By D. H. R. BARTON D.Sc. F.R.S. (THE UNIVERSITY GLASGOW) and R. C. COOKSON M.A. PH.D. (BIRRBECK COLLEGE LONDON W.C.1) Introduction.-The word “ conformation ” first used by W. N. Haworth,l can be defined in several ways. One of the most general definitions is as follows the conformations of a molecule are those arrangements in space of the atoms of the molecule which are not superposable upon each other. Such a definition includes arrangements of atoms in which angle strain has been introduced though this is not normally of great importance. Thus a’lmost all molecules will have theoretically an infinite number of con- formations. It is fortunate that the complexities which might arise from such considerations are minimised by the fact that in general only a few of the possible conformations are energetically preferred.I n the German language the word “ constellation ” of more recent origin,2 is used in the same sense,3 and in the literature of chemical physics the term “rotational isomer ”. For the vast majority of molecules the energy barriers between different conformations are too low to allow the separation of pure conformational isomers a t normal temperatures in the liquid or vapour phase but sufficient intramolecular congestion can raise the barrier enough to make such a separation possible experimentally. Familiar examples are provided by resolvable diphenyl derivative^.^ Tri-o-thymotide has recently been shown to undergo spontaneous resolution into conformational enantiomorphs of low optical ~tability.~ The Preferred Conformations of Some Simple Hydrocarbons In the last two decades many of the powerhl methods of modern chemical physics including electron and X-ray diffraction infrared Raman and microwave spectroscopy and statistical mechanics have been used to demonstrate the existence and nature of preferred conformations in simple W.N. Haworth “ The Constitution of Sugars ” E. Arnold and Co. London 1929 p. 90. Ebel “ Stereochemie ” Ed. Freudenberg Deuticke 1932 p. 825. Prelog J . 1950 420. Turner and Harris “ Organic Chemistry ” Longmans Green and Co. pp. 605 et seq. ; Shriner Adams and Marvel “ Organic Chemistry ” ed. Gilman 2nd edn. Wiley 1943 Vol. I pp. 343 et seq. 5 Baker Gilbert and Ollis J. 1952 1443 ; Powell Nature 1952,170,155 ; Newman and Powell J . 1952 3747. For examples of stable conformational isomers in other ring systems see inter al.Bentley and Robinson J . 1952 947 ; Bell ibid. p. 1527 ; Wittig and Zimmermann Chem. Ber. 1953 86 629 ; Hall and Turner J . 1955 1242 and references given there. 44 BARTON AND COOKSON CONFORMATIONAL ANALYSIS 45 molecules. For more detailed summaries of this background the reader is referred to the reviews by McCoubrey and Ubbelohde and by Mizushima.' Ethane.-Fig. 1 illustrates the energy of a molecule of ethane as a function of conformation in this case depending only on the relative orienta- tion of the two methyl groups. The molecule has the maximum energy FIGS. 1 and 2 I H..'- HyH "'" H (HI (1) 0" ( 2 ) 60" Views down the C-C bond of ethane when the set of three hydrogen atoms attached to the near carbon atom eclipse those attached to the far carbon atom when viewed down the C-C bond (1).The energy is at a minimum when each C-H bond bisects the angle formed by two C-H bonds of the other carbon atom (2). The barrier to rotation is estimated n-Butane.-The conformation of n-butane with lowest energy is the fully transoid staggered conformation (3). The two other minima in the potential as about 2.8 kcal. per mole. r! M C M C Fully eclipsed Skew or gauche energy curve (Pig. 2) correspond gauche conformations (4). It has (Me) (5) (4) Me to the Me Me (H) Eclipsed Staggered two enantiomorphous skew or (6) (3) been calculatedg that the energies of McCoubrey and Ubbelohde Quart. Rev. 1951 5 364. 7 Mizushima " The Structure of Molecules and Internal Rotation ' I Academic * Inter al. Kistiakowsky Lacher and Stitt J. Chem. Phys. 1939 7 289 ; Pitzer Pitzer Discuss.Furuday SOC. 1951 10 66 and references there cited. Press 1954. Chenz. Rev. 1940 27 39 ; McCoubrey and Ubbelohde ref. 6. 46 QUARTERLY REVIEWS the fully eclipsed (5) eclipsed (6) and skew (4) conformations are about 3.6 2.9 and 0.8 kcal. per mole respectively greater than that of the staggered conformation (3). The last value is in excellent agreement with the tempera- ture-dependence of the appropriate Raman lines. lo I n aliphatic compounds the most stable conformation is usually that in which the substituents on adjacent tetrahedral carbon atoms adopt the fully staggered conformation [as (3)j the two largest groups (or in qualification the two most strongly repelling dipoles) taking up the 180" arrangement. Eclipsed conformations are always avoided wherever possible.cyc1oHexane.-The conclusion by Sachse 11 and by Mohr 12 that cyclo- hexane can exist in only two conformations free from angle-strain has long been accepted by chemists. That the chair conformation (7) is more stable than the boat (8) is attested by much physical evidence including infrared l3 and Raman l4 spectroscopy and electron diffraction,15 and by thermo- dynamic considerations.16 17 Derivatives of cyclohexane always tend to take up the chair conformation whenever this is stereochemically possible (see detailed discussion later). The Decalins.-If boat conformations are accepted as less stable than chair conformations then both cis- and trans-decalin have unique preferred eonformations illustrated in (9) and (10) respectively. Hassel and his collaborators 1 5 l 8 9 l9 have shown by electron diffraction that in the vapour phase molecules of cis-decalin do indeed exist in the two-chair conformation (9) rather than in the long unchallenged two-boat conforma- lo Szasz Sheppard and Rank J .Chem. Phys. 1948 16 704. 11Sachse Ber. 1890 23 1363; 2. phys. Chem. 1892 10 203. 12 Mohr J. prakt. Chem. 1918 98 315. 13 Rasmussen J. Chem. Phys. 1943 11 249 and papers there cited. 1 4 Kohlrausch and Wittek 2. phys. Chem. 1941 48 B 177 ; Gerding Smit and l5 Hassel and Viervoll Acta Chena. Xcand. 1947 1 149 and papers there cited. Is Aston Schumann Fink and Doty J . Anzer. Chem. SOC. 1941 63 2029. 1 7 Beckett Pitzer and Spitzer ibid. 1947 69 2488. 18 Hassel Tidsskr. Kjenzi Bergvesen Met. 1943 3 91. 19 Bastiansen and Hassel ibid. 1946 6 70 ; Nature 1946 157 765.W'estrik Rec. Trav. chirn. 1942 61 561. BARTON AND COOKSON CONFORMATIONAL ANALYSIS 47 tion (11).12 the conformation (9) for subst,ituted cis-decalins.20 2 1 y 22 23 Much chemical evidence has also been adduced in favour of Reasons for the Existence of Preferred Conformations The interaction energy between two non-bonded atoms is weakly attrac- tive up to distances of approximat,ely the sum of the van der Waals radii of the atoms concerned. At distances less than this the interaction energy soon becomes repulsive and a high exponential function of interatomic distance. 24 In the fuily staggered conformation of a straight-chain aliphatic compound the distances between non-bonded atoms approximate quite closely to the sum of the van der Waals radii. Any rotation about a C-C bond (see above) brings the atoms attached to the main chain closer together and calls into play destabilising (repulsive) forces.In general it can be said that the energy of a particular conformation of a molecule depends on the presence or in specially formed cases (see above) the absence ofrepulsive interactions between non-bonded atoms. To a certain extent these repulsive interactions can be modified by bond-angle deformation. Modification by bond extension is not important because of the relatively large energy requirements for such a process. In the absence of direct physical evidence the most stable conformation of a molecule may then be selected by the principle of minimisation of (repulsive) non-bonded interactions. We would emphasise however that simple considerations of this kind are subject to qualification when other intramolecular forces due to hydrogen bonding or electrostatic effects (dipole or integral change interactions) come into play (see p.81). Consequences of the Existence of Preferred Conformations In the last five years the importance of the existence of preferred con- formations in organic molecules has become widely recognissd under the title of " conformational analysis ',. The present Review sets out to sum- marise with the aid of the appropriate background the more important aspects of the subject from the point of view of the organic chemist.25 The fundamental tenet of conformational analysis is that the physical and chemical properties of a molecule can be related to its preferred con- formation. The more important applications of this concept may be classified as follows (i) Phenomena that are a direct consequence of a preferred conforma- tion (a) Physical properties such as specific absorption bands in the 2o Barton and Miller J .Amer. Chem. SOC. 1950 72 1066. 21 Barton Experientia 1950 6 316 ; J. 1953 1027. 2 2 Beyler and Sarett J . Anzer. Chem. SOC. 1952 74 1406. 23 W. G. Dauben Tweit and Mannerskantz ibid. 1954 76 4420. 2 4 See for example C. K. Ingold " Structure and Mechanism in Organic Chemistry " Cornell Univ. Press Ithaca New York 1953 Chapter I1 ; Hughes Quart. Rev. 1948 2 107. 2 5 For more fully documented accaunts of certain phases of the problems see Orloff Chem. Reu. 1954 54 347 and Klyne in " Progress in Stereochemistry " Butterworths London 1954 Vol. I Chapter 11. 48 QUARTERLY REVIEWS ultraviolet or infrared region.( b ) Chemical effects dominated by steric com- pression such as ester hydrolysis (reaction rates) and relative stabilities of epimers (equilibria). (ii) Phenomena due to the interplay of conformational preferences and the geometrical requirements of the transition states of reactions (usually colline- arity or coplanarity of participating centres). Some Illustrations from Aliphatic Chemistry The Relative Stabilities of erythro- and threo-Isomers.-Let us consider the non-bonded interactions in an erythro-compound (12) as compared with those in its threo-diastereoisomer (13). In these and later formulz S M and L denote substituents attached to the same carbon atom which are \ c I I S-t-M I S-F-M I S-C-M i M-C-S I I i SYM I My k M y 5 ys y ... . I . .- .- -. ,. .. * -. .. - . ' S M**"*' "= "S s. "L L,'" ''.'M M/-*' *.. L L L L (14) (15) (16) (17) respectively smallest medium and largest in effective size. Each substituent will repel the two adjacent substituents on the next carbon atom; the sum of these repulsive energies for the stable staggered conformation (14) of the erythro-compound and for the three more stable conformations (15) (16) and (17) of the threo-isomer may be represented as follows (14) 2(L S) + 2(L M) + 2(M S) (15) 2(L M) + 2(L S) + S S + M M (16) 2(L:M) + 2(M:S) + L L + S S (17) 2(L S) + 2(M S) + L L + M M The differences in compression energy between the stable conformation of the erythro-compound and the three more stable conformations of the threo-compound are then (14) - (15) = 2(M S) - (M M + S S) (14) - (16) = 2(L (14) - (17) = 2(L M) - ( L L + M M) S) - ( L L + S S ) Since repulsion is a high exponential function of interatomic distance (M M) > (M S) (L L) > (L S) and (L L) > (L M) so that in each case (14) - (15) (14) - (16) and (14) - (17) should be a negative quantity.BARTON AND COOKSON CONFORMATIONAL ANALYSIS 49 In general then for diastereoisomeric pairs of non-polar compounds in which differences in free energy are mainly due to differences in compression energy the erythro-isomer should be more stable than the threo-isomer. The relative stabilities have been established of many pairs of meso- and racemic and of erythro- and threo-isomers either by direct equilibration or by introduction of the second asymmetric carbon centre in a reaction that is known to result (from its mechanism) in a mixture approximating to the equilibrium mixture.In many cases the expected stability order is found experimentally. For example equilibration 26 of the racemic succinic acids (18 ; R = R’ = alkyl aryl halogen or OH) gives the more stable meso-acids (19 ; R = R’). Similarly the threo-acids (18 ; R + R’ = alkyl aryl halogen or OH) isomerise to the erythro-acids (19 ; R + R’). Racemic stilbene dichloride 27 and dibromide 28 isomerise to equilibrium mixtures composed chiefly of the meso-dihalides. In many examples of this kind however the situation is complicated by dipole interactions which also favour the isomers which should be more stable from the conformational point of view. 502H 702H I I Cram and Abd Elhafez 29 have tabulated the products of reduction of ketones (20 ; R = aryl R’ = OH or NH,) and of oximes (21 ; R = aryl R’ = OH or NH,) with sodium and alcohol or with sodium amalgam reagents that are accepted as producing the more stable e~imer.~O The erythro(or meso)-isomer (22 and 23) is always formed in greater amount.p R R p I I I I I I I I R’-C-H R’-C-H R’-~-H R’-~-H l - I I C z N O H H2N-$-H I - I C-0 HO-C-H R I ! R k R (20) (22) (21) (23) 26 Fourteen examples are tabulated by Wagner-Jauregg (quoting Wolf) in “ Stereo- chemie ” ed. Freudenberg Deuticke 1932 p. 873 ; see also Linstead and Whalley J. 1954 3722. 27 Zincke Annalen 1879; 198 135. 28 Wislicenus and Seeler Ber. 1895 28 2693 ; Buckles Steinmetz and Wheeler J. Arner. Chem. SOC. 1950 72 2496 ; Abd Elhafez and Cram ibid.1953 75 339. 20 Idem ibid. 1952 74 5828. 30 Barton and Robinson J. 1954 3045. D 50 QUARTERLY REVIEWS Rates of Elimination from erythro- and threo-Compounds.-In the debromination of 1 2-dibromides by iodide ion symbolised below,31 the meso-dibromide (24) always 32 reacts more rapidly than the DL-isomer (25). Since the transition state of an E2 reaction such as this requires that the four centres concerned in the reaction should be coplanarY2OJ 3 3 ~ 34 with the two C-Br bonds antiparallel the conformation required for the meso- dibromide is (24) that for the m-dibromide (25). In the transition state M i truns-But - 2 -ene M cis- But - 2 -ene arising from (24) each methyl group is becoming eclipsed by a hydrogen atom whereas in that arising from (25) the two methyl groups are becoming eclipsed.The transition state is therefore of higher energy with respect to the ground state for (25) than it is for (24). Correspondingly meso-stilbene dibromide is debrominated by iodide ion about one hundred times as fast as the racemic isomer.32~ 35 Cram's Rule of Asymmetric Induction.-From consideration of the pro- ducts of proved configuration resulting from reactions involving addition to a carbonyl group adjacent to an asymmetric carbon atom Cram and ' Abd Elhafez 29 were able to propose the following rule " In reactions of the type (26) + (27) that diastereoisomer will predominate which would be 0 0 2 formed by the approach of the entering group [R'] from the less hindered side of the double bond [of the carbonyl group] when the rotational con- formation of the C-C bond is such that the double bond is flanked by the 31 Winstein Pressman and Young J .Amer. Chern. Xoc. 1939 61 1645. 32 Young Pressman and Coryell ibid. p. 1640. 3 3 Young Abs. Papers 8th Nat. Org. Chem. Symposium Amer. Chem. Soc. St. 34 Dhar Hughes Ingold Mandour Maw and Woolf J. 1948 2711. 35 For a pertinent case involving dehydrochlorination see Cram and Abd Elhafez Louis Dec. 1939 p. 92. J . Amer. Chem. SOC. 1952 74 5851. BARTON AND COOKSON CONFORMATIONAL ANALYSIS 51 two least bulky groups [S and MI attached to the adjacent asymmetric centre.’’ Thus in the addition of Grignard reagents or reduction by lithium aluminium hydride the carbonyl-oxygen atom being co-ordinated to the metal atom (MgX or AlH,) becomes effectively the largest group and thus orients itself between S and M.The approach of the R’ group is then directed as would be expected from that side of the molecule to which S is attached rather than M. Prelog’s Approach to Asymmetric ,Synthesis.-Prelog and his collabora- tors 36 have submitted to conformational analysis the extensive experimental results of McKenzie and his school on the configurational course of the addit’ion of Grignard reagents to phenylglyoxylic esters of asymmetric alcohols and have been able to relate the sign of rotation of the atrolactic acid produced on hydrolysis with the configuration of the original asymmetric alcohol. The most stable conformation of the phenylglyoxylic ester was considered to be (28) in which the two carbonyl groups are planar and anti to each other with the two larger groups of the asymmetric alcohol (29) skew to the ester-carbonyl oxygen.Addition of methylmagnesium halide will be more HO I I HO-5-S I c (29) 502H I Me-$ -OH Ph I I rapid from the less hindered side of the carbonyl group leading to a pre- ponderance of (30) over its diastereoisomer. Thus the formation of a partially racemic but lzvorotatory atrolactic acid showing an excess of the enantiomorph (31) indicates that the alcohol has the absolute configuration represented by (29). Conversely a dextrorotatory atrolactic acid indicates that the parent alcohol has the opposite configuration to (29). This method has been applied 36 to the determination of the absolute configuration of several groups of natural products. Relative Rates of Cyc1isation.-The rate of any reaction proceeding 36 Prelog Helv.Chirn. Acta 1953 36 308 ; Prelog and Meier ibid. p. 320 ; W. G. Dauben Dickel Jeger and Prelog ibid. p. 325 ; and later years. 52 QUARTERLY REVIEWS through an approximately planar transition state or intermediate whatever the ring size and whether or not the final product is cyclic will in general be smaller for the diastereoisomer reacting through the transition state or intermediate where large groups are eclipsed than for the isomer where smaller groups are eclipsed. ( a ) 1 2-cis-Cyclisations. Thus the reaction of form (32) via (33) will be slower than the reaction of form (34) via (35).37 Well-authenticated examples of this effect in a five-membered transition state are the rates of (32) (33) X L' s' (34) (35) acyl migration in 1 2-a~mino-alcohols such as erythro- (36) (slow) and threo- (37) (Fa&) ephedrine 389 39 and the slower rate of condensation with acetone of 9h I HO-C-H MeNH-q -H I I 1 de (36) meso- (38) than of racemic (39) dihydroben~oin.~~ The readier 9h 9 h Ph I I I HO-C-1 I HO-C-H HO-C-H H-7- I HO-$-H d e i h 6h I H-$ -NHMe I I I I I I I (37) (38) ? 9h I HO-C-H I HO-C-H Me-(:-H I H-G-Me I I 6 h I Gh (39) 37 Some of the links in the ring may be partial bonds.Z may represent one or more ring atoms (or conceivably none) X and Y are atoms or groups of the type referred to in the cursive text. 38 Welsh J . Arner. Chem. SOC. 1947 69 128 ; 1949 '71 3500. 39 Fodor Bruckner Kiss and Ohegyi J . Org. Chern. 1949 14 337 ; see also Close 40Hermans 2. phys. Chem. 1924 113 337. J . Orq. Chem. 1950 15 1131. BARTON AND COOKSON CONFORMATIONAL ANALYSIS 53 pyrolysis 29 of the xanthate of threo- (40) than of erythro-1 2-diphenyl- propan-1-01 (41) illustrates the operation of the effect in an approximately planar six-membered transition state.The relative rates of cyclisation of erythro- and threo-isomers may be reversed when the transition state is non-planar. Wendler 41 has suggested that the diastereoisomer of 3-acetamido-1 3- diphenylpropan-1-01 which undergoes N + 0 acetyl migration in acid solution has the configuration (42) on the grounds that the intermediate (43) in the migration suffers from no serious non-bonded interactions. On the other hand the other diastereoisomer (44) does not undergo migration 42 because this would involve an intermediate (45) with a strong non-bonded repulsion between phenyl and hydroxyl (or methyl) groups.43 (b) 1 3-cis-Cyclisations.H H (44) (45) In a trans-cyclisation with displacement of an adjacent substituent through a planar cyclic transition state symbolised in (34) + (46) as a general formulation of neighbouring-group partici- pation,44 the diastereoisomer (34) which reacted more rapidly in the cis- (c) 1 2-trans-Cyclisations. eyclisation now passes into a product (46) where the large groups are eclipsed. It reacts therefore more slowly than its isomer (32) where the product has only eclipsing of the large and the small groups. The reality of this effect for a three-membered transition state or inter- 4l Wendler Experientia 1953 9 416. 42 Stiihmer and Frey Arch. Phurm. 1953 286 8. 43 The hydroxyl group and one of the phenyl groups of the oxazine (45) are both 44 Winstein Morse Grunwald Schreiber and Come J .Arner. Chern. SOC. 1952 axial (see p. 55). 74 1113. 54 QUARTERLY REVIEWS mediate has been demonstrated in relative rates of phenyl-group migration (through phenonium-ion intermediates) for appropriate pairs of diastereo- isomers.44 45 Curtin in particular has emphasised the role of such a " cis- effect " in controlling the relative rates of competing pinacolic aryl migrations of 1 2-diary1 systems.46 To emphasise the contrast between cis- and trans-1 2-cyclisations let; us consider the reactions of the two diastereoisomeric 2-amino-1 2-diphenyl- ethanols. Thionyl chloride cyclises the N-formate of the erythro-isomer (47 ; R = H) via the chlorosulphite and with inversion to the oxazoline salt (48 ; R = H) under conditions that do not affect the formate (49) of H Ph .py-i Ph H 0 soc12 * H Ph pheNH+ " 0 4 (49) (50) the threo-isomer.*' On the other hand the threo-N-acetate (49 ; R = Me) undergoes N +- 0 acetyl migration through cis-cyclisation to the oxazolidine (50) on treatment with alcoholic hydrogen chloride conditions which leave the N-acetate (47 ; R = Me) of the erythro-isomer unchanged.39 Some Illustrations from Alicyclic Chemistry Examination of a model of the chair conformation of cyclohexane shows Six of the that the C-H bonds are of two geometrically different types.axis 4 5 Cram ibid. p. 2152. 46 Curtin and Crew ibid. 1955 77 354 and earlier papers ; for a review see Curtin 47 Weijlard Pfister Swanezy Robinson and Tishler J . Arner. Chem. SOC. 1951 For related transformations see inter &a Elliott J.1949 589 ; 1950 62 ; Rec. Chem. Progr. 1954 15 111. 73 1216. Ann. Reports 1953 50 277. BARTON AND COOKSON CONFORMATIONAL ANALYSIS 55 bonds lie parallel to the threefold symmetry axis of the ring as in (51) and have been designated “ axial ”.48 The other six bonds radiate out from the ring as in (52) and have been named “ equatorial ”.48 In the chair conformation of cyclohexane each equatorial hydrogen atom is skew to the four hydrogen atoms on the two adjacent carbon atoms and about 2-5 A distant from each of them [l 2-H H interactions ; see (53)J Each axial hydrogen atom is flanked by two equatorial hydrogen atoms attached to the adjacent carbon atoms in the same geometrical relation [l 2-H H interactions ; see (54)l. Each axial hydrogen atom is also about 2-5 ,& away from the other two axial hydrogen atoms on the same side of the ring [l 3-H H interactions ; see (54)l.All other H H or C H interactions are relatively unimportant. H.. (53) (54) Preferred Conformations of Simple cycloHexane Derivatives.-A study of accurate scale models shows that any substituent larger than hydrogen in an axial conformation is closer to the two axial hydrogen atoms (1 3-inter- actions) on the same side of the ring than the same substituent in an equa- torial conformation is to the adjacent equatorial and axial hydrogen atoms (1 %interactions). In consequence the stronger (repulsive) 1 3-interactions R (59) (60) (61) (62) dominate the energy relations and a substituent in general prefers to adopt an equatorial rather than an axial conformation.The extensive investigations of Hassel and his collaborators 49 on the 48 Barton Hassel Pitzer and Prelog Nature 1953 172 1096. “ Equatorial ” and “axial ” replace the earlier synonyms K and E (Hassel Tidsskr. Kjemi Bergvesen Met. 1943 3 32) and “equatorial ” and “polar ” (Beckett Pitzer and Spitzer J . Amer. Chem. SOC. 1947 69 2488). 49 Reviews Hassel and Ottar Acta Chem. Scand. 1947 1 929 ; Hassel Research 1950 3 504; Quart. Rev. 1953 7 221. 56 QUARTERLY REVIEWS electron diffraction of cyclohexane derivatives in the vapour phase-con- ditions where intermolecular interactions are a t a minimum-have revealed that a cyclohexane derivative normally exists predominantly in the chair conformation which has the maximum number of substituents equatorial. Thus monosubstituted cyclohexanes are to be represented 50 as (55) rather than (56) truns-1 2-disubstituted cyclohexanes as (57) rather than (58) cis-1 3-disubstituted derivates as (59) rather than (60) and trans-1 4- compounds as (61) rather than (62).Such alternative chair conformations can be interconverted by merely passing the cyclohexane ring through a planar or equivalent conformation. The interconversion of the two chair conformations through such an inter- mediate having angle-strain and higher non-bonded interactions generates an energy barrier between the two of only a few kcal./mole too small to allow their separation in the liquid or vapour state by the classical methods of organic chemistry. Since rates of thermal equilibration of conformational isomers are therefore generally much greater than rates of chemical reactions a reaction proceeding by a mechanism involving a transition state having geometrical requirements better satisfied by a less stable conformation may well follow a path mainly through that less stable c~nformation.~~ Mono- cyclic cyclohexane derivatives 52 are therefore from the point of view of reaction mechanism always subject to conformational ambiguities.Con- densed cyclohexane systems discussed below in which ring conversions are more difficult or geometrically impossible do not suffer from the same objections. Winstein and Holness 6 3 have recently ensured conformational homogeneity in an ingenious manner by using cis- (63) and trans-tert.- butylcyclohexanol (64). The very bulky tert. -butyl grouping guarantees the absence of any significant proportion of conformations with this group axial.Condensed cyclo Hexane Ring Systems.-The most stable conformation of a cyclohexane ring system is that with the greatest number of chair rings (see above). In most cases this principle enables an unambiguous con- formation to be deduced for such systems. Thus trans-anti-trans-per- hydrophenanthrene is to be represented as (65) the trans-a/B steroids (66) as (67),209 21* 5 4 and oleanane (68) the most important parent hydrocarbon 60 These and similar formulae in the present article are stylised representations of the true conformations. They are not intended to be accurate perspective or orthogonal projections. 5l Eliel Experientia 1953 9 91. 5 2 For an excellent review of monocyclic cyAohexane derivatives see Orloff Chem.5 3 Winstein and Holness J . Amer. Chem. SOC. 1955,77 5562. We are most grateful 6 4 Barton and Rosenfelder J . 1961 1048. Rev. 1954 54 347. to Professor Saul Winstein for informing us of his results before their publication. 57 of the pentacyclic triterpenoids as ( 69).55 All these conformational assign- ments are quite unambiguoixs and are supported by a wealth of chemical evidence. 56 BARTON AND COOKSON CONFORMATIONAL ANALYSIS R H A H H I g d d H I A ti An interesting demonstration 57 58 of the destabilising effect of boat conformations is provided by the fact that trans-syn-cis-perhydroplien- anthrene (70) where all rings can adopt chair conformations is more stable than the trans-syn-trans-isomer (71) where one ring is forced to assume a boat conformation. 55 Barton and Holness J .1952 78 ; Barton J. 1953 1027. 66 Conformations deduced from minimisation of non-bonded interactions refer ideally only to isolated molecules the conformations of which are entirely controlled by intra- molecular forces. In the crystal lattice it is conceivable that intermolecular forces could become dominant and that different conformations might be favoured. Fortu- nately in most cases it appears that the ideal conformation of the isolatedmolecule is also preserved in the crystal. For impressive examples see Carlisle and Crowfoot Proc. Roy. Sac. 1945 A 184 64 (cholesteryl iodide) ; Fridricksons and Mathieson J. 1953 2159 (lanostenyl iodoacetate) ; Carlisle and Abd El Rehim Chem. and Ind. 1954 579 (methyl oleanolate iodoacetate). 67 Linstead and Whetstone J. 1950 1428.6* Johnson Experientia 1951 7 315. 58 QUARTERLY REVIEWS Calculation of Energy Differences between Geometrical Isomers.-The calculation of energy differences by semiempirical procedures 59 can be extended to provide the relative stabilities of conformational isomers.60 The calculations led to the observed stability orders of chair- > boat-cyclo- hexane 2-chair trans-decalin > %chair cis-decalin 2-chair cis-decalin > %boat cis-decalin. A more convenient method has been developed,61 based upon Pitzer's treatment 62 of the methylcyclohexanes. This simple empirical approach assigns an interaction energy of 0.8 kcal./mole to each skew interaction as in n-butane (4). cis-Decalin for example has three skew interactions that do not occur in trans-decalin and is therefore ca.2-4 kcal./mole less stable in excellent agreement with experiment .63 The calculations have been extended 61 to include perhydro-phenanthrenes and -anthracenes and the calculated relative stabilities of some of the isomers have been confirmed experimentalIy.6lS G4 Reaction Rates and Equilibria controlled by Steric Compression.-(a) Relative stabilities of epimers. At a given secondary carbon atom in a cyclo- hexane ring system a substituent being necessarily larger than a hydro- gen atom is more stable in an equatorial than in the corresponding axial conformation. In a rigid fused ring system in which conformational inter- conversion is impossible the axial or equatorial conformation of a substituent depends directly on its configuration. 21 The most thoroughly explored ring system which tests this generalisation is the trans-A/B steroid nucleus (66) (67).In formula (72) the experimentally determined configuration (a or p) and the conformation (e or a) 65 of the more stable epimeric secondary alcohol is indicated ab every relevant position of the nucleus ; in every case the equatorial alcohol is more stable than its axial epimer.21 66 SimiIarly 69 Dostrovsky Hughes and Ingold J. 1946,173 ; Westheimer and Mayer J. Chem. Phys. 1946 14 733 and references there cited ; Hill ibid. 1948,16 938 and references there cited. 6o Barton J. 1948 340. 6 1 Turner J . Amer. Chern. SOC. 1952 74 2118 ; Johnson ibid. 1953 75 1498. 6 2 Beckett Pitzer and Spitzer ibid. 1947 69 2488. 63 The experimental value is 2-12 kcal./mole for the liquid phase at 25' (Davies An approximate correction 6 1 64Robins and Walker J.1954 3960; 1955 1789; Chem. and Ind. 1955 772. 6 5 e* represents an equatorial conformation with respect to ring c. 66 See also Barton in LettrB Inhoffen and Tschesche " Ober Sterine Gallensauren and Gilbert J . Amer. Chem. SOC. 1941 63 1585). to convert the calculated value to the same state yields a figure of 2.07 kcal. und verwandte Naturstoffe " Enke Stuttgart 1956. BARTON AND COOKSON CONFORMATIONAL ANALYSIS 69 compounds with enolisable substituents such as alkoxycarbonyl 67 on equilibration with base yield mainly the equatorial epimers. Br I H H (73) (74) The relative stabilities of several steroidal 1 2-dihalides have been determined.G8 The diaxial cholestene 2/3 3cc-dibromide (73) for example rearranges to the diequatorial dibromide (74).This is a case where the forces of steric compression are opposed by those of dipole repulsion. No doubt form (73) is destabilised particularly by the large 2P-axial bromine ratio to axial 10-methyl interaction for cydohexene trans-dibromide is rather more stable in the diaxial than in the diequatorial conforn~ation.~~ It is interesting that the all-equatorial chloride bromide (76) rearranges presum- ably by interchange of the bromine atoms and con formational inversion to (76) the chlorine atoms taking up axial positions in preference to the larger bromine atoms.70 A (75) Cl (76) Reduction of ketones or oximes with sodium and alcohol has long been known to give mixtures of epimeric alcohols or ainines of approximately the same composition as direct equilibration.More recently it has been sug- gested that all reductions by alkali metals and proton donors proceeding through carbanions give predominantly the most stable product Included are reductions of conjugated dienes q9-unsaturated ketones aromatic 67 Klyne " Progress in Stereochemistry " Butterworths London 1954 Vol. I p. 37 et sep. 68 Alt and Barton J. 1954 4284. 69 Bastiansen and Hassel Tidsskr. Kjemi Bergveseiz Met. 1946 6 96 ; Larnaudie Compt. rend. 1953 236 909 ; Kwestroo Meijer and Havinga Rec. Trnv. ch,im. 1954 73 717 ; Kozima Sakashita and Maeda J . Arner. Chem. Xoc. 1954 76,1965 ; Bender Flowers a8nd Goering ibid. 1955 77 3463. 'O Andersen Hassel and Lunde Acta Chem. Scand. 1952 6 966. 71 Barton and Robinson J. 1954 3045 ; in view of the interesting results recently reported by Cram Allinger and Langemann (Chem.and Ind. 1955 919) the reservation must be made that the carbanions must be sufficiently long lived to exercise their stereochemical preference. See also Johnson Bannister Bloom Kemp Pappo Rogier and Szmuszkovicz J . Amer. Chem. Xoc. 1953 75 2275 ; Arth Poos Lukes Robinson Johns Fleurer and Sarett ibid. 1954 76 1715. 60 QUARTERLY REVIEWS systems and halides. For exarnple,'l 5cc-chlorocholestane (77) and 5/3 6a- dibromocholestane (78) both afford cholestane (79) with the more stable trans-a/B fusion on reduction with lithium and liquid ammonia. Ci (77) &'- H (79) Br Ejr The uniform production of products of more stable configuration is most simply explained by assigning to the carbanion a definite but easily inverted tetrahedral configuration and by assigning to the electron pair steric require- ments somewhat larger than hydrogen but smaller than other substituents.Since at a given secondary carbon atom in a cyclohexane system the steric compression of an equatorial is less than that of an axial hydroxyl group one would expect equatorial hydroxyl groups to be more easily esterified than their axial epimers and the same relation to hold for the hydrolysis of t,heir esters. For the same reason an ester of an equatorial carboxylic acid should be hydro- lysed more rapidly than its axial epimer. Using the kM'LS-A/B steroid nucleus for illustration formula (80) shows which epimeric ester is more rapidly hydrolysed.21s e6 Agreement with expectation is complete. A closer correlation of hydrolysis rates with the relative magnitudes of non-bonded interactions can be secured without ( b ) Relative rates of esteriJication and hydrolysis.66 (c) Relative rates of solvolysis. It is difficult to compare the rates of pure X,1 replacement reactions of cyclohexyl compounds owing to the ease with which competing reactions take place. It would be expected that the extra steric compression to which axial substituents are subjected would be a t BARTON AND COOKSON CONFORMATIONAL ANALYSIS 61 least one factor leading to faster solvolysis rates relative to those of the corresponding equatorial substituents. This is probably the case for the XN1 solvolysis 7 2 of neomenthyl chloride (81) relative to menthyl chloride (82). For SN2 replacement processes axially oriented substituents should be replaced more rapidly than equatorially oriented substituents.This is because the approach to the back of an equatorial substituent is hindered by the axial groups [see formula (83)]. For the epimer approach of the reagent is hindered only by 2-substituents [see (84)l. Gallagher and Long 73 H (85) described an example of the expected difference when they showed that the axial llj3-bromine atom of methyl 3a-acetoxy- 1 lj3-bromo- 12-oxocholanate (85) was replaced with inversion by OH- more rapidly than was the equatorial 1 la- bromine of its epimer. ( d ) Rates of oxidation of secondary alcohols. The relative rates of oxidation of epimeric pairs of secondary alcohols to the ketone by chromic acid or by hypobromous acid are just the reverse of the relative rates of hydrolysis of their carboxylic esters.The equatorial cholestan-3/3-01 for example is oxidised to cholestanone more slowly than is the axial cholestan- 3a-01.'~ This is understandable if the rate-controlling step is not the forma- tion of the corresponding chromate or hypobromite but attack of some nucleophilic species on the hydrogen atom 21 after the ester has been formed. Westheimer et ~ 1 . ~ ~ have shown that this is indeed the case for certain chromic acid oxidations. Chromic acid oxidation has recently been shown to be subject to steric acceleration ; 76 that is the more hindered the alcohol the greater is the 7 2 Hughes Ingold and Rose J . 1953 3839 ; see also unpublished results cited in ref. 53. 79 Gallagher and Long J . Bid. Chem. 1946 162 521. 7 4 Vavon and Jacubowicz Bull. SOC. chim. France 1933 53 581.7 6 Westheimer et al. J . Amer. Chem. SOC. 1949 71 25 ; 1951 73 65 ; 1952 '74 76 Schreiber and Eschenmoser Angew. Chem. 1955 67 278 ; Helv. Chim. Acta 4383 4387. 1955 38 1529. 62 QUARTERLY REVIEWS release of compression energy in the transition state and the faster the reaction. Steric acceleration of solvolysis is a well-known phenomenon 77 and one can envisage in similar terms the acceleration of carbonium-ion rearrangements through the release of com- pression energy accompanying the rearrangement. For example the con- version of a high-energy boat conformation into a lower-energy chair con- formation through some carbonium-ion rearrangement might be regarded (e) Conformational driving forces. HO & @ I HO (86) (87) as a reaction assisted by a conformational driving force.An interesting example 78 is possibly provided by the acid-catalysed rearrangement of euphenol(86) to isoeuphenol (87). Analogous conformational driving forces no doubt play a part in the anomalous trans-annular reactions of medium- sized rings.79 (f) Dissociation constants. fGane and Ingold 8O developed measure- ment of dissociation constants into a powerful method of investigating the conformations of symmetrical acyclic dicarboxylic acids in solution. In the cyclohexane series 81 82 the cis- and trans-1 2-dicarboxylic acids show a large difference in ApK, the difference in the pK,'s of the first and the second dissociation constants. The smaller ApK,value for the trans-1 2-acid is due to the greater separation between the integral charges in the dianion APE cis trans cycZoHexane-1 2-dicarboxylic acid .. 1.80 1.15 cycZoHexane- 1 3-dicarboxylic acid . . 0.76 0.82 whichtadopts because of electrostatic repulsion the diaxial(88) rather than the diequatorial conformation (89). In the dianion from the cis-acid for which only one chair conformation is possible (go) the charges are closer together 77 Hughes Quart. Rcw. 1951 5 245 ; F. Brown Davies .Dostrovsky Evans and Hughes Nature 1951 167 987 ; H. C. Brown and Fletcher J . Amer. Chem. SOC. 1949 71 1845 ; Bartlett et al. ibid. 1955 77 2801 2804 2806. 78 Barton McGhie Pradhan and Knight Chem. and Ind. 1954 1325 ; J . 1955 876 ; see also Arigoni Viterbo Dunnenberger Jeger and Ruzicka Helv. Chim. Acta 1954 37 2306. 79 Prelog J . 1950 420 ; Cope Fenton and Spencer J . Amer. Chem. SOC. 1952 74 5884 ; Prelog and Schenker HeZw.Chim. Acta 1952 35 2044 ; Prelog Schenker and Kiing ibid. 1953 36 471. 80 Gane and Ingold J. 1931 2153 ; Ingold ibid. p. 2179. 81 Speakman J. 1941 490 and references there cited. 82 Barton and Schmeidler J . 1948 1197. BARTON AND COOKSON CONFORMATIONAL ANALYSIS 63 in fact the same distance apart as in (89). I n the 1 3-diacids the difference co; COi H co; H (88) (89) (90) in ApK is smaller and for similar reasons the trans-acid has the larger value. This would be expected from the conformations (91) and (92) the negative charges being closer in the trans- than in the cis-acid. Analysis H (91) (92) (93) of dissociation-constant data 8 2 for the tricarboxylic acid (93) shows that here the balance of electrostatic factors makes (94) the preferred conforma- tion for the trianion rather than the alternative (95).Similarly for epimeric amino-acids )NH+ C02- ApK is greater for the epimer in which the two poles are closer together in the preferred con- format ion .s3 To the extent that ionisation of acids and bases is determined by the degree of solvation of the ions in otherwise non-polar molecules axial acids and amines would be expected to be weaker than their equatorialepimers owing to greater hindrance to solvation in the axial ions. Amongst epimeric pairs of cyclohexanecarboxylic acids the stronger are those that can take up an equatorial c~nformation.~~ Even a t position 3 of cholestane the least hindered in the entire steroid nucleus the equatorial S/I-dimethylamino- derivative is 0.20 pK unit stronger than the axial 3a-epimer in 50% aqueous tert.-butyl alcohoLS5 I- - - - - -1 83 See p. 80. 84Dippy S. R. C. Hughes and Laxton J . 1954 4102. 8 5 Bird and CookEon Chem. and Ind. 1955 1479. 64 QUARTERLY REVIEWS Physical Properties.-(a) Infrared spectra. The most far-reaching and valuable correlations between the conformations of molecules and their physical properties relate to light absorption.86 The frequency of the C-0 stretching vibration of secondary alcohols is always higher for the equatorial than for the axial epimer. show this band at about 1040 cm.-l and axial ones at about 1000 cm.-l. Page 88 has recently found that for epimeric pairs of acetoxy- and methoxy-steroids the steroid C-0 stretching frequency in the 1150-1000 cm.-l region is higher for the equatorial than for the axial epimer. The higher frequency of equatorial C-X stretching vibrations is probably quite general.These regularities are obviously of use for the assignment of configuration to alcohols derived from conformationally rigid system .89 Even the C-D stretching frequency of appropriate deutero-compounds depends on whether the deuterium is axial or e q u a t ~ r i a l . ~ ~ R. N. Jones and his collaborators 91 have shown that introduction of an equatorial a-halogen atom into a cyclohexanone increases the carbonyl stretching frequency by about 20 cm.-l whereas an axial a-halogen atom scarcely affects the frequency. Equatorial and axial a-halogeno-cyclohexanones also differ characteristically in their ultraviolet absorption spectra. The effects are just the reverse of those in the infrared region. Thus an equatorial a-bromine atom shifts the weak absorption band that occurs in all saturated ketones a t about 280 mp to slightly shorter wavelengths but an axial a-bromine atom produces a marked shift to longer wavelengths with a three- or four-fold increase in intensity.92 Rather similar behaviour is shown by a-hydroxy- and a-acetoxy-ket~nes.~~ The approximate shift (AA) in wave- length of the weak band caused by a-substitution may be summarised as in the annexed Table.Equatorial alcohols usually ( b ) Ultraviolet spectra. AR(mp) a-Substituent c1 . . - 7 + 22 Br . . - 5 3- 28 . - 1 2 + 17 + 10 OH . OAc . . - 6 It should be noted that acetylation of an equatorial a-ketolShifts the maxi- s~ For a more detailed review see Braude and Waight " Progress in Stereo- chemistry " ed. Klyne Butterworths 1954 p.126 ; for the relation between con- formation and ultraviolet absorption of cyclic conjugated dienes see Braude Chem. and Ind. 1954 1557. Jones Humphries Herling and Dobriner J . Amer. Chem. SOC. 1951 73 3215 ; Cole Jones and Dobriner ibid. 1952 74 5571 ; Fiirst Kuhn Scotoni and Giinthard Helw. Chim. Acta 1952 3!5 951. s8 Page J. 1955 2017. For example see Cole J. 1952 4969 ; Aebi Barton Burgstahler and Lindsey J. 1954 4659. For correlation of infrared spectra and conformation of carbohydrates see Barker Bourne Stacey and Whiffen ibid. p. 171 ; Barker Bourne Stephens and Whiffen ibid. pp. 3468 4211. Corey Sneen Danaher Young and Rutledge Chem. and Ind. 1954 1294. R. N. Jones Ramsay Herling and Dobriner J . Amer. Chem. SOC. 1952,74 2828. B 2 Cookson J. 1954 282.93 Cookson and Dandegaonker J . 1955 352. BARTON AND COOKSON CONFORMATIONAL ANALYSIS 65 mum to longer wavelengths whereas acetylation of the axial epimer shifts it to shorter wavelength^.^^ 94 (c) Adsorption and partition. As one would expect from simple mechani- cal considerations a compound with an equatorial polar group is in general more strongly adsorbed on a chromatographic column than its epimer with an axial substituent. 21 Examples where this order of elution from alumina is followed are the various 2 3-dihalogenocholestanes 95 and the dihydro- lysergic acids.96 However exceptions have been noted,97 perhaps because surface forces may compel adsorbed molecules to adopt conformations that are not preferred in solution. Least deviation from the rule would therefore be expected in partition chromatography and indeed Savard 98 observed that axial steroidal alcohols travel faster on paper than their equatorial epimers .The von Auwers-Skita rules,9g accord- ing to which the cis-compound (of a pair of cis-trans-isomers) has the higher refractive index and density break down when applied to 1 3- disubstituted cyclohexanes where the reverse order holds. loo The rules are also frequently ambiguous when applied to polycyclic systems such as steroids. A modified rule is valid according to which for substituted cyclo- hexanes the stereoisomer with the more axial substituents has a higher refractive index and density.lo1 Kelly lo2 has proposed a generalised version of the rule stating that for isomeric cyclohexane and tetrahydropyran derivatives which are similarly substituted on corresponding ring carbon atoms the refractive indices and densities are inversely related to con- formational stability.InJEuence of Conformation on Reactions with Particular Xtereoelectronic Requirements.-(a) Diaxial bimolecular elimination. The transition state of lowest energy for an ionic E2 reaction (elimination of X and Y) requires 21 669 lo3 the four centres concerned to be in one plane as in (96). In conformationally rigid ring systems based on cyclohexane this geometrical (d) Refractive index and density. 9 4 Baumgartner and Tamm Helv. Chim. Acta 1955 38 441. 96Alt and Barton J. 1954 4284. g6 Stoll Petrzilka Rutschmann Hofmann and Giinthard Helv. Chim. Actu 1954 97 Brooks Klyne and Miller Biochem. J. 1953 54 212. 98 Savard J. Biol. Chem. 1953 202 457.99 Von Auwers Annalen 1920 420 89 ; Skita Ber. 1920 53 1792. 100 Mousseron and Granger Bull. SOC. chim. Prance 1938 5 1618 ; 1946 218 ; 13eckett Pitzer and Xpitzer J . Amer. Chem. SOC. 1947 69 2488 ; Goering and Serres ibid. 1952 74 5908 ; Noyce and Denney ibid. p. 5912 ; Haggis and Owen J. 1953 ~$08 ; Darling Macbeth and Mills ibid. p. 1364. 37 2039. 101 See Allinger Experientia 1954 10 328. 102 Kelly personal communication. 1°3 Young Pressman and Coryell J . Amer. Chem. SOC. 1939 61 1640 ; Winstein Pressman and Young ibid. p. 1645 ; Dhar Hughes Ingold Mandour Maw and Woolf J. 1948 2117 ; Barton and Miller J . Amer. Chem. SOC. 2950 72 1066 ; Barton and Rosenfelder J. 1951 1048; for an interesting application in Wolff-Kishner reduc- tions of a-ketols see Turner Anliker Helbling Meier and Heusser Helv.Chim. Acta 1955 38 41 1. E 66 QUARTERLY REVIEWS requirement is satisfied by 1 2-trans-diaxial substituents but not by 1 2-trans-diequatorial or of course by 1 2-cis-substituents. This rule which replaces the less demanding rule of " trans-elimination " has been demonstrated especially in the debromination of 1 2-dibromides by iodide ion 1°4 as in (97). I X Br t 1- An example where the course of a reaction is controlled by preference for diaxial elimination concerns yohimbic acid (98) and its 16-epimer corynanthic acid.105 Base induces the elimination of sulphuric acid from the sulphuric acid ester of yohimbic acid [see (99)] but of carbon dioxide and sulphuric acid from the corresponding ester of corynanthic acid [see (loo)]. I n each case the reaction follows the course of diaxial elimination.aT%H H/' I6 HOZC**'* OH (98) QH k,c&o H &&\) -S%-b H-jf,) - q - 6 (99) (100) Many more examples of preferred diaxial elimination could be quoted. However Bordwell and his colleagues lo6 have recently shown that elimina- tion of a cis-hydrogen atom rather than of an alternative trans-one may 104 In simple aliphatic compounds alternative mechanisms are possible Hine and 106 Cookson Chem. and I n d . 1953 337 ; Janot Goutarel le Hir Amin and Prelog 106 Weinstock Peerson and Bordwell J . Amer. Chem. SOC. 1954 76 4748 ; Bord- Brader J . dmer. Chem. SOC. 1955 77 361. Bull. SOC. chim. France 1952 1085. well and Kern ibid. 1955 7'7 1141 ; Bordwell and Peterson ibid. p. 1145. BARTON AND COOKSON CONFORMATIONAL ANALYSIS 67 take place when the former is made sufficiently acidic by a strongly electron- attracting group.Thus trans-2-toluene-p-sulphonylcyclohexyl toluene-p- sulphonate (101) undergoes second-order base-induced elimination of toluene-p-sulphonate ion to give the cyclohexene (102) rather than (103). The cis-isomer nevertheless still reacts more rapidly (by trans-elimination and presumably through the diaxial conformation) than the trans-isomer to give the same product. The converse of the well-established rule (see above) of preferred diaxial elimination in ionic-type reactions would be preferred diaxial addition. It has recently been demonstrated 95 that electrophilic addition of halogen proceeds in this manner in systems where conformations are unambiguous. For example bromine adds to cholest- 2-ene (104) to give mainly the diaxial 2a 3a-dibromocholestane (105) pre- sumably through the 2cc 3a-bromonium ion (106).Similarly addition of hypobromous acid to cholest-2-ene affords mainly the diaxial 3a-bromo- SP-hydroxycholestane (107). (c) Diaxial ring opening and closing in cyclohexene oxides. Either electrophilic or nucleophilic opening of epoxides affords mainly the diaxial product.108 This is exemplified g5 for ring A of the steroids by the reaction (b) Diaxial electrophilic addition. H H H 1 Br (105) ' J 107 Banerji Barton and Cookson unpublished work ; preferred diaxial addition l08 Fiirst and Plattner Abs. Papers 12th Int. Congr. Pure Appl. Chem. New York of hypohalous acid is a general rule. 1951 p. 405 ; see also Barton J . 1953 1027. 68 QUARTERLY REVIEWS of cholest-2-ene a-epoxide (108) with hydrogen bromide to give the diaxial 2/3 3a-bromohydrin (109) rather than the alternative trans- but diequatorial 3p 2a-bromohydrin and of the 2p 3P-epoxide (110) to give the 3a 2p- bromohydrin (107).The expected faster ring closure of diaxial than of diequatorial halogeno- hydrins has also been observed.lo7 Thus under standard conditions the times required for 90% reaction in the base-induced epoxide formation from the diaxial bromohydrins 3a-bromocholestan-2~-ol (107) and 2P-bromo- cholestan-3a-01 (109) are respectively 1.5 and 3-5 minutes. In contrast 40 hours are required before the reaction of the diequatorial Ba-bromo- cholestan-3p-01 (11 1) has reached 90% of completion. By analogy with preferred diaxial ring closure to ethylene oxides (see above) neighbouring-group participation of the kind extensively studied in simpler systems by Winstein and his school log should proceed more readily when the trans-1 2-sub- stituents concerned are both axial than when they are both equatorial.The replacement reactions of the various 2 3-halogenohydrins of cholestane vindicate this view.95 For example treatment of the 2g 3a-hromohydrin (109) with thionyl chloride gave via the bromonium ion (112) mainly the diaxial 2/3 3a-bromochloride (113) whilst the 3a 2p-bromohydrin (107) afforded via (106) mainly the diaxial 3a 2/3-chlorobromide (1 14). Under the same conditions the diequatorial bromohydrin (1 11) was recovered (a) Diaxial neighbouring-group participation. CI (114) unchanged. The diaxial chlorohydrins in which the less nucleophilic chlorine atom is known to provide a smaller driving force,l1° showed partici- pation of chlorine in replacement of hydroxyl only in reaction with the strongest electrophilic reagent phosphorus pentachloride.The phenomena referred to in this section and also the relative ease of 109 Winstein and his collaborators J . Amer. Chem. SOC. many papers under the 110 Winstein and Grunwald J. Amer. Chem. SOC. 1948 70 828. title " The Role of Neighbouring Groups in Replacement Reactions " 1942 et seq. BARTON AND COOKSON CONFORMATIONAL ANALYSIS 69 epoxide formation (see above) demonstrate again the important geometrical preference for four coplanar reaction centres. l1 I n cyclohexane compounds rearrangements in which the ring size is unchanged take place most readily when the centres involved are coplanar [as in (1 15)] that is when the group eliminated and the migrating carbon atom are both axial.21 c6 The epimeric 17P-amin0-17a-hydroxy- 17a-methyl-~-homo-steroids provide (e) DiaxiaE rearrangements.reaction an interesting with and nitrous illustrative acid.l12 contrast The 17a-~-hydroxy-group in their modes of y/Jm '$ displaces nitrogen from the intermediate diazonium ion (1 16) to give the oxide (117) as expected from the discussion above. But the diazonium ion (118) from the 17aa-hydroxy-com- pound in which the antiparallel axial groups are N,+ and Me undergoes methyl migration and ketonisation to give the oxide (119). (115) (f) Ring contraction. Under the compulsion of the same stereo-electronic force which is the theme of much of this Review rearrangements of cyclohexane compounds involving ring contraction proceed most easily when the eliminated group is equatorial for only then are the requisite 1 2-bonds coplanar and antiparallel.is the dehydration of the epimeric triterpenoid 3a- and 3a-alcohols. The axial 3a-alcohols (120) are dehydrated simply to the A2-compounds ( l z l ) whilst the equatorial 3~-alcohols ( 122) undergo on treatment with phosphorus pentachloride dehydration with ring contraction to give rearranged olefins 1,123). Use of infrared spectroscopy to determine An instructive example 21 (9) Halogenation of ketones. ll1 This can also be discussed in terms of " orbital overlap " ; see Brutcher and Roberts Abs. Papers 126th h e r . Chem. SOC. Meeting New York 1954 p. 52-0. Change in ring size introduces a further factor ; see idem Abs.Papers 127th Amer. (:hem. SOC. Meeting Cincinnati 1955 p. 39-K. 112 Klyne and Shoppee Chem. and Ind. 1952 470 ; Cremlyn Garmaise and Hhoppee J. 1953 1847 ; see also Ramirez and Stafiej J . Amer. Chem. SOC. 1955 77 134. 70 QUARTERLY REVIEWS configuration (see above) has established 113 that in the bromination of cyclohexanones the axial cc- bromo-ketone is always formed more rapidly than the equatorial epimer. This is attributed 113 to the favourable geo- metrical arrangement of the morbitals of the enol for overlap with the enter- HO H __I___) ing-bromine vacant orbital in the transition state for axial addition compared with equatorial addition [see (124)l. An alternative explanation based on preferred diaxial electrophilic addition (see above) to the enolic ethylenic linkage followed by elimination of hydrogen bromide [see (125)] cannot yet be wholly excluded.?r+ I Br 113 Corey Experientia 1953 9 329 ; J . Amer. Chem. Xoc. 1954 76 175. BkRTON AND COOKSON CONFORMATIONAL ANALYSIS 71 (h) Reaction of amines with nitrous acid. Another useful generalisa- tion 114 115 is that equatorial amines are converted by nitrous acid into alcohols of the same configuration but axial amines yield mostly olefin with some inverted alcohol. The course of the reaction differs somewhat then from that typical of acyclic amines where the intermediate diazonium ion decomposes by the 8,l mechanism to form olefin and alcohol mostly of ~~ ~~ -~ 1 1 4 Mills J. 1953 260 ; Bose Experientia 1953 9 256 ; cf. Barton and Rosen- felder J.1951 1048. 115 W. G. Dauben Tweit and Mannerskantz J. Amer. Chem. SOC. 1954 76 4420 ; w. G. Dauben and Jiu ibid. p. 4426 ; W. G. Dauben Tweit and MacLean ibid. 1955 77 48 and references there cited. 72 QUARTERLY REVIEWS inverted configuration.116 Thus the products of deamination of the four trans-decalyl-amines are set out in the annexed chart. The two equatorial amines (126) and (127) give equatorial alcohols only (128) and (129) respectively. The two axial amines (130) and (131) afford mostly olefin accompanied by equatorial alcohol 115 as indicated. Nitrous acid converts trans-2-amino-1 -phenylcyclohexanol (132) into pro- ducts derived from 1 -phenylcycZohexene oxide (133). The cis-isomer (134) which by analogy with acyclic compounds would have been expected to undergo mainly phenyl migration to yield 2 -phenylcycZohexanone gives (134) instead 99% of the ring-contracted cyclopentyl phenyl ketone.The differ- ence in behaviour may be attributed 117 to the fact that the conformation required for migration of the phenyl group is one of high energy relative to (134). Some Illustrations from Heterocyclic Chemistry Fortunately the substitution of hetero-atoms such as nitrogen and oxygen for one or more of the carbon atoms of a cyclohexane ring causes as will be 1 Bond lengths (A) 1 Bond angles I I I- t 1 c-c 1.54 c-c-c 109" G N c-N-c I c-0 I clear from the data summarised in Table 1 only slight distortion of the ring. Consequently the generalisations that have emerged in the discussion of cyclohexane chemistry can be carried over (with slight modification where necessary) to the heterocyclic analogues.Since the p-orbitals of nitrogen 116 Ingold " Structure and Mechanism in Organic Chemistry " G . Bell London 1953 p. 397. 11' Curtin and Schumukler J. Arner. Chem. Soc. 1955 77 1105. 11* Maccoll in " Progress in Stereochemistry " ed. Klyne Butterworths 1964 p. 361. BARTON AND COOKSON CONFORMATIONAL ANALYSIS 73 and oxygen appear to maintain an approximately tetrahedral distribution,ll9 the analogy becomes even more complete. Tetrahydropyran for example is represented loy (135) and piperidine by (136) and (137) in which by analogy with the stereochemistry of carbanions (see p. 78) the former may be expected to predominate.120 The four stereoisomeric dihydrolysergic acids ( 138) probably represent the piperidine compounds that have been most thoroughly investigated from the conformational point of view.121 Reduction of lysergic acid (139) with sodium and butanol gives Alkaloid Chemistry.-(a) The lysergic acids.(138) (139) mainly dihydrolysergic acid-I. This is therefore the most stable of the four stereoisomers of (138) and may be assigned the conformation and stereo- (140) (141) chemistry depicted in partial formula (14O) with all substituents on the piperidine ring equatorial. The expression (141) must then represent the 8-epiiner dihydroiso-lysergic acid-I for vigorous alkaline hydrolysis of derivatives of the latter acid affords the equatorial epimer (140). The methyl ester of the equatorial dihydrolysergic acid-I (140) is hydrolysed more rapidly than that of the axial acid (141). Nitrous acid deamination 119 C.A. Coulson " Valence " Oxford Univ. Press 1952 p. 209. 120 For illustrations of piperidine ring conformations see Leonard Thomas and Gash J . Amer. Chern. SOC. 1955 77 1552 and references there cited ; Przybylska and Barnes Acta Cryst. 1953 6 377 ; Lindsey and Barnes ibid. 1955 8 227 ; Visser Manassen and de Vries ibid. 1954 7 288. 121 Stoll Petrzilka Rutschmann Hofmann and Giinthard Helv. Chim. Acta 1964 37 2039 and references there cited ; see also Stenlake J. 1955 1626. 74 QUARTERLY REVIEWS -lysergic acid-I (e) (140) . . . . . -isolysergic acid-I (a) (141) . . . . -lysergic acid-I1 (a) (144) . . . . -isolysergic acid-I1 (e) (142) . . . . of the primary amine produced by Curtius degradation of the equatorial acid (140) proceeds to give the corresponding alcohol with retention of configura- tion.I n contrast the epimeric axial amine obtained in the same way from the axial acid (141) suffers elimination on deamination. The two remaining isomeric dihydrolysergic acids dihydro- and dihydro- iso-lysergic acid-11 must have rings c and D cis-fused. Of the two alterna- tive conformations possible for this ring system that having the large aromatic group equatorial [as in (142)] is preferred to that where it is axial [as in (143)l. Dihydroisolysergic acid-I1 is the more stable epimer and thus 3-00 4.80 4-61 3-41 k (142) (143) has the equatorial carboxyl group (142). In confirmation the methyl ester of this acid is more rapidly hydrolysed than is that of dihydrolysergic acid-I1 ( 144). Table 2 shows that the difference between the first and the second dis- sociation constants of the two pairs of amino-acids is less for the equa- torial epimers than for the axial epimers where the dissociable groups are closer together.TABLE 2 I Dihydro-acid ( b ) The tropune alkaloids. These alkaloids comprise another thoroughly investigated group,122 though the bridged piperidine ring system (145) intro- 122 For excellent and comprehensive reviews see Stoll and Jucker Angew. Chem. 1954 66 376 and especially Fodor Acta Chirn. Acad. Sci. Hung. 1955 5 380 ; Experi- entia 1955 11 129. BARTON AND COOKSON CONFORMATIONAL ANALYSIS 75 Esters of tropine The duces some atypical features in conformational behaviour. (146) are hydrolysed more slowly 123 than those of y-tropine (147). Me (14s) (146) (147) axial hydroxyl group of tropine (146) can be isomerised to the equatorial configuration in y-tropine 124 (147) and the axial carboxyl group of ecgonine (148) to the equatorial one of y-ecgonine 125 (149).C02H Me I H n I (148) (149) ( .150) Carbohydrate Chemistry .-Much of the chemistry of the pyranose sugars has been placed on a rational basis by the admirable work of Reeves,12* who has investigated the conformations of sugars by examining their capacity for complex formation with " cuprammonium " solutions and their rates of reaction with lead tetra-acetate. I n all cases where it is geometrically possible the pyranose ring adopts the chair conformation (lSO) which is also preferred in the crystal lattice.127 The complexes formed by 1 2-glycols in cuprammonium solution contain the copper atom linked to the two oxygen atoms in a five-membered ring.The geometrical requirements are therefore similar to those of the cleavage of 1 2-glycols with lead tetra-acetate in which a cyclic complex of PbIV is an intermediate at least for glycols that react relatively rapidly.128 Although the data for lead tetra-acetate concern relative rates and those l23 Sixma Siegmann and Beyerman Proc. k. ned. Akad. Wetenschap. 1951 54 B 452; Siegmann and Wibaut Rec. Trav. chim. 1954 73 203; Hromatka Csoklich and Hofbauer Monatsh. 1952,83 1323. For a discussion see Fodor ref. 122 ; Nickon and Fieser J . Amer. Chem. Soc. 1952 '94 5566. For a discussion of the occurrence of the boat piperidine conformation in tropanes see inter aE. Zenitz Martini Priznar and Nachod ibid. p. 5564 ; Sparke Chem. and Ind. 1953 749 ; Archer and Lewis ibid.1954 853. 124 WillstBtter Ber. 1896 29 936. 125 Findlay J. Amer. Ch,ern. SOC. 1953 '75 4624 ; 126 Reeves ibid. 1949 71,215 2116 ; 1950 72 1499 ; 19574 6 4595 ; Adv. Carbo- hydrate Chem. 1951 6 107. 12' Inter aE. Cox Goodwin and Wagstaff J. 1935 1495 ; Cox and Jeffrey Nature 1939 143 894; Astbury and Dsvies ibid. 1944 154 84; Beevers and Cochran Proc. Roy. SOC. 1947 A 190 257 ; McDonald and Beevers Acta Cryst. 1950 3 394. 1954 76 2855. 12* Criegee and Buchner Ber. 1940 73 563. 76 QUARTERLY REVIEWS for cuprammonium complexing concern equilibria the close parallel between the two reactions summarised in Table 3 is thus not unexpected. For both reactions the most favourable arrangement of the glycol is when the two alcohol groups are in one plane. This is satisfied by a cis-1 2-glycol on a five-membered ring or on the " sides " of a boat six-membered ring.129 For both reactions a cis(e,a)-glycol is more favourable than a trans(e,e)- glycol because it allows a closer approach to coplanarity (see below).I n almost every case the conformation indicated by the behaviour of the pyranoses and their derivatives with cuprammonium agreed with that re- quired for a minimisation of non-bonded interactions (niaximum number of equatorial substituents). Ring form Furanose P yranose (chair) Pyranose (boat) TABLE 3 CiS trans cis (e,a) trans (e,e) trans (a,a) cis (side of boat) Rate of reaction with lead tetra-acetate Instantaneous Slow Rapid Slow Very slow Instantaneous Extent of reaction with ciiprammonium solution Large None Medium Small None Large _____ The formation of acetals by carbohydrate molecules and related poly- hydric alcohols can also be rationalised in a satisfactory manner by con- formational considerations.130 The chemistry of the inositols also lends itself to conformation& analysis.131 Other Six-membered Heterocyclic Rings.-Several other heterocyclic analogues of cyclohexane have been shown by physical methods to exist in H lz9 The point could also be illustrated by the complex-formation of glycols with boric acid so extensively studied by Boescken and his colleagues (Boeseken Adv. Carbo- hydrate Chem. 1949 4 189). 130 Barker and Bourne J. 1952 905 ; Barker Bourne and Whiffen ibid. p. 3865 ; Mills Chem. and Ind. 1954 633. 131 Chargaff and Magasanik J . Biol. Chem. 1948 175 939 and references there cited ; Magasanik Franzl and Chargaff J .Amer. Chem. Soc. 1952 '74 2618 ; Angyal and MacDonald J. 1952 686 ; Posternak and Reymond Helv. CJzim. Acta 1953 36 260. BARTON AND COOKSON CONFORMATIONAL ANALYSIS 77 the chair conformation. Thus the preferred conformations of 1 4-dichloro- piperazine 132 and of 2 4 6-trimethyl-1 3 5-triazacydohexane (“ alde- hyde ammonia ”) 133 are (151) and (152) respectively. As in the cyclohexane series cis-1 4-dioxan-2 5-dicarboxylic acid (153) is less stable than the trans-acid (154) and the cis-2 6-diacid (155) is more stable than the trans- analogue 134 (156). a-Tristhioacetaldehyde (157) is as expected less stable than the /3-isomer lZ5 (158). Conformations of Six-membered Rings containing Trigonal Atoms.-The successive substitution of trigonal for tetrahedral carbon atoms in a six- membered ring results in general in an increasing approach of the ring to planarity.All possible combinations of tetrahedral and trigonal ring atoms cannot be considered here ; each system requires separate evaluation and generalisations about the properties of equatorial and axial bonds in saturated rings can be applied only with reserve. In cycZohexanone (159) the single trigonal carbon atom introduces only slight distortion relative to cyclohexane itself though it obviously modifies some of the non-bonded interactions for example an cc-equatorial substituent is almost eclipsed by the carbonyl group. In cyczohexane-1 4-dione (160) the interactions are still further modified and tlhe dipole moment is compatible with the presence of about 10% of the boat conformation.136 The two conformations of cyczohexene not involving (a) cycloHexanones. ( b ) cycloHexenes. 132 Andersen and Hassel Acta Chem. Scand. 1949 3 1180. 133 Lund ibid. 1951 5 678. 134 Summerbell and Stephens J . Amer. Chem. Soc. 19454 76 731 6401. 135 Baumann and Fromm Ber. 1891 24 1428 ; Chattaway and Kellett J. 1930 1352 ; Hassel and Viervoll Acta Chem. Scand. 1947 1 149. 136 Le FBvre and Lo FBvre J. 1935 1696. For an investigation of the importance of boat conformations in androstane- and actiocholane-3 l7-dioneY as determined by (lipole-moment measurements see Nace and Turner J . Arner. Chem. Soc. 1953 75 4063. 78 QUARTERLY REVIEWS appreciable angle strain are shown schematically in (161) and ( 162).13' The conclusion by Beckett Freeman and Pitzer 138 from thermodynamic data that the " half-chair " conformation (161) is more stable than the " half-boat " (162) by 2.7 kcal.per mole is supported 139 (with the usual -o k7A0 e!+T+-e a' Q (159) (160) (161) reservations see above) by the X-ray analysis of several crystalline sub- stances all of which have the half-chair conformation (161). I n this con- formation the ethylenic linkage keeps the two trigonal and allylic carbon atoms approximately in one plane. The two non-allylic methylene groups are completely staggered with respect to each other so that their C-H bonds resemble normal axial and equatorial bonds in cydohexane. The bonds attached to the two allylic carbon atoms approximate only to axial and equatorial bonds and may be described as quasi-axial (a') and quasi- equatorial (e') [see (161)l.The few examples available 137 suggest that substituents are usually more stable in equatorial and quasi-equatorial than in axial and quasi-axial conformations. 140 (c) cycloHexene oxides. Electron diffraction 141 shows that the six- A i7 membered ring of cyclohexene oxide adopts a conformation (163) similar to the half-chair conformation of cyclohexene itself. (d) cycloHexenone and cyclohexenyl cations anions and radicals. In cyclohex-2-enone the introdubtion of a third trigonal carbon atom destroys See also Raphael and Stenlake ibid. 1953 1286 ; Orloff Chern. Rev. 1954 54 409 ; Corey and Sneen J . Amer. Chem. SOC. 1955 77 2505. 137 Barton Cookson Klyne and Shoppee Chern. and Ind. 1954 21. 138 Beckett Freeman and Pitzer ibid. 1948 70 4227. 139 Inter al.Carlisle and Crowfoot Proc. Roy. SOC. 1945 A 184 64 ; Pastern& Acta Cryst. 1951 4 316 ; Lasheen ibid. 1952 5 593 ; Bastiansen and Markali Acts Chem. Scand. 1952 6 442 687 ; cf. Lindsey and Barnes Acta Cryst. 1955 8 227. This conformation seems to have been first suggested for tetralin by Mills and Nixon (J. 1930 2510) and for cyclohexene by Boeseken and Stuurman (Proc. k. ned. &ad. Wetenschap. 1936 39 2). 140 But see Goering Blanchard and Silversmith J . Amer. Chem. Xoc. 1954 76 5409. 141 Ottar Acta Chem. Scand. 1947 1 283. For a discussion of the reactions of cyclohexene oxides and anhydro-sugars in terms of this conformation see inter al. Cookson Chern. and I n d . 1954 223 1512; Angyal ibid. p. 1230. BARTON AND COOKSON CONFORMATIONAL ANALYSIS 79 any resemblance to boat or chair conformations.If all the ring carbon atoms except are approximately coplanar the ring can adopt only one conformation (164) which is free from angle strain. cycZoHexeny1 cations anions and radicals may be expected to have similar conformations. Owing to the greater overlap possible between (say) the p-orbital of a nucleophilic species and the partially vacant orbital of a cyclohexenyl cation in the transition state for quasi-axial addition than for quasi-equatorial addition one might expect that addition to a cyclohexenyl cation would give usually as the kinetically controlled product the quasi-axial isomer. Thus in the geometrically equivalent tetralin series X,1 hydrolysis of the bromide from podophyllotoxin 1429 143 (165) gives the quasi-axial epi-podophyllo- toxin 144 (166) although it is the less stable epimer.OH H &LH / oco OH OMe H (165) (166) Conformational Anomalies.-Whilst the principles of conformational analysis as outlined in this Review serve to correlate a great many experi- mental facts in a satisfactory manner there are certain cases where anomalies of behaviour have been observed. In the sequel some of the more important of these are discussed. (a) Polysubstituted cyclohexanes. Attention has been directed 145 to certain 2 2 6 6-tetrasubstituted cyczohexyl compounds in which the usual stability order of epimers is reversed. A simple consideration of non-bonded interactions shows that if a 2 2 6 6-tetrasubstituted cycbhexyl compound R R X (167) (168) (169) is more stable in the axial (167) than in the equatorial form then the same reversed stability order must on the assumption of fixed conformations 142 Hartwell and Schreclier J .Amer. Chem. SOC. 1951 73 2909. 143 Schrecker and Hartwell ibid. 1952 74 5676. 144 Idem ibid. 1953 75 5916. 1 4 5 Barton Chem. and Irzd. 1953 664. 80 QUARTERLY REVIEWS hold for (168) and (169). All other R-substituted cychhexyl compounds should however follow the usual stability order. An interesting example in compounds of the class (169) has been discussed Me by Fodor and his collaborators,122 whose evidence suggests 146 that in tropane derivatives the methyl group is axial as in (170). All the discussion in this Review on the relative stabilities of epimers and of alternative conformations refers strictly only to non-polar molecules in which the forces between atoms close together but not directly linked are repulsive and inversely proportional to a high power of interatomic distance.I n addition to such repulsive forces induced by the approach of the outer electronic orbitals of different atoms account must be taken of normal (attractive or repulsive) electrostatic forces between integral charges or dipoles which vary inversely with a much lower power of distance.147 An example where conformation appears to be deter- mined more by integral charge interaction than by non-bonded repulsions has already been cited on p. 63 [see (SS)]. Physical evidence shows that the trans- 1 2-dihalogeno-cyclohexanes consist of comparable coiicentrations of the diequatorial and diaxial con- f o r m a t i o n ~ ~ ~ ~ an anomaly usually attributed to dipole repulsion in the diequatorial conformations.The Raman spectra of the trans- 1 4-dihalo- genocyclohexanes have been interpreted 149 as showing the presence in solvents such as ether of a preponderating concentration of the diaxial conformation although only the expected diequatorial conformation exists in the crystal. Conformational equilibria are of course a function of temperature and especially for polar substances of the polarity of the medium. In spite of reservations with regard to 1 2-dihalogenocycb- hexanes the preferred conformations of most polyhalogenocycZohexanes of the benzene hexachloride type appear to be those with the maximum number of equatorial substituents.l5O 151 146 This argument accepts of course that the bulk of a methyl group should be greater than that of an electron pair.One must note however that X-ray diffraction (Visser Manassen and de Vries Acta Cryst. 1954 7 288) of tropine hydrobromide reveals that the methyl group is equatorial not the hydrogen atom. As always in the crystalline state the conformation may be influenced by intermolecular forces. 147 Ingold " Structure and Mechan&n in Organic Chemistry " G. Bell London 1953 Chap. 111. 148 Inter al. Bastiansen and Hassel Tidsskr. Kjemi Bergvesen Met. 1946 6 96 ; Larnaudie Compt. rend. 1953 236 909 ; Tulinskie Di Giacomo and Smyth J . Amer. Chem. Soc. 1953 75 3552 ; Kozima Sakashita and Maeda ibid. 1954 76 1965 ; Bender Flowers and Goering ibid. 1955 77 3463 ; Kwestroo Meijer and Havinga Rec. Truv. chim. 1954 73 717. ( b ) Dipoles and integral charges. (170) +iP 149 Kozima and Yoshino J .Amer. Chem. SOC. 1953 '95 166. 150 Bastiansen Ellefsen and Hassel Acta Chem. Scund. 1949 3 918 ; Norman ibid. 1950 4 251 ; van Vloten Kruissink Strijk and Bijvoet Acta Cryst. 1950 3 139. 151 The reactions of these substances illustrate well the stereospecificity of addition and elimination. See inter al. Cristol Hause and Meek J . Amer. Chem. Xoc. 1951 73 674 ; Hughes Ingold and Pasternak J. 1953 3832 ; Kolka Orloff and Griffing, BARTON AND COOKSON CONFORMATIONAL ANALYSIS 81 The delicate balance between dipolar and steric repulsions is nicely illustrated by the following example. 152 2-Bromocyclohexanone exists in solution in the axial (171 ; R H) rather than in the equatorial conforma- tion (172 ; R = H) in which repulsion between the C=O and CeBr dipoles would be greater.The 4 4-dimethyl derivative however exists in + (171) (172) (173) the conformation with the bromine equatorial (172 ; R = Me) since the methyl group introduces a large repulsive 1 3-interaction into the axial conformation (171). It is possible for intramolecular hydrogen bonds to introduce sufficiently powerful attractive forces to modify or reverse normal conformational preferences. Thus in the 0-H stretching region of the infrared spectrum cis-cyclohexane-1 3-diol in dilute solution in carbon tetrachloride shows evidence 153 of strong internal hydrogen bonding due to the diaxial conformation (173). cis-Substituents on a cyclohexane ring one of which must be equatorial and one axial are separated by the same distance as trans-substituents when both of these are equatorial.So a t first sight cyclisation reactions involving either cis- or diequatorial trans-substituents might be expected to take place equally easily. However for all reactions requiring an approximately coplanar transition state and not involving replacement of the cyclohexane sub- stituents cis(e,a)-compounds react more rapidly than trans- (e,e)-compounds. Familiar examples are the reactions of 1 2-diols with lead tetra-acetate 154 and with periodic acid,155 and of 1 2-amino-alcohols with lead tetra- acetate,156 and acyl migration in 1 2-amin0-alcohols.~~~ Hassel and Ottar,158 in a slightly different context first drew attention to the different response of a chair ring to the two different types of distortion imposed on it by bringing cis(e,a)- or trans(e,e)-substituents into more nearly coplanar positions.The distortion induced by forcing adjacent equatorial J . Amer. Chem. Xoc. 1954 76 3940 ; Riemschneider Monatsh. 1955 86 101 ; Cornu- bert and Rio Bull. SOC. chim. France 1955 60 and previous papers by the several authors. (c) Hydrogen bonds. (d) Differential reactivity of cyclohexane- 1 2-diols. 152 Corey J . Amer. Chem. SOC. 1953 75 2301 3297 ; 1954 76 175. 153 Kuhn ibid. 1952 74 2492 ; 154 Criegee Kraft and Rank Annalen 1933 507 184 ; 155 Price and Knell J . Amer. Chem. SOC. 1942 64 552. 156 McCasland and Smith ibid. 1951 73 ,5164 ; Posternak Helv. Chim. Acta 15' Fodor and Kiss Nature 1949 164 917 ; J . Arner. Chern. SOC. 1950 72 3495. 15* Hassel and Ottar Acta Chem. Scand. 1947 1 929. 1954 76 4323. Prelog Schenker and Giinthard Helv.Chim. Acta 1952 35 1598. 1950 33 1597. F 82 QUARTERLY REVIEWS and axial substituents more nearly into the same plane (174) leads to a flattening of the ring and an increase in endocyclic valency angles the axial atoms or groups move further away from one another. The whole movement resembles an incipient conformational inversion and requires W H A (174) H A (175) little energy. On the other hand forcing two equatorial bonds more nearly into the same plane (175) entails a reduction in the separation of the axial atoms or groups and therefore much more energy is required. In agreement with this approach normally only cis(e,a)-cycbhexane- 1 2- diols condense with acetone to form isopropylidene derivatives. 159 lS0 For an interesting exception see Angyal and Macdonald J .1952 686.