Mechanism one-electron oxidation of β-, γ-, and δ-hydroxyalkyl sulfides. Catalysis through intramolecular proton transfer and sulfur-oxygen bond formation

Article abstract of DOI:10.1021/ja970683r

The mechanism of photoinduced electron transfer between sulfur-containing alcohols and the 4-carboxybenzophenone (CB) triplet state in aqueous solution was investigated using laser flash photolysis and steady-state photolysis techniques. Bimolecular rate constants for quenching of the CB triplet state by five hydroxyalkyl sulfides, 2-(methylthio)ethano1 (2-MTE), 2,2'-dihydroxydiethyl sulfide (2,2'-DHE), 3-(methylthio)propanol (3-MTP), 3,3'-dihydroxydipropyl sulfide (3,3'-DHP), and 4-(methylthio)butanol (4-MTB), with varying numbers of OH groups and varying locations with respect to the sulfur atom, were determined to be in the range (3.3-4.8) x 10⁹ M^-1 s^-1 for neutral solutions. The intermediates identified were the CB ketyl radical anion (CB^.-), the CB ketyl radical (CBH^.), and intermolecularly (S.^..S)-bonded radical cations. An additional absorption band at approximately 400 nm in the transient spectra for some of the hydroxyalkyl sulfides was assigned to the intramolecularly (^.S-O)-bonded species (only for hydroxyalkyl sulfides which could adopt a five-membered ring structure). The spectra of appropriate (S.^..S)⁺ and (^.S-O) intermediates for the hydroxyalkyl sulfides were determined from complementary pulse radiolysis studies in acid and neutral aqueous solutions of the hydroxyalkyl sulfides, respectively. The observation of ketyl radical anions and intermolecular (S.^..S)-bonded radical cations of the hydroxyalkyl sulfides was direct evidence for the participation of electron transfer in the mechanism of quenching. Quantum yields of formation of intermediates from flash photolysis experiments and quantum yields of formaldehyde formation from the steady-state measurements were determined. The values of these quantum yields indicated that the diffusion apart (escape of the radical ions) of the charge-transfer complex, formed as a primary photochemical step, is a minor photochemical pathway (with a contribution of ~5-25% depending on the numbers of OH groups). Competing processes of proton transfer and back electron transfer within the CT complex gave significant contributions to these yields. Detailed mechanisms for the CB-sensitized photooxidation of sulfur-containing alcohols are proposed, discussed, and compared with that for the ^.OH-induced oxidation. One striking feature of the mechanisms is that there is a catalytic influence of neighboring groups on the radical reaction pathways during one-electron oxidation of the hydroxyalkyl sulfides in comparison to comparable reactions of nonsubstituted alkyl sulfides. Support for the mechanisms came in part from an analysis of observed solvent isotope effects on radical quantum yields.

Full text of DOI:10.1021/ja970683r

8000
J. Am. Chem. Soc. 1997, 119, 8000-8011
Mechanism of One-Electron Oxidation of â-, γ-, and
δ-Hydroxyalkyl Sulfides. Catalysis through Intramolecular
Proton Transfer and Sulfur-Oxygen Bond Formation
Krzysztof Bobrowski,*,†,‡ Gordon L. Hug,*,‡ Bronislaw Marciniak,*,§,‡
Brian Miller, and Christian Scho1neich*,
Contribution from the Institute of Nuclear Chemistry and Technology, 03-195 Warsaw, Poland,
Radiation Laboratory, UniVersity of Notre Dame, Notre Dame, Indiana 46556, Faculty of
Chemistry, A. Mickiewicz UniVersity, 60-780 Poznan, Poland, and Department of Pharmaceutical
Chemistry, UniVersity of Kansas, 2095 Constant AVenue, Lawrence, Kansas 66047
ReceiVed March 3, 1997X
Abstract: The mechanism of photoinduced electron transfer between sulfur-containing alcohols and the 4-carboxy-
benzophenone (CB) triplet state in aqueous solution was investigated using laser flash photolysis and steady-state
photolysis techniques. Bimolecular rate constants for quenching of the CB triplet state by five hydroxyalkyl sulfides,
2-(methylthio)ethanol (2-MTE), 2,2′-dihydroxydiethyl sulfide (2,2′-DHE), 3-(methylthio)propanol (3-MTP), 3,3′-
dihydroxydipropyl sulfide (3,3′-DHP), and 4-(methylthio)butanol (4-MTB), with varying numbers of OH groups
and varying locations with respect to the sulfur atom, were determined to be in the range (3.3-4.8) × 109 M-1 s-1
for neutral solutions. The intermediates identified were the CB ketyl radical anion (CB•-), the CB ketyl radical
(CBH•), and intermolecularly (S S)-bonded radical cations. An additional absorption band at approximately 400
nm in the transient spectra for some of the hydroxyalkyl sulfides was assigned to the intramolecularly (•S-O)-
bonded species (only for hydroxyalkyl sulfides which could adopt a five-membered ring structure). The spectra of
appropriate (S S)+ and (•S-O) intermediates for the hydroxyalkyl sulfides were determined from complementary
pulse radiolysis studies in acid and neutral aqueous solutions of the hydroxyalkyl sulfides, respectively. The observation
of ketyl radical anions and intermolecular (S S)-bonded radical cations of the hydroxyalkyl sulfides was direct
evidence for the participation of electron transfer in the mechanism of quenching. Quantum yields of formation of
intermediates from flash photolysis experiments and quantum yields of formaldehyde formation from the steady-
state measurements were determined. The values of these quantum yields indicated that the diffusion apart (escape
of the radical ions) of the charge-transfer complex, formed as a primary photochemical step, is a minor photochemical
pathway (with a contribution of ∼5-25% depending on the numbers of OH groups). Competing processes of proton
transfer and back electron transfer within the CT complex gave significant contributions to these yields. Detailed
mechanisms for the CB-sensitized photooxidation of sulfur-containing alcohols are proposed, discussed, and compared
with that for the •OH-induced oxidation. One striking feature of the mechanisms is that there is a catalytic influence
of neighboring groups on the radical reaction pathways during one-electron oxidation of the hydroxyalkyl sulfides
in comparison to comparable reactions of nonsubstituted alkyl sulfides. Support for the mechanisms came in part
from an analysis of observed solvent isotope effects on radical quantum yields.
Introduction
compounds.2-5 The attacking oxidant usually induces formation
of a distinct intermediate at the sulfide site, and the nature of
this intermediate circumscribes the scope of any subsequent
participation of neighboring groups. The nature and strength
of the coupling of the neighboring group to any attacking
oxidizing species and/or to the intermediate moiety formed at
the sulfide site affects the ultimate course of the sulfide
oxidation. Reactions that start as simple one-electron oxidations
can, through interactions with neighboring groups, end up as
transient species that are susceptible to carbon-carbon bond
fragmentation.
Organic sulfides play an important role as structural elements
in many materials and biological macromolecules. Because of
their high susceptibility to oxidation any attempt to maintain
the integrity of such sulfides requires a detailed understanding
of the underlying mechanisms as well as the physico-chemical
parameters which control oxidation reactions. A field which
would particularly benefit from such knowledge is the biotech-
nological production of recombinant proteins where methionine
residues often suffer oxidation during processing, purification,
and storage.1
An illustration of how these themes play out can be seen in
the oxidation of the dipeptide threonylmethionine (Thr-Met) by
two different one-electron oxidants, the hydroxyl radical, •OH,
In proteins, methionine residues are potentially surrounded
by a variety of neighboring groups such as carboxylate, amino,
hydroxy, and amide functionalities most of which have been
shown to affect sulfide oxidations in simple organic model
•- 6
•
and the sulfate radical anion SO4
.
The reaction with OH
(2) Glass, R. S.; Hojjatie, M.; Wilson, G. S.; Mahling, S.; Go¨bl, M.;
Asmus, K.-D. J. Am. Chem. Soc. 1984, 106, 5382-5383.
(3) Asmus, K.-D.; Go¨bl, M.; Hiller, K.-O.; Mahling, S.; Mo¨nig, J. J.
Chem. Soc., Perkin Trans. 2 1985, 641-646.
† Institute of Nuclear Chemistry and Technology.
‡ University of Notre Dame.
§ A. Mickiewicz University.
University of Kansas.
(4) Mahling, S.; Asmus, K.-D.; Glass, R. S.; Hojjatie, M.; Wilson, G. S.
J. Org. Chem. 1987, 52, 3717-3724.
X Abstract published in AdVance ACS Abstracts, August 15, 1997.
(1) Manning, M. C.; Patel, K.; Borchardt, R. T. Pharm. Res. 1989, 6,
903-918.
(5) Steffen, L. K.; Glass, R. S.; Sabahi, M.; Wilson, G. S.; Scho¨neich,
C.; Mahling, S.; Asmus, K.-D. J. Am. Chem. Soc. 1991, 113, 2141-2145.
S0002-7863(97)00683-5 CCC: $14.00 © 1997 American Chemical Society

One-Electron Oxidation of â-, γ-, and δ-Hydroxyalkyl Sulfides
J. Am. Chem. Soc., Vol. 119, No. 34, 1997 8001
leads to a sequence of reactions: (i) formation of a hydroxy-
sulfuranyl radical, >•S-OH, (ii) coupled proton/electron transfer
from the N-terminal amino group to the hydroxysulfuranyl
radical, and (iii) formation of a nitrogen-centered radical cation
which induces a carbon-carbon fragmentation of the Thr side
and Baciocchi et al.15 obtained kH/kD ) 1.2-1.4 for the
hydroxide/water-catalyzed fragmentation of arylalkanol radical
cations. However, no solvent isotope effect was observed for
the decomposition of hydroxysulfuranyl radicals of 2-MTE and
2,2′-DHE.13 For this reason, it was suggested that the mech-
anism of this reaction did not involve intermediary radical
cations. Instead it was proposed that the mechanism involved
a hydrogen transfer between the â-hydroxy group and the
oxygen of the hydroxysulfuranyl moiety via a cyclic transition
state (left-hand side of reaction 4) to yield an alkoxy radical
(reaction 4).
•-
chain. In contrast, the oxidation by SO4 results in the
formation of a relatively stable cyclic S-O bond that is formed
through interaction of an initially formed sulfur radical cation
>S•+ with the C-terminal carboxylate group. The difference
between these two one-electron oxidation pathways appears to
originate from the initial interaction of the two oxidizing species
with the sulfur, i.e., hydroxysulfuranyl radical vs sulfur radical
cation. Subsequently these initially formed intermediates at the
sulfide site interact with different neighboring groups, i.e.,
N-terminal amino group vs C-terminal carboxylate group. Other
examples for such mechanisms exist, i.e., the oxidation of
•
•- 10
γ-glutamylmethionine and analogs by OH,7-9 SO4
,
and
S-alkylglutathiones by OH11 and triplet 4-carboxybenzophe-
•
The latter, RS(CH2)2O•, would subsequently suffer fragmenta-
tion of the carbon-carbon bond (reaction 5).
none.12
Hydroxyalkyl sulfides provide conceptually simpler models
for understanding neighboring-group participation in the oxida-
tion of the sulfides than threonylmethionine. However, even
with these compounds complications arise. For instance, when
•OH radicals reacted with 2-(methylthio)ethanol (2-MTE) and
2,2′-dihydroxydiethyl sulfide (2,2′-DHE),13 an initially formed
hydroxysulfuranyl radical immediately fragmented to yield
alkylthiomethyl radicals and formaldehyde (reaction 1).
RS(CH2)2O• f •CH2SR + CH2dO
(5)
However, it was found13 that the hydroxyl radical-induced
fragmentation only yielded 40-50% formaldehyde, but R-(alkyl-
thio)alkyl radicals, •CH2SR, were produced quantitatively. The
inequality in the yields of R-(alkylthio)alkyl radicals and
formaldehyde can be accounted for by an intramolecular
hydrogen transfer from the CR-H bond of the substituent R in
the alkoxyl radical RS(CH2)2O•.
RS•(OH)CH2CH2OH f •CH2SR + H2O + CH2dO (1)
No intermediary sulfur radical cations were observed at neutral
to alkaline pH. On the other hand, when sulfur radical cations
or their dimeric three-electron bonded complexes were generated
at low pH (reaction 2),
RS(CH2)2O• f R′C•HSCH2CH2OH
(6)
This reaction is in competition with the CR-Câ fragmentation
of RS(CH2)2O• (reaction 5), and it resolves the discrepancy in
yields because it produces R-(alkylthio)alkyl radicals but no
formaldehyde. In reaction 6, R- ) R′CH2-.
•OH + 2RSCH2CH2OH + H+ f
[(R)(HOCH2CH2)S S(CH2CH2OH)(R)]+ + H2O (2)
The present paper investigates (i) the general mechanism of
one-electron oxidation of hydroxy-substituted aliphatic sulfides,
(ii) the neighboring-group participation on the formation of
radical cationic intermediates, and (iii) the mechanism of
carbon-carbon bond fragmentation subsequent to one-electron
oxidation of 2-(methylthio)ethanol and 2,2′-dihydroxydiethyl
sulfide. This paper provides further evidence that indeed typical
one-electron oxidants such as triplet 4-carboxybenzophenone
(CB) react via a different mechanism than hydroxyl radicals.
The key step during the triplet 4-carboxybenzophenone-induced
oxidation is the initial formation of a charge-transfer complex,
as now shown for various substituted thioethers.16-22 Within
this charge-transfer complex formed from the CB triplet and
the hydroxy-substituted alkyl sulfides, it will be shown that one
reaction involves proton-transfer from the γ-hydroxy-substituted
they were quite stable toward fragmentation, as will be shown
in this paper. The dimeric complex, symbolized as (S S)+,
exists in equilibrium 3 with its monomeric form.
(S S)+ a RSCH2CH2OH + R(S•+)CH2CH2OH (3)
In general, these â-hydroxy substituted radical cations, such as
R(S•+)CH2CH2OH in equilibrium 3, are expected to undergo
fragmentation of the carbon-carbon bond subsequent to base-
catalyzed deprotonation of the â-hydroxy group. As evidence
for this, Lucia et al.14 reported a solvent kinetic isotope effect
of kH/kD ) 2.0 for the pyridine-catalyzed fragmentation of
radical cations from 2-(N-phenylamino)-1,2-diphenylethanol,
(6) Scho¨neich, C.; Zhao, F.; Madden, K. P.; Bobrowski, K. J. Am. Chem.
Soc. 1994, 116, 4641-4652.
(15) Baciocchi, E.; Bietti, M.; Putignani, L.; Steenken, S. J. Am. Chem.
Soc. 1996, 118, 5952-5960.
(7) Bobrowski, K.; Scho¨neich, C.; Holcman, J.; Asmus, K.-D. J. Chem.
Soc., Perkin Trans. 2 1991, 353-362.
(16) Guttenplan, J. B.; Cohen, S. G. J. Org. Chem. 1973, 38, 2001-
2007.
(8) Bobrowski, K.; Scho¨neich, C.; Holcman, J.; Asmus, K.-D. J. Chem.
Soc., Perkin Trans. 2 1991, 975-980.
(9) Bobrowski, K.; Scho¨neich, C. Radiat. Phys. Chem. 1996, 47, 507-
510.
(10) Bobrowski, K.; Scho¨neich, C. Unpublished results.
(11) Bobrowski, K.; Pogocki, D.; Hug, G. L.; Scho¨neich, C. 1997,
manuscript in preparation.
(12) Bobrowski, K.; Hug, G. L.; Marciniak, B. Unpublished results.
(13) Scho¨neich, C.; Bobrowski, K. J. Am. Chem. Soc. 1993, 115, 6538-
6547.
(17) Bobrowski, K.; Marciniak, B.; Hug, G. L. J. Photochem. Photobiol.,
A 1994, 81, 159-168.
(18) Marciniak, B.; Bobrowski, K.; Hug, G. L.; Rozwadowski, J. J. Phys.
Chem. 1994, 98, 4854-4860.
(19) Bobrowski, K.; Marciniak, B.; Hug, G. L. J. Am. Chem. Soc. 1992,
114, 10279-10288.
(20) Marciniak, B.; Bobrowski, K.; Hug, G. L. J. Phys. Chem. 1993,
97, 11937-11943.
(21) Bobrowski, K.; Hug, G. L.; Marciniak, B.; Kozubek, H. J. Phys.
Chem. 1994, 98, 537-544.
(14) Lucia, L. A.; Burton, R. D.; Schanze, K. S. J. Phys. Chem. 1993,
97, 9078-9080.
(22) Marciniak, B.; Hug, G. L.; Bobrowski, K.; Kozubek, H. J. Phys.
Chem. 1995, 99, 13560-13568.

8002 J. Am. Chem. Soc., Vol. 119, No. 34, 1997
Bobrowski et al.
Table 1. Rate Constants for Quenching of the CB-Triplet by the Sulfur-Containing Alcohols in Aqueous Solution at pH ) 6.8 (Phosphate
Buffer)
alcohol
formula
H3CSCH2CH2OH
HOCH2CH2SCH2CH2OH
H3CSCH2CH2CH2OH
HOCH2CH2CH2SCH2CH2CH2OH
H3CSCH2CH2CH2CH2OH
abbreviation
kq × 10-9 a (M-1 s-1
)
2-(methylthio)ethanol
2-MTE
2,2′-DHE
3-MTP
3,3′-DHP
4-MTB
4.8 ( 0.1
3.3 ( 0.1
4.4 ( 0.2
3.6 ( 0.2
4.2 ( 0.1
2,2′-dihydroxydiethyl sulfide
3-(methylthio)propanol
3,3′-dihydroxydipropyl sulfide
4-(methylthio)butanol
a Errors are twice the standard deviations from the least-squares fits.
purifier (LabClear, Oakland, CA), and irradiated. The consumption
of CB and the formation of pinacol were determined by means of
HPLC, employing an SGE polymer-coated C18 column (4 × 250 mm),
an isocratic mobile phase consisting of 40% (v/v) acetonitrile and 0.1%
(v/v) trifluoroacetic acid, and UV detection at 254 nm. The quantum
yields of formaldehyde were determined by dividing the absolute yields
of formaldehyde by the initial yields of the CB triplet state. The yields
of the CB triplet state were deduced from the pinacol yields obtained
during photoirradiation of N2-saturated aqueous solutions containing
0.1 M 2-propanol in water and 2 × 10-2 M phosphate buffer, pH 7.0.
The quantum yields were calculated on the basis that each CB triplet
should yield 0.93 equiv of pinacol and lead to the consumption of 1.86
equiv of the CB ground state.
sulfides which catalyzes the oxidation under formation of
intramolecular sulfur-oxygen bonds.
Experimental Details
Materials. 4-Carboxybenzophenone (CB) and the hydroxyalkyl
sulfides, 2-(methylthio)ethanol (2-MTE), 2,2′-dihydroxydiethyl sulfide
(2,2′-DHE), 3-(methylthio)propanol (3-MTP), 3,3′-dihydroxydipropyl
sulfide (3,3′-DHP), and 4-(methylthio)butanol (4-MTB), obtained from
Aldrich Chemical Company, were of the purest commercially available
grade and were used as received. (2,4-Dinitrophenyl)hydrazine was
from Eastman Chemicals. Formaldehyde was purchased from Fisher
Chemical (Fair Lawn, NJ) as a 37% (w/w) aqueous solution. HPLC
grade acetonitrile was supplied by Fisher Chemical (Fair Lawn, NJ).
Water was purified by a Millipore Milli-Q system. Deuterium oxide
(D2O) was supplied by DOE, Savannah River State.
Laser Flash Photolysis Experiments. The nanosecond laser flash
photolysis setup has been described generaly elsewhere23,24 and also
as specifically configured in this work.25 The concentration of CB was
2 × 10-3 M in all experiments. The concentrations of alcohols were
in the range 2 × 10-5-2 × 10-4 M in the quenching experiments, but
the concentrations of these quenchers were 0.02 M for the quantum
yield determinations and for the experiments that involved recording
transient spectra in the time range following complete quenching
(>99.5%) of the CB triplet state. Transient absorption spectra and
quantum yield measurements were performed using a flow-cell system.
Typically five to seven laser shots were averaged for each kinetic trace.
The pH of the solutions was adjusted by adding sodium hydroxide and/
or HClO4. All solutions were deoxygenated by bubbling high-purity
argon through them.
The time of irradiation was chosen so that <10% CB conversion
occurred during the formation of H2CdO.
HPLC Analysis of Formaldehyde. The analysis of formaldehyde
was carried out by derivatization with (2,4-dinitrophenyl)hydrazine
followed by HPLC analysis according to two different procedures. A
1-mL aliquot of an irradiated solution was reacted for 10 min with 1
mL of a solution of 10-2 M (2,4-dinitrophenyl)hydrazine in 10-2
M
HCl. Subsequently the reaction mixture was extracted with 1 mL of
an n-hexane/CH2Cl2 (80:20, v/v) mixture by stirring for 5 min. The
HPLC analysis was performed employing a Shimadzu HPLC, equipped
with a UV detector. The organic layer, containing the hydrazone, was
separated from the aqueous phase, and a 5 µL aliquot was injected
into an HPLC column (SGE Hypersil C18, 5 micron, 250 × 4.6 mm)
which was eluted isocratically with acetonitrile/water, 60:40 (v/v), at
a flow rate of 1 mL/min. The hydrazones were monitored by UV
detection at 345 nm. Under the HPLC conditions the formaldehyde
hydrazone eluted with tR ) 3.9 min. In a second procedure, 20 µL of
the reaction mixture were directly (without extraction) injected into a
Varian RPC 18 column (5 micron, 4.6 × 150 mm) which was eluted
isocratically with a mobile phase consisting of 40:60 acetonitrile/water
(v/v), containing 0.1% trifluoroacetic acid. Under these conditions the
retention time of the formaldehyde hydrazone was 9.5 min. Both
methods gave identical results. Calibration was achieved by conducting
the same procedure with authentic formaldehyde samples.
Pulse Radiolysis Experiments. Pulse radiolysis experiments were
performed with 10 ns pulses of high-energy electrons from the Notre
Dame 7 MeV ARCO LP-7 linear accelerator and the Notre Dame Titan
Beta Model TBS-8/16-1S electron linear accelerator. Absorbed doses
were on the order of 4-6 Gy (1 Gy ) 1 J/kg). N2O-saturated 10-2
M
solutions of potassium thiocyanate were used as the dosimeter, taking
a radiation chemical yield of 6.13 and a molar extinction coefficient
of 7580 M-1 cm-1. A description of the pulse radiolysis setup, data
collection system, and details of dosimetry has been reported else-
where.26,27 The data acquisition subsystem has been updated28 and
includes a Spex 270M monochromator, a LeCroy 7200A transient
digitizer, and a PC-AT compatible computer. The software was written
within LabWindows of National Instruments.28 The experiments were
carried out with continuously flowing solutions.
Steady-State Photolysis Experiments. The irradiations were
performed in quartz cells with a Rayonet photoreactor equipped with
300 nm lamps. A solution of ferrioxalate was used as an actinometer,29
and the intensity of the incident light was determined to be 6.29 ×
10-6 einstein dm-3 s-1. Solutions containing CB (5 × 10-4 M) and a
hydroxyalkyl sulfide (6 × 10-3 M) at appropriate pH’s were purged
with oxygen-free nitrogen, passed through an OxiClear disposable gas
Results
Rate Constants of Quenching the CB Triplet State by
Hydroxyalkyl Sulfides. Five hydroxyalkyl sulfides (Table 1)
were used as quenchers of the CB triplet state in aqueous
solution at pH 6.8 (phosphate buffer). The CB triplet state is
always in its carboxylate form at pH > 6.3 (pKa ) 4.5 for CB),
and the hydroxyalkyl sulfides are in their nonionic form at
neutral and alkaline pH.
The triplet quenching rate constants kq were measured by
monitoring the triplet-triplet (T-T) absorption decays (kobs
)
of CB at wavelengths of 420, 480, and 540 nm. The latter
corresponds to the spectral maximum of the CB triplet, but
interference by the growth of absorbing electron-transfer
intermediates made the other wavelengths more favorable for
monitoring triplet decays in most instances.
(23) Encinas, M. V.; Scaiano, J. C. J. Am. Chem. Soc. 1979, 101, 2146-
2152.
(24) Nagarajan, V.; Fessenden, R. W. J. Phys. Chem. 1985, 89, 2330-
2335.
(25) Hug, G. L.; Marciniak, B.; Bobrowski, K. J. Phys. Chem. 1996,
100, 14914-14921.
(26) Janata, E.; Schuler, R. H. J. Phys. Chem. 1982, 86, 2078-2084.
(27) Schuler, R. H. Chem. Educ. 1985, 2, 34-37.
(28) Wang, Y. 1993, Unpublished report.
The pseudo-first-order rate constants, kobs, were calculated
using eq I which takes into account a concomitant, underlying
first-order growth of a photoproduct’s absorption30
(29) Murov, S. L.; Carmichael, I.; Hug, G. L. Handbook of Photochem-
istry, 2nd ed.; Dekker: New York, 1993.
(30) Encinas, M. V.; Wagner, P. J.; Scaiano, J. C. J. Am. Chem. Soc.
1980, 102, 1357-1360.

One-Electron Oxidation of â-, γ-, and δ-Hydroxyalkyl Sulfides
A - A∞
J. Am. Chem. Soc., Vol. 119, No. 34, 1997 8003
>•S-OH + S< a (>S S<)+ + OH-
(10)
ln
) -kobs
t
(I)
(
)
A0 - A∞
The yields of (S S)+ dimeric radical cations were assumed
to be equal to the yields of the primary >•S-OH adducts where
these yields are equal to 80% of the total available •OH radicals
(Gtotal ) 3.0 radicals per 100 eV absorbed).3,33 The molar
absorption coefficients reported in Table 2 are based on 10
separate determinations at the spectral maxima in question and
are based on G80% ) 2.4 radicals per 100 eV of energy absorbed
in N2-saturated solutions of the hydroxyalkyl sulfides at pH 1.
They compare well with the ꢀ’s (4000-9000 M-1 cm-1) of other
(R2S SR2)+ radical cations.4,31,33-37 Only for 3,3′-DHP was
the molar absorption coefficient corrected for contributions from
the (•S-OH)+ transient (vide infra).
where A0, A, A∞ are the absorbance changes at time 0, t, and
infinity, respectively.
The quenching rate constants, kq, were determined from a
linear least-squares fit of kobs vs [Q] plots employing the pseudo-
first-order relation
kobs ) (τT)-1 + kq[Q]
(II)
where τT is the lifetime of the CB triplet in the absence of
quencher. The resulting quenching rate constants obtained for
all the hydroxyalkyl sulfides are listed in Table 1 along with
the estimated uncertainties in the measurements.
The spectra of (•S-O) transients were obtained from N2O-
saturated aqueous solutions of 3-MTP and 3,3′-DHP (1 × 10-4
M) at pH 6.8. The higher pH and relatively low concentrations
of the hydroxyalkyl sulfides were used to avoid competitive
production of (S S)+ radical cations that might be formed in
competition with the intramolecularly bonded (•S-O) radicals,4
as shown in reaction 8. Under these experimental conditions,
the radiation chemical yield of the hydroxyl radicals, available
for reaction with the thioethers, was G(•OH) ) 5.3 (based on
the formula given by Schuler et al.38 which relates the G-value
Pulse Radiolysis. Pulse radiolysis experiments were per-
formed in order to obtain spectra of the isolated dimeric radical
cations (S S)+ and the various radicals with sulfur-oxygen
bonds. Experiments were done with each of the appropriate
hydroxyalkyl sulfides. Both types of sulfur-centered radicals
are possible electron-transfer products from the quenching
reactions of the CB triplet with the various hydroxyalkyl
sulfides. The spectra of these radicals were then used as possible
components in the analysis of the transient spectra following
the triplet-quenching events (vide infra).
•
of solute radicals generated by OH radicals to the product of
the rate constant for the reaction of •OH with the solute and the
solute concentration). Assuming the yield of intramolecularly
sulfur-oxygen bonded radicals to be equal to the yield of the
>•S-OH adduct () G80%), G ≈ 4.25 can be estimated. Molar
absorption coefficients based on this yield are reported in Table
2.
(S S)+ dimeric radical cations were generated in N2-saturated
solutions with high concentrations, 0.01-0.1 M, of the hy-
droxyalkyl sulfides at pH 1. A solution of low pH (acidified
with perchloric acid) was used to ensure a high concentration
of the protons to promote the elimination of water3,31 according
to reaction 7.
As reported previously,4,39 the oxidation of the alcohols
containing the thioether functionality leads to a radical species
which exists in acid-base equilibrium, as shown in reaction 8.
The acidic forms of the (•S-O) radicals formed from 3-MTP
and 3,3′-DHP absorb at λmax ) 390 nm with ꢀ’s of ≈1000 and
>•S-OH + H+ f >S•+ + H2O
(7)
These conditions were employed in order to severely limit
competing reactions of the initial hydroxysulfuranyl radicals
>•S-OH which can lead to other transients.13 In particular,
the excess protons would favor reaction 7 over reaction 4
followed by reactions 5 and 6 (for 2-MTE and 2,2′-DHE).
Likewise, reaction 7 would be favored over reaction 8 (for
3-MTP and 3,3′-DHP).
1450 M-1
cm-1. Their basic forms absorb at λmax ) 400 nm
with ꢀ ) 3250 and 2700 M-1 cm-1, respectively. For the
present discussion it is noteworthy that the acidity of the
hydroxyl group in the radicals with sulfur-oxygen bonds is
greatly enhanced over that of the unoxidized hydroxyalkyl
sulfides (see Discussion).
Transient Absorption Spectra of Intermediates during
Photolysis. Aqueous solutions of 4-carboxybenzophenone and
the various hydroxyalkyl sulfides that were used in the flash
photolysis and steady-state photolysis experiments were first
examined spectroscopically for any evidence of ground-state
association. The absorption spectra of these mixtures were
shown to be equal to those expected by adding the spectra from
separate solutions of ketone and quencher. No evidence for
ground-state association was found under the experimental
conditions used.
In the following, the type of radical in the middle structure of
reaction 8 will be referred to as (•S-OH)+, whereas the
deprotonated type on the far right-hand side of reaction 8 will
be referred to as (•S-O). (In model systems, the S-O bond in
the protonated radical species appears to be closer to being
hydroxysulfuranyl in nature as opposed to a two-center three-
electron bond.32)
High concentrations of the hydroxyalkyl sulfides facilitated
the formation of dimer radical cations through reaction of a >S•+
radical with a nonoxidized hydroxyalkyl sulfide (reaction 9)
The fate of the reactions after the laser flash were followed
by analyzing the optical absorption spectra of the transients. A
flow system was used, and individual kinetic traces were
(33) Hiller, K.-O.; Masloch, B.; Go¨bl, M.; Asmus, K.-D. J. Am. Chem.
Soc. 1981, 103, 2734-2743.
(34) Go¨bl, M.; Bonifacic, M.; Asmus, K.-D. J. Am. Chem. Soc. 1984,
106, 5984-5988.
>S•+ + S< a (>S S<)+
(9)
(35) Mohan, H.; Mittal, J. P. Radiat. Phys. Chem. 1991, 38, 45-50.
(36) Mohan, H.; Mittal, J. P. J. Chem. Soc., Perkin Trans. 2 1992, 207-
212.
and through the water-assisted reaction13 of the hydroxysul-
furanyl radical (>•S-OH) with a nonoxidized hydroxyalkyl
sulfide (reaction 10).
(37) Chaudhri, S. A.; Mohan, H.; Anklam, E.; Asmus, K.-D. J. Chem.
Soc., Perkin Trans. 2 1996, 383-390.
(38) Schuler, R. H.; Hartzell, A. L.; Behar, B. J. Phys. Chem. 1981, 85,
192-199.
(31) Bonifacic, M.; Mo¨ckel, H.; Bahnemann, D.; Asmus, K.-D. J. Chem.
Soc., Perkin Trans. 2 1975, 675-685.
(32) Carmichael, I. 1997, private communication.
(39) Glass, R. S.; Hojjatie, M.; Petsom, A.; Wilson, G. S.; Go¨bl, M.;
Mahling, S.; Asmus, K.-D. Phosphorus Sulfur 1985, 23, 143-168.

8004 J. Am. Chem. Soc., Vol. 119, No. 34, 1997
Bobrowski et al.
Table 2. Spectral Data of Intermolecular (S S)+ Type Dimeric
Radical Cations and Intramolecular (•S-O) Type Radicals for the
Sulfur-Containing Alcohols in Aqueous Solutions Obtained by the
Pulse Radiolysis Technique
(S S)+ a
(•S-O)d
alcohol
λ
max (nm) ꢀ (M-1 cm-1
)
λ
max (nm) ꢀ (M-1 cm-1
)
2-MTE
2,2′-DHE
3-MTP
3,3′-DHP
4-MTB
475
500
480
500
480
8100b
8100b
8400b
7300c
8250b
e
e
400
400
g
e
e
3250f
2700f
g
a [2-MTE, 3-MTP, 3,3′-DHP, 4-MTB] ) 0.01 M; [2,2′-DHE] ) 0.1
M, pH ) 1, N2-saturated. b Taking G(S S)+ ) 2.4 which is 80% of
•
the available OH radicals, total G(•OH) ) 3.0, standard deviation of
measurements (10%. c Calculated value taking G(•OH) ) 2.4 (80%
•
of the available OH radicals) as the sum of (S S)+ and (•S-O)
formation and assuming that the spectra of (S S)+ for 3,3′-DHP and
2,2′-DHE have the same shape (see text). d [3-MTP, 3,3′-DHP] ) 10-4
M, pH 6.8, N2O-saturated. e Cannot form stable cyclic (•S-O)-bonded
intermediates. f Taking G(•S-O) ) 4.25 which is 80% of the available
•OH radicals, total G(•OH) ) 5.3; standard deviation of measurements
(10%. g Does not form stable (•S-O)-bonded intermediates (see text).
collected at 10 nm intervals. Time-resolved spectra were
generated from these sets of traces, choosing appropriate time
delays and appropriate time windows. Spectra, so generated,
were resolved into component transients by a linear regression
technique40 of the form
n
∆A(λ ) ) ꢀ (λ )R
i
(III)
∑
j
i
j
i)1
where ∆A(λj) is the observed absorbance change of the
composite spectrum and ꢀi(λj) is the molar absorption coefficient
of the ith species at the jth wavelength of observation. The
linear regression coefficients, ai, correspond to cil (where ci is
the concentration of the ith transient and l is the optical
pathlength of the monitoring light). Further details of this
method have been described elsewhere.20 The spectra of the
CB triplet, CB ketyl radicals CBH•, and CB ketyl radical anions
CB•- were taken from previous work22,41 with renormalized
values of the molar absorption coefficients (ꢀ535 ) 6250 M-1
cm-1, ꢀ570 ) 5200 M-1 cm-1, and ꢀ660 ) 7660 M-1 cm-1 for
3CB*, CBH•, and CB•-, respectively).42,43 The component
spectra {ꢀi(λj)} of the transients from the hydroxyalkyl sulfides
were generated by pulse radiolysis (as mentioned above) for
this work.
Flash excitation of a solution of 4-carboxybenzophenone (2
× 10-3 M) and the various hydroxyalkyl sulfides (0.02 M)
resulted in the appearance of the absorption corresponding to
various transients depending on the time delay, the structure of
the hydroxyalkyl sulfide used, and the pH of the solution. High
concentrations of the hydroxyalkyl sulfides were used to rapidly
quench more than 99% of the CB triplets.
Figure 1. Resolution of the spectral components in the transient
absorption spectra following the quenching of CB-triplet by 2-MTE
(0.02 M) at pH 6.9 (without buffer) taken (a) 1.2 µs, (b) 14.1 µs and
(c) 31.6 µs after the flash.
(CBH•) of CB, and that at 480 nm is assigned to the
intermolecularly S S-bonded radical cation.
Spectra taken near pH 10 show a transient absorbing at 660
nm which is assigned to the ketyl radical anion (CB•-). For
3
the quenching of CB* by the hydroxyalkyl sulfides 3-MTP
and 3,3′-DHP, one additional intermediate could be identified
from the composite spectra. As shown in Figure 2, the presence
of an intermediate absorbing in the short-wavelength region
around 400 nm is required to get a reasonably good composite
fit with the experimental data at pH 6.8 for 3-MTP.
That intermediate is suggested to be an intramolecularly (•S-
O)-bonded species (see species on the far right of reaction 8).
Formation of this intermediate requires interaction of the
unpaired electron of the sulfur with a free electron pair of the
oxygen atom of the hydroxyl group. Its formation is assisted
by a suitable conformational arrangement (five-membered ring).
Intramolecular (•S-O)-bonded radicals derived from 3-MTP
and 3,3′-DHP (see structures in reaction 8) were characterized
in the Pulse Radiolysis section above. The presence of
intramolecular (•S-O)-bonded radicals was previously con-
firmed in pulse radiolysis2-5,44,45 and laser flash photolysis18
experiments with sulfur-containing acids and amino acids. At
pH ) 9.8 resolutions of the experimental transient absorption
spectra also led to the presence of the intermolecular S S-
bonded radical cations and the intramolecular (•S-O)-bonded
radicals, but at this pH both species had much shorter lifetimes
than did the corresponding species at lower pH.
When the CB triplet was quenched by the hydroxyalkyl
sulfides, the resulting transient absorption spectra had spectral
features reminiscent of the CB-triplet, the ketyl radical anion,
and the ketyl radical. A typical transient absorption spectrum
of this type, obtained for 2-MTE at pH 6.9, is presented in Figure
1. The transient at 570 nm is associated with the ketyl radical
(40) Bevington, P. R. Data Reduction and Error Analysis for the Physical
Sciences; McGraw-Hill: New York, 1969.
(41) Inbar, S.; Linschitz, H.; Cohen, S. G. J. Am. Chem. Soc. 1981, 103,
7323-7328.
(42) Hurley, J. K.; Sinai, N.; Linschitz, H. Photochem. Photobiol. 1983,
38, 9-14.
(43) Hurley, J. K.; Linschitz, H.; Treinin, A. J. Phys. Chem. 1988, 92,
5151-5159.
(44) Mohan, H. J. Chem. Soc., Perkin Trans. 2 1990, 1821-1824.
(45) Bobrowski, K.; Holcman, J. J. Phys. Chem. 1989, 93, 6381-6387.

One-Electron Oxidation of â-, γ-, and δ-Hydroxyalkyl Sulfides
J. Am. Chem. Soc., Vol. 119, No. 34, 1997 8005
Φ ) ∆ApꢀT/∆ATꢀp (IV)
where ∆AT is the absorption change in the actinometer im-
mediately after the flash due to the CB triplet at 535 nm
measured under the condition of no quenching and ∆Ap is the
product’s absorption change extrapolated to the “end-of-pulse”
value under the condition of nearly complete quenching. ꢀT
and ꢀp are the corresponding molar absorption coefficients. The
molar absorption coefficients ꢀT of the CB triplet and ꢀp for
CBH• and CB•- radicals were normalized22,43 as mentioned
above. The values of ꢀp for the (S S)+ dimeric radical cations
derived from all the hydroxyalkyl sulfides, and the (•S-O)-
bonded radicals of 3-MTP and 3,3′-DHP were taken from Table
2.
Since the absorption spectra of CBH•, CB•-, (S S)+ dimers,
and (•S-O)-bonded intermediates overlap, it was necessary to
resolve the spectra before the quantum yields could be com-
puted. Time-resolved transient spectra were generated from
kinetic traces collected at 37 distinct wavelengths, as described
in the preceding section. Spectral resolutions were made at the
appropriate points in time. From this procedure, a concentration
profile is produced for each of the transients, because each
spectral resolution yields one complete set of transient concen-
trations at a particular time. The primary quantum yields could
be then estimated by extrapolating the concentration profiles
to zero time. For each quantum yield, five concentrations were
determined at uniform time increments over an approximately
1 µs time interval after the laser flash. The concentrations were
then extrapolated back to the flash by a least-squares linear fit
to the five concentrations giving extrapolated concentrations,
c0. From eq IV, the quantum yields of the transient are then
simply
Figure 2. Resolution of the spectral components in the transient
absorption spectra following the quenching of CB-triplet by 3-MTP
(0.02 M) at pH 6.8 (without buffer) taken (a) 780 ns, (b) 11.7 µs, and
(c) 31.2 µs after the flash.
Φ ) c0/cT
(V)
where cT is the initial triplet concentration in the actinometer.
The resulting quantum yields are shown in Table 3.
In the case of the hydroxyalkyl sulfide 4-MTB, the composite
transient spectra at pH 6.1 and 10.0 (Figure 3) can be resolved
into a contribution from the intermolecularly (S S)-bonded
radical cation (λmax ) 480 nm) along with a contribution from
the ketyl radicals and ketyl radical anions. Since no extra
transition band was needed to fit the transient spectra even at
short wavelengths, the presence of an intramolecularly (•S-
O)-bonded species was ruled out, although the six-membered
ring structure should be thermodynamically possible.
Quantum Yields for Formation of Ketyl Radicals, Ketyl
Radical Anions, (S S)+ Radical Cations, and (•S-O)-
Bonded Radicals. The quantum yields for the transients could
be obtained in most circumstances. These experiments were
done under conditions where the quenching of the triplet of CB
was complete (>99%), as calculated from the quenching rate
constants in Table 1. Since the intersystem crossing yield of
CB is unity,41 the efficiency of a particular process in the
quenching event is equal to the quantum yield of the process.
To measure the quantum yields, the relative actinometry
method46,47 was used. For this application, matched optically
flat cells containing equal concentrations of CB were used. In
the cell used as the actinometer, no quenchers were present,
whereas in the sample cell the quencher concentration was
typically 20 mM. In the actinometer, the CB triplet’s absor-
bance was monitored at 535 nm immediately after the flash.
The quantum yields were calculated according to eq IV
This procedure was more elaborate than the previous
method17-19,21 of simply solving eq III for n transients using n
wavelengths. The more elaborate technique was used, because
it was found that when the (•S-O)-bonded transient was present,
the n-wavelength method was very sensitive to the absorbance
at 400 nm. To compensate for this sensitivity, the spectral
resolution method of determining the transients’ concentrations
uses the whole spread of wavelengths. For cases not involving
(•S-O), both the n-wavelength method and the spectral resolu-
tion technique gave almost identical results. The spectral
resolution method was used to obtain all reported quantum yields
in Table 3.
Supplementary measurements of the quantum yields for all
species (CBH•, CB•-, (S S)+, and (•S-O)) were made in D2O.
(For simplicity, CBH• will be used for both the deuteronated
and protonated ketyl radical.) Under these conditions the
alcohol groups of all the hydroxyalkyl sulfides studied undergo
instantaneous H/D exchange with the D2O solvent (confirmed
by 1H-NMR as described in ref 13). All quantum yields
measured in D2O are summarized in Table 3.
Similarities with and Contrasts to Dimethyl Sulfide. Since
neighboring groups actively participate in the one-electron
photooxidation of organic thioethers,18,19,21,22,48 we have em-
ployed dimethyl sulfide (DMS) as a reference system for the
hydroxyalkyl sulfides. Such an approach has two important
advantages for this type of mechanistic study. (i) The initial
step during one-electron oxidation is expected to be the same
(46) Lutz, H.; Breheret, E.; Lindqvist, L. J. Phys. Chem. 1973, 77, 1758-
1762.
(47) Carmichael, I.; Hug, G. L. J. Phys. Chem. Ref. Data 1986, 15,
1-250.
(48) Hug, G. L.; Marciniak, B.; Bobrowski, K. J. Photochem. Photobiol.
A 1996, 95, 81-88.

8006 J. Am. Chem. Soc., Vol. 119, No. 34, 1997
Bobrowski et al.
Figure 3. Resolution of the spectral components in the transient absorption spectra following the quenching of CB-triplet by 4-MTB (0.02 M) at
pH 6.3 (without buffer) taken (a) 1.6 µs, (b) 53 µs after the flash, and at pH 10.1 (c) 780 ns after the flash and (d) 53 µs after the flash.
Table 3. Quantum Yields of Product Formation in the CB-Sensitized Photooxidation of Sulfur-Containing Alcohols in Neutral (pH 6.7) and
Alkaline (pH 10) Solutionsa
b,c
c,d
d
d
d
•
•-
•
•-
+
•
alcohol
2-MTE
solvent
ΦCBH +CB
ΦCBH
ΦCB
Φ(S
Φ( S-O)
S)
H2O
D2O
H2O
D2O
H2O
D2O
H2O
D2O
H2O
D2O
0.29 (0.31)
0.28 (0.31)
0.35 (0.37)
0.35
0.58 (0.55)
0.47 (0.51)
0.62
0.17 (0.19)
0.19 (0.22)
0.29 (0.30)
0.31
0.35 (0.33)
0.26 (0.28)
0.46
0.12 (0.12)
0.09 (0.09)
0.06 (0.07)
0.04
0.23 (0.22)
0.21 (0.23)
0.16
0.11 (0.11)
0.08 (0.09)
0.05 (0.06)
0.04
0.21 (0.21)
0.22 (0.26)
0.22
e
e
e
e
2,2′-DHE
3-MTP
0.32 (0.32)
0.16 (0.13)
0.50
0.34
e
3,3′-DHP
4-MTB
0.51
0.33
0.28
0.35
0.20
0.19
0.16
0.13
0.09
0.24
0.12
0.08
e
a Numbers in parantheses are for alkaline solutions. b Sum of ΦCBH + ΦCB
.
c Deuteronated ketyl radical formed in D2O. d Based on spectral
•
•-
mix (see text). e Not observed (see text).
Table 4. Quantum Yields of Product Formation in the
CB-Sensitized Photooxidation of Dimethyl Sulfide (DMS) in
Neutral and Alkaline Solutions
Table 5. Decay Rate Constants of (S S)+ Radical Cations of
2-MTE Following the Quenching of CB Triplets
solvent
pH/pD
buffer
kfasta (s-1
)
a
b
b
b
+
•
•-
•
•-
pH
6.7
9.9
solvent
ΦCBH +CB
ΦCBH
ΦCB
Φ(S
S)
H2O
D2O
H2O
D2O
H2O
D2O
6.82
6.97
8.38
8.55
9.91
10.0
none
none
none
none
none
none
8.7 × 104
6.7 × 104
5.5 × 104
4.8 × 104
1.3 × 105
1.0 × 105
H2O
D2O
H2O
D2O
0.29
0.23
0.30
0.25
0.12
0.13
0.11
0.13
0.17
0.10
0.19
0.12
0.20
0.12
0.22
0.15
a Sum of ΦCBH + ΦCB
.
b Based on spectral mix (see text).
•
•-
a Fast component from biexponential fit to concentration profiles
from spectral resolutions.
in DMS and the hydroxyalkyl sulfides, i.e., formation of the
charge-transfer complex, involving a sulfur of the thioether
group and an oxygen of the carbonyl group of CB. (ii) DMS
lacks other functionalities which could affect reactions within
the CT-complex.17 Thus, any differences in quantum yields of
the quenching products observed for the hydroxyalkyl sulfides
as compared to DMS can then be ascribed to an involvement
of the hydroxyl group(s). With DMS as the CB-triplet quencher,
measurements were made of the quantum yields for all the
transient species (CBH•, CB•-, and (S S)+) in H2O and D2O.
The values are summarized in Table 4.
above). The only difference was that the time scale needed to
be expanded to get a reasonable sampling of points for the
relatively complex decay curves. For this, about 25 time points
were needed. Three different fitting functions were used: single
exponential, biexponential, and mixed (parallel first- and second-
order decay). The single exponential fits were poor, but there
was not much difference in the goodness-of-fit, ør2, between
the biexponential and mixed fits. The fast components of the
biexponential decays of the (S S)+ radical cations of 2-MTE
are listed in Table 5. The values are tabulated so as to show
any variation with pH or deuteration of the solvent. The first-
order component of the mixed decays differed in absolute
magnitude from the fast components of the biexponential fits,
Decay Kinetics of (S S)+. The decay of the (S S)+-
bonded intermediates was followed using the same time-resolved
spectral-resolution methods as those used for determining the
concentrations of transients for the primary quantum yields (see

One-Electron Oxidation of â-, γ-, and δ-Hydroxyalkyl Sulfides
J. Am. Chem. Soc., Vol. 119, No. 34, 1997 8007
In contrast, pulse-radiolytically generated •OH-adducts of
2-MTE and 2,2′-DHE do not yield measurable yields of (S S)+
radical cations in neutral and alkaline solutions.13 These •OH-
adducts, thus, appear to react by a different mechanism which
confirms our previous assessment that they may decompose by
the elimination of water. The water could be formed through
a rapid intramolecular proton or hydrogen transfer from the
adjacent hydroxyl group. The latter mechanism leads to the
formation of (alkylthio)ethoxy radicals which subsequently
undergo R,â-fragmentation into formaldehyde and an R-(alkyl-
thio)alkyl radical (see reaction 5). This will be discussed further
below.
3. Mechanism of Formation of Transients: Isotope
Effects. In general, the formation of the three-electron bonded
dimers [R2S SR2]+ in triplet-quenching events involves reac-
tions 11 and 12, followed by association of the generated radical
cation R2S•+ with a second nonoxidized sulfur (reaction 12).
However, since we observe solvent isotope effects on the yields
of radical cations that vary with the nature of the thioether, it
appears that the mechanism of reaction is more complex. In
the following, different possible mechanisms will be considered
for each of the sulfides investigated. The isotope effects will
be rationalized qualitatively rather than quantitatively since
geometries and charge distributions in transition state(s) and/or
intermediates are at present largely unknown.
Table 6. Quantum Yields Formaldehyde Formation in the
CB-Sensitized Photooxidation of Sulfur-Containing Alcohols in
Neutral and Alkaline Solution
alcohol
solvent
buffer
pH 6.7
pH 8.0
pH 10
2-MTE
2-MTE
2,2′-DHE
2,2′-DHE
H2O
D2O
H2O
D2O
none
none
none
none
0.052
0.021
0.074
0.021
0.111
0.019
0.070
0.009
0.096
0.025
0.068
0.015
but the trends in variation with the two solvent parameters (pH
and deuteration) were similar.
Formaldehyde Quantum Yields. Quantum yields of form-
aldehyde (H2CdO) formation were measured in the steady-state
irradiation of CB solutions (0.5 mM) with 2-MTE and 2,2′-
DHE present at a concentration of 6 mM at various pH’s. These
quantum yields are listed in Table 6.
Discussion
1. Quenching Rate Constants. The rate constants kq for
quenching of the CB triplet state by hydroxyalkyl sulfides (Table
1) are high (approaching one-third of diffusion controlled) and
comparable with those for sulfur-containing carboxylic acids,18
amino acids,19,20 and methionine-containing peptides.22 This
would put the quenching rate constants at the plateau value of
Rehm-Weller plots, corresponding to diffusion-limited reac-
tions. On the other hand, for aliphatic alcohols, which lack a
sulfur moiety, the kq’s are 4 orders of magnitude lower49 than
the kq’s obtained for the sulfur-containing alcohols. This
supports the idea that the quenching of the CB triplet state by
hydroxyalkyl sulfides involves an interaction at the sulfur atom.
The mechanism for this triplet quenching process by other
sulfur-containing organic compounds was established previously
as an electron transfer from the sulfur atom to the triplet species.
2. Transient Spectra. The observed electron-transfer
intermediates included CB•-, CBH•, intermolecularly (S S)-
bonded radical cations, and, in the case of 3-MTP and 3,3′-
DHP, intramolecularly (•S O)-bonded radicals. 3-MTP and
3,3′-DHP differ from the other alcohols in this work in that
they can form conformationally favorable five-membered (•S-
O)-bonded ring structures. This permits the interaction of an
unpaired electron on the sulfur with a nonbonded electron pair
on the oxygen atom of the hydroxyl group. The CB•-, CBH•,
and (S S)+ radical cations of 2-MTE13 and 2,2′-DHE35 transient
spectra have also been previously reported in the literature.
Resolutions of the transient spectra from flash photolysis
demonstrated that formation of the (S S)+ radical cations of
2-MTE and 2,2′-DHE is not restricted to the low pH region.
In the following discussion, the “product” species will usually
refer to the ion-pair (IP) state. The total secondary kinetic
solvent isotope effect can be written as50,51
φR
∏
i
kH
i
)
(VI)
kD
φIP
∏
j
j
where φR and φIP represent the fractionation factors of various
hydrogenic sites in the reactant form and in the ion pair,
respectively. The isotopic fractionation factor of a particular
hydrogenic site in a chemical species is defined as the ratio of
that site’s preference for deuterium over protium relative to the
similar preference for a single hydrogenic site in a solvent
molecule. The isotopic fractionation factor of a hydrogenic site
attached to a neutral oxygen is defined as one. Some other
common examples are φ ) 0.69 for hydrogenic sites attached
to a positively-charged oxygen and φ ) 0.5 for hydrogenic sites
attached to a negatively-charged oxygen.50
Dimethyl Sulfide (DMS). For better comparison between
the schemes, all reactions in the schemes have been assigned
numbers which directly relate them to the schemes (i.e., 1.1,
1.2 for Scheme 1; 2.1, 2.2 for Scheme 2; etc.). The first step
in the one-electron oxidation process of DMS represents the
formation of charge-transfer complex 3a, shown in reaction 1.1
in Scheme 1. In the course of reaction 1.1 two neutral reactants
(neglecting the 4-carboxy group of CB which does not
participate in the reaction) associate to yield the dipolar charge-
transfer complex 3a. According to the mechanism previously
proposed,17 complex 3a either suffers back electron transfer to
yield ground state CB and DMS (reaction 1.2, Scheme 1),
undergoes hydrogen transfer (reaction 1.3, Scheme 1), or
diffuses apart to yield product radical ions (>S•+ and CB•-).
The yield of the solvated product [(H3C)2S S(CH3)2]+ (2a)
shows a product solvent isotope effect of [2a]H O/[2a]D O ) 1.7
•
This is in contrast to the OH-induced oxidation of these
hydroxyalkyl sulfides where (S S)+ radical cations were only
observed at pH e 2.13 Data generated from flash photolysis
are displayed in Figure 1 and in Table 3 which show that the
(S S)+ radical cations of 2-MTE and 2,2′-DHE are formed
even in the pH range 6.9-10. The explanation for this
contrasting behavior lies in that the CB-sensitized oxidation of
2-MTE and 2,2′-DHE leads initially to the sulfur-centered
radical cations >S•+
3CB* + R2S f CB•- + R2S•+ (1)
(11)
which subsequently associate with an unoxidized hydroxyalkyl
sulfide, yielding the (S S)+ radical cations.
2
2
(50) Schowen, R. L. In Progress in Physical Organic Chemistry;
Streitwieser, A., Jr.; Taft, R. W., Eds.; Wiley: New York, 1972; Vol. 9, pp
275-332.
(51) Schowen, K. B. J. In Transition States of Biological Processes;
Gandour, R. D.; Schowen, R. L., Eds.; Plenum Press: 1978; pp 225-283.
R2S•+ + SR2 a [R2S SR2]+ (2)
(12)
(49) Scaiano, J. C. J. Photochem. 1973, 2, 81-118.

8008 J. Am. Chem. Soc., Vol. 119, No. 34, 1997
Bobrowski et al.
Scheme 1
a nonoxidized sulfide to yield the dimeric sulfide radical cation
2a (reaction 1.6, Scheme 1), which is observable through its
spectrum with λmax ) 465 nm. Since there is no rate
determining proton transfer step associated with reactions 1.4
and 1.5, eq VI50,51 predicts a normal solvent isotope effect of
kH/kD > 1.0 for the overall electron-transfer process. Such a
prediction is based on the competition of reactions 1.4 and 1.5
with reactions 1.2 and 1.3 where we do not expect any solvent
isotope effect on the individual reactions 1.2 and 1.3. This is
confirmed for reaction 1.3 by the lack of a solvent isotope effect
on the ketyl radical yield (CBH•) (Table 4).
As a result of the solvation, the fractionation factors of the
hydrogen atoms of the solvating water molecules will change
from φR ) 1.0 to φIP < 1.0. It is difficult to predict the
magnitude of the expected solvent isotope effect since we do
not know (i) the partial charges δ- on CB and δ+ on the sulfide
in the solvated ion pair 4a, (ii) the partial charges on the
solvating water molecules, and (iii) the fractionation factor of
the proton shared between water and the negatively charged
oxygen atom of the developing ketyl radical anion, CBδ-, in
the solvated ion pair 4a.
2-(Methylthio)ethanol (2-MTE) and 2,2′-Dihydroxydiethyl
Sulfide (2,2′-DHE). The formation of the radical cation
complexes of 2-MTE, 2b, and 2,2′-DHE, 2c, shows product
Scheme 2
isotope effects of [2b]H O/[2b]D O ) 1.4 and [2c]H O/[2c]D O
)
2
2
2
2
) 1.25, respectively (calculated from Table 3). They are smaller
than the corresponding product isotope effect of the (S S)+
yield in DMS of 1.7. This can be rationalized on the basis of
a mechanism displayed in Scheme 2.
The first steps of the reaction of 3CB* with the hydroxyalkyl
sulfides 2-MTE and 2,2′-DHE are exactly analogous to those
involving DMS and yield the charge-transfer complexes 3b/3c
(reaction 2.1, Scheme 2). Subsequently, also in analogy to
DMS, solvation of the charge-transfer complex 3b/3c (reaction
2.4, Scheme 2) competes with back electron transfer (reaction
2.2) and hydrogen transfer (reaction 2.3). However, there is
an additional channel open to the charge-transfer complexes 3b/
3c in comparison to 3a in the DMS case. For 3b/3c, instead of
solvating through the interaction of an oxygen from a solvent
molecule, the sulfur atom in the developing radical cation may
also interact with the oxygen of the terminal hydroxy group(s)
(5b/5c) of 2-MTE or 2,2′-DHE, (reaction 2.7, Scheme 2).
Separated radical ions can escape (reaction 2.8, Scheme 2) from
5b/5c to yield 1b/1c. Species 1b/1c may also directly result
from the separation of the solvated ion pair 4b/4c by reaction
2.5. An alternate fate of 5b/5c is a proton-transfer step within
the complex 5b/5c and escape of the resulting radicals by
reaction 2.9 in Scheme 2. This proton transfer and the escape
of CBH• may be particularly important in the pH range below
the pKa ) 8.241 of CBH•.
The key to understanding the differences between the solvent
isotope effects seen in the two thioethanols and DMS lies in
the competing reaction paths 2.4 vs 2.7 in Scheme 2.
A
comparison of the structures of 5b/5c with those of 4a, 4b, and
4c reveals that 5b/5c has one less proton on the oxygen directly
associated with the sulfur. Since the fractionation factor for a
at pH 6.7 (calculated from Table 4). This is a rather surprising
observation since products and reactants in the reaction leading
to radical ions do not have exchangeable protons. This
observation can be rationalized by postulating the formation of
the solvated ion pair (4a) (reaction 1.4, Scheme 1). The degree
of solvation will determine the probability of a successful
separation of the ion pair 4a into solvated product ions >S•+
(1a) and CB•- (reaction 1.5, Scheme 1) which is in competition
to reactions 1.2 and 1.3 of Scheme 1. It is noteworthy that the
product solvent isotope effect for CB•- at pH 6.7 was similar
to that for the (S S)+ transient ([CB•-]H O/[CB•-]D O ) 1.7,
proton is φH ) 0.69, the denominator ∏jφIP in eq VI should
+
j
be larger for the formation (reaction 2.7, Scheme 2) of 5b/5c
(one proton) than for the formation (reaction 2.4, Scheme 2) of
4b/4c (two protons). With a larger denominator in eq VI for
reaction path 2.7, Scheme 2, its solvent isotope effect should
be smaller than that of reaction path 2.4. Therefore, to the extent
that the reaction channels 2.4 and 2.7 are effectively competing,
the overall solvent isotope effect should be smaller than going
exclusively through reaction path 2.4 which is the analog to
2
2
calculated from Table 4). Subsequently, 1a will associate with

One-Electron Oxidation of â-, γ-, and δ-Hydroxyalkyl Sulfides
J. Am. Chem. Soc., Vol. 119, No. 34, 1997 8009
the only path available to DMS in reaction 1.4, Scheme 1. In a
similar line of argument, the extra hydroxyethyl group in 2,2′-
DHE is likely to enhance the competition of reaction 2.7 vs
reaction 2.4 in comparison to 2-MTE and, thus, lower the overall
solvent isotope effect of 2,2′-DHE compared to 2-MTE. Both
of these expectations are verified in the experiments. These
conclusions are based on the assumption that the positive partial
charge on the sulfur and the oxygen atoms are similar for 4a,
4b/4c, and 5b/5c.
Scheme 3
3-(Methylthio)propanol (3-MTP) and 3,3′-Dihydroxy-
dipropyl Sulfide (3,3′-DHP). For 3-MTP and 3,3′-DHP, there
is no solvent isotope effect, within experimental error, on the
yields of sulfide radical cations 2d/2e. However, there is a
significant normal solvent isotope effect on the yields of a
sulfur-oxygen bonded intermediate 7d/7e (Scheme 3). The
solvent isotope effects in the product yields are [7d]H O/[7d]D O
2
2
) 2.0 for 3-MTP and [7e]H O/[7e]D O ) 1.5 for 3,3′-DHP
2
2
(calculated from Table 3). In principle, these isotope effects
can be rationalized by a mechanism displayed in Scheme 3. It
was proposed to be as closely analogous as possible to the
mechanism presented for the ethanol derivatives in Scheme 2.
However, certain critical differences apply which are of
significance in understanding solvent isotope effects on the
quantum yields of transients. In general, formation of five-
membered sulfur-oxygen (and sulfur-nitrogen) bonded inter-
mediates such as 6d/6e and 7d/7e is fast, leading to an
intramolecular bond formation involving the developing sulfide-
radical-cation moiety. In addition, these five-membered sulfide-
oxygen bonded intermediates are relatively stable for the
propanol derivatives, in contrast to the four-membered rings in
the ethanol derivatives.13 It is of particular significance to this
discussion that sulfur-oxygen transients involving hydroxyl
moieties can deprotonate4 and that the acidity of the hydroxyl
group is greatly enhanced over the hydroxyl group in the
unoxidized hydroxyalkyl sulfide. The sulfur-oxygen bonded
intermediates derived from 3-MTP (6d, 7d) and 3,3′-DHP (6e,
7e) exist in the acid-base equilibrium (reaction 3.13, Scheme
3) with pKa ≈ 4.7.52
This could be due to the competition with reaction 3.9 having
a large normal solvent isotope effect. The apparent small
inverse solvent isotope effect could arise if reaction 3.9
dominates for the H2O reaction, but for the D2O reactions,
reaction 3.9 is supressed relatively more than 3.4 or 3.11,
possibly making them competitive with 3.9.
It is of note that the solvent isotope effects for the propanol
derivatives are generally larger than for the ethanol derivatives,
i.e., compare 1.4 for 2-MTE vs 2.0 for 3-MTP and 1.25 for
2,2′-DHE vs 1.5 for 3,3′-DHP. If deprotonation would already
occur to a significant extent within the solvated ion pair 5d/5e,
we would expect larger normal isotope effects due to the fact
that the resulting proton H3O+ would appear in the denominator
3
2
+
In order to understand the solvent isotope effects, it is
important to note from Table 3 that ΦCBH• ≈ Φ(•S-O) and that
+
of eq VI as (φH ) in contrast to (φH ) for DMS (two water
+
protons in 4a) and (φH ) for the ethanol derivatives (one proton
S+
ΦCB•- ≈ ΦS
for the propanol derivatives. From this it is
of the terminal hydroxyl group in 5b/5c). An alternative
explanation for the larger isotope effects in the propanol
derivatives could be related to contributions from primary
isotope effects due to a rate-determining proton transfer (within
structures 5d/5e) in the course of reactions 3.9 and 3.14 of
Scheme 3 which directly lead to 7d/7e and formation of ketyl
radicals CBH•. This path would likely be enhanced over the
analogous path 2.9 of Scheme 2, because the fast formation
and stability of 5d/5e of the thiopropanols in comparison to
5b/5c of the thioethanols enables proton transfer within the ion
pairs.
reasonable to conclude that the formation processes of CBH•
and (•S-O) intermediates are coupled and that the formation
of (S S)+ intermediates is the main pathway of >S•+ radical
cation decay. For Scheme 3 to accommodate these observations,
certain conclusions can be drawn about the rate constants in
Scheme 3: (a) the equilibrium 3.12 between 6d/6e and 1d/1e
would have to be very slow, (b) reaction 3.11 would also be
slow, and (c) hydrogen-transfer reaction 3.3 from 3d/3e would
be slow. Such an alignment of equilibria and rate constants
S+
would account for ΦCBH• ≈ Φ(•S-O) and ΦCB•- ≈ ΦS
in
the propanol derivatives. Deviations from the equalities between
these quantum yields might be accounted for by some or all of
the rate processes in (a), (b), and (c) being nonzero, see below.
With such restrictions on the rate processes in Scheme 3, it
can be seen that CB•- and (S S)+ are formed mainly from the
sequence of reactions 3.4/3.5/3.6; whereas CBH• and (•S-O)
are mainly formed in the sequence of reactions 3.7/3.9/3.14.
The normal solvent isotope effect on the formation of CBH•
and (•S-O) would come through equilibrium 3.7 and/or a rate-
determining proton transfer in reaction 3.9, see calculation
below. On the other hand, the expected normal solvent isotope
effect on the formation of (S S)+ and CB•- was not observed.
4-(Methylthio)butanol. In general, the considerations for
2-MTE and 2,2′-DHE will also apply to 4-(methylthio)butanol
which shows a similar solvent isotope effect on the formation
of sulfide radical cations as 2-MTE.
4. Quantum Yields of Transients. Although the initial step
in the mechanisms (Schemes 2 and 3) is the same for all the
hydroxyalkyl sulfides, the primary quantum yields of the various
intermediates differ significantly among the hydroxyalkyl
sulfides. For 2-MTE, 2,2′-DHE, and 4-MTB, the total quantum
yields for ketyl radical (CBH•) and ketyl radical anion (CB•-)
formation are roughly half the analogous yields for 3-MTP and
3,3′-DHP (Table 3). According to the mechanisms in Schemes
2 and 3, these quantum yields represent the combined yields of
two processes, namely the diffusion apart of the charge-transfer
(52) Wisniowski, P.; Hug, G. L.; Bobrowski, K. Manuscript in prepara-
tion.

8010 J. Am. Chem. Soc., Vol. 119, No. 34, 1997
Bobrowski et al.
complex (radical ion escape) and the intramolecular proton
transfer within the complex. This means that the quantum yield
for back electron transfer would be about 60-70% for 2-MTE,
2,2′-DHE, and 4-MTB and 30-35% for 3-MTP and 3,3′-DHP.
The quantum yields indicate that, contrary to quenching events
of sulfur-containing amino acids19-21 and carboxylic acids,18
back electron transfer is an efficient process in quenching events
involving hydroxyalkyl sulfides (particularly for 2-MTE, 2,2′-
DHE, and 4-MTB). The efficiency of quenching by 2-MTE,
2,2′-DHE, and 4-MTB is comparable to the analogous quench-
ing by unsubstituted alkyl sulfides.17
Scheme 4
The individual contributions to the yield of primary photo-
products can be separately characterized as being due in part to
the escape of radical ions from the charge-transfer complex and
to a somewhat greater extent to a proton-transfer reaction within
the charge-transfer complex (see the values of ΦCBH• in Table
3). The quantum yields for escaping radical ions vary from
about 10% for 2-MTE, 2,2′-DHE, and 4-MTB to about 25%
S+
for 3-MTP and 3,3′-DHP (see the values of ΦCB•- and ΦS
in Table 3). There are significantly higher values of ΦCBH• for
3-MTP and 3,3′-DHP compared to the other sulfur-containing
alcohols. These higher yields can be explained by the ability
of 3-MTP and 3,3′-DHP to form intramolecularly (•S-O)-
bonded species. From Table 3, it can be seen that the expected
transfer) because from the analogous equation to eq VI50
KH/KD ) φ(HO-)φ(H2Oδ+ 2δ+/φ(8b/8c) )
(0.5)(0.69)2δ+/1<1 (VIII)
S+
equality ΦCB•- + ΦCBH• ) ΦS
+ Φ•S-O does not quite
)
hold for 3,3′-DHP. This may be due to an error in the smaller
than usual molar extinction coefficient of 7300 M-1 cm-1 for
the (S S)+ species of 3,3′-DHP.
where δ+ is the positive charge shared by the oxygen in
structure 2b/2c. In addition, reaction 4.17 has an experimentally
determined solvent isotope effect of unity.13 If reaction
sequence 4.18/4.19 would yield formaldehyde with higher
efficiency than reaction sequence 4.16/4.17, a change of the
solvent from H2O to D2O would cause a normal solvent isotope
effect on the formaldehyde yields. However, a more likely
contribution to the normal solvent isotope effect arises from
the fact that in D2O, due to the less efficient reaction 4.18
(includes a primary kinetic isotope effect), deprotonation of 2b/
2c’s CR-H bond (which does not exchange protons with the
solvent) may excel (reaction 4.20, Scheme 4). This process
does not lead to formaldehyde.
5. Mechanism of Carbon-Carbon Bond Fragmentation
for Radical Cations of 2-MTE and 2,2′-DHE. There are
several key features that are important to consider for a
discussion of the mechanism of carbon-carbon bond fragmen-
tations. (i) The initial yields of 2b, the (S S)+ radical cations
of 2-MTE, are twice as high as the initial yields of 2c, the
(S S)+ radical cations of 2,2′-DHE (Table 3). (ii) There are
small normal solvent isotope effects on the order of kH/kD
≈
1.3 on kfast (Table 5). (iii) A computation of various ratios, Ω,
of formaldehyde-to-(S S)+ quantum yields
Ω ) ΦH CO/Φ(S
(VII)
+
S)
2
At neutral pH, the base-catalyzed reactions 4.16 and 4.18
would have to proceed with water as the base. At pH <
pKa(CBH•) ) 8.2 the developing CB•- within the ion pair 5b/
5c (Scheme 2) could theoretically act as a base to accept protons
from the terminal hydroxyl group of 2-MTE and 2,2′-DHE
leading to CBH• and the monomeric radical cation -OCH2-
CH2S•+-R by reaction 2.9 of Scheme 2. This zwitterion is
susceptible to CR-Câ fragmentation analogous to reaction 4.19
for the dimerized radical zwitterion. These considerations, along
with the arguments in the preceding paragraph for the high pH
range, can explain the normal solvent isotope effect on
formaldehyde formation.
However, it should be noted from Scheme 4 that it is
estimated that 8b/8c forms formaldehyde in only 50% yield.
This is based on gamma radiolysis yields of formaldehyde
compared to •OH yields (or >•S-OH) in aqueous solutions of
2-MTE at pH 6.13 If the unknown reaction channel of 8b/8c
that does not yield formaldehyde is more efficient at pH 10,
then the inverse solvent isotope effect would be less competitive,
and the normal solvent effect could increase with pH as
observed. Another possible explanation for this phenomenon
is that the base-catalyzed reaction sequence 4.18/4.19 is more
efficient at high pH than the reaction 2.9 is at neutral pH.
The quantum yields of formaldehyde can in some cases
slightly exceed the initial quantum yields of (S S)+ as in the
neutral
H2O
for 2-MTE from Tables 3 and 6 give the following: Ω
)
neutral
D2O
pH10
H2O
pD10
) 0.28. For
D2O
pH10
0.47, Ω
2,2′-DHE, Ω
) 0.25, Ω
) 0.87, and Ω
neutral
neutral
) 1.5, Ω
) 0.53, and Ω
) 1.1. The
H2O
D2O
H2O
observation that the quantum yields of formaldehyde slightly
exceed the initial quantum yields of (S S)+ for 2,2′-DHE
requires some discussion (vide infra). (iv) From Table 6, it
can be seen that there is an increasing solvent isotope effect on
the formation of formaldehyde with increasing pH. For 2-MTE,
[H2CO]H O/[H2CO]D O ) 2.5 at pH 6.7 and [H2CO]H O/[H2-
2
2
2
CO]D O ) 3.8 at pH 10. Similarly for 2,2′-DHE, [H2CO]H O
[H2CO]D O ) 3.5 at pH 6.7 and [H2CO]H O/[H2CO]D O ) 4.5
/
2
2
2
2
2
at pH 10.
These observations can be partially rationalized by the
mechanisms in Schemes 2 and 4. Increasing yields of form-
aldehyde can be expected with increasing pH based on an
increasingly favorable competition of hydroxide ion attack (for
reactions 4.16 and 4.18, Scheme 4) with reverse electron transfer
(reaction 4.15, Scheme 4) which does not lead to formaldehyde.
At any given pH, formaldehyde will be produced either via
reaction sequence 4.16/4.17 or 4.18/4.19. By analogy to the
results of Baciocchi et al.15 and Lucia et al.,14 reaction 4.18 is
expected to have a normal primary solvent isotope effect
between 1.2 and 2.0. Reaction 4.16 should exhibit an inverse
secondary solvent isotope effect (no rate-determining proton

One-Electron Oxidation of â-, γ-, and δ-Hydroxyalkyl Sulfides
J. Am. Chem. Soc., Vol. 119, No. 34, 1997 8011
case 2c from 2,2′-DHE. These observations warrant some
discussion. The higher frequency of CR-Câ bond breakage for
2,2′-DHE produces larger quantities of R-(alkylthio)methyl
radicals. Disproportionation of the latter leads to additional
formaldehyde via reactions 13 and 14.
at neutral pH, solvated ion pairs 5b/5c can be formed which
can undergo transformation either to CBH• and RS•+CH2CH2O-
or 1b/1c and CB•- (Scheme 2). Since 1b/1c does not undergo
instantaneous fragmentation, it is possible to observe 2b/2c in
CB-photosensitized oxidation at all pH values and in pulse
radiolysis experiments at pH 1. The 1b/1c cannot be generated
by pulse radiolysis at neutral solution using the •OH radical as
an oxidant, which accounts for the lack of 2b/2c in pulse
radiolysis experiments at neutral pH.
How can we describe the potential hydrogen transfer process
mechanistically? In its reaction with 2-MTE, the hydroxyl
radical will be initially trapped by the sulfur atom but probably
exists in equilibrium 15.
2•CH2SR f -CH2SR + +CH2SR
(13)
(14)
+CH2SR + H2O f H+ + CH2O + RSH
Similar reactions also operate for 2-MTE. However, since the
total formaldehyde yields are lower than the initial yields of
2b, we cannot assess how much processes analogous to reactions
13 and 14 contribute to the formaldehyde yields in the 2-MTE
system.
CH3SCH2CH2OH + •OH a CH3S•(OH)CH2CH2OH (15)
•
6. Proton Transfer vs Hydrogen Transfer in the OH
Radical-Induced Oxidation Mechanism of 2-MTE and 2,2′-
DHE. We are now left with the question whether the hydroxyl-
radical-induced fragmentation of 2-MTE and 2,2′-DHE proceeds
via proton transfer and sulfide radical cations (reaction 4.17a,
Scheme 4) or via hydrogen transfer (reaction 4.17b, Scheme
4), as originally proposed.13 During the hydroxyl-radical-
induced fragmentation, less formaldehyde yields were observed
for 2,2′-DHE than for 2-MTE at pH 6.13 This is in contrast to
the one-electron oxidation-induced fragmentation at pH 6.7 in
this work. This feature alone may render some support for the
originally proposed hydrogen-transfer mechanism.
If a cyclic structure such as 8b were to exist, homolytic
dissociation (the reverse of the oxidation reaction) of the
hydroxysulfuranyl radical would yield a hydroxyl radical in the
immediate vicinity of the terminal hydroxyl group. In general,
hydrogen abstraction by •OH from hydroxyl groups is possible,53
and its probable occurrence would be enhanced by the spatial
closeness of the reactants in these circumstances.
Acknowledgment. The work described herein was supported
by the Office of Basic Energy Sciences of the U.S. Department
of Energy (K.B., G.L.H., and B.M.), the Committee of Scientific
Research, Poland (Grant No. 2 P303 049 06) (B. M. and K.
B.), a Self-Fellowship (B. M.), and the Association for
International Cancer Research (Ch.S.). The authors thank
Professor K.-D. Asmus and Dr. I. Carmichael for instructive
conversations. This paper is Document No. NDRL-3988 from
the Notre Dame Radiation Laboratory.
From the previous discussion it can be seen that both proton
transfer and hydrogen transfer appear to operate in the mech-
anisms of formaldehyde formation in the radiolytic and pho-
tolytic oxidation of the thioethanols. Which of the two processes
dominate depends on the nature of the oxidant, and, in particular,
it depends on the reactive intermediates involved in the oxidation
•
process. In pulse radiolysis experiments using OH radicals,
JA970683R
hydroxysulfuranyl radicals are formed, and both hydrogen-
transfer and proton-transfer mechanisms can occur. In contrast
in the CB-photosensitized oxidation of 2-MTE and 2,2′-DHE
(53) Asmus, K.-D.; Mo¨ckel, H.; Henglein, A. J. Phys. Chem. 1973, 77,
1218-1221.