Supramolecular photochemistry in β-cyclodextrin hosts: A TREPR, NMR, and CIDNP investigation

Article abstract of DOI:10.1021/la904788t

A systematic investigation of the photochemistry and ensuing radical chemistry of three guest ketones encapsulated in randomly methylated β-cyclodextrin (β-CD) hosts is reported. Dibenzyl ketone (DBK), deoxybenzoin (DOB), and benzophenone (BP) triplet states are rapidly formed after photolysis at 308 nm. Time-resolved electron paramagnetic resonance (TREPR) spectroscopy, steady-state NMR spectroscopy, and time-resolved chemically induced nuclear polarization (TR-CIDNP) experiments were performed on the ketone/CD complexes and on the ketones in free solution for comparison. The major reactivity pathways available from these excited states are either Norrish I α-cleavage or H-atom abstraction from the interior of the CD capsule or the solvent. The DOB triplet state undergoes both reactions, whereas the DBK triplet shows exclusively α-cleavage and the BP triplet shows exclusively H-atom abstraction. Radical pairs are observed in β-CDs by TREPR, consisting of either DOB or BP ketyl radicals with sugar radicals from the CD interior. The TREPR spectra acquired in CDs are substantially broadened due to strong spin exchange. The electron spin polarization mechanism is mostly due to S-T₀ radical pair mechanism (RPM) in solution but changes to S-T-RPM. in the CDs due to the large exchange interaction. The TR-CIDNP results confirm the reactivity patterns of all three ketones, and DOB shows strong nuclear spin polarization from a novel rearrangement product resulting from the α-cleavage reaction.

Full text of DOI:10.1021/la904788t

pubs.acs.org/Langmuir
© 2010 American Chemical Society
Supramolecular Photochemistry in β-Cyclodextrin Hosts: A TREPR, NMR,
and CIDNP Investigation
Olesya A. Krumkacheva,† Vitaly R. Gorelik,† Elena G. Bagryanskaya,*,† Natalia V. Lebedeva,‡ and
Malcolm D. E. Forbes*,‡
†International Tomography Center, Institutskaya 3a, Novosibirsk 630090, Russia, and ‡Caudill Laboratories,
Department of Chemistry, CB #3290, University of North Carolina, Chapel Hill, North Carolina 27599-3290
Received December 18, 2009. Revised Manuscript Received February 13, 2010
A systematic investigation of the photochemistry and ensuing radical chemistry of three guest ketones encapsulated in
randomly methylated β-cyclodextrin (β-CD) hosts is reported. Dibenzyl ketone (DBK), deoxybenzoin (DOB), and
benzophenone (BP) triplet states are rapidly formed after photolysis at 308 nm. Time-resolved electron paramagnetic
resonance (TREPR) spectroscopy, steady-state NMR spectroscopy, and time-resolved chemically induced nuclear
polarization (TR-CIDNP) experiments were performed on the ketone/CD complexes and on the ketones in free solution
for comparison. The major reactivity pathways available from these excited states are either Norrish I R-cleavage or
H-atom abstraction from the interior of the CD capsule or the solvent. The DOB triplet state undergoes both reactions,
whereas the DBK triplet shows exclusively R-cleavage and the BP triplet shows exclusively H-atom abstraction. Radical
pairs are observed in β-CDs by TREPR, consisting of either DOB or BP ketyl radicals with sugar radicals from the CD
interior. The TREPR spectra acquired in CDs are substantially broadened due to strong spin exchange. The electron
spin polarization mechanism is mostly due to S-T0 radical pair mechanism (RPM) in solution but changes to S-T-
RPM in the CDs due to the large exchange interaction. The TR-CIDNP results confirm the reactivity patterns of all
three ketones, and DOB shows strong nuclear spin polarization from a novel rearrangement product resulting from the
R-cleavage reaction.
Introduction
detected due to confinement.9 Supramolecular host-guest com-
plexes sometimes show significant reactivity between host and
guest,10 which can complicate potential applications such as drug
delivery or nanoencapsulation of dyes for imaging and printing
processes.
The structure, reactivity, and dynamics of free radicals in
confined environments are a topic of substantial interest to the
biochemical, photochemical, and nanotechnological commu-
nities.1 For molecular recognition investigations,2 synthesis and
chemical modification of nanoassemblies,3 and control of pro-
duct yields and/or stereochemistry in small molecule reactions,4
several different types of encapsulation strategies have been
employed. Examples of such nanoreactors include micelles,
reverse micelles, vesicles, zeolites, nanocapsules, nanotubes, and
calixarenes.5 In photochemical reactions, encapsulation can sub-
stantially alter the nature of the excited states involved6 and the
structures of the ensuing reactive intermediates, which are often
free radicals.7 Encapsulation can cause product partitioning of
these intermediates to deviate from those observed in free solu-
tion,8 and in some cases previously unobserved products are
The β-cyclodextrin (β-CD) molecule (Chart 1) has been a
popular choice for encapsulation studies because of its general
compatibility with biological solutions and tissues, its widespread
availability, and its low cost.11 It consists of an intramolecularly
linked cycle of seven β-glucose rings with acetal bonds between
the individual sugar units, and its macromolecular shape is that of
a tapered cylinder. In aqueous solution, host molecules are
incorporated into the interior of the cylinder due to the hydro-
phobic effect. The solubility of these complexes in aqueous
solution can be improved by partial methylation of the CD
hydroxyl groups, and several researchers have taken advantage
of this strategy using singly or multiply methylated CD struc-
tures.12 The origin of this increased solubility is thought to be due
to the fact that methylation increases the hydrophobicity of the
interior of the CD much more than the exterior. Because most
guests are relatively nonpolar organic molecules, an increase in
the hydrophobicity of the interior pushes the equilibrium associa-
tion constant of the complex to higher values.
*To whom correspondence should be addressed. E-mail: elena@tomo.nsc.
ru (E.G.B.); mdef@unc.edu (M.D.E.F.).
(1) Rebek, J., Jr. Acc. Chem. Res. 2009, 42, 1660–1668.
(2) Kaifer, A. E. Acc. Chem. Res. 1999, 32, 62–71.
(3) Liu, X.; Basu, A. J. Am. Chem. Soc. 2009, 131, 5718–5719.
(4) (a) Chung, W.-S.; Turro, N. J.; Silver, J.; le Noble, W. J. J. Am. Chem. Soc.
1990, 112, 1202–1205. (b) Mortko, C. J.; Garcia-Garibay, M. A. J. Am. Chem. Soc.
2005, 127, 7994–7995. (c) Rekharsky, M.; Inoue, Y. J. Am. Chem. Soc. 2000, 122,
4418–4435. (d) Lem, G.; Kaprinidis, N. A.; Schuster, D. I.; Ghatlia, N. D.; Turro, N. J. J.
Am. Chem. Soc. 1993, 115, 7009–7010.
(5) Zheng, R.; Liu, G. Macromolecules 2007, 40, 5116–5121.
(6) (a) Cox, G. S.; Turro, N. J. J. Am. Chem. Soc. 1984, 106, 422–424. (b)
Berberan-Santos, M. N.; Choppinet, P.; Fedorov, A.; Jullien, L.; Valeur, B. J. Am.
Chem. Soc. 2000, 122, 11876–11886.
The photochemistry of small molecules encapsulated in CDs
and the free radical chemistry resulting from photochemical
excitation in the interior have been studied by several research
groups with a variety of different physical methods. For example,
Singh et al. examined the photochemistry of several substituted
R-phenyl ketonesand found thattheproductratios fromNorrishI
(7) Rao, V. P.; Zimmtt, M. B.; Turro, N. J. J. Photochem. Photobiol., A 1991, 60,
355–60.
(8) Gibb, C. L. D.; Sundaresan, A. K.; Ramamurthy, V.; Gibb, B. C. J. Am.
Chem. Soc. 2008, 130, 4069–4080.
(10) Park, S. K.; Jung, Y. D.; Jung, J. J.; Park, S.; Lee, M.-H. React. Funct.
Polym. 2004, 58, 93–101.
(9) Arumugam, S.; Kaanumalle, L. S.; Ramamurthy, V. J. Photochem. Photo-
biol., A 2007, 185, 364–370.
(11) D’Souza, V. T.; Lipkowitz, K. B. Chem. Rev. 1998, 98, 1741–1742.
(12) Harada, A. Acc. Chem. Res. 2001, 34, 456–464.
Langmuir 2010, 26(11), 8971–8980
Published on Web 03/01/2010
DOI: 10.1021/la904788t 8971

Article
Krumkacheva et al.
Chart 1
(R-cleavage) and Norrish II (intramolecular H-atom abstraction)
reactions were only moderately altered by inclusion in β-CDs.13 It
is somewhat surprisingthat inclusion in the CD would have only a
minor effect on the outcome of intramolecular reactions. Reac-
tions involving the photoexcited triplet state of naphthoquinone
in CDs have been reported by Takamori et al.14 They observed
carbon-centered radicals from intermolecular H-atom abstrac-
tion reactions as an exclusive reaction pathway. A report by
Lehmann and Bakker examined alkyl ketone photoreactivity in
β- and γ-CDs.15 Again, only H-atom abstraction was observed,
although there was some selectivity in the abstraction site.
Rao and co-workers reported that the photochemistry of
alkyldibenzyl ketones included in CDs is dramatically different
in solution compared to the solid state.16 In the 1990s, Lehmann
et al. used steady-state electron paramagnetic resonance (EPR)
spectroscopy to examine the thermal and photochemical reac-
tions of several peroxides included in β-CD.17 They found
significant reactivity between the alkoxy radicals produced on
heating and the sugar ring interior and exterior of the CD,
depending on sample preparation conditions. Other steady-state
EPR work on stable radicals in CDs has demonstrated a different
reactivity. For example, Beckett et al. studied the tetracyano-
naphthoquinone radical anion and found that it reacts with the
CD interior to create a diamagnetic product.18 Also, Lucarini and
Pedulli found that the 31P hyperfine interactions of phosphorus-
containing hydrazyl radicals were dramatically lowered when the
radicals were included in β-CD.19
obfuscate the interpretation of results and lead to incorrect
assignments of intermediates and mechanisms. This problem is
compounded by the lack of structural resolution of optical spectra
in the condensed phase. Magnetic resonance (EPR and NMR)
spectroscopy offers high structural resolution and high sensitivity,
and one can often make unambiguous assignments of free radical
intermediates and final products. Additionally, time-resolved
methods such as chemically induced nuclear spin polarization
(CIDNP)20 and time-resolved EPR (TREPR)21 spectroscopy,
with time resolution on the order of 100 ns, are ideal experimental
techniques for observation of the primary photochemical events
in host-guest chemistry. Indeed, TREPR and CIDNP have been
applied routinely to study radical reactions in other nanoreactors
such as micelles and reverse micelles.22 To the best of our knowl-
edge, there exist only two previous reports of TREPR investiga-
tions of photochemical reactions in CDs,14,15 and only one
involving the use of the powerful CIDNP experiment in either
steady-state or time-resolved mode.23
Background
Given the limitations of the previous studies described above,
we were motivated to carry out a more systematic study of
photoexcited triplet states in β-CDs, in particular those that could
react by either R-cleavage or intramolecular H-atom abstraction
reactions. It is also possible to design experiments involving triplet
states that could undergo both reactions competitively. In this
regard, we report here a combined NMR, TREPR, and TR-
CIDNP study of the radical chemistry that takes place after UV
laser flash photolysis of the three ketones shown in Chart 2 when
they are encapsulated in β-CDs in aqueous solution at room
temperature: deoxybenzoin (DOB), dibenzyl ketone (DBK), and
benzophenone (BP). All three molecules were selected for their
structural simplicity and for their known rapid formation of a
reactive excited triplet state upon UV excitation.
Clearly, previous EPR studies on CD radical chemistry have
lacked systematic structural variation. In many of these previous
investigations, the steady-state spectroscopic and kinetic methods
employed do not always account for the primary mechanistic
events due to secondary reactions or rearrangements. An addi-
tional problem is that transient optical absorption spectra often
exhibit broad lines that can be difficult to assign. Rearrangements
of reactive intermediates and secondary photolysis processes can
(20) (a) Harbron, E. J.; Forbes, M. D. E. Encycl. Chem. Phys. Phys. Chem. 2001,
2, 1389–1417. (b) Spin Polarization and Magnetic Effects in Radical Reactions;
Salikhov, K. M., Molin, Y. N., Sagdeev, R. Z., Buchachenko, A. L., Eds.; Elsevier:
Amsterdam, 1984.
(21) Forbes, M. D. E. Photochem. Photobiol. 1997, 65, 73–81.
(22) (a) White, R. C.; Gorelik, N. V.; Bagryanskaya, E. G.; Forbes, M. D. E.
Langmuir 2007, 23, 4183–4191. (b) Shakirov, S. R.; Lebedeva, N. V; Gorelik, V. R.;
Tarasov, V. F.; Bagryanskaya, E. G. Appl. Magn. Reson. 2006, 30, 535–548. (c) Lopez,
J. J.; Carter, M. A. G.; Tsentalovich, Y. P.; Morozova, O. B.; Yurkovskaya, A. V.; Hore,
P., J. Photochem. Photobiol. 2002, 75, 6–10. (d) Tarasov, V. F.; Bagranskaya, E. G.;
Shkrob, I. A.; Avdievich, N. I.; Ghatlia, N. D.; Lukzen, N. N.; Turro, N. J.; Sagdeev, R. Z.
J. Am. Chem. Soc. 1995, 117, 110–118.
(13) Singh, S.; Usha, G.; Tung, C.-H.; Turro, N. J.; Ramamurthy, V. J. Org.
Chem. 1986, 51, 941–944.
(14) Takamori, D.; Aoki, T.; Yashiro, H.; Murai, H. J. Phys. Chem. A 2001, 105,
6001–6007.
(15) Lehmann, M. N.; Bakker, M. G. J. Chem. Soc., Perkin Trans. 2 1997, 2131–
2134.
(16) Rao, B. N.; Syamala, M. S.; Turro, N. J.; Ramamurthy, V. J. Org. Chem.
1987, 52, 5517–5521.
(17) Lehmann, M. N.; Bakker, M. G.; Patel, H.; Partin, M. L.; Dormady, S. J. J.
Inclusion Phenom. Mol. Recognit. Chem. 1995, 23, 99–117.
(18) Beckett, J. L.; Hartzell, C. J.; Eastman, N. L.; Blake, T.; Eastman, M. P. J.
Org. Chem. 1992, 57, 4173–4179.
(23) Petrova, S. S.; Kruppa, A. I.; Leshina, T. V. Chem. Phys. Lett. 2004, 385,
40–44.
(19) Lucarini, M.; Pedulli, G. F. J. Org. Chem. 2000, 65, 2723–2727.
8972 DOI: 10.1021/la904788t
Langmuir 2010, 26(11), 8971–8980

Krumkacheva et al.
Article
Chart 2
Hyperfine interactions and g-factor differences between the
members of each geminate RP will induce mixing between the
initial RP triplet state and its corresponding RP singlet state, from
which geminate recombination can occur to give a nuclear spin
polarized product. In free solution, triplet RPs will also escape in-
to the bulk and form random or out-of-cage products (Scheme 2):
for DOB, this includes benzyl, diphenylethane, and the starting
material. Interestingly, the rearranged product paramethylbenzo-
phenone (PMB) also appears in this photochemistry. The PMB
molecule is created via recombination of RP1 where the benzoyl
radical attacks the 4-position of the benzyl radical, followed by
intramolecular H-atom transfer. Because the latter reaction is
slow on the time scale of TR-CIDNP, it is the vinyl protons of the
initially formed intermediate compound (Scheme 2, bottom left)
that appear in the CIDNP spectrum. Such a rearrangement has
been observed previously in steady-state photolysis experiments
involving DBK.27
Scheme 1
The bulk of this description of the photochemistry in Schemes 1
and 2 has been focused rather intentionally on the DOB triplet
state, as it exhibits the widest range of chemical reacitvity. It
should be noted that the DBK molecule follows exclusively the
R-cleavage reaction pathway and BP undergoes only H-atom
abstraction. Radical structures from the triplet states of these
ketones are also depicted in Scheme 2. All of the products (or
radicals) in Scheme 2 have been observed in our TR-CIDNP (or
TREPR) experiments. Their role in the mechanistic photochem-
istry in Scheme 1, with regard to inclusion in CDs versus free
solution, will be discussed in detail below.
Experimental Section
Scheme 1 outlines the photochemistry and possible radicals
formed in free solution and in the interior of CDs for the triplet
states of the ketones shown in Chart 1. Upon laser excitation,
these molecules have both (n,π*) and (π,π*) excited states
available to them. In DOB, the carbonyl group simultaneously
carries aryl and benzyl character, and the two lowest energy triplet
states may be close in energy for this ketone. This is important
because triplet (n,π*) states tend to undergo Norrish I R-cleavage
reactions, while triplet (π,π*) states are more likely to abstract H
atoms from the solvent or other surrounding media.24 If these two
states are strongly mixed, as they are, for example, in acetophe-
none,25 both types of reactivity can be observed. For such
ketones, the partitioning of the two reaction pathways may be
very sensitive to the local environment.
Intersystem crossing from the excited singlet state to the
molecular triplet state is expected to be fast,26 and all three triplet
states should be completely formed shortly after the laser flash,
well before the detection of radicals by TREPR or reaction
products by TR-CIDNP. For this reason, it is assumed that all
radical pairs (RPs) produced in our photochemical reactions will
be triplet-borne. For DOB, the Norrish I R-cleavage reaction will
produce benzoyl and benzyl radicals, and H-atom abstraction
from a suitable hydrogen donor will lead to a ketyl radical and a
donor radical. As indicated in Scheme 1, we will abbreviate the
two radical pairs as RP1 for the pair created from Norrish I
R-cleavage and RP2 for the pair created from H-atom abstrac-
tion. For CD-encapsulated DOB (DOB/CD), the hydrogen
donor is the CD host, in particular the interior facing H’s at the
3 and 5 positions of the sugar rings (Chart 3).
Materials. DOB (Aldrich) was purified by sublimation. Ran-
domly methylated β-CD (Fluka) was used as received. To prepare
the complex, DOB (8 mM) was added to an aqueous solution of
β-CD (2 mM). The solution was sonicated for 30 min, stirred for
6 h, and then filtered immediately prior to experiments. The
concentration of DOB in the complex was 1 mM, as measured
by UV and NMR spectroscopy. Deuterated solvents for the
NMR and CIDNP spectra, CD3CN 99% and D2O 98%, were
purchased from ISOTOPE and used as received. High purity
solvents 2-propanol, acetonitrile, cyclohexane, and distilled water
were used in the TREPR experiments.
TREPR Experiments. The TREPR signals28 were obtained
using a double-gated integrator (Stanford Research Systems) in
direct detection mode over 2-4 min. Low temperature TREPR
experiments werecarried out ona JEOL JES RE-1Xspectrometer
equipped with a liquid nitrogen variable temperature quartz
dewar. The temperature was maintained at 100 K. Samples were
prepared in single tubes that underwent five freeze-pump-thaw
cycles. The EPR signal was passed from the preamplifier of the
microwave bridge to the two boxcar gates as the magnetic field
was swept. Traces were collected at each field point and then
subtracted from the off-resonance signal. For solution TREPR
experiments, the sample was flowed through the microwave
resonator to prevent heating and sample depletion, while nitrogen
gas was bubbled continuously through a reservoir of typical
volume 20 mL. A more complete description of the apparatus
can be found in ref 28.
NMR and CIDNP Studies. 1H NMR experiments were per-
formed on an Avance 200 MHz Bruker DRX spectrometer. A
standardπ/2 pulse sequencewasusedwith64scans. NMR spectra
were measured at room temperature. The DOB concentration
(24) Turro, N. J.; Scaiano, J. C.; Ramamurthy, V. Principles of Molecular
Photochemistry: An Introduction; University Science Books: New York, 2008.
(25) Wagner, P. J.; Hammond, G. S. Adv. Photochem. 1968, 5, 21–156.
(26) Mcgarry, P. F.; Doubleday, C. E.; Wu, C. H.; Staab, H. A.; Turro, N. J. J.
Photochem. Photobiol., A 1994, 77, 109–117.
(27) Kraeutler, B.; Turro, N. J. Chem. Phys. Lett. 1980, 70, 270–275.
(28) Lebedeva, N. V.; Forbes, M. D. E. In Carbon-Centered Radicals and
Radical Cations; Forbes, M. D. E., Ed.; Wiley Series on Reactive Intermediates in
Chemistry and Biology; Wiley: New York, 2010; Volume 3, Chapter 14, pp 323-355,
(in press).
Langmuir 2010, 26(11), 8971–8980
DOI: 10.1021/la904788t 8973

Article
Krumkacheva et al.
Chart 3
Scheme 2
was 1 mM in the complex and 3 mM in the CD3CN solution. The
concentration of randomly methylated β-CD was 2 mM in the
aqueous solution. In TR-CIDNP experiments,29 the sample was
purged with argon and sealed in a standard NMR Pyrex tube and
then irradiated by an excimer laser (308 nm, 30 mJ) inside the
NMR probe. CIDNP experiments were carried out with the usual
pulse sequence: presaturation-laser pulse-evolution time-
detection pulse-free induction decay. A complete description of
the technical details for this experiment can be found in ref 29.
Because the background signals in CIDNP spectra are suppressed
by the presaturation pulses, only signals of the polarized products
appear in the CIDNP spectra. The number of scans was 32.
Results and Discussion
NMR Analysis of the Ground State DOB/CD Complex.
In order to understand the photochemistry of the DOB/CD
complex and how DOB behaves differently in CDs than in free
solution, the structure of the complex itself was investigated in the
absence of light using 1H NMR spectroscopy. As mentioned
above, random methylation means that C2, C3, and C6 can each
have substitutents that are either -OH or -OCH3 (Chart 1). This
can affect the chemical shifts of neighboring protons, some of
which may be involved in the photochemical H-atom abstraction
reaction. Therefore, care should be taken to identify these signals
if they can be resolved in the NMR spectrum. Figure 1a shows
NMR data acquired at 400 MHz for randomly methylated β-CD,
while Figure 1c shows the NMR spectrum of pure DOB in free
(29) (a) Miller, R. J.; Closs, G. L. Rev. Sci. Instrum. 1981, 52, 1876–1885. (b)
Closs, G. L.; Miller, R. J.; Redwine, O. D. Acc. Chem. Res. 1985, 18, 196–202. (c)
Closs, G. L.; Miller, R. J. J. Am. Chem. Soc. 1979, 101, 1639–1641.
8974 DOI: 10.1021/la904788t
Langmuir 2010, 26(11), 8971–8980

Krumkacheva et al.
Article
Figure 2. (a) DarkNMRspectrumofa 10mM solutionofDOBin
CD3CN, (/ = H2O, // = CD2HCN). (b) CIDNP spectrum ob-
tained after photolysis of DOB in CD3CN. In this and all subsequent
CIDNP spectra, the time delay between the laser pulse and the rf-
pulse is 1 μs. Assignment of the CIDNP signals is as follows
(superscript letters indicate polarized NMR transitions): starting
material DOB (I)=PhA-CO-CH2B-PhC, diphenylethane (IV) =
(PhCH2A)2, rearrangement product PMB (III) = Ph-CO-
A
C6H5-CH2
.
in-cage products are the starting material DOB and the rearran-
gement product PMB, while the major escape product is diphe-
nylethane. The CH2 group of the benzyl radical has the largest
hyperfine coupling constant (2AH(CH2) = -16.28 G); therefore,
the most intense CIDNP signals are observed for these protons
(4.34 ppm) of DOB. The phase of this CIDNP signal (emissive) is
in agreement with Kaptein’s rules.20a Taking into account the
g-factors for the benzoyl radical (g = 2.0006) and benzyl radical
(g = 2.0026) (Δg > 0) as well as the negative sign of the HFI
constants for the CH2 group,30 the sign of the observed CIDNP is
calculated to be Γn = μεΔgAH < 0 (μ is the spin multiplicity of the
precursor, which is þ1 for a triplet, ε represents the in-cage or the
escape product, and is þ1 for the in-cage product). Thus, the
phase is predicted to be emissive, which is exactly what is observed
experimentally for these protons.
Another fairly intense peak (2.91 ppm) assigned to the CH2
groups of the escape product diphenylethane formed by the
recombination of two benzyl radicals. The absorptive phase of
this CIDNP signal is also predicted by Kaptein’s rules (μ = 1, ε =
-1, Δg > 0, AH < 0). We have also observed CIDNP for the
meta and para protons of the phenyl groups (AH < 0) (7.23 ppm)
of the starting material DOB and diphenylethane. The structure
of the CIDNP spectrum is complicated due to overlapping of
signals of these two carriers, which contribute polarizations of
opposite phase. An emissive CIDNP signal from the CH2 group
of PMB was observed at 4.98 ppm. In summary, these CIDNP
results show unambiguously that, in free solution, the dominant
reaction pathway for 3DOB* is Norrish I R-cleavage. The
structures of the products formed and their nuclear polarization
phases are those expected from the in-cage versus escape reactions
shown in Scheme 2.
Figure 1. (a) NMR spectrum of randomly methylated β-CD aqu-
eous solution; (b) NMR spectra of the encapsulated DOB and CD
(solid line) in H2O, and the (uncomplexed) materials are presented
for comparison (dashed line); (c) NMR spectrum of DOB in
CD3CN solution: β-CD (I), DOB (II) = PhA-CO-CH2C-PhB.
The asterisk marks a transition due to the residual solvent.
solution. To facilitate comparison, expanded chemical shift
regions for the two DOB phenyl group protons and the aliphatic
CD protons are shown in Figure 1b, which is the NMR spectrum
of the host/guest complex. There are two different chemical shifts
observed for the (exterior) H1 protons (5.28 and 5.05 ppm) of the
CD. As mentioned above, these arise due to the possible presence
of different substituents at the neighboring C2 position (-OH vs
-OCH3). The unresolved signals between 3.3 and 4.02 ppm are
also due to protons in the CD, but a more detailed analysis of this
region is difficult because of the overlap of signals from the two
structural types.
The 1H NMR spectrum of DOB in CD3CN (Figure 1c) consists
of fivesignals inthe aromatic region (7.4-8.1 ppm) and one signal
due to the CH2 group (4.34 ppm). The two aromatic rings of DOB
have different chemical shifts. The first aromatic ring is proximal
to the CH2 group (multiplets centered at 7.23, 7.34, and 7.56
ppm). The second ring is distal to the CH2 group (multiplets
centered at 7.74 and 7.9 ppm). The chemical shifts of the protons
in both CD and DOB are changed after complexation due to
modification of the local environment of the protons. The change
in the chemical shift of the first and second aromatic rings upon
complexation is evidence of significant interaction between the
CD cavity and both aromatic rings. Analyses by NMR for DBK
and BP ground states in β-CDs were also carried out and showed
similar features (spectra not shown).
CIDNP of DOB in Free Solution. After laser flash photo-
lysis of DOB in acetonitrile, CIDNP signals are observed, as
shown in Figure 2. The nuclear polarization is formed exclusively
from the benzyl and benzoyl radicals formed after R-cleavage. In
acetonitrile solution, the contribution from RP2 (radical pairs
from H-atom abstraction) is negligible due to the very slow rate of
this process in acetonitrile. For 3DOB* photolysis, the expected
CIDNP of Ketone/CD Complexes. The solubility of DOB
in D2O is very low, with a maximum of 0.5 mM. For this reason,
we did not observe CIDNP during the photolysis of pure DOB in
D2O solution. We therefore assume that all CIDNP signals
observed in our experiments with host/guest complexes arises
from photochemistry involving CD-included DOB molecules.
The TR-CIDNP signals from the products formed after laser
flash photolysis of DOB/CD complexes are shown in Figure 3.
Emissive polarization is observed for the CH2 (4.33 ppm) and
(30) Paul, H.; Fischer, H. Helv. Chim. Acta 1973, 56, 1575–1594.
Langmuir 2010, 26(11), 8971–8980
DOI: 10.1021/la904788t 8975

Article
Krumkacheva et al.
assigned to the H2 proton, that is, from the radical created by
H-atom abstraction at H3 (Chart 3, left side).
The phase of the CIDNP polarization for the CD protons (3.38
ppm) should be positive according to Kaptein’s rules for this
geminate radical pair (ε = 1, μ = 1, Δg < 0, AH > 0). However,
in the experiment, we observed a negative sign of polarization for
these protons. It should be noted that Kaptein’s rules were
developed to predict the phase of S-T0 radical pair mechanism
polarization. When the sign of the nuclear polarization does not
follow the sign of the hyperfine coupling constant in the inter-
mediate radicals, a possible explanation is that the polarization is
formed via the S-T- mechanism20 or due to a mutual effect of
magnetic nuclei.33 The necessary condition for Kaptein’s rules to
violated is the following: ΔgB0 < |A1|, | A2|, where A1 and A2 are
the hyperfine constants of radical 1 and radical 2, and B0 is the
magnetic field of the NMR spectrometer. In our case, this
condition is not met and we should not expect a violation of
Kaptein’s rules for our system.
Compared to free solution, the intensity of the CIDNP signal
for the escape product diphenylethane (2.9 ppm) is decreased in
CDs. There are two reasons for this observation. First, the
concentration of benzyl radicals is lower because H-atom abstrac-
tion has become an alternative reaction pathway. The product
yield of diphenylethane from R-cleavage is expected to decrease.
Second, even if R-cleavage takes place, radicals must escape from
the CD into the bulk solution in order to form this product. The
long residence time of the radicals in the CD means that nuclear
spin relaxation will take place before recombination, leading to a
decrease in the CIDNP intensity. This is in good agreement with
the conclusions of Petrova et al.23 in a CIDNP study of DBK
photolysis in unmodified β-CDs. To confirm this for our system,
we ran a similar CIDNP experiment, that is, DBK irradiated in
randomly methylated β-CDs. The results are shown in Figure 4.
No nuclear spin polarization is observed for the CD protons
(compare the chemical shift region from 3.0 to 3.5 ppm in
Figure 3b vs that in Figure 4b). The CIDNP results in Figure 4
strongly suggest that Norrish I R-cleavage is by far the dominant
reaction pathway for DBK included in CDs.
To make further sense of these reactivity pathways, CIDNP
results were also obtained for BP included in CDs, and these
results are shown in Figure 5. The BP triplet state does not
undergo the R-cleavage reaction, and so for this ketone the
H-atom abstraction reaction becomes the dominant pathway.
The intense nuclear polarization observed for the CD protons in
Figure5b strongly supports this statement. It shouldbe noted that
the phase of the polarization is also emissive, indicating that
S-T- mixing is also taking place in this system. The observed
CIDNP signals for the CD protons are also broader than those in
the case of photolysis of DOB in CD (Figure 3b), and the signals
seem to arise from several different protons rather than just one
site. This indicates that BP is much more reactive toward the
interior of the CD and will abstract protons from the H3 as well as
the H5 position, and possible even the -OCH3 protons.
Figure 6 shows two dark NMR spectra of BP/CD solutions,
one acquired before photolysis and one immediately after 10 min
of photolysis. These spectra show the disappearance of the BP
phenyl ring signals and the growth of a recombination product via
the reaction of CD and BP ketyl radicals. These signals were used
to assist in the assignment of Figure 5b above. Positive but weak
CIDNP polarization was observed in the recombination product.
The mechanism of formation of this polarization is not clear and
will be the subject of future investigations.
Figure 3. (a) Dark NMR spectrum of DOB in CD in D2O. (b)
CIDNP spectrum obtained after photolysis of DOB with CD in
aqueous solution. CIDNP assignments are made for the following
products: starting material DOB (I) = PhA-CO-CH2B-PhC,
β-CD (II), PMB (III) = Ph-CO-C6H5-CH2A, diphenylethane
(IV) = (PhCH2A)2.
phenyl (7.26 and 7.3 ppm) protons of DOB and for the CH2
protons (5.1 ppm) of the rearrangement product PMB. Observa-
tion of the rearrangement product PMB in CDs suggests that,
after R-cleavage, there is substantial mobility of at least one
member of the radical pair. Unfortunately, it is not possible to tell
from this experiment if the mobility is translational or rotational.
Positive CIDNP is observed for the CH2 protons of DPE. In
addition, polarized signals were observed for the protons of the
CD molecule at 3.2 ppm. From this observation, we can conclude
that the CD host participates in this photochemistry, most likely
via the H-atom abstraction reaction outlined in Scheme 1.
For analysis of CIDNP signal phases, g-factors and hyperfine
coupling constants for the intermediate radicals are required.
Such parameters for the DOB ketyl radical are not available from
the literature. However, we can use parameters for radicals of
similar structures, one of which is shown in Table 1.30 From the
TREPR data (see below), we have determined that the g-factor
for the ketyl radical is 2.0030, which matches well with the
literature data for the model radical. Other parameters used in
the analysis are also listed in Table 1, including those from the two
sugar radical structures used as models for the CD radicals.31,32
H-atom abstraction can take place from two accessible sites in
the CD interior, leading to two different radical structures that are
illustrated in Chart 3. The abstraction sites are the H3 or H5
protons of the sugar rings, of which there are a total of seven of
each type in a given CD structure. The expected magnetic
parameters for these radicals should be consistent with literature
parameters for sugar radicals. The CD radical on the left side of
Chart 3 is formed if H3 is abstracted from the β-CD. Sugar
radicals of this type have the following parameters: g = 2.0031
and two different hyperfine coupling constants, AH = 29.5 G and
AH = 27.4 G. The CD radical on the right side of Chart 3 is
formed when H5 is the abstraction site. Sugar radicals of this type
have g = 2.0031 and three different coupling constants: AH
=
33.3 G, AH = 9.9 G, and AH = 7.1 G. From comparison to the
values of the chemical shifts in the dark NMR spectrum of
randomly methylated β-CDs, the observed CIDNP signal is
(31) Davidson, R. S.; Wilson, R. Mol. Photochem. 1974, 6, 231–238.
(32) Gilbert, B. C.; King, D. M.; Thomas, C. B. J. Chem. Soc., Perkin Trans. 2
1981, 1186–1199.
(33) Salikhov, K. M. Chem. Phys. 1982, 64, 371–380.
8976 DOI: 10.1021/la904788t
Langmuir 2010, 26(11), 8971–8980

Krumkacheva et al.
Article
Table 1. Magnetic parameters for Simulations in Figures 7 and 9 and in CIDNP Analyses
The TR-CIDNP data presented above confirm that, in CDs, the
photochemical reaction mechanism of 3DOB* is significantly
altered compared to that observed in free solution. It moves away
from the R-cleavage reaction observed for 3DBK* and toward the
H-atom abstraction reaction observed for BP. The data reveal that
abstraction of the H3 CD proton is taking place and that there is
subsequent formation of a “tight” radical pair experiencing a large
electron spin-spin exchange interaction. This leads to formation
of CIDNP due to the S-T- mechanism, vide infra.
carried out several experiments using X-band TREPR spectro-
scopy to detect directly the free radical intermediates involved in
this photochemistry. When DOB is photoexcited in a solvent that
is a good H-atom donor, such as 2-propanol, the TREPR
spectrum in Figure 7 is obtained. Multiplet polarization (E/A)
from the radical pair mechanism (RPM) of CIDEP is observed
for both the ketyl radical and the solvent radical. Spectral
simulation identifies the signal carriers as the familiar 2-propanol
radical (seven-line spectrum with AH = 19.48 G) and the
DOB ketyl radical (center). Clearly, H-atom abstraction is the
major reaction pathway observed in free solution for DOB in
2-propanol.
Time-Resolved EPR in Free Solution. The TREPR experi-
ment can be very informative when comparing reaction pathways
in free solution versus confined media. In this regard, we have
Langmuir 2010, 26(11), 8971–8980
DOI: 10.1021/la904788t 8977

Article
Krumkacheva et al.
Figure 7. Time-resolved EPR spectra detected 300 ns after laser
flash photolysis of DOB in 2-propanol solution. The black solid
line corresponds to the experimental data, and the gray line to a
simulated spectrum of the 2-propanol radical (outer lines; see text
for parameters) and the DOB ketyl radical (unresolved lines in
center).
Figure 4. (a) Dark NMR spectrum of DBK in CD in D2O (/ =
residual H2O or HDO). (b) CIDNP spectrum obtained after
photolysis of DBK with CD in aqueous solution. Assignments
are made as follows: starting material DBK (I) = PhA-CH2
B
-
CO-CH2B-PhA, β-CD (II), PMB (III) = Ph-CH2-CO-
C6H5-CH2A, diphenylethane (IV) = (PhCH2A)2.
of the TREPR spectrum in Figure 7 corresponds to a g-factor of
2.0030. From Figure 7, we can see that the values for the hyperfine
coupling constant of the model radical Ph-C•(OH)-CH2-CH3
provide a good fit of the ketyl radical spectrum in terms of overall
spectral width and relative intensities.
It is noteworthy that the E/A pattern from RPM CIDEP is
completely symmetrical in Figure 7. The central line of the
2-propanol radical is split into emissive and absorptive compo-
nents from its -OH hyperfine splitting (0.7 G). This is a rather
striking and unusual result for a triplet state reaction, which almost
always shows some net polarization from the tripet mechanism
(TM) of CIDEP. Based on this result, we can state with certainty
that the TM is not operating in this system. Because the lifetime of
the 3DOB* is quite long in this solution (∼170 ns), the electron spin
state populations in the triplet manifold have completely relaxed to
their Boltzmann levels, which are in general undetectable by
TREPR. The same observation has been made for 3BP* in similar
solvents. Similar results to those in Figure 7 were obtained during the
photolysis of DOB in cyclohexane (spectrum not shown). This
solvent was chosen to better mimic the hydrophobic interior of the
CD moiety.
Figure 5. (a) NMR spectra of BP in CD in D2O (/ = residual H2O
or HDO). (b) CIDNP spectrum obtained after photolysis of BP in
CD in aqueous solution. (I) (Ph)2A-CO, (II) β-CD, (III) recombi-
nation product of β-CD and BP ketyl radicals.
In Figure 7, there is no evidence for the presence of benzyl or
benzoyl radicals from the R-cleavage reaction, which differs from
the CIDNP results. However, the NMR experiments were carried
out in acetonitrile, which is a relatively poor H-atom donor
compared to 2-propanol. Therefore, the rate of H-atom abstrac-
tion is substantially increased in the TREPR experiment, and this
reaction pathway dominates. In the absence of any TM contribu-
tion, RPM is the only available polarization mechanism. How-
ever, the hyperfine coupling constant differences for benzyl and
benzoyl radicals are in general quite small, and this means that
any RPM CIDEP generated by these radicals is likely to be weak.
TREPR of DOB/CD Complexes in Water. X-band
TREPR spectra recorded at several delay times after 248 nm
laser flash photolysis of DOB/CD complexes are shown in
Figure 8. While it is possible that a different wavelength of
excitation of a molecule can lead to a change in the mechanism
of a photochemical reaction due to formation of higher energy
states, we have confirmed that, for DOB, the radicals produced
and the mechanism of CIDEP formation are both independent of
the wavelength of photolysis (248 vs 308 nm). These spectra are
tentatively assigned to one or more radicals from the H-atom
Figure 6. Dashed line: Dark NMR spectrum of BP/CD in D2O
before photolysis in aqueous solution. Solid line: NMR spectrum
of BP/CD in D2O after 10 min of photolysis (/ = residual H2O or
HDO). (I) = (Ph)2A-CO, (II) = β-CD, (III) = product of recom-
bination of β-CD and BP ketyl radicals.
The phase of RPM CIDEP is defined by the sign of the
exchange interaction (J < 0) and the spin multiplicity of the
precursor μ = 1. Multiplet polarization of phase E/A is predicted
according to Γ =-μJ > 0 and is indeed observed. As mentioned
above, spectral parameters for the DOB ketyl radical were not
found in the literature. We simulated its spectrum using the
parameters for the structurally similar Ph-C•(OH)-CH2-CH3
radical (Table 1), which has been fully characterized.30 The center
3
abstraction reaction between DOB* and the randomly methy-
lated CD. The spectra are about 22 G wide (FWHM), and their
8978 DOI: 10.1021/la904788t
Langmuir 2010, 26(11), 8971–8980

Krumkacheva et al.
Article
Figure 8. TREPR spectra detected after laser flash photolysis of
DOB with CD in aqueous solution using 300 ns sampling gates at
the delay times indicated. Transitions below the baseline are in
emission.
line shape does not depend on the delay time. The most reason-
able explanation for the broad line shape is an increase in electron
spin relaxation time T2 due to a strong exchange interaction in
RP2 consistingof the ketyl and CD radicals. It is also possible that
hyperfine anisotropy is appearing in the spectra due to an increase
in the rotational correlation time of the ketyl radical when it is
confined to the CD. The time dependence of this TREPR signal
can be described using a monoexponential decay with a char-
acteristic decay constant τ = 3.1 μs.
Figure 9. (a) StickplotEPRspectrumofthe DOBketyl radical, (b)
TREPR spectrum of the experimentally observed TREPR spec-
trum detected after laser flash photolysis of DOB with CD in
aqueous solution at a delay time of 300 ns (from Figure 8), and (c)
stickplot EPR spectrum of the type I CD radical. Magnetic
parameters for simulations are listed in Table 1. See text for
discussion.
To confirm this assignment, the experimental spectrum ac-
quired at 300 ns delay (Figure 8) was compared with two
calculated TREPR spectra for RP1 and RP2. Figure 9 shows
the experimental spectrum from Figure 8 (300 ns delay time) with
simulations for different radical pairs: (a) RP2, that is, the DOB
ketyl radical and the typeI CD radical (Chart 3, left), and (b) RP1,
the benzoyl radical and the benzyl radical. The type I CD radical
is used in the simulations to remain consistent with the CIDNP
results above. Magnetic parameters used in the simulations were
taken from Table 1 for all species. From Figure 9, we conclude
that RP2 (DOB ketyl radical and type I CD radicals), with a line
width of 5 G, gives the best fit to the experimental spectrum.
The net emissive polarization in the spectra shown in Figure 8
can be formed via two distinct mechanisms: the triplet mechanism
(TM) or an S-T- contribution to the RPM. As previously
discussed, in order to observe TM CIDEP from radical reactions
involving DOB, the rates of the chemical reactions leading to the
radicals (i.e., the inverse of the triplet lifetime) must compete with
electronic spin relaxation in the triplet state (1-10 ns). It is known
from the literature that the lifetime of 3BP* in CD aqueous solution
(350 ns) does not change much from its lifetime in cyclohexane
(300 ns).34 The molecular structures of DOB and BP are similar,
and we can assume that their rate constants for H-atom abstraction
from β-CD should also be similar. The BP/CD CIDNP results
above, coupled with our solution TREPR experiments presented
here for DOB, support our assertion that TM polarization is
unlikely to be generated in reactions of DOB in CDs.
Figure 10. (a) Stickplot TREPR spectrum of the BP ketyl radical,
(b) experimentally observed TREPR spectrum detected after laser
flash photolysis of BP with CD in aqueous solution at a delay time
of 600 ns, and (c) stickplot TREPR spectrum of the type I CD
radical. Magnetic parameters for simulations are listed in Table 1.
As a rule, when the S-T- mechanism of CIDEP is operative,
the observed spectra are asymmetric, because S-T- flip-flop
transitions are allowed only for definite spin levels; thus, EPR
lines corresponding to T- levels with Ri projection of nuclear spins
are the only ones that become polarized. However, the observed
CIDEP spectrum in Figure 8 has a symmetrical line shape. It is
possible that fast nuclear relaxation of the radical electronic spin
levels is taking place, which can happen when there are numerous
nuclei coupled to both electrons. Figure 10b shows the TREPR
spectrum obtained when BP/CD complexes are irradiated. The
same broad emissive signal as observed in Figure 8 for DOB/CD
is observed here.
The CIDNP results in Figure 5 clearly indicate that H-atom
abstraction from the CD cavity interior by 3BP* is indeed taking
place during the photolysis of the BP/CD complex. This means
that the CIDEP spectrum has contributions from the BP ketyl
radical and at least type I CD radicals if not both. The FWHM of
the observed signal (2.8 mT) in Figure 10b is less than the FWHM
(34) Monti, S.; Flamigni, L.; Martelli, A.; Bortolus, P. J. Phys. Chem. 1988, 92,
4447–4451.
Langmuir 2010, 26(11), 8971–8980
DOI: 10.1021/la904788t 8979

Article
Krumkacheva et al.
of the type I CD radical (5 mT), but is larger than the FWHM of
the BP ketyl radical (1.1 mT). It should also be noted that the
observed FWHM in Figure 10b is about two times less than the
predicted spectral width of the type I CD radical. This indicates
that the exchange interaction in this RP is substantially larger
than the HFI constants of the type I CD radical. The S-T-
emissive phase of the observed CIDEP also points to the existence
of a large exchange interaction in this system, comparable to the
Zeeman interaction. We did not perform an accurate simula-
tion of the TREPR spectra, as this needs to take into account, for
each individual radical pair system, the value of the exchange
interaction and its degree of modulation by radical motion inside
the CD cavity. We know little about these parameters at the
moment.
It is interesting to note that the asymmetry typical of S-T-
mixing is clearly present in the BP/CD (Figure 10b) spectrum, in
contrast with the DOB/CD CIDEP (Figure 8) spectrum. This
could be due to the difference in spin relaxation times. For the
DOB ketyl radical, the spin relaxation time is expected to be less
than that for BP ketyl radical, because of the former’s larger HFI
constants (see Table 1).
Another possible interaction that can affect the shape of EPR
spectrum and create CIDEP is the electron dipole-dipole inter-
action. If the distance between the two radical centers in RP2 does
not exceed several angstroms, then the value of the dipole-dipole
interaction will be in the range of 1-10 mT. As shown by
Shushin,35 in the absence of an exchange interaction, the dipole-
dipole interaction itself can generate multiplet polarization with
opposite phase to that contributed by S-T0 RPM CIDEP. We
did not observe any multiplet polarization in our spectrum, and
we can conclude either that the contribution from this polariza-
tion is negligible or that it is canceled by S-T0 RPM CIDEP. Of
course, we expect a large exchange interaction to be present in
RP2, which might lead to a different polarization pattern than
that predicted by Shushin. Such a theoretical consideration is
outside the scope of the present paper. It should be noted that, in
previous reports of TREPR spectroscopy of flexible biradicals
experiencing large exchange interactions and S-T- mixing, the
same asymmetry appears in the spectra.
Figure 11. X-band TREPR spectrum acquired 500 ns after 308
nm laser flash photolysis of the DOB/CD complex in a frozen
aqueous solution at 100 K. Note the half-field transition near B0 =
100 mT, which is the signature of a randomly oriented molecular
triplet state.
field transition is observed at about 100 G, highly indicative of a
randomly oriented powder pattern triplet state EPR spectrum.38
From the approximate D value of 2000 G (based on the spectral
width) and estimated rotational correlation time of β-CDs in
water (τr ∼ 3 ns), we can conclude that the lifetime of the triplet
molecule is substantially longer that its electron relaxation time.
Thus, we can state conclusively that we are not observing
molecular triplet states in our TREPR spectra of DOB/CD
complexes in water.
Conclusions
We have examined in detail the photophysics and photochem-
istry of DOB, DBK, and BP in permethylated β-CD inclusion
complexes and compared these reactions to free solution using
time-resolved magnetic resonance (TREPR, TR-CIDNP) tech-
niques. The results clearly show that, in confined conditions such
as the interior of β-CDs, the reaction pathways for the DOB
triplet state include both R-cleavage and H-atom abstraction.
This correlates well with the behavior of DOB in free solution;
however, the partitioningof the reactionpathways depends onthe
presence or absence of a H-atom donor. Many of the free radicals
involved in this photochemistry have been characterized for the
first time, as have their products. The rearrangement product of
the radicals from R-cleavage (PMB) has been conclusively identi-
fied by NMR spectroscopy for the first time. The observation of
this rearrangement implies that there is significant mobility of the
radicals in the CD interior, but whether it is translational or
rotational freedom remains unclear. The spin polarization me-
chanism observed in DOB/CD and BP/CD complexes is attrib-
uted to S-T- RPM CIDEP, with no contribution from the TM in
the case of either DOB or BP. The exchange interaction between
the members of RP1 must be large, quite possible comparable to
the Zeeman interaction at X-band (∼0.34 mT).
We have also studied CIDEP during the photolysis of 3DBK*
in CDs using TREPR (spectra not shown). It is known that the
lifetime of the triplet DBK molecule is very short in solution at
room temperature36 and that the only reaction pathway to
radicals is R-cleavage. In cyclohexane solution, we observed a
strong positive CIDEP polarization due to the TM, in agreement
with literature data37 and a small contribution from the RPM
(E/A). We did not observe any CIDEP during the photolysis
of aqueous solutions of DBK in CDs. The most reasonable
explanation for this is the fast recombination of radicals formed
after R-cleavage.
TREPR of DOB and DOB/CD Complexes in Frozen
Solutions. The signal carrier in Figure 8 could be the partially
averaged DOB molecular triplet state. To rule out this possibility,
an X-band TREPR spectrum of the DOB/CD complex in frozen
aqueous solution was obtained at 100 K. The spectrum we
obtained at a delay time of 500 ns, using flash photolysis at 308
nm and 300 ns gates, is shown in Figure 11. A very strong half-
Acknowledgment. M.D.E.F. thanks the National Science
Foundation for support of this research through Grant No.
CHE-0809530. E.G.B., O.A.K., and V.R.G. thank the Russian
Foundation for Basic Research through Grant 08-04-00555,
Russian Federal Agency for Education (GK 1144), and Scientific
School 3604.2008.3. We thank Dr. V. F. Tarasov for fruitful
discussions.
(35) (a) Shushin, A. I. Chem. Phys. Lett. 1991, 183, 321–326. (b) Shushin, A. I.
Chem. Phys. Lett. 1999, 260, 261–270.
(36) Arbour, C.; Atkinson, G. H. Chem. Phys. Lett. 1989, 159, 520–525.
(37) Turro, N. J.; Paczkowski, M. A.; Zimmt, M. B.; Wan, J. K. S. Chem. Phys.
Lett. 1985, 114, 561–565.
(38) Tominaga, K.; Yamauchi, S.; Hirota, N. J. Phys. Chem. 1990, 94, 4425–
4431.
8980 DOI: 10.1021/la904788t
Langmuir 2010, 26(11), 8971–8980