Effect of nitroxide radicals on chemically induced dynamic electron polarization of spin-correlated radical pairs in aqueous micellar solutions of sodium dodecyl sulfate

Article abstract of DOI:10.1007/s11172-008-0184-7

The multispin systems consisting of spin-correlated radical pairs (SCRPs) and stable nitroxide radicals, localized in micelles of sodium dodecyl sulfate (SDS), were studied by ESR and pulse laser photolysis techniques. In all the systems studied, the stab

Full text of DOI:10.1007/s11172-008-0184-7

Russian Chemical Bulletin, International Edition, Vol. 57, No. 7, pp. 1416—1427, July, 2008
1416
Effect of nitroxide radicals on chemically induced dynamic electron
polarization of spinꢀcorrelated radical pairs in aqueous micellar
solutions of sodium dodecyl sulfate
V. R. Gorelik,a V. F. Tarasov,b S. R. Shakirov,a and E. G. Bagryanskayaa
aInternational Tomography Center, Siberian Branch of the Russian Academy of Sciences,
3a ul. Institutskaya, 630090 Novosibirsk, Russian Federation.
Fax: +7 (383) 333 1399
bN. N. Semenov Institute of Chemical Physics, Russian Academy of Sciences,
4 ul. Kosygina, 119991 Moscow, Russian Federation.
Fax: +7 (495) 938 2484. Eꢀmail: spinchem@chph.ras.ru
The multispin systems consisting of spinꢀcorrelated radical pairs (SCRPs) and stable nitroxide
radicals, localized in micelles of sodium dodecyl sulfate (SDS), were studied by ESR and pulse laser
photolysis techniques. In all the systems studied, the stable nitroxide radicals exert no effect on
the shape of the ESR spectra of the SCRPs (in particular, on the shape of their antiphase structure)
and on the decay kinetics of the ESR signal of the SCRPs. In the SDS micelles, the electron spin
polarization transfer from the nonequilibrium electron spin states of the molecular triplets (SCRP
precursors) is the most efficient mechanism of generation of the electron spin polarization in nitroxide
radicals. The experimental data also show that the nitroxide radicals and SCRP radicals are most
probably distributed uniformly in the micellar phase.
Key words: electron spin resonance, chemically induced dynamic electron polarization, micelles,
micellized radical pairs, spinꢀcorrelated radical pairs, electron spin polarization transfer.
Micelles represent an example of nonuniform suꢀ
translational diffusion coefficients obtained by the nuꢀ
merical simulation of the ESR spectra of the SCRPs with
analogous dependences of the correlation time of rotaꢀ
tional diffusion of various paramagnetic probes shows
that the both dependences are identical, i.e., the assumpꢀ
tion on the uniform diffusion of radicals does not contraꢀ
dict the experimental data. However, in this case,
the rotational correlation times are an averaged characꢀ
teristic of the whole micellar phase rather than, e.g., of
the Stern layer in which the paramagnetic probe is locaꢀ
lized.8 In addition, it can be assumed in the microreactor
theory that the SCRP radicals diffuse in certain concentꢀ
ric layers of micelles, for instance, in the Stern layer or
in the palisade layer. If the hydrophilicities of the probe
and SCRP partners are approximately equal, the coinciꢀ
dence of the aboveꢀmentioned results is quite expected
and cannot be interpreted in favor of this or another point
of view.
pramolecular media widely used in the applied and sciꢀ
entific areas.1 The micellar phase is inhomogeneous,1—4
which provides various models for solubility1—3,5 of orꢀ
ganic molecules in micelles. According to the most popꢀ
ular model, hydrophilic compounds are localized in the
Stern layer, whereas hydrophobic compounds are preꢀ
dominantly localized in the micelle core. Of course, the
assertion about a certain localization of molecules in the
micellar phase is idealization, because any molecule inꢀ
corporated into the micellar phase diffuses. An opposite
case of idealization is an assumption that the diffusion
coefficient of a dissolved molecule is independent of its
position in the micellar phase.6 ESR spectroscopy of
stable nitroxide radicals is widely used to study the local
molecular mobility of various media,7 including micelꢀ
lar media.8 The interpretation of ESR spectra8—10 deꢀ
pends substantially on the accepted model of paraꢀ
magnetic probe distribution in the nonuniform micelꢀ
lar phase.
In the present work, we suggest another approach to
the problem of investigation of the localization of organic
molecules and radicals in the micellar phase. Populaꢀ
tions of the electron spin states of excited molecular
triplets are known to be primarily nonequilibrium. When
colliding with nitroxide radicals, spinꢀpolarized triplets
can transfer their electron spin polarization to the radiꢀ
The microreactor model proposed11,12 for the deꢀ
scription of spin selective processes in micellized spinꢀ
correlated radical pairs (SCRPs) assumes the uniform
diffusion of radicals over the whole micelle volume.
A comparison of the temperature dependences13 of the
Published in Russian in Izvestiya Akademii Nauk. Seriya Khimicheskaya, No. 7, pp. 1390—1401, July, 2008.
1066ꢀ5285/08/5707ꢀ1416 © 2008 Springer Science+Business Media, Inc.

CIDEP in the presence of nitroxide radicals
Russ.Chem.Bull., Int.Ed., Vol. 57, No. 7, July, 2008
1417
cals. This mechanism of chemically induced dynamic
electron polarization (CIDEP) is named the electron
spin polarization transfer (ESPT).14 The absence of poꢀ
larization transfer in systems for which the latter is effiꢀ
cient in uniform solvents could indicate the specific loꢀ
calization of organic molecules (triplets) and radicals in
the micellar phase, which prevents collisions of the tripꢀ
letꢀexcited molecules with the radicals.
The electron spin polarization transfer in the micellar
phase was first observed in Ref. 15. It was speculated that
the polarization of the nitroxide radicals appears due to
the spin polarization transfer from the net polarized raꢀ
dicals formed upon the photodissociation of 1ꢀhydroxyꢀ
cyclohexyl phenyl ketone.
Another purpose of the presented study was the folꢀ
lowing. If the micelle containing the SCRPs includes the
nitroxide radical, then the relatively longꢀlived threeꢀ
spin system is formed. Such threeꢀspin systems are very
interesting from the viewpoint of spin communication,
and one of its variants has earlier been discussed in deꢀ
tail.16 Moreover, these systems correspond to the criteria
of applicability of Purtov´s theory17 for the spin dynaꢀ
mics of threeꢀspin systems to a considerably greater extent
than the systems organized by radical pairs (RPs) and
the nitroxide radical in usual homogeneous organic solꢀ
vents. It is important that the increase in the probability
of recombination of triplet RPs in the zero magnetic field
induced by the nitroxide radicals was observed18 just in
the latter systems. Later this phenomenon was named
spin catalysis.19 It would be very informative to observe
the effect of nitroxide radicals on ESR spectra of SCRPs
from the viewpoint of both the applicability and testing
conditions for the existing theories of spin catalysis and
the formulation of experimental conditions that are most
adequate to these theories.
the sensitivity of the method is insufficient. Detected difꢀ
ferences due to CIDEP are typically of the two types. The
first variant takes place when the spin states
|αβ;χ> and |βα;χ> of the pair (see Fig. 1) are populaꢀ
ted equally, while the populations of the spin states
|αα;χ> and |ββ;χ> are different. This situation appears
when CIDEP in the SCRPs is determined by the triplet
mechanism.20,21 In this case, all the spectral lines in the
TR ESR spectrum are polarized equally, regardless of the
configuration of magnetic spins of the pair. If the populaꢀ
tions of the levels |αα;χ> and |ββ;χ> are the same and
the populations of the levels |αβ;χ> and |βα;χ> are differꢀ
ent, the SCRPs are polarized via the radicalꢀpair SТ0
mechanism.21 In this case, the sign of polarization of the
lowꢀfield parts of the ESR spectrum is opposite to that of
the highꢀfield parts.
The relaxation shifts of the resonance frequencies22
(ΔS and ΔT, see Fig. 1) appear due to the stochastic moduꢀ
lation of the spinꢀspin exchange interaction in the SCRPs
induced by radical diffusion. Let us call this relaxation
type the internal relaxation. The shift of the resonance
frequencies results in the situation when no compensaꢀ
tion (for example, transitions |αα;χ> ↔ |βα;χ> and
|αβ;χ> ↔ |ββ;χ> in radical А) occurs even if the spin states
r
a
T+χ
T0χ
T–χ
αα;χ
TBχ
SBχ
SAχ
αβ;χ
βα;χ
TAχ
ββ;χ
Main characteristics of ESR spectra of SCRPs. Timeꢀ
resolved ESR spectroscopy (TR ESR) of SCRPs was not
practically discussed in the Russian scientific literature.
Therefore, we will briefly present the main characterisꢀ
tics of the TR ESR spectra of the SCRPs detected in
the time interval from several tens of nanoseconds to
several tens of microseconds.
Sχ
TA
b
2/TS
ΔS
TB
In liquid lowꢀviscosity solvents generation of CIDEP
is mainly dictated by the electron exchange interaction,
whereas the dipole spinꢀspin interaction is usually neꢀ
glected. The energy levels of the RPs calculated in
the highꢀfield approximation with the spinꢀspin exchange
interaction between the radicals exponentially decaying
with an increase in the r distance between them for
an arbitrary configuration of the magnetic nuclear spins χ
(χ = χAχB, where χA and χB are the nuclear spin configuraꢀ
tions of radicals A and B of the pair) are shown in Fig. 1.
The Boltzmann differences in the populations of the RP
spin states are not usually detected, because the magnetic
field modulation cannot be used and, as a consequence,
SB
ΔT
ωA
ωB
2/TT
SA
Fig. 1. Diagram of the energy states of the SCRPs at different
distances r between the pair radicals A and B (a) and the idealized
TR ESR spectrum of the SCRPs at the spinꢀadiabatic occupation of
the SCRPs and the equilibrium triplet state of the precursor (b);
and Δ are the spectral shifts of the S and T components of
the APS, respectively; 2/TS and 2/TT are their widths.
Δ
S
T

1418
Russ.Chem.Bull., Int.Ed., Vol. 57, No. 7, July, 2008
Gorelik et al.
|αα;χ>, |ββ;χ> and |αβ;χ>, |βα;χ> are occupied equally in
pairs. The appearance of shifts is due to the spinꢀcorrelaꢀ
ted state of the pair, i.e., to the fact that at any moment the
spin state of the SCRP is the coherent superposition of
the basis spin states. In the absence of net polarization,
the superposition of two spectral lines for the case of slow
exchange results in the signal, whose shape coincides
with that of the signal of the first derivative of the absorpꢀ
tion signal. This shape of the resonance signal is named
the antiphase structure (APS).23 Spectral shifts can subꢀ
stantially exceed the linewidths within the fast exchange
when the resonance lines are exchangeꢀnarrowed (preꢀ
cisely this limit was used for the numerical simulation of
the ESR spectrum of the SCRPs in Fig. 1). The latter case
is often observed for the spinꢀcorrelated biradicals.24
At the positive (negative) difference in the energies of
the spin states |αβ;χ> and |βα;χ>, these states correlate22
with the triplet |T0> (singlet |S>) and singlet |S> (tripꢀ
let |T0>) states of the contact pair, respectively. If the
first case takes place, the component of the APS
Results and Discussion
Based on the model "similarꢀinꢀsimilar," we can asꢀ
sert that TEMPO should be localized near the Stern layer
with the NO group "immersed" into this layer.8—10 Acꢀ
cording to the concepts of localization, the TENIO radiꢀ
cal should be situated closer to the center of the micelles,
perhaps, in the palisade layer due to the hydrophobic
character of four ethyl groups. The NTR1 and NTR2
nitroxide radicals are the hydrophilic positively charged
radical ions. Therefore, they should be localized in the
aqueous layer that is directly adjacent to the Stern layer.
|αα;χ>
↔ |βα;χ> is named singlet (S) and the
|ββ;χ> ↔ |αβ;χ> — component is named triplet (T). Just
this case is presented in Fig. 1 (ΔS and ΔT are the resoꢀ
nance shifts of the singlet and triplet components, respecꢀ
tively). For the slow exchange (the resonance shifts of the
components are smaller than their widths 2/TS and 2/TT),
the assignment of the ESR signals to an individual radical
still has the physical sense. For instance, the S compoꢀ
nent |αα;χ> ↔ |βα;χ> belongs to radical А.
In the cases when the difference between the resoꢀ
nance frequencies is high, the highꢀfield approximation
can be invalid and the difference in the populations of
the RP spin states depends substantially on the nuclear
configuration. This mechanism is named the radicalꢀpair
SТ mechanism.21 The ESR spectra of the micellar
SCRPs in which the SТ– mechanism is efficient have
been analyzed25 in detail.
(2,4,6ꢀTrimethylbenzoyl)diphenylphosphine oxide
(TMBDPO), 2,2ꢀdimethoxyꢀ2ꢀphenylacetophenone (Irgaꢀ
kur 651), [13СО]ꢀα,α´ꢀdimethyldeoxybenzoin (DMDB),
2ꢀmethylꢀ1,4ꢀnaphthoquinone (MNQ), benzophenone,
xanthone, and benzyl are localized in the micellar phase,
which follows from the optical absorption spectra. Keꢀ
tones TMBDPO,26 Irgakur 651,27 and DMDB
(see Ref. 28) photodissociate in the triplet excited state
to form the triplet SCRPs (Scheme 1) consisting
of 2,4,6ꢀtrimethylbenzoyl and phosphinoyl radicals
(SCRP1), benzoyl and α,α´ꢀdimethoxybenzyl radicals
(SCRP2), and benzoyl and cumyl radicals (SCRP3),
respectively.
Experimental
In all cases, we used 0.1 Мaqueous micellar solutions of sodium doꢀ
decyl sulfate (SDS). The micelle concentration was ~10–3 mol L–1
.
Some uncertainty in the micelle concentration is related to
the dependence of the aggregation number on the solubilizate
concentration.
Stable nitroxide radicals, viz., 2,2,6,6ꢀtetramethylpiperidinꢀ1ꢀ
oxyl (TEMPO), 3,4ꢀdimethylꢀ2,2,5,5ꢀtetraethylimidazolinꢀ1ꢀoxyl
(TENIO), ammonium N,N,Nꢀtrimethyl(2,2,6,6ꢀtetramethylꢀ1ꢀ
oxypiperidinꢀ4ꢀyl) iodide, and ammonium N,N,Nꢀtrimethyl(2,2,6,6ꢀ
tetramethylꢀ1ꢀoxypiperidinꢀ4ꢀyl)methyl sulfate (NTR1 and NTR2,
respectively) were used.
Aqueous micellar solutions of the ketones under study were
irradiated in the presence and absence of the nitroxide radicals
directly in the ESR cavity with the light of an NdꢀYAG laser
(λ = 355 nm). ESR spectra were acquired at the outlet of a preꢀ
amplifier (12 MHz) using a LeCray digital oscillograph.
Benzophenone29 and MNQ30 are photoreduced in
SDS micelles to form the SCRP4 and SCRP5 of the corꢀ
responding ketyl radical and alkyl radical of the detergent.
As follows from the studies of benzophenone photoreꢀ
duction in homogeneous solutions of dodecyl alcohol31
in acetonitrile and in SDS micelles,32 hydrogen atom
abstraction from different СН2 fragments of the hydroꢀ
carbon chain of both the alcohol and detergent occurs
with approximately the same efficiency, except for the
–
—СН2—OSO3 fragment, which is stable to photooxidaꢀ
tion. This indirectly indicates in favor of the uniform
distribution of benzophenone over the whole micellar

CIDEP in the presence of nitroxide radicals
Russ.Chem.Bull., Int.Ed., Vol. 57, No. 7, July, 2008
1419
Scheme 1
Diꢀtertꢀbutylphenol dissolved in the micellar phase was
used for xanthone as a hydrogen atom donor. In this
case,33 the SCRP6 consists of the ketyl radical of xanꢀ
thone and the phenoxy radical of diꢀtertꢀbutylphenol
(Scheme 2).
TR ESR of the SCRPs in micelles in the absence of
stable nitroxide radicals. The TR ESR spectrum of the
SCRP1 (Fig. 2) is formed by two signals.25 The doublet
with a splitting of 36.5 mT and g = 2.0036 belongs to
the phosphinoyl radical (see Scheme 1). The central sigꢀ
nal is assigned to the 2,4,6ꢀtrimethylbenzoyl radical
(g = 2.0007).
SCRP1
Initially (see Fig. 2, a) the both components of the
doublet and the central signal are positively polarized.
Since none of the APS components is emission, the poꢀ
pulations of the "internal" spin states |αβ;χ> and |βα;χ> are
lower than that of the spin state |ββ;χ> and higher than
that of the |αα;χ> state. In the considered case, the intense
polarization of the radicals via the triplet mechanism is
a consequence of high anisotropy of occupation of the
spin states34 and the high dissociation rate35 of the triplet.
As the observation time tobs increases, the intensity of
the lowꢀfield component (mZ(31P) = +1/2) of the doublet
signal decreases substantially more rapidly than the inꢀ
tensity of the highꢀfield component (mZ(31P) = –1/2).
This occurs because the internal relaxation rate in the
subsystem with the overall spin Fz = –1/2 (two electron
spins plus the 31P nuclear spin) is much higher21,25 than
that in the subsystem with Fz = +1/2. As a result, at
tobs > 250 ns the emissionꢀpolarized S component of
the APS appears in the lowꢀfield and central signals
(see Fig. 2, b). The T component of the APS of the
highꢀfield signal becomes negative considerably later
(see Fig. 2, c) at tobs > 500 ns. To this time, the emissive S
component of APS predominates in the lowꢀfield signal,
because to this moment the total population of the subꢀ
system Fz = –1/2 becomes less than the population of the
|αα;α> spin state. This evolution of the spectral shape is
possible only in the case when the internal relaxation rate
exceeds the rate of external one.
SCRP2
SCRP3
phase. No similar detailed experimental studies were carꢀ
ried out for MNQ; however, there is no evidence to beꢀ
lieve that the photooxidation of the detergent moleꢀ
cules by this quinone is more stereospecific.
SCRP4
SCRP5
Scheme 2
Products

1420
Russ.Chem.Bull., Int.Ed., Vol. 57, No. 7, July, 2008
Gorelik et al.
mZ(31P) = +1/2
mZ(31P) = –1/2
time (see Fig. 3, b). It is well seen that the APS compoꢀ
nents have different resonance frequencies, i.e., the obꢀ
servation of the APS is due, in fact, to the spectral shifts
caused by the internal spin relaxation because of the
electron spin Heisenberg exchange between unpaired
electrons of the radicals.
A
a
E
A
Since the HFC constant on the carbonyl carbon atom
(А(13С=О) = 12.5 mT) in the benzoyl radical substanꢀ
tially exceeds the collision frequencies of the SCRP3
radicals in SDS micelles,28 the spectral lines of the
SCRP3, by analogy to the SCRP1, can be attributed to
individual radicals. The ESR spectrum of this pair is
shown in Fig. 4. The doublet signal (g = 2.0006), each
component of which is antiphaseꢀsplit, is assigned to
the benzoyl radical (see Scheme 1). The multiplet signal
(g = 2.0026) in the center of the spectrum belongs to the
cumyl radical. Since the dissociation rate of DMDB in
the tripletꢀexcited state is lower than the spin relaxation
rate, the contribution of the triplet mechanism to the
electron polarization of the SCRP3 radicals is small. In
this case, the initial population of the "internal" spin
states |αβ,χ> and |βα,χ> (see Fig. 1) is twofold lower than
the populations of the "external" spin states |αα,χ> and
|ββ,χ>. The high difference between the populations and
rather large spectral shifts of the S and T components of
the APS provide the high APS intensity in the TR ESR
spectra of the SCRP3 due to the high HFC constant. In
spite of the high HFC constant in the benzoyl radical,
the contribution of the radicalꢀpair ST– mechanism proꢀ
portional to the squared HFC constant (i.e., by an order
of magnitude lower than that in the case of the SCRP1) is
small in the magnetic field of the spectrometer. As
b
c
E
A
2 mT
E
B
Fig. 2. The TR ESR spectra of the SCRP1 observed upon the
photodissociation of TMBDPO in SDS micelles at tobs = 120 (a),
290 (b), and 480 ns (c). The intensities of the spectra are amplified
by 3 (b) and 12 times (c); А is adsorption (absorption, positive ESR
signal), and E is emission (radiation, negative ESR signal).
The evolution of the shape of the lowꢀfield APS is
shown in Fig. 3. For the short observation time, the T
component prevails (see Fig. 3, a), whereas the S compoꢀ
nent of the APS predominates at the long observation
a result, the APS shape is timeꢀindependent. At long tobs
,
a
b
the signal of the benzoyl radical is a superposition of the
APS and the signal of the free (escaped from the SCRP)
radical (the doublet of the poorly resolved triplets). It is
important that the ESR linewidth of the free radical
mZ(13CO) = +1/2
mZ(13CO) = –1/2
Z
c
2 mT
B
4 mT
B
Fig. 3. The lowꢀfield component (mZ(31P) = +1/2) in the ESR
spectrum of the phosphinoyl radical at tobs = 136 (a), 312 (b), and
500 ns (c). The triplet (a) and singlet (c) components of the APS
have different resonance frequencies.
Fig. 4. The ESR spectra of the SCRP3 observed upon the photolysis
of DMDB in SDS micelles at tobs = 500 ns (solid line) and 1.6 μs
(dotted line, the intensity of the spectrum is amplified by 1.8 times).

CIDEP in the presence of nitroxide radicals
Russ.Chem.Bull., Int.Ed., Vol. 57, No. 7, July, 2008
1421
(see Fig. 4, dotted line) is much narrower than the APS
components of the same radical in the SCRP3 composiꢀ
tion. The same is observed for the SCRP1. In this case,
both the free radicals and SCRP radicals are localized in
the micellar phase. Thus, in the SCRP1 and SCRP3 the
rate of external relaxation induced by the anisotropic
HFC is much lower than the internal relaxation rate
induced by the exchange interaction. Thus, the division
of the total width of the APS into additive contributions is
not quite valid, because Redfield´s theory of external
relaxation is applicable for the SCRP case only within the
slow and weak exchange.
In the case of the SCRP4 and SCRP5 formed upon the
photoreduction of benzophenone and MNQ, the ESR
spectra (Fig. 6, a and b) are approximately the same,
representing a superposition of the antiphaseꢀsplit ESR
spectra of the alkyl radicals (products of SDS photooxiꢀ
dation) and the ESR spectra of the corresponding ketyl
radicals of benzophenone and MNQ. For the both
SCRPs, the exchange is slow for the external components
of the spectrum of the alkyl radicals (the corresponding
differences in the resonance frequencies exceed the colliꢀ
sion frequencies of the radicals) and fast for the cenꢀ
tral part.22
In the case of the SCRP2 (Fig. 5), the HFC constants
in the benzoyl and α,α´ꢀdimethoxybenzyl radicals are
low and do not exceed 0.6 mT. That is why, the electron
spin exchange in the SCRP2 is fast for the majority of
nuclear configurations. Therefore, it is impossible to asꢀ
sign the ESR signals to any individual radical of the pair.
However, the times of external transverse relaxation in
the SCRP2 radicals lie in the microsecond range. This
results in a monotonic change in the spectral linewidths,
depending on the observation time, which impedes an
analysis and interpretation of the ESR spectrum of
the SCRP2. The radical pairs are negatively polarized
according to the triplet mechanism, i.e., the populaꢀ
tions of the triplet levels |αα;χ> exceed those of the
|ββ;χ> levels. Because of the weak intensity of the HFC,
the spectral shifts of different APS components are
low and the rates of the processes that proceed via the
radicalꢀpair SТ_ mechanism are negligible. Due to
this, the external relaxation rate is faster than the rate
of equalization of the populations of the "internal"
spin states of the SCRP2. As a result, the SCRP2 have no
APS, whose components possess opposite signs of poꢀ
larization.
Since the photoreduction rate in the micellar phase is
rather low (~107 s–1 for benzophenone), the contribution
of the triplet polarization mechanism to the ESR spectra
of these pairs is small. Therefore, the pairs have the
distinctly pronounced APS (see Fig. 6, dashed lines). At
long time delays (more than 2 μs), the signals of the free
alkyl radicals polarized via the radicalꢀpair ST0 mechaꢀ
nism predominate in the spectrum.
The TR ESR spectra of the SCRP6 formed upon the
photooxidation of diꢀtertꢀbutylphenol with xanthone in
SDS micelles are presented in Fig. 7, a. The spectra were
obtained at tobs = 560 ns and represent a superposition of
the unresolved ESR signals distributed in a narrow interꢀ
val of resonance fields. Due to this, the spin exchange is
a
a
4 mT
B
b
E
b
×2
0.1 mT
B
Fig. 5. The TR ESR spectra of the SCRP2 observed upon the
photolysis of 2,2ꢀdimethoxyꢀ2ꢀphenylacetophenone in SDS miꢀ
celles at tobs = 260 ns (a) and 1.12 μs (b).
Fig. 6. The TR ESR spectra of the SCRP4 (a) and SCRP5 (b) at
CTEMPO = 1 (a) and 2 mmol L–1 (b). Dashed lines show the spectra
of the SCRP4 and SCRP5 in the absence of TEMPO; tobs = 320 ns.

1422
Russ.Chem.Bull., Int.Ed., Vol. 57, No. 7, July, 2008
a
Gorelik et al.
A
Effect of the stable nitroxide radicals on the CIDEP of
the SCRPs in micelles. The addition of stable nitroxide
radicals to aqueous micellar SDS solutions of the consiꢀ
dered ketones provides two main effects: CIDEP generaꢀ
tion in the stable nitroxide radicals and a decrease in the
intensity of the TR ESR signals of the SCRPs (or in other
words, CIDEP quenching in the SCRP radicals). It is
important to underline several factors.
E
b
1. We found indications to the specificity of localizaꢀ
tion of the ketones and radicals, including the stable
nitroxide radicals, in none of the experimental examples.
In all experimental systems, CIDEP generation in the
nitroxide radicals occurs with approximately the same
efficiency.
2. The TEMPO radical is chemically active during
the photolysis of all ketones under study. This is exꢀ
pressed in the fact that TEMPO is consumed either in
the reactions with tripletꢀexcited ketone or with radicals
of the corresponding SCRPs.
1 mT
B
Fig. 7. a. The TR ESR spectrum observed upon the photolysis of
xanthone (2 mmol L–1) in aqueous micellar solutions of SDS
and diꢀtertꢀbutylphenol (2 mmol L–1) in the presence of TEMPO
(2 mmol L–1) (solid line) and absence of TEMPO (dashed line).
b. Difference in the TR ESR spectra in the presence and absence
of TEMPO.
3. The shape of the TR ESR spectra of the SCRPs is
independent of the presence of TEMPO. This is valid for
both the SCRPs in which no APS is observed (Fig. 9, a
(SCRP1 at short observation times) and Fig. 10 (SCRP2))
and the SCRPs in which the APS predominates in
the spectrum (see Fig. 9, b (the lowꢀfield component of
the TR ESR spectrum of the SCRP1 at intermediate
tobs values), Fig. 11 (SCRP3 at tobs < 1 μs), Fig. 6, a and b
(SCRP4 and SCRP5, respectively)). In the case of
the SCRP6, the shape of the complicated ESR signal
also remains unchanged, because the observed signal
(see Fig. 7) is a superposition of the ESR signals of the
SCRP6 and TEMPO.
4. A comparison of the linewidths of the polarꢀ
ized TEMPO radical at low concentrations, when the
occupation of micelle is less than unity, shows that
the linewidth of the polarized radical is equal to that of
the thermolyzed TEMPO. In other words, no additionꢀ
al broadening of the resonance lines occurs for the
TEMPO. At high TEMPO concentrations, the ESR lineꢀ
widths of the spinꢀpolarized nitroxide radicals increase
with an increase in the TEMPO concentration in micelꢀ
lar SDS solutions substantially more slowly than the linꢀ
ewidths of the equilibrium radicals, which is caused,
most likely, by a decrease in the TEMPO concentration
during photolysis.
fast. The presence of the APS is proved exclusively by the
numerical simulation of the spectra. Since the rate of
2ꢀbutylphenol photooxidation is much higher than that of
the photooxidation of the detergent molecules with benꢀ
zophenone or MNQ, the contribution of the net polarizaꢀ
tion via the triplet mechanism to the CIDEP of the SCRP6
is noticeably higher than that for the SCRP4 and SCRP5.
The emissive and poor TR ESR spectrum observed for
the laser photolysis of benzyl in SDS micelles is shown in
Fig. 8 (curve 1). The spectrum contains no characteristic
signals of the alkyl radicals of SDS. Presently we cannot
identify the observed signal. Probably, it belongs to the
triplet of benzyl but, perhaps, to the radical products of
photooxidation of unknown admixtures. Further we will
neglect this weak signal and believe that the photolysis of
benzyl in SDS micelles is not accompanied by SCRP
formation.
1
5. The kinetics of the TR ESR intensities of the SCRP1
normalized to the maximal intensity is shown in Fig. 12.
The normalized decay kinetics of the TR ESR signals
(CIDEP quenching kinetics) is independent of the presꢀ
ence of TEMPO (see Fig. 12). This is valid for other
studied SCRPs. The listed conclusions also concern the
systems consisting of the SCRP1 and stable radicals
TENIO, NTR1, and NTR2.
2
2 mT
B
Fig. 8. The TR ESR spectrum observed upon the laser photolysis of
benzyl (2 mmol L–1) in aqueous SDS micelles in the absence (1)
and presence of TEMPO (1 mmol L–1) (2).
To explain CIDEP generation in the nitroxide radiꢀ
cals in micellar solutions during the photolysis of alkylꢀ

CIDEP in the presence of nitroxide radicals
Russ.Chem.Bull., Int.Ed., Vol. 57, No. 7, July, 2008
1423
a
A
E
b
c
4 mT
B
Fig. 9. The CIDEP of TEMPO upon the photolysis of TMBDPO (6 mmol L–1) in SDS micelles at CTEMPO = 1.5 (solid line) and 6 mmol L–1
(dotted line) and tobs = 120 (a), 290 (b), and 490 ns (c).
aromatic ketones, it was suggested15 that the nonequilibriꢀ
um magnetization is transferred from the SCRP radiꢀ
cals to the nitroxide radicals due to the spinꢀspin exꢀ
change interaction between the radicals of the pair and
the nitroxide radical. However, this explanation is not
general as minimum, because the CIDEP of the nitroxꢀ
ide radicals is also observed when no RPs are formed in
the micelles. In fact, the laser photolysis of benzyl in SDS
micelles produces the broad and lowꢀintensity ESR sigꢀ
nal (see Fig. 8, curve 1) with g = 2.0045 at the intensity
maximum, which possibly belongs to tripletꢀexcited benꢀ
zyl. At the same time, the addition of TEMPO to micellar
solutions of benzyl results in the easily detected signal of
the negatively polarized TEMPO radicals (see Fig. 8,
curve 2). Analogously, the strong emissive net polarizaꢀ
tion of TEMPO was observed36 for the photolysis of
a
c
2
3
2
E
1
2
b
3
1
1
0.5 mT
B
2 mT
B
Fig. 10. The CIDEP of TEMPO upon the photolysis of 2,2ꢀdimethoxyꢀ2ꢀphenylacetophenone (6 mmol L–1) in SDS micelles at
CTEMPO = 2 (1), 4 (2), and 6 mmol L–1 (3) and tobs = 260 ns (a) and 1.12 μs (b). c. Fragments of the TR ESR spectra observed at tobs = 3.6 μs
in the absence (1) and presence (2) of TEMPO (4 mmol L–1).

1424 Russ.Chem.Bull., Int.Ed., Vol. 57, No. 7, July, 2008
Gorelik et al.
I
A
a
0.9
0.8
0.7
0.6
0.5
0.4
E
2
4 mT
1
B
0
1
2
3
[TEMPO]/mmol L–1
Fig. 11. The CIDEP of TEMPO (2 mmol L–1) upon the photolysis
of DMDB (2.4 mmol L–1) in SDS micelles at tobs = 500 ns. Dotted
line is the TR ESR spectrum observed under the same conditions
but in the absence of TEMPO. The intensity of the spectrum in
the presence of TEMPO is amplified by 4.6 times.
I
1.0
0.8
0.6
0.4
0.2
0
b
xanthone and bromopyrene in SDS micellar solutions,
which also form no RPs.
All the systems under study are qualitatively identical
in the sense than the CIDEP in the stable radicals appears
with approximately the same efficiency, regardless of the
structure of both the radicals and ketones and correꢀ
sponding SCRPs. This observation makes it possible to
neglect the specificity of solubilizate localization in the
micellar phase, i.e., to advocate that the compounds disꢀ
solved in the micellar phase experience pair collisions
with the frequency (Z) determined by the equation
0
0.2
0.4
0.6
0.8
1.0
1.2 t/ms
Fig. 12. a. Relative net TR ESR intensities (I) of the highꢀfield signal
(1) and optical absorption at the maximum (2) of the phosphinoyl
radical upon the photolysis of TMBDPO (6 mmol L–1) in SDS
micelles at different TEMPO concentrations. b. Kinetics of the highꢀ
field ESR signal normalized to the maximum intensity value for the
phosphinoyl radical upon the photolysis of TMBDPO (6 mmol L–1
in SDS micelles in the absence (solid line) and presence (dotted line)
of TEMPO (4 mmol L–1).
Z = 3RDη /VF,
fil
where R is the collision radius, D is the mutual diffusion
coefficient, VF is the free volume of the micellar phase,
and ηfil is the number of micelle filling with the collision
partner. For the SCRP radicals, ηfil = 1. For the nitroxide
radicals, ηfil is usually identified with the mean value
)
obtained by the Poisson distribution. Thus, one can acꢀ
quenchers of both the triplet and singlet photoexcited
electronic states.37 The CIDEP can be generated in the
nitroxide radicals according to the radicalꢀtriplet pair
mechanism (RTPM).38 If the polarization on the nitrꢀ
oxide radicals in all the SCRPs under study was negaꢀ
tive, the predomination of the RTPM in CIDEP generation
could be assumed; the more so, the viscosity of the micelꢀ
lar SDS phase is relatively high. However, in the case
of TMBDPO, the polarization of the nitroxide radiꢀ
cals is positive (see Fig. 9). Therefore, it is improbaꢀ
ble that the RTPM is the predominant mechanism
in the generation of nitroxide radical polarization in miꢀ
cellar solutions. The positive polarization could appear
due to the quenching of the excited singlet state of
TMBDPO, as it occurs for coronene quenching in benꢀ
zene.39 However, this assumption contradicts the emisꢀ
sion polarization of TEMPO during the photolysis of
other ketones.
–1
cept Z = 6•107η
s
as the standard characteristic. Note
fil
that VF is determined by the geometry of the micelle and
radicals and also by the hydrophobic properties of
the solubilizate because of the equilibrium of the diffuꢀ
sion flows directly at the interface of two phases (aqueous
and micellar) and the dependence of the region of diffuꢀ
sion roving of the solubilizate on its hydrophilic properꢀ
ties. The polar radicals can diffuse to the regions with
the increased water content, and the region of nonpolar
radical diffusion is restricted by the more hydrophobic
(smaller in geometric sizes) part of the micellar phase.
This, particularly, results in the dependence of the HFC
constants in the nitroxide radicals on their polarity8 rathꢀ
er than their specific localization.
Several different physicochemical spin selective proꢀ
cesses involving the stable nitroxide radicals can occur in
the considered systems. First, the nitroxide radicals are

CIDEP in the presence of nitroxide radicals
Russ.Chem.Bull., Int.Ed., Vol. 57, No. 7, July, 2008
1425
Collisions of the radicals with the spinꢀpolarized
molecular triplets can result in polarization generation in
the stable radicals via the ESPT mechanism.14 It is conꢀ
ventional but useful to distinguish the ESPT and RTPM
mechanisms, because the time of CIDEP generation via
the ESPT mechanism is restricted by a substantially shortꢀ
er interval than that in the RTPM, viz., the relaxation
time of the electron spin states of the molecular triplet. In
the case of the ESPT mechanism, the signs of polarizaꢀ
tion on the nitroxide radical and RP radicals polarized
via the triplet mechanism should coincide, i.e., this mechꢀ
anism qualitatively agrees with the experimental data.
Quantitatively it also does not contradict the experimenꢀ
tal results. In fact, the relaxation rate of populations
of the isotropically rotating triplet with the zeroꢀfield
splitting parameters, which satisfy the inequality
|DZFS| >> |EZFS|, is the following20:
radicals with the αꢀelectron spin. Since no reaction occurs
upon the collision of two radicals with the likely oriented
spins, an excess of βꢀspins of the radicals of the pair will
result in the situation when predominantly the nitroxide
radicals with the αꢀelectron spin would be involved in
the accepting reaction, i.e., the "reaction" polarization
transfer will take place. This polarization mechanism is
not less probable than the transfer of the nonequilibrium
magnetization from an ensemble of the polarized radiꢀ
cals of the pair to an ensemble of the nitroxide radicals.
However, in the case of the SCRP3—SCRP5 when the net
polarization of the radicals is low, polarization generaꢀ
tion due to this mechanism cannot be efficient. Neꢀ
vertheless, the nitroxide radicals turned out to be spinꢀ
polarized.
Moreover, the accepting reaction of one of the partꢀ
ners by the SCRP should produce free radicals (partners
by the SCRP of the accepted radical) polarized according
to the radicalꢀpair ST0 mechanism. In the cases of the
SCRP4 and SCRP5, it can be difficult to detect this effect,
because the widths of the components of the APS and free
radicals are comparable. In the cases of the SCRP1 and
SCRP3, free radicals can easily be distinguished from
the radicals in pair by substantially narrower resonance
lines as well. However, no increase in the rate of formaꢀ
tion of the free radicals polarized via the radicalꢀpair SТ0
mechanism is observed in the presence of the nitroxide
radicals in these systems. At the same time, accepting
unambiguously takes place (see above). Therefore,
the rate of this process is appreciably lower than the
internal relaxation rate in the SCRPs, i.e., the collision
rate of the pair radicals in SDS micelles. Based on this,
we may conclude that the radicalꢀpair ST0 mechanism
can be neglected for random pairs formed by the nitrꢀ
oxide radical and one of the SCRP radicals; the more so,
this mechanism occurs only at low concentrations of the
nitroxide radicals.
The exchange electron spinꢀspin interactions of
the pair radicals with the nitroxide radical can polarize
the latter, if the pair radicals are net polarized. Indeed,
among the eight spin states of the threeꢀspin system for
the fixed configuration χ of the nuclear spin subsystem in
a strong magnetic field under the conditions when the
external longitudinal paramagnetic relaxation can be neꢀ
glected, the population of the two spin states |ααα;χ> and
|βββ;χ> remains unchanged. The other six states can be
grouped according to the projection of the total electron
spin: |αββ;χ>, |βαβ;χ>, |ββα;χ> (Mz = –1/2) and |ααβ;χ>,
|αβα;χ>, |βαα;χ> (Mz = +1/2). The pair exchange interacꢀ
tions between the pair radicals and the interactions of the
pair radicals with the nitroxide radical (JAB(rAB), JBC(rBC),
and JCA(rCA), respectively) equalize the populations of
the states inside these groups but induce no "flipꢀflop"
transitions of the electron spins between the spin states
belonging to different groups. Let us assume that the pair
1/T1 = (2/15)D2 •[4τ /(1 + 4ω
τ
2) +
c
2
ZFS
c
0
2
+ τ /(1 + ω
τ
2)],
c
0
c
where τc is the correlation time, and ω0 is the frequency of
the ESR spectrometer in the Xꢀrange.
For the case34 of TMBDPO, |DZFS| = 0.185 cm–1
.
Accepting τc ≈ 100 ps as an approximate estimate for
TMBDPO in SDS micelles, for ω0 = 6•1011 s–1 we obtain
that the longitudinal relaxation time in the electronꢀexciꢀ
ted triplet of TMBDPO is approximately 1/T1 = 3•108 s–1.
For a nitroxide radical concentration of 6•10–3 mol L–1
,
the average micelle filling is ηfil = 4. Using these esꢀ
timates, we have that the fraction of depolarized
triplets due to collisions with the nitroxide radical is
(ZηfilηpolT1)/(1 + ZηfilηpolT1), where ηpol is the polarizaꢀ
tion coefficient equal to ~0.4 for the diffusionally conꢀ
trolled quenching of the triplet. This value is well consisꢀ
tent with the experimental data (see Fig. 12, a).
The relative absorbance values of the phosphinoyl
radical at the wavelength λ = 355 nm (at the maximum of
the optical absorption spectrum) measured at the kinetic
curve maximum (immediately after a laser pulse) and the
relative net intensities of the TR ESR highꢀfield compoꢀ
nent of the SCRP1 at the maximum of its kinetic curve are
shown in Fig. 12. In the both cases, the values were
obtained relative to the corresponding signals in the abꢀ
sence of TEMPO. The almost coincidence of the results
of different experiments and the agreement with the theoꢀ
retical estimate indicate that the ESPT mechanism is
predominant for CIDEP generation in the nitroxide raꢀ
dicals, if assuming the simultaneous polarization transfer
and triplet state quenching.
In the nitroxide radicals, CIDEP can also appear due
to spin selective reactions of SCRP radical accepting by
the nitroxide radicals. Let us assume that the radicals of
the pair are positively polarized, i.e., the number of raꢀ
dicals with the βꢀelectron spin exceeds the number of

1426
Russ.Chem.Bull., Int.Ed., Vol. 57, No. 7, July, 2008
Gorelik et al.
is born from the molecular triplet polarized state, so that
the population of the |T0> spin state of the triplet is 1/3
and those of the spin states |T+> and |T–> are 1/3 – η and
1/3 + η, respectively, where η is the net polarization via
the triplet mechanism. If the spin exchange is efficient,
the system reaches intermediate equilibrium within the
time 2/Z ≈ 30 ns. It is easy to calculate the population of
each group existing in the intermediate equilibrium state.
If N nitroxide radicals are localized in a micelle, then
the intensity of the ESR spectra of each SCRP radical
is proportional to 2η/(2 + N) and the intensity of the
ESR signal of the nitroxide radicals is proportional to
2ηN/(2 + N). Using the standard approximation of the
Poisson distribution of the solubilizate in micelles, we
obtain that the intensities of the pair radicals (IA and IB)
and nitroxide radicals (IC) are proportional
thyldeoxybenzoin already in the triplet state due to, e.g.,
ketone enolization in the encounter complex. It is noteꢀ
worthy that the lifetime of the triplet state of dimeꢀ
thyldeoxybenzoin is approximately two orders of magniꢀ
tude longer than the lifetime of the SCRPs in benzene,
i.e., the quenching of the triplet state of the ketone by
the nitroxide radical is a knowingly more probable event
than the generation of the threeꢀspin system.
Thus, among possible processes the electron spin poꢀ
larization transfer from nonequilibrium molecular triplets
to the nitroxide radical is most efficient. All other procesꢀ
ses, including spin catalysis, which affect the decay kineꢀ
tics of radicals to this or another extent, play a substanꢀ
tially smaller role in electron spin polarization generation
of the nitroxide radicals in micellar multispin systems.
The lifetimes of the triplet excited states of the keꢀ
tones under study lie in a rather wide interval: from
100 ps (TMBDPO) to microseconds (benzyl). In the
cases of TMBDPO and 2,2ꢀdimethoxyꢀ2ꢀphenylaceꢀ
tophenone, these lifetimes are shorter than the relaxation
times of the nonequilibrium spin states of the correꢀ
sponding triplets. Nevertheless, the numerical estimaꢀ
tions show that the assumption on the uniform distribuꢀ
tion of the radicals over the micelle volume is quite
adequate to the experimental observations. Of course,
this does not exclude a possibility of specific localization
of the radicals in micelles, although it seems rather
strange, because this localization was found for none of
13 radicals with different chemical structures. However,
the experimental data obtained show that this specificity
is unnecessary to assume.
The absence of any influence of the nitroxide radicals
on the linewidths of the SCRP radicals and the width of
the APS components even at relatively high concentraꢀ
tions of the nitroxide radicals suggests the following:
either the exchange interaction between the nitroxide
radicals and the pair radicals is low, which explains,
particularly, the relatively low efficiency of accepting of
the SCRP radicals, or the quenching of the excited triplet
states of the considered ketones by the nitroxide radicals
is so efficient that no threeꢀspin systems are formed.
,
.
The time of establishing the quasiꢀequilibrium polaꢀ
rizations IA, IB, and IC is determined by the collision
frequency of the radicals in micelles, i.e., lies in an interꢀ
val of only several tens of nanoseconds. This time interval
is much shorter than the characteristic time of exchange
by the nitroxide radicals of micelles limited by bimolecꢀ
ular collisions of the micelles with each other or between
the radicals and micelles. Due to this, the presence of
the third spin in the micelle exerts no qualitative effect on
the character of the ESR spectrum of the net polarized
radicals, decreasing only its intensity. However, if this
process is efficient, it should be accompanied by the addiꢀ
tional exchange broadening of the spectral lines of both
the SCRP radicals and the nitroxide radical. As shown by
the analysis of the shape of the TR ESR spectra, no addiꢀ
tional broadening is observed even for the very narrow
resonance lines of the α,α´ꢀdimethoxybenzyl radical
(see Fig. 10, c, SCRP2).
Thus, of the set of probable spin selective processes,
namely, spin polarization transfer from the spinꢀpolaꢀ
rized triplet states, spin polarization due to the RTPM,
spin selective accepting, CIDEP due to the radicalꢀpair
ST0 mechanism in random radical pairs, and exchange
interaction with net polarized radicals, only the first one
is completely consistent with the experimental results.
The efficiency of other processes is substantially lower.
In addition, these processes would contradict the experiꢀ
mental facts in several cases. Therefore, the observed
increase in the probability of recombination of the SCRPs
of the benzoyl and secꢀphenetyl radicals18 in the zero
magnetic field contradicts our experimental observations
and conclusions. Perhaps, the high concentration of the
nitroxide radical induces the racemization of dimeꢀ
This work was financially supported by the Russian
Foundation for Basic Research (Project Nos 05ꢀ03ꢀ32197,
06ꢀ03ꢀ91362ꢀYaF, and 07ꢀ03ꢀ00948).
References
1. Surfactants in Solutions, Eds K. L. Mittal, B. Lindeman, Pleꢀ
num, New York, 1984, 1—3; Surfactants in Solutions, Eds K. L.
Mittal, P. Botherel, Plenum, New York, 1986, 4—6; A. I.
Rusanov, Mitselloobrazovanie v rastvorakh poverkhnostnoꢀakꢀ
tivnykh veshchestv [Micelle Formation in Solutions of Surfacꢀ
tants], Khimiya, St. Peterburg, 1992 (in Russian).
2. F. ReissꢀHusson, V. Luzzati, J. Phys. Chem., 1964, 68, 3504;
F. M. Menger, Acc. Chem. Res., 1979, 12, 111; M. F. Vitha, A. J.
Dallas, P. W. Carr, J. Phys. Chem., 1996, 100, 5050; B. L.

CIDEP in the presence of nitroxide radicals
Russ.Chem.Bull., Int.Ed., Vol. 57, No. 7, July, 2008
1427
Bales, L. Messina, A. Vidal, M. Peric, O. R. Nascimento,
J. Phys. Chem. B, 1998, 102, 10347.
3. M. Aamodt, M. Landgren, B. Jönsson, J. Phys. Chem., 1992,
96, 945.
4. Y. Chevalier, C. Chachaty, J. Phys. Chem., 1985, 89, 875;
A. Belmajdoub, N. Mahieu, P. Tekely, D. Canet, J. Phys.
Chem., 1992, 96, 1011.
21. A. L. Buchachenko, R. Z. Sagdeev, K. M. Salikhov, Magnitnye
i spinovye effekty v khimicheskikh reaktsiyakh [Magnetic and
Spin Effects in Chemical Reactions], Nauka, Novosibirsk, 1978
(in Russian); U. E. Steiner, T. Ulrich, Chem. Rev., 1989, 89, 51;
K. M. Salikhov, Yu. N. Molin, R. Z. Sagdeev, A. L. Buchachenꢀ
ko, Spin Polarization and Magnetic Effects in Radical Reactions,
Elsevier, Amsterdam, 1984.
5. M. Almgren, F. Grieser, J. K. Thomas, J. Am. Chem. Soc., 1979,
101, 279; R. Nagarajan, M. A. Chaiko, E. Ruckenstein, J. Phys.
Chem., 1984, 88, 2916; N. R. Jagannathan, K. Venkateswaran,
F. G. Herring, G. N. Patey, D. C. Walker, J. Phys. Chem., 1987, 91,
4553; B. L. Bales, R. Zana, J. Phys.Chem. B, 2002, 106, 1926.
6. M. van der Auweraer, C. Dederen, C. PalmansꢀWindels, F. C.
DeSchryver, J. Am. Chem. Soc., 1982, 104, 1800; B. L. Bales,
M. Almgren, J. Phys. Chem., 1995, 99, 15153; M. E. Daraio,
P. F. Aramendia, E. San Román, Chem. Phys. Lett., 1996, 250,
203; R. Ranganathan, C. VautierꢀGiongo, B. Bales, J. Phys.
Chem. B, 2003, 107, 10312.
7. A. L. Buchachenko, A. M. Vasserman, Stabil´nye radikaly
[Stable Radicals], Khimiya, Moscow, 1973 (in Russian); A. N.
Kuznetsov, Metod spinovogo zonda [Spin Probe Method], Nauꢀ
ka, Moscow, 1976 (in Russian).
8. A. M. Vasserman, Usp. Khim., 1994, 63, 391 [Russ. Chem. Rev.,
1994, 63, 373 (Engl. Transl.)].
9. M. F. Ottaviani, P. Baglioni, G. Martini, J. Phys. Chem., 1983,
87, 3146; P. Baglioni, M. F. Ottaviani, G. Martini, J. Phys.
Chem., 1986, 90, 5878; S. Ristori, M. F. Ottaviani, D. Lenti,
G. Martini, Langmuir, 1991, 7, 1958.
10. P. J. Bratt, H. Choudhurry, P. B. Chowdhurry, D. G. Gillies,
A. M. L. Krebber, L. H. Sutcliffe, J. Chem. Soc., Faraday
Trans., 1990, 86, 3313; S. Ristori, G. Martini, Langmuir, 1992,
8, 1937; B. L. Bales, Ch. Stenland, Chem. Phys. Lett., 1992, 200,
475; B. L. Bales, Ch. Stenland, J. Phys. Chem., 1993, 97, 3418;
B. L. Bales, Ch. Stenland, J. Phys. Chem., 1995, 99, 15163.
11. V. F. Tarasov, A. L. Buchachenko, V. I. Mal´tsev, Zh. Fiz.
Khim., 1981, 55, 1921 [J. Phys. Chem. USSR, 1981, 55
(Engl. Transl.)].
22. V. F. Tarasov, M. D. E. Forbes, Spectrochim. Acta, Part A,
2000, 56, 245.
23. A. I. Shushin, Chem. Phys. Lett., 1997, 275, 137; A. I. Shushin,
Chem. Phys. Lett., 1998, 282, 413.
24. M. D. E. Forbes, G. L. Closs, P. Calle, P. Gautam, J. Phys.
Chem., 1993, 97, 3384.
25. V. F. Tarasov, H. Yashiro, K. Maeda, T. Azumi, I. A. Shkrob,
Chem. Phys., 1998, 226, 253.
26. H. Hayashi, Y. Sakaguchi, M. Kamachi, W. Schnabel, J. Phys.
Chem., 1987, 91, 3936.
27. H. Fisher, R. Baer, R. Hany, J. Chem. Soc., Perkin Trans. 2,
1990, 787.
28. V. F. Tarasov, I. A. Shkrob, E. N. Step, A. L. Buchachenko,
Chem. Phys., 1989, 135, 391.
29. J. C. Scaiano, E. B. Abuin, L. C. Stewart, J. Am. Chem. Soc.,
1982, 104, 5673.
30. Y. Sakaguchi, H. Hayashi, J. Phys. Chem., 1984, 88, 1437.
31. F. R. Hewgill, G. M. Proudfoot, Austr. J. Chem., 1977, 30, 695.
32. R. Breslow, S. Kitabatake, J. Rothbard, J. Am. Chem. Soc.,
1978, 100, 8156.
33. V. R. Gorelik, K. Maeda, H. Yashiro, H. Murai, J. Phys. Chem.
A, 2001, 105, 8011; H. Murai, N. Ishiwata, K. Kuwata,
Y. Sakaguchi, H. Hayashi, Chem. Phys. Lett., 1993, 203, 482.
34. I. V. Koptyug, N. D. Ghatlia, G. W. Sluggett, N. J. Turro,
S. Ganapathy, W. G. Bentrude, J. Am. Chem. Soc., 1995,
117, 9486.
35. S. Jokusch, I. V. Koptyug, P. F. McGarry, G. W. Sluggett, N. J.
Turro, D. M. Watkins, J. Am. Chem. Soc., 1997, 119, 11495.
36. A. Kawai, A. Shikama, M. Mitsui, K. Obi, Bull. Chem. Soc. Jpn,
2001, 74, 1203.
12. L. Sterna, D. Ronis, S. Wolfe, A. Pines, J. Chem. Phys., 1980,
73, 5493.
13. V. F. Tarasov, L. L. Yasina, I. I. Aliev, A. M. Vasserman,
M. D. E. Forbes, Khim. Fiz., 2002, 21, 107 [Russ. Chem. Phys.,
2002, 21 (Engl. Transl.)].
14. T. Imamura, O. Onitsuka, K. Obi, J. Phys. Chem., 1986, 90,
6741; J.ꢀI. Fujisawa, K. Ishii, Y. Ohba, M. Iwaizumi, S. Yamauꢀ
chi, J. Phys. Chem., 1995, 99, 17082; V. F. Tarasov, I. A.
Shkrob, A. D. Trifunac, J. Phys. Chem. A, 2002, 106, 4838.
15. W. S. Jenks, N. J. Turro, J. Am. Chem. Soc., 1990, 112, 9009.
16. K. M. Salikhov, A. J. van der Est, D. Stehlik, Appl. Magn.
Reson., 1999, 16, 101.
17. I. M. Magin, V. S. Shevel´kov, A. A. Obynochny, A. I. Kruppa,
T. V. Leshina, Chem. Phys. Lett., 2002, 357, 351; I. M. Magin,
P. A. Purtov, A. I. Kruppa, T. V. Leshina, Appl. Magn. Reson.,
2004, 26, 155; I. M. Magin, P. A. Purtov, A. I. Kruppa, T. V.
Leshina, J. Phys. Chem. A, 2005, 109, 7396.
37. G. Porter, M. R. Wright, Discuss. Faraday Soc., 1959, 27, 18;
G. J. Hoijtink, Mol. Phys., 1969, 2, 114; O. L. Gijzemann, F. Kaꢀ
ufmann, G. Porter, J. Chem. Soc., Faraday Trans. 2, 1973, 69,
727; V. A. Kuzmin, A. S. Tatikolov, Chem. Phys. Lett., 1977, 51,
45; V. A. Kuzmin, A. S. Tatikolov, Chem. Phys. Lett., 1978, 53,
606; A. R. Watkins, Chem. Phys. Lett., 1980, 70, 262; K. R.
Naqvi, V. P. Wid, Chem. Phys. Lett., 1976, 41, 570; A. R.
Watkins, Chem. Phys. Lett., 1974, 29, 526; A. S. Tatikolov, P. P.
Levin, T. A. Kokrashvili, V. A. Kuzmin, Izv. Akad. Nauk SSSR,
Ser. Khim., 1983, 517 [Bull. Acad. Sci. USSR, Div. Chem. Sci.,
1983, 32, 465 (Engl. Transl.)]; T. Suzuki, K. Obi, Chem. Phys.
Lett., 1995, 246, 130; G. J. Hoijtink, Mol. Phys., 1960, 3, 67; J. N.
Murrell, Mol. Phys., 1960, 3, 317; G. W. Robinson, J. Mol.
Spectrosc., 1961, 6, 58; Y.ꢀN. Chiu, J. Chem. Phys., 1972, 56, 4882.
38. C. Blättler, H. Paul, Chem. Phys. Lett., 1990, 166, 375; V. V.
Tarasov, Ph. D. Thesis, University of Zürich, Zürich, Switzerꢀ
land, 1999; A. Blank, H. Levanon, J. Phys. Chem. A, 2000, 104,
794; A. I. Shushin, J. Chem. Phys., 1993, 99, 8723; A. I.
Shushin, Chem. Phys. Lett., 1999, 313, 246; F. J. Adrian, Chem.
Phys. Lett., 1994, 229, 465.
18. E. N. Step, A. L. Buchachenko, N. J. Turro, J. Am. Chem. Soc.,
1994, 116, 5462; A. L. Buchachenko, E. N. Step, V. L. Ruban,
N. J. Turro, Chem. Phys. Lett., 1995, 233, 315.
19. A. L. Buchachenko, V. L. Berdinsky, J. Phys. Chem., 1996,
100, 18292.
39. A. Kawai, K. Obi, J. Phys. Chem., 1992, 96, 52.
20. P. W. Atkins, G. T. Evans, Chem. Phys. Lett., 1974, 25, 108;
P. W. Atkins, G. T. Evans, Mol. Phys., 1974, 27, 1633.
Received October 9, 2007;
in revised form January 25, 2008