752
COMMUNICATIONS TO THE EDITOR
substitution in either the CH:{- or the -OH groups. ‘l’lius the formation o f radiolytic hytlrogeu froill ethanol niust be primarily :L ratlical process ( 1 ) since molecular hydrogen forniation by a iiioiiomolecular process (2) would necessarily be affected by deuterium substitution in a t least two of these groups; bimolecular processes (such as ( W ) )are excluded by the observed distribution of deuterium in the radiolytic hydrogen (Table I) . 4 (a)
(Is)
+
CI-IaCH2OH ---+ CHIC13013 H Ii C ~ H B+ O 11, CZH~O
+
+
CIIaCHcOH + I-12 ( a ) CH3CHO or ( b ) CHn=CHOII or
+
which differ only slightly in bond energy, and i t 1 the liquid phase where the high collision frequency should prevent equilibration of energy among the various molecular degrees of frecdoni. I t is further apparent that the primary processes leading i o loss of hydrogen atoms are sirni1:ir i t i tlic C O I I TABLE I1 Mass number
(1)
50 49 48 47 46 45 44 43 42 41
(2)
( c ) CHg=CH? or CH,-CH-OH 0 ‘’
(d)
I
CH3-CH-OH,
etc.
I n the radical process (11, the hydrogen atoms must originate almost exclusively from -CH2-. A similar conclusion may be deduced from the mass patterns of these deuteroethanok (Table 11). In the patterns of both CH3CH20H and in CD3CHzOH the peak corresponding to the loss of one hydrogen atom is about twice the size of the parent mass peak. I n the pattern for CD3CH20H the peak corresponding to the loss of a deuterium atom is small-but in the pattern f o r CH3CDzOH the peak corresponding to the loss of one hydrogen atom is very small while the peak corresponding to the loss of a deuterium atom is as large a s the parent mass peak. Thus here also the loss of a hydrogen atom from CH3CDZOH occurs largely from the -CD2- group, despite the opposition offered to this preference by the deuterium isotope effect5 The deuterium content of the radiolytic hydrogen produced from these three alcohols considered in conjunction with the selectivity in hydrogen atom formation ( l a ) discussed above demonstrates that there must also be a pronounced selectivity in the ensuing abstraction reactions ( l b ), 4 I t is thus apparent that the production of these radiolysis products from ethanol is not a random process even under conditions where the random process is most favored, i.e., for the formation of hydrogen atoms from carbon-hydrogen bonds (4) For t h e alcohol CHaCDzOH. a pure radical process involving only D atoms from t h e -CDz- and involving only abstraction from the C H r and -OH would produce gas with a maximum H percentage of 50%. Any contribution by the bimolecular process (2d) or by abstraction from t h e - C D r would decrease the percentage of H. The observed percentage of XI was 58% (explained by assuming a small contribution of H atoms from the CHr and OH). Similar deductions may be made from t h e d a t a on CHCHnOD radiolyses. ( 5 ) Examination of the mass spectra of CHaCHDCHa (Turkevich, Friedman, Solomon and Wrightson, Tnxs JOURNAL, 70, 2638 (1948)), of CHsCDzCHa (Condon, McMurray and Thornton, J . Chcm. P h y s . , 19,1010 (1951)). and of CDICHZCDZ(Condon, THISJ O I J R N A L , ~ ~4675 , ( 1951)) shows t h a t electron impact removes a secondary hydrogen 13-17 times as readily as a primary hydrogen. Similarly electron impact removes a tertiary hydrogen 55 times as readily as a primary hydrogen from 2-methylpropane (Condon, McMurray and Thornton, loc.
