STEREOCHEMISTRY OF CYCLIC COMPOUNDS Ring Formation.—An overwhelming proportion of the known cyclic compounds contain five- or six-numbered rings. The special tendency to the formation of such rings, thus indicated, is well illustrated by the hydroxy-acids of the general formula When n=3 or 4, but only then, these acids lose water spontaneously, forming cyclic internal esters (lactones) of the structure the cis-form would alone be possible.
Substitution has a remarkable effect in promoting ring closure. Thus tetramethylsuccinic acid, XIX., passes into its cyclic anhy dride, XX., with far greater readiness than succinic acid itself, Such ring formation can thus only occur when a 5- or 6-membered compound will be formed.
It is possible to form three- and four-membered rings, but they tend to spring open under the influence of reagents, the tendency being more marked in the 3- than in the 4-rings. Thus ethylene, XV., which we can regard as a 2-ring, cyclopropane, XVI., and cyclobutane, XVII., form a series of diminishing unsaturation.
and dimethylmaleic acid, XXI., shows so great a tendency to ring closure that when liberated from its salts it passes spontaneously into its anhydride, XXII. The larger the substituent the greater seems to be its effect. Thus the dimethyl-lactone, XXIII., can only be produced indirectly, and once opened it cannot be re formed, but the analogous dipropyl-lactone, XXIV., is formed spontaneously from the corresponding hydroxy-acid, XXV. :— For ethylene combines with all the halogens, even with iodine, when ethylene di-iodide XV.a is formed; cyclopropane, XVI., is stable towards iodine but reacts with bromine in the cold, form ing trimethylene dibromide; cyclobutane, XVII., is not opened by bromine, but is converted by catalytic reduction with hydrogen (under conditions under which cyclopentane and cyclohexane are perfectly stable) into n-butane, XVII.a.
These facts constitute a striking confirmation of the tetrahedral theory, for the angle at which the axes of a regular tetrahedron are inclined to one another is 109° 28', and this is very nearly the same as the angle between adjacent sides of the regular pentagon (io8°). Thus we have an immediate explanation of the ready formation of 5-membered carbon rings; in them the valencies linking each carbon atom to its neighbour on either side are inclined at an angle which hardly differs from the natural angle between carbon valencies. In cyclobutane, however, the
angle of deflection of the valencies from their natural inclination, (109° 44', is considerable; in cyclopropane it is larger, 2(1°9° 44'; whilst in ethylene it amounts to 44'.
This theory correlating the relative instability of 3- and 4-rings with the strain due to deflection of the valencies from their natural directions is known as Baeyer's Strain Theory (1885). In six- and higher-membered rings there is apparently no strain ; the strain is relieved by buckling of the ring. Thus cyclohexane may have either of the strain free arrangements.
Such non-planar configurations, if suffi ciently permanent, should cause molecular dissymmetry in appropriate derivatives. It has been impossible to demonstrate this dissymmetry; hence the molecules of these buckled rings are probably in a state of flux between their various possible configurations on account of their thermal agitation. Convincing evidence of the non-planar configuration of the cyclohexane ring is however supplied by the discovery (Hiickel, Annalen, 1925, 441, I) that decahydronaph thalene, XVIII., can exist in a trans- as well as in a cis-modifica tion, for if the two cyclohexane rings in this compound were plane It has long been held probable that the regular tetrahedral configuration obtains only in a carbon atom attached to four like radicals, as in methane or carbon tetrachloride, and that when the radicals are not all alike the angles between the valencies may be rendered unequal. Further, if the angle between one pair of valencies is increased by the space demands or the mutual repul sion of the two attached radicals, the angle between the other pair may suffer a correlated decrease. Many facts have been accumu lated by Thorpe and Ingold which can be interpreted as supporting the view that enhanced tendency to ring closure may arise from such modifications of the natural valency angle.