Kinetic isotope effect for hydrogen abstraction by .OH radicals from normal and carbon-deuterated ethyl alcohol and methylamine in aqueous solutions

Article abstract of DOI:10.1021/jp027790e

Selective H/D kinetic isotope effects (KIEs) have been determined for the various routes by which ^.OH radicals react with ethanol (CH₃CH₂OH and CD₃CD₂OH) and methylamine (CH₃NH₂ and CD₃NH₂) in H₂ Osolutions. The KIEs have been evaluated from overall rate constants and the yields in which the individual primary radicals were generated in the ^.OH reaction with these compounds. The analytical method applied for the yield determinations was redox titration with suitable scavengers, namely, methyl viologen and Fe(CN)₆^3- for the reducing radicals (CH₃C^.HOH/CD₃C^.DOH and ^.CH₂NH₂/^.CD₂NH₂ ) and I^- and hydroquinone for the oxidizing radicals (CH₃CH₂O^./CD₃CD₂O^. and ^.NHCH₃/^.NHCD₃). The numerical results obtained also include, besides yields relative to total available ^.OH, absolute rate constants for most of these scavenging reactions. For the alcohol, the major process (almost 90%) is H/D abstraction from the Ca bond with KIE = 1.96. For methylamine, abstraction from C_?± H/D occurs with only 37% (H) and 26% (D) but at a similar KIE = 1.86. The remainder, denoting the major process in this case, accounts for the formation of aminyl radicals. The secondary KIEs for N-H cleavage, and O-H cleavage in the case of the alcohol, are close to unity, reflecting the expected negligible influence of C-H/D substitution in the attached alkyl groups. Abstraction of ?2-C-attached H/D in ethanol shows, also in qualitative agreement with expectation, a larger KIE = 3.4 than that for ?±-C H/D. An interestingly high KIE a?? 50 was obtained for the 1,2 hydrogen shift that converts the CH₃CH₂O^. and CD₃CD₂O^. oxyl radicals into the corresponding ?±-hydroxyethyl radicals CH₃C^.HOH and CD₃C^.DOH. All of the results are discussed in light of existing literature data on relevant H/D isotope effects, the influence of solvent relative to gas phase, the selectivity of the ^.OH attack, and other mechanistic considerations. Specifically, the mechanism by which the ^.OH reacts with the amine seems likely to involve a transient, caged aminium/hydroxide ion pair.

Full text of DOI:10.1021/jp027790e

7268
J. Phys. Chem. B 2003, 107, 7268-7276
•
Kinetic Isotope Effect for Hydrogen Abstraction by OH Radicals from Normal and
Carbon-Deuterated Ethyl Alcohol and Methylamine in Aqueous Solutions†
Marija Bonifacˇic´,*,‡ David A. Armstrong,§ Igor Sˇtefanic´,| and Klaus-Dieter Asmus
i
Department of Physical Chemistry, Ruder BosˇkoVic´ Institute, Bijenicˇka c. 54, 10000 Zagreb, Croatia,
Department of Chemistry, UniVersity of Calgary, Calgary, Alberta, Canada T2N 1N4, CulVer Academies,
1300 Academy Road, CulVer, Indiana 46511, and Faculty of Chemistry, Adam-Mickiewicz-UniVersity,
Grunwaldzka 6, 60-780 Poznan, Poland
ReceiVed: December 20, 2002; In Final Form: February 27, 2003
•
Selective H/D kinetic isotope effects (KIEs) have been determined for the various routes by which OH
radicals react with ethanol (CH3CH2OH and CD3CD2OH) and methylamine (CH3NH2 and CD3NH2) in H2O
solutions. The KIEs have been evaluated from overall rate constants and the yields in which the individual
•
primary radicals were generated in the OH reaction with these compounds. The analytical method applied
for the yield determinations was redox titration with suitable scavengers, namely, methyl viologen and
Fe(CN)63- for the reducing radicals (CH3C•HOH/CD3C•DOH and •CH2NH2/•CD2NH2) and I- and hydroquinone
for the oxidizing radicals (CH3CH2O•/CD3CD2O• and •NHCH3/•NHCD3). The numerical results obtained also
•
include, besides yields relative to total available OH, absolute rate constants for most of these scavenging
reactions. For the alcohol, the major process (almost 90%) is H/D abstraction from the CR bond with KIE )
1.96. For methylamine, abstraction from CR H/D occurs with only 37% (H) and 26% (D) but at a similar KIE
) 1.86. The remainder, denoting the major process in this case, accounts for the formation of aminyl radicals.
The secondary KIEs for N-H cleavage, and O-H cleavage in the case of the alcohol, are close to unity,
reflecting the expected negligible influence of C-H/D substitution in the attached alkyl groups. Abstraction
of â-C-attached H/D in ethanol shows, also in qualitative agreement with expectation, a larger KIE ) 3.4
than that for R-C H/D. An interestingly high KIE ≈ 50 was obtained for the 1,2 hydrogen shift that converts
the CH3CH2O• and CD3CD2O• oxyl radicals into the corresponding R-hydroxyethyl radicals CH3C•HOH and
CD3C•DOH. All of the results are discussed in light of existing literature data on relevant H/D isotope effects,
•
the influence of solvent relative to gas phase, the selectivity of the OH attack, and other mechanistic
considerations. Specifically, the mechanism by which the •OH reacts with the amine seems likely to involve
a transient, caged aminium/hydroxide ion pair.
Introduction
and there is now substantial evidence that ET competes with H
abstraction in substrates with parent ion reduction potentials
Hydroxyl radicals (•OH) are formed as short-lived intermedi-
ates in several different processes. These include radiolysis and
photolysis and Fenton and Fenton-like reactions as well as
reactions of NOx species. The hydroxyl radicals are thus
produced in atmospheric and in metabolic processes in living
organisms. •OH is one of the most aggressive oxidative species
known and is implicated in oxidative damage in living cells.
Its reactions with organic substrates typically occur by addition,
H-atom abstraction, or electron transfer (ET). The large HO-H
bond-dissociation enthalpy of water ensures that H abstraction
is generally exothermic. Thus, activation energies for H-
abstraction reactions are low, and the rates are relatively rapid.
In terms of ET, the small size of •OH means that its reorganiza-
tion energy is expected to be large, and this process is therefore
likely to compete only with H abstraction in cases were the
difference in reduction potentials is appreciable. However, the
reduction potential is also quite high (E°(•OH/OH-) ) 1.9 V 1),
e1.5 V. Nitrogen-containing compounds (NR), which have
E°(+•NR/NR) values below this, are typical cases where this
may occur, trimethylamine being a case in point.2
One of the problems relating to the elucidation of mechanisms
•
of reactions of OH is that the H abstraction and ET processes
are often difficult to differentiate. This arises from the fact that
electron transfer can be followed by the rapid loss of a proton
from the one-electron-oxidized substrate, thus giving the same
product as H-atom abstraction.2 Such a mechanistic problem
prevailed, for example, in the oxidation of simple amino acids
in their anionic form where several different radical products
•
have been observed as result of the OH reaction (e.g., a total
of six in the case of alanine).3-5
To distinguish between the various routes and mechanisms
by which these radicals were formed in the amino acid system,
time-resolved esr and redox titrations have successfully been
applied.5-7 These pulse radiolysis studies, supported by some
laser flash photolysis, revealed the concurrent action of the •OH
radical as a hydrogen-abstracting agent as well as an electrophile
interacting with the lone electron pair at the amino acid’s
nitrogen. Although much of the overall mechanism could be
disclosed, some of the mechanistic details remained obscured
† Part of the special issue “Arnim Henglein Festschrift”.
* Corresponding author. E-mail: bonifacic@rudjer.irb.hr. Tel: 385-1-
4680093. Fax: 385-1-4680245.
‡ Rudher Bosˇkovic´ Institute.
§ University of Calgary.
| Culver Academies.
Adam-Mickiewicz-University.
10.1021/jp027790e CCC: $25.00 © 2003 American Chemical Society
Published on Web 04/15/2003

•
H/D Abstraction by OH from Ethanol and Methylamine
J. Phys. Chem. B, Vol. 107, No. 30, 2003 7269
and called for additional support through, for example, H/D
isotope effect studies. However, it became quite clear during
the early stages of such investigations that the complexity of
the amino acid system would still leave certain ambiguities. For
quantitative assignments and interpretations, it seemed manda-
tory to first obtain reliable and unambiguous data from even
simpler systems.
In view of all of this, we therefore decided to conduct a series
of experiments on H/D isotope effects concerning the reaction
of •OH with simple alcohols and amines. The reactions in focus
are eqs 1 and 2 for ethyl alcohol and 3 and 4 for methylamine:
Concerning either mechanistic possibility, H/D isotope effects
should provide further insight. Although we are not aware of
any study on the effect of the deuteration of aliphatic amines,
overall or total rate constants for abstractions from all possible
sites are available for the reactions of hydroxyl radicals with
CH3CH2OH11 and CD3CD2OH.12 However, just taking the ratio
of these two rate constants would not be sufficient for the present
purposes because they include reactions 11 and 12, which take
place simultaneously with reaction 1.13 A very similar situation
can be expected for the deuterated molecule, reactions 13 and
14, but again, no determination of individual products has been
reported.
•
•OH + CH3CH2OH f CH3 CHOH + H2O
(1)
(2)
(3)
(4)
•OH + CH3CH2OH f •CH2CH2OH + H2O
•OH + CH3CH2OH f CH3CH2O• + H2O
(11)
(12)
•OH + CD3CD2OH f CD3 CDOH + HDO
•
•OH + CH3NH2 f •CH2NH2 + H2O
•OH + CD3NH2 f •CD2NH2 + HDO
•OH + CD3CD2OH f •CD2CD2OH + HDO (13)
•OH + CD3CD2OH f CD3CD2O• + H2O
(14)
In addition to an experimental determination of absolute rate
constants and yields, we will briefly discuss some attempts to
put our results, particularly those referring to isotope effects,
into a theory-based perspective.
Finally, the results presented here for simple primary alcohol
and amine molecules could also be of relevance for the
elucidation of processes in biological or other organic systems.
For example, the H/D isotope effect has recently been applied
in an attempt to identify hydroxyl radical-like species within
the metal-binding site of a protein.14
For ethyl alcohol, such reactions can proceed only via the
hydrogen atom abstraction mechanism, and the alternative
electron/proton-transfer mechanism can be excluded. In contrast,
for methylamine, both mechanisms are possible: (i) a direct
hydrogen abstraction as indicated by reactions 3 and 4 and (ii)
an electron transfer from nitrogen leading to the N-centered
aminium radical cation, reaction 5, followed by a rapid loss of
H+/D+ from the neighboring carbon atom, reaction 6 or 7.
•OH + NH2CH3/NH2CD3 f +•NH2CH3/+•NH2CD3 + OH-
(5)
Experimental Section
+•NH2CH3 f •CH2NH2 + H+
+•NH2CD3 f •CD2NH2 + D+
(6)
(7)
All investigations have been conducted with aqueous solutions
with water purified by the Serv-A-Pure Co. system. Absolute
ethyl alcohol (200 proof, Aaper Alcohol and Chemical Com-
pany), CD3CD2OD (Aldrich, anhydrous, 99+ atom % D),
methylamine (Aldrich, 40 wt % solution in water), methylamine-
d5 deuteriochloride (Aldrich, 98+ atom % D), and other
chemicals (Aldrich, Fluka) were used as received from the
vendors.
The latter mechanism gains in likelihood in view of the relatively
low reduction potential of the general couple (+•NH2-R/NH2-
R), E° ≈1.3 V.8 It also should be recognized in this context
that the aminium cation is formed together with a strong base,
OH-. Effective deprotonation should thus already be facilitated
within the initial solvent cage. Deprotonation of an aminium
radical cation with the loss of a proton from CR-H has been
Pulse radiolysis was performed with an 8-MeV Titan Beta
model TBS-8/16-1S LINAC at the Notre Dame Radiation
Laboratory with pulses of 2.5-ns duration and doses per pulse
in the range of 2-10 Gy. A description of the pulse radiolysis
setup, data collection, and processing can be found elsewhere.15
All solutions were freshly prepared before each experiment.
Solutions containing deuterated compounds were always used
within a period of 4 h after preparation of the solution. This
ensured that there was no exchange of carbon-bound D atoms
with H atoms from water. Oxygen- and nitrogen-bound deute-
rium, however, suffered fast H/D exchange (i.e., CD3CD2OD
and CD3ND2‚DCl were present as CD3CD2OH and a mixture
shown to occur in the case of +•N(CH3)3 (and it also occurs
2
in the oxidative C-H activation of certain hydrocarbons9).
However, the proton loss is expected to occur more readily from
the acidic N-H position and to result in the respective aminyl
radicals, reactions 8 and 9.
+•NH2CH3 h •NHCH3 + H+
+•NH2CD3 h •NHCD3 + H+
(8)
(9)
+
of CD3NH2/CD3NH3 at the pH ranges applied in this study).
Solutions were generally deaerated (N2) and saturated with N2O.
Under these conditions, approximately 90% of all primary
species available for reaction with solutes were OH radicals.
The same aminyl radicals may, however, also be produced by
direct H-atom abstraction, reaction 10.
•
The remaining ∼10% are H• atoms, and for their yield, a value
of G(H•) ) 0.06 µmol J-1 was used throughout the study,
independent of the solute concentrations.16 The yield of scav-
engable hydroxyl radicals, G(•OH), was calculated by taking
into account the total scavenging capacity in each solution
(concentration multiplied by rate constant for all substrates
reacting with hydroxyl radicals) according to a generally
•OH + NH2CH3/NH2CD3 f •NHCH3/•NHCD3 + H2O
(10)
Both the C-centered •CH2NH2 and N-centered •NHCH3 radicals
have, in fact, been proposed as products of the reaction of
hydroxyl radicals with CH3NH2 in basic aqueous solution10
without, however, a quantitative determination of their yields.
•
applicable formula.17 The total concentration of OH radicals

7270 J. Phys. Chem. B, Vol. 107, No. 30, 2003
Bonifacˇic´ et al.
per pulse applied in our present investigation of N2O-saturated
systems amounts to about 1-6 µM. Dosimetry was performed
with thiocyanate solutions as described earlier.18 All experiments
have been conducted at room temperature.
The accuracy of radiation chemical experiments is generally
considered to be about (10%. This also applies to our present
set of data. Error limits given for the specific numerical values
refer only to the standard deviation of the mean of a series of
single measurements taken (usually three for each individual
system/concentration).
Results and Discussion
Overall Reaction Rate Constants. Because none of the
radicals involved in the reactions studied (i.e., hydroxyl or
radicals derived from ethyl alcohol or methylamine) shows any
significant optical absorption in the range suitable for accurate
direct measurements of their kinetics, the rate constants were
determined using the competition kinetics method. The reference
reactant was the ferrocyanide ion, Fe(CN)64-, which is oxidized
Figure 1. Ratio of optical densities at 420 nm in the absence (OD0)
and presence of ethyl alcohols (OD) versus the reacting substrates
concentration ratio obtained in pulse-irradiated, N2O-saturated aqueous
•
by OH radicals with a rate constant of k17 ) 1.0 × 1010 M-1
4-
s-1 11
.
solutions containing 1 mM Fe(CN)6 and various amounts of CH3-
CH2OH (O) or CD3CD2OH (4). Dose per pulse: 3 Gy. Experimentally
obtained values are corrected for the amount of scavengable hydroxyl
radicals for each solution. (For details, see text.)
•OH + CH3CH2OH f products
•OH + CD3CD2OH f products
(15)
(16)
(17)
TABLE 1: Relative Yield of Radicals Formed in Reactions
•
with OH
•OH + Fe(CN)64- f Fe(CN)63- + OH-
radical
(R-C)
radical
(â-C)
radical
(O/N)
compound
Grel
Grel
Grel
CH3CH2OH CH3 CHOH 0.89 •CH2CH2OH 0.07 CH3CH2O• 0.04
•
3-
Ethyl Alcohol. The yields of Fe(CN)6 generated via
reaction 17 in neutral, N2O-saturated aqueous solution contain-
ing 1 mM Fe(CN)64- were determined by measuring its optical
absorption at 420 nm in the absence (OD0) and in the presence
(OD) of various amounts of normal or deuterated ethyl alcohol
(1-13 mM). Absorption intensities were measured only during
the few first microseconds after the pulse to allow all of the
primary reactions (i.e., 15 or 16 and 17) to be completed while
•
CD3CD2OH CD3 CDOH 0.88 •CD2CD2OH 0.04 CD3CD2O• 0.08
CH3NH2
CD3NH2
•CH2NH2
•CD2NH2
0.37
0.26
•NHCH3
•NHCD3
0.63
0.74
TABLE 2: Total and Partial Rate Constants for the Attack
•
of OH on Specific Atoms and Kinetic Isotope Effects for
H/D Abstractiona
ktotal
kR-C
kâ-C
0.17
0.05
kO/N
3-
still avoiding possible second-order back reduction of Fe(CN)6
CH3CH2OH 2.40
CD3CD2OH 1.24
2.14
1.09
0.10
0.10
via the simultaneously formed R-hydoxyalkyl radicals. The
possible error caused by the latter reaction was also kept to a
minimum because of the relatively low dose per pulse (ca. 3
Gy).
kH/kD ) 1.93 kH/kD ) 1.96 kH/kD ) 3.4 kH/kD ) 1.0
CH3NH2
CD3NH2
3.60
2.75
1.33
0.715
2.27
2.04
kH/kD ) 1.11
The overall rate constants k15 ) (2.40 ( 0.06) × 109 M-1
s-1 and k16 ) (1.24 ( 0.03) × 109 M-1 s-1 for the •OH reactions
with CH3CH2OH and CD3CD2OH, respectively, have been
calculated from the slopes of the straight lines obtained by
plotting ([OD0]/[OD])corr versus [alcohol]/[Fe(CN)64-] according
to expression I.
kH/kD ) 1.31 kH/kD ) 1.86
a Units: 109 M-1 s-1
.
) 1.9 × 109 M-1 s-1 and k16 ) 1.2 × 109 M-1 s-1 have been
used. Applying the somewhat higher rate constant k15 obtained
in the present study did not lead to any significant change. It is
further noted that for the range of scavenger concentrations
applied in this study the corrections remained small (i.e., the
rate constants obtained with the corrected values were only
5-7% higher than those from uncorrected data).
[OD0]
[OD]
k15,16
[alcohol]
) 1 +
(I)
4-
(
)
k17
(
)
[Fe(CN)6
]
corr
where([OD0]/[OD])corr ) ([OD0]/[OD])measured × (G(•OH)alcohol
/
For CD3CD2OH, the overall rate constant obtained in this
work is identical to that published earlier.12 The agreement is
not as good for the fully hydrogenated CH3CH2OH. In fact,
the rate constant for reaction 15 has been measured repeatedly
by different research groups with reported values between (1.6-
2.2) × 109 M-1 s-1, the recommended value being 1.9 × 109
G(•OH)0). G(•OH)0 is the calculated yield of scavengable
hydroxyl radicals for the solution containing only 1 mM
Fe(CN)64-, and G(•OH)alcohol refers to solutions containing, in
addition, various concentrations of ethyl alcohol.
The corresponding plots are shown in Figure 1, and the rate
constants are listed in the second column of Table 2. The
correction of the measured [OD0]/[OD] ratios was necessary
because the yield of hydroxyl radicals available for reactions is
not constant but increases with the total scavenger capacity of
substrates (alcohol plus Fe(CN)64-) in the solution.16,17 For the
first calculation of the respective G(•OH) available for reaction
in each solution, the originally reported rate constants11 of k15
M-1 s-1 11
An inspection of conditions and procedures in the
.
original publications revealed that probably no corrections had
been made for the change of G(•OH) at different scavengers
concentrations. Such corrections would, however, result in up
to 10% higher rate constants, which would bring the earlier
reported values of k15 very close to the value determined in
this study.

•
H/D Abstraction by OH from Ethanol and Methylamine
J. Phys. Chem. B, Vol. 107, No. 30, 2003 7271
Methylamine. Basically, the same procedure as described
above was applied for the evaluation of the corresponding
•
overall rate constants k18 and k19 for the reaction of OH with
normal and deuterated methylamines. Experiments were, how-
ever, performed at pH 10.8-11.0. This pH range was selected
to push equilibrium 20 as much as possible to the right-hand
side while still not allowing appreciable dissociation of the
•OH radicals (pKa(•OH) ) 11.9). Also, only the actual concen-
trations of unprotonated amine, as calculated from the amine’s
pKa and the pH of the solution, were used for the determination
of the rate constants. (For both the normal and C-deuterated
methylamine, a pKa of 10.65 was employed.) The reaction of
hydroxyl radicals with protonated amine, CH3NH3+, can be
neglected at these high pH values. Because of its low rate
constant (k ) 3.5 × 107 M-1 s-1 19), this reaction contributes
less than 1% to the overall rate. The same was assumed to apply
to the deuterated amine cation, CD3NH3+, for which a rate
Figure 2. Formation of MV•+ as a function of time measured at 600
nm in a pulse-irradiated N2O-saturated aqueous solution containing 0.1
M CD3CD2OH and 0.5 mM MV2+. Optical absorption intensity is
expressed as the yield of MV•+ (G/µmol J-1). Dose per pulse: 1.4 Gy.
•
constant with OH radicals has neither been reported in the
literature nor determined in this work.
•OH + CH3NH2 f products
•OH + CD3NH2 f products
CH3NH3+ h CH3NH2 + H+
(18)
(19)
(20)
on such titrations but identified their yields simply as the
•
difference between the total OH/H• yield and the sum of the
radicals that were quantitatively accessible via redox titration.
For yields and rate constants of the oxidizing radicals, two
different reactions were applied. The titration of the aminyl
radicals was performed using hydroquinone (E0 ) 0.023 V for
the Q•-/Q2- couple)21 (eq 25/26).3-5
Actual measurements were conducted in N2O-saturated
solutions containing 0.5 mM Fe(CN)64- and various concentra-
tion of CH3NH2 (0.2-1.3 mM) or CD3NH2 (0.4-3.5 mM). The
respective rate constants were determined from plots similar to
•NHCH3/•NHCD3 + HQ- f NH2CH3/NH2CD3 + Q•-
(25/26)
the ones shown in Figure 1 and were calculated to be k18
)
(3.6 ( 0.1) × 109 M-1 s-1 and k19 ) (2.75 ( 0.06) × 109 M-1
s-1. They are also listed in the second column of Table 2. Only
the value for k18 can be compared with previously reported data
The analysis was based on the formation kinetics and yields of
the semiquinone radical anion, Q•- (λmax ) 427 nm, ꢀ ) 7200
M-1 cm-1). Actual measurements were performed at pH 11,
where the hydroquinone exists in all three possible protonation
states, H2Q, HQ-, and Q2-, in the ratio of 6:72:23, respectively.
Rate constants determined this way, therefore, represent a
weighted value for this composition.
The alkoxyl radicals were analyzed by applying iodide as a
selective scavenger.13 In this case, the overall reaction sequence
(eqs 27/28 and 29) leads to the formation of the optically
detectable I2•- radical anion (λmax ) 380 nm; ꢀ380 ) 9400 M-1
cm-1).22
that range from (1.8-5.7) × 109 M-1 s-1 10,20
.
Yields and Individual Rate Constants. The yields of the
individual radicals produced in the overall oxidations of the
alcohols and amines (reactions 1-14) were determined by the
same selective redox titration that was applied in related previous
studies of ours on the •OH-induced oxidation of aliphatic amino
acids.3,5 Thus, the strong reducing character of all C-centered
R-hydroxy and R-amino radicals allowed their convenient
titration with methyl viologen (MV2+, E0 ) -0.45 V for the
MV2+/MV•+ couple),1 as stoichiometrically outlined in the
general reactions 20-24. Yields were quantitatively determined
by measurements of the optical absorption of MV•+ at 600 nm
CH3CH2O•/CD3CD2O• + I- f
and using ꢀ600 ) 12 820 M-1 cm-1 3
. Rate constants were
CH3CH2O-/CD3CD2O- + I• (27/28)
evaluated from the formation kinetics of MV•+ at varying MV2+
concentrations. Whenever possible, the latter were kept at e0.5
mM to avoid the direct reduction of MV2+ by hydrated electrons.
(For respective rate constants, see ref 11.)
•-
I• + I- h I2
(29)
The actual results obtained concerning all of the radicals of
interest are presented and discussed in the following sections
under the headlines of their respective molecular precursors.
C-Deuterated Ethyl Alcohol. Figure 2 shows a typical trace
obtained at 600 nm in pulse-irradiated N2O-saturated solution
•
•
CH3 CHOH/CD3 CDOH + MV2+
f
MV•+ + CH3CHO/CD3CDO + H+ (21/22)
containing, in this case, 0.11 M CD3CD2OH and 0.5 mM MV2+
.
•CH2NH2/•CD2NH2 + MV2+
f
The signal intensity due to the optical absorption of MV•+ has
been normalized for the dose and the MV•+ molar extinction
coefficient. This allows one to display the optical signal directly
in terms of an absolute yield, G(MV•+)/µmol J-1, as a function
of time. Two kinetically distinguishable processes can be
realized: the majority of the MV•+ formation has been
completed within a few microseconds, and an additional
generation occurs on a longer time scale (∼50 µs). The
combined total yield amounts to G(MV•+) ) (0.682 ( 0.010)
MV•+ + CH2NH/CD2NH + H+ (23/24)
The less-reducing â-hydroxyalkyl radicals generated in reac-
tions 11 and 13 (i.e., •CH2CH2OH and •CD2CD2OH) cannot be
quantified via a corresponding reaction with MV2+. Their
3-
oxidation affords a more potent redox partner such as Fe(CN)6
However, these reactions are very slow, hampering an unam-
biguous quantitative determination. We therefore did not embark
.

7272 J. Phys. Chem. B, Vol. 107, No. 30, 2003
Bonifacˇic´ et al.
µmol J-1. It did not change (within indicated error limit) when
the MV2+ concentration was lowered to 0.1 or 0.3 mM. Under
the experimental conditions, the only radicals available for
TABLE 3: Rate Constants for the Various Redox Titration
and Rearrangement Reactions
reaction
rate constant
-
reaction with CD3CD2OH were •OH and H• (conversion of eaq
CH3C•HOH + MV2+ f MV•+ + . . . .
(2.1 ( 0.1) × 109 M-1 s-1
•
CD3C•DOH + MV2+ f MV•+ + . . . .
(2.4 ( 0.2) × 109 M-1 s-1
to OH via reaction with N2O occurs with t1/2 ≈ 3 ns). The
•
absolute yield of scavengable OH radicals for this alcohol
•CH2NH2 + MV2+ f MV•+ + . . . .
•CD2NH2 + MV2+ f MV•+ + . . . .
(3.6 ( 0.3) × 109 M-1 s-1
3.6 × 109 M-1 s-1
concentration was calculated to 0.647 µmol J-1 17 For H• atoms,
.
G ) 0.06 µmol J-1 was taken.
CH3CH2O• + I- f I• + . . . .
4 × 109 M-1 s-1a
Hydrogen atoms are less reactive than •OH radicals and thus
most likely more selective. Accordingly, they should preferen-
tially undergo a H/D abstraction reaction from the most labile
CH3N•H + HQ- f Q•- + . . . .
(1.1 ( 0.1) × 108 M-1 s-1
CH3CH2O• f CH3C•HOH
CD3CD2O• f CD3C•DOD
CD3CD2O• + CD3CD2OH f
CD3CD2OD + CD3C•DOH
(5 ( 2) × 106 s-1
(1.1 ( 0.1) × 105 s-1
2 × 105 M-1 s-1a
•
R-C position in the alcohol, reaction 30, producing CD3 CDOH
as the major radical. An abstraction from the â position will
presumably be negligible considering that the rate constant of
this reaction is as low as k30 ) 2.8 × 106 M-1 s-1 23
.
(See ref
a Estimate.
44.) The overall reaction of the •OH radical (eq 16a), however,
occurs almost 3 orders of magnitude faster (k16 ) 1.24 × 109
M-1 s-1, see also Table 2).
anism includes, as the first step, the formation of alkoxyl radicals
CD3CD2O• (eq 14). These are oxidizing species and therefore
not reactive toward MV2+. But they are known to undergo an
intramolecular 1,2 hydrogen shift readily (possibly solvent-
assisted), thereby being transformed into the MV2+-reducing
H• + CD3CD2OH f CD3 CDOH + HD
•OH + CD3CD2OH f product radicals +
(30)
•
•
CD3 CDOD radicals (eq 31). The same result would be achieved
by a second-order hydrogen-abstraction reaction from the parent
molecule (eq 32).16 (In the H2O solvent. both of the -OD
functional groups formed in eqs 31 and 32 will, of course,
undergo fast D f H exchange.)
HDO/H2O (16a)
The total yield of MV•+ (i.e., the yield of MV2+ reducing
•
radicals) due to the reactions of OH radicals alone with CD3-
CD2OH accordingly amounts to
•
CD3CD2O• f CD3 CDOD
(31)
G(MV•+)•OH ) G(MV•+)total - G(H•) ) 0.68 - 0.06 )
0.62 µmol J-1
•
CD3CD2O• + CD3CD2OH f CD3CD2OD + CD3 CDOH
(32)
or in relative units
The overall first-order rate constant of (1.1 ( 0.1) × 105 s-1
noted above for the slow process therefore corresponds to (k31
+ k32[CD3CD2OH]). We did not make an attempt to evaluate
the individual values for k31 and k32 by an extra set of
experiments at different CD3CD2OH concentrations. In view
of some published rate constants for similar reactions, namely,
G(MV•+)•OH
G(•OH )
0.62
0.647
G(MV•+)•OH, relative
)
)
) 0.96.
In other words, 96% of the hydroxyl radicals reacting with
CD3CD2OH end up as radicals capable of reducing MV2+. The
remainder of 4% (i.e., of radicals that are not able to reduce
MV2+) is ascribed to the formation of â-hydroxy-C-centered
•
2.6 × 105 M-1 s-1 for CH3O• + CH3OH f CH2OH + CH3-
OH 26 and 1.1 × 106 M-1 s-1 for (CH3)3CO• + CH3CH2OH f
CH3 CHOH + (CH3)3COH),27 the estimated contribution of
•
radicals CD2CD2OH formed by reaction 13. (See ref 45.)
k32[CD3CD2OH] amounts to ∼2 × 104 s-1, leaving reaction 31
as the clearly dominating process. In any case, the rate-
determining step for the slow MV2+ reduction is attributable to
•
•
The reducing species are considered to be CD3 CDOH
radicals. Not all of them are, however, the immediate result of
a D-atom abstraction by the OH radicals at CR (eq 2), as
the conversion of CD3CD2O• to CD3 CDOD, irrespective of the
•
•
becomes apparent upon inspection and analysis of the formation
kinetics of MV•+ displayed in Figure 2. Two processes are
evident. The first is complete in ∼1-2 µs whereas the second
slower process has a first-order growth rate of (1.1 ( 0.1) ×
105 s-1. Only the fast process was dependent on [MV2+] in a
way that is characteristic for a pseudo-first-order reaction. Thus,
actual mechanism.
A quantitative evaluation of the overall small yield of the
slow process was achieved with solutions containing an MV2+
concentration of 2 mM, which allows a clear and unambiguous
distinction of the two processes. The absolute yield of the slow
part reflecting the yield of CD3CD2O• radicals amounted to
(0.049 ( 0.006) µmol J-1. In units relative to G(•OH), this
corresponds to G(CD3CD2O•)relative ) (0.08 ( 0.01). In other
words, about 8% of all •OH radicals end up as alkoxyl radicals.
(The presence of 2 mM MV2+ lowers G(•OH) by up to ca. 5%
because of its competition with N2O for eaq-. This was taken
in account although it did not significantly influence the final
result.)
•
only this fast process can be assigned to the CD3 CDOH radicals
resulting from reaction 2 and their subsequent reaction with
MV2+ (eq 20). The total yield of this contribution (i.e., the yield
•
of directly generated CD3 CDOH radicals) corresponds to G )
•
0.573 µmol J-1 or, on a relative scale, to 88% of the OH
radicals. An analysis of the MV•+ formation kinetics in solutions
of varying MV2+ concentration (0.1-0.5 mM) gave a bi-
molecular rate constant of k22 ) (2.4 ( 0.2) × 109 M-1 s-1
.
Normal Ethyl Alcohol. In a pulse-irradiated N2O-saturated
This and other rate constants for radical-induced titration
reactions are listed in Table 3.
solution containing 0.1 M CH3CH2OH and 0.1-0.4 mM MV2+
,
the formation of MV•+ (traced at 600 nm) occurred by just one
exponential process, the rate of which was proportional to the
MV2+ concentration. For this alcohol, the separation of
The minor but clearly distinguishable slower process in Figure
•
2 is associated with CD3 CDOH radicals formed via a different,
CH3 CHOH from the CH3CH2O• radical, similar to the proce-
•
additional route. It is important to recognize its exponential but
completely [MV2+]-independent kinetics. The assumed mech-
dure described above for CD3CD2OH/ MV2+ system, is not

•
H/D Abstraction by OH from Ethanol and Methylamine
J. Phys. Chem. B, Vol. 107, No. 30, 2003 7273
possible because of the presumably much faster intramolecular
transformation of CH3CH2O• to CH3 CHOH (eq 33). By
•
comparison with some related literature data28 and the value
we derive below from our own experiments, this is indeed the
case.
CH3CH2O• f CH3 CHOH
(33)
•
The bimolecular rate constant for the reduction of MV2+ by
CH3 CHOH has been measured to be k21 ) (2.1 ( 0.1) × 109
•
M-1 s-1. ([MV2+] in these experiments was varied from 0.1 to
0.5 mM). This rate constant is somewhat lower than the
previously published early pulse radiolysis value of 4 × 109
M-1 s-1 29, which by comparison with several published values
in the range of (2.1-3.5) × 109 M-1 s-1 30 for the corresponding
reaction of the even better reductant (CH3)2C•OH seems too
high. We therefore attribute the higher accuracy to our present
value of k21.
•-
Figure 3. Plot of reciprocal relative yield of I2 (1/Grel formed only
in the reaction with CH3CH2O•; for details, see text) versus reciprocal
iodide concentration as measured at 380 nm. Pulse-irradiated, N2O-
saturated, neutral aqueous solution contained iodide and 0.86 M CH3-
CH2OH.
The total yield of MV•+ was determined to be G(MV•+)total
) (0.684 ( 0.004) µmol J-1. From calculations similar to those
for the CD3CD2OH system, the yield due to the reactions of
•
only OH radicals with CH3CH2OH is equal to
From the slope of this reciprocal plot (slope ) intercept ×
k/k27), it is possible to estimate the overall rate constant k for
those CH3CH2O• radical reactions that compete with the iodide
oxidation. Taking k27 ≈ 4 × 109 M-1 s-1 for the latter (i.e., a
value similar to the reported one for the oxidation of iodide by
CH3O• radicals, see above), we calculate k ≈ (5 ( 2) × 106
s-1. This is very close to 8 × 106 s-1 reported for the 1,2 H-atom
shift of alkoxyl radicals28 such as CH3CH2CH2O•, and accord-
ingly, we assign this rate constant essentially to reaction 33.
There may be a small contribution by the second-order process
(eq 34) that, on the basis of known rate constants for similar
reactions,26,27 may not exceed 10%.
G(MV•+)•OH ) G(MV•+)total - G(H•) ) 0.684 - 0.06 )
0.624 µmol J-1
or in relative units
G(MV•+)•OH
G(•OH )
0.624
0.67
G(MV•+)•OH, relative
)
)
) 0.93
(i.e., 93% of all hydroxyl radicals reacting with CH3CH2OH
•
produce CH3 CHOH (partially via CH3CH2O• radicals)). The
remainder of 7% is due to the formation of â-hydroxy-C-
•
centered radicals CH2CH2OH formed in reaction 11.
•
CH3CH2O• + CH3CH2OH f CH3CH2OH + CH3 CHOH
•
The fraction of OH radicals leading to CH3CH2O• radicals
was determined by measuring the absolute yield of I2•- in N2O-
saturated, neutral aqueous solutions containing 0.86 M CH3-
CH2OH and iodide in the range of 3.3 to 10 mM. Upon pulse
(34)
Finally, it seems appropriate at this point to comment on some
earlier data13 on the relative yields of abstraction in light of the
present results. The relative yield of oxyl radicals CH3CH2O•
had been found to be 2.5% (i.e., practically identical to our
present finding). The 84.3% yield reported in this earlier study
irradiation, I2 was formed in increasing yield as [I-] was
•-
increased, and values of G ) 0.035 up to 0.070 µmol J-1 were
measured within the employed iodide concentration range. These
values had to be corrected, however, for the small (but, in the
correlation with the total yield of I2•- measured, still significant)
•
for initially generated reducing CH3 CHOH radicals is, however,
•-
amount of I2 formed directly in the reaction of iodide with
slightly lower than our present value of 89%. Accordingly, the
•OH radicals. The corrective terms were calculated from the
known alcohol and iodide concentrations in each system and
•
relative yield of the redox inert CH2CH2OH (calculated as a
difference) is higher in this earlier report (13.2 vs 7%). Possible
•
•
the respective rate constants for the OH reactions. The thus-
reasons for the slight discrepancy in the CH3 CHOH yield may
•-
corrected G(I2
)
values were then related to the total yield
be the higher accuracy of the pulse radiolysis system employed
in our present study. Furthermore, no mathematical dependence
of scavengable •OH yields as a function of scavenger concentra-
tion had been established at that time. Also, the possible
contributions of H•-atom reactions with the alcohol and the
scavenger used for titrating reducing species (tetranitromethane)
were not fully recognized. Given all of the possible uncertainties,
there is no point in reevaluating the old system because the
two important values (i.e., the relative yields of oxyl and
R-hydroxyethyl radicals) are not very different. Nevertheless,
we recommend the data from our present study for further
applications.
Methylamine. The yields of reducing, C-centered R-amino-
methyl radicals in N2O-saturated solutions containing 0.5 mM
MV2+ and 0.12 M normal methylamine, pH 10.8, or 0.05 M
deuterated methylamine, pH 10.9, were again determined by
measuring the yields of produced MV•+ at 600 nm. The values
oxyl
•
•-
of OH radicals (i.e., Grel ) G(I2
)
oxyl/G(•OH)) and plotted
according to
1
Grel
1
k
1
)
1 +
•-
-
G0(I2
)
oxyl, relative
(
(
))
k27 [I ]
as 1/Grel versus the reciprocal iodide concentration in Figure 3.
From the intercept (i.e., at infinite iodide concentration),
G (I2
)
oxyl, relative ) Grel(CH3CH2O•) ) 0.043 ( 0.005 has been
•-
0
•
obtained. This means that about 4% of all OH radicals lead
initially to the formation of CH3CH2O• radicals (reaction 12).
•
Accordingly, the relative yield of CH3 CHOH radicals pro-
•
duced via reaction 1 amounts to G(CH3 CHOH)•OH, relative
)
G(MV•+)•OH, relative - G(I2
)
) 0.93 - 0.04 ) 0.89
•-
oxyl, relative
(i.e., 89% of the •OH radicals directly yield the R-hydroxyethyl
radical).

7274 J. Phys. Chem. B, Vol. 107, No. 30, 2003
Bonifacˇic´ et al.
obtained were 0.308 and 0.227 µmol J-1, respectively. As in
the case of normal ethanol, the formation of MV•+ occurred
via an exponential [MV2+]-dependent process. The measured
yields include the 0.06 µmol J-1 contribution by H atoms, which
were considered to react exclusively with neutral methylamine
molecules and to yield only the reducing R-aminomethyl
radicals. The yields of scavengable hydroxyl radicals in each
of the above solutions were calculated to be G(•OH)total ) 0.67
•
and 0.63 µmol J-1, respectively. Thus, the reaction of OH
•
radicals with CH3NH2 leads to the formation of CH2NH2
radicals with a yield of
G(MV•+)•OH ) G(MV•+)total - G(H•) ) 0.308 - 0.06 )
0.248 µmol J-1
or, expressed in relative units,
G(MV•+)•OH
G(•OH)
Figure 4. Plot of reciprocal semiquinone yield expressed as 1/G(Q•-
)
0.248
0.67
G(MV•+)•OH, relative
)
)
) 0.37
versus [H2Q]-1 measured at 427 nm in pulse-irradiated, N2O-saturated
aqueous solution at pH 11.0 containing hydroquinone and 0.12 M
methylamine. Upper inset: absorption (arbitrary units) versus the time
trace at 427 nm obtained in the same solution containing 0.42 mM
hydroquinone. Lower inset: plot of kobs for semiquinone formation
versus [H2Q] in the same solutions as in the main plot.
•
•
Accordingly, 37% of all OH radicals yield CH2NH2 radicals
whereas the remainder of 63% is assigned to the formation of
oxidizing NHCH3 radicals. For the system with CD3NH2, the
•
numerical values are
feature for all employed H2Q concentrations, that the signals
were not stable after the growth period. They always showed
some second-order decay that, tentatively, we assign to the back
reduction of the semiquinone radical by the methylamine-
derived, reductive radicals produced simultaneously. The pseudo-
first-order rate constants for the semiquinone formation were,
therefore, obtained by fitting the experimental data to an
expression with a sequential growth and decay. Results are
shown in the bottom inset of Figure 4 as kobs versus hydro-
quinone concentration. From the slope of the obtained straight
line, a rate constant of (1.1 ( 0.1) × 108 M-1 s-1 was calculated
and assigned to reaction 25. This rate constant seems to be
characteristic for the oxidation of hydroquinone at pH 11.0 by
alkylaminyl radicals, in general, because it is very similar to
the values reported previously for aminyl radicals derived from
G(MV•+)•OH ) G(MV•+)total - G(H•) ) 0.227 - 0.06 )
0.167 µmol J-1
or, expressed in relative units,
G(MV•+)•OH
G(•OH)
0.167
0.63
G(MV•+)•OH, relative
)
)
) 0.26
•
•
In this case, 26% of all OH radicals generate CD2NH2, and
the remainder of 74% is assigned to the formation of oxidizing
•NHCD3 radicals.
The respective rate constants for the MV2+ reduction by •CH2-
NH2 have been measured to be k23 ) (3.6 ( 0.3) × 109 M-1
s-1. The same number was also obtained for the corresponding
reaction of •CD2NH2, though only from a single-point measure-
ment. Our present measurement nicely confirms our previous
3.3 × 109 M-1 s-1 value for the reduction of MV2+ by
•CHRNH2 radicals that were formed as the result of an oxidative
decarboxylation of glycine anions (at pH 10.6)5 or methionine.31
The formation of the oxidizing N-centered radicals was
additionally proven and quantitatively measured for solutions
alanine and glycine (1.1 × 108 and 0.8 × 108 M-1 s-1
,
respectively).3,4
The yield of semiquinone radicals, calculated from the
measured maximum absorbances by using the known molar
extinction coefficient of 7200 M-1 cm-1 for Q•- and applying
corrections for the slight decay, increased with the concentration
of hydroquinone. This indicates the occurrence of additional
•
reactions of NHCH3 radicals in competition with reaction 25.
To analyze for this, the reciprocal Q•- yield has been plotted
versus the reciprocal hydroquinone concentration in Figure 4.
From the intercept of the straight line, G(Q•-) ) 0.404 µmol
J-1 is obtained at infinite hydroquinone concentration. In units
relative to the yield of scavengable •OH radicals (calculated to
be 0.675 µmol J-1 in the investigated solution), it is G(Q•-)relative
) G(•NHCH3)relative ) 0.404/0.675 ) 0.60. This is very close
to the value of 0.63 deduced from the results presented above
using methyl viologen as a scavenger for the reducing radicals.
Such a good fit also confirms that the procedures applied in
this study for yield determinations are meaningful and correct.
Finally, it might be interesting to stress the relatively high
stability of both aminyl radicals, •NHCH3 and •NHCD3. Under
the experimental conditions used in this study, it seems that
they decay only by second-order radical-radical reactions as
indicated by the slight disappearance of MV•+ on the time scale
of ∼200 µs. In particular, it seems that neither the first-order
nor the second-order transformation into C-centered radicals,
•
containing CH3NH2. The NHCH3 radicals generated in this
system were titrated by using hydroquinone as a scavenger
following the procedure that was successfully applied for the
quantification of similar NH-R radicals formed in the OH-
induced oxidation of amino acid anions.3 Actual pulse-radiolysis
measurements were performed with N2O-saturated, pH 11.0
solutions containing 0.12 M methylamine and hydroquinone in
the concentration range of 0.2-1.2 mM. The relevant reaction
•
•
•NHCH3 + HQ- f NH2CH3 + Q•-
(25)
has already been introduced in an earlier section (and discussed
with respect to the hydroquinone’s pK values).
The formation of the semiquinone radicals was followed at
the maximum of their optical absorbance at 427 nm. A typical
time-resolved trace is shown as the upper inset in Figure 4 for
a total H2Q concentration of 0.42 mM. It is noted, as a common

•
H/D Abstraction by OH from Ethanol and Methylamine
J. Phys. Chem. B, Vol. 107, No. 30, 2003 7275
reactions 35 and 36, plays any significant role on this time scale.
SCHEME 1
•NHCH3 f •CH2NH2
(35)
•NHCH3 + CH3NH2 f •CH2NH2 + CH3NH2 (36)
This parallels correspondingly low reactivities, with respect to
these processes, of aminyl radicals derived from amino acids
(e.g., glycine, •NH-CH2-CO2-).3,4 (Such radicals decay,
however, also by additional intramolecular homolytic C-C bond
rupture liberating •CO2-, a reaction not possible, of course, for
the methylamine-derived radicals.)
Table 2 summarizes the overall rate constants (ktotal) for the
reactions of •OH radicals with CH3CH2OH, CD3CD2OH, CH3-
NH2, and CD3NH2. Also displayed are the partial rate constants
for the formation of all radicals detected as products. These rate
constants were obtained by multiplying ktotal by the relative
yields of the respective radicals listed in Table 1. Rate constants
for the various redox titration and rearrangement reactions,
which were measured and are new, are summarized in Table 3.
Kinetic Isotope Effects. The kinetic isotope effects (KIEs
An important aspect of the present work is a comparison of
the partitioning of OH attack on the aliphatic groups versus
•
the OH or NH2 components in the ethanol and methylamine
•
•
or values of kH/kD) for the OH reactions have been given in
molecules, respectively. As pointed out earlier, ET to OH
Table 2 for the total and for each of the partial rate constants.
KIEs for total rates of reaction with •OH are not available from
the literature for CH3NH2, and those derived for ethanol from
the earlier solution data give values of 1.3 to 1.8 depending on
which result is used for k15. The earlier gas-phase value of 2.632
is more precise, and it could be significant that the present
solution result of 1.93 (Table 2) lies below it. This may be an
indication that the barrier heights of the present reactions are
lowered in solution because of polar transition states33 or the
fact that tunneling is reduced.34 An elucidation of these points
requires high-level theoretical calculations. Further discussion
is centered on KIEs of the individual or partial reactions, the
weighted sums of which make up the total KIEs.
(reaction 5) is feasible in the case of methylamine, and the
C-centered radicals may be formed subsequently by proton loss
(reactions 6 or 7). This is particularly evident when one
considers that the initial product of reaction 5 is a closely spaced
[HO-...+•NH2CH3/+•NH2CD3] ion pair as shown in Scheme 1.
Proton transfer to the strong base HO- may occur before
outward diffusion, which takes several picoseconds (rate kdiff).7
The overall process via this route that includes reaction 5
followed by reaction 6 has been denoted as k3,ET in Scheme 1
to indicate its alternative character to the direct abstraction route
(reaction 3). (For the deuterated system, the reference reactions
are 4 and 7 instead of 3 and 6.) In any case, if that were the
•
dominant mode of reaction of OH, then deuteration of the C
As is to be expected,35 the highest isotope effect applies to
the abstraction of a primary hydrogen atom from an unsubsti-
tuted alkyl group (“unsubstituted” meaning connected only to
an alkyl group or to hydrogen). In our study, this applies to the
terminal methyl group in ethanol for which KIE ) 3.4 has been
found. This value can be compared to 4.6 for the H/D KIE in
ethane in the gas phase.36 Again, the solution result is smaller.
However, a relatively high error limit must be allowed in the
present result because of the low relative yields of the â-hydroxy
radicals, which are the basis for the calculation of the kâ-C rate
constants. For secondary hydrogens in unsubstituted methylene
groups, a recent review35 lists KIE ) 2.6 both in the gas phase
and in solution. This is to be compared with 1.96 in Table 2
for the H/D abstraction from the CR position in ethanol (yielding
the R-hydroxyalkyl radical). Here, the yields, rate constants, and
consequently the KIE could be determined at high accuracy,
and the difference from the pure aliphatic system must be taken
as real. A plausible rationale is the C-H/C-D bond-weakening
effect exerted by the R-positioned hydroxy group37 in combina-
tion with the reported35 and generally accepted fact that the KIEs
decrease with decreasing bond-dissociation energies.
center should have a negligible (or minimal) effect on the value
of ktotal. This is because any possible reduction in the yield of
C-centered radicals due to the KIE for the overall electron-
transfer route (and expressing itself in k3,ET-H and k3,ET-D for
the hydrogenated and deuterated systems, respectively) should
be made up by a compensating increase in aminium deproto-
nation. Instead, what is seen in Table 2 is a very significant
reduction in ktotal (KIE ) 1.31) coupled with a small reduction
(rather than increase) in kN (KIE ) 1.11). In fact, apart from
the latter effect, the behavior for methylamine very closely
parallels that of ethanol, for which ET to •OH is expected to be
negligible;46 viz., in both systems, deuteration causes a very
significant reduction in kR-C and a very small or negligible effect
on kO/N. One may therefore conclude that in ethanol and
methylamine the C-centered radicals are formed by H-abstrac-
tion reactions (i.e., via reaction 3) with no significant contribu-
tions from reactions 5 and 6 or 7 (k3,ET route in Scheme 1) and
their counterparts in ethanol. The discussion in the preceding
section, showing that the present kR-C KIEs can be rationalized
on the basis of literature values for known abstraction reactions
in related systems, is in accord with this.
The KIE ()1.86) for the H/D abstraction from the methyl
group in methylamine can be explained in a similar manner by
the influence of the amino group. Indeed, the fact that it is
slightly lower than that for the abstraction of the secondary
hydrogen in ethanol conforms to the difference in the BDEs
Because the rapid proton exchange between the N-D bonds
in the original CD3ND2 and the solvent H2O precludes any
primary isotope effect in the attack of •OH on the amino group,
the present study gives no information on the relative roles of
reaction 5 followed by reaction 8 or 9 (k10,ET route in Scheme
1) as opposed to reaction 10. However, one remarkable aspect
of the present data is the approximately 20-fold-higher (reducing
to 10-fold higher on a per H atom available basis) rate constant
for H abstraction from N-H compared to that for O-H in our
investigated systems (compare kN and kO in Table 2). If direct
for the CR-Η bonds in ethanol (397 kJ mol-1 37
) and methylamine
(393 kJ mol-1).38 However, the barrier heights in both of these
systems are low enough that tunneling may be very important,
and detailed conclusions will require a careful theoretical
analysis.

7276 J. Phys. Chem. B, Vol. 107, No. 30, 2003
Bonifacˇic´ et al.
(14) Zhao, F.; Ghezzo-Scho¨neich, E.; Aced, G. I.; Hong, J.; Milby, T.;
Scho¨neich, C. J. Biol. Chem. 1997, 272, 9019.
(15) Hug, G. L.; Wang, Y.; Scho¨neich, C.; Jiang, P.-Y.; Fessenden, R.
W. Radiat. Phys. Chem. 1999, 54, 559.
(16) von Sonntag, C. The Chemical Basis of Radiation Biology; Taylor
and Francis: New York, 1987.
(17) Schuler, R. H.; Hartzell, A. L.; Behar, B. J. Phys. Chem. 1981, 85,
192.
H abstraction is involved, then the deciding parameter would
be the difference in BDEs (423 kJ mol-1 for MeNH-H 38 versus
437 kJ mol-1 for EtO-H 40), which would affect the barrier
heights. However, ET is also a factor that could enhance the
rate for MeNH2. The KIE for kO in Table 2 is 1.0. The KIE of
1.11 for kN is close to, if not within, the experimental error limits.
However, a secondary KIE of this magnitude might arise in
ET (reaction 5) or in an H abstraction (reaction 10).
(18) Janata, E.; Schuler, R. H. J. Phys. Chem. 1982, 86, 2078.
(19) Getoff, N.; Schwoerer, F. Int. J. Radiat. Phys. Chem. 1970, 2, 81.
All of the rate constants discussed above find their reflection
in the various yields of radicals and products. Accordingly, there
is no point in entering into a similarly detailed analysis of yields
with basically the same arguments again. However, the rate
constant data in Table 3 are worth a comment. The first four
are for ETs from radicals to MV2+. For the first two, the radicals
are CH3C•HOH and CD3C•DOH, and for the second two, •CH2-
NH2 and •CD2NH2. Since D is a slightly better electron-density-
releasing substituent than H,41 the overall electron density at
oxygen or nitrogen should be higher in the deuterated radicals,
and one might expect faster ET from those species. The rate
for CD3C•DOH is marginally larger than for CH3C•HOH, but
overall, the results show that the effect is small or negligible.
Two other comments concern the intramolecular conversion
of the ethoxyl radicals to their respective R-hydroxyethyl radicals
(eqs 31 and 33). Our experiments reveal a KIE of about 50
(see first-order rate constants in Table 3). Although our rate
constant for the 1,2 H-atom shift of CH3CH2O•, reaction 33,
was similar to that of the CH3CH2CH2O• alkoxyl radical from
reference 28, we were unable to find KIEs for the 1,2 H shift
of CH3CH2O• or similar alkyl alkoxyl radicals. Thus, our
presently measured value appears to be the first of this kind.
The very high KIE is an indication that tunneling probably plays
a significant role indeed in the solvent-assisted mechanism
proposed earlier for this process.28,42 It has, in fact, been noted
in a recent publication43 that “some of the largest deuterium
isotope effects that have been reported involve unimolecular
H-atom shift in radicals.” Other than this 1,2 H-atom shift, the
intermolecular hydrogen transfer (eqs 32 and 34) shows a
probably very normal KIE (actual numbers related to literature
values26,27 vary from 1.5 to 6.1 but must allow very high error
limits).
(20) Wigger, A.; Gruenbein, W.; Henglein, A.; Land, E. J. Z. Natur-
forsch., B: Chem. Sci. 1969, 24, 1262.
(21) Armstrong, D. A.; Sun, Q.; Schuler, R. H. J. Phys. Chem. 1996,
100, 9892.
(22) Hug, G. L. Natl. Stand. Ref. Data Ser.; U.S. National Bureau of
Standards, Report No. 69, 1981.
(23) Mezyk, S. P.; Bartels, D. M. J. Phys. Chem. 1997, 101, 1329.
(24) Solar, S.; Solar, W.; Getoff, N.; Holcman, J.; Sehested, K. J. Chem.
Soc., Faraday Trans. 1 1982, 78, 2467.
(25) Burchill, C. E.; Ginns, I. S. Can. J. Chem. 1970, 48, 2628.
(26) Ellison, D. H.; Salmon, G. A.; Wilkinson, F. Proc. R. Soc. London,
Ser. A 1972, 328, 23.
(27) Paul, H.; Small, R. D., Jr.; Scaiano, J. C. J. Am. Chem. Soc. 1978,
100, 4520.
(28) Gilbert, B. C.; Holmes, R. G. G.; Laue, H. A. H.; Norman, R. O.
C. J. Chem. Soc., Perkin Trans. 2 1976, 1047.
(29) van Leeuwen, J. W.; van Dijk, C.; Veeger, C. Eur. J. Biochem.
1983, 135, 601.
(30) Neta, P.; Grodkowski, J.; Ross, A. B. J. Phys. Chem. Ref. Data
1996, 25, 709.
(31) Hiller, K.-O.; Masloch, B.; Go¨bl, M.; Asmus, K.-D. J. Am. Chem.
Soc. 1981, 103, 2734.
(32) Wallington, T. J.; Dagaut, P.; Kurylo, M. J. J. Phys. Chem. 1988,
92, 5024.
(33) Roberts, B. P.; Steel, A. J. J. Chem. Soc., Perkin Trans. 2 1994,
2155.
(34) Garrett, B. C.; Schenter, G. K. In Structure and ReactiVity in
Aqueous Solution; Cramer, C. J., Truhler, D. G., Eds.; American Chemical
Society: Washinton, DC, 1994; p 122.
(35) Fokin, A. A.; Schreiner, P. R. Chem. ReV. 2002, 102, 1551.
(36) Tully, F. P.; Droege, A. T.; Koszykowski, M. L.; Melius, C. F. J.
Phys. Chem. 1986, 90, 691.
(37) Dyke, J. M.; Groves, A. P.; Lee, E. P. F.; Zamanpour Niavaran,
M. H. J. Phys. Chem. 1997, 101, 373.
(38) Lias, S. G.; Bartmess, J. E.; Liebman, J. F.; Holmes, J. L.; Levin,
R. D.; Mallard, W. G. J. Phys. Chem. Ref. Data 1988, 17, Suppl. No. 1.
(39) Pearson, R. G. J. Am. Chem. Soc. 1986, 108, 6109.
(40) Berkowitz, J.; Ellison, G. B.; Gutman, D. J. Phys. Chem. B 1994,
98, 2744.
(41) March, J. In AdVanced Organic Chemistry; Wiley & Sons: New
York, 1985; p 18.
Acknowledgment. D.A.A. acknowledges financial support
from the University of Calgary and NSERCC. K.-D.A. is
currently on sabbatical leave from the Department of Chemistry
and Biochemistry of the University of Notre Dame.
(42) Elford, P. E.; Roberts, B. P. J. Chem. Soc., Perkin Trans. 2 1996,
2247.
References and Notes
(43) Zhao, Y.; Lu, Y.; Parker. V. D. J. Chem. Soc., Perkin Trans 2
2001, 1481, and references therein.
(1) Wardman, P. J. Phys. Chem. Ref. Data 1989, 18, 1637.
(2) Das, S.; von Sonntag, C. Z. Naturforsch., B: Chem. Sci. 1986, 41,
505.
(3) Bonifacˇic´, M.; Sˇtefanic´, I.; Hug, G. L.; Asmus, K.-D.; Armstrong,
D. A. J. Am. Chem. Soc. 1998, 120, 3139.
(4) Bonifacˇic´, M.; Armstrong, D. A.; Carmichael, I.; Asmus, K.-D. J.
Phys. Chem. B 2000, 104, 643.
(5) Sˇtefanic´, I.; Bonifacˇic´, M.; Asmus, K.-D.; Armstrong, D. A. J. Phys.
Chem. A 2001, 105, 8681.
(6) Hug, G. L.; Fessenden, R. W. J. Phys. Chem. A 2000, 104, 7021.
(7) Hug, G. L.; Bonifacˇic´, M.; Asmus, K.-D.; Armstrong, D. A. J. Phys.
Chem. B 2000, 104, 6674.
(8) Yu, D.; Rauk, A.; Armstrong, D. A. J. Am. Chem. Soc. 1995, 117,
1789.
(9) Bockman, T. M.; Hubig, S. M.; Kochi, J. K. J. Am. Chem. Soc.
1998, 120, 2826.
(10) Getoff, N.; Schwoerer, F. Int. J. Radiat. Phys. Chem. 1971, 3, 429.
(11) Buxton, G. V.; Greenstock, C. L.; Helman, W. P.; Ross, A. B. J.
Phys. Chem. Ref. Data 1988, 17, 513.
(44) Methyl viologen is quoted to react with hydrogen atoms at pH 1
with k ) 6 × 108 M-1 s-1, forming a product that does not absorb at 600
nm.24 This would mean that, under the conditions of our experiment, H•
radicals would also be partially scavenged by methyl viologen (ca. 25, 40,
and 55% for solutions containing 0.11 M CD3CD2OH and 0.1, 0.3, and 0.5
mM MV2+, respectively). H• radicals are produced only at low yield in
N2O-saturated aqueous solutions at pH > 3 (ca. only 10% of G(•OH)).
Considering the very low error limit for the analysis of MV•+, the decrease
in total G(MV•+) caused by this reaction should, nevertheless, be clearly
detectable. Because this was not the case and the yield of reduced MV2+
stayed constant at all [MV2+] employed, we concluded that in the system
under investigation all H• radicals reacted exclusively with CD3CD2OH.
(45) Possible conversion of the less-reducing •CD2CD2OH into strongly
•
reducing CD3 CDOH radicals via the reaction •CD2CD2OH + CD3CD2OH
f CD3 CDOH + CD3CD2OH is very slow (k ) 16 M-1 s-1 for normal
•
ethyl alcohol25) and does not occur on the ∼200-µs time scale of this
experiment.
(46) Pearson39 estimated Eo(MeOH•+/MeOH) ) 2.75 V, and Eo(EtOH•+
/
EtOH) will be about 0.3 V less. With Eo(•OH/OH-) ) 1.90 V,1 ET from
(12) Anbar, M.; Meyerstein, D.; Neta, P. J. Chem. Soc. B 1966, 742.
(13) Asmus, K.-D.; Mo¨ckel, H.; Henglein, A. J. Phys. Chem. 1973, 77,
1218.
EtOH to •OH will be about 0.5 V endergonic, about the same as for nitrate
•
or sulfate ion oxidation by OH.