Laser flash photolysis kinetic studies of enol ether radical cations. Rate constants for heterolysis of α-methoxy-β-phosphatoxyalkyl radicals and for cyclizations of enol ether radical cations

Article abstract of DOI:10.1021/ja0177399

Two series of enol ether radical cations were studied by laser flash photolysis methods. The radical cations were produced by heterolyses of the phosphate groups from the corresponding α-methoxy-β-diethylphosphatoxy or β-diphenylphosphatoxy radicals that

Full text of DOI:10.1021/ja0177399

Published on Web 04/12/2002
Laser Flash Photolysis Kinetic Studies of Enol Ether Radical
Cations. Rate Constants for Heterolysis of
r-Methoxy-â-phosphatoxyalkyl Radicals and for Cyclizations
of Enol Ether Radical Cations
John H. Horner,* Elsa Taxil, and Martin Newcomb*
Contribution from the Department of Chemistry, UniVersity of Illinois at Chicago,
845 West Taylor Street, Chicago, Illinois 60607
Received December 11, 2001
Abstract: Two series of enol ether radical cations were studied by laser flash photolysis methods. The
radical cations were produced by heterolyses of the phosphate groups from the corresponding R-methoxy-
â-diethylphosphatoxy or â-diphenylphosphatoxy radicals that were produced by 355 nm photolysis of
N-hydroxypryidine-2-thione (PTOC) ester radical precursors. Syntheses of the radical precursors are
described. Cyclizations of enol ether radical cations 1 gave distonic radical cations containing the
diphenylalkyl radical, whereas cyclizations of enol ether radical cations 2 gave distonic radical cation products
containing a diphenylcyclopropylcarbinyl radical moiety that rapidly ring-opened to a diphenylalkyl radical
product. For 5-exo cyclizations, the heterolysis reactions were rate limiting, whereas for 6-exo and 7-exo
cyclizations, the heterolyses were fast and the cyclizations were rate limiting. Rate constants were measured
in acetonitrile and in acetonitrile solutions containing 2,2,2-trifluoroethanol, and several Arrhenius functions
were determined. The heterolysis reactions showed a strong solvent polarity effect, whereas the cyclization
reactions that gave distonic radical cation products did not. Recombination reactions or deprotonations of
the radical cation within the first-formed ion pair compete with diffusive escape of the ions, and the yields
of distonic radical cation products were a function of solvent polarity and increased in more polar solvent
mixtures. The 5-exo cyclizations were fast enough to compete efficiently with other reactions within the ion
pair (k ≈ 2 × 109 s-1 at 20 °C). The 6-exo cyclization reactions of the enol ether radical cations are 100
times faster (radical cations 1) and 10 000 times faster (radical cations 2) than cyclizations of the
corresponding radicals (k ≈ 4 × 107 s-1 at 20 °C). Second-order rate constants were determined for reactions
of one enol ether radical cation with water and with methanol; the rate constants at ambient temperature
are 1.1 × 106 and 1.4 × 106 M-1 s-1, respectively.
Scheme 1
Enol ether radical cations are among the simplest members
of the radical cation family, and they are readily produced in
chemical or electrochemical oxidations of closed-shell enol
ethers. These intermediates also are produced in non-oxidative
reactions involving heterolytic fragmentation of an R-alkoxy
oxidative conditions.5,6 Despite the biochemical and potential
radical containing a â-leaving group (Scheme 1). Radical
heterolyses, producing radical cations and anions,1,2 are impli-
cated in DNA degradation reactions where a C4′ sugar radical
eliminates phosphate to give an enol ether radical cation3 that
can oxidize guanosine, and this entry to DNA radicals was used
by Giese’s group in studies of electron (or hole) transfer in
DNA.4 The radical heterolysis entry to radical cations also has
good synthetic potential because it permits formation of
intermediates that are highly reactive in radical-like reactions
without exposing other sensitive groups in a molecule to
synthetic relevance of enol ether radical cations, however,
reactions of these species are poorly understood in terms of both
the mechanisms and kinetics of their reactions.7
The lack of kinetic information for enol ether radical cations
in general and the absence of calibrated reactions that can be
used in competition kinetic experiments8 requires that kinetics
be determined directly, but direct UV monitoring is difficult
because the chromophores of simple enol ether radical cations
are far into the UV with λmax near 200 nm.7 In such a situation,
one can use a reporter (or probe) reaction wherein a detectable
* Corresponding author. E-mail: men@uic.edu.
(1) Behrens, G.; Bothe, E.; Eibenberger, J.; Koltzenburg, G.; Schulte-Frohlinde,
D. Angew. Chem. 1978, 90, 639.
(5) Crich, D.; Huang, X. H.; Newcomb, M. J. Org. Chem. 2000, 65, 523-
529.
(6) Crich, D.; Ranganathan, K.; Huang, X. H. Org. Lett. 2001, 3, 1917-1919.
(7) Bernhard, K.; Geimer, J.; Canle-Lopez, M.; Reynisson, J.; Beckert, D.;
Gleiter, R.; Steenken, S. Chem. Eur. J. 2001, 7, 4640-4650.
(8) Newcomb, M. Tetrahedron 1993, 49, 1151-1176.
(2) Gilbert, B. C.; Norman, R. O. C.; Williams, P. S. J. Chem. Soc., Perkin
Trans. 2 1980, 647-656.
(3) Steenken, S.; Behrens, G.; Schulte-Frohlinde, D. Int. J. Radiat. Biol. 1974,
25, 205-210.
(4) Giese, B. Acc. Chem. Res. 2000, 33, 631-636.
9
5402
J. AM. CHEM. SOC. 2002, 124, 5402-5410
10.1021/ja0177399 CCC: $22.00 © 2002 American Chemical Society

Enol Ether Radical Cations
A R T I C L E S
Scheme 2
Scheme 3
product is formed in competition with the reaction that is being
measured, and the desired kinetics are extracted from the sum
of the rate constants for both processes. Probe reactions for
reactive transients often are bimolecular processes, and a
bimolecular protocol for enol ether radical cation detection was
reported.9 In this work, we describe the preparation and kinetic
studies of unimolecular enol ether radical cation reporter
reactions that were developed for applications in laser flash
photolysis (LFP) kinetic methods.10,11 The unimolecular ap-
proach offers advantages in terms of the reproducibility of the
reporter reaction kinetics and in the broad scope of reactions
that can be studied, and our results also provide information
about the processes occurring in the ion pair produced in the
radical heterolysis reactions.
Scheme 4 a
Results
The experimental requirements for enol ether radical cation
clocks for LFP applications are that the radical cations are
produced efficiently and that the reporter reaction provides a
readily detectable signal. Desirable, but not requisite, properties
are that the initial reactions giving the radical cation clocks are
fast and that a method for determining yields exists. We were
able to incorporate all of these properties into the clocks by
using N-hydroxypyridine-2-thione (PTOC) ester12 precursors to
R-alkoxy radicals that contain â-phosphate leaving groups. We
designed two series of clocks (1, 2) that ultimately produce
diphenylalkyl radicals that contain a strong long-wavelength
chromophore centered in the range 330-335 nm.
a Key: (a) t-BuO2CCH(OMe)CH3, LDA. (b) (RO)2P(O)Cl. (c) TFA,
CH2Cl2. (d) 2,2′-Dipyridyl disulfide bis-N-oxide, Bu3P.
radicals 5. Heterolytic fragmentation of the phosphate groups
in radicals 5 gave radical cations 1. Cyclization of radical cations
1 gave the distonic radical cations 6 that were detected at 330-
335 nm. Radical cations 1a and 1b were produced from both
â-diethylphosphatoxy and â-diphenylphosphatoxy radicals, and
radical cation 1c was produced from the â-diphenylphosphatoxy
radical. The leaving group can have an effect on the kinetics
and yields of products, and we used a labeling that identifies
the phosphate group. For example, radical 5b-E is the â-diethyl
phosphate radical that gives radical cation 1b, whereas radical
5b-P is the â-diphenyl phosphate radical that gives the same
radical cation, 1b.
The reaction sequence for radical cations 2 (Scheme 3) is
similar to that for radical cations 1. The difference in this
sequence is that distonic radical cations 9, formed by cyclizations
of 2, further react by cyclopropylcarbinyl radical ring openings
to give the detectable diphenylalkyl radicals 10. We used
â-diethyl phosphate radicals as the precursors to radical cations
2 because the synthetic route to the â-diphenyl phosphate
precursors was precluded (see below).
Preparation of Precursors. The synthetic sequence for the
precursors to radicals 1 is shown in Scheme 4. Hydroxyalkyl-
ation of the enolate from the tert-butyl ester of 2-methoxy-
propanoic acid with the known aldehydes 11 gave R-methoxy-
â-hydroxy esters 12 that were not isolated but were phos-
phorylated with either diethyl- or diphenylphosphoryl chloride
The reaction sequence for radical cations 1 is shown in
Scheme 2. Laser flash photolysis (LFP) of the PTOC esters 3
gave acyloxyl radicals 4 that decarboxylated “instantly” on the
nanosecond time scale to give the R-methoxy-â-phosphatoxy
(9) Newcomb, M.; Miranda, N.; Huang, X. H.; Crich, D. J. Am. Chem. Soc.
2000, 122, 6128-6129.
(10) A portion of this work has been communicated. See ref 11.
(11) Horner, J. H.; Newcomb, M. J. Am. Chem. Soc. 2001, 123, 4364-4365.
(12) Barton, D. H. R.; Crich, D.; Motherwell, W. B. Tetrahedron 1985, 41,
3901-3924.
9
J. AM. CHEM. SOC. VOL. 124, NO. 19, 2002 5403

A R T I C L E S
Horner et al.
Scheme 5 a
a Key: (a) RO2CCH(OMe)CH3, LDA. (b) (EtO)2P(O)Cl. (c) NaOH or
KOH, EtOH, H2O. (d) 2,2′-Dipyridyl disulfide bis-N-oxide, Bu3P.
to give phosphates 13-E and 13-P, respectively. In each case,
one diastereomer of 13 was obtained by column chromatogra-
phy. We assign (2RS,3RS) stereochemistry to the major dia-
stereomer on the basis of the analogous reactions reported by
Heathcock,13 but the stereochemistry was not further studied
because it is not important; a chiral center is lost in formation
of the radical, giving the same product radical from each
diastereomer. The esters were converted to the parent carboxylic
acids 14 by reaction with TFA. Acids 14 were then converted
to the PTOC esters 3 by conventional methods.
The synthetic sequence for the precursors to radical cations
2 (Scheme 5) was similar to that employed for the precursors
to 1. The intermediate methyl esters or 2,4-dimethoxybenzyl
esters 17 were saponified to give the â-diethyl phosphate
carboxylic acids 18. We were not able to prepare â-diphenyl
phosphate precursors to radical cations 2. The â-diphenyl
phosphates were not stable to basic saponification conditions
and, thus, required TFA hydrolysis of tert-butyl esters, but the
vinylcyclopropane moiety in these compounds was not stable
to TFA treatment.
Spectra and Yields. Figure 1A shows a time-resolved
spectrum after laser irradiation of PTOC ester 3b-E (precursor
to radical cation 1b) that is representative of the spectra from
reactions of radical cations 1. The signal growing in with λmax
at 332 nm is from the diphenylalkyl radical 6b, and the position
of the absorbance is similar to that of the diphenylmethyl
radical.14 The bleaching centered at 360 nm is from destruction
of the PTOC ester precursor. The absorbance centered at 490
nm is from the byproduct of the photolysis reaction, the pyridine-
2-thiyl radical for which the molar extinction coefficient has
been reported.15 Figure 1B shows the time-resolved spectrum
following photolysis of PTOC ester 7b (precursor to radical
cation 2b) that has the same features as above, in this case,
from the final product, distonic radical cation 10b.
The insets in Figure 1 show kinetic traces recorded at 335
and 490 nm. The initial intensity of the 490 nm signal from the
pyridine-2-thiyl radical was used as a standard for measuring
the ultimate yields of diphenylalkyl radicals formed from
radicals 1 and 2. We used molar extinction coefficients of ꢀ )
3 200 for the pyridine-2-thiyl radical15 and ꢀ ) 20 000 for the
diphenylalkyl radicals. An internal check on the relative
Figure 1. Time-resolved spectra from reactions of radicals 1b (A) and 2b
(B). The bleachings centered at 360 nm are from destruction of the PTOC
ester precursors. The insets show the kinetic traces at 335 (λmax for
diphenylalkyl radicals) and 490 nm (λmax for the 2-pyridinethiyl radical).
Table 1. Percent Yields of Product Distonic Radical Cations 6 and
10a
b
c
c
c
radical
ACN
0.5% TFE
1.0% TFE
2.5% TFE
5a-E
5a-Pc
8a
5b-E
5b-P
8b
50
85
65
14
34
25
12
44
98
80
42
61
80
22
65
100
92
100
95
70
80
100
60
20
65
80
90
47
10
5c-P
8c
a Yields of products 6 and 10 determined by comparison of the ultimate
diphenylalkyl radical absorbance to the initial thiopyridyl radical absorbance.
The estimated accuracy is (10% of the value given. b ACN ) acetonitrile.
c TFE ) 2,2,2-trifluoroethanol. The percentage of TFE in ACN is listed.
c 15% yield observed in THF.
extinction coefficients was possible because, in several cases,
the yields of diphenylalkyl radicals reached values close to 100%
in high polarity solvent.
Table 1 lists the yields of distonic radical cations 6 and 10
as a function of solvent. Because the leaving group has an
influence on the yields, we identified the specific â-phospha-
toxyalkyl radicals that were produced in the initial photolysis
reactions. Reactions were conducted in acetonitrile (ACN) and
in ACN solutions containing 2,2,2-trifluoroethanol (TFE). A
solvent polarity effect was apparent with yields consistently
increasing for all radical cations when the solvent polarity was
increased. Figure 2 shows the effect for radical 5b-E that is
especially dramatic. A leaving group effect also was apparent
in that the product yields from the diphenylphosphatoxy radicals
5a-P and 5b-P were consistently greater than those from the
diethylphosphatoxy radicals 5a-E and 5b-E, respectively, even
though the same radical cations (1a and 1b, respectively) were
formed from these pairs. As discussed later, the reduced yields
(13) Heathcock, C. H.; Pirrung, M. C.; Young, S. D.; Hagen, J. P.; Jarvi, E. T.;
Badertscher, U.; Ma¨rki, H.-P.; Montgomery, S. H. J. Am. Chem. Soc. 1984,
106, 8161-8174.
(14) Chatgilialoglu, C. In Handbook of Organic Photochemistry; Scaiano, J.
C., Ed.; CRC Press: Boca Raton, FL, 1989; Vol. 2, pp 3-11.
(15) Alam, M. M.; Watanabe, A.; Ito, O. J. Org. Chem. 1995, 60, 3440-3444.
9
5404 J. AM. CHEM. SOC. VOL. 124, NO. 19, 2002

Enol Ether Radical Cations
A R T I C L E S
Figure 2. Kinetic traces at 330 nm for reaction of radical 5b-P in
acetonitrile containing 2,2,2-trifluoroethanol in the amounts shown on the
labels. In each case, the 330 nm signal was normalized against the intensity
of the 490 nm signal (from the pyridine-2-thiyl radical) obtained in the
same experiment.
Figure 3. Arrhenius functions for reactions determined in this work. The
reactions of radicals 5a-E and 8a involve rate-limiting heterolyses, whereas
the reactions of radical cations 1b and 2b involve rate-limiting cyclization.
in the lower polarity solvents must be due to reactions of the
first-formed radical cation with the phosphate anion in the ion
pair produced in the heterolysis reaction.
Table 2. Arrhenius Functions for Reactions of Radicals 5 and 8
temp
range
The greatest yield of distonic radical cation product found in
ACN containing no TFE was formation of 6a in 85% yield from
reaction of radical 5a-P. When the reaction of 5a-P was
conducted in the lower polarity solvent THF, the yield of 6a
was reduced to about 15%. In discussion, we will argue that
diffusive escape from the initially formed ion pair most likely
does not occur in THF, and the yield of distonic radical cation
6a reflects competition within the ion pair between cyclization
of the radical cation 1a and reactions that consume it.
Kinetics. In the sequences of reactions shown in Schemes 2
and 3, the initial photolytic cleavages and the subsequent
decarboxylation reactions are faster than the temporal resolution
of the nanosecond kinetic spectrometer. In the case of radicals
2, the ring opening “reporter” reactions of distonic radical
cations 9 have rate constants (krep) of about 4 × 1011 s-1 at
ambient temperature.16 Therefore, the “slow” reactions for which
kinetics are obtained in both series of radical cations 1 and 2
are the heterolysis reactions of the R-methoxy-â-phosphatoxy-
alkyl radicals 5 and 8 (khet) that produce the enol ether radical
cations, the cyclizations of radical cations 1 and 2 to distonic
radical cations 6 and 10, respectively, (kcycle) or a convolution
of these processes.
LFP kinetic studies were performed in ACN and in solutions
of ACN containing TFE. Arrhenius functions were determined
for several cases (Figure 3 and Table 2). Table 3 contains rate
constants at ambient temperature. For cases where Arrhenius
functions were determined, the rate constants in Table 3 are
those calculated for 20 °C from the Arrhenius function; for other
cases, the kinetic values are those measured at (20.0 ( 0.2) °C.
As developed in the Discussion, the rate-limiting processes
for radical cations 1a and 2a were the heterolysis steps, whereas
the 6-exo and 7-exo cyclization reactions were the rate-limiting
processes for radical cations 1b, 2b, 1c, and 2c in ACN
containing TFE. For radical 8b at low reaction temperatures,
the kinetic traces showed a convolution of two processes (Figure
4), and the kinetics of each were obtained by deconvolution
c
radical
solventa
Arrhenius functionb
rate-limiting heterolysis reactions
(10.32 ( 0.22) - (3.66 ( 0.30)/θ
(9.71 ( 0.10) - (2.94 ( 0.12)/θ
5a-E
ACN
ACN
10 to 51
-30 to 47
-30 to 30
8a
0.5% TFE (9.10 ( 0.11) - (1.57 ( 0.14)/θ
rate-limiting cyclization reactions
1b (from
5b-E)
1b (from
5b-P)
ACNd
(9.9 ( 0.5) - (3.7 ( 0.6)/θ
20 to 51
2.5% TFE (10.99 ( 0.14) - (4.76 ( 0.19)/θ -10 to 50
ACN
(10.90 ( 0.17) - (4.57 ( 0.23)/θ -10 to 50
0.5% TFE (11.03 ( 0.36) - (4.77 ( 0.50)/θ
1.0% TFE (10.96 ( 0.32) - (4.69 ( 0.45)/θ
10 to 49
10 to 50
2.5% TFE (10.68 ( 0.25) - (4.29 ( 0.33)/θ -10 to 50
2b (from 8b) ACNe
(10.47 ( 0.17) - (4.23 ( 0.21)/θ -32 to 50
2.5% TFE (10.52 ( 0.12) - (3.82 ( 0.15)/θ -32 to 52
a ACN ) acetonitrile; TFE ) 2,2,2-trifluoroethanol. The percentage of
TFE in ACN is listed. b Errors given at 2σ. θ ) 2.303RT in kcal/mol.
c Temperature range studied in °C. d Undeconvoluted data; the heterolysis
and cyclization reactions have similar rate constants. e Data deconvoluted
to correct for slow heterolysis; the rate constants are for the cyclization
reaction.
Table 3. Rate Constants for Reactions of Radicals 5 and 8 at
Ambient Temperaturea
b
b
b
b
radical
ACN
0.5% TFE
1.0% TFE
2.5% TFE
5a-E
5a-Pc
8a
5b-E
5b-P
8b
3.9 × 107
>2 × 108
3.3 × 107
1.4 × 107
3.1 × 107
2.1 × 107
3.5 × 105
8 × 107
1.1 × 108
>2 × 108
>2 × 108
>2 × 108
2.7 × 107
3.0 × 107
4.7 × 107
1.2 × 105
1× 105
>2 × 108
8.5 × 107
2.4 × 107
3.0 × 107
3.9 × 107
2.8 × 105
>2 × 108
2.6 × 107
2.9 × 107
4.5 × 107
1.4 × 105
2 × 105
5c-P
8c
a In units of s-1. Values in bold face are rate constants at 20 °C calculated
from the Arrhenius functions in Table 2. Other rate constants are measured
values in the range (20 ( 0.2) °C. The instrumental limit is 2 × 108 s-1
.
b ACN ) acetonitrile; TFE ) 2,2,2-trifluoroethanol. The percentage of TFE
in ACN is listed. c Rate constant in THF >2 × 108 s-1
.
(data are given in the Supporting Information). The Arrhenius
function and rate constant reported for this radical in Tables 2
and 3 are for the 6-exo cyclization reaction. Convolution of the
heterolysis and cyclization kinetics was expected for radical
5b-E reacting in ACN as discussed later, but due to the low
(16) Newcomb, M.; Johnson, C. C.; Manek, M. B.; Varick, T. R. J. Am. Chem.
Soc. 1992, 114, 10915-10921.
9
J. AM. CHEM. SOC. VOL. 124, NO. 19, 2002 5405

A R T I C L E S
Horner et al.
propylcarbinyl radicals have rate constants of about 5 × 1011
s-1 at ambient temperature.16
The solvent effects on the kinetics for reactions of 5a-E and
8a (diethyl phosphate leaving group, 5-exo cyclizations) indicate
that the rate-limiting processes are the heterolysis reactions. The
kinetics for these two radicals are quite similar in ACN and in
0.5% TFE in ACN, and they dramatically accelerate with
increasing solvent polarity such that the reactions are faster than
the instrument limit in 2.5% TFE.
Radical 5a-P gives the same radical cation as radical 5a-E
(i.e. 1a), but for 5a-P the kinetics for this reaction were too
fast to measure in ACN demonstrating that the heterolysis, and
not the cyclization step, was rate-limiting. The heterolysis
reaction was still too fast to measure when radical 5a-P was
allowed to react in the low polarity solvent THF. The increased
rate constant for reaction of 5a-P in comparison to 5a-E is
consistent with the decreased basicity of the diphenyl phosphate
group compared to that of the diethyl phosphate group; that is,
diphenylphosphoric acid is about 1 pKa unit more acidic than
diethylphosphoric acid.18
Figure 4. (A) Time-resolved growth spectrum at 20 ns intervals from 61
ns after photolysis of PTOC ester 7b in acetonitrile at 21 °C. The spectral
trace at 41 ns was subtracted from each spectrum to give a zero baseline.
(B) Convoluted kinetic trace at 335 nm following photolysis of 7b in
acetonitrile at -20 °C. The initial bleaching is from destruction of the
precursor.
yield of product radical 6b, the signal-to-noise was not adequate
for deconvolution. For this case, we report the apparent first-
order rate constant of the reaction and note that this is a lower
limit.
The heterolysis rate constants for radicals 5a-E, 5a-P, and
8a are in qualitative agreement with those found previously. In
reactions of radicals 19-E and 19-P,9 the enol ether radical cation
produced is less stable than radical cations 1 and 2 in that it
lacks one alkyl group, and the rate constants for heterolyses of
radicals 19 determined by LFP studies of solutions in ACN were
smaller than those found in the present work.9 Consistent with
observations here, heterolysis of radical 19-P was considerably
faster than heterolysis of 19-E. Radical 20, which gives an enol
ether radical cation with a substitution pattern the same as 1
and 2, was found to heterolyze in MeOH with a rate constant
To demonstrate the use of the radical cation clocks in LFP
kinetic studies, we determined rate constants for reactions of
radical cation 1c with water and with methanol at ambient
temperature in acetonitrile containing 2.5% TFE. PTOC ester
3c-P was employed as the precursor. Reactions were conducted
with varying amounts of water or methanol. The observed rate
constant (kobs) for reaction of 1c is given by eq 1 where kcycle is
the first-order rate constant for cyclization of 1c, kTrap is the
second-order rate constant for reaction of the radical cation with
water or methanol, and [Trap] is the concentration of water or
methanol in a particular run. Assuming that the rate constant
for cyclization 1c is essentially constant over the concentration
range of water or methanol employed, a plot of kobs versus [Trap]
will have a slope of kTrap. The rate constant for reaction of water
with 1c was (1.1 ( 0.1) × 106 M-1 s-1, and the rate constant
for reaction of methanol with 1c was (1.4 ( 0.4) × 106 M-1
s-1. We note that these rate constants could be for reaction of
water and methanol as nucleophiles, as bases that deprotonate
the radical cation, or a combination of nucleophilic capture and
deprotonation.
exceeding 3 × 109 s-1 19
. We return to these kinetic values below
when we discuss the multiple reactions that comprise the
“heterolysis reaction”.
The Arrhenius log A terms for 5a-E and 8a deserve special
note. If the entropy of activation were zero at 298 K, then the
log A term would be 13.1. Fragmentation reactions can have
positive entropies of activation giving log A values of up to
16,20 but that would be expected for gas-phase reactions or
homolytic cleavages in solution that do not produce charge. The
heterolytic cleavage reactions we are studying here do create
charge in the transition state, and the small log A terms found
indicate a high degree of organization in these cleavages. Highly
defined organization in the cleavage of radicals 5a and 8a might
be required for efficient charge-stabilization in the incipient
radical cations and phosphate anions, and similar log A terms
were found for â-phosphatoxy and â-acetoxy radicals that
heterolyzed to give styrene radical cations.21-24 It is also possible
kobs ) kcycle + kTrap [Trap]
(1)
Discussion
Rate-Limiting Reactions. A series of fast processes are
involved in the reactions studied in this work (Schemes 2 and
3), but the phosphate heterolysis (khet) reactions and cyclization
reactions (kcycle) are the “slower” processes for which rate
constants were measured. The photochemical cleavages of the
N-O bonds in the PTOC esters are instantaneous reactions on
the nanosecond time scale, and the decarboxylation reactions
of acyloxyl radicals to give alkyl radicals occur in less than 1
ns.17 In the case of radical cations 2, cyclopropylcarbinyl radical
ring openings follow the cyclization reactions that give distonic
radical cations 9, and the ring openings of 2,2-diphenylcyclo-
(18) Peppard, P. F.; Mason, G. W.; Andrejasich, C. M. Inorg. Nucl. Chem. 1965,
27, 697-709.
(19) Gugger, A.; Batra, R.; Rzadek, P.; Rist, G.; Giese, B. J. Am. Chem. Soc.
1997, 119, 8740-8741.
(20) Benson, S. W. Thermochemical Kinetics, 2nd ed.; Wiley: New York, 1976.
(21) 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.
(17) Bockman, T. M.; Hubig, S. M.; Kochi, J. K. J. Org. Chem. 1997, 62, 2210-
2221.
9
5406 J. AM. CHEM. SOC. VOL. 124, NO. 19, 2002

Enol Ether Radical Cations
A R T I C L E S
that the solvent is highly organized in the TS of the heterolysis
to accommodate the formation of charge.
case of alkyl radicals 21 and 22 where the diphenylethene
moiety in 21 results in about a 100-fold increase in the rate of
cyclization in comparison to 22. The origin of this effect is not
apparent, but it clearly is not a result of an offset between
activation energies and entropies of activation for the cycliza-
tions. One possible explanation is that the diphenylethenyl group
inherently accelerates the cyclization by providing a more stable
radical product as it does for simple radical cyclizations, but a
steric interaction of the highly substituted radical cation center
with this moiety works to retard the cyclization. Some support
for this conjecture is found in the cyclizations of tertiary R-ester
and R-amide radicals 23b that are 2 orders of magnitude slower
than cyclizations of the analogous secondary radicals 23a,
apparently due to steric effects involving the planar radical
center.26,29
For reactions in the “b” series (the 6-exo cyclizations of
radical cations 1b and 2b), the cyclization reactions and not
the heterolyses were rate limiting in ACN solutions containing
TFE. These reactions were found to be slower than those in
the “a” series in the same solvent mixture, and the solvent
polarity effects on the kinetics were minor. The good match in
the rate constants from radicals 5b-E and 5b-P, both of which
give radical cation 1b, demonstrates that the heterolyses steps
were not involved in the kinetics.
In ACN without TFE, however, the observed rate constants
for the diethyl phosphate-containing radicals 5b-E and 8b were
convolutions of the rate constants for heterolysis and cyclization.
The diphenylphosphatoxy radical 5b-P in ACN must heterolyze
much faster than the measured rate by analogy to radical 5a-P,
and the observed rate constant for 5b-P in ACN is that for the
cyclization of radical cation 1b. In the case of radical 8b reacting
to give radical cation 2b, we could deconvolute the two rate
constants, but this was not possible for radical 5b-E because
the yield was low, and the signal-to-noise was poor. Thus, the
reported kinetic values for 5b-E in ACN are from first-order
solutions to data that are convoluted. Because the convolution
of similar rate constants results in an apparent first-order rate
constant that is smaller than the actual values, the reported rate
constant for “cyclization” of radical 1b in ACN of 1.4 × 107
s-1 at 20 °C is a lower limit; the actual value will be close to
Only limiting values could be obtained for the kinetics of
the 7-exo cyclizations of enol ether radical cations 1c and 2c
because these cyclization reactions were slow enough such that
bimolecular reactions were important in the overall rates of
reaction. Competing reactions are indicated for radical cation
2c by the low yields of product 10c and the fact that the observed
rate constants increased when the reactions were conducted with
increasingly dilute solutions (data not shown). The yields of
distonic radical cation product 6c from radical cation 1c are
similar to the product yields from radical cation 1b produced
from the phosphate radical precursor 5b-P, but the observed
decrease in rate constants as the solvent polarity was increased
suggests that second-order reactions of 1c with phosphate anion
are increasingly important in the lower polarity solvent mixtures.
Because the observed rate constant is the sum of all rate
constants for reactions of 1c, the values we obtained must be
upper limits for the rate constants for cyclizations.
2 × 107 s-1 25
.
The Arrhenius parameters for the reactions in the “b” series
are consistent with those expected for 6-exo cyclizations if one
uses the results of radical cyclizations as a guide. Specifically,
the entropic terms (log A) for cyclizations of 1b and 2b are
quite similar to those for analogous carbon-centered radi-
cals.8,26,27 It is perhaps more interesting to focus on the activation
energies for these cyclization reactions that are the key features
in the kinetics. The enol ether radical cation cyclizations were
expected to be much faster than the cyclizations of analogous
radicals, and that is the case for radicals 1b and 2b, which
cyclize at ambient temperature about 100 and 10 000 times faster
than radicals 21 and 22, respectively.27,28 If one considers the
cyclization reaction to be a typical “radical reaction”, the radical
cations might be regarded as super-radicals.
Ion-Pair Reactions. The ultimate yields of diphenylalkyl
radical products 6 and 10 provide information about reactions
occurring in the ion pairs initially produced by the heterolysis
reactions. The 6-exo cyclizations of 1b and 2b were slow enough
such that cyclization occurred for diffusively free radical cations
only (see above). If one focuses on the yields of products from
the 6-exo cyclization reactions in Table 1, one sees that the
yields appreciably increased as the solvent polarity increased
by addition of TFE. This solvent polarity effect on yield reflects
increasing amounts of radical cation that successfully escape
from the initially formed ion pairs as the solvent polarity was
increased.
Scheme 6 shows the possible reactions in the ion pair. The
initial heterolysis reaction produces an ion pair with a dialkyl
phosphate anion positioned next to the enol ether radical cation.
Collapse of the ion pair will return the â-phosphate radical or
a rearranged radical 24. If the original radical is produced in an
ion pair collapse reaction, it will simply heterolyze again.
Radical 24 might heterolyze back to the ion pair rapidly, but
there is evidence that â-alkoxy-â-phosphatoxyalkyl radicals such
as 24 will be more stable than R-alkoxy-â-phosphatoxyalkyl
radicals. For example, computed energies for radicals 26 and
27 at the B3LYP/6-31G* level indicate that 27 is more stable
than 26 by about 3 kcal/mol due to an anomeric effect that
stabilizes radical 27.30 If radical 24 does not heterolyze rapidly,
One surprising kinetic result is that radical cations 1b and
2b have comparable rate constants for cyclization, unlike the
(22) Choi, S. Y.; Crich, D.; Horner, J. H.; Huang, X. H.; Newcomb, M.; Whitted,
P. O. Tetrahedron 1999, 55, 3317-3326.
(23) Whitted, P. O.; Horner, J. H.; Newcomb, M.; Huang, X. H.; Crich, D.
Org. Lett. 1999, 1, 153-156.
(24) Newcomb, M.; Horner, J. H.; Whitted, P. O.; Crich, D.; Huang, X. H.;
Yao, Q. W.; Zipse, H. J. Am. Chem. Soc. 1999, 121, 10685-10694.
(25) Simulated convoluted data for consecutive reactions, both with rate constants
of 3 × 107 s-1, give an apparent rate constant of 1.8 × 107 s-1 when solved
as a first-order process.
(26) Newcomb, M.; Horner, J. H.; Filipkowski, M. A.; Ha, C.; Park, S. U. J.
Am. Chem. Soc. 1995, 117, 3674-3684.
(27) Newcomb, M.; Choi, S. Y.; Horner, J. H. J. Org. Chem. 1999, 64, 1225-
1231.
(29) Musa, O. M.; Choi, S. Y.; Horner, J. H.; Newcomb, M. J. Org. Chem.
1998, 63, 786-793.
(30) Horner, J. H. Unpublished results.
(28) Beckwith, A. L. J.; Moad, G. J. Chem. Soc., Chem. Commun. 1974, 472-
473.
9
J. AM. CHEM. SOC. VOL. 124, NO. 19, 2002 5407

A R T I C L E S
Horner et al.
Scheme 6
Table 4. Normalized Relative Rate Constants for Reactions
Occurring in Ion Pairs in ACNa
radical
cation
c
phosphate
cyclization
deprotonationb
escape
D/E
1b
1b
2b
1a
1a
2a
(EtO)2P(O)O-
(PhO)2P(O)O-
(EtO)2P(O)O-
(EtO)2P(O)O-
(PhO)2P(O)O-
(EtO)2P(O)O-
86
66
75
46
15
35
14
34
25
8
8
12
6.1
2.0
3.0
46
77
53
a See discussion in text for the method. b Deprotonation to give allyl
radical 25 and/or collapse to radical 24. c Ratio of deprotonation and/or
collapse in the ion pair to diffusive escape.
We thus envision two solvent polarity effects at play in the
reactions we studied. Increasing solvent polarity increased the
rate of heterolysis of the â-phosphate radicals as indicated in
the strong polarity effect observed in the kinetics of reactions
of 5a-E and 8a. In addition, increased solvent polarity resulted
in more efficient formation of free ions, most likely by
accelerating the ion pair escape reaction.
Another ion pair reaction is obviously important for the 5-exo
cyclizations. The yields of products from these cyclizations in
ACN are much greater than the yields of products from the
corresponding 6-exo cyclizations. If our model of diffusive
escape competing with reactions within the contact ion pair is
correct, then the 5-exo cyclizations of 1a and 2a must compete
efficiently with other reactions of the radical cation in the ion
pair, thus resulting in increased yields of detectable diphenyl-
alkyl radicals. That is, the 5-exo cyclizations must be occurring
partially within the contact ion pair as shown in Scheme 6. We
note that either deprotonation or nucleophilic capture of the
distonic radical cation product following the cyclization in the
ion pair would give products that still contain the diphenylalkyl
radical moiety and will be indistinguishable by UV spectroscopy
from the distonic radical cation product formed by cyclization
of the diffusively free radical cation.
its formation will serve as a trapping reaction that leads to lower
yields of detectable products because cyclization of 24 would
be much too slow to be important on the time scale of the kinetic
measurements.27,28
Alternatively, deprotonation of the radical cation by the
phosphate anion within the ion pair is possible. This reaction
would give allylic radical 25 and phosphoric acid, and the
process will definitely lead to a reduction in yield of detected
diphenylalkyl radical products. A 4-exo cyclization of radical
25 produced from the “a” series would be too slow to measure
by LFP methods, and a 5-exo cyclization of radical 25 produced
from the “b” series will be much slower than a conventional
5-exo radical cyclization, which has a rate constant of about 2
× 105 s-1 at 20 °C for the 5-hexenyl radical.
The relative yields for the two pairs of â-diethyl phosphate
radicals (5a-E vs 5b-E, and 8a vs 8b) and that for the
â-diphenylphospate radicals (5a-P vs 5b-P) can be used to
establish the relative rate constants for processes occurring in
the ion pair in ACN (Table 4). From the yield data for the 6-exo
cyclizations, we established the relative rate constants for
deprotonation and/or collapse to radical 24 versus diffusive
escape from the ion pairs. The ratio of these rate constants (D/
E) was then used to calculate the approximate amount of 5-exo
cyclization product formed from diffusively free radical cations
1a and 2a. The remaining yields of cyclic product for radical
cations 1a and 2a are ascribed to 5-exo cyclizations occurring
in the ion pair. This crude analysis indicates that (1) deproton-
ation by the diethyl phosphate anion and/or collapse to radical
24 is about three times as fast as deprotonation by and/or
collapse of the diphenyl phosphate anion, (2) the relative amount
of diffusive escape was consistently small in ACN, and (3) the
5-exo cyclizations of radical cations 1a and 2a had similar rate
constants as expected from the results of 6-exo cyclizations of
1b and 2b where cyclization rate constants were measured.
We can add a semiquantitative aspect to the analysis in Table
4 by assuming that 5-exo cyclizations of the radical cations 1a
and 2a are about 50 times faster than 6-exo cyclizations of
radical cations 2a and 2b; this difference in rate constants for
5-exo and 6-exo cyclizations has been found in a number of
From the data in this work, it is not possible to identify which
of the above processes is more prevalent in reducing the yields
of distonic radical cation products, but previous results suggest
that both reactions are likely to occur. For example, phosphate
dianion in water traps enol ether radical cations to give â-alkoxy-
â-phosphate radicals, and R-alkoxy-â-phosphate radicals are not
detected;7 this could be a kinetic effect, but it is possible that
the latter were formed reversibly, and the former were more
stable. In addition, the yields of radical cations formed from
heterolyses of â-phosphate radicals 19 are lower in ACN than
in ACN containing TFE, much like the solvent effect on yield
seen here, but deprotonation is not possible for this system,
suggesting that collapse to a less reactive â-alkoxy-â-phosphate
radical occurred.9 Deprotonation reactions of enol ether radical
cations by dialkyl phosphate anions are known to occur,
however.31 The fact that the yields are more severely reduced
in reactions of â-diethyl phosphate radicals than in reactions of
â-diphenyl phosphate radicals appears to be more consistent
with a deprotonation reaction because diethyl phosphate is a
stronger base,18 although diethyl phosphate might also react
more rapidly in the nucleophilic addition reaction that gives
radical 24.
(31) Giese, B.; Beyrich-Graf, X.; Erdmann, P.; Petretta, M.; Schwitter, U. Chem.
Biol. 1995, 2, 367-375.
9
5408 J. AM. CHEM. SOC. VOL. 124, NO. 19, 2002

Enol Ether Radical Cations
A R T I C L E S
LFP kinetic studies of radical cyclizations.26,27,29,32-34 If that
approximation holds for the enol ether radical cations, then the
5-exo cyclizations and deprotonation by the diethyl phosphate
anion and/or ion pair collapse have rate constants at ambient
temperature in the range 1-2 × 109 s-1. The deprotonation by
the diphenyl phosphate anion and/or ion pair collapse and
diffusive escape from the ion pair have rate constants that are
Scheme 7
smaller, in the range 2-4 × 108 s-1
.
We note that the rate constants estimated above are for
reactions in ACN. The rate constants for the 5-exo radical cation
cyclizations apparently will be changed little in solvents of
different polarity, but the rate constants for all of the other
reactions are expected to vary as the solvent polarity changes.
Specifically, one expects that an increase in solvent polarity
will result in an increase in the rate constant for diffusive escape
and a reduction in the relative rate constants for deprotonation
or nucleophilic capture of the radical cation.
Fast cyclization of radical cation 1a within the ion pair is
almost certainly the reason that distonic radical cation product
6a was detected in 15% yield when radical 5a-P was allowed
to react in THF. That is, ion pair escape in this low polarity
solvent is not expected to be efficient. For example, we
previously observed that a highly stabilized p-methoxystyrene
radical cation produced by a â-phosphate heterolysis reaction
did not give a diffusively free species in THF, although it did
in polar solvents including ACN.23 From the results for 1a
produced from 5a-P in ACN, we estimated that the cyclization
was about 5 times as fast as deprotonation or collapse to radical
24, but in THF the cyclization reaction apparently is only 1/6
as fast as deprotonation or collapse to 24.
heterolysis step that gives a contact ion pair, possibly equilibra-
tion of a contact ion pair with a solvent-separated ion pair, and
diffusive escape from the ion pair (Scheme 7). Different
experimental methods might measure different steps in the
heterolysis reaction, and that appears to be the case for various
studies of â-phosphate radicals.
In our previously reported intermolecular approach for
measuring heterolysis of radicals 19, the technique provided
rate constants for formation of diffusively free radical cations.9
If equilibrations of the radical-cation-anion pairs with the
â-phosphate radicals occur in the systems studied here, then
the rate constants for heterolysis of radicals 5a-E and 8a will
be greater than those found when diffusively free radical cations
are formed because radical cations 1a and 2a cyclize partially
within the ion pair. This equilibration effect might explain why
the rate constants for heterolysis of 5a-E and 8a are apparently
orders of magnitude greater than those for heterolysis of radical
19-E.
A different situation was encountered when Gugger et al.
studied the rate of heterolysis of radical 20.19 In that work, the
method involved competition between trapping 20 by PhSH and
formation of the radical cation that eventually reacted with the
solvent, methanol.19 The polarity of MeOH as evaluated on the
ET(30) solvent polarity scale (55.4)35 is greater than that for
ACN (45.6)35 and for solutions of 0.5% TFE in ACN (deter-
mined here as ET(30) ) 52). Therefore, the heterolysis reactions
of 5a-E and 8a in MeOH should be somewhat faster than we
can measure with our technique. Nonetheless, the rate constant
for heterolysis determined for 20 (k > 3 × 109 s-1) appears to
be almost an order of magnitude faster than what might be
expected for 5a-E and 8a in the same solvent if a linear
correlation between log k and ET(30) exists as found in other
radical heterolyses.23 It is interesting to speculate that the
apparent fast heterolysis of radical 20 might involve a favorable
stereoelectronic effect that is enforced by the structure of the
radical such that the entropy of activation for the heterolysis is
considerably reduced from that observed with 5a-E and 8a.
If the rate constant for the cyclization has little or no solvent
polarity effect in THF as is the case for the reactions in ACN
and ACN containing TFE, the 30-fold change in relative rate
constants upon proceeding from ACN to THF reflects a strong
acceleration of the deprotonation and/or ion pair collapse
reactions as the solvent polarity was decreased, consistent with
the generalizations made above. That is, poorer solvation
enhanced nucleophilicity and/or basicity of the phosphate.
Extending the crude kinetic approximations made for reactions
in ACN to the reaction in THF, we estimate that the ion pair
collapse and/or deprotonation reaction has a rate constant at
ambient temperature of ca. 3-6 × 1010 s-1
.
Thus, in ACN the ion pair has a lifetime on the order of 1
ns, but in the less polar solvent THF, the ion pair lifetime is
reduced to ca. 20 ps. These values are important for mechanistic
understanding and synthetic applications of radical cation
reactions that employ radical heterolyses as the entries. Many
synthetic applications of radical reactions involve the use of
tin hydride and are conducted in low polarity solvents such as
benzene or toluene. In these solvents, one should not expect
that diffusively free radical cations would be produced from
â-phosphate radicals. Conversely, the production of diffusively
free radical cations from heterolyses of C4′ DNA radicals and
their models in water is expected to be quite efficient.
An Enol Ether Radical Cation Kinetic Scale. As we
demonstrated with water and MeOH trapping, the reporter LFP
method can be used for direct determinations of rate constants
for reactions of enol ether radical cation 1c. These types of LFP
studies can be performed with a wide range to nucleophiles,
bases, and reductants to establish a kinetic scale for enol ether
radical cations that would be useful for mechanistic understand-
ing and in the design of synthetic applications. We caution,
however, that radical cation 1c is not an optimal system for
these kinetic studies. The inherent cyclization reaction is slow
enough such that bimolecular reactions leading to decay
complicate the measurements. The 6-exo cyclizations of 1b and
Measured Kinetics of Heterolysis Reactions. The “het-
erolysis reaction” is composed of multiple processes, the
(32) Johnson, C. C.; Horner, J. H.; Tronche, C.; Newcomb, M. J. Am. Chem.
Soc. 1995, 117, 1684-1687.
(33) Horner, J. H.; Martinez, F. N.; Musa, O. M.; Newcomb, M.; Shahin, H. E.
J. Am. Chem. Soc. 1995, 117, 11124-11133.
(34) Musa, O. M.; Horner, J. H.; Shahin, H.; Newcomb, M. J. Am. Chem. Soc.
1996, 118, 3862-3868.
(35) Reichardt, C. Chem. ReV. 1994, 94, 2319-2358.
9
J. AM. CHEM. SOC. VOL. 124, NO. 19, 2002 5409

A R T I C L E S
Horner et al.
2b, with rate constants of about 3 × 107 s-1 at ambient
temperatures, are not useful for these types of measurements
because they are too fast. An optimal unimolecular reaction for
LFP studies probably would have a rate constant of about 1 ×
106 s-1 at ambient temperature.
Laser Flash Photolysis. Acetonitrile (HPLC grade) used in LFP
studies was purchased from Fisher and used as received. 2,2,2-
Triflouroethanol was stirred over CaSO4 and NaHCO3, filtered, and
distilled. LFP studies were performed with an Applied Photophysics
LKS-50 kinetic spectrometer using a Nd:YAG laser at 355 nm.
Solutions of the PTOC precursor having an absorbance of 0.2 to 0.5 at
355 nm were placed in a thermally jacketed addition funnel and sparged
with helium. The temperature of the solution was adjusted by circulation
of a water/ethylene glycol 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 measurements below
10 °C, the addition funnel and flow cell were placed in a nitrogen-
purged box fitted with quartz windows.
Conclusion
The unimolecular reporter method for studies of enol ether
radical cations and their progenitor radicals developed here
provides both kinetic and mechanistic insights. The reactions
are complex, and one observes strong solvent polarity effects
in the formation of diffusively free radical cations from radical
heterolyses. The radical reactions of the radical cations, such
as the cyclizations we studied, are highly accelerated, whereas
the cation reactions, such as trapping with water, are greatly
retarded. These characteristics can be employed advantageously
in synthesis because a radical-type reaction of the radical cation
gives either a distonic radical cation or a radical plus a cation
product, and the reactivities of the products will be diminished
for the radical and enhanced for the cation. Thus, it is likely
that cascade synthetic conversions involving enol ether radical
cations will be useful. We expect that applications of these
intermediates will increase when kinetic scales are available and
mechanistic understanding is improved.
Kinetic Data Analyses. For kinetic runs, several traces typically
were averaged to improve signal-to-noise. Fitting the experimental data
to a first-order exponential growth using nonlinear least-squares methods
was performed with Applied Photophysics software. The analysis of
consecutive reactions (radical 8b in ACN) was performed with
Sigmaplot.36
Acknowledgment. We are grateful to the NSF (CHE-
0296027) for financial support.
Supporting Information Available: Synthetic details and
tables of kinetic data (PDF). This material is available free of
charge via the Internet at http://pubs.acs.org.
Experimental Section
Syntheses. The synthetic work is described in the Supporting
Information. Intermediate products typically were characterized by
NMR spectroscopy and HRMS. The PTOC esters are thermally unstable
and were characterized by NMR spectroscopy and, in some cases,
electrospray HRMS.
JA0177399
(36) Sigmaplot, ver. 2.0; Jandel Scientific Software: San Rafael, CA, 1994.
9
5410 J. AM. CHEM. SOC. VOL. 124, NO. 19, 2002