J. Phys. Chem. 1983, 87, 3961-3966
isomer as discussed previously, followed by the radiative decay of the energy acceptor. Fuchs et al.14 measured the vacuum-ultraviolet excited fluorescence and concluded that the upper singlet levels in liquid benzene essentially decay by autoionization when their excitation energy is larger than a characteristic critical value of 7 eV. They observed efficient recombination luminescence which indicates that charge separation is followed by geminate recombination, without spin relaxation, from the lowest charge-transfer state to the S1state. It is worthwhile noting that the broad fourth peak at -9 eV in the excitation spectrum is the mirror image of the large negative peak with the minimum at -9 eV in the transmission spectrum. Because the large negative peaks in the transmission spectra result from the ionization of molecules in the films: the fourth peak in the excitation spectrum may be attributed to the recombination fluorescence. I t is noteworthy that the profile of the luminescence excitation spectrum shown in Figure 2 is very similar to the trapped-electron spectrum for benzene measured in the gas phase.3 In other words, the peak intensities in the luminescence excitation spectrum roughly correspond to the relative cross sections for the resonant electron impact excitations for levels of lBpU,lBlu ('E2,), and lE1,. On the contrary, as shown in Figure 2, there is a marked difference in the relative peak intensities between the luminescence excitation spectrum and the electron transmission spectrum. The most characteristic difference is the large intensity of peak 5 in the electron transmission spectrum. Although peak 5 is attributed to the electronic excitation of lA1, lBzuas mentioned previously, there seem to be some other contributions to this peak. The large peak in the electron transmission spectra dIt/dVi vs. eVi corresponds to the sharp increase in the transmitted electron current through the film. In our previous papers>15 it was suggested that the profile of the conduction bands must be reflected in the transmission spectra to some extent and
-
(14) C. Fuchs, F.Heisel, and R. Voltz, J. Phys. Chem., 76,3867 (1972). (15) K. Hiraoka and M. Nara, Chem. Phys. Lett., 94, 589 (1983).
3961
the peaks might correspond to the rising parts of the high density-of-states conduction bands.15 In principle, any kind of vacant orbitals of free molecules can be candidates for the formation of the unoccupied conduction bands when molecules are condensed and form a molecular crystal. Jordan and Burrow studied the temporary anion states of unsaturated hydrocarbons in the gas phase by electron transmission spectroscopy.16 They observed the vertical temporary anion states of benzene at 1.15 and 4.85 eV. In the electron transmission spectrum of benzene (Figure 2), a small but sharp peak 1 and a large peak 5 are observed at 1.2 and 4.8 eV. Such a good agreement suggests that peaks 1 and 5 correspond to the rising part of the high density-of-states conduction bands which are formed by these vacant molecular orbitals of benzene. The large intensity of peak 5 is likely to be due to the strong overlap of vacant molecular orbitals which makes the interaction between these orbitals strong, and high density-of-states conduction bands can be formed in the solid phase. A further investigation in this respect is now in progress in our laboratory. In addition to the bands at 280,303, and 395 nm, a weak band at 233 nm appears in Figure 1. In order to characterize this band, the electron impact excitation spectrum for this emission, L vs. eVi, was measured by using a sharp cutoff band-pass filter with A,, (half-width) of 210 (30) nm. As shown in Figure IC,the onset of the excitation spectrum corresponds to that of peak 6 (So S2transition) in the electron transmission spectrum. This suggests that the 233-nm band is due to the S2 So fluorescence. Lipsky et al. observed the S2-* So fluorescence of benzene at 230 nm in the vapor and neat liquid p h a ~ e s . ' ~ JThe ~ location of the weak band observed in this experiment also argues for S2 fluorescence assignment. Registry No. Benzene, 71-43-2.
-
-
(16) K. D. Jordan and P. D. Burrow, Acc. Chem. Res., 11, 341 (1978). (17) F.Hirayama, T. Gregory, and S. Lipsky, J. Chem. Phys., 58,4696 (1973). (18) T. Gregory, F.Hirayama, and S. Lipsky, J. Chem. Phys., 58,4697 (1973).
Free-Radical Concentration in Doped Sulfur. Theory and Experiment Stephen J. Kennedy, John C. Wheeler,' Chemlstry Department, 8-0 14, University of California -Sari Dlego, La Jolla, California 92093
Christopher Osuch, and E. Wassermant Allied Corporation, Corporate Technology, Morristown, New Jersey 07960 (Received: February 3, 1983)
We present here a comparison between theory and some recent measurements of the ESR spectrum of liquid sulfur doped with various impurities. The data are taken near the polymerization transition temperature of pure sulfur and are interpreted as a measure of the concentration of the radical chain ends. By means of a simple equilibrium theory of Tobolsky-Eisenberg (TE) type we explain the observed signal intensity of the doped samples, which coincides with that of pure sulfur at high temperatures but does not exhibit the abrupt drop in intensity below the transition temperature as does pure sulfur.
Introduction Liquid sulfur undergoes a transition at 160 o c from a (low temperature) liquid consisting almost entirely of s8 'Present address: E. I. duPont de Nemours and CO.,C e n t r d Research and Development Department, Wilmington, DE 19898.
rings to a (high temperature) mixture of high molecular weight polymer in equilibrium with Sa. At this temperature dramatic variations occur in the heat capacity, the viscosity, the fraction of polymerized material, the density, and the concentration of unpaired electrons (as measured by electron spin resonance (ESR)).
0022-3654/83/2087-3961$01.50/00 1983 American Chemical Society
3962
The Journal of Physical Chemistry, Vol. 87,
Kennedy et al.
No. 20, 1983
Electron spin resonance measurements were first made on liquid sulfur by Gardner and Fraenke1.l These measurements were later repeated by Van Aken2 and Koningsberger and DeNeef.3 In all three cases the results are in reasonably good agreement with the results from calculations of free chain ends based on the Tobolsky-Eisenberg (TE) theory for liquid ~ u l f u r . A ~ scaling theory recently proposed for sulfur is also in agreemenL5 In the theory it is assumed that the ESR signal from sulfur is due to the unpaired electrons a t the ends of sulfur chains in making the comparison to experiment. All of these ESR measurements were made at temperatures above 170 "C, about 10 OC above the polymerization transition. Below this temperature the signal becomes very weak. Both of the theories mentioned above agree that there should be a sharp drop in the concentration of radicals as the temperature is lowered through the transition at 160 OC, from approximately lo4 to lo-', M. Recently we have made measurements both on pure sulfur and on sulfur doped with various dopants such as Cl,, Biz, and I,. The data for pure sulfur a t high temperatures agree with the previous experimental results. Again the signal vanishes at low temperature, presumably due to the very low radical concentrations. However, in the present measurements the signal detection has been extended down to 158 "C and shows the beginning of the expected sharp drop in signal intensity. Somewhat surprisingly, the ESR signal from the doped sulfur is essentially the same as that for pure sulfur above the transition temperature for added dopant concentrations in the range of 1-5% by weight, even though we expect the total concentration of chains to be larger by a factor of about lo4 in the doped samples. Below the transition temperature of pure sulfur, the ESR signal of the doped samples does not drop precipitously like that of pure sulfur, but simply continues to decrease steadily with decreasing temperature in a manner very similar to that observed above 160 O C . Similar behavior is observed for all Br,-doped samples in the range 1-5% by weight. In this paper we present these results and show that they are precisely what is predicted by a simple extension of the T E theory to doped sulfur together with one simplifying assumption which, as we will show, obviates the need to estimate any parameters other than those already specified in the T E theory of pure sulfur. We will also offer a heuristic explanation for the observed behavior (based on an actual and an effective dissociation of radical ends) that provides some additional physical insight. Essentially identical results are obtained from a generalization of the nonclassical scaling theory of pure sulfur to doped sulfur.6 The results of that generalization for ESR are shown here in Figure 2 for comparison with the classical TE theory; the details of that model in connection with the heat capacity of doped sulfur will be presented elsewhere.
general. We consider the following equilibria: Sa = *S8* Kl = exp[(TAS, - AH,)/RT]
*S8,*
+ s8 = *SE(,+~)* Kp = exp[(TAS, - A H p ) / R r ] Br, = 2Br*
K,
*SEn* + l/,Br2 = *S8,Br
*S8,Br
2Kc
+ ll2Br, = BrS8,Br
'/&,
(1)
where * denotes an unpaired electron spin, and *Sa,*, *SBnBr,and BrS8,Br are chains ( n 1 1)while S8stands for closed sulfur S8 rings. We will use the parameters from TE, that is, AH / R = 1595 K, ASp/R = 3.6908, AH1/R = 16507 K, and A&/R = 11.575,where R is the gas constant. In the last two equations the equilibrium constants are related by the symmetry factors required to give the equilibrium constant for the reaction
*SEn*+ BrS8,Br = 2*S8,Br K =4 (2) The assumption that K is independent of n can be justified in terms of the Fyory-Huggins theory of polymer solutions, as f i s t pointed out by Scott,' provided one takes as concentrations the number of moles per total moles of Ss. From these equilibria, it follows that the concentrations of the various species are related by the equations (*S,,*) = K1Kpn-'(S8)n (BrS8,*) = 2K,(Br2)1/2(*S8n*) (BrS8,Br) = K,2(Br2)(*S8,*) (Br*) = [K2(Brz)]1/2 (3) where we have used molecular species in parentheses to represent concentrations in moles per total moles of Sa. In these same units the total (number) concentration of sulfur poly'mers m
C [(*S8,*)+ (BrS8,*) + (BrS,,Br)]
x,
(4)
n=l
is given by (5) while the fraction of S8 units incorporated in chains is m
4
2
C n=1
n[(*S8,*) + (BrS8,*)
+ (BrS8,Br)]
The concentration of sulfur radicals is m
x, =
C [2(*S8,*) + (BrS,,*)I
n=l
Theory
Since we will be comparing our theoretical results to experiments done on sulfur doped with bromine, we will use bromine in our equations, although the theory is more (1) D. M. Gardner and G. K. Fraenkel, J . Am. Chem. Soc., 7 8 , 3279 (1956). (2) J. E. Van Aken, Physica, 39, 107 (1968). (3) D. C. Koningsberger and T. DeNeef, Chem. Phys. L e t t . , 4, 615 (1970). (4) A. V. Tobolsky and A. Eisenberg, J . Am. Chem. Soc., 81, 780 (1959). (5) J. C. Wheeler, S. J. Kennedy, and P. Pfeuty, Phys. Reo. Lett., 45, 1748 (1980). (6) S. J. Kennedy and J. C. Wheeler, J . Phys. Chem., to be published.
(7) It is this concentration that is measured by the ESR experiments. Conservation of sulfur requires that 4 + (Sa) = 1 or
while conservation of Br, requires that (7) R. L. Scott, J . Phys. Chem., 69, 261 (1965).
The Journal of Physical Chemistry, Vol. 87, No. 20, 1983 3963
Free-Radical Concentration in Doped Sulfur
f/2CK2(Br2)11/2 = (BrJO (9) where (Br,)O is the initial concentration of Br2 added to the system (assumed small). Given all of the equilibrium constants and the initial value (Br2)0,eq 8 and 9 could be solved for (Br,) and (Sa),and then x,, f#J, and x, determined from eq 5-7. However, the equilibrium constants K2 and K , are not well determined. At this point we introduce the assumption that virtually all of the bromine is reacted with sulfur. This assumption appears to be consistent with the measurements on sulfur. It has the advantage that it renders the calculation straightforward and, more importantly, relieves us of the need to estimate K2 and K,. Our assumption requires that (BrJ
T,,this is by no means the case for other properties. The total polymer concentration x, and the mean chain length or degree of polymerization, P, are drastically different above T, for doped and undoped sulfur. From eq 12, we see that x, changes from about K11/2 lo4 to about (Br2)0= As a consequence, the degree of polymerization, P 4 / x , changes from about (f#J/Kl)-/2 lo5 to about f#J/(Br2)0 10. This results in the dramatic decrease in viscosity observed by Bacon and Fanelli8 and Schenkgupon doping sulfur above
-
-
-
TP, Results and Discussion The theoretical prediction for the temperature dependence of x, is shown in Figure 1 for (Br2)0= 0% , 1% , and 5% by weight. These curves were obtained by an iterative solution of the second of eq 12, first setting the right-hand side equal to (Br2)0and then correcting by successive values (8) R. F. Bacon and R. Fanelli,
J. Am. Chem. SOC.,65, 639
(9) J. Schenk, Physica, 23, 325 (1957).
(1943).
3964
The Journal of Physical Chemistry, Vol. 87, No. 20, 1983
Kennedy et al.
1
-1 -100
F /
-120.-
14
',
_1 1
t
r
-
1 120
4
0
140
1
.
I
,
,
,
I
180
160
t
,
,
,
I
200
Afl w 200G
I
220
"c
Figure 2. Same as Figure 1 except curves are the prediction of the nonclassical theory. I n this theory AHJR = 1360 K as in ref 5 while A S J R has been shifted to 10.75 to produce somewhat better agreement with the data above T,.
of (K,(S8)[1- ( S 8 ) ] ] ' I 2 .The value of Kc(Br2)'i2can then be evaluated from the first of eq 12 and is determined to be on the order of lo4, consistent with our assumptions. As noted above, very similar results are obtained by a generalization of the nonclassical theory based on the n 0 vector model. For completeness these curves are presented in Figure 2. The details of this calculation together with applications to the heat capacity of doped sulfur, where the differences between the two theories are marked, will be presented elsewhere.6 In Figure 1 the experimentally determined values of x , for pure sulfur and sulfur doped with 1.0970, 1.4%, and 6.3% Br, by weight are shown. The experimental determinations of x , were made by using a dual-cavity spectrometer. While no radicals could be observed in pure sulfur much below the transition temperature of 160 O C , the doped samples exhibited a signal at significantly lower temperatures. Above a temperature of about 170 " C , within experimental uncertainty, all samples show the same dependence of radical concentration on temperature. Moreover, the absolute spin concentrations of the runs with 1.09% and 1.4% Br, are in good agreement with that observed for pure sulfur and predicted by theory. The higher absolute concentration a t highest dopings may be of significance, but could be an experimental artifact, and is not treated by the theory. Below this temperature the spin concentration of the undoped samples is observed to decrease more rapidly than that of doped ones. Overall the fit to the experimental data is well within 1order of magnitude, which is comparable to the uncertainty in determining absolute spin concentration via the ESR technique. In addition, there is some lack of certainty in the value of AS, employed here. A change in AS1 would cause an overall shift in the theoretical curve, either up or down. Our relative measurements on the other hand are significantly better, and indeed the shape of the theoretical curve matches the data very well. A more serious problem was revealed in measurements on sulfur doped either with I2 or with fairly high concentrations of Br, (11.4% by weight). In these cases the agreement of theory and experiment was not as good. While the assumption that all dopant molecules have reacted with sulfur to form -S-X type bonds and that the only species contributing to the ESR signal are free sulfur
-
Figure 3. ESR spectra showing the existence of a second species in the presence of either I, or high concentrations of Br,: (A) pure sulfur, 233 O C ; (B) sulfur plus 11.4% Br,, 229 OC;(C) sulfur plus 1.3% I,, 221 O C (dotted line shows base line; the sloping base line is an instrumental artifact).
chain ends seems reasonably justified for the bromine case, when I2 is the dopant (or at high Br2 concentrations) additional complicationsarise. This is apparent from spectra such as those in Figure 3 which shows the signal obtained from samples containing Iz, or a large amount of Br,. At this point we do not know the exact nature of the additional radical species which are present in these samples, although it was often possible to measure the "usual" sulfur peak in their presence. When this was done the intensity of the sulfur signal was observed to behave in a manner qualitatively similar to the case where low Br2dopant levels were used. A more fundamental complication is that the theory also assumes that the only species present in molten sulfur are S8 rings and linear polymers. There is considerable experimental evidence due to Steudel and co-workers that below T several intermediate-sized rings are also quite stable.'OP Moreover, there is evidence of a rapid interconversion between Sa, S7, and s&'' In addition, on the basis of several different theoretical models,12-15there is an indication that near Tplarge closed-looppolymers may be important. The situation regarding the importance of this effect is a t present unclear, but we do not expect the presence of ring polymers to significantly affect the results for ESR measurements predicted here. The dramatic change in x , below Tpupon doping sulfur, together with its similarity to pure sulfur and near independence of (Br2)0above Tp,can perhaps be better understood by the following heuristic argument. In pure (10)(a) R. Steudel and H.-J. Maude, Angew. Chem., Int. Ed. Engl., 18,152(1979);(b) R. Steudel, H.-J. Maude, D. Rosenbauer, H.Mockel, and T. Freyhodt, ibid.,20,394 (1981);(c) R. Steudel and H.-J. Maude, ibid.,16, 112 (1977);(d) R. Steudel, 2. Anorg. Allg. Chem., 478, 139 (1981);(e) R. Steudel and H.-J. Mausle, ibid., 478,156 (1981);(0 H.-J. Maude and R. Steudel, ibid.,478,177 (1981). (11)F.N.Tebbe, E. Wasserman, et al., J. Am. Chem. SOC.,104,4971 (1982). (12)R. E.Harris, J. Phys. Chem., 74,3102 (1971). (13)P.Pfeuty and J. C. Wheeler, Phys. Lett. A , 84,493 (1981). (14)R. Cordery, Phys. Reu. Lett., 47,457 (1981). (15)B.Duplantier and P. Pfeuty, J . Phys. A: Gen. Phys., 15,L127 (1982).
The Journal of Physical Chemistry, Vol. 87, No. 20, 1983 3965
Free-Radical Concentration in Doped Sulfur
sulfur below Tpradicals are produced (within our model) by the opening of a sulfur Sa ring and, since chain propagation is not favorable, the concentration of radicals is proportional to that of *Sa*units which is Kl(Ss), so that x,
a
Kll
(18)
Above T p ,however, chain propagation is favored and the average polymer length becomes very great. Now the radical ends produced by the opening of an Sa ring can separate and move more or less independently of one another, effectively “dissociating”. We might immediately anticipate, as is typically the case in a dissociation process, that the concentration of “dissociated” radicals varies proportionally to the square root of the effective dissociation constant Kl1I2
x,
(19)
Indeed the concentration of radicals is given (cf. eq 7) in pure sulfur by x , = 2K1(Sa)/[1- Kp(Sa)I
(20)
Below Tp,Kp < 1and (Sa)N 1so eq 18 is regained. Above T we may write the equilibrium constant for the “dkociation” of Sato form two independent radical ends as
Xr2/(SJ = 4K1’(Sa)/[l - Kp(Sa)]2= 44K1
(21)
where we have used eq 6 to obtain the second equality. This gives X,
= 2[Ki(S~)$1~’~
(22)
in agreement with eq 19, since both (Sa)and 4 are of order unity. In doped sulfur an entirely different mechanism is at work above and below Tpbut the effect is quite analogous to the situation for pure sulfur above T,,.As we have indicated previously, we expect the majority of polymers in doped sulfur, above and below the transition, to be of the “doubly capped” type BrSanBr, with a farily high concentration, about equal to (BrJO, while most of the ESR signal comes from “single capped” polymers, BrSs,*. The equilibrium between double and single capped polymers is controlled by the breaking of an S-S bond. We could expect that this process will have an equilibrium constant proportional to K1 and that it will produce two radical ends that can separate and move independently. On this basis the concentration of radicals would be given by x,
a
K1’/’
(23)
both above and below Tp. This argument can be made more quantitative by considering the following typical exchange reaction: BrSa(n+m-l)Br + *Sa* BrSa,* + BrSa,* (24) The equilibrium constant for this reaction is expected to be about unity, and will be exactly 4 if the bonding energies and entropies are independent of chain length: (BrSa,*) = 4K1(S8)(BrSa(n+m-l$r) (25) Summing this equation on both n and m and rearranging the right-hand side, one obtains (making use of the inequalities 13) the relation x? = 4Ki(Ss)4
(26)
which is identical with the exact expression in eq 12. Below Tp,most of the “polymers”are short chains and only n and m close to unity contributes to the sum. As a consequence, C$ x, (Br,)O and
- -
x,
2[K1(S8)(Br2)0]1/2
(27)
Well above Tp,however, the average chain length is greater so that many values of n and m contribute to the sums and 4 N 1- K;’, so that x , is essentially identical with its value in pure sulfur x,
N
2[K1(Sa)(1- Kp-1)]1/2
(28)
Thus, both above and below the polymerization temperature, x , is indeed proportional to Kl1I2,but the proportionality constant differes by a factor of about [ (Br,)O]‘ I z in the two cases. Since (BrJO is of order lo-, in the experiments described here, there is only about 1 order of magnitude difference between x , below and above Tp in shifting from eq 27 to eq 28. This change is comparable to that resulting from the temperature dependence of K1 over the same temperature interval so that only a steady smooth decrease is observed. If much smaller values of (Br2)0were used (and if the sensitivity of ESR measurements were sufficiently great) a more dramatic drop in intensity would be observed. Only if (Br2)Owere extremely small, on the order of K1 10-l2, would the behavior approach that of pure sulfur. Thus, in both pure sulfur and doped sulfur it is the breaking of an S-S bond that creates the radicals. The difference arises from the fact that only above Tpare these radical ends dissociated in pure sulfur, whereas in doped sulfur (with dopant concentrations much greater than Kl1i2)the radical ends can be considered independent both above and below the transitions.
-
Experimental Section Sulfur was first prepared according to the method of Bacon and Fanelli.16 Approximately 5 mL of this material was then transferred to a bulb on a grease-free vacuum line, boiled under 1 atm of argon, and cooled to room temperature. The gases were pumped away and the sulfur was carefully melted under vacuum. After the volatiles released by melting were removed, the system was repressurized with argon and the cycle was repeated again. After three of these cycles, corresponding to about 8 h of boiling time, the sulfur was vacuum distilled into an adjacent bulb, boiled once more for about 1 h, and degassed. The ESR samples were prepared from this material, which showed essentially no signal at g = 2.002. The pure sulfur sample was prepared by vacuum distilling the sulfur into a quartz tube of nominal 4-mm 0.d. and sealing the tube off. The amount of sulfur was chosen so as to give a sample length of about 4.5 cm. A sample containing a known amount of Brp was prepared in a second quartz tube which had been cut from the same piece of tubing as that used for the pure sulfur sample, thereby assuring identically sized samples. The part of the vacuum line with the sample tube was removable via an O-ring sealed joint. Into this could be distilled some sulfur, purified as above, the weight of which ( - 3 g) was determined with an analytical balance. Bromine was condensed into the apparatus which was then weighed again to determine the exact amount of Br, added. The sulfur was melted and thoroughly mixed, and 4.5 cm of the sample tube was filled with the mixture. This sample was then also sealed off with a torch. Several other samples, prepared by weighing purified sulfur and the dopant into small precleaned test tubes, mixing, and then transferring to quartz tubes, gave results similar to those of the more carefuly prepared samples; however, it was not always possible to prevent the con(16)R. F. Bacon and R. Fanelli, Ind. Eng. Chem., 34, 1043 (1942).
3966
J. Phys. Chem. 1983, 87,3966-3971
tamination of the sample with organic material, which resulted in the formation of a black substance with a g = 2.002. The I,-doped samples were also prepared in this fashion. Bromine and iodine were obtained from Baker, and both were used without further purification. The ESR spectra were obtained on a Varian E-12 spectrometer, equipped with a dual cavity. One cavity contained a reference sample while the other, fitted with a V-4510 variable-temperature accessory, contained the sulfur sample. Temperatures were measured by placing a chromel-alumel thermocouple into the stack of the Dewar just above the active portion of the cavity and are accurate to k0.5 "C. During an experiment the temperature of the sample was raised to >200 "C and then lowered in increments, allowing about 15 m for equilibration before taking data a t each new temperature. The two samples used for quantitative comparisons were prepared in such a way that they occupied close to identical volumes in the cavity. The level of sulfur in the tubes was chosen such that the samples extended throughout the entire active region of the cavity and the two samples were made from the same ground and polished tube. A sample of DPPH dissolved in benzene was used as a reference to allow for correction of any change in the spectrometer's sensitivity during, or between, runs. To obtain an estimate of the signal intensity in absolute concentration units a Varian Co. 0.00033% pitch in KC1 sample containing 1013 spins per centimeter of length was placed in the sample portion of the cavity and compared to the reference signal under conditions equivalent to those used for examination of the sulfur samples. Peak areas were obtained by numerically integrating the equation for a Lorentzian line of the measured heights and widths. Relative areas are corrected both for error in the attenuator of the spectrometer 100-kHz receiver and for Curie law variation in signal strength with temperature. Estimates of error in the relative spin concentrations vary with signal intensity from i 2 % or 3% a t temperatures
-
above 200 "C to 15-2070 a t 160 "C. Absolute spin concentrations are believed to be within a factor of -5 of the true value.
Conclusion In conclusion we have observed that the presence of halogen dopants in sulfur produces a marked change in the ESR signal intensity below the polymerization transition temperature. Our measurements indicate the presence of free radical ends in doped sulfur, above and below Tp,in concentrations comparable to those above Tp in pure sulfur. We also have obtained experimental evidence for the expected sharp decrease in the concentration of free radical ends below T in pure sulfur. The simple equilibrium theory propose3 here provides an adequate description for the observed ESR signal intensity in both pure sulfur and sulfur doped with bromine. Our original assumption that all of the bromine dissociates and bonds to sulfur appears to be justified for Brz concentrations of about by the good agreement with the qualitative features of the data. We have also offered, in the way of a heuristic explanation, the observation that the radical ends can be interpreted as dissociated in pure sulfur above Tpwhere the polymer chains are very long while the radical ends in doped sulfur are free and independent both above and below Tp. Acknowledgment. This research was supported by NSF Grants CHE 81-19247 and CHE 78-25188. Helpful conversations with R. L. Scott are gratefully acknowledged. N o t e Added in Proof. A referee has informed us that the thesis of K~ningsberger'~ discusses the ESR of halogen-doped liquid sulfur. The observations are similar to those reported here. We thank Professor R. Steudel for making a copy of the thesis available. Registry No. S, 7704-34-9; Br2, 7726-95-6; 12, 7553-56-2. (17) D. C. Koningsberger, Thesis, T. H. Eindhoven, The Netherlands, 1971.
Intermediates in the Reduction of 5-Halouracils by ea;
'
Edwln Rivera and Robert H. Schuler' Department of Chemistry, Catholic University of Puerto Rico, Ponce, Puerto Rico, and Radiation Laboratory and Department of Chemistry, University of Notre Dame. Notre Dame, Indiana (Received: February 7, 1983)
The radical anions of 5-bromo- and 5-iodouracil produced by the reaction of ea; with the halouracils have been observed in nanosecond pulse radiolysis experiments by their optical absorption at 330 nm. They decay with half-periods of 7.0 f 0.5 and 1.7 f 0.3 ns. These short lifetimes contrast with the much longer periods observed for the electron adducts to 5-chloro-and 5-fluorouracil (4.9 and >15 ps). The uracilyl radical produced by halide elimination from the radical anions has an extinction coefficient not greater than 150 M-' cm-' at 330 nm and no appreciable absorption at wavelengths longer than 310 nm but reacts with 5-bromouracil to form an addition product which has a broad absorption at 365 nm. The rate constant for this latter reaction, (2.66 f 0.09) X IOs M-' s-l , provides a reference against which to compare the rates of competing reactions. The absolute rate constant for abstraction of hydrogen from tert-butyl alcohol by uracilyl radical has been determined to be (2.3 f 0.2) X lo7 M-' s-' . It is clear that uracilyl radical reacts very rapidly by both addition and abstraction processes, reflecting the u character of this radical.
The radiation chemistry of 5-bromouracil was examined by a number of investigators during the late 1960s2 be0022-3654/83/2087-3966$0 1.50/0
cause of interest in using this moiety to localize the site of radiation damage in radiobiological studies by selective 0 1983 American Chemical Society
