294 REPORTS ON THE STATE OF SCIENCE, ETC. 



Absorption Spectra and Chemical Constitution of Organic 

 Compounds.— -Report of Committee (Professor I. M. Heilbron, 

 Chairman; Prcftssor E. C. C. Baly, Secretary; Professor A. W. 

 Stewart). Brawn vp hy the Secretary. 



In the report of the Committee presented in 1920 (B.A. Report, 1920, p. 222) 

 the origin of the absorption bands exhibited by compounds was discussed, and 

 a theory was put forward, based on the energy quantum theory, which gives an 

 explanation of the phenomenon. Since 1920 a considerable amount of experi- 

 mental work has been carried out which very materially supports this theory, 

 and it becomes necessary to review the whole question of absorption spectra, 

 because the situation has very materially changed during the last few years. 

 Thei'e is no doubt that the original hypothesis as at first conceived of a direct 

 correlation between the absorption bands shown by a substance in the visible 

 or ultra-violet region of the spectrum and the constitution of that compound 

 has been proved to be untenable. Since this hypothesis still seems to be accepted 

 in its original form by some experimenters, it is advisable to state the reasons 

 which militate against it. 



The hypothesis had its origin in the work of Hartley and Dobbie on a few 

 substances, sucli as isatin, carbostyril, and o-oxycarbanil, where it was found 

 possible to determine the constitution of the parent substance by comparison of 

 its absorption spectrum with those of its two methyl derivatives. Since it was 

 found that the absorptive power of the substance agreed with that of only 

 one of the two methyl derivatives, it was concluded that the constitution of 

 the parent substance was the same as the constitution of that methyl derivative. 

 From these observations there grew up a theory which postulated that tliere is 

 a direct correlation between the primary structure of a compound and the 

 absorption band it exhibits in the ultra-violet or visible region of the spectrum, 

 and indeed that the absorption band can be taken as an index of its primary 

 structure. Although there can be no doubt that there must exist some relation 

 between the composition of a molecule and the frequencies of the light which 

 it absorbs, it is certainly not a direct one in the above sense — that is to say, 

 there is certainly no law which directly connects a given primary structure with 

 a particular absorption band in that exceedingly minute section of the spectrum 

 known as the visible and ultra-violet. The whole region of the spectrum, within 

 which a substance is known to exert characteristic absorption, extends from 

 A.=0.1 /i and A = 3,000 /'., whilst the ultra-violet and visible region lies between 

 0.22/^ and 0.7fi /x, which is less than l/5000th part of the whole. Apart from 

 any other question, it seems somewhat arbitrary to base a fundamentally 

 important conclusion on the observation of one alisorption band in tliat tiny 

 region, especially when it is remembered that that absorption band is only one 

 of many exhibited by tlie substance throughout the whole spectrum. Then, 

 again, there is the fact that a very large number of compounds do not exhibit 

 any absorption band in that region, but only in the little known region lietween 

 A = .22 f). and A = .l p.. This is ignored by the upholders of the absorption- 

 constitution correlation, who only concern themselves with those particular 

 substances whicli happen to exert alisorption in the region between the wave- 

 lengths 0.76/1 and 0.22/;.. 



The arguments against this theory, which are based on actual observations, 

 are overwhelmingly strong, and the more important of these may be given in 

 detail. In the first place there are many substances which exhibit different 

 absorption bands in the ultra-violet, according to whether they exist in the 

 liquid or vapour state. Amongst the examples of this type of compound may 

 be mentioned pyridine, piperidine, and benzaldehyde. If there existed a direct 

 correlation between structure and absorption in the ultra-violet, it would be 

 necessary to attribute two different structures to each of these compounds. Some 

 recent results (Baly and Duncan, Trans. C'hem. Soc, 121, 1008 (1922)) give 

 strong evidence that the same is also true of ammonia, but the observation has 

 not as yet been directly made, since the absorption bands of the gas and liquid 

 phases of ammonia lie in the extreme ultra-violet region between A=0.2/x and 



I 



