THE FORMATION AND STABILITY OF SPIRO-COMPOUNDS. 1579 CLXXIV.-The Formation and Stability of Compounds. Part 111. spiro- Compounds c y clo Pentane. By OSCAR BECHER and JOCELYN FIELD THORPE. spiro-f+om IN Part I of this series (Beesley Ingold and Thorpe T. 1915, 107 1080) attention was directed to the fact that when the normal angle between two of the valencies of a carbon atom is changed as the result of their inclusioln in a ring groups attached by means of the other two valencies apparently take up an altered relative position. The hypothesis which was suggested in this connexion was that when two of the valenciee a and b of a carbon atom are maintained by their participation in a ring a t some inclination other than the normal the two remaining valencies, c and d will assume directions which enclose an angle differing from the normal in the opposite sense.In particular if 2p be the angle formed by the valencies a and b the angle 20 between the directions taken up by the valencies c and d will be deter-mined by the condition that these directions are equally inclined to each other and to the directions occupied by the valencies u and b . Thus the angles between the lines of action of the follow-ing pairs of valencies namely a and c b and c a and d b and d , are all equal to 28. From this condition it follows that i f 2P is greater than the normal angle (namely the angle 2tan- v S = 109O28/16// subtended a t the centre of a regular tetrahedron by one cf its sides) then 20 will be less than this amount and vice versa.A general equation from which 28 can be calculated when 2/3 is known may very readily be obtained by the following method. 3 H 1680 BECXER AND THORPE THE FORMATION AND Let a small sphere (see figure) be described around the carbon atoni as centre and let the points A R C and D in which tho lines of action of the valencies a b c and d cut the sphere be joined by great circles. Let M and N be the two points of inter-section of the great circle A B with the great circle CD M being nearest A and B and N nearest C and B. This construction involves the description of a number of right-angled spherical triangles and1 a consideration of any of them such as for example, the triangle BC’M shosws that The solution of this equation is C O ~ 28 =COS p . COS (r - e).cos e = j ( J c 0 ~ 2 p + s -COS p) fo’r all real values of the angles. The gelneral character of this solution is exhibited in the following table in which 28 is calcu-lated for a number of values of 2P which the different alicyclic structures may be‘supposed to determine : Ring. cycZoPropane ......... cy cl oBu t ane ......... cycZoPentane ......... cycloHexano ......... cycloHeptano ......... cycZoOc tam ......... No ring ............... 28. 60” 0’ 0” 90 0 0 108 0 0 120 0 0 128 34 17 135 0 0 109 28 16 2P. 116“ 55’ 66” 112 58 44 109 46 13 107 14 58 105 16 3 103 40 38 109 28 16 From this table it will be seen that in one case namely that corresponding with the cyclopentane ring the angle 28 differs from the normal angle by some minutes of arc only.This case, therefore is of particular interest since groups attached to two of the valencies of a carbon atom the remaining t’wo valencies o STABILITY OF SPIRO-COMPOUNDS. PART III. 1581 which are bound in a cyclopeiitane ring should differ but little in their reactions and particularly in their interactions from corresponding groups in a similar structure froni which the cyclopentane ring is absent. The present communication deals with this case the comparison instituted being between 68-dimethylglutaric acid (I) on the one hand and cyclopentane-1 1-diacetic acid ( I T ) on the other and it is shown that the] reactions of desrivatives of the latter acid follow very closely the analogies presented by the corresponding deriv-atives of the former.I n Part I the result of an experimental study of cgdohexaiie-1 1-diacetic acid (111) was described and it was shown that the (111.) behaviour of certain derivatives of this substance differed very notably from that of the corresponding nienibers of the dimethyl-glutaric series. The principle points of contrast presented themselves in coniiexion with : (a) The unstable character which the dibromo-ester of cycTo-hexane-l 1-diacetic acid (IV) possesses owing to the tendency i t has to eliminate ethyl bromide and to pass into the bromoclactone ester (V) whereas the corresponding dibronlo-ester of 60-dimethyl-glutaric acid (VI) is stable. C HBr*CO,Et CH(CO,Et)*? Me CH Br CO,Et C6'310<~~Br*~0,Et pFllo<CHBr--CO Me>C<CHBr*CO,Et ( b ) Ths great stability of aydohexanesyirocyclopropanedicarb-oxylic acid (VII) which for example resists the action of con-centrated hydrochloric acid a t 240° whereas the corresponding dimethylcyclopropanedicarboxylic acid (caronic acid) (VIII) is transformed into terebic acid (IX) by 5 per cent.acid a t 200O. (IV. 1 W e ) (VI.) CH*CO,H Me CH=CO,H O--CMe, C6Hlo<~Irl co,a M e>'<h *CO,R ~ o ~ c H ~ ~ H*CO,II (VII.) (VIII.) PX.) C,H,,:CH*CO,H (X.) (c) The tendency to tho formation of cyclohexylideneacetic acid (X) when the bromo-lactone ester is treated with alkalis 1582 BECKER AND THORPE THE FORMATION AND Clearly the fact that groups can be eliminated from the acetic acid residues more readily in the cyclohexanediacetic series than in the dimethylglutaric series and also the fact that the products formed by the establishment of a bond between the residues are more stable in the cyclohexane series are both in full accord with the idea that the acetic acid residues are closer together in the cyclohexanediacetic series than in the dimethylglutaric series as the calculation of 2P fo'r the cyclohexane ring indicatw.On the other hand the relation which the conversion of the bromo-lactone ester into cyclohexylideneacetic acid large quantities of which are produced under widely varying conditions,+ bears to molecular structure is not very clear although it is doubtless connected with the greater stability which according to the hypothesis here put forward the cyclohexane ring should bestow on the semi-cyclic double bond when comparison is made with the double bond in dimethylacrylic acid CMa2:C€I*C0,H.On the basis of the hypothesis outlined one would therefore expect that in regard to the elimination of the groups from the acetic acid residues and also in regard to the stability of the products in which a bond has been established between the acetic acid residues and as regards the tendency to the formation of an unsaturated monobasic acid the derivatives of cyclopentanediacetic acid would resemble those of dimethylglutaric acid rather than those of cyclohexanediacetic acid in spite of the fact that in mole-cular weight and in cyclic structure cyclopentanediacetic acid comes closer to the latter. Actually the expectation has been realised. I n particular it is shown in the present paper that: (u) As regards the stability of its dibromo-ester cyclopentane-diacetic acid resembles dimethylglutaric acid more closely than i t resembles cyclohexanediacetic acid.Thus on distilling et Ivy! aa'-dibromocyclopentccne-l 1 -diacetnte (XI) under diminished pressure the major portion passes over unchanged some 10 per cent. only being changed into the bromo-ester (XII). In similar circumstances the cyclohexane bromo-ester is completely converted into the bromo-lactone ester and ethyl bromide. Ethyl dibromo-dimethylglutarate in small quantities distils practically without * Under certain conditions however the cyclohexylideneacetic acid itself becomes unstable and passes quantitatively into the isomeride with the double bond in the ring.This is therdore the product actually isolated in the circumstances (Zoc. cit.) STABILITY OF SPIRO-COMPOUNDS. PART ITS. 1583 decomposition although the corresponding bromo-lactone ester can be obtained by repeated slow distillation. ( 6 ) cycZoPentanespirocyclopropane-1 2-dicarboxylic acid (XIII) does not possesses the same general stability as the corresponding cyclohexane derivative. Thus whilst the latter is stable to con-centrated hydrochloric acid at 240° the former is rapidly demm-posed even a t 200° by 5 per cent. acid. Under these conditions, the dimethyl analogue caronic acid is also decomposed. $?H,*CH CH CO,H CA,*CH >c<bH*CO,H (XIII. ) (XIV.) $JH,*CH CH(CO,E)*C C H,*C H >c<CH(OH)--CO (XV.) (c) cycZoPentylideneacetic acid (XIV) can be isoiated from the product obtained by hydrolysing the bromo-lactone ester (XII) with alkalis.It is however apparently formed in very much smaller amount than is the case with the correspo'nding cyclo-hexylidene derivative. I n the dimethylglutaric series this curious reaction has not yet been observed a t all and indeed it may be peculiar to cyclic compounds. I n view of these results the chemistry of cyclopentanediacetic acid and its derivatives may be said to come well into line with the general view which forms the present working hypothesis underlying the experiments recorded in this series of papers. Ethyl ad-dibromocyclopentane-1 1-diacetate (XI) is produced by the action of phosphorus pentabromide and bromine on the anhydride of cyclopentane-1 1-diacetic acid (XIX) and subse-quent treatment with alcohol.From this a considerable quantity of the bromo-lactone ester (XII) can be abtained by repeated dis-tillation. The lactone on hydrolysis with 25 per cent. aqueous potassium hydroxide yields two compounds namely cyclopentyl-ideneacetic acid (XIV) (Wallach Annalen 1902 323 159 ; 1906, 347 324; Harding and Haworth T. 1910 97 493) and the lnctonic acid of aa'-di~ydro~~cyc1a~)entane-l I -&acetic acid (XV). The similarity between the open-chain series and that contain-ing the five-membored ring is also shown in the behaviour of the diacetic acids towards mono-bromination. Thus the mono-bromination of cyclopentanediacetic acid leads to the formation of ethyl a-bromocyclopentane-1 I-diacetic acid (XVI) in the same manner as dimethylglutaric acid yields ethyl a-bromodimethyl-glutarate.It will be noted that the a-brominated ester of cyclo 1584 BECKER AND THORPE THE FORMATION AND hexane-1 1-diacetic acid could not be obtained (T. 1915 107, 1084). I n the cases of both dibromination and monobromination besides the neutral esters obtained one and the same acid (XVII) is pro-QH,-CH CHBr*CO,Et CH,. CH2>C<CH2* Copt (XVI.) (XVII. ) 7 El,* C H C H( C0,H) ? CH,*CH2>c<CH,--- CO (XVIII.) ducecl. The acid product of broniination as well as the neutral ester on treatment with highly concentrated potassium hydroxide solution yields in addition to the lactone of a-hydroxycyclo-pentanediacetic acid (XVIII) both the trans- and cis-modifications of the spiro-acid (XIII) and (XX).trans-cycloPentanespirocyclopropane-1 2-dicar boxylic acid on distillation under atmospheric pressure yields the anhydride (XIX) of its cis-isomeride from which the cia-acid (XX) can be obtained in the usual manner. H* C* CO,H H*C*CO,H I CH,*CH, I (XIII.) (XX. ) FH,*CH CH*CO (XIX.) CH,*CE€[ >C<bH.C(J>O It should be added for the benefit of those who may wish to prepare the substances described in this paper that many difficul-ties were encountered owing to the remarkable lack of tendency to crystallise which they exhibit. At one tinlei it was thought that the research would have to be abandoned for this reason but, owing to its importance in relation to the general scheme of work a t present being undertaken in these laboratories it was necessary that every effort should be made to bring it to a successful con-clusion.As will be seen from the experimental portion this was ultimately accomplished although in some cases the compounds described became crystalline only after keeping for several months. EXPERIMENTAL. Dibromination of cycloPentane-l 1-diacetic Acid. Dibromination was effected by treating 23.2 grams of the acid, prepared by Kon and Thorpe's method (T. 1919 115 700) wit STABILITY OF SPIRO-COMPOUNDS. PART 111. 1585 115 grams of phosphorus pentabromide and after a clear solution had been obtained by gentle heating on the water-bath adding gradually 43 grams of bromine. The interaction was completed by heating until the halogen had disappeared and when cold the dibromo-acid bromide was poured into' 500 C.C.of well-cooled absolute alcohol. The oil which was precipitated by water was extracted by ether and the ethereal extract thoroughly shaken with dilute aqueous sodium carbonate solution to remove acid products. The ether was then evaporated. Ethyl aaf-dibromo-cyclopentane-1 1-diacetate (XI) boils at 211-212°/30 mni. and is a colourless fairly mobile oil (Found Br = 39-39. C,3H,,0,Br, requires Br = 40.0 per cent .). aaf-l)ibronzocyclopcii t a n e - 1 I-clincetic acid, C4H, C( CHBr* CO,H),, is obtained as a white crystalline precipitate when the dibromo-acid bromide obtained in the above experiment is poured into formic acid. It separates from formic acid in small prisms which melt at 177O (Found Br=46*43.C,HI2O4Br requires Br = 46.5 per cent.). Lactone of Ethyl a-Bromo-a-hydroxycyclopentane-1 1-diacetate (XII). The mixed cis- and trans-lactones of this formula are produced by the repeated distillation of the dibromo-ester whereby ethyl bromide is also formed. The mixture boils a t 220-222O/25 mm., and is a colourless viscid oil (Found Br=27.71. Cl1H,,O4Br requires Br= 27.5 per cent.). As the lactones showed no tendency to crystallise we were unable to1 separate the cis- and trans-isomerides. IXydrolysis of the BTomo-lac ton e. (a) cycloPentylicZeneace t ic Acid (XIV).-This acid can be prepared in small yield by hydro-lysing 12 grams of the lactone by means of 60 C.C. of a 25 per cent. aqueous potassium hydroxide a t the boiling point for two holurs.The clear solution is evaporated1 to a small bulk and acidified when the unsaturated acid separates and can be purified by recrystallisation from water. It melts at 52O (Found: C = 66.53 ; B= 7.94. (b) The lactonic acid of aa-dihy~Zrox~cycloerLtane-l 1-diacetic acid (XV) is obtained from the filtrate after the separation of cyclopentylideneacetic acid by extracting it after saturation with ammonium sulphate with ether. Ths product left when the ether is evaporated is a gum which solidified only after being kept f o r some months. It crystallises from benzene in small colourless Calc. C = 66.7 ; H = 7.9 per cent.) 1586 BECKER AND THORPE THE FORMATION AND prisms which melt a t 139-140O (Found C=53.88; H=5.98. C,H,,O requirw C = 54-0 ; H = 6.0 per cent.).The melting point of this compound would suggest that it is the trans-modification corresponding with the trans-modification of the higher homologue (T. 1915 107 I l O O ) which melts at 145O. Mon>obrornination of cycloPentnne-1 1 I-diacetic A cid. Twenty-six grams of cydopentanediacetic anhydride prepared by treating the acid with acetyl chloride (T. 1919 115 700), were mixed with 80 grams of phosphorus pentabromide and 28 grams of bromine added the process of bromination being the same as that already described in the case of dibromination. Ethyl aa'-Bromo-cyclopentane-1 1-diacetate (XVI).-The ester was distilled under diminished pressure and obtained as a nearly colourless oil which, however was shown by analysis to contain a small quantity of unbrominated ethyl ester o r possibly the corresponding lactone ethyl ester.Ultimately i t was found that a comparatively pure product could be obtained by using 9 per cent. excess of bromine (that is 31 grams instead of 28). The ester produced in this way boiled a t 192O/17 mm. (Found Br=25-43. C,3H210,Br requires Br=24*9 per cent.). (2) The Acid Product of Bromination. Ethyl Hydrogen a-Bromocyclopentane-l 1-diacetic A cid (XVII) .-This acid ester is identical with the one produced by acidifying the sodium carbonate extract from the dibromination experiment. I n the present instance it is obtained in the same manner from the mono-brominated product. It cannot be distilled without decomposition, but gave a fairly good analysis in the crude state (Found: Br =26*12.C,,Hl,O,Br requires Br =27.2 per cent.). (a) The Lactone of a-Bydroxycyclo-pentane-1 I-diacetic Acid (XVIII) .-By the action of boiling aqueous sodium carbonate on the above acid product o'f bromin-ation the lactone is obtained as sole product. Eight grams of the crude acid bromo-ester were hydrolysed by boiling for six hours with a solution containing 4.5 grams of anhydrous sodium carbonate in 45 C.C. of water. The clear solution obtained in this way was extracted repeatedly with ether the ethereal solution yielding 5 grams of a viscid gum on evaporation. This material distilled a t 228-230°/15 mm. but several months elapsed before it com-menced to crystallise. It is freely soluble in all the usual organic solvents with the exception of light petroleum in which it is insoluble and can be purified by treatment with the 40-60° (1) The Neutral Product of Brominution.The Action of Alkalis STABILITY OF SFIRO-COMPOUNDS. PART IIT. 1587 fraction of this solvent when it is obtained as a white crystalline powder melting a t 69-70° (Found C = 58.59 ; H = 6.65. C,H,,O, requires C = 58.7 ; H = 6-6 per cent.). The silver salt of the corresponding dibasic acid is precipitated when silver nitrate solution is added to a solution of the ammonium salt which has been prepared by heating a solution of the lactone in excess of dilute ammonia (Found Ag = 51.90. C,R120,Ag, requires Ag = 51.9 per cent.). t rails-cyclo Pentariespirocyclopropn e-1 2 -riicar boxyl ic A cid (XIII) , This spiro-acid is best prepared by the action of very concen-trated aqueous potassium hydroxide on the acid bromo-ester but i t can also be prepared from the neutral bromo-ester if the right conditions are observed.(a) From the Acid Bromo-ester.-A solution containing GO grams of potassium hydroxide in 50 C.C. of water was heated until the temperature reached 150° when 19 grams of the acid product of bromination were cautiously added as rapidly as possible and the vigorous reaction was allowed to subside. When cold the solid mass was dissolved in water and acidified with hydrochloric acid when crystals of the spiro-acid separated on cooling. (b) From the Neutral Bromo-ester.-Fifteen grams of the neutral monobromwster were mixed with an alcoholic potassium hydroxide solution containing 15 grams of the hydroxide in 9 C.C.of water and 60 C.C. of absolute alcohol and heated for ten hours. The solution freed from alcohol yielded the spiro-acid on acidification. The acid crystallises from water in colourless plates melting a t 2 1 1 O (Found C =58-79 ; H = 6.46. C,13,204 requires C=58.7; H = 6 - 6 per cent.). The silver salt is a white. crystalline powder (Found Ag=54.19. C,H,,04Ag2 requires Ag = 54.2 per cent.). The dz’anilide C 4 H 8 C < X ~ ~ 0 0 ~ ~ ~ was prepared by heating the acid with excess of aniline a t 200° for two hours and recrystal-lising the product from dilute alcohol. It forms white silky needles melting a t 289O (Found C = 75.34 ; H = 6-71. C,,H2,O2N, requires C = 75.4 ; H = 6.6 per cent.).cis-cycloPentanespirocyclopropane-1 2-dicar boxylic Acid (XX) . The filtrate from the trans-acid was extracted repeatedly with ether the ethereal solution yielding a gum on evaporation which became part,ly solid on keeping. This was found to be a mixtur 1588 THE FORMATION AND STABILITY OF SPIRO-COMPOUNDS. of the lactone (XVIII) and the cis-acid Irom which the latter could be separated by spreading on a porous plate. The cis-acid is ho'wever best prepared by distilling the trans-acid under diminished pressure and treating the anhydride of the cis-acid which is then formed with water. The cis-syiro-acid crystallises from a small quantity of water in white flatteneld prisms which melt a t 170O. It is much more readily soluble in water than the trans-isomeride (Found C= 59.03 ; H == 6.55.C,H,,O requires C=58*7; H=6.6 per cent.). The anhydride of the cis-acid (XIX) was prepared by the action of acetyl chloride on the cis-acid and also as mentioned above, by distilling the trans-acid under diminished pressure. As it showed no tendency to crystallise it was characterised by con- by treating H*CO*NHPh version into the anilic acid C,B,:C<y CH*CO,H it in benzene solution with the calculated quantity of aniline. It crystallises from dilute alcohol in small needles which melt a t 1 8 7 O (Folund C= 69-40 ; H = 6.62. C,,H,,O,N requires C = 69.5 ; H=6*6 per cent.). The Stability of the trans-Modification.-It was soon seen that the trans-acid was very much less stable towards concentrated hydrochloric acid a t 250° than the corresponding cyclohexane derivative for under these conditions it was found to undergo complete decomposition. The action of 5 per cent. acid was there-fore tried and it was fcund that whereas at 150° the acid remained unchanged even on prolonged heating at 200° it was completely decomposed in the course of an hour. Under the conditions, therefore which transform caronic acid into terebic acid the trans-cyclopentanespiro-acid is broken down into several products, of which free carbon is the chief. The cis-acid behaved in the same way although there was evidence that it was partly trans-formed into the trans-acid prior to decomposition. Otherwise the trans-acid like similar acids of the series is remarkably stable. It can for example be boiled for a short time with acid perman-ganate without change and is stable to alkaline permanganate in the cold. I n conclusion we wish to express our thanks to Mr. C. K. Ingold and to Mr. G . A. R. Kon for much help in this very difficult piece of experimental work during the absence of one of us abroad. THE IMPERIAL COLLEUE OF SCIENCE AND TECHNOLOUY, SOUTH KENSINGTON. [Rrcriued O C ~ O ~ P Y Sfith 1920.