COM~IUNCATIONS TO THE EDITOR
1624 Table I : Properties of (e-) in Several Hydrocarbon Polymers Polymera
% crystallinity
Linear polyethylene (Phillips Petroleum Co., Marlex 6050) Branched polyethylene (DuPont de Nemours Co., Alathon 1414) Isotactic polypropylene (Phillips Petroleum Co.) Atactic polypropylene (Hercules, Inc.) Isotactic poly(4-methylpentene-I) (Imperial Chemical Industries, Ltd.) Polyisobutylene (Enjay Chemical Co.)
Initial u(e-),d eleotrons/100 eV
A H ~ ~Gs , ~
-82*
0.46
3.4
w45*
0.12
3.3
-ioc
0.17
2.9
0
lo00 nm), indicating that it is due to (e-). The evidently position of maximum absorption, , , ,A, lies beyond 2000 nm, but this region was inaccessible with our present equipment. Thus, , ,A is strongly red-shifted relative to the alkane glasses wherein A,, lies between 1600 and 1700 nm.4 Ekstrom and Willardg have pointed out that A,,, increases with decreasing polarity of the trapping medium, but polarity effects clearly cannot be responsible for the effect noted here since both poly (4-methylpentene- 1) and alkane glasses are nonpolar. The nature of the electron trap in hydrocarbons is not yet known, but evidence seems to be accumulating in support of a void or cavity model. I n terms of this model, our results suggest electron trapping in a void that is larger than in alkane glasses. An electron in a large void would be expected to interact less with its surroundings than one in a smaller void, thereby leading to a larger T1 as we observe. Similarly, the smaller AH,, values for polymers follow if the line The Journal of Physical Chemistry
width is due mainly to unresolved hyperfine interactions between the electron and surrounding protons at the void boundary. Finally, if a simple “particle in a box” description can be applied to an electron in a void,g the red shift of the optical absorption of (e-) in poly(4-methylpentene-1) relative to that in glasses implies a larger void since the optical transition energy decreases with increasing void radius.
Acknowledgment. We thank Dr. J. Lin for assistance with the optical measurements and Dr. W. W. Parkinson for discussions and encouragement. We are grateful to the various companies listed in Table I for their gifts of the polymer samples. (7) A. T. Bullock and L. H. Sutcliffe, Trans. Faraday Soc., 6 0 , 2112 (1964). (8) K. Tsuji and F. Williams, ibid., in press. (9) A. Ekstrom and J. E. Willard, J . Phys. Chem., 72, 4599 (1968). (10) To whom correspondence should be addressed.
REACTOR CHEMISTRY DIVISION OAKRIDGENATIONAL LABORATORY OAKRIDGE,TEXNESSEE 37830 DEPARTMENT OF CHEXISTRY OF TENNESSEE UNIVERSITY KXOXVILLE, TENSESSEE 37916
RONALD 31.K E Y S E R ~ ~
FFRAXCON WILLIAMS
RECEIVED FEBRUARY 20, 1969
Carbon-13 Magnetic Resonance.
Phenyllithium
and Phenylmagnesium Bromide Sir:
The proton magnetic resonance parameters of phenyllithium, phenylmagnesium bromide, and pyridine have been determined and exhibit closer resemblance to one another than to values for covalently substituted benzenes.l The proposal was made that these spectral similarities imply that these species have
COMXUNICATIONS TO THE EDITOR
1625
similar structures. In addition, it was concluded that the unusual features in the H-2,6 shifts arise mainly from the paramagnetic terms in the Ramsey equation2 and that their magnitudes are a measure of the ionic character of the carbon-metal bonds. Such effects should also dominate carbon-13 shifts in these species with the added advantage that other effects which contribute significantly to proton shifts, such as ring currents3 and electric fieldsj4are probably less important in the interpretation of carbon-13 shifts. We have determined the natural abundance carbon13 spectra (15.1 NHz) of phenyllithium and phenylmagnesium bromide5 using proton decoupling techniques6 in conjunction with a time averaging device for spectral accumulation. The spectra consist of two unit intensity and two doubly intense resonances, which may be attributed to C-4 and C-1 and C-2,6 and C-3,5, respectively. The ambiguity in assigning the C-2,6 and C-3,s resonances was removed by selective proton decoupling of the C-2,6 proton multiplet. The quaternary carbon atoms C-1 were readily assigned due to the absence of large proton-induced splittings in the observed resonance peak. The observed shifts are summarized in Table I along with the corresponding values for pyridine and the pyridinium cation taken from the literature.
Table I : Carbon-13" and Nitrogen-14b Chemical Shifts in Phenylmetallics, Pyridine, and Pyridinium Ion Compound
CeHjLi CsHJfgBr Pyridine' Pyridiniumc
N
c-1 -43.16 -35.84
68d 181d
C-2,6
c-3,5
c-4
-12.88 -11.37 -21.72 -13.94
2.40 2.68 4.59 -0.45
3.12 3.97 5.04 -19.84
a P p m with respect to benzene. * Ppm with respect to nitromethane. ' Reference 9, carbon-13 shifts only; see also ref 8. Reference 12.
The large downfield shifts of C-1 (-40 ppm) and C-2,6 (-11 ppm) in these compounds are immediately striking. We interpret these data as indicating that the carbon-metal bonds in these compounds are sufficiently ionic that the carbon-13 spectra reflect information about the phenyl carbanion.' It is therefore instructive to compare these observed shifts to those in the isoelectronic pyridine system.8-10 While it is appreciated that the phenyl metallics probably exist as solvated aggregates, the basis of this comparison is considered valid. Protonation of pyridine results in changes of 7.78, -5.04, and -12.42 ppm for the 2, 3, and 4 carbons.9 These changes compare favorably with those for the corresponding carbons in the phenyl carbanion (12.1, -2.5, and -3.5 ppm)" relative to benzene, while the
39.5 ppmll upfield shift at C-1 is similar in character to the 113 ppm upfield shiftI2 l 3 noted for nitrogen-14 on protonation of pyridine. Theoretical studies14 have shown that the paramagnetic screening term in the Saika and Slichter equationI5 dominates the carbon-13 chemical shift. The principle factors affecting this term are (a) charge polarization, (b) variations in the n-bond order, and (c) the average excitation energy required for the magnetic field to mix higher energy paramagnetic levels into the groundstate description of the molecule. In pyridine8-10 the charge polarization term predicts a downfield shift for the nitrogen and a-carbon atoms (C-2,6) on protonation. Both centers become more positive as a proton is added. The predicted shift is opposite in direction to that observed and thus the charge polarization term cannot be dominant. The nitrogen-a-carbon (N-C2,J n-bond order in pyridine decreases on protonation. Other factors remaining constant, this mill result in an upfield shift at C-2,6. However, the calculated magnitudes of this effect are too small (