J . Phys. Chem. 1987,91, 3033-3036
3033
Photoreduction of Benzophenone by Acetonitrile: Correlation of Rates of Hydrogen Abstraction from RH with the Ionization Potentials of the Radicals R' Yousry M. A. Naguib, Colin Steel,* Saul G . Cohen,* Department of Chemistry, Brandeis University, Waltham, Massachusetts 02254
and Michael A. Young Research Division, Polaroid Corporation, Cambridge, Massachusetts 021 39 (Received: December 15, 1986)
Triplet benzophenone ('K) abstracts a hydrogen from acetonitrile with a rate constant of 130 f 30 M-' s-l. Despite this low rate constant acetonitrile is not really an "inert" solvent and at low light fluxes, where T-T annihilation is not a major fate of the triplet, a major decay,path for 'K is hydrogen abstraction with resulting pinacol (K2H2)formation ( c $= ~ 0.1~ ~ ~ at t = 0). Both K2H2formation and 'K lifetime rapidly decrease with irradiation due to the light absorbing transients (LAT's) which are formed along with K2H2from ketyl radicals (KH'). The rate constants per hydrogen (kH)for abstraction from R-H by the electrophilic 'K correlate well with the ionization potentials (IP) of the corresponding radicals (R').
Introduction The photoreduction of benzophenone (K) is a prototypical photochemical reaction's2 and a wide variety of reductants including hydrocarbons,614 and alcohol^^^-^^ have been employed. In the presence of moderate concentrations of good hydrogen donors like benzhydrol (KH,), hydrogen abstraction to form diphenyl ketyl radicals (KH') can be the major fate of triplet benzophenone (3K). The bulk of KH', so formed, subsequently dimerizes to form benzopinacol (K2H2)19
'K
-
+ KH2 KH' + K H ' KH' + KH' K2H2 -+
(1)
(24
Acetonitrile is often used as a solvent for carrying out photochemical reactions in part because of its good spectral and solubility properties. It is also possible to obtain long n-r triplet lifetimes at room temperature using this solvent, provided low excitation intensities are employed so that triplet-triplet annihilation is negligible; in other words, it is a relatively "inert" solvent. Thus, for benzophenone, row of 50-100 ps can easily be obtained and by taking great care with sample purity and preparation we have been able to obtain lifetimes as long as 180 p ~ . ~But, ' this (1) Cohen, S. G.; Parola, A.; Parsons, G. H. Chem. Rev. 1973, 73, 141-161. (2) Scaiano, J. C. J. Photochem. 1973/74,2,81-118. Wagner, P. J. Top. Curr. Chem. 1976. 66. 1-52. (3) Inbar, S;Gnschitz, H.; Cohen, S. G. J . Am. Chem. SOC.1981, 103, 1048-1054; 1980, 102, 1419-1421. (4) Shaefer. C. G.; Peters K. S. J . Am. Chem. SOC.1980,102,7567-7568. (5) Simon, J. D.; Peters, K. S. J. Am. Chem. SOC.1981,103,6403-6406; 1982, 104, 6542-6547; J. Phys. Chem. 1983, 87, 4855-4857. (6) Abbott, G. D.; Phillips, D. Mol. Photochem. 1977,8, 289-309. (7) Berger, M.; Camp, R. N.; Demetrescu, I.; Giering, L.; Steel, C. Isr. J. Chem. 1977, 16, 311-317. (8) Giering, L.; Berger, M.; Steel, C. J . Am. Chem. SOC.1974,,96, 953-958. (9) Wagner, P. J.; Truman, R. J.; Scaiano, J. C. J . Am. Chem. SOC.1985, 107, 7093-7097. (10) Encinas, M. V.; Scaiano, J. C. J . Am. Chem. SOC.1981, 103, 6393-6397. (1 1) Hammond, G. S.; Baker, W. P.; Moore, W. M. J . Am. Chem. SOC. 1961, 83, 2195-2199. (12) Walling, C.; Gibian, M. J. J . Am. Chem. Soc. 1%5,87, 3361-3364. (13) Bell, J. A.; Lnschitz, H. J . Am. Chem. SOC.1963, 85, 528-532. (14) Beckett, A.; Porter, G. Trans. Faraday SOC.1963, 59, 2038-2050. (15) Porter, G.; Wilkinson, F. Trans. Faraday Soc. 1961,57, 1686-1691. (16) Moore, W. M.; Hammond, G. S.; Foss, R. P. J. Am. Chem. SOC. 1961, 83, 2789-2794. (17) Moore, W. M.; Ketchum, M. D. J . Phys. Chem. 1964,68, 214-217. (18) Topp, M. R. Chem. Phys. Lett. 1975, 32, 144-149. (19) Schuster, D. I.; Karp, P. B. J . Photochem. 1980, 12, 333-344. (20) Gramain, J-C.; Remuson, R. J. Org. Chem. 1985, 50, 1120-1 122. (21) Chilton, J.; Giering, L.; Steel, C. J. Am. Chem. SOC.1976, 98, 1865-1870.
0022-3654/87/2091-3033$01.50/0
is hard to reproduce on a routine basis. Such lifetimes may be compared with a 2.7-ps lifetime in a hydrocarbon such as isooctane22and a 700-ps lifetime in an inert fluorocarbon solvent.23 However, when benzophenone is irradiated in acetonitrile, as in any other hydrogen-containing solvent, the observed lifetime rapidly decreases with the concomitant buildup of products which absorb in the 350-nm region.2' These products have sometimes been called light-absorbing transients ( L A T ' S ) ~ because ~ ~ * ~ they ~~ are not stable and break down even in the dark and especially in the presence of oxygen. Because they are formed only when a hydrogen donor is present, it is reasonable to assume that they are formed by ring coupling of the ketyl radicals either with themselves, reaction 2b, or with radicals (S') derived from the solvent (SH), reaction 4. 'K SH KH' S' (3)
+
- -+ -
+ KH' KH' + S'
KH'
Q
Q
(2b)
(4)
Photoproducts of the type Q (and/or Q ) are expected to have lower triplet energies than benzophenone, and so, to be diffusion-controlled quenchers for the latter. Acetonitrile may not therefore be such an "inert" solvent as is commonly thought. Because of the importance of benzophenone and acetonitrile in organic photochemistry and because we were interested in the reduction of dyes by KH' in a~etonitrile,'~ we thought it of some interest to study reaction 3 and its consequences in greater detail.
Possible forms of Q and Q' w i t h X = KH or S
Results and Discussion Abstraction of Hydrogen from Acetonitrile. In the absence of added reductant (eq 1) and at low light flux, where T-T an(22) Clark, W. D. K.; Litt, A. D.; Steel, C. J. Am. Chem. SOC.1969,91, 541 3-5415. (23) Parker, C. A.; Joyce, T. A. J. Chem. SOC.,Chem. Commun. 1968, 749-751. (24) BickstrBm, L. J.; Appelgren, K. L.; Niklasson, R. J. V. Acta Chem. Scad. 1965, 19, 1555-1565; 1966, 20, 2617-2620. (25) Schenk, G. 0.;Cziesla, M.; Eppinger, K.; Matthias, G.; Pape, M. Tetrahedron Lett. 1967, 193-198. (26) a) Cohen, S. G.; Cohen, J. I. Tetrahedron Lett. 1968,4823-4826. b) Cohen, S. G.; Cohen, J. I. Isr. J. Chem. 1968,6, 757-767. (27) Filipescu, N.; Minn, F.L. J . Am. Chem. SOC.1968,90, 1544-1547. (28) Wagner, P. J. Mol. Photochem. 1969, 1 , 11-87. (29) Scaiano, J. C.; Abuin, E. B.; Stewart, L. C. J . Am. Chem. SOC.1982, 104, 5613-5679. (30) Naguib, Y. M. A.; Cohen, S. G.; Steel, C. J . Am. Chem. Soc. 1986, 108, 128-133.
0 1987 American Chemical Society
Naguib et al.
3034 The Journal of Physical Chemistry, Vol. 91, No. I I , 1987
0.10
c
R
0.08
v)
N
0.06 c
E
!n c
model
! t I 9
0.04
0010-
1
-
f 0006
k, [SH] (s") 3000 2500
d 0.02
0 002 0.00
20
0
60
40
80
I
1 0
I
100
02
Figure 1. Absorbance as a function of time following laser flash: (a) experimental trace (-); (b) model values (A). [K] = 1.1 X 1V3M, laser energy = 5.7 mJ at 355 nm, and cell path length = 1.0 cm. k3[SH] = 2500 s-l.
nihilation is unimportant, )K decays by a variety of (pseudo) first-order processes, reactions 6 and 7 in Scheme I. Reaction
SCHEME I
Figure 2. Absorbance as a function of time following laser flash: (a) experimental trace (-); (b) model curves for different values of k,[SH]. 0.016
KI.b_3K K
+h
--
3K 3K KH.
+ SH
+
KH.
(5) ~ p
K
+ S'
-
+
K
+Q
(8)
+ k7 + k3[SH] = k")
The longest experimentally observed lifetime in acetonitrile is 180 so k3[SH] 5 5.6 X lo3 s-'. Since [SH] = 19 M, this sets an upper limit for k3 of 290 M-' s-l. This value is remarkably low in comparison with other compounds having comparable C-H bond strengths (vide infra)., At laser intensities, we must also include T-T annihilation 3K 3K 4 K K (9 ) gszl and
+
!
-
-
I
-
-
O'Oo4
'IA 1
\
-
-
{A
(3)
6 is radiative emission (k6 = 140 s-1)23331 and reaction 7 comprises all those pseudo-first-order decay processes due to impurity quenching, etc. Reaction 3 is hydrogen abstraction from the solvent, in this case, acetonitrile. The observed lifetime in the absence of quencher Q is then k6
I
K2H2
3K + Q
1/7,,bsd =
0.008
(7)
(2b)
I
-
-
Q
I
- I 0.012
(6)
(2e)
I
I
-
0.3 ms, [3K] has decayed almost to zero and that essentially all the signal is due to KH'. The level of absorbance in this region is determined by k3[SH] and so can be used to determine the latter. Again setting [SH] = 19 M, we get k3 = 130 f 30 M-I s-], which is consistent with the low-intensity phosphorescence emission measurements mentioned above. We shall discuss the possible reasons for this very low rate constant for hydrogen abstraction below, but we shall first display some of the consequences of this reaction. Variation in Triplet Lifetime with Irradiation. As mentioned in the Introduction, even when benzophenone is irradiated at quite modest intensities (Isbs 2 lo-' einstein/(L.s) at 365 nm in our experiments) the observed lifetime drops rapidly. To first approximation, this can be modelled by considering the sequences of reactions shown in Scheme I. The buildup of Q with time in reaction 2b results in a decrease in the observed lifetime
+
+
1/70bsd(t) = k") ik,[Q(t)I
(11)
We used the above values for (ks+ k7) and k3. Rate constant k, was set at the diffusion-controlled value, Le., 1.5 X 1Olo M-I
1967: .. . , n r -297. - -
(32) Linschitz, H.; Steel, C.; Bell, J. A. J . Phys. Chem. 1962, 66, 2574-2577, and ref. 34. (33) Yekta, A.; Turro, N. J. Mol. Photochem. 1972, 3, 307-322. (34) Bensasson, R.V.; Gramain, J-C. J. Chem. SOC.,Faraday Trans. I 1980, 76, 1801-1810.
(35) Hurley, J. K.; Linschitz, H., private communication. (36) Hurley, J. K.;Sinai, N.; Linschitz, H. Phorochem. Phofobiol. 1983, 38, 9-14. (37) Obtained from spectrum in ref 3 and extinction coefficients in ref 35.
The Journal of Physical Chemistry, Vol. 91, No. 11, 1987 3035
Photoreduction of Benzophenone by Acetonitrile
I
0
I
20
I
I
I
40
I
60
I
1 80
lrradlatlon T h e (mln)
Figure 4. Variation in triplet benzophenone lifetime with irradiation time. [K]= 1.15 X lo-' M in acetonitrile, Iabs= 2.2 X lo-' einstein/L.s); model curve (-); experimental points (A).
somewhat more reactive than ethane. In a similar vein, 2-propanol and 2-methylpropane have comparable tertiary C-H BDEs and comparable reactivities. Thus for methyl radicals the general trend in rate contsants is in line with BDEs and there is little evidence for strong polar effects on the reactivities. In the case of K, the ratios kH(methanol)/kH(ethane) and k ~ ( 2 - p r o p a n o l ) / k ~ ( 2 methylpropane) are enhanced by about 4.6 compared to the corresponding ratios for CH3'. However kH(acetonitrile)/kH(ethane) decreases by a factor of almost 900 in going from CH3' to 3K. This low intermolecular reactivity of 3K toward acetonitrile agrees in a pleasing way with the work of Wagner et aL41 From studies on intramolecular hydrogen abstraction by aromatic carbonyls they found that replacement of methyl by cyano at the reactive site lowered the reactivity by a factor of 30. It was therefore suggested that the inductive electron-withdrawing effect of the cyano group more than offsets the resonance stabilization. This resonance is the same effect which stabilizes the 'CH2CN radical and so lowers the BDE of H-CH2CN compared to HCHzCH3.39 In rationalizing trends in reaction if the data do not conform to a simple BDE pattern, additional effects such as "polar contributions" (11) to the transition state are often invoked.
-
S-I, and for k2 we used the literature value of 1 X lo8 M-' s-' 35 {-d 6 -6 H. +R ) with a = 0.4; see eq 2. For simplicity we did not consider any cross coupling reactions like (4). The model and experimental I I1 data are shown in Figure 4. The initial rapid drop in experimental In this respect Hammett up correlations* have been extensively lifetime is fitted quite well, but at longer times the model lifetimes used to establish the importance of polarity in the transition state are shorter than the experimental values. This could easily come of radical reactions. Here we consider a related approach which about if there is some thermal2I and/or p h o t o c h e m i ~ a inl ~ ~ ~ ~ also ~ ~ ~adds ~ insight to the problem, viz., showing that there is a stability of the quencher so that it does not continue to build up correlation between kH and the IP of R'. The latter is a measure as indicated by the simple model. of the tendency of R' to donate an electron to the attacking Benzopinacol Yield. The same model used to predict the electrophilic "radical" -0' ('K) and so is a measure of the convariation in triplet lifetime can also be used to model the pinacol tribution of I1 to the transition state (TS). The energy difference yield as a function of time. This is shown in Figure 5 as the dashed between I1 and the products (-OH + R') is approximately equal curve along with the experimental points. Notice that the model to (IP(R') + D(0-H) - EA(0') - C) where IP(R') is the ionestimates the observed yields well in the early stages, but that the ization potential of the radical R', D(O-H) is the strength of the calculated values are smaller than the observed yields later in the 0-H bond in ketyl, EA(0') is the electron affinity of the oxy reaction. This comes about because the model does not predict radical formed on removing H from ketyl, and C is the Coulombic Tobsd(t)very well at these later stages (see Figure 4). However, energy released on bringing -0-and R+ from infinity to the if experimental values of T,,bsd(t) are used to predict the pinacol distance apart in the complex. For a series of substrates R-H, yields, the situation is much better. It is easy to show that, even only IP(R') will differ. In Figure 6 we have plotted the IP's of if the quencher undergoes some complex decay, the yield of pinacol radicals R'47-49 against kH.8'14'42'43These data for 3K can be as a function of time is given by
In Figure 5 we show, as the full line, the curve generated by the function [K2H2(t)] = 2 X lo4 &Tobsd(t) dt. The curve now fits the data well over the whole range. In these experiments [SH] = 19 M and Iabs = 2.1 X einstein/(L.s) so that (1 - a ) k 3 = 100 M-I s-l. For a = 0.4, this makes k3 = 165 M-' s-l, in agreement with the kinetic measurements above. Reorganizing eq 12, we see that the quantum yield of pinacol formation over an interval [O,t] is
so that initially + K I H I ( 0 ) i= 0.1 for robsd(0) = l/k(') = 100 ks. As the reaction progresses, + K 2 H 2 ( t ) decreases because of the build up in Q. Thus, after 30 min of irradiation the experimental quantum yield was 0.03 (see Figure 5). At this point, the experimental lifetime had dropped from its initial value of 100 to 24 ks; the measured concentration of benzopinacol was [K2H2] = 9.8 p M and eq 11 yielded [Q] = 2.1 WM(2% conversion). Rate Constants f o r Hydrogen Abstraction by Triplet Benzophenone. A variety of compounds have been used in studying hydrogen abstraction by 'K and it is instructive to compare some of these data with comparable results for methyl radicals (Table I). Here kH refers to the rate of abstraction per active H. Compared to ethane, both methanol and acetonitrile have somewhate lower primary C-H bond dissociation energies (BDE)38,39 and, in the case of.methy1 radical reactions,m both are
(38) King, K. D.; Gcddard, R. D. Int. J . Chem. Kinet. 1975,7, 837-855. (39) Data taken from: CRC Handbook of Chemistry and Physic, 64th ed.; CRC Press: Boca Raton, FL, 1983-1984;pp F187-F189. (40) (a) Trotman-Dickenson, A. F.;Milne, G. S. Tables of Bimolecular Gas Reactions; National Bureau of Standards, U S . Government Printing Office: Washington, DC, 1967. (b) Gray, P.; H e r d , A. A.; Jones, A. Chem. Rev. 1971, 71, 247-294. (41)Wagner, P. J.; Kemppainen, A. E. J . Am. Chem. SOC.1972, 94, 7495-7499. (42) Cohen, S. G.; Litt, A. D. Tetrahedron Lett. 1970,837-840. (43)Turro, N. J. Modern Molecular Photochemistry; Benjamin-Cummings: Menlo Park, CA, 1978;p 374. (44)Tedder, J. M. Tetrahedron 1982,38, 313-329. (45) For substituent effects on BDE, see: (a) Zavitsas, A. A,; Pinto, Z. A. J. Am. Chem. Soc. 1972,94,7390-7396.(b) Zavitsas, A. A. J . Am. Chem. SOC.1972,94,2779-2789. (46) For substituent effects on polar transition state see: (a) Walling, C. Free Radicals in Solution; Wiley-Interscience: New York, 1957. (b) Kochi, J. K., Ed. Free Radicals; Wile-Interscience: New York, 1973;Vol. 1 and 2. (c) Pryor, W. A.; Davis, W. H. J . Am. Chem. SOC.1977,99,6365-6372.(d) Davis, W.H.; Gleaton, J. H.; Pryor, W. A. J . Org. Chem. 1977,42, 7-12. (e) Tanner, D. D.; Samal, P. W.; Ruo, T. C-S.; Henriquez, R. J . Am. Chem. SOC.1979,101, 1168-1175. (0 Minisci, F.;Citterio, A. Advances in Free Radical Chemistry; Heyden: Philadelphia, PA, 1980; Vol. 6,pp 65-153. (47) (a) Schultz, J. C.; Houle, F. A.; Beauchamp, J . L. J . Am. Chem. SOC. 1984,106, 3917-3927. (b) Houle, F.A,; Beauchamp, J. L. J . Am. Chem. SOC.1978, 100, 3290-3294. (48) (a) Franklin, J. L.; Dillard, J. G.; Rosenstock, H. M.; Hesson, J. T.; Drax, L. K.; Field, F. H. Ionization Potentials, Appearance Potentials and Heats of Formation of Gaseous Positive Ions; US.Department of Commerce, National Bureau of Standards: Washington, DC, 1969. (b) IPSof tH,OH, CH(CH3)OH,and C(CH&OH were calculated from appearance potentials (AP) of their corresponding carbonium ions" and BDE of the C-H bond being broken.39 (49) (a) Lossing, F. P. J . Am. Chem. SOC.1977, 99,7526-7530. (b) Potapov, V. K.; Sorokin, V. V. High Energy Chem. (USSR) 1972,343-346.
Naguib et al.
3036 The Journal of Physical Chemistry, Vol. 91, No. 11, I987
T -
-e:: f
20
16 12
a 4
I
I
0
7.0
6.0 0
40
20
60
80
Irradiation Tlme (min)
Figure 5. Benzopinacol yield as a function of irradiation time. [K]= 1.15 X lo-' M in acetonitrile, I,, = 2.2 X lo-' einstein/(L.s); experiA); model curve (---); model curve (-) using eq 12. mental points (0. TABLE I: Rate Constants per Hydrogen for Hydrogen Abstraction bv Methvl Radicals and bv Tridet Benzoihenone relative kHb substrate BDE." kcal/mol CH," ,Kd CH3CH2-H 98 1.o 1.oe ~~
HOCHZ-H NCC H 2-H (CH313C-H HO( CH3)zC-H
94 93 92 91
4.5 7.9 173 120
21 0.0091 119 375
0.0
9.0
10.0
11.0
I P of Rodicol ( V I
Figure 6. Rate constants per hydrogen, kH, as a function of ionization potential, IP, of radical R' (log kH, IP): (CH3),C (5.76, 6.7); c-CsH,, (4.78, 7.7); (CH3)3CCH2(3.55, 8.3); C,H5CH2 (5.10, ?.2); CH$N (1.64, 10.9); CH,OH (5.00, 7.6); CH2CH3 (3.68, 8.8); CH(CH3)OH (5.48, 6.7); and C ( C H 3 ) 2 0 H(6.26, 6.4).
contributions to the TS. It is significantly higher, for example, than for jK (CH3)& (E,,, = 4.4 k ~ a l / m o l )where ~ polar contributions (11) should be more important because of the lower IP of (CH3)$CH2 (Figure 6).
+
Experimental Section Benzophenone (Aldrich) was recrystallized from ethanol and further purified by vacuum sublimation. Acetonitrile (Fisher spectrograde) was dried over anhydrous potassium carbonate, then "References 38 and 39. bAbsolute rate constants per H (M-I d) refluxed over phosphorus pentoxide, distilled (bp 81 "C), and kept for CH,' at 437 K and for ,K at 298 K may be obtained by multiplyover molecular sieves 4-8 mesh. Samples (3 mL) were degassed ing by 130 and 4800, respectively. CReference40. "References 7, 8, by four freeze-thaw cycles at torr and then irradiated 42, and 43 except for acetonitrile entry which is this work. 'kH for (reactangular 1 cm path length cells) with stirring at 365 nm with ethane has not been determined directly but can be estimated from the a 200-W mercury-xenon arc and a Bausch and Lomb high-inkH for 2,2-dimethylpropaneassuming the same ratio for the kH's aptensity monochromator. The intensity of the incident light was plies as for the corresponding CH,' abstractions. measured with a vacuum diode (Hamamutsu R847) which had interpreted as reflecting increased contribution of I1 to the TS been previously calibrated against potassium f e r r i ~ x a l a t e . A ~~ as one goes from cyanomethyl to 2-hydroxy-2-propyl. Thus for general description of the lifetime apparatus has been given reaction with 3K, acetonitrile is unreactive because of the very previ~usly.~' low contribution of I1 toward stabilizing the TS and thus in Degassed benzophenone/acetonitrilesolutions (1.1 mM) conlowering the activation energy. From this viewpoint there must tained in 1 cm path length Pyrex cells were excited with a single be only a minimum polar contribution to the TS in the case of unfocused 355 nm, 15 ns, -6 mJ pulse from a frequency-tripled acetonitrile. Methanol is almost 2000 times more reactive than Molectron MY-32 Nd:YAG laser. The resulting transient abacetonitrile, although both have similar BDE's, because of the sorption was detected with a computer-controlled kinetic specrelatively low IP of 'CH,OH and hence larger polar contribution trometer, coaxial with the laser pump beam, consisting of a (11) to the TS. shuttered 100-W tungsten source monochromator (Instruments It is worthwhile noting that linear relationships have been SA, Inc. H-1OV) and photomultiplier (Hamamatsu R928) with observed between log ki,, the interaction rate constant, of aromatic a 200-ohm load resistor and dynode string wired for reduced gain ketones and the IP's of a series of donor molecules. Here k,, is and high linearity.52 The detector signal was stored by a Tektronix essentially the quenching constant of the triplet ketone by the 7A13/7912AD transient digitizer and subsequently processed by donor. The donors include amines, thio and oxy compounds, and Tektronix LSI-11 and IBM PC computers. The cumulative laser certain benzene derivatives with lone pair or easily ionizable exposure of each sample was limited to less than 25 mJ of incident electrons. In these cases the rate constants are fairly large (L106 energy (Le., four shots); no shot-to-shot variation in the decay M-' s-') and it is believed that reaction proceeds predominantly kinetics was observed. via initial charge transfer (CT) from the donor to the k e t ~ n e ' . ~ , ~ HPLC analyses were carried out with a Waters 440 Chromatograph equipped with a 214-nm wavelength extender to in[-0- H-R"] crease the sensitivity to benzopinacol detection. A 5-hm reCT verse-phase column (Waters Nova-Pak GIB) and 40% aqueous In the present study, with compounds of generally lower reactivity acetonitrile gave good resolution of the components of the pholog kH correlates with the IP of the radical formed and high IPS, tolysis mixture. Benzopinacol was identified by coinjection of on hydrogen abstraction reflecting the fact that now reaction is authentic samples and by its signal ratio at 254 and 214 nm. controlled by factors which influence the activation energy for Registry No. Hz, 1333-74-0; benzophenone, 119-61-9; acetonitrile, hydrogen abstraction, as discussed above. 75-05-8; benzopinacol, 464-72-2. A significant number of Arrhenius preexponential factors have been determined for triplet carbonyl hydrogen abstractions7and from these it may be estimated that the A factor for 3K + CH3CN (SO) Hatchard, C. G.; Parker, C. A. Proc. R . SOC.London A 1956, 235, 518. is ~ 2 . X4 lo7 M-I sd. In conjunction with k3 = 130 M-' s-' this (51) Steel, C.; Thomas, T. F. J . Chem. SOC.,Chem. Commun. 1966, yields E,,, E 7 kcal/mol. Such an activation energy is ap900-902. proaching that for a typical abstraction of primary C-H by methyl (52) Porter, G.; West, M. A. Techniques in Chemisfry;Wiley-Interscience: radicals (E,,, E 10 kcal/mol)40 where there are also minimal polar New York, 1974; p 414.