Electron Transfer Reactions and Mobile Holes in Radiolysis of Squalane. Time-Resolved FDMR Study

Article abstract of DOI:10.1021/jp9532475

Fluorescence detected magnetic resonance (FDMR) is applied to study ion-molecule reactions in squalane (2,6,10,15,19,23-hexamethyltetracosane).Several time-resolved pulsed FDMR experiments are suggested to measure rate constants of scavenging by aromatic solutes.These constants vary from 5E8 to 1.6E9 mol^-1 dm³ s^-1.The ion-molecule reactions of radical cations are somewhat slower than reactions of radical anions.The scavenging of the alkane hole is significantly faster than ion-molecule reactions of molecular cations.The mobile hole decays on a time scale of several tens of nanoseconds.These FDMR results are in accord with our recent work on transient absorption spectroscopy in radiolysis of squalane.

Full text of DOI:10.1021/jp9532475

J. Phys. Chem. 1996, 100, 14681-14687
14681
Electron Transfer Reactions and Mobile Holes in Radiolysis of Squalane. Time-Resolved
FDMR Study†
I. A. Shkrob and A. D. Trifunac*
Chemistry DiVision, Argonne National Laboratory, Argonne, Illinois 60439
ReceiVed: NoVember 2, 1995; In Final Form: April 11, 1996X
Fluorescence detected magnetic resonance (FDMR) is applied to study ion-molecule reactions in squalane
(2,6,10,15,19,23-hexamethyltetracosane). Several time-resolved pulsed FDMR experiments are suggested to
measure rate constants of scavenging by aromatic solutes. These constants vary from 5 × 108 to 1.6 × 109
mol-1 dm3 s-1. The ion-molecule reactions of radical cations are somewhat slower than reactions of radical
anions. The scavenging of the alkane hole is significantly faster than ion-molecule reactions of molecular
cations. The mobile hole decays on a time scale of several tens of nanoseconds. These FDMR results are
in accord with our recent work on transient absorption spectroscopy in radiolysis of squalane.
Introduction
electrophylic gases (e.g., CO2) were added to prevent the
formation of A•-. Recently, it has been realized that many
commonly used electron scavengers exhibit complex chemistry
of their own: they quench the excited states of alkanes,5 their
anions can transfer electrons to the solute,6 and they can
dissociate or lose charge. As a result, instead of the simplifica-
tion of the reaction kinetics, one achieves its further complica-
tion. There is a need for alternative ways of studying ET
reactions occurring in radiolysis.
Despite rapid progress in the field, the radiation chemistry
of alkanes is still poorly understood.1,2 One of the obstacles is
lack of identification of short-lived reaction intermediates.
While optical detection is difficult due to weak and featureless
absorption, other detection techniques are insufficiently fast. The
traditional approach is to use probe molecules. This method is
rewarding in the studies of electron transfer (ET) reactions.
In the presence of aromatic scintillator molecules (A), the
primary charges, the solvent hole (RH•+) and the thermalized
electron (et-), are scavenged, and the recombination of the solute
Here we address this problem using pulsed time-resolved
fluorescence detected magnetic resonance (FDMR). With this
technique, the resonance microwave (µw) field is applied to
modulate the recombination yield of spin-correlated radical ion
pairs. The decrease in the fluorescence yield is studied as a
function of the external magnetic field and the delay time of
the µw pulses.3,7
FDMR has four advantages as compared to the detection of
the delayed fluorescence. First, FDMR is selective to radical
ions and no other species. Energy transfer reactions may yield
fluorescence, but they do not result in FDMR. Second, only
geminate pairs yield FDMR. Random encounters of radical ions
(including those occurring in spurs) do not contribute to the
FDMR signal. Third, the involved radical ions can be charac-
terized through their EPR signatures. Fourth, rapid phase
relaxation in some radical ions (such as CO2•-) precludes their
observation by FDMR.7 Within 20-30 ns after the formation
of CO2•-, the spin alignment between the radical ion partners
required for the formation of FDMR is lost. Thus, FDMR is
“blind” to nongeminate reactions, energy transfer, and the
involvement of certain radical ions. This results in considerable
simplification of the observed kinetics.
ions yields the delayed fluorescence from A*:2,3
1
et- + A f A•-
(1)
RH•+ + A f RH + A•+
(2)
(3)
(4)
(5)
RH•+ + A•- f RH + 1,3A*
A
•+ + A•- f A + 1,3A*
et- + A•+ 1,3A*
1A* f A + hν
f
(6)
This fluorescence and the absorption of aromatic solute ions
A•( are easy to detect, which explains the popularity of the
scavenging techniques.
Unfortunately, the solute molecules are involved in reactions
other than reactions 1-6. At low solute concentration the
scavenging is slow, while at high concentration the energy
transfer from the singlet alkane can be a significant source of
1A*.4 Absorption spectra of aromatic monomer and dimer
cations, anions, proton adducts, radicals, and triplets frequently
overlap.1 Multiple pairs and cross-recombination between the
ions in different spurs make the kinetics less tractable. All these
factors introduce ambiguity in the interpretation of the data. To
simplify the kinetics, many researchers added a large quantity
of electron (or hole) scavengers. For example, haloalkanes and
However, there is a drawback; FDMR is a slow technique.
The detection is delayed by the 10-50 ns that is required for
the spin mixing and the µw excitation of the geminate pairs.3,7
In nonviscous alkanes (η ≈ 1 cP), many fast ET reactions are
complete by this time. In this work, we study FDMR in the
viscous hydrocarbon squalane (η ≈ 20 cP 1) for which the
diffusion-controlled reactions occur on the FDMR time scale.
Several time-resolved experiments were used to examine
electron and proton transfer reactions by means of pulsed
FDMR. We find that the rate constants of the ET reactions
fall in the range (0.5-1.6) × 109 mol-1 dm3 s-1. These rates
are insufficient to explain the fast generation of the solute radical
cations as observed by FDMR and transient absorption spec-
† Work performed under the auspices of the Office of Basic Energy
Sciences, Division of Chemical Science, US-DOE, under contract number
W-31-109-ENG-38.
* To whom correspondence should be addressed.
X Abstract published in AdVance ACS Abstracts, August 15, 1996.
S0022-3654(95)03247-3 CCC: $12.00 © 1996 American Chemical Society

14682 J. Phys. Chem., Vol. 100, No. 35, 1996
Shkrob and Trifunac
Figure 1. (a) FDMR spectra observed from a 5 × 10-4 mol dm-3 solution of perylene in N2-saturated squalane (50 ns µw pulse, 200 ns boxcar
gate). The emission at λ > 460 nm was collected. The spectra were normalized at the center; the delay time t is indicated next to the traces. The
broad signal is from a long-lived olefin ion of squalane, SQ•+. (b) Pulse-sweeping FDMR experiment with the same solution (50 ns µw pulse, 50
ns boxcar gate). The kinetics were collected at zero and 3.5 mT offset field; for comparison, the latter trace is normalized. The broken curve is
the convolution of the zero-offset curve with exp(-kt), k ) 2.6 × 105 s-1
.
troscopy.1 From the latter, the scavenging of the holes was
found to proceed with rate constants ∼(5-7) × 109 mol-1 dm3
s-1. The present work supports the identification of the mobile
cation species as the squalane radical cation.
Results and Discussion
Squalane Ions. The radical cation partner of A•- (dubbed
SQ•+) as it appears in our FDMR spectra resembles a Gaussian
line with the second moment σ of the spectrum ∼2.52 mT
(Figure 1a). This ion must be very stable since in (1-10) ×
10-6 mol dm-3 perylene solutions the broad signal from SQ•+
persists for 15 µs after the electron beam pulse. For inhomo-
geneously broadened lines
Experimental Section
The scintillators 2,5-diphenylcarbazole (PPO), biphenyl-d10
(d10Bh), anthracene-d10 (d10An), and octafluoronaphthalene (F8-
Np) were used as received; naphthalene-h8 (h8Np), perylene (Pe),
and rubrene were twice sublimed in Vacuo; N,N,N′,N′-tetra-
methyl-p-phenylenediamine (TMPD) was recrystallized from
ethanol and sublimed three times (all from Aldrich); perdeu-
terated naphthalene (d8Np), p-terphenyl, and nitrobenzene (d5-
NB) were supplied by Cambridge Isotope Labs; triethylamine
(TEA, 99+%), tetramethylethylene (TME), and benzonitrile
(BN) were used as received from Aldrich. n-Hexane and
n-hexadecane were passed through 1.5 m silica gel columns.
The quality of purification was tested by measuring absorption
at 200-220 nm. The methods for squalane purification are
given in ref 1.
σ2 ) (1/3)
A I (I + 1)
(7)
∑
k
k k k
where Ak is the hyperfine coupling constant (hfcc) and Ik is the
spin of the kth nucleus.8 For aromatic ions, the spectrum width
σ can be calculated from the reported EPR data.9 For example,
for d10An•+ and d10An•- the second moments are 0.19 and 0.16
mT, respectively, while for Pe•+ σ ) 0.58 mT. Thus, for
deuterated solutes σ(A•() , σ(SQ•+), and the [SQ•+ A•-] and
[A•+ A•-] pairs can be readily distinguished spectrally.
We found that SQ•+ reacts with aromatic solutes with
ionization potentials (IP) less than 7.7 ( 0.1 eV (all IP used
are the gas-phase values):
Radiolysis was performed with 12 or 25 ns electron beam
pulses from a 3 MeV van de Graaff accelerator operated at 5-10
nC per pulse. The solutions were purged with pure nitrogen or
carbon dioxide prior to and during the experiment. The details
of our FDMR spectrometer may be found in ref 7. Three types
of time-resolved FDMR experiments were performed. In the
first, the delay time td of a 30-80 ns µw pulse was stepped in
increments of 10-50 ns relative to the radiolytic pulse. The
FDMR signal (the difference in the fluorescence with and
without the resonant µw field) was sampled within 50-100 ns
after the µw pulse using a boxcar averager (see diagram in
Figure 1b; we will refer to this experiment as the pulse-sweeping
FDMR). In the second, the delay time td was fixed and the
FDMR kinetics (the difference between the fluorescence traces
with and without the µw field) was sampled using a digital
oscilloscope (10 ns/point). In the third experiment, the magnetic
field was swept for a fixed delay time td of the µw pulse and
the FDMR spectrum was sampled with a boxcar averager (a
typical FDMR spectrum is shown in Figure 1a). If not indicated
otherwise, the µw field was 0.28 mT. All experiments were
performed at 23 °C.
SQ•+ + A f SQ + A•+
(8)
Following this reaction, the fraction of the [SQ•+ A•-] pairs
decreases, while the fraction of the [A•+ A•-] pairs increases.
Consequently, the broad signal from SQ•+ decays faster than
the FDMR signal from A•+ and A•- (the narrow line at the
spectrum center, Figure 1a).
When the FDMR kinetics is probed with a short µw pulse
(10-50 ns), the decay of the signal at the spectrum center
follows Λ(t), the geminate decay of radical ion pairs.7,10 The
signal from SQ•+ obtained with a µw pulse applied at t ) td
decays as Λ(t) exp(-k8[A]td), where k8 is the rate constant of
reaction 8. Comparing the pulse-sweep FDMR kinetics
sampled at the spectrum center and 3.5 mT off-center (where
the signal is from SQ•+), one can find the rate constant of
reaction 8 by deconvolution of the off-center trace (Figure 1b).
For anthracene and perylene we obtained 5 × 108 mol-1 dm3
s-1. When 0.1 mol dm-3 TME was added to the solution, the
broad signal changed. It is known that at this high concentration
of TME the radical cations of TME aggregate, giving a broad

Mobile Holes in Radiolysis of Squalane
J. Phys. Chem., Vol. 100, No. 35, 1996 14683
featureless FDMR signal.11 In perylene-TME solutions, this
broad signal decayed with the rate constant 1.2 × 109 mol-1
scavenge this ion. This ionization energy is too low for an
alkane radical cation.
dm3 s-1
.
Radiolysis of 0.01 mol dm-3 squalene in n-hexane (with 10-4
mol dm-3 biphenyl-d10) yields a FDMR spectrum very similar
to that observed in neat squalane. This spectrum exhibits the
wings of the same width as SQ•+ and a narrower signal (with
fwhm ≈ 1.5 mT) superimposed on the signal from d10Bh•- (σ
≈ 0.12 mT). The scavenging of squalene•+ follows the same
dependence on the solute IP as that of SQ•+. In naphthalene-
d8 solutions, the signal from squalene•+ grows with the delay
time of the µw pulse at any concentration of naphthalene-d8.
Apparently, squalene scavenges naphthalenes•+. The IP of most
olefins, dienes, and trienes is in the range from 8.1 to 9 eV.17
The lowest IP known for a diene (7.5 eV) was found for 2,5-
dimethyl-2,4-hexadiene.17 This olefin is similar to squalene in
structure. Double bonds in squalene are located near the
branching points and are not conjugated. The spin density in
squalene•+ must be localized at the single double bond. The
same applies to any olefin ion formed on fragmentation of
squalane, so the similarity between the FDMR spectra for
squalene•+ and SQ•+ is expected. An estimate of σ for various
olefin ions of squalane using the EPR parameters reported for
smaller ions of similar structure yields 2.2-2.7 mT.9 This is
close to the observed spectrum width of 2.52 mT.
If the delay time of the µw pulse is fixed rather than swept,
the normalized FDMR traces obtained at the spectrum center
and off-center are identical. Though SQ•+ ions are fully
scavenged by aromatic solute, the signal from SQ•+ was still
observed. FDMR spectra reflect the relatiVe abundances of the
pairs that exist at the moment of application of a short µw pulse.
Our result also implies that the diffusion coefficients of SQ•+
and A•+ are close; otherwise the trace obtained at the center
would decay faster than that at an off-center position.
Let us discuss the identity of SQ•+. The broad FDMR signal
observed in our experiments resembles that observed by Okazaki
et al., who used a continuous-wave FDMR.12 It was speculated
that the broad signal is from the squalane hole involved in rapid
resonant charge transfer,
RH•+ + RH f RH + RH•+
with k9 = 107 mol-1 dm3 s-1 12
(9)
.
Charge hopping (9) mixes
nuclear subensembles. As the rate of the hopping increases, it
first broadens and then narrows the FDMR signals. Tadjikov
et al. suggested that even at 77 K the charge hopping is so fast
that it narrows the FDMR spectra.13 This would explain the
observed 20% increase in the spectrum width on dilution of
squalane with 3-methylheptane.13 The dilution reduces the rate
of charge hopping and broadens the FDMR signal. It was
estimated that k9 ≈ 8 × 107 mol-1 dm3 s-1 at 77 K and 3 ×
108 mol-1 dm3 s-1 at 290 K.13,14 However, the broadening
observed on dilution (ca. 0.2 mT) could well be due to matrix
effects. In the spectra reported by Okazaki et al., the FDMR
signal at 290 K was 2 times broader than the signal obtained at
45-150 K.12 If, following Tadjikov et al., one tries to explain
this broadening through a faster reaction (9), it is necessary to
assume that even at 290 K the hopping is still in the
line-broadening regime. The only way to account for this
anomalous result is to suggest that the two groups observed
two different radical cation species. From the study on the
damping of the ∆g beats, Veselov et al. suggested that at 290
K the FDMR signal of the solvent hole in squalane was ∼0.6
mT wide.14 The signal from SQ•+ is 4 times wider.
From our transient absorption study1 we deduced that
radiolysis of squalane results in the formation of two radical
cation species, a mobile ion-I and a normally diffusing ion-II.
The latter undergoes a transformation on the microsecond time
scale and yields a stable ion-III, whose lifetime exceeds 50 µs.
Ion-I and ion-II are formed in the very early time frame. Ion-I
fully decays in 40 ns; our estimate of its lifetime in a free state
is ∼25 ns. This time is shorter than the time of spin mixing
and µw-induced spin-flip required for the development of
FDMR. No ion with a lifetime that short can be observed with
FDMR, continuous-wave and time-resolved alike. The scav-
enging of ion-I occurs with rate constants (5-7) × 109 mol-1
dm3 s-1, while typical ET reactions of solute cations are 4-5
times slower. Scavenging of normally diffusing ion-II and ion-
III by aromatic solutes follows the pattern of SQ•+. We believe
that these ions and SQ•+ are the same species.
It does not seem likely that SQ•+ is an impurity ion. In
purified squalane, the half-life of et- is 0.43 µs; aromatic solutes
scavenge the electron with a rate constant of (3-5) × 1011 mol-1
dm3 s-1. Hence, the conjugated double bond impurity can
account for no more than 10-5 mol dm-3. The most abundant
impurity in squalane must be the products of partial hydrogena-
tion of squalane. From the extinction coefficient of squalane
at 220 nm, we estimated that the purified solvent contained no
more than 7 × 10-4 mol dm-3 of the olefin impurity. Thus,
the impurity can account for no more than 10% of the secondary
ions.
The spectrum width σ of the squalane hole observed in low-
temperature solid is 1.2 mT.13,15 Using k9 from ref 14, we
calculated that the time T2 of the electron exchange relaxation
due to reaction 9 would be 20-50 ns (T2 ) σ2τres is the
-1
residence time of the hole at room temperature, 1.7 ns). This
relaxation rate is insufficient for the narrowing of the FDMR
signal within the duration of short µw pulses in our time-
resolved experiments. At best, such a slow hopping would
introduce a line-broadening.16 Experimentally, the FDMR signal
neither narrows nor widens when the length of a weak µw pulse
(0.1 mT) increases from 20 ns to 2 µs. There is no indication
that SQ•+ undergoes rapid resonant charge transfer (reaction
9).
It seems very likely that SQ•+ is a molecular ion, probably
an olefin radical cation formed on fragmentation of the parent
hole RH•+.3 Addition of 0.1-0.5 mol dm-3 cis- and trans-
decalins, toluene, and norbornane does not result in the
disappearance of the broad signal from SQ•+, although decalins
are known to scavenge the squalane hole at 77 K.15 Naphthalene
and biphenyl (IP ) 8 and 7.9 eV, respectively17) also do not
scavenge the signal from SQ•+. Only solutes with IP < 7.8 eV
We turn now to the FDMR studies of the ion-molecule
reactions of aromatic ions. In these experiments we studied
electron transfer from solute ions to solute molecules. For
radical anions, the kinetics measurements were complicated by
an unexpected effect: the second scavenger affected the initial
distribution of the solute ions.
Effect of CO2 on the FDMR Kinetics. In the presence of
electron scavenger CO2 the reaction scheme (1)-(6) must be
extended:

14684 J. Phys. Chem., Vol. 100, No. 35, 1996
Shkrob and Trifunac
•-
et- + CO2 f CO2
(10)
(11)
(12)
(13)
RH•+ + CO2•- f R• + •HCO2
A
•+ + CO2•- f 1,3A* + CO2
CO2•- + A f CO2 + A•-
Reaction 13 occurs with pyrene and perylene; with anthracene
and PPO it is not energetically favorable. The mobility of CO2•-
is ∼10 times higher than that of aromatic ions.
In the presence of 0.037 mol dm-3 CO2 (saturated solution)
most of the geminate pairs generated in radiolysis are [SQ•+
CO2•-]. These pairs do not contribute to the scintillator
emission. The [A•+ CO2•-] pairs yield strong emission from
1A* on recombination. However, the FDMR spectra sampled
in N2- and CO2-saturated solutions are identical; nothing
indicates the involvement of the [A•+ CO2•-] pairs in the
formation of FDMR. The EPR spectrum of CO2•- in water is
indicative of the fast relaxation with T2 ≈ 10-20 ns.18 With
this short T2 time the [A•+ CO2•-] pairs (and all subsequent
pairs) lose their spin correlation by the time of FDMR
acquisition. Therefore, the FDMR obserVed in CO2 solutions
is from the [SQ•+ A•-] and [A•+ A•-] pairs.
Since both the aromatic solute and CO2 compete for et-, the
yield of [SQ•+ A•-] pairs decreases on addition of CO2. Taking
into account the difference in the concentration of A and CO2,
one could expect that virtually no FDMR signal would be
observed in the saturated CO2 solutions. We compared the
kinetics of fluorescence and FDMR in N2- and CO2-saturated
solutions of biphenyl, PPO, and An in n-hexane, cyclohexane,
n-hexadecane, and squalane. The results may be summarized
in the following way:
Figure 2. Time dependences of FDMR for 5 × 10-14 mol dm-3
anthracene-d10 (a) and perylene (b) in N2- and CO2-saturated squalane
([CO2] ) 0.036 mol dm-3). The emission at 300-400 nm (a) and
460-600 nm (b) was collected; the traces were obtained on the top of
the resonance lines from the aromatic radical ions. A 12 ns electron
pulse was applied at t ) 0; a 30 ns µw pulse (the oscilloscope trace of
the diode signal is shown next to the traces) was applied (a) at td ) 50
ns and (b) td ) 50 and 450 ns. For the latter delay time the traces
were multiplied by 3. The dotted curves are the CO2 traces normalized
to the N2 traces at t ) td. As seen from these traces, on the saturation
of squalane with CO2 the decay kinetics becomes faster and the FDMR
signal decreases several times.
(i) The addition of CO2 increases the yield of the delayed
fluorescence.
(ii) In CO2-saturated solutions the decay of the FDMR signal
is significantly faster than in N2-saturated solutions (Figure 2).
(iii) The FDMR effect always decreases with addition of CO2
to scintillator solutions. For a given [A], the reduction factor
(γ) is the same for all scintillators (cf. Figure 2a,b); it increases
with the delay time of the µw pulse (Figure 2b) and correlates
with the bulk viscosity of the solution. In CO2-saturated
squalane ([d10An] ) 5 × 10-4 mol dm-3) the reduction in the
FDMR signal sampled with the µw pulse applied at 50 ns was
γ ≈ 3.7 times (compare with γ ) 12 for td ) 1.5 µs), in
n-hexadecane γ ) 11 times, and in n-hexane we did not find
any FDMR signal. This reduction is concentration-dependent:
for example, for [d10An] ) 5 × 10-4 mol dm-3 (n-hexadecane)
γ ) 11, while for [d10An] ) 5 × 10-3 mol dm-3 solution γ )
2.2. In squalane, the FDMR signal in CO2-saturated solution
can be observed even at [A] ) 10-5 mol dm-3 (γ ≈ 35).
To account for the observed behavior, we used a Monte Carlo
simulation program4,19 using the set of parameters from ref 1.
Figure 3a shows the time dependence of the emission from the
[SQ•+ A•-] pairs, which are the main contributors to the FDMR.
The simulation reproduces the values γ and its time dependence
quite well.
without 0.037 mol dm-3 CO2 using the Monte Carlo program
(Figure 3b). In low-viscosity solutions, the broadening of Π(r)
relative to p(r) has a minor effect on the recombination: by the
time of the FDMR observation the distribution is broadened by
diffusion of the ions. In viscous solutions, the broader distribu-
tion causes a lesser number of pairs to recombine by the
observation time. Following the WAS approach,21 the reduction
factor γ can be estimated as fe(c) × [A]/c, where c ) [A] +
[CO2] and fe(c) is the fraction of scavenged electrons fe(c) )
[Rc]0.6/(1 + [Rc]0.6). At [CO2] . [A] the factor γ c-0.6, as
was observed experimentally (see below).
Ion-Molecule Reactions of Aromatic Solutes. The rate
constants of ET reactions measured with our time-resolved
FDMR method are compiled in Tables 1 and 2. Here we
consider several experimental schemes that have been suggested
to determine these rate constants.
The gas-phase electron affinity (EA) of benzonitrile is ∼0.1
eV higher than EA of biphenyl and ∼0.25 eV lower than EA
of perylene. Thus, benzonitrile accepts an electron from Bh•-,
Bh•- + BN f Bh + BN•-
(14)
while BN•- transfers an electron to perylene:
It is frequently assumed that the initial distribution Π(r) of
radical ions follows the distribution p(r) of thermalized electrons.
For dilute solutions of electron scavengers this assumption is
incorrect, since the radial distribution of the electrons widens
with time.20 To illustrate this point, we calculated Π(r) for 5
× 10-4 mol dm-3 anthracene solution in squalane with and
BN•- + Pe f BN + Pe•-
(15)
Since the IP of benzonitrile, nitrobenzene, and CCl4 are more
than 10 eV,17 these solutes do not scavenge any of the involved

Mobile Holes in Radiolysis of Squalane
J. Phys. Chem., Vol. 100, No. 35, 1996 14685
Figure 4. (a) FDMR decay traces obtained in N2-saturated 10-3 mol
dm-3 squalane solution of biphenyl-d10 with (from the top to the bottom)
1.5, 3, 6, 12, and 18.5 × 10-3 mol dm-3 triethylamine. A 25 ns electron
beam pulse was applied at t ) 0, the µw pulse was applied at td ) 350
ns, and the emission at 300-400 nm was collected. The FDMR kinetics
were obtained on the top of the resonance line from the radical cations
of biphenyl-d10. (b) The concentration dependence of the first-order
rate constant k of the FDMR decay. The constant k was found by
deconvolution of the FDMR traces with exp(-kt); the [TEA] ) 1.5 ×
10-3 mol dm-3 trace was used as the standard.
Figure 3. (a) Monte Carlo simulation of the emission yield from
recombination of geminate pairs of SQ•+ and A•- for [A] ) 5 × 10-4
mol dm-3
.
The mobility of et- was 0.016 cm2/(V s), p(r)
r2
exp(-r/bE), bE ) 2.06 nm; the rate constant of scavenging et- by CO2
and aromatic solute was taken as 3 × 1011 mol-1 dm3 s-1; the
fluorescence time of A was 1.5 ns; lifetime of the primary squalane
ion is 25 ns; and the rate constants of scavenging the primary and
secondary ions of squalane by A were taken as 6.5 × 109 and 1.5 ×
109 mol-1 dm3 s-1, respectively. The trace for a CO2-saturated squalane
(0.037 mol dm-3) was normalized at t ) 50 ns. The ascending line is
the time dependence of the ratio γ of the emission yields from the N2-
and CO2-saturated solutions. (b) The distribution Π(r) of radical anions
at the moment of electron scavenging. The traces were simulated for
the CO2- and N2-saturated solutions of 5 × 10-4 mol dm-3 aromatic
scintillator. In the CO2 solution, the distribution of the anions (both
addition of 0.01-0.1 mol dm-3 benzonitrile, the FDMR spectra
from perylene solutions undergo several changes.
First, the fwhm of the FDMR spectra at td ) 50 ns decreases
from 1.49 to 1.2 mT. This change indicates that most electrons
were scavenged by benzonitrile, whose radical anion has a
narrower line than Pe•-. For longer delay times td of the µw
pulse, the line at the center of the spectrum broadens and finally
becomes as broad as that of Pe•- (reaction 15). The rate of
this broadening suggests that k15 ≈ 1.8 × 109 mol-1 dm3 s-1
.
Second, on addition of benzonitrile the decay of FDMR
becomes faster. When [BN] > 0.01 mol dm-3, further addition
of the scavenger does not change the decay kinetics while
decreasing the FDMR signal as γ ≈ [BN]-0.6 (see above).
Apparently, most of the thermalized electrons are rapidly
scavenged (i.e., Π(r) ≈ p(r)) and the decay kinetics do not
change upon further increase in the scavenger concentration.
Addition of TMPD to the perylene solution causes a notable
acceleration of the FDMR decay. TMPD scavenges A•+ and
SQ•+,
CO2 and A•-) is considerably narrower.
•-
TABLE 1: Rate Constants of Electron Transfer Reactions
of Radical Anions in Squalane at 23 °C (×109 mol-1 dm3
s-1
)
donor•-
EA, eVa
0.15
acceptor
EA, eV rate constant
biphenyl
benzonitrile
perylene
CCl4
nitrobenzene-d5
perylene
0.25
1.6
1.8
1.6
1.5
1.6
∼0.7
perylene
∼0.7
2.0
2.0
benzonitrile
0.25
∼0.7
TMPD + A•+ f TMPD•+ + A
(16)
a Gas-phase EA is taken from ref 27.
without scavenging et- and A•-. Due to the low IP of TMPD,
no fluorescence is formed when TMPD•+ recombines with a
radical anion whose parent molecule has gas-phase EA > 0.35
( 0.05 eV.23 In the presence of TMPD the decay of the FDMR
signal from 1Pe* follows the geminate decay curve Λ(t)
multiplied by exp(-k16[TMPD]t). The deconvolution yielded
TABLE 2: Rate Constants of Electron Transfer and Proton
Transfer for Radical Cations in Squalane at 23 °C (×109
mol-1 dm3 s-1
)
IP,
rate
acceptor•+
SQ•+
tetramethylethylene
perylene
biphenyl
triethylamine
eVa
donor
IP, eV constant typeb
∼7.8 perylene
6.9
6.9
6.2
7.5
7.5
0.5
1.4
1.5
0.6
0.26
ET
ET
ET
ET
PT
k16 ) (6 ( 0.5) × 108 mol-1 dm3 s-1
.
8.3 perylene
A similar experiment was performed with d10Bh/BN solution
(10-3 and 1.5 × 10-2 mol dm-3, respectively). The emission
of biphenyl-d10 was observed at 300-400 nm. On addition of
benzonitrile the central line broadens from 0.7 to 1.2 mT
following reaction 15; this broadening proceeds with the rate
constant 1.8 × 109 mol-1 dm3 s-1. In another experiment, we
applied the µw pulse at td ) 0.35 µs and collected the FDMR
traces for different concentrations of triethylamine added to the
biphenyl-d10 solution (Figure 4a). Triethylamine (IP ) 7.5 eV)
6.9 TMPD
7.95 triethylamine
7.5 triethylamine
a Gas-phase IP is taken from refs 17 and 27. b Type of the reaction;
ET, electron transfer; PT, proton transfer.
radical cations. To discriminate against the fluorescence from
benzonitrile, we observed the emission at λ > 460 nm.22 On

14686 J. Phys. Chem., Vol. 100, No. 35, 1996
Shkrob and Trifunac
scavenges both SQ•+ and Bh•+. Apparently, the recombination
of the [TEA•+ Bh•-] pairs yields 1Bh* since the resonance lines
from TEA•+ appear in the spectra. For [TEA] < 2 × 10-3 mol
dm-3 the FDMR signal observed at td ) 0.35 µs increases with
[TEA]; for higher concentrations of triethylamine the signal
decreases. In the latter range the FDMR decay kinetics become
considerably faster, but the spectrum shape does not change.
This indicates proton transfer between TEA•+ and TEA.24 The
dependence of the rate constant of the FDMR decay on [TEA]
exhibits two linear ranges (Figure 4b). For the first one (<8 ×
10-3 mol dm-3) the slope is close to 6 × 108 mol-1 dm3 s-1
,
which corresponds to the rate constant of charge transfer; in
the second range the slope is 2.6 × 108 mol dm-3, which
corresponds to the rate constant of proton transfer.
CCl4 readily scavenges both electrons and aromatic solute
anions. Its radical anion rapidly dissociates into CCl3• and Cl-
precluding the formation of FDMR:
•- + CCl4 f A + CCl4
CCl4•- f •CCl3 + Cl-
(17)
(18)
•-
A
Reactions 17 and 18 accelerate the decay of FDMR considerably
more than the distribution narrowing effect discussed above.
To measure the rate constant of reaction 17, we saturated the
Pe/CCl4 solutions with 0.037 mol dm-3 CO2. At this concen-
tration of CO2, the addition of (1-10) × 10-3 mol dm-3 CCl4
should not introduce any additional narrowing effect. We
Figure 5. (a) Normalized FDMR spectra from a N2-saturated squalane
solution of 8.8 × 10-4 mol dm-3 perylene with (i-iii) 3 × 10-3 mol
dm-3 and (iv) 1.5 × 10-2 mol dm-3 of nitrobenzene-d5. The bold traces
i and iv were obtained at td ) 50 ns; traces ii and iii were obtained at
td ) 100 ns and td ) 300 ns, respectively (50 ns µw pulse, 50 ns boxcar
gate, λ > 460 nm). For higher concentration of nitrobenzene-d5 and
longer delay times td the intensity of the broad signal from SQ•+ relative
to the central line rapidly decreases. The shoulder to the left of the
spectrum center is from nitrobenzene-d5; this signal increases with td.
Both changes are due to reaction 19. (b) Pulse-sweeping FDMR
experiment performed at the spectrum center and 3.5 mT off-center
(shown by arrows in the spectra) for [Pe] ) 3.6 × 10-4 mol dm-3 and
[NB] ) 3.5 × 10-3 mol dm-3. The dashed line is the zero-offset trace
obtained k17 ) 1.6 × 109 mol-1 dm3 s-1
.
Perdeuterated
nitrobenzene also reacts with Pe•-,
Pe•- + NB f Pe + NB•-
(19)
and yields a narrow resonance line with g ) 2.0059 (Figure
5a). On addition of nitrobenzene, the [SQ•+ Pe•-] pairs (which
fluoresce at λ > 460 nm) are transformed into [SQ•+ NB•-]
pairs which fluoresce in the UV. At the same time, both the
[Pe•+ Pe•-] and [Pe•+ NB•-] pairs yield emission at λ > 460
nm. This causes a fast decay of the signal in the wings of the
FDMR spectrum. By comparative analysis of the pulse-
sweeping FDMR kinetics at the spectrum center and in the wings
convoluted with exp(-kt), k ) 5.3 × 106 s-1
.
a spectrum from the [SQ•+ BN•-] pairs was observed. On
addition of TEA, a set of resonance lines from TEA•+ emerged.
If the scavenging rate of the solvent hole was diffusion-limited,
then only 2% of the holes would be scavenged in 50 ns. The
actual conversion is an order of magnitude greater.
(e.g., Figure 5b) we found k19 ) 1.5 × 109 mol-1 dm3 s-1
.
Mobile Radical Cations in Squalane. Our pulse radiolysis
results suggest that one of the cations formed in ionization of
squalane (ion-I) is ∼7 times more mobile than other cations. It
was found that the scavenging of ion-I by aromatic solutes yields
solute radical cations and that the absorption spectrum of ion-I
resembles that of the solvent hole stabilized at 72 K.25 Both of
these observations point to the radical cation of squalane.
In this section we consider the FDMR evidence that the
mobile ions-I are indeed the radical cations in geminate pairs.
Although the lifetime of ion-I is too short to be detected by
means of magnetic resonance (∼25 ns1), the fast generation of
aromatic radical cations can be observed by FDMR.
It can be argued that the fluorescence bands of TEA and BN
overlap and that TEA has greater luminescence yield per
recombination. These two factors would bias FDMR toward
the observation of the emission from [TEA•+ BN•-] pairs. To
exclude the interference of the TEA emission, we performed
FDMR experiments with solutes whose fluorescence bands do
not overlap with that of TEA: perylene (λ > 460 nm) and
rubrene (λ > 540 nm). It is known that the fluorescence of
TEA exhibits a tail to 400 nm in moist alkane solutions.26 We
found no evidence that this happens in pure squalane. No
1
FDMR was observed at λ ) 290 nm, where TEA* has its
emission maximum. Apparently, the recombination of TEA•+
1
A similar FDMR experiment has been performed by Lef-
cowitz and Trifunac,24 who tried to observe the mobile solvent
radical cation in cyclohexane. Contrary to their expectations,
the growth of the signal from the solute radical cations at td >
50 ns was not faster than diffusion-controlled.24 This was
interpreted as an indication of the short lifetime of the mobile
species.
In our experiments we observed the growth of FDMR from
TEA•+ as a function of [TEA] in the benzonitrile/triethylamine
solution. Benzonitrile emits at 300 nm;22 its fluorescence was
collected using a Corning 3-54 filter. Due to its high IP,
benzonitrile does not scavenge the holes. With no TEA present,
and Pe•- does not yield TEA*.24
The use of the donor-acceptor solutes presents two compli-
cations: first, perylene and rubrene scavenge TEA•+; second,
unlike benzonitrile, perylene scavenges SQ•+ and the solvent
holes. Therefore, we used dilute (1-20) × 10-6 mol dm-3
solutions. For [Pe] ) 2 × 10-5 mol dm-3 the lifetime of the
normally diffusing cation (such as TEA•+ and SQ•+) is ∼90
µs. Since the mobile cation is short-lived, its scavenging by
perylene accounts for less than 0.05% of the total decay. On
addition of ca. 10-3 mol dm-3 TEA the lifetime of SQ•+ is ∼3
µs, and the lifetime of TEA•+ is ∼6 µs. Although the quality
of FDMR spectra is poor, the signal from TEA•+ was clearly

Mobile Holes in Radiolysis of Squalane
J. Phys. Chem., Vol. 100, No. 35, 1996 14687
cations are scavenged at slower rates ((0.5-1) × 109 mol-1
dm3 s-1). Proton transfer reactions are even slower. The
scavenging of the mobile ions proceeds with a rate constant
(5-7) × 109 mol-1 dm3 s-1 1
.
Our identification of the mobile ions in squalane as the solvent
radical cations was supported by (i) the absorption spectrum
and (ii) the fast generation of solute radical cations on
scavenging.1 In this paper we demonstrate (iii) the participation
of the mobile ions in the formation of FDMR.
We believe that the high mobility of the squalane radical
cations is due to fast resonant charge transfer with τres ≈ 200
ps.1 For this short τres, the FDMR spectrum of the solvent hole
must be a single narrow line. However, at room temperature
the hole is too short-lived to be observed by FDMR. The narrow
FDMR signals observed by Tadjikov et al.13 are from impurity
ions.
Acknowledgment. We thank Dr. C. D. Jonah for his help
with the Monte Carlo simulations, Dr. M. C. Sauer, Jr., for
constructive criticism, and Dr. D. W. Werst and Mr. R. Lowers
for their operation of the accelerator.
References and Notes
(1) Shkrob, I. A.; Sauer, M. C., Jr.; Trfiunac, A. D. J. Phys. Chem.
1996, 100, 5993.
(2) Willard, J. E. In Radiation Chemistry: Principles and Applications;
Farhataziz, Rodgers, M. A. J. Eds.; VCH: New York, 1987; p 395.
(3) Werst, D. W.; Trifunac, A. D. J. Phys. Chem. 1991, 95, 3466.
(4) Sauer, M. C., Jr.; Jonah, C. D.; Naleway, C. A. J. Phys. Chem.
1991, 95, 730.
(5) Sauer, M. C., Jr.; Jonah, C. D.; Le Motais, B. C.; Chernowitz, A.
C. J. Phys. Chem. 1988, 92, 4099.
(6) Sauer, M. C., Jr.; Jonah, C. D. J. Phys. Chem. 1992, 96, 5872.
(7) Shkrob, I. A.; Trifunac, A. D. J. Chem. Phys. 1995, 103, 551.
(8) Poole, C. P., Jr. ESR: A ComprehensiVe Treatise on Experimental
Techniques; Wiley Interscience: New York, 1967.
(9) Landolt-Bornstein. Magnetic Properties of Free Radicals; Springer-
Verlag: New York, 1977.
(10) Smith, J. P.; Trifunac, A. D. J. Phys. Chem. 1981, 85, 1645.
(11) Desrosiers, M. F.; Trifunac, A. D. J. Phys. Chem. 1986, 90, 1560.
(12) Okazaki, M.; Nunome, K.; Matsuura, K.; Toriyama, K. Bull. Chem.
Soc. Jpn. 1990, 63, 1396.
(13) Tadjikov, B. M.; Melekhov, V. I.; Anisimov, O. A.; Molin, Yu. N.
Radiat. Phys. Chem. 1989, 34, 353.
(14) Veselov, A. V.; Bizyaev, V. L.; Melekhov, V. I.; Anisimov, O.
A.; Molin, Yu. N. Radiat. Phys. Chem. 1989, 34, 567.
(15) Werst, D. W.; Percy, L. T.; Trifunac, A. D. Chem. Phys. Lett. 1988,
153, 45.
(16) Shkrob, I. A.; Werst, D. W.; Trifunac, A. D. J. Phys. Chem. 1994,
98, 13262.
(17) Levin, R. D.; Lias, S. G. Ionization Potential and Appearance
Potential Measurements; 1971-1981 NRSDS-NBS 71.
(18) Chawla, O. P.; Fessenden, R. W. J. Phys. Chem. 1975, 79, 2693.
(19) Schmidt, K. H.; Sauer, M. C., Jr.; Lu, Y.; Liu, A. J. Phys. Chem.
1990, 94, 244.
Figure 6. FDMR spectra from N2-saturated solution of squalane with
2 × 10-5 mol dm-3 perylene and (i) no, and (ii) 5 × 10-3 and 7 ×
10-4 mol dm-3 triethylamine (50 ns µw pulse, 50 ns boxcar gate, λ >
460 nm). Traces i and ii were obtained at td ) 50 ns; for other traces
td is indicated in the figure. The lines from TEA•+ are marked with
arrows. These lines emerge on the top of the broad signal from SQ•+
from the earliest observation times. The spectra were normalized by
the signal at the spectrum center.
observed from the earliest detection time (Figure 6) and
indicated 20-30% conversion of the holes.
Another type of FDMR experiment was performed with
perylene/nitrobenzene-d5 solutions. The emission from 1Pe* at
λ > 460 nm was collected. Due to reaction 20, for [d5NB] )
0.06 mol dm-3 the lifetime of Pe•- is 10 ns. The recombination
of the [SQ•+ d5NB•-] pairs does not yield the fluorescence at λ
> 460 nm. Thus, the only pair that yields FDMR at λ > 460
nm and td > 50 ns is [Pe•+ d5NB•-]. Experimentally, the FDMR
signal from Pe•+ can be observed even for [Pe] ) 2.4 × 10-5
mol dm-3. At this concentration, the charge transfer from SQ•+
would take 50 µs. The FDMR signal linearly increases with
[Pe] (2 × 10-5 to 10-3 mol dm-3). This result implies the rapid
formation of Pe•+ competing with the decay of the squalane
hole. To summarize, our FDMR data support the View that
the mobile ions reacting with triethylamine and aromatic solutes
are short-liVed solVent radical cations of squalane.
(20) Hummel, A. AdV. Radiat. Chem. 1974, 4, 1.
(21) Warman, J. M.; Asmus, K.-D.; Schuler, R. H. J. Phys. Chem. 1969,
73, 931.
(22) Berlman, I. B. Handbook of Fluorescence Spectra of Aromatic
Molecules, 2nd ed.; Academic Press: New York, 1971.
(23) Werst, D. W. Chem. Phys. Lett. 1996, 251, 315.
(24) Lefkowitz, S. M.; Trifunac, A. D. J. Phys. Chem. 1984, 88, 77.
(25) Teather, G. G.; Klassen, N. V. J. Phys. Chem. 1981, 85, 3044.
(26) Muto, Y.; Nakato, Y.; Tsubomura, H. Chem. Phys. Lett. 1971, 15,
147.
Conclusion
Radiolysis of squalane yields a mobile solvent hole and
normally diffusing olefin ions. Only the latter can be observed
directly by FDMR. The mobile ion is scavenged by aromatic
solutes significantly faster than other cations. In squalane, the
fastest ET reactions of normally diffusing ions have the rate
constants ∼(1.6-1.8) × 109 mol-1 dm3 s-1. Most aromatic
(27) Lias, S. G.; Bartmess, J. E.; Liebman, J. F.; Holmes, J. L.; Levin,
R. D.; Mallard, W. G. J. Phys. Chem. Ref. Data 1988, 17, 1.
JP9532475