摘要:

QUARTERLY REVIEWS PHOTOCHEMICAL REARRANGEMENTS AND RELATED TRANS- FORMATIONS By P. DE MAYO and S. T. REID (DEPARTMENT OF CHEMISTRY THE UNIVERSITY OF WESTERN ONTARIO LONDON CANADA) 1. Scope WITHIN the last few years a number of Reviews on various aspects of photochemistry both mechanisticf and preparati~e,~.~ have appeared. The present Review is concerned with photochemically induced rearrange- ments and related transformations which occur in condensed phases. Although incidental addition of solvent may occur in these processes this Review does not in general include intermolecular reactions such as dimerisation* or the addition of large rn01ecules,~s~~~ or of oxygene or other small molecule^.^ Photo-oxidative processes are excluded as are also rearrangements of free radicals or carbenes2s3 which may efficiently be produced by non-photochemical methods.The Sus rearrangement of diazo-ketones has been adequately covered e l s e ~ h e r e . ~ ~ ~ All the sub- stances considered contain chromophores having absorption bands in the ultraviolet region and hence the rearrangements are induced by light of these wavelengths. 2. Simple Aldehydes and Ketones Aldehydes and ketones absorb weakly in the 290 mp region of the ultraviolet with an extinction of 10-30;8 stronger absorption also occurs below 200 mp that at shortest wavelength being analogous to that occurring in the spectra of the related ~lefins.~ The weak band in the car- bony1 spectrum is attributed to an n+n* transition of a non-bonding electron on the oxygen atom to an anti-bonding n-orbital in the multiple Inter al.Simons Quart. Rev. 1959 13 3; C. Reid ibid. 1958 12 205; Wyman Chem. Rev. 1955 55 625. Schonberg “Praparative Organische Photochemie,” Springer-Verlag Berlin 1958. Mayo “Advances in Organic Chemistry” Interscience Publ. Inc. New York 1960 Vol. 11 p. 367. Mustafa Chem. Rev. 1952 51 1. Schonberg and Mustafa Chem. Rev. 1947,40 181. Bergmann and McLean Chem. Rev. 1941,28 367; Bateman Quart. Rev. 1954,8 147; Schenck Angew. Chem. 1957 69 579; Nickon and Bagli J. Amer. Chenz. Sac. 1961,83 1498. ’ Schenck and Schmidt-Thomke Annalen 1953,584 199. * Bielecki and Henri Ber. 1913,46 3627. Wheland “Resonance in Organic Chemistry” Wiley New York 1955 p. 278. 393 394 QUARTERLY REVIEWS bond to give a singlet.lo It is this absorption and consequent formation of an energised state which appears to be responsible for most of the photo- chemical transformations of saturated aldehydes and ketones in condensed phases.After absorption the energy may be disposed of by re-emission by photo-chemical reactions such as dissociation rearrangement and reaction with the medium by intersystem crossing and by internal conversion.ll In the photochemical transformations to be described there is as yet little evidence to indicate the nature of the initial reacting species and studies successfully directed to this end for condensed-phase reactions have been comparatively few.l2,l3 In the following discussion therefore no distinction will in general be made. (a) The Cleavage of Cyclic Ketones.-Irradiation of acyclic ketones in the vapour phase leads to bond cleavage between the wcarbon atom and the carbonyl group followed by intermolecular reactions of the derived free radicals in ways dependent on the temperature.14 Cyclic ketones behave similarly but the diradicals so formed may subsequently undergo a number of intramolecular reactions.One such reaction apparently only observed in the gas phase is the extrusion of carbon monoxide and recyclisation to the hydrocarbon con- taining one carbon atom less.15 A recent example is the conversion of the ketone (1) into the hydrocarbon (2).16 In solution the intermediate diradical formed by the initial cleavage may recyclise without extrusion of carbon monoxide and such a process is probably involved in the epimerisa- tion of androsterone (3) to lumiandrosterone (4),17 and of other related transformations.l* Such a process may also take place in the gas phase but no suitably substituted carbonyl compound appears to have been irradiated under these conditions.lo Mulliken J. Chem. Phys. 1935 3 564; McMurray ibid. 1941 9 231; see also Sidman Chem. Rev. 1958,58 689. l1 Kasha Discuss. Faraday SOC. 1950,9 14. l2 Inter al. Hammond and Moore J. Amer. Chem. SOC. 1959 81 6334; Hammond Leermalkers and Turro ibid. 1961 83 2395 2396; Cundall and Milne ibid. 1961,83 3902. l3 See the discussion of photo-reactions of anthracene by Hochstrasser and Porter Quart. Rev. 1960 14 146. l4 Davis Chem. Rev. 1947 40 201. l5 Benson and Kistiakowsky J. Amer. Chem. SOC. 1942 64 80; Blacet and Miller ibid. 1957 79 4327. l6 Cremer and Srinivasan Tetrahedron Letters 1960 No. 21 24. l7 Butenandt and Poschmann Ber. 1944,77 394.l8 Inter al. Butenandt et al. Beu. 1941 74 1308; 1942 75 1931; 1944 77 392; Barton Campos-Neves and Scott J. 1957,2698; Bots Rec. Trav. chim. 1958,77,1010. DE MAY0 AND REID PHOTOCHEMICAL TRANSFORMATIONS ETC. 395 A second mode of reaction of the carbonyl-alkyl diradical is exempli- fied by the conversion of cyclohexanone into hex-5-ena1.19~20~21 It has also been observed in the irradiation of camphor (5) which gives as the main product campholenaldehyde (6).22923 It has been suggested that the ketone (7) is another of the products but decisive evidence for this structure is lacking.23 The mechanism of this transformation also has not been rigidly established although it has been reasonably proposed that direct transfer of a hydrogen atom from a /%carbon atom to the carbonyl group takes place possibly before diss~ciation.~~ A third mode of reaction of the same diradical has been observed.Irradiation of cyclic ketones in reactive solvents (water alcohol aqueous acetic acid) results in solvolysis of the ring. This reaction first reported by Ciamician and Silber,l 9 9 2 5 has recently found application in terpenoid chemistry. It is exemplified by the conversion of lanostanone (8) into the acid (9),26 and the partial synthesis of dihydronyctanthic acid,26 and has been applied in the elucidation of the structures of dammarenolic acid26 and ~aleranone.~’ In all cases the cleavage takes place to give the expected more stable alkyl radical. Recent investigation of the mechanism of this l9 Ciamician and Silber Ber. 1908,41,1071; 1909,42,1510; 1913,46,3077. 2o Kharasch Kuderna and Nudenberg J.Org. Chew. 1953 18 1225. 21 Srinivasan J. Amer. Chem. SOC. 1959,81 5541. 22 Ciamician and Silber Ber. 1910 43 1340. 23 Srinivasan J. Amer. Chem. SOC. 1959 81 2604. 24 Srinivasan J. Amer. Chem. SOC. 1959 81 1546. 25 Ciamician and Silber Ber. 1907 40 2415. 26 Arigoni Barton Bernasconi Djerassi Mills and Wolff J. 1960 1900. 2i Krepinsky Romanuk Herout and Sorm Tetrahedron Letters 1960 No. 7 9. 396 QUARTERLY REVIEWS reaction has indicated that the diradical formed initially undergoes an intramolecular hydrogen transfer to give the corresponding keten which then reacts with the solvent to give the final product.28 The transformations just described therefore? all represent alternate modes of decomposition of the same diradical and such a scheme is illustrated for the case of Fig.I . cyclohexanone in Fig. 1. (b) Reactions involving y-Hydrogen Atoms.-Irradiation of carboiiyl compounds in the presence of substances capable of giving stable radicals by hydrogen abstraction (e.g. diphen~lrnethane,~~ cycl~hexene,~~ ace- naphthene31) is known to lead to the formation of alcohols. These are formed by addition of the hydrocarbon radical to the carbonyl compound (or a partially reduced intermediate derived from it).30 Aliphatic ketones undergo the same reaction intramolecularly if a hydrogen atom is available on the y-carbon atom pentan-2-one (lo) for example gives the cyclobutanol (12) presumably through the diradical (1 l).32 Disposition of the electrons in an alternative sense results in decomposition of the intermediate (11) to acetone and ethylene.33 In the irradiation of ketones these two reactions in general accompany each other but the corresponding cyclisation of aldehydes has not been reported.The formation of the CH CH (1 3) 28 Quinkert personal communication. 23 Paternb and Chieffi Gazzetta 1909,39,415. 30 Mayo Stothers and Templeton Canad. J. Chem. 1961,39,488. 31 Mayo and Stoessl unpublished work. 32 Yang and Yang J. Amer. Chem. SOC. 1958,80,2913. 83 Davis and Noyes J. Amer. Chem. SOC. 1947 69 2153; Manning ibid. 1957 79 5151; Brunet and Noyes Bull. SOC. chim. France 1958,121; Srinivasan J. Amer. Chem. Soc. 1959 81 5061 ; see however Pitts J. Chem. Educ. 1957,34 112. DE MAY0 AND REID PHOTOCHEMICAL TRANSFORMATIONS ETc. 397 alcohols has been depicted as a four-centre ~earrangernent,~~ as in (1 3) but later work35 has shown that the reaction is not stereospecific as was originally and that for instance 3,3-dimethoxypregnan-20-one (14) gives two isomers of the alcohol (15).Several steroids have been found to undergo this reaction together with the concomitant decomposi- tion leading in this case to the products (16) and (17) and acetone. The expected cyclobutanol(l9) is formed on irradiation in ethanol of a steroid 20-ketone containing a 21-acetoxyl group (as in 18). In addition however the acetal(20) is CH -0Ac G Other proximity reactions involving intramolecular hydrogen abstraction by a radical formed by photolysis have been used in synthesis. Irradiation of the nitrite group3' results in cleavage followed by an intermolecular addition of the NO radical to form an oxime.38,39 In this way a three-step synthesis of aldosterone 21-acetate (22) from corticosterone acetate (21) has been achieved.39 However the reaction takes a different course in the case of 17P-nitrites and leads to hydroxamic acidsY4O presumably by 34 Buchschacher Cereghetti Wehrli Schaffner and Jeger Helv.Chim. Acta 1959 35 Yang and Yang Tetrahedron Letters 1960 No. 4 10. 36 Wehrli Cereghetti Schaffner and Jeger Helv. Chim. Acta 1960,43,367. 37 The photolysis of t-butyl nitrite to give acetone and nitrosomethane has been reported; Coe and Doumani J. Amer. Chem. Soc. 1948,70,1516. 38 Inter al. Naylor and Anderson J. Org. Chem. 1953 18 115. 39 Barton Beaton Celler and Pechet J. Amer. Chem. SOC. 1960 82 2640; Nuss- baum CarIon Oliveto Townley Kabasakalian and Barton ibid. 1960 82 2973; Barton and Beaton ibid.1961,83,750; Barton arid Beaton ibid. 1960,82,2641. 40 Robinson Gnoj Mitchell Wayne Townley Kabasakalian Oliveto and Barton J. Amer. Chem. SOC. 1961 83 1771. 42 2122. 398 QUARTERLY REVIEWS CO.CH,OAc 0 &- carbon-carbon bond cleavage of the radical formed by photolysis of the nitrite. This is illustrated below for the 17P-nitrite of 3a-acetoxy-SOL-andro- stan-17/3-01 (23). The existence of a C1,-radical as intermediate is sup- ported by the fact that 13-~iso-compounds are also formed. Related Q-NO transformations have been achieved with hyp~halites.~~ Azides such as n-butyl and n-octyl a i d e undergo photochemical transformation into pyrrolidines and this has been used in a synthesis of cones~ine.~~~ The Hofmann-Loffler-Freytag reaction as applied for instance to the synthesis of the pyrrolidine ring in dihydroc~nessine~~~ can also be carried out photo- chemically.A proximity effect of a different kind has been observed in the conversion of cyclodecanone (24) into cis-9-decalol (25).43 Hex-5-en-Zone (26) *l Akhtar and Barton J. Amer. Chem. SOC. 1961 83 2213; Walling “Free Radicals 4a (a) Barton and Morgan Proc. Chem. SOC. 1961 206; (b) Corey and Hertler J. 43 Barnard and Yang Proc. Chem. SOC. 1958 302. in Solution,” Wiley New York 1957 p. 386. Amer. Chem. SOC. 1959 81 5209. DE MAY0 AND REID PHOTOCHEMICAL TRANSFORMATIONS ETC. 399 however is converted into the bicyclic ether (27) by photochemical addition,44u the intermolecular equivalent of which has been known for some time.44b 3. Conjugated Monocyclic Systems A number of photochemical transformations take place at least formally by electron redistribution within the centres constituting an unsaturated ring system.These reactions may be divided somewhat arbitrarily on the basis of the size of the ring and the nature of the n- electron system involved. For the most part only systems containing extended conjugation have been investigated. The small but important group of cross-conjugated transformations will therefore be considered separately (p. 410). (a) Even-numbered Rings.-(i) Six-membered rings. This group con- sists of substances containing the system indicated in (28). On irradiation these are converted in general and as pointed out by Barton,45 into extended systems such as (29) or substances derived from (29) by chemical reaction with the medium.Exceptions to this behaviour appear to be of two kinds. The first is dimerisation and so does not come within the scope of this Review. It occurs generally when one of the ethylenic linkages forms part of a fused aromatic ringY4 but has recently been reported to occur with a-pyridone (30). The product does not have the cyclobutane structure originally and recent work in these laboratories and in those of the University of Alberta has established structure (31) as correct. 44 (a) Srinivasan J. Amer. Chem. Suc. 1960,82,775; (6) Paternb and Chieffi Gazzetta 46 Barton Helv. Chim. A m 1959 42 2604. 46 Taylor and Paudler Tetrahedron Letters 1960 No. 25 1. 1909,39 341. 400 QUARTERLY REVIEWS although the available evidence does not exclude one formed by 1,2:1,4- or 1,4 1,4-addition.It is very probable that at least in monocyclic systems the predominance of cleavage or of dimerisation depends on the concentra- tion of the irradiated solution. The second reaction which may intervene is carbon bridging (p. 414). The conversion of carbocyclic systems containing the grouping (28) into (29) has been observed with simple hydrocarbons such as cyclo- h e ~ a d i e n e ~ ~ s ~ ~ and a-~hellandrene,~~ but undoubtedly the most studied reaction of this type is that occurring in the irradiation of ergosterol (32) to give the vitamin D precursor pre-ergocalciferol(33). The latter is further transformed under the influence of light and of heat the scheme presented in Fig. 2 being that of Havinga and his group.48 The subject has been reviewed el~ewhere.~*~~~ The conversion of lumisterol (35) into pre- ergocalciferol(33) is a parallel transformation and equivalent transforma- tions have been carried out in other steroid systems including those lacking the 10-methyl An analogous transformation of the triterpenoid Fig.2 &*I7 10 HO ‘ Ergosterol (32) c- Heat HO”“ 8” hv L c- F;le Pre-ergoca lci ferol (33) Vitamin 0 (36) Tachysterol (34) HO Lumisterol (35) 47 deKock Minnaard and Havinga Rec.Truv. chim. 1960 79 922. 48 Havinga deKock and Rappoldt Tetrahedron 1960 11 276. 4s Fieser and Fieser “The Steroids,” Reinhold Publ. Corp. New York 1959 p. 146; Inhoffen and Irmscher Fortschr. Chem. Urg. Naturstofe 1959 17 70; Lythgoe Proc. Chem. SOC. 1959 141. Velluz and Amiard Bull. SOC. chim. France 1955,205; Velluz Goffinet and Amiard Tetrahedron 1958,4,241.DE MAYO AND REID PHOTOCHEMICAL TRANSFORMATIONS ETC. 401 methyl dehydroursolate (37) to compounds (38) and (39) has been re- ported.51 Investigation of the mechanism of cleavage has at present revealed no evidence for a triplet state (phosphorescence) on the irradiation of ergosterol in isopentane-ethanol at 80"~. Fluorescence was observed however suggesting that collapse to the triene may take place from a singlet state.48 The ergosterol (and dehydrocholesterol) system appears to be the only re- ported example of the cyclisation of a triene.* In the case of the ergosterol series the same stationary state is reached by prolonged irradiation of each of the isomers involved but it is not clear why (33) gives (32) whereas (34) is believed to give (35). A further stereochemical specificity which is not mechanistically clear is revealed in the irradiation of pyrocalciferol (40) and isopyrocalciferol(4 1).These substances differ from ergosterol only in the stereochemistry of the ring fusions but irradiation in these cases leads (as is general in seven-membered rings where no cleavage reaction is available) to bond formation as in (42).52 This has been interpreted on the assumption that collapse to the cyclic valency tautomer from the initial excited singlet is preferred. The anti-isomers (32) and (35) take in this (40) (4 0 (42) argument the abnormal course because of the increased energy of the ring fusions whilst (40) and (41) are not so inhibited.52 This however does not explain why a simple substance such as cyclohexadiene does not form the bridged tautomer.The explanation may well be a subtle one since the cleavage reaction is a photoequilibrium whilst the formation of the bicyclo- hexane system is not since no absorbing chromophore remains. The observed photochemical epimerisation of isodehydrocholesterol (43) to coprosta-6,8-dienol (44)63 provides further evidence for the existence of * The formation of cyclohexadiene by the irradiation of cis trans-l,3,5-hexatriene has now been reported (Srinivasan J. Amer. Chem. SOC. 1961 83,2806). 61 Autrey Barton and Reusch Proc. Chem. SOC. 1959 55. 52 Dauben and Fonken J. Amer. Chem. Soc. 1959,81,4060. 63 Windaus Linsert and Eckhardt Annaleiz 1938,534,22; Windaus and ZiihlsdorfF ibid. 1938,536 204. 402 QUARTERLY REVIEWS such an equilibrium as it can easily be envisaged as proceeding via the cyclodecatriene.In systems such as that in (44) in which the equilibrium is displaced towards the diene (presumably because of medium-ring forma- tion) it would be of interest to seek for carbon-bridging by prolonged irradiation Replacement of X or Y in structure (28) by a trigonal atom does not affect the general course of the reaction. 6-Substituted cyclohexa-3,4- dienones (45) are converted in the first place into ketens (46) which then react with a base which may be water an amine e t ~ . ~ ~ The product (47) obtained may be further transformed by cis-trans-isomerism (or such isomerism may take place in the keten itself) and in addition the amides and acids obtained may be cyp- or ,&+unsaturated. This may be because of initial protonation at either the a- or the y-position in different substances because of isomerisation under the influence of the stronger bases used or because of shift of the double bond under the influence of light as recently reported.55 The presence of substituents in the remaining position next to the carbonyl group inhibits the cleavage so that compound (48) does not undergo ring fission even in the presence of aniline.This has been attri- buted to steric hindrance to the formation of the intermediate trans-keten (51). Instead a slow aromatisation takes place to give a phenol (50) and the scheme shown has been proposed.54 In contrast the irradiation of 6-acetoxy-6-methyl-2,4-dienone (45; R1 = AcO R2 = Me) in anhydrous ether gives slowly o-cresol with acetoxyl eliminati~n.~~ A modified rationalisation of these aromatisations would be that the intermediate (49) is formed directly from the cis-keten (as in 52).This avoids the necessity of invoking a photochemically produced intermediate with charge separation and in addition rationalises the different behaviour of the ketone (45; R1 = AcO R2 = Me) since the keten derived from the latter will exist predominantly in the trans-form to which the process indicated in (52) cannot apply. Aromaticity in this case is then achieved by the alternative homolysis of the acetoxyl group. 64 Barton and Quinkert J. 1960 1. bb Levina Kostin and Gembitskii J. Gen. Chem. (U.S.S.R.) 1959 29 2421. DE MAYO AND REID PHOTOCHEMICAL TRANSFORMATIONS ETC. 403 OH Me (50 Me An example in which one of the ethylenic linkages in the cyclohexa- dienone (45) is present in an enol has been recorded in the racemisation of usnic acid (53).54 A final example of interest is the preparation of the cyclo- octatetraene derivative (55) by addition of dimethyl acetylenedicarboxylate to benzene.It is presumably formed through the adduct (54) which then rearranges. 56 CO,Me (54) (55) Substances of type (28) where X is a hetero-atom behave similarly. The pyran-ionone equilibrium (56),57 the methanolysis of the a-pyrone (57),68 and the rearrangement of certain spiropyrans (e.g. 58) to give coloured open-chain substances (59)59 are representative examples. Both X and Y may be hetero-atoms in (28). Unsaturated sultones on irradiation in methanol or ether-benzylamine give the expected sulphonic acid deriva- tives this is exemplified by the conversion of (60) into (62).60 By analogy s6 Grovenstein and Rao Tetrahedron Letters 1961 No.4 148. 57 Biichi and Yang J . Amer. Chem. SOC. 1957 79 2318. 58 Esterle Girling Mayo and Wiley unpublished observations. 59 Heiligman-Rim Hirshberg and Fischer J. 1961 156. 6o Henmo Mayo Sattar and Stoessl Proc. Chem. Soc. 1961 238. 404 QUARTERLY REVIEWS (58) (cis or trans) (59) with the work just described this presupposes the intermediacy of a “sulphen” (61). Although certain of these diene systems are further extended by conjugation with a carbonyl group it seems probable that a similar photochemical process is involved-one in which the non-bonded electrons on the carbonyl-oxygen atom are not concerned. The partial conversion of benzene into fulvene and the analogous transformations of toluene isopropylbenzene and anisole,61 are the only examples known of photochemically induced bond cleavage in a 6- membered aromatic ring.A photochemically induced Fries rearrangement of catechol monoacetate (63) to (64) and (65) has been claimedG2 to be intramolecular. (ii) Other even-membered rings. Although cyclobutene is the four-centre analogue of the six-centre systems discussed in the previous section it does not absorb light above 220 mp. However if an absorbing system is introduced similar behaviour is observed. Again no satisfactory evidence is available as to whether the keten (or its equivalent) is formed directly from a singlet or through other intermediate species. As example may be quoted the conversion of the cyclobutenone (66) into an acid (67)*5 61 Angus Blair and Bryce-Smith J.1960 2003. 62 Anderson and Reese Proc. Chem. SOC. 1960 21 7. DE MAYO AND REID PHOTOCHEMICAL TRANSFORMATIONS ETC. 405 and of the butenedione (68) into the ester (69).s3 The six-centre equivalent to the latter transformation does not appear to have been reported yet such o-quinones readily undergoing addition with unsaturated system6 or alternatively reduction to the quin01.~~ Only one example appears to have been recorded of an eight-centre rearrangement. This is the transformation of a bicyclic compound (70) in a six-centre rearrangement to the cyclo-octatriene (71) which is further transformed into the open-chain tetraene (72)45 that was however only characterised spectroscopically. (70) (70 (72) (b) Seven-membered Rings.-No rearrangement product from an un- saturated five-membered ring has been reported.45 The rearrangement of seven-membered rings has in contrast been observed in many instances both with two and with three units of unsaturation.(i) Rings containing two endocyclic double bonds. Cyclohepta- 1,3-diene itself (73; R = R’ = H) certain of its derivatives (73; R = OH R’ = H or OMe) and the ketone (79 are converted on irradiation into substituted bicyclo [3,2,0]heptanes (74) and (76).s53as Eucarvone (77) on irradiation in alcohol or acetic acid gives a similar bicyclo-compound (78),s736s and this n 0 OMe is further transformed by irradiation in alcohol in a photoequilibrium into an isomer This further transformation presumably proceeds by cleavage between the a-carbon atom and the carbonyl group (p. 394) followed by bond formation at the alternative position in the allylic 63 Mallory and Roberts J.Amer. Chem. Soc. 1961,83 393 (footnote 6). 64 Ciamician and Silber Ber. 1901 34 1530. 65 Chapman and Pasto Chem. and Ind. 1961 53; Rigaudy and Courtot Tetrahedron 66 Dauben and Cargill Tetrahedron 1961,12 186. 67 Buchi and Burgess J. Amer. Chem. SOC. 1960 82,4333. 6B Hurst and Whitham Proc. Chem. SOC. 1961 116. Letters 1961 No. 3 95. 406 QUARTERLY REVIEWS radical (79). Irradiation in acetic acid gives as a second product a bicyclo [2,2,1 Iheptanone (8 l) which is the result of a more complex re- arrangement:* perhaps through the route indicated although the specific function if any of the solvent is not clear. (ii) Rings containing three endocyclic double bonds. Cycloheptatriene66 and certain of its oxygenated derivative^^^ behave analogously to seven- membered dienes.Amongst these derivatives are y-tropolone methyl ether (82)69 and colchicine (83).'O The latter gives three products two of which a- and ,&lumicolchicine are stereoisomers (84). Such evidence as was recently presented for (85) as the structure of a-lumicolchicine is not d OMe yet ~ompelling,~~ but it appears that interconversions are possible.* The irradiation of a-tropolone (86; R = R' = R" between the isomers = H) itself has been more closely studied72 and the situation found more intricate than had * Recent work by Chapman and Smith (J. Amer. Chem. SOC. 1961 83 3914) has established that a-lumicolchicine is a dimer of /3-lumicolchicine. 69 Chapman and Pasto J. Amer. Chem. SOC. 1958,80 6685 70 Forbes J. 1955 3864; Gardner Brandon and Haynes J.Amer. Chem. SOC. 1957,79 6334. 71 Schenck Kuhn and Neumuller Tetrahedron Letters 1961 No. 1 12. 72 Dauben Koch and Thiessen J. Amer. Chem. SOC. 1960,82 6087; Dauben Koch Chapman and Smith ibid. 1961 83 1768. DE MAYO AND REID PHOTOCHEMICAL TRANSFORMATIONS ETC. 407 at first been thought. Irradiation of the tropolone in methanol gave the ester (89; R = Me) and irradiation of the methyl ether in water gave the same substance. However irradiation of the methyl ether in methanol gave in the first place a bicyclo[3,2,0]heptadienone (87; R = Me R’ = R’ = H) which on further irradiation disappeared with the concomitant formation of an isomer (88; R = OMe R’ = R” = H). On addition of water the latter was then transformed into the monocyclic ester (89; R = Me).The final stages appear mechanistically reasonable particularly in the case of the tropolone itself. Here the intermediate corresponding to (88; R = R = R” = H) would be the ketone (90) and this might be expected to afford the keten (p. 396) which would then react with the solvent. The detailed mode of conversion of (87) into (88) is as yet unclear but insight has been gained by the fact that the tropolones (86; R = Me R’ = Me R” = H) and (86; R = Me R’ = H R” = Me) give first (87) and then (88) (R’s as before). A six-centre rearrangement with formation of the norcaradiene system has apparently not been observed among cycloheptatriene~,~~,~~ perhaps because of its ready reversibility. Such a rearrangement may be involved in the conversion of the benzotropolone methyl ether (91) into the naphthoic ester (92).75 4.A Heteroanndar Olefink System The only heteroannular diene whose irradiation has been reported so far is that contained in the cholestane derivative (93). In ethanol this gave the 3,5-cyclo-ether (94),7s and the stereochemistry of the cyclopropane ring is the reverse of that formed in solvolysis of cholesterol derivatives. It may be envisaged as formed formally from a species* such as (95). * No implication is intended as to the parity of the electrons designated. 73 Forbes and Ripley Chem. and Ind. 1960,589. 74 For a discussion of the norcaradiene problem see Doering and Knox J. Amer. 75 Forbes and Ripley J. 1959 2770. Chem. SOC. 1957,79 352. Dauben and Ross J. Amer. Chem. SOC. 1959 81 6521. 408 QUARTERLY KEVIEWS Suprasterol-11 for which the structure (96) has been is obtained by overirradiation of calciferol(36); its genesis may involve a vinylogously equivalent process.Dehydroergosterol (97; R = H) contains the same chromophore as does calciferol with the modification that the cisoid H 00.' portion now forms part of a six-membered ring. The product (98) of irradiation of the acetate (97; R = Ac) is derived by cleavage of the 9,lO-bond and bond redi~tribution.~~ However cleavage of the bond to C9H17 & \ -&I Ac -ace..*< 1:: ;@ RO (97) (99 give a diradical is inadmissable as an initiating step since the reaction has been shown to be stereospecific. Dehydrolumiergosterol acetate (97; a-Me at C(lo) R = Ac) on irradiation gives a product differing from (98) in the configuration of the methyl group; and in both photochemical transformations the configuration is inverted.Bond migration and bond cleavage must therefore be concurrent,79 but the nature of the intermediate species if any is conjectural. 5. Intramolecular Addition Reactions Whilst there is a very large number of examples of substances which contain ethylenic linkages which by forming part of an absorbing con- jugated system can be induced either to dirneri~e*~*~ or to add to a suitable 77 Dauben Bell Hutton Laws Rheiner and Urscheler J. Amer. Chern. SOC. 1958 80 41 16. 78 Barton and Kende J. 1958,688. *O Cookson and Hudec Proc. Chem. SOC. 1959 11. Barton Bernasconi and Klein J. 1960 511. DE MAYO AND REID PHOTOCHEMICAL TRANSFORMATIONS ETC. 409 unsaturated ~ u b s t r a t e ~ ~ ~ ~ l on irradiation similar intramolecular reactions are far more restricted in scope.The requirements are of course that at least one of the n-electron systems should be activated by the light used and that the two systems should be able to come close together. Confor- mational rigidity is not required. Irradiation of bicyclo [2,2,1 ]hepta-2,5-diene-2,3-dicarboxylic acid (99) in ether gives a good yield of the quadricycloheptane isomer .. ...... (99) &:2 - &:IH (roo) The double bonds are sufficiently close for homoallylic interaction to be unexceptional the interaction is reflected in the abnormal ultraviolet spectrum of the acid (99) (Amax. 243 mp) in comparison with that of maleic acid (Amax. 210 The adduct (101) of cyclopentadiene andp-benzo- quinone undergoes similar internal addition to give a “near cage” molecule (lO2).** A single ethylenic linkage when substituted by chlorine has sufficient absorbance (€198 9OOO) for activation and the insecticide “iso- drin” (103) is also converted into a cage isomer (104).85 On the other hand carvone has considerable conformational flexibility the most suitable C‘ conformation for internal addition being that represented in (105).However since even in this arrangement bonding distance appears too great it may be that the particular excited species involved has greater flexibility for on irradiation carvone camphor (106) is obtainedg6- probably the earliest example of this type of reaction. Verbenone (107) having a chromophore similar to that of carvone but lacking the isolated Inter al. Ayer and Buchi U.S.P. 2,805,242; Angus and Bryce-Smith J.1960 4791 ; Bryce-Smith and Vickery Chem. and Ind. 1961,429. 82 Cristol and Snell J. Amer. Chem. SOC. 1958 80 1950. 8s Ley and Wingchen Ber. 1934 67 501. 84 Cookson Crundwell and Hudec Chem. and Ind. 1958 1003; see also Yates and 85 Cookson and Crundel Chem. and Ind. 1958 1004. 86 Ciamician and Silber Ber. 1908 41 1928; Sernagiotto Guzzetta 1917 47 153 ; Eaton Tetrahedron 1961,12,13. 1918,48,52; Buchi and Goldman J. Amer. Chem. Soc. 1957,79,4741. 410 QUARTERLY REVIEWS ethylene linkage rearranges on irradiation to chrysan thenone ( 108).87 This transformation is reminiscent of the racemisation of a-pinene under thermal conditions,88 a process which presumably proceeds through a monocyclic diradical. A similar discrete species may be involved in the conversion (107)+( 108).6. Cross-conjugated Dienones Of all the cross-conjugated systems available only that expressed in the cipher (109; R = R’ = H or Me) has so far been examined. The system RD R‘ (109) has maximal absorption (above 200 mp) at about 240 mp so it seems that since many of the transformations to be discussed occur in Pyrex vessels absorption in the low-intensity band at about 320 mp (n+sr*) is responsible for the transformation. These investigations were conducted with the terpenoid santonin (1 and the transformations found were amongst the earliest photo- chemical rearrangements discovered more recent work has extended the study to steroid derivatives. The irradiation of santonin (1 10) in aqueous acetic acid gives isophoto- santonic lactone (1 1 1)90 and photosantonic acid (1 13).91392 Irradiation in ethanol or dioxan or in water containing one equivalent of alkali gives a substance lumisantonin unknown to the original workers.89 This has been Hurst and Whitham J.1960 2864. Fuguitt and Hawkins J. Amer. Chem. Soc. 1947 69 319. 8g Simonsen and Barton “The Terpenes,” Cambridge Univ. Press 1952 Vol. 111 Barton Mayo and Shafiq J. 1957,929. 91 van Tamelen Levin Brenner Wolinsky and Aldrich J. Arner. Chern. SOC. 1958 ga Barton Mayo and Shafiq J. 1958,3314. p. 292. 80 501. DE MAYO AND REID PHOTOCHEMICAL TRANSFORMATIONS ETC. 41 1 shown to have the structureg3 and stereochemistryg* shown in (112). Lumisantonin is converted into isophotosantonic lactone (1 1 1) by the action of hot aqueous acetic acid (in the absence of light) or into an acid (1 13) by irradiation in cold acetic acid.There must presumably be therefore a direct and an indirect route to (1 11) from (1 10). Transformations precisely analogous to the formation of lumisantonin have been achieved by the OH oq co O L Q H02C * (J ‘3) (112) o-co 0-co irradiation of 1,2-dehydr0-4-methyltestosterone (1 14; R = R” = Me R’ = H) and 1,2-dehydr0-2,4-dimethyltestosterone~~ (1 14; R = R’ = R” = Me) in dioxan. A transformation analogous to the formation of isophotosantonic lactone has been carried out on prednisone acetate (115).96 A number of further observations have been made. The transformation of santonin (1 10) into the lactone (1 11) occurs independ- ently of the configurations at position 6 and 11 .97 Further whereas the isophotosantonic lactone type of rearrangement (in acetic acid) is not apparently affected to any major degree by substitution at position 4 the conversion of the lumisantonin type is irradiation of the compound (1 14; R = R = H R” = Me) in dioxan gave at least eight products.Replace- ment of the 10-methyl group (R”) by acetoxyl led to loss of the acetyl group and formation of the phenol (1 16).98 93 Barton Mayo and Shafiq J. 1958 140; Arigoni Bosshard Bruderer Biichi Jeger and Krebaum Helv. Chim. Acta 1957 40 1732; see also Cocker Crowley Edward McMurry and Stuart J. 1957,3416. 94 Barton and Gilham J. 1960,4596. 96 Weinberg Utzinger Arigoni and Jeger Helv. Chim. Acta 1960 43 236. g6 Barton and Taylor J. 1958 2500. g7 Barton Proc Chem. SOC. 1958 61. g8 Warszawski Schaffner and Jeger Helv. Chem. Acta 1960 43 501. 412 QUARTERLY REVIEWS A simple general rationalisation for these changes would be that irradia- tion of santonin produces an excited species which can collapse to lumisan- tonin or to isophotosantonic lactone depending on the conditions.In both reactions a bond between Co) and C(6) is formed but to form lumin- santonin (112) the electrons constituting the 1,lO-bond move to form a bond between C(4) and C(lo) whilst to form the lactone (1 11) the 5,lO-bond must be cleaved. Such a species may be formally represented as (1 17) though there is as yet no evidence as to nature or number of the intermedi- ate species involved. In keeping with this the transformation to lumisan- tonin (1 12) is stereospecific and involves inversion of the angular methyl A similar species may be involved in the rearrangement of dehy- droergosterol(97; R = H).79 Collapse of the intermediate (1 17) under the influence of acetic acid may then give isophotosantonic lactone (1 1 1).Approach of water to Coo) from the or-side with cleavage of the $10-bond leads to the stereochemistry implied by (118). This differs from the ~ t e r e o c h e m i ~ t r y ~ ~ ~ ~ ~ previously suggested only at C(1),99 and places this hydrogen and the hydroxyl group cis and a. On other grounds it has been suggested that these should be cis.1oo but with the interpretation that the hydroxyl group should be inverted. In the absence of the 4-methyl group irradiation in dioxan is more com- plex. From 1,2-dehydrotestosterone acetate (114; R = R‘ = H R” = Me) as mentioned eight substances have been isolated and structures (119; R = H and 120-124) have been attributed to six of these.101J02 One (121) is of the lumisantonin type.Whilst it is not yet clear what the 99 Djerassi Osiecki and Hen J. Org. Chem. 1957 22 1361. loo Huffman Experientia 1960 16 120. lol Dutler Bosshard and Jeger Helv. Chiin. Acta 1957 40 494. Io2 Utzinger Dutler Weinberg Arigoni and Jeger Angew. Chem. 1959 71 80. DE MAYO AND REID PHOTOCHEMICAL TRANSFORMATIONS ETC. 413 nature of the intermediates is fission of the 9,lO-bond seems necessary. The situation is rendered clearer by consideration of the products of the irradiation of prednisone acetate (115) in ethanol. In this substance the 9,lO-bond is further weakened by the presence of the 11-keto-group. The formation of the main product lumiprednisone acetate (125) was ration- alised as indicated below.96 In addition neoprednisone acetate (126) was formed under prolonged irradiation presumably by alternate cleavage of (125) to the less stable radical.Irradiation in dioxan gives a phenol (127) equivalent to (124) (or its inverted isomer 128). If it be assumed that a substance equivalent to (126) is formed in the irradiation of the dehydro- I androsterone derivative (1 14; R = R = H R = Me) then the ketone (120) may be formed from it by a “lumisantonin” rearrangement. The genesis of compounds (121) and (124) then needs no further comment. The spiran (1 19 ; R = H) may be formed from (120) or (121) as indicated in (1 29),* and the conversion of (1 19) to (122) and (123) finds analogy in the conversion (1 30) + (1 3 1) reported by S t a ~ d i n g e r . ~ ~ ~ No methyl migrations are involved in any of these transformations and 1,Zshifts may be con- ceived as occurring by dissociation and addition.lo5 * Umbellulone containing the same chromophore as (1 12) is quantitatively converted into thymol on irradiation,lo3 whereas (1 12) is not aromatised easily presumably because of the blocking groups.loS Wheeler and Eastman J. Amer. Chem. SOC. 1959 81 236. lo* Staudinger and St. Bereza Annalen 191 1,380 243. lo5 Berson Olsen and Walia J. Amer. Chem. Soc. 1960 82 5O00. 414 QUARTERLY REVIEWS The differing behaviour in the absence and presence of the 4-methyl group may be the result of quite small energy differences leading to sig- nificant changes in the photochemical equilibria. Possibly the eclipsed methyl interaction (similar to that in cis-buteneloG) in the spiran (119; R = Me) together with the influence of the 6-hydrogen atom may suffice to suppress formation of this compound and hence of the compounds (122) and (123) derived from it.A “lumisantonin” rearrangement has been reported recently to occur on irradiation of 3/3-acetoxylanosta-5,8-dien-7-one.107 The final rearrangement in this group to be discussed is the conversion of lumisantonin (1 12) into photosantonic acid (1 13). Ketonic cleavage of the former as in the case of lanostanone (8) would give a diradical (132) which could be converted by electron redistribution into a keten-carbene (133); rearrangement of this and reaction with the solvent (ethanol or aqueous acetic acid) would then give the lactone (1 1 3).91 (112) - - ‘a} (113) 7. Oxygen-transfer Rearrangements A number of rearrangements are known in which a transfer of an oxygen atom originally attached to nitrogen takes place (a) o-Nitrobenzaldehyde rearrangement.-This group of reactions is the largest of this class and the most studied.The earliest example was the conversion of o-nitrobenzaldehyde into o-nitrosobenzoic acid observed by Ciamician and Silber in 1901.108 The reaction appears general for ortho- substituted nitrobenzene derivatives having an a-hydrogen atom and typical examples are the conversions (134) -+ (135) and (the mechanism being assumed by analogy) (136) -+ (137).lo9 (136) (137) lo6 Turner “Theoretical Organic Chemistry,” Kekuld Symposium Butterworths lo7 Barton McGhie and Rosenberger J. 1961 1215. lo8 Ciamician and Silber Ber. 1901 34 2040. loo Tanasescu Bull.SOC. chim. France 1926,1718; Berson and Brown J. Amer. Chem. Scientific Publn. London 1959 p. 75. SOC. 1955 77 447. DE MAYO AND REID PHOTOCHEMICAL TRANSFORMATIONS ETC. 4 15 Since at least in the case of o-nitrobemaldehyde itself the transforma- tion will take place in the crystal this rearrangement should be intra- molecular. In a number of studies,110 however not much has been found to clarify the problem. It seems nevertheless very probable that the first step is the abstraction of the benzyl hydrogen by the photoactivated nitro- group to give a species which may be represented as (138). Electron redistribution then leads to the keten (139) from which o-nitrosobenzoic acid is obtainable as indicated in (140). Support for this view is available in that ortho-substituted benzophenones have been shownlll to be enolised under similar conditions presumably by an equivalent process.* The latter represents a particular case of the general phenomenon of carbonyl y-hydrogen abstraction already referred to. In the presence of alcohol (as solvent) the corresponding ester is obtained by intermolecular reaction with the keten and a similar process can be written for the other related transformations. In particular the conversion (141) -+ ( 142)112 would be envisaged as proceeding through the keten imine whilst the discrete 0 0 (140) intermediate indicated in the conversion of (136) into (137) becomes unnecessary. It is interesting that an apparently similar transformation has been reported for the arsenic-containing compound (143) which is converted into the arsonic acid ( 144).l13 (143) (144) * The cyclopropane by-product obtained in the irradiation of /3-ionone5' has been shown in these laboratories to be in fact the alternative conjugated diene formed by similar hydrogen transfer.Tanasescu Bull. Soc. chim. France 1926,1443; Leighton and Lucy J. Chem. Phys. 1934,2 756,760. ll1 Yang and Rivas J. Amer. Chem. SOC., 1961 83 2213. 112 Sachs and Kempf Ber. 1902 35 2704. 113 Karrer Ber. 1914 47 1783. 416 QUARTERLY REVIEWS Irradiation of o-nitrostilbenes (145; R = NMe, OH or OMe) if following the course indicated for other o-nitrobenzene derivatives might be expected to give the ketones (146) through intermediate allenes. The products isolated the isatogens (147),114 may be derived from such ketones by cyclisation and oxidation. No such oxidation is required in the photo- chemical conversion (148) + (149),115 and this seems to require a modified mechanism.02N \ NOz O,N QTJ \ Og - oicw - Y (149) O' (148) The conversion of o-nitrosobenzaldehyde into the azo-compound (1 50) is an example of this transformation at a lower oxidation (b) Nitrone rearrangement.-Irradiation of nitrones (1 5 1) gives by internal addition oxazirans (1 52) which being unstable frequently rearrange spontaneously to arnides.ll7 With certain N-phenylnitrones irradiation in benzene gives the oxaziran whereas irradiation in ethanol gives the amide.114,118 The oxaziran in benzene solution is converted back into the nitrone in the dark within 24 hours. The oxazirans obtained are where they have been compared identical with those prepared by oxida- tion of the Schiff's bases with peracetic acid.ll9 The fused oxaziran (154) prepared by the action of light on 5,5-dimethyl-l-pyrroline 1-oxide (1 53) appears to be remarkably stable,120 although it is thermally rearranged to 5,5-dimethylpyrrolid-2-one on prolonged heating.Other examples of this Pfeiffer and Kramer Ber. 1913 46 3655; see also Pfeiffer Ber. 1912 45 1819; Tanasescu Bull. SOC. chim. France 1927 1074; Krohnke Krohnke and Vogt Chem. Ber. 1953,86 1500. 116 R i d and Willc Annalen 1954,590,91. 117 Kamlet and Kaplan J. Org. Chem. 1957 22 476; Krohnke Annalen 1957 604 203. 118 Splitter and Calvin J. Org. Chem. 1958 23 651. llD Emmons J. Amer. Chem. SOC. 1957,79 5739. 12* Bonnett Clark and Todd J. 1959 2102. 114 Splitter and Calvin J. Org. Chem. 1955 20 1086. DE MAYO AND REID PHOTOCHEMICAL TRANSFORMATIONS ETC.4 17 transformation include the conversion of 9-acridyl-N-phenylmtrone (1 55) into the anilide (156),121 and of the quinoxaline derivative (157) into the hydroxyquinoxaline (1 58),122 whilst the conversion of azoxybenzenes into phenolic azo-compounds12s may be a vinylogous extension of this reaction. 0- I CO-NHPh CH=N+Ph If so the transformation of the bromoazoxy-compound (159) into the phenol (1 60) may be as represented here. We are indebted to Professor R. B. Woodward (Harvard) for stimulating discussions. 121 Chardonnens and Heinrich Helv. Chim. Acta 1949 32 656; Mikhailov and Ter- 122 Landquist J. 1953 2830. 12s Badger and Buttery J. 1954 2243. Sarkisyan Bull. Acad. Sci. U.S.S.R. 1954 559.

ISSN:0009-2681    

DOI:10.1039/QR9611500393

出版商:RSC    

年代:1961

数据来源: RSC

摘要:

NITROSATION DIAZOTISATION AND DEAMINATION By J. H. RIDD (CHEMISTRY DEPARTMENT UNIVERSITY COLLEGE LONDON) THE three reactions linked by the title of this Review can be considered as stages on the common reaction path R-NH + R-NH-NO + R.N:N.OH -+ R*N$ -f R++ Deamination products . . . . . (1 1 With secondary amines the reaction stops at the nitrosamine stage; with aromatic amines the reaction effectively stops at the diazonium ion stage; and with primary aliphatic amines the reaction extends to the formation of a wide variety of deamination products. However although this scheme accounts for most of the observations it appears possible that other mechanisms of deamination may be significant ; these complications are briefly discussed after the main evidence for path (1) has been presented.The total material included in the title is too great for a single review; the following discussion is therefore limited to those observations that throw light on the mechanism of the reactions concerned. Only the simpler aromatic and aliphatic amines are considered to exclude the many complications which arise from special structural features. The discovery of deamination by Pirial in 1846* (in the conversion of aspartic into malic acid) slightly preceded the isolation of the aliphatic amines2 but was soon applied both to them and to the aromatic a~nines.~ The deamination of the aliphatic amines was shown to give not only the corresponding alcohol but a considerable range of other products ; thus the reaction of n-propylamine gave n-propyl alcohol isopropyl alcohol propene and some of the nitroso-derivative of di-n-pr~pylamine.~ The formation of isopropyl alcohol was first explained as due to hydration of the propene but Linnemann showed5 that propene did not react with water under the conditions of deamination.Further work showed that the rearrangements in deamination could include the carbon skeleton as in the formation of t-pentyl alcohol from neopentylaminee and in the inter- conversion of alicylic rings termed the Demjanov rearrangements.‘ The first observed reaction of aromatic amines with nitrous acid was the conversion of aniline into phen01,~ but this was followed by the isolation * The 1846 publication was in a little known Italian journal (ZZ Cimento) but the results were soon made generally available by publication1 elsewhere. R. Piria Ann. Chem.Phys. 1848 22 160; Annalen 1848 68 343; cf. report in J. 1849 1 342. A. Wurtz Annalen 1849 71 330. A. W. Hofmann AnnaZen 1850,75 356. * V. Meyer and F. Forster Ber. 1876 9 535; L. Henry Comp. rend. 1907,145,899. E. Linnemann Ber. 1877 10 11 11. M. Freund and F. Lenze Ber. 1890 23,2865; 1891 24,2150. C. K. Ingold “Structure and Mechanism in Organic Chemistry,” Bell Ltd. London 1953 p. 486. 41 8 RIDD NITROSATION DIAZOTISATION AND DEAMINATION 4 19 by Griess8 of diazotised picramic acid (I) (a neutral diazo-oxide) and a little later by the isolation of true diazonium salts.g Many of the reactions of diazonium ions and the controversies over the structure of diazo- compounds have been well summarised by SaunderslO and by Zollinger,ll and form one of the best known fields of organic chemistry.(I) Before discussing the mechanism of the reactions it is useful to compare the early structural assignments with the results of more recent investiga- tions. One of the early problems concerned the difference between the highly reactive diazonium salts and the less reactive diazo-compounds (both with the empirical formula Ar.N,X). The explanation is now known to lie in the difference between an ionic and a covalent bond but Arrhe- nius’s theory of ionic dissociation was not proposed until 1887 and so the difference in reactivity had then to be expressed as a difference in the valencies of the atoms concerned. Of the several proposed structures for the diazonium salts that (11) first put forward by Blomstrand12 and later supported by Hantzsch13 is most in accord with modern views (cf.I11 for the diazo-compounds). If the N-X bond in structure (11) is written as ionic then this structure becomes equivalent to (IVa) the valence-bond structure frequently used to represent the diazonium ion. Other structures contri- buting to the resonance hybrid include (TVb) and structures with the (IV b) + (IVC) positive charge on the aromatic ring (e.g. IVc). The contribution of structures of type (IVc) increases the order of the C-N bond and thus stabilises the diazonium ion against decomposition by loss of nitrogen; such structures can also be considered to give rise to the strong electron- withdrawal by the -Nz+ group.14 The related structures in aliphatic P. Griess Annalen 1858,106 123; 1860 113,201. P. Griess Annalen 1861 120 125. lo K. H. Saunders “The Aromatic Diazo-Compounds,” Arnold and Co.London l1 H. Zollinger “Chemie der Azofarbstoffe,” Birkhauser Verlag Bade 1958. l2 C. W. Blomstrand “Chemie der Jetztzeit,” C. Winter Heidelberg 1869 p. 272. l3 A. Hantzsch Ber. 1895,28,1743. l4 E. S. Lewis and M. D. Johnson J. Amer. Chem. SOC. 1959,81 2070. 1949. 420 QUARTERLY REVIEWS diazonium ions (e.g. H+CH2= N+= N-) would involve hyperconjugation,* a less effective form of electron release. The two isomeric diazotates formed by diazonium salts in alkaline media were formulated by Hantzschls as geometrical isomers (Va and b) by analogy with the oximes; the cis(syn)-compound is formed rapidly and At-N II KO-N (Val Ar-N II (Vb) N-OK is converted into the trans(anti)-compound on storage. This geometrical isomerisation was not at first accepted but has been supported with reference to the diazotate ions by modern physical s t ~ d i e s .~ ~ ~ ~ ~ In contrast there is still a certain amount of disagreementl8 concerning the formation of the corresponding diazohydroxides (Ar.N:N*OH) and the isomeric primary nitrosamine (Ar-NH-NO). These compounds have not been isolated and recent work suggests that the syn-diazo-hydroxide is never formed to a significant extent in the reaction of diazonium saltsll with a1 kali . By the end of the nineteenth century the products of aliphatic deamina- tion were known in considerable detail and the main structural features of the compounds formed on aromatic diazotisation were at least partly understood. However the factors determining the reaction rate had not been studied and the mechanism of the complex rearrangements in aliphatic deamination were unknown.The main part of this Review is concerned with the more recent work on these two topics. The ratedetermining stage It is difficult to describe the development of the kinetic studies in strict chronological order. The kinetic form of diazotisation depends very much on the conditions but this was not appreciated by the early workers and so several different kinetic forms were reported and considered to be generally valid ; this led to considerable controversy. The initial nitrosation stage [cf. reactions (l)] is now recognised to be rate-determining at low acidities both for diazotisation and for aliphatic deamination ; the kinetic complexity therefore arises from the several mechanisms of nitrosation some of which include two potentially rate-determining steps.In the follow- ing historical survey the earlier papers are assigned according to the predominant nitrosation mechanism to which they refer. Only the more * In Mulliken’s terrninol~gy,~~ this would be sacrificial hyperconjugation and therefore less important than the hyperconjugation in the corresponding carbonium ions (isovalent hyperconjugation). N. Muller and R. S. Mulliken J. Arner. Chem. SOC. 1958 80 3489. E. S. Lewis and H. Suhr J. Arner. Chem. SOC. 1958 80 1367; R. J. W. LeFkvre R. Roper and I. H. Reece J. 1959 4104 cf. B. A. Porai-Koshits Tetrahedron 1960 11 30. l6 A. Hantzsch Ber. 1894 27 1702. l7 R. J. W. LeFhre and J. B. Sousa J. 1955 3154. RIDD NITROSATION DIAZOTISATION AND DEAMINATION 42 1 important papers have been considered; a very detailed account of the early work is available elsewhere.1° The Nitrous Anhydride Mechanism.-Under the conditions of the early kinetic studies this is the predominant mechanism for the diazotisation of amines similar in basicity to aniline in dilute perchloric acid (<O*~M) dilute sulphuric acid ( < O ~ M ) and in very dilute hydrochloric acid (<O.~M); it is also the predominant mechanism for the deamination of aliphatic amines.The evidence for this mechanism has developed gradually from 1899 until the present day. In 1899 Hantzsch and Schiimann19 published the first kinetic results on diazotisation. They showed that the diazotisation of aniline and several similar amines in dilute (-0.002M) hydrochloric acid was overall of the second order and since there were only two reactants they assumed that the order was unity with respect to each and expressed their results by equation (2) * Rate = k[Ar.NH,+] [HNO,].(2) They did not study the detailed variation of the reaction rate with acidity and so they had no direct evidence that the protonated amine was the reacting species but they did find that all their amines reacted at about the same rate. Since the amines were almost completely protonated this equivalence of reaction rates is most simply explained by assuming that the conjugate acids are the reacting species and that these are of equal reac- tivity. Other workers,20 using greater concentrations of hydrochloric acid confirmed the second-order kinetics but found marked differences in the reactivity of different amines.In contrast to equation (2) the deamination of methylamine was shown by Taylorz1 to give third-order kinetics in a form which can be expressed by equation (3) The same kinetic form was obtained for the reaction of ammonia with nitrous acid21 and for the reaction of several other amino-compounds including dimethylamine.z2 The last result suggests that the kinetic form of equation (3) can be generally associated with a rate-determining nitro- sation. Some years later an equivalent kinetic equation was obtained by * In all the kinetic equations in this Review the concentration terms refer to the molecular or ionic concentration of the stated species not to the stoicheiometric con- centration unless [as for the conditions of eqn. (2)] this is effectively equivalent to the former. l* A.Hantzsch and M. Schumann Ber, 1899 32 1691. 2o J. Boeseken W. F. Brandsma and H. A. J. Schoutissen Pruc. Acad. Sci. Amrev- dam 1920,23,249. 21 T. W. J. Taylor J. 1928 1099. 22 T. W. J. Taylor and L. S. Price J. 1929 2052. Rate = k[R.NHz] [HNO2l2. (3) 422 QUARTERLY REVIEWS S ~ h m i d ~ ~ for the diazotisation of aniline in sulphuric acid at acidities (4.2M) somewhat higher than those used by previous workers. Schmid did not refer to the earlier work leading to equation (2). It is not easy to reconcile equation (2) with equation (3) and so the early theories of diazotisation were based on the acceptance of one of these kinetic forms and on the rejection of the other. Several groups of workers considered that the second molecule of nitrous acid implied in equation (3) was in some way “spurious” and that only the kinetic form of equation (2) need be considered.This point of view was put forward by Earl and Hills24 with reference to diazotisation and by Dusenbury and with reference to deamination ; both arguments were illustrated by results which were considered to conflict with equation (3). However later workers have confirmed this kinetic form for both diazotisation26 and deamina- t i ~ n . ~ ’ Two interpretations were based on the third-order equation. Kenner2* suggested a mechanism involving the nitrosation of the free amine followed by a slow proton transfer to a nitrite ion (Scheme 1). This would give the Ar.NH + H,NO,+ + Ar.NH,.NO* + H,O Ar.NH,.NO+ + NO,- -+ Ar-NH-NO + HNO Slow AreNH-NO -+ ArN,+ Fast Fast SCHEME 1. correct kinetic form for two nitrous acid molecules are effectively involved in the transition state but the slow proton loss implies that diazotisation should be subject to general base-catalysis.Austin has since that the acetate ion cannot take the place of the nitrite ion in aliphatic deamina- tion and although some buffer catalysis is observed in diazotisation this is associated with a more complex kinetic form (see below). Hammett30 suggested that the third-order form arose from nitrosation by nitrous anhydride (Scheme 2). This interpretation has now been shown to be correct and has also led to the reconciliation of equations (2) and (3); however it was not immediately accepted and is not even mentioned in some later Reviews.31 23 H. Schmid 2. Electrochem. 1936 42 579. 24 J. C. Earl and N. G. Hills J. 1939 1089.2s J. H. Dusenbury and R. E. Powell J. Amer. Chem. SOC. 1951,73 3266 3269. 26 L. F. Larkworthy J. 1959 31 16. 27 (a) A. T. Austin E. D. Hughes C. K. Ingold and J. H. Ridd J. Amer. Chem. Soc. 28 J. Kenner Chem. and Ind, 1941 19 443. 29 A. T. Austin Ph.D. Thesis London 1950. 30 L. P. Hammett “Physical Organic Chemistry,” McGraw-Hill Inc. New York 1940 31 H. H. Hodgson and W. H. H. Norris J.-Soc. Dyers and Colourists 1949 65 226; 1952,74,555; (b) G. J. Ewing and N. Bauer J. Phys. Chern. 1958,62,1449. p. 294. J. C. Earl Research 1950 3 120. RIDD NITROSATION DIAZOTISATION AND DEAMINATION 423 2HN0 + N,O + H,O Fast Ar-NH + N203 -+ ApNH,-NO+ + NO,- Slow Ar-NH,.NO+ -+ ArN,+ Fast SCHEME 2. Unlike the mechanism outlined in Scheme 1 and also unlike those suggested by other the mechanism of Scheme 2 has as its first stage a potentially slow inorganic reaction.This distinction provides a simple way of establishing the truth of Hammett’s suggestion. The evidence for this slow stage has been obtained by decreasing the a ~ i d i t y . ~ ~ ~ ~ With reactant concentrations of N ~ O - ~ M in @002~-perchloric acid the concentration of the free amine is sufficient to react with the nitrous anhydride before a significant proportion of the nitrous anhydride can undergo hydrolysis the rate-determining stage then shifts to the rate of formation of nitrous anhydride and the kinetic form is as equation (4) The reaction rate is then effectively independent of the concentration of the amine and also over a limited range of basicity independent of the nature of the amine.Equations (3) and (4) together indicate that the nitrosating agent is nitrous anhydride and that the later stages of equation (1) do not influence the reaction rate.33 The acidity used in establishing equation (4) is that used by Hantzsch and SchUmann,lg although these authors used hydrochloric acid and would thereby introduce a small amount of chloride ion catalysis. There is little doubt that Hantzsch and Schumann were essentially observing diazotisation according to equation (4) but since they worked with equal concentrations of the two reactants they could not distinguish between equations (2) and (4). The analysis of their results in terms of equation (4) explains both the kinetic form and the equal reactivity of the amines. Studies of lSO-exchange between nitrous acid and water have provided further support for the intermediate formation of nitrous anhydride.The equilibrium concentration of nitrous anhydride in these solutions is very small and so in the absence of the amine the nitrous anhydride formed according to equation (4) should rapidly undergo hydrolysis to nitrous acid exchanging one oxygen atom with the medium in the overall process. The rate of 180-exchange between nitrous acid and water there- fore provides a maximum value* for the possible rate of formation of nitrous anhydride. Bunton Llewellyn and Stedman35 have shown that * A maximum value because exchange could occur by other mechanisms besides the intermediate formation of nitrous anhydride ; one other me~hanism,~ is known to become important when the nitrite ion concentration is low.32 E. D. Hughes C. K. Ingold and J. H. Ridd J. 1958 65. 33 E. D. Hughes C. K. Ingold and J. H. Ridd J. 1958 88. 34 C. A. Bunton and G. Stedman J. 1959 3466. 35 C. A. Bunton D. R. Llewellyn and G. Stedman Clzern. SOC. Special Publ. 1957 Rate = k[HN0,I2. (4) No. 10 113; J. 1959 568. 424 QUARTERLY REVIEWS at very low acidities and at high concentrations of nitrite ion the rate of this oxygen exchange is second-order with respect to nitrous acid and in fair agreement with the rate of diazotisation according to equation (4). More recently diazotisation and oxygen exchange have been compared under identical conditions ;3s the reaction rates are then almost identical. Hence the nitrosation of the amine is preceded by a reaction in which two molecules of nitrous acid come together with the breaking of one N-0 bond to form an intermediate the concentration of which is proportional to [HN0,I2 (cf.eqn. 3). These observations establish the intermediate as nitrous anhydride. The absence of the amine concentration from equation (4) was first in studies of diazotisation in buffer solutions containing an excess of sodium nitrite; the plot of the percentage reaction against time is then linear because the concentration of molecular nitrous acid in the solution is maintained constant by the ready prototropic equilibration. However the reaction in buffer solutions is complicated by base-catalysis of the formation of nitrous anhydride; this apparently ~ p e r a t e s ~ ~ ~ through the sequence of reactions shown in Scheme 3 (where B- is the buffer base) and leads to the kinetic form shown in equation (5).Rate = k[HN0,]2[B-] HNO + H+ @ H,NO$ Fast H,NO,+ + B- F=* NOB + H,O Fast NOB + NO,-+ N,O + B- Slow APNH + N,O -+ ArN,+ Fast SCHEME 3. The timing of the stages in Scheme 3 is that observed for diazotisation in phthalate buffers. In acetate buffers and with higher concentrations of nitrite ions there is some evidence that the kinetic form of equation (6) is obtained corresponding to a change whereby the second stage of Scheme 3 becomes rate-determining:39*40 Rate = k[HNO,] [B-] [H+] Other have suggested that equation (6) corresponds to a general 36 C. A. Bunton J. E. Burch B. C. Challis and J. H. Ridd unpublished work. 87 E. D. Hughes C. K. Ingold and J. H. Ridd Nature 1950,166,642. 39 G. Stedman J. 1960 1702. *O J. 0. Edwards J. R. Abbott H. R. Ellison and J.Nyberg J. Phys. Chew. 1959 E. D. Hughes and J. H. Ridd J. 1958 70. 63 359. RIDD NITROSATION DIAZOTISATION AND DEAMINATION 425 acid-catalysed formation of the nitrosonium ion but this is inconsistent" with experiments involving 150-exchange.41 The direct nitrosation of the amine by the mixed anhydrides (NOB) may make a small contribution to the reaction rate. Most of the available results on the reaction of amines with nitrous anhydride according to equation (3) are summarised in Table 1. They show that for the aromatic amines there is a fair correlation between reactivity and basicity (cf. ref. 26) but that this does not hold if the aliphatic amines are included; the latter are far less reactive than their high basicity would suggest. The equilibrium constant for the formation of nitrous anhydride ( K = [N,0,]/[HN0,]2) is known42 to be about 0.2 at 20" and so the true rate coefficient for the interaction of the free amine with mole- cular nitrous anhydride can be estimated.For aniline this rate coefficient [defined by equation (7)] is abont lo7 mole-l sec.-l 1. at 25"; i.e. it is considerably less than the encounter rate of the two species.43 This result is of interest in connexion with the reaction of the amines with nitrosyl halides described below. Rate = K[R*NH,][N,Q,] (7) TABLE 1. Nitrosation by nitrous anh~dride.21,22,26,44 Aniline deriv. Temp. 10-5k (eqn. 3)a Amineb Temp. 10-5k (eqn. 3)& P-H 25 O 27 Ammonia 25 O 0.04 P-H 0 3-11 Methylamine 25 4-8 p-OMe 0 5.56 Propylamine 25 2.8 p-c1 0 0.92 Dimethylamine 25 4.0 p-NMe,+ 0 0.14 a Rate coefficients in mole-2 sec.-l L2.Some authors have emphasised the experimental difficulties in studying these reactions because of the decomposition of nitrous acid (cf. ref. 276). These figures are therefore likely to be less accurate than those for the aromatic amines. The Nitrosyl Halide Mechanism-Nitrosation by the nitrosyl halides becomes important first in the presence of rather low concentrations of halide ions ; thus nitrosation by nitrosyl chloride is the predominant mechanism for the diazotisation of aniline in concentrations of hydro- * The interpretation of equation (6) by a new reaction path is also inconsistent with an interesting kinetic principle. The transition from equation (5) to equation (6) is apparently brought about by increasing the concentration of nitrite ions and therefore also that of nitrous acid.If the order with respect to a chemical species is reduced when the concentration of that species is increased this cannot result from the incursion of a new mechanism of lower order because the relative contribution of such a mechanism would be decreased by the change in concentration. The change in kinetic form must therefore result from a shift of the rate-determining step to an earlier stage of he same reaction path. 41 C. A. Bunton and M. Masui J. 1960 304. 42 T. A. Turney J. 1960,4263; C. A. Bunton and G. Stedman J. 1958,2440. 43 Cf. S. W. Benson "The Foundations of Chemical Kinetics," McGraw-Hill Cob 44 W. Schmid Monazsh. 1954 85,424. Inc. New York 1960 p. 497. 426 QUARTERLY REVIEWS chloric acid exceeding about O.~M.* These mechanisms can also operate for aliphatic deaminati~n,~~ although they must be unimportant in the feebly acidic solutions usually employed.Several of the early papers contain evidence that diazotisation is catalysed by hydrochloric acid,4s but the kinetic form of this catalysis was first elucidated by SchmidQ7 in 1937. He found that the effect of the halide ion (X-) was to add the kinetic term shown in equation (8) (where X = Br or C1) to the background rate of diazotisation given by equation (3) Rate = k [Ar-NH,] [H+] [HNO,] [X-1. For bromide ions the catalytic coefficient [k in eqn. (S)] is more than 100 times as great as for chloride ions. Hammett30 pointed out that this cata- lytic term is equivalent to the concentration product [ArsNH,] [NOXI and that the catalysis can therefore be considered to operate as shown in Scheme 4.As with the nitrous anhydride mechanism (Scheme 2) the first stage of Scheme 4 is a potentially rate-determining inorganic reaction in which the amine is not involved. However it has not yet been found possible to HNO + X- + H+ + NOX + H,O NOX + Ar-NH -+ Ar-NH,.NO+ -+ ArN$ Fast Slow Fast SCHEME 4. adjust conditions such that the rate of formation of nitrosyl chloride is rate-determining in diazotisation,? but the corresponding stage can be made rate-determining in bromide-ion catalysis49 and more easily in iodide-ion catalysis;49 the kinetic form is then as equation (9) Rate = k[HN02] [H+] [X-1. (9) This equation has been inter~reted~~ as a rate-determining attack of the halide ion on the nitrous acidium ion (H,NO,+). It is interesting that the normal oxidation of iodide ions by nitrous acid is not observed when diazotisation occurs under the conditions of equation (9) ; this suggests that the oxidation also involves nitrosyl iodide and that the nitrosyl iodide reacts so rapidly with the amine that the alternative possibility of homo- lysis (leading to oxidation) is not observed.* This value is for diazotisation with lO”~-nitrous acid. With greater concentrations of nitrous acid the relative importance of the nitrous anhydride mechanism is increased because the corresponding kinetic term involves [HN02]-2. -t This has been achieved in the related reaction of nitrous acid with azide ions4* (see ref. 48). 46 H. Schmid and R. Pfeifer Monatsh. 1953 84 829 842. ‘II E.g. H. A. J. Schoutissen J. Amer. Chenz.SOC. 1936 58 259. 47 H. Schmid 2. Electrochem. 1937 43 626; H. Schmid and G. Muhr Ber. 1937 40 G. Stedman J. 1959 2949. 40 E. D. Hughes and J. H. Ridd J. 1958 82. 70,421. RIDD NITROSATION DIAZOTISATION AND DEAMINATION 427 Some evidence is available on the reaction of molecular nitrosyl halides with amines. The rate coefficients of equation (8) contain implicitly the equilibrium constants for the formation of the nitrosyl halides and these equilibrium constants are known for both nitrosyl chloride and nitrosyl bromide at several temperatures. From the rate coefficients [eqn. (S)] and the equilibrium constants Schmid and his co-worker~~~ have obtained the true rate coefficients for the reaction of the free amines with the molecular nitrosyl halides i.e. the rate coefficients (k) in equation (10) Rate = k [Ar-NH,] [NOXI.(10) Some of these values for the reaction of amines with nitrosyl chloride are given in Table 2. Schmid et al. stressed the dependence on the basicity of the amine but this is very slight (for comparison p-toluidine is more basic than o-chloroaniline by a factor of 250) and the most interesting features of these results lie in their similarities to encounter reactions. The values of the rate-coefficients approach that expected for the frequency of bi- molecular encounters in aqueous solution (leading to a rate coefficient of sec.-l l.2)43 and the activation energy calculated from the temperature dependence of k [eqn. (lo)] is also consistent with this inter- pretation. The value obtained for the diazotisation of aniline is 4560 cal. mole-l and that for the diazotisation of p-chloroaniline is 4950 cal.mole-l. In comparison the temperature-dependence of the viscosity of aqueous solutions gives diffusion-controlled reactions43 an apparent activation energy of about 4 kcal. mole-l at 25". The results for the reaction of amines with nitrosyl bromide are less complete but very similar ; for aniline the rate coefficient k[of eqn. (lo)] is 3.2 x lo9 sec.-l l.,. Although the calculation of these rate coefficients requires several approximations the consistency of the results makes it probable that the reaction of these amines with the nitrosyl halides approaches closely to a diffusion-controlled process. TABLE 2. Nitrosation by nitrosyl chloride.50 Amine 10-9k (eqn. 1O)a Amine 10-9k (eqn. 10)a o-Chloroaniline 1.16 o-Toluidine 2.44 m-Chloroaniline 1-63 rn-Toluidine 2.70 p-Chloroaniline 1.89 p-Toluidine 3.0 Aniline 2.60 a In mole-1 sec.-l 1.at 25". Acid-catalysed Mechanisms of Nitrosation.-If the mechanisms of diazotisation were limited to those just described the reaction of most amines with nitrous acid would occur very slowly in molar concentrations of perchloric or sulphuric acid for the perchlorate ion and the sulphate H. Schmid and E. Hallaba Monatsh. 1956,87,560; H. Schmid and M. G. Fouad ibid. 1957 88 631 ; H. Schmid and C. Essler ibid. p. 1 1 10. 428 QUARTERLY REVIEWS ion do not form covalent nitrosating agents and as implied in equation (3) the rate of diazotisation by the nitrous anhydride mechanism decreases rapidly with acidity because of the protonation of the free amine. In fact for amines such as aniline the reaction rate first decreases and then increases with acidity; the subsequent increase comes from the incursion of new acid-catalysed mechanisms of diazotisation.The first of the acid-catalysed mechanisms to be identifieds1 had the kinetic form shown in equation (1 1) Rate = k[Ar.NH2] [HNO J [H+] ; (1 1) this result has been inter~reted~~ as a rate-determining reaction of the free amine with the nitrous acidium ion (H2N02+) (Scheme 5) HNO + H+ ~5 H,NO,+ Fast At-NH + H,NO$ -+ Ar.NH,-NO+ -+ AN,+ Slow Fast SCHEME 5. Of several other possible interpretations that involving a reaction of the free amine with an equilibrium concentration of the nitrosoilium ion (formed by the reaction HNO + H+ f NO+ + H20) was rejected33 from a consideration of the rate of lsO-exchange between nitrous acid and water.In rejecting this mechanism it is implied that the ON-OH bond is not completely broken until the interaction with the amine. This mechan- istic detail has recently received support from a study of the reaction of azide ions with nitrous The reaction has the kinetic form of equa- tion (12) and apparently occurs as shown in Scheme 6; the suggested intermediate (N3.NO) can be isolated at low temperature^.^^ Rate = k [N3-] [HNO,] [H+] HNO + H+ F= H,NO,+ Fast H,N02+ + Na- N,.NO -j N + NZO Slow Fast SCHEME 6. The rate coefficient of equation (12) is about ten times greater than that observed with a number of amines according to equation (1 1); the nucleo- philic powers of azide ions and of such amines as o-chloroaniline towards the nitrous acidium ion are therefore not very different.When the azide reaction is carried out in water containing an excess of l80 the nitrous oxide produced does not contain the equilibrium concentration of isotopic oxygen;34 this shows that the nitrosation stage is not preceded by the 61 E. D. Hughes C. K. Ingold and J. H. Ridd J. 1958 77. €3. W. Lucien J. Amer. Chem. SOC. 1958 80,4458. RIDD NITROSATION DIAZOTISATION AND DEAMINATION 429 formation of the nitrosonium ion in a fast equilibrium step. This conclu- sion should apply also to the amine reaction. The nitrous acidium ion appears to be a more reactive nitrosating agent than nitrous anhydride for L a r k ~ o r t h y ~ ~ has shown that the rate of reaction according to equation (1 1) is much less sensitive to the basicity of the amine than that according to equation (3).Diazotisation by the nitrous acidium ion is therefore observed most easily with the less basic amines equation (1 1) accounts completely for the diazotisation of p-nitroaniline in concentrations of perchloric acid between 0.01 and O - ~ M . ~ ~ At higher acidities certain new factors complicate the kinetic form and the reaction rate increases above that given by equation (1 1). The analysis of kinetics in terms of reaction mechanisms becomes considerably more difficult at acidities above 0 . 5 ~ ; the conventional assumptions concerning acidity functions are proving inadequate,* but there is as yet no simple alternative approach. Diazotisation should be more suitable than most reactions for mechanistic analysis because the protonation of the primary aromatic amines should generally follow Ho (since H is defined from examples of this equilibrium) and because the equilibrium concentration of the nitrosonium ion in perchloric acid media has been to follow J o t .Nevertheless the following discussion is restricted to the two main factors involved. One of these appears to be a medium effect caused equally by perchloric acid and sodium perchlorate. With the less basic amines this is the only important factor up to acidities of 3 . 0 ~ ; thus the diazotisation of p-nitro- aniline in aqueous perchloric acid containing sodium perchlorate to maintain the ionic strength (and hence the medium effect) constant follows equation (11) with the substitution of h for [H+] up to [H+] = OM.^^ The exponential dependence of this medium effect on the con- centration of sodium perchlorate and the dependence on the nature of the metal ion57 show that the effect is not an example of specific perchlorate- ion catalysis analogous to the specific halide-ion catalysis discussed above.The other factor is the incursion of a new kinetic term of the form given in equation (13):57 Rate = k [Ar-NH,+] [HNO,]ho (1 3) This term has only been observed with the more basic amines and it has been interpreted as the effective nitrosation of the protonated amine in * This is shown most clearly in recent discussions of protonation and hydrogen- isotope exchange in aromatic rings (cf. ref. 54). -f The acidity function nomenclature in this Review follows that of Paul and Long. (ref. 56). 63 L. F. Larkworthy J. 1959 3304. 54 A. J. Kresge and Y.Chiang Proc. Chem. SOC. 1961 81. 66 N. C. Deno H. E. Berkheimer W. L. Evans and H. J. Peterson J. Amer. Chem. O6 M. A. Paul and F. A. Long Chenz. Rev. 1957,57 1. G7 B. C. Challis and J. H. Ridd Proc. Chem. SOC. 1961 173. SOC. 1959,81 2344. 430 QUARTERLY REVIEWS such a way that the proton being displaced is still present in the transition The details of this mechanism remain to be established. The above interpretation is not a return to the old idea that the protonation of the amine is necessary before diazotisation will occur reaction takes place much more readily through the free amine but if the amine is almost entirely protonated the interaction of the nitrosating agent with the protonated amine can apparently become significant. Nitrosation by the nitrosonium ion would be expected to become more important as the acidity is increased and the incursion of this mechanism should be first observed with the less basic amines for the nitrosonium ion should be the most reactive nitrosating agent.The kinetic form of the reaction of benzamide with nitrous acid in strong sulphuric acid has been explained in this way.58 Diazotisation at High Acidities.-Diazotisation in about 60 % perchloric acid and that in about 60 % sulphuric acid have to be considered separately partly because the stoicheiometric nitrous acid is then essentially present as ionised nitrosonium and partly because the kinetic form is very different from any previously discussed. In the perchloric acid media diazotisation of aniline p-toluidine and p-nitroaniline follows equation (14)?O the reaction rate therefore decreases very rapidly with acidity (e.g.by a factor of 10 between 9111- and 9*5~-perchloric acid). The results in sulphuric acid media are similar although the decrease with acidity (when expressed by the H acidity function) is slightly greater. In contrast to diazotisation in dilute acid the reaction at high acidities shows a large solvent isotope effect with deuterium60 (kH/kD = 10). It is difficult to reconcile the kinetic form and the isotope effect observed at high aciditieswith any mechanism for the rate-determining nitrosation of the amine; it seems probable that a rapid reversible nitrosation is followed by a later slow stage involving a proton transfer to the medium and one such mechanism is shown in Scheme 7. The isotope effect could then be derived partly from a change in the position of the initial equilibrium and partly from the subsequent slow stage.Rate = k[Ar.NH,+] [NO+]ho-2; (14) At-NHf + NO+ + Ar.NH,.NO+ + H+ Fast Ar-NH,-NO+ -+ Ar.NH.NO + H+ Slow Ar.NH.NO -+ ArN$ Fast SCHEME 7. s8 H. Ladenheim and M. L. Bender J. Amer. Chem. Suc. 1960,82 1895. 59 K. Singer and P. A. Vamplew J. 1956 3971; N. S. Bayliss and D. W. Watts 6o B. C. Challis and J. H. Ridd Pruc. Chem. Suc. 1960 245. Austral. J. Chem. 1956 9 319. RIDD NITROSATION DIAZOTISATION AND DEAMINATION 43 1 It is easy to see why the rate-determining step should change in this way as the acidity is increased. The rate of proton transfer from positive nitrogen to the medium is known to decrease with increasing acidity;61 thus at high acidities the N-protons of the anilinium ion contribute a separate peak to the proton magnetic resonance spectrum.62 In contrast the rate of displacement of the nitrosonium ion from Ar-NH,-NO+ by a proton may well increase with acidity for the transition state associated with the nitrosation of the protonated amine must provide a path for the corresponding reverse reaction.The proportion of the protonated nitros- amine (Ar-NH,.NO+) reverting to the reactants should therefore increase with acidity and this would tend to change the rate-determining step in the direction discussed above. This change in the rate-determining step makes it possible to understand why some of the more basic amines undergo C-nitrosation instead of diazotisation at somewhat higher acidities.63 A slow proton transfer at high acidities must provide a barrier to diazotisation analogous to that of the N-methyl group in the reaction of secondary amines with nitrous acid.It is therefore reasonable that at high acidities a process analogous to the Fischer-Hepp rearrangement (leading to C-nitrosation) should be observed with primary aromatic amines. Nitrosation Mechanisms General Survey.-The suggested mechanisms for diazotisation and deamination at low acidities ([H+] < 0 . 5 ~ ) can be simply illustrated by the network of reaction paths in Scheme 8;33 in this representation the symbol X- stands for any Bronsted base in the medium except the amine molecule the nitrite ion and bases derived from the solvent. The six numbered arrows correspond to the six types of rate- NOX +NH,Ar.NO -+ Ar-N,+ H+ HN02+ H,NO$ - Fast 3 7 /Ar'NH Fast \4 NO,-\ IJ./ N203 SCHEME 8. determining step detected in the kinetic study and the numbers designate the corresponding kinetic equations. This network of reaction paths has been of value in interpreting the kinetics of other nitrosations as well as diazotisation and deamination. The corresponding forms of a number of the kinetic equations listed in Scheme 8 have been observed by Stedman39 in the reaction of nitrous acid C. G. Swain J. T. McKnight M. M. Labes and V. P. Kreiter J. Amer. Chem. SOC. 1954,76,4243. 62 B. N. Figgis personal communication. 63 L. Blangey Helv. Chim. Acta 1938 21 1579. 432 QUARTERLY REVIEWS with hydrazoic acid and by Bunton Dahn and their co-worke~s~~ in the oxidation of ascorbic acid by nitrous acid (this appears to involve a preliminary 0-nitrosation).These related studies confirm the complexity of nitrosation kinetics and extend the range of possible substrates. The first stage of Scheme 8 the protonation of nitrous acid appears necessary for all nitrosation reactions. The nature of the next stage depends on the concentrations of the different nucleophiles present in the solution and on the discrimination shown by the nitrous acidium ion. Rate co- efficients for the reaction of different nucleophiles with the nitrous acidium ion are given in Table 3 in terms of the rate coefficient (k) of equation (1 5 ) [this generalises equations (4) (9) and (1 1) to include all possible sub- strates ( S ) ] Rate = k[S] [HNO,] [ H + ] . (1 5 ) Some of the values of k in Table 3 may require correction because of uncertainties in the equilibrium constants used in the calculations (cf.ref. 48) but the results are sufficient to show that the nitrous acidium ion does not discriminate markedly between different nucleophiles. When the values of k are considered as a function of the basicity or nucleophilicity of the substrate the results for neutral nucleophiles* appear to be approaching a limiting value and the results for negative nucleophiles may also be near a limiting value the latter limit being about 10 times the former. TABLE 3. Nitrosation by the nitrous acidium ion. Neutral substrates o-Chloroaniline 175 p-Ni troaniline 161 o-Ni t roaniline 145 Ascorbic acid 63 Hydrazoic acid 33-7 Waterb 4 2,4-Dinitroaniline 3.7 k (eqn. 15)u Ref 33 53 53 64 65 34 53 Negative substrates h i d e ion 2340 48 Acetate ion 2200 39 Ascorbate ion -2000 64 Nitrite ion 1893 33 Thiocyanate ion 1460 48 Iodide ion 1370 33 Bromide ion 1170 33 Chloride ion 975 48 Nitrate ionC 0.26 66 k (eqn.15)a Ref. =In sec.-l at 0". bFrom lPO-exchange. cThis value refers to 25". It is interesting to speculate on whether this difference between the two limits comes largely from the effect of charge on the encounter rate of the two reactants (the difference is of the right order; cf. ref. 67) but the little * The value for aniline is not given because of complications arising from the nitrous anhydride mechanism but this value cannot be sigqificantly greater than that for a-chloroaniline. 64 C. A. Bunton H. Dahn and L. Loewe Nature 1959,183,163; H. Dahn L. Loewe E. Liischer and R. Menasd Helv. Chim. Acra 1960,43 287 and following papers.66 E. Abel and H. Schmid 2. phys. Chem. 1928 136 430 and preceding papers. O7 P. Debye Trans. Electrochem. Soc. 1942,82,265. G. Stedman J. 1959,2943. RIDD NITROSATION DIAZOTISATION AND DEAMINATION 433 information at present available on activation energies does not permit a definite decision. Indeed very little is known about the details of these nitrous acidium ion reactions. Lidstone6* has suggested that N-protonation is involved making the nitrous acidium ion isoelectronic with formic acid but it is still possible that the proton transfer is part of the rate-determining step. The Productdetermining stage The stability of aromatic diazonium ions enables them to be considered as the end-product in the reaction of aromatic amines with nitrous acid. However in the aliphatic series the corresponding diazonium ions are only hypothetical intermediates; the final products are a mixture of compounds derived by substitution elimination and rearrangement.The kinetic studies show that the first stage of deamination involves N- nitrosation it remains to consider the conversion of the resulting nitros- amine into the final products. Until recently primary aliphatic nitrosamines were unknown but monomethylnitrosamine has now been obtained by reaction of methyl- amine with an ethereal solution of nitrosyl chloride at low temperature~.~~ The compound was identified from the similarity of its ultraviolet spectrum to that of dimethylnitrosamine and was shown to decompose into di- azomethane at higher temperatures. This mode of decomposition cannot be generally important for the simpler aliphatic amines" in water or acetic acid for the direct substitution products are formed without significant isotopic exchange between the a-hydrogen atoms and the In aqueous media at normal temperatures the primary nitrosamine almost certainly rearranges to give first the diazohydroxide then the diazonium ion and finally a carbonium ion (cf.eqn. 1). The carbonium ion interpretation of deamination was being discussed as early as 1928,71 when very little detailed information was available on the properties of carbonium ion reactions. By the early 1950's the correct- ness of the carbonium ion interpretation was generally accepted and since then the main interest has come from an analysis of the differences between deamination and related S N ~ reactions.In discussing the present position it is convenient first to survey those characteristics of deamination that suggest a carbonium ion interpretation and then to discuss some special features which distinguish deamination from other carbonium ion reactions. Evidence for Carbonium Ion Intermediates.-Since the nitrosation is * This decomposition by proton loss is important with amines containing an activat- ing group (C=O C z N etc.) in the a-position the diazo-compound is then the major product. g8 A. G. Lidstone Chem. andlnd. 1959,1316. 70 A. Streitwieser and W. D. Schaeffer J. Amer. Chem. SOC. 1957 79 2888. 71 J. W. Baker K. E. Cooper and C. K. Ingold J. 1928,426. E. Muller H. Haiss and W. Rundel Chem. Ber. 1960 93 1541. 434 QUARTERLY REVIEWS rate-determining the kinetic method cannot be used to elucidate the mechanism of later reaction stages; the evidence for carbonium ion intermediates has therefore to come from the nature and stereochemistry of the products.The multiplicity of these products is illustrated for one typical amine (n-butylamine) in Scheme 9; the results suggest the inter- mediate formation of a carbonium ion that can react with a nucleophile rearrange or eliminate a proton. A recent discussion of the deamination of meth~lamine~~ has led to a similar conclusion. The stereochemistry of the direct substitutions shows that the reaction usually involves racemisation accompanied by some inversion,73 a result indicating the intermediate formation of a carbonium ion followed rapidly by reaction with a nucleo- phile while one side of the carbonium ion is still partly shielded by the departing nitrogen molecule.As might be expected the amount of race- misation increases with the stability of the carbonium ion ;'O the formation of [1-2H] butyl acetate from [l-2H]butylamine by nitrous acid in acetic acid leads to 3 1 % racemisation (69 % inversion) while the corresponding reaction of secondary butylamine leads to 72 % of racemisation. Retention Me * [CH2] ,OH Me. [CH2];Cl (5.2; Me.CH2CH=CH (36.5 %> Me*CH2CH *Me (2.8 yo) I Cl SCHEME 9. Product composition in the deamination of n-butylamine in aqueous hydrochloric A later analysis7o of the olefin fraction for deamination in acetic acid has shown that 29 % of the olefin is but-2-ene (9 % cis; 20 % trans); this but-Zene ispresumably derived from the rearranged carbonium ion.of configuration has been observed in the deamination of a-amino-a~ids,~~ as in other S N ~ reactions involving a neighbouring carboxyl group.75 The rearrangements of the carbon skeleton observed in deamination (e.g. the conversion of neopentylamine into t-pentyl alcohol) generally correspond to those in other SN1 reaction^.^ Where several products are formed the comparison of product composition with that from related S N ~ reactions has sometimes provided evidence for the same carbonium 72 A. T. Austin Nature 1960 188 1086. 73 P. Brewster F. Hiron E. D. Hughes C. K. Ingold and P. A. D. S . Rao Nature 74 F. C. Whitmore and D. P. Langlois J. Amer. Chem. SOC. 1932,54,3441. 76 Ref. 7 p. 383. 1950,166 179. RIDD NITROSATION DIAZOTISATION AND DEAMINATION 43 5 ion intermediate ; examples include some reactions of t-butyl com- p o u n d ~ ~ ~ and the aqueous deamination* of allylic amine~.~’ Related comparisons involving the products formed by elimination are more difficult because of the reaction of nitrous acid with the olefins formed in deamination but the reaction of t-butylamine with nitrous acid has been shown to give the same yield of olefin as the decomposition of the corresponding dimethylsulphonium salt under similar conditions ;79 for other substrates some discrepancies have been r e p ~ r t e d .~ ~ ~ ~ ~ To summarize the comparison of deamination and S,l reactions provides strong evidence that some deaminations involve a carbonium ion intermediate and presumptive evidence that such intermediates are generally formed. Special Characteristics of Deamination.-The evidence for carbonium ion intermediates makes it particularly interesting to study those reactions where significant differences exist between the products of deamination and those of other SN1 reactions e.g.solvolysis of toluene-p-sulphonates. In part these differences appear to arise from a lower discrimination in the deamination so that rearrangement products that are unimportant in solvolysis become significant in deamination. Thus the deamination of n-butylamine gives a considerable amount of secondary substitution pro- ducts owing to a hydrogen migration (Scheme 9) but no secondary product is obtained on solvolysis of the corresponding toluene-p-sulphon- ate although studies with [ l-2H]butylamine show that deamination and solvolysis involve similar amounts of racemisation.70 Primary carbonium ions rearrange to give secondary products under more drastic conditions (e.g. in Friedel-Crafts reactions) but normally discriminate in favour of direct substitution and elimination. Another apparent example of this lack of discrimination comes from the migration aptitudes of the p - methoxyphenyl and the phenyl group :81 in the pinacol rearrangement the former migrates more rapidly by a factor of 500 but in pinacolic deamina- tion the factor is only 1.56. However this lack of discrimination which suggests that the carbonium ions derived from deamination are unusually reactive is not a sufficient explanation of the differences observed. This is shown most clearly by the results of Cram and McCartys2 for the acetic acid deamination of l-methyl- 2-phenylpropylamine and the acetolysis of the corresponding toluene-p- * These allylic deaminations in acetic acid involve less rearrangement than do acetolyses of the corresponding halides.The results have been explained in terms of a “hot” carbonium ion in the deaminati~n,~~ but this result may also be influenced by ion- pair equilibria (cf. p. 440). 76 L. G. Cannell and R. W. Taft Abstracts of papers presented at the 129th Meeting of the American Chemical Society Dallas 1956 p. 46-~. 77 R. H. DeWolfe and W. G. Young Chem. Rev. 1956,56,753. 78 D. Semenow Chin-Hua Shih and W. G. Young J. Amer. Chem. Soc. 1958 80 5472. 79 J. S. Burgess Ph.D. Thesis London 1953. A. Streitwieser J. Org. Chem. 1957 22 861. D. Y. Curtin and M. C . Crew J. Amer. Chem. Soc. 1954 76 3719.82 D. J. Cram and J. E. McCarty J. Amer. Chem. Soc. 1957 79 2866. 436 QUARTERLY REVIEWS sulphonate. These compounds exist as two diastereoisomers termed erythro* (VI) and threo (VII) each of which contains three p-groups (Ph Me H) capable of undergoing rearrangement during deamination or solvolysis. From the symmetry of the system the rearrangement of the phenyl group gives the same structural product as the non-rearranged material but the rearrangement does have stereochemical consequences which permit its observation. Acetolysis of the toluene-p-sulphonates X = NH or p-C6H4Me.SOi0 appears to be accelerated by the phenyl group and to lead to a symmetrical “phenonium” ion (illustrated for the threo-isomer by VIII). Some hydrogen migration is observed giving (IX) but none of the methyl migration Me H I I Me,CH-C-OAc I I Et-C-OAc (X) Ph (‘X) Ph leading to (X).In contrast in the deamination of the threo-amine (VII; X = NHd more methyl than phenyl migration occurs (ratio 1-5:1) and the amount of hydrogen migration is also increased. This result cannot simply arise from a lack of discrimination in the rearrangement for in the deamination of the erythro-amine phenyl migration exceeds methyl migration by a factor of eight. The above example is only one of a number of unusual rearrangements observed in dea~nination,~~ but these can be rationalised by assuming that the mobility of groups in deamination is largely controlled by the conforma- tions of the diazonium ion and that the group trans to the diazo-nitrogen atom in the most stable conformation is particularly mobile (ground-state control of migration).The three conformations of the threo-diazonium ions H (XIII) * The erythrodiastereoisomer is that which in one eclipsed conformation has at least 83 Cf. D. Y. Curtin and M. C. Crew J. Amer. Chem. Soc. 1955,77,354; L. S. Ciereszko two sets of similar substituents in line. and J. G. Burr ibid. 1952 74 5431. RIDD NITROSATION DIAZOTISATION AND DEAMINATION 437 are (XI)-(XIII) conformation (XI) should be the most stable for each bulky group is flanked on one side by a hydrogen atom. The methyl group is then in the favoured position for rearrangement. A similar argument applied to the erythro-isomer would place the phenyl group in the most favoured position for rearrangement. The importance of this ground-state control can be linked to the low activation energy for the breakdown of the aliphatic diazonium ion.In the solvolysis of the toluene-p-sulphonates the mobility of the /!I-substituents presumably facilitates the ionisation to an extent which is energetically large in comparison with the energy differences between the ground-state conformations ; the relative stability of these conformations is then less revelant and the greater mobility of the phenyl groups is sufficient to determine the product. In comparison with ionisation of the toluene-p- sulphonate the activation energy for heterolysis of the diazonium ion should be very small and the contribution from the mobility of the /!I- substituents should be correspondingly decreased ; it is therefore reason- able that reaction should occur in the most probable ground-state confor- mation rather than in that which in the sulphonates is energetically the most favourable.The carbonium ion from (IX) should first assume conformation (XIV) but this can then rearrange to (XV) by partial rotation about the central bond. Although the methyl group should be the most mobile in conforma- tion (XIV) (for the a-electrons of the C-Me bond are well placed to overlap with the vacant p-orbital on the adjacent carbon) this would not be true in conformation (XV). Hence the above explanation requires that migration of the methyl group either occurs at the same time as the C-N,+ heterolysis or follows this process before rotation about the centre C-C bond becomes significant. Me Other studies have provided evidence that some rotation can occur about the central C-C bond after the loss of nitrogen.One example con- cerns the pinacolic deamination of 2-amino-1,l -diphenylpropan-1-01 (XVI); in this reaction the rearranged carbonium ion is stabilised by proton loss from the hydroxy-group to form the ketone (XVII). The overall HOCPh2CH2-NH2 -+ HO*CPh2CH2+ + Ph*COCH,Ph ow (XVII) reaction was shown by Bernstein and Whitmore 84 to involve about 88 % of inversion at the migration terminus. More recent work by Benjamin R4 H. I. Bernstein and F. C. Whitmore J. Amer. G e m . Soc. 1939 61 1324. 438 QUARTERLY REVIEWS Schaeffer and Collins85 has involved labelling of the amine (XVI) in one phenyl group with 14C. The compound then exists as two diastereo- isomers one of which is shown in its most stable conformation as (XVIII). As before the reaction was observed to occur with about 88 % of inversion H HO@; HOQT Me - MQ -Me HO Pf (XVIII) ;ti* (XIX) h* P h C0.C.a .:H Me Ph * .COC.-.H Th Me Me -@- (XXI I I) (XXI) Ph A* (XXI I) to give the ketone (XXII) and 12 % of retention to give the isomer (XXIII).However the interesting new result is that in the reaction of the dia- stereoisomer (XVIII) the ketone formed with inversion involved essentially complete migration of the labelled phenyl group while that formed with retention correspondingly involved migration of the unlabelled phenyl group. The formation of the latter product indicates that the lifetime of the carbonium ion (XIX) is sufficiently long to permit some part of the material to undergo the partial rotation to (XX). However the absence of labelled- phenyl migration in the ketone of retained configuration indicates that the carbonium ion does not have time to reach conformation (XXI).It appears therefore that the deamination of these compounds produces a reactive open* carbonium ion whose lifetime is rather less than the period of rotation about the central C-C bond. This conclusion concerning reactive open carbonium ions has not been universally accepted and it has been ~ u g g e s t e d ~ ~ ~ ~ ~ that a considerable part of the rearrangement and substitution in deamination can occur at the same time as the loss of nitrogen i.e. that the substitution is at least partly an SN2 reaction of the diazonium ion. The ready breakdown of aliphatic diazonium ions is considered so to reduce the differences in the activation energy for competing reactions that direct displacement by solvent can become comparable with the unimolecular formation of a carbonium ion.The Reviewer considers that this emphasis on the bimolecu- lar nature of the reaction may give a misleading picture of the transition state because the argument is based on energetics alone and does not consider the probability factors involved. Since ground-state control operates in the selection of the rearranging group so that deamination * Z.e. not a bridged structure analogous to (VIII). 85 B. M. Benjamin H. J. Schaeffer and C. J. Collins J. Arner. G e m . Soc. 1957 79 61 60. RIDD NITROSATION DIAZOTISATION AND DEAMINATION 439 occurs in the most populated conformation a similar kind of ground-state control should operate as far as the orientation of the local solvent molecules is concerned.The deamination is therefore unlikely to require the special orientation of solvent molecules associated with an SN~ substitution or an E2 elimination (cf. ref. 82). Benjamin et al. pointed out that the “topside” rearrangements (e.g. the formation of XXIII) are difficult to reconcile with a mechanism involving a synchronous rearrange- ment,85 and on the whole it seems best to consider all these reactions as involving an intermediate carbonium ion. The formation of olefins in deamination has been discussed in terms of an E2 reaction,so but the arguments can be easily rephrased in terms of ground-state control of elimination from an intermediate carbonium ion. The material quoted above was selected to illustrate how the special facility of carbonium ion formation by the loss of nitrogen can apparently modify the relative proportions of the products formed.A more detailed consideration of the nature and stability of carbonium ions formed under different conditions belongs more to the general field of carbonium ion chemistry ; such matters have been discussed in recent publications.86 There is one other factor that particularly concerns deamination namely the possible influence of ion pairs involving diazonium ions in non- aqueous solvents. Such ion pairs have been discussed by Huisgen and Rii~hardt,~’ but recently new evidence has been obtained by White and AufdermarshS8 from an oxygen-18 study of the decomposition of N- nitrosoamides (ON-NR-CO-R’ where R is aliphatic). These nitrosoamides are known to decompose by a preliminary rate-determining rearrangement to the corresponding diazo-e~ter~~ (R-N,-OCOR’) which can then undergo a number of reactions depending on the conditions.In such solvents as acetic acid one of these reactions involves the formation of the ester R’CO,R by an intramolecular path requiring mainly retention of con- figuration. These features would be consistent with a cyclic SN~ transition state but the oxygen-isotope studies show that the two oxygen atoms become partly equivalent before the final product is formed. The degree of mixing depends on the conditions and on the compound in a way that suggestP the ion-pair equilibrium (XXIV) shown in Scheme 10. The products with retained and inverted configuration involve essentially the same degree of mixing of the oxygen isotopes; it appears that some rotation of R+ can occur in the ion-pairs (XXV) without further equilibra- tion of the oxygen atoms.86 V. F. Raaen and C. J. Collins J . Amer. Chem. SOC. 1958 80 1409; C. J. Collins W. A. Bonner and C. T. Lester ibid. 1959 81 466; D. Bethel1 and V. Gold Quart. Rev. 1958,12 173; M. S. Silver J. Amer. Chem. SOC. 1961 83 3482. 88 E. H. White and C. A. Aufdermarsh J. Amer. Chem. SOC. 1958 80 2597; 1961 83 1179. 8 9 R. Huisgen and C. Ruchardt Annalen 1956 601,21 and preceding papers; E. H. White J. Amer. Chem. SOC. 1955 77 6014. R. Huisgen and C. Riichardt Annalen 1956,601,l. 440 QUARTERLY REVIEWS NO 0 + RO-g-R' R '"0-5- R' '"0 0 SCHEME 10. The reaction paths in Scheme 10 do not fully indicate the complexity of these nitrosoamide reactions but the evidence88 that the loss of nitrogen involves a diazonium-carboxylate ion-pair supports earlier conc1usions~7 on the importance of ion-pair equilibria in these reactions and suggests that similar equilibria may influence the course of deamination in media of low polarity.Deamination could lead first to a diazonium-hydroxide ion- pair but in acetic acid such ion-pairs would be converted into diazonium acetate ion-pairs by proton transfers. The subsequent formation of a carbonium ion would then occur in the neighbourhood of a hydroxide or an acetate ion. This may explain the extensive formation of unrearranged products in the acetic acid deamination of allylic arnine~'~ and other results previously explained in terms of cyclic S N i transition states.90 It seems increasingly necessary to distinguish between deamination in water and in solvents of low polarity.The deamination of alicyclic amines has not been included in this Review but two recent papers on this subject add to the general knowledge of deamination mechanisms. Streitwieser and Coverdalegl have shown that the aqueous deamination of cis-2-deuteriocyclohexylamine proceeds with almost complete retention of configuration so that the carbonium ion if formed reacts before reaching conformational equilibrium. This might suggest a cyclic S N ~ transition state but Boutle and Buntong2 have shown that this is unlikely since the oxygen in the final cyclohexanol is not that present in the initial nitrosation of the amine. An intermediate carbonium ion therefore appears more probable. From the above discussion it appears that the simple representation of deamination by equation (1) is still adequate to explain the great majority of the experimental results although consideration must be given to the different mechanisms of nitrosation to the special facility of carbonium go P.S. Bailey and J. G. Burr J. Amer. Chem. Soc. 1953 75 2951 ; cf. discussion in ref. 80. D1 A. Streitwieser and C. E. Coverdale J. Amer. Chern. SOC. 1959 81 4275. s2 D. L. Boutle and C. A. Bunton J. 1961,761. RIDD NITROSATION DIAZOTISATION AND DEAMINATION 44 1 ion formation by loss of nitrogen and to the influence of diazonium salt ion-pairs in media of low polarity. The author thanks Sir Christopher Ingold F.R.S. Professor E. D. Hughes F.R.S. and many of his colleagues for their comments on this manuscript. He also acknowledges several helpful discussions with Dr. G. Stedman.

ISSN:0009-2681    

DOI:10.1039/QR9611500418

出版商:RSC    

年代:1961

数据来源: RSC

摘要:

THE ACTINIDE OXIDES By L. E. J. ROBERTS (ATOMIC ENERGY RESEARCH ESTABLISHMENT HARWELL DIDCOT) To regard the elements of atomic number higher than actinium as forming an actinide (5f) series is chiefly of value in correlating and explaining the chemistry of the later members and the emergence of the group valency of three.l The chemistry of the oxides illustrates the general tendency for higher valencies to be preferred by the earlier members of the series; the ideal formulz of the main oxide phases are given in Table 1. The uranium TABLE 1. Ideal formuliz of important oxide phases. Tho Pa0 UO NpO PuO Am0 - Ac203 - - - - Pu203 Am203 Cm203 Tho PaO UO NpO PuO AmO CmO (Pa02.3) u4°9 Pa205 u30 NP308 uo3 and plutonium oxides are of great technical interest in the nuclear-energy industry and many papers given at the Second Geneva Conference on the Peaceful Uses of Atomic Energy (1958) summarise various aspects of technological work.The present Review is concerned with the structures of individual oxides and of the mixed oxide phases formed from them and with the surface chemistry and the mechanisms of some of their reactions. The oxidation reactions have received particular attention since they may illustrate some features of solid-state reactions of general importance. Monoxide phases The formation of monoxides by all the elements from thorium to americium has been reported, usually on the basis that an oxide film with the sodium chloride structure has been identified on a specimen of metal. The structure definitely established by the relative intensities of the X-ray diffraction lines is typical of most semi-metallic MX phases irrespective of the metal structure or its atomic r a d i ~ s .~ An early report of the preparation of americium monoxide by reduction of the sesquioxide with hydrogen has not been c~nfirmed.~ Many attempts to prepare uranium monoxide by the direct reaction of the dioxide with metallic uranium have failed and it is possible that the 1 Katz and Seaborg “The Chemistry of the Actinide Elements,” Methuen London 1957. Zachariasen Acta Cryst. 1952 5 19. Rundle Acta Cryst. 1949 1 180. 4 Penneman and Asprey Proc. Geneva ConJ 1956,7 355. 442 ROBERTS THE ACTINIDE OXIDES 443 “UO” phase always contains some carbon or nitr~gen;~ similar phases have been made by heating uranium carbide with the dioxide above 1600° and one nearly pure specimen had the composition UCo,5,00.p3.6 The monoxide phase does not appear during the oxidation of solid metallic uranium at low temperatures but it does appear on the outer surface not at the metal-oxide interface when partially oxidised specimens are annealed.This unusual sequence suggests that nucleation of uranium monoxide is possible only when it is catalysed by nitride or carbide at a free surface.’ The deliberate addition of a little carbon resulted in the formation of much more of the monoxide phase in such quantity that carbon could have been only a minor constituent. It seems certain that pure or nearly pure monoxide phases can be pre- pared but that they are semi-metallic interstitial compounds similar to and miscible with the MC and the MN phase and are not truly the oxides of the elements in a bivalent state.The lower oxides said to occur in the black residues left when metallic thorium or uranium is dissolved in hydrochloric acid,s are not related to the monoxide phases described here; their character is not yet clearly established. Dioxides and related phases The dioxides are the most important and most characteristic oxides of the early members of the actinide series. They are high-melting refractory oxides with the fluorite (CaF,) structure (Fig. la); the regular but not Fig. la. Fig. I b . FIG. 1. (a) The fluorite structure MO,. Small circles = M; large circles = 0. (b) The perovskite structure AB03. Full circle = A; small open circles = B; large open circles = 0. linear contraction of the unit cell with increasing atomic number is an expected effect of the increasing number offelectrons.The relative stability of the dioxides is characteristic of the series. Thorium dioxide is the only regular oxide of thorium. The final oxides formed on pyrolysis in air of Rundle Baenziger Wilson and McDonald J. Arner. Chem. SOC. 1948,70,99. Vaughan Melton and Gerds U.S.A.E.C. Report BMI-1175 1957. Williams and Westmacott Rev. Metallurgie 1956 189. Katzin J. Amer. Chem. SOC. 1958 80 5908; Young J Inorg. Nuclear Chem. 1958 Asprey Ellinger Fried and Zachariasen J. Amer. Chem. SOC. 1955,77,1707. 7 418. 444 QUARTERLY REVIEWS proactinium and uranium compounds are Pa,O and U,08 but uranium dioxide may be formed by reducing the higher oxides with hydrogen at 300-600" while Pa,O is reported to be more stable being reduced to the dioxide in hydrogen at 1500".Ignition of neptunium and plutonium salts in air gives the dioxides. Plutonium dioxide is partially reduced in hydrogen at high temperatures americium dioxide loses oxygen when heated to 800" in a vacuum and curium dioxide is only formed under 1 atm. or more of oxygen. The dioxides are convenient starting materials for the prepara- tion of quadrivalent compounds.1o The relevance of the surface chemistry of these oxides to their general chemistry is a consequence of the small particle size or high specific surface of many preparations; as a result surface reactions can account for measurable changes in the overall stoicheiometry. (a) The Oxidation of Uranium Dioxide.-The uranium oxide in equili- brium at low temperatures with oxygen at atmospheric pressure is the trioxide U03 and above 700" U308 but the formation of either is apparently inhibited by a kinetic barrier.The trioxide can be prepared by oxidation at low temperatures in 1 atm. of oxygen of uranium dioxide of large surface area,ll whereas a highly crystalline oxide UO or U,O, can be oxidised to trioxide only under high oxygen pressures. The formation of U30 is occasionally noticed when dioxide of small particle size is exposed to air at room temperature but this is probably due to a sudden local rise in temperature and a highly crystalline dioxide is oxidised at low temperatures to yield fluorite-type oxides U308 being formed as a second step at temperatures above 200".12 The oxidation of uranium dioxide at low temperatures is interesting since the mobile species is oxygen and the structure of the products very closely related to that of the dioxide itself.The first process is a chemisorp- tion measurable at -183" which is rapid in the presence of an excess of oxygen,13 but a slow process of chemisorption follows rapid physical adsorption when the oxygen is added in small increments.14 The heat of adsorption decreases regularly with surface coverage from an initial value of 50 kcal./mole to values typical of physical adsorption when 50-80 % of the surface uranium ions have reacted with an oxygen The chemisorption is succeeded by a slow process which eventually leads to the absorption of about 5 times as much oxygen and can be followed at all temperatures between -138" and SO0.l6 The extent of the oxidation at these temperatures is simply proportional to the surface area lo Smiley and Brater Progr.Nuclear Energy Series 111 1958 2 107; and ref. 1. l1 Boulle Jarg and Berges Compt. rend. 1951,233 1281. l2 Jolibois Compt. rend. 1947 224 1395; Aronson Roof and Belle J. Chem. Phys. l3 Roberts J. 1954 3332; 1955 3939. l4 Ferguson and McConnell Proc. Ro-v. Soc. !p57 A 241 67. l5 McConnell and Roberts "Chemisorption ed. W. E. Garner Butterworths l6 Anderson Roberts and Harper J. 1955 3946. 1957 27 137. Scientific Publ. London 1956 218. ROBERTS THE ACTINIDE OXIDES 445 of the dioxide; this process results in all preparations of uranium dioxide containing an excess of oxygen after exposure to air. The few measurements available indicate that the density increases as oxidation proceeds after an initial fall which may plausibly be ascribed to the formation of the chemisorbed layer.The absorptive capacity of the uranium dioxide for oxygen can be regenerated by annealing at a temperature where it is known that migration of oxygen into the lattice does occur and therefore oxidation at room temperature probably produces an oxidised layer on each particle of uranium dioxide. Infrared lines characteristic of amorphous trioxide U03 have been observed when very small (100 A) particles are oxidised at room temperature,17 but the density and X-ray evidence for rather larger particles that are oxidised slowly (so that the process is isothermal) are not in accord with the formation of trioxide suggesting rather penetration of the lattice by oxygen and formation of a layer of one of the tetragonal pseudo-cubic phases containing interstitial oxygen.16,18 The mechanism of such a low-temperature reaction is of considerable interest; the energy liberated by the chemisorptive process may cause a rearrangement of the oxygen ions.19 Bulk oxidation occurs above 80" by a process having an activation energy of 20-25 kcal./mole; oxidation in air proceeds to a limit between U02.33 and U02.38 finally causing some distortion of the lattice from cubic to tetragonal symmetry and a contraction of the unit cell although the surface area is unchanged and individual particles are virtually un- altered.16*21 The regular increase in the density of a given sample of uranium dioxide as oxidation proceeds shows that the product contains oxygen in interstitial sites in the uranium dioxide lattice.20s21 Kinetic results indicate a diffusion-controlled and not a phase-boundary reaction but do not distinguish between diffusion into the dioxide down a concen- tration gradient and diffusion through a thickening layer of product as the rate-determining step.22J2 A distinction is possible in principle by X-ray studies but difficult to make because of overlapping diffraction lines.The profiles of the lines on X-ray powder photographs have been interpreted in terms of a concentration gradient throughout the uranium dioxide particle causing a regular contraction of the lattice,21 and as showing the initial formation of a skin of tetragonal oxide U O Z . ~ ~ . ~ ' Two recent studies with the X-ray diffractometer an inherently more powerful tool provide evidence of the formation of a tetragonal phase having cla = 1.01 or 0-99 in the early stages of the reaction with the eventual formation of a tetra- gonal phase having c/a = 1-03 when the composition has reached u02.33- U02.3,.17J8 This is probably the best account to date of the oxidation of l7 Hoekstra Santoro and Siegel J.Inorg. Nuclear Chem. 1961 18 166. l9 Anderson and Gallagher 4th Internat. Symp. Reactivity of Solids 1960. 2o Blackburn Weissbart and Gulbransen J . Phys. Chew. 1958 62 902. 21 Anderson Symp. Peaceful Uses of Atomic Energy in Australia 1958. 22 Alberman and Anderson J. 1949 5303. Belbeoch Piekarski and Perio J. Nuclear Materials 1961 3 60. 446 QUARTERLY REVIEWS rather large uranium dioxide particles at ordinary pressures but it is to be remembered that different mechanisms may hold for different particle sizes18 and that oxidation at low pressures ceases at compositions below U0,.2,.16 The oxidation rate is instantly changed by changing the ambient oxygen pressure as is the case for the room-temperature oxidation.16 Since the heat of chemisorption covers the whole range 4-55 kcal./mole it is clear that some of the surface oxygen will be reversibly absorbed above loo" and this oxygen may be concerned with the rate-determining concentration gradient.(b) Stable Phases in the Uranium Dioxide-Oxygen System.-The phases formed directly during the oxidation of uranium dioxide below 250" are unstable at high temperatures; the stable phases at high temp- eratures are a non-stoicheiometric U02+. phase the cubic or pseudocubic u4og phase and the orthorhombic U,O,- phase which is discussed below.The system does not easily lend itself to the usual anneal-and- quench techniques because of the extremely rapid oxygen diffusion in UO,+ ;23 the apparent diffusion coefficient characteristic of all interstitial oxygen atoms in UO,,. is -3.10-8 cm.2/sec. at 750". G r o n v ~ l d ~ ~ studied the U02-U409 system by high-temperature X-ray measurements from 20" to 900" and his results are shown in Fig 2; a Composition O/U FIG. 2. Portion of U-Oz phase diagram. Circles denote X-ray results (taken from Gronvold J. Inorg. Nuclear Chem. 1955 1 357). metallographic study by Schaner is in reasonable agreement at low 0:U ratios.25 Higher solubilities of oxygen in uranium dioxide below 400" 23 Belle Geneva Conf. 1958 paper P/2404. 24 Gronvold J. Inorg. Nuclear Chem. 1955 1 357. 25 Schaner J.Nuclear Materials 1960 2 110. ROBERTS THE ACTINIDE OXIDES 447 were reported by Vaughan et aZ.26 from high-temperature X-ray evidence; however their materials were originally annealed at 450" and it is possible that the U@Q phase is slow to nucleate at 450" for 0:U ratios approaching 2.0. The phase boundaries at higher temperatures and higher 0:U ratios are more reliably determined by measuring equilibrium oxygen pressures and applying the phase rule the pressure at a given temperature is independent of composition when two solid phases are present. Oxygen pressures have been measured from 950" to 1150" by an effusion method,27 from 880" to 1080" by a high-temperature e.m.f. method,28 and from 1000" to 1450" by direct tensimetric The agreement between these three investigations is very good except that Blackburn2' places the upper limit of the U,OQ phase at a higher 0:U ratio than that in Fig.2. There may be a small region of homogeneity of the U40g phase at low temperatures.26 At least two tetragonal phases based on distortions of the fluorite structure in the composition range UO,. 25-U02.40 can be preserved indefinitely below 500" although their thermodynamic stability is still open to doubt. These are the y1 phase,* having a = 5.381 A c = 5.556 A c/a = 1.030 stable from 180" to about 400" and the y phase having a = 5.406 A c = 5.491 A c/a = 1.016 stable from 200" to about 550". The composition of the y1 phase has been variously given as uo2.33 t a U02.38 and that of the y phase as U02.30 to uo2.35.17*18321730 There is no doubt that the y1 phase is richer in oxygen than yz since the reaction y1 + y + U308 occurs between 360" and 460"; similarly the y phase must be richer in oxygen than U@Q since the reaction y z -+ U409 + U308- occurs above 550".Both phases may exist over a range of composi- tion. Hoekstra et a1.l' point out that the tetragonal phases should be described in terms of a body-centred tetragonal cell. It also seems certain that metastable cubic phases can be prepared at temperatures below those at which the U02+z phase is ~table.~l,~O Cubic phases with 0:U < 2.0 have also been reported on the evidence ofcell edges greater than those for U02 itself. While it seems possible that the dioxide could lose oxygen at high temperatures under very reducing conditions the compositions of these phases have not been established by analysis.A UO,- phase would be expected to be re-oxidised very rapidly in air which may explain why similar results have not been recorded more often. Density measurements prove conclusively that the U,O structure is related to that of U02 by the inclusion of interstitial oxygen 1 oxygen atom per unit cell (see Fig. la). X-Ray powder patterns show all the strong lines of the dioxide lattice displaced because the unit cell * The notation is that used by Perio. 26 Vaughan Bridge and Schwartz U.S.A.E.C. Report BMI-1241. 27 Blackburn J. Phys. Chem. 1958,62 897. 28 Aronson and Belle J. Chem. Phys. 1958 29 151. 2 9 Roberts and Walter J . Inorg. Nuclear Chem. in the press. 30 Anderson Bull. SOC. chim. France 1953 781 ; Perio ibid. p. 840. ias contracted 448 QUARTERLY REVIEWS (a = 5.438 A; cf.a = 5.468 for UO,) and a large number of super- structure lines at high angles.,* It is clear that the uranium atoms must have remained very close to the f.c.c. fluorite positions and that the true unit cell of the structure must be very large. Entropy values reveal that u4og is highly ordered compared with the U02+z phase.29 The positions of the oxygen atoms cannot be deduced readily from X-ray data because of the large difference between the scattering factors of uranium and oxygen atoms. From neutron diffraction data Andresen et. aL31 proposed a tetra- gonal structure for U409 with a = 2 ,/(2a) and c = 2a where a = 5.438 A. In principle the positions of the oxygen atoms can be determined in this way. However the resolving power of neutron diffractometers may be insufficient to reveal the structure of a very large unit cell from work on powders and a single-crystal study would be preferable.Perio and his have recently studied single crystals of u4og by X-ray methods and report that the structure is in fact cubic with a = 4a and have proposed a model with the interstitial oxygen atoms disposed in an ordered fashion on positions equivalent to the ($+$) position in the original cell (see Fig. la). Most measurements of the densities of the tetragonal y1 and y2 phases show that these phases like U@g contain interstitial oxygen atoms relative to the UO structure ;17920921 the low densities occasionally reported,33 which would indicate cation vacancies in the structure may be due to the presence of a little U308 in the annealed specimens.The tetragonal symmetry indicates that the ordering of the interstitial oxygen atoms is different from that in U409 and that the cations have moved further from the fluorite positions. Different formulations of the true unit cells have been p r o p ~ s e d ~ ~ ~ ~ but the structures have not been determined; if the true unit cells are even larger than that of U,O, their resolution will be a formidable undertaking. The difficulty of attaining single homogeneous phases on cooling material to room temperature has marred much work on the electrical and magnetic properties of the system. It seems clear however that U02+z is a metal-deficient p-type semiconductor while U409-2 and U30,-2 (the y1 phase) are metal-excess n-type semiconductors ; this confirms the view that these are regular ordered (c) The NpO,-O2 Pu02-02 and Pa0,-0 Systems.-The behaviour of neptunium and plutonium dioxide towards oxygen bears no resemblance to that of uranium dioxide.Both are inert in oxygen plutonium dioxide treated with atomic oxygen gave no evidence of higher oxide formation nor did neptunium dioxide when heated in oxygen or nitrogen dioxide.36 31 Andresen Enlarged Symp. on Reactor Materials Stockholm Oct. 1959. 32 Belbeoch Piekarski and Perio C.E.A. Report 1960. 33 Gronvold and Haraldsen Nature 1948 162 69. 34 Willardson Moody and Goering J. Inorg. Nuclear Chenz. 1958 6 19; Vaughan Bridge and Schwartz U.S.A.E.C. Report BMI-1241 1957. s6 Gruen Koehler and Katz J. Amer. Chem. SOC. 1949 73 1478. ROBERTS THE ACTINIDE OXIDES 449 This is the more remarkable in the case of neptunium dioxide since a higher oxide NPSO, is well known and is probably the stable oxide in air below about 200"; the Np308 structure can be preserved in high pressures of oxygen up to 500".Np,O decomposes to give Np02 directly in low oxygen pressures above 400" with no evidence of any intermediate 0xide.,~9~' Both neptunium and plutonium dioxide preparations do absorb some oxygen when heated in oxygen after having been fully reduced by exposure to hydrogen or carbon monoxide at 500-800".38 The reaction is measur- able above 200" and oxygen contents as high as PuO,,, have been reported.39 The absorption is greater the larger the surface area and the oxygen absorbed never exceeds an amount equivalent to a mon01ayer;~~ it may be noted that an apparent change of about 0.07 in the 0:Pu ratio is caused by the adsorption of a monolayer of oxygen by oxide particles of 0.05 p diameter-that is a preparation of surface area 10 m.2/g.and larger specific surfaces are often recorded for preparations of these refractory oxides. The size of the unit cell showed no significant change on oxidation and reduction. These observations all suggest that these are surface processes; such a chemisorption operative only above 200" is presumably an activated process quite different from the adsorption of oxygen by a reduced surface of uranium dioxide resembling rather the high-temperature chemisorption found on thorium The properties of protactinium dioxide are of great interest. Three cubic fluorite-type oxides have been rep~rfed.~ The dioxide is a black solid obtained by reducing higher oxides in hydrogen at 1550" and the composi- tion must be close to PaO since the cell constant (a = 5.505 &falls between those of the dioxides of thorium and uranium.Diprotactinium pentox- ide (a = 5.455 A) is prepared by heating a protactinium(v) hydroxide precipitate to 500" in air and its formula has been deduced from the white colour. Another black cubic oxide of intermediate composition is pre- pared by heating the pentoxide to 1800" in vacuo as well as a black tetra- gonal oxide having a = 3.835 A c = 5.573 A. Densities could notbe determined on the 50 pg. samples used. However the contraction from dioxide to pentoxide 0.050 for a change in oxidation number of one unit is similar to the initial contraction of UOz to U40 (0.030 A for an oxidation number increase 4 + 4.5) and it seems probable that the cubic pentoxide is related to protactinium dioxide by inclusion of interstitial oxygen.The unit cell of the tetragonal phase may be expressed as the face- centred cell with a = 5.42 A c = 5-573 A c/a = 1.03 which is similar t o 38 Gruen and Katz J. Amer. Chem. SOC. 1950,71 2106. 37 Roberts and Walter U.K.A.E.A. Report A.E.R.E. R-3624 1961. 38 Roberts Adwick Rand Russell and Walter Geneva Conf. 1958 paper P/26. 40 Jackson and Rand U.K.A.E.A. Report A.E.R.E. R-3636. 41 Roberts and Walter personal communication. 42 Sellers Fried Elson and Zachariasen J. Amer. Chem. Soc. 1954,76,5935. Drummond and Welch J. 1957,4781. 450 QUARTERLY REVIEWS the tetragonal y phases in the uranium-oxygen system. The fluorite phases in the protactinium-oxygen system may therefore resemble the similar phases of the uranium-oxygen system but the precise course of the oxidation of protactinium dioxide is not yet clear; both analytical and density data are needed.(d) Related Mixed Oxide Phases.-The actinide dioxides form complete solid solutions with one another and with cerium dioxide and extensive cubic solid solutions with zirconium dioxide. The cell constant of the solid solutions varies linearly with the molar composition indicating that the cations are distributed at random on the cation s i t e ~ . ~ * * ~ ~ The oxidation properties of such solid solutions which contain uranium dioxide are similar to those of this oxide itself. Solid solutions of uranium dioxide in thorium dioxide have been most intensively studied ; chemisorp- tion is measurable at - 183 O on surfaces prepared by crushing and two kinetically distinct processes occur one at 0-25’ and one above 100”.Prolonged oxidation at high temperatures of solid solutions con- taining more than 60 moles % of uranium dioxide yields U,O,-like structures but only cubic phases are formed by oxidising solid solutions more dilute in this dioxide; these cubic phases contain oxygen in interstitial up to a limiting composition UYThl-,02.33 if y 2 0.5. The cell constant contracts at first on oxidation but passes through a minimum when the oxidation number of the uranium is about 5.0 and then expands as oxidation proceeds. The behaviour is very regular; the change in cell constant consequent upon oxidation per mole of uranium ions (i.e. Aa/y for U,Th,-,O + U,Th,-,O,,,) depends only on the oxidation number reached by the uranium and is independent of y.45 Identical values of Aa/y for a given oxidation number are found from Gronvold’s values2* for the U02+ phases when y = 1.At high temperatures equili- brium oxygen pressures can be measured over the oxidised cubic phases and the thermodynamic behaviour of the U,Th,-,O,+ system is very regular from y = 1 to y = 0.01 with the heat of solution of oxygen in the crystals falling as the uranium dioxide concentration falls below 50 %.38946 There is no indication that plutonium(1v) and neptunium(1v) show any more tendency to oxidise as components of solid solutions than as dioxide. Tho,-PuO and Ce0,-PuO solid solutions do not absorb extra oxygen when heated in air and Np0,-UO solid solutions behave similarly to Tho,-UO solid solutions the neptunium(1v) apparently acting as an inert diluent.The behaviour of these “regular” solid solutions of the MO structure towards oxygen thus confirms and extends the general properties of the 43 Mulford and Ellinger J. Amer. Chem. SOC. 3958 80 2023; Wolten ibid. 1958 80 4772; Mulford J. Phys. Chem. 1958 62 146; Rudorff and Valet 2. anorg. Chem. 1953 271 257; Duwez J. Amer. Ceram. SOC. 1957 40 321. 44 McConnell J. 1958 947. 45 Anderson Edgington Roberts and Wait J. 1954 257. 46 Aronson and Clayton J. Chern. Phys. 1960 32 749. ROBERTS THE ACTINIDE OXIDES 45 1 pure dioxides. Some new features are shown by the “anomalous” solid solutions of fluorite-structure formed by dissolving oxides of bivalent and tervalent metals in the actinide dioxides. Extensive solid solutions are formed with many of these oxides but not with oxides of metals of small ionic radius-magnesium beryllium aluminium.Always solid solutions are formed with an intact cation sub-lattice and charge balance maintained by vacancies appearing in the anion ~ub-lattice.~~ Oxygen can be incorporated to fill the anionic vacancies by forming uranium(v) or uranium(v1) in place of uranium(1v). Since the diminution of the lattice energy caused by the formation of vacancies must be con- siderable it is not surprising that some oxides such as magnesium oxide which do not form appreciable solid solutions with uranium dioxide can dissolve in the latter to about 35 moles % if the conditions are oxidising enough to enable the complete fluorite lattice M02 to be ~ e t a i n e d .~ ~ ~ There is a considerable tendency to form a complete M02 lattice even where a genuine uranium@) solid solution can be formed. Solid solutions of uranium dioxide and yttrium trioxide with the uranium as uranium(rv) could only be obtained at 1000” by finally reducing in sealed tubes with metallic uranium and could only be preserved in the complete absence of oxygen. The brown preparations absorbed oxygen rapidly in air at tempera- tures of -20” to +20” and eventually formed black fluorite-type phases with a contracted cell and a composition close to M02.0.49 This rapid bulk oxidation at room temperature and below occurred even with coarsely crystalline preparations and is entirely different from the limited oxidation of the outer layers of uranium dioxide at these temperatures.The cell constants for the U02-Y203 UO,-Y203-0 and MgO-U02-0 systems are summarised in Fig. 3 ; it may be noted that the cell constant for the last system at 33 moles % of MgO agrees with that given for MgU,O, a fluorite-type oxide prepared by the decomposition of magnesium uran- ate~.~O This is close to the solubility limit for magnesium oxide and the formal charge number of the uranium is 5.0. It is also extremely difficult to oxidise the solid solution Uo. 5Yo. 502 beyond this composition at which the oxidation number of the uranium is 5.0. These facts may indicate but do not prove the existence of uranium(v) in these compounds. These M02 structures can be oxidised beyond this composition and in air the uranium-rich materials may be further oxidised to U308-like structures or to U30 itself.The cubic solid solutions of U308 and another oxide are special cases of this general type and the oxidation number of the uranium cannot be assumed to remain 5-33 (equiv. to u308).51 The cubic 47 Lang Knudsen Fillimore and Roth N.B.S. Circular 568 1956; Budnikov Tresviatsky and Kushakovsky Geneva Conf. 1958 paper P/2193. 48 Anderson and Johnson J. 1953 1731. 49 Anderson Ferguson and Roberts J. Inorg. NucZear Chem. 1955,1,340; Ferguson and Fogg J. 1957,3679. 50 Hoekstra and Katz J. Amer. Chem. Soc. 1952,74 1683. 51 Hund and Peetz,Z. anorg. Chem. 1952,267,189; 271,6; Hund Peetz and Kotten- hahn ibid. 1955 278 184. 452 QUARTERLY REVIEWS oxides M02+z contain interstitial oxygen and the cube cell-edge contracts below the MO,. value but not as sharply as when anionic vacancies are being filled.Fig. 3 includes data for Np02-Y20 solid solutions; these can be seen to resemble the UO,-Y,O behaviour (curve No. 1) rather than that of the U,Y,,,O solid solutions (curve 2).38 The cell constants of the Np02-Y,03 solid solutions do not change when they are fired in air or reduced in contact with metallic uranium.52 The neptunium is then in the neptu- niumtrv) state and this experiment shows that the resistance of neptu- nium to oxidation in fluorite-type oxides does not depend on the space available and the contrast between the behaviour of these anion-deficient solid solutions of neptunium and uranium is even sharper than that between the dioxides of these elements themselves. I00 5 0 Moles % of U02 or Np02 FIG. 3. Cell dimensions of cubic phases 1 U02-Y01.a; 2 UyY1-,02; 3 U,Mgl-,02; 4 NP02-YOl.6.The quadrivalent actinide oxides are also capable of forming structures of the perovskite type (Fig. lb) with alkaline-earth oxides. Perovskite or distorted perovskite structures have been reported for CaUO, SrUO, BaU0,,47 BaTh03,53 and B~PuO,.~* The structure of the CaUO phase is open to doubt and it has bees proposed that it has the rare-earth “type C” structure-essentially a fluorite structure with an ordered arrangement of anion vacancies.55 Lower oxides M,O,-MO Here are found the closest analogies between the oxide systems of the actinide and the lanthanide elements. The rare-earth oxides commonly exist as the hexagonal “A” form or the cubic “C” form which is closely 52 Rand and Jackson U.K.A.E.A. Report A.E.R.E. R-3635. Smith and Welch Acta Cryst.1960 13 653. 54 Russell Harrison and Brett U.K.A.E. A. Report A.E.R.E. R-3044 1960. 65 Alberman Blakey and Anderson J. 1951 1352. ROBERTS THE ACTINIDE OXIDES 453 related to the fluorite structure; the oxides with larger cations have C-type polymorphic forms at low temperature^.^^ Known structures in the actinide series are as in Table 2. The sesquioxide Pu203 is prepared by TABLE 2. Ac2Q Type A a = 4.07 A c = 6.29 A Type A a = 3.84 A c = 5.96 A pu01.6 Type c Type A a = 3.82 A c = 5.97 A Type c Cm203 Type c a = 11.04A a = 11-03 A a = 11-00 A h2°3 Am203 reducing plutonium dioxide PuO with metallic plutonium in a closed crucible at 15OO” Am20 by reducing the dioxide with hydrogen and Cm203 by heating the product of the decomposition of the oxalate at 600” in a vacuum while curium dioxide is made by heating the last material in 1 atm.of oxygen and cooling the product slowly. The “type C” form of plutonium sesquioxide is not a polymorphic form of Pu,O but a separate oxide with a composition about PuO,.~,~’ that the ordered type-C structure can tolerate some additional oxygen has been demonstrated in several studies of mixed-oxide phases. Phases other than PuO, P U O ~ . ~ and the hexagonal Pu203 do not occur in samples of compo- sition Pu,03-PuO quenched from high temperatures but equilibration experiments show that one more stable phase of composition about PuO,., exists above 600°;58 here is another example of a very rapid phase transformation. The “A” form of americium sesquioxide was obtained by reduction above 800” and the “C” form by reduction at 600°.59 Decomposition of americium dioxide in a vacuum begins at about 700” and has been fol- lowed as far as Am01,85; one logp-l/Tplot showed a very distinct break at 1270” and a composition AmO1.,, but X-ray examination of the intermediate oxides has not been reported.60 The black oxide formed by igniting curium oxalate in air has the fluorite structure with a larger cell than that of the dioxide and is presumably oxygen-defi~ient.~ Although the facts reported are few it seems that the oxide systems between M203 and M02 do not resemble closely the corresponding rare-earth oxides where a series of ordered intermediate phases occurs at low temperatures with larger ranges of homogeneity at high tem- peratures.61 It is worth considering whether the radiation damage self- inflicted on the actinide oxides by a-decay may be expected to hinder the 5~ Roth and Schneider J.Res. Nut. Bur. Stand. 1960,64A 309. 57 Holley Mulford Huber Head Ellinger and Bjorklund Geneva Conf. 1958 58 L. E. Russell personal communication. 5B Templeton and Dauben J. Amer. Chem. Soc. 1953 75 4560. 6o Asprey U.S.A.E.C. Report UCRL-329 (Revised) 1949. paper P/701. Bevan J . Inorg. Nuclear Chem. 1955,1,49; Brauer and Gingerich Angew. Chem. 1957 69 480. 454 QUARTERLY REVIEWS formation of ordered phases. The half-lives of the isotopes concerned are 839Pu 2~4.10~; 2 4 1 ~ 470; 244Cm 18 years. Approximate calculation of the effect of a-particle decay in which the recoiling heavy nucleus dis- places many atoms from their rest positions by collision shows that the percentages of the atoms in the dioxides so displaced in one month are 0.1 % for 239Pu 5 % for 241Am and 100 % for 244Cm.62 Such results can hardly affect the structures of plutonium oxides but may be important in the curium-oxygen system.A series of double oxides containing plutonium(rI1) has been reported BaPuO, PuAlO, PuUO, and PuCrO,; all have the perovskite structure and their formation apparently stabilises the plutonium(II1) state since they are relatively readily formed under reducing condition^.^^ Oxides MO,. 5-M02. Triuranium octaoxide U30, is the oxide formed by the ignition of uranium compounds in air or 15 cm. of oxygen at temperatures below goo" followed by slow cooling. It is orthorhombic with a density 8.3 g./c.c. much lower than that of uranium dioxide. An analogous neptunium phase Np02.64f o.03 can be prepared by treating neptunium hydroxides with nitrogen dioxide at 300-450" or by heating neptunium(v) and neptunium(v1) compounds in air at 275-450°.37 A second form of Pa205 also white but orthorhombic and not cubic was prepared fortuitously during attempted fluorination of a protactinium oxide presumably by hydrolysis and was identified as Pa205 from its colour and by analogy with the structures of Nb20 and Ta205? These compounds must be quite closely related structurally; evidence from the decomposition of U30 and UO, and the heat treatment of U30, discloses a large family of orthorhombic or pseudo-hexagonal phases whose interrelations are very complex.Their elucidation is difficult for the reasons already discussed namely the difficulty of fixing oxygen positions in the presence of uranium atoms particularly when the true unit cell is very large and the need for precise measurement of composition and the attainment of equilibrium.The main orthorhombic phases are listed in Table 3 the small pseudo- cell being always given together with the true cell when this has been reported. No attempt has been made to include all the observations when agreement between those of different observers is reasonable. The data for the hexagonal a-UO phase are included for comparison. Zachariasen drew attention to the close similarity between the uranium positions in U30 and in the wU03 phase. This is brought out by Fig. 4a which shows the c axis projection of the assumed unit cell of U308 with the hexagonal unit cell of a-UO outlined (dotted).Along the c-axis perpendicular to the plane shown in Fig. 4a stretch endless chains of 62 D'Eye and Roberts U.K.A.E.A. Report A.E.R.E. C/M 306. 63 Sellers Fried Elson and Zachariasen J. Amer. Chem. SOC. 1954 76 5935; Holser Acru Cryst. 1956,9 196. ROBERTS THE ACTINIDE OXIDES 455 -U-0-U-0-atoms in a-UO, it is assumed that certain of these oxygen atoms between the uranium-oxygen layers are missing in U,O, these are indicated by a dotted ring. For true hexagonal symmetry the ratio of the axes a,lb of the equivalent orthorhombic pseudo-cell would be 43y i.e. 1.732 as is obvious from Fig. 4a the actual ratio for U,08 is 1-69. This TABLE 3. Cell dimensions of orthorhombic U,O,-like phases. Composition UP8 a (A) 6-717 6.723 6.878 7-05 6.815 6.75 1 6.73 6.579 6.92 b (A> 3.99 ( x 3) 3.975 ( x 3) (3.971) 3-81 ( ~ 3 ) (3.938) 3-97 ( x 8) 3-96 ( ~ 8 ) 4.080 4.02 c <A) a 4.150 4.143 ( x 2) 4-168 4.143 ( X 2) 4.136 4-143 (X2) 4.143 (>< 2) 4.182 4.18 Ref.67 65 66 66 5 37 63 b C a The factor by which a dimension of the pseudo-cell must be multiplied to give the dimension of the true cell is given in parentheses. Zachariasen Acta Cryst. 1948 1 265. Siegel Acta Cryst. 1955 8 617. structure was altered in detail by a neutron diffraction study by Andre~en.~~ In this structure (Fig. 4b) every uranium atom is bonded to an oxygen directly above and below and the oxygen atoms in the uranium sheets are -a - a 0 Oxygen 0 OXYg en @ Uranium @Uranium FIG. 4. c-Axis projections of (a) previously assumed unit cell of U308 (b) derived unit cell ofU30s. (Reproduced by permission from Acta Cryst.1958,11,612.) 64 Andresen Acta Cryst. 1958 11 612. 456 QUARTERLY REVIEWS rearranged giving 6 oxygen atoms surrounding an UI atom and 7 surround- ing an Un atom. Andresen points out that the dimensional changes in going from a-U03 to U308 refer to the a and b axes and not the c axis. The uranium-oxygen distance in the chains along the c axis is the shortest of all 2.07 A and these oxygen atoms are sometimes referred to as the "uranyl" oxygen atoms for reasons to be discussed. A rather similar structure was arrived at by Chodura and MalyYg5 from X-ray measurements of single crystals grown at 1 lOO"t5. The composition range of the U30s phase itself is narrow at low temperatures. There is general agreement that a closely related phase which we shall term the UOi.6 phase is formed at slightly lower oxygen content.The phase limits are not precisely known; they were determined66 as U02.56 to U02.65 on the basis of annealing U02-U308 mixtures in evacuated quartz tubes at 1200". It seems certain that a single crystal of this phase was studied by Rundle et d5 and showed the large unit cell given in Table 3. The P-U3O8 structure was obtained by heating crystals of U02.64 in oxygen at 750" and the trigonal U308 or U308-% phase by heating U30s in a closed capillary tube. The latter (trigonal) phase has a truly hexagonal arrangement of uranium atoms. These phases are probably related by different ordering in the uranium-oxygen sheets. The Np308 phase has been prepared apparently free of other phases with compositions of to NpO,.,,; it should probably be referred to as the N P O ~ .~ phase. Some superstructure lines have been observed and the true cell is probably large; the relation to the u30 structure could probably only be found by neutron-diffraction methods ; the ratio a/b for Np308 is low and the c value larger than in u308.37 Phases of this general type can certainly contain cations of more than one kind; they have been prepared by the oxidation of cubic mixed oxide phases (see p. 450) and by the reduction of some uranates. Uranium trioxides and the uranates Uranium is the only one of these elements to form oxides with all the metal atoms in the sexivalent state. At least five crystalline forms of uranium trioxide are known and the usual one made by heating the hydrated peroxide in air at 400" is amorphous.The preparation and inter- conversion of the various forms has recently been summarised by Hoekstra and Siegels7 and their notation is used here. The heats of formation of all five crystalline forms are very close ranging from -293 to -295 kcal./ mole; all other forms can be converted into the yellow y-form by heating in oxygen at 650'140 atm. ; the y-form itself is usually prepared by igniting uranyl nitrate hexahydrate in air at 400-600". 65 Chodura and Maly Geneva Conf. 1958 paper P/2099. 66 Hoekstra Siegel Fuchs and Katz J. Phys. Chem. 1955 59 136. 67 Hoekstra and Siegel Geneva Conf. 1958 paper P/1548; J. Inorg. Nuclear Chenz. 1961 18 154. ROBERTS THE ACTINIDE OXIDES 457 The crystal structures of only the a- and the &form are well known though the oxygen positions even in the a-trioxide have been deduced from steric considerations.The hexagonal crystal structure of this (a = 3.963 c = 4.16 A) has already been described. The 8-form is cubic with the ReO structure which is related to the perovskite structure (Fig. 1 b) by the omission of the central "A" cation. This form of uranium trioxide has been prepared by dehydrating the monohydrate in air at 415" and can apparently exist over a range of composition. The original report6* gave the composition as U02.82 with cell edge a = 4.146 & 0.005 A but the stoicheiometric form can be prepared if sufficient time is allowed. This is the only structure in which uranium is co-ordinated to six oxygen atoms with which it forms equivalent bonds; the &form can then be regarded as made up of three-dimensional U-0-U chains similar to those along the c axis of the hexagonal a-trioxide form.The structure of molydenum trioxide is somewhat similar.69 It is clear that the crystal structures of the other forms of uranium trioxide are considerably more complex; that of the stable y-form has been reported as orthorhombic and as mono~linic.~~ The decomposition of uranium trioxide to U308 is normally irreversible since re-oxidation can be effected only in high pressures of oxygen. Many intermediate phases have been identified the course of the decomposition depending on the type of trioxide and on the ambient oxygen p r e s ~ u r e . ~ ~ ~ The structures observed are related to the a-U03 and a-U,O structures. One well-characterised phase is the UO,, phase which is stable over a 150" temperature range in air; it is orthorhombic with a = 6.90 A b = 3.91 A c = 4.15 A which is the characteristic c-axis length for the U-0-U-0 chains.fiydrated Uranium Trioxides.-The product of the reaction between uranium trioxide and water at room temperature after air-drying is U0,,2H20. The thermal decomp~sition~~ of the dihydrate in air and in vacuum yields a-UO,,H,O of variable composition down to at least U03,0.8H20 and U03,0.5H20. /%UO,,E€,O was prepared by heating the anhydrous trioxide with water at 180" and by hydrolysis of uranyl acetate solution at l10°.73 These compounds have been termed uranic acids but the older nomenclature is adhered to here in order not to suggest a simple relation between the structures of these compounds and those of the uranates in which radicals such as UO,,- cannot be identified.However infrared spectra do indicate the presence of hydroxyl groups in U03,2H,0 and in a-U03,H20.74 68 Wait J. Znorg. Nuclear Chein. 1955 1 309. 6D Hagg and Magneli Rev. Pure Appl. Chem. (Australia) 1954 4 235. 'O Perio Bull. Soc. chim. France 1953,776; Connolly Acta Cryst. 1959,12,949. 71 Wait U.K.A.E.A. Report A.E.R.E. R-3623. 72 Hiittig and von Schroeder 2. anorg. Chem. 1922 121 243; Gentile Talley and Collopy J. Znorg. Nuclear Chem. 1959 10 114; Dawson Wait Alcock and Chilton J. 1956 3531. 73 Bergstrom and Lundgren Acta Chem. Scand. 1956 10 673. 74 Deane U.K.A.E.A. Report A.E.R.E. R-3411. 458 QUARTERLY REVIEWS The structure of only the P-UO3,H2O phase has been described in detail.73 It is orthorhombic a = 6.29 b = 5.64 c = 9.93 A; the uranium atoms lie on a face-centred array forming UO,(OH) layers with linear UO groups perpendicular to the sheets.This appearance of uranyl (UO,) groups as distinct from U-0-U chains is repeated in the uranate structures discussed below. The Uranates.-Uranates have been prepared by addition of alkali to aqueous solutions of uranyl salts by the decomposition of double salts by heating the oxide U03 or U308 with oxides hydroxides or carbonates and by the reaction of U308 with alkali or alkaline-earth chlorides in the air. The last reaction has been used for making single crystals. All the uranates are insoluble in water. They are more stable thermally than uranium trioxide; thus MgU2O7 is stable to 850" in a vacuum and Li2U04 to 1300" in air although the most stable form of uranium trioxide de- composes at 700" in air.Recent studies by Russian authors and by J. S. Anderson and his collaborators have considerably extended our knowledge of the alkali- metal uranate~.~~ Polybasic uranates such as Li4U05 are most easily prepared for the lower members of the series and the more complex polyuranates are more easily prepared the larger the alkali-metal atom. Cs,U,O, and Rb,U6OI9 are known but the lithium series apparently stops at Li2U3Ol0. Many structural data exist for the alkali-metal uranates for the uranates of Group I1 elements and for lead uranates. Detailed structural results have been published for CaUO, MgU04,76 and BaU04.77 The CaUO structure (Fig. 5) is related to that of a-uranium FIG. 5. Layers and chains of atoms observed in uranate structures shown in somewhat idealised form (a) CaUO,; (b) BaUO,; (c) MgUO,.0 U atom with OI above and below; 0 On. (Reproduced by permission from Acta Cryst. 1954,7 795.) 75 Efremova Ippolitova Simanov and Spitsyn Proc. Acad. Sci. (U.S.S.R.) 1959,124 115; Kovba Ippolitova Simanov and Spitsyn ibid. 1958 120 465; J. G. Allpress Thesis Melbourne 1960; D. G. Kepert M.Sc. Thesis Melbourne 1960. 76 Zachariasen Acta Cryst. 1948 1 281; 1954 7 788. 77 Samson and SillCn Arkiv Kemi Min. Geol. 1947 25 A 1. ROBERTS THE ACTINIDE OXIDES 459 trioxide; each uranium atom is bonded to two “primary” 01 atoms perpendicular to the layer shown in Fig. 5 and to the six “secondary” OII atoms shown which are alternately above and below the plane illus- trated the distances being U-01 = 1.91 A U-011 = 2-29 A. The UO2(O2) layers form infinite nets with the calcium atoms between them; uranyl groups can be identified instead of the U-0-U chains of a-U03.One form of CdUO is similar,78 and also one form each of LizU04 Na2U0, and K2U04. More analogies with other structures can be drawn when uranium has the more usual co-ordination number of six. This is so in the BaUO structure which is pseudotetragonal with the uranium co-ordinated by two 01 and four 011 atoms the oxygen octahedra sharing corners to make infinite nets of the type shown in Fig. 5 with barium atoms between the (U02)02 layers. Similar structures have been reported for one poly- morphic form of the alkali-metal mon~uranates~~ and for PbU04.79 In the a-Li2U04 structure which is similar to that of K2NiF4 and other oxides such as La2Ni0, the lithium ions lie between the 01 atoms and U-01-Li- Li-Or-U strings are formed along the c-axis of the tetragonal cell.In the MgU0 structure each uranium atom is again bonded to two 01 and four On atoms but the oxygen octahedra share two edges to form infinite strings (Fig. 5); one form of CdU0,is similar.78 The U-01 and U-On distances in these compounds are not constant but fall into a regular sequence when the number of bonds made to a given atom is taken into account.80 Zachariasen has suggested that polyuranates may be formed by sharing the 01 atoms which lie above and below the orthouranate layers or chains and a structure in which this occurs has been suggested for MgU3OIo and ZnU30,0.78 At high temperatures or under mildly reducing conditions the uranates lose oxygen and other structures are found.Compounds formally involving uranium(v) have been reported to have the perovskite structure (e.g. NaUO, KU0375) where uranium-oxygen chains are formed in all directions. This structure is often found for AB03 compounds when the A2B04 compound has the K,NiF structure. U308-like compounds orthorhombic with the characteristic c-spacing of 4.14 A have been observed in the decomposition of alkali-metal uranates and in the CdO- U03 system.78 Finally under more reducing conditions fluorite phases are often formed of the type already described (p. 451). The uranate structures show clearly the tendency to form collinear symmetrical MOz groups common to all the actinides from uranium to americium but to few other elements. No pure oxide phase of this type has been described for elements other than uranium though Np,08 is Ippolitova Simanov Kovba Polonina and Beresnikova Radiokhimiya 1959 6 660.7Q Frondel and Barnes Acta Cryst. 1958 11 562. so Zachariasen and Plettinger Acta Cryst. 1959 12 526. 460 QUARTERLY REVIEWS probably a related structure and the oxyfluoride KAmO,F, has the CaUO structure with (ArnO,)+ groups.81 In view of the extra stability of the uranates over uranium trioxide it is possible that neptunates and plutonates could be made. The author thanks many friends and colleagues for information and discussions and in particular Professor P. L. Robinson and Dr. E. Wait for reading this manuscript. *l Asprey Ellinger and Zachariasen U.S.A.E.C. Report AECD-35-54.

ISSN:0009-2681    

DOI:10.1039/QR9611500442

出版商:RSC    

年代:1961

数据来源: RSC