Investigation of non-Rehm-Weller kinetics in the electron transfer from trivalent phosphorus compounds to singlet excited sensitizers

Article abstract of DOI:10.1002/poc.3170

Singlet excited states (¹S* and ¹S ^+*) of neutral and monocationic sensitizers, S and S ⁺, respectively, were quenched by electron transfer (ET) from a variety of trivalent phosphorus compounds (Z₃P). The quenching rate constants k_q, which are equal to the rate constants k_ET of the ET from Z₃P to ¹S^* or ¹S^+*, were determined by the Stern-Volmer method. The logarithm of k_ET was plotted against free-energy change ΔG⁰ of the ET. The plot deviated upward from the line predicted by the Rehm-Weller (RW) theory in the endothermic region, the deviation being larger in the ET to a neutral acceptor ¹S^* than in the ET to a cationic acceptor ¹S^+*. Such a kinetic behavior is in sharp contrast to that observed in the ET from amines (R ₃N), where the ET to either neutral or cationic acceptor takes place according to the RW prediction. The ET from a donor, Z₃P or R ₃N, to a neutral acceptor ¹S^* is a charge-separation type, during which electrostatic attraction between the donor and the acceptor is generated, whereas the ET to a cationic acceptor ¹S^+* is a charge-shift type, which results in neither electrostatic attraction nor repulsion. Difference in kinetics-energetics relationship by the type of ET, which is not recognized in the ET from R₃N donor, becomes "visible" when Z ₃P is used as a donor. Copyright 2013 John Wiley & Sons, Ltd. The rate constants k_ET of electron transfer from trivalent phosphorus compounds to singlet photoexcited sensitizers were determined by the Stern-Volmer method. Logk_ET-ΔG⁰ plots were found to deviate upward from the line predicted by the Rehm-Weller theory, with deviation being larger in ET to neutral acceptors than in ET to cationic acceptors. Copyright

Full text of DOI:10.1002/poc.3170

Special Issue Article
Received: 12 April 2013,
Revised: 1 June 2013,
Accepted: 11 June 2013,
Published online in Wiley Online Library: 22 July 2013
(wileyonlinelibrary.com) DOI: 10.1002/poc.3170
Investigation of non-Rehm–Weller kinetics
in the electron transfer from trivalent
phosphorus compounds to singlet
†
excited sensitizers
a
a
Shinro Yasui * and Munekazu Tsujimoto
1
*
1
+*
+
Singlet excited states ( S and S ) of neutral and monocationic sensitizers, S and S , respectively, were quenched by
q
electron transfer (ET) from a variety of trivalent phosphorus compounds (Z P). The quenching rate constants k ,
3
1
*
1 +*
which are equal to the rate constants kET of the ET from Z P to S or S , were determined by the Stern–Volmer
method. The logarithm of kET was plotted against free-energy change ΔG of the ET. The plot deviated upward from
the line predicted by the Rehm–Weller (RW) theory in the endothermic region, the deviation being larger in the ET to
a neutral acceptor S than in the ET to a cationic acceptor S . Such a kinetic behavior is in sharp contrast to that
observed in the ET from amines (R N), where the ET to either neutral or cationic acceptor takes place according to
the RW prediction. The ET from a donor, Z P or R N, to a neutral acceptor S is a charge-separation type, during
3
0
1
*
1 +*
3
1
*
3
3
which electrostatic attraction between the donor and the acceptor is generated, whereas the ET to a cationic
1
+*
acceptor S is a charge-shift type, which results in neither electrostatic attraction nor repulsion. Difference in
kinetics–energetics relationship by the type of ET, which is not recognized in the ET from R N donor, becomes
3
“
3
visible” when Z P is used as a donor. Copyright © 2013 John Wiley & Sons, Ltd.
Supporting information may be found in the online version of this paper.
Keywords: trivalent phosphorus; electron transfer; Rehm–Weller theory; Stern–Volmer analysis; charge separation; charge shift
in ET from Z P, but there are in fact several reports on amine
INTRODUCTION
3
0
[9–17]
donors R N giving deviating logk –ΔG plots.
Such a kinetic
3
ET
Electron transfer reactions (ET) are one of fundamental chemical
reactions, which occur ubiquitously in inorganic and organic
chemistry as well as in biological systems. Mechanistic aspects of
ET have long been studied by both experimental and theoretical
chemists to formulate dependency of ET rate constant (kET) on
behavior has been interpreted to result from a highly exothermic
reaction following the ET step. High exothermicity is supplied by
[
12–15]
follow-up reactions such as cleavage of a covalent bond,
[
16]
[17]
formation of a covalent bond, and second ET. It should be
emphasized that these follow-up reactions are carried out by
0
•+
free-energy change (ΔG ; opposite sign of driving force) of the
reduced acceptors but not by amine radical cations R
from the ET. As for our ET from Z
3
N
resulting
[
1]
[2]
3
+
P, the ET step generates trivalent
phosphorus radical cation Z P , which subsequently undergoes
3
ET. The Rehm–Weller (RW) and the Marcus theory are definitely
successful examples.
•
an ionic reaction with a nucleophile such as water and alcohol in
the solvent, to eventually afford the pentavalent oxo-compound
Studies on ET kinetics have been developed mainly using
compounds with second row elements such as amines (R N) or
3
[
3–5]
18
Z
3
P = O.
Use of O labeled water has clearly shown ionic
alkoxybenzenes (ArOR) as
a donor. Third row element
•+
[18]
reaction occurring between Ar
3
P
and water.
P are exothermic enough to make the logkET–ΔG plots
deviate upward.
The ionic reac-
counterparts of these donors, namely, phosphines and sulfides,
respectively, also undergo ET under certain conditions. Never-
theless, kinetics of the ET from these compounds have been
studied less often. There seems to be a naive belief that ET from
compounds with third row elements occurs in a similar fashion
as the ET from compounds with second row elements. Third
row elements have d-orbitals so that the compounds with these
elements might behave differently in the ET.
•
+
0
tions of Z
3
*
Correspondence to: Shinro Yasui, Institute of Human and Environmental
Sciences, Tezukayama University, Gakuen-Minami, Nara, 631-8585, Japan.
E-mail: yasui@tezukayama-u.ac.jp
We have found that a variety of trivalent phosphorus
†
This article is published in Journal of Physical Organic Chemistry as a Special
issue: In Memoriam, edited by Michael Page (University of Huddersfield,
Chemistry, Queensgate, Huddersfield, HD1 3DH, United Kingdom).
compounds (Z P; Z = alkyl, aryl, OR; R = aryl, alkyl) undergo ET to
3
[
3–8]
various types of electron-deficient compounds in the dark.
Analyzing the ET processes kinetically mainly based on UV–vis
This manuscript is dedicated to the memory of Professor Rory More O'Ferrall.
spectroscopy, we have found that ET from Z P examined always
3
0
give logk –ΔG plots deviating upward from the line predicted
by the RW theory in the endothermic region (ΔG > 0). Plots
ET
a S. Yasui, M. Tsujimoto
Institute of Human and Environmental Sciences, Tezukayama University,
Gakuen-Minami, Nara, 631-8585, Japan
0
deviating from the RW prediction are not events observed only
J. Phys. Org. Chem. 2013, 26 1090–1097
Copyright © 2013 John Wiley & Sons, Ltd.

ELECTRON TRANSFER FROM TRIVALENT PHOSPHORUS COMPOUNDS TO EXCITED SENSITIZERS
Me
The Stern–Volmer (SV) method is widely used to collect kinetic
data of ET processes. Quenching constant k , which can be
Me
P
P
PPh2
q
Ph(3-n)P(OR)n
3
X
3
X
obtained from the slope of the SV plot together with the lifetime
of the excited sensitizer, is identical to ET rate constant kET
provided that the quenching occurs through ET mechanism.
Previously, we performed the SV analysis on the ET quenching
Me
n = R =
1
a
X =
b; p-MeO
X =
1l;
1
Me
1
1g; o-Me
1h; p-Me
1m; 1 Et
1
n;
1o;
p;
1q;
2
2
3
3
3
Me
Et
Me
Et i
Pr
1c; o-Me
1d; m-Me
1e; p-Me
+
1
+*
of rhodamine 6G (Rho ) in the singlet excited state, Rho , by
1
[
19]
Z
k
3
P in aqueous acetonitrile.
To investigate dependency of
1f; H
0
0
ET on ΔG over a wide range of ΔG , we selected a variety of
P (1) whose oxidation potentials spread in a range of more
Bu3P
1r;
1i; p-F
Z
3
1j; p-Cl
1k
ꢀ
1
0
than 1.5 eV (145 kJmol ). Then, we found that logkET–ΔG plot in
the endothermic region deviates upward from the RW prediction.
Recently, we performed the SV analysis on the ET quenching
of several neutral sensitizers, namely, 9,10-dicyanoanthracene
1
R
2
R
N
R3
(
OR)n
a; R1 = R2 = R = Ph
3
2
2
2
(OR)n
(
DCA) and 9-cyanoanthracene (CA), in the singlet excited states,
1
2
3
b; R = Ph, R = R = Me
1
1
[20,21]
3a; 1,2-(MeO)2
b; 1,3-(MeO)2
c; 1,4-(MeO)2
3d; 1,2,4-(MeO)3
3e; 1-MeO, 4-Me
DCA* and CA*, respectively, by Z
3
P in acetonitrile.
The ET
1
2
3
c; R = Ph, R = R = H
3
3
1
1
0
1
2
3
to DCA* or CA* exhibited upward deviation in the logkET–ΔG
2d; R = R = R = Et
1
2
3
plot as well, but degree of the deviation was larger than the
2e; R = R = R = n-Bu
1
+*
f; R = R = R3 = Hex
1
2
deviation observed in the ET to a cationic acceptor, Rho . That
2
2
2
0
1
2
3
3f; 1,4-(C H15O)2
g; R = R = Et, R = H
7
is, logkET–ΔG plots in the ET from Z
3
P separate depending on an
1
2
3
h; R = R = n-Bu, R = H
acceptor. The upward deviation can be interpreted primarily by
1
2
3
•
+
2i; R = R = cyclo-Hex, R = H
assuming that the resulting cation radical Z3P reacts with water
in the solvent highly exothermically. However, if the upward
deviation of the plot results solely from the follow-up reaction
Chart 1. Quenchers used in this work
•+
of Z P , the degree of the deviation should be the same
3
irrespective of an acceptor.
Charge distribution changes during the ET in different ways
depending on whether the acceptor is cationic or neutral. ET
+
EtHN
Me
O
NHEt
+
Me
N
from a neutral donor such as Z P to a neutral acceptor, where
3
CO Et
2
Me
a neutral–neutral pair becomes a cation–anion pair, generates
electrostatic attraction between the donor and the acceptor.
On the other hand, ET from a neutral donor to a cationic
acceptor, where a cationic center moves from the acceptor to
the neutral donor, results in neither electrostatic attraction nor
+
Ac
Rho+
CN
O
repulsion. This difference in the type of ET may be responsible
0
for the separate logk –ΔG plots. In sharp contrast to the ET
N
ET
from Z
3
P, we found that ET from a donor with second row
N or ArOR to both a cationic acceptor,
CN
CA
CN
Me
element such as R
3
DCA
MA
1
+* [19]
1
1
Rho , and a neutral acceptor, DCA* or CA*, gives a single
0
[20]
logkET–ΔG plot that traces well the RW prediction.
shows that the separate plots appearing in the ET from Z
originates from a property of phosphorus atom in Z P.
We here discuss the origin of separate logkET–ΔG plots
observed in the ET from Z P considering change of charge
distribution within the redox pair during the ET. At some stage of
the discussion, results from theoretical computations will be
consulted. The quenchers examined in this study, trivalent phospho-
This fact
Chart 2. Sensitizers used in this work
3
P
3
0
Materials
3
Sensitizers (Tokyo Chemical Industry Co.) and trivalent phosphorus
compounds 1 (Tokyo Chemical Industry Co. or Aldrich) were commer-
cially available and purified before use if necessary. In the SV analyses,
acetonitrile of fluorescence analysis grade (Nacalai Tesque) was used
without purification.
rus compounds Z
3 3
P (1), amines R N (2), and alkoxybenzenes ArOR
(
3), are listed in Chart 1. Cationic sensitizer, 10-methylacridinium salt
+
(
Ac ), and a neutral sensitizer, 10-methylacridon (MA), were also
subjected to the examination in this study to obtain additional data.
Sensitizers used in this work are given in Chart 2.
Electrochemical measurements
–3
The solution of a sample (5.0 × 10 M) and tetraethylammonium
tetrafluoroborate (0.10 M) as a supporting electrolyte in acetonitrile was
subjected to RDE measurement with a rotating disk platinum electrode
+
(
1000 rpm) as a working electrode and Ag/Ag (in a solution of silver
EXPERIMENTAL
nitrate (0.01 M) and tetraethylammonium perchlorate (0.1 M) in acetonitrile)
as a reference electrode. Values of half-wave potentials E1/2 were read on the
voltammograms obtained.
Instruments
UV–visible spectra were recorded on a Shimadzu UV-2200A spectropho-
tometer. Fluorescence from excited sensitizers was monitored on a
Shimadzu RF-5000 spectrofluorophotometer. Rotating disk electrode
SV analysis
(
RDE) voltammetry was carried out with a BAS RDE-1 using an ALS
q
The quenching rate constants k were determined according to the
procedure reported previously.
[19]
electrochemical analyzer Model 620A.
J. Phys. Org. Chem. 2013, 26 1090–1097
Copyright © 2013 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/poc

S. YASUI AND M. TSUJIMOTO
(
)
Computation
and S , showing that the SV equation (2) holds in quenching
1
( ) *
of S
by 1. Representative examples are given in Fig. 2.
Density functional theory (DFT) were performed with Gaussian 09 package
at the B3LYP level of theory with a pertinent basis set such as 6-31G(d),
[
22]
I0=I ¼ 1 þ kqτ0½1ꢁ
(2)
6-31 + G(d), or 6-311 + G(d).
Molecular structures were visualized
using GaussView 5.0 (Gaussian, Inc.).
The lifetime τ
0
of the fluorescence of each sensitizer is
available from literature as given in Table 1, with which the
RESULTS
quenching rate constant k was determined based on Eqn (2)
q
(
)
for each combination of 1a–r and S . Table 2 reports the kq
values determined here under the term of ET rate constant kET,
since k is equal to k in this system (vide infra).
Electrochemical measurements
Electrochemical measurements on trivalent phosphorus com-
pounds 1a–r were performed with a RDE in acetonitrile at
q
ET
+
2
5 °C using Ag/Ag as a reference electrode. The half-wave
Mechanism of the fluorescence quenching
1
( )
potentials E1/2( S *) of the sensitizers in the singlet photoex-
cited state
according to Eqn (1).
[
19]
1
(
)
As has been shown previously,
trivalent phosphorus compounds 1 with Rho results in quantita-
tive formation of the corresponding oxo-compound Z P = O (2).
This observation shows the existence of trivalent phosphorus rad-
steady-state photolysis of
S
* (where ( ) = 0 or +) were calculated
+
3
E1=2ð1Sð Þ*Þ ¼ E1=2ðSð Þ0Þ þ ΔE0;0ðSð ÞÞ
(1)
•+
•+
ical cation 1 along the reaction pathway. The radical cation 1
(
)
once formed undergoes easily the ionic reaction with a small
where E1/2(S ) is a half-wave potential of a sensitizer in the ground
0
(
)
amount of water contained in the solvent to eventually give 2.
state, and ΔE0,0(S ) is the zero–zero photoexcitation energy. The
values of E1/2(S ) are available from literature. ΔE (S ) values
were estimated by the absorption maximum of S and the
emission maximum of
summarized in Table 1.
+
1
+*
(
0
)
( )
Thus, the singlet excited state of Rho , Rho , is quenched
0,0
(
•+
)
through the ET from 1 to generate 1 . Meanwhile, considerations
+
1
(
)
of the energy levels of Rho exclude the possibility of singlet-
S *. Electrochemical parameters are
1
+*
1
+*
singlet energy transfer from Rho to 1 to quench Rho . We
have also shown that the ET takes place from several types of phos-
+
[29–31]
phines (Z P; Z = alkyl, aryl) to Ac under the photo-irradiation.
3
SV analysis
The ET from triarylphosphines (Z P; Z = aryl) to the singlet
3
(
)
1
*
[32]
When the solution of S in acetonitrile was excited at the
wavelength of the absorption maximum of each sensitizer, the
characteristic fluorescence was observed. The wavelengths of
excitation and emission maxima of the fluorescence are given
also in Table 1. The intensity of the fluorescence from each sen-
sitizer decreased upon the addition of 1, indicating that S
quenched by 1. Examples are shown in Fig. 1. Ratio I /I, where I0
and I are the intensity of the fluorescence from a sensitizer
observed in the absence and presence of 1, respectively, was
plotted against the concentration of 1. A linear line with the
intercept being unity was obtained for each combination of 1
photoexcite state of neutral acceptor DCA is well known. In
fact, 355 nm-laser flash photolysis on the solution of
triarylphosphines Ar P in the presence of DCA and biphenyl gave
3
a transient UV absorption of the corresponding radical cation
•
+ [32]
Ar P .
The close similarity in structure between DCA and CA
3
1
( ) *
1
*
is
predicts that CA is also quenched by 1 through the ET
mechanism. MA in the singlet excited state, MA , is also a good
one-electron acceptor.
the sensitizers S
1
*
0
[
27,33]
In conclusion, singlet excited states of
1
( ) *
examined here are quenched through the ET
from 1. That is, the quenching constants k obtained by the SV
q
method represent the ET rate constants kET.
Table 1. Photochemical and electrochemical properties of sensitizers
(
)
a
+
1
()
b
+
Sensitizer
Absorption
Emission maximum of the
τ
0
/ns
E1/2(S
0
) /V vs Ag/Ag
E
1/2( S *) /V vs Ag/Ag
maximum/nm
fluorescence/nm
+
c
d
Rho
Ac
525
358
375
380
545
482
430
438
410
3.0
– 1.10e
1.22
2.22
1.69
1.10
0.84
+
e
32.9e
– 0.73
f
DCA
CA
12.7
– 1.23
e
g
11.5i
– 1.95j
h
MA
397
6.1
– 2.23
a
+
Half-wave potential of the sensitizer in the ground state (S ). Adjusted to the values against Ag/Ag .
0
b
Half-wave potential of the sensitizer in the singlet excited state. Calculated by Eqn (1).
Ref. 23.
c
d
e
Ref. 19.
Ref. 24.
Ref. 18.
f
g
h
i
Ref. 25.
Ref. 26.
Ref. 27. Values measured under deaerated conditions.
j
Measured by RED in this work.
wileyonlinelibrary.com/journal/poc
Copyright © 2013 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2013, 26 1090–1097

ELECTRON TRANSFER FROM TRIVALENT PHOSPHORUS COMPOUNDS TO EXCITED SENSITIZERS
ꢀ
7
ꢀ3
Figure 1. Decrease in intensity of the emission from a singlet excited sensitizer upon the addition of 1. (A): [DCA] = 1.6 × 10 M, [1q] = 0, 1.51 × 10
,
ꢀ3
ꢀ3
ꢀ7
ꢀ3,
ꢀ3
ꢀ3
ꢀ2
ꢀ2
ꢀ7
4
5
.53 × 10 , 8.84 × 10 M. (B): [CA] = 3.4 × 10 M. [1f] = 1.99 × 10 3.98 × 10 , 7.95 × 10 , 1.19 × 10 , 1.99 × 10 . (C): [MA] = 4.2 × 10 M, [1b] = 0,
.36 × 10 , 1.04 × 10 , 1.44 × 10 M. A sharp emission band with the short wavelength in each spectrum is an artificial one resulting from the
ꢀ3
ꢀ2
ꢀ2
excitation
0
ArOR (3) to the acceptors examined here gives logkET–ΔG plot
that traces the RW prediction.
Theoretical computations
To test a possibility of adduct formation between 1 and an
acceptor during the ET, theoretical calculations were applied.
Thus, gas-phase potential energy surface of an adduct 4a formed
between 1f and DCA (Chart 3) was computed as a function of P–O
length at the DFT B3LYP/6-311+ G(d) level of theory. Similar
computation was made on a model adduct 4b between MA and
PH instead of 1f for the sake of economy of computation at the
3
DFT B3LYP/6-31 + G(d). Energy profiles of 4a and 4b are given in
Figs. 4a and 4b, respectively.
The optimized structures of Ph P , Bu P , and (MeO) P at
•
+
•+
•+
3
3
3
Figure 2. Stern–Volmer plots for quenching of MA (□), CA (♦), DCA (Δ),
and Ac I (●) by 1e
+
-
the B3LYP/6-31G(d) showed that a positive “hole” is located
almost exclusively on the phosphorus atom in these radical
cations (Figure S1).
In addition, the UV–visible spectra of the sensitizers examined
here did not change upon the addition of a large excess
amount of 1 either in the dark or under the irradiation,
suggesting that no chemical reaction takes place between 1
and the sensitizers.
DISCUSSION
1
()*
Logarithm of kET in the ET from Z
3
P (1) to S is plotted against
0
ΔG in Fig. 3. Various types of Z P, namely, phosphines,
3
phosphinites, phosphonites, and phosphites, construct a single
line with respect to ET to each particular acceptor. This fact indi-
cates that the ET from 1 takes place according to a common
mechanism. In fact, theoretical calculations on radical cations
0
k –ΔG profile
ET
0
The free-energy change ΔG of ET step for each combination of
1
( ) *
donor (1) and acceptor ( S ) is given by Eqn (3).
from different types of
3
Z P, namely, triarylphosphines,
trialkylphosphines, or trialkyl phosphites, show a positive charge
located almost exclusively on the phosphorus atom. In other
words, Z P (1a–r) examined in this work can be taken as a series
3
ΔG ¼ E1=2ð1Þ–E1=2ð Sð Þ*Þ
0
1
(3)
of compounds despite of divergent structures.
0
The logarithm of kET0was plotted against ΔG in Fig. 3. In an
exothermic region (ΔG < 0), logk is constant with k being
An important finding in Fig. 3 is that the plots deviate upward
from the line predicted by the RW theory in the endothermic
ET
–1 –1
ET
10
0
diffusion limited (≈ 2 × 10
M
s ). No Marcus’ “inverted
region (ΔG > 0). The conventional Marcus equation also fails
region” is observed even in a highly exothermic region (up to
to reproduce the observed plot with any values of parameters.
0
0
0
ΔG = –1.8 eV). In the endothermic region (ΔG > 0), logk drops
There have been several reports in which logk –ΔG plots
ET
ET
0
1 ( ) *
as ΔG increases. Importantly, the plot for the ET from 1 to S
deviate upward from the RW prediction. This behavior of kinetics
is brought about when radical species generated during the ET
step undergo highly exothermic follow-up reactions such as
in the endothermic region deviates upward from the prediction
by the RW equation. On the other hand, the ET from R N (2) and
3
J. Phys. Org. Chem. 2013, 26 1090–1097
Copyright © 2013 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/poc

S. YASUI AND M. TSUJIMOTO
Table 2. Quenching of the singlet excited sensitizers by 1
+
a
0
9
–1 –1 b
Sensitizer
Z
3
P (1)
E
1/2(1)/V vs Ag/Ag
ΔG /eV
k
ET/10 M s
+
c
Rho
o-Tol P (1c)
0.88
1.03
1.19
1.19
1.10e
1.21
1.28e
1.49
1.49
1.87
1.83
1.71
0.39
0.88
1.19
1.32
0.39
0.88
1.32
0.39
0.65
0.88
1.03
1.19
1.32
1.28
1.49
1.87
1.83
0.39
0.65
0.88
1.02
1.03
1.19
1.14
1.11
1.21
1.32
1.28
1.49
1.87
1.83
1.71
0.65
1.03
1.19
1.19
1.14
1.28
1.83
1.71
– 0.37
– 0.19
– 0.03
– 0.03
– 0.12
– 0.01
0.06
0.27
0.27
0.65
0.61
5.39
8.37
7.78
8.35
7.65
5.12
5.01
1.26
2.39
3
c
p-Tol
3
P (1e)
c
Ph P (1f)
3
Ph P (1f)
3
c
c
c
c
c
Bu P (1k)
3
Ph
Ph
2
2
POMe (1l)
POEt (1m)
PhP(OMe) (1n)
2
PhP(OEt) (1o)
2
c
c
(
(
(
MeO)
EtO) P (1q)
Pr O) P (1r)
3
P (1p)
0.0762
0.0456
0.159
3
i
c
0.49
3
+
+
f
Ac
Mes P (1a)
– 1.83
– 1.34
– 1.03
– 0.90
– 1.83
– 1.34
– 0.90
– 1.30
– 1.04
– 0.81
– 0.66
– 0.50
– 0.37
– 0.41
– 0.20
ꢀ0.18
0.14
– 0.71
– 0.45
– 0.22
– 0.08
– 0.07
0.09
0.04
0.01
0.11
0.22
0.18
0.39
0.77
0.73
0.61
– 0.19
0.19
0.35
0.35
0.30
0.44
0.99
0.87
19.1
3
o-Tol
Ph P (1f)
p-Cl-Ph P (1j)
3
P (1c)
19.4
21.0
13.2
18.6
17.9
18.3
19.6
16.2
16.3
16.5
15.3
13.7
13.7
3
3
g
Ac
Mes
o-Tol
3
P (1a)
P (1c)
3
p-Cl-Ph P (1j)
3
DCA
Mes P (1a)
3
p-An
o-Tol
3
P (1b)
P (1c)
3
p-Tol P (1e)
3
Ph P (1f)
3
p-Cl-Ph
Ph POEt (1m)
PhP(OEt) (1o)
3
P (1j)
2
8.26
6.08
6.10
20.0
12.7
9.97
2
(
MeO) P (1p)
3
(
EtO)
3
P (1q)
P (1a)
p-An P (1b)
CA
Mes
3
3
o-Tol
m-Tol
p-Tol
3
P (1c)
3
P (1d)
P (1e)
11.4
10.1
10.2
10.4
10.3
12.3
3
Ph P (1f)
3
Ph
Ph
2
2
(o-Tol)P (1g)
(p-Tol)P (1h)
p-F-Ph P (1i)
3
p-Cl-Ph P (1j)
6.83
3
Ph
PhP(OEt)
2
POEt (1m)
(1o)
9.03
5.04
0.180
0.666
1.45
4.75
4.17
1.41
2
(
(
(
MeO) P (1p)
3
EtO) P (1q)
3
i
Pr O)
3
P (1r)
P (1b)
MA
p-An
3
p-Tol P (1e)
3
Ph P (1f)
3
d
Ph
Ph
3
2
P (1f)
(o-Tol)P (1g)
1.54
1.49
Ph POEt (1m)
0.933
0.0511
0.247
2
(
(
EtO)
Pr O)
3
P (1q)
P (1r)
i
3
a
Half-wave potentials of 1. Measured by RDE.
Determined by the Stern–Volmer method in acetonitrile under the aerobic conditions unless otherwise noted.
b
c
Data from Ref. 24.
d
Determined under an argon atmosphere.
e
f
ox
ox
[28]
Based on the peak oxidation potentials E
p
determined by cyclic voltammetry; calculated by assuming E1/2 = E
p
– 0.03.
Iodide salts.
g
Tetrafluoroborate salts.
wileyonlinelibrary.com/journal/poc
Copyright © 2013 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2013, 26 1090–1097

ELECTRON TRANSFER FROM TRIVALENT PHOSPHORUS COMPOUNDS TO EXCITED SENSITIZERS
o
1
+*
1
+*
1
+*
Figure 3. Dependency of logkET on free-energy change ΔG in the ET from a series of 1 to Rho (○), Ac (iodide salt) (□), Ac (tetrafluoroborate salt) (+),
1
*
1
*
1
*
1
+*
1
*
DCA (▼), CA (Δ), and MA (♦). Symbols X denote the points for ET from 2 and 3 to Rho and CA altogether. A solid line represents the prediction by
the RW theory
+
+
We studied kinetics of the ET from tributylphosphine (1k) to
“rigid” and “flexible” viologens in the dark in acetonitrile
PPh
-
PH3
NC
3
O
[
35]
0
containing methanol.
We found in the ET that logk –ΔG
ET
plots deviate upward from the RW prediction and the deviation
is larger in the former than in the latter. To explain the deviating
behavior in logk –ΔG plots, we considered the reaction of Z P
-
N
0
ET 3
•+
CN
Me
with a nucleophile quantitatively. Thus, we proposed a modified
Marcus equation (Eqn (4)) that includes free energy B gained by
P–O bond formation between the resulting radical cation 1k
4
a
4b
Chart 3. Adducts between a sensitizer and Ph
3
P or PH
3
•+
and methanol. The value of B was calculated to be ca. 2 eV based
on a thermochemical cycle, with which Eqn (4) well accommo-
dated the deviating plots by taking 3.2–4.5 eV of reorganization
[
12–15]
[16]
[17]
cleavage
or formation of a covalent bond or second ET.
To date, we have examined kinetically ET from various types of Z P
to various types of acceptors to find that any logk –ΔG plot for ET
3
[36]
0
energy λ.
ET
from Z P deviates upward from the theory in the endothermic
"
ꢀ
ꢁ #
3
2
0
[
6–8,19]
•+
ΔG þ λ þ B
region.
Trivalent phosphorus radical cation Z3P generated
kET ¼ Zexp ꢀ
(4)
during the ET step reacts with a nucleophile such as water or
alcohol such exothermically that the plot deviates upward. DFT
4kBTλ
•
+
3
calculations with B3LYP/6-31G(d) predict that the reaction Z P
–
•
ꢀ1
ꢀ1
+
OH → Z
3
P -OH is 818.9kJmol and 786.2kJmol exothermic
However, this treatment cannot explain the separate plots in
Fig. 3. While B value should be independent of charge on the
acceptor in principle, whether the acceptor is neutral or cationic
makes logk –ΔG relationship in the endothermic region
separate into two lines. The ET from 1 to a neutral acceptor,
when Z = Bu and Ph, respectively. Importantly, the lengths of P–O
•
bonds in optimized structures of Z
3
P -OH are 1.81 Å and 1.70 Å,
0
when Z = Bu and Ph, respectively, which are well within an
ET
[
34]
expected value for a P–O single covalent bond.
Figure 4. Energies of adducts computed for 4a at the DFT B3LYP/6-311 + G(d) (A) and for 4b at the DFT B3LYP/6-31 + G(d) (B) as a function of the P–C
distance designated in the structures in Chart 3
J. Phys. Org. Chem. 2013, 26 1090–1097
Copyright © 2013 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/poc

S. YASUI AND M. TSUJIMOTO
1
* 1
*
1
*
CA , DCA , or MA , is charge-separation (CS) type, whereas the
1
+*
ET from 1 to a cationic acceptor, Rho , is charge-shift (CSh)
type (Scheme 1). In the former ET, electrostatic attraction within
the redox pair is being generated, which likely lowers the energy
of the final stage in the ET step to contribute to the acceleration
of the ET rate. In CSh type-ET, neither electrostatic attraction nor
repulsion occurs during the ET, and this kind of acceleration is
Scheme 2. Another type of charge shift ET, during which a neutral-cation
pair becomes a cation-cation pair.
0
logkET–ΔG plots, even though electrostatic repulsion is generated
[6,7]
during the ET (another CSh-type ET; Scheme 2). The observation
[37]
•+
not operative.
3
suggests that a reaction of Z P with a nucleophile contributes to
0
Multiple RW plots have been reported as to the ET quenching
upward deviation of logkET–ΔG plots more significantly than
another factor, namely, electrostatic attraction within the resulting
radical ion pair.
1
of DCA* by n-donors such as aliphatic amines and by π-donors
[
38–42]
such as aromatic hydrocarbons.
The quenching by the
0
former donors gives logkET–ΔG plot horizontally shifted to
the positive direction relative to the plot for the quenching by
the latter donors, which has been interpreted to result from a
For the sake of comparison, we also performed the SV analysis
1
*
1
+*
3
on ET from R N (2) and ArOR (3) to CA and Rho under the
otherwise identical conditions to those for the ET from 1. The
ET from 2 and 3 to either neutral or cationic acceptors takes
1
more efficient interaction of the former donors with DCA* than
with the latter ones in the ground state.
[
41,42]
[19,20]
Farid and Gould
place with approximately obeying the RW equation (Fig. 3)
suggested more drastic interaction between a donor and an
acceptor. They examined CS-type ET to neutral acceptors, singlet
excited aromatic cyanides (including CA and DCA ), and
For the RW equation to be held, the pair of the oxidized donor
[
1]
and the reduced acceptor must disappear rapidly. Usually,
back ET occurring from the latter to the former in the ground
state contributes to the rapid disappearance of the pair. This is
the case for the ET from 2 and 3. Back ET from the reduced
1
*
1
*
CSh-type ET to cationic acceptors, excited pyrylium salts,
[
43]
0
occurring in endothermic region.
In these ETs, logkET–ΔG
•+
•+
plots deviate upward from the RW prediction, with the
degree of the deviation being larger in the former ET than
in the latter ET. To explain their observation, they proposed
acceptor to 2 or 3 is rapid enough to make the ET kinetics
[
1,44]
follows the RW prediction.
In turn, the consideration here
•
+
highlights peculiar reactivity of the radical cation 1 . Radical
•
+
“
a bonded exciplex mechanism”, which postulates the forma-
cation 1 may have slightly longer lifetime allowing it to interact
with the reduced acceptor in the ground state. This is quite
understandable because the radical cations resulting from 2 or
tion of a covalent bond during the exciplex stage. They
claimed that pyridine donor effectively forms a “bonded
exciplex” with aromatic cyanides and pyrylium salts, making
the ET rate higher than the RW prediction. Their observation,
•+
•+
3, namely, R N or ArOR , respectively, cannot undergo an ionic
3
reaction with a nucleophile which requires expansion of the
covalency. No further reaction of the reduced acceptor such as
based on which they have presented their model, is similar
0
•–
•
to ours in a sense that logk –ΔG plots deviate upward in the en-
CA and Rho faster than the back ET is possible, either.
ET
dothermic region, and the deviation is larger in CS-type ET than in
CSh-type ET. However, they interpreted the origin of the separate
plots simply as a result of retardation of the ET to pyrilium cation
by steric hindrance of this acceptor without taking into consider-
ation difference in the change of charge distribution during ET.
We performed DFT B3LYP/6-311 + G(d) computation to
evaluate energy which would be generated by forming an
adduct 4a between 1f and DCA as a function of the distance
between the phosphorus atom in 1f and the 9C carbon in
Another issue in this study is to figure out why the Marcus
inverted region is absent. The absence of the Marcus inverted
region is often met when reorganization energy λ is large
enough, namely, when ET occurs at a long distance because λ
is a function of a distance between a donor and an acceptor.
At this stage of our research, we cannot tell whether the distance
1
()*
between 1 and S is long enough or not during the ET. We will
discuss this point elsewhere in near future.
3
DCA. Energy of a model adduct 4b between PH and MA was
also computed as a function of P–9C distance. The results are
shown in Figs. 4a and b, respectively. There is no energy gain
by forming an adduct, and, furthermore, unlike in Gould's
calculations on ET from pyridine, energy change of each adduct
in our system, whether a real one or a model, hardly shows an
appreciable inflection point. The results rule out a possibility that
DCA or MA forms a covalent adduct with 1 in the ground state.
In other words, interaction between 1 and the sensitizer is not
CONCLUSION
We examined kinetics of ET from a variety of trivalent phosphorus
3
compounds Z P to neutral and monocationic sensitizers in the sin-
1
*
1 + *
glet excited states, S and S , respectively. The rate constant kET
was determined based on the SV method, and logkET was plotted
0
0
against free-energy change ΔG of the ET step. LogkET–ΔG plots in
the endothermic region deviate upward from the prediction by
the RW theory, which is interpreted to result from facile reaction
of the resulting radical cation Z
water contained in the solvent. The degree of the deviation is
3
larger in the ET from Z P to S than in the ET to S . In the former
ET, the radical ion pair generated by the ET is stabilized through
electrostatic attraction, which contributes to further acceleration
0
the origin of the upward deviation of logkET–ΔG plot.
•
+
•+
In conclusion, here, the reaction of Z
3
P
with a nucleophile is a
3
P
with a nucleophile such as
0
factor to make logkET–ΔG plots deviate upward from the RW predic-
tion. In addition to this factor, electrostatic interaction being gene-
rated with the ET proceeding contributes to acceleration of the ET
1
*
1 + *
rate. Interestingly, we have found that ET from Z P to iron(III)
3
complexes occurring thermally also exhibits upward deviation of
of the ET rate. On the other hand, nitrogen counterparts of Z P,
3
namely, amines R N, as well as alkoxybenzenes ArOR, obey the
3
1
*
1 + *
RW theory in the ET to both S and S . Kinetic behavior of R3N
and ArOR different from that of Z P is understandable taking
3
•+
•+
into account that the resulting radical cations R3N and ArOR
cannot undergo nucleophilic reaction which requires expansion
of covalency.
Scheme 1. Charge separation ET (top) and charge shift ET (bottom).
wileyonlinelibrary.com/journal/poc
Copyright © 2013 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2013, 26 1090–1097

ELECTRON TRANSFER FROM TRIVALENT PHOSPHORUS COMPOUNDS TO EXCITED SENSITIZERS
[
[
25] T. Niwa, K. Kikuchi, N. Matsusita, M. Hayashi, T. Katagiri, Y. Takahashi,
T. Miyashi, J. Phys. Chem. 1993, 97, 11960.
26] The solution of MA was excited not at the absorption maximum
Acknowledgement
This work is financially supported by a Tezukayama Research
(
4
397 nm) but at 380 nm to avoid overlap of the fluorescence at
10 nm with artificial emission resulting from the excitation.
27] S. Fukuzumi, K. Ohkubo, J. Am. Chem. Soc. 2002, 124, 10270–10271.
Grant 2012.
[
[
ox
28] The value of the peak oxidation potential E
p
for reversible
REFERENCES
oxidation is larger by 0.03 V than the value of half-wave poten-
[
[
[
1] D. Rehm, A. Weller, Isr. J. Chem. 1970, 8, 259–271.
2] R. A. Marcus, J. Phys. Chem. 1968, 72, 891–899.
3] S. Yasui, M. Fujii, C. Kawano, Y. Nishimura, K. Shioji, A. Ohno, J. Chem.
Soc., Perkin Trans. 2 1994, 177–183.
4] S. Yasui, K. Shioji, M. Tsujimoto, A. Ohno, J. Chem. Soc., Perkin Tans. 2
tial E1/2 when one electron is transferred at 25°C. Since the
anodic oxidation of 1 exerts irreversible character, the differ-
ox
ence between E
p
and E1/2 would be larger than 0.03, meaning
that the values under discussion in this table are slightly larger
than true values of E1/2
[
.
1
999, 855–862.
5] S. Yasui, K. Shioji, M. Tsujimoto, A. Ohno, Heteroatom Chem. 2000,
1, 152–157
[29] S. Yasui, K. Shioji, M. Yoshihara, T. Maeshima, A. Ohno, Tetrahedron
[
Lett. 1992, 33, 7189–7192.
[30] S. Yasui, K. Shioji, A. Ohno, M. Yoshihara, J. Org. Chem. 1995, 60,
2099–2105.
1
[
[
6] S. Yasui, K. Itoh, M. Tsujimoto, A. Ohno, Chem. Lett. 1998, 27, 1019–1020.
7] S. Yasui, K. Itoh, M. Tsujimoto, A. Ohno, Bull. Chem. Soc. Jpn. 2002,
[31] Trivalent phosphorus compound Z
such as alkyl diphenylphosphinite (Ph
phenylphosphonite (PhP(OR) ) undergoes nucleophilic attack on
3
P having at least one P-O bond
7
5, 1311–1318.
8] S. Yasui, M. Tsujimoto, K. Shioji, A. Ohno, Chem. Ber. Recueil 1997,
30, 1699–1707.
9] G. B. Schuster, J. Am. Chem. Soc. 1979, 101, 5851–5853.
2
POR) or dialkyl
[
2
+
1
the 9-carbon of Ac in the dark to form a phosphonium adduct,
[
which subsequently decomposes in an Arbuzov fashion to eventu-
[
[
10] J.-M. Savéant, Acc. Chem. Res. 1993, 26, 455–461.
11] C. Costentin, M. Robert, J.-M. Savéant, J. Am. Chem. Soc. 2005, 127,
ally give phosphine oxide AcP(=O)Z
philic attack on the tetrafluoroborate salt of Ac took place with
triarylphosphines with electron-releasing substituents, 1b (p-An P)
2
(Z = Ph, OR). Likewise, nucleo-
+
12154–12155.
3
1
+*
[
[
[
12] M. S. Workentin, R. L. Donkers, J. Am. Chem. Soc. 1998, 120, 2664–2665.
13] R. L. Donkers, M. S. Workentin, J. Am. Chem. Soc. 2004, 126, 1688–1698.
14] S. Antonello, K. Daasbjerg, H. Jensen, F. Taddei, F. Maran, J. Am.
Chem. Soc. 2003, 125, 14905–14916.
15] S. Antonello, M. Crisma, F. Formaggio, A. Moretto, F. Taddei, C.
Toniolo, F. Maran, J. Am. Chem. Soc. 2002, 124, 11503–11513.
16] M. J. Thomas, C. S. Foote, Photochem. Photobiol. 1978, 27, 683–693.
17] E. Pelizzetti, E. Mentasti, E. Pramauro, Inorg. Chem. 1976, 15,
3
and 1e (p-Tol P), in competition with the ET quenching of Ac .
[32] S. Yasui, S. Tojo, T. Majima, J. Org. Chem. 2005, 70, 1276–1280.
[33] S. Fukuzumi, N. Satoh, T. Okamoto, K. Yasui, T. Suenobu, Y. Seko, M.
Fujitsuka, O. Ito, J. Am. Chem. Soc. 2001, 123, 7756–7766.
[34] D. G. Gilheany, in: The Chemistry of Organophosphorus Compounds
(Ed.: F. R. Hartley), Willy, New York, 1990, Vol. 1, Chap. 2, 14–15.
[35] S. Yasui, K. Itoh, A. Ohno, N. Tokitoh, Org. Biomol. Chem. 2006, 4,
2928–2931.
[
[
[
2
898–2900.
18] M. Nakamura, M. Miki, T. Majima, J. Chem. Soc., Perkin Trans. 2 2000,
447–1452.
[36] These values of λ are reasonable considering that the ET is accom-
panied by significant geometrical change of acceptor viologens. In
fact, a larger λ is adopted to the ET to flexible cyclic viologens than
to the ET to rigid cyclic viologens. The former might undergo geo-
metrical change more largely than the latter.
[
1
[
[
[
19] S. Yasui, M. Tsujimoto, K. Itoh, A. Ohno, J. Org. Chem. 2000, 65, 4715–4720.
20] S. Yasui, M. Tsujimoto, J. Phys. Org. Chem. 2008, 21, 541–543.
21] The Karl-Fischer analysis has shown that acetonitrile contains ca. 0.1
[37] The observed separate plots were fitted by curves calculated based
on Fig. 4 by assuming arbitrarily a single value of B for various types
%
water even after repeated distillation on calcium hydride. S. Yasui,
K. Shioji, A. Ohno, M. Yoshihara, Chem. Lett. 1993, 22, 1393–1396.
22] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb,
J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson,
H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino,
G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda,
J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai,
T. Vreven, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. Bearpark,
J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi,
J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar,
J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox,
J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann,
O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin,
K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J.
Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas, J. B. Foresman,
J. V. Ortiz, J. Cioslowski, D. J. Fox, Gaussian 09, Revision A.02, Gaussian,
Inc., Wallingford CT, 2009.
3
of Z P used in this work, which resulted in a larger λ in CSh-type ET
[
than in CS-type ET (see Figure S2). This may be unreasonable
considering movements that might occur in the redox pair and in
the surrounding solvent during the ET. A quantitative interpretation
of the kinetic data in the present work would be possible if Eq. 4 is
further modified by introducing a parameter with respect to
electrostatic interaction and/or repulsion in the redox pair.
[38] P. Jacques, E. Haselbach, A. Henseler, P. Suppen, J. Chem. Soc.
Faraday Trans. 1991, 87, 3811–3813.
[39] P. Jacques, D. Burget, J. Photochem. Photobiol. A: Chem. 1992, 68,
165–168.
[40] P. Jacques, X. Allonas, J. Photochem. Photobiol. A: Chem. 1994,
78, 1–5.
[41] T. N. Inada, K. Kikuchi, Y. Takahashi, H. Ikeda, T. Miyashi,
J. Photochem. Photobiol. A: Chem. 2000, 37, 93–97.
[42] A. I. Burshtein, K. L. Ivanov, Phys. Chem. Chem. Phys. 2002, 4,
4115–4125.
[23] E. E. Lessing, A. Von Jena, M. Reichert, Chem. Phys. Lett. 1975, 36,
5
17–522.
[43] R. S. Farid, I. R. Gould, J. Org. Chem. 2007, 72, 6970–6981.
[44] R. Ballardini, G. Varani, M. T. Indelli, F. Seandola, V. Balzani, J. Am.
Chem. Soc. 1987, 100, 7219–7223.
[
24] K. Kikuchi, C. Sato, M. Watabe, H. Ikeda, Y. Takahashi, T. Miyashi,
J. Am. Chem. Soc. 1993, 115, 5180–5184.
J. Phys. Org. Chem. 2013, 26 1090–1097
Copyright © 2013 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/poc