 702. 93 Birch and Donovan Austral. J. Chem. 1953 6 360 and subsequent papers. 9 4 Bu’Lock and Leadbeater Biochem. J. 1956 62 476. 95 Idem unpublished results. 96 Bu’Lock Jones and Leeming J. 1956 3767. 392 QUARTERLY REVIEWS -C=C- groups and which uses acetyl- and succinyl-coenzyme A as precursors. 4CH,*CO*SA + HO,C*CH,*CH,*CO*SA -1 [HOzC*CH,*CO*CH2*CO*CH2*C0 CR,*CO *CH,*CH,*CO,H] -1 H02C *C=C *C_C *C=C CH,*CO -CH,*CH,*CO ,H HO,C *C~C-CFC*CH=C=CH.CO *CH,*CH,.CO,H \- coa CH,*CrC*C=C*CH==C=CH.CH( OH) CH,*CH,*CO,H H *C=C *C=C *CH=C=CH*CH (OH) *CH,.CH,*CO ,H Such a synthesis from acetate and succinate would be promoted by the inhibitory action of malonate on normal succinate oxidation ; allene forma- tion at the keto-acid stage would be facilitated by the gain in conjugation (with the carbonyl group).The alternative reactions a t the final stage- reduction of the carbonyl group Sccompanied either by decarboxylation or by reduction of carboxyl to methyl-lead to nemotinic and odyssic acid respectively. C and C, Compounds.-Evidence of a rather different kind lends some support to a not dissimilar scheme for the biogenesis of acetylenic compounds in Polyporus anthracophibus ; 72 the compounds produced by this fungus are apparently interconvertible by way of smaller molecules.~7 All the compounds are readily derivable from the four compounds shown below and these (the equivalence of -CO*CH2- and -CG& units being again assumed) can be derived in the manner indicated from the common metabolic intermediates acetate succinate and fumarate CH .cH=J =cH-c~; =C e~ kc .CH= ;=CH .C O,H 5 acetate HO,CGH=CH*C~~rC*Cr~zC-CH=!=CH*CO,H fumarate + 3 acetate H02C*CH,~CH,C~~rC~C~~zC*CH=~=CH*C0,H succinate + 3 acetate HOzC-CH=CH*C~~zC*CH=~=CH*CO,H fumarate + 2 acetate Further t F ansf or mat ions Whilst it is clear that speculation about the ultimate origin of other natural acetylenes must await the results of further experimental work a comparison of the structures of various polyacetylenes from the same or related sources serves to draw attention to some other points of possible biogenetic interest.Thus the co-occurrence in Clitocybe diatreta of the amide (XXVII) and the nitrile (XXVIII) suggests a possible biogenesis for the somewhat 8' Bu'Lock Leadbeater and Turner unpublished results.BU'LOCK NATURAL ACETYLENES 393 uncommon cyanide group (the production98 of free hydrocyanic acid by the polyacetylene-producing fungus B.841 may be noted in this connection). In the same way the occurrence in related species of the tridecanol derivative (XIII) the phenylheptanol derivative (XIV) and the (phenylpropy1)furan derivative (11) all isomeric with each other suggests the possible operation of two interesting transformations. One the apparent conversion of an acetylenic alcohol into a furan derivative 5 5 CH-CH II II R.C~C-C%JCH=CH*CH,.OH + R.CH,*CEC*C CH \O/ is paralleled by a simpler reaction in ~ i t r o .~ ~ The second correlates a hexa- dienynyl residue and a phenyl group as seen also in the pair of Coreopsis hydrocarbons (XVI) and (XIX) and might suggest a biological variation of the Diels-Alder reaction A more cautious interpretation would be to regard it as evidence for a route to aromatic compounds closely related to one leading to straight-chain unsaturated compounds. Such a route would clearly be different from the well-established shikimic acid route. On the other hand if the previous hypotheses concerning the biosynthesis of acetylenes are generally applic- able such a route to phenyl derivatives would start from the same kind of precursors as are involved in the route from acetate to hydroxybenzene derivatives as postulated by I n considering the co-occurrence of the thiophen junipal and p-anisalde- hyde in DBdalea juniperina Birkinshaw and Chaplen 71 suggested that both might arise from a common non-cyclic precursor.The route to the thiophen," which can be formally written as R-CZC-CEC-R' + H,S -+ ~--(lltR' S had already been suggested by Chdlenger and Holmes in connection with the occurrence of ad-terthienyl with its unbranched chain of 12 carbon atoms in certain Composits. loo I n addition to such hypothetical transformations which could involve common precursors rather than acetylenic compounds themselves there is some direct evidence that the natural acetylenic compounds are not always gt3 Robbins Rolnick and Kavanagh Mycologia 1950 42 161. g9 Heilbron Jones Smith and Weedon J. 1946 54. loo Challenger and Holmes J. 1953 1837 ; Zechmeister and Sease J .Amer. Ghem. * Cf. the isolation from Compositae of a thiophen corresponding to the addition of SOC. 1947 69 273. hydrogen mlphide to the hydrocarbon (XX) (Sorensen unpublished work). 394 QUARTERLY REVIEWS mere end-products of metabolism. Thus the polyacetylenes from Basidio- mycete B.841 are actively broken down by the fungus mycelium,94 whilst in Polyporus anthracophilus there is a complicated interaction between the breakdown and re-synthesis of the various acetylenic metabolite^.^' In higher plants seasonal variations in the relative amounts of different acetylenic constituents as observed in ananthe crocata and Cicuta virosa 63 may indicate a similar state of affairs. Recently Eimhjellen lol has described work with certain strains of Enterobacteriaceae which will utilise (non-natural) acetylenic compounds such as acetylenedicarboxylic acid as sole carbon source.In these organisms acetylene breakdown follows a pathway similar to that put forward here as a route for their synthesis. Thus acetylenedicarboxylic acid is converted into oxaloacetic acid apparently by hydration and by way of the enol-form and not via fumaric acid by hydrogenation C0,H C0,H C0,H I I I C C-OH CO Ill + II + I -+ etc. C C H CH2 1 I I CO,H CO,H C0,H The conversion into oxaloacetic acid can be brought about by cell-free extracts but the reversibility of the reaction has not yet been demonstrated. Other strains will utilise butyne-1 4-diol and propynal as sole carbon sources. The oxidation of butyne-1 4-dio1 hexa-2 4-diyne-1 6-diol and propynol by soil bacteria has also been described,102 propynol being stated to afford hexa-2 4-diyne-1 6-dioic acid.However under the experimental conditions described (1 month's incubation in a.ir a t 22') this product might perhaps be expected to result from autoxidation and non-enzymic oxidative coupling. The author is especially grateful to all those who have helped by friendly criticism and by making available results at present unpublished particularly Professor E. R. H. Jones (Oxford) and Professor N. A. Sorensen (Trondheim). lol Eimhjellen personal communication ; Biochem. J. 1956 64 4 ~ . 102Hanaoka Harada and Takizawa J . Agric. Chent. SOC. Japan 1952 26 151 (Chern. Abs. 1954 48 10,114).

ISSN:0009-2681    

DOI:10.1039/QR9561000371

出版商:RSC    

年代:1956

数据来源: RSC

摘要:

THE CHEMISTRY OF THE AROMATIC HETEROCYCLIC N-OXIDES By A. R. KATRITZKY M.A. D.Phil. B.Sc. (THE DYSON PERRINS LABORATORY OXFORD UNIVERSITY) IN general little interest was taken in the chemistry of heterocyclic N-oxides" until about 15 years ago. The discovery that the antibiotics iodinin and aspergillic acid were respectively a phenazine dioxide (I) and the cyclic hydroxamic acid tautomer (11) of a pyrazine oxide (111) attracted some a t t e n t i ~ n ~ but more important to the development of the subject was the determination of the dipole moment of pyridine 1-0xide.~ This showed 0- 0-. Q I Q+ I 0- 0- M+ 0 that in addition to (IV) the canonical forms (V) and (VI) must make important contributions to the resonance hybrid. Ochiai predicted and con- firmed that this would facilitate electrophilic substitution in the 4-position ; following this lead Japanese workers have completed a very large amount of work 5 on heterocyclic N-oxides since 1943.In Holland den Hertog inde- pendently discovered the ready nitration of pyridine l-oxide and did much to develop the field.6 Colonna's work in Italy should also be mentioned.' It has become apparent that structures (VII) and (VIII) also contribute to the pyridine l-oxide resonance hybrid ; the fact that the N-oxide function is strongly polnrisable in both directions is of considerable theoretical Clemo and McIlwain J. 1938 479; Clemo and Daglish J. 1950 1481. Dutcher and Wintersteiner J . Biol. Chern. 1944 155 359 ; cf. Dunn Gallagher See inter aZia papers by Spring Shaw and Landquist and their co-workers. Linton J . Amer. Chem.Soc. 1940 62 1945. Summarised to 1953 by Ochiai J. Org. Chem. 1953 18 534. Latest paper den Hertog van Ammers and Schukking Rec. Trav. ckirn. 1955 Colonna Ga,zzetta 1956 86 705. Newbold and Spring J. 1949 S 127. '74 1171. * Throughout this Review the representation N+-O- has been used for N-oxides The former has the advantage that comparison with other in preference to N+O. canonical forms is clearer (cf. IV-VIII). 395 396 QUARTERLY REVIEWS interest,s and in this respect the Nf-0- group of an N-oxide is similar to a nitroso-group attached to a benzene ring.9 The ability of the N-oxide group both to accept or to donate electrons is clearly shown by a comparison of dipole moments of alkyl phenyl 4-pyridyl and (4-pyridyl 1-oxide) compounds. l o Attention is directed to compilations on amine oxides,11 derivatives of pyridine 1 2 quinoline and isoquinoline,l3 acridine l4 furazans and isatogens,l5 and quinoxalines and cinnolines.16 Preparation of Aromatic Heterocyclic N-oxides This is the most generally applicable method of preparation.The most convenient oxidising agent is peracetic acid (i.e. a mixture of glacial acetic acid and 30% hydrogen peroxide) ; monoper- phthalic acid and perbenzoic acid have also been used. Hydrogen peroxide alone is usually without effect. Direct ~ x i d a t i o n . ~ l1 An aromatic nitrogen-heterocyclic compound is converted less readily than a tertiary amine into the N-oxide or than a sulphide into the sulphoxide but more readily than a monosubstituted ethylenic bond into an epoxide. Quinine (IX) gives successively the quinuclidine mono-N-oxide and the di-N-oxide whereas 2-methylthiopyridine gives the corresponding sulph- oxide (X).17 The reaction is subject to steric hindrance and gives anomalous results with some quinoxaline derivatives l1 electron-attracting groups interfere.1Ga Cyclisation of a dicarbonyl compound with hydroxylamine.Glutaconic dialdehyde (XI) is cyclised by ammonia t o pyridine and by hydroxylamine Pyrimidines give mono-oxides only.l* See e.g. Jaff4 J . Amer. Chem. SOC. 1954 76 3527. Robinson Chem. and Ind. 1925 44 456. lo Katritzky Randall and Sutton J. in the press. l1 Culvenor Rev. Pure Appl. Chem. (Australia) 1953 3 83. l2 Cislak Ind. Eng. Chem. 1955 47 800. l3 Elderfield " Heterocyclic Compounds " Wiley New York 1952 Vol. IV l4 Albert " The Acridines " Arnold London 1951 p.144. l5 Smith Chem. Rev. 1938 23 193. l6 Simpson " Condensed Pyrazine and Pyridazine Ring System " Interscience New York 1953 pp. 9 54 232 311 348. lBa Landquist J . 1953 2816 ; Landquist and Stacey ibid. p. 2822. l7 Shaw Bernstein Losee and Lott J . Amer. Chem. SOC. 1950 72 4362. l8 Ochiai and Yamanaka Pharm. Bull. (Japan) 1955 3 175. pp. 121 237. KATRJTZKY HETEROCYCLIC N-OXIDES 397 to pyridine 1 -oxide l9 and homophthalaldehyde (XII) similarly gives iso- quinoline 2-oxide. 2o Glutaconic acid with hydroxylamine gives the cyclic hydroxamic tautomer (XIII) of 2 6-dihydroxypyridine 1 -oxide and other examples of this type of reaction are known.21 (XIV) Cyclisation of a substituted hydroxyZamine.ll In the preparation of quinolines by reductive c y clisa tion of /3 - o - ni t r o p hen ylpr opionic acid deriva - tives the corresponding quinoline 1 -oxide is often obtained as a by-product.The quinoline is formed by the cyclisation of an intermediate amine the N-oxide by cyclisation of the corresponding hydroxylamine. I n this way quinaldine l-oxide (XV) has been prepared from the ketone (XIV) and derivatives of 1 -hydroxyquinol-2-one (XVI) and 2-aminoquinoline l-oxide from o-nitrophenylacrylic esters and nitriles l19 22 respectively ; a-o-nitro- benzoyl-ketones are reduced to 4-hydroxyquinoline l-oxides 22a and the reaction has also been extended to afford 2 9-diazaphenanthrene 9-0xides.~3 The similar preparation of 5-membered ring compounds is exemplified by the mild reduction of o-nitro-azo-compounds to benzotriazole l-oxides 23a and o-nitroanilides to benziminazole l-oxides.23b The reaction between or-amino- nitriles and oxo-aldoximes to give 2-aminopyrazine l-oxides (XVII) 24 and the use of benzyloxyurea in the Traube pyrimidine synthesis to afford derivatives of pyrimidine 1 -oxide 25 are related reactions. OH 0- 0- (XVI I) (XVI I I) (x I x) l9 Baumgarten Merllinder and Olshausen Ber. 1933 66 1802. 2o Schopf Hartmann and Koch Ber. 1936 69 2766. 21 Ames and Grey J. 1955 631 3518 ; Nielsen Elming and Clauson-Kaas Actn 2a Taylor and Kalenda J . Org. Chenz. 1953 18 1755. 225 Gabriel and Gerhard Ber. 1921 54 1067 1613 ; Cornforth and James Biochem. 23a Elbs J. prakt. Chem. 1924 108 209 ; Ross and Warwick J. 1956 1724. 23b Niementowski Ber. 1910 43 3012 and refs. therein. a 4 Sharp and Spring J. 1951 932. z5 Lott and Shaw J . Amer.Chem. SOC. 1949 '71 70. Chem. Scand. 1955 9 9 30. ,7. 1956 63 124. 23 Hansen and Petrow J . 1953 350. 398 QUARTERLY REVIEWS Internal cyclisation of a nitro-compound. In this group of reactions the N=O bond of a nitro-compound undergoes a reaction of the carbonyl- addition type. a- Amino-o-nitrophenylacetonitriles (as XVIII) are cyclised to indazole l-oxides (e.g. XIX) by treatment with alkali,26 and o-nitro- phenylurea derivatives (XX ; X = 0 S NH NPh) give benzotriazine 1 -oxides (XXI) with sodium hydroxide. 27 The preparation of isatogens (indolenin-3-one 1 -oxides) (XXIII) by heating 1 -0-nitrophenylvinyl chlorides (XXII) with pyridine l5 is probably of the same type. - (XXIV) 0- The Wohl-Aue reaction,ll in which phenazine $oxides (XXIV) are prepared by heating aromatic amines with nitro-compounds has probably a related mechanism initiated by nucleophilic attack of Ph*NH- on the nitrobenzene ; this is supported by the isolation of p-nitrodiphenylamines as by-products.28 N-Hydroxyacridones are formed together with acridones by the reaction of o-nitrobenzaldehydes with benzene.29 R-f-FR R.$+N*OH - F ? ~ - ~ - O - 0 N N\ -Rq‘ N RC=O YHO OH OH ‘0’ ‘0- R (XXV) (XXV 1) (xxv I I) Miscellaneous cyclisations.o- Aminophenyl ketoximes are cyclised by nitrous acid to 4 5-benzo-1 2 3-triazine l - o x i d e ~ . ~ ~ ~ 1-0xa-2 5-diazole 2-oxides (XXVI) are obtained by the mild oxidation of dioximes (XXV) and the spontaneous dimeriaation of nitrile oxides (RNCO),15 which can also undergo an alternative dimerisation to l-oxa-2 4-diazole %oxides or trimerise to s-triazine 1 3 5 - t r i o ~ i d e s .~ ~ ~ Certain y-bromo-nitro-com- pounds cyclise to isooxazoline 2-oxides. 29C Aldehydes rea,ct with cc-diketone monoximes to give oxazole 3-oxides (XXVII).30 Reactions of Aromatic Heterocyclic N-Oxides Because of the possibility of electron movement both into the hetero- cyclic ring from the Nf-0- group and in the opposite direction as shown for e.g. pyridine l-oxide by the contribution of canonical forms (IV-VIII) 26 Behr J . Amer. Ghem. SOC. 1954 76 3672. 27 Wolf Wilson Pfister and Tishler ibid. p. 4611 ; Rrndt Ber. 1913 46 3522 ; 28 Wohl Ber. 1903 36 4135. 2 9 Albert, “ The Acridines ” Arnold London 1951 pp. 99-101. 29a Meisenheimer Senn and Zimmermann Ber. 1927 60 1736 ; cf. Ockenden and Schofield J. 1953 1915. 29b Wieland Ber. 1909 42 803.29c Smith and Scribner J. Anzer. Chem. SOC. 1956 78 3412 and refs. therein. 30 Selwitz and Kosak ibid. 1955 77 5370. Arndt and Rosenau Ber. 1917 5Q 1245. KATRITZKY HETEROCYCLIC A’-OXIDES 399 to the resonance hybrid N-oxides show great variety in their chemical reactions. (XXV I I I) (XX I x> (XXN (xxx I) (XXXII) CXXXIII) An electrophilic reagent (E) may attack the oxygen (as in XXVIII) or the y-carbon atom (as in XXIX). (A reaction of this type a t the %-carbon atom is hardly ever found being presumably prevented by the powerful adverse inductive effect of the neighbouring positively charged nitrogen atom.) Acids alkyl halides and certain Lewis acids receive electrons a t the oxygen atom giving respectively salts quaternary salts and co-ordination compounds. The species responsible for nitration inercuration and some other electrophilic substitutions receive electrons from 1 he y-carbon atom in the transition state and finally displace a proton from this point giving y-nitro-N-oxides etc.A nucleophilic reagent (Nu) may attack the oxygen (as in XXX) the cc-carbon (as in XXXI) or the y-carbon atom. Reducing agents and some (kher electron donors such as phosphorus trjchloride supply an electron pair to the oxygen atom and cause deoxygenation. Some powerful nucleo- philes e.g. Grignard reagents attack the a-carbon atom giving an inter- mediate (XXXII) which spontaneously loses a proton and an oxide ion thus affording the corresponding a-substituted deoxygenated heterocyclic compound (XXXIII). Weaker nucleophilic agents e.g. chloride cyanide 2nd acetoxy-ions are able to attack the cc- or the y-carbon atom if the N-oxide first forms a co-ordinated intermediate with an electron acceptor (as in XXVIII) this happens when the N-oxide reacts with sulphonyl chloride henzoyl chloride-potassium cyanide acetic anhydride etc.In certain of these reactions it is possible that a-substitution results from reaction of the co-ordinated intermediate by way of a cyclic transition state. I n reactions with nucleophilic reagents heterocyclic N-oxides sometimes behave similarly to the nitrones R*N+(*O-):CR, as emphasised by Colonna. 31 Finally there is a group of reactions in which a substituent is attached to a carbon atom /3 or 6 to the N+-0- group i.e. either on the @-carbon atjom of the ring or on the first carbon atom of an a- or y-side-chain.The iirechanism here is not clear ; in certain cases a free-radical attack has been suggested. These various reactions are now discussed in more detail ; it will be noted that most of the illustrations are taken from the pyridine and the qriinoline field which is because much more work has been done there than with other N-oxides. N-Oxides Electrophilic Attack on Oxygen (cf. XXX) .-Salt formation. 3l Colonna Boll. Sci. Fuc. Chirn. ind. Bologna 1940 4 134 ; Chern. Abs. 1940 34 7290. 400 QUARTERLY REVIEWS form stable salts with strong acids unless other negative groups are present. The basicity is however considerably less than that of the corresponding deoxygenated compound the pK values 32 of the conjugate acids of e.g. pyridine pyridine 1-oxide and 4-nitropyridine I -oxide are respectively 5-29 0.79 and - 1.7.The basicities of a series of substituted pyridine l-oxides were found 32 to conform to the Hammett equation with p = 2.09. Picrolonates are convenient for the characterisation of N-0xides.~2a N-Oxides with some difficulty give quaternary salts. These are often converted by alkali into the corresponding deoxygenated base and aldeh~de,~ 33 just as are the quaternary salts from aliphatic N-oxides. The mechanism of this reaction (XXXIV) may be compared Quaternary salts. R,N+TO-CH - H r- OH HCrO ?_ 0 - CH2-H P OH (xxx I V) (XXXV) OMe (XXXV I> OMe (XXXV I I) with the oxidation of alcohols with chromic acid (XXXV). In favourable cases the quaternary salt with alkali gives a pseudo-base e.g. (XXXVI) -+ (XXXVII). 34 Quinoxaline 1 -oxide adds methyl iodide solely on the 4-nitrogen atom.lBa Co-ordination compounds. Very little is known about their formation from heterocyclic N-oxides. Pyridine 1 -oxide and sulphur trioxide give C5H,N*0*S03.35 Pyridine 1 -oxide is nitrated by mixed acid at 100" to give 4-nitropyridine l-oxide in very good yield.5 If the reaction is carried out at 150" some 2-nitropyridine is also obtained ; presumably this is formed by deoxygenation of 2-nitro- pyridine l-oxide (cf. below).5 This nitration a to the N+-0- group is excep- tional ; if the y-position is occupied nitration is usually either not effected (e.g. in 4-alkylpyridine l-oxides),36 takes place in another ring (e.g. in isoquinoline 2-oxide in the 5-position 37) or when a very strongly electron- releasing group is in the Q- or the y-position nitration takes place in the 16-position (as e.g.in 4-hydroxypyridine l-oxide 5 38). Electrophilic Attack on the Ring (XXXI) .-Nitration. 32 Jaff6 and Doak J . Amer. Chem. Xoc. 1955 77 4441. 32a Katritzky J. 1956 2404. 33 Ochiai Katada and Naito J . Pharm. SOC. Japan 1944 64 210 ; Chem. A h . 34 Lehrnstedt and Dostal Ber. 1939 72 1071. 35 Baumgarten and Erbe Ber. 1938 71 2603. 36 Ishikawa J. Pharm. SOC. Japan 1945 65 6 ; Chem. Abs. 1951 45 8529. 37 Ochiai and Ishikawa J . €'harm. SOC. Japan 1945 65 4A 17 ; Chem. Abs. 38 Naito J . Pharm. SOC. Japan 1947 67 246 ; Chem. Abs. 1951 45 9541. 1951 45 5154. 1951 45 8527. KATRITZKY HETEROCYCLIC N-OXIDES 401 The position substituted in quinoline l-oxide (XXXVIII ; R = H) varies ; at 10" 5- and 8- and at 70" mainly the 4-nitro-derivative are formed ; at still higher temperatures initial deoxygenation again directs substitution into the benzenoid ring.5 The nitrations of some other quinoline l-oxides (XXXVIII ; R = Me C1 Br) show a similar temperature dependence but in the ether (XXXVIII ; R = OMe) substitution is entirely at the 5-position and in the derivative (XXXVIII ; R = NO,) entirely at the 4-positi0n.~~ The directive power of the N+-0 - group is very great compared with that of other substituents in the same ring; 2- and 3-ethoxypyridine 1-oxide are both nitrated exclusively in the 4-position.4O However 2-hydroxy- pyridine 1-oxide [or its hydroxamic acid tautomer cf. (11)] is nitrated in (XLI tlhe 5-p0sition,~~ and in 3 5-dimethoxy- 3-bromo-5-methoxy- and 3 5- diethoxy-pyridine 1 -oxide substitution even occurs in the 2-position probably partly for steric reasons.61 *1 In phenazine &oxide (XXIV) nitration occurs more easily than in the deoxygenated parent and the nitro-group enters the 3-position ; similar hehaviour is found with chlorophenazine 5-oxides but in methoxy-compounds the methoxyl group directs the orientation of nitrati~n.~la Little is known about more complicated systems ; 4-phenylcinnoline l-oxide on nitration gives four different products none of which has been orientated,l6 and 7 8-benzoquinoline 1-oxide is nitrated in the positions shown in (XXXIX).42 3 4-Benzocinnoline 1-oxide was originally con- sidered 42a to be nitrated mainly in the 2- and partly in the 3-positionY but clipole-moment evidence 42b showed that the major product could not be the 2-isomer and chemical evidence showed that it was not the 1- 3- 8- or 10- isomer.42c It is of interest that both 2 :'5-dimethyl-3 6-diphenylpyrazine aeNaito J .Pharm. SOC. Japan 1948 68 209; Chern. Abs. 1953 47 8075; * O den Hertog Kolder and Comb6 Rec. Traw. chim. 1951 70 591. 41 den Hertog Henkens and Dilz ibid. 1953 72 296. 41a Otomasu Pharm. Bull. Japan 1954 2 283 ; 1956 4 117. 421wai J . Pham. SOC. Japan 1951 '71 1291. 42a King and King J. 1945 824. 42b Calderbank and Le FBvre J. 1951 649. 42C Arcos Arcos and Miller J. Org. Chem. 1956 21 652. Okamoto J . Pharm. SOC. Japan 1951 '71 727. 402 QUARTERLY REVIEWS (XL) and the corresponding 1 4-dioxide give m-nitrophenyl compounds on nitrati~n.~ Other electrophilic substitutions. Pyridine 1 -oxide is mercurated in the 4-position under mild conditions ; quinoline l-oxide however forms the 8-mer~uri-derivative~4~ probably through a cyclic transition state.45 Bromina- tion of pyridinel -oxide could not be effected,46 but 4-bromoquinoline l-oxide was prepared by direct hal~genation.~~ Treatment of pyridine 1 -oxide with 20% fuming sulphuric acid and mercuric sulphate at 220-240" gave 3-sulphopyridine l-oxide in 51% yield but no substitution could be N e - q induced under milder conditions ; 46 presumably under the forc- ing conditions the co-ordination complex with SO is formed (see above) in which the activating character of the Nf-0- group is lost (cf.A). Although the Friedel-Crafts reaction fails with pyridine 1 -oxide 46 the derivative (XLI) cyclises into the 4-position giving the oxo-ester (XLII). The deoxygenated analogue cyclises less readily and into the 2-position.48 Nucleophilic Attack on Oxygen (cf. XXXII).-The ease of reduction of heterocyclic N-oxides varies. Some phenazine and quinoxaline N-oxides liberate iodine from acidified potassium iodide l1 23 but pyridine and quinoline l-oxides are generally very resistant to reduction as shown by the values of their reduction potentials (C,H,NO - 1.2786 ; cf. Me,NO - 004562).~ Other groups in the molecule can often be reduced selectively without simultaneous loss of the N-0 group (see below). Iron-acetic acid,40 41 49 and sodium dithionite 24 remove the N-oxygen atom efficiently. Heterocyclic #-oxides may also be deoxygenated by an oxygen acceptor. This method is most useful when it is desired to leave unaffected other groups susceptible to reduction.Phosphorus trichloride converts 4-nitro- pyridine 1 -oxide into 4-nitropyridine. The method fails with 4-nitroquino- line l-oxide but phosphorus tribromide may then be used.5 An early example of this type of reaction in the benzopyrazole series is recorded.50 Triphenyl phosphite has also been used as an oxygen accept0r.~1 4 3 Beech J. 1955 3094. 4 4 Ukai Yamamoto and Hirano J . Pharm. Soc. Japan 1953 73 823; Ghem. 4 5 Yamamoto Hirano and Yotsuzuka Pharrn. Bull. (Japan) 1955 3 105. 4 6 Mosher and Welsh J . Amer. CherrL. SOC. 1955 77 2902. 4 7 Ochiai and Okamoto J . Pharrn. SOC. Japan 1947 67 87 ; Ghem. Abs. 1951 4 g den Hertog and Comb& Rec. Trav. chirn. 1951 70 581. 5 0 Reissert and Lemmer Ber. 1926 59 351. 51 Hamana J. Pharm. SOC. Japan 1955 75 139. Abs.1954 48 9946 (cf. also J . Pharm. SOC. Japan 1955 75 490). 45 9538. 48 Murray and Hauser J . Org. Chem. 1954 19 2008. KATRITZKY HETEROCYCLIC N-OXIDES 403 Pyridine 1-oxides are deoxygenated in poor yield by heating them alone or in concentrated sulphuric acid.11 Hydrogen peroxide will also cause deoxygenation if re-oxidation is prevented by steric hindrance - e.g. 1 2 - benzophenazine 7 12 - dioxide gives the 7 - monoxide Nucleophilic Attack on the Ring (cf. X X X I I I and XXXVI).-Grignard reagenis. As mentioned above Grignard reagents convert the group -N+(O-):CH- into -N:CR- ; e.g. quinoline 1-oxide and phenyl- magnesium bromide give 2-phenylquinoline 53 Good results are also obtained with benzoquinolines 54 but pyridine 1 -oxides give poorer yields.13 Here the yields may be improved by using the complex between pyridine 1-oxide and benzoyl chloride.Action of sulphuryl chloride or phosphorus oxychloride. l1 As already indicated nucleophilic attack on the ring is enhanced by previous co-ordina- tion of the oxygen atom. Pyridine 1-oxide and sulphuryl chloride give a mixture of 2- and 4-chloropyridine which can be separated because the weakly basic 2-chloropyridine does not form a p i ~ r a t e . ~ ~ The reaction has heen used most where the formation of two products is prevented. A 2- or ;L 4-substituted quinoline 1 -oxide is converted into the corresponding 2 4-disubstituted quinoline ; 56 isoquinoline 2-oxides give exclusively I -chloroisoquinolines 5' (only one true ortho-position) ; and the N-mono- oxide ( X L I V ) gives only the a-chloro-compound probably because the y-position is sterically hindered.58 Phosphorus oxychloride reacts with phenazine N-oxides where there is no free a- or y-position chlorine being introduced into the side rings.S9 Although x- and y-nuclear attack on complexes such as (XLV) pre- ponderated it has been shown that reactions of this type give some /%nuclear 60 and side-chain 61 chlorination. This reagent converts N-oxides (without an alkyl group tc or y to the N+-0- see below) into a-0x0-compounds i.e. - -CH:N( 0)- becomes -CO*NH-. Ochiai has emphasised the formal similarity to the last reaction by writing ( X L V I ) and ( X L V I I ) as inter- mediates but it is possible that the reaction involves a cyclic transition state ( X L V I I I ; of course the electrons can also be considered to move in the opposite direction).This reaction is known in the benziminazole and quinoline series,5 62 etc. ( X L I I I ) . 52 Action of acetic anhydride. 5 2 Pachter and Kloetzel J . Amer. Chem. SOC. 1951 73 4958. 53 Colonna and Risaliti Gazzetta 1953 83 58. 6 4 Colonna and Fatutta ibid. p. 622. 6 5 Bobranski Kochanska and Kowalewska Ber. 1938 '71 2385. 5 6 Reitsema Chem. Rev. 1948 43 58. 5 7 Robinson J . Amer. Chem. SOC. 1947 69 1939. s B Kermack and Tebrich J. 1945 375. 6 9 Postovskii and Abramova Zhur. obshchei Khim. 1954 24 485 ; Chem. Abs. 6 o Gouley Moersh and Mosher J . Amer. Chem. SOC. 1947 69 303. 61 Kato J . Pharm. SOC. Japan 1955 75 1236 1239. 6 2 Montanari and Risaliti Gaxxetta 1953 83 278. 1!)55 49 6273. c c 404 QUARTERLY REVIEWS Quinoline l-oxide and 2-bromopyridine give the pyridone (XLIX) ; 63 it is suggested that this reaction involves an analogous cyclic transition state (L).With these reagents qujnoline 1 -oxide gives 2-cyanoquinoline probably by way of the intermediate (LI). This discloses an obvious analogy to the Reissert reaction and other examples Benzoyl chloride and patassium cyunide. are given in a comprehensive review of the latter reaction.64 Benzoyl chloride and potassium hydroxide with quinoline 1 -oxides give carbo- styrils 547 a2 and diethyl sodiomalonate with preformed benzoyloxyquino- linium chloride gives diethyl 2-quinolylmalonate. l3 Recently examples of these reactions in the pyrimidine series have been described. l8 Reactions leading to Substitution /I or 6 to the N+-O-.-Pachter 65 showed that the reaction between quinaldine 1 -oxide benzoyl chloride and sodium hydroxide gave 2- benzoyloxymethylquinoline (LII) .Acetic anhydride acetoxylates the alkyl group of u- or y-alkylpyridine l-oxides e.g. (LIII ; Takeda and Hamamoto J . Pharm. Soc. Japan 1953 '73 1158 ; Chem. A h . 1954 48 12748. 64McEwen and Cobb Chem. Rev. 1955 55 543. 6 5 Pachter J . Arner. Chem. Xoc. 1953 75 3026. KATRITZKY HETEROCYCLIC N-OXIDES 405 X = H Me or OAc) + (LIV).66-e9 With 2-picoline l-oxide (LIII ; X = H) some 6-methylpyrid-2-one is also formed.66 The reaction can be used to make hydroxymethyl- and formyl-pyridines by hydrolysis of the acetates (LIV; X = H Me or O A C ) . ~ ~ 4-Picoline 1 -oxide gives in addition to 4-a~etoxymethylpyridine~ some 3-hydroxy-4-methylpyridine (LV). It was suggested that these bot>h came from a common anhydro-base intermediate (LVI).69 A free-radical mechan- ism has also been suggested for this reaction.70 A mixture of products appears to be generally formed; thus lepidine l-oxide gives 2- and 3-hydroxylepidine as well as 4-quinolylmethan01.7~ Another reaction giving P-substitution is the production of b-hydroxy- compounds from N-oxides with toluene-p-sulphonyl chloride ; e.g. iso- quinoline 2-oxide is converted into 4-toluene-p-sulphonyloxyisoquinoline ; Ochiai has suggested 7 2 that the compounds (LVII) and (LVIII) are inter- mediates in this reaction. N-Oxides with substituted groups AZkyZ groups. In the a- or y-position to the N+-0- groups alkyl groups show enhanced reactivity ; thus 2- and 4-methylpyridine 1 -oxide undergo Claisen condensations with ethyl oxalate 73 and give cyanine dyes.74 When a or y to the Nf-0- these groups readily undergo nucleophilic replacement by amines sodium alkoxides phenoxides and thio-derivatives.6 49 l 7 75 Attempts at replace- ment with carbanions have been less suc~essful.~~ Halogen atoms in phenazine N-oxides are more reactive than those in ~henazines.~~ 4-Nitro- pyridine 1 -oxides are converted into 4-chloro(or -bromo)pyridine 1 -oxides by treatment with acetyl chloride,5 or with boiling hydrochloric (hydrobromic) acid.5 49 Heating them with acetic anhydride and dimethylaniline (acceptor for nitrous acid) gives 4-hyclroxypyridine 1 -oxides.The nitro-group in 2 6-dimethyl-3-nitropyridine l-oxide does not show this r e a ~ t i v i t y . ~ ~ Catalytic reduction of 4-nitropyridine 1 -oxides in neutral media affords the corresponding 4-aminopyridine l-oxide ; in acid the "-0- group is also lost.Reduction under other conditions leads to azo- azoxy- and hydrazo- compounds Nityo-groups and halogen atoms. 6 6 Kobayashi and Furukawa Pharrn. Bdl. (Japun) 1963 1 347 ; 6* Bullitt and Maynard ibid. p. 1370. 69 Berson and Cohen ibid. 1955 77 1281. 70 Boeckelheide and Harrington Chem. und Ind. 1955 1423. 71 Kobayashi Furukawa Akimoto and Hoshi J . Pharm. SOC. Japan 1954 74 791 ; Chem. Abs. 1955 49 11659; cf. also Kato J . Pharm. SOC. Japan 1955 75 1233. 7 2 Ochiai Abs. Papers XIVth Internat. Congr. Pure Appl. Chem. Zurich 1955 p. 375; cf. Ochiai and Ikehara Pharm. Bull. (Japan) 1955 3 454. 73Adams and Miyano J . Amer. Chem. SOC. 1954 76 3168. 7 4 Takahashi and Satake J . Pharm.SOC. Japan 1952 72 1188; Chem. Abs. 7 5 Katritzky unpublished work. 76 Nakayama J . Pharm. SOC. Japan 1951 71 1391. 77 Pachter and Kloetzel J . Amer. Chenz. Xoc. 1952 74 971. Chem. Abs. 87 Boeckelheide and Linn J . Amer. Ghem. SOC. 1954 76 1286. 1965 49 10948. 1953 47 7500. 406 QUBRTERLY REVIEWS 2-Nitropyridine l-oxides have been made by oxidation of 2-amino- pyridine l-oxides with Caro's acid.78 Amino- and hydroxy-groups. Comparison of ultraviolet spectra shows that 2-hydroxypyridine l-oxide exists ip the tautomeric cyclic hydroxamic acid form i.e. l-hydroxy-2-pyridone 7 9 3 80 (cf. I1 and 111) but 2-amino- 75 and probably 4-amino- and 4-hydroxy-pyridine l-oxide exist mainly in the N-oxide form.813 82 2- and 4-Hydroxypyridine l-oxide may be prepared by dealkylation of 2- and 4-alkoxypyridine l-oxide ; 5 9 79 80 2- and 4-amino-derivatives are made respectively by oxidation of a 2-acylaminopyridine followed by removal of the acyl gr0up,~3 75 and by reduction of the readily available 4-nitropyridine l-oxides.Reaction of 4-hydroxypyridine 1 -oxide with diazomethane gives a mixture of Lmethoxypyrid-4-one and 4-methoxypyridine 1 -oxide.*l 4- Aminopyridine 1 -oxide with methyl iodide gives 4-amino- 1 -methoxy- (LI XI ( LX) pyridinium iodide. 33 4-Aminopyridine 1 -oxides can be diazotised and the diazonium compounds subjected to the various Sandmeyer reactions etc. ; the diazo-group can also be replaced by hydrogen.38 2-Aminopyridine l-oxide can also be diazotised in dilute a ~ i d . ~ 5 2-Ethoxycarbonylaminopyridine 1 -oxides (as LIX) lose ethanol when heated forming bicyclic products (as LX).84 2-Hydroxy- and 2-amino-N-oxides give intense red and blue colours respectively with ferric chloride ; 239 24 79 under forcing alkaline conditions the 2-amino- may be converted into a 2-hydro~y-group.~~ The importance of N-oxide intermediates in synthetic work is likely to increase ; thus in recent syntheses of ricinine 85 and alstyrine,86 in each case the pyridine 4-substituent was introduced by nitration of a l-oxide and further transformation of the nitro-group.This Review was written during the tenure of an Imperial Chemical Indus- tries fellowship. '*Brown Abs. 128th meeting Amer. Chem. Soc. 1955 p. 10-0. 79 Shaw J . Amer. Chem. SOC. 1949 71 67. *O Cunningham Newbold Spring and Stark J. 1949 2091. 81 Ochiai and Hayashi J . Pharm. SOC. Japan 1947 67 151 ; Chem. Abs. 1951 82 Jaffd J . Amer. Chem. SOC. 1955 77 4445; but see also Hayashi J . Pharm. 8 3 Adams and Miyano J . Amer. Chem. SOC. 1954 76 2785. 84 Katritzky J. 1956 85 Taylor and Crovetti J . Amer. Chem. SOC. 1956 78 214. 96 Lee and Swan J. 1956 771. 45 9540. SOC. Japan 1951 71 213.

ISSN:0009-2681    

DOI:10.1039/QR9561000395

出版商:RSC    

年代:1956

数据来源: RSC

摘要:

THE STEREOCHEMISTRY OF SUB-GROUP VIB OF THE PERIODIC TABLE By S. C. ABRAHAMS 9h.D. (LABORATORY FOR INSULATION RESEARCH MASSACHUSETTS INSTITUTE OF TECHNOLOGY CAMBRIDGE MASSACHUSETTS U.S.A. AND CHEMISTRY DEPART- MENT THE UNIVERSITY GLASGOW) THE properties of the elements in sub-group VIB of the Periodic Table oxygen sulphur selenium tellurium and polonium undergo striking transi- tions as the group is ascended. The chemical behaviour of the group passes from that of the typical non-metals oxygen and sulphur to the typical metal polonium. A systematic change is also found in the structure of the elements from diatomic molecules through ring and chain molecules to a simple lattice composed of polonium atoms. A corresponding transition in electrical properties accompanies the structural evolution for oxygen and sulphur are insulators selenium and tellurium are semiconductors and polonium shows metallic conduction.The interrelation between structure and conductivity has already been discussed. Within this sub-group chemical bonds to one two three four and six other atoms are known and this complexity in .bond formation has attracted sustained interest over many years. In the last decade a great increase in both theoretical and particularly experimental knowledge regarding the nature of the various kinds of chemical bond formed by the atoms of sub- group VIB has become available. The present Review is an attempt to tiring together the most important information concerning the geometrical and electronic configuration of these atoms in their different bonded states. The Stereochemistry of Oxygen (a) Two-bonded Oxygen.-The mechanism of bond formation by bivalent oxygenis well exemplified by the case of the water molecule.TheH-O-H angle would be 90" if pure p orbitals alone were used by the oxygen atom. Experimentally this angle is found to be 104.5" (Table l) which could be the result of an admixture of some s with the oxygen p orbitals. A full account of the electronic nature of the water molecule has been given by C10u1son.2 I n general the amount of s-character in the bonds formed by Bivalent oxygen will be determined by the nature of the other atoms present. The s-admixture could vary from small percentages giving bond angles close to go" to that required for sp2 hybridisation. An examination of Table 1 reveals that the values for the oxygen valency angle measured in a variety of molecules fall within the limits of error into two fairly distinct groups.With the exception of ozone the angle is 1 von Hippel J . Chem. Phys. 1948 16 372. 2 Coulson " Valence " Oxford Univ. Press Oxford 1952. 407 408 QUARTERLY REVIEWS TABLE 1 . Oxygen valency angle Molecule Ethylene oxide . . . . . . Trimethylene oxide. . . . . Mercury diethylene oxide . . . Fluorine monoxide . . . . . Hydrogen deutero-oxide . . . Water . . . . . . . . Dimethyl peroxide . . . . . Furan . . . . . . . Divinyl ether . . . . . . Diethylether . . . . . . Mercuric oxide . . . . . . Chlorine monoxide . . . . . Dimethyl ether . . . . . . 1 4-Dioxan . . . . . . . Benzofurazan . . . . . . Potassium ethyl sulphate . . . Ozone . . . . . . . . p-Dimethoxybenzene .. . . Di-p-bromophenyl ether . . . Di-p-iodophenylether . . . . Diphenyl ether . . . . . . Angle 61.6" & 0.1" 96" & 6" 101-5" f 1-5" 103.8" 1.5" 104.0" & 0.5" 104.5" & 0.1" 105.1" 6 0.1" 105" & 3" 106.2" 108*1" 107" & 3" 112" f 2" 108" $ 3" 109*8" 110*S" * 1" 110" 108" f 5" 112" f 5" 112" i 12" 116.8" :I; 0.5" 121" & 2" 123" 1" 123" j 2" 123" -& 1" 124" f 5" 94.5" & 3" 114" j- 4" Method * M .TV . E.D. X-Ray E.D. Spec. spec. Y Y E.D. M.W. E.D. Y t 9 9 X-Ray and N.D E.D. D.M. E.D. XyRa y M.W. X-Ray 9 ? D.M. Re.€. 3 4 5 6 7 8 9 10 4 11 12 13 4 4 14 15 16 17 4 18 19 20 21 fSt 24 25 * Abbreviations used in this and in all subsequent tables are D.M. dipole moment ; E.D. electron diffraction ; M.W. microwave ; N.D. neutron diffraction ; Spec. spectroscopy ; X-Ray X-ray diffraction.t A number of references to earlier determinations are given in this paper. Cunningham Boyd Myers Gwinn and Lo Van J . Chem. Phys. 1951 19 676. Quoted by Allen and Sutton Acta Cryst. 1950 3 46. Bernstein and Powling J. Chem. Phys. 1950 18 685. Ibers m d Schomaker J. Phys. Chem. 1953 57 G99. Darling and Dcnnison Phys. Rev. 1940 57 128. ti Grdenid ibid. 1952 5 367. 8 Strandberg J . Ghem. Phys. 1949 17 901. [Inc. N.Y. lo Herzberg "Infrared and Rarnan Spectra " 1945 p. 489 D. van Nostrand Co. l1 Bak Hansen and Rastrup-Andersen Discuss. Faraday Xoc. 1955 19 30. l2 Almenningen Bastiansen and Hansen Acta Chsm. Xcand. 1955 9 1306. l3 Barricelli and Bastiansen ibid. 1949 3 201. I4Roth Acta Cryst. 1956 9 277. 15Dunitz and Hedberg J . Ainer. Clzem. Soc. 1950 72 3108.l6 Gibbs J. Chein. Phys. 1954 22 1460. l7 Hassel and Viervoll Acta Chem. Scand. 1947 1 149. 2o Trambarulo Ghosh Burrus and Gordy J. Chern. Phys. 1953 21 851 ; Hughes 21 Goodwin Przybylslia and Robertson Acta Cryst. 1950 3 279. 2 2 Toussaint Mem. Soc. roy. Sci. Likge 1952 12 1. 23 Plieth Z. Naturforsch. 1947 2a 409. 24 Toussaint Bull. SOC. Sci. Likge 1946 15 86. 25Coop and Sutton J. 1938 1869. Luzzati Acta Cryst. 1951 4 193. lsJarvis ibid. 1953 6 327. ibid. 1956 24 131. ABRAHAMS STEREOCICEMISTRY OF SUB-GROUP VIB 409 r- CI - about 109.5" or less unless one or both of the attached groups are aromatic in which case the angle is very close to 120". It hence appears that in these non-aromatic molecules use is made of the oxygen 2s orbitals to approach sp3 (tetrahedral) hybridisatlion.Where an aromatic group is involved the hybridisation appears to be sp2 (trigonal). Both in ozone and in the aromatic ethers various ionic canonical forms are likely to contribute to the final electronic structure e.g. (I) and (11). In ethylene oxide which appears to be an example of strain in the classical sense p 2 bonding has been assumed for the oxygen atom with " bent " bonds lying along the arc tangents to the carbon and oxygen orbitals. (b) Higher VuEencies of Oxygen.-The literature gives but few reports of critical measurements of molecular dimensions for compounds containing oxygen exhibiting a valency higher than two. Recently the crystal struc- ture of trischloromercurioxonium chloride has been solved 2 6 p 27 and shown to consist of the ions (111). The two studies are essentially in complete a,greement and show J + c I- (1 I I> that the planar trischloromercurioxonium ion possesses trigonal symmetry the 0-Hg-C1 group being nearly linear (0-Hg-C1= 175").The unsubstituted hydroxonium ion OH3+ has been studied 28 by infrared spectroscopy and reported to possess C3 symmetry with angles close to tetrahedral. This ion is hence very similar to NH with which it is iso- electronic. These determinations of the symmetry of two OR,+ ions are thus in disagreement possibly because of the difference in size between hydrogen and mercury. There are no unambiguous examples of oxygen forming four simultaneous covalent bonds. The cases of basic beryllium acetate and the zinc blende- type oxides have been discussed by Wells.29 (c) The Oxygen-Oxygen Dihedral Angle.-A fundamental stereochemical property of oxygen that has received considerable attention is the dihedral angle formed by the planes containing the oxygen-oxygen bond and severally each of the other oxygen valency bonds.A discussion of this 26&avni6ar and Grdenib Acta Cryst. 1955 8 275. 27 Weiss Nagorsen and Weiss 2. nnorg. Chem. 1953 274 151. 28Ferriso and Hornig J . Chem. Phys. 1955 23 1464. 29 Wells " Structural Inorganic Chemistry " Oxford U.P. 1945. 410 QUARTERLY REVIEWS angle (4 in Fig. 1) was given for the case of H,O by Penney and Sutherland 30 who assumed that only oxygen p orbitals were used in bonding. F I G . 1 Stmcture of the hydrogen peroxide molecule. They showed that the predominant steric factor governing the dihedral angle is not the interaction between hydrogen atoms nor that between OH bonds but is the repulsive interaction between the unpaired electrons on each oxygen atom.This repulsion is greatest when the orbitals containing the lone pairs are parallel. They calculated the molecule to have a minimum potential energy for 4 about 100". Lassettre and Dean 31 took additional interactions into account and did not make the assumption that p orbitals alone would be used. They found that the combined effect of all the inter- action energies resulted in an equilibrium dihedral angle that lay in the range 94-1 13". A crystallographic study 32 of the hydrogen peroxide-urea addition complex led to the assignment of 106" & 2" for the dihedral angle and a later X-ray study 33 of solid hydrogen peroxide a t -20" gave q5 = 94" & 1.5".Gigukre 34 quotes the value 80" & 20" for hydrogen peroxide based upon a spectroscopic study. Disulphur decafluorodioxide has now been shown 35 by an electron- diffraction study to possess a peroxide structure with a S-O-S bond angle of 105" & 3" and a dihedral angle 4 of 107" & 5". A microwave study 36 of hydrogen peroxide gave the height of the potential barrier as 0.3 kcal./mole a sinusoidal approximation being used for the hindering potential while a subsequent paper 37 showed that the then available microwave data were equally applicable to a model in which there is a high cis- and a low trans-barrier to internal rotation such as 8 and 0.6 kcal./mole respectively. Luft 38 places t'he energy barrier Vo 30Penney and Sutherland Trans. Faradny Soc. 1934 30 808.31Lassettre and Dean J. Chem. Phys. 1949 17 317. 3 2 Lu Hughes and Giguere J. Amer. Clzem. Xoc. 1941 63 1507. 33Abrahams Collin and Lipseomb Acta Cryst. 1951 4 15. 34 GiguGre Bull. Xoc. chim. France 1954 21 720. 35Harvey and Bauer J . Amer. Chem. SOC. 1954 76 859. 36Massey and Bianco J . Chem. Phys. 1954 22 442. 37 Massey and Hart ibid. 1955 23 942. 38 Luft ibid. 1953 21 179. ABRAHAMS STEREOCHEMISTRY OF SUB-GROUP VZB 41 1 a t 6 < Vo < 18 lrcal./mole and has recently reviewed 39 the spectroscopic and thermodynamic data for hydrogen peroxide. (d) The Oxygen-Oxygen Bond.-The oxygen-oxygen bond has now been extensively studied in a variety of different circumstances. The nature of t,his bond in the oxygen molecule has been discussed by Coulson,2 in terms of moleculas orbital theory.The length of the oxygen-oxygen bond in the ground state for 0 is given in Table 2 and very closely corresponds to a pure double bond. In ozone it has been suggested 2O that the bond of length 1.278 A possesses 50% double-bond character the canonical struc- tures (IV) and (V) being considered most important with small contributions (IV) ( V) (VI) (VI I> from (VI) and (VII). Although the preparation of several alkali ozonides has been rep~rted,~O no structural information was given. Giguhe and Harvey 41 have claimed that the ozonide ion does not exist. The oxygen-oxygen bond length in the HO radical determined 42 with Ihe aid of Badger’s rule has been taken as indicating a bond similar to that observed in the superoxide ion. Magnetic-susceptibility measurements 43 have shown that the superoxide ion 0,- contains a three-electron bond.‘Chis bond of length 1.28 8 may be regarded as derived from the oxygen molecule by filling one of the two vacant antibonding pn orbitals. Thus as in the case of ozone the oxygen-oxygen bond length of 1-28 8 corresponds t o 50% double-bond character. The identity of bond types (d) and (e) has been demonstrated 44 on the basis of the magnetic susceptibilities - 18 x 10-6 for 0 2 - in Na,O and - - 17.7 x 10-6 for H,02. This view is strengthened by the identity in Ibond lengths and also by a consideration of the electronic structures for each type may be considered as derived from the oxygen molecule by filling both vacant antibonding pn orbitals thus producing a single bond. The lmgth of a single oxygen-oxygen bond may hence be regarded as 1.49 8.The bond in 02+ of length 1.12 8 in the 2L!g state can be considered to possess approximately 150y0 double-bond character with only one electron in an antibonding orbital. A smooth bond order-bond length curve may now be constructed as in Fig. 2 passing through each point discussed above mithin the overall error in that point. (e) The Carbon-Oxygen Bond-Bond-length and related information is now available for bonds between oxygen and many other atoms. However apart from the oxygen-oxygen bond the only bonds for which enough data Bond type (e) in Table 2 may be taken as a standard single bond. 39 Luft Monatsh. 1955 86 528. 40 Whaley and Kleinberg J . Amer. Chem. SOC. 1951 73 79. 41 Giguhre and Harvey ibid. 1954 76 5891. 42 GiguZtre J . Chem. Phys. 1954 22 2085.43Klemm and Sodomann 2. unorg. Chem. 1935 225 273. 44 Neiding and Kazarnovskii Doklady Akad. Nuuk S.S.S.R. 1950 74 735. 412 QUARTERLY REVIEWS TABLE 2. Oxygen-oxygen bond types and lengths Bond 0-0 -0-0 [O-01- [O-012- -0-o- c0-01+ Example 0 2 0 3 HOZ IX-KO P-NaO BaO Ca0,,8H20 Sr02,8H20 Ba0,,8HZO Li,O HZO 2 SF,*O,-SF O2+ Bond length (A) 1.2074 & 0.0001 1-278 & 0.003 1.30 1-33 f 0.06 1.31 0.03 1.28 & 0-07 1.28 0.02 1-47 1.49 f 0.04 1-48 1.49 1.48 1.3 f 0.1 1-47 f 0.02 1-49 f 0.02 1.49 & 0.01 1.47 & 0.03 1,1227 f 0.0001 Ref. 45 20 42 46 47 48 49 50 51 52 52 52 53 54 33 34 35 55 0 I I I I I I 750 100 50 0 Doubledond character (%I 1.70 I FIG. 2 Variation of the oxygen-oxygen bond length with bond character. 4 5 Babcock and Herzberg Astrophys. J. 1948 108 167.4 6 Templeton and Dauben J . Amer. Chem. SOC. 1950 '72 2251. 47 Zhdanov and Zvonkova Doklady Akad. Nauk S.S.S.R. 1952 82 743. 48Kasatochkin and Kotov Zhur. Tekhn. Pix. 1937 '7 1468. 49 Abrahams and Kalnajs Acta Cryst. 1955 8 503. 5O Butuzov Doklady Akud. Nauk S.S.S.R. 1947 58 1411. 51 Abrahams and Kalnajs Acta Cryst. 1954 '7 838. 5 2 Harr Thesis Syracuse University N.Y. 1952. 53 Fehhr von Wilucki and Dost Chem. Ber. 1953 86 1429. 54 Gigusre and Schomaker J . Amer. Chem. SOC. 1943 65 2025. 5 5 Herzberg " Spectra of Diatomic Molecules " 2nd edtn. Macmillan & Co. London 1950. ABRAHAMS STEREOCHEMISTRY OF SUB-GROUP VI 413 are known to establish bond length-bond order relations are probably the sulphur-oxygen and carbon-oxygen bonds. Of these the former bond is discussed on p.427 and only the latter bond will be treated here. The ability of bonds to oxygen to partake of double-bond character in varying degree is well illustrated by the carbon-oxygen bond. Cox and Jeffrey 56 have summarised the data for 22 different molecules containing single carbon-oxygen bonds and find the mean carbon-oxygen distance to be 1.437 8. The corresponding double-bond distance based on deter- minations for eleven different molecules results in a mean C=O value of 1.185 A Cox and Jeffrey suggest taking 1.44 A as the standard single carbon-oxygen bond length but feel that 1.19 A for the standard double- bond length is less reliable in view of the postulate 57 that t'he double-bond length may vary with the polarity of the bond. Vaughan and Donohue 58 have proposed C-0 = 1.42 A and C=O = 1.20 8.TABLE 3. Recent carbon-oxygen bond-letagth determinations Molecule Carbon monoxide . . Carbonyl chloride . . Anthraquinone . . . Carbonyl selenido . . Carbonyl sulphide . . isoCyanic acid . . . Methyl isocyanate . . Acetone . . . . . p-Dimethoxybenzerre . Furan . . . . . . Diketen . . . . . %tthyl ether . . . . 1 4-Dioxan . . . . Methylether . . . . Methanol . . . . . Ethanol . . . . . Bond length (A) 1.131 -& 0-005 1.166 0.002 1.15 $ 0.02 1.1588 * 0.0001 ~1.1637 & 0.0005 1.17 3 0.01 1.184 & 0.02 1.18 & 0.03 1.23 2 0.03 1.24 -f 0.03 1-36 5 0.02 1.372 1.377 1.15 -k 0.02 1-22 -I- 0.03 1.24 :k 0.06 1.39 .k 0.06 1.43 $ 0.02 1.44 0-03 1.46 4 0.03 1.42 f 0-03 1.43 0.008 1-44 4 0.01 1.48 & 0.04 1-48 0.04 Method M.W. XIkay M.W. 9 9 X-Ray E.D.7 9 3 X-Ray M.W. E.D. X-Ray E.D. 9 Y , X-Ray M.W. E.D. Y Y 9 Ref. 59 60 61 62 63 64 65 66 67 68 69 70 21 11 12 7 1 71 4 4 4 72 73 4 74 74 5 6 Cox and Jeffrey Proc. Roy. Xoc. 1951 207 A 110. 57Walsh Trans. Faraday Soc. 1947 43 60. 58Vaughan and Donohue Acta Cryst. 1952 5 530. 59 Gilliam Johnson and Gordy Phys. Rev. 1950 78 140. 6o Robinson J. Chem. Phys. 1953 21 1741. 61 Zaslow Atoji and Lipscomb Acta Cryst. 1952 5 833. 6 2 Sen Indian J . Phys. 1948 22 347. 63 Strandberg Wentink and Hill Phys. Rev. 1949 75 827. 6 4 Strandberg Wentink and Kyhl ibid. p. 270. 6 5 Jones Shoolery Shulman and Yost J . Chem. Phys. 1950 18 990. 6 6 von Dohlen and Carpenter Acta Cryst. 1955 8 646. 414 QUARTERLY REVIEWS Recent carbon-oxygen bond-length determinations are in good agree- ment with the value 1.43 A for a single bond and with the smaller value of about 1.17 A for the double bond (see Table 3).The orders of bonds of intermediate length are a t present insufficiently established to allow the construction of a relationship other than linear between bond order and length. The Stereochemistry of Sulphur (a) Two-bonded 8uZphur.-Sulphur in common with all the elements of sub-group VIB has the valency electron configuration (ns) 2( npJ (np,) (np,) 2 where n = 3 for sulphur. Unlike oxygen however the 3d orbitals are available for sulphur bond formation in an expansion of the outer octet to a decet or even a duodecet of electrons. The experimental values for the sulphur valency angle in which two other atoms only are linked to the atom under consideration are listed in Table 4.The general similarity in bond angle for sulphur in this table and oxygen in Table 1 indicates that the bond mechanism for sulphur and oxygen cannot be very different. Thus for the S-S-S bond angle the measured values vary only from 103" to 108" with a mean value of 106" (di-iododiethyl trisulphide is an exception). The angular range used is suggestive of a considerable s-admixture in the sulphur bonds with an approach to sp3 hybridisation. The bond angle in S has been discussed 75 on the basis of nuclear quadruple resonance measurements. If it is assumed that the observed angle of 1074" is also the angle between the axes of the two orbitals originating a t the sulphur atoms about 20y0 of s character is estimated in the bond. An alternative view is that if the bond angle is not identical with the inter- orbital angle then nearly pure p orbitals could be used and the increased size of the bond angle over 90" could be attributed to a pivoting of the p orbitals.One interesting difference between the behaviour of sulphur and oxygen is that apart from SO, the 2-bonded sulphur valency angle appears never tb exceed 109-5" even when aromatic groups are linked to the sulphur atom. The small valency angle of H2S (92.1") appears a t first sight to be a good example of the use of pure p orbitals. Burrus and Gordy 76 have suggested that this angle as well as the nuclear couplings can be accounted for by postulating roughly equal (ca. 15% of each) s and d contributions to the bonding orbitals. The admixture of both s and d orbitals tends to have 67 Eyster Gillette and Brockway J .Amer. Chem. SOC. 1940 62 3236. 68 Allen Bowen Sutton and Bastiansen Trans. Faraday SOC. 1952 48 991. 68 Bauer quoted in Ann. Rev. Phys. Chem. 1953 4 245. 'OKimura and Kurita J . Ghem. SOC. Japan 1951 72 396. 'lKatz and Lipscomb Acta Cryst. 1952 5 313. 7 2 Tauer and Lipscomb ibid. p. 606. 731vash and Dennison J . Ghem. Phys. 1953 21 1804. 7 4 Kimura J . Chem. SOC. Japan 1950 71 18. 76Dehmelt Phys. Rev. 1963 91 313. 70Burr~a and Gordy ibid. 1953 92 274. ABRAHAMS STEREOCHEMISTRY OF SUB-GROUP VIB TABLE 4. Xulpphur valency angle Compound Angle Value Method (A) Sulphur linked only to two other sulphur atoms (CF,),S . . . . . . . . . . BaS,O ,,2H20 . . . . . . . . BaSeS,O ,2H,O . . . . . . CH,*SO,*S*S*SO,.CH . . . . . . CH3*5,*CH . . . . . . . . . BaS,O 6 ,2H,O (orthorhombic) .. . * . . . BaS,,H,O . . . . . I . BaS60,,2H,O (triclinic) . . . . . s . . . . . 1 . . . . cs2s6. . . . . . . . . I*C2H4*S,.C,H4.1' . . . . . . . C6H5*SO,.s*S02.b~& . . . . . . 103.8" :k 3" 103" f 2" 103" f 2" 104" f 3" 104" 5" 104" f 2" 104.5" f 1" 106.5" Ij 1" 107" & 3" 107.8" f 0.5" 108.8" f 2" 113" f 2" E.D. X-Ray 9 Z.G. X-Ray 7 9 $ 9 9 9 9 1 Y (B) Sulphur linked only t o two other atoms C,H,S . . . . . H S . . . . . . HDS. . . . . CH,*SH . . . . . (I'C,H4),S . . . . As486 . . . sc1 . . . . . . S,N4 . . . . . . P,S . . . . . . As,S,. . . . . . ( p -CH,*C ,H4.S0 ,*s) ,Te BaSes,0,,2H2O . . (CH,),S,. . . . . (C6H5*S0,*S),Te . . 8e(SCN) . . . . PAS . . . . . . !NH4)2(S203)2Te . . c-s-c H-S-H H-S-D s-s-c C-S-H AS-S-AS c1-S-cl AS-S-AS N-S-N P-s-P S-5-Te S-S-Se s-s-c S-S-Te S-S-Te Se-S-C P-s-P 8-s-c1 S-S-Te s-s-c 8-s-c s-s-c s-s-c c-s-c P-s-P c-s-c C-S-H 0-s-0 65.8" f 0.1" 92.1" f 0.2" 98" f 10" 99.4" f 0.5" 93.3" * 0.2" 100" f 2" 102" & 3" 102" f 3" 102" f 3" 100-3" 103" & 3" 103" f 3" 104" f 5" 104" f 2" 104" 104" & 5" 104" & 1" 104.5" & 2.5' 105" f 3" 105.4" f 3" 105.6" f.3" 107" & 1" 107" & 3" 109" f 2" 109" f. 1" 109.5" & 1" 113" f 2" 119.5" f 0.5" 119.0" & 0.5" M.W. 9 9 X k M.W. E.D. Raman X-Ray 3 Y ? 9 9 Y 9 9 9 E.G. X-Ray E.D. X-kay E.D. X-Ray E.D. M.W. > Y 9 7 77Bowen Trans. Paraday SOC. 1954 50 452. '~Foss Furberg and Zachariasen Acta Chem. Scand. 1954 8 459. 7*Foss and Tjomsland ibid. p. 1701. 81Donohue and Schomaker J. Chem. Phys. 1948 16 92. 82Foss and Zachariasen Acta Chern. Scand. 1954 8 473.a3Abrahams Acta Cryst. 1954 7 423. Sorum ibid. 1953 7 1. 415 Ref. 77 78 79 80 81 82 83 84 85 86 87 88 3 76 89 90 91 92 93 94 95 96 97 79 81 98 99 100 101 4 103 77 77 103 104 105 101 103 106 107 108 416 QUARTERLY REVIEWS a cancelling effect on the bond angle but adds up to produce the observed asymmetry in the molecular electric field. (b) PoEy-sulphur Chains.-An outstanding characteristic which distin- guishes the behaviour of sulphur from that of oxygen is the ability to form long chains. Thus the stable state of elementary sulphur consists of s8 rings and compounds such as S,,,CI are known.log There has been no confirmed report of a branched poly-sulphur chain although rings helices and other non-branched forms are well established. A review of the X-ray electron-diffraction and spectroscopic data has been given ll0 which shows S-S formation is very unlikely.The only cases in which sulphur forms a single (one only) bond t o sulphur appear to be in ions such as [0,S-Sl2- and as termind members of poly-sulphide chains. In orthorhombic sulphur 86 the puckered eight-membered ring has the sym- metry 62m within the limits of experimental error. In the S62- ion8' the chain is helical while the [S,06]z- ion 8 2 possesses a plane of symmetry passing through the middle atom of the sulphur chain. The sulphur chain in this ion is hence equivalent to an s8 ring with three linked sulphur atoms missing. The sulphur-sulphur bond (Table 5 ) varies in length from 1-89 to 2.39 A. The length chosen for a single bond by Pauling 111 was 2.08 A and indeed that -s\ - s 8 4 Mathieson and Robertson J .1949 724. 8 6 Foss and Tjomsland Actu Chem. Scand. 1955 9 1016. 86Abrahams Actu Cryst. 1955 8 661. 87 Abrahains and Grison ibid. 1953 6 206. Dawson and Robertson J. 1948 1256. 89 Bird and Townes Phys. Rev. 1954 94 1203. 90Donohue J. Amer. Chem. SOC. 1950 72 2701. 91 Solimene arid Dailey Phys. Rev. 1953 91 464. 92Lu and Donohue J . Amer. Chem. SOC. 1944 66 818. 9 3 Stammreich Forneris and Sone J . Chem. Phys. 1965 23 972. 941to Morimoto and Sadanaga Actu Cryst. 1952 5 775. 95C1ark J. 1952 1615. 96 van Houten Vos and Wiegers Rec. Trav. chim. 1955 74 1167. 97Foss and i)yum Actu Chem. Scund. 1955 9 1014. 980yum and FOSS ibid. p. 1012. 99 Foss and Larssen ibid. 1954 8 1042. loo Ohlberg and Vaughan J . Amer. Chem. Xoc. 1954 76 2649. lolVos and Wiebenga Actu Cryst.1955 8 217. lo2Foss and Vihovde Actu Chem. Scund. 1954 8 1032. lo3Toussaint Bull. SOC. chim. Belg. 1945 54 310. lo4Stevenson and Beach J . Amer. Chem. SOC. 1938 60 2872. lo5 Blackmore and Abrahams Actu Cryst. 1955 8 329. l o 6 Rouault and Gallagher Phys. Rev. 1949 75 1319. l o 7 Sirvetz J. Chem. Phys. 1951 19 938. 108 Crable and Smith ibid. p. 602. lO9Feh6r and Baudler 2. anorg. Chem. 1952 267 293. 110 FOSS Acta Chem. Scund. 1950 4 404 ; Woodrow Carmack and Miller J . Chem. ll1 Pauling " Nature of the Chemical Bond " 2nd edtn. Cornell Univ. Press Phys. 1951 19 951 ; Minoura J. Chem. SOC. Jupun 1952 73 244. Ithaca N.Y. 1940. ABRAHAMS STEREOCHEMISTRY OF SUB-GROUP VIB 41 7 this value is identical with the average of the 43 bond lengths given in Table 5. If 2-08 8 is accepted as a single-bond length the measured value in orthorhombic sulphur (S,) is short and this bond must hence possess some double-bond character.to account for the small heat of reaction of opening the S ring. Koch 113 suggested that in addition to the conventional canonical forms which can contribute to the structure of the S molecule an alternating no-bond double-bond structure might be important. The only sulphur-sulphur bonds which have been measured as signi- ficantly longer than 2.08 8 the value proposed as a standard single-bond length occur in decafluorine disulphide and sodium dithionite. Dunitz 114 has discussed the [i3204]2- ion sulphur-sulphur bond length in terms of Pauling's relationship,115 -4R(n) = 0.353 log, n for which n is 0.36. I n the case of F,S*SF, n is 0.69.Both bonds are hence substantially less than single ,bonds. Huggins 116 has recently calculated the length of the sulphur-sulphur single bond to be 2.053 & 0.02 fi on the basis of a new relation between bond energies and radii. The sulphur-sulphur bond length in S in the 3C,- (ground) state has been spectroscopica'lly determined to be 1.887 A. As in the case of 0 in t,he ground state this bond corresponds fa'irly closely to a double bond. Intermediate points between 2.08 and 1.89 A for the lengths of single and double sulphur-sulphur bonds respectively are available. The short I)onds of length 2.02 A in the tetra- and hexa-sixlphide ions have been I)ostulated 83 to correspond to about 50-33% of double-bond character 1 hus defining the shape of the sulphur-sulphur bond order-bond length carve.It may be noticed in Table 5 (B) that a variation in bond length along poly-sulphur chains has been measured only if four or more linked sulphur atoms are present. In this Table the observed variations are close to the cstimated errors in the individual bond lengths and hence may not be significant. For the six examples reported the sulphur-sulphur bonds are alternately short and long for a chain containing an odd number of such 1)onds. If an even number of sulphur-sulphur bonds are present in the chain the two central bonds appear to be of equal length and the equivalent of a plane of symmetry is introduced into the chain. I n both cases chemi- cally equivalent bonds are of identical length within the limits of error. The outermost bonds are the short bonds unless oxygen atoms are linked to the terminal sulphur atoms when it is the outer sulphur-oxygen bonds that are short.(c) Dihedral Angle.-The dihedral angle in sulphur as in oxygen is determined primarily by the pn electron repulsion with the greatest re- pulsion between the unshared pairs of electrons on adjacent sulphur atoms. This suggestion had previously been made 112 Powell and Eyring J . Amer. Chem. SOC. 1943 65 648. 113 Koch J . 1949 408. l14Dunitz Acta Cryst. 1956 9 579. 115 Pauling J . Amer. Chem. SOC. 1947 69 542. 116 Huggins ibid. 1953 '75 4126. 418 QUARTERLY REVIEWS TABLE 6. Sulphur-sulphur bond lengths I Compound I Bond length (a) I Method I Ref. I (A) Without bond length variations s . . . . . . . . Na,S,03,5H,0 . . . . . S8 . . . . . . . (c~,i,s . . . Cl,N,O,Hl8S2,2Hz0 * .. . $z:H,);S . . . . . . ( CF3),S2 . . . . . . . (CF3)2S3 * . . . . . (C,H5*S02),S . . . . . C1,SZ (C,H,*SO,*S),Te . . . . NaK,Cl,(S,06)2 . . . . NaK,C1,(S,06) . . . . (CH,*C6H,*SO2.S),Te . . . (NH,),TeS,O . . . . . BaSeS40,,2H,0 . . . . (CH,*SO,*S),Te . . . . SzFlo . . . . . . . . . . . . . . . . . . . K2S206 . . . . . Na,S,O . . . . . . Na2S,0,,2H;0 . . . . . 1.887 1.97 f 0.06 2.037 f 0.005 2.04 f 0.02 2.04 & 0.005 2.05 f. 0.02 2.05 f 0.04 2-053 * 0.019 2.065 f 0.016 2-07 & 0.02 2.07 & 0.10 2.08 & 0.03 2.08 f 0.04 2.08 & 0.04 2.11 -j= 0.04 2.11 f 0.03 2.13 f 0.04 2.14 f 0.03 2.14 & 0.02 2.16 f 0-02 2.21 f 0.03 2.389 f 0.010 Spec. X-Ray E.6. X-Ray E.D. X-Ray E.D. x’-itay E.D. X-Ray 117 118 86 81 119 104 88 77 77 84 4 98 120 120 97 99 79 102 121 122 123 114 (B) With variations in the bond length BaS,,H,O .. . . . . cs,s . . . . . . . BaS40,,2H,0 . . . . . BaS,0,,2H20 (triclinic) . . BaS60,,2H,O (orthorhombic) (CH,*SO,),S . . . . . 2-02/2-07/2.02 1*99/2*10/2*03/2* 12/2*03 2.10/2*02/2*13 2.12/2-04/2.04/2* 10 2.1 4/2-04/2.04/2.14 2*10/2*06/2*10 & 0.025 f 0.03 f 0.03 -& 0.04 & 0.03 * 0.03 ~ 83 87 78 85 82 80 * NN’-Diglycyl-L-cystine dihydrste. Pauling 12* has deduced a relationship between the dihedral angle and the valency angle in sulphur rings for various numbers of sulphur atoms in the ring. For orthorhombic sulphur Pauling’s relationship predicts a dihedral ll’Ikenoue J . Phys. SOC. Japan 1953 8 646. llsTaylor and Beevers Actu Cryst. 1952 5 341. ll9 Yakel and Hughes ibid. 1954 7 291. Stanley ibid. 1953 6 187.Martinez Garcia-Blanco and Rivoir ibid. 1956 9 95. lZ1 Idem ibid. in the press. 123Harvey and Bauer J. Amer. Chem. Xoc. 1953 75 2840. 124 Pauling Proc. Nut. Acad. Sci. 1949 35 495. ABRAHAMS STEREOCHEMISTRY OF SUB-GROUP VIB 419 angle of 99-0" based on the observed bond angle of 107.8" This prediction may be compared with the observed value of 99.3" (see Table 6). The sulphur dihedral angle has now been experimentally measured in a number of cases and these results are tabulated in Table 6. It is f o h d that the angles range from 74" to 110" and hence are distributed on both sides of 90". The establishment of the sulphur dihedral angle a t about 90" leads to the result that pentasulphide chains or groups such as X-S-S-S-Y may possess two isomeric forms. Poss 125 has pointed out that if the isomer is of the cis-form i.e.with the two terminal S-S or S-X S-Y bonds rotated through about 90" on the Same side of the plane of the three central atoms the group may be considered as derived from the S8 ring. If the isomer is of the trans-form i.e. with the terminal atoms on opposite sides of the central plane a helix will be formed which can be right- or left-handed. Of the compounds in Table 6 it is noteworthy that those crystallising in the cis-form (88 and S5062-) have dihedral angles greater than go" while those in the trans-form (S62- and di-iododiethyl trisulphide) have dihedral angles less than 90". TABLE 6. Sulphur dihedral angle I Compound CszS . . . . . . . BaS,,H,O . . . . . . (I*C,H,),S . . . . . . H,S . . . . . . . (CH,-SO,),S . . . . . BaS,06,2H,0 .. . . . Cl,S . . . . . . . (CH3)2s3 . . . . . . . . . S 8 * C10N406k18S2,2HZ0 * . . . BaS,06,2H,0 (triclinic) . . BaX5O,,2H,O (orthorhombic) Dihedral angle 74.0" 75.6" 82" ca. 90" ca. 90" 90" 92 " 93" 99.3" 101" 106.5" 110" Method X-Ray 9 9 Spec. X-Ray E.6. x'kay 2 ) Y 9 Ref. 87 83 90 126 80 78 4 81 86 119 85 82 * NN'-Diglycyl-L-cystine dihydrate. (d) Heterocyclic Sulphur.-The valency angle of two- bonded sulphur does not necessarily remain unchanged when constraints are placed on it such as when the sulphur atom forms part of a ring molecule. One of the earliest heterocycles to be investigated was thiophen in which the C-S-C angle is not about 105" as might be expected on the basis of Table 4 (B) but 91" [see Table 7 (A)]. The authors 127 suggested that some of the sulphur 3d orbitals might be used in valency-bond structures of the forms (VII1)-(X) etc.; Longuet-Higgins 128 extended this argument employing lZ5Foss Acta Chem. Scand. 1953 7 1221. 126Wilson and Badger J. Chem. Phys. 1949 17 1232. lZ8 Longuet-Higgins Trans. Paraday SOC. 1949 45 173. Schomaker and Pauling J. Amer. Chem. Soc. 1939 61 1769. D D 420 QUARTERLY REVIEWS the molecular-orbital method and computed the mobile bond orders in thiophen to be C-S = 0.59 C(a)-C(3) = 0.73 and C(3)-C(4) = 0.61. The + (Vlll) (I x) (X) bond order-length curve of Fig. 3 being used the predicted carbon-sulphur bond length is then 1.73 A in excellent agreement with the experimental value of 1.74 8. Direct calculation of the C-S bond order in thiophen as 0.58 has also been r e ~ 0 r t e d . l ~ ~ An accurate determination of the molecular structure by modern methods has not yet been made although the crystal structure of the high-temperature disordered form has been studied.130 Double- bond character (%) FIG.3 Variation of the carbon-sulphur bond length with bond character. There have not been any systematic attempts to explain the variations in the sulphur valency angle Table 7 (A) from 91" to 101" in the syst,em =CR-S-CR= and from 65.8" to 1066" in the system -CR,-S-CR,-. In contrast to those in Table 7 (A) the values given for the sulphur valency angle in Table 7 (B) as well as the nitrogen-sulphur bond length show no variation. The interesting case of the ethylene sulphide molecule with a sulphur bond angle of 65~8"~ has been compared with cyclopropane for which Coulson and Moffitt 131 suggested that the bonds are bent and lie along the arc tangents to the bonding orbitals.de Hem J. Amer. Chem. Xoc. 1954 76 4802. l3OAbrahams and Lipscomb Acta Cryst. 1952 5 93. 131 Coulson and Moffitt Phil. Mag. 1949 40 1. ABRAHAMS STEREOCHEMISTRY OF SUB-GROUP VIB 42 1 TABLE 7 (A). Sulphur valency angle and carbon-sulphur bond length in heterocyclic molecules Molecule Molecule Valency angle N-S Bond length (A) Method 1 Ref. 1 (XI) . . . . . (XII) . . . . . (XIII). . . . . (XIV). . . . . (XV) . . . . . (XVI). . . . . (XVII) . . . . (XVIII) . . . . (XIX). . . . . Valency angle 658" 91" 91" 91.2" 98.3" 99.0" 101" 104" 106.5" C-S Bond length (A) 1.819 -& 0.001 1.74 5 0.03 1.84 5 0-03 1.72 & 0.013 1.74 rf 0.013 1.752 f 0.017 1.78 & 0.05 1.85 & 0.07 1.81 & 0.03 1.810 f 0.010 Method M.W.E.D. X-Ray 3 Y ? Y 1 9 9 E.D. Ref. 3 127 133 134 134 135 136 137 138 17 TABLE 7 (B). Sulphur ualency angle and nitrogen-sulphur bond length in heterocyclic molecules (XX) . . . . . 102O 1.60 -& 0.05 X-Ray 18 (XXI). . . . . ~ 102" ~ 1.60 f 0.05 ~ , ~ 95 ~ A spectroscopic and thermodynamic study 132 of tri- tetra- and penta- methylene sulphide indicates that these molecules are all coplanar in contrast to 1 4-dithian which.assumes a chair configuration 136 and p-dithiin with a boat configuration.137 (e) The Carbon-Sulphur Bond.-Enough reliable experimental and theoretical work is now available to derive a bond order-bond length curve for the carbon-sulphur bond. for the single carbon-sulphur bond length from a list of eleven molecules containing a l3 Scott Finke Hubbard McCullough Katz Gross Messerly 'Pennington and Waddington J .Amer. Chem. SOC. 1953 75 2795. laaEeles Acta Cryst. 1956 9 365. la4 Cox Gillot and Jeffrey ibid. 1949 2 356. 13s Jeffrey ibid. 1951 4 58. laa Brahde Acta Chern. Scand. 1954 8 1145. Cox and Jeffrey 56 obtained a mean value of 1.812 la6 Marsh ibid. 1955 8 91. Howell Curtis and Lipscomb ibid. 1954 7 498. 422 QUARTERLY REVIEWS formal single bond. This value is also the sum of the Pauling ll1 covalent radii of sulphur and carbon. Huggins 116 has proposed a new system of interatomic distances and has calculated the C-S single bond to be of length 1.83 & 0.02 8. The mean of the first six entries in Table 8 all of which should be close to formal single carbon-sulphur bonds involving only two bonds to the sulphur atom is 1-83 8.The weighted mean of all these bond lengths is 1-82 8 and this value may hence be taken as the length of a pure single carbon-sulphur bond. The determination of the length of a pure double carbon-sulphur bond offers greater difficulty. The sum of the Pauling radii is 1.61 8 and among the recent data in Table 8 there is no molecule with an unambiguous double carbon-sulphur bond ; e.g. in O=C=S the molcule is assumed 139 to have the resonance forms O=C=S 0-C-S O=C-S with a respective impor- tance of 58 14 and 28%. Molecular-orbital calculations 1*0 on thiophthen show that the experimental 13* carbon-sulphur bond lengths of 1-72 and 1-74 8 correspond to bond orders of 1.54 and 1.49 respectively. By assuming 66 a single-bond length of 1-81 8 and a linear bond order-length relation the length of a double bond was demonstrated to be about 1-65 8.The entries in Table 8 however suggest that the Pauling value of 1.61 A might be closer to the length of a pure double carbon-sulphur bond. It is clearly very desirable to measure this length in a well-defined double bond. If we take 1-61 8 for the double 1-82 8 for the single and the data based on thiophthen for the intermediate carbon-sulphur bond lengths as in Fig. 3 the results in Table 8 yield some interesting information. Thus di-p-tolyl and di-p-bromophenyl sulphide each possess about 40% of double- bond character in the C-S bond. The environment for the sulphur atom in these molecules is similar to that in thiophen or thiophthen namely =CH-S-CH= for which pd hybridisation was assumed (p.419). Related evidence supporting the view that the bonds in these molecules have con- siderable double-bond character is provided by the tendency to planarity in the molecules. Thus in di-p-tolyl sulphide the angle between the normals to the two benzene rings is only 56" resulting in a 3-19 A contact between nearest carbon atoms in the different benzene rings. If this angle were go" the closest contact of this kind would be 3-99 8 demonstrating that the C-S bond here cannot be single with cylindrical symmetry. Simil- arly in the case of di-p-bromophenyl sulphide the two benzene rings are each rotated about 36" out of the Br-S-Br plane resulting in a dihedral angle between the two aromatic rings of about 60". A similar effect is found in the sulphoxides for dimethyl sulphoxide has a pure single bond on the basis of Fig.3 while diphenyl sulphoxide in which the sulphur atom has a thiophen-like environment again has about 3540% of double-bond character. In the case of the sulphones no such effect appears for the C-S bond lengths in dimethyl and di-p-bromophenyl sulphone are equal within the error of observation to a single bond. - + - I - - 139Townes and Dailey J. Chem. Phys. 1949 17 782. 140Evans and de Heer Acta Cryst. 1949 2 363. ABRAHAMS STEREOCHEMISTRY OF SUB-GROUP VIB 423 The carbon-sulphur bond length [see also Table 7 (A)] TABLE 8 . I Molecule I C-S Bond length (A) I Method I Ref . I CH. *SH . . . . . . . . . (CH. *S)4C . . . . . . . . (CF.). S . . . . . . . . . (CF.).S. . . . . . . . . . (CF.).S. . . . . . . . . . (I.C,H,),S . . . .. . . . cs . . . . . . . . . TeCS . . . . . . . . . HNCS . . . . . . . . . ocs . . . . . . . . . . cs. . . . . . . . . [Hg(SCN).]['Clu(en). ] . . . . . NH. SCN . . . . . . . . (NH. *CS *) . . . . . . . . . (XXII) . . . . . . . . . (XXIII) . . . . . . . . (CH.*SO.).C.C.N.CH. . . . . . /3-CH,.C4H,*S0 t . . . . . . (p.Br*C.H.). S . . . . . . . (C.H.). SO . . . . . . . . (p.CH.*C. H4)2S . . . . . . (C.H..SO. ).S . . . . . . . (C.H..SO.). Se . . . . . . . CH.. CO .SH . . . . . . . (CH,.SO,-S),Te . . . . . . (p -Br*C . H.) . s . . . . . . . . NaSO..CH. *OH . . . . . . (CH.*SO.).S. . . . . . . . (CH.).SO. . . . . . . . . (CH.). SO . . . . . . . . (p.Br.C.H.).SO. . . . . . . NN'-Diglycyl-L-cystine dihydrate . 1.808 1.81 f 0.02 1.828 f 0-015 1.829 f 0.017 1.848 & 0.015 1-86 4 0.05 1.5349 & 0.0002 1.557 & 0.010 1.557 f 0.010 1.561 & 0.002 1.5586 & 0.0005 1.56 & 0.02 1.57 f 0.10 1.59 1.663 1.68 f 0.02 1.708 & 0.008 1.726 & 0.007 1.770 f 0.009 1.744 & 0.017 1.75 & 0.03 1.75 5 0.03 1.760 & 0.015 1-76 5 0.02 1.77 & 0.05 1.77 f 0.05 1.78 & 0.02 1.80 & 0.02 1.80 & 0.06 1.80 & 0.04 1-82 1.838 & 0.011 1.84 & 0.04 1.87 0.017 M.W. X-Ray E.D. 9 9 .. .. M.W. .. .. 9 9 .. E.D. X-Ray Y Y 9 7 .. .. .. .. .. .. .. 9 9 .. .. E.6. $-Ray E.5. X-Ray .. .. 91 142 77 77 77 90 143 144 145 146 64 17 147 148 149 150 151 152 152 135 105 103 153 84 154 80 155 156 102 103 157 158 103 119 * en = Ethylenediamine . t /3-Isoprene sulphone . 1 4 1 Tarbell and Harnish. Chem . Rev. 1951. 49. 1 . 142Perdock and Terpstra. Rec . Truv .chim. 1943. 62. 687 . 14. Mockler and Bird. Phys . Rev. 1955. 98. 1837 . 144Hardy and Silvey. ibid. 1954. 95. 385 . lP5 Beard and Dailey. J . Chem . Phys. 1950. 18. 1437 . 346 Dousmanis. Sanders. Tomes. and Zeiger. ibid. 1953. 21. 1416 . 147 Scouloudi. Acta Cryst. 1953. 6. 651 . 143Zvonkova and Zhdsnov. Zhur . Fiz . Khim. 1949. 28. 1495 . l49 Long. Markey. and Wheatley. Actu Cryst. 1954. 7. 140 . l5O Penfold. ibid. 1953. 6. 707 . lS1 Wheatley. ibid. p . 369 . lS2 Idem. ibid. 1964. 7. 68 . 424 QUARTERLY REVIEWS X-S-X Y-S-O I * Molecule The case of dimethanesulphonyl disulphide (CH,*SO,),S is interesting for the C-S bond is again shorter than single (about 30% of double-bond character) although the sulphur atom is in a different environment from the examples just cited. However in this molecule the sulphur-sulphur bond lengths vary as shown in Table 5 (B) and the short carbon-sulphur bond confirms the view that in such molecules the outermost bonds are the short ones.An excellent review of the literature dealing with the cleavage of the carbon-sulphur bond in bivalent sulphur compounds has been given by Tarbell and Harnish.141 (f) Expansion of th Sulphur Valency-electron 0ctet.-In the two preceding sections pd hybridisation involving the sulphur 3d orbitals was postulated in the group %-S-C' Considerable additional evidence supporting the premise of an expansion in the sulphur valency-electron octet to a decet or duodecet has now been presented,159 based largely upon ultraviolet absorption spectra measurements. (g) Three-bonded Sulphur.-The stereochemistry of sulphur in the sulphoxide grouping 'S-0 once a matter for much debate is now relatively well established.The three sulphur bonds form a shallow pyramid and the molecular dimensions of the measured sulphoxides are given in Table 9. The nature of the orbitals used by sulphur in the / \- / rs0 (A) ~ Method I Ref. TABLE 9. Dimensions in the sulphoxide group 106-8" & 0.1" 106" f 1" 108" & 2" 107" i. 5' 106" f 6" 106.2' & 0.7" 92-8" -& 0.1" 114" f 2" * 96" & 2" 100" 5" - 97.3" f 1.0" 1.412 f 0.001 1.45 & 0.02 1.45 (assumed) 1.47 0.03 1.47 & 0.03 1.473 f 0.015 I * Angles sma.ller than 106" were not used in the models tried. 160 161 162 157 4 153 l5 Abrahams and Grenville-Wells M.I.T. Laboratory for Insulation Research 154Furberg and Oyum Acta Chenz. Scand.1954 8 42. 155Gordy J . Chem. Phys. 1946 14 560. 156Lister and Sutton Trans. Faraday SOC. 1939 35 495. 157 Bastiansen and Viervoll Acta Chem. Scand. 1948 2 702. 158 Truter J . 1955 3064. 15D Eastman and Wagner J . Amer. Chem. SOC. 1949 71 4089 ; Price and Morita ibid. 1953,75,4747 ; Bordwell and Andersen ibid. p . 6019 ; Rothstein J. 1953 3991 ; Cilento and Walter J . Amer. Chem. Soc. 1954 76 4469 ; Jaff6 J . Chenz. Phys. 1954 22 1430 ; Mangini and Passerini Gazzetta 1954 84 606. 160Ferguson J . Amer. Chem. Soc. 1954 76 850. l6lPalrner ibid. 1938 60 2360. 162 Stevenson and Cooley ibid. 1940 62 2477. Tech. Rep. No. 105 1956. ABRAHAMS STEREOCHEMISTRY O F SUB-GROUP VIB 425 sulphoxide group has not yet been determined. In diphenyl sulphoxide 153 the angle between the benzene rings and the C-S-C plane is 81.9" in agree- ment with the suggestion made for diphenyl sulphone 113 159 163 that the sulphur 3d and the carbon Zp orbitals overlap resulting in the planes of the aromatic rings becoming normal to the C-S-C plane.A striking feature of Table 9 is the constancy of the X-S-0 angle whereas the X-S-X angle varies from about 93" to 114". In the case of the corresponding sulphides this variation in X-S-X angle has been in part correlated with the change in admixture of s p and possible d orbitals. The presence of the non-orthogonal pyramidal bond arrangement appears to be evidence for some hybridisation rather than for pure p orbitals' being used by sulphur in the sulphoxide group. In the case of SO, the arrangement of the 3-bonded sulphur atom is quite different.In the gas state this molecule has zero dipole moment suggesting a planar molecule with a trigonal arrangement of oxygen atoms. This model has been confirmed by electron-diffraction methods 161 which show the 0-S-0 bond angle to be 120" & 2". In the solid state both the ice- l G 4 and the asbestos-like form of sulphur trioxide have the molecules linked together through S-0-S groups the resulting four sulphur bonds being arranged approximately tetrahedrally. Siebert,l66 using Raman spectra has shown that the group symmetry for the (CH,),S+ ion is C3v i.e. is pyramidal in shape with the sulphur atom a t the apex. (h) Four-bonded Sulphur.-An excellent review of the preparative chemistry of 4-covalent sulphur has been given by Suter.167 I n the 4-bonded state sulphur is very stable and may be observed in many different groupings.One of the commonest is the sulphate group and Wells 29 has summarised the earlier investigations demonstrating the tetrahedral con- figuration of the four sulphur-oxygen bonds. A recent confirmation of this model has been made l9 for potassium ethyl sulphate in which the mean 0-S-0 angle is 109". A similar arrangement is found 118 in Na,S,0,,5H20 with bond angles in the range 104-1 15". Recent work on the [S2Q6]2- ion,l20 the [S30,,]2- ion,lG8 and the [SO,]2- ion 169 has confirmed the earlier assignment of a tetrahedral bond distribution for 4-bonded sulphur. The molecular structure of several sulphones has now been elucidated and the results are summarised in Table 10. With the exception of /3-isoprene sulphone the carbon (or halogen)- sulphur-oxygen angle has very nearly the same constant value of about 107" in the sulphoxides and in the sulphones.The F-S-F bond angle is A related bond arrangement is found in the sulphones. 163Koch and Moffitt Trans. Faraday Soc. 1951 47 7. 164 Westrik and MacGillavry Rec Trav. chim. 1941 60 794. 166 Idem Acta Cryst. 1954 7 764. 166 Siebert 2. anorg. Chem. 1952 271 65. 167 Suter " The organic chemistry of sulfur " J. Wiley Sons Inc. N.Y. 1944. 16* Eriks and MacGillavry Acta Cryst. 1953 '7 430. 169 Larson and Helmholz J. Chem. Phys. 1954 22 2049. 426 QUARTERLY REVIEWS identical at 92.8" in thionyl and in sulphonyl fluoride and a strong similarity also exists between other corresponding pairs of angles in the sulphones and sulphoxides (Tables 9 and 10). It may be noticed that the 0-S-0 angle in the sulphone group is invariably the largest sulphur bond angle TABLE 10.Dimensions in the sulphone group A Molecule x-s-0 x-s-x F,SO . . . . Cl,SO . . . (CH,),SO . . . /3-CHS.C,H,*SO * . (p-Br-C,H,),SO . CH,.N:C:C (p-I.C,H,),SO,. . (SO,.CH,)a 'r 107.1" f 0.5" 106.5" & 2" 105" f 3 " 108.7" & 5" 111" &4" 107.7" f 1" 99.6" f 1.5" 92.8" f0.5" 11 1.2" f 2" 115" &15" 98.3" & 1.5" 100" h0.5" 106" f 2 " 106*8" f0.4" * &Isoprene sulphone. t N-Methyl-2 2-dimethylsu A 0-s-0 129.6" &0.5" 119.8" & 5" 125" f15" 112.9" f 1.5" 131" 1 3 " 111" +4" 118.4" f 0.4" 1.37 &0.01 1.43 f 0 . 0 2 1.43 f 0 . 0 2 1,436 f0.017 1.54 10.05 1.433 f0.006 - lhonylvinylideneamine . Method M.W. E.D. X'-'Ray , 7 3 3 Ref. 173 161 4 135 103 174 152 in a given molecule and also is always greater than the tetrahedral value of 109.5" in contrast with the 0-S-0 angle in the [S,O,]2- ion 114 of 108.2" and in the [HO*CH,*SO,]- ion 158 of 108.5".The range in the C-S-C angle from 98" to 115" conflicts with the prediction 170 made on the basis of ring-closure experiments that the limits for this angle were 75" and 90" (cf. Table 7). The sulphone and sulphoxide groups are indeed so similar that it has recently been shown 171 that diphenyl sulphoxide can dissolve in diphenyl sulphone to form a continuous series of solid solutions which retain the diphenyl sulphone crystal structure in proportions up to 90% of diphenyl sulphoxide. At this concentration the bonds to nine in every ten pairs of sulphone oxygen atoms are replaced by one sulphur-oxygen bond and a lone pair of electrons which latter hence appears to have very nearly the same orientation as the missing sulphur-oxygen bond.The dipole moment of the molecule (C6H5)$12 has been measured 172 as 4.4 D. The analogous molecules (p-CH3*C6H4),SeX, where X is Br or C1 and (CH3),TeC1 have been shown to possess an unusual bond distribution and it mould be valuable to know if a similar bond distribution can also be exhibited by sulphur attached to four univalent groups. (i) Xix-bonded 8uZphur.-Very few compounds containing 6-bonded sulphur have been examined and their molecular shape determined. In the case of SF, a very stable molecule the shape is reported 175 on the basis of electron-diffraction experiments to be that of a regular octahedron l 7 O Luttringhaus and Buchholz Ber.1939 '72 2057. l 7 l Abrahams and Silverton Acta Cryst. 1956 9 283. 1V3 Jensen 2. anorg. Chem. 1943 250 245. f73Fristrom J . Chem. Phys. 1952 20 1. 17'Keil and Plieth 2. Krist. 1955 106 388. 176Brockway Rev. Mod. Physics 1936 8 231. ABRAHAMS STEREOCHEMISTRY OF SUB-GROUP VIB 427 with S-F = 1.58 Duffey 176 points out that there are no unshared electrons in the sulphur valency shell in SF, and predicts that the octahedral structure is more stable than the corresponding trigonal prism even if the ibmount of s-character were allowed to change. Another example of 6-bonded sulphur is disulphur decafluoride and Harvey and Bauer lZ3 have determined this structure to be (XXIV) with a 0.03 8. regular octahedral bond distribution and S-F = 1.56 0.02 8. The remain- ing example is disulphur decafluor~dioxide.~~ Here again all bond angles are right angles except S-0-0 which is 105” (cf.Table l) with S-F = 1-56 & 0.02 A. (j) The XuZphur-Oxygen Bond.-The nature of the sulphur-oxygen bond in various molecules has been the subject of much discussion largely c:entring around the question of whether the bond is a covalent double bond or a co-ordinate link or whether both bond types contribute. A review of much of the work in this field up to 1945 is given by Phillips Hunter and Sutton 177 who regard the bond as primarily double in nature. Wells 178 has criticised this paper and concluded that a t the time of writing there was no satisfactory explanation for this short bond. Jeffrey and Stadler 179 have tabulated the values of reliable S-0 distances in several oxides oxy- acids and acid salts of sulphur and show that this bond length appears to lle very constant at about 1.44 8.Moffitt 180 has used a molecular-orbital method to analyse the electronic structure of this bond. He finds that the sulphur-oxygen bond is largely tlouble in character and that the 3d orbitals of the sulphur atom are of importance in the formation of this bond. Moffitt’s bond-order assignments combined with the observed bond lengths in Tables 9 10 and 11 lead to it mean value of 1.425 8 for the double sulphur-oxygen bond length. A sulphur-oxygen bond that may be regarded as close to single has been measured l9 in potassium ethyl sulphate the anion of which has the structure (XXV). If this bond of length 1-60 8 is selected as a standard single bond WXV) and the value 1.43 8 as a standard double-bond length the resulting linear relationship closely fits Moffitt’s predicted bond orders for sulphur dioxide the sulphones and the sulphoxides.176Duffey J . Chem. Phys. 1950 18 128 and 510. 1 7 7 Phillips Hunter and Sutton J. 1945 146. 178 Wells J . 1949 55. laOMoffitt Proc. Roy. Soc. 1950 200 A 409. 178 Jeffrey and Stadler J. 1951 1467. 428 QUARTERLY REVIEWS A discussion of the nature of the sulphur-oxygen bond has recently been given by Simon and Kriegsmann,lsl who compute the length of the single bond to be 1.69 A based on Raman spectra observations. TABLE 11. Sulphur-oxygen bond lengths (see also Tables 9 and 10) Compound SO (ice-form) . . . . (C&50S0,)2Se . . . . BaS40,,2H,0 . . . . SO (asbestos form) . . (NH4),TeS406 . . .. (CH,-SO,*S),Te . . . BaS50,,2H,0 . . . . so . . . . . . . KZS206 - . . . . /?-CH,*C,H,*SO . . . (NH4),SO3-N,O . . . I<,SO,*NH,. . . . . K2NH(S03) . . . . Na,S,06,2H,0 . . . . C,H,*SO,K. . . . . HSO,-NH . . . . . (CH,*SO,),S . . . . NaK,Cl,(S,O,). . . . NaK,C1,(S,06) . . . Na,S,03,5H,0 . . . . NaSO,*CH,*OH . . . Li,SO,,H,O . . . . Na,S,O . . . . . Length (A) 1.40 & 0.05 1.60 & 0.05 1.41 & 0.05 1.41 & 0.04 1.63 f 0.04 1.43 f 0.04 1.43 f 0.06 1.43 f 0.04 1.4321 & 0.0005 1.432 f. 0.005 1.43 & 0.015 1.43 & 0.05 1.436 f 0.017 1.44 f 0.01 1.44 & 0.03 1.447 f 0.005 1.46 f 0.05 1-46 f 0.03 1-60 f 0.03 1.48 & 0.03 1-48 -j= 0.05 1-48 & 0.04 1-48 0.05 1.48 0.06 1-50 f 0.01 1.50 & 0.02 1.505 & 0.017 1.41 3 0.04 Method Ref. 164 164 154 78 165 165 99 102 82 108 107 182 121 135 179 179 183 122 19 19 184 80 120 120 118 158 169 114 The Stereochemistry of Selenium (a) Two-bonded Selenium-The general resemblance in bond formation between 2-bonded oxygen and sulphur is paralleled by an even greater similarity by that between sulphur and selenium.Both elements show marked allotropy with formation of eight-membered rings and long chains. Hexagonal selenium however has an important characteristic not yet found in a sulphur allotrope being a semiconductor whereas sulphur behaves a,s an insulator. The angle between the bonds linking a selenium atom with two other atoms varies over a rather small range. The Se-Se-Se bond angle has a mean value of 1039" (Table 12) as compared with the corresponding average sulphur angle of 106.2". The angle in H,Se of 91" can probably be accounted for by a mechanism similar to that applied to the 92.1" angle in H,S where l B 1 Simon and Kriegsmann 2.phys. Chem. 1955 204 369. l 8 2 Post Schwartz and Fankuchen Acta Cryst. 1952 5 372. lB3 Jeffrey and Jones ibid. 1956 9 283. 184Kitnd~ and King J . Arner. Chem. SOC. 1951 73 2315. ABRAHAMS STEREOCHEMISTRY OF SUB-GROUP VIR 429 an admixture of s and d orbitals with the sulphur p orbitals has been pro- posed.'6 The bond angles measured in the remaining compounds have an average value of about 104" [excluding SeO and (CH,),Se where the experimental errors are probably rather large] and in no case does this angle exceed the tetrahedral bond angle of 109.5". (6) The Xelenium-Xelenium Bond .-Although polyselenium chains of infinite length are known such as the helices in the hexagonal selenium allotrope there has been no evidence to suggest the formation of branched chains.The mean selenium-selenium bond length of 2.34 A in the elemen- tary allotropes (Table 13) may be regarded as corresponding to a pure single bond for there is no indication of double-bond character in the Se ring molecules (cf. p. 417). The Se molecule with bond length 2.19 A can tentatively be taken as containing a standard double bond. A linear bond length-bond order relation based upon these two points niay then be assumed for the selenium-selenium bond. TABLE 12. Selenium valency angle Compound H2Se . . . SeO . . . Bases40 ,,2H2d (CH3)2Se Se(SeCN) . . Se( SCN)2 ( p-C1*C,HJ2Se2 ' Hexagonal Se . (CF,),Se . . (C,H4*S0,),Se . a-Se . . . P-Se .. ( p-CH3.b6H4) ,Se (C6H6)2Se2 . * . . . . . . . . . . . . . . . . . . I . . . . . . . . . . . . . . . . . . . . . . . Angle 91" 90" 0.5" 98" f. 2" 98" f 10" 101" f. 3" 101" f 2" 101" f 3" 101.1" & 0.6" 103.6" & 2" 104.4" rf 5" 105.1" 4 2" 105.3" & 2.3" 105-7" f 0.8" 106" rf 2" 106" f. 2" Method Spec. X-Ray E.5. X-Ray Y Y Y Y ED. X-Ray Y7 9 7 Y Y Y Ref. 185 186 186 187 79 188 100 189 190 77 154. 191 192 193 194 (c) The Selenium Dihedral Angle.-There have not been many determina- tions of the dihedral angle between the bonds Se-X and Se-Y in the grouping X-Se-Se-Y. The data presented in Table 14 suggest that this angle is close to but can lie on either side of go" which is the angle to be expected if it is determined primarily by pn electron repulsion as appears to be the case for oxygen (p.409) and sulphur (p. 417). lE5Palik J. Chem. Phys. 1955 23 980. 186 McCullough J. Amer. Chem. Soc. 1937 59 789. lS7 Goldish Hedberg Marsh and Schomaker ibid. 1955 77 2948. lag McCullough Kruse and Christofferson U.S. Office of Ordnance Research Con- tract DA-04-495-ord-305 1955. lgO Grison J . Chem. Phys. 1951 19 1109. lglBurbank Acta Cryst. 1951 4 140. 192 Marsh Pauling and McCullough ibid. 1953 6 71. 19 Blackmore and Abrahams ibid. 1955 8 323. lg4Marsh ibid. 1952 5 458. Aksnes and Foss Acta Chem. Xcand. 1954 8 1787. 430 (XXVI) . . . . (XXVII) . . . (XXVIII) . . . QUARTEIRLY REVIEWS TABLE 13. Selenium-selenium bond length 80" -l 7" 1.86 * 0.10 95" 2 4" 1-83 -& 0.04 (Se-N length) 96" 1.96 Compound (XXIX). . . 98.6" & 2.0" Se . . . . . . Se(SeCN) .. . . . (c6HS),se2 . . . . . (p-C1*C&4),Se2 . . . (CF,),Se . . . . . a-Se . . . . . . Hexagonal Se . . . . p-se . . . . . . 2.01 0.03 Bond length (A) X-Ray 9 ) 9 9 9 9 2-19 5 0.03 2-29 & 0.01 2.33 & 0.03 2.333 5 0.008 2.335 f 0.032 2.34 f 0.02 2.34 f 0.014 2.36 5 0.04 198 18 199 200 Method E.D. X-Ray Y Y E.G. X-Ray 99 Y TABLE 14. Selenium dihedral angle Ref. 195 194 189 188 77 191 192 190 Compound Dihedral angle Method (p-Cl*C6H4),Se2 . . . . (C6H5),Se . . . . . (C,H,),Se . . . . . Se(SeCN) . . . . . Hexagonal Se . . . . u-Se . . . . . . . P-Se . . . . . . . 74" 82" 83" 94" 101" 102" 102" X-Ray D.M. X-Ray 9 Y Y 9 Ref. 188 194 196 189 190 191 192 The dihedral angle for the mixed -Se-S- grouping in BaSeS40,,2H,O has been determined79 as log" and in Se(SCN) it is 79O.100 (d) Heterocyclic ,Selenium.-A review of the organic chemistry of selenium has been given by Campbell Walker and Coppinger.lg7 Measurements of TABLE 15.Selenium valency angle and Se-C bond length in selenium heterocycles (XXV I) (XXV I I> (XXVI I I) txx IX) Molecule 1 Se valency angle 1 Se-C Bond length (A) I Method 1 Ref. lQ5 Maxwell and Mosley Phys. Rev. 1940 57 21. lo6 Rogers and Campbell J. Amer. Chem. SOC. 1952 74 4742. lQ7 Campbell Walker and Coppinger Chem. Rev. 1952 50 279. 108von Eller Compt. rend. 1954 239 1043. lee Wood and Williams Nature 1942 150 321. 200Marsh and McCullough J . Arner. Chem. Soc. 1951 73 1106. ABRAHAMS STEREOCHEMISTRY OF SUB-GROUP VIB 431 molecular constants have been reported for very few heterocyclic selenium compounds and these data are summarised in Table 15.The Raman and infrared-absorption spectra of liquid selenophen C4H4Se have been studied by Gerding Milazzo and Rossmark,201 and shown to resemble that of thiophen. An examination of this spectrum appears to indicate that the selenophen molecule is not coplanar but bent with a single plane of symmetry present passing through the selenium atom and the single carbon-carbon bond in contrast to planar thiophen. The tentative interpretation given of the vibrational frequencies suggests that confirmation of the selenophen symmetry must await a more complete study. An X-ray study of selenanthren 199 suggests that this molecule also is not coplanar but is bent about the Se-Se axis as the line of fold through an angle of 127". (e) Three- bonded Selenium.-In those 3-bonded selenium compounds for which molecular constants are available the data imply that the three selenium bonds form a pyramid somewhat similar to the configuration of sulphur in the sulphoxides (Table 9).The possibility of a pyramidal bond arrangement of this nature had previously been discounted 202 since although organic sulphoxides had been optically resolved separation into enantio- morphs had not been achieved with compounds of the form R,R,SeO. A pyramidal configuration was described 186 for solid SeO in which the O-Se-0 bond angles were reported to be go" 98" 98" in each group. In F,SeO Rolfe and Woodward203 report a Raman spectrum that closely resembles that of C12Se0 F2S0 and C1,SO. On the basis of the number of Raman fundamentals observed they predict a pyramidal molecule (cf.p. 424). The structure of H,SeO is reported 204 also to consist of pyramidal SeO groups with a mean O-Se-0 bond angle of 100" -+ lo" joined together by four hydrogen bonds per SeO group. Bryden and McCullough 205 have established the crystal structure of benzeneseleninic acid and this too is found to have a pyramidal structure with an O-Se-0 bond angle of 103.5" & 0.7" and the C-Se-0 bond angles equal to 98-7" & 0.9". The crystal structure of diphenyl selenoxide is isomorphous with that of diphenyl sulphoxide 153 which is known to have a pyramidal configuration (Table 9) ; hence it may be predicted that in this molecule the selenium atom again has a pyramidal bond distribution. (f) Four-bonded Selenium.-The stereochemistry of sulphur and selenium attached to four other atoms or groups depends upon the valency of these groups.The case of 4-bonded sulphur with bivalent groups attached has already been discussed (p. 425) and this bond distribution is entirely differ- ent from that of selenium with four univalent groups attached. The best determinations of molecules of this latter type are those of di-p-tolylselenium 201Gerding Milazzo and Rossmark Rec. Trav. chirn. 1953 72 957. 202 Gaythwaite Kenyon and Phillips J . 1928 2280. 203Rolfe and Woodward Trans. Paraday Soc. 1955 51 778. 204 Wells and Bailey J. 1949 1282. 2os Bryden and McCullough Acta Cryst. 1954 7 833. 432 QUARTERLY REVIEWS dichloride and dibromide,206 which have the molecular configurations (XXX) and (XXXI). In the dichloride the dihedral angle between the (xxx) (xx x I> C-Se-C plane and the C1-Se-C1 plane is 85" & 2" in the dibromide the corresponding angle is 87" 2".Along the selenium-halogen bonds the selenium radius is 1.40 A which is identical with the octahedral covalent radius assigned ll1 to this atom. However the selenium-carbon bonds of 1.93 and 1.95 8 in the two molecules (Table 17) are close to the normal single-bond radius for selenium of 1.17 8. McCullough and Marsh 206 have suggested that while the selenium-carbon bond is of the 4s4p2 type (although the C-Se-C bond angle is intermediate between 120" corresponding to this suggestion and go' which is indicative of a higher padmixture or a com- bination of d with the s,p hybrid) the selenium-halogen bond probably involves 5s orbitals. Pauling 206 has suggested that 4d and 5s orbitals may enter equally into hybridisation with the 4p electrons in forming this axial bond while Palmer 207 has proposed the use of sp3d2 hybridisation.There is a striking resemblance between the structure of the two molecules just described if the halogen atoms are disregarded and that of di-p-tolyl selenide.lg3 Thus the C-Se-C valency angles are nearly identical (cf. Table l2) the selenium-carbon bond distances are identical a t 1.93 8 and the dihedral angle between the plane of the benzene rings and of the C-Se-C plane is 35-40' for both structures. There are no other unequivocal determinations of the structure of $-bonded selenium compounds of this type. Lister and Sutton 208 suggested that SeCl has a tetrahedral bond distribution but recognised that the available electron-diffraction data were not decisively in favour of this model.However in the solid state SeCl, SeBr, and TeBr are isomor- phou~.~O~ It is likely that TeBr has a structure similar to that of TeCI, which is known to have an unsymmetrical structure on account of its large dipole moment.172 Hence the prediction of a regular tetrahedral bond distribution in SeCl may be in error. In the liquid state SeCl has been claimed to have the structure [SeCl,]+Cl- on the basis of the Raman spectrum.210 An electron-diffraction investigation 211 of SeF has indicated C, sym- metry with 4 angles 212 of 104" 5 5" and two of 120" & 10". The Raman 206McCullough and Marsh Actu Cryst. 1950 3 41. 207 Palmer Endeavour 1953 12 124. 208 Lister and Sutton Trans. Faruday SOC. 1941 37 393. *09 Brink private communication.210 Gerding and Houtgraaf Rec. Trav. chim. 1954 '73 737. 211Bowen Nature 1953 172 171. 212 Lachmmn {bid. p. 499. ABRAHAMS STEREOCHEMISTRY OF SUB-GROW VIB 433 spectrum of SeF is also reported 213 to agree with C, symmetry and it is concluded that this spectrum is most consistent with a trigonal bipyramidal arrangement the lone pair of electrons occupying an equatorial position. In the case of selenic acid,214 the structure is reported to consist of tetra- hedral SeO groups (mean O-Se-0 bond angle 110" &- 5") linked together by a system of 0-H-0 bonds with four such bonds per SeO unit. In this structure with bivalent atoms attached to selenium the [SeO,] 2- ion ibppears to haye a bond distribution identical with that of the [S0,l2- ion. (g) Six-bonded Selenium.-The only 6- bonded selenium molecule for which the molecular parameters have been determined 215 is SeF which like SF, has an octahedral bond configurat,ion.(h) The Selenium-Oxygen Bond.-There are insufficient data for a bond length-order relation to be constructed. Table 16 summarises the available ineasurements on this bond. TABLE 16. Xelenium-oxygen bond lengths I Compound H,SeO . . . . . C,H,*SeO,H . . . H,SeO . . . . . SeO . . . . . Bond length ( A ) 1.61 4 0.05 1.707 & 0-015 1.765 5 0-015 1.74 f 0.02 1.76 f 0.08 1-61 0.03 Method Ref. 214 205 205 204 186 216 The difference in the two observed Se-0 bond lengths in benzeneseleninic acid is significant (4 times the estimated standard deviation in these bonds) and is caused by one of the oxygen atoms forming a hydrogen bond. (i) The Xelenium-Carbon Bond .-A considerable variation in the length of the selenium-carbon bond has been reported (Table 17).As in the case TABLE 17. Selenium-carbon bond lengths (see also Table 15) Molecule OCSe . . . . . . Se(SeCN) . . . . . C6H5*Se0,H . . . . (p-CH,.C,H4),Se . . . (C6H,),Se2 . . . . . (p-C1'C6H4),Se . . . (p-CH,*C6H,),SeC1 . . (CF,),Se . . . . . (p-CH,.C,H4),SeBr . . (CF,),Se . . . . . (CH,),Se . . . . . Bond length (A) 1.7090 & 0.0001 1.83 f 0.10 1.903 -& 0.021 1.93 0.03 1.93 0.05 1.93 & 0.06 1.93 0.03 1.934 f 0.018 1-95 & 0.03 1.958 & 0.022 1-977 & 0.012 Method M.W. X-Ray 9 9 9 9 E.D. X-Ray E.D. 7 9 Ref. 63 189 20.5 193 194 188 206 77 206 77 187 _______ 213 Rolfe Woodward and Long Trans. Faraday SOC. 1953 49 1388. 214 Bailey and Wells J.1951 968. 216 Palmer and Elliott J . Amer. Chent. SOC. 1938 60 1852. Maxwell J. Opt. SOC. Amer. 1940 30 374. 434 QUARTERLY REVIEWS of the selenium-oxygen bond it appears that a corresponding variation in bond character is also present but the difficulty of assigning the correct multiplicity to these bonds does not permit an unambiguous bond order- length curve to be proposed although linear interpolation between the value 1-97 A for a single-bond and 1.71 A for a double-bond length allows a reason- able interpretation for the data in Table 17. The Stereochemistry of Tellurium (a) Two-bonded Tellurium.-In the elementary state tellurium does not exhibit the structural complexity observed in sulphur and in selenium. A single stable form is known consisting of infinite helices of tellurium atoms.in a hexagonal crystal structure. Relatively few measurements of the tellurium bond angle involving two bonds only to the tellurium atom have been reported and the available results are presented in Table 18. the exception of TeO,* the tellurium valency angle involving two TABLE 18. Tellurium bond angle ~ Compound (p-C1*C6H4),Te . . . . TeBr . . . . . . (CH,*S02*S),Te . . . . (p-CH,*C6H4),Te . . . . (NH4),TeS,06 . . . . HexagonalTe . . . . TeO . . . . . . . Bond angle Method 94" f- 1" 98" f 3" 100" f 3" 101" & 2.7" 103" & 3" 103.7" &- 2" 132" & 5" With bonds Ref. 188 217 102 218 99 190 219 is rather constant with an average value of 100". A close similarity t o selenium is indicated by the formation of continuous series of solid solutions of selenium in telluri~m.1~0 (b) The TeZlurium-TeZZurium Bond .-The meagre published (see Table 19) data concerning the length of this bond hardly justify any conclusions.TABLE 19. Tellurium-tellurium bond length ~ Compound FeTe . . . . . . AuAgTe,* . . . . . Hexagonal Te . . . . (p-C1.C6H4),Te2 . . . Te . . . . . . . Bond length (A) 2-90 & 0.03 2.87 & 0.10 2.82 & 0.02 2.702 & 0.005 2.59 & 0-02 * Sylvanite. Method ~ Ref. X-Ray 9 9 9 9 E.D. 220 22 1 190 188 195 217 Rogers and Spurr J . Amer. Chem. SOC. 1947 69 2102. Blackmore and Abrahams Acta Cryst. 1955 8 317. ,19 Stehlik and Balak Coll. Czech. Chem. Comm. 1949 14 595. 220 Grmvold Haraldsen and Vihovde Acta Chem. Scund. 1954 8 1927. *,lTunell and Pauling Acta Cryst. 1952 5 375. * Te has a distorted octahedral bond arrangement.ABRAHAMS STEREOCHEMISTBY OF SUB-GROUP V ~ B 435 (c) Dihed4 Tellurium Angle.-The values of those dihedral angles involving tellurium which have been measured are summarised in Table 20. This angle like the corresponding angles for oxygen sulphur and selenium is blose-to 90". TABLE 20. I C,ompound (C,H,*SO,.S),Te . . . (CH,*SO,*S),Te . . . (CH,.C,H,-SO,.S),Te . . (NH,),TeS,O . . . . (p-C1*C6H,),Te . . . Hexagonal Te . . . . Dihedral tellurium) angle Atoins defining angle Value SSTe /STeS SSTe/STeS SSTo /ST& SSTe/STeS CTeTe/TeTeC TeTeTe/TeTeTe 79" 81" 86" 86"' 95" 99" 101" Ref. 98 102 97 99 188 190 (d) Higher Valencies of Tellurium.-The stereochemistry of three-bonded tellurium does not yet appear to have been investigated. Jensen 172 has measured the dipole moment of di-p-tolyl telluroxide as 3.93 D and suggests that this indicates a semipolar Te-0 bond.An electron-diffraction study 223 on TeCl was reported to be compatible with a distorted trigonal bipyraniid with the Cl-Te-C1 angle equal to f3" Lachinan 223 has pointed out that if five of the six angles in this molecule are 93" & 3" then the sixth angle will be 171" 9". A similar hond distribution has been described for TeI in a preliminary X-ray study.224 A study of the Raman spectrum 210 of solid TeC1 is reported to suggest strongly the presence of [TeCl,]+ ions (cf. SeCI, p. 432) as also does a corre- sponding study 225 on TeCl,,AlCl,. The crystal structure of g-dimethyltellurium dichloride was reported 189 to be similar to that of the diarylselenium dihalides (XXX) and (XXXI). I n %his molecule the C1-Te-C1 bond angle is about 172" the C-Te-C bond angle is near 110" and the C-Te-C plane is normal to the C1-Te-C1 axis.McCullough 226 has suggested that in compounds of the type R,(S,Se,Te)Hal, the structure is that of a trigonal bipyramid with the halogen atoms a t the apices the three equatorial positions being occupied by the R groups and t he unshared electron pair. Tellurium hexafluoride like the sulphur and the selenium analogue has an octahedral bond distribution. Brockway 175 reports the Te-F bond tktance to be 1.82 & 0.03 8. Although sP6 and SeF6 are very inert and stable to hydrolysis TeF is easily hydrolysed. ilhat this is due to the powibility of the tellurium atoms' accepting a pair of oxygen electrons in a vacant 4f orbital. An octahedral bond distribution has also been reported 228 for tellurium in the anion of Cs,TeBr,.2 2 2 Stevenson and Schomnker J . Anaer. Chenz. Xoc. 1940 62 1267. 223Lachmann J. Chem. Phys. 1954 22 1459. 2 2 4 Blackmore Abrahams and Kalnajs Acta Cryst. 1956 9 295. 2 2 5 Qerding and Houtgraaf Rec. Trav. china. 1954 73 759. 2 2 6 McCullough Acta C'ryst. 1953 6 746. 227 Kimball J. Chem. Phys. 1940 8 188. 228Bagnall D'Eye and Freeman J. 1955 3959. 3". It is suggested 2 2 7 7 E E 436 QUARTERLY REVIEWS (e) The TeZZurium-Carbon Bond.-The tellurium-carbon bond length appears to have been reported only in the three following cases. In (p-C1*C,H4)2Te it is 2.16 & 0.14 in (p-CH,*C,H,),Te 2-05 0.05 and in TeCS the length is measured 144 as 1-904 & 0.010 A in a microwave determination. The Stereochemistry of Polonium Polonium is a soft low-melting grey metal like lead and exhibits no covalent bonding in the elementary state.Beamer and Maxwell 229 first measured the lattice constants of a- and /3-polonium and pointed out that the simple cubic structure has the symmetrical properties associated with true metallic bonding as is also indicated by its other physical properties. The crystal structures of several polonium compounds have now been determined. The formation of the oxide PoO has been confirmed,230 and the Po-0 distance measured 231 as 2.44 A corresponding to an ionic contact between PO,+ and 02-. Zinc polonide ZnPo has a face-centred cubic structure of the zinc blende type,232 and lead polonide PbPo belongs t o the NaCl structure type.232 Platinum polonide PtPo, has been shown 232 to crystallise in the hexagonal system with the Cd(OH) structure type and NiPo also is Some polonium halides have been investigated such as PoC1 and P0Clp,2339 234 and PoBr and P0Br*.23~9 228 The only polonium halide for which the crystal structure is known is the tetrabromide.In this crysta1,228 the polonium is in octahedral co-ordination with bromine the Po-Br distance being about 2.8 A. The stereochemistry of polonium in some complex ions has also been reported. Thus the crystal structure of ammonium hexachloropolonite 234 is isomorphous with that of (NH,),PtCl, with a Po-C1 distance of 2-38 A indicating this bond to be largely covalent for the radius of the Po4+ ion is 231 1.04 Similarly ammonium hexabromo- polonite is isostructural with the chloro-complex 228 the Po-Br separation of 2.60 A again corresponding primarily to a covalent bond.Czsium hexabromopolonite Cs,PoBr, has been found 228 isostructural with Cs,TeBr and the Po-Br and Te-Br distances are respectively 2.64 and 2.61 A. and of C1- is ll1 1-81 8. It is a pleasure to thank Dr. J. C. Speakmail and Dr. E. Grison for reading this manuscript and Professor A. von Hippel and Professor J. M. Robertson for their interest and encouragement. The support of the U.S. Office of Naval Research the Army Signal Corps and the Air Force under ONR Contract N5 ori-07801 and the tenure of an I.C.I. Research Fellowship are acknowledged. 229 Beamer and Maxwell J . Chern. Phys. 1919 17 1293. 230 Martin J . Phys. Chew,. 1954 58 911. 231 Bagnall and D’Eye J. 1954 4295. a32 Unpublished results by the Mound Laboratory Monsanto Chemical Co. Miamis- 233 Joy 125th Amer. Chem. SOC. Meeting Kansas City 1954. 234 Bagnall D’Eye and Freeman J . 1955 2320. 236 Joy 126th Amer. Chem. SOC. Meeting New York 1954. burg Ohio U.S.A.

ISSN:0009-2681    

DOI:10.1039/QR9561000407

出版商:RSC    

年代:1956

数据来源: RSC

摘要:

SULPHUR NITRIDE AND ITS DERIVATIVES By MARGOT GOEHRING (ANORGANISCHE ABTEILUNG CHEMISCHES INSTITUT UNIVERSITAT HEIDELBERG) THE chemistry of the sulphur nitrides began a long time ago. As early as 1835 Gregory discovered a sulphur nitride formed by reaction of sulphur monochloride (S2C12) and ammonia. Today inany sulphur-nitrogen com- pounds are known which are derived from this nitride and their chemistry has been thoroughly investigated. It shows hardly any analogy with that of the nitrogen-oxygen compounds. It is governed by the great stability of the linkage between nitrogen and sulphur by a tendency to formation of negatively charged ions and by the ease of polymerisation which can lead to large molecules containing long-chain or six-membered or eight-membered ring systems. Tetrasulphur Tetranitride S,N,.-In 1835 it was found that reaction between sulphur monochloride and ammonia gave besides other substances a compound which contained only sulphur and nitrogen ; 1s but it was not until 1850 that the exact composition of this compound was worked out and not until 1896 that a determination of the molecular weight gave the formula S,N4.39 4 5 Tetrasulphur tetranitride S,N, is solid a t room temperature.It forms orange-yellow crystals of the class C, and space group P2,/a6 It is diamagnetic and strongly endothermic. The melting point is 178". There are two convenient methods for the preparation of tetrasulphur tetranitride. First one can obtain it from sulphur chlorides and ammonia it is particularly advantageous to use a sulphur chloride in which sulphur and chlorine are in the atomic ratio 1 3 or 1 4.' This method affords tetrasulphur tetranitride in good yield ; but the reaction is complicated in its detail and it is not known what the intermediate stages are.A second method for the preparation of tetrasulphur tetranitride which however is hardly feasible for preparative purposes uses the disproportionation of sulphur occurring when elementary sulphur is dissolved in lj quid ammonia with gentle warming and the blue solution is then set aside at room tenz- perature :8 10s 4- 4NH3 + 6HZS + S,N . - ( 1 ) 'M. Gregory J. Phurnb. 1835 21 315; 22 301. E. Soubeiran Ann. Chim. Phys. 1838 67 71. J. M. Fordos and A. GOlis Compt. rend. 1850 31 702. A. Andreocci 2. unory. Chem. l899 14 246. R. Schenck Annulen 1896 29Q 171. D. Clark J. 1952 1615; cf.M. Buerger Amer. Minerul. 1936 21 675. H. Jonas and W. Knauff " Naturforschung und Medizin in Deutschland " 1939-46 Dietrichsche Verlagsges. Wiesbaden Vol. 23 p. 195. * 0. Ruff and E. Geisel Ber. 1905 38 2659. 437 438 QUARTERLY REVIEWS The constitution of tetrasulphur tetranitride has been investigated by many authors. Chemical investigation 9 10 11 has shown that the oxidation number of the sulphur in this compound is + 3. Further it has been shown that hydrolysis of tetrasulphur tetranitride always gives ammonia or ammonium ions as well as sulphur oxy-acids which is in good agreement with the fact that the electronegativity of nitrogen is greater than that of sulphur. l2 In weakly alkaline solution the hydrolysis occurs according to equation (2) but in strongly alkaline solution according to equation (3) 2S,N + 60H- + 9H,O + 2S30,2- + Sz0,2- + SNH .(2) lo S,N4 + 60H- + 3H,O + 2S032- + S,032- + 4NH3 . (3) l3 The result of the hydrolysis is typical for a substance in which sulphur has the oxidation number + 3 as this can then easily undergo dismutation half into S2+ and half into S4+. It cannot be reconciled with an assumption that in tetrasulphur tetranitride one is dealing with a nitrile of dithionous acid H,S,04; for it is impossible that reaction (2) could occur by way of S2042- as intermediate. l o On reduction of tetrasulphur tetranitride with tin(I1) chloride,13 l4 or with dithionite s2042-,15 the sulphur with oxidation number + 3 is reduced to sulphur with oxidation number + 2 and tetra- sulphur tetraimide S,(NH), results. It has been shown by chemical l 6 as well as by physical l7 methods that this reduction product of tetrasulphur H-Y-S-lf-H Y'S'? S ? N=S=N 7 s H-N-S-N-H ( 1 ) (I I) tetranitride which is a white solid has formula (I).Since this tetraimide is easily oxidised back to tetrasulphur tetranitride-for example by chlorine 15-it is reasonable to assume that tetrasulphur tetranitride has formula (1I),l8 lo which is nicely in line with the results of reduction and of hydrolysis. Nevertheless it became evident that formula (11) which contains sulphur bound in two different ways does not correctly reproduce the valency states of tetrasulphur tetranitride. It was shown that the sulphur atoms in the tetranitride molecule cannot be differentiated either chemically or physically. Attempts to effect this differentiation were undertaken l9 by labelling with 0.Ruff and E. Geisel Ber. 190.1 37 1573. lo M. Goehring Chem. Ber. 1947 80 110. l 1 V. Murthy Proc. Judien Acad. Sci. 1953 37 A 23. lo L. Pauling " The Nature of the Chemical Bond " Cornell Univ. Press New l 3 A. Meuwsen Ber. 1929 62 1959. l4 H. Wolbling 2. anorg. Chem. 1908 57 281. l5 E. Pluck Diplomarbeit Heidelberg 1956. M. H. M. Arnold J . 1938 1596. l7 E. R. Lippincott and M. C. Tobin J . Amer. Chem. SOC. 1951 73 4990. la M. H. M. Arnold J. A. C. Hugill and J. M. Mutson J. 1936 1645. l9 M. Goehring and J. Ebert 2. Naturforsch. 1955 lob 241. York 1940. GOEHRING SULPHUR NITRIDE AND ITS DERIVATIVES 439 35S. Tetrasulphur tetranitride was prepared by reaction of compounds which contained only S4+ with compounds which contained only S2+.The S4+ was labelled with 35S. Then the tetrasulphur tetranitride molecule was cleaved into compounds which contained only either S4+ or S2+. If formula (11) were strictly correct then the radioactivity should be found only in the compounds with S4+. In fact however the activity was found to be equally divided between the S4+ and the S2+ compounds which were formed on the fission of the tetrasulphur tetranitride. This experiment shows then that in tetrasulphur tetranitride the sulphur atoms cannot be differentiated as S4+ and S2+ atoms as formula (11) would demand but that an equalisation of valency must take place in the molecule. This result is supported by investigation of the Ka X-ray emission spectrum of the sulphur in tetra- sulphur tetranitride. The position of the Ka doublet of sulphur depends on the oxidation number of the sulphur in the compound under investigation and on its electron demand.21 If a single compound contains sulphur atoms with differing electron demands then Ka doublets are observed in more than one position.Such duplication can be observed for instance for S20,2- or S3062-,22 2O For tetrasulphur tet'ranitride only a single sulphur Ka doublet appears and its position corresponds to that which one would expect for a substance containing sulphur with the oxidation number + 3.20 Thus it is clear that in the tetrasulphur tetranitride molecule all the sulphur atoms present must be in the same valency state. These results require that tetrasulphur tetranitride shall have the electronic formulz (IIa-f). Resonance between the limiting formulae would account for the impossibility of differentiating the sulphur atoms Other formulae such as (IIg-i) could be brought into the resonance system although (IIk) appears less likely as a result of the Ka spectrum.ISR-51 IS-Fi=Sl 15-ij-Sl IS-&3 lS=irr-2;1 +I Id1 111 Ih l+4 11 A N I ly II y I I -_ II ls-N=SI rS=N-SI s-Q-5I I_S-N-SI IS-N=Y (1 1 a) ( I 1 b) (I I c ) (I Id) (I I el (Ilf 1 (I lgl (I1 h) (11 i ) (I1 k) The view that one is dealing with a resonance system in tetrasulphur tetranitride which ha'd been concluded from chemical dnta,l0# l* has been verified by determinations of structure. Chia-Si Lu and Donohue 23 were 2o A. Faessler and M. Goehring Naturwiss. 1962 39 169. 21 I d e m ibid. 1943 31 567. 2 2 A. Faessler 2. Physik 1931 72 734. 23 Chia-Si Lu and J.Donohue J. Arner. Chern. Soc. 1944 66 818. 440 QUARTERLY REVIEWS also led to formula (IIa-i) by the results of a study of electron diffraction in tetrasulphur tetranitride vapour. Since the molecule was shown to be far from planar resonance formulze such as (IIe) and (IIf) in which d-levels participate are essential for interpretation of the complete system. X-Ray investigations by D. Clark have underlined this result ; for Clark found in tetrasulphur tetranitride a S-N distance of 1.62 A and this corresponds to a single link with considerable double-bond character (theoretical for S-N 1.74 A and for S N 1.54 A). Since in tetrasulphur tetranitride the distance between any two non-adjacent sulphur atoms i.e. sulphur atoms which are not connected only to nitrogen was found to be only 2-58 A whilst the van der Waals radius and a single S-S bond require 3.7 and 2.08 A respec- tively it appears likely that forms such as (IIg-i) have a certain weight.This structure determination leads to the molecule shown in the Figure in which the nitrogen atoms lie in one plane while the sulphur atoms form a slightly distorted tetrahedron. The angle a t the sulphur is 102" that a t nitrogen 115". A different structure has been proposed by von Hassel and Vierroll 24 in which the sulphur atoms lie in a plane and the nitrogen atoms form a tetrahedron this would bring in formula (IIk) which has been discussed also as a consequence of infrared and Raman spectra.25 It has been demonstrated with certainty a t any rate that tetrasulphur tetranitride contains a wavy 8-ring with N-S links ; linkages are perhaps possible between sulphur atoms which are bound not only to nitrogen or between nitrogen atoms which are bound not only to sulphur.Disulphur Dinitride S,N, and Polysulphur Nitride (SN) ,.-The ring system of tetrasulphur tetranitride can be cleaved easily. If one heats tetrasulphur tetranitride vapour a t 300"/0.01 mm. and cools the issuing gases strongly one obtains an easily volatile white solid soluble in organic solvents with the formula S,N2 the molecular weight of which has been established cryoscopically. 26 This disulphur dinitride is stable only a t low temperature. Even a t room temperature it polymerises to a substance (SN), which is brass-coloured in mass dark blue in thin layers insoluble in organic solvents and obviously of high molecular weight.If atmospheric moisture is present then dimerisation to S,N accompanies the polymerisation. The dimerisation is quantitative if one treats solutions of disulphur dinitride in inert organic solvents with traces of alkali metals sodium hydroxide z 4 0. Hassel and H. Viervoll Tidskr. Kerrei Bergvesen Met. 1943 3 7. 25 E. R. Lippincott and M. C. Tobin J . C'hem. Phys. 1953 21 1559. 26 M. Goehring and D. Voigt, Natzcrwks. 1953 40 482 ; 2. anorg. Chem. 1956 285 181. GOEHRING SULPHUR NITRIDE AND ITS DERIVATIVES 441 Reaction (4) is thus reversed potassium cyanide or sodium carbonate. and in addition disulphur dinitride can undergo reaction (5).* The polysulphur nitride which is formed by reaction ( 5 ) is diamagnetic and forms fibre-like crystals with a metallic sheen and shows the remarkable property that it is a semiconductor.At 25" the specific resistance of the pressed powder extrapolated to infinite pressure is 0.01 ohm cm. The resistance of the powder decreases with rising temperature. This pheno- menon can be explained if one assumes that in polysulphur nitride the bonds are not fixed as shown in formula (IV) or (V) but that the electrons of the double bonds can move a t will through the whole macromolecule. If one considers the compounds (SN), (SN), and (SN), one sees that with increasing molecular weight an increasing number of limiting formulae is required for complete description of the electronic state of the compound. Increase in the possibilities for resonance deepens the colour. It also stabilises the compound disulphur dinitride is very sensitive to shock and decomposes into its elements explosively above 30" ; tetrasulphur tetra- nitride is sensitive to shock and decomposes explosively also when heated rapidly to above 130" ; polysulphur nitride apparently decomposes explo- sively neither on shock nor when heated.Reactions of Tetrasulphur Tetranitride with Ammonia and with Alcohols. -Cleavage of the tetrasulphur tetranitride ring system into two molecules of disulphur dinitride can be effected not merely by heat but apparently also by many chemical reactions This occurs for instance on treatment with ammonia or with alcohols. Ruff and Geisel showed that tetrasulphur tetranitride and ammonia give an ammoniate of the composition S,N,,2NH3. In a similar reaction disulphur dinitride gives an ammoniate S,N,,NH3.27 However according to the X-ray patterns and the absorption spectra it now appears that these two ammoniates are identical.The volatile disulphur dinitride can be sublimed from these ammoniates even a t room temperatures 27 W. Gesierich Diplomarbeit Heidelberg 1954. * This reaction explains an old experiment by F. P. Burt (J. 1910 97 1171) who found a blue film in the reaction vessel after heating tetrasulphur tetranitride on a silver gauze. 442 QUARTERLY REVIEWS so it must be a,ssumed that tetrasulphur fetranitride is cleaved during its reaction with ammonia In liquid ammonia this amnioniate behaves as a proton-donor like a normal inorganic acid amide. Accordingly solutions of the ammoniate (VI) in liquid ammonia conduct an electric current. With potassium amide which in consequence of the reaction NH,- + Hf = NH is a base in liquid ammonia a reaction analogous to neutralisation occurs ; in accordance with reaction (6) one obtains a mixture of yellow solid pyrophoric and extremely reactive potassium salts K(NS) and K,(N,S) .28 This interpretation of the reaction assumes fission of the ammoniate (VI) into an imide (VII) of sulphoxylic and an imide (VIII) of orthosulphurous acid.3 KNH - K (N,§) + K(N§) + 3NH . . . . . . . . . . . . ( 6 ) The reactions of disulphur dinitride and of tetrasulphur tetranitride with alcohols .are analogous to those with ammonia. Formation of a product S2N2,CH3*OH analogous to the compound (VI) can be demonstrated by means of the absorption spectra of the s0lutions.~7 Metal Thionitrosylates from Tetrasulphur Tetranitride.-n"etal thionitro- sylates can be prepared from the ammoniate of tetrasulphur tetranitride by way of the imjde (VII).Thus reaction with lead iodide gives the red Pb(NS)2,9 29 with thallium nitrate the red-brown Tl(NS) or the ochre- coloured Tl(NS),5Tl(NS), with silver nitrate the red-brown Ag(NS), and with cuprous chloride the brown CU(NS),.,~ For preparation of these thionitrosylates one can use solutions of the imide (VI) in liquid ammonia or in alcohol. I n these reactions with metals of a low oxidation number thionitrosylates of metals having a higher oxidation number can be formed- as for example in the case of the thallium compound where TllI1(NS) is formed from T11(N03) this is due to the strong oxidising properties of the compound (VIII) which is formed from the ammoniate (VI) and can oxidise the metal ions whilst being itself reduced to the compound (VII).Metal Thionitrosylates from Tetrasulphur Tetrahide.-Tetrasulphur tetraimide (I) S,(NH), the reduction product of tetrasulphur tetranitride also readily gives metal thionitrosylates. I n these reactions in contrast to those with tetrasulphur tetranitride the metal does not change its oxidation number. Mixing tetrasulphur tetraimide with cuprous chloride in solution 28 W. Berg and M. Goehring 2. anorg. Chem. 1954 275 273. 29 M. Goehring J. Weiss and G. Zirker 2. anorg Chem. 1955 278 1. GOEWRING SULPHUR NITRIDE AND ITS DERIVATIVES 443 in pyridine gives a precipitate of the brown-black [Cu(NS)], and with silver nitrate the red-brown [Ag(NS)],. 29 A mercurous compound [Hg,(NS),] 3O can also be obtained.From mercury of oxidation number + 2 one can prepare the compounds [Hg(NS),] and Hg5(NS),.31 However it is not certain whether the ring system of tetrasulphur tetraimide (I) is still retained in these compounds. It is also questionable whether the metal is bound to the nitrogen or to the sulphur of the NS group. The fact that silver iodide and [C,H,*NS] (IX) are formed on reaction between [Ag(NS)] and ethyl iodide makes it perhaps probable that the metal is bound to nitrogen in these strongly polarised substances. (I x ) ~ ~ A more definite statement can be made about the structure of another complex compound which can be obtained from tetrasulphur tetraimkle. On reaction of lithium aluminium hydride with tetrasulphur tetraimide the white solid very explosive Li[Al(NS),] is formed.33 Since gentle hydrolysis of this compound regenerates tetrasulphur tetraiinide it is reasonable to assume that the product retains the ring system of the tetraimide (I).Thionitrosyl Complex Compounds with Metals of the Eighth Transition Group.-It has been found that very stable complex compounds can be obtained from tetrasulphur tetranitride and metals of the eighth transition group. These complex compounds have the composition [M(NS),]. So far nickel palladium platinum cobalt and iron have been incorporated as central atom in compounds of this type.34 35 Three methods have been used for preparation of these complex com- pounds. First an anhydrous halide of a metal of oxidation number + 2 can be treated in alcoholic solution with tetrasulphur tetranitride the yield is particularly good if dithionite is added as reducing agent.Secondly metal carbonyls can be treated with tetrasulphur nitride in an inert solvent ; and thirdly finely divided metals can be shaken with a solution of disulphur dinitride. The first process is particularly suitable for the preparation of the nickel palladium and cobalt compounds ; the second particularly for the preparation of the nickel cobalt and iron compounds; but the third process gives only small yields. The platinum compound can be obtained by treatment of chloroplatinic acid with tetrasulphur tetranitride in hot dimethylformamide. 35 30 M. Goehring and G. Zirker ibid. 1956 285 70. 31 A. Meuwsen and M. Losel ibid. 1953 271 221. 32 F. Lengfeld and J. Stieglitz Rer. 1895 28 2742. 3 3 M. Goehring and G. Zirkar 2. Nuturforsch.1955 lob 58. 3 4 M. Goehring K.-W. Daum and J. Weiss ibid. p. 298. 35 E. Fluck M. Goehring and J. Weiss 2. unorg. Chem. 1956 287 61. 444 QUARTERLY REVIEWS These metal thionitrosyls are crystalline and very stable. The iron compound appears almost black the nickel and the cobalt compound are deep violet the palladium compound is red and the platinum compound is dark blue. All are soluble in organic compounds and insoluble in water. Nickel tetrathionitrosyl 36 is diamagnetic cobalt tetrathionitrosyl 37 has a magnetic moment of 1.90 Bohr magnetons and iron tetrathionitrosyl 38 has a moment of 2.94 magnetons. In this compound cobalt thus apparently has one unpaired d-electron ; and iron has two unpaired d-electrons. X-Ray investigations by Lindqvist and Weiss 38a have shown that there me S,N groups (111) in the compounds acting as ligands.These complex compounds have no known analogues in the chemistry of tlhe metals of the eighth transition group. The peculiarities particularly the resonance system of the (NS) group are apparently able to stabilise otherwise unusual valency states of these metals. Halogenation of Tetrasulphur Tetranitride.-Treatment of tetrasulphur tetranitride with halogens affords " thiazyl halides ". The ring system of tetrasulphur tetranitride remains intact on cautious fluorination with silver fluoride which yields the compound (XI).39 This compound S4N4F4 is colourless and solid. Besides this product there have also been obtained the volatile SN,F and SNF. Chlorination of tetrasulphur tetranitride yields the chloride (XII) which is a yellow solid ; its molecular weight corresponds to the trimeric formula (XII).40 Finally bromination gives a bronze- coloured solid bromide (XIII) whose molecular weight is not yet known.41 These compounds contain sulphur with oxidation number + 4. Accord- ingly treatment of the trimeric chloride with concentrated hydrochloric acid gives all the sulphur as sulphur dioxide.42 It has been shown that this compound undergoes reactions in which the original six- membered ring system remains intact as well as reactions in which it is destroyed. Thus for instance the chloride is converted into a pale-yellow SO adduct (XIV) if it is cautiously treated with sulphur trioxide. If this Of these halides the chloride has been most investigated. 36 M. Goehring and A. Debo 2. a7torg.Chem. 1953 273 319. 37 K.-W. Daum M. Goehring and J. Weiss ibid. 1955 278 260. 38 M. Goehring and K.-W. Daum ibid. 1055 282 83. 3aa J. Lindqvist and J. TVeiss Angew. C?2em. 1956 in the press. 0. Glemser R. Schroder and H. Haeseler ibid. 1955 279 28. 40 A. Meuwseri Rer. 1931 64 2311. 4 1 A. Clever and W. Muthmann Ber. 1896 29 340. 4 2 A. Meuwsen Colloquium of the Inorg. Chem. Section Int. Union Pure Appl. Chem. Niinster (Westphalia) Sept. 2-6 1954 p. 130. GOEHRING SULPHUR NITRIDE AND ITS DERIVATIVES 445 adduct is heated sulphur dioxide is given off and the ring system of the chloride is oxidised to that of sulphanuric chloride (XV). 43 .CI +3SQ (x I I> (x v) The ring system is also retained on oxidation with ethyl hy-pochlorite and the substance (XVI) is obtained by an obscure renction.42 H Ow 1) However other oxidants lead to complete destruction of the original ring system.For instance nitrogen dioxide affords a compound [NO]2S207,42 as it does in an analogous reaction with tetrasulphur t e t ~ a n i t r i d e ~ ~ but nitric oxide gives the dark green compound S,N,Cl whose constitution is still unknown. Tetrasulphur tetranitride can be re-formed from the thiazyl chloride the reaction (7) can be made to occur q~antitatively.~~ N C1-f 's-Cl H-YS-Y-H ly=S-p N 84 f 3 !j 7 - 6 19 19 4- 12HC1 .....(7) y H-N-S-N-H Ik-FNI I 4 I CI Ammonolysis apparently also destroys the ring system (XII). A red substance is formed which with mercuric iodide in solution with liquid ammonia gives the olive-green compound Hg(N,S) which can also be obtained from the ammonia of tetrasulphur tetranitride being then produced from the intermediate product (VIII).45 [ClSN] + 3NH + 3HC1 + [H,N*SN] -+ 3HC1 + 3HN:S:NH S(NH) + Hg2+ -+ Hg(H,S) + 2H+ It is thus shown that a thiazyl halide can serve as starting material for This the preparation of derivatives of the iinide of orthosulphurous acid.43 M. Goehring and H. Malz 2. Naturforsch. 1954 9b 567. 4 4 A. Meuw-sen and S. Kriiger 2. anorg. Chem. 1938 236 221. 45 W. Berg M. Goehring and H. Malz ibid. 1956 283 13. 446 QUARTERLY REVIEWS imide (VIII) which is derived from sulphur dioxide (XVII) by replacement of oxygen by the isosteric NH group has not yet been itself obtained pure ; (v111) (xv I I) but its potassium salt can be prepared.45 This salt K,N,S is obtained on t'reatment of [BrSN] with ammonia and potassium amide.Further an n-butyl derivative of this amide S(NC4Hg)z can be ~repared.~6 Thiotrithiazyl Compounds.-The most stable compounds which are derived from tetrasulphur tetranitride belong to a peculiar group with the composition [S4N3]X where X may be halogen HSO, or NO,. This class of compound was discovered 47 as early as 1880. The most accessible is t)hiotrithiazyl chloride S4N,Cl which is obtained as a yellow solid on reaction for example of tetrasulphur tetranitride with hydrogen chloride 48 or with sulphur mono~hloride,~~ or from [NSCl] with sulphur mon0chloride.5~ Reaction of tetrasulphur tetranitride with sulphur monochloride or other acid halides proceeds by way of a green intermediate product S3N,Cl. Thiotrithiazyl chloride is soluble only in concentrated acids.Attempts have been made to determine the structure of the thiotrithiazyl compounds from chemical and physical data,51 but no proposal has yet been fully confirmed. Tetrasulphur Dinitride.-Tetrasulphur tetranitride reacts when heated with sulphur best in carbon disulphide solution in an autoclave forming a further sulphur nitride.52 After purification by distillation in a high vacuum this product corresponds in analysis and molecular weight with the formula S4N,.53 It is a dark-red substance which has an obnoxious smell and melts a t 23". Its diamagnetism shows beyond doubt that tetra- sulphur dinitride does not dissociate a t room temperature into smaller (XVII I> (XI X I fragments *S,N* of radical character like the formally analogous dinitrogen tetr0xide.5~ Detailed discussion 54 of the results of hydrolysis of tetra- 4 6 G.Weis Diplomarbeit Heidelberg 1955. 47 E. Demargay Compt. rend. 1880 91 1066. 48 A. G. MacDiarmid N a t u r e 1949 164 1131. 49 W. Muthmann and E. Seitter Ber. 1897 30 627. 50 A. Meuwsen Ber. 1932 65 1724. 51 Cf. e.g. ref. 50 and M. Goehring and D. Schuster 2. anorg. Chem. 1953 271 281. 5 2 IV. Muthmann and E. Clever ibid. 1897,13,200 ; F. L. Usher J . 1925,127 730. 53 A. Meuwsen 2. anorg. Chen?. 1951 266 250. 54 M. Goehring H. Herb and W. Wissemeier ibid. 1952 267 238. GOEHRING SULPHUR NITRIDE AND ITS DERIVATIVES 447 sulphur dinitride and of its absorption spectrum has led to proposal of the formula (XVIII). This structure appears the more likely since sulphur- nitrogen chemistry includes a whole series of compounds containing the skeleton (XIX).Tetrasulphur dinitride of formula (XVIII) takes its place in this group by replacement of oxygen with sulphur. Sulphur-Nitrogen Compounds with a Cyclic Skeleton containing Sulphur Nitrogen and Oxygen.-From tetrasulphur tetranitride one can obtain tri- sulphur dinitrogen dioxide on reaction with thionyl chloride particularly in presence of sulphur dioxide.55 This product S3N202 forms yellow crystals which are easily soluble in organic solvents and are not wetted by water. It is readily hydrolysed to ammonium trithionate; in presence of a little water it decomposes according to the equation (8) 2S,N202 -+ 2x0 + S,N . * (8) Trisulphur dinitrogen dioxide which very probably has the formula (XX) contains as can be shown 56 by labelling with 35S two sulphur atoms derived from tetrasulphur tetranitride and one sulphur atom derived from thionyl chloride.According to reaction (8) the sulphur atom which is derived from the thionyl chloride is given up as sulphur dioxide and accordingly s that which is bound to the two oxygen atoms. Trisulphur dinitrogen dioxide is smoothly oxidised by sulphur trioxide to a white solid substance S,N205 which is soluble in organic solvents and can readily be sublimed ; it has the formula (XXI).57 It is now interesting to ask why the single quadrivalent sulphur atom in trisulphur dinitrogen dioxide is not oxidised while the two other sulphur atoms are oxidised. Comparison with (XVII) shows that this single sulphur atom is in the same valency state as in sulphur dioxide (XVII). But as is shown by experiments with 35S sulphur dioxide is not oxidised by sulphur trioxide whereas the ion is so ~xidised.~* Sulphur trioxide appears to be an almost specific oxidising agent for trebly bound sulphur of oxidation number + 4.A reaction of this type was noted earlier in the forma'tion of the compound (XV) from the chloride (XII). The compound S3N205 which is a substituted cyclic amide of pyro- sulphuric acid is also accessible directly from tetrasulphur tetranitride. 57 The nitrogen atom of tetrasulphur tetranitride acting as donor can add on sulphur trioxide so that the compounds S4N4,2S03 and S,N4,4S03 are 66 M. Goehring and J. Heinke 2. anorg. Chem. 1953 272 297. 66 Idem ibid. 1955 278 53. 57 M. Goehring H. Hohenschutz and J. Ebert ibid. 1954 276 47. 58 J. L. Huston J. Amer. Chem. SOC. 1951 72 3049.448 QUARTERLY REVIEWS f0rmed.5~ The adduct (XXTI) can then react further with sulphur trioxide. The resulting S,N,05 is able to react with pyridine as a true derivative of pyrosulphuric acid forming the adduct (XXIII). On treatment with water it undergoes hydrolysis as shown in the annexed scheme. If throughout the whole reaction one uses sulphur trioxide labelled with 36S the activity is divided as shown in the scheme and this at the same time constitutes the proof of formula (XXI). Sis (xx I 1 I) HN YH2 02!i* 'SO,H The cyclic skeleton of compounds (XX) and (XXI) appears to be favoured energetically in the chemistry of sulphur-nitrogen compounds. It is also possible to build in a six-membered ring of this kind if nitrogen with oxidation number - 1 is present.60 NN'-Dimethoxysulphamide (XXIV) which is 0 B 3 S2 CH30-( 'I$0CH3 soa CH&)-$/ )-ocy cH30-f >-OtH 0s 702 - 02s\ 7 2 0 Y b oc2Hs (XXIV) (x xv) - (XXVI) obtained from sulphuryl chloride and hydroxylamine methyl ether reacts with sulphur trioxide to give a compound (XXV) which contains this ring system.This product also is a colourless solid soluble ' in organic solvents and readily sublimable. Naturally the coin- IN@ \INi" pound (XXV) like other compounds with the same skeleton is "{-) sensitive to hydrolysis. Solvolysis yields an ester (XXVI). Treating Hg,(NS) with thionyl chloride affords a compound S,N,0,61 but it is an open question whether this belongs to the 59 M. Goehring H. Hohenschutz a,nd R. Appel 2. Naturfomch. 1954 9b 678. Goehring and H. K. A. Zahn Chem. Ber. 1956 89 179. 61 A.Meuwsen and M. Losel 2. anorg. Chem. 1953 271 221. 0 (XXvII) GOEHRING SULPHUR NITRIDE AXD ITS DERIVATIVES 449 present series of compounds. It is a mobile orange-red oil and it is possible that it has formula (XXVII) being then an oxygen analogue of tetrasulphur dinitride. Sulphur Hmides.-As mentioned earlier tetrasulphur tetranitride is readily reduced to tetrasulphur tetraimide S,(NH),. This compound crystallises as colourless needles of the class Czh ; the melting point is 152". The oxidation number of the sulphur in this compound is + 2 as is shown by the reaction with hydrogen iodide. l o Since the infrared spectrum clearly shows NH frequencies but no SH frequencies,17 formula (I) is firmly estab- lished. However this formula also follows from chemical reactions for instance that with formaldehyde which gives the substance (XXVIII) and that with pheiiyl isocyanate which gives the substance (XXIX) whence hydrochloric aeid affords phenylurea.H H H-tiJ-S-y-H HO-CHFY-§-Y-CH;OH 'C-N-S-N-C+ H-N-S-N-H HO-CHFN-S-N-CH~OH C,H,- N ,N-C,H C6H5-N H ;-c6H5 @ 5 5 5 s o\c-k-5-k-C." / \ 5 5 (1) (XXVl I I) ( X X I X ) Thus this sulphur imide is formally derived froni elementary sulphur S four of the sulphur atoms in the S8 molecule are replaced by NH groups. It has been further found that a second sulphur imide can be obtained which is also derived from S, namely a heptasulphur imide in which one sulphur atom of the 53 molecule is replaced by one NH group. Hepta- sulphur imide (XXX) is formed on reaction of sulphur nionochloride with ammonia. This compound was early discovered but it was first given the formula S6NH,,62 or S,,N3H,,18 and it was not until 1942 that it was shown 63 to be S,NH.The compound forins colourless crystals of the class D2h16 and the space group Pbnm and melts a t 113.5" ; it has been thoroughly investigated physically and chemically. 64 The infrared spectrum 30 shows NH bands. This a t once indicates formula (XXX) which is confirmed by the result of hydrolysis. Heptasulpliur irnide forms stable mercuric com- pounds Hg(NS,) with mercuric mercury 6 5 and Hg,(NS7)2 with mercurous mercury. 3O Like tetrasulphur tetraimide heptasulphur monoimide forms a stable hydroxymethyl derivative (XXXI) with formaldehyde ; further it is readily acetylated and benzoylated giving compounds (XXXII). 66 The nitrogen acts as donor atom towards sulphur trioxide ; so an N-sulphonic acid can be obtained froin heptasulphur imidc by means of sulphur tri0xide.6~ 6 2 A.K. Macbeth and H. Graham Proc. Roy. Irish Acad. 1923 36 31. 63 M. H. M. Arnold B.P. 544,577/1942; M. H. M. Arnold and W. E. Perry G 4 M. Goehring H. Herb and W. Koch 2. anorg. Chem. 1951 264 137. 6 5 A. Meuwsen and F. Schlossnagel ibid. 1953 271 226. 6 6 M. Goehring and W. Koch Z . Naturforsch. 1952 Yb 634. 67 M. Goehring and H. Hohenschutz Naturwiss. 1953 40 291. U.S.P. 2,364,414/1944 ; 2,372,046/1945 ; 2,382,845/1945. 450 QUARTERLY REVIEWS 7-S-7 S L S 4 s-s-7 7 N-SO,H S-S-4 H (xxx) ( X X X I ) (XXXI I) (xx x I 1 I) Thus as in other sulphur-nitrogen compounds one finds in these sulphur imides an eight-membered ring system as the skeleton. Here we shall only mention that such eight-membered rings occur also with higher valency stages of sulphur provided that it is linked to nitrogen. Thus in the H 02S-N-S02 I t HY YH 03-N-SO H (XXIV) chemistry of sulphur with the oxidation number + 6 a tetrameric sulphimide (XXIV) has been found in the form of derivatives-a silver salt and an N-methyl derivative-that is to be considered formally as a direct oxidation product of tetrasulphur tetraimide.6s However there is no genetical relation between these compounds (XXIV) and (I). R. Appel and M. Goehring 2. anorg. Chem. 1953 271 171.

ISSN:0009-2681    

DOI:10.1039/QR9561000437

出版商:RSC    

年代:1956

数据来源: RSC

摘要:

REACTIONS IN SOME NON-AQUEOUS IONISING SOLVENTS By V. GUTMANN DIPL.ING.DR.TECHN. (VIENNA) PH.D.( CANTAB.) (THE TECHNICAL UNIVERSITY OF VIENNA AUSTRIA) XONISING solvents are inherently polar and dissolve many ionic and covalent compounds to give conducting solutions. Water is undoubtedly the most important member in this group but its usefulness as a medium for chemical reactions is limited by its chemical properties. Thus it reacts with the chlorides of non-metals and of many transition elements and cannot be used as a medium to facilitate their reactions in solution whereas phosphorus oxychloride,l for example has been found to be a suitable solvent for this group of compounds. Non-aqueous solvents are found among widely different classes of chemical compound. 2-4 Thus the ionising properties of liquid ammonia and sulphur dioxide are well known.Like water and ammonia a number of hydrides such as hydrazine hydrogen fluoride and hydrogen cyanide behave as ionising solvents. Anhydrous acids such as sulphuric nitric or acetic acid are also ionising solvents as are alcohols and acid amides. :Further groups include a number of covalent halides e.g. bromine(1n) fluoride arsenic( 111) chloride oxides of non-metals molten iodine and certain organic compounds such as pyridine or nitrobenzene. Properties of Ionising Solvents.-Both the physical properties and the general chemical character determine the ionising power of a polar solvent. It is usually greater when the dielectric constant and the enthalpies of solvation are high and when the solvent molecules are associated as shown hy high heats of vaporisation L (in kcal.mole-1) in Table 1 which also shows inter alia the dielectric constant 8 and the specifio conductivity K (in ohm-l cm.-l). Ionising solvents dissolve certain compounds with the formation of conducting solutions and their tendency to solvate solutes or ions is fre- quently exhibited by the formation of crystalline compounds containing solvent molecules (solvates). Most ionising solvents possess a self-con- ductivity which is usually small and is to be attributed to the presence of Folvent ions (self-ionisation). It is usually possible t o recognise substances which behave as acids or bases,6 and ionic reactions may be carried out which may be either metathetical or acid-base in type. Gutmann &err. Chern. Ztg. 1955 56 126. Audrieth and Kleinberg " Non-Aqueous Solvents " John Wiley and Sons Inc.Jander " Die Chemie in wasserahnlichen Losungsmitteln " Springer Berlin Moeller " Inorganic Chemistry " John Wiley and Sons Inc. New York 1952. Franklin " The Nitrogen System of Compounds " Reinhold Publ. Gorp. New New York 1953. Gottingen Heidelberg 1949. York 1935. 6 Cady and Elsey J. Chem. Educ. 1928 5 1425. F F 451 452 QUARTBRLY REVIEWS TABLE 1. Some physical properties of certain ionising solvents Solvent H,O . . NH . HF . . H C N . . HNO . H,SO . CH,.OH . H.CO,H . AcOH BrF . . IF . . ASP . . NH . r j ~ H-CO-NH AsCi,. . IC1 . . POCl . NOCl. . SeOCI . soc1,. . so,ci . I . . . so . . NSO . . HgBr . - -____ Ions produced by self-ionisation H,O+ OH- NH,+ NH,- N,H,+ N,H,- H,F+ HF,- H,CN+ CN- H,NO,+ NO,- CH,*OH,+ CH,-O- H-CO ,H ,+ H.CO.0 - AcOH,+ AcO- HCO*NH,+ H.CO*NH- BrF,+ BrF,- IF,+ IF,- AsF,+ AsF,- H,SO,+ HSO4- ASCi,+ AsCi,- I+ IC1,- POCl,+ c1- NO+.C1- SO,Cl+ c1- HgBr+ HgBr,- I+ 1,- so++ so,- NO+ NO,- Classification of Ionising M.p. 0" - 78 2 - 85 - 13.4 -41.4 - 98 10.4 8.4 16.6 2.5 -9 9.6 - 8.5 27.2 1-2 -61.5 9.8 - 104.5 -54.1 - 18 238 113.6 - 75.7 -11.3 B.p. 100" -33.5 1133 19.5 25 86 274 64.8 100.8 118 193 127 98 57.8 130.2 97.4 105.8 - 6.5 178 75.7 69.5 320 183 - 10 21.1 Solvents.-In E (temp.) 81.1 (18") 22.0 ( -34' 51.7 (25") 83-6 (0") 123.0 (16") Unknown ca. 85 (20") 31.2 (20") 57.0 (25") 115.5 (20") Unknown 36.2 (25") Unknown 12.8 (20") Unknown 13.9 (22") 18.2 (-12' 46.2 (20") 9.1 (20") 9.2 (20") 9.8 (240") 11.1 (1lSO) 13.8 (15") 2.4 (18") 9.7 (IS0) x (temp.) 6 x 10-8 (25") 2 x (25") 4 x 10-1' ( -7SC 1 x 10-5 ( - 3 ~ 5x10-7 (0") 9x10-3 (00) 2 x 10-2 ( 18") 2 ~ 1 0 - 9 (25") 4~ 10-9 (25") s x 10-3 (25") 2~ 10-5 (250) 2 x 10-5 (25") 1x10-7 (200) 5x10-3 (350) 2 x 10-8 (20") 2~ 10-5 (200) 3~ 10-9 (20") 2 x 10-8 (20") 2 x 10-4 (2420) 1x10-7 (00) 6 x ~ O - ~ (25") 6 x (20") 3 ~ 1 0 - ~ (-20' 9 x (140") 1 x (17") L 9.72 5.64 10.7 6.5 6.74 7.25 8.42 5.38 5-81 10 10.1 8.57 12.64 9.95 8.06 6.14 7.60 6.48 6.68 20.01 10.39 5-96 9.11 considering first acid-base reactions a distinction is usually made between protonic and non-protonic solvents.The former contain hydrogen in an ionisable form and the cations produced by auto-ionisation are considered to be solvated protons. Typical protonic soZvents are water liquid ammonia 2-5 hydrazine,2 hydrogen cyanide,3 hydrogen sulphide -sulphuric acid,2 nitric acid formic and acetic acids,* alc~hols,~ and amides.1° Liquid hydrogen fluoride 2 is frequently also included.The following are typica.1 certain oxides of non-metals (liquid sulphur dioxide and liquid dinitrogen tetroxide 12) some covalent fluorides [bromine(m) fluoride l3 iodine(v) fluoride l4 arsenic(II1) fluoride,l5 and fluorosulphonic acid 16] chlorides [iodine(I) c h l o r i d ~ l ~ arsenic(II1) chloride l8 and probably also antimony(II1) chloride] oxychlorides (selenium oxychloride l9 carbonyl Non-protonic solvents do not contain hydrogen. l1 Audrieth and Ogg " The Chemistry of Hydrazine " John Wiley and Soils Inc. Maass and Jander Portschr. chem. Porsck. 1953 2 619. Williams Chem. Rev. 1931 8 303. New York 1951. l o Walden " Elektrochemie nicht-wassriger Losungen " Barth Leipzig 1924.l1 Spandau and Gutmann Angew. Chem. 1952 64 93. l2 Addison and Thompson J. 1950 S211 218. l3 Woolf and EmelBus J. 1949 2865 ; for a review see Gutmann Arzgeu?. Chem. 14EmelBus and Sharpe J. 1949 2206; Woolf J. 1950 3678. l 6 Woolf and Greenwood J. 1950 2200. l7 Gutmann 2. anorg. Chem. 1951 264 156. 1950 62 312. l6 Woolf J. 1955 433. l9 Smith Chem. Rev. 1938 23 165. Idem ibid. 1951 266 331. GUTMANN NON-AQUEOUS SOLVENTS 453 chloride 2o thionyl chloride 21 sulphuryl chloride phosphorus oxychloride 1 nitrosyl chloride 23) bromides [mercury(II) bromide 24 aluminium bromide,25 iodine bromide 26] and molten iodine.27 Some organic compounds have been mentioned among protonic solvents but others are non-protonic such as acetic anhydride 28 or diethyl ether.29 I n addition organic ionising solvents are known which appear to have no self-ionisation in the pure liquids.Among these nitrobenzene pyridine ethylenediamine acetone and acetonitrile may be mentioned. Acids and Bases.-Acids may be defined as solutes which increase the concentration of cations characteristic of the pure solvent and bases as solutes which increase that of anions characteristic of the pure solvent.6 Thus each solvent with self-ionisation may be regarded as a parent of acids and bases. According to the self-ionisations of water and liquid ammonia (2H20 + H30+ + OH- and 2NH3 + NH,+ + NH2-) the hydroxyl and amide ions are considered formally analogous with respect to the correspond- ing solvent systems being characteristic for basic solutions. Similarly hydroxonium and ammonium compounds are analogous since they behave as acids in the respective solvents by increasing the concentration of solvent cations.In an analogous way each solvent with self-ionisation may be considered as a parent for a system of compounds. An alternative description of acid-base behaviour is that due to Brsnsted 30 and L o w ~ Y . ~ ~ These definitions can be applied to protonic systems only since proton-transfer reactions are considered as responsible for both the auto-ionisation of the amphoteric solvent molecules and for most acid-base reactions in their solutions. Acids and bases are defined as proton-donors and proton-acceptors respectively. Detailed discussions of the concept of acids and bases in non-aqueous systems have been given by various authors 2 - 6 ~ 11 30-39 and so need not be discussed further.2o Germann J . Amer. Chem. Soc. 1925 47 2461. 21 Spandau and Brunneck 2. anorg. Chew,. 1953 270 201 ; 1955 278 197. 2 2 Gutmann Monatsh. 1954 85 393 404. 2 3 Burg and McKenzie J . Amer. Chem. SOC. 1952 74 3143. 24 Jander and Brodersen 2. anorg. Chem. 1050 261 261 ; 1950 262 33. 25 Jander and Zschaage ibid. 1953 272 53. 26 Gutmann Monatsh. 1951 82 156. 27 Jander and Bandlow 2. phys. Chem. 1943 A 191 321. 28 Jander Rusberg and Schmidt 2. anorg. Chern. 1948 255 238; Schmidt 29 Jander and Kraffczyk ibid. 1955 282 121 ; 1056 283 217. 30 Brransted Rec. I’rav. chim. 1923 42 718. 31 Lowry J. Soc. Chem. Ind. 1923 42 43. 32 Ebert and Konopik Osterr. Chem. Ztg. 1949 50 184. s3 Gutmann and Lindqvist 2. phys. Chem. 1953 203 250. 34 Gutmann Svensk Kern.Tidskr. 1955 68 1. 35 Lindqvist Acta Chem. Xcand. 1955 9 73. 36 Bjerrum Angew. Chem. 1951 63 527. 37 Bell Quart. Rev. 1947 1 113 ; Ann. Reports 1934 31 71. 38 Luder and Zuffanti “ The Electronic Theory of Acids and Bases ” John Wiley and Sons Inc. New York 1946. 39 Bradley J. Chem. Educ. 1950 27 208. Wittkopf and Jander ibid. 1948 256 113. 454 QUARTERLY REVIEWS Some Reactions in Liquid Ammonia.-Liquid ammonia is one of the best-known non-aqueous ionising solvents and serves to illustrate some general points. Numerous compounds are soluble to give conducting solu- tions. According to its self-ionisation (2NH3 + NH4+ + NH,-) ammonium salts behave as acids and amides as bases. Solutions of protonic acids in liquid ammonia contain NH4+ ions and thus show acidic properties.Owing to the high proton affinity of the solvent they will be stronger acids than in water e.g. ammonium acetate is a strong acid in liquid ammonia although acetic acid is rather weak in water ; acetamide is a weak base in water but shows acidic properties in liquid ammonia CH3*CO*NH + NH + NH4+ + CH,*CO*NH-. The two types of solvent ions may combine on mixing of acidic and basic solutions (neutralisation reaction). Solvent molecules are produced and the ions remaining in the solution may give a salt e.g. NH4+Cl- + K+NH,- = K+C1- + 2NH3. Solvolytic reactions occur very readily in liquid ammonia solution e.g. SO,Cl + 4NH3 + SO,(NH,) + 2NH4C1 or TiC1 + 8NH3 + Ti(NH,) + 4NH4Cl. Metathetical reactions are also possible. Thus insoluble barium chloride is formed from mixtures of soluble silver chloride and barium nitrate.Many examples of complex formation are also known.40 Weak acids or bases may react both with strong bases and with acids and thus show amphoteric properties. This is illustrated by the behaviour of silver amide in liquid ammonia AgNH + Ag+ + NH,- (basic reaction) and AgNH + 2NH3 + NH4+ + [Ag(NH,),]- (acid reaction). The re- actions may be studied by preparative conductometric or potentiometric methods. The glass electrode may be used 41 and a number of polaro- graphic investigations have also been reported.42-44 Liquid ammonia is unique in its ability to dissolve alkali and alkaline- earth metals without oxidation. The solutions show extremely high con- ductivities owing to the presence of electrons which are more or less sol- ~ a t e d .~ These solutions are therefore very powerful reducing agents. 45-47 Zintl and his co-workers 4*9 49 found that solutions of alkali metals in liquid ammonia were able to reduce many compounds to the free elements to intermetallic compounds or to homopolyatomic anionic complexes con- taining the reduced elements. For example with lead iodide the compound Na,[Pb(Pb),] has been obtained. The formation of compounds containing the elements with unfamiliar *O Schmitz-Dumont Angew. Chem. 1950 62 560. 41 Heyn and Bergin J. Amer. Chem. SOC. 1953 75 5120. 4 2 Laitinen and Nyman ibid. 1948 70 2241 3002. 4 3 Laitinen and Shoemaker ibid. 1950 72 663 4975. 44 McElroy and Laitinen J . Phys. Chem. 1953 57 564. 4 5 Kraus J. Amer. Chem. SOC. 1908 30 653 1197 1323 ; 1921 43 749 ; Chem. 4 6 Watt Chem.Rev. 1950 46 289. 47 Fernelius and Watt ibid. 1937 20 195. 48 Zintl Goubeau and Dullenkopf 2. phys. Chem. 1931 A 154 1. 49 Zintl and Harder ibid. p. 47. Rev. 1931 8 251. GUTMANN NON-AQUEOUS SOLVENTS 455 oxidation states is one of the lesser known features of chemistry in liquid ammonia solution. 50 For example tetracyanonickelate(I1) is reduced by potassium in liquid ammonia a t - 33" to the red cyanonickelate(I),Sl which is slowly reduced further a t 0" to the yellow cyanonickelate(O) K,[Ni( CN),].52 No cyanonickelate(1) is observed when cyanonickelate( II) is reduced by excess of potassium.52 Similarly the compounds K,[CO(CN),]~~ and K,[Pd(CN),]54 are produced which contain the transition metal in the zero oxidation state. On the other hand the reduction of the cyano-complex of chromium(m) gives a product containing chromium(I) and that of the corresponding compound of manganese yields %I mixture of complex cyanides containing Mn(1) and Mn(0).55 The reduction of bromopentamminoiridium(1n) bromide leads to the interesting compound Ir(NH3),,56 which is insoluble in liquid ammonia.There is also evidence for a less stable ammine of platinum(0).57 Carbonyl compounds of the transition metals with an oxidation number of zero 58 may be further reduced to carbonyl metallates. These are salt-like compounds containing the metals with negative oxidation numbers in the complex anions e.g. Na[Col-(CO),] or Na2[Fe2-(C0)4].59 Other types of reduction reactions in liquid ammonia have been reviewed elsewhere.462 479 60 Reactions in Glacial Acetic Acid.-Although acetic acid is a poorer ionising solvent than formic acid it has found many applications in analytical chemistry mostly for acid-base titrations.These are carried out with great ease accuracy and elegance. The pH scale of acetic acid covers 12 units 6 1 against only 6 for formic acid which is furthermore less easily accessible in the pure state.62 Compounds which are very weak bases in water and thus cannot be titrated in this medium become strong bases in glacial acetic acid.8 63 64 Although the strength of perchloric acid in acetic acid is smaller than in water accurate titrations with organic bases are readily carried Absolute perchloric acid solutions in glacial acetic acid are easily prepared by mixing aqueous perchloric acid with the calculated quantity of acetic anhydride in glacial acetic acid.The solutions can be standardised against sodium carbonate or sodium phthalate. The end-point may be found potentiometrically or preferably by the use of 50 Colton J. Chem. Educ. 1954 31 527. 51 Eastes and Burgess J . Amer. Chem. SOL 1942 64 1187. 5 2 Watt Hall Choppin and Gentile ibid. 1954 76 373. 53 Hieber and Bartenstein Natrmwiss. 1952 13 300. 54 Burbage and Fernelius J . Amer. Chem. SOC. 1943 65 1484. 55 Davidson and Kleinberg J. Phys. Chem. 1953 57 571. 56 Watt and Mayfield J . Arner. Chem. Soc. 1953 75 6178. 57 Watt Walling and Mayfield ibid. 6175. 58 Hieber Angew. Chem. 1942 55 1 ; Emel6us and Andersson '' Modern Aspects 59 Behrens and Weber 2. anorg. Chem. 1955 281 190. 6o Birch Quart. Rev. 1950 4 69. 61 Jander and Klaus J . Inorg. Nucl. Chem. 1955 1 334.6 2 Lange 2. phys. Chem. 1940 A 187 27. 63 Hall Chem. Rev. 1931 8 191. G 5 Auerbach Drug Standards 1951 19 127 of Inorganic Chemistry " Routledge London 1952. 6 4 Hammett ibid. 1933 13 61, 456 QUARTERLY REVIEWS colour indicators such as crystal-violet which gives a sharp colour change from blue to green. The sharpness of the colour change is suppressed by the presence of water which therefore should be excluded. Among the compounds which can be titrated with an accuracy of & 0.2% are aniline pyridine and other nitrogenous bases.66 Tertiary aliphatic amines can be estimated in the presence of primary and secondary amines,67 since the latter can be converted into neutral amides by addition of acetic anhydride.G8 It seems of particular interest that amino-acids can also be determined with great ease in glacial acetic acid.69-71 Amino- sulphamides have been estimated by potentiometric methods.72 7 3 Alkali salts of weak acids such as picrates citrates or tartrates yield the respective acetates in acetic acid and thus may be easily titrated with perchloric acid in this medi~m.6~ Furthermore derivatives of pyrrole and chlorophyll 74 may be determined as well as polypeptide~.~5 Apart from acid-base titrations various addition substitution and redox reactions have been found of analytical interest.Iodine numbers of fats and essential oils may be determined 76 and bromine may be used t o titrate organic compounds which can form bromo-derivatives.76 For the titration of phenol with bromine the addition of sodium acetate has been recommended. Redox reagents are chromium( VI) oxide sodium perman- ganate bromine titanium(rr1) chloride and chromium(I1) salts.77 78 The titrations are usually carried out in perchloric acid solutions and in an inert atmosphere but traces of water are tolerable.Reactions in Bromine (111) Fluoride.-Bromine(1r) fluoride dissolves a number. of fluorides and is a powerful fluorinating agent. Many elements oxides halides or salts of oxy-acids are converted into fluorides in which the highest valency states are often found.79 It is an ionising solvent for fluorides and has a self-ionisation in the pure liquid state,13 8o which may be explained in terms of fluoride-ion transfer processes between solvent molecules (autofluoridolysis) 3 3 9 34 F- r- - J. BrF + BrF + BrF,+ + BrF,- base 2 acid 1 acid 2 base 1 6 6 Blumrich and Bandel Angew.Chem. 1941 54 373. 67 Wagner Brown and Peters J . Amer. Chem. SOC. 1947 69 2609. 68 Haslam and Hearn Analyst 1944 69 141. 69 Harris Biochem. J . 1935 29 2820. 70Nadeau and Branchen J . Amer. Chem. Soc. 1935 57 1363. 7 1 Toennis and Callan J . Bid. Chem. 1938 125 259. 7 2 Markunas and Riddick Analyt. Chem. 1951 23 337. 7 3 TomiEek Coll. Czech. Chem. Comm. 1948 13 116. 7 4 Conant Chow and Dietz J . Amer. Chem. Xoc. 1934 56 2185. 75 Harris J . Biol. Chem. 1929 84 296. 76 TomiEek and Dolezal Acta Pharm. Int. 1950 1 31. 77 TomiEek and Heyrovsky Coll. Czech. Chem. Comm. 1950 15 997. 78 TomiEek and Valcha ibid. 1951 16 113. 79 Sharpe and Emelkus J. 1948 2135. 80 Banks Emelkus and Woolf J. 1949 2861. GUTMANN NON-AQUEOUS SOLVENTS 457 Fluorides of the alkali and alkaline-earth metals as well as silver nitrosyl and nitryl fluorides act as bases by donating fluoride ions to the solvent K F + BrF = K+ + BrF4- Fluorides of other elements may accept fluoride ions from the solvent and thus show acidic properties BrF + SbF = BrF2+ + SbF,-.Other acids are the fluorides of boron gold(m) silicon germanium(1v) tin(Iv) titanium(Iv) phosphorus( v) arsenic( v) vanadium(v) niobium(v) tantalum( v) bismuth( v) ruthenium(v) and platinum(Iv) as well as hydrogen fluoride 81 and sulphur trioxide. Thus a fluoride-ion donor may be regarded as a base a.nd a fluoride-ion acceptor as an acid. Compounds have been found which appear to contain solvent ions e.g. I<+BrF4- or BrF,+SbF,-. Complex fluorides are formed by neutralisation reactions e.g. KBrF4 + BrF,SbF + KSbF + 2BrF, which may simply be written Mf + SbF,- + KSbF,.Such compounds are easily prepared by making use of both the fluorinating and ionising properties of bromine(II1) fluoride compounds which will yield the acidic and basic fluorides by fluorination are mixed in the desired proportions and allowed to react with excess of bromine( III) fluoride. A mixture of potassium chloride and antimony( 111) oxide for example may be used to obtain potassium hexafluoroantimonate while a mixture of gold and silver in equimolar proportions will yield the insoluble silver tetrafluoroaurate AgAuF, by treatment with bromine(II1) fluoride.82 Hexafluorovanadate~,~~ hexafluororuthenates(v),84 pentafluoro- manganates(1v),~5 complex oxyfluorides of rhenium,8a and a number of nitrosonium and nitroiiium compounds 87 88 have been prepared for the first time by analogous reactions and many other complex fluorides may easily be obtained by this methodS6-91 (see Table 2).I .1. r-- 3. Product TABLE 2. Examples of cmpbex formation in bromine trijluoride Ref. Reactants for I I acid basic solution solution Au Ag v,o NOCl VCl BaCI VCl AgCl Nb CaC1 GeO NOCl AgAuF AgVF NOVF Ca(NbF 6 ) Ba (VF 6 12 (NO) ,GeF 6 82 83 83 85 90 87 s1 Rogers and Katz J . Amer. Chenz. I Reactants for acid basic solution solution B2°3 NaCl As203 NO SO Ru Mn(IO,) KCl Cr2O3 AgCl NKog Yoc. 1952 74. 1375. Product NaBF NO ,AsF NO,SO,F KRuF KMnF AgCrOF Ref. 89 88 89 84 85 85 s2 Sharpe J. 1950 2901. as Sharpe and Woolf J. 1951 798. 86 Peacock J. 1955 602. 87 Woolf J. 1950 1053. Woolf and EmeMus ibid.p. 1050. 89 Emel6us and Woolf ibid. p. 164. O0 Gutmann and Emelkus ibid. p. 1046. a3 Emel6us -and Gutmann J. 1949 2979. Hepworth Peacock and Robinson J. 1954 1197. 91 Sharpe {bid. p. 8444. 458 QUARTERLY REVIEWS Reactions in Solutions of Some Chlorides and 0xychlorides.-Self-ionisa- tion of the pure liquid seems to occur in iodine(1) chloride,17 arsenic(rr1) chloride,ls carbonyl chloride,20 nitrosyl chloride 23 92 phosphorus oxy- chloride,l 93 selenium oxychloride thionyl chloride 21 sulphuryl chloride 22 and possibly also in antimony(rr1) chloride. The equilibria may be regarded as due to chloride-ion-transfer reactions between solvent molecules e.g. c1- I J. ICl + IC1 + I+ + IC1,- AsC1 + AsCl + AsC1,+ + AsC1,- base 2 acid 1 acid 2 base 1 Accordingly an acid may be defined as a chloride-ion acceptor and a base as a chloride-ion donor.These solvents give conducting solutions with various solutes from which compounds between solute and solvent may be isolated (solvates). Examples are the compounds KCl,ICl (CH,),NC~,~ASC~,,~~ Ca(A1Cl,),,2COCl2,95 FeC1,,2NOCl,96 SbC1,,POC1,,97 ZrC1,,2POC13,9s and FeC1,,2SeOC12.gg Some of them may contain the solvent molecules bound in complex ions,34 e.g. I<+IC1,-,lOO POCl,+SbCI,- lol or possibly [ ( CH,),N] + [As,Cl,,]- @,O,Cl] + [FeC14]-.96 Chlorides may behave as acids or bases (Table 3) and neutralisation reactions lead to the formation of chloro- complexes. Thus chloroanti- monates lo2 chlorostannates l o 3 chloroaluminates lo4 and chlorozirconates 98 have been isolated from the solutions. Other complex compounds such as chlorovanadates(Iv) lo5 seem to be formed in the solutions but cannot be obtained free from acidic and basic components.It is interesting that usual colour indicators can be used to follow neutralisations between anhydrous chlorides in thionyl chloride 21 or phosphorus 0xychloride.105~ Table 3 gives examples investigated by conductimetric titrations or preparative investiga- tions ; the relative basicities for chlorides in phosphorus oxychloride deter- mined potentiometrically l o 5 ~ increase in the order SbCl, FeCl, ZrCl, SbCl, POCl, SnCl, TaCl, NbCl, AlCl, PCI, TiCl, Me,NCl pyridinium chloride Et,NCl 9 2 Lewis and Wilkins J. 1955 56. 9 3 Gutmann Monatsh. 1952 83 164. v4 Lindqvist and Anderson Acta Chenz. Scand. 1954 8 128. 95 Germann and Gagos J . Phys. Chem. 1924 28 965 ; Germann and Timpany 96 Addison and Lewis Quart.Rev. 1955 9 124. 97 Gutmann 2. anorg. Chem. 1952 269 279. 98 Gutmann and Himml unpublished results. 99 Wise J . Arner. Chern. SOC. 1923 45 1233. loo Wyckoff ibid. 1920 42 1100. Io1 Maschka Gutmann and Sponer Monatsh. 1955 86 52. Io2 Gutmann Research 1950 4 336. lo3 Idem 2. anorg. Chem. 1952 270 179. lo4 Germann and Birosel J . Phys. Chenz. 1925 29 1469. lo5 Gutmann Monatsh. 1951 82 473. 1a5aGutmann and Marringer 2. anorg. Chenz. in the press. J . Amer. Chern. SOC. 1925 47 2275. GUTMANN NON-AQUEOUS SOLVENTS 459 TABLE 3. Examples of acids and bases in chlorides and oxychlorides (A = acid B = base Am = amphotwic i = insoluble or sparingly soluble.) Solute Pyridine . . KCl. . . . R,NCl . . . AgCl . . . BaC1 . . . AZCI . . . FeC1 .. .. SnC1 . . . TiC1 . . . ZrC1 . . . VCl,. . . . TeC1 . . . SbC1 . . . PCI,. . . . I C l B E B i Am ( ? ) A A A Am A AsCI B i (?) B i A A A A i A Am B A SbC1 B B B i COCI B B B B A 1 NOCl B i B A A A Am A POCI B B B i Am A Am Am Am A Am Am A SOCI B B B 1 A A A Am Am A B Am A SO,Cl I3 B i A A A i A B A SeOCI B B B i B A A A A Am A These solvents can be used as chlorinating agents ; e.g. iodine(1) chloride converts various elements into the chlorides lo6 and vanadium(rr1) chloride may be obtained either from finely divided vanadium powder and iodine(1) chloride lo7 or from vanadium(m) oxide and thionyl chloride. lo8 Thionyl chloride reacts less readily with metals but the rate is considerably increased by addition of acids such as aluminium or ferric chlorides.109 Thionyl chloride converts vanadium pentoxide into the oxytrichloride and zir- conium(rv) oxide into the compound ZrCI,,SOCl,.Aluminium chloride has been removed from Friedel-Crafts mixtures by the use of phosphorus oxychloride.110 Derivatives of the unknown arsenic( v) chloride e.g. PCl,,AsCl and AsCl,,SbCl, have been obtained by passing chlorine through the solutions of phosphorus(v) or antimony(v) chloride in arsenic(m) chloride.111 These may contain the [AsClJf ion which is known to exist in the compound [AsCl,]+[AsF,]-. 112 This interesting compound is formed 112 when chlorine is passed through liquid arsenic(1n) fluoride at 0". It has recently been discovered *that thionyl and carbonyl chloride are excellent dehydrating agents since with water they give exclusively volatile products,l13 e.g.SOC1 + H20 + SO + 2HC1. By this method the com- mercially available " anhydrous " hydrogen fluoride can be freed from the last traces of water.2 Since water of crystallisation may also react anhydrous chlorides can be produced from the corresponding hydrates. This method is particularly useful for preparing anhydrous chlorides from hydrates which are decomposed by heat. lo6 Gutmann 2. anorg. Chem. 1951 264 169. 107 Idem Monatsh. 1950 81 1155. lo8 Hecht Jander and Schlapmann 2. anorg. Chem. 1947 254 255. lo9 Hubbard and Luder J . Amer. Chem. SOC. 1951 73 1373. 110 Dye ibid. 1948 70 2595. 112 Kolditz 2. anorg. Chew. 1955 280 313. 113 Hecht ibid. 1947 254 37. ll1 Gutmann Monatsh. 1951 82 473. 460 QUARTERLY REVIEWS Reactions in Molten Mercury(11) Bromide.-Anhydrous mercury(I1) bromide when molten is an excellent solvent for various clasBes of com- pound.24 Its self-conductivity is attributed 24 to ionisation according to the equation 2HgBr = HgBrf + HgBr,- which may be considered as due to bromide-ion transfer between solvent molecules. 33 34 Mercuric salts of perchloric snlphuric nitric and phosphoric acids show acidic properties in this solvent e.g. Hg(C104) + HgBr + HgBr+ + 2ClO,- while bromides of the electropositive metals are typical bases. Halides of mercury also behave as bases in molten mercury(I1) bromide,l14 since they produce bromide ions by reaction with solvent molecules e.g. HgO + HgBr + Hg,OBr+ + Br-. Ionic reactions between acids and bgses may lead to insoluble products ; e.g. thallium(1) sulphate is formed from niercury(rr) sulphate and thallium( I) bromide.Similarly anhydrous copper(=) sulphate can be prepared by using a copper(I1) halide. Perchlorates nitrates and phosphates of many other elements can be prepared in a similar manner. By allowing mercury(I1) oxide to react with the sulphate in mercury(n) bromide solution a red insoluble product of composition (HgO),HgSO is formed. 114 Analogous compounds are formed from the sulphide selenide and telluride of mercury. Mercury(I1) bromide acts as a good dehydrating agent for compounds containing water of crystallisation which is volatile a t temperatures corre- sponding with the liquid range of mercury@) bromide ; 24 e.g. anhydrous salts have been prepared from the hexahydrate of mercury@) perchlorate and the dihydrate of mercury(I1) nitrate. Many conductometric titrations have been carried out in this medium.l15 For potentiometric work the gold electrode has been found to give reproducible potential values.116 It has been suggested that conductometric and potentiometric methods could be used for the estimation of electrolytes in molten mercury(rr) bromide; 116 e.g.mercury(@ oxide may be readily titrated with mercury(=) perchlorate. Some Reactions in Liquid Sulphur Dioxide.-The ionising properties of liquid sulphur dioxide were first observed by Walden.l17 Jander and his school 39 118 have based the interpretation of their results on the self- ionisation of liquid sulphur dioxide according to the equation ZSO + SO++ + SO,- Although it is well established that sulphites behave as bases in its solutions,11g no typical acid is known with certainty Thionyl compounds are very poor electrolytes l 2 0 and do not exchange with solvent cations.121 It seems therefore more likely that SOClf and Cl- ions but only negligibly small concentrations of SO++ ions are produced in these solutions.I n the presence of aluminium chloride however slow exchgnge is observed,122 which might be explained by the formation of SO(AlCl,),. 114 Jander and Brodersen 2. anorg. Chein. 1950 262 33. 115 Idem, ibid. 1951 264 92. 1 1 7 Walden Ber. 1899 32 2862. 118 Jander and Ullmann 2. anorg. Chem. 1937 230 405. 119 Johnson Norris and Huston J . Amer. Chem. SOC. 1951 73 3052. 120 Jander and Wickert 2. phys. Chena. 1936 A 178 57. Iz1 Herber Norris and Huston J . Amer. Chem. SOC. 1954 76 2015. l Z 2 Masters and Norris ibid. 1955 77 1346. 116 Idem 2. analyt.Chem. 1951 133 146. GUTMANN NON-AQUEOUS SOLVENTS 461 This might behave as a weak acid in liquid sulphur dioxide but its reaction with sulphites has not yet been studied. Sulphur dioxide is a very useful solvent for the preparation of various compounds. Niobium and phosphorus oxytrichlorides are formed from the pentachlorides while tungsten(vI) chloride yields the insoluble oxytetra- chloride WCl + SO + WOCZ + SOCl,. The reaction 2KBr + SOC1 + SOBr + 2KCl in liquid sulphur dioxide offers a useful way to produce thionyl bromide. Solutions of other thionyl compounds such as the thio- cyanate or the acetate have been obtained but it is not possible to isolate these compounds in the solid state.l18 Tetra- methylammonium hexachloroantimonate is easily ~repared.1,~ By the reaction of acetyl chloride with antimony(v\ chloride lZ4 acetyl hexachloro- antimonate is formed while acetyl tetrafluoroborate is produced from acetyl fluoride and boron(rn) f l ~ 0 r i d e .l ~ ~ It reacts with potassium acetate t o give insoluble potassium tetrafluoroborate. When antimony(v) chloride is added to a solution of nitrosyl chloride in liquid sulphur dioxide a bright yellow conducting solution is formed from which nitrosonium hexachloroantimonate can be obtained by evaporation of the solvent. The solution can be used to prepare a number of other nitrosonium compounds by metathetical reactions. For example nitrosonhm hexafluorophosphate NOPF, is precipitated on addition of tetramethylammonium hexafluorophosphate,126 and in an analogous manner nitrosonium nitroprusside (NO),[Fe(NO) (CN),] can also be 0btained.l2~1 128 The solution of nitrosyl fluoride in sulphur dioxide contains the compound SO,,NOF which can be is01ated.l~~ It reacts with fluorides to form the respective nitrosoniurn fluoro-complexes such as the fluoroborate or the fluorosilicate.130 Boron(m) chloride is converted into the fluoride e.g. BC1 + 3SO,,NOF = BF + 3NOC1 + 3so, and the chlorides of phos- phorus arsenic and antimony yield nitrosonium salts of the hexafluoro-acids e.g. AsCI + 6S02,NOF = NOASP + 2N0 + 3NOC1 + 6SO,. With sul- phur trioxide nitrosonium fluorosulphonate NO(SO,F) is formed.130 Nitronium hexachloroantimonate gives a conducting solution in liquid sulphur dioxide.131 Metathetical reactions with tetramethylammonium salts may lead to the soluble tetramethylammonium hexachloroantimonate and insoluble nitronium compounds such as the perchlorate or the tetra- fluoroborate,131 e.g.NO,SbCI +- R4NC10 + NO,ClO + R,NSbCI,. It should be noted that solutions of nitryl chloride in sulphur dioxide do not give these reactions. Another interesting group of compounds the alkali fluorosulphinates 123 Jander and Hecht 2. anorg. Chem. 1943 250 308. Seel and Bauer 2. Naturforsch. 1947 2,b 397 ; Seel 2. anorg. Chem. 1943 Sulphur dioxide is an excellent medium for complex formation. 252 24. 1 2 5 Idem ibid. 1943 250 331. 127 Seel ibid. 1950 261 81. 129 Seel and Meier ibid. 1953 274 196. 130 Seel and Massat ibid. 1955 280 185. 131 Seel N6gr&di and Posse ibid. 1952 269 197. lZ6 Seel and Gossl ibid. 1950 263 253. lZ8 Seel and Walassis ibid. p. 85. 462 QUARTERLY REVIEWS E”*S02M has been obtained from solutions of alkali fluorides in sulphur dio~ide,l3~3 133 These are isomorphous with the corresponding isosteric chlorates 134 and have been shown to fluorinate complex compounds or oxychlorides ; e.g.thionyl chloride or arsenic(m) chloride is converted into the respective fluoride.134 Many other reactions in liquid sulphur dioxide such as the occurrence of redox reactions the sulphonation of aromatic compounds and the possi- bility of carrying out Friedel-Crafts reactions 135 have been reviewed Conclusion.-Although only a few examples of recent advances of the chemistry in non-aqueous ionising solvents have been reviewed here it is quite apparent that many new applications in preparative and analytical chemistry are opening up. It may well prove more difficult to obtain a detailed picture of the nature of these solutions.Hardly any precise physicochernical investigations have been carried out and indeed these are usually more difficult than in aqueous solutions and it is this aspect of the subject which merits special attention. 132 See1 and Meier 2. anorg. Chem. 1953 274 202. 133 Seel and Riehl ibid. 1955 282 293. 134 Seel Jonas Riehl and Langer Angew. Chem. 1955 67 32. 135 ROSS Percy Brandt Gebhert Mitchell and Yolles I d . Eng. Chem. 1942 34 3 9 l1 924.

ISSN:0009-2681    

DOI:10.1039/QR9561000451

出版商:RSC    

年代:1956

数据来源: RSC

摘要:

HOMOGENEOUS REACTIONS OF MOLECULAR HYDROGEN IN SOLUTION By J. HALPERN (THE UNIVERSITY OF BRITISH COLUMBIA VANCOUVER CANADA) MOLECULAR hydrogen ( H2) is a relatively unreactive subst>ance. Its homo- geneous reactions both in the gas phase and in solution are normally characterised by large activation barriers which reflect its high dissociation energy (ca. 103 kcal./mole) and the strong repulsion forces associated with its closed-shell electronic configuration. The great majority of the known catalysts which lower the activation energies of hydrogenation reactions are solids. Common examples are the transition metals including Ni Co Pt and Pd and certain metallic oxides such as CuO Cr203 ZnO and their mixtures. Despite their widespread application and the fact that their action has been widely studied the detailed mechanism by which such catalysts function is still not fully understood.Theories have been advanced which relate the activities of hydrogenation catalysts to their lattice spacing l but for the most part such theories still await critical confirmation. The limited progress which has been made in this field particularly a t the level of quantitative theory reflects in large measure the complexity of these catalyst systems and the inherent difficulties associated with the study of heterogeneous reactions. In view of this recent demonstrations that some substances notably the salts of certain metals possess the property of activating molecular hydrogen homogeneously in solution and thus functioning effectively as homogeneous hydrogenation catalysts have evoked considerable interest.Because the understanding of properties of molecules and ions in solution is generally in a more advanced state than that of solids it might be expected that the study of such systems will lead to a more detailed knowledge of the nature of the catalytic process than has been provided thus far by the study of heterogeneous catalysts. To date more than fifteen systems have been discovered and studied in which H is activated catalytically and undergoes homogeneous reaction in solution. The principal ones together with the available kinetic data about each are listed in Table 1. I n discussing these systems it is con- venient to divide them into two categories the first of which includes the reactions which proceed in organic solvents and the second reactions in aqueous solution.and to their electronic properties 23 Beeck Rev. Mod. Physics 1945 1'7 61. Trapnell Quart. Rev. 1954 8 404. ( a ) Beeck Discuss. Paraday Soc. 1950 8 118; ( b ) Dowden Research 1948 1 239; J . 1950 242. 463 Catalytic species CuOAc . . . . CuOAc . . . . AgOAc . . . . Co,(CO) * . . . Ethylene -platinous @u(OAc) . . . . Cu++o . . . . . Agf . . . . . Hg++ . . . . . Hg,++. . . . . Mn0,- . . . . Ag+ + MnO,- . . chloride Co(CN.) . . . . mi€,-. . . . . O H - . . . . . TABLE 1. Summary of homogeneous hydrogenation reactions Solvent Quinoline Pyridine Pyridine Benzene ether etc. Toluene acetone Aqueous HOAc Aqueous HClO Aqueous HClO Aqueous Aqueous HClO Aqueous .HClO Aqueous HC104 Water Water Ammonia HC104 Reaction studied Reduction of CuII or quinone ; para-H con version CUII * CUI AgI Ago Hydroformylation hydrogenation? etc .C2H4 $- HZ + C2H6 CuII .CuI or Cr,O,=+ Cr+++ Cr20,=+ Cr+++ etc. Cr20,=+ Cr+++ Hg++ + Hg,++ Hg,++ + HgO MnO,- + MnO MnO,- + MnO H absorption para-HZ conversion or D exchange para-H conversion or D exchange Temp. range 25-117 O 100 25-78 90-200 < O 80-140 80-1 40 30-70 65-100 65-100 30-70 30-60 25 80-1 10 - 50 Kinetics for - d[K,]/dt 13-16 12-14 24 26 15 18 20 14 9 23 A S 3 (e.u.) ( - 20) ( - 2 5 ) - 7 - 10 - 22 - 12 - 10 - 17 - 26 ( - 7 ) Refs. 5 6 7 9 9 10 11,12 13 15 16 17 21 23 23 25 25 27 28 34 35 Kinet,ics given are for low HC10 concentration (0-0.025~ HALPERN REACTIONS OF MOLECULAR HYDROGEN 465 Reactions in organic solvents Cuprous Salts.-The first clear-cut demonstration that H can be activated homogeneously in solution is due to C a l ~ i n .~ In 1938 he showed that quinoline solutions of cupric acetate and cupric-salicylaldehyde coniplex could be reduced homogeneously by H a t temperatures of about 100" the reaction proceeding in an autocatalytic manner as shown in Fig. 1. The 1 I I I I 0 50 100 150 200 250 Time (min.) FIG. 1 No. 2 2 millinzoles CuII salt; No. 6 1 nzillimole ; No. 9 3 milli,moles. (Reproduced by permission from Calvin Trans. E'aradny SOC~. 1938 34 1181.) Hydrogenation of cupric-salicylaldehyde conaplex in quinoline at 105" ; 396 mm. H,. breaks in the H absorption curves correspond approximately to the stoicheiometric requirements for the complete reduction of CuII to CuI. The slower subsequent absorption of H reflects the reduction of C d to metallic Cu which was observed to separate out on prolonged standing.Calvin showed that the rate of reaction of H depends only on the amount of CuII reduced and concluded that the cuprous salt which is produced in the reaction is the catalytic species responsible for activating H,. Con- firmation of this was provided by the demonstration that other substrates such as p-benzoquinone could also be hydrogenated homogeneously in the presence of dissolved cuprous acetate. The rate of H absorption depends only on the concentration of the cuprous salt and is the same whether the substrate being reduced is a cupric salt or benzoquinone. When para-H was used to reduce cupric acetate conversion was observed only after the reduction of CuII to CuI had proceeded to ~ompletion.~ This conversion Calvin Trans.Faraday SOC. 1938 34 1181. 466 QUARTERLY REVIEWS has recently been shown to occur through the same activated intermediate as the hydrogenation reaction^.^ Originally the rate of activation of H in this system was reported to be of first order in the concentration of dissolved H and between first and second order in the total concentration of the cuprous salt with an apparent activation energy of about 15 kcal./mole.6 To account for these kinetics it was suggested that a dimer of the cuprous salt (probably in the form of a quinoline-containing complex) is the effective catalyst in this system and that the reaction proceeds by the following mechanism 2CuI =+ (CUI) . * ( l a ) The observed kinetics require that k > k,’ > k > kp. However in the light of recent work which suggests that when due account is taken of the inhibitory effects of impurities such as water and acetic acid the rate of reaction appears to be exactly of second order in the total cuprous salt concentration Calvin and Wilmarth 7 have questioned the evidence supporting the dimeric configuration of the catalyst species and have proposed in place of reactions (la) and (lb) a single termolecular rate-determining step ie.2CuI + H y=+ 2Cu1,H . ‘ (2) followed by reactions similar to (lc) and (Id). The earlier conclusion that the activated complex involves two Cu atoms still applies. The catalytic activity varies with the nature of the cuprous salt.7 In quinoline solution the activities of the acetate salicylaldehyde and 4-hydroxysalicylaldehyde are similar and somewhat higher than those of the stearate and benzoate.On the other hand the cuprous nitrobenzoates and nitrosalicylaldehydes as well as the cuprous complexes of certain Schiff’s bases are inactive. These results have been interpreted to indicate that the catalytic activity increases with the basicity of the anion of the cuprous salt. However it seems difficult to extend this generalisation to all the salts and in the case of the substituted Schiff’s bases at least steric effects also appear to be involved.’ Wright and Weller have shown that ethylene- diamine and ethylenediaminetetra-acetic acid inhibit the catalytic activity of cuprous acetate in the reduction of both p-benzoquinone and cupric acetate. However it is not clear whether these reagents interfere with the ki ki Wilmarth and Barsh J. Amer.Chem. SOC. 1953 75 2237. 6 (a) Calvin ibid. 1939 61 2230; 7 Calvin and Wilmarth ibid. 1956 78 1301 ; Wilmarth and Barsh ibid. p. 1305. Wright and Weller ibid. 1954 76 3345. (b) Weller and Mills ibid. 1953 75 769. HALPERN REACTIONS OF MOLECULAR HYDROGEN 467 initial activation of H by cuprous acetate or with a subsequent step in the reduction of the substrate. The catalytic activity of a given cuprous salt is also affected by altering the solvent. Weller and Mills 6b examined the hydrogenation of cupric acetate and of quinone in the presence of cuprous acetate in various solvents chiefly organic bases. They found that pyridine quinoline and i t number of their alkylated derivatives as well as dodecylamine support the catalytic activity of cuprous acetate. On the other hand no catalytic activity was observed in a number of other solvents including dimethylaniline dipentylamine diethanolamine 8-hydroxyquinoline indole formamide and dibutyl phthalate.The results provide some indications that the catalytic activity increases with the base strength of the solvent and that it is also influenced by the solvent's chelating tendency and by steric considerations ; however the data are insufficient to permit any quantitative correlations on these points. that in contrast to those in quinoline the kinetics of activation of H by cuprous acetate in pyridine and dodecylamine ihre of first order in the catalyst concentration. Coupled with molecular weight determinations which show cuprous acetate to be substantially tmassociated in pyridine these results suggest that the effective catalyst is a monomeric cuprous species.The reason for the apparent differences in kinetics and mechanism between quinoline and these solvents is not wholly clear. They have been attributed to steric considerations involving differences in the sizes of the solvent molecules. However related studies on some of the other systems to be described in this Review make it appear likely that differences in the basicity of the solvents are also involved. Silver Salts.-The reduction of silver acetate by H in pyridine solution to metallic silver has been reported to proceed at temperatures as low as 25°.9 10 The rate is apparently homogeneously determined and is of first order in the concentration of dissolved H and of silver acetate. The mechanism which has been advanced to explain these kinetics involves lieterolytic dissociation of the H molecule in the rate-determining step i.e.AgOAc + H --+ AgH + HOAc (rate-determining) . (3a) AgH + AgOAc .- 2Ag + HOAc (fast). - ( 3 b ) !The formation of AgH as an intermediate seems plausible and energetically consistent with the observed activation energy of about 14 kcal./mole. Thus the fluoride reacts about five times as rapidly as the sulphate and acetate while solutions of silver trifluoroacetate chloride perchlorate and nitrate are inactive. These marked variations in reactivity have been ascribed to differences in basicity of the anions and the correlation thus obtained seems to provide support for the postulated rate-determining step (3a). The hydrogenation of silver acetate has also been observed in dodecylamine solution.It has recently been shown The reactivity varies with the nature of the silver salt.10 Wright Weller and Mills J . Phys. Chem. 1955 59 1060. l3 Wilmarth and Kapauan J . Amer. Chem. SOC. 1966 78 1308. UCI 468 QUARTERLY REVIEWS Dicobalt Octacarbony1.-Early work on the " 0x0 " or hydroformylatioln reaction in which an olefin reacts with H and CO in the presence of a cobalt metal catalyst i.e. CHR:CH + H + CO + CH,RCH,*CHO . - (4) led Adkins and Krsek11 to suggest that this reaction is homogeneously catalysed. Several other workers 1 2 have since confirmed this conclusion which is supported by a variety of evidence including the insensitiveness of the reaction to the presence of carbon monoxide and of sulphur compounds which normally poison metallic cobalt catalysts.The conditions usually employed in the " 0x0 '' synthesis involve tem- peratures between 90" and 200" and partial pressures of 50-200 atm. each of carbon monoxide and hydrogen in the presence of metallic cobalt. Jn addition to hydroformylation a number of other reactions involving H, such as hydrogenation of olefins and reduction of alcohols have been observed to take place under these conditions.12 These reactions are all homogeneously catalysed and proceed in a variety of organic solvents including ether benzene ethanol methylcyclohexane etc. Apparently the solvent dues not play as important a r81e or exert as specific an influence in these systems as in the case of the cuprous and silver salt-catalysed reactions described earlier. It is probable that under the conditions of these reactions metallic cobalt reacts with carbon monoxide forming a number of carbonyls including dicobalt octacarbonyl i .e.2co + 8CO + CO,(CO) ( 5 ) Evidence has been obtained to indicate that Co,(CO) [or the related cobalt hydrocarbonyl HCo( CO),] is the effective catalyst in these reactions and that it functions homogeneously.11 l2 In particular it has been dernon- strated that molecular H is split readily by the reaction CO,(CO)~ + H + 2HCo(CO) . - (6) Thus it is likely that Co,(CO) is responsible for the catalytic activation of H and that the final products arise from the reaction of HCo(CO) with the substrate. More work must be done before the detailed kinetics and mechanisms of these reactions are understood. It seems not unlikely that the catalytic activity of Co,(CO) (i.e.the ease with which it splits H homogeneously) is related to the presence of an unoccupied low-lying delocalised electronic orbital into which electrons from the H molecule can be transferred readily. Ethsleneplatinous Chloride.-The hydrogenation of ethyleneplatinous chloride in toluene or acetone solution has recently been investigated by Flynn and H~1burt.l~ It was shown that a t temperatures below - lo" in the presence of an excess of ethylene ethane is formed without the 11 Adkins and Krsek J . Amer. Chem. Soc. 1948 70 383 ; 1949 '71 3051. l2 Wender Levine and Orchin &id. 1950 72 4375 ; Wender Orchin and Storch ibid. p. 4842 ; Orchin Adv. Catalysis 1953 5 385 ; Wender Sternberg and Orchin J . Arner. Chem. Soc. 1953 '75 3041. l3 Flynn and Hulburt ibid. 1954 76 3393 3396.HALPERN REACTIONS OF MOLECULAR HYDROGEN 469 accompanying deposition of metallic platinum. has been proposed for this reaction The following mechanism (PtC12C2H4)2 + 2C,H4 + 2PtC12(C,H4) . (7a) 2PtC12(C2H,) + 2H2 + (PtC12C,H,) + 2C2H . * (Tb) Although the evidence for a homogeneous reaction of H in this system appears fairly convincing the detailed kinetics and the nature of the catalyst species remain to be established. Reactions in Aqueous Solution Cupric Salts.-In 1909 Ipatieff and Werchowsky 14 observed that aqueous solutions of cupric acetate were reduced by hydrogen under rela- tively mild conditions to cuprous oxide. This reaction has since been demonstrated l5 to occur homogeneously in solution the kinetics being of first order in the concentrations of cupric acet'ate and of hydrogen as shown 0 24 36 48 60 Erne (min.1 FIG.2 Pirst-order rate plots for the reduction of cupric acetate &a aqueous solution at various H partial pressures ; 130". Partial pressures (atm.) of hydrogen 0 6-8 0 13.6 c7 20.4 0 27.2 A 34.0. (From Dakers arid Halpern Canad. J . Chew. 1954 32 969.) in Pig. 2. Hydrogen is apparently activated by interaction with a molecule of cupric acetate. This behaviour is in marked contrast to that reported earlier for the reduction of cupric salts in quinoline solution (Fig. l) where only the cuprous species exhibits catalytic activity. In aqueous solution cupric acetate was also shown l6 to catalyse the homogeneous hydrogenation of other dissolved substrates such as CrzO,=. l4 Ipatieff and Werchowsky Ber. 1909 42 2078. l5 Halpern and Dakers J .Ckem. Phys. 1954 22 1272 ; Dakers and Halpern ('anad. J . Ghern. 1954 32 969. l6 Peters and Halpern ibid. 1955 33 356. 470 QUARTERLY REVIEWS It was subsequently found that this catalytic activity is not confined to cupric acetate but is exhibited by a variety of other cupric salts and complexes. Typical of the results obtained are the rate plots in Fig. 3 0 20 40 Time (rnin.1 FIG. 3 60 EfSect of cupric perchlorate on the rate of reaction of hydrogen with various reducible substrates ; 110' ,- 20 atm. H,. Oxidant Cu( ClO,) (nioles/l.) 0.10 0.20 0.30 Cr,O,=. . - 0 0 0 10,- . . A A A Ce4+ . - 0 0 (From Peters and Halpern J . Phys. Chem. 1955 59 793.) showing the reduction of various substrates by H in the presence of different amounts of cupric perch10rate.l~ The rate at which H reacts is seen to be essentially independent of the nature or concentration of the substrate.Providing that the hydrogen-ion concentration did not exceed about 0.1~1 the kinetics for the various cupric salts were all of the form - d[HJ / dt = Ic[H,][CuII] . * (8) The apparent catalytic activity reflected in the magnitude of k varied considerably from one cupric salt to another,l* as shown in Table 2. The perchlorate system was subjected to a detailed kinetic study.17 19 In this medium no complexing of Cu+' is believed to occur and hence the simple Cu++ ion constitutes the catalytic species. At low HC10 concentra- tions (< O.~M-H+) the kinetics are essentially expressed by equation (8) l7 Peters and Halpern J. Phys. Chem. 1955 59 703. l8 Ident Canad. J . Chem. 1956 34 554.lo Halpern Macgregor and Peters J. Php. Chem. 1956 60 in the press. HALPERN REACTIONS O F MOLECULAR HYDROGEN 47 1 TABLE 2. Effect of cornplexing agents on the catulytic activity of Cu++ in aqueous solution Medium Butyrate . . Propionate . Acetate . . Suiphate . . Probable cupric species Cu(O.COPri) Cu(O*COEt) Cu(OAc) cuso Re!ative catalytic activity 150 150 120 6.5 1 Medium Chloride. . . Perchlorate. . Glycine . . . Ethylenediamine Probable. copric species CUC1,' cu++ CuGl Cu( EDA) + + Relative :atalytic activity 2.5 1 < 0.5 0.1 and are consistent with the assumption of a simple bimolecular rate-deter- mining step involving one H molecule and one Cu++ i0n.l' However with increasing hydrogen-ion concentration a marked decrease in rate was noted as well as a tendency for the kinetic dependence on the Cu++ con- centration to shift from first toward second order.The following mechanism has been proposed l9 to account for these results ki lLl ka Cu++ +H =+ CuH+ + H + . * (9a) CuH+ +Cut.+ -+ 2Cu-+ + H+ . * ( 9 4 2Cu+ + Substrate -+ Products + 2Cu++ . * (9c) Application of the steady-state treatment to this mechanism leads to fast the following rate expression 'Chis equation becomes equivalent to (8) at low hydrogen-ion concentrations Imt correctly predicts a decrease in rate as well as a shift from first- to second-order dependence on [Cu++] as the hydrogen-ion concentration increases. At 110" the value of k-l/k was found l9 to be 0.25. These considerations demonstrate how a change in the apparent order of the reaction with respect to the catalyst concentration may occur without a real change in the mechanism of the H activation process.This is of interest in view of the fact that a number of other catalyst systems (including ouprous acetate which has already been described) have been reported in which a shift from first to second order in the catalyst concentration is observed on changing the solvent. The suggestion that CuH+ is the activated intermediate in the Cu++- catalysed reactions has been further supported 2o on the grounds that ( ( 4 ) a similar intermediate has been postulated to explain the activation of H by cuprous salts in quinoline (i.e. equation 2) ( b ) energetically the formation of CuH+ seems more plausible than that of the other intermediates such as Cu+ and Cuo which might be considered to result from the reaction of an H molecule with a Cu++ ion and (c) the promoting influence of various 2o Halpern Adw.Catalysis 1956 9 in the press. 472 QUARTERLY REVIEWS anions on the catalytic activity of Cu++ which increases in the same order as their basicity (Table 2) is readily explained by assigning to them the r61e of stabilising the H+ ion which is released in the initial step. A schematic pot'ential-energy diagram depicting the activation process according to the above mechanism and showing the formation of CuHf as a,n intermediate,2* is presented in Fig. 4 (a). The activated complex corresponds to the crossing point of the two curves and probably has a configuration resembling [Cu++**-H-***H+]. The mechanism entails essenti- ally the displacement of a hydride ion from H to the catalyst both electrons involved in the Cu+-H bond being contributed by the H molecule.I n this light the r61e of Cu++ may be considered to be essentially that of I Cu+++Hz - Cu - H distance Ag - H distance (a) (6) FIG. 4 Schematic potential-energy diagrams for the activation of H in aqueous solution by (a) Cu++ and (b) Ag+. an electron acceptor suggesting that its cat,alytic activity may be related to the presence of a low-lying (Le. 3d or hybrid) unoccupied orbital into which the H electrons can enter. The lowering of the catalytic activity of Cuff on chelation with glycine and ethylenediamine (Table 2) probably reflects the fact that in the chelate complexes these orbitals are used in forming covalent bonds with the ligands. This phenomenon parallels the well-known poisoning of heterogeneous metallic catalysts by electron- donating substances such as sulphur compounds.Silver Salts.-Sa,lts of silver have also been observed 2 l to activate H homogeneously in aqueous solution and to catalyse the hydrogenation of substrates such as Cr,O,- a t temperatures as low as 30". The kinetics are of first order in the concentration of dissolved H and of second order 21 Webster and Halpern J . Php. Chem. 1956 60 280. HALPERN REACTIONS O F MOLECTJLAR HYDROGEN 473 in Ag+ as shown in Fig. 5 . In this respect the behaviour of Ag+ in aqueous solution differs from that of cuprous or silver salts in pyridine 9 3 lo but n - 'G 75 % g 70 - I 4 Q v L Qo 0 - Y .L s5 cc & m I 0 0 0.05 0.70 0 0.005 0.010 [Ag+] (mole L-') [Ap-C]* (mo1e2 (.-'I FIG. 5 Effect of silver salts on the rate of the reaction between H and Cr20,= in aqueous solution.Temperatures 0 40" (C104-) 0 50" (C104-) A 70" (Clod-) 50" (NO,-). (From Webster and Halpern J . Phys. Chene. 1956 60 280.) resembles that of cuprous salts in q~inoline.~ 6 p The most probable mechanism parallels that which has been proposed for the latter system i e . 2Ag+ + H + 2AgH+ (rate-determining) . (lla) The normal termolecular kinetics (Table 1 ) coupled with energetic considerations lend support to this mechanism. The activation path by which AgH+ is formed is shown 2O schematically in Fig. 4 (b). The H molecule is depicted as splitting homolytically through an activated com- plex whose configuration probably resembles [Ag+***H***H***Ag+]. A slight deviation from the second-order dependence on the Agf concentration in the direction of lower order,21 suggests that there may also be a small contribution to the observed activation of H through an alternative path i.e.Ag+ + H --+ AgH + H+. . * (12) This resembles the path which had been proposed earlier (equation 3a) for the first-order activation of H by silver salts in pyridine where it is probably favoured because of the greater basicity of the solvent. There is an important parallel here with the Cuff-catalysed reactions in aqueous solution 2AgH+ + Substrate + Products + 2Ag+ (fast) . . ( l l b ) 474 QUARTERLY REVIEWS where the apparent kinetic dependence on the Cu++ concentration also shifts from first to second order as the acidity (i-e. hydrogen-ion concentra- tion) of the solution is increased. Mercuric and Mercurous Salts.-The reduction of mercuric salts by H has been observed 22 23 to proceed homogeneously in aqueous solution at temperatures as low as 60".In perchlorate medium the reaction proceeds in two stages ( A and B in Fig. S) the first corresponding to the reduction of Hg++ to Hg,++ and the second to the reduction of Hg,++ to metal. -25 0 25 50 150 250 350 459 Tme (min.1 FIG. 6 Reduction of mercuric perchlorate by H in 0.05w-HC10 ~olution at 74.8" ; 4.0 atm. H,. In ( a ) 0 = [Hg++] 0 = [Hg2++]. In ( b ) 0 = log [Hg++] 0 = log [Hg,++] A = log{[Hg++] + [k2/(21c - k2)][Hg~-+]"). (From Korinek and Halpern J Phys. Chena. 1956 60 285.) The transition between the two stages corresponding to the first appearaace of metallic Hg and the constant ratio of [Hg,++] to [Hg+f-] during the second stage are governed by the thermodynamics of the process Hg (1.) + Hg++ + Hg,++ - (13) whose equilibrium constant K is about 70 a t 60°.It has been demonstrated 23 that both Hg++ and Hg,++ contribute t'o the activation of H, the kinetics a t any point being given by - d[H21/ dt = hCH,l[Hgf+l + ~,[H,1[~~2++1 * * (14) Halpern Korinek and Peters Research 1954 7 61s. 23 Korinek and Halpern J . Phys. Chem. 1956 60 285, HALPERN REACTIONS OF MOLECULAR HYDROGEN 475 The linear plots shown in Fig. 6 (b) are based on functions obtained through integration of this expression. The kinetics (Table 1) support the assumption of two independent and additive rate-determining steps in which H molecules are activated by homogeneous reaction with Hg++ and Hg2++ respectively. It seems reasonable by analogy with the suggestions made earlier for Cu++ and Ag+ that the active intermediate which is formed in this step is HgHf.How- ever on energetic grounds the formation of Hg atoms by a two-electron transfer from H, i.e. Hg+++H -+ HgO +2H+ . (15) Hg,++ + H -+ 2Hg0 (or Hg,) + 2H+ . - (16) also appears very favourable and probably constitutes the simplest rate- determining process consistent with the observed kinetics. 20 23 It is of interest that in contrast to Cu++ the presence of various com- plexing anions markedly reduces the reactivity of Hg++ toward H, probably reflecting the much greater stability of the mercuric complexes. The rates of reaction with H of mercuric salts have been observed 24 to decrease in the following order perchlorate nitrate > acetate > chloride > bromide.Permanganate Reduction.-The reduction of MnO,- by H in acid solution i.e. MnO,- +$H +H+ + MnO + 2H,O . * (17) has recently been shown to be a homogeneous reaction.25 The kinetics are of first order each in the concentrations of MnO,- and H, suggesting that one H molecule and one MnO,- ion participate in the rate-determining step. On energetic grounds,20 the formation of MnV1 (with the simultaneous formation of a H atom) appears inconsistent with the observed activation energy of 14 kcal./mole and it is more probable that MnV is formed in the first instance according to or Under the conditions which were used MnV is known to disproportionate readily to give the observed product MnO,. It is of interest that MnV can be formed either by a two-electron transfer from H to MnO,- (equation 18) or by the transfer of an oxygen atom from MnO,- to H (equation 19).By using 180-labelled KMnO, it has been demonstrated 26 that a mechanism of the latter type applies in the perman- ganate oxidation of benzaldehyde. However it is not necessarily preferred The rather low value of the activation entropy of the reaction - 17 e.u. Table 1) finds readier explanation on the basis of the alternative mechanism since the activated complex would probably be more ionic and hence more highly hydrated than the reactants in this case. In the presence of Ag+ a kinetic contribution to the reduction of MnO,- e. 2 4 Korinek and Halpern Canad. J . Chem. 1956 34 1372. 26 Webster and Halpern Trans. Faraday SOC. in the press. a6JViberg and Stewart J . Amer. Chem. SOC. 1956 '7'7 1786.476 QUARTERLY REVIEWS of the form k[H,][Ag+][MnO,-] was observed. This has been attributed 25 t o the activation of H through an alternative mechanism involving the formation of MnVI and AgH+ as intermediates i.e. Mn0,- + Ag+ + H -+ MnO,' + AgH+ + H+ . (20) followed by fast reactions to give the observed product MnO,. This path is favoured by a very low activation energy (9 kcal./mole Table 1). Niscellaneous Hydrogenation Reactions.-It has been observed 2' that aqueous solutions of cobaltous cyanide readily absorb H a t room tenipera- ture the total amount taken up corresponding approximately to one hydrogen atom per cobalt atom. The reduction of certain dissolved sub- strates such as cinnamic acid can also be effected in this system apparently through a homogeneous reaction with H2.The rate of H uptake has been reported to be of first order in H and of second order in CoII and to attain a maximum value when the mole ratio of CN- to CoII in the solution is approximately 4-5. Winfield 28 has concluded that the catalytic species is a dicobalt complex with the possible configuration [( CN),Co-C=N + CO(CN),]-~. The ability t o activate H homogeneously has also been attributed to solutions of certain complex salts of rhodium and palladium.29 However the nature of the catalysts in these systems remains to be established. Kaneko and Wadsworth 30 recently reported that the hydrogenation of aqueous solutions of Co(NH3),S0, to yield metallic Co can be effected a t temperatures of 150-245" in the presence of colloidal graphite or of a dis- solved quinol. In the latter case the reaction is presumably homogeneous and although some metallic cobalt is present in the system its catalytic activity does not appear to be significant.The proposed mechanism which still awaits confirmation implies that H reacts homogeneously with the quinone to form a quinol intermediate which in turn reduces the cobaltous salt. In contrast to the carbonyl and cyanide complexes discussed earlier t,he ammine complexes of cobalt do not appear to activate H homogeneously. At temperatures of up to 150" no tendency to activate H homogeneously in aqueous acetate or perchlorate solution could be detected 31 for the following metal ions Cat+ Mg++ Zn++ Mn++ Co++ Ni++ Cd++ Pb++ Al+++ Fe+++ U02+f and VO,-. Base-catalysed Exchange Reactions.-In 1936 Wirt z and Bonhoeffer 3 2 observed that potassium hydroxide catalysed the homogeneous exchange of H2 with the deuterium in heavy water a t temperatures of about 100".The validity of this result was subsequently questioned by Abe 33 who attributed the catalysis to traces of impurities such as colloidal iron oxides. However The analogy with dicobalt octacarbonyl is striking. 27 Iguchi J . Chern. SOC. Japun 1942 63 1752 634. 28 Winfield Rev. Pzcre Appl. Chem. (Australia) 1955 5 217. 29 Iguchi J . Chem. SOC. Japun 1939 60 1287 ; Sibata and Matumoto ibid. 1939 3'JKaneko and Wadsworth J . Phys. Chem. 1956 66 457. 31 Peters and Halpern unpublished results. 32 Wirtz and Bonhoeffer 2. phys. Chern. 1936 177 A 1. 33 Abe Sci. Papers I n s t . Phys. Chem. Res. (Tokyo) 1941 38 287. 60 1173. HALPERN REACTIONS OF MOLECULAR HYDROGEN 477 a recent careful re-investigation of this system by Wilmarth and his co-workers 34 appears to have confirmed the essential validity of the earlier work; catalysis of the homogeneous conversion of para-H by aqueous alkali solutions was also demonstrated.The kinetics of both the exchange and the conversion reaction are of first order in OH- and in dissolved H or that the reaction may occur either through a hydride ion i.e. D + OH- + D- +DOH HOH+D- + OH-+HD D,. It has been suggested 34 the intermediate formation of (rate-determining) . . (21a) (fast) . . (21b) or through a concerted attack on D of an OH- ion and an H,O molecule which may be depicted 32 as I 7 i - - - - - - - - ' ---- HO-/H+ + D-; ;D+ + -OH 3 HO- + HD +DOH . (22) The formation of D- (or H-) as an intermediate appears to be energetically consistent with the observed activation energy 34 of about 23 kcal./mole.The two mechanisms depicted above become equivalent if step (21b) is very fast and the lifetime of the intermediate D- correspondingly short. That this is in fact the case is indicated by the failure to detect any reduction of a dissolved substrate such as Cr0,2- under these conditions. 31 An analogous NH,-catalysed para-H conversion and D,-NH exchange in liquid ammonia a t - 50" has also been observed.35 The kinetics parallel those of the OH-cabalysed reaction in water although the activation energy is presumably much lower (estimated a t 10 kcal./mole). An analogous mechanism has been proposed for this system. Conclusions Among the important conclusions arising from these studies on homo- geneous hydrogenation reactions in solution is that contrary to views once commonly held hydrogenation catalytic activity does not necessarily depend on either an electronic band structure or a geometrical arrangement of atoms characteristic of the solid state.Thus conclusive evidence has been obtained in recent years that a number of simple substances for the most part metal ions or metal- containing compounds can activate H homogeneously in solution. At present it appears difficult to advance a single " mechanism " for the activation process which will apply without serious qualification to all the reaction systems that have been studied. Thus there are indications that H can be activated by a variety of processes including homolytic or heterolytic fission of the H-H bond or electron transfer from the H molecule to the catalyst and that superficially similar catalysts or even a given catalyst under different conditions or in different solvents can activate H by different mechanisms.34 Claeys Dayton and Wilmarth J . Chem. Phys. 1950 18 759 ; Wilmarth I)ayton and Plournoy J . Amer. Chem. SOC. 1953 '75 4549. 35 Wilmarth end Dayton {bid. p. 4553. 47 8 QUARTERLY REVIEWS There seems to be some basis for dividing the systems studied into two categories. In the first group which includes CuI salts in quinoline AgI and HgI salts and cobaltous cyanide in aqueous solution and Co,(CO) in organic solvents the activation process requires the presence of two metal atoms. I n these systems the mechanism may most reasonably be depicted as involving the homolytic dissociation of H, i.e.2M (or M,) + H + 2MH . * (23) The second category includes CuII HgII and Milo,- salts in aqueous solution as well as CuI and AgI salts in pyridine. In these systems only one metal atom or ion seems to participat'e in the activation process which probably involves either the tlransfer of the hydrogen electrons to the catalyst i.e. M +H + M'+2H+ . - (24) or the heterolytic dissociation of H, i.e. M +H + MH- + H + . ( 2 5 ) In these cases a suitable base usually the anion of the metal salt or a solvent molecule is required to stabilise the hydrogen ions which are released in the activation process. This probably explains the important r61e of the solvent and of complexing anions in these systems. I n connection with the above classification of catalysts particular interest attaches to the studies on Cu++ in aqueous solution which illustrate how an apparent transition from one type of mechanism to the other can occur for a given catalyst with a change in the properties of the solvent.It is of interest that a heterolytic dissociation mechanism similar to that represented by equation (25) has been proposed to explain the heterogeneous catalytic activation of H by oxide semiconductors such as zinc oxide 36 and the biological activation of H by the enzyme hydr~genase.~' The latter may be considered as a special and very interesting case of a homo- geneous hydrogenation catalyst. Its chemical structure and the detailed mechanism by which it functions still remain to be elucidated. If any property emerges which appears to be common to all the species which have been observed to activate H homogeneously and which may well prove to be a prerequisite for both homogeneous and heterogeneous hydrogenation catalysts it is that the catalyst must have a high electron affinity.This generally implies the presence of low-lying unoccupied elec- tronic orbitals or bands. The process of activation of H appears to involve in each case some measure of displacement (not necessarily complete transfer) of electrons from the H molecule to the catalyst.38 The explana- tion for this probably lies in the fact that the formation of an activated complex involving an electronically saturated molecule such as H, generally involves the promotion of electrons into anti-bonding orbitals. Hence a lowering of the activation energy is to be expected if the activated complex is coupled with a suitable electron acceptor.An interpretation of catalytic 36 Parravano and Boudart Adv. Catalysis 1955 7 47. 37 Krasna and Rittenberg J . Amer. Chew,. SOC. 1954 76 3015. 38 Halpern and Peters J. Chem. Phys. 1955 23 605. HALPERN REACTIONS OF MOLECULAR HYDROGEN 479 activity along these lines has been proposed by Eyring and Smith.39 The specific suggestion that heterogeneous catalysis may involve the transfer of electrons from the adsorbed reactant to electronic bands or levels in the catalyst has also been advanced previously 3b and supported by a variety of experimental evidence. 40 However the detailed mechanism by which this may be achieved seems to vary widely from system to system. An apparent exception to the classification of homogeneous catalysts and the generalised interpretation of catalytic activity outlined above is the base-catalysed exchange reaction of H in water or ammonia which does not involve any metal ion or metal-containing compound.However it seems likely that the activation of H in this system occurs by a special mechanism involving only proton shifts without the formation of any active reducing intermediate. According to this view OH- and NH,- are not hydrogenation catalysts but only bring about an exchange reaction of H with the solvent in which the H (or HD) molecule is regenerated. 38 Eyring and Smith J. Phys. Chem. 1952 56 972. 40 Couper and Eley Discuss. Furday SOC. 1950 8 172 ; Dowden and Reynolds ibid. p. 184 ; Schwab ibid. p. 166 ; Kemball Proc. Roy. SOC. 1962 214 A 413.

ISSN:0009-2681    

DOI:10.1039/QR9561000463

出版商:RSC    

年代:1956

数据来源: RSC

摘要:

THE LOCATION OF HYDROGEN ATOMS IN CRYSTALS By R. E. RICHARDS M.A. D.PHIL. (FELLOW OF LINCOLN COLLEGE OXFORD AND UNIVERSITY DEMONSTRATOR IN CHEMISTRY) THE structures of molecular crystals have been studied for many years by the powerful method of X-ray diffraction. The contribution of hydrogen atoms to the scattering pattern is however usually so small that the location of these atoms has been inferred from the positions of the heavy atoms and determination of distance between hydrogen atoms has been out of the question. Recently the technique of measurement and the method of interpretation of X-ray patterns has permitted the location of hydrogen atoms with moderate accuracy but the task is formidable. Since the last war new methods have been used to find the positions of hydrogen atoms.The diffraction of electrons by crystals can now be used to determine crystal structures and the method has already proved itself capable of locating hydrogen atoms with an accuracy of 0.03-0-lk Neutron beams of high intensity can now be obtained from an atomic " pile " and the diffraction patterns obtained when such a beam is scattered by a crystal can be used to determine the positions of hydrogen and deuterium atoms. Radiofrequency spectroscopy can now be used to study the positions of protons in crystals ; the method is restricted to relatively simple struc- tures but where it can be applied it is quick the interpretation is easy and the accuracy is as good as that of most diffraction methods. Jn this Review the principles of these methods are described and their applications are illustrated with typical examples.Radiofrequency spectro- scopy in so far as it can be applied to this problem is discussed at greater length than the other methods not because it is more important but because the principles of the method are less familiar to most chemists than those of diffraction methods. Diffraction Methods Crystals constitute a three-dimensional diffraction grating and will scatter a beam of waves of suitable wavelength. The resulting diffraction pattern depends upon the wavelength of the beam of radiation and on the spatial distribution of the scattering elements in the crystal. The process by which the detailed crystal structure is obtained from the diffraction pattern is beyond the scope of this Review ; a very clear account for the case of X-ray diffraction has been given by Jeffrey and Crui kshank.Jeffrey and Cruikshank Quart. Rev. 1953 7 335 480 RICHARDS HYDROGEN ATOMS I N CRYSTALS 481 It is not often possible to obtain the crystal structure directly from the diffraction patterns but trial structures must be postulated and the intensi- ties and positions of the maxima in the diffraction pattern calculated from them compared with experiment. When the correct structure has been found there follows a series of successive refinements in which the atomic co-ordinates are adjusted to give closer and closer agreement between the calculated and observed diffraction patterns. This procedure is often a laborious and formidable task and the extent to which the refinements of the trial structure are carried depends on the amount of detail required.X-Ray Dflraction.-The scattering of X-rays arises primarily from their interaction with the extra-nuclear electrons of the atoms and the structure arrived a t is a particular distribution of electron density in the crystal. This is often presented in the form of maps showing sections or projections of the crystal lattice with contours of equal electron density. The positions of the atoms are usually assumed to coincide with the regions of maximum electron density. The electron density in the region of a hydrogen atom is extremely low and diffuse in comparison with that near other atoms. The hydrogen atom is associated with fewer electrons than any other namely only one and even this is to some extent shared with the atom to which it is covalently bonded. If tlhe link has appreciable ionic character with the hydrogen atom the positive end of the dipole there is virtually no maximum of electron density a t the proton.The large amplitude of thermal vibration of the light hydrogen atom often makes the situation even less favourable. These considerations suggest that X-rays are unlikely to provide a sensitive method of locating hydrogen atoms. Indeed hydrogen atoms are often ()mitted altogether from the calculations although it is sometimes remarked that their inclusion in assumed positions improves significantly the agreement between the calculated and the observed patterns. In spite of these difficulties it is sometimes possible to locate hydrogen atoms if the most refined experiments and methods of interpretation are used. Intensities are re-measured with a Geiger counter instead of a photo- graphic plate and much greater accuracy is achieved although the measure- ments are long and laborious.These measurements are used to refine the model obtained from a conventional analysis and finally a so-called difference synthesis is carried out. The distribution of electron density due to the heavy atoms only is calculated from the most refined model and in effect subt'racted from the actual distribution of electron density obtained from t>he experimental results. Maps are then obtained of the variation of electron density due to the hydrogen atoms only. Good examples of electron density maps for hydrogen atoms only obtained by difference syntheses are given by Penfold in a description of X-ray studies on the structure of 2-pyridone.The positions of the hydrogen atoms are clearly defined and the estimated accuracy of the carbon-hydrogen and nitrogen-hydrogen bond lengths is & 0.1 8. Other examples of work 2 Penfold Acta Cryst. 1953 6 591. 482 QUARTERLY REVIEWS of this kind include the determination of the structures 3 4 of B4H1 and B5H11 in which the boron-hydrogen bond lengths are determined with an error of -+ 0.04 to The positions of the hydrogen atoms in salicylic acid have been determined with an accuracy of -J= 0.1 A and in boric acid the 0-H distance has been found to be 0.88 8 though the author says this must be too low. Pig. 1 shows a projection of the salicylic acid crystal obtained by Cochran 5 by X-ray diffraction. The contours show 0.1 A. I FIa. 1 The electron distribution in a single molecule of salicylic acid from which the contributions of carbon and oxygen atoms have been subtracted.Contours at every 0.1 d-= zero contours omitted negative contours in broken line. (Reproduced by kind permission from Cochran.6) the distribution of electron density due to the hydrogen atoms only the contributions of the heavy atoms having been removed by carrying out a difference synthesis. The positions of the heavy atoms are shown by dots and the arrangement of the hydrogen atoms can be seen from the contours. It is clear that the hydrogen atoms in the hydrogen bonds between the carboxyl groups lie nearer to one oxygen atom than to the other. Electron Dif€raction.-The method of electron diffraction has been developed recently in Russia and in Australia 8 for the study of crystal 3 Lipscomb J .Chim. phys. 1949 46 252. 5 Cochran Acta Cryst. 1953 6 260. 7 Pinsker Trudy Institouta mineralnovo syria N.T.I. 1936 109 ; Idem J . Chem. Phys. 1954 22 614. Zachariasen ibid. 1954 ’7 305. Uspekhi khim. 1939 8 ; Pinsker “ Diffraction des Blectrons ” Editions de l’Acad6mie des Sciences de l’U.R.S.S. 1949 ; Pinsker Travaux de Z’lnstitut de Cristallographie Livreison 10 Communications au I11 Congres International de Cristallographie 1954 p. 91 ; Wein- stein ibid. p. 115 ; Weinstein and Pinsker ibid. p. 145. Cowley Actn Cryst. 1953 6 516. RICHARDS HYDROGEN ATOMS IN CRYSTALS 483 structures. A highly collimated beam of electrons accelerated through 50-70 kv is diffracted by small very thin single crystals or by oriented polycrystalline preparations. Although there are certain difficulties con- cerned with the theoretical interpretation of the diffraction patterns it is often possible to carry out a full three-dimensional Fourier analysis in a manner analogous to that used in X-ray diffraction.1 The electron beam gives diffraction patterns by virtue of its wave properties but because it consists of charged particles it interacts with regions of high electrostatic potential.Thus the contour maps obtained give the variation of electrostatic potential through the crystal instead of electron density as in the case of X-rays. The electrostatic potential in the crystal is determined by the distribution of the positively charged nuclei and of the electron clouds which surround them. The nuclei provide the points of maximum potential and the more diffuse the electron distribution around the nucleus the steeper will be the equipotential contours due to t,he positive nucleus.This is the opposite situation to that found for X-rays and it means that whereas the scattering factor of hydrogen for X-rays is about one-fifteenth of that of carbon for electrons its scattering factor is only about one-fifth that of carbon. Any ionic character of the link to the hydrogen atom which draws the electrons from the proton into the bond will make the hydrogen atom more readily detected by electron scattering whilst it is made more difficult by X-ray diffraction. By taking advantage of this Pinsker and TatarinovaO studied the crystal structure of paraffin and found the carbon-hydrogen distance to he 1-17 -j- 0.05 8. Cowley lo has studied the structure of boric acid and finds the oxygen-hydrogen distance to be 1.00-1.05 A (cf.Zachariasen 6 ) . Weinstein 7 studied the crystal structure of dioxopiperazine and was able to determine the carbon- and nitrogen-hydrogen distances with an error of approximately -+ 0.03 A and Lobatchev l1 determined the structure of urotropine by electron diffraction and quotes the carbon-hydrogen distance ;LS 1.17 & 0.1 A. Neutron DifPraction.12-A beam of neutrons of momentum mu where 'IIZ and v are respectively the mass and velocity of the neutrons behaves as though it possesses wave properties of wavelength il = h/mv. Neutron beams of wavelength about 1 A are therefore diffracted by crystals and the patterns obtained can be used to derive information about the crystal structure. Sufficiently intense neutron beams can only be obtained from an atomic pile ; a collimator is built into the pile and there emerges a beam of neutrons ivhich have a range of ve10cities.l~ This collimated beam of neutrons is made nearly monochromatic by allowing it to be diffracted by a large crystal and then collecting a narrow beam of the diffracted neutrons a t 9 Pinsker and Tatarinova Acta Physicochinz.U.R.S.X. 1936 5 381. lo Cowley Acta Cryst. 1953 6 516 522 846. 11 Lobatchev Travaux de l'lnstitut de Cristallographie Livraison 10 p. 167. l2 Bacon " Neutron Diffraction " Oxford Univ. Press 1955. l3 Bacon and Thewlis Proc. Roy. SOC. 1949 A 196 50. HH 484 QUARTERLY REVIEWS a suitable angle. The monochromatic neutron beam is then diffracted by the sample which may be a powder or single crystal and the diffraction pattern is measured.A boron trifluoride neutron-counter is used to measure the neutron intensity as a function of diffraction angle. Neutron scattering by diamagnetic materials occurs mainly by inter- action of the neutrons with the atomic nuclei in the crystal and the atomic scattering factors for neutrons vary only by a factor of about three over the Periodic Table in striking contrast with the atomic scattering factors for X-rays which depend on the atomic number. This implies that the hydrogen atoms in a crystal contribute as effectively to the scattering pattern as most other elements and hence their positions can be determined with approximately equal accuracy. It happens that hydrogen atoms in molecules give rise to a very high proportion of incoherent seattering (this is due largely to the very different interactions which occur when the neutron and scattering proton have parallel nuclear spins as compared with when The effective wavelength chosen is about 1.2 8.0' 90' 180' FIG. 2 Fourier synthesis of the neutron scattering density projected on the (001) plane of KH,PO at room temperature. Continuous contours are positive broken lines are negative and zero contours are dotted. The elongated contours due to the hydrogen atoms in the hydrogen bond between the two oxygen atoms are clearly visible. (Reproduced by kind permission from Pease and Bacon.lsa) they have antiparallel spins) and whilst this is not a serious inconvenience for work with single crystals it is usual to use deuterated materials for work on powdered samples because deuterium atoms give a much weaker RICHARDS HYDROGEN ATOMS IN CRYSTALS 485 incoherent scattering than hydrogen atoms.However although neutron diffraction appears to be ideally suited to the location of hydrogen atoms in crystals various difficulties of interpretation and the reiatively low resolu- tion at present available require that an X-ray investigation of the crystal should be carried out first. The neutron-diffraction results can then be used to fill in the details not obtainable from the X-ray patterns. Measure- ments have so far been confined to relatively simple niolecules. Peterson and Levy 14 studied a single crystal of heavy ice a t - 50" C and found the oxygen-deuterium bond length to be 1-01 A with the DOD angle nearly tetrahedral ; they showed l5 from single crystal-measurements on potassium hydrogen difluoride that the proton was midway & 0.1 A between the fluorine atoms in the HI?,- ion.The same authors l6 have determined t,he structure of ammonium chloride from single-crystal measurements and found the nitrogen-hydrogen distance to be 1.03 & 0.02 8 and they have also made a detailed study of all four phases of powdered deuteroammonium Ixomide. 17 FIG. 3 Projection on (001) of the density of neutron scattering due to hydrogen atoms in KH,P04 wt - 180". The small circles mark the p o s i t i o ? ~ of the K P and 0 atoms the contribu- lions of which to the scattering pattern have been removed. The contours are as in Fig. 2. The ordered arrangement of th.e hydrogen atoms was maintained by applying a n electric field to the crystal during the measurements.(Reproduced by kind permission f r o m Pease and Bawn.l8b) l 4 Peterson and Levy Phys. Rev. 1953 92 1082. l 5 Idem J . Chem. Phys. 1952 20 704. l6 Idem Phys. Rev. 1952 86 766. 17 Levy and Peterson J. Amer. Chem. SOC. 1953 75 1536. 486 QUARTERLY REVIEWS Measurements directed to the location of hydrogen atoms in hydrogen- bonded substances also include the study l8 by Pease and Bacon of potassium dihydrogen phosphate KH,PO,. In this substance the hydrogen atom in the 0-H-0 hydrogen bond is considerably elongated along the 0-0 axis. Fig. 2 shows a projection of this crystal on the (001) plane and the contours of the hydrogen atoms lying between the two oxygen atoms show this asymmetry very clearly. The hydrogen atoms are indicated by the negative contours lying between the well-defined oxygen atoms.This elongation may be due to thermal vibrations of the hydrogen atom along the 0-0 axis or to a disordered arrangement of the hydrogen atoms among pairs of possible positions closer to one oxygen atom than t o the other. If the latter is true then the two positions cannot be further from the centre of the 0-0 line than & 0.175 A. Below - 150" the crystal becomes ferro- electric and when an electric field is applied the hydrogen atoms are then no longer elongated and lie asymmetrically between the oxygen atoms in an ordered arrangement. This is shown clearly in Fig. 3 which is projected on exactly the same plane as Fig. 2 although the contours represent the result of a difference synthesis in which the contributions of the heavy atoms have been removed leaving only the contributions 'of the hydrogen atoms.The positions of the heavy atoms are shown by the small circles. Spectroscopic Methods Vibrational Spectra.-The atoms in a molecule can vibrate in a number of ways known as fundamental vibrational modes and the frequencies of these vibrations are determined by the masses of the atoms and the forces between them. Most of these vibrational modes involve very complicated motions of all the atoms and the particular values of the frequencies are often highly characteristic of a particular molecule. In certain cases how- ever some of the modes of vibration are highly localised in the molecule and involve the motion of atoms in a particular group. These vibrations which are localised in a particular group within the molecule often occur a t characteristic frequencies in all sorts of molecules which contain this particular grouping.These so-called " characteristic group frequencies " are very widely used for analysis and structural diagn0~is.l~ The vibrational mode which involves the stretching vibration of an X-H bond is particularly characteristic of the X atom and its environment because owing to the great difference in mass between X and H the motion consists largely of the vibration of the hydrogen atom against the heavy X atom which remains nearly stationary. The masses of other atoms attached to X do not therefore have much effect on the vibration frequency of the X-H bond. The study of these characteristic frequencies in a crystal can therefore be used to decide to which atom a particular hydrogen atom is joined.If the infrared radiation is polarised before it is allowed to fall on a single crystal so that the electric vector of the radiation vibrates in only one direction with respect to the crystal axes then those molecular vibrations l 8 Pease and Bacon ( a ) Proc. Roy. SOC. 1953 A 220,397 ; ( b ) Nature 1954,173,443. 19 Bellamy " The Infra-red Spectra of Complex Molecules " Methuen London 1954. RICHARDS HYDROGEN ATOMS IN CRYSULS 487 which involve a change of dipole moment along this direction will absorb more strongly than those for which the change of dipole moment is per- pendicular to the direction of polarisation of the radiation. This effect can be used to study the directions in which certain bonds point with respect to the crystal axes.For example Ambrose and Elliott 2O have examined natural proteins and oriented films of synthetic polypeptides in this way. I n the a-fibres the nitrogen-hydrogen stretching vibration a t 3300 cm.-l absorbs more strongly when the radiation is polarised along the fibre axis This indicates that the nitrogen-hydrogen bond lies parallel to the poly- peptide chain in agreement with the folded-chain structures attributed to these substances. On the other hand the ,&fibres show strongest absorption of the nitrogen-hydrogen stretching frequency when the radiation is polarised a t right angles to the fibre axis indicating that the nitrogen-hydrogen bonds now lie perpendicular to the fibre axis as required by an extended- chain structure. Nuclear Resonance.-Nuclear resonance spectra 21 arise when transitions are induced by radiation among energy levels which become available to a nucleus when it is placed in a magnetic field.This branch of spectroscopy provides a useful method of locating hydrogen atoms in certain crystals for which the arrangement of hydrogen atoms is not too complicated. The process which gives rise to a nuclear resonance spectrum will first be described and then the interactions which occur between nuclei will be discussed so far as they help in the location of hydrogen atoms. Many atomic nuclei behave as though they possess a " spin " i.e. they have an angular momentum p . It can be shown classically that a spinning uniform sphere of mass M and charge e has associated with it a magnetic moment pe/2Mc which arises from the circulation of electric charge.For i%ctual particles this is not quite true and we write where g is known as the nuclear g factor and must be obtained experimentally for each nucleus. The quantum theory requires that the angular momentum p of a particle must be given by where h is Planck's constant and I is a " spin" quantum number which may have integral or half-integral values and is a characteristic of the particle. Actual magnetic moment z= pa = g.pe/2Mc . * (11 p = ( h / 2 n ) d [ I ( I + 1)1 * . (2) Therefore from equations (1) and (a) where po is a unit of magnetic moment called the nuclear magneton. We may therefore regard the nucleus as a short bar magnet of moment given hy equation ( 3 ) . 2o Ambrose and Elliott Proc. Roy. SOC. 1951 -4 205 47 ; 1951 A 208 75; 21 Bloembergen Purcell and Pound Phys.Rev. 1948 '73 679 ; Bloch ibid. 1946 1951 A 206 206; 1952 A 211 490. 70 460 ; Andrew " Nuclear Magnetic Resonance " Cambridge Univ. Press 1955. 488 QUARTERLY REVIEWS If this magnet is placed in a uniform magnetic field H, it experiences a torque like a compass needle but on account of its spin angular momentum it precesses about the field Ho just as a toy gyroscope suspended on a vertical pivot precesses about the Earth's gravitational field. The potential energy of the nuclear magnet in the field H clearly depends on its orientation in the field and apart from an additive constant is given by The quantum theory demands that this energy shall be quantised and the condition is that the component of angular momentum in the direction of the field is mh/2n where m = I I - 1 I - 2 .. . 1,0 - 1 - 2 . . . - I . Thus from equations (l) (2) and (3) the component of pa along the field pm is restricted to the values and the potential energy of the nucleus in each of these levels must be from equations (4) and (5) Since there are (21 + 1) values of m there are (21 + 1) energy levels defined by equation (6) and each corresponds to a different orientation of the nuclear magnet in the field Ho. In each level the nucleus precesses about the field maintaining the appropriate orientation which gives the allowed - H x component of pa along H,. * (4) mgpo =PnlPo * * ( 5 ) U(m) = - Hopmpo = - Homgpo * ' (6) FIG. 4 Energy levels of a nucleus of spin = & in a field H,. projection of p or pa along the field. In the case that I = 8 (for example for 1H 19F or 31P nuclei) m can be + 8 or - 4 pa is (gp,d3)/2 and ,urn = + igp or - +gp (Fig.4). The half-angle of precession is then c0s-~(d3/3) or c0s-l (- 1 / 3 / 3 ) . Clearly the maximum component of pa along the field is ~ O / % n ( m a x ' ) IgpO = pp07 since I is the maximum value of m. The dimensionless quantity p is what is usually called " the magnetic moment " of the nucleus. If transitions are induced by radiation among these energy levels which become available to the nucleus in a magnetic field then energy must be RICHARDS HYDROGEN ATOMS IN CRYSTALS 489 absorbed or emitted and the resulting spectrum is called the nuclear reson- ance spectrum. The frequency of radiation required to do this is such that r3 U = hv, where d U is the separation of two adjacent energy levels and v is the resonance frequency since a selection rule applies which limits A m to 1.Thus we have from equations (6) h v = AU,,+,+l = U(m) - U(m + 1) = H,g,u,[m + 1 - m] This frequency v, turns out to be identical with the precession frequency of the nucleus in the field H as given by Larmor’s theorem and a detailed classical calculation 21 of the interaction of the precessing nucleus with an oscillating magnetic field leads to the conclusion that the nucleus will be tipped from a given orientation by a radiation field rotating at the Larmor frequency in phase with the nuclear precession. Such a field is generated by electromagnetic radiation oscillating with its vector a t right angles to H,. I n a magnetic field of 10,000 gauss the radiation needed falls in the radiofrequency region of the spectrum.For example for protons in a field of 10,000 gauss v is about 42.6 Mc./sec. ; for fluorine v is about 40.1 Mc./sec. and for nitrogen v is about 3.1 Mc./sec. The transition probability is the same for Am = + 1 or - 1 so that for net absorption of energy to occur the lower energy levels must be more populated than the upper ones. The levels are so closely spaced that this excess of population is very small. For example for protons in a field of 10,000 gauss a t room temperature a Boltzman distribution would lead to an excess of population in the lower energy level of only 3 or 4 nuclei in every million and it is on these that we must rely for our measurements ! The absorption spectrum can be measured by keeping the applied field H constant and finding the frequency a t which energy is absorbed or alternatively by using a fixed frequency and varying the separation of the energy levels by altering the value of H,.The spectrum can be plotted as intensity of absorption against frequency or against applied field strength but for a given nucleus it is easy to convert one scale into the other by using equation (7). For isolated nuclei the absorption spectrum would be a single very sharp line a t a frequency given by equation (7) but in a crystal the inter- nuclear interactions modify this. A nucleus of spin I = $ (for example hydrogen) and moment ,upo may occupy two orientations in a magnetic field H,. This nucleus generates in its neighbourhood a weak local magnetic field of its own which may be resolved into a static component in the direction of H and a rotating component (arising from its precession about H,) a t right angles to H (Fig.5). The static component of the local field in the direction of H is & (pp0/r3)(3 cos2 8 - 1) at a point distant r from the nucleus and of co-ordinate angle 8 to the direction of H,. The sign depends on the momen- tary orientation of the nucleus in H (Fig. 5 ) . If the nuclei are grouped in pairs at a distance r each one will therefore find itself in 8 field = How = PPoHoP * ( 7 ) H & (ppo/r3>(3 cos2 8 - 11, 490 QUARTERLY REVIEWS and a double absorption line would be expected with the doublet separation equal to (2pp0/r3)(3 cos2 I9 - 1). The rotating component of the local field generated by one nucleus rotates a t the Larmor frequency which is itself nearly equal to the resonance frequency of a nearby nucleus because they are in closely similar fields H,.Each nucleus of the pair will therefore experience a rotating field near the resonance frequency polarised a t right angles to H,. These are just the conditions required to induce transitions among the allowed energy levels. Two neighbouring nuclei which occupy adjacent energy levels could therefore exchange a quantum of energy through this interaction of the rotating components of their local fields and simul- taneously exchange energy levels and orientations in the field H,. This so-called spin-spin interaction involves exchange of energy only among the nuclei and no energy is gained or lost by the surroundings. This spin exchange has the effect of limiting the life-time of a nucleus in a given energy level and leads through the Heisenberg uncertainty principle t o a further magnetic interaction which gives a broadening about one-half of that caused Rotating c o m ~ onen t FIG.5 Local fields generated by a precessing nucleus. by the static components of the local fields.22 In the particular case of a pair of like nuclei the separation of the doublet is increased by this quantum mechanical effect t o (3p,u,/r3)(3 cos2 I9 - 1). Of course if the nuclei are unlike so that their precession frequencies are different the second broaden- ing mechanism is negligible. A crystal which contains protons grouped in pairs would therefore be expected to show a nuclear resonance spectrum which is a doublet broadened to some extent by second-order interactions between relatively distant pairs of nuclei.The separation of the doublet depends on the distance between the adjacent nuclei and on the angle which the line joining them subtends to the field. The separation of the doublet therefore varies with the orienta- tion of the crystal in the field. Pake 22 studied the proton resonance of a single crystal of gypsum as a function of crystal orientation in the magnetic field. The two water molecules have different orientations in the crystal so two pairs of fines were observed and from their separation and angular dependence it was possible to measure the H-H distances and the angles 2 2 Pake J . Chem. Phys. 1948 16 327. RICHARDS HYDROGEN ATOMS IN CRYSTALS 49 1 they subtend to the crystal axes. Fig. 6 shows the absorption curves obtained for the proton resonance of a single crystal of gypsum for various orientations of the applied magnetic field H, in the (001) plane of the .. QLO" (H' a/ong [ 7001 1 +540 FIG. 6 Absorption curves of proton resonance in a single crystal of gypsum for various orientations of the crystal with respect to the magnetic $eld €5,. (Reprodu.ced by kind permission f rom Puke. 2 crystal. The angle between H and the [loo] direction is stated with each spectrum. At some settings e.g. 54" the two pairs of lines due to the two distinguishable water molecules in the crystal are clearly separated whereas at others they coalesce. Fig. 7 shows how the separation of the doublet FIG. 7 Variation of the separation of the doublet peaks due to each water molecule in the unit cell of gypsum as a function of the angle C$ between H and [loo].(Reproduced by kind permission from Pake.22) 492 QUARTERLY REVIEWS peaks varies with the angle q5 between H and [loo]. The two different water molecules in the unit cell give rise to two curves which differ only in phase. The circles are experimental points and the squares and triangles are calculated for an H-H distance of 1.58 A. A simpler example is the study of the structure of urea by Andrew and H ~ n d r n a n . ~ ~ The crystal structure of urea was studied by X-ray methods 2 4 and the heavy atoms were located readily but it was not possible to decide whether the planes of the amino-groups lay in the same plane as the rest of the urea molecule or a t right angles to it. The two molecules of urea in each tetragonal unit cell are both arranged with the C=Q bonds parallel to the tetragonal axis.I f the crystal is placed in a magnetic field H, with the tetragonal axis parallel to the field it turns out that the proton pairs of the amino-groups are oriented a t the same angle 8 to the field and on the basis of the planar model for the urea molecule 8 would be 31" & l&" whilst for the non-planar model 8 would be 90". By use of the value for THE the H-H distance of 1.803 & 0.015 A obtained from separate measurements of the second moment of the proton resonance (see below) the doublet separation ~,,!L,,!L,Y-~(~ cos2 8 - 1) is 8.7 5 0.5 gauss for the planar model (0 = 31") and 7 - 2 gauss for the non-planar model (8 = 90"). The observed value of 8.9 & 0.2 gauss confirms the planar model and determines the orientation of the amino-groups in the crystal.Further information can be obtained by making use of a relation due to Van V l e ~ k . ~ ~ For a crystal containing any arrangement of nuclei the mean square-width or second moment of the nuclear resonance absorption line is given by the expression (dH2)AV = (3/2)1(1 + l)g2,uo2. N - l z [(3 cos2 8,k - 1)2rjk-6] j)k where N is the number of nuclei at resonance in the unit cell Oj is the angle between the jk vector and the applied field H, If and gf refer to nuclei of a different kind from those a t resonance a,nd rj is the distance between the j - t h and k-th nuclei. The second moment or mean-square width of an absorption line can be obtained readily from the experimental measurements by numerical or graphical methods since (4H2), = -co liof(H).dH2.dH/ -a l+of(&):dH where f(H) is the absorption line intensity as a function of H and AH is the deviation from the centre of the symmetrical resonance line.I n the case of urea Andrew and Hyndman 23 computed the expected a3 Andrew and Hyndman Discuss. Faraday SOC. 1955 19 195. 24Vaughan and Donohue Acta Cryst. 1952 5 530. 25 Van Vleck Php. Rev, 1948 '74 1168. RICHARDS HYDROGEN ATOMS IN CRYSTALS 493 values of the second moment of the proton resonance for various values of r H H and various orientations of .the crystal in the field using the planar and the non-planar model. The variation of second moment with orienta- tion in the field provided very convincing confirmation of the planar structure and the value of the H-H distance was found from the experimental second moments to be 1.803 0.015 A corresponding to an N-H distance of 1.046 0.01 A a.nd an HNH angle of 119.1" 2".If a finely powdered solid is used instead of a single crystal the absorption line obtained is an average resulting from superposition of the lines from all orientations of the crystal axes. The line shape can be calculated for a powder 22 with its protons grouped in pairs and is shown in Fig. 3. The broken line shows the shape calculated for nearest-neighbour interactions only. Relatively distant nuclei also exert a weak broadening effect and this weak effect is usually assumed to have a gaussian form exp [ - ( ~ l H ) ~ / 2 8 ~ ] where p2 characterises the amount of this so-called intermolecular broadening. I I I \ I I I 1 H FIG. 8 Continuous line i s obtained by applying a gaussian broadening function to the broken line to take account of intermolecular interactions.Broken line Absorption curve for a n isolated pair of nuclei of I = 8 (Pake z2). In Fig. 8 the continuous line is the result of applying such a broadening function to the curve calculated from nearest-neighbour interactions only. Similar calculations can be made for other simple configurations of nuclei and the line shape for powdered crystals with nuclei grouped in equilateral t'riangles is shown 26 in Fig. 9. The second moment of the absorption line for a powdered sample can be calculated from equation (8) by replacing the terms in (3 cos2 8 - 1)2 by their average values over space. For a powder the second moment then becomes - {AH2)AV = (6/6)1(1 + l)g2pO2. N - l Z rj,-6 + (4/15),~,,~N-~ j)k The configurations of protons in the crystals of powdered substances 26Andrew and Bersohn J .Chem. Phys. 1950 18 159. 494 QUARTERLY REVIEWS H FIG. 9 Bersohn 26). Broken curve Absorption curve for isolated triangles of nuclei of I = & Continuous curve is obtained by applying a gaussian broadening function (Andrew and as in Fig. 8. have been studied in this way for a variety of crystals. For example the proton resonance spectrum of infusible white precipitate has been studied by Deeley and Richards.27 This material is precipitEtd from mercuric FIG. 10 Absorption curves expected for diflerent structures ( a ) tetrahedral groups of nuclei (I) ( b ) pairs of nuclei (11) (c) triangles of nuclei (111). (Reproduced by kind permission from Deeley and 27 Deeley and Richards J.1954 3697. RICHARDS HYDROGEN ATOMS I N CRYSTALS 495 chloride solutions by ammonia under the right conditions and various structures have been assigned to it. Rammelsberg 28 formulated it as NHg2C1,NH,C1 (I) Franklin 29 as NH2HgC1 (11) and Britton and Wilson 3O and Glasson and Gregg 31 as xHgO,(l - x)HgC12,2NH (111). Lipscomb 32 interpreted X-ray diffraction patterns in terms of chains of mercury atoms and amino-groups with chloride ions packed between the chains. I n structure (I) the protons are grouped in tetrahedra in (11) in pairs and in (111) in equilateral triangles and in Fig. 10 the calculated absorption curves for these three structures are drawn approximately to scale. The second moments of the proton resonances expected for the three structures are respectively about 50 gauss2 20 gauss2 and 38 gauss2.Experimental measurements gave curves which could be fitted closely to the pair structure (11)) with a second moment of 18.6 gauss%. From the second moment and quantitative comparison of the line width with theoretical curves the hydrogen-hydrogen distance was found to be 1.688 A. This distance corresponds to an N-H distance of 1.03 A with a tetrahedral HNH angle as in the ammonium ion and provides convincing confirmation of the chain structure ( A ) . Similar experiments in which absorption line shapes have NH2 + + CH ,=C-CH 2 I 1 0-CO ( B ) /NH2\ / \ Hg \ /Hg NH2 + ( A ) /Hg been used to identify the configurations of hydrogen atoms in crystals have shown that nitric sulphuric perchloric and chloroplatinic acid monohydrates crystallise as oxonium (H,O+) salts but oxalic acid dihydrate does not ; 33 that hydrazine nitrate and sulphate contain the W2H,++ ion but the oxalate the N,H,+ ion ; 34 that diketen crystallises 35 in the structure (B) and that the mono- and the di-hydrate of boron trifluoride crystallise as true hydrates and not as oxonium ions as has been suggested.36 If the positions of the heavy atoms in a crystal have been found by X-ray methods it is sometimes possible to use the second moment of the nuclear resonance spectrum to locate the hydrogen atoms even when their configuration does not give a line with useful structure.This method is particularly applicable to problems in which only one parameter remains unknown. For example in an ionic crystal containing XH ions for which the positions of the X and other atoms are known only the X-H distance remains to be determined if the ion is assumed to be tetrahedral.Tho second moment depends on the inverse sixth powers of the distances and z8 Rammelsberg J . prakt. Chem. 1888 38 558. 29 Franklin J. Amer. Chem. SOC. 1907 29 35. 30 Britton and Wilson J. 1933 601 1045. 3 l Glasson and Gregg J. 1953 1493. 3 3 Richards and Smith Trans. Paraday SOC. 1951 47 1261 ; 34 Pratt and Richards ibid. 1953 49 744. 35 Ford and Richards Discuss. Faraday SOC. 1955 19 193. 36 Idem J. 1956 3870. 32 Lipscomb Acta Cryst. 1951 4 266. 1952 48 307. 496 QUARTERLY REVIEWS so the major contribution to it comes from the H-H and X-H interactions within the ion and a small proportion arises from inter-ionic effects. The X-H distances in the NH,+ i0n,37 the BH,- i0n,~8 and the PH4+ ion 39 have been determined in this way and the method is best illustrated by a more detailed consideration of the work 37 on ammonium chloride.The second moment of the proton resonance a t - 195" was found to be 49.5 & 0.5 gauss2. The crystal structure is cubic and the positions of the nitrogen and chlorine atoms are known. The intermolecular contribution to the second moment from interactions between hydrogen atoms in diflerent ions are calculated from equation (9) an approximate value of the N-H distance of 1.04 A being used and found to be 6.5 gauss2. The contribution of the 14N and 35Cl and 37Cl atoms are now computed and since their nuclear moments are small this contribution is found to be only 0-1 gauss2. When these two intermolecular broadening terms are subtracted from the total experimental value we are left with the contribution from intramolecular terms only namely 42.9 gauss2.A correction must now be applied which may be very important for symmetrical ions which are not very rigidly fixed in the lattice. Although the measurements are made a t low tempera- tures the ammonium ion still undergoes a zero-point vibrational motion (a) of the hydrogen atoms with respect to themselves and to the nitrogen atom and more important (b) of torsional oscillation of the ion as a whole in the force field which holds it in position in the crystal. Correction for (a) (Deeley and Richards 40) is needed because the broadening of the lines depends on the mean value of r3 and not on the cube of the mean value of r and these are different for a vibrating bond.The correction is usually very small however and will be ignored here. Correction for ( b ) is needed because if the ion is rocking in some way in the crystal this motion will to some extent reduce the local fields generated by neighbouring nuclei by an averaging effect and the resultant reduction in the intramolecular contribu- tion to the second moment can be calculated in terms of the amplitude of this zero-point rocking motion. E'or ammonium chloride the frequency of this torsional motion is known from spectroscopic measurements 4 1 to be 390 cm.-l so that the amplitude can be calculated. This motion is found to have reduced the intramolecular second moment by the factor 0.913 so that the correct value should have been 47.0 gauss2. By using this value and equation (9) the nitrogen-hydrogen distance is found to be 1.038 & 0.004 8.If the nuclear resonance spectrum of more than one type of atom can be observed then additional parameters can be obtained. In this way Deeley and Richards 4O measured the second moments of the proton and fluorine resonance spectra of hydrazine fluoride. The positions of the nitrogen and fluorine atoms are known from X-ray measurements and this leaves two unknown parameters in the crystal structure namely the 37 Gutowsky Pake and Bersohn J. Chem. Phys. 1954 22 643. 39 Pratt and Richards Trans. Paraday Soc. 1954 50 670. 40 Deeley and Richards ibid. p. 560. 41 Wagner and Hornig J. Chem. Fhys. 1950 18 296 305. Ford and Richards Discuss. Paraday SOC. 1955 19 230. RICHARDS HYDROGEN ATOMS I N CRYSTALS 497 N-H distance and the H...F distance in the strong hydrogen bonds formed.The second moments of the proton and fluorine resonances provide the two necessary measured parameters and the N-H distance was found to be 1.07 (cf. the N-H distance in ammonium chloride) and the H.-.F distance to be 1.54 A. The sum of these 2-62 A is equal to the N..*F distance found by X-ray methods and confirms that the hydrogen atom in this hydrogen bond lies along the N...F line. If the crystal structure of the solid is not known the intermolecular broadening can sometimes be estimated without a very serious loss of accuracy but in certain cases a better method can be used which was devised by Andrew and E a d e ~ . ~ ~ Experimental values of the proton second moment were obtained for benzene and for 1 3 5-trideuterobenzene.The deuteron has a much smaller magnetic moment than the proton and the substitution reduces the intra- and inter-molecular contributions by very different factors both of which can be calculated. From the two measured second moments the inter- and intra-molecular contributions can be separ- ately evaluated. Conclusion The merits of these methods for the location of hydrogen atoms may be summarised as follows. X-Ray dijfraction. Measurements with the most refined techniques must be used and the experimental work may be considerable. The interpretation of the diffraction patterns must be carried out in the greatest possible detail and is usually a very lengthy and complicated operation. The accuracy is low but the method can be applied to complicated structures as well as to simple ones.Measurements must be made on carefully prepared specimens of correct thickness. The protons make a more important con- tribution to the diffraction pattern than in the case of X-rays so that better accuracy can be achieved with less elaborate methods of interpretation. Certain difficulties occur in the interpretation of the patterns in unfavourable cases but the method is promising. The experiments are difficult to carry out and require expensive and elaborate equipment. Neutron-beam intensities a t present available are low so that the method has very limited resolving power. When used in conjunction with other diffraction methods hydrogen atoms can be located with moderate accuracy but this will undoubtedly be improved when more intense neutron sources become available. Use of polarised radiation can give qualitative informa- tion about the positions of hydrogen atoms. The measurements are easy to perform and take only a few hours. The interpretation is relatively easy but useful information can only be derived for rather simple structures. In cases where it can be applied the accuracy is good compared with that of the other iiiethods. Electron diffraction. Neutron diffraction. Infrared spectra. Nuclear resonance methods. 42Andrew and Eades Proc. Roy. SOC. 1953 A 218 537.

ISSN:0009-2681    

DOI:10.1039/QR9561000480

出版商:RSC    

年代:1956

数据来源: RSC