Dynamics of ionization reactions of β-substituted radicals, substituent and solvent effects

Article abstract of DOI:10.1021/ja971996p

The dynamics of reactions of carbon-centered 1-arylalkyl radicals with bromine or chlorine attached to the carbon adjacent (β) to the radical center have been examined using nanosecond laser flash photolysis. The primary reaction of the radicals containing the electron-donating 4-methoxy group on the phenyl ring is highly dependent on the solvent composition. In weakly ionizing solvents such as acetonitrile, the radicals decay in a second-order manner indicating that coupling of two radical centers is the primary mode of radical decay. However, when the ionizing ability of the solvent is increased by addition of water, 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP), or 2,2,2-trifluoroethanol (TFE), heterolysis of the β-substituent becomes the dominant mode of decay. The occurrence of the heterolysis reaction is demonstrated unambiguously by direct observation of the radical cation produced as the primary, heterolysis product. The rate constants for and yield of the heterolysis reaction are found to be dependent on both the solvent ionizing ability and radical structure. In neat water or neat HFIP the reactions become extremely fast and occur with rate constants in the 10⁷ s^-1 to ≤ 10⁸ s^-1 range. For the β-bromophenethyl and β-bromo-4-methyphenethyl radicals, no heterolysis is observed even under strongly ionizing conditions, indicating that the rate constant for ionization is strongly influenced by the substituent on the phenyl ring. For radicals with an additional β-phenyl substituent, rapid heterolysis takes place leading to the formation of the stilbene radical cation. The formation of a radical/radical cation equilibrium was observed under the appropriate conditions only for the 4-methoxyphenethyl radical derivatives.

Full text of DOI:10.1021/ja971996p

10652
J. Am. Chem. Soc. 1997, 119, 10652-10659
Dynamics of Ionization Reactions of â-Substituted Radicals.
Substituent and Solvent Effects
Frances L. Cozens,* Melanie O’Neill, Roumiana Bogdanova, and Norman Schepp
Contribution from the Department of Chemistry, Dalhousie UniVersity, Halifax,
NoVa Scotia, Canada, B3H 4J3
ReceiVed June 17, 1997X
Abstract: The dynamics of reactions of carbon-centered 1-arylalkyl radicals with bromine or chlorine attached to
the carbon adjacent (â) to the radical center have been examined using nanosecond laser flash photolysis. The
primary reaction of the radicals containing the electron-donating 4-methoxy group on the phenyl ring is highly
dependent on the solvent composition. In weakly ionizing solvents such as acetonitrile, the radicals decay in a
second-order manner indicating that coupling of two radical centers is the primary mode of radical decay. However,
when the ionizing ability of the solvent is increased by addition of water, 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP),
or 2,2,2-trifluoroethanol (TFE), heterolysis of the â-substituent becomes the dominant mode of decay. The occurrence
of the heterolysis reaction is demonstrated unambiguously by direct observation of the radical cation produced as
the primary, heterolysis product. The rate constants for and yield of the heterolysis reaction are found to be dependent
on both the solvent ionizing ability and radical structure. In neat water or neat HFIP the reactions become extremely
fast and occur with rate constants in the 107 s-1 to g 108 s-1 range. For the â-bromophenethyl and â-bromo-4-
methyphenethyl radicals, no heterolysis is observed even under strongly ionizing conditions, indicating that the rate
constant for ionization is strongly influenced by the substituent on the phenyl ring. For radicals with an additional
â-phenyl substituent, rapid heterolysis takes place leading to the formation of the stilbene radical cation. The formation
of a radical/radical cation equilibrium was observed under the appropriate conditions only for the 4-methoxyphenethyl
radical derivatives.
Introduction
leaving group to generate a more stable radical12-14 are two
well-known reactions of â-substituted radicals. In addition,
these radicals undergo solvolysis reactions involving either acid-
catalyzed or uncatalyzed heterolysis of the carbon-leaving group
bond to generate a radical cation that is then trapped by solvent
(or another nucleophile).15 These heterolysis reactions are
thought to be important for radicals generated enzymatically
and in the bond cleavage step of radical-induced DNA strand
cleavage. As a result, significant effort has been directed toward
studying the reactions of these simple radicals. ESR15 and
product studies4,7 have been especially useful to establish the
existence of the solvolysis reaction by detecting the presence
of substitution products. Results from pulse radiolysis studies
have provided some information about the dynamics of the
reactions of these radicals, especially with respect to second-
order rate constants for the acid-catalyzed dehydration of a
variety of 2-hydroxy-1-alkoxy alkyl radicals15 as well as rate
constants for the acid-catalyzed reactions of 1-alkoxy radicals
with phosphate groups at the 2-position.16
Carbon-centered radicals with leaving groups such as OH,
OR, NH2, or halogens attached to the carbon adjacent (â) to
the radical center play an important role in a wide variety of
biological and chemical processes. For example, radicals with
a hydroxy, ammonium, or phosphate group at the â-position
are thought to be primary intermediates in some enzyme
catalyzed reactions, such as the reduction of diols and ethano-
lamines by diol dehydrase and ethanolamine deaminase, re-
spectively, and the ribonucleotide reductase catalyzed conversion
of ribonucleotides to deoxyribonucleotides.1-4 In addition, 4′-
deoxyribonucleotide radicals produced by hydrogen atom
abstraction from the R-carbon of a sugar moiety of DNA have
been postulated to be important intermediates in radical induced
cleavage of DNA strands.5-8 Furthermore, radicals possessing
a â-acetate or phosphate group have been implicated as
important intermediates in oxidative lipid damage.9
Homolytic cleavage of the carbon-leaving group bond to
generate an alkene10,11 and rearrangement by a 1,2 shift of the
Much less is known about the dynamics of the uncatalyzed
heterolysis of the carbon-leaving group bond in â-substituted
radicals. A limited number of rate constants for the heterolysis
of alkyl radicals with leaving groups such as halides, sulfate,
and acetate in the â-position have been measured in wholly
X Abstract published in AdVance ACS Abstracts, October 1, 1997.
(1) Frey, P. A. Chem. ReV. 1990, 90, 1343-1357.
(2) Stubbe, J. Biochemistry 1988, 27, 3893-3900.
(3) Stubbe, J. J. Biol. Chem. 1990, 265, 5329-5332.
(4) Lenz, R.; Giese, B. J. Am. Chem. Soc. 1997, 119, 2784-2794.
(5) Behrens, G.; Koltzenburg, G.; Schulte-Frohlinde, D. Z. Naturforsch.
1982, 37c, 1205-1227.
(6) Koltzenburg, G.; Behrens, G.; Schulte-Frohlinde, D. J. Am. Chem.
Soc. 1982, 104, 7311-7312.
(11) Scaiano, J. C.; Barra, M.; Krzywinski, M.; Sinta, R.; Calabrese, G.
J. Am. Chem. Soc. 1993, 115, 8340-8344.
(12) Tanner, D. D.; Law, F. C. P. J. Am. Chem. Soc. 1969, 91, 7535-
7537.
(13) Barclay, L. R. C.; Lusztyk, J.; Ingold, K. U. J. Am. Chem. Soc.
1984, 106, 1793-1796.
(7) Giese, B.; Beyrich-Graf, X.; Burger, J.; Kesselheim, C.; Senn, M.;
Schaefer, T. Angew. Chem., Int. Ed. Engl. 1993, 32, 1742-1743.
(8) Schulte-Frohlinde, D.; von Sonntag, C. Sugar Lesions in Cell
Radiobiology; Wiley-Liss: New York, 1990; pp 31-42.
(9) Mu¨ller, S. N.; Batra, R.; Senn, M.; Giese, B.; Kisel, M.; Shadyro, O.
J. Am. Chem. Soc. 1997, 119, 2795-2803.
(14) Beckwith, A. L. J.; Duggan, P. J. J. Am. Chem. Soc. 1996, 118,
12838-12839.
(15) Davies, M. J.; Gilbert, B. C. In AdVances in Detailed Reaction
Mechanisms; Coxon, J. M., Ed.; JAI Press: London, 1991; Vol. 1, pp 35-
82.
(10) Scaiano, J. C.; Barra, M.; Calabrese, G.; Sinta, R. J. Chem. Soc.,
Chem. Comm. 1992, 1418-1419.
(16) Behrens, G.; Koltzenburg, G.; Ritter, A.; Schulte-Frohlinde, D. Int.
J. Radiat. Biol. 1978, 33, 163-171.
S0002-7863(97)01996-3 CCC: $14.00 © 1997 American Chemical Society

Dynamics of Ionization Reactions of â-Substituted Radicals
J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10653
Scheme 1
Figure 1. Transient absorption spectra obtained immediately after 266-
nm laser irradiation of 1-acetoxy-2-chloro-1-(4-methoxyphenyl)ethane
1a in (b) nitrogen-saturated acetonitrile and (O) oxygen-saturated
acetonitrile.
aqueous solution.6,17 However, in those studies, simple het-
erolysis is not necessarily the only reaction available to the
radical,5 and the measured rate constants may include significant
contributions from other competing processes such as solvolysis
by an SN218 mechanism or rapid homolysis of the â-group.11
One way to identify heterolysis as an important mode of reaction
for these radicals is to provide direct evidence that the decay
of the radicals leads to the formation of a radical cation as the
primary product.19-21 In recent years, substituted styrene and
stilbene radical cations have been shown to be readily detectable
in solution using fast reaction techniques such as laser flash
photolysis.22-26 The relatively long lifetime of these radical
cations opens up the possibility that the ionization of â-substi-
tuted radicals can be studied directly by monitoring the
formation of the substituted styrene radical cations under
appropriate conditions.
Figure 2. Transient absorption spectra obtained after 266-nm laser
irradiation of 1-acetoxy-2-chloro-1-(4-methoxyphenyl)ethane 1a in
nitrogen-saturated 83% TFE/17% AcN. Spectra were recorded ([) 0.3
µs, (O) 1.2 µs, and (b) 4 µs after the laser pulse. The inset shows the
time-resolved decay at (b) 305 nm and growth at (O) 360 nm and (0)
600 nm.
ability of the solvent, and the stability of the initially formed
cationic intermediate.
Results
In the present work, we describe results concerning the
reactions of 2-halo-1-arylalkyl radicals 2a-f and the 2-halo-
1,2-diphenylethyl radicals 4a,b, generated by laser photolysis
of the precursors 1a-f and 3a,b in Scheme 1. Our results show
that heterolysis of the â-group to give styrene or stilbene radical
cations does take place in solution and that this reaction
dominates other processes in ionizing solvents such as water,
2,2,2-trifluoroethanol (TFE), and 1,1,1,3,3,3-hexafluoro-2-pro-
panol (HFIP). The heterolysis reactions are often very fast
proceeding with rate constants exceeding 108 s-1 but are
otherwise influenced in the same way as other SN1 type reactions
with respect to the nature of the leaving group, the ionizing
Heterolysis in Acetonitrile Mixtures. As shown in Figure
1, 266-nm laser photolysis of 1-acetoxy-2-chloro-1-(4-methox-
yphenyl)ethane 1a in nitrogen-saturated acetonitrile (AcN) gives
a transient species with an absorption maximum near 300 nm.
The transient decays in a second-order manner and is long-
lived under nitrogen-saturated conditions but is completely
quenched by the addition of oxygen, Figure 1. This behavior
and the location of the absorption maximum near 300 nm27 are
consistent with identification of the transient as the â-chloro-
4-methoxyphenethyl radical 2a produced by photohomolysis,
eq 1.
(17) Koltzenburg, G.; Bastian, E.; Steenken, S. Angew. Chem., Int. Ed.
Engl. 1988, 27, 1066-1067.
(18) Zipse, H. Angew. Chem., Int. Ed. Engl. 1994, 33, 1985-1988.
(19) Photocurrent20 and CIDNP21 techniques have recently been used to
show that model compounds of 4′-DNA undergo heterolysis.
(20) Giese, B.; Erdmann, P.; Giraud, L.; Gobel, T.; Petretta, M.; Schaefer,
T. Tetrahedron Lett. 1994, 35, 2683-2868.
(21) Gugger, A.; Batra, R.; Rzadek, P.; Rist, G.; Giese, B. J. Am. Chem.
Soc. Accepted for publication.
In a nitrogen-saturated TFE/AcN mixture containing 83% (by
volume) TFE, laser photolysis of the same precursor also gives
the â-chloro radical 2a immediately after the laser pulse, Figure
2. However, in this solvent, the radical decays much more
rapidly than in acetonitrile with a first-order rate constant of
2.3 × 106 s-1. In addition, as the radical decays, a second
transient species with absorption maxima at 360 and 600 nm is
(22) Brede, O.; David, F.; Steenken, S. J. Chem. Soc., Perkin Trans. 2
1995, 23-32.
(23) Johnston, L. J.; Schepp, N. P. J. Am. Chem. Soc. 1993, 115, 6564-
6571.
(24) Johnston, L. J.; Schepp, N. P. In AdVances in Electron Transfer
Chemistry; Mariano, P. S., Ed.; JAI Press: Greenwich, CT, 1996; Vol. 5,
pp 41-102.
(25) Ebbesen, T. J. Phys. Chem. 1988, 92, 4581-4583.
(26) Lewis, F. D.; Dykstra, R. E.; Gould, I. R.; Farid, S. J. Phys. Chem.
1988, 92, 7042-7043.
(27) Tokumura, K.; Ozaki, T.; Nosaka, H.; Saigusa, Y.; Itoh, M. J. Am.
Chem. Soc. 1991, 113, 4974-4980.

10654 J. Am. Chem. Soc., Vol. 119, No. 44, 1997
Cozens et al.
Figure 3. Time resolved growth at 600 nm observed upon 266-nm
laser irradiation of 1-acetoxy-2-chloro-1-(4-methoxyphenyl)ethane 1a
in nitrogen-saturated TFE/AcN mixtures containing (9) 33%, (0) 50%,
([) 67%, (o) 83%, and (b) 100% TFE.
concurrently formed, Figure 2. The absorption spectrum of this
transient species is identical to that obtained previously for the
radical cation of 4-methoxystyrene,23 and we can therefore
identify the transient produced in 83% TFE as the 4-methoxy-
styrene radical cation 5a, eq 2.
As shown in the inset of Figure 2, the rate constant for the
decay of the â-chloro radical at 305 nm closely matches the
rate constant for the growth of the radical cation at 360 and
600 nm, indicating that the radical cation is produced from a
reaction of the photogenerated radical. Moreover, formation
of the radical cation is completely quenched in oxygen-saturated
83% TFE. Since the reactivity of the radical cation is known
to be insensitive to oxygen concentration,23 the disappearance
of the radical cation can be attributed to trapping of the radical
by oxygen. These results therefore provide strong evidence that
the radical cation is produced upon reaction of the radical,
namely by heterolysis of the carbon-chlorine bond adjacent to
the radical center, eq 2.
Figure 4. Growth of radical cation at 600 nm upon 266-nm irradiation
of (a) 1-acetoxy-2-chloro-1-(4-methoxyphenyl)ethane 1a; (b) 1-acetoxy-
2-chloro-1-(4-methoxyphenyl)propane 1b; (c) 1-acetoxy-2-bromo-1-
(4-methoxyphenyl)ethane 1c; (d) 1-acetoxy-2-bromo-1-(4-methoxyphe-
nyl)propane 1d; and (e) the radical cation at 475 nm upon 266-nm
irradiation of 1,2-dichloro-1,2-diphenylethane 3a as a function of (b)
water, (O) HFIP, or ([) TFE content in nitrogen-saturated acetonitrile.
after complete formation of the radical cation. Thus, in 100%
TFE, the optical density at 600 nm after the radical cation is
fully formed is approximately 15 times greater than in 33% TFE.
Ionization of the chloro group from the â-chloro-4-methox-
yphenethyl radical 1a was also examined in HFIP/AcN mixtures
and in water/AcN mixtures. In HFIP/AcN mixtures containing
from 33% HFIP to 83% HFIP, time-resolved growth of the
radical cation was observed, with the rate constant increasing
from 3.5 × 105 s-1 in 33% HFIP to 2.3 × 107 s-1 in 83% HFIP.
In 100% HFIP, the rate constant for radical cation formation
exceeded the 1 × 108 s-1 time-resolution of our laser system.
Similar results were obtained for reaction in water/AcN
mixtures. In this solvent system, the rate constant for radical
cation growth increased from 1.3 × 106 s-1 in 33% water to
3.0 × 107 s-1 in 83% water. In 100% aqueous solution, the
rate constant for the growth was again beyond the resolution
of the laser system. These data are summarized in Figure 4a.
The fact that formation of the heterolysis product is rapid in
83% TFE but is not observed in neat acetonitrile suggests that
the rate constant for radical cation formation is strongly
influenced by the nature of the solvent. The effect of solvent
is clearly illustrated by the data in Figure 3 which shows the
time-dependent change in optical density at 600 nm due to
radical cation formation upon â-heterolysis of the radical 1a in
various TFE/AcN mixtures. In the mixed TFE/AcN solvent
containing 33% TFE, a small amount of radical cation growth
is observed with a first-order rate constant of approximately
5.1 × 105 s-1. As the amount of TFE is increased to 50%, a
more substantial growth is observed with a larger rate constant
of 6.4 × 105 s-1. As the TFE content is then increased to 66%
and 83%, the rate constant continues to increase to 1.1 × 106
s-1 and 2.3 × 106 s-1, and reaches a value of k ) 1.1 × 107
s-1 in neat TFE. The growth curves in Figure 3 also show that
the increase in the rate constant for radical cation formation as
a function of TFE content in acetonitrile is accompanied by a
significant increase in the maximum optical density at 600 nm
Results similar to those described above were observed upon
266-nm laser flash photolysis of 1-acetoxy-2-chloro-1-(4-
methoxyphenyl)propane 1b. In nitrogen-saturated acetonitrile,
only the â-chloro-4-methoxyphenylpropyl radical 2b, eq 2, with
an absorption maximum at 310 nm and a slow, second-order
decay was observed. Upon the addition of ionizing solvents
like water, HFIP or TFE to the acetonitrile, the radical decay
became more rapid and was accompanied by the formation of
the radical cation of anethole 5b with absorption maxima at
380 and 600 nm, indicating that radical 2b also undergoes
â-heterolysis of chloride ion under ionizing conditions, eq 2.
As for the reaction of radical 2a described above, the rate
constant for the formation of the anethole radical cation from
radical 2b was measured in water/AcN, HFIP/AcN, and TFE/
AcN mixtures. In each case, the observed rate constant
increased with increasing water, HFIP or TFE content, and
eventually became too fast to measure. These results are
summarized in Figure 4b.

Dynamics of Ionization Reactions of â-Substituted Radicals
J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10655
Figure 7. Absorption spectra immediately after 266-nm laser irradiation
of 1-acetoxy-2-bromo-1-(4-methylphenyl)ethane 1e in (b) nitrogen-
saturated acetonitrile and (O) oxygen-saturated acetonitrile.
Figure 5. Absorption spectra immediately after 266-nm laser irradiation
of 1-acetoxy-2-bromo-1-(4-methoxyphenyl)ethane 1c in (b) nitrogen-
saturated acetonitrile and (O) oxygen-saturated acetonitrile.
Figure 8. Absorption spectra after 266-nm laser irradiation of 1,2-
dichloro-1,2-diphenylethane 3a in 33% TFE/67% AcN. Spectra were
recorded ([) 20 ns, (O) 50 ns, (b) 180 ns, and (0) 400 ns after the
laser pulse.
Figure 6. Absorption spectra after 266-nm laser irradiation of
1-acetoxy-2-bromo-1-(4-methoxyphenyl)ethane 1c in nitrogen-saturated
50% water/50% AcN. Spectra were recorded ([) 0.16 µs, (O) 0.63 µs
and (b) 1.25 µs after the laser pulse.
measured in the same solvents as described above for the
â-chloro radicals 2a and 2b. The results are summarized in
Figure 4c,d.
Laser irradiation of the bromo derivatives, 1-acetoxy-2-
bromo-(4-methoxyphenyl)ethane 1c or 1-acetoxy-2-bromo-(4-
methoxyphenyl)propane 1d, in nitrogen-saturated acetonitrile
led to the formation of the â-bromo radicals 2c or 2d, eq 3,
with an absorption maximum at 350 nm, Figure 5. While the
position of the maximum is unusual for benzylic radicals, it
matches closely that obtained for the â-bromo-4-methoxyphe-
nylpropyl radical previously generated.23 In dry, nitrogen-
â-Bromo radicals 2e and 2f with absorption at 330 nm, Figure
7, are produced in nitrogen-saturated acetonitrile by laser
irradiation of 1e and 1f. However, unlike the 4-methoxy
derivatives discussed above, no evidence for the formation of
the radical cation of styrene or 4-methylstyrene that would be
produced by heterolysis of the â-bromo group was observed as
the radicals decayed in TFE or HFIP, nor did the lifetime of
the radicals change with solvent ionizing ability. It has
previously been shown that the radical cations of 4-methylsty-
rene and styrene are sufficiently long-lived in these solvents to
be observed.23 The fact that the radical cations were not
detected, together with the observation that the lifetimes of the
radicals remained essentially unchanged upon going from
acetonitrile to HFIP, indicates that the heterolysis of these
radicals is too slow to compete with other reactions of the
radical.
A different result is obtained for the â-substituted 1,2-
diphenylethyl radicals 4a and 4b, which differ from the
unreactive radical 2f by the presence of a phenyl group instead
of a hydrogen at the carbon undergoing ionization. The time-
dependent absorption spectrum generated upon 266 nm irradia-
tion of 1,2-dichloro-1,2-diphenylethane 3a in 33% TFE/AcN
is shown in Figure 8. After the laser pulse, a strong absorbance
band at 475 nm and a weaker band at 700 nm that grow in a
first-order manner with equal rate constants of 1.3 × 107 s-1
are clearly observed. These absorption bands can be confidently
assigned to the trans-stilbene radical cation 6 on the basis of
its known absorption spectrum.28 The growth of the stilbene
saturated acetonitrile, there was no evidence that radicals 2c or
2d undergo heterolysis to give the 4-methoxystyrene radical
cation 5a or the anethole radical cation 5b. However, in TFE,
HFIP, or water mixed with acetonitrile, loss of bromide clearly
takes place and leads to the formation of the radical cations, eq
3, with absorption maxima near 350 and 600 nm. The growth
of the 4-methoxystyrene radical cation by heterolysis of the
â-bromo radical 2c in 50% aqueous acetonitrile is shown in
Figure 6. In this case, it is difficult to match the decay of the
radical with the growth of the radical cation, since the major
absorption band of the radical is almost at the same position as
the 350 nm band of the radical cation. Observed rate constants
for radical cation formation from radicals 2c and 2d were
(28) Spada, L. T.; Foote, C. S. J. Am. Chem. Soc. 1980, 102, 391-397.

10656 J. Am. Chem. Soc., Vol. 119, No. 44, 1997
Cozens et al.
radical cation is completely quenched upon the addition of
oxygen, indicating that the radical cation is produced by
heterolysis of the â-chloro radical generated by laser irradiation
of the 1,2-dichloro-1,2-diphenylethane 3a, eq 4.
Figure 9. Relationship between the observed rate constants for the
formation of radical cations at 600 nm upon 266-nm irradiation of (b)
1-acetoxy-2-chloro-1-(4-methoxyphenyl)ethane 1a; (9) 1-acetoxy-2-
bromo-1-(4-methoxyphenyl)ethane 1c; (O) 1-acetoxy-2-chloro-1-(4-
methoxyphenyl)propane 1b; (0) 1-acetoxy-2-bromo-1-(4-methoxyphenyl)-
propane 1d; and at 475 nm upon 266-nm irradiation of (4) 1,2-dichloro-
1,2-diphenylethane 3a and solvent ionizing ability (Y-values) of
methanol/water mixtures.
In neat acetonitrile, the stilbene radical cation was not
observed, and we expected to observe a transient due to the
presence of the â-chloro-1,2-diphenylethyl radical 4a. Absorp-
tion in the 300-330 nm region that could be due to the radical
was detected. However, while this absorption was quenched
by oxygen as expected for a radical species, the absorption was
weak with no defined maximum, and confident assignment of
this absorption to the radical 4a could not be made.
As shown in Figure 4e, the observed rate constant for stilbene
radical cation formation from the â-chloro radical 4a showed a
similar solvent effect to that measured for the reactions of
radicals 2a-d, with the rate constant increasing from 6.1 ×
106 s-1 in 17% aqueous acetonitrile to greater than 1 × 108 s-1
in 66% aqueous acetonitrile. An increase in the rate constant
for stilbene radical cation formation was also observed as TFE
content in acetonitrile was increased, Figure 4e.
the radical cation. This situation whereby both the formation
of the â-substituted radical by nucleophilic addition to a radical
cation and decomposition of the radical by â-heterolysis are
rapid processes leads to the possibility that an equilibrium state
containing the â-radical and the radical cation, eq 5, can be
generated in the presence of an appropriate amount of a suitable
nucleophile, either chloride or bromide.
The trans-stilbene radical cation was also produced upon
photolysis of 1,2-dibromo-1,2-diphenylethane 3b. In this case,
formation of the radical cation was an exceptionally fast process
in ionizing solvents, taking place with rate constants near 108
s-1 even in water/acetonitrile mixtures containing as little as
10% water. As a result of the large rate constants associated
with this reaction, detailed solvent effects were not investigated.
Ionization in Water/Methanol Mixtures. Solvent ionizing
abilities as measured by Y values are not readily available for
acetonitrile mixtures but are known for water/methanol mix-
tures.29 The rate constants for the formation of the heterolysis
product of the radicals 2a-d and 4a were therefore measured
in water/methanol mixtures from 0% to 70% water. In each
case, the rate constants for growth of the radical cations
increased as a function of water content. The results are given
in Figure 9 as the rate constants plotted against Y values.
Determination of Radical/Radical Cation Equilibrium
Constants. It has previously been demonstrated that substituted
styrene radical cations often undergo efficient and rapid addition
of nucleophiles to the â-position to produce the same type of
â-substituted radicals generated in the present work.23,30-34 On
the other hand, the results described above provide clear
evidence that these radicals rapidly undergo the reverse reaction
under certain conditions and lose the nucleophile to regenerate
Assuming that heterolysis is the dominant reaction of the
radical, measuring the rate constant for radical cation formation
as a function of the concentration of either chloride or bromide,
depending on the identity of the â-substituent, is one method
for evaluating rate constants for the formation of the equilibrium
state. Under these conditions, the observed rate constant for
the decay of the radical represents the rate constant for the
approach to equilibrium, and the rate equation for the decay of
the â-substituted radical and the formation of the radical cation
will have the following form, eq 6.
kobs ) khet + kcomb[X-]
(6)
According to this rate equation, the observed rate constant
for the formation of the radical cation should increase as a
function of the concentration of the common-ion, with the
intercept representing the rate constant for heterolysis, khet, and
the slope representing the rate constant for the addition of the
nucleophile to the radical cation, kcomb
.
As shown in Figure 10, the rate constant for the formation
of the 4-methoxystyrene radical cation upon reaction of
â-chloro-4-methoxyphenethyl radical in 50% aqueous acetoni-
trile did indeed increase as the concentration of chloride was
increased from 0.0015 to 0.007 M. A good linear relationship
is observed, and from the slope a second-order rate constant of
3.6 × 108 M-1 s-1 for the addition of chloride to the radical
cation of 4-methoxystyrene can be established. In addition, the
optical density at 600 nm after the radical cation was fully
formed decreased significantly as the concentration of chloride
increased. This is consistent with the formation of a radical/
radical cation equilibrium, with the radical cation component
(29) Fainberg, A. H.; Winstein, S. J. Am. Chem. Soc. 1956, 78, 2770-
2777.
(30) Maroulis, T.; Shigemitsu, Y.; Arnold, D. R. J. Am. Chem. Soc. 1978,
100, 535-541.
(31) Neunteufel, R. A.; Arnold, D. R. J. Am. Chem. Soc. 1973, 95, 4080-
4081.
(32) Majima, T.; Pac, C.; Nakasone, A.; Sakurai, H. J. Am. Chem. Soc.
1978, 103, 4499-4508.
(33) Farid, S.; Mattes, S. L. J. Am. Chem. Soc. 1986, 108, 7356-7361.
(34) Arnold, D. R.; Du, X.; Chen, J. Can. J. Chem. 1995, 73, 307-318.

Dynamics of Ionization Reactions of â-Substituted Radicals
J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10657
Scheme 2
of radical cation are not taking place. Thus, the rate constants
measured in the ionizing solvents reflects predominantly (if not
exclusively) the heterolysis reaction of the radical.
A significant observation from these results is that ionization
of bromide and chloride from the â-substituted radicals is an
extremely fast reaction, with the first-order rate constants
exceeding 1 × 108 s-1 in highly ionizing solvents such as neat
water or HFIP. The radical moiety therefore has an enormous
rate accelerating effect on the rate constants for the heterolysis
of the primary or secondary halides. For example, the proto-
typical secondary halide, isopropyl bromide, undergoes sol-
volysis in water (at 50 °C) with a rate constant35 of about 1 ×
Figure 10. Relationship between the rate constants for the formation
of the radical cations upon 266-nm laser irradiation of (b) 1-acetoxy-
2-chloro-1-(4-methoxyphenyl)ethane 1a and chloride concentration and
(O) 1-acetoxy-2-bromo-1-(4-methoxyphenyl)ethane 1c and bromide ion
concentration in nitrogen-saturated 50% water/50% acetonitrile.
10-5
s-1 that is at least 1013 times smaller than the rate constant
decreasing as the concentration of the chloride ion increases.
Similar results were obtained by monitoring the rate constant
for the growth of the 4-methoxystyrene radical cation by
ionization of the â-bromo-4-methoxyphenethyl radical as a
function of bromide concentration. A reasonably linear plot is
again observed, Figure 10, with the slope corresponding to the
addition of bromide to the radical cation of kcomb ) 3.1 × 109
for ionization of bromide from the secondary â-bromo radical
2d studied in the present work. The magnitude of the
accelerating effect can also be illustrated by comparing the rate
constant for SN1 ionization of 4-methoxyphenethyl chloride of
approximately 1000 s-1 in 20% aqueous acetonitrile36 to the
rate constant of 4.6 × 106 s-1 for ionization of chloride from
the â-chloro-4-methoxyphenylpropyl radical 2b in 17% aqueous
acetonitrile. In this case, insertion of the CH• group between
the anisyl group and the carbon bearing the chloride leaving
group causes a 5000-fold increase in the rate constant for
heterolysis.
A simple explanation for the observed rate accelerating effect
of the radical center is that the positive charge formed upon
ionization is not localized at the carbon undergoing ionization
but is instead delocalized into the adjacent, half-filled p-orbital
of the radical center, Scheme 2.
Thus, in a similar manner to the mechanisms by which a lone
pair of electrons of oxygen in R-alkoxymethylhalides or
π-electrons of unsaturated carbons in R-vinylmethyl halides
increase rate constants for solvolysis by stabilizing the incipient
carbocation, the half-filled orbital can interact directly with the
σ-bond undergoing cleavage to stabilize the positive charge
formed in the heterolysis reaction. In fact, the acceleration
provided by the radical center is similar to that provided by
resonance donating groups such as oxygen or vinyl groups, as
exemplified by the 1013-fold increase in the rate constant for
solvolysis of methoxymethyl chloride compared to solvolysis
of n-propyl chloride.37
Further evidence for the direct participation of the radical
center in the heterolysis reaction comes from the effect of
substituents on the aromatic ring. In our experiments, we could
find no evidence for heterolysis of the â-bromo-4-methylphen-
ethyl radical 2e or the â-bromophenethyl radical 2f, indicating
that heterolysis of these radicals not bearing the highly electron-
donating 4-methoxy group takes place with a rate constant <105
s-1 even in highly ionizing solvents such as HFIP.38 This upper
limit is considerably smaller than rate constants of >108 s-1
measured for the heterolysis of the corresponding 4-methoxy
substituted radical 2c in neat HFIP and water. The heterolysis
process is therefore very sensitive to the substituents on the aryl
ring, with the electron-donating 4-methoxy group accelerating
M-1 s-1
.
When similar experiments were carried out with the â-chloro
and the â-bromo-4-methoxyphenylpropyl radicals in 50% aque-
ous acetonitrile, the observed rate constants for formation of
the anethole radical cation by heterolysis of the â-halide were
not affected by halide concentrations up to 0.05 M. This
suggests that the rate constants for heterolysis of the radicals
under these conditions, khet ) 1.1 × 107 s-1 for the â-chloro
derivative and khet ) 5.2 × 107 s-1 for the â-bromo derivative
(Figure 4c,d), are considerably more rapid than the rate constant
for the addition of the corresponding halide to the radical cations,
even at halide ion concentrations up to 0.05 M.
Discussion
Observation of the aryl alkene radical cations in the present
work provides unambiguous evidence that the â-substituted
arylethyl radicals 2a-d and 4a,b undergo heterolytic cleavage
of the carbon-leaving group bond in ionizing solvents. Thus,
the observed rate constants for the formation of the radical
cations measured in the present work must include a term for
the rate constant for the heterolysis reaction of the radical.
However, the observation of the formation of the radical cations
does not necessarily rule out the possibility that the observed
rate constants also contain components for other reactions
available to the radical, such as rapid homolysis or SN2
substitution. Since homolysis is unlikely to be strongly
influenced by solvent polarity, the dramatic increase in the
magnitude of the rate constant for the reaction of the radical as
solvent ionizing ability increases can be attributed to an
increased rate constant for the heterolysis process. Thus, in
solvents where the rate constant for radical cation formation is
significantly greater than the reaction of the radical in neat
acetonitrile, it seems reasonable to conclude that homolysis does
not contribute significantly to the measured rate constants. In
addition, a close correspondence is seen between the increase
in the relative yield of the radical cation as measured by the
absorbance at 600 nm and the magnitude of the rate constant
for the radical reaction as the ionizing ability of the solvent is
increased. This clearly suggests that competing processes such
as homolysis or SN2 substitution which would decrease the yield
(35) Laughton, P. M.; Robertson, R. E. Can. J. Chem. 1956, 34, 1714-
1719.
(36) Extrapolated from Table 1 in: Richard, J. P.; Jencks, W. P. J. Am.
Chem. Soc. 1984, 106, 1383-1396.
(37) Streitwieser, A., Jr. SolVolytic Displacement Reactions; McGraw-
Hill: New York, 1962.

10658 J. Am. Chem. Soc., Vol. 119, No. 44, 1997
Cozens et al.
heterolysis of the â-substituent. Since substituents have only a
small effect on benzyl radical stability, the large substituent
effect must be due to stabilization of the positive charge formed
as the heterolysis reaction proceeds to the transition state. Given
that the radical center sits between the phenyl ring and the site
of ionization, the large substituent effect indicates that the
electron-donating stabilizing effect of the phenyl substituent is
transmitted effectively through the radical center and is con-
sistent with electron-donation from the radical center to the
carbon-leaving group bond undergoing heterolysis.
Cation stabilizing groups such as methyl or phenyl attached
directly to the carbon undergoing ionization also play an
important role in determining the overall reaction rate constant.
The â-methyl substituted radicals 2b and 2d undergo heterolysis
with rate constants that are consistently larger than the het-
erolysis rate constants for the corresponding â-unsubstituted
radicals 2a and 2c. The rate acceleration caused by the â-methyl
group can be attributed, at least in part, to the electron-donating
ability of the methyl groups that stabilize the positive charge
formed at the carbon undergoing ionization. It should also be
noted that the oxidation potential of anethole is about 100 mV
lower than that for 4-methoxystyrene,39 indicating that the
presence of the methyl group at the â-position has a substantial
stabilizing effect on the radical cation formed by ionization of
the leaving group. Thus, faster ionization in the presence of
the â-methyl group may also be a reflection of the increased
stability of the radical cation.
The presence of a phenyl group at the â-position has an even
more dramatic effect on the rate constants for ionization than a
methyl group. No ionization of the â-bromophenethyl radical
2f was observed, even in a highly ionizing solvent like HFIP.
On the other hand, the â-chloro- (4a) and the â-bromo (4b)
substituted 1,2-diphenylethyl radicals, which have a phenyl
group at the â-position, undergo rapid ionization. In fact,
heterolysis of bromide from radical 4b was the fastest reaction
measured in the present work, and ionization of chloride from
4a takes place with rate constants of 8.5 × 107 s-1 in 50%
aqueous acetonitrile and 3.9 × 107 s-1 in 50% TFE/50% AcN
that are considerably larger than the rate constants for ionization
of chloride from the 4-methoxyphenethyl and the 4-methox-
yphenyl propyl radicals. Since the oxidation potential for
stilbene40,41 is approximately 100-200 mV higher than the
oxidation potentials of 4-methoxystyrene and anethole, the
relatively faster rate constants for heterolysis to give the stilbene
radical cation cannot be explained solely on the basis of radical
cation stability. Instead, the large rate enhancement must result
from stabilization of the positive charge by the phenyl group
as the heterolysis reaction proceeds to the transition state.
The greater leaving ability of bromide ion relative to chloride
ion in simple SN1 reactions is well-known in physical organic
chemistry. This well-established order is clearly followed for
the reactions of the â-halo radicals studied in the present work,
which shows that in any given solvent mixture, ionization of
the â-bromo radicals takes place with rate constants that are
5-50 times greater than ionization of the corresponding â-chloro
radicals.
Effect of Solvent on Rate Constant for Ionization. Rate
constants for nucleophilic substitution of alkyl halides are
strongly influenced by solvent, with Y parameters for solvent
ionizing ability being derived from the effect of solvent on the
rate constants for rate-limiting ionization of substrates such as
adamantyl derivatives. In general, the rate constants for the
heterolysis of the â-substituted radicals studied in the present
work behave as expected with respect to increasing ionizing
ability of the solvent,42 with heterolysis being fastest in
acetonitrile solutions containing water followed by solutions
containing HFIP and then TFE.43 Interestingly, however, the
magnitude to which the rate constants are influenced by ionizing
ability of the solvent is strongly dependent on the structure of
the radical. For example, the rate constant for the loss of
bromide from the â-bromo-4-methoxyphenylpropyl radical 2d
increases 2.5-fold upon going from 50% TFE/50%AcN to 100%
TFE, whereas the rate constant for the same reaction of the
â-bromo-4-methoxyphenethyl radical 2c increases by almost 20-
fold under identical conditions.
A more accurate assessment of the effect of solvent ionizing
ability on the rate constants for the ionization reaction of the
â-substituted radicals can be obtained from Winstein plots of
the data obtained in methanol/water mixtures. These plots of
the rate constant vs Y values29 for each methanol/water mixture
are shown in Figure 9. All the plots are reasonably linear but
have slopes that change substantially depending on the nature
of the leaving group and the presence or absence of alkyl or
phenyl groups at the â-position. Thus, m ) 0.72 calculated
from the slope of the plot of the rate constant for heterolysis of
the â-chloro-4-methoxyphenethyl radical 2a as a function of Y
is considerably larger than m ) 0.54 from the slope of the same
plot for the corresponding â-bromo-substituted radical 2c.
Similarly, the m value of 0.21 for the heterolysis of the â-chloro-
4-methoxyphenylpropyl radical 2b is slightly larger than the m
value of 0.18 measured for the â-bromo derivative 2d. In
addition, ionization of chloride or bromide from the 4-meth-
oxyphenethyl radicals 2a and 2c to give the 4-methoxystyrene
radical cation is significantly more sensitive to solvent ionizing
ability than formation of the more stable anethole radical cation
produced by ionization of chloride or bromide from the
4-methoxyphenylpropyl radicals 2b and 2d.
This variation in m value can be explained by noting that the
observed rate constants for heterolysis of the carbon-bromine
bond are substantially larger than those for the ionization of
chloride. The higher reactivity for the â-bromo derivatives
indicates an earlier transition state that should be less sensitive
to environmental effects such as solvent polarity. The decrease
in the m value upon addition of a â-methyl group is also
consistent with the higher reactivity and earlier transition state
induced by the cation stabilizing influence of the â-methyl
group.
On the other hand, despite the fact that heterolysis of the
1,2-diphenylethyl radical 4a is more rapid than heterolysis of
the other radicals studied in the present work, it shows
considerable sensitivity to solvent ionizing ability, with m )
0.62. In this case, the positive charge at the carbon undergoing
(38) Note that failure to observe the radical cations from the â-bromo-
4-methylphenethyl radical does not necessarily indicate that ionization is
not taking place. Ionization may be occurring, but, given the higher oxidation
potential of 4-methylstyrene, rapid electron-transfer of the radical cation/
bromide ion pair may take place to give the neutral 4-methylstyrene and a
bromine atom. If this reaction does occur, the initial ionization step would
still be solvent dependent, and the observed decay of the radical would
increase in ionizing solvents. Since the decay of the â-bromo-4-meth-
ylphenethyl radical and the phenyl unsubstituted derivative remained
unchanged upon going from acetonitrile to water, ionization followed by
electron-transfer is not likely to be a dominant process under our conditions.
(39) Workentin, M. S.; Schepp, N. P.; Johnston, L. J.; Wayner, D. D.
M. J. Am. Chem. Soc. 1994, 116, 1141-1142.
(42) An increased rate constant for â-heterolysis of model compounds
of 4′-DNA radicals upon going from methanol to 1:1 acetonitrile:water has
recently been demonstrated.4
(43) Y(H2O) ) 4.1; Y(HFIP) ) 3.8; Y(TFE) ) 1.8. Y values taken from:
March, J. AdVanced Organic Chemistry, 4th ed.; John Wiley & Sons: New
York, 1992; p 361.
(40) Lewis, F. D.; Bedell, A. M.; Dykstra, R. E.; Elbert, J. E.; Gould, I.
R.; Farid, S. J. Am. Chem. Soc. 1990, 112, 8055-8064.
(41) Lewis, F. D.; Kojima, M. J. Am. Chem. Soc. 1988, 110, 8664-
8670.

Dynamics of Ionization Reactions of â-Substituted Radicals
J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10659
Conclusion
Table 1. Rate Constants and Equilibrium Constants for the
â-Halo-4-methoxyphenethyl Radical-4-Methoxystyrene Radical
Cation Equilibrium in 50% Aqueous Acetonitrile (25 °C)
The primary mode of decay of the radicals 2a-d and 4a,b
under ionizing conditions is heterolytic cleavage of the â-chloro
or â-bromo substituent to generate the corresponding radical
cations. The ionization reactions are remarkably fast, taking
place with rate constants in the 107 s-1 to g 108 s-1 range in
neat HFIP or neat water. These large rate constants highlight
the ability of the radical centers to interact with the positive
charge formed as the heterolysis process proceeds from the
initial state to the transition state. The kinetics for the ionization
of the â-substituted radicals are otherwise controlled by factors
such as solvent ionizing ability, nature of leaving group and
electronic effects in the same way as typical SN1 ionization.
However, the sensitivity of the heterolysis reactions to solvent
ionizing ability is distinctly dependent on the nature of the
leaving group and the structure of the radical cation product.
Investigations into the reactivity of these kinds of radicals with
different substituents at the â-position as well as efforts to detect
SN2 substitution reactions are currently in progress.
halide
kcomb/M-1 s-1
khet/s-1
Keq/M
Br-
Cl-
3.1 × 109
7.0 × 106
0.0023
0.0050
3.6 × 108
1.8 × 106
ionization may not be delocalized into the adjacent benzyl
radical moiety to the same degree as delocalization into the
strongly resonance donating 4-methoxybenzyl radical moiety
of radicals 2a-d. Less delocalization of charge in the transition
state would leave the ionization reaction more dependent on
solvent stabilization of the charge and would therefore increase
the sensitivity of the rate constant to solvent ionizing ability.
Note that the m values derived from the slopes of the plots
in Figure 9 are all significantly less than 1. Small m values
sometimes indicate significant nucleophilic assistance and
substantial SN2 character in the substitution reaction. This is
unlikely in our experiments, since the initial cationic product
from the ionization reaction is observed directly. The small m
values therefore reflect the remarkably high velocity of the
ionization processes which suggests a very early transition state
and a corresponding decrease in the sensitivity to solvent
ionizing power. Similar small m values for slower ionization
reactions that have early transition states due to strain have been
noted.44
Experimental Section
Materials. The radical precursors, 1-acetoxy-2-chloro-1-(4-meth-
oxyphenyl)ethane 1a, 1-acetoxy-2-chloro-1-(4-methoxyphenyl)propane
1b, 1-acetoxy-2-bromo-1-(4-methoxyphenyl)ethane 1c, 1-acetoxy-2-
bromo-1-(4-methoxyphenyl)propane 1d, 1-acetoxy-2-chloro-1-(4-me-
thylphenyl)ethane 1e, and 1-acetoxy-2-chloro-1-(phenyl)ethane 1f, were
prepared by bromination or chlorination of the appropriate styrene in
glacial acetic acid. 1,2-Dichloro- and 1,2-dibromo-1,2-diphenylethane,
3a and 3b, respectively, were prepared by either chlorination or
bromination of trans-stilbene. All materials were characterized by 1H-
and 13C-NMR (Bruker 250 MHz) using CDCl3 as the solvent.
Acetonitrile (AcN) was spectroscopic grade (Omnisolve, BDH). 2,2,2-
Trifluoroethanol (TFE) and 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP)
were purchased from Aldrich and used as received. Methanol and water
were doubly distilled prior to use.
Laser Flash Photolysis. The nanosecond laser flash photolysis
system at Dalhousie University is of standard design.46,47 The samples,
typically 2.5 mL, were contained in 7 × 7 mm2 laser cells made out of
Suprasil quartz tubing and bubbled with a slow stream of dry nitrogen
for 20 min prior to laser irradiation. The excitation source was the
fourth harmonic from a Continuum Nd:YAG NY-61 laser (266 nm; e
8 ns/pulse; e 15 mJ/pulse). Transient signals from a monochromator/
photomultiplier system were initially captured by a Tektronix 620A
digital oscilloscope and transferred to a Power Macintosh computer.
The experiments are computer controlled using the Power Macintosh
and software written using the Labview programming language. The
absorbance of the substrates at 266 nm was ≈0.3 in the 7 × 7 mm2
laser cells. The kinetic experiments were carried out in individual static
samples; in these experiments, each sample was exposed to a maximum
of three laser pulses. The spectra were obtained under solvent flow
conditions to ensure that there was no build-up of photolysis products
during the laser experiments.
Radical/Radical Cation Equilibrium. Nucleophiles are
well-known to add to the â-position of aryl alkene radical cations
to generate â-heteroatom substituted radicals that then go on to
give final products.30,31 The results from the present work show
that in some cases, such as in polar, ionizing solvents, addition
of the nucleophile will be reversible, especially under circum-
stances where the nucleophile is also a good leaving group.
These results imply that the efficiency of product formation by
nucleophilic addition to radical cations will not only be related
to the rate constant for nucleophilic addition but also to the
magnitude of the reverse heterolysis reaction. In addition,
unusually slow rate constants for addition of nucleophiles to
radical cations may be, in part, because of a rapid reversible
ionization.45 For example, the lifetime of the anethole radical
cation has been found to be completely insensitive to bromide
and chloride ion at concentrations up to 1.0 M.23 While the
rate constant for addition of these nucleophiles to the anethole
radical cation should be considerably less than diffusion
controlled in aqueous solution, the fact that ionization of chloride
and bromide takes place with rate constants >1 × 108 s-1
indicates that the anethole radical cation would not be quenched
by these nucleophiles, even if the second-order rate constants
for nucleophilic addition are as fast as 1 × 107 M-1 s-1
.
Acknowledgment. We gratefully acknowledge the Natural
Science and Engineering Research Council of Canada (NSERC)
for financial support of this research through an equipment and
operating grant (F.L.C.). We would also like to thank Dalhousie
University for support of our research program. F.L.C. is the
recipient of an NSERC-WFA.
Equilibrium constants for the radical/radical cation equilib-
rium reaction determined by combination of the rate constants
for addition, kcomb, of halides to the radical cation and heteroly-
sis, khet, of the halides are shown in Table 1. These equilibrium
constants indicate that at chloride concentrations up to 0.005
M and bromide concentrations up to 0.0023 M, the dominate
species present at equilibrium will be the radical cation, with
the nucleophile adduct being the less favored species.
Supporting Information Available: Tables of observed rate
constants (3 pages). See any current masthead page for ordering
and Internet access instructions.
JA971996P
(44) Lowry, T. H.; Richardson, K. S. Mechanism and Theory in Organic
Chemistry, 3rd ed.; Harper & Row: New York, 1987; p 391.
(45) Trampe, G.; Mattay, J.; Steenken, S. J. Phys. Chem. 1989, 93, 7157-
7160.
(46) Scaiano, J. C.; Tanner, M.; Weir, D. J. Am. Chem. Soc. 1985, 107,
4396-4403.
(47) Scaiano, J. C. J. Am. Chem. Soc. 1980, 102, 7747-7753.