Fourier transform EPR study and quantum chemical calculations of dihydrofuran radicals formed by triplet sensitized electron transfer in aqueous solution

Article abstract of DOI:10.1039/b109064j

The radicals formed by electron transfer from 2,3-dihydrofuran, 2-methyl-4,5-dihydrofuran and 2,3-dimethyl-4,5-dihydrofuran to the laser induced triplet state of anthraquinone-2,6-disulfonic acid were studied in aqueous solution by Fourier transform electron paramagnetic resonance (FT EPR) in the nanosecond time-scale. With 2,3-dimethyl-4,5-dihydrofuran and 2-methyl-4,5-dihydrofuran as electron donor the radical cations were observed directly, whereas with 2,3-dihydrofuran as electron donors the radicals observed are successor radicals of the radical cations. The assignment of the radical structures was done by simulation of the experimental EPR spectra and was supported by quantum chemical density functional theory (DFT) calculations. In the experiments with 2-methyl-4,5-dihydrofuran only one OH^--adduct could be determined although other so far unknown radicals contribute to the measured spectra. With 2,3-dihydrofuran two OH^--adduct radicals and the deprotonated radical cation were detected together with contributions from a dimer radical. The rate constants of the electron transfer to the anthraquinone-2,6-disulfonate triplet and of the formation of the successor radicals from the primary 2,3-dihydrofuran radical cation were determined, respectively. The hyperfine coupling constants calculated by quantum chemical DFT method agree quite well with the experimental one.

Full text of DOI:10.1039/b109064j

View Article Online / Journal Homepage / Table of Contents for this issue
Fourier transform EPR study and quantum chemical calculations of
dihydrofuran radicals formed by triplet sensitized electron transfer in
aqueous solution
a
J. Geimer, K. Bernhard, S. Naumov and D. Beckert
b
c
a
a
Universit a¨ t Leipzig, Interdisziplin a¨ re Arbeitsgruppe Zeitaufgel o¨ ste Spektroskopie,
Permoserstrasse 15, D-04318 Leipzig, Germany
Max-Planck-Institut f u¨ r Strahlenchemie, Stiftstrasse 34–36, D-45470 M u¨ lheim a.d. Ruhr,
Germany
b
c
Institut f u¨ r Oberfl a¨ chenmodifizierung, Permoserstrasse 15, D-04318 Leipzig, Germany
Received 5th October 2001, Accepted 22nd January 2002
First published as an Advance Article on the web 9th April 2002
The radicals formed by electron transfer from 2,3-dihydrofuran, 2-methyl-4,5-dihydrofuran and 2,3-dimethyl-
4
,5-dihydrofuran to the laser induced triplet state of anthraquinone-2,6-disulfonic acid were studied in aqueous
solution by Fourier transform electron paramagnetic resonance (FT EPR) in the nanosecond time-scale. With
,3-dimethyl-4,5-dihydrofuran and 2-methyl-4,5-dihydrofuran as electron donor the radical cations were
2
observed directly, whereas with 2,3-dihydrofuran as electron donors the radicals observed are successor radicals
of the radical cations. The assignment of the radical structures was done by simulation of the experimental EPR
spectra and was supported by quantum chemical density functional theory (DFT) calculations. In the
ꢀ
experiments with 2-methyl-4,5-dihydrofuran only one OH -adduct could be determined although other so far
unknown radicals contribute to the measured spectra. With 2,3-dihydrofuran two OH -adduct radicals and the
ꢀ
deprotonated radical cation were detected together with contributions from a dimer radical. The rate constants
of the electron transfer to the anthraquinone-2,6-disulfonate triplet and of the formation of the successor
radicals from the primary 2,3-dihydrofuran radical cation were determined, respectively. The hyperfine
coupling constants calculated by quantum chemical DFT method agree quite well with the experimental one.
acceptor. The dihydrofurans are oxidized and the primary
Introduction
radical cations transform to different successor radicals and
dimer radicals which are interesting in a more practical point
of view. The detection of the 2,3-dihydrofuran radical cation
The oxidation of enolether was studied with different methods
like pulse radiolysis, laser photolysis and electrochemical
methods. Enolether radical cations are of interest because they
belong to the simplest p-type radicals and because they are
intermediate transients in the heterolysis of DNA. More
recently, the spectroscopic properties and the reactivity of eno-
lether radical cations in aqueous solutions were reviewed by
Steenken et al. (see references therein). In the mentioned
papers the redox properties and the reactivities of the enolether
radicals are considered in detail. However, the structure of the
primary radical cations and their successor radicals, respec-
tively, are of interest too. Therefore, in this paper we report
about time-resolved FT EPR investigations of photochemical
generated dihydrofuran radicals and some of its methylated
derivates in aqueous solutions in the nanosecond time
range.
By the method of sensitized electron transfer from a donor
molecule to the laser induced triplet state of an acceptor mole-
cule radical ion pairs and their successor radicals are gener-
ated. Here, the Fourier transform electron paramagnetic
resonance (FT EPR) is the appropiate method to study the
radical structure and radical transformation a few nanose-
conds after radical formation. We have already shown how
this method was successfully used in studying radicals of var-
ious pyrimidine-like DNA-bases.
8
was successfully done by Knolle et al. in low temperature
Freon-matrices by cw-EPR. In liquid solution at room tem-
perature the detection of the dihydrofuran radical cations by
EPR was not possible till now. Gilbert et al. studied the suc-
9
cessor radicals of the 2,3-dihydrofuran radical cation under
various experimental conditions by steady state EPR experi-
ments. These results were helpful in radical assignment and
will be compared to our time resolved EPR results.
1
,2
The mechanism of radical generation is shown in Scheme 1.
The extinction coefficient of the dihydrofurans is very small
compared to 2,6-AQDS at 308 nm, therefore the light absorp-
tion is due to 2,6-AQDS alone. By the efficient intersystem
crossing (ISC) spin-polarized triplet states AQDST1 are
formed. This triplet state is the starting point of various reac-
tions as shown in the scheme. In the presence of an adequate
electron donor (DH) like DHF, the dominant reaction path-
way is the electron transfer from the donor to the triplet state
of the 2,6-AQDS. This triplet sensitized electron transfer pre-
serves the triplet polarization in the formed radicals if the radi-
cal formation process is faster than the spin lattice relaxation
time T of the triplet precursor. The EPR-spectra of the radi-
3
3
1
cals are therefore emissively polarized with enhanced intensi-
ties due to chemically induced dynamic electron polarization
4
–7
In this paper we want to
1
0,11
describe experiments with 2,3-dihydrofuran (DHF), 2-
methyl-4,5-dihydrofuran (2-MeDHF) and 2,3-dimethyl-4,5-
dihydrofuran (2,3-Me DHF) (see Scheme 1) as electron donors
(CIDEP) effects.
The intensity of the triplet polarized
EPR-spectra is proportional to the rate constant ket of the
radical formation process. Because of the fast electron transfer
process between the different dihydrofurans and the triplet
2
and anthraquinone-2,6-disulfonic acid (2,6-AQDS) as electron
1
738
Phys. Chem. Chem. Phys., 2002, 4, 1738–1745
This journal is # The Owner Societies 2002
DOI: 10.1039/b109064j

View Article Online
Scheme 1
state of 2,6-AQDS, and therefore the high intensity of the
Theoretical method
emissively polarized EPR-spectra, we are able to measure
EPR spectra of the formed radicals about 10 ns after their for-
mation. Apart from the triplet mechanism the EPR intensities
are modified by a second polarization mechanism, the radical
pair polarization which leads to emissively polarized low field
lines and absorptively polarized high field lines. The spectra
are therefore in most cases asymmetric in the intensity of high
and low field lines.
Quantum chemical calculations were done using DFT Hybrid
method with 6-31G(d) and 6-311+G(d,p) stan-
2
0–22
B3LYP
2
3
dard basis sets (Gaussian 98 ). B3LYP/6-31G(d) method
has shown very reasonable results in the calculation of trans-
formations and hyperfine structure (hfs) coupling constants
8
of DHF and 2,5-DHF radicals in freon matrices. The equili-
brium geometries of the structures investigated were optimized
without any restriction. The energy of the systems, the electro-
nic structures, the atomic spin densities and the Mulliken
atomic charge distributions for the neutral radicals and the
radical cations were calculated both for gas phase and aqueous
solution. The calculation of the hfs is helpful for the identifica-
Experimental
2
4,25
tion of the transients.
The isotropic hyperfine coupling
The 308 nm XeCl line of an excimer laser (Lambda Physik,
LPX 105 ESC) with 10–15 mJ per pulse with a pulse length
of about 15 ns was used. The jitter of the laser pulse relative
to the trigger pulse is in the order of 1 ns and therefore, the
FT EPR spectra can be measured after a few nanoseconds.
constants were calculated as part of the population analysis.
To test the influence of the polarized continuum on molecular
structure the geometry optimisations were also made in water
using two self-consistent reaction field (SCRF) models: Onsa-
2
6
27,28
1
2
ger (SCRF ¼ Dipole) and COSMO (SCRF ¼ CPCM)
The FT EPR equipment has been described previously. In
the experiments the power of the microwave amplifier was 20
W with a length of the p/2 pulses of 48 ns or 1 kW with a
p/2 pulse of 16 ns. The excitation width was about ꢁ14
MHz or ꢁ40 MHz, which is not enough to detect the EPR-
spectra of the dihydrofuran radicals in one registration run
at fixed field position. The spectra where therefore measured
in a step-wise fashion by variation of the magnetic field offset,
and up to 30 spectra were added together to show the whole
which use a more reliable cavity as union of a series of inter-
locking atomic spheres.
Results
As indicated in Scheme 1 the electron transfer from dihydro-
furans to the 2,6-AQDS triplet is the primary step of radical
generation. Unambiguous evidence of this initiation reaction
is the presence of both radical ions, the radical anion 2,6-
2
EPR-spectrum of the DHF, 2-MeDHF and 2,3-Me DHF radi-
cals. Because of the deadtime of about 100 ns all the spectra
were extrapolated with the linear prediction singular value
ꢀ
ꢃ
+
ꢃ
AQDS and the radical cation DH , respectively. The FT
1
3
ꢀ
ꢃ
decomposition method (LPSVD) to correct the base line.
The chemicals anthraquinone-2,6-disulfonic acid (2,6-
AQDS) and DHF where purchased from Aldrich and were
used without further purification. 2-Methyl-4,5-dihydrofuran
and 2,3-dimethyl-4,5-dihydrofuran were synthesized as
described in refs. 14–19. A solution of about 20 mM KOH
EPR spectrum of the radical anion 2,6-AQDS is always in
the center of the spectra with a strong emissive polarization.
This strong emissive signal is not presented in all spectra
shown in this paper since it has been discussed previously.
In the FT EPR experiments only the measurements with 2-
MeDHF and 2,3-Me DHF as electron donor were successful
2
9
2
(
from Merck) in water from a milli-Q-plus system (Millipore)
with respect to detecting the radical cation.
In Fig. 1 the FT EPR spectrum of 1 mM 2,6-AQDS with
2,3-Me DHF in aqueous solution at pH 12 and a delay time
2
was used in the pH range between pH 11 to 12. The measure-
ments at pH 9.2 were performed in a buffer solution of 20 mM
disodium hydrogen phosphate dihydrate from Merck.
To remove oxygen the samples were bubbled with argon
(
The sample flows through the EPR tube (optical path length
of about 1.5 mm) at a rate of about 1–2 ml min to avoid
enrichment of reaction products. The probe reservoir and glass
tube system before the resonator was a double wall system
which allows the variation of the temperature in the range
between 5 C and 70 C. For all the measurements shown
in this work we chose a constant temperature of about
of 40 ns is shown. The emissive low field lines and the absorp-
tive high field lines indicate the E*/A character of the 2,3-
+
ꢃ
99.99%) 20 min before and during the whole experiments.
Me
2
DHF spectrum which is a superposition of triplet and
radical pair polarization. The spectrum of the radical anion
ꢀ1
ꢀ
ꢃ
2,6-AQDS (not shown in the figure) is emissively polarized.
+
ꢃ
The experimental 2,3-Me DHF spectrum is compared with
2
the simulated one which was calculated with the hfs coupling
constants listed in Table 1 [radical (1)]. Our simulation pro-
gram considers also the second order splittings of equivalent
nuclei (two protons in positions 4, 5 and methyl groups in posi-
tion 2, 3) which influence the line positions and line numbers in
ꢂ
ꢂ
ꢂ
1
0 C.
Phys. Chem. Chem. Phys., 2002, 4, 1738–1745
1739

View Article Online
noise ratio of these spectra and the loss of lines by spin-polar-
ization effects and line superpositions, respectively, demands a
careful and critical analysis to determine the structure of the
radicals contributing to Fig. 2. The first point is to look at dif-
ferences between the spectra at 40 ns and 2 ms which shows
that in the low field range some small lines diminish and some
increase with time. This indicates a contribution of the radical
+
ꢃ
cation 2-MeDHF to the 40 ns spectrum. The increasing lines
can be assigned to successor radicals of the radical cation.
Furthermore, we include in the spectra simulation the results
of quantum chemical DFT calculations which were so success-
+
ꢃ
ful in the case of 2,3-Me
dure three radicals were identified from the spectra in Fig. 2,
2
DHF . In a time consuming proce-
+
ꢃ
the radical cation 2-MeDHF (2), the deprotonated radical
cation (3) and the C3-OH-addition radical (4). The hfs cou-
pling constants of (2), (3) and (4) obtained by spectra simula-
tion are summarized in Table 1, where the agreement with the
calculated hfs values is notable. The radical structure gener-
ated by deprotonation at C5 could be excluded with the calcu-
lated hfs parameter of this structure because an agreement
between the two parameter sets is also qualitatively impossible.
To demonstrate the quality of the simulation procedure the
lowest line group of radical (4) is presented in Fig. 3. Because
of the low signal to noise ratio no kinetic data were extracted
from the spectra.
ꢀ
ꢃ
Fig. 1 FT EPR spectrum of the radical cation 2,3-Me
after 40 ns with 1.0 mM 2,6-AQDS and 20 mM 2,3-Me DHF in aqu-
2
DHF (1)
2
eous solution, pH 12. The simulated spectrum with hfs coupling con-
stants in Table 1 is represented for comparison.
+
the 2,3-Me DHF spectrum considerably. The g value was
determined by comparison with the g value of the radical
anion 2,6-AQDS (g ¼ 2.00412) in the same spectrum. The
hfs coupling constants calculated by DFT methods using the
Onsager and COSMO models to consider the solvent influence
with e ¼ 78 are included in Table 1. The good agreement with
ꢃ
2
ꢀ
ꢃ
In Fig. 4 the FT EPR spectrum of the radicals from DHF at
the delay time of 40 ns is shown. This spectrum is the superpo-
sition of at least three different radicals, the deprotonated radi-
cal cation (5), the C5-OH-addition radical (6) and the C4-OH-
addition radical (7). In the simulated spectrum in Fig. 4 the
intensity ratio of the radicals (5), (6) and (7) of 3 : 1 : 1 was
the experimental results is notable.
ꢃ
The linewidth in the 2,3-Me
2
DHF+ spectrum was deter-
mined to Do1/2 ¼ (1.4 ꢁ 0.2) MHz which corresponds to
2eff ¼ 220 ns. Assuming a lifetime controlled linewidth we
estimate for the lifetime t of the radical cation 2,3-
Me DHF a limiting value trc > 220 ns. Although this is near
the limit of the FT EPR lifetime resolution of transient radical
it would be very interesting to study the kinetics of conversion
T
+
ꢃ
rc
used. The radical cation DHF is not detectable because the
lifetime is too short (see below). The structures (5), (6) and
+
ꢃ
2
(
7) and their hfs coupling constants obtained in the simulation
procedure are shown in Table 2. The radical identification was
done by comparison of the experimental hfs coupling con-
stants with the theoretical values calculated by DFT methods
+
ꢃ
of the radical cation 2,3-Me DHF to successor radicals.
2
Unfortunately, there was not enough 2,3-Me
available for these experiments.
In Fig. 2 the FT EPR spectra of the radicals from 2-MeDHF
2
DHF substance
9
and the data given by Gilbert et al. This assignment is sup-
ported by measurements at different delay times and pH values
in buffered solutions. In Fig. 5 the FT EPR spectrum of the
at delay times of 40 ns and 2 ms are shown. The low signal to
Table 1 Structure and experimental and B3LYP/6-31G(d) calculated in water with SCRF ¼ Dipole [*SCRF ¼ CPCM(COSMO) in parenthesis
for comparison] EPR parameters of methylated dihydrofuran radicals. The calculated DFT values of radical (4) are taken for the most stable struc-
ture from the two possible conformations (cf. Table 3). ** - calculated averaged value by OH-group rotation is 0.16 mT
EPR parameter/mT
Radical structure
Atomic spin density
This work
DFT calc.
(1)
(2)
(3)
(4)
S(O) ¼ 0.218
a(2H,4) ¼ 3.42
a(3H,3) ¼ 1.84
a(2H,5) ¼ 0.82
a(3H,2) ¼ 0.57
g ¼ 2.0045
3.46
1.89
1.02
0.57
S(C2) ¼ 0.135
S(C3) ¼ 0.561
S(C4) ¼ ꢀ0.044
S(C5) ¼ ꢀ0.006
S(O) ¼ 0.223
a(2H,4) ¼ 4.10
a(2H,5) ¼ 1.21
a(1H,3) ¼ 1.60
a(3H,2) ¼ 0.52
4.27
1.22
S(C2) ¼ 0.087
S(C3) ¼ ꢀ0.637
S(C4) ¼ ꢀ0.042
S(C5) ¼ ꢀ0.005
S(O) ¼ 0.082
ꢀ1.62
0.52
a(2H,5) ¼ 3.93
a(1H,4) ¼ 1.21
a(1H,3) ¼ 0.21
a(3H,2) ¼ 1.21
3.64
1.57
S(C2) ¼ 0.486
S(C3) ¼ ꢀ0.205
S(C4) ¼ 0.620
S(C5) ¼ ꢀ0.063
S(O) ¼ 0.135
0.30
ꢀ1.34
a(3H,2) ¼ 2.33
a(1H,3) ¼ 1.31
a(1H,4) ¼ 0.13
a(1H,4) ¼ 0.13
a(1H,5) ¼ 0.31
a(OH,3) ¼ 0.06
1.90(1.90)*
0.84(1.81)
0.06(0.06)
0.20(0.19)
0.54(0.55)
0.63(0.59)**
S(C2) ¼ 0.816
S(C3) ¼ ꢀ0.023
S(C4) ¼ 0.003
S(C5) ¼ ꢀ0.009
1
740
Phys. Chem. Chem. Phys., 2002, 4, 1738–1745

View Article Online
Fig. 4 FT EPR spectrum of the DHF successor radicals from the
+
ꢃ
radical cation DHF . Sample: 1 mM 2,6-AQDS with 30 mM DHF
in aqueous soulution, pH 12. Delay time: 40 ns. Parameter of the simu-
lated spectrum and designation of the radicals see text and Table 2.
Fig. 2 Low field part of the FT EPR spectra with 1.0 mM 2,6-AQDS
and 40 mM 2-MeDHF as donor in aqueous solution, pH 12 at delay
times 40 ns and 2 ms.
concentrations. The increase of the emissive triplet polariza-
tion within 50 ns is due escape of the radical anion from the
primary radical ion pair generated by the electron transfer
and it can be described by first order kinetics with
deprotonated radical cation (5) is shown at pH 6.0 (delay time:
96 ns). In neutral solution the deprotonation reaction of the
radical cation DHF dominates over the nucleophilic addi-
tion of OH or H O at positions 4 and 5. On the other hand,
+
ꢃ
7
ꢀ1
ꢀ
k
esc ¼ (7.5 ꢁ 0.5) ꢄ 10 s . The second characteristic feature
2
3ꢀ
of the polarization time profiles is the decrease of the polariza-
tion with increasing quinone concentration and a delayed rise
of the polarization in the time range of about 1 ms. A similar
the nucleophilic addition of the phosphate anion PO4 at
position 5 dominates at pH 9.2 in a buffered solution with 20
mM disodium hydrogen phosphate (Fig. 6). The hfs para-
meters of the phosphate adduct radical (8) are given in Table
ꢀ
ꢃ
behavior was observed for 2,6-AQDS with triethylamine as
electron donor and was interpreted by the reaction of the elec-
tron donor with free quinone molecules and quinone com-
plexes, respectively. The fast increase of the polarization is
due to reaction of the radical ion pairs with free quinone
and the slow one by the escape from the self-exchange interac-
tion in quinone complexes. The influence of the quinone com-
plexes with DHF as electron donor is only small and can be
neglected in the quantitative analysis to determine the electron
transfer rate constant. The triplet polarization factor was
determined as a function of the DHF concentration and the
results are presented in Fig. 10. By comparison with the results
obtained with triethylamine as donor the electron transfer rate
constant from DHF to the 2,6-AQDS triplet was estimated to
3
1
2
. The additional P-hfs coupling of the phosphate group at
position 5 gives the characteristic spectral pattern which is
compared in Fig. 7 with the low field line group of the C5-
OH-addition radical (6). Whereas in the buffered solution the
phosphate adduct radical dominates, in the pure aqueous solu-
tion the OH-adduct radicals are detected.
In Fig. 4 with (9) an additional radical is indicated which
increases with higher DHF concentration (Fig. 8). This dublet
of dublets spectrum is assigned to the DHF dimer radical in
9
accordance with Gilbert et al. (9). The OH coupling cannot
be resolved because of the large linewidth.
The kinetic studies of the photoreduction of 2,6-AQDS by
DHF consist of two parts, the study of the primary electron
transfer and the experiments to the build up-kinetics of the
radical cation successor radicals.
9
ꢀ1 ꢀ1
kel ¼ (1.6 ꢁ 0.2) ꢄ 10
M
9
s . This value is considerably
ꢀ1 ꢀ1
s ) measured by optical meth-
lower than that (5 ꢄ 10 M
1
ods.
The kinetics of the generation of the C5-OH-adduct radical
and the deprotonated radical cation were measured with the
To study the kinetics of the primary electron transfer the tri-
ꢀ
ꢃ
plet polarization of the radical anion 2,6-AQDS was mea-
sured as a function of the quinone and of the donor
concentration. In Fig. 9 the polarization time profiles of the
lowest field line groups near by B
0
¼ 340 mT (cf. Fig. 4)
ꢀ
ꢃ
whereas the C4-OH-adduct radical was detected with a line
close to the radical anion. With a 16 ns p/2 pulse we are able
to detect these line groups with the shortest delay time of 8 ns.
The FT EPR amplitudes of the C5-OH-adduct radical and the
deprotonated radical cation are presented in Fig. 11 in the time
range up to 100 ns. The amplitudes are not normalized to the
whole spectra intensities because the line polarization by the
radical pair mechanism causes different intensities which can-
not be estimated with a good reliability. The results for the
C4-OH-adduct radical are not included in Fig. 11 to avoid
confusion in the representation. The rise of the EPR amplitude
of all three successor radicals can be described by first order
radical anion 2,6-AQDS are shown for different quinone
7
ꢀ1
kinetics with the rate constant k
clearly different from the escape rate constant of the radical
r
¼ (5 ꢁ 1) ꢄ 10 s . This is
ꢀ
ꢃ
anion 2,6-AQDS from the primary radical ion pair. From
these two rate constants the decay rate constant of the radical
+
ꢃ
8
cation DHF can be estimated to be kdecay ¼ (1.5 ꢁ 0.5) ꢄ 10
ꢀ1
s
value given in ref. 1 for pH 5. This indicates that the reactivity
at pH 12. This result is higher by a factor of 10 than the
Fig. 3 Comparison of experimental and simulated spectrum of the
low field line group of the C3-OH addition radical of 2-MeDHF (delay
time: 2 ms).
+
ꢃ
ꢀ
of DHF
1
with the strong nucleophile OH is very high
ꢀ1 ꢀ1
s ).
0
(k ꢅ 10
M
Phys. Chem. Chem. Phys., 2002, 4, 1738–1745
1741

View Article Online
Table 2 Structure and EPR parameters of the 2,3-DHF radicals. Results from this work and comparison with data from literature and calculated
at B3LYP/6-31G(d) and SCRF ¼ Dipole (*SCRF ¼ CPCM for comparison) level. Hfs values of the radicals (6) and (7) are taken for the most
stable structure and for radical (8) is taken the averaged values of the two possible conformations (cf. Table 3)
EPR parameter/mT
Radical structure
Atomic spin density
This work
DFT calc.
Literature
S(O) ¼ 0.270
8
S(C2) ¼ ꢀ0.007
S(C3) ¼ ꢀ0.036
S(C4) ¼ 0.575
S(C5) ¼ 0.110
a(2H,3) ¼
a(2H,2) ¼
a(1H,4) ¼
a(1H,5) ¼
4.18
1.51
4.27
1.38
1.44
0.53
ꢀ1.49
ꢀ0.42
(
5)
6)
S(O) ¼ 0.100
9
S(C2) ¼ ꢀ0.064
S(C3) ¼ 0.610
S(C4) ¼ ꢀ0.220
S(C5) ¼ 0.530
a(2H,2) ¼ 3.54
a(1H,3) ¼ 1.33
a(1H,5) ¼ 1.33
a(1H,4) ¼ 0.24
3.77
ꢀ1.55
ꢀ1.49
0.34
3.50
1.34
1.34
0.25
9
(
S(O) ¼ 0.135
a(1H,3) ¼ 4.23
a(1H,3) ¼ 3.15
a(1H,4) ¼ 2.19
a(1H,5) ¼ 1.70
a(1H,2) ¼ 0.06
a(1H,2) ¼ 0.036
4.97(4.56)*
2.50(2.84)
4.10
S(C2) ¼ ꢀ0.007
S(C3) ¼ ꢀ0.036
S(C4) ¼ 0.575
S(C5) ¼ 0.110
2.15
2.15
ꢀ2.29(ꢀ2.42)
1.80(1.11)
ꢀ0.09(ꢀ0.08)
ꢀ0.04(ꢀ0.01)
9
(7)
S(O) ¼ 0.270
a(1H,5) ¼ 1.69
a(1H,4) ¼ 1.05
a(1H,2) ¼ 0.33
a(1H,2) ¼ 0.11
a(1H,3) ¼ 0.11
a(1H,3) ¼ 0.12
g ¼ 2.0038
ꢀ1.49(ꢀ1.51)*
0.90(0.90)
0.61(0.61)
0.06(0.06)
0.19(0.19)
0.07(0.07)
1.66
S(C2) ¼ ꢀ0.007
S(C3) ¼ ꢀ0.036
S(C4) ¼ 0.575
S(C5) ¼ 0.110
1.09
0.34
0.12
0.12
(8)
a(1H,3) ¼ 4.23
a(1H,3) ¼ 3.15
a(1H,4) ¼ 2.22
a(1H,5) ¼ 1.55
a(1P,5) ¼ 0.12
a(1H,2) ¼ 0.06
a(1H,2) ¼ 0.03
g ¼ 2.0034
4.51(4.74)*
2.39(2.43)
S(O) ¼ 0.270
S(C2) ¼ ꢀ0.007
S(C3) ¼ ꢀ0.036
S(C4) ¼ 0.575
S(C5) ¼ 0.110
ꢀ2.08(ꢀ2.26)
1.80(1.99)
ꢀ0.18(ꢀ0.20)
ꢀ0.05(ꢀ0.06)
ꢀ0.03(ꢀ0.04)
8
(9)
a(1H,4) ¼ 3.30
a(1H,5) ¼ 1.40
3.07
1.79
Discussion
The FT EPR detection of the radical cation of the methyl-sub-
stituted dihydrofurans and their successor radicals yields direct
evidence for their structure and decay kinetics in the nanose-
cond time-scale. The stability of the dihydrofuran radical
9
cations increase with methylation at positions C2 and C3.
Although the radical cation of DHF was not detected directly
+
ꢃ
the lifetime of DHF can be estimated to be in the order of 10
ns at pH 12, compared to 83 ns in dilute aqueous solution at
1
+
ꢃ
pH ꢅ 5. Therefore, the FID of DHF decays in the deadtime
of our FT EPR system.
The decay of the radical cations is due to two main chan-
nels—deprotonation and nucleophilic addition of H O/OH
2
and/or PO in aqueous buffered solutions. Our results from
nanosecond FT EPR experiments agree with the results
ꢀ
3ꢀ
4
obtained by steady-state EPR measurements of Gilbert et
9
ꢃ
al. They found the C4-OH-adduct by addition of OH as
the main radical at pH 8.0 and small contributions from the
deprotonated radical cation. Lowering of pH leads to an
Fig. 5 FT EPR spectrum of the deprotonated radical cation (5) after
96 ns with 1.0 mM 2,6-AQDS and 100 mM DHF at pH 6.0.
1
742
Phys. Chem. Chem. Phys., 2002, 4, 1738–1745

View Article Online
ꢀ
ꢃ
Fig. 9 Polarization time profiles of the radical anion 2,6-AQDS in
depencence on the 2,6-AQDS concentration (0.1ꢆ ꢆ ꢆ1.0 mM) with 30
mM DHF at pH 12.
Fig. 6 FT EPR spectrum of the nucleophilic adduct radical with
phosphate anion (8) after 2.0 ms with 1.0 mM 2,6-AQDS and 100
mM DHF in aqueous solution, pH 9.2 (20 mM phosphate buffer).
Fig. 10 Reciprocal triplet polarization of the radical anion 2,6-
AQDS vs. reciprocal donor concentration (DHF and triethylamine)
to determine the electron transfer rate constant.
ꢀ
ꢃ
increase of the intensity of the deprotonated product together
with an appearance of contributions from the C5-OH-adduct.
By radical generation by the triplet-sensitized electron transfer
and FT EPR detection in the nanosecond time-scale we found
a favoured formation of the C5-OH-adduct radicals. This dif-
ferent weighting of the reaction channels can be caused by dif-
ferent radical generation pathways. Unfortunately, the signal-
to-noise ratio in the nanosecond time-scale experiments is too
low to study this interesting problem.
Fig. 7 Comparison of the low field line group of the nucleophilic
+
ꢃ
adduct radicals of the radical cation DHF . Sample as in Fig. 6 with
0 mM phosphate buffer at pH 9.2 after 40 ns (a) and 2 ms (b), in non-
buffered solution after 40 ns (c) and at pH 12 (with KOH) after 40 ns
d).
2
(
9
Gilbert et al. also detected additionally line groups from a
dimer radical formed after reaction of the radical cation with
DHF parent molecules at pH 3.0. After oxidation with Cl
ꢀ
ꢃ
2
Fig. 8 FT EPR spectra of the DHF successor radicals in dependence
on the DHF concentration (delay time: 40 ns and pH 12). Additional
broad lines appear at high DHF concentration (see text).
Fig. 11 Generation kinetics of the successor radicals of the radical
+
ꢃ
cation DHF . Sample: 1 mM 2,6-AQDS with 30 mM DHF in aqu-
eous solution, pH 12. The lines were calculated with first order kinetics.
Phys. Chem. Chem. Phys., 2002, 4, 1738–1745
1743

View Article Online
at pH < 1.0 they found the dimer radical but contributions
from the C5-OH-radical too. In low temperature radiolysis
of different dihydrofurans in Freon matrices, Knolle et al.
has shown only a relatively small change of the geometrical
parameters (for the C3-OH-2-MeDHF radical (4) a maximal
8
˚
ꢂ
change of 0.003 A in bond length and 1.0 in dihedral angle).
This relatively small change in molecular geometry can be
understood in terms of the inherent limitation of the Onsager
approach for the molecules with small dipole moment (all
molecules studied have small dipole moments up to 3.6 debye).
Therefore, we expect only small solvent effects on the spin and
charge distributions in comparison to vacuum. In contrast, for
radical (8) (dipole moments up to 11.6 debye) the changes of
the geometrical parameters are essential larger (a maximal
found the dimer radical cation (9) with two hfs coupling con-
stants (see Table 2) similar to the values estimated from our
time resolved experiments. Because the line groups of the
dimer radical in our time resolved measurements disappear
in a time shorter than 2 ms we attribute these lines to the dimer
radical cation (9). This assignment is supported by the strong
increase of the dimer radical intensity with the dihydrofuran
concentration (Fig. 8).
˚
ꢂ
The radical identification was essentially supported by com-
parison of our experimental hfs coupling constants with the
theoretical values calculated by quantum chemical DFT meth-
ods. The agreement between the experimental and theoretical
data is surprisingly good (Tables 1 and 2). In previous studies
change of 0.027 A in bond length and 1.6 in dihedral angle).
To prove this limitation of Onsager model the geometries, spin
density distributions and hfs coupling constants calculated in
water were compared with solvent effects obtained from
2
7,28
COSMO Polarized Continuum Model (SCRF ¼ CPCM).
4
,5
of radicals of pyrimidine-type bases we have shown that the
geometry and spin density distribution of the radical cations
and their deprotonated successor radicals are strongly influ-
enced by the solvent. Solvent models on polarizable continuum
dielectrics provide reliable solvent contributions. The solvent
influence on geometrical parameters is especially important
This method uses more realistic shape of the cavity than the
Onsager model (SCRF ¼ Dipole). Both SCRF methods result
in similar solvent effects on the spin density distribution for
the radicals with small dipole moment but differ significantly
for radical (8). Therefore, in Tables 1 and 2 the results of
calculations with the Onsager and COSMO model are pre-
sented.
ꢀ
2ꢀ
for OH and OPO3 adduct radicals, where the ring is not
planar and geometry distortions are significant. The compari-
son between vacuum and water for the radicals (4), (6) and (7)
ꢀ
32ꢀ
Quantum chemical calculations on OH and OPO
adduct radicals are complicated by the fact that these species
Table 3 Comparison of B3LYP /6–31G(d) and two SCRF methods (Onsager SCRF ¼ Dipole and SCRF ¼ CPCM (COSMO)) calculated and
0
2ꢀ
- adduct radicals,
experimental values for proton hyperfine splittings a(H) [mT]. f [ ] - dihedral angles between a and b protons for OH- and PO
DE(kcal mol ) - relative energy between two structures
4
ꢀ
1
Dipole
CPCM
I
Structure
SCRF model
I
II
Average
II
Average
Exp
(
4)
a(3H,2)
a(1H,3)
a(1H,4)
a(1H,4)
a(1H,5)
a(1H,5)
a(OH,3)
f
1.90
0.84
0.06
0.20
0.54
0.01
1.92
1.91
1.90
ꢀ0.01
0.07
0.51
0.01
0.40
1.9
1.88
1.89
1.89
ꢀ0.01
0.12
0.52
0.01
0.36
2.33
1.31
0.11
0.13
0.31
–
2.97
0.81
0.06
0.19
0.55
0.02
0.59
2.98
ꢀ0.07
0.16
0.47
ꢀ0.08
0.04
0.50
ꢀ0.03
ꢀ0.03
0.63(*)
ꢂ
0.16
ꢂ
0.13
ꢂ
0.06
ꢂ
64.0
0.0
58.9
+1.2
65.4
0.0
54.4
+1.0
ꢀ
1
1
1
DE/kcal mol
(6)
(7)
(8)
a(1H,3)
a(1H,3)
a(1H,4)
a(1H,5)
a(1H,2)
a(1H,2)
f
4.97
2.50
4.71
2.38
4.89
2.44
4.56
2.84
4.94
2.51
4.75
2.67
ꢀ2.4
1.86
0.06
0.01
4.23
3.15
2.19
1.70
0.06
0.036
ꢀ2.29
1.80
ꢀ2.40
1.60
0.07
ꢀ2.35
1.70
0.08
ꢀ2.42
1.11
0.08
ꢀ2.38
2.67
0.04
ꢀ0.09
ꢀ0.04
0.02
ꢂ
ꢀ0.01
0.01
ꢂ
0.01
ꢂ
ꢂ
55.2
60.8
+0.8
50.2
0.0
71.6
ꢀ0.07
ꢀ
DE/kcal mol
0.0
a(1H,5)
a(1H,4)
a(1H,2)
a(1H,2)
a(1H,3)
a(1H,3)
f
ꢀ1.49
0.90
0.61
0.06
0.19
0.07
65.6
0.0
ꢀ1.34
3.00
0.56
ꢀ1.36
1.94
0.59
0.04
0.14
0.05
ꢀ1.51
0.90
0.61
0.06
0.19
0.07
ꢀ1.37
3.09
0.55
ꢀ1.44
1.99
0.58
0.01
0.06
0.06
1.69
1.05
0.33
0.11
0.12
0.11
ꢀ0.03
ꢀ0.08
0.04
ꢀ0.04
ꢀ0.08
0.04
ꢂ
ꢂ
ꢂ
55.5
+1.5
65.3
0.0
54.3
+1.0
ꢀ
DE/kcal mol
a(1H,3)
a(1H,3)
a(1H,4)
a(1H,5)
a(1P,5)
a(1H,2)
a(1H,2)
f
4.48
1.77
4.54
2.55
4.51
2.39
4.78
2.17
4.70
2.68
4.74
2.43
4.23
3.23
2.22
1.55
0.12
0.06
0.03
ꢀ2.04
1.43
ꢀ2.12
2.17
0.19
ꢀ2.08
1.80
0.18
ꢀ2.30
1.73
ꢀ2.22
2.26
ꢀ2.26
1.99
ꢀ0.18
ꢀ0.05
ꢀ0.06
ꢀ0.23
ꢀ0.07
ꢀ0.04
ꢀ0.18
ꢀ0.05
0.00
ꢀ0.20
ꢀ0.06
ꢀ0.04
ꢀ0.04
ꢀ0.05
ꢀ0.02
0.03
ꢂ
ꢂ
ꢂ
ꢂ
76.8
80.2
ꢀ0.1
64,9
78.4
+6.0
ꢀ1
DE/kcal mol
0.0
0.0
1
744
Phys. Chem. Chem. Phys., 2002, 4, 1738–1745

View Article Online
resolved FT EPR spectra was backed up by quantum chemical
DFT calculations, the experimental and calculated data agree-
ing surprisingly well.
Acknowledgement
Prof. S. Steenken (Max-Planck-Institut f u¨ r Strahlenchemie,
M u¨ lheim) is acknowledged for stimulating discussions. Finan-
cial support by the Deutsche Forschungsgemeinschaft, by the
Fonds der Deutschen Chemischen Industrie and the Volkswa-
gen Foundation (project I/72040) is gratefully acknowledged.
References
1
2
3
K. Bernhard, J. Geimer, M. Canle-Lopez, J. Reynisson, D. Beck-
ert, R. Gleiter and S. Steenken, Chem. Eur. J., 2001, 7, 4640.
M. Bietti and S. Steenken, in Handbook of Electron Transfer, ed.
V. Balzani, Wiley-VCH, Weinheim, 2001, pp. 494–579.
L. Kevan and M. K. Bowman, Modern Pulsed and Continuous-
Wave Electron Spin Resonance, Wiley-Interscience, New York,
Fig. 12 Theoretical data of the hfs coupling constant a(OH,3) of the
C3-OH-2-MeDHF radical (4) and change of the energy of the radical
in dependence of the torsion angle of the OH group (details see text).
1
990.
4
5
J. Geimer, K. Hildenbrand, S. Naumov and D. Beckert, Phys.
Chem. Chem. Phys., 2000, 2, 4199.
are nonplanar. Recently, it has been shown that 5,6-dihydro-6-
thymyl and 5,6-dihydro-5-thymyl and N1-substituted cyto-
3
0
S. Naumov, A. Barthel, J. Reinhold, F. Dietz, J. Geimer and D.
Beckert, Phys. Chem .Chem. Phys., 2000, 2, 4207.
J. Geimer and D. Beckert, J. Phys .Chem. A, 1999, 103, 3991.
J. Geimer, O. Brede and D. Beckert, Chem. Phys. Lett., 1997, 276,
411.
W. Knolle, I. Janowski, S. Naumov and R. Mehnert, J. Chem.
Soc., Perkin Trans. 2, 1999, 2447.
B. C. Gilbert, R. O. C. Norman and P. S. J. Williams, J. Chem.
Soc., Perkin Trans. 2, 1980, 647.
K. A. McLauchlan, Appl. Magn. Res., 1996, 11, 357.
2
5
sines exist in two conformations. The study of geometrical
structures of radicals (4), (6), (7) and (8) yields two stable con-
formers I and II too (cf. Table 3, independently of the basis set
and SCRF model used) with widely different b-splitting. If the
vibration averaging by the out-of-plane motion is taken into
account, for radical (8) a better agreement with the experiment
can be obtained (see Table 3). The results calculated with the
COSMO model are closer to the experimental values than
those by the Onsager model.
6
7
8
9
0
1
11 L. T. Muus, P. W. Atkins, K. A. McLauchlan and J. B. Pedersen,
Chemically Induced Magnetic Polarisation, Reidel Publ., Dor-
drecht, 1977.
The calculated hfs coupling constants are in reasonable
agreement with the experimental values as obtained from
simulation that an assignment of these coupling constants to
proton positions in the radicals can be made unambiguously.
The only discrepancy consists in the calculated hfs coupling
constants for the OH proton of OH-adducts which differ dis-
tinctly from the experimental values. To resolve this discre-
pancy we have tested the influence of a rotation of the OH-
group along the C–O bond on the calculated hfs value for
the C3-OH-2-MeDHF radical (4). The results for the coupling
constant a(OH,3) and the change of the relative energy in
dependence on the torsion angle are shown in Fig. 12. Due
to the small activation barrier of 2.6 kcal mol the rotation
of the OH-group in the radical (4) is probable and the interac-
tion with the spin on C2 changes strongly. The value a(OH,3)
varies between ꢀ0.307 mT and +0.624 mT depending on the
torsion angle. For a free rotation of the OH group we get an
average value a(OH,3)av ¼ 0.158 mT whereas the experimental
value is a(OH,3)exp ¼ 0.06 mT.
1
2
T. Kausche, J. S a¨ uberlich, E. Trobitzsch, D. Beckert and K. P.
Dinse, Chem. Phys., 1996, 208, 375.
D. S. Stephenson, Nucl. Magn. Reson. Spectrosc., 1988, 20, 515.
D. J. Collins and A. M. James, Aust. J. Chem., 1989, 42, 223.
1
1
3
4
15 A. Schmitt and H. U. Reißig, Chem. Ber., 1995, 128, 871.
16 O. Temme, S. A. Taj and P. G. Andersson, J. Org. Chem., 1998,
6
3, 6007.
D. E. McGreer, N. W. K. Chiu and M. G. Vinje, Can. J. Chem.,
965, 43, 1398.
1
7
1
1
1
8
9
K. Gollnik and K. Knutzen-Mies, J. Org. Chem., 1991, 56, 4017.
K. Bernhard, PhD Thesis, University of Duisburg, Duisburg,
2001.
20 C. Lee, W. Yang and R. G. Parr, Phys. Rev. B, 1988, 37, 785.
ꢀ
1
21
22
23
A. D. Becke, J. Chem. Phys., 1993, 98, 5648.
A. D. Becke, J. Chem. Phys., 1996, 104, 1040.
M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A.
Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery, R.
E. Stratmann, J. C. Burant, S. Dapprich, J. M. Millam, A. D.
Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi, V. Bar-
one, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C. Adamo, S.
Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala, Q. Cui, K.
Morokuma, D. K. Malick, A. D. Rabuck, G. K. Raghavachari,
J. B. Foresman, J. Cioslowski, J. V. Ortiz, A. G. Baboul, B. B.
Stefanov, G. Liu, A. Liashenkop, P. Piskorz, I. Komaromi, R.
Gomperts, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham,
C. Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe, P.
M. W. Gill, B. Johnson, W. Chen, M. W. Wong, J. L. Andres,
C. Gonzales, M. Head Gordon, E. S. Replogle and J. A. Pople,
Gaussian 98, Rev. A7, 1998.
Conclusions
Using time resolved FT EPR and triplet anthraquinone-2,6-
disulfonate as the oxidant the radical cation and their succes-
sor radicals were studied for dihydrofuran and two methylated
derivates. The radical cations of 2,3-dimethyl-4,5-dihydro-
furan and 2-methyl-4,5-dihydrofuran were detected for the
first time in aqueous solutions. The successor radicals of the
dihydrofuran radical cation are the deprotonated cation and
the nucleophilic OH-adduct radicals which are generated with
2
4
B. Engels, L. A. Eriksson and S. Lunell, Adv. Quan. Chem., 1996,
7, 297.
2
25 S. Naumov, K. Hildenbrand and C. v. Sonntag, J. Chem. Soc.,
Perkin Trans. 2, 2001, 1648.
2
6
M. W. Wong, M. J. Frisch and K. B. Wiberg, J. Am. Chem. Soc.,
991, 113, 4776.
27 V. Barone and M. Cossi, J. Phys. Chem. A, 1998, 102, 1995.
1
7
ꢀ1
the same first order rate constant k ¼ (5 ꢁ 1) ꢄ 10 s . The
electron transfer rate constant from dihydrofuran to the
anthraquinone-2,6-disulfonate triplet was determined to be
2
2
8
9
S. Miertus, E. Scrocco and J. Tomasi, Chem. Phys., 1981, 55, 117.
J. Geimer and D. Beckert, Appl. Magn. Res., 2000, 18, 505.
30 F. Jolibois, J. Cadet, A. Grand, R. Subra, N. Rega and V. Barone,
J. Am. Chem. Soc., 1998, 120, 1864.
9
ꢀ1 ꢀ1
s . The interpretation of the time
kel ¼ (1.6 ꢁ 0.2) ꢄ 10 M
Phys. Chem. Chem. Phys., 2002, 4, 1738–1745
1745