Using cyclodextrins to encapsulate oxygen-centered and carbon-centered radical adducts: The case of DMPO, PBN, and MNP spin traps

Article abstract of DOI:10.1021/jp100777u

We present electron spin resonance (ESR) experiments that describe the interaction of β-cyclodextrin (β-CD) with spin adducts of three spin traps: 5,5-dimethyl-1-pyrroline N-oxide (DMPO), N-tert-butyl-α- phenylnitrone (PBN), and 2-methyl-2-nitrosopropane

Full text of DOI:10.1021/jp100777u

J. Phys. Chem. A 2010, 114, 6217–6225
6217
Using Cyclodextrins to Encapsulate Oxygen-Centered and Carbon-Centered Radical
Adducts: The Case of DMPO, PBN, and MNP Spin Traps
Mariana Spulber and Shulamith Schlick*
Department of Chemistry and Biochemistry, UniVersity of Detroit Mercy,
4001 West McNichols Road, Detroit, Michigan 48221
ReceiVed: January 26, 2010; ReVised Manuscript ReceiVed: April 29, 2010
We present electron spin resonance (ESR) experiments that describe the interaction of ꢀ-cyclodextrin (ꢀ-CD)
with spin adducts of three spin traps: 5,5-dimethyl-1-pyrroline N-oxide (DMPO), N-tert-butyl-R-phenylnitrone
(PBN), and 2-methyl-2-nitrosopropane (MNP). The focus was on spin adducts of oxygen-centered radicals
trapped by DMPO and PBN and on carbon-centered radical adducts trapped by MNP. The radicals were
generated by reaction with hydroxyl radicals and the spin adducts studied were DMPO/OH and PBN/OH,
MNP/CH2COOH generated in CH3COOH, and MNP/CF2COOH in CF2HCOOH. Di-tert-butyl nitroxide
((CH3)3C)2NO (DTBN) was also detected in experiments with MNP as the spin trap. A range of interactions
of the spin adducts and DTBN with ꢀ-CD was identified. The presence of ꢀ-CD led to significant stabilization
of DMPO/OH and PBN/OH but to a negligible effect on the 14N hyperfine splitting of the adducts, aN, indicating
that the N-O group is outside the ꢀ-CD caVity. An increase of aN was detected for DTBN and MNP/CH2COOH
in CH3COOH in the presence of ꢀ-CD, a result we assigned to bonding at the rim of the host. Experiments
with methylated ꢀ-CD (Me ꢀ-CD) provided support for this conclusion. A different type of complexation
was detected for DTBN and MNP/CF2COOH in CF2HCOOH: for specific host concentrations both “in” and
“out” species were detected. We suggest that the hydrophobicity of the fluorinated adduct leads to insertion
of the adduct inside the host cavity. Calculation of the association constant Ka indicated the competition
between DTBN and the adduct for inclusion in the host. For MNP as spin trap, the two nitroxide radicals
(adduct and DTBN) have the same type of interaction with the host: at the rim in acetic acid, and inside the
host cavity in CF2HCOOH. Experiments with DTBN in the absence of the spin trap and of adducts illuminated
the effect of the local polarity and of the pH on the hyperfine splittings and indicated that the presence of
acetic acid encourages rim complexation.
Introduction
of aN by interaction with the host has become an important
indicator of guest-host complexation in numerous ESR studies.
The ability of cyclodextrins (CDs) to form inclusion com-
plexes by noncovalent bonding with a variety of guests has
become an exciting field of research, with important applications
in chromatography, drug delivery, and catalyst stabilization.1,2
CDs are cyclic oligosaccharides containing six (R-CD), seven
(ꢀ-CD), or eight (γ-CD) D-(+) glucopyranose units attached
by R-(1,4) glucosidic bonds, with lipophilic inner cavities and
hydrophilic outer surfaces.11 In aqueous solutions the slightly
nonpolar cyclodextrin cavity contains water that can be replaced
by guests with lower polarity.3
Kotake and Janzen have studied the effect of pH on the inclusion
process and assigned the lower degree of inclusion at higher
pH values to the protons generated by dissociation of the OH
groups; the same argument was also used to explain that CDs
are poor hosts for ionic species.6 The same authors detected
“multimodal” inclusion: in the case of a nitroxide with several
functional groups, different complexes were generated, depend-
ing on the group that became the guest.7 The choice of an
interesting nitroxide led to the detection, with excellent resolu-
tion due to D-enrichment, of different ESR spectra for the “in”
and “out” nitroxides (inside and outside the CD host), with
assignment based on the aN values.8 More recently, it was
demonstrated that pulse ESR methods such as electron spin-echo
envelope modulation (ESEEM) and double electron-electron
resonance (DEER) can lead to additional structural information
compared to continuous wave (CW) ESR.9
CD inclusion complexes with nitroxide radicals have been
studied early on by electron spin resonance (ESR) spectroscopy
and take advantage of the ability of this method to detect the
effect of the local environment on the guest with great specificity
and sensitivity. Atherton and Strach have investigated the
association of di-tert-butyl nitroxide ((CH3)3C)2NO (DTBN)
with R-CD and ꢀ-CD at 296 K and calculated the equilibrium
constant for ꢀ-CD, taking into consideration that the solution
is not ideal.4 Martinie et al.5 have demonstrated the complexation
of R-CD and ꢀ-CD with several nitroxides by ESR. In both
papers4,5 the nonpolar environment was reflected in lower
hyperfine splitting (hfs) from the 14N nucleus, aN, in guests
enclosed in the CDs, compared to aqueous solutions. Lowering
NMR studies have added important details on the relative
location of guests in R-, ꢀ-, and γ-CDs.10 An NMR study of a
series of R-phenyl-N-tert-butylnitrones (PBNs) in ꢀ-CD has
determined the stoichiometry of the guest-host complex (1:1)
and the corresponding binding constants.11 These NMR studies,
which focus on the host as modified by the guest, offer
complementary information to the ESR study of CD-nitroxide
complexes, whose focus is on the guest.
* Corresponding author. Telephone: 1-313-993-1012. Fax: 1-313-993-
1144. E-mail address: schlicks@udmercy.edu (S. Schlick).
10.1021/jp100777u  2010 American Chemical Society
Published on Web 05/12/2010

6218 J. Phys. Chem. A, Vol. 114, No. 21, 2010
Spulber and Schlick
CHART 1: Spin Traps Used in the Present Study
fluorinated adduct MNP/CF2COOH. These latter adducts were
generated by attack of HO• radicals on acetic acid (CH3COOH,
AA) and difluoroacetic acid (CF2HCOOH, DFAA), respec-
tively.19 DTBN was also generated in experiments with MNP
as the spin trap, and its complexation by ꢀ-CD was detected.
Most experiments were performed with ꢀ-CD, but additional
clarifications were obtained by using methylated ꢀ-CD (Me
ꢀ-CD). In addition, the presence of both adduct and DTBN in
systems with MNP as the spin trap necessitated experiments of
DTBN/ꢀ-CD (in the absence of MNP); these experiments
illuminated the roles of AA and DFAA in complex formation
and validated our conclusion that in some systems the guest
was located on the rim of the host. Some of the aspects studied
in this paper, for example, fluorinated spin adducts and the effect
of the local polarity, have not been discussed in the existing
literature.
In contrast to the extensive literature on ESR studies of
CD-nitroxide complexation, there are only few reports on the
interaction between CDs and nitroxide spin adducts. These
systems are more complicated than the CD-nitroxide systems
because they also include the spin trap, the radical source, the
attacking species (typically HO• radicals), and in some cases
additional paramagnetic species that can also exhibit “in” and
“out” ESR spectra. Despite this complexity, study of spin
adducts in CDs and the ability of CDs to increase the lifetimes
of spin adducts have been reported recently. Tordo and
co-workers have described the increased stability of nitrone
adducts of the superoxide radical anion and their inclusion
complexes with randomly methylated or selectively methylated
ꢀ-CD.12,13 Of particular interest for the present study is the
inclusion of MNP adducts of alkyl radicals in ꢀ-CD: Encapsula-
tion led to broader line widths in the ESR spectra of the adducts,
but the higher solubility of MNP in water and the increased
lifetimes of the adducts were major advantages.14 In an
interesting recent approach, Han et al. described the interaction
of a ꢀ-CD-nitrone conjugate with the superoxide radical anion
and used molecular modeling to predict the stability of adducts
and their longer lifetimes in the presence of the host.15
Experimental Section
Materials. DMPO, PBN, MNP, ꢀ-CD, and Me ꢀ-CD were
purchased from Sigma-Aldrich and used as received. Hydrogen
peroxide, H2O2, was obtained from Sigma-Aldrich as a 3%
aqueous solution. Deionized “ultrapure” water with low con-
ductivity and <10 ppb total organic carbon was provided by a
Millipore Model Direct-Q UV system and was used in all
experiments. MNP was supplied by Sigma-Aldrich as a
diamagnetic dimer, which is known to dissociate in aqueous
solutions26
(CH3)3CsNdNsC(CH3)3 f 2(CH3)3CsNdO
The major motivation in initiating our study of nitroxide
complexation by CDs has emerged from spin trapping studies
in our laboratory whose goal was to detect and identify oxygen-
centered and carbon-centered free radicals generated during the
fragmentation of polymeric membranes used in fuel cells and
of corresponding model compounds.16-19 Ex situ experiments
in the laboratory have been performed by generating hydroxyl
radicals, HO•, via UV irradiation of hydrogen peroxide.20 Direct
ESR detection of radicals was possible only in a limited number
of cases, by lowering the temperature in order to increase the
lifetimes of unstable intermediates.16 Most systems required spin
traps in order to scavenge the short-lived radicals, a process
that leads to the formation of the more stable nitroxide radical
adducts.21,22 The spin traps used in our studies were 5,5-
dimethyl-1-pyrroline N-oxide (DMPO), N-tert-butyl-R-phe-
nylnitrone (PBN), and 2-methyl-2-nitrosopropane (MNP), Chart
1, and identification of adducts was facilitated in some cases
by the spin trapping database.23 In the study of model com-
pounds19 and in the in situ study of a fuel cell inserted in the
resonator of the ESR spectrometer24,25 we have described the
detection and identification of carbon-centered radicals (CCRs)
and oxygen-centered radicals (OCRs) as DMPO and MNP
adducts. In some cases the detection of spin adducts was
hampered because of their low stability. For these reasons we
explore here the effect of CDs on oxygen-centered radical
adducts of DMPO and PBN, and on carbon-centered radical
adducts of MNP. The focus in the present study was on the
effect of CD on the lifetimes of spin adducts, the variation of
hfs due to encapsulation, the location of the nitroxide in the
complex, the effect of the pH, and the importance of guest
hydrophobicity on the interactions CD-guest.
V
V
O
O
Sample Preparation. The DMPO/OH and PBN/OH spin
adducts were obtained by mixing 0.1 mL of aqueous solution
of the spin trap (1 × 10-3 M DMPO and 2.5 × 10-3 M PBN,
respectively) with 0.01 mL of 3% H2O2. The pH of each solution
was adjusted to the range 5-7 with 0.1 M NaOH calibrated
solution; this procedure was performed because nitroxide
radicals are less stable in acidic media.27 DMPO and PBN
adducts were generated at a neutral pH, to enhance their stability.
The generation and decay (“time scans”) of the DMPO and
MNP adducts were followed by in situ UV irradiation of the
appropriate solutions (in the presence and in the absence of
ꢀ-CD) at 300 K in quartz capillary tubes placed inside the ESR
resonator, using a 300 W ozone-free Xe arc equipped with a
water filter (Oriel).
Attack of hydroxyl radicals on acetic acid leads to the
•
formation of the CH2COOH radical; the corresponding MNP
adduct has aN ) 15.7 G and aH ) 8.6 G (2H) at pH ) 7.19 In
the case of CF2HCOOH as the model compound, the •CF2COOH
radical is generated and the corresponding MNP adduct has aN
) 13.4 G and aF ) 22.6 G (2F) at pH ) 7; aN is significantly
lower than for MNP/CH2COOH because of the higher fluorine
1
electronegativity, and aF is much higher than that for the H
•
hfs in CH2COOH, 22.6 vs 8.6 G.19 The MNP/CCR adducts
were prepared by mixing 0.1 mL aqueous solution of MNP (2
× 10-4 M) with 1 mL AA or DFAA (2M) and 0.05 mL of
H2O2 3%. The pH of the solution was adjusted to 5 with NaOH,
in order to facilitate the inclusion process that is enhanced at
lower pH values.6,28 All mixtures were UV irradiated at room
temperature in quartz capillary tubes for 10 min with a low-
pressure mercury source (Mineralight Model PCQX1). The
stability of the MNP adducts prepared as described above is
high, typically 3 days at ambient temperature; see next section.
As seen below, we have determined the various effects of
ꢀ-CD presence: From the much higher stability of DMPO/OH
and PBN/OH adducts in the presence of ꢀ-CD in aqueous
solutions at 300 K, to the different types of complexes of ꢀ-CD
with the protiated adduct MNP/CH2COOH and with the

Cyclodextrins to Encapsulate Radical Adducts
J. Phys. Chem. A, Vol. 114, No. 21, 2010 6219
Figure 1. (A) Generation of the DMPO/OH adduct and its decay at 300 K in the absence (line 1) and in the presence of several ꢀ-CD concentrations:
0.5 × 10-3 M (line 2), 1 × 10-3 M (line 3), 2 × 10-3 M (line 4), 3 × 10-3 M (line 5), 3.5 × 10-3 M (line 6). The red arrow indicates the signal
whose intensity was monitored in the time scan. (B) The half-life of the DMPO/OH adduct as a function of ꢀ-CD concentration, determined at the
half-maximum intensity of the signal shown by the red arrow after stopping the UV irradiation. The solid line was obtained by linear fit.
The appropriate amounts of ꢀ-CD were added to an aliquot of
the spin adduct, and the mixture was transferred to capillary
tubes for ESR measurements. The MNP adducts are more stable
and complexation by ꢀ-CD was performed at pH 5; the
hyperfine splittings under these conditions are slightly different
compared to those published in ref 19; see next section.
ESR Measurements. Spectra were recorded at 300 K using
Bruker X-band EMX spectrometers operating at 9.7 GHz and
100 kHz magnetic field modulation, and equipped with the
Acquisit 32 Bit WINEPR data system version 3.01 for acquisi-
tion and manipulation, and the ER 4111 VT variable temperature
units. The microwave frequency was measured with the Hewlett-
Packard 5350B microwave frequency counter. The hyperfine
splittings of the spin adducts were determined by fitting the
spectra using the WinSim (NIEHS/NIH) simulation package;29
the fitting also determines the relative intensity of each
component for spectra that consist of a superposition of
contributions from different adducts or radicals. The weak
satellites from 13C, indicated by arrows in Figure 11, were not
included in the simulations shown in Figures 4A, 11, and 12.
Typical acquisition parameters for all ESR spectra were sweep
width 100-160 G, microwave power 2 mW, time constant 20.48
ms, conversion time 41.94 ms, 2048 points, modulation
amplitude 0.2-0.4 G, receiver gain 5 × 104, and 20-40 scans.
irradiation; in the presence of the host, however, the signal can
be detected for time intervals that depend on ꢀ-CD concentra-
tion. The decay rate of the DMPO/OH spin adduct is slower
with the increase of host content; for example, 2500 s after the
initial start of the irradiation, the concentration of the spin
adducts is 2-5 times higher in the presence of ꢀ-CD than in
its absence.
Figure 1B presents the half-life of DMPO/OH, τ1/2, deter-
mined at the half-maximum intensity of the signal shown in
the inset of Figure 1A; encapsulation of the DMPO/OH adducts
leads to an increase of their half-life from 700 s the absence of
ꢀ-CD to 1600 s for the highest host concentration, [ꢀ-CD] )
3.5 × 10-3 M.
The complexation process between PBN/OH spin adducts and
ꢀ-CD was studied at 300 K for host concentrations in the range
(1-100) × 10-3 M. The presence of CD resulted in a 2-10-
fold increase in the adduct intensity, but aN and aH splittings
were unchanged, aN ) 15.1 G and aH ) 3.4 G.12b The major
effect of the host presence is the increased adduct stability, as
clearly seen in the time scans shown in Figure 2A. In the absence
of ꢀ-CD the decay of the adduct is fast and no signal was
detected after 100 s, in agreement with the literature.30 It is
interesting to note that the variation of τ1/2 with host concentra-
tion is different for the two spin traps: Figure 1B shows a linear
dependence for DMPO/OH, but Figure 2B shows an abrupt
increase of τ1/2 for PBN/OH in a narrow range of ꢀ-CD
concentrations, (15-30) × 10-3 M.
Results and Discussion
DMPO and PBN as Spin Traps. The effect of ꢀ-CD on the
stability and hyperfine splittings of the DMPO/OH adduct was
studied at 300 K, and the ꢀ-CD concentrations were in the range
(0.5-3.5) × 10-3 M. Changes in the values of aN with
increasing ꢀ-CD concentration were within experimental error,
≈0.1 G, and only slight increase in the line widths, typically
0.2 G, were detected when the ESR spectra were measured in
the temperature range 290-320 K. However, a significant
increase in the stability of the adduct was measured: The time
scans presented in Figure 1A show the intensity of the adduct
during UV irradiation and the decay after the irradiation was
stopped, for the indicated ꢀ-CD concentrations; the red down-
ward arrow in the inset points to the monitored signal. We note
the higher intensity of the adduct during its generation, indicating
the formation of the inclusion complex, and its slower decay
in the presence of ꢀ-CD; both effects are more pronounced for
the higher ꢀ-CD concentrations. In the absence of ꢀ-CD the
adduct intensity decays immediately after stopping the UV
The greater stability of the DMPO/OH and PBN/OH adducts
and the negligible effect of the host on aN and aH values of the
adducts are taken as indicators that the nitroxide group is outside
the host cavity.31 In Figure 3 we present the proposed structures
for the partial inclusion of these adducts in ꢀ-CD: “in” for the
methyl groups of DMPO and for the benzene ring of PBN,7
but “out” for the N-O groups, which are exposed to the solvent.
To obtain the structure shown in Figure 3, the structures of the
adducts were optimized using the Gaussian 03 software followed
by insertion in the CD, based on the picture of complexation
deduced from the ESR spectra (nitroxide group outside the host).
The MNP/CH2COOH Adduct and DTBN in Acetic Acid.
Attack of hydroxyl radicals on acetic acid (AA) in aqueous
solutions at 300 K and pH ) 5 leads to the generation of the
MNP/CH2COOH adduct with aN ) 15.4 G and aH ) 8.3 G
(12%), and of the DTBN nitroxide radical with aN ) 16.7 G
(88%), as seen in Figure 4A. The aN value for DTBN is lower

6220 J. Phys. Chem. A, Vol. 114, No. 21, 2010
Spulber and Schlick
Figure 2. (A) Generation of the PBN/OH adduct and its decay at 300 K in the absence (line 1) and the presence of various ꢀ-CD concentrations:
10-2 M (line 2), 2.5 × 10-2 M (line 3), 5 × 10-2 M (line 4), 7.5 × 10-2 M (line 5). The red arrow indicates the signal whose intensity was
monitored in the time scan. (B) The half-life of the PBN/OH adduct as a function of ꢀ-CD concentration, determined at the half-maximum intensity
of the signal shown by the red arrow after stopping the UV irradiation. The solid line was obtained by connecting the points by a ꢀ spline line.
Figure 3. Proposed inclusion of the DMPO/OH adduct (A) and of the PBN/OH adduct (B) in ꢀ-CD. Note the position of the N-O group outside
the host cavity for both adducts. The structures of the adducts were optimized using the Gaussian 03 software, and the adduct was inserted in the
CD, based on the picture of complexation deduced from the ESR spectra (nitroxide group outside the host). See text.
than that measured in water, where a value of 17.2 G was
reported.4 This difference will be discussed below, in experi-
ements with DTBN/ꢀ-CD in the absence of the adduct.
Changes in the stability, relative intensity, and hfs were
measured for ꢀ-CD concentrations in the range (1-100) × 10-4
M. At the lower host concentrations, [ꢀ-CD] < 5 × 10-4 M,
changes in the relative intensities of the two spectral components
were detected: a decrease for DTBN and an increase for MNP/
CH2COOH, accompanied by very slight changes of aN and aH
values ((0.1 G). For higher host concentrations, the relative
intensity of the adduct continued to increase; even more
important were the pronounced increases in aN values for both
nitroxides. The ESR spectra for [ꢀ-CD] ) 10-3 M are shown
in Figure 4B: aN for both DTBN and the MNP/CH2COOH
increased from 16.7 to 17.3 G for DTBN and from 15.4 to 6.1
G for MNP/CH2COOH. These results are clear indications that
both the adduct and DTBN in the presence of the host are in a
more polar environment. We propose that these guests are
located at the rim of the host where the higher polarity is due
to the presence of OH groups.32
(1-30) × 10-4 M, and the ESR spectra at two Me ꢀ-CD
concentrations (10-4 M and 3 × 10-3 M) are presented in Figure
S1 (Supporting Information). The aN value for the MNP/
CH2COOH adduct is the same for the two Me ꢀ-CD concentra-
tions, aN ) 16.1 G, and equal to that in the presence of ꢀ-CD,
concentration 10-3 M (Figure 4B); this value of aN is higher
than that for the adduct in water, thus confirming the rim location
proposed in ꢀ-CD. The aN values for DTBN, however, followed
a different pattern: 16.7, 16.9, and 16.3 G for [Me ꢀ-CD] ) 0,
10-4, and 3 × 10-3 M, respectively. The lower aN value at the
highest Me ꢀ-CD concentration suggests inclusion in the host
cavity. A possible explanation for this behavior is the higher
hydrophobicity of both Me ꢀ-CD (compared to ꢀ-CD) and of
DTBN compared to the MNP/CH2COOH adduct. The subtle
effects of host and guest polarities seem to dictate the com-
plexation site: at the rim for the adduct, and inclusion for DTBN
for a high concentration of the less polar host, Me ꢀ-CD.
The stability of the MNP/CH2COOH adduct increased
significantly upon addition of the ꢀ-CD host. Figure 5 presents
the intensity of the adduct in the absence of ꢀ-CD and for
[ꢀ-CD] ) 1 × 10-3 M. The initial concentration of the adduct
is larger by a factor of ≈2, and the adduct is detectable for
≈150 h compared to 70 h in the absence of ꢀ-CD. The plots
Experiments with Me ꢀ-CD as host were performed in order
to better understand the local polarities and to validate the
proposed location for the spin adduct and DTBN in the inclusion
complexes. The concentrations of Me ꢀ-CD were in the range

Cyclodextrins to Encapsulate Radical Adducts
J. Phys. Chem. A, Vol. 114, No. 21, 2010 6221
Figure 6. Experimental and simulated ESR spectra of the MNP/
CF2COOH adduct and of DTBN detected at 300 K and pH 5: (A)
[ꢀ-CD] ) 0, (B) [ꢀ-CD] ) 10-3 M. The hyperfine splittings of the
adducts are shown on the right and their relative intensity on the left.
Figure 4. Experimental and simulated ESR spectra of the MNP/
CH2COOH adduct and of DTBN detected at 300 K and pH 5: (A)
[ꢀ-CD] ) 0, (B) [ꢀ-CD] ) 10-3 M. The hyperfine splittings of the
adducts are shown on the right and their relative intensity on the left.
concentrations of Me ꢀ-CD, the DTBN radical becomes
dominant in the system; this increase of the DTBN relative
intensity is accompanied by its inclusion in the Me ꢀ-CD cavity.
We note that at pH ) 7 the addition of ꢀ-CD had no effect on
the hfs of both adduct and DTBN: no evidence for complexation
was detected.
MNP/CF2COOH and DTBN in CF2HCOOH (DFAA). The
results for the MNP/CF2COOH adduct generated when aqueous
solutions of CF2HCOOH are exposed to hydroxyl radicals are
shown in Figures 6-9. Figure 6A presents experimental ESR
spectra, corresponding simulated ESR spectra, and their decon-
volution into contributions from the fluorinated adduct (aN )
13.5 G and aF ) 22.4 G (2F), 52%), and from DTBN (aN )
17.1 G, 48%). The aN value for DTBN is higher in the DFAA
system compared to that in AA (aN ) 16.7 G, Figure 4A). At
this stage we tentatively assign this difference to the higher
polarity in aqueous DFAA compared to aqueous AA; this
conclusion will be validated by experiments with DTBN in the
absence of the adduct; see below.
Figure 5. The stability of the MNP/CH2COOH adducts at 300 K in
the absence (black solid square) and in the presence (red open circle)
of ꢀ-CD, concentration 10-3 M. The solid line was obtained by a spline
line through the experimental points.
Figure 6B shows the results for a host concentration of 10-3
M. Addition of the host leads to the detection of two types of
MNP/CF2COOH adducts and DTBN radicals, with different aN
values. For the adduct we assign the two species to the “in”
and “out” adducts on the basis of their aN values, 12.7 G for
“in” and 13.4 G for “out”. We notice that one type of adduct is
identical to that detected in the absence of ꢀ-CD (aN ) 13.5 G,
also show that the rate of decay of the adduct intensity in the
presence and in the absence of the host are similar.33
The relative intensity of MNP/CH2COOH increased signifi-
cantly in the presence of ꢀ-CD, for example to 90% for [ꢀ-CD]
) 10-3 M. A similar behavior was also observed in the case of
Me ꢀ-CD, up to a host concentration of 10-3 M. At higher

6222 J. Phys. Chem. A, Vol. 114, No. 21, 2010
Spulber and Schlick
adduct (Figure 5): The initial concentration of the adduct is
larger by a factor of ≈2, and the adduct is detectable for >150
h compared to 70 h in the absence of ꢀ-CD, and as in the case
of MNP/CH2COOH, the plots also show that the rate of change
of the adduct intensity in the presence and in the absence of
the host are similar.33
Table 1 shows the variation of the “in” and “out” species for
MNP/CF2COOH and DTBN as a function of ꢀ-CD content. It
is interesting to note that at the highest ꢀ-CD concentration all
the DTBN is “in” (and 17% of the total ESR signal intensity),
while only 73% of the total adduct is “in” (and 61% of the
total ESR signal intensity). Additional important details on the
effect of ꢀ-CD were observed: First, DTBN “in” was detected
at a lower ꢀ-CD concentration compared to the spin adduct.
Second, the large increase of both “in” species in a narrow range
of ꢀ-CD concentration was detected, (2.5-10) × 10-4 M for
DTBN and (5-8) × 10-4 M for the spin adduct, as shown in
Figure 8.
Figure 7. The stability of the MNP/CF2COOH adducts at 300 K in
the absence (black solid square) and (2) in the presence (red open circle)
of ꢀ-CD, concentration 10-3 M. The solid line was obtained by a spline
line through the experimental points.
The association constants, Ka, as a function of ꢀ-CD
concentration (Table 2) may be estimated based on the reason-
able assumption that the concentration of ꢀ-CD is higher than
those of the nitroxides.34 For these conditions:
Ka ) r1/(ro[CD])
(1)
In eq 1 r1 is the concentration of the “in” species and ro is
the concentration of the “out” species, as shown in Table 1.
The values are plotted in Figure 9 for both MNP/CF2COOH
and DTBN. We note that DTBN is a strong competitor for the
adduct in terms of inclusion in the host. For [ꢀ-CD] ≈ 9 ×
10-4 M, Ka for DTBN increases dramatically, while Ka for the
adduct starts to decrease. These results reinforce the idea that
the two nitroxide radicals are competitors for inclusion in the
host.
The attachments of the two spin adducts, MNP/CH2COOH
in acetic acid and MNP/CF2COOH in fluoroacetic acid, to the
ꢀ-CD host are different. The ESR spectra presented in Figures
4 and 6, together with the variations of aN values determined
by the simulations of the spectra, suggest rim attachment for
the MNP/CH2COOH adduct and inclusion of the fluorinated
adduct MNP/CF2COOH inside the host cavity. It seems reason-
able to suggest that the hydrophobicity of the fluorinated adduct
drags the adduct into the host cavity. The contrasting complex-
ation for the two adducts is shown in Figure 10A for the MNP/
CH2COOH adduct and in Figure 10B for the MNP/CF2COOH
adduct; Figure 10 was generated as described for the structures
shown in Figure 3. It is appropriate to look at these different
interactions with ꢀ-CD in light of the recent paper that
distinguishes between “inclusion complexation” (in the host
cavity) and “encapsulation interaction”, when the guest is
between the wider rims of the cavity of two CD molecules or
on the rim of one host molecule.35 The first is illustrated by the
adduct and DTBN in CF2HCOOH and the second by the adduct
and DTBN in CH3COOH.
Figure 8. Variation of the % relative intensities for MNP/CF2COOH
“in” and for DTBN “in” as a function of ꢀ-CD concentration. The
relative intensities were calculated separately in % total intensity for
the each nitroxide. See Table 1.
Figure 9. The association constant Ka at 300 K of DTBN and of
the MNP/CF2COOH adduct as a function of ꢀ-CD concentration.
See text.
The DTBN Radical in CH3COOH and in CF2HCOOH.
A surprising result of the present study was that for a given
model compound (AA or DFAA), the two nitroxide radicals
(the spin adduct and DTBN) have the same type of interaction
with the host: at the rim in AA, and inside the host cavity in
DFAA. Moreover, in the absence of the host, the aN value for
DTBN is higher in the DFAA system (17.1 G) compared to
that in AA (16.7 G). This result emphasizes the complexity of
a system that includes the radical source, the spin trap, the
adduct, hydroxyl radicals, and DTBN, in addition to the
aF ) 22.4 G), the “out” adduct. Similar arguments apply for
the DTBN radicals: “in” with aN ) 16.6 G, “out” with aN )
17.1 G. The stability of the MNP/CF2COOH adduct increases
significantly upon addition of the host. Figure 7 presents the
intensity of the spin adduct in the absence of ꢀ-CD and in the
presence of host concentration of 10-3 M. The effect of host
addition is similar to that detected for the MNP/CH2COOH

Cyclodextrins to Encapsulate Radical Adducts
J. Phys. Chem. A, Vol. 114, No. 21, 2010 6223
TABLE 1: Hyperfine Splittings and Relative Intensities of the “out” and “in” DTBN Radical and MNP/CF2COOH for the
Indicated ꢀ-CD Concentrations
DTBN “out”
DTBN “in”
MNP/CF2COOH “out”
MNP/CF2COOH “in”
[ꢀ-CD] × 104/M
aN/G
%
aN/G
%
aN/G
aF/G
%
aN/G
aF/G
%
0
1
17.1
17.1
17.1
17.1
17.1
17.1
17.1
17.1
17.1
17.1
17.1
17.1
48
58
56
54
42
38
26
18
11
4
13.4
13.4
13.4
13.4
13.4
13.4
13.4
13.4
13.4
13.4
13.4
13.4
13.4
13.2
22.4
22.4
22.4
22.4
22.4
22.4
22.4
22.4
22.4
22.4
22.4
22.4
22.4
22.5
52
42
44
46
46
40
37
38
33
30
28
27
25
22
1.5
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
10
20
30
16.7
16.7
16.7
16.6
16.6
16.6
16.6
16.6
16.6
16.6
12
22
37
26
29
21
22
22
15
17
12.7
12.7
12.7
12.6
12.6
12.7
12.7
23.5
23.5
23.4
23.4
23.4
23.4
23.5
18
27
45
47
49
60
61
3
2
TABLE 2: Association Constants, Ka, of DTBN and
MNP/CF2COOH for the Indicated ꢀ-CD Concentrations
Ka/M-1
column) are higher than those for the simpler system in the
absence of acetic acid, possibly because of the lower pH value.
All results presented in this section reinforce our assumption
mentioned above for data in the absence of host: the presence
of the acids AA and DFAA, with pKa values of 4.8136 and 1.34,37
modify the local polarity and lead to a lower aN value in the
AA solutions. We conclude that for the system host/DTBN two
effects are at play: the local polarity (lower for AA than for
DFAA) in the absence of host, and the insertion of DTBN in
the host cavity in the absence of the adduct.
An important extrapolation can be made for rationalizing the
results for both the adduct and DTBN shown in Figures 4 and
6: In the presence of the host, the more hydrophobic adduct
(MNP/CF2COOH) is inserted in the host, while the less
hydrophobic adduct (MNP/CH2COOH) is attracted to the polar
rim.
[ꢀ-CD] × 104/M
DTBN
adduct
2
3
4
5
6
7
8
9
10
20
30
952
1447
2846
2407
3766
6562
8148
11000
789
1169
1875
1865
1815
1200
924
cyclodextrin, and encouraged us to explore the behavior of
DTBN in the absence of other competitors but in the presence
of the cyclodextrin host. Ref 4 reported the insertion of DTBN
in ꢀ-CD and the detection of “in” and “out” ESR signals; as
expected, the “in” signal had a smaller value of aN. We are
aware of one study of DTBN in R-CD, which suggested that
the N-O group is outside the host cavity.31 At this stage it
seemed reasonable to expect that the γ-CD cavity is too large
for the DTBN radical. Therefore the additional experiments were
performed with ꢀ-CD as the host.
The results are shown in Figures 11 and 12. Figure 11 presents
the ESR spectra of aqueous DTBN, at pH ) 7.2. In the absence
of ꢀ-CD, Figure 11A shows aN ) 17.2 G; the same aN value
was measured for low ꢀ-CD concentration, 10-4 M. The “in”
and “out” DTBN species with aN values of 16.6 and 17.2 G,
respectively, were measured in the presence of high host
concentration, Figure 11B. These results are in agreement with
the data cited in ref 4. Moreover, the association constants given
in Table 3 (second column) are similar to those presented in
ref 4.
Contrasting results were detected in the presence of acetic
acid, Figure 12. In the absence of the host, Figure 12A indicates
aN ) 16.8 G, as in the presence of the adduct (Figure 4A). For
a low ꢀ-CD concentration, aN becomes 17.1 G, a sign for a
higher local polarity; this value of aN is identical to that
presented in Figure 4B and is similarly taken to describe a rim
location for the nitroxide. In the presence of high host
concentrations, 10-3 M, Figure 12B, two DTBN species with
aN values of 16.6 and 17.2 G, respectively, were measured. We
propose that the aN values represent “in” and “rim” locations
of DTBN. The association constants given in Table 3 (third
Conclusions
We have presented experiments that describe the potential
of ꢀ cyclodextrin (ꢀ-CD) to interact with spin adducts of 5,5-
dimethyl-1-pyrroline N-oxide (DMPO), N-tert-butyl-R-phe-
nylnitrone (PBN), and 2-methyl-2-nitrosopropane (MNP). The
focus was on spin adducts of oxygen-centered radicals trapped
by DMPO and PBN, and on carbon-centered radicals trapped
by MNP. We have chosen to study the DMPO/OH and PBN/
OH adducts, the MNP/CH2COOH adduct in CH3COOH, and
the MNP/CF2COOH adduct in CF2HCOOH. The radical di-
tert-butyl nitroxide ((CH3)3C)2NO (DTBN) was also generated
in experiments with MNP as the spin trap, and its presence
provided the opportunity to explore the effect of ꢀ-CD in the
presence of two nitroxides. The interactions of the nitroxides
(adducts and DTBN) with ꢀ-CD were examined in terms of
their stability and of the variation of the 14N hyperfine splitting
of the nitroxides, aN, which is known to decrease at a location
of lower polarity and increase when the N-O group is
H-bonded.
A range of interactions of the spin adducts and DTBN with
ꢀ-CD was detected. The presence of ꢀ-CD led to significant
stabilization of the DMPO/OH and PBN/OH adducts, but to
negligible variations of the hyperfine splittings and line widths.
The half-life of the DMPO/OH adduct depends linearly on the
host concentration and increases from 700 s in the absence of
the host to 1600 s for [ꢀ-CD] ) 3.5 × 10-3 M. The half-life of
the PBN/OH adduct shows a major increase in a narrow range
of ꢀ-CD concentrations, (1.5-3.0) × 10-2 M. The increased
stability suggests complexation with the host; the negligible

6224 J. Phys. Chem. A, Vol. 114, No. 21, 2010
Spulber and Schlick
Figure 10. Proposed locations of the MNP/CH2COOH adduct (A) and of the MNP/CF2COOH adduct (B) in ꢀ-CD. Note the position of the N-O
group close to the rim of the host for MNP/CH2COOH and inside the host caVity for MNP/CF2COOH. As for Figure 3, the structures of the adducts
were optimized using the Gaussian 03 software, and the adduct was inserted in the ꢀ-CD, based on the picture of complexation deduced from the
ESR spectra (rim or “in”). See text.
Figure 12. ESR spectra of DTBN at 300 K at pH 5 in the presence of
acetic acid (2 M): (A) in the absence of ꢀ-CD (top spectrum), and for
Figure 11. ESR spectra of DTBN at 300 K at pH 7.2: (A) in the
absence of ꢀ-CD (top spectrum), and for [ꢀ-CD] ) 10-4 M; (B) for
[ꢀ-CD] ) 10-3 M. Downward arrows point to 13C satellites in (A),
which were not considered in the simulated spectra.
[ꢀ-CD] ) 10-4 M; (B) for [ꢀ-CD] ) 10-3 M.
polarity is higher. Experiments with methylated ꢀ-CD (Me
ꢀ-CD) provided support for this conclusion.
A different type of complexation was detected for both DTBN
and MNP/CF2COOH in CF2HCOOH: for specific host concen-
trations both “in” and “out” species were detected. It seems
reasonable to suggest that the hydrophobicity of the fluorinated
effect on aN indicates that the N-O group is outside the ꢀ-CD
cavity. This effect is assigned to partial enclosure of the adducts.
For both DTBN and MNP/CH2COOH, the presence of ꢀ-CD
leads to an increase of aN, a result that was interpreted as due
to the location of the guests at the rim of the host, where the

Cyclodextrins to Encapsulate Radical Adducts
J. Phys. Chem. A, Vol. 114, No. 21, 2010 6225
(6) Kotake, Y.; Janzen, E. G. J. Am. Chem. Soc. 1989, 111, 7319–
TABLE 3: Association Constants, Ka, of DTBN in ꢀ-CD in
the Absence (Middle Column) and in the Presence (Last
Column) of Acetic Acid
7323.
(7) Kotake, Y.; Janzen, E. G. J. Am. Chem. Soc. 1992, 114, 2872–
2874.
(8) Franchi, P.; Pedulli, G.-F.; Lucarini, M. J. Phys. Chem. B 2008,
112, 8706–8714.
(9) Ionita, G.; Florent, M.; Goldfarb, D.; Chechik, V. J. Phys. Chem.
B 2009, 113, 5781–5787.
(10) Eastman, M. P.; Brainard, J. R.; Stewart, D.; Anderson, G.; Lloyd,
W. D. Macromolecules 1989, 22, 3888–3892.
(11) Bardelang, D.; Clement, J.-L.; Finet, J.-P.; Karoui, H.; Tordo, P.
J. Phys. Chem. B 2004, 108, 8054–8061.
(12) (a) Karoui, H.; Rockenbauer, A.; Pietri, S.; Tordo, P. Chem.
Commun. 2002, 3030–3031. (b) Bardelang, D.; Rockenbauer, A.; Karoui,
H.; Finet, J.-P.; Tordo, P. J. Phys. Chem. B 2005, 109, 10521–10530.
(13) Cle´ment, J.-L.; Tordo, P. Advances in Spin Trapping. In Electron
Paramagnetic Resonance; Gilbert, B. C., Davies, M. J., Murphy, D. M.,
Eds.; A Specialist Periodical Report; Royal Society of Chemistry: Cam-
bridge, 2007; Vol. 20, pp 29-49 and references therein.
(14) (a) Luo, L. B.; Han, D. Y.; Wu, Y.; Song, X. Y.; Chen, H. L.
J. Chem. Soc., Perkin Trans. 1998, 2, 1709–1714. (b) Song, X.; Chen, Y.;
Chen, H. New J. Chem. 2001, 25, 985–988.
(15) Han, Y.; Tuccio, B.; Lauricella, R.; Villamena, F. A. J. Org. Chem.
2008, 73, 7108–7117.
(16) Kadirov, M. K.; Bosnjakovic, A.; Schlick, S. J. Phys. Chem. B
2005, 109, 7664–7670.
Ka/M-1
[ꢀ-CD] × 104/M
DTBN/ꢀ-CD
DTBN/AA/ꢀ-CD
7
8
561
612
636
619
583
4624
4493
2812
2131
1886
1571
2606
10
20
30
40
50
60
70
adduct drags the adduct inside the host cavity. Calculation of
the association constant Ka indicated the competition between
DTBN and the adduct for inclusion in the host cavity: For
[ꢀ-CD] ≈ 9 × 10-4 M the Ka for DTBN increases dramatically,
while the Ka for the adduct plateaus and starts to decrease. These
results reinforce the idea that the two nitroxide radicals are
competitors for inclusion in the host.
(17) Bosnjakovic, A.; Schlick, S. J. Phys. Chem. B 2006, 111, 10720–
10728.
In a given system containing MNP as spin trap, the two
nitroxide radicals (adduct and DTBN) have the same type of
interaction with the host: at the rim in acetic acid and inside
the host caVity in CF2HCOOH.
Experiments with DTBN in the absence of the spin trap and
of adducts illuminated the effect of the pH and of the local
polarity on the hfs and indicated that the presence of acetic acid
encourages rim complexation.
(18) Lund, A.; Macomber, L.; Danilczuk, M.; Stevens, J.; Schlick, S. J.
Phys. Chem. B 2007, 111, 9484–9491.
(19) Danilczuk, M.; Coms, F. D.; Schlick, S. Fuel Cells 2008, 8 (6),
436–452.
(20) Bednarek, J.; Schlick, S. J. Phys. Chem. 1991, 95, 9940.
(21) Janzen, E. G. Acc. Chem. Res. 1971, 4, 31–40.
(22) Lagercrantz, C. J. Phys. Chem. 1971, 75, 3466–3475.
(23) Spin Trapping Database, http://EPR.niehs.nih.gov.
(24) Danilczuk, M.; Schlick, S.; Coms, F. D. J. Phys. Chem. B 2009,
113, 8031–8042.
(25) Danilczuk, M.; Coms, F. D.; Schlick, S. Macromolecules 2009,
42, 8943–8949.
Acknowledgment. This research was supported by grants
(26) Madden, K. P.; Taniguchi, H. J. Am. Chem. Soc. 1991, 113, 5541.
(27) Gaffney, B. J. In Spin Labeling, Theory and Applications; Berliner,
L. J. Ed.; Academic Press: New York, 1976; Chapter 5, pp 1830-238.
(28) Stella, V. J.; Rao, V. M.; Zannou, E. A.; Zia, V. AdV. Drug DeliVery
ReV. 1999, 36, 3–16.
from the Polymers Program of the National Science Foundation.
Supporting Information Available: Experimental and
simulated ESR spectra of the MNP/CH2COOH adduct and of
DTBN detected at 300 K and pH 5 in the presence of methylated
ꢀ-CD (Me ꢀ-CD) as the host. This material is available free of
charge via the Internet at http://pubs.acs.org.
(29) The software WinSim can be accessed at: http://www.niehs.nih.gov/
research/resources/software/tools/index.cfm.
(30) Kotake, Y.; Janzen, E. G. J. Am. Chem. Soc. 1991, 113, 9503–
9506.
(31) Eastman, M. P.; Freiha, B.; Hsu, C. C.; Lum, K. C.; Chang, C. A.
J. Phys. Chem. 1987, 91, 1953–195.
(32) Vujosevic, D.; Dilger, H.; McKenzie, I.; Martyniak, A.; Scheuer-
mann, R.; Roduner, E. J. Phys. Chem. B 2007, 111, 199–208.
(33) We are grateful to a reviewer for this observation.
(34) Kotake, Y.; Janzen, E. G. J. Am. Chem. Soc. 1989, 111, 5138–
5140.
References and Notes
(1) Szejtli, J. Chem. ReV. 1998, 98, 1743–1753.
(2) Hedges, A. R. Chem. ReV. 1998, 98, 2035–2044.
(3) (a) Spulber, M.; Pinteala, M.; Fifere, A.; Moldoveanu, C.; Man-
galagiu, I.; Harabagiu, V.; Simionescu, B. C. J. Inclusion Phenom.
Macrocyclic Chem. 2008, 62, 135–142. (b) Miron, L.; Mares, M.; Nastasa,
V.; Spulber, M.; Pinteala, M.; Fifere, A.; Harabagiu, V.; Simionescu, B. C.
J. Inclusion Phenom. Macrocyclic Chem 2009, 63, 159–162, and references
therein.
(4) Atherton, N. M.; Strach, S. J. J. Chem. Soc., Faraday Trans. 2
1974, 357–360.
(5) Martinie, J.; Michon, J.; Rassat, A. J. Am. Chem. Soc. 1975, 97,
1818–1823.
(35) Song, L. X.; Bai, L.; Xu, X. M.; He, J.; Pan, S. Z. Coord. Chem.
ReV. 2009, 253, 1276–1284.
(36) Kim, H.-S.; Chung, T. D.; Kim, H. J. Electroanal. Chem. 2001,
498, 209–215.
(37) Ye, Y. K.; Stringham, R. W. J. Chromatogr., Sect. A 2001, 927,
47–52.
JP100777U