Kinetics of radical heterolysis reactions forming alkene radical cations

Article abstract of DOI:10.1021/ja046089g

Rate constants for heterolytic fragmentation of β-(ester)alkyl radicals were determined by a combination of direct laser flash photolysis studies and indirect kinetic studies. The 1,1-dimethyl-2-mesyloxyhexyl radical (4a) fragments in acetonitrile at ambient temperature with a rate constant of k_het > 5 × 10⁹ s^-1 to give the radical cation from 2-methyl-2-heptene (6), which reacts with acetonitrile with a pseudo-first-order rate constant of k = 1 × 10⁶ s^-1 and is trapped by methanol in acetonitrile in a reversible reaction. The 1,1-dimethyl-2-(diphenylphosphatoxy)hexyl radical (4b) heterolyzes in acetonitrile to give radical cation 6 in an ion pair with a rate constant of k_het = 4 × 10⁶ s^-1, and the ion pair collapses with a rate constant of k ≤ 1 ± 10⁹ s ^-1. Rate constants for heterolysis of the 1,1-dimethyl-2-(2,2- diphenylcyclopropyl)-2-(diphenylphosphatoxy)ethyl radical (5a) and the 1,1-dimethyl-2-(2,2-diphenylcyclopropyl)-2-(trifluoroacetoxy)ethyl radical (5b) were measured in various solvents, and an Arrhenius function for reaction of 5a in THF was determined (log k = 11.16-5.39/2.3RT in kcal/mol). The cyclopropyl reporter group imparts a 35-fold acceleration in the rate of heterolysis of 5a in comparison to 4b. The combined results were used to generate a predictive scale for heterolysis reactions of alkyl radicals containing β-mesyloxy, β-diphenylphosphatoxy, and β-trifluoroacetoxy groups as a function of solvent polarity as determined on the E_T(30) solvent polarity scale.

Full text of DOI:10.1021/ja046089g

Published on Web 10/23/2004
Kinetics of Radical Heterolysis Reactions Forming Alkene
Radical Cations
John H. Horner,* Laurent Bagnol, and Martin Newcomb*
Contribution from the Department of Chemistry, UniVersity of Illinois at Chicago,
845 West Taylor Street, Chicago, Illinois 60607
Received July 1, 2004; E-mail: men@uic.edu
Abstract: Rate constants for heterolytic fragmentation of â-(ester)alkyl radicals were determined by a
combination of direct laser flash photolysis studies and indirect kinetic studies. The 1,1-dimethyl-2-
mesyloxyhexyl radical (4a) fragments in acetonitrile at ambient temperature with a rate constant of khet
>
5 × 109 s-1 to give the radical cation from 2-methyl-2-heptene (6), which reacts with acetonitrile with a
pseudo-first-order rate constant of k ) 1 × 106 s-1 and is trapped by methanol in acetonitrile in a reversible
reaction. The 1,1-dimethyl-2-(diphenylphosphatoxy)hexyl radical (4b) heterolyzes in acetonitrile to give radical
cation 6 in an ion pair with a rate constant of khet ) 4 × 106 s-1, and the ion pair collapses with a rate
constant of k e 1 × 109 s-1. Rate constants for heterolysis of the 1,1-dimethyl-2-(2,2-diphenylcyclopropyl)-
2-(diphenylphosphatoxy)ethyl radical (5a) and the 1,1-dimethyl-2-(2,2-diphenylcyclopropyl)-2-(trifluoro-
acetoxy)ethyl radical (5b) were measured in various solvents, and an Arrhenius function for reaction of 5a
in THF was determined (log k ) 11.16-5.39/2.3RT in kcal/mol). The cyclopropyl reporter group imparts a
35-fold acceleration in the rate of heterolysis of 5a in comparison to 4b. The combined results were used
to generate a predictive scale for heterolysis reactions of alkyl radicals containing â-mesyloxy, â-diphen-
ylphosphatoxy, and â-trifluoroacetoxy groups as a function of solvent polarity as determined on the ET(30)
solvent polarity scale.
â-(Ester)alkyl radicals react in rearrangement, substitution,
and elimination reactions that are analogous to reactions of
closed-shell molecules, although considerably more facile.1,2
During the past decade, studies by several groups reduced the
viable mechanisms for these reactions to two basic types:
concerted reactions and reactions involving heterolytic frag-
mentations that gave transient radical cations.1 The consensus
view in recent years has moved toward heterolytic fragmentation
reactions, especially in cases where the putative radical cation
intermediate is stabilized as in a styrene radical cation or an
enol ether radical cation. For â-(ester)alkyl radicals lacking the
stabilizing groups, the mechanistic picture is less clear. None-
theless, Crich and co-workers have developed synthetic methods
employing â-(phosphatoxy)alkyl radicals with the premise that
alkene radical cations are formed in ion pairs, and their results,
especially in studies of stereoselectivity,3,4 support the frag-
mentation mechanism.
reactivity toward heterolysis was achieved by using â-mesyl-
oxy-, â-(diphenylphosphatoxy)-, and â-(trifluoroacetoxy)alkyl
radicals and by employing a “reporter group” moiety that rapidly
diverts the radical cation intermediate and prevents ion pair
recombination. Kinetic studies of â-mesyloxy- and â-phos-
phatoxyalkyl radical heterolysis reactions in water were reported
previously,5 although, in principle, those reactions could have
involved concerted eliminations.6 Our results provide a predic-
tive kinetic scale for alkyl radical heterolysis reactions in various
organic solvents. They provide no support for concerted reaction
pathways and, thus, imply that heterolysis reactions giving
radical cations are the predominant pathways for all â-(ester)-
alkyl radical reactions.
Results and Discussion
We describe the transients formed in radical heterolysis
reactions in terms of the ion pair model developed for heterolysis
reactions of closed-shell molecules7 and the analogous radical
ion pair model for transients formed by photoinduced electron-
transfer (PET) reactions.8,9 The initial radical heterolysis reaction
gives a contact ion pair (CIP) comprised of a radical cation and
an anion (Scheme 1). The CIP can collapse to the original radical
For â-(ester)alkyl radicals that heterolyze to give styrene
radical cations and enol ether radical cations, laser flash
photolysis (LFP) studies have provided valuable mechanistic
insights. We report here a combination of direct LFP and indirect
kinetic studies of heterolytic fragmentation reactions of â-(ester)-
alkyl radicals that give simple alkene radical cations. High
(5) Koltzenburg, G.; Behrens, G.; Schulte-Frohlinde, D. J. Am. Chem. Soc.
1982, 104, 7311-7312.
(1) Beckwith, A. L. J.; Crich, D.; Duggan, P. J.; Yao, Q. W. Chem. ReV. 1997,
97, 3273-3312.
(6) Zipse, H. Acc. Chem. Res. 1999, 32, 571-578.
(7) Winstein, S.; Clippinger, E.; Fainberg, A. H.; Robinson, G. C. J. Am. Chem.
Soc. 1954, 76, 2597-2598.
(2) Crich, D. In Radicals in Organic Synthesis; Renaud, P., Sibi, M. P., Eds.;
Wiley-VCH: Weinheim, 2001; Vol. 2, pp 188-206.
(3) Crich, D.; Ranganathan, K. J. Am. Chem. Soc. 2002, 124, 12422-12423.
(4) Crich, D.; Shirai, M.; Rumthao, S. Org. Lett. 2003, 5, 3767-3769.
(8) Gould, I. R.; Farid, S. Acc. Chem. Res. 1996, 29, 522-528.
(9) Yabe, T.; Kochi, J. K. J. Am. Chem. Soc. 1992, 114, 4491-4500.
9
10.1021/ja046089g CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 14979-14987
14979

A R T I C L E S
Scheme 1
Horner et al.
Scheme 2
Figure 1. Time-resolved spectra from reactions of radicals 4c (A) and 4a
(B) in acetonitrile over ca. 14 and 2.5 µs, respectively. The spectra are
baseline corrected by subtraction of an early time trace; growing peaks have
positive absorbances, whereas decaying peaks have negative absorbances.
The insets show the kinetic traces at 490 nm (from decay of radical 2) and
at 242 nm for the first 4 µs of the reactions.
or to a rearranged radical, or it can solvate to give a solvent-
separated ion pair (SSIP). In turn, the SSIP can return to the
CIP or further solvate to give diffusively free ions (FI).
Nucleophiles, bases, and reducing agents might react with
radical cations at any stage in the sequence.
Mesylate Heterolysis and Radical Cation Trapping. Figure
1A shows the time-resolved spectral behavior when radical 4c
was produced in acetonitrile. A relatively small signal with λmax
≈ 234 nm grew in with a rate constant equal to those for decay
of the signals that have λmax ) 278 nm and λmax ) 490 nm.
The 490 nm signal is known to arise from radical 2,11 and we
assign the decaying signal at 278 nm to radical 2 also. The
growing signal at 234 nm is due to 2,2′-dipyridyl disulfide,
formed by self-coupling of radical 2, and this assignment is
consistent with the UV-spectrum of 2,2′-dipyridyl disulfide,
which has λmax at 236 nm. Note that there is no indication of a
UV spectrum that can be attributed to radical 4c.
LFP production of radical 4a in acetonitrile was followed by
growth of a strong signal with λmax ) 242 nm in addition to
signal decay at λmax ) 278 nm and at λmax ) 490 nm (Figure
1B). The 242 nm signal grew in faster than the other signals
decayed, indicating the presence of a new transient. Limited
previous studies13,14 reported that alkene radical cations have
weak UV-visible absorbances, but several lines of evidence
indicate that the spectrum observed here was not from alkene
radical cation 6. Trapping studies showed that radical 4a could
not be trapped by PhSH (see below), which requires that the
rate of heterolysis of radical 4a was more than 3 orders of
magnitude faster than signal growth at 242 nm. The 242 nm
The â-(ester)alkyl radical precursors used in this work were
PTOC esters10 1 (Scheme 2), which were prepared from the
corresponding carboxylic acids by conventional methods.
Synthetic details for the precursors are given in the Supporting
Information. Laser photolysis (355 nm) of PTOC esters in He-
sparged flowing solutions cleaved the weak N-O bond to give
the pyridine-2-thiyl radical (2) and an acyloxyl radical (3) that
decarboxylated rapidly (subnanosecond) to give the desired
radical (Scheme 2). Byproduct radical 2 has a long-wavelength
absorbance with λmax ) 490 nm,11 and the initial absorbance
from 2 can be used to estimate yields from radical reactions in
some cases. Radical 2 appears to decay mainly by self-coupling
in a diffusion-controlled process. PTOC esters 1 also were used
as precursors in preparative reactions. The reactions were
initiated by visible-light photolysis, which cleaves the PTOC
esters, the chain sequence was propagated by reactions of
transients with a thiol, and thiyl radicals reacted with another
molecule of PTOC ester to continue the chain reaction.
Precursors 1a-1c gave the series of alkyl radicals 4 that was
studied in LFP and preparative reactions. The radicals contain
an excellent â-mesylate leaving group (4a), a good â-phosphate
leaving group (4b), and a poor â-methoxide leaving group (4c).
Radical 4c was expected not to heterolyze and served as a blank
system for spectroscopy. Precursors 1d and 1e gave radicals 5
containing â-phosphate and â-trifluoroacetate leaving groups
and a reporter-group12 moiety for direct detection in LFP studies.
signal was essentially completely suppressed by 1 × 10-3
M
diisopropylethylamine in CH3CN, consistent with diffusion-
controlled reaction with (undetectable) radical cation 6 before
it could react to give the species absorbing at 242 nm. In
addition, the 242 nm signal was persistent in acetonitrile, even
in the presence of high concentrations of methanol (see below).
(10) The acronym PTOC is for pyridine-2-thioneoxycarbonyl. PTOC esters were
originally developed by Barton’s group and are also known as Barton esters.
See: Barton, D. H. R.; Crich, D.; Motherwell, W. B. Tetrahedron 1985,
41, 3901-3924.
(11) Alam, M. M.; Watanabe, A.; Ito, O. J. Org. Chem. 1995, 60, 3440-3444.
(12) Newcomb, M.; Tanaka, N.; Bouvier, A.; Tronche, C.; Horner, J. H.; Musa,
O. M.; Martinez, F. N. J. Am. Chem. Soc. 1996, 118, 8505-8506.
(13) Mehnert, R.; Brede, O.; Cserep, G. Radiat. Phys. Chem. 1985, 26, 353-
363.
(14) Clark, T.; Teasley, M. F.; Nelsen, S. F.; Wynberg, H. J. Am. Chem. Soc.
1987, 109, 5719-5724.
9
14980 J. AM. CHEM. SOC. VOL. 126, NO. 45, 2004

Kinetics of Radical Heterolysis Reactions
A R T I C L E S
Scheme 3
The spectrum also cannot result from the mesylate anion formed
by heterolysis, which has no major absorbance at wavelengths
longer than 220 nm. Alkene radical cations are known to react
with acetonitrile,15 and we ascribe the new signal to distonic
radical cations 7 formed by the addition of acetonitrile to radical
cation 6 (Scheme 3). Thus, radical cation 6 apparently absorbed
too weakly to be detected, but adduct 7 could be observed. Note
that we show one isomer for radical cation 7 in the scheme,
but two isomers are expected from trapping at C2 and C3 of 6;
DFT calculations indicate that the tertiary radical 7 shown in
the scheme will be about 3 kcal/mol more stable than the
isomeric secondary radical.16
The rate constant for formation of product 7 was relatively
insensitive to solvent polarity. A series of reactions of radical
4a in acetonitrile containing increasing amounts of 2,2,2-
trifluoroethanol (TFE) was performed, and the rate constant for
formation of 7 increased only slightly from k ) 1 × 106 s-1 in
neat acetonitrile to k ) 2 × 106 s-1 in a solution of acetonitrile
containing 5% TFE, which is high in polarity as evaluated by
its ET(30) solvent polarity parameter.17 The modest solvent
polarity effect is different than the effects found in heterolysis
reactions of radicals that give styrene and enol ether radical
cations,18-20 and more dramatic solvent polarity effects also were
found in studies of radicals 5 discussed below, where the rate-
limiting step is the initial radical heterolysis reaction. The small
solvent effect is consistent with the proposal that the rate-limiting
step was the reaction of radical cation 6 with acetonitrile, that
is, kACN ) 1 × 106 s-1. As a corollary, equilibria involving
radical cation 6 and radical 4a or ion pairs containing the radical
cation apparently were only slightly perturbed by changes in
solvent polarity.
Figure 2. (A) Kinetic traces at 242 nm for reactions of 4a in acetonitrile
containing (from the top) 0, 0.0001, 0.001, 0.01, 0.1, and 1.0 M MeOH.
(B) Ratio of products in methanol competition experiments versus methanol
concentration; the values for the regression line are discussed in the text.
confirmed in preparative reactions. When radical 4a was
generated in the presence of MeOH and tert-butyl mercaptan,
2-methyl-2-methoxyheptane and 2-methyl-3-methoxyheptane
were formed in 90% yield and a 1:3.5 ratio (2-MeO:3-MeO).
At the qualitative level, the methanol results indicate that
MeOH intercepted radical cation 6 in competition with aceto-
nitrile trapping. Importantly, the signal at 242 nm that did form
in the presence of methanol was relatively persistent, decaying
on the millisecond time scale at a rate similar to that for decay
in the absence of methanol. This demonstrates that the transient
giving rise to the 242 nm signal cannot be radical cation 6,
which should react relatively rapidly with methanol.
At the quantitative level, the methanol-trapping results were
complex. The rate constant for signal formation at 242 nm was
not significantly altered when methanol was present, even when
the amount of signal formed was considerably reduced. This
kinetic behavior requires that the methanol reaction involved
the formation of another transient in equilibrium with radical
cation 6 and that the new transient gave products in a rate-
limiting reaction that has a rate constant comparable to that for
reaction of 6 with acetontirile.21 We propose that adduct 8 is
produced in a fast equilibrium reaction, and deprotonation of 8
to give methoxy-substituted radical 9 is the rate-limiting step
(Scheme 3). We show one product from the methanol-trapping
reaction, but both the secondary and the tertiary alcohol products
were formed in preparative reactions. This model is the same
as that deduced by Arnold and co-workers for reactions of alkene
radical cations with MeOH in acetonitrile solution,15 and it is
consistent with expectations that the loss of neutral MeOH from
intermediate 8 will be an exceedingly fast reaction.
The interpretation of the results with 4a was supported by a
series of LFP studies in which methanol was added to the
reaction mixtures. In these reactions, the intensity of the signal
at 242 nm was reduced as the concentration of MeOH increased
(Figure 2A), indicating the formation of a methanol-trapping
product, and the formation of methanol-trapped products was
(15) Mangion, D.; Arnold, D. R. Acc. Chem. Res. 2002, 35, 297-304.
(16) See Supporting Information.
(17) Reichardt, C. Chem. ReV. 1994, 94, 2319-2358.
(18) Choi, S. Y.; Crich, D.; Horner, J. H.; Huang, X. H.; Martinez, F. N.;
Newcomb, M.; Wink, D. J.; Yao, Q. W. J. Am. Chem. Soc. 1998, 120,
211-212.
(21) If reactions of radical cation 6 involved a simple competition between
acetonitrile and MeOH trapping, then the observed rate constant for
formation of 7, which equals that for decay of 6, must increase as a function
of MeOH concentration. For example, if 75% of 6 was trapped by MeOH,
then the observed rate constant would be 4 times that found in the absence
of MeOH. In practice, the observed rate constant was essentially unchanged
when the signal at 242 nm was decreased by more than 75%.
(19) Choi, S. Y.; Crich, D.; Horner, J. H.; Huang, X. H.; Newcomb, M.; Whitted,
P. O. Tetrahedron 1999, 55, 3317-3326.
(20) Horner, J. H.; Taxil, E.; Newcomb, M. J. Am. Chem. Soc. 2002, 124, 5402-
5410.
9
J. AM. CHEM. SOC. VOL. 126, NO. 45, 2004 14981

A R T I C L E S
Horner et al.
The methanol concentration effect on the intensity of the 242
nm signal clearly was not linear. As shown in Figure 2A, the
242 nm signal growth was noticeably decreased at concentra-
tions of MeOH as low as 1 × 10-4 M, but it was not completely
suppressed by 0.1 M MeOH. This behavior is consistent with
aggregation of methanol in acetonitrile in a relatively unreactive
form; presumably, the nucleophilicity of hydrogen-bonded
MeOH is reduced considerably from that of monomeric MeOH
in acetonitrile. In such a case, the ratio of products 9 and 7
would be described by eq 1, where the rate constants are those
labeled in Scheme 3, the value KMeOH is a composite of the
equilibrium constant for formation of the methanol adduct and
the dissociation constant(s) for aggregated MeOH, and X is the
effective aggregation number for MeOH. Equation 2 is the
logarithmic form of eq 1.
Reactions of radical 4a in acetonitrile or THF in the presence
of thiophenol confirmed that the heterolysis reaction giving
radical cation 6 was quite rapid. Thiophenol reacts with alkyl
radicals by H-atom transfer with k ≈ 1 × 108 M-1 s-1 24,25
,
but
the only product (>98%) found in preparative reactions of
radical 4a in the presence of up to 1 M PhSH was 2-methyl-
2-heptene, apparently formed by reduction of radical cation 6
by PhSH.26,27 The absence of the H-atom trapping product
2-methyl-3-mesyloxyheptane, which was demonstrated to be
stable to the reaction conditions, establishes a lower limit for
the rate constant for the heterolysis reaction of 4a of khet > 5 ×
109 s-1. It also suggests that collapse of the CIP to return radical
4a was slower than (presumably) diffusion-controlled reduction
of 6 by PhSH and/or that the equilibrium for the heterolysis
reaction favored the CIP.
The accumulated results with â-mesylate radical 4a show that
heterolysis in CH3CN to give radical cation 6 is exceedingly
fast, with a subnanosecond lifetime for 4a. Fast heterolysis
reactions of R-methoxy-â-mesylate,28 R-aryl-â-mesylate,29 and
â-aryl-â-mesylate30 radicals in organic solvents were previously
reported, but these radicals were precursors to stabilized enol
ether and styrene radical cations. In pulse radiolysis studies in
water,5 the rate constant reported for reaction of the radical from
propyl mesylate was 2 × 105 s-1, whereas the rate constants
for reactions of the radicals from butyl mesylate and isobutyl
mesylate were >106 s-1, which reflected the kinetic limit of
the method.5 It appears that the rate constant reported for
heterolysis of the radical from propyl mesylate is anomalous.
The limiting value for the rate constant for heterolysis of 4a,
khet > 5 × 109 s-1, suggests that the equilibrium reaction for
forming the CIP in CH3CN is exergonic in acetonitrile. That
conclusion is based on the assumption that the upper limit for
-1
[9]/[7] ) kdeprotKMeOH[MeOH]1/X(kACN
)
(1)
log([9]/[7]) )
log(KMeOHkdeprot/kACN) + 1/X log([MeOH]) (2)
Using the model of Scheme 3, we calculated the relative
amount of adducts 7 formed for each reaction from the observed
intensity of the 242 nm signal and assigned the remainder to
products 9. Figure 2B is a plot of the log of the product ratio as
a function of the log of MeOH concentration (e.g., eq 2). The
linearity of the plot over a 4 orders of magnitude change in
MeOH concentration is noteworthy. The slope is 0.36 ( 0.03,
indicating that the effective aggregation number for MeOH in
acetonitrile was about 3. The intercept of the plot is 0.56 (
0.06, which is the logarithm of the composite of the rate
constants and KMeOH. Because the observed rate for formation
of the 242 nm signal was relatively constant when MeOH was
present, kdeprot ≈ kACN, in which case KMeOH ≈ 3.6 M-1, where
the rate constant for collapse of the CIP is krec < 4 × 109 s-1
,
which is the rate constant for the highly exothermic collapse in
acetonitrile of the CIP comprised of the diphenylmethyl cation
and the chloride anion.31 In the latter case, the rate of CIP
collapse is controlled by the energy required to reorganize the
acetonitrile solvent; that is, the reaction occurs in the so-called
“solvent polarization caging regime”.31 If the equilibrium
between radical 4a and its CIP in acetonitrile favors the CIP,
then accumulation of diffusively free ions is expected because
K
MeOH is a combination of the equilibrium constant for complex
formation and dissociation constant(s) for aggregated MeOH.22
As noted above, reversible complexation of alkene radical
cations with methanol in acetonitrile solvent was previously
deduced by Arnold and co-workers from the products formed
in photochemical NOCAS (nucleophile-olefin combination,
aromatic substitution) reactions that involve alkene radical cation
intermediates.15 Reversible adduct formation as shown in
Scheme 3 also is consistent with stereoselectivity studies by
Crich and co-workers who observed racemization in cyclization
of an alkene radical cation from a chiral precursor containing a
hydroxy group, whereas the analogous amine-containing radical
cation cyclized stereoselectively.4 Reactions of MeOH with
styrene radical cations23 and with enol ether radical cations20
were reported to be second-order processes with rate constants
on the order of 106-107 M-1 s-1, but it is possible that the
nonlinear concentration effect of MeOH was not apparent in
these studies due to a limited range of MeOH concentrations
studied.
(24) Franz, J. A.; Bushaw, B. A.; Alnajjar, M. S. J. Am. Chem. Soc. 1989, 111,
268-275.
(25) Newcomb, M.; Choi, S. Y.; Horner, J. H. J. Org. Chem. 1999, 64, 1225-
1231.
(26) Reaction of PhSH with enol ether radical cations (presumed to occur by
electron transfer) to give enol ether products was previously shown to be
faster than reactions of the enol ether radical cations with aqueous solvent.
See: Giese, B.; Burger, J.; Kang, T. W.; Kesselheim, C.; Wittmer, T. J.
Am. Chem. Soc. 1992, 114, 7322-7324. Giese, B.; Beyrich-Graf, X.;
Burger, J.; Kesselheim, C.; Senn, M.; Schafer, T. Angew. Chem., Int. Ed.
Engl. 1993, 32, 1742-1743.
(27) A referee noted that the reduction of radical cation 6 by PhSH could be an
inner-sphere process involving nucleophilic capture of the radical cation
by PhSH, deprotonation of the adduct to give the â-phenylthio radical,
and elimination of phenylthiyl radical to give the observed alkene product.
The distinction between an outer-sphere and an inner-sphere reduction of
the alkene radical cation is not critical for the studies reported here, but
ancillary studies appear to support an outer-sphere process. Specifically,
substitution of thioanisole for thiophenol in matched reactions resulted in
only a ca. 20% decrease in the amount of alkene 11 formed in the reaction,
and reaction of radical 4a with MeOH in the presence of t-BuSH gave ca.
90% yield of methyl ethers from methanol trapping. For the latter case,
t-BuSH would be expected to be a better nucleophile than PhSH.
(28) Taxil, E.; Bagnol, L.; Horner, J. H.; Newcomb, M. Org. Lett. 2003, 5,
827-830.
(22) Consistent with this model, MeOD suppressed the 242 nm signal less than
MeOH when similar concentrations were compared. For example, at 0.1
M MeOD, the 242 nm signal was reduced by 30%, whereas the signal was
reduced by 55% with 0.1 M MeOH. Isotope effects are expected both in
the aggregation of MeOD (stronger hydrogen-bonding interactions) and in
the rate of deprotonation of intermediate 8 (slower deprotonation), and the
two isotope effects should work in concert to reduce production of the
product radical 9.
(29) Lancelot, S. F.; Cozens, F. L.; Schepp, N. P. Org. Biomol. Chem. 2003, 1,
1972-1979.
(30) Bagnol, L.; Horner, J. H.; Newcomb, M. Org. Lett. 2003, 5, 5055-5058.
(31) Peters, K. S.; Li, B. L. J. Phys. Chem. 1994, 98, 401-403.
(23) Johnston, L. J.; Schepp, N. P. J. Am. Chem. Soc. 1993, 115, 6564-6571.
9
14982 J. AM. CHEM. SOC. VOL. 126, NO. 45, 2004

Kinetics of Radical Heterolysis Reactions
A R T I C L E S
Scheme 5
is irreversible under the experimental conditions.32 Thus, the
reduction of the radical cation appears to be faster than CIP
collapse when one considers the entire concentration range of
PhSH studied. At the low concentrations of PhSH (<0.1 M),
however, an apparent curvature in the plot suggests that the
collapse reaction was competitive with the radical cation
reduction reaction. If one assumes that reduction of radical cation
6 by PhSH is a diffusion-controlled process with a rate constant
of kET ≈ 1 × 1010 M-1 s-1, one can set a limiting value on the
rate constant for CIP recombination krec < 1 × 109 s-1. If the
apparent curvature in the plot is real, and the infinite dilution
value for the ratio of [10]/[11] is 2.5, then the rate constant for
collapse of the CIP solved from this intercept32 is krec ) 1 ×
109 s-1 (again assuming a diffusion-controlled reaction between
the radical cation and PhSH). The equilibrium population of
the CIP from radical 4b in acetonitrile is small (e0.4%) in any
event. In less polar solvents, the heterolysis will be slower, and
the population of the CIP will be smaller.
Figure 3. Ratios of phosphate 10 to alkene 11 observed in reactions of
0.02 M PTOC ester 1b with PhSH in CD3CN.
Scheme 4
solvation of the CIP in acetonitrile and diffusive escape from
the SSIP will have rate constants in the range of (2-30) × 108
s-1 8,20,31
which are comparable to the expected CIP collapse
,
rate constant.
The PhSH trapping results for 4b are consistent with the LFP
results. The equilibrium constant between radical 4b and the
CIP apparently favors the neutral radical. If the rate constant
for escape from the CIP is on the order of 1 × 108 s-1, then the
rate of formation of diffusively free radical cation 6 could be
Phosphate Heterolysis. LFP generation of â-phosphate
radical 4b in acetonitrile gave time-resolved spectra similar to
those obtained with methoxy radical 4c. The growth of signal
at λ ) 242 nm (5 × 105 s-1) was somewhat faster than signal
decay at λ ) 278 nm (3 × 105 s-1), but the intensity of the 242
nm signal growth relative to the 278 nm signal decay was
comparable to that observed with 4c. Thus, the LFP results
indicate that adduct 7 was not formed in appreciable amounts.
about 10% of the rate of heterolysis of 4b, or ca. 4 × 105 s-1
.
Radical-radical coupling reactions of 4b in the LFP experiments
will compete with the formation of diffusively free 6, resulting
in limited formation of acetonitrile adduct 7.33
Direct Measurements of Radical Heterolysis Rate Con-
stants. Radicals 5 were produced in LFP studies using PTOC
esters 1d and 1e as the precursors. Radicals 5 incorporate a
diphenylcyclopropane reporting group12 adjacent to the leaving
group, and in related R-methoxy-â-substituted radicals, precur-
sors to enol ether radical cations, the same reporter group was
found to promote heterolysis as well as provide a UV-observable
product.34 Heterolysis reactions of radicals 5 will give radical
cation 12, which was expected to rearrange rapidly to distonic
radical cations 13 and/or 14 (Scheme 5). The diphenylalkyl
radical chromophore in 13 will have an absorbance with λmax
≈ 335 nm, whereas the diphenylalkyl cation chromophore in
Preparative reactions of 4b in the presence of thiophenol,
however, demonstrated that the heterolysis reaction giving
radical cation 6 in an ion pair was relatively fast and permitted
an estimation of the rate constant for heterolysis. In a series of
reactions, radical 4b was produced in the presence of varying
concentrations of PhSH in acetonitrile-d3, and the products were
analyzed by NMR spectroscopy. Both 3-(diphenylphosphatoxy)-
2-methylheptane (10) and 2-methyl-2-heptene (11) were formed
in these reactions. The alkane phosphate product 10 was formed
by H-atom trapping of radical 4b by PhSH, and alkene 11 was
formed by reduction of radical cation 6 by PhSH (Scheme 4).26,27
The ratio [10]/[11] is plotted against the concentration of PhSH
in Figure 3. Note that the ratios have relatively large errors due
to the small amounts of alkene formed. It is possible that some
of alkene 11 was consumed by addition of PhSH, but we did
not observe thiophenol addition products.
(32) Newcomb, M. Tetrahedron 1993, 49, 1151-1176.
(33) Schulte-Frohlinde and co-workers reported rate constants for the heterolysis
of two â-phosphate radicals that gave isobutene radical cation in water of
(1-3) × 104 s-1.5 The studies involved pulse radiolysis of aqueous solutions
to generate the hydroxyl radical, which abstracted H-atom from the
substrate, and time-resolved conductivity measurements. From our results,
one would predict that reaction of radical 4b in water would be much faster,
k ≈ 4 × 108 s-1. The diphenyl phosphate group in 4b is more reactive
than the dialkyl phosphate groups in the earlier study, and a more stable
trisubstituted radical cation is formed by heterolysis of 4b. Possibly most
importantly, the rate constant for reaction of 4b is that for formation of
radical cation in an ion pair, whereas the early study measured the rate of
acid formation from reaction of the radical cation with water.
The slope of the plot in Figure 3 is 23 ( 4 M-1 (error at 2σ).
The slope is equal to kH/khet, and, by using a rate constant for
reaction of radical 4b with PhSH24,25 of 1 × 108 M-1 s-1, one
obtains a rate constant for heterolysis of 4b of khet ) 4 × 106
s-1. The intercept of the plot in Figure 3 is effectively zero,
and a zero intercept would result when the heterolysis reaction
(34) Newcomb, M.; Miranda, N. J. Org. Chem. 2004, 69, 6515-6520.
9
J. AM. CHEM. SOC. VOL. 126, NO. 45, 2004 14983

A R T I C L E S
Horner et al.
Table 1. Rate Constants for Heterolysis Reactions of Radicals 5
at Ambient Temperaturea
solvent
ET(30)
5a
5b
cyclohexane
toluene
30.9
33.9
37.4
38.5
45.6
0.64
7.6
17
1.5
4.2
2.7
THF
TFTb
21
acetonitrile
140c
43
a Observed rate constants at 22 ( 2 °C in units of 106 s-1
.
b TFT )
trifluorotoluene. c The rate constant for reaction of 5a in acetonitrile at 20
°C was calculated from a measured rate constant of 4.6 × 107 s-1 at -20
°C and a log A term of 11.16; see text.
from the enol ether reaction show decay of Chl* and growth of
the chloranil radical anion. Subtraction of the spectra from the
enol ether oxidation reaction from those obtained in the
oxidation of alkene 15 gave difference spectra showing two
absorbances from reactions of radical cation 12 (Figure 4) that
are similar to those in the spectra obtained from the reaction of
radicals 5 in acetonitrile.
Figure 4. Product spectra from reactions of radical 5a in TFT (b) and in
acetonitrile (0) and from oxidation of alkene 15 by the chloranil triplet in
acetonitrile (2). The spectra were scaled to arbitrary absorbance units and
offset for clarity. Note that the spectra from radical 5a are for the products
that form in the reactions; the spectrum of byproduct radical 2 and the initial
bleaching effect were subtracted from the experimental spectra.
Scheme 6
Ultrafast ring opening of the reporter group in radical cation
12 was demonstrated by reaction of â-phosphate radical 5a in
the presence of PhSH in acetonitrile. From the LFP results (see
below), radical 5a heterolyzes in acetonitrile at ambient tem-
perature with a rate constant of k ) 1.4 × 108 s-1. In the
preparative reaction with 1 M PhSH, we obtained no detectable
amount of alkene 15 from trapping of radical cation 12.
Assuming a conservative limit of 10% for the detection of 15
and diffusion-controlled reduction of alkene radical cation 12
by PhSH, the ring opening of 12 would have a rate constant k
> 1 × 1011 s-1 at ambient temperature. That limit is larger
than the rate constants for dynamic phenomena in ion pairs,
including collapse of the CIP in CH3CN. Therefore, the observed
rate constants for reactions of radicals 5 apparently are the rate
constants for the initial heterolysis reactions that produce 12 in
a CIP. It is possible that reactions of radicals 5 proceeded
directly to ring-opened products, that is, reaction by concerted
cleavage of the leaving group and a cyclopropane bond, but
that would not complicate the kinetic analysis.
14 was expected to absorb with λmax ≈ 450 nm.35,36 The 2,2-
diphenylcyclopropylcarbinyl radical cleaves with a rate constant
of 5 × 1011 s-1 at ambient temperature,37 and ring opening of
radical cation 12 was expected to be even faster. Thus, it was
anticipated that the ring-opening reporting reaction was likely
to compete with reactions of radical cation 12 in an ion pair.
Radicals 5 were studied in a variety of solvents. Figure 4
shows the final product spectrum for reaction of 5a in trifluo-
romethylbenzene (trifluorotoluene, TFT), which is typical for
the reactions in low polarity solvents. The signal with λmax
)
334 nm is from distonic radical cation 13 or a product from 13
in which the cation moiety has reacted with solvent. The ultimate
signal intensity at λ ) 334 nm was compared to the initial signal
intensity at λ ) 490 nm (from byproduct radical 2) to estimate
yields of 13, which were essentially quantitative for reactions
in THF and TFT and ca. 75% for reactions of 5a in cyclohexane.
In the medium polarity solvent acetonitrile, the signal intensity
at λ ) 334 nm from reactions of radicals 5 was reduced, and a
new signal with λmax ≈ 460 nm was observed (Figure 4). In
the reaction of radical 5b in CH3CN at ambient temperature
and in reactions of 5a in CH3CN at low temperatures, the two
signals grew in with comparable rates, suggesting that they arose
from a mixture of products 13 and 14.
The identities of products 13 and 14 were confirmed by
oxidation of alkene 15 in acetonitrile by the chloranil triplet
(Chl*), a strong oxidant formed by 355 nm photolysis of
chloranil (Scheme 6). Triplet chloranil oxidized alkene 15 in a
diffusion-controlled reaction, and time-resolved spectra were
obtained. For the analysis of these spectra, we subtracted time-
resolved spectra obtained when Chl* was produced in the
presence of dihydropyran in an otherwise matched reaction. The
enol ether and its radical cation have no appreciable absorbance
in the UV-visible range studied here, and time-resolved spectra
The rate constants for heterolysis of radicals 5 in various
solvents at ambient temperature are listed in Table 1. In low
polarity solvents, product 13 was formed, and a mixture of 13
and 14 was formed in acetonitrile. The rate constant for reaction
of radical 5a in acetonitrile at ambient temperature was greater
than the temporal limit of our unit. To determine a value for
the reaction at 20 °C, rate constants were measured at low
temperatures and used with the log A term found for reaction
of 5a in THF (see below) to calculate the rate constant for
reaction at ambient temperature.
The heterolysis reaction of radical 5a in acetonitrile provides
data for evaluation of the kinetic effect of the cyclopropane
group. Radical 5a reacted 35 times faster than radical 4b, and
the accelerating effect of the cyclopropane in 5a is nearly an
order of magnitude greater than the acceleration due to the
phenyl group in radical 16, where we take the rearrangement
rate constant for isomerization of 16 to its benzylic radical
product in acetonitrile (kobs ) 1.8 × 107 s-1 18,19
)
to be the rate
constant for heterolysis. In fact, radical 5a reacted faster than
radical 17, which gives the corresponding radical cation with a
rate constant of khet ) 8 × 107 s-1 in CH3CN at ambient
(35) Faria, J. L.; Steenken, S. J. Phys. Chem. 1993, 97, 1924-1930.
(36) Johnston, L. J. Chem. Commun. 1998, 2411-2424.
(37) Newcomb, M.; Johnson, C. C.; Manek, M. B.; Varick, T. R. J. Am. Chem.
Soc. 1992, 114, 10915-10921.
9
14984 J. AM. CHEM. SOC. VOL. 126, NO. 45, 2004

Kinetics of Radical Heterolysis Reactions
A R T I C L E S
Figure 5. Rate constants for heterolysis reactions of radicals 5a (b) and
5b (0) at ambient temperature as a function of solvent polarity. The rate
constant for reaction of 5a in cyclohexane was not included in the regression
line.
Figure 6. Temperature-dependent plots for heterolysis of radical 5a in THF
(b) and in acetonitrile (9). The regression lines are the Arrhenius functions
given in the text.
temperature.38 The kinetic effect of the cyclopropane in the
heterolysis of the radical mirrors the behavior in heterolytic
fragmentations of closed-shell molecules, where a cyclopropyl
group has a greater accelerating effect than a phenyl group.39,40
Table 1 also lists the ET(30) solvent polarity values for the
various solvents. The ET(30) solvatochromic values correlate
with many polar reactions,17 including radical heterolysis
reactions,18,19,30,41 and can be used to estimate solvent polarity
effects. Plots of log k versus ET(30) are shown in Figure 5. For
the â-phosphate radical 5a, the rate constant for reaction in
cyclohexane clearly is out of line with the other rate constants,
and the reduced yield of 13 from reaction of 5a in cyclohexane
suggests that there were unknown reactions in this solvent.
Excluding the cyclohexane rate constant, the slope for the data
for 5a is 0.109 ( 0.004. For the heterolysis of radical 5b, the
regression line has a slope of 0.12 ( 0.02. In heterolysis
reactions of â-aryl-â-(ester)alkyl radicals that give styrene
radical cation products, plots of log k versus ET(30) had slopes
of 0.109 ( 0.00430 and 0.128 ( 0.003.41 The similarity in the
values for the different systems suggests a generality in
transition-state structures and degrees of polarization for various
radical heterolysis reactions.
The heterolysis reaction of radical 5a in THF was studied
over the temperature range from -39 to 51 °C (Figure 6). These
data give an Arrhenius function of log k ) (11.16 ( 0.12) -
(5.39 ( 0.15)/θ, where θ ) 2.3RT in kcal/mol, and errors are
at 2σ. Limited low-temperature studies (-40 to -4 °C) of 5a
in acetonitrile gave a low-precision Arrhenius function with a
similar entropic term: log k ) (11.0 ( 0.4) - (3.8 ( 0.4)/θ.
The entropic terms found here are similar to those found in a
number of radical heterolysis reactions that give styrene radical
cations and enol ether radical cations.18-20,30 Although one might
anticipate that the fragmentation will be entropy releasing,
radical organization in the transition state and solvent reorga-
Figure 7. Estimated rate constants at ambient temperature for heterolysis
of â-(ester)alkyl radicals that give trisubstituted alkene radical cations: (A)
mesylates (lower limit), (B) diphenyl phosphates, (C) trifluoroacetates. The
gray bars are the ET(30) values for common organic solvents (ACN )
acetonitrile).
nization to accommodate charge formation apparently result in
negative entropies of activation.
Kinetic Model for Radical Heterolysis Reactions in
Organic Solvents. One objective of this research was to provide
a predictive background for radical heterolysis reactions that
give unsubstituted alkene radical cations in typical organic
solvents. Correlations of the rate constants for heterolysis of
radicals 5 with solvent polarity provide the most important
information in that regard. The magnitudes of the solvent effects
on the rates of these reactions were similar to those found for
radical heterolysis reactions that give styrene and enol ether
radical cations, and it seems likely that solvent effects for
radicals lacking the reporter group moiety also will be similar.
For a working model, we assume that the correlations shown
in Figure 5 will apply for simple â-(ester)alkyl radicals; that is,
that rate constants will correlate with ET(30) values with slopes
of 0.115. The limit for the rate constant for heterolysis of
mesylate radical 4a in acetonitrile (khet > 5 × 109 s-1) and the
rate constant for heterolysis of phosphate radical 4b in aceto-
nitrile (khet ) 4 × 106 s-1) provide anchor points on a kinetic
scale for these types of alkyl radicals. The average ratio of rate
constants for reaction of radicals 5a and 5b in the common
solvents studied was 5; therefore, we assume that â-trifluoro-
acetate alkyl radicals will react 0.2 times as fast as â-diphenyl
phosphate alkyl radicals, or khet ) 8 × 105 s-1 in CH3CN. The
kinetic values are collected in Figure 7, which provides estimates
for the rate constants for heterolysis reactions of the various
radicals at ambient temperature as a function of solvent polarity
on the ET(30) scale.17
(38) Bales, B. C.; Horner, J. H.; Huang, X. H.; Newcomb, M.; Crich, D.;
Greenberg, M. M. J. Am. Chem. Soc. 2001, 123, 3623-3629.
(39) Brown, H. C.; Peters, E. N. J. Am. Chem. Soc. 1977, 99, 1712-1716.
(40) Kirmse, W.; Krzossa, B.; Steenken, S. J. Am. Chem. Soc. 1996, 118, 7473-
7477.
(41) Whitted, P. O.; Horner, J. H.; Newcomb, M.; Huang, X. H.; Crich, D.
Org. Lett. 1999, 1, 153-156.
9
J. AM. CHEM. SOC. VOL. 126, NO. 45, 2004 14985

A R T I C L E S
Scheme 7
Horner et al.
fragmentation of the radical, and the CIP is favored at
equilibrium. The rate constants at ambient temperature for
diffusive escape from various CIPs in acetonitrile are in the
range of (2-30) × 108 s-1 8,20,31
,
and these rate constants,
combined with the favorable equilibrium for CIP formation, will
result in large concentrations of diffusively free ions. Collapse
of the CIP produced from â-phosphate radical 4b in acetonitrile
was estimated to have a rate constant on the order of krec ) 1
× 109 s-1, however. This collapse rate constant could be up to
an order of magnitude faster than escape from the CIP in
CH3CN. CIP escape rate constants in solvents less polar than
acetonitrile will be slower, and it is likely that effectively no
diffusively free ions can form from â-phosphate radicals in low-
polarity solvents.
The predictions in Figure 7 appear to be reasonable. For
example, Ingold and co-workers studied the rearrangement of
the â-trifluoroacetoxy radical shown in Scheme 7.42 The rate
constant for rearrangement in CF2ClCFCl2 at 75 °C obtained
by kinetic ESR spectroscopy was k ) 7 × 104 s-1. Using a log
A term of 11.2 and adjusting for the fact that the method
typically gave rate constants that were too small by a factor of
2, the estimated rate constant for the rearrangement at 20 °C
would be k ≈ 1 × 104 s-1. From Figure 7, the predicted rate
constant for heterolysis of â-trifluoroacetoxy radicals that give
trisubstituted alkene radical cations in CF2ClCFCl2, ET(30) )
33.2,17 at 20 °C is k ≈ 3 × 104 s-1. It would appear that the
rearrangement in Scheme 7 proceeds by initial heterolysis and
ion pair recombination.
Conclusion
Radical heterolysis reactions offer an attractive synthetic entry
to radical cations, but little was known about the kinetics of
radical heterolysis reactions that give simple alkene radical
cations. The present work resulted in kinetic values that provide
a foundation for understanding these reactions and might be
used for designing synthetic conversions. An overriding theme
is the remarkably low barriers for the charge-forming fragmen-
tation reactions, even though the alkene radical cation products
do not contain stabilizing aryl or ether groups. The â-mesylate
radical heterolysis reaction in acetonitrile is faster than ion pair
collapse, and this feature will result in the rapid accumulation
of diffusively free radical cations. The â-phosphate radical
heterolysis is a relatively fast process that should compete
favorably with H-atom transfer in typical synthetic radical chain
reaction applications. The estimated rate constants for radical
heterolysis in various solvents shown in Figure 7 can be
improved by detailed studies of specific radicals, but we believe
they will provide useful starting points for synthetic planning
purposes.
The estimated rate constants in Figure 7 can be used in
combination with known rate constants for radical trapping
reactions to predict the relative amounts of radical trapping and
heterolysis that would be expected in a typical radical chain
reaction. â-Mesylate radicals heterolyze so rapidly that they
should fragment faster than H-atom trapping by tin hydride even
in a low polarity solvent such as benzene, and they are an
obvious choice if one wishes to generate an alkene radical cation
for an intermolecular reaction. â-Trifluoroacetate radicals in
benzene, on the other hand, would be largely reduced by tin
hydride (kH ) 2 × 106 M-1 s-1 43
)
at 0.1 M concentration.
In addition to the predicted rate constants at ambient
temperature for heterolytic fragmentations, one can estimate
temperature effects on these reactions. The log A values for
heterolysis of radical 5a found here for reactions in THF and
acetonitrile (log A ) 11.2) are quite similar to those found in
other radical heterolysis reactions that gave styrene radical
cations and enol ether radical cations,18-20,30 and it seems
reasonable to assume that this entropic term will be ap-
proximately correct for most radical heterolysis reactions. The
log A value reflects net entropy demand in the fragmentation
transition state,44 but intermolecular reactions typically have
larger negative entropies of activation, resulting in log A terms
of ca. 9. The result is that bimolecular reactions will be less
sensitive to temperature effects than the heterolysis reactions.
One important point is that the rate constants for heterolytic
fragmentation are not equivalent to the rate constants for
formation of diffusively free radical cations. In the case of an
isomerization reaction or an intramolecular nucleophilic capture
reaction, the radical cation might react in the CIP, but
intermolecular capture is limited by diffusional processes and
cannot compete with CIP collapse. CIP partitioning between
collapse to the starting radical or its isomer and diffusive escape
will be controlled largely by the stability of the ions and the
solvent polarity. For the CIP produced from â-mesylate radical
4a, CIP collapse in acetonitrile was slower than heterolytic
Experimental Section
The synthetic procedures for compounds used in this work are
described in the Supporting Information.
Laser Flash Photolysis Studies. Acetonitrile (HPLC grade) used
in LFP studies was purchased from Fisher Scientific and used as
received. LFP studies were performed with Applied Photophysics LKS-
50 and LKS-60 kinetic spectrometers using Nd:YAG lasers with
irradiation at 355 nm. Solutions of the PTOC precursor having an
absorbance of 0.2-0.5 AU at 355 nm were placed in a thermally
jacketed addition funnel and sparged with a stream of helium for 15-
30 min. The temperature of the solution was adjusted by circulation of
a water/methanol solution from a regulated bath through the jacket.
The solutions were allowed to flow through a quartz cell (1 cm × 1
cm i.d.). Temperatures were measured with a thin wire copper/
constantan thermocouple placed in the flowing stream immediately
above the irradiation region of the flow cell. For studies conducted
below 10 °C, the addition funnel and flow cell were placed in a
nitrogen-purged box fitted with quartz windows. For kinetic runs,
several traces typically were averaged to improve signal-to-noise. Curve
fitting of kinetic traces was achieved with Applied Photophysics
software supplied with the spectrophotometers.
Photolysis of 1b in CD3CN with Thiophenol. A stock solution of
the PTOC ester (1b) was prepared in CD3CN. A solution of thiophenol
in CD3CN (1 mL) contained in a 5 mm NMR tube or a 1-mL volumetric
flask was deoxygenated for 2 min with a stream of nitrogen from a
syringe needle inserted into the tube through a rubber septum cap. A
stock solution of PTOC ester precursor in CD3CN (100 or 200 µL)
(42) Barclay, L. R. C.; Lusztyk, J.; Ingold, K. U. J. Am. Chem. Soc. 1984, 106,
1793-1796.
(43) Chatgilialoglu, C.; Newcomb, M. AdV. Organomet. Chem. 1999, 44, 67-
112.
(44) At ambient temperature, log A ) 13.1 when ∆Sq ) 0.
9
14986 J. AM. CHEM. SOC. VOL. 126, NO. 45, 2004

Kinetics of Radical Heterolysis Reactions
A R T I C L E S
was added, and nitrogen was passed through the solution for an
additional minute. The initial thiophenol concentration ranged from 0.04
to 0.8 M, and the PTOC ester concentration was 0.02 M. The solutions
were irradiated with a 150 W floodlamp for 15-20 min. The ratio of
for 3 min with a stream of nitrogen. A solution of 1a (0.1 M, 100 uL)
in CD3CN was added. The solution was deoxygenated for a further 3
min and irradiated for 40 min with a 150 W flood lamp. A solution (9
mL) of ethylbenzene (0.128 mg/mL) in THF was added, and the
resulting mixture was analyzed by gas chromatography (DB-5 column).
Response factors were determined with authentic samples of the two
ethers. Yields of 70% ( 3% and 20% ( 0.7% were obtained for
2-methoxy-2-methyl-2-heptane and 3-methoxy-2-methyl-2-heptane,
respectively.
1
10 to 11 was determined by H NMR spectroscopy by integration of
the signals for H-3 of 10 (δ 4.96) and of 11 (δ 5.13). The error in the
integrals was estimated to be (10% of the measured value, and errors
in the ratios were determined by standard error propagation. In Figure
3, the [10]/[11] ratio was plotted against the average thiophenol
concentration estimated as [PhSH]average ) 0.5([PhSH]initial + [PTOC-
]
final). The results of the studies are listed in Table S2 in the Supporting
Acknowledgment. This work was supported by a grant from
the National Science Foundation (CHE0235293).
Information.
Photolysis of 1a in CD3CN with Thiophenol. The procedure was
the same as that described above for 1b. 2-Methyl-2-heptene (11) was
the only product observed by NMR spectroscopy. Yields of 11
determined by GC (DB-5 column) with ethylbenzene as an internal
standard were 61-76%.
Supporting Information Available: Synthetic methods, de-
tailed kinetic results, and NMR spectra. This material is available
free of charge via the Internet at http://pubs.acs.org.
Photolysis of 1a in Methanol with tert-Butyl Thiol. A solution
(0.9 mL) of tert-butyl thiol (0.05 M) in methanol was deoxygenated
JA046089G
9
J. AM. CHEM. SOC. VOL. 126, NO. 45, 2004 14987