COMMUNICATIONS TO THE EDITOR
1566
ence as that of Hunt and Taube, his method would give a value of 1.0425 a t 15'. A value of 1.045 f 0.003 a t 15' seems to be compatible with both experimental and theoretical data. A similar sort of argument points to a value of 1.040 f 9.003 a t 25'.
Vol. 67
Acknowledgment.-This work was supported by the National Science Foundation under Grants G-5411 and G-17422. hl. G. wishes to express his gratitude to the United States Educational Commission in the United Kingdom for a travel grant.
COMMUNICATIONS TO THE EDITOR CATIONIC ;CIBBILITIES IK FUSED CESIUM KITRATE AND THALLOUS SITRATE
Sir : In the course of tlie past few years many papers dealing with the determination of transport numbers of ions in pure ionic melts have been published. Most of the experiments were made with porous plug cells1-4 in which the plugs act as reference frames for the ionic velocities. I n our present work we used the experimental technique of electrophoresis on thin layers for the determination of single ionic mobilities in molten salts. The layer consisted of fine alumina powder sprayed on a sintered non-porous alumina support strip (30 cm. X 1.5 cm. X 2 mm. thick). The thickness of the layer was about 10 mg./cm.2. It was impregnated with the pure salt under test and doped at one end with a small amount of radioactive cations. The cell terminals consisted of two platinum electrodes immersed in the melt contained in crucibles. The electric connection between the crucibles and the strip was achieved by asbestos paper bridges. The potential gradient along the alumina strip during the experiments was measured by two auxiliary platinum wire electrodes in contact with the strip at both ends. The mobility data reported are referred to this potential gradient. Under normal experimental conditions the field strength was 3-6 v./cm. and the running time was 1-2 hr. The current was 10-45 ma., which corresponds to a maximum Joule heat of 0.2 w./cm.2. This value is small enough to avoid temperature differences along the strip. When the experiment was completed, the strip was cooled to room temperature and the activity distribution scanned by a G.M. window counter in which the aperture was 0.5 mm. I n Table I we compare a few results obtained for alkali nitrate melts with those obtained by other authors who determined the porous plug transport numbers in the same systems. We also give original results for cesium nitrate a t 450' and thallous nitrate a t 250'. The mobilities in the third column are the results of runs carried out a t different potential gradients. (1) B. B. Owens and F. R. Duke, Ames Laboratory, Iowa State College, Unclassified Report USAEC ISC-992. ( 2 ) R. W. Laity and F. R. Duke, J . Electrochem. Sac., 205, 197 (1958); "Metals Reference Book," Butterworths Scientific Publications, London,
1955, pp. 614-627. (3) A. Klemm, Discussiuns Faradag Sac., 83, 203 (1961). (4) E. D. Wolf and F. R. Duke, Ames Laboratory, Iowa State University, Unclassified Report USAEC 19-334.
TABLE I CATIONIC MOBILITIES OF PUREFCSED SALTS u x 104, T, Salt
CSNO~
Tlxoa
oc.
450
250
cm.2 v.-1 sec.-1
1.56 1.69 1.64 1.08 0.99 1,04
Av.
Previous work
1 . 6 3 zt 0.07
...
1.05iO.06
...
1.11
NaNOB
350
3.86 3.90 3.84
3.87 zt 0.07
3.86 i 0.05a
KNOa
350
2.04 2.08 2.12
2.08 f 0.06
2.21 f 0.11"
AgN03
250
Reference I .
2.57 2.68 2.57 f 0.11 2.47 Reference 2.
2.87 f 0.1gb
We feel that electrophoresis on thin layers in fused salts is an accurate and useful method for determining electrical transport properties of ionic melts. DIPARTIMENTO MATERIALI CHIMICA ALTE TEMPERATURE C. C. R. EURATOM ISPRA-VABESE, ITALY RECEIVED JANUARY 29, 1963
S. FORCHERI C. MONFRINI
PHOTOCHEXICAL REACTION OF HYDROGES BROMIDE WITH OLEFINS AT LOW TEMPERATURE
Sir: A recent electron spin resonance study of the interaction between ultraviolet irradiated hydrogen bromide and various olefins1 has been interpreted in terms of radicals formed by addition of a bromine atom to the double bond in such a manner that bonding is effectively symmetrica1.l The alternative reaction considered was hydrogen atom addition, but this was excluded on the grounds that the spectra were not those expected for alkyl radicals, and, for reaction with cyclopentene and 2-butene, spectra were identical when deuterium bromide was used instead of hydrogen bromide. (1) P. I. Abell and L. H. Piette, J . A m . Chem. Sac., 84,916 (1962).
COMMUNICATIONS TO THE EDITOR
July, 1963
These clear-cut results, which appear to constitute a n important advance in our understanding of the mechanism, have been widely accepted and quoted.2 However, examination of the spectra reported has suggested an interpretation not considered originally, which, if correct, invalidates the conclusions outlined above. Accepting that photolysis results in the formation of hydrogen and bromine atoms in close contact with olefin molecules, there seems to be a variety of possible products which might remain trqpped. Hydrogen atoms were detected in some instances, but in very low relative concentration. Addition of hydrogen has been eliminated, a t least in two instances, but attack on allylic hydrogen to give the corresponding allyl radicals is a possibility which was not considered. Bromine atoms, if trapped as such, would not contribute to the observed spectrum. Unsymmetrical “sigma” addition would give radicals whose spin resonance spectra are not expected to resemble those detected. Symmetrical addition ill be discussed below. Certainly it would be most surprising if hydrogen atoms were unable to react with the olefins, but that relatively stable bromine atoms reacted very rapidly. It should be stressed that the fate of the photolytic hydrogen atom is a trivial matter for the reaction in fluid solution, since a chain mechanism is involved13but that it is probably of equal importance to that of the bromine atom in a solid a t 77’K. It is postulated that the radicals detected by Abell and Piette’ were allylic. This would explain the results of experiments with deuterium bromide, and since these experiments were confined to 2-butene and cyclopentene, the spectra of radicals obtained from these olefins will be discussed in detail. The spectrum reported for 2-butene radicals consists of seven poorly resolved lines having roughly a binomial distribution of intensities and a separation of between 10.8 and 12.6 gauss. This spectrum is in reasonable accord with expectation for the radical CH2CH= CHCHz. Thus, to a first approximation, the unpaired electron will be distributed equally 011 C1 and Ca, so that hyperfine coupling to three or-protons and three @-protonsshould be observed. Their isotropic coupling constants will be approximately equal in magnitude and about half those normally observed-that ik, about 12 gauss. The spectrum reported for the radical derived from cyclopentene consists of eleven lines of alternating intensities, separated by about 9.4 gauss, the more prominent set of five lines having approximately a binomial distribution of intensities. This result is in good accord with expectation for the cyclopentenyl radical, I. Here there are two equivalent a-protons H
H I
1567
tion relative to the plane of the radical. Coupling to the “central” allylic proton should be small and hence simply contributes to the line widths. The coupling constant of 9.4 gauss is assigned to the a-protons, and that of 18.8 gauss to the @-protons. The latter constant should be related to the maximum value Q ~ H of about 50 gauss by the relationship apK = Q ~ Hcos2 8, 8 being the angle between the radical plane and the plane
7
through the relevant C-C unit. Hence a value of about 17 gauss is predicted, if the small negative spindensity on the “central” carbon is ignored. The apparently low value of the a-proton coupling then can be understood in terms of this large delocalization onto the @-protons. However, this theory leaves open the problem of the marked line-width alternation. A simple explanation would be that the a-proton coupling constant is not 9.4 gauss, but slightly greater, say about 11 gauss. This would have the effect of replacing the odd numbered lines by closely spaced doublets, without affecting those with even numbers. As the line widths are too great for resolution of these doublets, the effect on the derivative spectrum would be to reduce drastically the apparent intensities of the odd numbered lines, as is found experimentally. These deductions are reinforced by comparing the results with those of Fessenden4 for the cyclopentyl radical. The (aand @-protoncoupling constants of 21.6 and 35.3 gauss* are very close to twice the values now assigned to the cyclopentenyl radical. Results for some of the other compounds studied can also be understood adequately in terms of the present theory, but these spectra were generally less well resolved and since no studies were made with deuterium bromide addition of hydrogen cannot be ruled out. The major drawback of the bridging bromine atom concept is that there is no indication, in the spectra discussed above, of any anisotropy in the g-values. Appreciable anisotropy would be expected for the bromiite atom adducts, together with a shift in gav to values somewhat greater than that of the free-spin. Unfortunately, g-values are not reported, but judging from one spectrum which displays also the doublet due l o trapped hydrogen atoms, the g-values are very close to 2.0023, as expected for simple alkyl or allyl radicals. Furthermore, there is no indication of the expected characteristic hyperfine coupling to bromine. Thus, despite the fact that there is good evidenee from other sources that the chain addition of hydrogen bromide to olefins proceeds via a bridged bromineolefin intermediate, it is concluded that the electron spin resonance results for radicals trapped during the first stages of the photolysis in the solid phase cannot be taken as evidence in support of this mechanism. Indeed, if the present postulate is correct, they are the result of a minor side process.
H
and four equivalent p-protons which should couple more strongly than usual because of their favorable orienta(2) See, for example, M. Szwrarc, Chewi. Soc. (London) Spec. Puhl., 16, 103 (1962). (3) D. H. Hey and W. A. Waters, Chem. Rev., Zl, 169 (1937).
(4) R. W. Fessenden, Blellon Institute Quarterly Report, No. 6041 (1961).
&I. DEPARTMEKT OF CHEMXSTRY LEICESTER UNIVERSITY LEICESTER, ENGLAND RECEIVED FEBRUARY 20, 1963
c. R.SYMOKS
