How solvent modulates hydroxyl radical reactivity in hydrogen atom abstractions

Article abstract of DOI:10.1021/ja903856t

The hydroxyl radical (HO) is a highly reactive oxygen-centered radical whose bimolecular rate constants for reaction with organic compounds (hydrogen atom abstraction) approach the diffusion-controlled limit in aqueous solution. The results reported herein show that hydroxyl radical is considerably less reactive in dipolar, aprotic solvents such as acetonitrile. This diminished reactivity is explained on the basis of a polarized transition state for hydrogen abstraction, in which the oxygen of the hydroxyl radical becomes highly negative and can serve as a hydrogen bond acceptor. Because acetonitrile cannot participate as a hydrogen bond donor, the transition state cannot be stabilized by hydrogen bonding, and the reaction rate is lower; the opposite is true when water is the solvent. This hypothesis explains hydroxyl radical reactivity both in solution and in the gas phase and may be the basis for a containment strategy used by Nature when hydroxyl radical is produced endogenously.

Full text of DOI:10.1021/ja903856t

Published on Web 02/10/2010
How Solvent Modulates Hydroxyl Radical Reactivity in
Hydrogen Atom Abstractions
Susan Mitroka, Stephanie Zimmeck, Diego Troya, and James M. Tanko*
Department of Chemistry, Virginia Polytechnic Institute and State UniVersity,
Blacksburg, Virginia 24061
Received May 12, 2009; E-mail: jtanko@vt.edu
Abstract: The hydroxyl radical (HO•) is a highly reactive oxygen-centered radical whose bimolecular rate
constants for reaction with organic compounds (hydrogen atom abstraction) approach the diffusion-controlled
limit in aqueous solution. The results reported herein show that hydroxyl radical is considerably less reactive
in dipolar, aprotic solvents such as acetonitrile. This diminished reactivity is explained on the basis of a
polarized transition state for hydrogen abstraction, in which the oxygen of the hydroxyl radical becomes
highly negative and can serve as a hydrogen bond acceptor. Because acetonitrile cannot participate as a
hydrogen bond donor, the transition state cannot be stabilized by hydrogen bonding, and the reaction rate
is lower; the opposite is true when water is the solvent. This hypothesis explains hydroxyl radical reactivity
both in solution and in the gas phase and may be the basis for a “containment strategy” used by Nature
when hydroxyl radical is produced endogenously.
Solvent polarity can have a huge effect on the kinetics of
reactions involving or forming charged species in solution. In
contrast, reactions of neutral radicals are much less sensitive to
solvent polarity effects, mainly because charged species are not
involved, and there is not a significant change in dipole moment
in the progression from reactants to transition state. Because of
this, other subtler solvent properties such as viscosity or internal
pressure can influence the rate of certain radical reactions; such
solvent effects are much more difficult to detect in polar
reactions because they are masked by the overwhelming effect
of solvent polarity.1
For example, solvent viscosity can affect the rate and product
distribution when radical caged-pairs (geminate or diffusive)
are involved. Internal pressure can influence rate if there is a
difference in the volume of the reactants compared to the
transition state (∆Vact * 0) and can influence the relative rate
of some radical reactions. However, these solvent effects are
generally small, with changes in rate or product distribution not
much greater than an order of magnitude.1 Consequently,
instances where solvent dramatically affects the rate or selectiv-
ity of reactions involving neutral radicals are rare and noteworthy.
The classic example of a significant solvent effect in a radical
reaction involves free radical chlorinations of alkanes conducted
in benzene solvent. The chain-carrying chlorine atom forms a
complex with benzene, lowering its reactivity and increasing
its selectivity (by nearly 2 orders of magnitude) in hydrogen
atom abstractions.1,2 A more recent example of a significant
solvent effect was reported by Ingold and co-workers, who
found that rate constants for hydrogen atom abstractions from
phenols were reduced in solvents where the phenol was
stabilized by hydrogen bonding.3 In this case, it was the
reactivity of the substrate, not the radical, that was diminished
as a result of a solute/solvent interaction.
The hydroxyl radical (HO•) is an extremely reactive oxidant,
important in chemistry, biology, medicine, materials, and the
environment. In biology, hydroxyl radical is recognized as the
most reactive of the so-called reactive oxygen species (ROS).
Cancer, arthritis, and Parkinson’s disease are but a few of the
ailments that are linked to hydroxyl radical. Despite it’s role in
disease, HO• is also a vital part of the body’s natural defense
mechanisms. Reactive oxygen species are produced endog-
enously as a means of destroying foreign antigens or abnormal
cells. It is often stated that in biological systems, hydroxyl
radical reacts with the first molecule it encounters.
Hydroxyl radical is also important in atmospheric chemistry
because of its ability to oxidize volatile organic pollutants and
has been referred to as the atmosphere’s “detergent.”4 Recently,
Vöhringer-Martinez et al. examined the role of water in the gas-
phase reaction of the hydroxyl radical with acetaldehyde,
reporting that a water concentration of 3% led to an increase in
the rate of hydrogen abstraction (abstraction taking place from
the aldehydic hydrogen).5 Their hypothesis that hydrogen
bonding to water in the transition state lowered the reaction
barrier raises several intriguing questions: Why is hydrogen
bonding to water more important in the transition state as
compared to reactants? Is this stabilization unique to substrates
with functional groups that are capable of hydrogen bonding
(such as the carbonyl group in acetalaldehyde)? Is this stabiliza-
tion also important for reactions of hydroxyl radical in solution?
In short, how general is this phenomenon, and can we
understand it on the molecular level?
(1) Tanko, J. M.; Suleman, N. K. Solvent Effects in the Reactions of
Neutral Free Radicals. In Energetics of Organic Free Radicals; Simo˜es,
J. A. M., Greenberg, A., Liebman, J., Eds.; Chapman & Hall: New
York, 1996; pp 224-293.
(3) Litwinienko, G.; Ingold, K. U. Acc. Chem. Res. 2007, 40, 222–230.
(4) Li, S.; Matthews, J.; Sinha, A. Science 2008, 319, 1657–1660.
(5) Vo¨hringer-Martinez, E.; Hansmann, B.; Hernandez-Soto, H.; Francisco,
J. S.; Troe, J.; Abel, B. Science 2007, 315, 497–501.
(2) Russell, G. A. J. Am. Chem. Soc. 1958, 80, 4987–4996.
9
10.1021/ja903856t  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 2907–2913 2907

A R T I C L E S
Mitroka et al.
Table 1. Rate Constants for Hydrogen Abstraction by HO• from
Various Organic Substrates in CH3CN and H2Oa
Until recently, pulse radiolysis was one of a few means of
generating hydroxyl radical in solution for the determination
of absolute reaction rates. Because this technique, by definition,
involves the radiolysis of water, solution-phase studies of
hydroxyl radical kinetics have largely been conducted in water.6
As a result, relatively little is known about the reactivity of HO•
in other solvents. Recent developments have made it possible
to study the hydroxyl radical in a nonaqueous environment
through the use of photolabile hydroxyl radical precursors.
N-Hydroxypyridine-2-thione (PSH), developed by Zard et al.,7
is one such precursor that, when photolyzed at 355 nm, cleanly
produces the hydroxyl radical. Because the radical byproduct
(pyrithyl radical) is relatively stable, decaying on a microsecond
time scale, this precursor is an ideal candidate for studying the
hydroxyl radical kinetics in solution via laser flash photolysis
(LFP).8
kH(H2O) × 10-7
kH(H2O)/
kH(CH3CN)
kH(CH3CN) × 10-7
substrate
(M-1 s-1
)
(M-1 s-1
)
(CH3)3CC(CH3)3
CH3(CH2)4CH3
CH3(CH2)5CH3
c-C6H11-CH3
c-C6H12
(CH3)2CHCH(CH3)2
CH3(CH2)3OH
CH3CH2OH
(CH3)2CHOH
CH3OH
(CH3)3COH
(CH3CH2)2O
CH3OC(CH3)3
CH3CH2OC(CH3)3
C4H8O (THF)
CH2Cl2
5.83 ((0.61)
4.48 ((0.79)
6.24 ((0.88)
4.22 ((0.28)
6.72 ((0.99)
12.0 ((3.90)
4.20 ((0.98)
8.28 ((3.20)
7.19 ((1.72)
5.90 ((0.67)
3.62 ((0.03)
4.56 ((1.00)
4.35 ((0.35)
1.78 ((0.63)
3.50 ((0.65)
7.63 ((0.89)
3.21 ((1.16)
8.41 ((2.32)
4.85 ((0.92)
5.09 ((0.96)
b
b
b
b
147
123
168
90
660
770
710
610
c
c
f
-
100
23
28
420 ((33.4)
193 ((10.4)
200 ((24.0)
e g j l p ee
- - , -
,
,
c d g i q s
, , -
,
,
c
g i k m n
, ,
-
, ,
16
95 ((4.4)
s t
,
78
37
355 ((127)
t
160
u V
,
126
117
1.2
3.5
1
225 ((88)
t
410
Platz et al.8b,c developed a method for “visualizing” the
hydroxyl radical, which does not have a convenient absorption
in the UV-vis: hydroxyl radical addition to the ortho and para
positions of trans-stilbene produces an adduct with an absorption
at 392 nm, allowing trans-stillbene to act as a viable spectro-
scopic probe for monitoring HO• kinetics. These workers also
noted a solvent effect on HO• reactivity. The rate constant for
addition to aromatics (trans-stilbene, benzene) was observed
to be lower in acetonitrile compared to water. Molecular orbital
calculations supported the notion that this was due to a polarized
transition state, reminiscent of the polar effect introduced by
Russell,9 Walling,10 and others11 for hydrogen atom abstraction
reactions (Vide infra). Other studies reporting Hammett param-
eters for HO• addition reactions to aromatic compounds have
given negative F+ values, consistent with the buildup of
negative charge on the hydroxyl moiety in the transition state.12
The strong dipole that is formed in the transition state has the
ability to be stabilized by the surrounding solvent.
Very little is known about the kinetics of hydrogen abstrac-
tions by HO• in nonaqueous solvents. Amplified by high level
molecular orbital calculations, the observations reported herein
provide answers to all of the questions posed above and a
comprehensive understanding of HO• reactivity in solution and
the gas phase, correcting what we now believe to be some of
the misconceptions pertaining to the chemistry of hydroxyl in
solution.
w
y
-
8.8 ((1.8)
11.3 ((2.7)
10.5 ((0.97)
d f i s
, , ,
(CH3)2CO
CHBr3
CHCl3
ClCH2CO2H
w z
,
w aa dd
,
-
0.35
0.85
1.74 ((1.4)
f
4.3
a The rate constants in water were obtained from the Notre Dame
Radiation Laboratory database (http://www.rcdc.nd.edu/index.html);
these values were verified by consulting the original papers. b Reference
17. c Reference 18. d Reference 19. e Reference 20. f Reference 21.
g Reference 22. h Reference 23. i Reference 24. j Reference 25.
k Reference 26. l Reference 27. m Reference 28. n Reference 29.
o Reference 30. p Reference 31. q Reference 32. r Reference 33.
s Reference 34. t Reference 35. u Reference 36. V Reference 37.
w Reference 38. x Reference 39. y Reference 40. z Reference 41.
aa Reference 42. bb Reference 43. cc Reference 44. dd Reference 45.
ee Reference 46.
8452A). Laser flash photolysis experiments were conducted using
an Applied Photophysic LKS.60 spectrometer using the third
harmonic of a Continuum Surelite I-10 Nd:YAG laser (4-6 ns
pulse, 355 nm). Transient signals were monitored by a Hewlett-
Packard Infinium digital oscilloscope and analyzed with the Applied
Photophysics SpectraKinetic Workstation software package (v.
4.59).
Laser Flash Photolysis (LFP). Substrates were prepared in
acetonitrile or in an acetonitrile/water cosolvent and deoxygenated
prior to photolysis. (Steady-state UV-vis spectra were recorded
to verify that N-hydroxypyridine-2-thione was the only species
absorbing at the excitation wavelength). In all LFP experiments, a
fixed concentration of trans-stilbene (0.0015 M) was utilized as a
spectroscopic probe, monitoring the signal buildup at 392 nm. Low
laser power (ca. 10-20 mJ) was used in all experiments to eliminate
any laser power dependency to the observed rate constants thereby
minimizing the contributions of radical-radical reactions and other
artifacts. Substrate concentrations were varied over a factor of at
least 10 over 5-7 separate experiments.
Experimental Section
Materials. All of the solvents and fine chemicals used in this
study were obtained from Sigma-Aldrich. Liquid substrates were
distilled and N-hydroxypyridine-2-thione was recrystallized prior
to use.
Apparatus. Steady-state UV-vis spectra were recorded on a
Hewlett-Packard diode array UV-vis spectrophotometer (HP
Calculations. Electronic structure calculations were performed
using either the Spartan 0413 molecular modeling software or
Gaussian 03.13
(6) Xu, G.; Chance, M. R. Chem. ReV. 2007, 107 (8), 3514–3543.
(7) Boivin, J.; Cre´pon, E.; Zard, S. Z. Tetrahedron Lett. 1990, 47, 6869–
6872.
Results and Discussion
(8) (a) Aveline, B. M.; Kochevar, I. E.; Redmond, R. W. J. Am. Chem.
Soc. 1995, 117, 9699–9708. (b) DeMatteo, M. P.; Poole, J. S.; Shi,
X.; Sachdeva, R.; Hatcher, P. G.; Hadad, C. M.; Platz, M. S. J. Am.
Chem. Soc. 2005, 127, 7094–7109. (c) Poole, J. S.; Shi, X.; Hadad,
C. M.; Platz, M. S. J. Phys. Chem. A. 2005, 109, 2547–2551.
(9) Russell, G. A. Reactivity, Selectivity, and Polar Effects in Hydrogen
Atom Transfer Reactions. In Free Radicals; Kochi, J. K., Ed.; John
Wiley & Sons, Inc.: New York, 1973; Vol. I pp 275-331.
(10) Walling, C. Free Radicals in Solution; John Wiley & Sons, Inc.: New
York, 1957.
Rate constants for the reaction of HO• with a variety of
substrates in acetonitrile are summarized in Table 1. Compared
to tert-butoxyl radical,14 rate constants for hydrogen abstraction
by hydroxyl radical are generally 2-3 orders of magnitude
greater. The high reactivity of HO• is accompanied by low
selectivity. For aliphatic hydrogens, the per-hydrogen reactivity
is approximately tertiary (13.9) > secondary (1.4) > primary
(11) Roberts, B. P. Chem. Soc. ReV. 1999, 28, 25–35.
(12) (a) Merga, G.; Rao, B. S. M.; Mohan, H.; Mittal, J. P. J. Phys. Chem.
1994, 98, 9158–9164. (b) Anbar, M.; Meyerstein, D.; Neta, P. J. Phys.
Chem. 1966, 70, 2660–2662.
(13) Frisch, M. J. et al. Gaussian 03, ReVision C.02; Gaussian, Inc.:
Wallingford, CT, 2004.
9
2908 J. AM. CHEM. SOC. VOL. 132, NO. 9, 2010

Modulation of Hydroxyl Radical Reactivity by Solvent
A R T I C L E S
Scheme 1
(1.0), based upon a multiple regression analysis of the results
(assuming the global rate constant to be the sum of contributions
from each type of hydrogen). For the alcohols, the hydrogen of
the hydroxyl group is about 3 times more reactive than a
primary, aliphatic hydrogen.
A reviewer has vehemently argued that the results reported
herein, as well as the earlier results of Platz et al., are not
attributable to hydroxyl radical. Rather, it was suggested that
at the low substrate concentrations used in these experiments
(millimolar), HO• would instantaneously abstract hydrogen from
acetonitrile solvent, present in large excess, forming •CH2CN,
and that it was the kinetics of •CH2CN that was observed by
laser flash photolysis. (It was also suggested that the photolysis
byproduct, PyrS•, might also be a contributor to the observed
kinetics.) However, as noted by others, acetonitrile is an
exceptional solvent in that it is “extremely unreactive towards
‘electrophilic’ oxygen-centered radicals.”15 The evidence against
potential roles for either •CH2CN or PyrS• is reviewed and
discussed in detail in Supporting Information.
To provide new evidence that eliminates •CH2CN from
consideration as a hydrogen atom abstractor under these
conditions, some of the laser flash experiments were repeated
using a solvent devoid of abstractable hydrogens, specifically,
Freon 113 (1,1,2-trichlorotrifluoroethane). In this solvent, the
rate constants for hydrogen abstraction from cyclohexane and
methanol were 7.3 ((0.5) × 107 and 9.2 ((1.1) × 107 M-1
s-1, respectively, nearly identical to those measured in aceto-
nitrile (Table 1).
Preparative-scale experiments were performed in acetonitrile,
using cyclohexane and 2,3-dimethylbutane as substrates. The
photoinitiated (350 nm) reaction of PSH with alkanes yields
the corresponding sulfides in good yield, presumably via the
chain mechanism depicted in Scheme 1.7
To ascertain relative reactivities of primary, secondary, and
tertiary hydrogens, competition experiments were conducted
using 2,3-dimethylbutane as the source for primary and tertiary
hydrogens and cyclohexane as the source for secondary
hydrogens. Rather than being present in millimolar concentra-
tions as with the laser flash experiments, the hydrocarbon
substrates were used as cosolvents. The derived relative reac-
tivities in acetonitrile, tertiary (15.2) > secondary (3.9) > primary
(1.0), compare favorably to the values estimated from the laser
flash results (Vide supra), consistent with the role of HO• as
the hydrogen abstractor.
(14) Finn, M.; Friedline, R.; Suleman, N. K.; Wohl, C. J.; Tanko, J. M.
J. Am. Chem. Soc. 2004, 126, 7578–7584.
(15) Avila, D. V.; Brown, C. E.; Ingold, K. U.; Lusztyk, J. J. Am. Chem.
Soc. 1993, 93, 466–470.
(16) Reichardt, C. SolVents and SolVent Effects in Organic Chemistry, 2nd
ed.; VCH: Weinheim, 1990.
(17) Rudakov, E. S.; Volkova, L. K.; Tret’yakov, V. P. React. Kinet. Catal.
Lett. 1981, 16, 333–337.
(18) Buxton, G. V.; Greenstock, C. L.; Helman, W. P.; Ross, A. B. J. Phys.
Chem. Ref. Data 1988, 17, 513–886.
(19) Willson, R. L.; Greenstock, C. L.; Adams, G. E.; Wageman, R.;
Dorfman, L. M. Int. J. Radiat. Phys. Chem. 1971, 211–220.
(20) Adams, G. E.; Boag, J. W.; Michael, B. D. Trans. Faraday Soc. 1965,
61, 1417–1424.
(21) Adams, G. E.; Boag, J. W.; Currant, J.; Michael, B. D. In Pulse
Radiolysis; Ebert, M., Keene, J. P., Swallow, A. J., Baxendale, J. H.,
Eds.; Academic Press: New York, 1965; pp 131-143.
(22) Motohashi, N.; Saito, Y. Chem. Pharm. Bull. 1993, 41, 1842–1845.
(23) Park, H.-R.; Getoff, N., Z. Z. Naturforsch. A: Phys. Sci. 1992, 47A,
985–991.
(24) Wolfenden, B. S.; Willson, R. L. J. Chem. Soc., Perkin Trans. 1982,
805–812.
Unfortunately, these experiments do not proVe that the
observed chemistry is attributable to hydroxyl radical. Rather,
they only eliminate other reasonable alternative explanations.
In addition to the aforementioned trans-stilbene adducts with
PyrS• and •CH2CN, Platz and co-workers also considered and
eliminated triplet stilbene and stilbene radical cation as species
giving rise to the 392 nm transient. Computational studies
provided no evidence for HO•/CH3CN complexes to explain
the diminished reactivity of hydroxyl radical in acetonitrile; our
results in Freon 113 add additional support to this conclusion.
On the basis of the laser flash results and the accompanying
product studies, hydroxyl radical emerges as the most likely
explanation for the observed chemistry.
Table 1 also summarizes rate constants for reactions of HO•
in water obtained from the literature. In cases where more than
one value was available, these were averaged and reported with
95% confidence limits. For hydrocarbons, the rate constants for
hydrogen abstraction by HO• are nearly 2 orders of magnitude
lower in acetonitrile than in water solvent. When the substrate
possesses an electronegative substituent (halogen, carbonyl, etc.),
this difference diminishes to much less than an order of
magnitude. However, the solvent effect is restored; the rate
constants in water are again about 2 orders of magnitude greater
than in acetonitrile when the substrate bears an electron-donating
group such as alkoxyl or hydroxyl.
(25) Matheson, M. S.; Mamou, A.; Silverman, J.; Rabani, J. J. Phys. Chem.
1973, 77, 2420–2424.
(26) Elliot, A. J.; McCraken, D. R. Radiat. Phys. Chem. 1989, 33, 69–74.
(27) Buxton, G. V. Trans. Faraday Soc. 1970, 66, 1656–1660.
(28) Baxendale, J. H.; Khan, A. A. Int. J. Radiat. Phys. Chem. 1969, 1,
11–24.
(29) Neta, P.; Dorfman, L. M. AdV. Chem. Ser. 1968, 81, 222–230.
(30) Heckel, E.; Henglein, A.; Beck, G. Ber. Bunsenges. Phys. Chem. 1966,
70, 149–154.
(31) Matthews, R. W.; Sangster, D. F. J. Phys. Chem. 1965, 69, 1938–
1946.
(32) Elliot, A. J.; Simsons, A. S. Radiat. Phys. Chem. 1984, 24, 229–231.
(33) Greenstock, C. L.; Ng, M.; Hunt, J. W. AdV. Chem. Ser. 1968, 81,
397–417.
(34) Thomas, J. K. Trans. Faraday Soc. 1965, 61, 702–707.
(35) Eibenberger, J. Vienna University, Vienna, Austria, 1980.
(36) Karpel Vel Leitner, N.; Papailhou, A.-L.; Croue, J.-P.; Peyrot, J.; Dore,
M. Ozone Sci. Eng. 1994, 16, 41–54.
(37) Mezyk, S. P.; Cooper, W. J.; Bartels, D. M.; O’Shea, K. E.; Wu, T.
J. Phys. Chem. A 2001, 105, 3521–3526.
(38) Haag, W. R.; Yao, C. C. D. EnViron. Sci. Technol. 1992, 26, 1005–
1013.
(39) Getoff, N. Radiat. Phys. Chem. 1991, 37, 673–680.
(40) Emmi, S. S.; Beggiato, G.; Casalbore-Miceli, G.; Fuochi, P. G. J.
Radioanal. Nucl. Chem., Lett. 1985, 93, 189–197.
(41) Lal, M.; Mahal, H. S. Radiat. Phys. Chem. 1992, 40, 23–26.
(42) Rezansoff, B. J.; McCallum, K. J.; Woods, R. J. Can. J. Chem. 1970,
48, 271–276.
(43) Anbar, M.; Meyerstein, D.; Neta, P. J. Chem. Soc. B 1966, 742–747.
(44) Chutny, B. Collect. Czech. Chem. Commun. 1966, 31, 358–361.
(45) Bednar, J.; Teply, J. Collect. Czech. Chem. Commun. 1960, 25, 842–
851.
The results in Table 2 extend these observations: In going
from neat acetonitrile to a 90% acetonitrile/water cosolvent, the
(46) Adams, G. E.; Boag, J. W.; Michael, B. D. Trans. Faraday Soc. 1965,
61, 492–505.
9
J. AM. CHEM. SOC. VOL. 132, NO. 9, 2010 2909

A R T I C L E S
Mitroka et al.
Table 2. Rate Constants for Hydrogen Abstraction by HO• from
311G* levels are reported. It should be noted that every level
of theory (AM1, B3LYP/6-311G*, UHF/6-311G*) employed
gave virtually the same qualitative results. As the reactants
approach the transition state, the hydrogen being transferred
becomes substantially more positive, and the oxygen becomes
more negative, consistent with the notion of a polarized
transition state. However, the degree of polarization is far greater
for CH4 and CH3OH compared to CHCl3, as expected on the
basis of the preceding discussion. The negative charge on the
oxygen of hydroxyl radical in the transition state means that
this oxygen can be a hydrogen bond acceptor.
Various Organic Substrates in 90% CH3CN/H2O and 100% CH3CN
kH(CH3CN/H2O) × 10-7 kH(CH3CN) × 10-7 kH (CH3CN/H2O)/
substrate
(M-1 s-1
)
(M-1 s-1
)
kH(CH3CN)
CH3(CH)4CH3
c-C6H12
(CH3)2CHCH(CH3)2
CH3CH2OH
CHBr3
11.3 ((1.04)
15.2 ((0.68)
16.8 ((0.81)
11.1 ((0.80)
8.33 ((0.84)
4.26 ((1.06)
4.48 ((0.79)
6.72 ((0.99)
12.0 ((3.90)
8.28 ((3.20)
8.41 ((2.32)
4.85 ((0.92)
2.5
2.3
1.4
1.3
1.0
0.9
CHCl3
rate constants increase by about a factor of 2 for the hydrocar-
bons but remain virtually unchanged for substrates with elec-
tronegative substituents. (Higher proportions of water could not
be used because of solubility problems.)
As noted, there are in principle two contributors to the
observed solvent effect. In addition to hydrogen bonding, it is
also possible that solvent polarity plays a role in stabilizing the
transition state (relative to the reactants). However, the following
analysis strongly suggests that hydrogen bonding interactions
in the transition state, rather than a simple solvent polarity effect,
is the etiology of the effect.
The free energy of solvation of a polar molecule in a polar
solvent can be estimated by the Kirkwood equation and applied
to the reaction between two polar molecules A + B f transition
state (ts) via activated complex theory.16 The magnitude of the
solvent effect depends on the dipole moment (µ) and radius (r)
of the transition state relative to reactants, and the dielectric
constant (ε) of the solvent as expressed in eq 1, where ln k0
refers to the rate constant in a solvent of dielectric constant of
unity, ε0 is the permittivity of vacuum, and N, π, R, and T have
their usual meanings.
Walling,10 Russell,2 and others11 have argued for the impor-
tance of a polar transition state for hydrogen atom abstraction
reactions. These results can be explained by taking their ideas
one step further (Figure 1): because oxygen is more electro-
negative than carbon and hydrogen, in the transition state for
hydrogen abstraction, electron density is pulled toward the
oxygen of the hydroxyl radical, giving it a partial negative
charge and a partial positive charge on the RH portion of
the transition state. This development of negative charge on the
oxygen of the hydroxyl radical affords the opportunity for the
solvent (H2O) to stabilize the transition state through its polarity
and/or ability to participate in hydrogen bonding. Table 2 shows
that for hydrocarbons even 10% addition of water has a
significant effect on the rate because the transition state is so
highly polarized.
µ2ts
µA2
rA3
µB2
rB3
1
N
ε - 1
2ε + 1
ln k ) ln k0 +
-
-
(1)
(
)
)
3
In contrast, when the substrate possesses electronegative
substituents (e.g., halogen, carbonyl), the transition state is less
polarized because the substituent competes for electron density;
there is less transfer of negative charge to the oxygen of the
hydroxyl radical and hydrogen bonding interactions are expected
to be weaker, thus explaining why there is little to no rate
enhancement in going from acetonitrile to water for these
substrates. It should be noted that the hydrogen on the hydroxyl
radical (not involved in the reaction) bears substantial positive
charge in the reactant, transition state, and product. Although
this hydrogen can also participate in hydrogen bonding, this
interaction does not affect the relative energies because the
charge on this hydrogen remains constant in the progression
from reactant to transition state to product.
For alcohols and ethers, where H-abstraction also occurs at
the R-carbon, we hypothesize that the electron-withdrawing
properties of oxygen in the substrate (manifested through an
inductive effect) are offset by the resonance stabilization
afforded by the lone pair of electrons (Scheme 2). This
resonance effect would thus allow the oxygen of the hydroxyl
radical to become highly negative so that hydrogen bonding
interactions would again become important. Presumably the
effect of oxygen diminishes with distance so that the ꢀ-hydro-
gens (and beyond) are aliphatic in nature, and the solvent effect
on their reactivity is similar to the alkanes.
(
4πε0 RT
rts
Using the dipole moments and radii for CH4, HO•, and (CH4/
HO•)q obtained from CCSD(T)/aug-cc-pVDZ calculations based
on MP2/aug-cc-pVDZ geometries (Vide infra), the rate is
actually predicted to be slightly greater in acetonitrile than water
(kCH CN/kH 0 ) 1.01). This result makes sense because one of
3
2
the reactants (hydroxyl radical) possesses a significant dipole
moment; the transition state has a slightly lower dipole moment
and a larger radius. The increased rate in water thus cannot arise
simply because it is a more polar solvent than acetonitrile, but
rather because it is able to stabilize the developing negative
charge on the hydroxyl radical in the transition state by acting
as a hydrogen bond donor. It should also be noted that the results
obtained in Freon 113 add further support to this hypothesis
because (a) the rate constants are the same as in acetonitrile,
and (b) Freon 113 has a substantially lower dielectric constant
than acetonitrile.
To further assess the contribution of dipolar interactions to
the observed rate enhancement in water, electronic-structure
theory calculations of the HO• + CH4 f H2O + CH3• reaction
barrier were calculated using implicit solvation models. CCSD(T)/
aug-cc-pVDZ calculations with the polarized continuum model
(PCM) using geometries and frequencies calculated at the MP2/
aug-cc-pVDZ level indicate that the reaction in water should
be slower than in acetonitrile or the gas phase. In effect, the
barrier for the PCM calculations using water as a solvent in the
PCM model (6.76 kcal/mol) is larger than the barrier when using
acetonitrile as a solvent (5.86 kcal/mol). Both barriers are larger
than the gas-phase barrier at that level (5.03 kcal/mol). These
predictions (based upon the PCM model, which does not account
for implicit hydrogen bonding) are in stark contrast with the
experimental results in water solvent. This “disparity” is easily
To test these hypotheses, and assess how atomic charges on
individual atoms vary in the progression from reactants f
transition state f product, molecular orbital calculations were
performed on the pertinent species for the reactions (a) CH4 +
HO• f CH3• + H2O, (b) Cl3CH + HO• f Cl3C• + H2O, and
(c) HOCH3 + HO• f HOCH2• + H2O at various levels of
theory. In Figure 2, the charges obtained from a natural
population analysis at the MP2(full)/aug-cc-pVQZ//UHF/6-
9
2910 J. AM. CHEM. SOC. VOL. 132, NO. 9, 2010

Modulation of Hydroxyl Radical Reactivity by Solvent
A R T I C L E S
Figure 1. Formation of a polarized transition state for hydrogen atom abstraction from a hydrocarbon by hydroxyl radical. When water is the solvent, the
free energy of the transition state is lowered relative to that of the reactants because of hydrogen bonding of water to the developing negative charge on the
hydroxyl moiety.
Scheme 2
As a final confirmation of this interpretation, the effect of
solvent water molecules on the minimum-energy reaction path
of the HO• + CH4 f H2O + CH3• reaction was calculated.
Geometry optimizations and frequency calculations were con-
ducted at the MP2/aug-cc-pVDZ level, and energies were refined
at the CCSD(T)/aug-cc-pVDZ level. Qualitatively, when water
molecules were added to the calculations for the HO•/CH4
reaction, the resulting transition state clearly shows that the water
molecules properly align so as to stabilize the developing
charges as hypothesized. Quantitatively, the magnitude of the
stabilization is consistent with the magnitude of the observed
kinetic solvent effect.
reconciled if the transition state for hydrogen atom abstraction
is stabilized by hydrogen bonding in water solvent.
Table 3 shows the calculated barriers and calculated relative
rate constants for HO• + CH4 f H2O + CH3• with various
water molecules hydrogen bonding to different sites of the HO•/
CH4 system either as donors or acceptors. One water molecule
acting as hydrogen-bond donor to the OH radical reduces the
barrier by 1.52 kcal mol-1. On the other hand, if the solvating
water molecule acts as hydrogen-bond acceptor, the barrier
increases by 0.32 kcal mol-1. If one water molecule acts as
hydrogen-bond donor and another one as hydrogen-bond
acceptor, the barrier decreases by 1.14 kcal mol-1. This decrease
is almost a perfect balance between the 1.52 kcal mol-1 decrease
and 0.32 kcal mol-1 increase in the barriers of the transition
states with only one molecule acting as donor or acceptor,
respectively, suggesting that the effect of individual water
molecules might be additive. Attempts to study the reaction in
Figure 2. Atomic charges for the reaction of hydroxyl radical with CH4,
CH3OH, and CHCl3 obtained from natural population analysis of the
reactants, transition states, and products at the MP2(full)/aug-cc-pVQZ//
UHF/6-311G* levels.
9
J. AM. CHEM. SOC. VOL. 132, NO. 9, 2010 2911

A R T I C L E S
Mitroka et al.
Table 3. Barriers of the HO• + CH4 f H2O + CH3• Reaction with Various Levels of Solvationa
a Energies calculated at the CCSD(T)/aug-cc-pVDZ level using geometries and harmonic frequencies obtained at the MP2/aug-cc-pVDZ level. The
barriers correspond to enthalpies of activation at 0 K.
which two water molecules act as hydrogen-bond donors failed
to locate the appropriate reagents conformation. In effect, even
though a transition state in which two water molecules are acting
as hydrogen-bond donors was located, optimizations of the OH
radical with two water molecules acting as hydrogen-bond
donors always led to the isomer in which one water molecule
acts as hydrogen-bond donor and the other is an acceptor.
Finally, the barrier of the HO• + CH4 f H2O + CH3•
reaction, with two water molecules acting as hydrogen bond
donors and one water molecule acting as hydrogen bond
acceptor, was calculated. The decrease in the barrier for this
solvation level with respect to the unsolvated system is 2.37
kcal mol-1 with a relative rate enhancement consistent with
experimental results. Various attempts to include additional
water molecules that are in direct contact with the HO• at the
transition state led to solvation of any of the three water
molecules forming the first solvation sphere, as expected.
the addition of HO• to aromatic substrates to explain a nearly
2 orders of magnitude diminution in rate in acetonitrile compared
to water. For these additions, calculations suggested that the
hydroxyl moiety was nearly anionic in the transition state,
providing a clear opportunity for water stabilization via hydrogen
bonding. The effect water has on modulating HO• reactivity
has enormous implications. In biological systems, this means
that HO• may be less reactive in the hydrophobic regions of a
cell than previously believed, i.e., hydroxyl radical does not
necessarily react with the first molecule it encounters. Indeed,
Nature may use this as a containment strategy when hydroxyl
radical is produced naturally within cells.
On the basis of rate measurements in water, reactions of HO•
were generally thought to be diffusion-controlled. Consequently,
it has been assumed that attempts to protect against the damaging
effects of HO• through the use of antioxidants are doomed to
failure because a successful defense strategy requires a degree
of selectivity to be demonstrated by the reactive radical. The
radical must react preferentially with antioxidant, rather than
the molecule or materials that one is trying to protect, in order
for any defense system to be effective. Our results demonstrate
that such selectivity might be attainable in a hydrophobic
environment.
Conclusions
The hypothesis that the hydroxyl radical reacts through a polar
transition state is fully supported by the experimental and
theoretical data obtained in this study. As a consequence of this
polarization, hydrogen bonding to water stabilizes the transition
state, resulting in larger rate constants when water is the solvent.
This explanation also predicts that the magnitude of the solvent
effect will decrease when electronegative substituents are present
on the R-carbon of the substrate. The solvent effect is also
significant when the substrate possesses electron-donating
substituents, presumably because of competing inductive (electron-
withdrawing) and resonance (electron-donating) effects.
These results extend the observations of Platz and co-work-
ers,8b,c who argued for a highly polarized transition state for
Finally, it should be stressed that the ideas presented herein
provide a molecular-level understanding of hydroxyl radical
reactivity both in solution and the gas phase.
Acknowledgment. This work was supported by the National
Science Foundation (CHE-0548129, CHE-0547543) and in part by
the Macromolecular Interfaces with Life Sciences (MILES) Integra-
tive Graduate Education and Research Traineeship (IGERT) of the
National Science Foundation under Agreement No. DGE-0333378
9
2912 J. AM. CHEM. SOC. VOL. 132, NO. 9, 2010

Modulation of Hydroxyl Radical Reactivity by Solvent
A R T I C L E S
in the form of fellowships to S.M. and S.Z.D.T. is a Cottrell Scholar
of Research Corporation. We thank Profs. T. Daniel Crawford and
Naushadalli K. Suleman for helpful discussions.
traces, kobs versus concentration plots, the absolute energies and
optimized geometries of all calculated structures, and the
complete Gaussian 03 citation. This material is available free
of charge via the Internet at http://pubs.acs.org.
Supporting Information Available: Discussion of the role of
•CH2CN and PyrS• in these experiments, details of the product
studies and competition experiments, representative transient
JA903856T
9
J. AM. CHEM. SOC. VOL. 132, NO. 9, 2010 2913