Two-dimensional emission quenching and charge separation using a Ru(II)-photosensitizer assembled with membrane-bound acceptors

Article abstract of DOI:10.1021/jp9710805

Novel syntheses of the bipyridine ligand 1, dcHb (dcHb = 4,4′-dicarboxy-2,2′-bipyridine), by anionic oxidation of 4,4′-dimethyl-2′,2-bipyridine (dmb) using molecular oxygen (4 atm), and of the sensitizer precursor 4, tris(4,4′-diethoxycarbonyl-2,2′-bipyridine)ruthenium(II) bis(triflate), from a chloride-free Ru(II) precursor 3b, Ru^II(DMSO)₄(triflate)_2-n(EtOH)_n (n = 0-2, DMSO = dimethyl sulfoxide, triflate = OSO₂CF₃) are reported. The anionic sensitizer Ru(dcb)₃^4- (5) was shown to bind to vesicles of lecithin when these were made positively charged by cationic bipyridinium electron acceptors. With cetylmethylviologen (CMV²⁺) as quencher, the time-resolved decay of the Ru(dcb)₃^4- emission followed a model for diffusion-controlled quenching in two dimensions. However, the diffusion coefficient obtained from a fit to the data was very small, (6±2)×10^-11 m² s^-1, comparable to values for amphiphiles in bilayers, even though Ru(dcb)₃^4- diffuses in the water region at the vesicle surface. The quantum yield of primary charge separation was 0.06±0.02, which is significant, if not high, despite the large electrostatic attraction between the reactants. Attempts were made to increase the charge separation yield by the use of a monocationic acceptor. Possible extensions of the system are discussed, such as charge separation across the vesicle membrane and the covalent linking of a donor to the sensitizer.

Full text of DOI:10.1021/jp9710805

7494
J. Phys. Chem. B 1997, 101, 7494-7504
Two-Dimensional Emission Quenching and Charge Separation Using a
Ru(II)-Photosensitizer Assembled with Membrane-Bound Acceptors
,
†
‡
§
§
Leif Hammarstr o1 m,* Thomas Norrby, Gunnar Stenhagen, Jerker M a˚ rtensson,
‡
†
Bj o1 rn A° kermark, and Mats Almgren
Department of Physical Chemistry, Uppsala UniVersity, Box 532, S-751 21 Uppsala, Sweden; Department of
Organic Chemistry, Royal Institute of Technology, S-100 44 Stockholm, Sweden; and Department of Organic
Chemistry, Chalmers UniVersity of Technology, S-412 96 G o¨ teborg, Sweden
X
ReceiVed: March 27, 1997; In Final Form: June 30, 1997
Novel syntheses of the bipyridine ligand 1, dcHb (dcHb ) 4,4′-dicarboxy-2,2′-bipyridine), by anionic oxidation
of 4,4′-dimethyl-2′,2-bipyridine (dmb) using molecular oxygen (4 atm), and of the sensitizer precursor 4,
tris(4,4′-diethoxycarbonyl-2,2′-bipyridine)ruthenium(II) bis(triflate), from a chloride-free Ru(II) precursor 3b,
II
Ru (DMSO)
4
(triflate)2-n(EtOH)
n
-
(n ) 0-2, DMSO ) dimethyl sulfoxide, triflate ) OSO
2
CF
3
) are reported.
4
The anionic sensitizer Ru(dcb)
charged by cationic bipyridinium electron acceptors. With cetylmethylviologen (CMV ) as quencher, the
3
(5) was shown to bind to vesicles of lecithin when these were made positively
2
+
4
-
time-resolved decay of the Ru(dcb)
dimensions. However, the diffusion coefficient obtained from a fit to the data was very small, (6 ( 2) ×
m s , comparable to values for amphiphiles in bilayers, even though Ru(dcb) diffuses in the water
3
emission followed a model for diffusion-controlled quenching in two
-
11
2
-1
4-
10
3
region at the vesicle surface. The quantum yield of primary charge separation was 0.06 ( 0.02, which is
significant, if not high, despite the large elecrostatic attraction between the reactants. Attempts were made
to increase the charge separation yield by the use of a monocationic acceptor. Possible extensions of the
system are discussed, such as charge separation across the vesicle membrane and the covalent linking of a
donor to the sensitizer.
Introduction
In attempts to achieve efficient photochemical charge separa-
tion, as well as to understand fundamental properties of
photoinduced electron transfer, several supermolecules have
been synthesized, having photosensitizer molecules covalently
1
linked to electron acceptors and/or donors. Recently, there has
also been an increased interest in systems where the reactants
are assembled by weaker forces, such as electrostatic attraction
or hydrogen bonds.2
We are currently constructing Ru(II)-polypyridine sensitizers
3
covalently linked to a series of electron donors, aiming
eventually at a catalytic donor function. In the present study,
we have investigated the possibility of assembling anionic
sensitizer-donor molecules of this kind, and cationic bipyri-
dinium electron acceptors that are hydrophobically bound to
the surface of lipid vesicles, by using the electrostatic attraction
between the reactants (Figure 1).4 Some favorable properties
were expected to follow from this organization of the reactants,
which is why we wanted to examine this system in some detail.
First, the electrostatic attraction to the bipyridinium-containing
vesicles may prove to be a versatile and synthetically simple
way of adding an acceptor functionality to such sensitizer-donor
supermolecules. Second, one may use acceptors with low
opposite charge, since the sensitizer would be attracted by the
collective electric field from several acceptors at the vesicle
surface. Thus, one may reduce the pairwise attraction between
the reactants, which lowers the cage-escape efficiency after
Figure 1. Schematic structure of the system investigated, where the
5
4-
photoinduced charge separation in bimolecular reactions, and
hydrophilic Ru(dcb)
containing the cationic acceptor cetylmethylviologen (CMV ).
3
is attracted to the surface of the vesicles
2
+
still assemble the reactants efficiently. Third, with the sensitizer
diffusing in the water region around the vesicle, the quenching
was expected to be more efficient than when both the Ru
complex and the acceptors are amphiphilic. In the latter case,
to the slow diffusion of amphiphilic molecules in lipid bilayers.7
Finally, the electron transfer from the donors to the photooxi-
dized Ru is expected to be slow in the donor-sensitizer molecules
mentioned above, and subsequent reactions, e.g., oxidation of
water, will be inhibited if the reductive equivalent and the donor
are not kept apart. Therefore, the vesicle membranes may be
6
quenching was found to be rather inefficient, presumably due
X
Abstract published in AdVance ACS Abstracts, August 15, 1997.
S1089-5647(97)01080-8 CCC: $14.00 © 1997 American Chemical Society

Emission Quenching and Charge Separation
J. Phys. Chem. B, Vol. 101, No. 38, 1997 7495
of use not only to assemble the reactants but also to further
stabilize the charge separation products by transmembrane
electron transfer.
SCHEME 1
We have previously studied the transmembrane transfer of
electrons from external electron donors to secondary acceptors
in the vesicle interior, using the amphiphilic bipyridinium
acceptors cetylmethylviologen (N-hexadecyl-N′-methyl-4,4′-
2+
bipyridinium, CMV ) and cetylbipyridine (N-hexadecyl-4,4′-
+
8
bipyridinium, CB ) as mediators. This reaction was in both
cases shown to proceed via the transmembrane diffusion of the
0
0
uncharged, reduced forms of the mediators, i.e., CMV and CB .
While CB was formed by direct one-electron reduction of CB
by the external reductants, CMV was formed in an indirect
way, via a disproportionation equilibrium: 2V S V + V .
The uncharged forms of the mediators diffused rapidly through
the vesicle membrane (τ ) 0.7 ms for CB ; a similar value
was inferred for CMV ) and reacted with the secondary
0
+
0
+
2+
0
0
8b
0
acceptors.
In the present system, reduced acceptors formed by photo-
induced electron transfer from the Ru sensitizer could possibly
be withdrawn from the external interface of the vesicle to the
interior, where their electrochemical energy might be utilized
in secondary reactions. The oxidized Ru complex left at the
interface may be reduced again by an electron donor, and thus
go through cycles of excitation and electron transfer to acceptor
molecules at the vesicle surface. The donor-sensitizer-
acceptor-membrane assembly we aim at would in this way
exhibit some analogies with a natural photosynthetic reaction
9
center, featuring a cyclic function of the sensitizer, transmem-
brane electron transfer, and a membrane-bound acceptor pool,
although there are obviously important differences as well.
In the present study, we focus on the quenching behavior of
the anionic Ru complexes and vesicle-bound bipyridinium
acceptors and on the organization of the reactants, as well as
on the efficiency of the primary charge separation. Therefore,
the donor could be excluded, and the simple anionic sensitizer
electron-poor ligand 2, 4,4′-diethoxycarbonyl-2,2′-bipyridine
diester) was facilitated by exchange of the rather strongly bound
(
chloride ligands for ethanol or triflate ions. Precipitation of
the chloride ions as AgCl(s) was achieved by refluxing of the
RuCl2(DMSO)4 species 3a with 2.05 equiv of silver triflate in
4-
ruthenium tris(4,4′-dicarboxylato-2,2′-bipyridine) (Ru(dcb)3
)
II
ethanol (12-15 h), yielding 3b, Ru (DMSO)4(triflate)2-n(EtOH)n
was used as a model of the donor-sensitizer molecules. Some
possible extensions and limitations are discussed.
(
n ) 0-2). Slow addition of the chloride-depleted 3b to a
refluxing ethanolic solution of 2 yielded 4. The formation of
the product was monitored by the appearance of the deep red
color of 4, and by thin-layer chromatography (TLC). It is very
important to fully remove all chloride ions, since they tend to
Results and Discussion
Syntheses. The dicarboxylic acid ligand 1, 4,4′-dicarboxy-
2
,2′-bipyridine (dcHb), has been prepared earlier from 4,4′-
II
form byproducts, e.g., the blue 7, trans-dichloroRu (diester)2
dimethyl-2,2′-bipyridine (dmb) by oxidation using potassium
II
(
(
point group symmetry D4h); green 8, cis-dichloroRu (diester)2
point group symmetry C2) (Figures 2 and 3) which are very
1
0
11
permanganate or by chromic acid which require laborious
isolation and purification methods. We here report a novel,
speedy, and transition-metal-free procedure for the synthesis of
stable kinetically; extended refluxing (5-7 days) of a reaction
solution containing 7 and 8 does not lead to conversion into 4
to any useful extent. Complexes displaying trans-bipyridineru-
thenium coordination geometry are very rarely encountered, e.g.
1
(Scheme 1), adopted from an earlier method reported for the
12
preparation of monocarboxylic acids from picolines, by anionic
oxidation of dmb with molecular oxygen (O2) at 4 atm pressure.
The oxygen atmosphere is obtained in a standard Fisher-Porter
1
7
Meyer et al. A small amount of a pink species 9 (not shown)
containing 12 protons which are inequivalent in NMR (point
group symmetry C1) was sometimes also formed. By NMR
and UV-vis (Figure 3) we have indications that this is yet
another Ru(II) species complex composed of two molecules of
diester 2 and two different anionic ligands. Finally, a late-
eluting dark purple species 10 was observed after chromatog-
raphy, which we ascribe to species 9 where one or several of
the ester groups have been hydrolyzed (UV-vis and NMR
spectra in Supporting Information).
5
00 mL reaction flask tested to 15 atm. CAUTION: Always
observe proper protective measures (e.g., eye protection, shield)
when working with oxygen under pressure. The synthesis
reported here of the octahedral ruthenium(II) complex (point
group symmetry13 D3) 4, tris(4,4′-diethoxycarbonyl-2,2′-bipy-
ridine)ruthenium(II) bis(triflate) is by a novel method based on
the ruthenium(II) precursor 3a, RuCl2(DMSO)4, readily available
from RuCl3‚3H2O and dimethyl sulfoxide (DMSO). In our
hands the procedure by Wilkinson et al.14 most likely yielded a
Basic Characteristics of Ru(dcb)84-. The absorption spec-
mixture of trans- (minor) and cis-dichloro (major) species.
1
5
4-
Alessio et al. more recently have reinvestigated the RuCl2-
DMSO)4 system; and this precursor in the synthesis of
ruthenium(II) complexes has found recent use, e.g., by Boss-
trum of Ru(dcb)3 in water shows a strong band in the
(
visible region, with a maximum at 466 nm, characteristic of
the lowest metal-to-ligand charge-transfer (MLCT) transition
1
6
18
mann et al.
The complexation reaction with the rather
in polypyridine complexes of Ru(II). The emission from the

7496 J. Phys. Chem. B, Vol. 101, No. 38, 1997
Hammarstr o¨ m et al.
attained much faster than the rate of decay of the excited
2
1
species. In the present study, however, one may conclude
that only the fully dissociated species has to be considered, since
the experiments were in all but a few cases performed at pH )
1
1.
Emission Quenching by MV . As for a large number of
2
+
4-
Ru(II)-polypyridine complexes, the excited state of Ru(dcb)3
was quenched by methylviologen (MV ).
2+ 5b,c,18b,22
The reduc-
tion of the emission intensity under steady-state illumination is
shown in Figure 4. For a purely dynamic quenching process,
the emission intensity is given by the well-known Stern-Volmer
2
3
equation:
I /I ) 1 + K [Q]
(1)
0
SV
where I0 and I are the emission intensities in the absence and
presence of quencher, respectively; KSV ) τ0kq, where τ0 is the
unquenched lifetime, kq is the second-order rate constant for
the quenching reaction, and [Q] is the concentration of quencher.
In the present system, the charged ground-state reactants form
ion pairs to some extent. If the emission intensity from excited
4
-
Ru(dcb)3 within ion pairs is negligible due to rapid static
quenching, eq 1 may be rewritten to include a combination of
2
4
dynamic and static quenching:
I /I ) (1 + K [Q])(1 + K [Q])
(2a)
(2b)
0
SV
A
Figure 2. trans- (7) and cis- (8) dichlororuthenium(II)(4,4′-diethoxy-
carbonyl-2,2′-bipyridine) byproducts which have been separated and
2
[
(I /I) - 1]/[Q] ) K + K + K K [Q]
identified. H-6 and H-6′ (H-5′, H-3′, -R′) denotes the two magnetically
0 SV A SV A
different pyridine moieties of each 4,4′-diethoxycarbonyl-2,2′-bipyridine
1
where KA is the ion-pair formation constant. Equation 2a
predicts an upward curvature in the I0/I vs [Q] plot, due to static
quenching, and this is also seen in Figure 4. The inset shows
the data plotted according to the linearized eq 2b, and from the
values of the slope and intercept of the fitted line, one obtains
ligand (pyridine or pyridine′), with regard to the assignment of the H
NMR spectra in the experimental part.
lowest MLCT state is centered around 645 nm (corrected) and
was found to decay with a lifetime of 690 ns in N2-purged water
1
9
at 293 K.
4
-1
2
the two solutions KSV ) 1.23 × 10 M and KA ) 9 × 10
The dissociation behavior of the corresponding acid form Ru-
-1
M , or vice versa. Independent measurements of the emission
lifetime vs [Q] resulted in single-exponential decay curves, since
+
(
dHcb)32 has been studied by Park and co-workers.20 They
reported the observation of two reversible deprotonation steps:
first two protons dissociate independently, forming a neutral
complex, and then, at higher pH the remaining four protons
4-
all emission emanated from the Ru(dcb)3 that were not
statically quenched. The lifetime data was plotted according
to
4
-
dissociate independently of each other, giving the Ru(dcb)3
species. According to their data, the pH values where 50% of
the protons in each of the two steps had dissociated were pH50%
τ /τ ) 1 + K [Q]
(3)
0
SV
4
-1
)
0.7 and 3.7. The corresponding values for the excited state
were pH*50% ) 1.9 and 4.1. These values were calculated with
(not shown) and gave a value KSV ) 1.22 × 10 M ,
10 -1 -1
corresponding to kq ) 2.57 × 10
M
s
(τ0 ) 475 ns in
the implicit assumption that the dissociation equilibrium was
air-saturated water).
II
]2+(OTf)
II
II
II
Figure 3. UV-visible spectra in CH
2
Cl
2
of 4, [Ru (diester)
3
2
; 7, [trans-Cl
2
Ru (diester)
2
]; 8, [cis-Cl
2
Ru (diester)
2
]; 9, Ru (diester)
2
(Cl)-
(
X) (a pink species, not completely identified). The spectra have been normalized for comparison, and the ligand-centered π-π* absorption below
00 nm is shown smaller than the actual amplitude.
3

Emission Quenching and Charge Separation
J. Phys. Chem. B, Vol. 101, No. 38, 1997 7497
TABLE 1: Values for the Diffusion Coefficient Obtained
a
from a Fit to Eq 4 for Different Vesicle Preparations
D (m s-1)c
2
-
16 b
preparation no.
d
e
[
Q] × 10
σ
tot
σ
q
-
2
(mC m-2)
(mC m-2)
(
m
)
I
II
III
IV
1
.1
.8
.8
.2
.5
.0
.7
.1
7.6
38
48
12
49
60
47
21
47
32
49
3.5
9
2
3
4
5
6
6
8
5.3
6.3
12
13
18
19
21
26
32
37
6.6
5.0
4.4
4.5
7.0
6.3
1
1
0.0
1.6
5.6
a
Note that the average value of D varies somewhat with vesicle
preparation, but there is no detectable covariation of the D and [Q]
b
2+
values. Surface concentration of the quencher CMV (molecules/
4
-
Figure 4. Quenching of the emission intensity from Ru(dcb)
3
by
2
c
d
m ). Lateral diffusion coefficient obtained from a fit to eq 4. Surface
charge density of the vesicles. Surface charge density contribution
2
+
MV , determined by steady-state fluorimetry. The data are plotted
according to eq 2a, and the line is a least-squares linear fit to the data
for low concentrations. Inset: The same data plotted according to eq
e
from the quencher only.
2
b, and the line is a least-squares linear fit to the data. Conditions:
the charged vesicles is strong, even at relatively low surface
concentrations of viologen. Although the Poisson-Boltzmann
equation may not be very accurate for multivalent ions, the
fraction of Ru(dcb)3 that is further away than a molecular
diameter (≈1 nm) from the vesicle surface is expected to be
very small. Therefore, the observed reduction of the emission
intensity as the concentration of viologen increased should be
due to an increasingly efficient quenching of Ru complexes
already bound to the vesicle surface and does not reflect an
increasing fraction of bound complex.
4
-
air-saturated water at 20 °C. [Ru(dcb)
)
3
] ) 7 µM, pH ) 11 ([NaOH]
1 mM).
4
-
Time-resolved emission studies were performed in order to
4
-
investigate the localization of Ru(dcb)3 and the quenching
2+
behavior. The concentration of CMV in the outer monolayer
of the vesicle membrane was varied from 0 to 10 mol %. The
surface charge density σtot was in most cases kept around 40
2
mC/m or above (see Table 1) by addition of N-cetylpyridinium
+
(
CPy ). This corresponds to at least one unit charge per six
4
-
Figure 5. Quenching of the emission intensity from Ru(dcb)
3
by
lipids and was done in order to ensure a high degree of
2
+
4-
+
CMV in lecithin vesicles, determined by steady-state fluorimetry.
Conditions as in Figure 4, with [lecithin] ) 3.0 mM. In this figure,
Ru(dcb)3 binding. Separate experiments showed that CPy
4-
did not quench the Ru(dcb)3 emission. Some representative
decay curves are shown in Figure 6. It is clear that the emission
decay cannot be described by a simple mono- or biexponential
function, as would have been expected for a system where
2
+
17
-2
[
CMV ] ) 140 µM corresponds to [Q] ) 1.0 × 10
m .
To conclude, these experiments show that the quenching of
4-
2+
excited Ru(dcb)3 by MV is diffusion controlled (kq ) 2.6
4
-
Ru(dcb)3 formed simple ion pairs. Not even a fit to three
exponentials gave satisfactory results. Instead, the curves were
fitted to a kinetic model for diffusion controlled quenching in
two dimensions:
10
-1 -1
×
10
M
s ), and that ion pairs are formed between the
reactants in the ground state, with a formation constant KA ) 9
2
-1
-3
×
10 M (in water at 293 K, ionic strength ) 2 × 10 M).
This information is needed for further discussion (see below).
Quenching by CMV2+ in Vesicles. The amphiphilic violo-
1
/2
2+
I(t)/I
0
(t) ) (1 - R)exp{-(t/τ
0
) - 7.44[Q]r(tD)
-
gen N-cetyl-N′-methylviologen (CMV ) binds completely to
vesicles of egg lecithin,8 with the viologen headgroup exposed
to the water phase and the cetyl chain in the hydrocarbon region
of the vesicle bilayer. Upon addition of CMV to a solution
of Ru(dcb)34 and uncharged (zwitterionic) vesicles of egg
lecithin, the emission from the Ru-complex was considerably
quenched, but the ratio I0/I did not follow a simple dependence
a
2.28[Q]Dt} + R exp{-t/τfree} (4)
2+
where τ0 is the unquenched emission lifetime, [Q] the surface
concentration of quencher (molecules per unit area), r the
interaction distance, and D the lateral diffusion coefficient. The
second term was included to describe the emission decay of
-
4
-
4-
on [Q] (Figure 5). One may have expected that Ru(dcb)3
those Ru(dcb)3 that were not bound to the vesicles, with
2+
would form ion pairs with one (or two) CMV molecules at
the vesicle surface. However, with a static quenching within
such ion-pairs, this would lead to a linear (or quadratic)
dependence of I0/I on [Q]. The data in Figure 5 instead suggest
a cooperative effect, where the attraction of Ru(dcb)3 to the
vesicle was determined, e.g., by the surface charge density of
the vesicles.
lifetime τfree, and R is a factor (0-1) weighing the contributions
from the two exponential functions. This equation (apart from
the second exponential function) was derived by Owen,26 and
the numerical factors were modified according to Caruso et al.27
The fitted curves are also shown in Figure 6, and the quality
4
-
2
of the fit was good, as judged from the ø values and the residual
plots. Separate experiments on a longer time scale than in
Figure 3 established that the value τfree ) 590 ns was
independent of [Q], and τfree was therefore constrained to this
Numerical calculations using the Poisson-Boltzmann equa-
2
5
4-
tion suggest that the binding of the tetravalent Ru(dcb)3 to

7498 J. Phys. Chem. B, Vol. 101, No. 38, 1997
Hammarstr o¨ m et al.
assumes an infinite rate of reaction at distances equal to the
2
6
reaction distance.
Nevertheless, the quenching process is
expected to be well within the diffusion-controlled limit, since
the diffusion coefficients were much smaller than those for free
molecules in water.
To determine if there was any quenching on a time scale too
short to be resolved by the single-photon-counting equipment
(static quenching), comparison was made between the intensity
values obtained in Figure 5 with those obtained by integrating
the time-resolved curves in Figure 6. For the highest value of
-
2
[
Q] ()11.6 m ) used in the latter experiments, I/I0 ) 0.07,
while in the steady-state measurements I/I0 ) 0.03 at the same
Q] (the steady-state value was corrected for the different τ0
[
with and without oxygen in this comparison). This would
indicate a small fraction of static quenching. However, also a
very small increase in R in the time-resolved measurements,
due to the irreversible photoreaction (see below), would account
for the difference between the values from steady-state and time-
resolved experiments. Thus, even though we cannot completely
exclude the presence of static quenching its contribution must
be smaller than 5% of I0 even at the highest quencher
concentrations.
Figure 6. Representative results from time-resolved single-photon-
4-
counting experiments on the emission quenching of Ru(dcb)
3
by
2
+
CMV in lecithin vesicles. The solid lines are fits to eq 4, and a
representative residual is shown (inset). The dotted lines indicate the
parts of the decay curves represented by the second exponential term
in eq 4, i.e., a single-exponential decay with lifetime ) τfree. Thus, the
difference between the solid and dotted lines are the contribution from
the two-dimensional quenching at the vesicle surface. In the figure,
curves with different values of R were chosen so as to obtain a clear
presentation with a good separation between the curves. Note though
that there was no clear dependence of R on [Q] for the entire set of
An unspecified, irreversible reaction occurred during the time-
resolved emission experiments, induced by the excitation light
(λ ) 327 nm), leading to a continuously increasing emission
data in Table 1. Conditions: N -purged water at 20 °C, pH ) 11
2
4
-
intensity during the data collection. This was also observed
when the samples were excited by the excimer/dye laser system
(λ ) 466 nm). At both excitation wavelengths, prolonged
excitation resulted in a blue coloration of the solution, charac-
(
[NaOH] ) 1 mM), [Ru(dcb)
3
] ) 25-30 µM, [lecithin] ) 6.5-7.0
16 16
-2
-2
mM; (from top) [Q] ) 2.8 × 10
m
, 4.2 × 10
m
, and 11.6 ×
1
6
-2
1
0
m . The surface charge density (σ) was kept approximately
+
constant by addition of CPy (see Table 1).
3
0
teristic of viologen radicals. This was most evident when
value in the fitting procedure. The reduction of τfree compared
to τ0 ()690 ns) is probably due to the encounter between free
Ru(dcb)34 and vesicles. In the fitting process, the values of
2
+
31
micelles of pure CMV (6 mM, which is above the cmc )
2
+
instead of vesicles with CMV , was used to bind the Ru
complexes. The accumulation of viologen radicals indicates
that a component of the system acts as a sacrificial donor,
preventing the recombination of viologen radicals with the
oxidized Ru complexes. It seemed that the increase in emission
intensity was due to an increasing fraction of Ru complexes
that are not bound to the vesicles. One possible reason would
be a photoinduced decarboxylation of Ru(dcb)3 , as a conse-
quence of irreversible electron transfer from the carboxylate
groups; either to the oxidized Ru in the same complex (the
-
[
Q] and r )1.1 nm28 were also kept constant. A 10% change
in either of these parameters resulted in only a 10% difference
in the value of D obtained. The quenching by MV2 was clearly
diffusion controlled (see above), and it is likely that this is the
+
+
case also for the vesicle-bound CMV , as required by the model
behind eq 4.
4
-
The values of the diffusion coefficient D obtained from the
fits are given in Table 1. The independence of the D values on
[Q] gives important support for the correctness of the physical
1
8
32a
oxidation is metal-centered ), in analogy with, e.g., oxalate,
model chosen for the quenching process. For example, the
results of Medhage and Almgren29 show that an attempt to fit
the data from a three-dimensional diffusion-controlled quenching
reaction to a two-dimensional model results in too high values
of D that also vary with the concentration of quencher, in
contrast to the present results. The D value was also indepen-
dent of σtot, as expected from the model. It appears therefore
2
+
or to CMV , although 466 nm is well outside the previously
reported viologen-carboxylate charge-transfer bands.32 This
would also explain the accumulation of viologen radicals.
However, we have not made any direct attempts to determine
the mechanism responsible for the irreversible, photoinduced
4
-
2+
changes observed in the Ru(dcb)3 /CMV system, and any
conclusion must therefore await further studies of the topic.
4-
that the major fraction of the Ru(dcb)3 molecules are bound
as counterions at the vesicle surface and that this fraction is
quenched by two-dimensional, diffusional encounter with
The irreversible photoinduced reaction would explain why
the fraction of unbound complexes is higher than expected from
the Poisson-Boltzmann calculations; R ) 0.05-0.2 for curves
with σtot > 30 mC/m, and somewhat higher for the lower values
of σtot. The escape of some Ru complexes from the vesicle
surface during the measurements does not seem to change the
time behavior of the emission decay from that predicted by eq
2+
CMV .
The average value of D varied between the different vesicle
preparations, and this type of variation is not uncommon in
quantitative kinetic studies in vesicular systems.8 The origin
of these variations could be slight variations in the vesicle
characteristics between the preparations, e.g., differences in size
4
. This is also expected if the Ru complexes may be regarded
as either free or bound to the vesicle surface. Only the relative
amplitudes of the two exponential terms in eq 4 would change
during the measurements. The value of R would depend on
several independent parameters, as, e.g., the number of photons
absorbed, and consequently no clear dependence on [Q] or σtot
can be observed.
(
and hence in radius of bilayer curvature) or other properties
that affect the lipid packing in the vesicle bilayer.
The data for the first 15 ns after the excitation showed a decay
that was more rapid than predicted by eq 4, and these data were
excluded from the fits. We ascribe this to quenching of excited
Ru(dcb)34- that were within, or close to, reaction distance with
a quencher molecule already at the moment of excitation. The
model behind eq 4 does not include this contribution but
Finally, it must be noted that the value of D, i.e., the sum of
the lateral diffusion coefficients for Ru(dcb)3 and CMV , is
4
-
2+

Emission Quenching and Charge Separation
J. Phys. Chem. B, Vol. 101, No. 38, 1997 7499
of the same order of magnitude as experimental values for lipids
and other amphiphilic molecules incorporated in liquid-crystal-
line bilayers or monolayers. These values are 2 orders of
obtained is reasonable. For a photoinduced electron-transfer
reaction, where the rate of reformation of the excited precursor
complex is negligible, ΦET ) k′-d/∑k , where k′ is the rate
7
i
-d
magnitude lower than the D values in water for free molecules
of the same size as Ru(dcb)34 . Although the quenching is more
efficient than in previous experiments with an amphiphilic Ru
constant for separation of the products from the solvent cage,
-
and ∑k is the sum of the rate constants for all reactions of the
i
5
products in the cage. In the present case, ∑k ) k′ + kBET
where the latter is the rate constant for back electron transfer
i
-d
,
6
complex, we expected an even higher efficiency, since
Ru(dcb)34 is diffusing in the water region around the vesicle.
The reason for the low value of D is not clear, but because of
its importance we take the liberty to speculate on its origin.
One could conceive that the Ru(dcb)34- would bind rather
tightly to the cationic amphiphiles and thus follow their lateral
-
to re-form the ground-state reactants.
In the present system, as opposed to Ru(bpy)32+/MV2+, there
is an electrostatic attraction between the products within the
solvent cage, and apparently a relatively low value of the
diffusion coefficients (see above). Consequently, the value of
k′-d is expected to be much lower, and therefore, the value of
+
movements in the bilayer. However, binding to CPy cannot
explain the apparent low value of D, since the same values were
2
+
2+
kBET must also be lower than for Ru(bpy)3 /MV in order for
ΦET to reach the observed value. Indeed, on energetic ground
the back electron transfer is expected to be in the Marcus
+
obtained also when no CPy was present (see Table 1). Nor
can Ru(dcb)34 be bound to the quencher molecules, since this
would result in “ion-pair quenching”, and the emission decay
curves would not follow the predictions of eq 4, as discussed
above.
-
3
3,34
inverted region
for both systems, and the value of ∆G° is
35
3-
+
about 0.2 eV more negative for the Ru(dcb)3 /CMV , which
is expected to give a significantly lower rate of back electron
Another possible explanation is that the encounter between
the reactants is in some way hindered by the lipid headgroups.
One may envisage that the reactants are bound at different depths
in the bilayer, that their encounter rate is limited by diffusional
fluctuations of the headgroups, or that they are forced to attain
an orientation relative to each other in which quenching is
relatively unfavorable. To some extent, this effect could be
accounted for within the two-dimensional model, by assuming
3
+
+
transfer than for Ru(bpy)3 /MV . The reorganization energy
may also be lower due to the less polar environment at the
vesicle, as compared to the homogeneous Ru(bpy)32 /MV
+
2+
system, which is also expected to reduce the rate of back-
electron transfer in the inverted region. Finally, the slightly
28b
larger reaction distance in the present system may reduce the
electronic coupling and contribute to the decrease in kBET. Thus,
it seems reasonable that the value of kBET is relatively low in
the present system, which compensates to a great extent the
likewise low value of k′-d, so that ΦET reaches a detectable
level.
a less than diffusion-controlled reaction rate. It would lead to
low apparent values of D. Medhage and Almgren29 have studied
the consequences of a decrease in reaction rate upon encounter
for a reaction in 0-3 dimensions. From their results it is clear
that for reactions in two dimensions that are not diffusion
controlled the values of D obtained in the data analysis, with
an erroneous assumption of diffusion control, are indeed lower
than the true ones. However, it is uncertain if D values can be
obtained that are 2 orders of magnitude lower that those
corresponding to diffusion control, without a clear deviation of
the emission time behavior from that predicted by eq 4. Further
experiments on similar systems are necessary in order to
determine the origin of the low D values presently obtained.
Charge Separation Yield. The mechanism for the quench-
The transmembrane electron transfer via the rate-determining
disproportionation of viologen radicals (see Introduction) occurs
on the time scale of 1s at low surface concentrations.5
Therefore, it cannot compete with the recombination of the
reduced acceptor and the oxidized sensitizer, which was found
to occur on the time scale of 1 ms for one reactant pair per
vesicle. To overcome this problem, one may add an electron
donor (see Introduction) and/or use a single-electron acceptor
that may diffuse through the membrane without the need for
+
further reaction, e.g., cetylbipyridine (CB , see Introduction).
4-
ing of Ru(dcb)3 by viologen is, as for many other Ru-
polypyridine complexes,22 electron transfer from the excited Ru
complex to the quencher. Flash-photolysis experiments were
made in order to measure the yield of charge separation, ΦET,
defined according to
+
Quenching by CB in Vesicles. Attempts were also made
4-
+
to quench the excited Ru(dcb)3 by cetylbipyridine (CB ) in
+
vesicles. The methylbipyridine (MB ) was found to quench
the emission from Ru(bpy)3 by the same mechanism as MV ,
i.e., by oxidative electron transfer. Also, it has been shown
that CB binds completely to lecithin vesicles and that the
reduced form (CB ) transfers electrons through the bilayer of
lecithin vesicles with a lifetime of 0.7 ms. Therefore, if CB
would accept an electron from the excited Ru(dcb)34 , the
system could be used to achieve a photoinduced charge
separation across the vesicle membrane, with a membrane-bound
acceptor pool and a secondary electron acceptor added to the
2+
2+
6
CMV+]
*Ru(dcb)34-]
+
[
1
f
max
0
ΦET
)
(5)
[
8b
+
t)0
-
4-
where f is the fraction of excited Ru(dcb)3 that is quenched,
-
[
*Ru(dcb)34 ]t)0 is the concentration of excited Ru-complexes
+
immediately after the excitation pulse, and [CMV ]max is the
maximum concentration of charge separation products (mea-
sured just after complete excited-state decay but before signifi-
+
vesicle interior. The advantages with CB as compared to
CMV are that with the former there will be no electrostatic
attraction between the charge separation products within the
2
+
cant recombination).
For a solution with 29 µM Ru(dcb)34-, 390 µM CMV2+, and
+
17
-2
solvent cage. Also, CB only has to be reduced in one step
7
[
0
(
.30 mM lecithin ([Q] ) 1.1 × 10 m ), the values
*Ru(dcb)34 ]t)0 ) 3.3 ( 0.3 µM and [CMV ]max ) 0.2 (
.05 µM were obtained from the flash-photolysis measurements
see Experimental Section, curves not shown). Inserted in eq
-
+
before it can diffuse across the membrane, and may therefore
compete with the diffusion-controlled recombination with the
sensitizer.
+
5
, with f ) 0.93 ( 0.03, these values give ΦET ) 0.04-0.09 (I
Unfortunately, addition of CB up to 10 mol % of the lipids
-
3
)
2 × 10 M). This may be compared with the value ΦET )
in the external monolayer of lecithin vesicles did not quench
2+
2+
28
4-
0
.38 for the couple Ru(bpy)3 -MV (in water, I ) 0.01 M).
Some estimates using conventional models of bimolecular
electron-transfer processes serve to show that the value of ΦET
the Ru(dcb)3 emission at pH g 5.0 (where the quencher is in
+
+
the unprotonated form CB ). At higher concentrations of CB
(≈15 mol %), the vesicles were unstable and precipitated within

7
500 J. Phys. Chem. B, Vol. 101, No. 38, 1997
Hammarstr o¨ m et al.
some minutes upon addition of Ru(dcb)34-. This was the case
The charge separation yield was significant, if not high, and
by development of the system it may be improved. The system
thus realizes photoinduced charge separation using sensitizers
preassembled with a membrane-bound acceptor pool. From
previous studies, it is clear that an extension to include charge
separation across the lipid membrane is possible. The system
may thus prove to be a convenient way to add an acceptor
function to sensitizer-donor molecules.
+
both when CB was added before and after vesicle preparation.
+
+
The reason for the lack of quenching by CB and MB is
likely their relatively low reduction potentials. The reduction
+
+
potential of CB and MB depends on pH, due to a protonation
equilibrium at the nonalkylated nitrogen, and the pKa values
+
for MB are 3.5 and 10.5 for the oxidized and reduced forms,
3
6
4-
respectively. Although the potentials of the excited Ru(dcb)3
(
≈
2
0
-0.70 V vs NHE ) and the quenchers should be equal at pH
7.5 these potentials involve protonation of the reduced
acceptor. In the quenching reaction, however, at pH g 5 the
protonation is much slower than the rate of dissociation of the
reactant encounter/successor complex. Therefore, quenching
is not exergonic unless MB or CB are protonated already in
the oxidized form (E° ) -0.45 vs NHE ). This is in agreement
with results from earlier experiments with Ru(bpy)3 , which
is a stronger photoreductant: the rate constant for quenching
by HMB /MB in homogeneous solution was equal to that of
Experimental Section
Methods. The H NMR and 13C NMR spectra of the ligands
1
1
and 2 and complexes 4-10 were recorded in deuterated
chloroform, methanol, dimethyl sulfoxide (DMSO), or deute-
rium oxide made alkaline with NaOH (using a solvent suppres-
sion routine decoupling the solvent hydroxy resonance at 4.75
ppm), on Bruker AM 400 (400 MHz proton, 100 MHz carbon)
or DMX 500 (500 MHz proton, 125 MHz carbon) instruments.
Chemical shifts (δ) are reported in parts per million (ppm)
downfield from TMS (tetramethylsilane). IR spectra were
recorded on an ATI Mattson Infinity Series FT-IR spectrometer
as KBr disks. Melting points were obtained on a B u¨ chi
apparatus and are uncorrected. Analytical TLC was performed
on precoated aluminum oxide gel 60 F254 plates (neutral, Merck)
with UV detection. Aluminum oxide gel (neutral, activated,
+
+
30
2
+
2
+
+
2
+
6
MV at pH < 3, but smaller and constant at pH > 4.5.
Possible Development of the System. It seems that the
system studied provides the desired organization of the reactants
at the vesicle surface, although the quenching was much slower
than expected. The large attractive field from the charged
acceptor/vesicle assembly should allow the use of Ru complexes
4
-
with a somewhat lower net negative charge than Ru(dcb)3
,
6
0 mesh, 95+%, Alfa) was used for preparative column
which would lend some freedom in the modification of the
complexes. Despite the relatively large electrostatic attraction
between the charge separation products within the solvent cage,
a measurable yield of charge separation was obtained, and a
reduction of the acceptor and sensitizer charge should give some
improvement. Also, if electron transfer from a donor to the
oxidized sensitizer could compete with the cage recombination,
one could obtain a higher charge separation yield. Thus, by
using the donor-sensitizer molecules discussed in the Introduc-
tion, one may realize a system with a donor-sensitizer-acceptor
pool that performs photoinduced charge separation in a lipid
membrane. The problem with the irreversible photoinduced
reaction could possibly be solved by using negatively charged
substituents other than carboxylic groups.
chromatography of the diester 2 and complexes 4 and 7-10.
The electronic absorption spectra were recorded on a Varian
Cary 5E UV-vis-NIR spectrophotometer at 25.0 °C ((0.1 °C).
Steady-state emission spectra were recorded on a Perkin-Elmer
LS-5 luminescence spectrometer or a SPEX Fluorolog 2
spectrofluorimeter. The electrospray ionization mass spectrom-
etry (ESI-MS) experiments were performed on a ZabSpec mass
spectrometer (VG Analytical, Fisons instrument). Electrospray
conditions were as follows: needle potential, 3 kV; acceleration
voltage, 4 kV; bath and nebulizing gas, nitrogen. Liquid flow
was 50 µL/min using a syringe pump (Phoenix 20, Carlo Erba,
Fisons instrument). Solvent compositions were 50% acetoni-
trile-50% water containing 1% acetic acid or 100% acetonitrile.
Accurate mass measurements were obtained by the use of poly-
ethylene glycol (PEG) as an internal standard.
To also achieve electron transfer across the vesicle membrane,
a monocationic acceptor should preferably be used, which
becomes uncharged after reduction by a single electron. In this
way, it could transverse the membrane directly, in contrast to
the viologens (see Introduction), and perhaps avoid recombina-
tion with the sensitizer. This acceptor could be a cetylbipyridine
that is modified by electron-withdrawing substituents, so that
it may quench Ru(dcb)34 by electron transfer. However, the
Small unilamellar vesicles were prepared by sonication with
a tip. A small sample of lecithin dissolved in chloroform was
transferred to a glass tube. The chloroform was removed by
overnight storage at low pressure, and 5 mL of water or buffer
was added. The lecithin solution was sonicated at room
temperature, using a Soniprep 150 from MSE Scientific Instru-
ments, Crawley, UK. This procedure has been demonstrated
to give a relatively monodisperse sample of unilamellar vesicles
-
substituent will make the molecule more polar, which will
reduce the rate of its transmembrane diffusion.8
b,37
This process
8
with an average radius around 15 nm.
will thus be less able to compete with recombination with the
oxidized sensitizer. With some fine-tuning of the acceptor
properties, it may be possible to optimize the quenching and
charge separation processes.
The surface concentration of quencher, [Q], was calculated
2
25
using a value of 0.71 nm for the surface area of the lecithin.
The headgroup area of the quenchers was relatively small, and
they were present at a mole fraction less than 0.1. Thus any
possible expansion of the vesicle surface due to the addition of
the quenchers was neglected in the calculation of [Q].
Conclusions
The photosensitizer Ru(dcb)34- was shown to bind strongly
to the surface of lipid vesicles when the latter were positively
charged by incorporation of amphiphilic bipyridinium acceptors.
All photophysical and photochemical measurements were
performed in milli-Q water at pH ) 11 (1 mM NaOH), at 20
( 2 °C. Some experiments with MB and CB were performed
in 2 mM phosphate buffer at different pH values.
+
+
2+
The quenching of the excited sensitizer by the viologen CMV
followed a model for diffusional quenching in two dimensions.
The value of the diffusion coefficient obtained is of the same
order of magnitude as for amphiphilic reactants bound in
bilayers. This is much smaller than the value expected, since
Ru(dcb)34 diffuses in the water region around the vesicle
interface.
Time-resolved emission experiments were performed using
single-photon counting; a mode-locked Nd:YAG laser was used
to pump a DCM dye laser. The output from the dye laser was
frequency doubled to 327 nm and used to excite the samples.
The instrumental response function fwhm ≈ 400 ps, as
determined with a scattering sample. The emission was
-

Emission Quenching and Charge Separation
J. Phys. Chem. B, Vol. 101, No. 38, 1997 7501
observed around 610 nm using an interference filter with 10
nm bandwidth.
through green to orange-yellow, the oxygen atmosphere was
released and the reaction solution quenched by pouring onto
ice (300 mL), hydrolyzing the excess of potassium tert-butoxide.
The slightly opaque, strongly alkaline, bright yellow water phase
was extracted with multiple fractions of CH2Cl2 until no increase
in volume of the separated organic phase was found (8-10 ×
Measurements of ΦET was conducted with a conventional
flash-photolysis setup. An ELI-94 excimer laser, operating with
XeCl (λ ) 308 nm) was used to pump a LT-1113 dye laser
(both from the Estonian Academy of Sciences) that excited the
5
0 mL). The first CH2Cl2 phase was checked for unreacted
sample (λexc ) 466 nm, <20 ns fwhm). A pulsed Xe lamp
provided the analyzing light, and the detection system employed
a Techtronix 7912AD digitizer and a personal computer for
conversion to absorption traces. Control experiments were
performed in order to exclude direct photoionization of the
starting material (none found) by TLC (Al2O3, hexanes:ethyl
acetate, 60:40, Rf ≈ 0.65). The pH of the water phase was
adjusted to ≈3.5 with 6 M HCl, at which a precipitate of 1
formed. After the precipitate had settled, it was collected by
suction filtration, washed with hot water at pH ≈ 3 (100 mL),
hot ethanol (100 mL, 96%) and finally hot acetone (100 mL)
in order to remove traces of DMF which initially made the
precipitate sticky. A yellow powdery solid was obtained (830
mg, 3.4 mmol, 63%), with an additional 3% obtained after the
evaporation of the washings, sufficiently pure for the subsequent
esterification. Mp dec above 290 °C (lit. mp above 325 °C1 ].
Ru(dcb)34 or photoreduction of excited bipyridinium acceptors.
-
4-
Since the excited-state spectrum of Ru(dcb)3 was not known,
the concentration of excited complexes created by the laser pulse
was calculated from the decrease in absorbance at 450 nm
(
shifted from λexc to reduce the intensity of scattered excitation
4-
4
-1
light). For the ground-state Ru(dcb)3 , ꢀ450 ) 2.3 × 10 M
0a
-
1
4
-1
-1 38a
cm (value at band maximum ꢀ464 ) 2.6 × 10 M cm
)
1H NMR (D2O made alkaline with NaOH): δ 8.68 (dd, J )
0.9, 5.2 Hz, 2H, H-6 and H-6′), 8.29 (dd, J ) 0.9, 1.5 Hz, 2H,
H-3 and H-3′), 7.77 (dd, J ) 1.5, 5.2 Hz; 2H, H-5 and H-5′).
COOH was not observed under the experimental conditions used
(NaOH dissolved in D2O giving a “pH” ≈ 12), with the solvent
hydroxy resonance decoupled at 4.75 ppm before each aquisition
and it was assumed that the excited state had roughly the same
value as for Ru(bpy)32 , ꢀ450 ) 5 × 10 M cm .
+
3
-1
-1 38b
Although
this assumption may be somewhat crude, the error in the value
4
-1
-1
of ∆ꢀ450 ) -1.8 × 10 M cm (excited state - ground state)
is probably minor. The initial excited-state decay (Figure 6)
could not be resolved with the flash-photolysis equipment.
Therefore, a solution without quencher, but otherwise identical,
was used to measure the concentration of excited states produced
by the laser flash. The concentration of separated products
created was measured at 600 nm, where the absorption from
13
using a Bruker-supplied NMR software routine. C NMR (D2O
made alkaline with NaOH): δ 175.5, 158.3, 152.3, 148.9, 125.9,
-
1
123.8. IR (KBr disk, (cm ), all strong except where noted):
2925 (m), 2456 (m), (O-H, H-bonded), 1719 (CdO), 1632,
1561, 1365, 1283, 1270.
+
all species but the reduced viologen (CMV ) is negligible. For
+
4
-1
-1
30
CMV , ꢀ600 ) 1.3 × 10 M cm was used. The fraction
f in eq 5 was obtained from a comparison of the areas under
the emission decay curves for quenched and unquenched
samples, measured by the single-photon-counting equipment.
2: 4,4′-Diethoxycarbonyl-2,2′-bipyridine (diester). Crude
4,4′-dicarboxy-2,2′-bipyridine (1, 855 mg, 3.5 mmol) was
suspended in EtOH (30 mL, 99.5%). Concentrated H SO (1
2
4
mL, 18 M) was added though the reflux condenser (Caution:
strongly exothermic). The mixture was refluxed under argon,
and the solid slowly dissolved (ca. 12 h) gradually forming a
dark clear solution as the reaction proceeded. The formation
of 2 was followed by TLC (Al2O3, hexanes:EtOAc 60:40, 2:
Rf ≈ 0.58, 1: immobile). The reaction solution was poured
onto a mixture of ice and ammonia (NH3(aq), 15 M, 2.4 mL,
36 mmol). The ethanol was removed from the aqueous phase
at the rotary evaporator, and 2 was extracted with CH2Cl2 (4 ×
50 mL). The combined CH2Cl2 phases were dried by shaking
with brine (saturated NaCl(aq)) and evaporated at the rotary
evaporator. The crude product (1.02 g, 3.4 mmol, 97%) was
recrystallized (the first time with charcoal) from ethanol to yield
straw colored needles; repeated recrystallization gave white to
off-white needles (556 mg, 1.9 mmol, 54%) mp 156-157 °C.
A sample purified by sublimation (12 h, ca. 140 °C, pump
vacuum ca. 1 Torr) yielded white filamentous needles, mp
2+
Materials. N-Cetyl-N′-methyl-4,4′-bipyridinium, (CMV ),
+
N-cetyl-4,4′-bipyridinium (CB ), and N-methyl-4,4′-bipyri-
+
dinium (MB ), all dichloride salts, were available from earlier
studies.8
a,b
Methylviologen (MVCl2, Sigma) and egg lecithin
(Grade 1, Lipid Products, Nutfield, Surrey, U.K.) were used as
received. All other chemicals were of highest available purity
and used as received. Ethanol (99.5%, spectroscopic grade)
and 96% (purum) were used as supplied by Kemetyl, Sweden.
Dichloromethane (CH2Cl2), acetone, acetonitrile, N,N-dimeth-
ylformamide (DMF), and methanol of 99+% HPLC grade were
supplied by Labscan (Dublin) and were used as received.
Sulfuric acid, concentrated 95-97%, of grade puriss was
supplied by Kebo AB, Sweden. Hexanes (bp 60-80 °C, Shell,
polymer grade) and ethyl acetate (EtOAc, puriss, Kebo AB,
Sweden) were distilled prior to use. 4,4′-Dimethyl-2,2′-bipy-
ridine (Aldrich, 99%), silver triflate (AgOSO2CF3 or AgOTf,
Lancaster, 98%), RuCl3‚3H2O (Aldrich, 41.71% Ru assay),
potassium tert-butoxide ((CH3)3COK, Aldrich 95% or Lancaster
1
0a
158.5-159 °C [lit. 159 °C ]. High-purity material also can
be obtained by column chromatography on Al2O3, eluent
hexanes:EtOAc 90:10, followed by recrystallization from EtOH.
9
7%), hexafluorophosphoric acid (HPF6, 60 wt %, soluble in
water, Aldrich) were used as received. Oxygen was supplied
by Air Liquide, Sweden.
Elemental analysis: Calculated for C H N O : C, 63.99;
H, 5.37; N, 9.33. Found: C, 63.86; H, 5.27; N, 9.45.
1
6
16
2
4
1
Analyses. The analyses (carbon, hydrogen, and nitrogen,
reported in mass %) were performed by Analytische Labora-
torien GmbH, D-51789 Lindlar, Germany.
H NMR (CDCl as solvent and internal standard): δ 8.95
3
(dd, J ) 0.9, 1.5 Hz, 2H, H-3 and H-3′), 8.87 (dd, J ) 0.9, 4.9
Hz, 2H, H-6 and H-6′), 7.91 (dd, J ) 1.5, 4.9 Hz, 2H, H-5 and
H-5′), 4.46 (q, J ) 7.2 Hz, 4H, (CO)O-CH2CH3), 1.45 (t, J )
1
: 4,4′-Dicarboxy-2,2′-bipyridine, (dcHb). 4,4′-Dimethyl-
,2′-bipyridine (1.00 g, 54 mmol) and potassium tert-butoxide
6.10 g, 54 mmol) were dissolved in DMF (100 mL) under argon
1
3
7
.2 Hz, 6H, (CO)O-CH2CH3). C NMR (CDCl3 as solvent
2
(
and internal standard): δ 165.2, 156.6, 150.1, 139.0, 123.3,
-
1
1
(
20.6, 62.0, 14.3. IR (KBr disk, (cm ), all strong): 1728
in a 500 mL Fisher-Porter reaction flask. Stirring for 10 min
at room temperature afforded the dark blue dianion. The argon
was replaced by oxygen and maintained at 4 atm. for 10-15
min. at room temperature, the solution eventually reaching ca.
CdO), 1557, 1374, 1363, 1289, 1254.
4: [Tris(4,4′-diethoxycarbonyl-2,2′-bipyridine)ruthenium(II)]
II
2+
Bis(triflate), [Ru (diester)3] (OTf)2. An ethanolic solution
1
4
4
5 °C. When the color of the solution had turned from blue
(99.5%,15 mL) of 3a, RuCl2(DMSO)4 (50.3 mg, 0.10 mmol,

7
502 J. Phys. Chem. B, Vol. 101, No. 38, 1997
Hammarstr o¨ m et al.
1
equiv) and AgOTf (54.7 mg, 0.21 mmol, 2.05 equiv) was
5H2O. A sample of 4 (18.2 mg, 0.014 mmol) was dissolved in
acetone (1 mL), and aqueous ammonia (6 mL, 1.5 M) was
added. Stirring for 30 min at room temperature afforded 5; the
progress of the hydrolysis was followed by TLC (vide supra).
The aqueous ammonia was removed at the rotary evaporator,
and the red solid was dissolved in water (5 mL). Addition of
a few drops of 0.1 M HPF6(aq) immediately precipitated 6.
Crystals were obtained by dissolving the precipitate in hot water
with a minimum amount of aqueous ammonia and then adding
stirred and refluxed (12 h) under argon. A clear yellow solution
of 3b was obtained after filtration through Celite (in air) to
remove the AgCl(s). The filtered solution of 3b was added
dropwise over a period of 75-95 min to a refluxing solution
of diester 2 (100 mg, 0.33 mmol, 3.2 equiv) in 20 mL of ethanol.
The solution gradually turned dark red, and the reaction was
followed by TLC (Al2O3, EtOAc:methanol:glacial acetic acid:
water, 15:5:1:1 parts by volume). After 1 h, the yellow 4 (Rf
≈
0.63) and a darker intermediate species (Rf ≈ 0.68) were
0.1 M HPF (aq) to the point where a precipitate was just visible.
Slow cooling of this solution afforded small red crystals of 6
(10.4 mg, 0.009 mmol, 66%) which were collected on a glass
frit and dried at the pump.
6
observed, the amount of 4 slowly increasing, and the darker
decreasing, over the next 24-48 h. (In case the chloride ions
were incompletely removed from the ruthenium precursor 3a,
the formation of byproducts including the blue 7 (Rf ≈ 0.98,
appears violet on the TLC plate or column gel) or green 8 (Rf
Elemental analysis: Calculated for Ru(dcHb)3(PF6)2‚5H2O
C36H34N6O17P2F12Ru): C, 35.63; H: 2.82; N, 6.92. Found:
(
≈
0.95, appears deep purple on the TLC plate or column gel)
C, 35.53; H, 2.84; N, 6.84.
and a pink species 9 (Rf ≈ 0.93) was also observed). After 48
h, the crude reaction mixture was filtered hot, then reduced to
one-quarter volume at the rotary evaporator, cooled in ice until
crystallization of 4 was complete (1-3 h), and then collected
on Celite. Washing with cold ethanol (99.5%) leaves the
crystalline 4 on the Celite together with any further traces of
AgCl (the washings were retained for further chromatography
yielding a small second crop; vide infra). The product 4 was
then dissolved from the Celite bed by washing with CH2Cl2,
yielding a bright-red, luminescent solution. Addition of diethyl
ether precipitated 4 as a bright red solid (76.6 mg, 0.058 mmol,
UV-Vis: (Figure 3) λmax 466 nm, for a dissolved sample,
4
-
-1
-1
Ru(dcb)3 , ꢀ 24 800 M cm (1 mM NaOH, pH ) 11). The
1
H NMR and 13C NMR of a dissolved sample was identical
with those of 5 (vide supra). IR (KBr disk, (cm )): 3110 (w),
977 (w), 2926 (w), 1716 (s) (CdO), 1612 (s), 1368 (s), 1313
-1
3
(s), 1263 (s), 1232 (s).
Chromatographic Separation of Compounds 4 and 7-10.
The purple ethanolic washings obtained during the isolation of
4
may be further purified, and a small second crop of 4 obtained,
including material where one or several of the ester groups of
have been hydrolyzed. This material was best converted
4
5
7%). Recrystallization from CH2Cl2 by vapor diffusion of
directly to 6 (vide supra). Careful chromatography using a very
short column and a multistep gradient afforded products (7-
diethyl ether in a closed vessel yielded larger crystals of X-ray
quality.39 The point group symmetry for the solution species
1
0), two of which have been identified (7 and 8). Conditions
is D3. Elemental analysis: Calculated for C50H48N6O18S2F6-
Ru: C, 46.19; H: 3.72; N, 6.46. Found: C, 45.93; H, 3.61; N,
were alumininum oxide, slurry-packed in hexanes, column
diameter 4 cm, height 1-1.5 cm. Gradient elution, with approx-
imately 50 mL of each solvent composition: hexanes:EtOAc:
6
.41.
+
II
+
ESI-MS: M , [Ru (diester)3(OTf)] , observed peak maxi-
1
00% hexanes, 25% EtOAc, 50% EtOAc; 75% EtOAc, 100%
mum mass 1151.183 m/z, simulated peak maximum mass for
C49H48N6O15F3SRu, 1151.191. For mass spectra, see Supporting
Information.
EtOAc; EtOAc:CH2Cl2: 12.5% CH2Cl2; 25% CH2Cl2; 50% CH2-
Cl2; 100% CH2Cl2; CH2Cl2:EtOH: 0.37% EtOH; 0.75% EtOH,
1
% EtOH; 1.25% EtOH; 1.5% EtOH; 3% EtOH; 6% EtOH;
-
1
-1
UV-Vis: λmax 466 nm, ꢀ 24 500 M cm (CH2Cl2), 22 400
CH2Cl2:MeOH: 1.25% MeOH; 2.5% MeOH; 5% MeOH; 10%
MeOH; 25% MeOH; 50% MeOH; 100% MeOH; 99% MeOH;
-1
-1
40
-1
-1
M
cm (acetone), lit. 23 300 M cm (EtOH:MeOH 4:1).
H NMR (CDCl3 as solvent and internal standard): 8.89 (dd,
1
1% 2 M NaOH. The volume used of each of the above mixtures
J ) 0.6, 1.7 Hz, 6 H, H-3), 8.16 (dd, J ) 0.6, 5.8 Hz, 6 H,
H-6), 8.06 (dd, J )1.7, 5.8 Hz, 6 H, H-5), 4.48 (q, J ) 7.2 Hz,
was continuously adjusted as the colored products were observed
leaving the column; the mixture suitable for the compound
eluting was added in sufficient volume (up to 300 mL), i.e.,
until the compound had left the column (to a sufficient extent).
Any remaining excess diester 2 was eluted in the hexanes:EtOAc
system. Next, the blue trans-dichloro compound 7 (light blue
on the gel, violet-blue in solution) was eluted, beginning by
12.5% CH Cl in EtOAc. This was followed by the cis-dichloro
1
2 H, (CO)O-CH2CH3), 1.42 (t, J ) 7.2 Hz, 18 H, (CO)O-
13
CH2CH3). C NMR (CDCl3 as solvent and internal standard):
1
1
62.9, 156.3, 154.3, 139.8, 128.1, 123.0, 63.2, 14.1.
IR (KBr disk, (cm ), all strong): 1726 (CdO), 1490, 1369,
271 and 1141 (C-F).
-
1
5
: Tetrasodium[tris(4,4′-dicarboxylate-2,2′-bipyridine)ruthe-
2
2
II
4-
nium(II)], [Ru (dcb)3] , Prepared in Situ. The established
compound (dark purple on the gel, dark green in solution) 8
forming a sharp band at 0.75% EtOH, being completely eluted
by sufficient volumes of 1% EtOH; the pink species 9 forming
a sharp band at 1.25% EtOH (separating from the yellow product
band) and being eluted at 1.5% EtOH. The yellow to bright
red (depending on the concentration) product 4 fractions were
obtained from around 2.5% MeOH and onward (late fractions
with one or more ester groups hydrolyzed). A late purple
fraction 10, likely material where one or several of the ester
groups of 9 have been hydrolyzed, was obtained from the last
MeOH and MeOH/NaOH mixtures.
41
method (e.g., Furlong et al. ) was used: A 2-3 mg sample of
was stirred in a few milliliters of 10 mM NaOH (pH ) 12)
4
for 2 h at room temperature. The conversion of 4 to 5 was
followed by TLC (vide supra) and was essentially complete
within 1 h. For the photophysical quenching studies this
solution was diluted to pH ) 11 (1 mM NaOH).
UV-Vis: λmax 466 nm for Ru(dcb)3 , ꢀ 22 000 M cm-1
4
-
-1
1 mM NaOH, pH ) 11), lit.3 26 000 M cm (1 M NaOH).
8a
-1
-1
(
The point group for the solution species is D3.
1H NMR (D2O made alkaline with NaOH; vide supra): δ
8
6
.81 (dd, J ) 0.6, 1.8 Hz, 6 H, H-3), 7.81 (dd, J ) 0.6, 5.8 Hz,
7: trans-DichloroRu(II)(4,4′-diethoxycarbonyl-2,2′-bipyri-
1
3
II
+
H, H-6), 7.62 (dd, J ) 1.8, 5.8 Hz, 6 H, H-5). C NMR
2 2 2 2
dine) , [trans-Cl Ru (diester) ]. ESI-MS: M , [Ru(diester) -
+
(
1
D2O made alkaline with NaOH): 173.6, 159.9, 154.4, 148.1,
28.8, 125.8.
: [Tris(4,4′-dicarboxy-2,2′-bipyridine)ruthenium(II)] Bis-
Cl(CH CN)] , one Cl replaced by acetonitrile in the spectrom-
3
eter; simulated peak maximum mass for C H N O ClRu:
3
4
35
5
8
6
778.122 m/z, observed 778.6 (not high resolution). For spectra,
see Supporting Information.
II
2+
(
hexafluorophosphate) Pentahydrate, [Ru (dcHb)3] (PF6)2‚-

Emission Quenching and Charge Separation
J. Phys. Chem. B, Vol. 101, No. 38, 1997 7503
TLC, Rf ≈ 0.98, appears violet on the TLC plate or column
gel (vide supra).
(dd, J ) 1.7, 5.7 Hz, 1 H, H-5), 7.53 (dd, J ) 1.8, 5.9 Hz, 1 H,
H-5), 4.58 (dq, J ) 2.4, 7.2 Hz, 4 H, (CO)O-CH2CH3), 4.43
(dq, J ) 2.4, 7.2 Hz, 4 H, (CO)O-CH2CH3), 1.52 (dt, J ) 2.4,
7.2 Hz, 6 H, (CO)O-CH2CH3), 1.39 (dt, J ) 1.2, 7.0 Hz, 6 H,
(CO)O-CH2CH3).
UV-Vis (see Figure 3): λmax at 463, 520, 592, and 682 nm
in CH2Cl2). The molar absorptivity ꢀ could not be determined
(
due to lack of material, but was estimated to be similar to that
of 8 (vide infra) by comparison of the relative intensities of the
π-to-π* transition and the MLCT transitions. No steady-state
emission was observed upon irradiation at 520, 592, or 682 nm.
Irradiation at 463 nm gave a very weak emission residue with
For NMR and UV-visible spectra of 10, please see Sup-
porting Information.
Acknowledgment. This work was supported by The Knut
and Alice Wallenberg Foundation and The Swedish Natural
Science Research Council. T.N. acknowledges a personal grant
from Vattenfall AB, Sweden. We thank Mikael Andersson,
Uppsala University, for assistance with some of the steady-state
emission studies and Anna B o¨ rje, Royal Institute of Technology,
for the 2D NMR study.
a band structure similar to that of 4, λmax em 630 nm. The point
group of the solution species is D4h. 1H NMR (CDCl3 as solvent
and internal standard): δ 9.49 (d, J ) 5.9 Hz, 4 H, H-6); 8.79
(
4
d, J ) 1.8 Hz, 4 H, H-3); 8.09 (dd, J ) 1.7, 5.9 Hz, 4 H, H-5);
.55 (q, J ) 7.1 Hz, 8 H, (CO)O-CH2CH3); 1.51 (t, J ) 7.1
Hz, 12 H, (CO)O-CH2CH3).
13C NMR (CDCl3 as solvent and internal standard): 164.0,
Supporting Information Available: ESI-MS spectra of
compounds 4, 7, and 8. UV-vis spectra comparing 9 and 10
and parts of the NMR spectra comparing compounds 4, 9, and
1
61.2, 154.8, 137.4, 122.9, 121.2, 62.6, 14.3.
8
: cis-Dichlororuthenium(II)(4,4′-diethoxycarbonyl-2,2′-bi-
II
+
pyridine)2, [cis-Cl2Ru (diester)2]. ESI-MS: M , [Ru(diester)2-
Cl(CH3CN)] , observed 778.655 m/z, simulated peak maximum
1
0 (6 pages). Ordering information is given on any current
+
masthead page.
mass for C34H35N5O8ClRu: 778.122 m/z, alternatively [Ru-
+
(
diester)2(OAc)(H2O)] , simulated peak maximum. Mass for
References and Notes
C34H37N4O11Ru: 779.151 m/z. For other ions and spectra, see
Supporting Information.
(1) See e.g. (and references therein): (a) Kurreck, H.; Huber, M. Angew.
Chem., Int. Ed. Engl. 1995, 34, 849. (b) Sauvage, J.-P.; Collin, J.-P.;
Chambron, J.-C.; Guillerez, S.; Coudret, C.; Balzani, V.; Barigelletti, F.;
De Cola, L.; Flamigni, L. Chem. ReV. (Washington, DC) 1994, 94, 993. (c)
Gust, D.; Moore, T. A.; Moore, A. L. Acc. Chem. Res 1993, 26, 198. (d)
Wasielewski, M. R. Chem. ReV. (Washington, DC) 1992, 92, 435. (e)
Balzani, V.; Scandola, F. Supramolecular Photochemistry; Ellis Horwood:
UK, Chichester, 1991.
TLC, Rf ≈ 0.95, appears deep purple on the TLC plate or
column gel (vide supra).
-
1
UV-Vis (see Figure 3) λmax at 430 nm (ꢀ )14 200 M
-
1
-1
-1
cm ) and 586 nm (ꢀ ) 14 100 M cm , both in CH2Cl2).
No steady-state emission was observed upon irradiation into
either of the visible bands. The point group of the solution
species is C2, and the assignment of the proton resonances is
(2) For recent examples of noncovalent, pairwise assembly of donors
and acceptors, see refs 2a-d and references therein. (a) Hayashi, T.;
Takimura, T.; Ogoshi, H. J. Am. Chem. Soc. 1995, 117, 7, 11606. (b) Kr a´ l,
V.; Springs, S. L.; Sessler, J. L. J. Am. Chem. Soc. 1995, 117, 7, 8881. (c)
Berman, A.; Izraeli, E. S.; Levanon, H.; Wang, B.; Sessler, J. L. J. Am.
Chem. Soc. 1995, 117, 8252. (d) Roberts, J. A.; Kirby, J. P.; Nocera, D. G.
J. Am. Chem. Soc. 1995, 117, 8051. (e) Benniston, A. C.; Harriman, A. J.
Am. Chem. Soc. 1994, 116, 11531.
supported by 2D NMR.
1H NMR (500 MHz, CDCl3 as solvent and internal stan-
dard): δ 10.47 (d, J ) 6.0 Hz, 2 H, H-6′), 8.86 (d, J ) 2.1 Hz,
2 H, H-3′), 8.69 (d, J ) 2.1 Hz, 2 H, H-3), 8.19 (dd, J ) 1.7,
6.0 Hz, 2 H, H-5′), 7.73 (d, 6.0 Hz, 2 H, H-6), 7.52 (dd, J )
1.7, 6.0 Hz, 2 H, H-5), 4.60 (q, J) 7.1 Hz, 4 H, (CO)O-CH′2-
(3) Sun, L.; Hammarstr o¨ m, L.; Norrby, T.; Berglund, H.; Davydov;
R.; Andersson, M.; B o¨ rje, A.; Korall, P.; Philouze, C.; Almgren, M.; Styring,
S.; A° kermark, B. J. Chem. Soc., Chem. Commun. 1997, 607.
CH3), 4.45 (q, J ) 7.1 Hz, 4 H, (CO)O-CH2CH3), 1.54 (t, J )
.1 Hz, 6 H, (CO)O-CH2CH′3), 1.41 (t, J ) 7.1 Hz, 6 H,
(4) Several examples exist where electrostatic forces have been used
7
(
to assemble sensitizers and acceptors at surfactant aggregates. For reviews,
see e.g.: (a) Fox, M. A. Photoinduced Electron-Transfer III; Top. Curr.
Chem.; Springer-Verlag: Berlin, 1991; Vol. 159, p 68. (b) Willner, I,;
Willner, B. Photoinduced Electron-Transfer III; Top. Curr. Chem.;
Springer-Verlag: Berlin, 1991; Vol. 159, p 153. (c) Rabani, J. In Photo-
induced Electron Transfer, Part B; Fox, M. A., Chanon, M., Eds.;
Elsevier: Amsterdam, 1988; p 642. (d) Gr a¨ tzel, M. Heterogeneous
Photochemical Electron Transfer; CRC Press: Boca Raton, FL, 1989. (e)
Matsuo, T. J. Photochem. 1985, 29, 41.
CO)O-CH2CH3). 1 C NMR (125 MHz, CDCl3 as solvent and
3
internal standard): 164.6, 164.0, 160.8, 158.5, 155.6, 152.9,
1
36.8, 135.5, 125.4, 124.9, 122.3, 122.0. 63.0, 62.9, 14.8, 14.6.
II
9: Ru (diester)2(Cl)(X). Another minor product, a pink
species 9 (Rf ≈ 0.93) of point group symmetry C1 was also
1
isolated. By the appearance of the H NMR spectrum, we
1
3
(5) (a) Rehm, O.; Weller, A. Ber. Bunsen-Ges. Phys. Chem. 1969, 73,
34. (b) Kalyanasundaram, K. Coord. Chem. ReV. 1982, 46, 159. (c)
Serpone, N. In Photoinduced Electron Transfer, Part D; Fox, M. A.,
conclude that it contains two inequivalent diester ligands ( C
NMR spectra could not be obtained because of lack of material).
TLC, Rf ≈ 0.93, appears pink on the TLC plate or column gel
8
Chanon, M., Eds.; Elsevier: Amsterdam, 1988; p 47.
(
6) Hammarstr o¨ m, L.; B o¨ rje, A.; Norrby, T., unpublished results.
(
4
vide supra).The UV-visible spectrum (see Figure 3, λmax at
(7) (a) Lange, Y. In The Physical Chemistry of Lipids, Handbook of
08, 547 nm) appears similar to that of 8, with two major bands
Lipid Research; Small, D. M., Ed.; Plenum Press: New York, 1986; Vol.
, p 523. (b) Blume, A. In Phospholipids Handbook; Cevc, G., Ed.; Marcel
in the visible region. The λmax of 9 is blue-shifted ca. 40 nm
compared to that of 8, suggesting that one of the chloride ligands
has been replaced by a ligand that better stabilizes the metal
d-orbitals (HOMO). The molar absorptivity ꢀ could not be
determined due to lack of material but was estimated to be
similar to that of 8 (vide supra) by comparison of the relative
intensities of the π-to-π* transition and the MLCT transitions.
4
Dekker: New York, 1993; p 455. (c) For monolayer values, see refs 7d
and 27 and references therein. (d) Li, L. Patterson, L. K. J. Phys. Chem.
1
995, 99, 16149.
(8) (a) Hammarstr o¨ m, L.; Almgren, M.; Norrby, T. J. Phys. Chem.
1992, 96, 6, 5017. (b) Hammarstr o¨ m, L.; Almgren, M.; Lind, J.; Merenyi,
G.; Norrby, T.; A° kermark, B. J. Phys. Chem. 1993, 97, 10083. (c)
Hammarstr o¨ m, L.; Berglund, H.; Almgren, M. J. Phys. Chem. 1994, 98,
9
588. (d) Hammarstr o¨ m, L.; Almgren, M. J. Phys. Chem. 1995, 99, 11959.
(9) (a) Deisenhofer, J.; Norris, J. R. (Eds) The Photosynthetic Reaction
1
H NMR (CDCl3 as solvent and internal standard): δ 10.41
Centre; Academic Press: San Diego, 1993; Vol. 1-2. (b) Deisenhofer, J.;
Michel, H. Angew. Chem., Int. Ed. Engl. 1989, 28, 8, 829. (c) Huber, R.
Angew. Chem., Int. Ed. Engl. 1989, 28, 848.
(10) (a) Launikonis, A.; Lay, P. A.; Mau, A. W.-H.; Sargeson, A. M.;
Sasse, W. H. F. Aust. J. Chem. 1986, 39, 1053-1062. (b) Sprintschnik, G.;
Sprintschnik, H. W.; Kirsch, P. P.; Whitten, D. G. J. Am. Chem. Soc. 1977,
(dd, J ) 0.7, 5.9 Hz, 1 H, H-6), 10.32 (dd, J ) 0.7, 5.9 Hz, 1
H, H-6), 8.83 (s, J ≈ 0.7, 1.8 Hz (distorted by overlap) 2 H,
H-3 and H-3); 8.71 (dd, J ) 0.7, 1.8 Hz, 1H, H-3), 8.68 (dd, J
)
0.7, 1.8 Hz, 1H, H-3), 8.18 (dd, J ) 1.6, 5.9 Hz, 1H, H-5,
overlapping with peak at δ 8.17), 8.17 (dd, J ) 1.6, 5.9 Hz,
9
9, 4947-4954.
11) (a) Garelli, N.; Vierling, P. J. Org. Chem. 1992, 57, 3046-3051.
(b) Oki, A. M.; Morgan, R. J. Synth. Commun. 1995, 25, 4093-4097.
1
H, H-5, overlapping with peak at δ 8.18), 7.81 (dd, J ) 0.7,
.9 Hz, 1 H, H-6), 7.66 (dd, J ) 0.7, 5.9 Hz, 1 H, H-6), 7.59
(
5

7504 J. Phys. Chem. B, Vol. 101, No. 38, 1997
Hammarstr o¨ m et al.
(
12) Bartok, W.; Rosenfeld, D. D.; A. Schriesheim, A. J. Org. Chem.
963, 28, 410-412. This type of oxidation is also described, without
experimental details, in: Potts, K. T. Bull. Soc. Chim. Belg. 1990, 99, 741-
68.
13) N.B. All point group notations refer to a simplified ligand group
(24) See e.g.: Lakowicz, J. R. Principles of Fluorescence Spectroscopy;
Plenum Press: New York, 1983.
(25) See e.g.: Israelachvili, J. Intermolecular and Surface Forces, 2nd
ed.; Academic Press: London, 1992.
1
7
(
(26) Owen, C. S. J. Chem. Phys. 1975, 62, 3204.
orbital approach, reflecting the effective symmetry experienced by the
hydrogens of the ligand system as observed in the H NMR spectra of the
solution species. See for instance: Kettle, S. F. A. Symmetry and Structure,
(27) Caruso, F.; Grieser, F.; Murphy, A.; Thistlethwaite, P.; Urquhart,
R.; Almgren, M.; Wistus, E. J. Am. Chem. Soc. 1991, 113, 4838.
(28) (a) Chiorboli, C.; Indelli, M. T.; Scandola, M. A. R.; Scandola, F.
J. Phys. Chem. 1988, 92, 156. (b) In ref 28a, the value of r ) 1.03 nm was
1
2
nd ed.; Wiley: Chichester, 1995.
14) Evans, I. P.; Spencer, A.; Wilkinson, G. J. Chem. Soc., Dalton
Trans. 1973, 204-209.
15) (a) Alessio, E.; Mestroni, G.; Nardin, G.; Attia, W. M.; Calligaris,
2+
2+
(
used for the couple Ru(bpy)3 -MV , and in the present study, substitution
-
of -H by -COO was assumed to contribute by roughly 0.1 nm. Due to
(
the membrane, the nonspherical viologen headgroups were probably not
randomly oriented in the encounter complexes, which makes the value r )
1.1 nm somewhat uncertain. However, for the purpose of the present
investigation this was not critical.
M.; Sava, G.; Zorzet, S. Inorg. Chem. 1988, 27, 4099-4106. (b) Alessio,
E.; Balducci, G.; Calligaris, M.; Costa, G.; Attia, W. M.; Mestroni, G. Inorg.
Chem. 1991, 30, 609-618.
(16) Bossmann, S. H.; Ghatlia, N. D.; Ottaviani, M. F.; Turro, C.; D u¨ rr,
(29) Medhage, B.; Almgren, M. J. Fluorescence 1992, 2, 7.
(30) Watanabe, T.; Honda, K. J. Phys. Chem. 1982, 86, 2617.
(31) Pileni, M.-P.; Braun, A. M. Photochem. Photobiol. 1980, 31, 423.
(32) (a) Prasad, D. R.; Hoffman, M. Z.; Mulazzani, Q. G.; Rodgers, M.
A. J. J. Am. Chem. Soc. 1986, 108, 5135. (b) Stramel, R. D.; Thomas, J. K.
J. Chem. Soc., Faraday Trans. 2 1986, 82, 799. (c) Jones, G., II; Zisk, M.
B. J. Org. Chem. 1986, 51, 947.
H.; Turro, N. J. Synthesis 1996, 1313-1318.
17) Coe, B. J.; Friesen, D. A.; Thompson, D. W.; Meyer, T. J. Inorg.
Chem. 1996, 35, 4575-4584 and referenced therein.
18) (a) Juris, A.; Balzani, V.; Barigelletti, F.; Campagna, S.; Belser,
(
(
P.; von Zelewsky, A. Coord. Chem. ReV. 1988, 84, 85. (b) Kalyanasundaram,
K. Photochemistry of Polypyridine and Porphyrin Complexes; Academic
Press: London, 1992.
(33) Marcus, R. A. Annu. ReV. Phys. Chem. 1964, 15, 155.
(34) (a) Ohno, T.; Yoshimura, A.; Prasad, D. R.; Hoffman, M. Z. J.
Phys. Chem. 1991, 95, 4723. (b) Yonemoto, E. H.; Riley, R. L.; Kim, Y.
I.; Atherton, S. J.; Schmehl, R. H.; Mallouk, T. H. J. Am. Chem. Soc. 1992,
114, 8081.
(19) A biexponential decay was reported in ref 20, with the shorter
lifetime between 1 and 20 ns at pH ) 0-7. This was attributed to a
protonation/deprotonation equilibration in the excited state. However, the
intermediate, uncharged form was a poorer emitter than both the fully
protonated and the fully dissociated forms. Therefore, such an equilibration
would have produced both decreasing and increasing components in the
pH interval examined by the authors. Also, the reported lifetime of 20 ns
is to short to be explained by a protonation at pH ) 7, and the fraction of
even singly protonated excited species is expected to be <<1% at pH ) 7
and therefore not be detectable in the emission experiments. Thus, it is
likely that the fast component of the biexponential decay reported in ref 20
may rather be explained by an impurity or an experimental artifact.
II/III
4-
(35) The Ru
reduction potential is 0.1 V more positive for Ru(dcb)3
,
than for Ru(bpy)32
difference.
+ 20
and the work term5 contributes to another 0.1 V
(36) Roullier, L.; Laviron, E. Electrochim. Acta 1978, 23, 773.
(37) (a) Cevc, G. In Phospholipids Handbook; Cevc, G., Ed.; Marcel
Dekker: New York, 1993; p 639. (b) Cevc, G.; Marsh, D. Phospholipid
Bilayers; Wiley: New York, 1987.
(38) (a) Dabestani, R.; Bard, A. J.; Campion, A.; Fox, M. A.; Mallouk,
T. E.; Webber, S. E.; White, J. M. J. Phys. Chem. 1988, 92, 1872. (b)
Yoshimura, A.; Hoffman, M. Z.; Sun; H. J. Photochem. Photobiol., A:
Chem. 1993, 70, 29.
(39) An X-ray investigation is being pursued presently.
(40) Cook, M. J.; Thomson, A. J. Chem. Br. 1984, 914-917.
(41) Furlong, N. D.; Wells, D.; Sasse, W. H. F. J. Phys. Chem. 1986,
90, 1107-1115.
(
20) Park, J. W.; Ahn, J.; Lee, C. J. Photochem. Photobiol., A: Chem.
995, 86, 89.
21) Ireland, J. F.; Wyatt, P. A. H. AdV. Phys. Org. Chem. 1976, 12,
31.
22) Hoffman, M. Z.; Bolletta, F.; Moggi, L.; Hug, G. L. J. Phys. Chem.
Ref. Data 1989, 18, 219.
23) Stern, O.; Vollmer, M. Phys. Z. 1919, 20, 183.
1
(
1
(
(