Kinetic studies of a fast, reversible alkene radical cation cyclization reaction

Article abstract of DOI:10.1021/jo061868w

The radical cation formed by mesylate heterolysis from the 1,1-dimethyl-7,7-diphenyl-2-mesyloxy-6-heptenyl radical was studied in several solvents. Computational results suggest that the initially formed acyclic radical cation is a resonance hybrid with p

Full text of DOI:10.1021/jo061868w

Kinetic Studies of a Fast, Reversible Alkene Radical Cation
Cyclization Reaction
John H. Horner and Martin Newcomb*
Department of Chemistry, UniVersity of Illinois at Chicago, 845 West Taylor Street,
Chicago, Illinois 60607
men@uic.edu
ReceiVed September 9, 2006
The radical cation formed by mesylate heterolysis from the 1,1-dimethyl-7,7-diphenyl-2-mesyloxy-6-
heptenyl radical was studied in several solvents. Computational results suggest that the initially formed
acyclic radical cation is a resonance hybrid with partial positive charge in both double bonds of 1,1-
diphenyl-7-methyl-1,6-octadiene (10). Thiophenol trapping was used as the competing reaction for kinetic
determinations. The acyclic radical cation rapidly equilibrates with a cyclic distonic radical cation, and
thiophenol trapping gives acyclic product 10 and cyclic products, mainly trans-1-(diphenylmethyl)-2-
(1-methylethenyl)cyclopentane (11). The rate constants for cyclization at ambient temperature were k )
(0.5-2) × 1010 s-1, and those for ring opening were k ) (1.5-9) × 1010 s-1. Laser flash photolysis
studies in several solvents show relatively slow processes (k ) (2.5-260) × 105 s-1) that involve rate-
limiting trapping reactions for the equilibrating radical cations. In mixtures of fluoroalcohols RfCH2OH
in trifluoromethylbenzene, variable-temperature studies display small, and in one case a negative, activation
energies, requiring equilibration reactions prior to the rate-limiting processes. Fast equilibration of acyclic
and cyclic radical cations implies that product ratios can be controlled by the populations of the acyclic
and cyclic species and relative rate constants for trapping each.
Introduction
attractive feature of radical cation formation by radical het-
erolysis reactions is that nucleophilic trapping can be faster than
molecular correlation times, permitting diastereoselective reac-
tions that preserve chirality in a putative achiral intermediate.10,11
Despite the potential utility of radical cations in synthesis,
some reaction types are not well developed. For example,
cationic and radical cyclization reactions that form five- or six-
membered ring carbocycles are well-known, but related cy-
clization reactions involving a radical cation adding to an alkene
are less well studied.12 A radical cation cycloaddition reaction13
can maintain conjugation that is present in the reactant, but a
5-exo or 6-exo cyclization reaction would convert a localized,
conjugated radical cation into a distonic radical cation. Such a
Radical cations are open-shell intermediates that offer unique
reactivity patterns, more reactive than a radical but more
controllable than a cation.1,2 Typically, radical cations are formed
from unsaturated compounds by chemical,3 electrochemical,4,5
or photochemical6 one-electron oxidation reactions, but they also
can be produced by heterolytic fragmentation of radicals
containing â-leaving groups.7-9 In synthetic applications, an
(1) Schmittel, M.; Burghart, A. Angew. Chem., Int. Ed. 1997, 36, 2550-
2589.
(2) Tanko, J. M.; Phillips, J. P. J. Am. Chem. Soc. 1999, 121, 6078-
6079.
(3) Bauld, N. L. Tetrahedron 1989, 45, 5307-5363.
(4) Moeller, K. D. Tetrahedron 2000, 56, 9527-9554.
(5) Reddy, S. H. K.; Chiba, K.; Sun, Y. M.; Moeller, K. D. Tetrahedron
2001, 57, 5183-5197.
(6) Mangion, D.; Arnold, D. R. Acc. Chem. Res. 2002, 35, 297-304.
(7) Beckwith, A. L. J.; Crich, D.; Duggan, P. J.; Yao, Q. W. Chem. ReV.
1997, 97, 3273-3312.
(8) Crich, D. In Radicals in Organic Synthesis; Renaud, P., Sibi, M. P.,
Eds.; Wiley-VCH: Weinheim, 2001; Vol. 2, pp 188-206.
(9) Crich, D.; Brebion, F.; Suk, D.-H. Top. Curr. Chem. 2006, 263, 1-38.
(10) Crich, D.; Shirai, M.; Rumthao, S. Org. Lett. 2003, 5, 3767-3769.
(11) Crich, D.; Ranganathan, K. J. Am. Chem. Soc. 2005, 127, 9924-
9929.
(12) See: Huang, Y. T.; Moeller, K. D. Org. Lett. 2004, 6, 4199-4202
and references therein.
(13) Bauld, N. L.; Bellville, D. J.; Harirchian, B.; Lorenz, K. T.; Pabon,
R. A., Jr.; Reynolds, D. W.; Wirth, D. D.; Chiou, H. S.; Marsh, B. K. Acc.
Chem. Res. 1987, 20, 371-378.
10.1021/jo061868w CCC: $37.00 © 2007 American Chemical Society
Published on Web 01/27/2007
J. Org. Chem. 2007, 72, 1609-1616
1609

Horner and Newcomb
SCHEME 1
SCHEME 2
product, which might react by internal electron transfer to give
the isomeric radical cation 4, but computational results suggest
that the most stable “acyclic” radical cation is species 5, which
is a resonance hybrid of 3 and 4 (see below). Irrespective of
the detailed structure of the acyclic radical cation, it can react
in 5-exo cyclization reactions to give distonic radical cations 6
and 7, and other cyclic transient radical cations, such as 8, can
be formed.
reaction might be thermodynamically disfavored unless the
distonic radical cation contains good cation- and/or radical-
stabilizing groups. Nonetheless, because localized and distonic
radical cations can demonstrate much different reactivities in
follow-up reactions including nucleophilic captures and second-
ary oxidations, carbocyclizations of alkene radical cations might
be successful even if thermodynamically disfavored when the
timings of the follow-up reactions are appropriate.
We have studied radical cation cyclization reactions to
understand the kinetics of their reactions better and with the
aim of developing radical cation probes14 or clocks15,16 that can
be used in indirect kinetic studies.17 We describe here kinetic
studies of an alkene radical cation 5-exo cyclization reaction
that provide rate and equilibrium constants for the cyclization
and reverse reaction in a variety of solvents. As one would
expect, the cyclization and ring-opening reactions of alkene
radical cations are several orders of magnitude faster than the
corresponding reactions of radicals.
The products obtained from preparative scale radical chain
reactions of precursor 1 in the presence of thiophenol are shown
in Scheme 2. The major products formed were mesylate 9, diene
10, and trans-cyclization product 11, and each of these was
identified by comparison to an authentic sample that was
prepared for characterization purposes. Several minor products
were detected in the reaction mixtures, and samples of two of
those, cis-cyclization product 12 and trans-tricyclic product 13,
also were independently prepared. The Friedel-Crafts cycliza-
tion product 13 was previously reported to be formed from
reaction of diene 10.20
Computational studies were used as a guide for understanding
the species formed in the reactions we studied. Enthalpies of
reactions forming radical cations 3-5 by oxidations of diene
10 were estimated by DFT calculations21 (B3LYP/6-31G*) via
the reactions outlined in Scheme 3. The enthalpy change for
oxidation of diene 10 to the delocalized system 5 was computed
from an isodesmic reaction of 10 with the ethylene radical
cation. For the localized radical cations 3 and 4, we estimated
that the enthalpy changes upon oxidation would be ap-
proximately equal to those for oxidations of the alkenes 14 and
15 and computed the enthalpy change for the isodesmic reactions
of these alkenes with the ethylene radical cation. From these
results (see Supporting Information), the resonance hybrid
Results
The system we studied is described in Scheme 1, which shows
possible transients, and Scheme 2, which shows products. The
PTOC ester18,19 1 was prepared from the corresponding car-
boxylic acid as described in the Supporting Information. Radical
precursor 1 reacted in laser flash photolysis (LFP) reactions and
in photoinitiated radical chain reactions to give the â-mesylate
radical 2. It is important to note that, although LFP and
photoinitiation methods were employed, the reactions studied
in this work are those of electronic ground-state species and
not photochemical reactions. Heterolytic fragmentation of the
mesylate group in 2 can give radical cation 3 as the initial
(20) Ishii, H.; Yamaoka, R.; Imai, Y.; Hirano, T.; Maki, S.; Niwa, H.;
Hashizume, D.; Iwasaki, F.; Ohashi, M. Tetrahedron Lett. 1998, 39, 9501-
9504.
(21) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.;
Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A.
D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi,
M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.;
Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick,
D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.;
Ortiz, J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi,
I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.;
Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M.
W.; Johnson, B. G.; Chen, W.; Wong, M. W.; Andres, J. L.; Head-Gordon,
M.; Replogle, E. S.; Pople, J. A. Gaussian 98, revision A.9; Gaussian,
Inc.: Pittsburgh, PA, 1998.
(14) Schepp, N. P.; Shukla, D.; Sarker, H.; Bauld, N. L.; Johnston, L. J.
J. Am. Chem. Soc. 1997, 119, 10325-10334.
(15) Newcomb, M.; Miranda, N.; Sannigrahi, M.; Huang, X.; Crich, D.
J. Am. Chem. Soc. 2001, 123, 6445-6446.
(16) Horner, J. H.; Taxil, E.; Newcomb, M. J. Am. Chem. Soc. 2002,
124, 5402-5410.
(17) Newcomb, M. Tetrahedron 1993, 49, 1151-1176.
(18) PTOC is an abbreviation for pyridine-2-thioneoxycarbonyl. PTOC
esters react in chain reactions to give acyloxyl radicals that rapidly
decarboxylate to give alkyl radicals. For details, see ref 19.
(19) Barton, D. H. R.; Crich, D.; Motherwell, W. B. Tetrahedron 1985,
41, 3901-3924.
1610 J. Org. Chem., Vol. 72, No. 5, 2007

ReVersible Alkene Radical Cation Cyclization Reaction
SCHEME 3
The approach in Scheme 4 indicated that distonic radical
cation 6 was approximately 6 kcal/mol more stable than distonic
radical cation 7. The gas-phase computational results cannot
be trusted to provide highly accurate energies for charged species
in solution, of course, but the prediction that 6 is more stable
than 7 is supported by experimental studies of a related distonic
radical cation species; radical cation 18 was favored over 19 in
low-polarity media, whereas 18 and 19 were of comparable
energies in a moderate-polarity solvent.26 Interestingly, the
absolute enthalpy for distonic radical cation 6 was only slightly
greater (2 kcal/mol) than that of the delocalized acyclic radical
cation 5, and this enthalpy ranking and difference closely match
the experimental results discussed below in that the equilibrium
between cyclization and ring opening slightly favored the acyclic
species.
SCHEME 4
Thiophenol Trapping Studies. Heterolysis of the mesylate
group in radical 2 was demonstrated in a series of trapping
studies that have been reported.27 In brief, radical 2 reacted
rapidly by heterolysis to give an acyclic radical cation that was
initially formed in a contact ion pair (CIP), and rate constants
for the heterolysis reaction that produces the CIP, for collapse
of the CIP back to neutral radical 2, and for solvation of the
CIP to give a solvent-separated ion pair (SSIP) or free ions were
determined.27 These processes are quite fast, and mesylate
heterolysis to give radical cations occurred with half-lives of
1-20 ns at ambient temperature.27
In preparative reactions studied in this work, radical precursor
1 was allowed to react in chain reactions in the presence of
thiophenol. Trapping studies were performed with varying
concentrations of PhSH in several solvents and solvent mixtures
at ambient temperature. PTOC ester 1 at ca. 10 mM concentra-
tion was allowed to react in radical chain reactions that were
initiated by visible light photolysis. It is important to note that
the reactions were for radicals and radical cations in the ground
state; we did not study photochemical reactions. Following the
reactions, products were identified by HPLC, NMR spectros-
copy, and GC mass spectrometry. The major products were
mesylate 9, diene 10, and trans-cyclization product 11, and
minor products also were obtained (Scheme 2). The yields of
various products were determined by quantitative HPLC. GC
could not be used for quantitative results because mesylate 9
decomposed thermally to give, in part, diene 10. The mesylate
product 9, diene 10, and cyclization product 11 accounted for
more than 90% of the products. Figure 1 shows a typical HPLC
result where minor products are apparent.
radical cation 5 was computed to be 4 kcal/mol more stable
than radical cation 4 and 23 kcal/mol more stable than 3.22
As a delocalized species, resonance hybrid 5 would be
expected to be lower in energy than either of the localized radical
cations, 3 and 4. The computational results indicate that the
positive charge in 5 is concentrated mainly (85%) on the
diphenylethene moiety. Both double bonds in diene 10 are
lengthened upon oxidation to 5. The phenyl-substituted C1-
C2 double bond in 5 is 4.0 pm longer than the corresponding
double bond in diene 10, whereas the C6-C7 double bond in
5 is only 1.7 pm longer than the C6-C7 double bond in 10.
Computations of the enthalpies of formation of the cyclic
distonic radical cations 6 and 7 were less straightforward than
those above because the radical and cation centers are not in
one moiety. We made the assumption that the distonic species
behaved as if the cation and radical centers were isolated from
each other and did not interact. In fact, these centers are close
enough such that they might interact with one another, but a
cancellation of errors was expected when the energetics of the
two distonic radical cations were compared. We calculated the
enthalpy of reactions for the closed shell compounds 16 and
17 reacting with the tert-butyl cation by hydride transfer to
estimate the proton affinities of the cationic intermediates shown
in Scheme 4. The bond dissociation energies (BDEs) of these
species were then approximated using BDE values from the
literature; specifically, we used C-H BDE values of 96 kcal/
mol for isobutane and 80 kcal/mol for 1,1-diphenylethane for
the estimates.23-25
For representative results of the trapping studies, we listed
in Table 1 the ratio of acyclic product 10 to the total amounts
of cyclic products obtained when 1 M thiophenol was employed.
Detailed results of the trapping studies are given in Tables S1-
S6 in the Supporting Information, and Figure 2 shows plots of
acyclic product 10 to cyclic product ratios for three sets of
studies. In Figure 2, four analyses for one or two reactions per
thiophenol concentration were used to give average results with
(22) We judge the results to be reliable at the semiquantitative level
because similar computations for isodesmic reactions of alkenes with known
ionization potentials with an ethylene radical cation had average errors of
2.2 kcal/mol; see Supporting Information.
(23) Rossi, M. J.; McMillen, D. F.; Golden, D. M. J. Phys. Chem. 1984,
88, 5031-5039.
(24) Berkowitz, J.; Ellison, G. B.; Gutman, D. J. Phys. Chem. 1994, 98,
2744-2765.
(26) Horner, J. H.; Bagnol, L.; Newcomb, M. J. Am. Chem. Soc. 2004,
126, 14979-14987.
(25) Halgren, T. A.; Roberts, J. D.; Horner, J. H.; Martinez, F. N.;
Tronche, C.; Newcomb, M. J. Am. Chem. Soc. 2000, 122, 2988-2994.
(27) Horner, J. H.; Lal, M.; Newcomb, M. Org. Lett. 2006, 8, 5497-
5500.
J. Org. Chem, Vol. 72, No. 5, 2007 1611

Horner and Newcomb
SCHEME 5
of the plots required that formation of the cyclic transient be
reversible in the time frame of thiophenol trapping of the radical
cations.17
The competition reaction sequence is shown in Scheme 5,
and the ratio of acyclic product 10 to cyclic products is described
by eq 1,17 where the rate constants are those shown in Scheme
5. Solution of the experimental results by nonlinear regression
FIGURE 1. Portion of HPLC trace (reversed-phase C18 column,
acetonitrile/methanol/water elution) from reaction of 1 in CH2Cl2 in
the presence of 2.0 M PhSH monitored at 222 and 260 nm. Products
9-11 and 13 are labeled with their compound numbers. Product 12
coelutes with 11 under these conditions, but 12 was shown to be a
minor component by GC mass spectrometry.
analysis gives ratios of rate constants that contain kcycle and kopen
.
These ratios will provide absolute rate constants for the
cyclization and ring-opening reaction in Scheme 5 if the rate
constants for the thiophenol reactions are known or estimated.
[10]/[cyclic products] ) (kT1kopen)/(kT2kcycle) +
(kT1[PhSH])/kcycle (1)
TABLE 1. Kinetics from Thiophenol Trapping Experimentsa
kcycle
kopen
solvent
benzene
PhCF3
CH2Cl2
ET(30)b [10]/[cyclics]c
(1010 s-1
)
(1010 s-1
)
We assume that the bimolecular trapping reaction of the
acyclic radical cation by PhSH is a diffusion-controlled reaction
based on the large difference in oxidation potentials for
2-methyl-2-butene28 and thiophenol29 of 1.2 V. Similar assump-
tions that PhSH trapping of radical cations would be a diffusion-
controlled process were made previously by others.26,30,31 The
diffusional rate constant for thiophenol, kdiff, estimated from the
von Smulochowski equation32 and the Stokes-Einstein equa-
tion33 is 2 × 1010 M-1 s-1 at ambient temperature. With this
value for kT1 in Scheme 5, the first-order rate constants for
cyclization of the radical cations can be calculated from the
slopes of the plots in Figure 2.
34.3
38.5
40.7
45.6
52.8
53.8
12 ( 3
6 ( 2
1.3 ( 0.2
1.8 ( 0.1
1.8 ( 0.8
7.0 ( 1.1
3.9 ( 0.3
9 ( 4
3.3 ( 0.1
5.8 ( 0.3
5 ( 1
CH3CN
0.5% TFE/TFTd
1% TFE/TFTd
0.87 ( 0.08 2.7 ( 0.4
0.56 ( 0.03 1.5 ( 0.2
6.2 ( 0.6
a Rate constants at ambient temperature with errors at 2σ; detailed results
b
are in the Supporting Information. ET(30) solvent polarity value from ref
c
34 for pure solvents and measured for mixtures of TFE in TFT. Average
ratio of acyclic product 10 to cyclic products for reactions conducted in
the presence of 1.0 M PhSH. dTFE ) 2,2,2-trifluoroethanol; TFT )
trifluorotoluene.
The rate constants for the cyclization reaction are in good
agreement with one another and vary little with solvent polarity
as judged by the ET(30) solvent polarity values,34 which are
listed in Table 1. The cyclization reactions are very fast, with
rate constants in the 1010 s-1 range. It is noteworthy that 5-exo
cyclization reactions of radicals have log A terms in their
Arrhenius functions in the range of 9.5-10.5.17,35,36 If the log
A values for cyclization of radical cation 5 are in a similar range,
then the entropy-limited rate constants for cyclization at any
temperature would be in the range of 3 × 109 to 3 × 1010 s-1
,
which is comparable to the experimental values in Table 1. Thus,
the cyclization reactions of the acyclic radical cation apparently
have very small or zero activation energies such that they are
highly or completely under entropic control.
FIGURE 2. Ratios of acyclic product 10 to cyclic products from
thiophenol trapping reactions in trifluorotoluene (b), methylene chloride
(9), and trifluorotoluene containing 1% 2,2,2-trifluoroethanol (2). The
lines are fits from eq 1.
(28) Schepp, N. P.; Johnston, L. J. J. Am. Chem. Soc. 1996, 118, 2872-
2881.
(29) Armstrong, D. A.; Sun, Q.; Schuler, R. H. J. Phys. Chem. 1996,
100, 9892-9899.
(30) Giese, B.; Burger, J.; Kang, T. W.; Kesselheim, C.; Wittmer, T. J.
Am. Chem. Soc. 1992, 114, 7322-7324.
errors at 2σ that are shown on the plots. The plots of thiophenol
trapping results demonstrate a competitive reaction sequence17
in which an acyclic radical cation is in equilibrium with one or
more transients that react to give cyclic products (Scheme 5).
The shallow slopes in the product ratio plots show that
thiophenol trapping of the acyclic radical cation(s) was competi-
tive with the cyclization reactions, and the nonzero intercepts
(31) Giese, B.; Beyrich-Graf, X.; Burger, J.; Kesselheim, C.; Senn, M.;
Schafer, T. Angew. Chem., Int. Ed. Engl. 1993, 32, 1742-1743.
(32) Schuh, H. H.; Fischer, H. HelV. Chim. Acta 1978, 61, 2130-2164.
(33) Edward, J. T. J. Chem. Educ. 1970, 47, 261-270.
(34) Reichardt, C. Chem. ReV. 1994, 94, 2319-2358.
(35) Newcomb, M.; Horner, J. H.; Filipkowski, M. A.; Ha, C.; Park, S.
U. J. Am. Chem. Soc. 1995, 117, 3674-3684.
(36) Johnson, C. C.; Horner, J. H.; Tronche, C.; Newcomb, M. J. Am.
Chem. Soc. 1995, 117, 1684-1687.
1612 J. Org. Chem., Vol. 72, No. 5, 2007

ReVersible Alkene Radical Cation Cyclization Reaction
TABLE 2. Rate Constants and Arrhenius Parameters from LFP
Experimentsa
Ea
solventb
ET(30)c
k
(20 °C) (s-1
)
log A
10.1
8.4 ( 0.3
8.3 ( 0.2
7.8 ( 0.1
7.9 ( 0.1
6.9 ( 0.1
(kcal/mol)
PhCF3 (TFT)
TFT/0.5% TFE
TFT/0.0125 M F2
TFT/0.025 M F2
TFT/0.05 M F1
TFT/0.1 M F1
TFT/0.2 M F1
CH2Cl2
38.5
52.8
48.5
49.8
51.8
52.5
53.2
40.7
45.6
55.6
2.5 ( 105
2.4 ( 106
5.8 ( 105
1.0 ( 106
2.6 ( 106
7.5 ( 106
2.6 ( 107
2.6 ( 106
4.7 ( 106
6.0 ( 106
6.3 ( 0.4
2.7 ( 0.4
3.4 ( 0.2
2.4 ( 0.1
2.0 ( 0.1
0.03 ( 0.2
6.3 ( 0.3 -1.5 ( 0.5
CH3CN
CH3CN/5% TFE
10.7 ( 0.2
11.1 ( 0.3
5.4 ( 0.2
5.8 ( 0.3
a Kinetic results for growth of a signal at 334 nm following irradiation
of PTOC precursor 2 with 355 nm laser light. bTFT ) trifluorotoluene;
TFE ) 2,2,2-trifluoroethanol; F1 ) 2,2,3,3,4,4,5,5-octafluoropentanol; F2
) 2,2,3,3,4,4,5,5,6,6,7,7,8,8,8-pentadecafluorooctanol. cMeasured ET(30)
values; see ref 34.
FIGURE 3. Time-resolved UV-visible spectrum from reaction of 1
in CH2Cl2 showing growth of diphenylalkyl radical signals with λmax
) 334 nm; no growth in absorbance was observed for longer
wavelengths than shown here. The time slices are at 207 ns (b), 351
ns (0), 607 ns (1), and 1.98 µs (4) after the laser pulse. Traces at the
indicated times were obtained by subtracting the raw spectra from the
spectrum at 130 ns. The inset shows the kinetic trace at 334 nm.
temperature as found here, the activation energy for the process
is in the range of 2-4 kcal/mol.
One ramification of the kinetic results for the cyclization and
ring-opening reactions is that the equilibrium between acyclic
and cyclic species should increasingly favor the acyclic species
as the temperature is raised. The cyclization is an entropy-limited
process that will not increase in rate with an increase in
temperature, whereas the ring opening has an activation energy
and will be faster at higher temperatures. This observation
presages the results of LFP studies discussed in the next section.
SCHEME 6
Laser Flash Photolysis (LFP) Studies. We conducted LFP
studies in methylene chloride, trifluorotoluene (TFT), and
acetonitrile and in TFT and CH3CN solutions containing 2,2,2-
trifluoroethanol (TFE) or other fluorinated alcohols to increase
solvent polarity. In these studies, PTOC ester 1 again was the
radical precursor. Irradiation of 1 with 355 nm laser light (third
harmonic of a Nd:YAG laser) gave an acyloxyl radical and the
pyridine-2-thiyl radical as first-formed intermediates. Subnano-
second decarboxylation of the acyloxyl radical gave â-mesylate
radical 2, which reacts reversibly by heterolytic fragmentation
of the mesylate group to give a contact ion pair.27 The contact
ion pair lifetimes are in the picosecond time domain,27 and these
species solvate to solvent-separated ion pairs and then to free
ions faster than the response times in our LFP studies. Thus,
the processes evaluated by LFP are mainly those for the free
radical cations. Moreover, the PhSH trapping results in the
previous section demonstrated that equilibration between acyclic
and cyclic distonic radical cations was established faster than
processes studied by LFP methods. As noted earlier, the
reactions are triggered by laser photolysis, but they are not
photochemical reactions, which would be much faster than one
can observe with the nanosecond-response photomultiplier tubes
we used.
Because plots of acyclic product 10 to cyclic product have
measurable intercepts (cf. Figure 2), it is possible to calculate
a ratio that contains the ring-opening rate constant for the cyclic
transient, kopen/kT2.17 If we assume that thiophenol reacts with
the cyclic distonic radical cation in diffusion-controlled reac-
tions, then kT2 ) 2 × 1010 M-1 s-1 at ambient temperature,
and we can calculate rate constants for kopen. Values thus
calculated are listed in Table 1. The apparent result is that the
ring-opening reaction is very fast, and the equilibrium between
an acyclic radical cation and a cyclic distonic radical cation
favors the acyclic species. The assumption that PhSH trapping
of the cyclic distonic radical cations is a diffusion-controlled
process is not as firm as the assumption that reduction of the
acyclic radical cation will be diffusion controlled because the
mode(s) of reaction of the distonic radical cations with PhSH
are not known. Therefore, it is possible that the rate constant
k
T2 in Scheme 5 is smaller than the diffusion-control limit, and
The first observed UV spectra in LFP studies displayed a
broad, weak absorbance at 490 nm from the pyridine-2-thiyl
radical.37 Destruction of the PTOC ester precursor should result
in bleaching in the region of 300-400 nm, and this was
observed for reactions in CH2Cl2. For other solvents, the
bleaching in the 300-400 nm range was small, suggesting the
“instantaneous” formation of a transient with weak absorbance
in this region.
accordingly, the equilibrium in Scheme 5 will be more favorable
for the cyclic product. Despite that caveat, the equilibrium
constant values in Table 1 appear to be consistent with both
LFP studies and computational results discussed above.
Whereas the cyclization reactions approach the entropic limit
for a 5-exo cyclization with a small or zero activation energy,
the ring-opening reactions have larger activation energies. The
entropic demand for ring-opening reactions is small, with log
A values on the order of 12 to 13. For a ring-opening reaction
with a rate constant in the range of (1-10) × 1010 s-1 at ambient
(37) Alam, M. M.; Watanabe, A.; Ito, O. J. Org. Chem. 1995, 60, 3440-
3444.
J. Org. Chem, Vol. 72, No. 5, 2007 1613

Horner and Newcomb
SCHEME 7
FIGURE 4. Rate constants from LFP studies at 20 °C as a function
of the concentration of fluoro alcohol added to TFT solutions.
Among the possible products formed “instantly” in the LFP
studies, 1,1-diarylethene radical cations have absorbances with
λmax ≈ 400 nm,38 diphenylalkyl radicals have λmax ≈ 335
nm,35,36,39,40 and diphenylalkyl cations have λmax ≈ 450 nm.26,38,41
We observed no strong signals in the region 300-450 nm in
any of the initial spectra in TFT, CH2Cl2, and CH3CN solvents,
but strong signals from diphenylalkyl radicals grew in with time
as discussed below. The absence of strong absorbances in the
initial spectra was somewhat surprising, but an independent
oxidation of diene 10 confirmed that the initially formed radical
cations do not absorb strongly. Specifically, we oxidized diene
10 in a photoinduced electron transfer (PET) reaction using a
chloranil triplet as the oxidant. Excitation of chloranil with 355
nm laser light gives the relatively long-lived triplet species,
which is a strong oxidant that reacts with alkenes in diffusion-
controlled processes.26 When the chloranil triplet was produced
by LFP in CH3CN solvent in the presence of diene 10, signals
for the chloranil triplet decayed as this species oxidized 10. New
signals from diphenylalkyl radicals, with λmax at ca. 330 nm,
grew in with time, but no strong signals for the radical cation
from 10 were observed in the first formed spectrum.42
The thiophenol trapping results require that the radical cations
be equilibrated with subnanosecond half-lives. On the slower
kinetic scale evaluated in the LFP studies, strong signals with
λmax ≈ 335 nm grew in with time. Figure 3 shows typical results.
The growing absorbances resembled those for diphenylalkyl
radicals,35,36,39,40 and the subsequent slow decay of these
transients also resembled the behavior of diphenylalkyl radi-
cals.35,36,40 From these spectroscopic results and the product
studies discussed above, reactions shown in Scheme 6 are
important. The acyclic species is in rapid equilibrium with a
cyclic transient that reacts relatively slowly as an acid, as an
electrophile, or in a rearrangement reaction to give diphenylalkyl
radical species such as 20, 21, and 8 that are observed
spectroscopically.
growth of the signal at 335 nm ranged from 2 × 105 s-1 to 6
× 106 s-1 at ambient temperature for trifluorotoluene (TFT),
methylene chloride, and acetonitrile.
Table 1 lists values for kcycle thus calculated for all solvents
evaluated except benzene, in which the data were too scattered
to give meaningful results. The errors in the values of kcycle are
at 2σ and reflect the precision of the experimental data.
The rate constants observed for CH3CN solutions are similar
to the rate constants found for reactions of a simple alkene
radical cation in that solvent.26 In those studies, the predominant
process was reaction of CH3CN as a nucleophile with the radical
cation, a reaction that was previously observed by Arnold when
alkene radical cations were produced in CH3CN in photo-
NOCAS reactions.43 For the system studied here, however,
acetonitrile capture of both the acyclic radical cation and the
cyclic, distonic radical cation is possible, and the latter reaction
should be fast. Thus, two predominant pathways for reactions
are possible in this solvent, as illustrated in Scheme 7, and these
cannot be differentiated from the LFP kinetic data. 5-Exo
cyclizations of 6,6-diphenyl-5-pentenyl radicals have rate
constants greater than 1 × 107 s-1 at ambient temperature,35,36
and the cyclization of radical 22 in Scheme 7 is not likely to be
rate limiting.
When fluorinated alcohols were added to TFT to increase
the solvent polarity, the rate constants for signal growth
increased considerably. A relatively linear plot of log kobs vs
the logarithm of the concentration of fluoro alcohol suggested
that the alcohols reacted as bases and/or nucleophiles (Figure
4). The slope of the log/log plot is 1.4, suggesting that partially
aggregated alcohol (average aggregation number of 1.4) was
involved in rate-limiting processes. We note that the rate
constants appeared to increase with solvent polarity (cf. reactions
in TFT and in CH2Cl2), and it is possible that the kinetic effect
of the fluoro alcohols was due at least in part to increasing
solvent polarity with increasing concentrations of alcohol.
Variable-temperature kinetic results conducted over the
temperature range -30 to 50 °C were quite unusual (Table 2
and Figure 5). The reaction in TFT appeared to be typical with
a log A value of 10.1 and an activation energy of 6.3 kcal/mol.
As the solvent polarity was increased by addition of increasing
amounts of fluorinated alcohols, however, both log A and Ea
values decreased dramatically and eventually reached values
that are not possible for a simple reaction (Table 2).
Rate constants for the processes followed in the LFP studies
are listed in Table 2. These reactions were orders of magnitude
slower than those evaluated in the thiophenol trapping studies.
The observed first-order or pseudo-first-order rate constants for
(38) Ramamurthy, V.; Lakshminarasimhan, P.; Grey, C. P.; Johnston,
L. J. Chem. Commun. 1998, 2411-2424.
(39) Chatgilialoglu, C. In Handbook of Organic Photochemistry; Scaiano,
J. C., Ed.; CRC Press: Boca Raton, FL, 1989; Vol. 2, pp 3-11.
(40) 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.
(41) Faria, J. L.; Steenken, S. J. Phys. Chem. 1993, 97, 1924-1930.
(42) An expectation that 1,1-diaryl alkene radical cations will absorb
strongly at ca. 400 nm might be based on limited data. For example, PET
oxidation of 1,1-diphenylethene and 1,1-diphenylbut-1-ene by the chloranil
triplet in CH3CN solution gave results similar to those with diene 10. Weak
signals centered at ca. 400 nm were observed.
The impossibly small log A terms and small or negative
activation energies require that at least one fast equilibrium
(43) deLijser, H. J. P.; Arnold, D. R. J. Org. Chem. 1997, 62, 8432-
8438.
1614 J. Org. Chem., Vol. 72, No. 5, 2007

ReVersible Alkene Radical Cation Cyclization Reaction
FIGURE 5. Arrhenius functions from LFP studies of 1 in TFT-
containing fluoro alcohols. b, TFT; O, TFT/0.0125 M n-C7F15CH2-
OH; 9, TFT/0.025 M n-C7F15CH2OH; 0, TFT/0.05 M n-C4HF8CH2OH;
2, TFT/0.10 M n-C4HF8CH2OH; 4, TFT/0.20 M n-C4HF8CH2OH.
FIGURE 6. Product ratios from reactions in CH2Cl2. The symbols
are experimentally determined ratios, and the lines are results from
kinetic simulations. The plots are for the ratio [acycle 10]/[cycle 11]
(b), the ratio [mesylate 9]/[cycle 11] (9), and the ratio [mesylate
9]/[acycle 10] (2).
SCHEME 8
presence of thiophenol, the â-mesylate radical 2 can be trapped
in an H-atom transfer reaction by PhSH to give mesylate product
9. The radical heterolysis reaction in competition with radical
trapping is a multistep process. Reversible fragmentation of the
mesylate radical gives a contact ion pair (CIP) containing acyclic
radical cations in equilibrium with cyclic radical cations. Both
radical cations can be trapped by PhSH in the ion pair, or they
can escape to give free ions (FI). Thiophenol trapping of the
acyclic radical cation gives diene 10, and trapping of the cyclic
radical cation gives mainly product 11. The equilibrium constant
for cyclization and ring opening of the radical cations is assumed
to be the same in the CIPs and free ions.
Figure 6 shows results from kinetic modeling for reactions
in CH2Cl2, which are typical. The symbols show experimental
product ratios as a function of thiophenol concentration, and
the lines show the predictions of those ratios from kinetic
modeling. Note that there is no analytical expression for the
ratio of mesylate 9 to cyclic product 11, but the simulated
behavior from the kinetic modeling is in excellent agreement
with the observed ratio of products at various PhSH concentra-
tions. In this example, the rate constants from modeling were
reaction precedes the rate-determining step. In such a situation,
the observed log A value is the total of the log A of the initial
reaction minus the log A of the back-reaction plus the log A of
the second reaction. Thus, again, the LFP results are consistent
with the dynamic situation suggested from the PhSH trapping
results where rapid equilibration of acyclic and cyclic radical
cations occurred. The distonic radical cation should be more
strongly solvated due to its localized charge, and as the
concentration of alcohol in the solvent is increased, the distonic
radical cation might have a more extensively organized solvent
shell, resulting in a ring-opening reaction with an increasing
log A value with increasing alcohol concentration. Because the
log A term for the back-reaction in the equilibration is a negative
value in the total log A expression, increasing the entropy of
the ring-opening reaction would reduce the overall entropy term
for the reactions monitored in the LFP studies.
Kinetic Modeling. One method for evaluating the kinetic
results for the radical cation reactions studied here and the
mesylate heterolysis reactions previously reported27 is to obtain
rate constants by kinetic modeling and compare those empirical
rate constants to the rate constants found with least-squares fits
of analytical rate expressions such as eq 1. We performed kinetic
modeling by numerical integration of the differential rate
equations for the reaction sequence shown in Scheme 8,44 which
reflects the complexity of radical heterolysis reactions. In the
khet ) 6.5 × 108, kcoll ) 1.3 × 1010, ksep ) 1.7 × 1010, kcycle
)
1.8 × 1010, and kopen ) 4 × 1010 s-1, where kH for PhSH
trapping of radical 2 was taken as 1 × 108 M-1 s-1 45,46 and
kET for PhSH trapping of any radical cation was assumed to be
2 × 1010 M-1 s-1. All of the rate constants obtained empirically
in kinetic modeling in this example and for other solvents were
within 1 standard deviation of the least-squares values for the
rate constants for radical heterolysis reactions27 and the rate
constants for cyclization and ring opening found in this work.
The modeling results not only lend support for the accuracies
of the rate constants determined by least-squares solutions of
the analytical rate expressions but also provide some insights
about the limits on rate constants. For example, the model cannot
be fit successfully under any circumstances when the radical
cation cyclization and ring-opening reactions are not fast enough
to establish an equilibrium between acyclic and cyclic radical
cations in competition with PhSH trapping. We note, however,
(45) Franz, J. A.; Bushaw, B. A.; Alnajjar, M. S. J. Am. Chem. Soc.
1989, 111, 268-275.
(46) Newcomb, M.; Choi, S. Y.; Horner, J. H. J. Org. Chem. 1999, 64,
1225-1231.
(44) PSI-Plot, version 7.01 for Windows; Poly Software International
Inc.: Pearl River, NY, 2002.
J. Org. Chem, Vol. 72, No. 5, 2007 1615

Horner and Newcomb
through a gentle stream of N2 from a syringe needle. The solutions
were then shielded from light. The PTOC ester solution (300 µL,
0.033 M) was added, and the solution was further deoxygenated
for 2-3 min as described above. The syringe needle was withdrawn,
and the solution was irradiated with a 150 W tungsten floodlamp
at a distance of ca. 0.5 m for 1-1.5 h. The reaction mixtures were
analyzed directly by HPLC. Mixtures of methanol, acetonitrile, and
water were used as eluant (typically 70:20:10, respectively) for 10
min, followed by a gradient (1 min) to a 70:30 mixture of methanol/
acetonitrile, followed by further elution to 30 min.
Laser Flash Photolysis studies were performed with an Applied
Photophysics LKS-50 laser kinetic spectrometer. The third harmonic
(355 nm, 7 ns duration, 40-60 mJ/pulse) from a Continuum Surelite
1 laser was used in all LFP studies. Data from the photomultiplier
were digitized using an Agilent infinium oscilloscope (model
54845a). When possible, oversampling (64/1) was used to improve
S/N. Samples of radical precursor 1 were prepared in the appropriate
solvent, and the concentration of 1 was adjusted to give an
absorbance of 0.2-0.5 at 355 nm. The samples (typically 250 mL)
were placed in a jacketed addition funnel and deoxygenated by a
slow flow of helium. The temperature of the solution was adjusted
by circulating a temperature-regulated methanol/water mixture
through the external jacket with a circulating bath. The sample was
equilibrated at the desired temperature for 10-20 min and then
allowed to flow through a flow cell at a rate of 5-20 mL per min.
The sample temperatures were measured by a thermocouple inserted
into the flow cell just above the laser and analyzing light paths.
For low-temperature data, the sample and cuvette were enclosed
in a box equipped with quartz windows that was purged with
nitrogen to minimize water condensation on optical surfaces.
that any error in the assumptions that thiophenol reacts with
both types of radical cations with diffusion-controlled rate
constants would result in an error in the absolute values for
rate constants obtained by modeling as well as those obtained
by least-squares fits of the analytical expressions.
Conclusion
Both thiophenol trapping studies and laser flash photolysis
kinetic studies indicate that the acyclic alkene radical cation
formed from solvolysis of the mesylate group in radical 2 rapidly
equilibrates with one or more cyclic radical cations. Compu-
tational results suggest that delocalized acyclic radical cation 5
is the lowest-energy species in the system studied here, and the
kinetic studies are consistent with this conclusion. The 5-exo
cyclization equilibrium is established faster than diffusion-
controlled second-order reactions, and therefore, the product
mixtures formed from this reaction manifold will be determined
largely by the relative rate constants of competing follow-up
reactions. Of considerable interest from the synthetic perspective,
the removal of the phenyl groups in the various radical cation
transients should have a relatively constant effect on all
transients, leading to the conclusion that 5-exo cyclization of a
simple 1,6-diene radical cation also should be fast with an
equilibrium constant close to unity.
Experimental Section
The preparation of PTOC ester 1 and authentic samples of
products from the radical reactions are described in the Supporting
Information.
Reactions of 1 and Analysis for 9-13. Reactions of PTOC ester
1 were conducted in 8 mm (i.d.) Pyrex reaction tubes that were
capped with rubber septa. The tubes were purged for at least 5
min with a continuous flow of N2, and solvent (TFT, DCM,
benzene, or ACN), thiophenol, and, in some cases, TFE were added
to the reaction tube to give a total volume of 700 µL. The solutions
were deoxygenated for 2-4 min by agitating them while passing
Acknowledgment. This work was supported by a grant from
the National Science Foundation.
Supporting Information Available: Synthetic details, tables
of product yields, tables of computational results, and NMR spectra.
This material is available free of charge via the Internet at
http://pubs.acs.org.
JO061868W
1616 J. Org. Chem., Vol. 72, No. 5, 2007