Photoinduced electron transfer at liquid/liquid interfaces. Part III. Photoelectrochemical responses involving porphyrin ion pairs

Article abstract of DOI:10.1021/ja992215i

The interfacial photoreactivity of water soluble porphyrin ion pairs was explored by photocurrent measurements at the water/1,2-dichloroethane interface. The anionic zinc meso-tetrakis(p-sulfonatophenyl)-porphyrin and cationic zinc tetrakis(N-methylpyridy

Full text of DOI:10.1021/ja992215i

J. Am. Chem. Soc. 1999, 121, 10203-10210
10203
Photoinduced Electron Transfer at Liquid/Liquid Interfaces. Part III.
Photoelectrochemical Responses Involving Porphyrin Ion Pairs
David J. Ferm´ın, H. Dung Duong, Zhifeng Ding, Pierre-Franc¸ois Brevet, and
Hubert H. Girault*
Contribution from the Laboratoire d’Electrochimie, Departement de Chimie, Ecole Polytechnique Fe´de´rale
de Lausanne, CH-1015 Lausanne, Switzerland
ReceiVed June 28, 1999
Abstract: The interfacial photoreactivity of water soluble porphyrin ion pairs was explored by photocurrent
measurements at the water/1,2-dichloroethane interface. The anionic zinc meso-tetrakis(p-sulfonatophenyl)-
porphyrin and cationic zinc tetrakis(N-methylpyridyl)porphyrin undergo spontaneous association in solution,
leading to electrically neutral surface active heterodimer species. Heterogeneous quenching of the porphyrin
ion pair was observed in the presence of the electron donor decamethylferrocene and the electron acceptor
tetracyanoquinodimethane. No substantial photoresponses were detected in the presence of only one of the
porphyrin species. Fundamental aspects on the dimer structure and photoreactivity were extracted from the
photocurrent dependence on the concentration of both porphyrins. Photocurrent spectra as well as analysis of
the association constant indicate that the porphyrin species are intimately associated as inner sphere ion pairs.
Relevant issues in connection to back electron transfer, heterodimer adsorption, and orientation at the interface
are also exposed.
Introduction
the four-electron reduction of oxygen.8 The heterodimer species
were adsorbed on glassy carbon electrodes. Interestingly, the
pair involving CoTMPyP and the free ligand H2TPPS showed
catalytic properties only for the two-electron mechanism.8 In a
previous contribution, we have introduced photocurrent re-
sponses at the polarized water/1,2-dichloroethane (DCE) inter-
face featuring the heterodimer ZnTPPS-ZnTMPyP.9 Under
similar conditions, this heterodimer exhibits photocurrent re-
sponses up to 50 times larger than those observed for the anionic
zinc tetrakis(carboxyphenyl) porphyrin (ZnTPPC).10-12 In the
present manuscript, the nature of these photoresponses is
examined in the presence of the hydrophobic quenchers deca-
methylferrocene (DCMFc) and tetracyanoquinodimethane
(TCNQ). The photocurrent dependence on the concentration of
each of the porphyrin species suggests that the association
process occurs in solution followed by adsorption of the
porphyrin pair at the liquid/liquid junction.
Naturally occurring porphyrin dimer plays an essential role
in photosynthetic centers in bacteria such as Rhodopseudomonas
Viridis.1,2 The special pair of bacteriochlorophylls exhibits
efficient charge transfer to quinone species within a fraction of
a nanosecond. Extensive research on the homogeneous photo-
chemistry and photophysics has been carried out mainly in
covalently linked porphyrin resembling the structure of the
special pair.3,4 An alternative type of dimer can be formed by
spontaneous association of cationic and anionic porphyrin
species as reported by Linschitz et al.5-7 Flash photolysis and
time-resolved EPR spectroscopy of the pair zinc meso-tetrakis-
[4-trimethylanilinium] porphyrin (ZnTTAP) and copper meso-
tetrakis(p-sulfonatophenyl) porphyrin (CuTPPS) suggested the
formation of localized triplet states upon illumination, featuring
lifetime of 30 µs in water/acetone mixtures.7 It has also been
considered that a small fraction of the dimer species exhibits
the appropriate porphyrin orientation for electron transfer
involving a singlet precursor.7
Experimental Section
The electrochemical cell employed for all experiments is represented
in Figure 1. The organic phase supporting electrolyte was bis(triphenyl-
phosphoranylidene) ammonium tetrakis (4-chlorophenyl) borate
(BTPPATPBCl). Details on the electrolyte preparation and photoelec-
trochemical set up have been given previously.9-12 The surface area of
the liquid|liquid junction was 1.53 cm2. Illumination was provided by
a green He-Ne laser (543 nm) or by a 450 W arc-Xe lamp in
The interfacial reactivity of this type of dimer is largely
unknown. Recently, D’Souza and co-workers have shown that
cobalt tetrakis(N-methylpyridyl)porphyrin (CoTMPyP) and
CoTPPS also form ion pairs with catalytic properties toward
(1) Deisenhofer, J.; Epp, O.; Miki, K.; Huber, R.; Michel, H. Nature
1985, 618.
(2) Chambron, J.-C.; Chardon-Noblat, S.; Harriman, A.; Heitz, V.;
Sauvage, J.-P. Pure Appl. Chem. 1993, 65, 2343-2349.
(3) Kalyanasundaram, K. Photochemistry of Polypyridine and Porphyrin
Complexes; Academic Press: London, 1992.
conjunction with
a grating monochromator. Sodium tetrakis-
(8) D’souza, F.; Hsieh, Y. Y.; Deviprasad, G. R. Chem. Commun. 1998,
1027-1028.
(9) Ferm´ın, D. J.; Duong, H.; Ding, Z.; Brevet, P.-F.; Girault, H. H.
Electrochem. Commun. 1999, 1, 29-32.
(10) Ferm´ın, D. J.; Ding, Z.; Duong, H. D.; Brevet, P. F.; Girault, H. H.
J. Chem. Soc., Chem. Commun. 1998, 1125-1126.
(11) Ferm´ın, D. J.; Ding, Z.; Duong, H.; Brevet, P.-F.; Girault, H. H. J.
Phys. Chem. B 1998, 102, 10334-10341.
(12) Ferm´ın, D. J.; Duong, H.; Ding, Z.; Brevet, P.-F.; Girault, H. H.
Phys. Chem. Chem. Phys. 1999, 1, 1461-1467.
(4) Balzani, V.; Scandola, F. Supramolecular Photochemistry; Hor-
wood: Chichester, 1991.
(5) Ojadi, E.; Selzer, R.; Linschitz, H. J. Am. Chem. Soc. 1985, 107,
7783-7784.
(6) van Willigen, H.; Das, U.; Ojadi, E.; Linschitz, H. J. Am. Chem.
Soc. 1985, 107, 7784-7785.
(7) Hugerat, M.; Levanon, H.; Ojadi, E.; Biczok, L.; Linschitz, H. Chem.
Phys. Lett. 1991, 181, 400-406.
10.1021/ja992215i CCC: $18.00 © 1999 American Chemical Society
Published on Web 10/14/1999

10204 J. Am. Chem. Soc., Vol. 121, No. 43, 1999
Ferm´ın et al.
Figure 1. Schematic representation of the electrochemical cell.
BTPPATPBCl stands for bis(triphenyl-phosphoranylidene) ammonium
tetrakis (4-chlorophenyl) borate.
Scheme 1. Water Soluble Porphyrin Species Employed as
Sensitizers at the Water |DCE Interfacea
f
a
The group R in the meso position corresponds to the functional
groups shown.
(carboxyphenyl) zinc porphyrin, sodium meso-tetra-sulfonatophenyl zinc
porphyrin and meso-tetra-N-methyl-4-pyridium zinc porphyrin tosylate
were purchased from Porphyrin Inc. The structures of these porphyrin
species are displayed in Scheme 1. All Galvani potential differences
Figure 2. Cyclic voltammograms obtained in the dark for 5 × 10-6
mol dm-3 ZnTPPS (a) and 5 × 10-6 mol dm-3 ZnTMPyP (b) at 10,
20, 40, 60, and 80 mV s-1. No quencher species was introduced in the
organic phase. Transfer responses associated with of 5 × 10-6 mol
dm-3 ZnTMPyP (c) in the presence of 5 × 10-7 (‚‚‚), 5 × 10-6 (s),
were estimated by taking the formal transfer potential of the standard
ions tetramethylammonium and tetrapropylammonium as ∆wo φ°′
)
+
TMA
0.160 V and ) ∆wo φ°′ -0.093 V.13
+
TPA
and 5 × 10-5 mol dm-3 (- - -) of ZnTPPS at 20 mV s-1
.
Results and Discussions
The magnitude of this response increases with increasing scan
rate, extending over a potential range of more than 200 mV at
80 mV s-1. This voltammetric signal is related to the adsorption
of the porphyrin species at the liquid/liquid junction. Adsorption
of ZnTPPS4- has been confirmed by differential capacitance
measurements. For a concentration of 5 × 10-5 mol dm-3, the
potential of minimum capacitance is shifted by 75 mV toward
positive potentials with respect to the value obtained for the
same interface in the absence of the porphyrin species, (10 (
5 mV). The shift in the potential of zero charge is approximately
3 times smaller than that observed for the porphyrin ZnTPPC4-
under similar conditions.11
Transfer Potential of the Porphyrins ZnTPPS and
ZnTMPyP. In the cell represented in Figure 1, the ionic
porphyrin species are initially present in the aqueous phase.
Upon polarization across the liquid/liquid junction, these por-
phyrins can be transferred from water to DCE. The formal
transfer potential for each of the ionic species is not only
determined by their charge but also by the difference in solvation
energy in both phases.14 The solubility of the porphyrins
ZnTPPS4-, ZnTPPC4-, and ZnTMPyP4+ is rather negligible in
DCE. In Figure 2a, the voltammetric responses associated with
the transfer of ZnTPPS4- exhibit rather sharp features as a
function of the scan rate. By convention, transfer of a negative
charges from water to the organic phase is associated with
negative currents. The separation between the transfer peaks at
slow scan rates approaches the expected 15 mV for a tetravalent
Rather complicated voltammograms were also obtained in
the presence of ZnTMPyP4+ as displayed in Figure 2b. Two
responses are clearly observed within the potential window. The
response at negative potentials corresponds to the transfer of
tosylate present as counterion of the cationic porphyrin. Cyclic
voltammograms in the presence of sodium tosylate show a
anion. The position of the transfer response establishes a formal
transfer potential of ∆wo φ°′
) -0.22 ( 0.01 V. At slow
4-
ZnTPPS
similar response with a formal transfer potential ∆wo φ°′
)
tosylate
potential sweep rates, it is also observed that the positive peak
is followed by a second smaller response at about -0.150 V.
-0.28 ( 0.01 V. At more positive potentials, the transfer of
ZnTMPyP4+ is observed featuring the convolution of two or
more responses in a narrow potential range. At 10 mV s-1, a
(13) Shao, Y. Ph.D. Thesis, University of Edinburgh: Edinburgh, 1991.
(14) Girault, H. H. In Modern Aspects of Electrochemistry; Bockris, J.
O. M., Conway, B. E., White, R. E., Eds.; Plenum Press: New York, 1993;
pp 1-62.
peak to peak separation of about 15 mV is observed, providing
a formal transfer potential of ∆wo φ°′
) 0.05 ( 0.03 V.
4+
ZnTMPyP

Interfacial Photoinduced Electron Transfer
J. Am. Chem. Soc., Vol. 121, No. 43, 1999 10205
As the scan rate is increased, postpeak features are also
developed. Transposing the analysis developed by Wopschall
and Shain for redox processes involving adsorbed species,15 the
presence of postpeak responses is connected to strongly adsorbed
reactant species. In this sense, these voltammetric features may
suggest that the transfer of ZnTMPyP4+ is also preceded by an
adsorption step. Contrary to the cases of ZnTPPC4- and
ZnTPPS4- studies of the excess charge associated with the
porphyrin adsorption by impedance measurements are rather
complex due to the proximity of the transfer potential range.
On the other hand, preliminary studies by potential modulated
fluorescence spectroscopy and optical surface second harmonic
generation in total internal reflection from the organic phase
suggest the presence of adsorbed ZnTMPyP4+ at potentials more
negative than the formal transfer potential.
The evolution of the transfer response associated with
ZnTMPyP4+ with increasing concentrations of ZnTPPS4- is also
shown in Figure 2c. As the concentration of the anionic
porphyrin is increased, the transfer of ZnTMPyP4+ effectively
shifts toward more positive potentials. The shape of the response
is also affected, revealing new features at either side of the
transfer signal. This interesting behavior suggests an association
equilibrium involving both porphyrins in the aqueous phase.
Considering the Nernst equation for the ZnTMPyP4+ transfer
Figure 3. Photocurrent transients originating from the photooxidation
of DCMFc in the presence of 5 × 10-5 mol dm-3 ZnTPPS and
ZnTMPyP at various Galvani potential differences. Photon flux was
6.4 × 1015 cm-2 s-1 at 543 nm.
presence of equimolar concentrations of ZnTPPS4- and ZnT-
MPyP4+ in the aqueous phase and the electron donor decam-
ethylferrocene (DCMFc) in DCE are displayed in Figure 3. It
is observed that the photocurrent increases as the applied Galvani
potential increases. These results resemble the behavior observed
for the heterogeneous quenching of ZnTPPC by ferrocene
derivatives.11,12 Negligible photoresponses are observed in the
absence of the organic phase quencher. Interestingly, no
substantial photocurrent is detected in the presence of only one
of the porphyrin species at the same experimental conditions.
This is a surprising result as both porphyrins exhibit rather
[ZnTMPyP]o
[ZnTMPyP]w
RT
4F
∆wo φ ) ∆wo φ°′
+
ln
(1)
4+
ZnTMPyP
(
)
and the association constant of both porphyrins in aqueous phase
KC
[dimer]w
KC )
(2)
[ZnTMPyP]w[ZnTPPS]w
positive redox potentials in the excited state (E* > 1 V). From
it follows that the shift of the formal transfer potential of
ZnTMPyP for a given concentration of ZnTPPS corresponds
to
ox
this behavior, it is evident that the photoactive species corre-
sponds to
a
heterostructure involving ZnTPPS4- and
ZnTMPyP4+
.
w
∆wo φ°′
- ∆o φ°′
)
4-
4+
ZnTMPyP4++ZnTPPS
ZnTMPyP
The overall photoresponses in Figure 3 are substantially
higher at potentials close to 0 V in comparison to the monomer
ZnTPPC4-. For instance, at 0.10 V the photocurrent in the
presence of ZnTPPS4- and ZnTMPyP4+ is approximately 40
times larger than in the case of ZnTPPC for the same photon
flux.11 Another relevant observation is related to the fact that
no enhancement of the photocurrent occurred in the presence
of the pair ZnTPPC-ZnTMPyP. Preliminary flash photolysis
studies do reveal quenching of the triplet lifetime for the pair
ZnTPPC-ZnTMPyP, indicating an effective interaction between
both monomers in solution. However, photocurrent responses
associated with the quenching of ZnTPPC are very little affected
by the presence of ZnTMPyP. As discussed in the next section,
the differences in the heterogeneous photochemistry between
ZnTPPS-ZnTMPyP and ZnTPPC-ZnTMPyP are connected
to the adsorption behavior of these porphyrins at the water/
DCE interface.
The features of the photocurrent transients in Figure 3, even
at potentials close to the edge of the polarizable window, are
consistent with a heterogeneous electron-transfer mechanism.
Effectively, the “instantaneous” photocurrent observed upon
illumination is proportional to the flux of electron injection to
the heterocomplex excited state.11,12 The flux of electron
injection is dependent on the (i) interfacial concentration of the
reactants, (ii) the photon flux to the interface, (iii) the optical
capture cross section of the sensitizer, and (iv) the competition
RT
4F
ln(KC[ZnTPPS]w) (3)
Considering the shift of the transfer signal as a function of the
ZnTPPS concentration in Figure 2c, the association constant
can be estimated as (3.0 ( 2.5) × 107 mol-1 dm3. Unfortunately,
the complexity of the transfer signal introduces considerable
uncertainties in the determination of the transfer potentials.
Interestingly, association constants of the same magnitude have
been reported for several porphyrin ion pairs in water/acetone
mixtures at room temperature.5 As shown in the next section,
photocurrent dependence on the concentration of both porphyrins
does confirm these results. The voltammetric responses in Figure
2c also illustrate the polarization window where heterogeneous
photoinduced electron-transfer responses can be studied without
significant perturbations arising from the transfer of porphyrin
species. From this figure, the potential window extends over
∼300 mV. Furthermore, no additional features were observed
in the presence of organic phase quenchers, indicating that the
ground state of the sensitizer is not involved in any electron-
transfer processes within the polarization window.
Photocurrent Responses Involving the Heterocomplex
ZnTPPS-ZnTMPyP. Photocurrent transient responses in the
(15) Bard, A. J.; Faulkner, L. R. Electrochemical Methods; Wiley: New
York, 1980.

10206 J. Am. Chem. Soc., Vol. 121, No. 43, 1999
Ferm´ın et al.
Figure 4. Photocurrent transients obtained in the presence of 5 × 10-5
mol dm-3 ZnTPPS and ZnTMPyP in the aqueous phase, and TCNQ
as the organic phase quencher. Negative photocurrents are associated
with the photoreduction of TCNQ. The positive overshoot in the off
transient is linked to back electron transfer to the oxidized ground state
sensitizer. Photon flux was 6.4 × 1015 cm-2 s-1 at 543 nm.
Figure 5. Schematic representation of a photoinduced electron transfer
from an aqueous soluble sensitizer (Sw) to an organic phase electron
acceptor (Qo) featuring a redox couple acting (Rss/Oss) as a super-
sensitizer (a). The rate constants ket, kd, kps, krec, and kss stand for electron
transfer, decay of the excited state, product separation, back-electron
transfer and supersensitization respectively. Product separation ef-
ficiency for the photoreduction of TCNQ as a function of the Galvani
potential difference in the presence and absence of an equimolar ratio
(2.5 × 10-4 mol dm-3) of the hexacyanoferrate redox couple (b). Other
experimental parameters as in Figure 4.
between electron transfer and decay of the excited state.11 In
the case of ionic porphyrin ZnTPPC4-, both the coverage of
the sensitizer at the liquid/liquid interface as well as the
phenomenological photoinduced electron-transfer rate constant
have been found to be dependent on the Galvani potential
difference.11,12 Mechanisms involving homogeneous quenching
followed by transfer of charged products can be effectively
excluded from this analysis as these processes feature essentially
zero initial photocurrent and slow rising photoresponses in the
time scale of 1-102 s.16-18
The product separation efficiency (Φps) can be defined as
the ratio between the steady state (Jss) and initial photocurrent
(J0)
The relaxation observed following the initial photocurrent and
the negative overshoot during the off transient are related to
back electron-transfer processes.11,12 Recombination is rather
dependent on the redox properties of the organic species. In
Figure 4, photocurrent strongly decreases after the initial
response in the presence of the electron acceptor TCNQ. In this
case, the photoinduced electron-transfer involves negative
photocurrents as the electron injection occurs from the porphyrin
heterodimer to TCNQ. It is also observed that the negative
photocurrent increases as the potential is shifted toward the
negative edge of the window. The positive overshoot in the off-
transient reflects a back electron transfer to the oxidized ground
state sensitizer. The competition between product separation
and recombination is not only dependent on the Galvani
potential difference,12 but it could also be affected by interaction
with supersensitizers.9 A schematic representation of the over-
all mechanism, including the supersensitization step, is shown
in Figure 5a. It is envisaged that the recombination step is
in competition with electron exchange between the photo-
induced intermediate species and the supersensitizer located
in the aqueous phase. In a previous report, it was demon-
strated that the photocurrent relaxation in Figure 4 is part-
ially suppressed in the presence of an equimolar concentra-
tion of the hexacyanoferrate couple acting as supersensi-
tizer.9
Φps ) Jss/J0
(4)
The effect of the Galvani potential difference and the super-
sensitizer [Fe(CN)6]4-/[Fe(CN)6]3- on the product separation
efficiency is contrasted in Figure 5b. It is observed that the
efficiency increases from ∼0.2 to 0.7 at -0.15 V upon adding
2.5 × 10-4 mol dm-3 of the hexacyanoferrate redox couple.
This effect clearly indicates that the photocurrent decay and
overshoot are indeed connected to a heterogeneous electron-
transfer process. Despite the noticeable increase of the steady-
state photocurrent in the presence of the supersensitizer, a
considerable fraction of the photoinduced intermediate species
still undergoes back electron transfer. Consequently, Φps is also
dependent on the Galvani potential difference under these
conditions. Analysis of the potential dependence of the electron
transfer and recombination rate constant for the monomer
ZnTPPC4- has suggested that both processes involve an
activated step which is related to redox and interfacial properties
of the quencher species.12
Interfacial Behavior of the Pair ZnTPPS-ZnTMPyP.
Although the previous analysis indicates that the photoactive
species involves both zinc porphyrins, the nature of this
heterocomplex should be addressed in more detail. In Figure 6,
photocurrent spectra recorded at various potentials are compared
to the absorption features of the free monomers ZnTPPS and
ZnTMPyP in solution. In these experiments, the relative input
photon flux was monitored by a photomultiplier tube connected
to a liquid light guide. Subsequently, photocurrent responses
were normalized to the output of the photomultiplier tube at
(16) Kotov, N. A.; Kuzmin, M. G. J. Electroanal. Chem. 1990, 285,
223.
(17) Kotov, N. A.; Kuzmin, M. G. J. Electroanal. Chem. 1992, 341,
47-60.
(18) Kotov, N. A.; Kuzmin, M. G. J. Electroanal. Chem. 1992, 338, 99.

Interfacial Photoinduced Electron Transfer
J. Am. Chem. Soc., Vol. 121, No. 43, 1999 10207
Figure 6. Photocurrent spectra associated with the photooxidation of
DCMFc by the heterodimer ZnTPPS-ZnTMPyP. Photocurrent re-
sponses were normalized to the input flux of photon at each wavelength.
Absorption features for the free monomer ZnTPPS and ZnTMPyP in
aqueous solution are also superimposed.
each wavelength.10 It is observed that the photocurrent responses
associated with the Q-band transitions are red-shifted respect
to the absorption signal of the free porphyrin monomers. This
behavior is rather different to the observed for ZnTPPC, in
which photocurrent and absorption maxima coincide within the
same wavelength range.10 As mentioned previously, photocur-
rent responses arising from the free monomers ZnTPPS and
ZnTMPyP4+ are negligible, therefore the photocurrent spectra
in Figure 6 reflect only the absorption behavior of the het-
erodimer species. The red shift of the Q-band suggests that the
dimer ZnTPPS-ZnTMPyP forms an inner-sphere ion pair.19
Other porphyrin pairs have also shown similar spectroscopic
behavior.5 In this type of ion pair, both species are in close
proximity losing the first solvation shell. In addition, inner-
sphere ion pairs can be regarded as dipolar polyatomic ions with
nonspherical charge distribution.20 In the case of ZnTMPyP-
ZnTPPS, the overall charge of the pair structure is zero.
The adsorption of the porphyrin heterocomplex at the water/
DCE interface can be followed by the concentration dependence
of the photocurrent signal. Photocurrent potential curves ob-
tained under chopped illumination at 12 Hz and lock-in detection
at various concentration of ZnTPPS and ZnTMPyP are displayed
in Figure 7. The real component of the photocurrent in the
presence of DCMFc was considerably larger than the quadrature
component, indicating that transient responses associated with
recombination and interfacial charging current are rather
negligible at this frequency.12 The potential dependence of the
real part of the photocurrent reflects increasing photocurrent
with increasing Galvani potential difference as illustrated in
Figure 3. A rather interesting behavior is also observed from
the photocurrent dependence on the monomer concentration.
The increase of the photocurrent with increasing concentration
of ZnTPPS and ZnTMPyP seems rather similar at all Galvani
potential differences. This trend is more clearly displayed in
Figure 8, where photocurrent isotherms are plotted at various
potentials. As a first approximation, this effect can be rational-
ized in terms of the association of both porphyrin species in
solution leading to electrically neutral photoactive species. The
photocurrent responses are effectively linked to adsorbed
heterodimer species which are in equilibrium with the total
concentration in solution. Due to the absence of a net charge
Figure 7. Real (a) and imaginary (b) component of the photocurrent
in the presence of DCMFc and various concentration of ZnTPPS and
ZnTMPyP. The light chopping frequency was 12 Hz. The photon flux
was 6.4 × 1015 cm-2 s-1. It is observed that the imaginary part of the
photocurrent is negligible with respect to the real part, indicating that
the photoresponses are effectively in phase at this frequency.
Figure 8. Photocurrent dependence on the concentration of ZnTPPS
and ZnTMPyP at various Galvani potential differences. Experimental
conditions as in Figure 7.
on the heterodimer species, the adsorption equilibrium remains
largely independent of the Galvani potential difference. Con-
sequently, from the phenomenological mechanism introduced
in previous work,11,12 the photocurrent response in the absence
of recombination is expressed as
â[dimer]w
max
photo
jphoto ) j
(5)
[
]
1 + â[dimer]w
max
photo
where j
is the photocurrent corresponding to a full compact
monolayer of the dimer species
ket
max
photo
j
) e
I0σNs
(6)
ket + kd
e is the electronic charge, ket the pseudo-first-order rate constant
for electron transfer, kd the rate constant of excited state
relaxation, I0 the photon flux, σ the capture cross section, and
Ns the surface density of dimer species in a close compact
monolayer. Equations 5 and 6 are associated with a mechanism
where photoresponses are determined by a competition between
relaxation of the excited state and electron transfer. The
parameter â corresponds to the Langmuir isotherm parameter.
The concentration of the dimer in solution is simply given by
(19) Nancollas, G. H. Interactions in electrolyte solutions; Elsevier:
Amsterdam, The Netherlands, 1966.
(20) Marcus, Y. Ion solVation; John Wiley and Sons Ltd.: Chichester,
1985.

10208 J. Am. Chem. Soc., Vol. 121, No. 43, 1999
Ferm´ın et al.
max
photo
A closer look into the potential dependence of j
revealed
that the increment of ket with increasing potential approaches a
Tafel behavior. The ratio ket/kd can be estimated by rearranging
eq 6
1
ket/kd )
(7)
max
(eI0σNs/j
) - 1
photo
In Figure 9b, this ratio is plotted as a function of the Galvani
potential difference assuming a value of Ns of 2.5 × 1014 cm-2
.
The exponential increment of ket/kd with increasing Galvani
potential can be described in terms of a Tafel relation with a
transfer coefficient close to R ) 0.5. These results have been
phenomenologically interpreted in terms of an activation barrier
for the electron-transfer process.11,12 In addition, taking the
lifetime of the triplet state as 30 µs after the report by Hugerat
et al.7 for a similar porphyrin ion pair, ket may lie in the range
of 104-105 s-1. The value of ket is within the same order of
magnitude to the estimated for the monomer ZnTPPC. Conse-
quently, these results seem to indicate that the larger photocur-
rent observed for the pair ZnTPPS-ZnTMPyP in comparison
to the porphyrin ZnTPPC is mostly connected to a higher
interfacial concentration of the porphyrin ion pair.
The value of ket, as extracted from the previous analysis,
corresponds to the integral of the electron-transfer rate over the
distance separating the adsorbed sensitizer and the organic redox
couple.11 Recent descriptions of the interfacial structure of ITIES
based upon lattice-gas approximation21 and molecular dynam-
ics22 suggest that the interfacial region between water and DCE
extends over an average distance of 1 nm. Taking this value as
a first approximation, the pseudo-first-order heterogeneous rate
constant is of the order of 10-2 cm s-1. Although this magnitude
is ∼10 times faster than previously reported rate constants for
heterogeneous electron transfer at ITIES in the dark,23-27 it is
still 2-3 orders of magnitude slower than the upper limit for
outer sphere adiabatic electron-transfer reactions from the
Marcus’ model.28-30
Figure 9. Potential dependence of the photocurrent associated with a
full monolayer of porphyrin heterodimer (a) obtained from the fittings
of the photocurrent isotherms (Figure 8). The ratio of the pseudo-first-
order electron transfer and decay of the excited rate constants as a
function of the Galvani potential difference (b) was estimated from eq
7, assuming 2.5 × 1014 cm-2 as the maximum surface density of dimer
species. Considering a Butler-Volmer dependence of ket, a transfer
coefficient of 0.5 is estimated.
the association equilibrium for the two porphyrin monomers as
defined in eq 2.
The dependence of the photocurrent on the concentration of
ZnTPPS and ZnTMPyP at various Galvani potential differences
is displayed in Figure 8. Fittings employing eq 5 are also shown
as continuous lines. Three parameters were adjusted for each
max
photo
According to the values obtained for â, the Gibbs energy of
Galvani potential difference, j
, KC and â. The parameters
adsorption (∆Gads) can be estimated as -22.8 ( 0.4 kJ mol-1
.
KC and â were found almost independent of the applied
potential. For instance, the association constant KC exhibited
an average value of (6 ( 1) ×107 mol-1 dm3 throughout the
potential range. This value is comparable to the preliminary
estimations based on the shift of the ZnTMPyP transfer potential
in the presence of ZnTPPS (cf. Figure 2). In turn, the increase
of the photocurrent with increasing potential is associated with
The magnitude of ∆Gads is approximately half of the value
reported for ZnTPPC under similar experimental conditions.11,12
This surprising result indicates that, although the adsorption
energy for ZnTPPC is larger, a smaller number of these
porphyrins can be accommodated at the liquid/liquid junction
in comparison to the heterodimer. The large adsorption energy
exhibited by ZnTPPC in comparison to ZnTPPS is also
responsible for the difference in photoreactivity of the corre-
sponding dimer structures. As mentioned earlier, no enhance-
ment of the photocurrent associated with the quenching of
ZnTPPC is observed upon addition of the cationic ZnTMPyP.
Although spectroscopic evidence suggests the formation of ion
pairs between ZnTPPC and ZnTMPyP in solution, the adsorp-
tion sites at the interface are expected to be mostly occupied
max
photo
max
photo
the parameter j
. The potential dependence of j
is
displayed in Figure 9a. This behavior reflects an increase of
the electron-transfer rate with respect to the relaxation of the
excited state. Another interesting aspect arising from this
max
photo
analysis is the rather high values obtained for j
in com-
parison to the responses observed for the monomer ZnTPPC.11
Taking a capture cross section of 10-17 cm2 at 543 nm for the
heterodimer, the photocurrent density is 4-5 times higher than
the value expected for a close packed layer of porphyrin
molecules lying coplanar to the interface.11 Assuming that
(21) Schmickler, W. J. Electroanal. Chem. 1997, 429, 123-127.
(22) Benjamin, I. Annu. ReV. Phys. Chem. 1997, 48, 407.
(23) Cheng, Y.; Schiffrin, D. J. J. Chem. Soc., Faraday Trans. 1993,
89, 199-205.
photocurrent arises from the first monolayer of adsorbed
max
photo
porphyrin species, the high values of j
can be taken as an
(24) Ding, Z.; Ferm´ın, D. J.; Brevet, P.-F.; Girault, H. H. J. Electroanal.
Chem. 1998, 458, 139-148.
evidence for a specific orientation of the adsorbed porphyrin
pairs at the interface. On the basis that porphyrin heterodimer
consists of an electrically neutral dipolar molecule, it could be
envisaged a kind of self-assembling mechanism for the adsorp-
tion process at the liquid/liquid junction. The interfacial orienta-
tion of heterodimer species may also play a fundamental role
on the kinetics of photoinduced electron transfer.
(25) Tsionsky, M.; Bard, A. J.; Mirkin, M. V. J. Phys. Chem. 1996, 100,
17881-17888.
(26) Tsionsky, M.; Bard, A. J.; Mirkin, M. V. J. Am. Chem. Soc. 1997,
119, 10785-10792.
(27) Shi, C.; Anson, F. C. J. Phys. Chem. B 1999, 103, 6283-6289.
(28) Marcus, R. A. J. Phys. Chem. 1990, 94, 4152-5.
(29) Marcus, R. A. J. Phys. Chem. 1990, 94, 1050-5.
(30) Marcus, R. A. J. Phys. Chem. 1991, 95, 2010-13.

Interfacial Photoinduced Electron Transfer
J. Am. Chem. Soc., Vol. 121, No. 43, 1999 10209
Figure 10. Photocurrent potential curves in the presence of 5 × 10-5
mol dm-3 of ZnTPPS and ZnTMPyP, and DCMFc as quencher at
various concentration of Li2SO4. Other experimental conditions as in
Figure 7. Photocurrent responses were effectively in-phase throughout
the potential range for all concentrations of the aqueous supporting
electrolyte.
by the anionic porphyrin instead of the dimer species. This
behavior shows that the interfacial photoreactivity of ion pair
dimers not only depends on the association of the porphyrin
species, but also on the adsorption properties of each of the
species involved.
Dependence of Photoresponses on the Ionic Strength of
the Aqueous Phase. Photocurrent responses for a concentration
of 5 × 10-5 mol dm-3 of the porphyrins ZnTPPS and ZnTMPyP
are displayed in Figure 10 for various concentrations of Li2-
SO4. It is observed that the overall photocurrent decreases with
increasing concentration of the supporting electrolyte. Similar
to the results in Figure 7, the imaginary component of the
photocurrent was negligible in comparison to the real compo-
nent. Photocurrent isotherm analysis revealed that the ∆Gads and
KC are dependent on the supporting electrolyte concentration.
Figure 11 shows that ∆Gads becomes more negative, and KC
decreases with increasing concentration of Li2SO4. The origin
of the ∆Gads dependence on the aqueous ionic strength is yet to
be rationalized. In principle, the increase of the supporting
electrolyte concentration may involve higher screening of the
local charge associated with the porphyrin units, enhancing the
surface assembling of adsorbed dimer species.
Figure 11. Dependence of the Gibbs energy of adsorption (a) and
ZnTPPS-ZnTMPyP association constant (b) as function of the aqueous
supporting electrolyte concentration. Both parameters were obtained
from photocurrent isotherms as exemplified in Figure 8.
where A is the constant derived from the Debye-Hu¨ckel theory,
i.e., 0.509 for aqueous solution at 25 °C, and C is an adjusting
parameter. Analysis of a large range of electrolytes have
provided values for C between 0.2 and 0.4 kg mol-1 19
The
.
solid line in Figure 11b corresponds to a fit based on eqs 8 and
9, assuming a value of 0.3 kg mol-1 for C. This result effectively
confirmed that the KC dependence on I is associated with
changes in the activity coefficients of the charged porphyrins.
Nevertheless, it should be mentioned that the previous ap-
proximation related to the activity coefficient of the dimer
species is not strictly valid in the light of the dependence of
∆Gads on the ionic strength.
It should also be concluded that the decrease in the photo-
current responses with increasing supporting electrolyte con-
centration (Figure 10) cannot be ascribed only to the dependence
of ∆Gads and KC on the ionic strength. Although the behavior
of these two parameters has clear effects on the photocurrent
isotherms, the estimated changes in total photocurrent are rather
small in comparison to the experimental trend. In turn, the
changes in the total photocurrent may reflect changes in the
quantum yield, i.e., the ratio between the rates of electron
transfer and relaxation of the excited state. It is well-established
that the lifetime of the triplet state for several porphyrins
decreases with increasing ionic strength of the medium.3
Decreasing photocurrent with increasing supporting electrolyte
concentration has also been observed for the monomer ZnTPPC.
Finally, from the value of KA ) (5.9 ( 0.8) × 108 mol-1
dm3, the distance between the ion pair centers (a) can be
The behavior of KC in Figure 11b can be rationalized in terms
of a decrease in the activity coefficient of the charged porphyrin
monomers with increasing concentration of the supporting
electrolyte. To quantify this effect, the association constant
involving the activities of the ionic species (KA) can be
expressed as
adimer
(aZnTPPY4-)(aZnTMPyP
KC
KA )
)
(8)
4+
4+
)
(γZnTPPS4-)(γZnTMPyP
)
where the activity coefficient of the heterodimer is assumed
independent of the supporting electrolyte concentration and close
to unity. Activity coefficients of highly charged species as a
function of the ionic strength (I) have been described by the
formula proposed by Davies in 193819
I1/2
I + I1/2
-log γz ) Az2
- CI
(9)
(
)

10210 J. Am. Chem. Soc., Vol. 121, No. 43, 1999
Ferm´ın et al.
estimated employing the expression derived by Fuoss for contact
ion pairs20
electron donor TCNQ. In this case, the recombination reaction
was partially suppressed by shifting the Galvani potential
difference toward more negative values, or by addition of the
hexacyanoferrate redox couple acting as supersensitizer.
The photocurrent dependence on the concentration of both
porphyrin species suggests that the association processes occur
in homogeneous phase, followed by adsorption of the ion pair
at the liquid/liquid junction. The adsorption step was effectively
independent of the Galvani potential difference, indicating the
electrically neutral nature of the ion pair. The Gibbs energy of
adsorption and the association constant of the ion pair were also
estimated. Although the adsorption energy was found smaller
in comparison to the porphyrin ZnTPPC, the photocurrent
response under similar condition was found up to 50 times larger
for the heterodimer. The difference in the magnitude of the
photoresponses mainly arises from a higher surface density in
the case of the ZnTPPS-ZnTMPyP ion pair. Experimental
evidence suggests that adsorbed heterodimer species are oriented
in a noncoplanar fashion with respect to the interface. On the
other hand, excitonic effects revealed by photocurrent spectra
as well as estimation of the distance separating the pair
ZnTPPS-ZnTMPyP indicate the formation of a contact (inner
sphere) ion pair. The overall analysis demonstrates that the
heterogeneous photoreactivity of sensitizers forming ion pairs
is determined not only by the photophysical properties of the
supramolecular structure but also by the interfacial behavior of
each of the species involved.
KA ) [(4 × 10-24Nπ/3)](a/nm)3 exp(b)
(10)
where N is the Avogadro’s constant and b corresponds to
b ) |zZnTPPS||zZnTMPyP|(e2/4πꢀok)T-1ꢀ-1a-1
(11)
Evaluation of eqs 10 and 11 provide a contact distance a )
5.8 Å. This distance is larger than the interplane separation
between the special pair of bacteriochlorophyll molecules in
the R. Viridis, i.e., 3.8 Å1. However, in the case of the naturally
occurring dimer, the macrocycles are covalently linked. The
estimated distance between ZnTPPS and ZnTMPyP, in conjunc-
tion with the excitonic effect revealed in the photocurrent spectra
(Figure 6) and the apparently noncoplanar adsorption of the
heterodimer at the interface point out that the monomeric units
are indeed face-to-face oriented. In this respect, this type of
heterodimer can be regarded as a spontaneously generated
supramolecular structure which not only exhibits similar
characteristics to synthetic and naturally occurring dimers, but
also is able to self-assemble at the water/DCE interface. The
correlation between interfacial orientation of the adsorbed
porphyrins and heterogeneous photoreactivity is currently under
investigation.
Acknowledgment. The authors are grateful for the financial
support of the Fonds National Suisse de la Recherche Scienti-
fique project 20-055692.98/1. The enlightening discussions with
Riikka Lahtinen from the Helsinki University of Technology
are also gratefully acknowledged. Many thanks to Lise Van
Long, Carole Claivaz, and Vale´rie Devaud for their contributions
and technical assistance. The Laboratoire d’Electrochimie is part
of the European Network ODRELLI (Organization, Dynamics
and Reactivity at Electrified liquid/liquid interfaces).
Conclusions
Water soluble anionic ZnTPPS and cationic ZnTMPyP
spontaneously associate in solution, generating heterodimer
species which undergoes photoinduced heterogeneous electron
transfer with hydrophobic redox species at the water/DCE
interface. No detectable photocurrent was observed in the
presence of only one of the porphyrin species. Photocurrent
responses involving the quenchers DCMFc and TCNQ were
found dependent on the Galvani potential difference. Strong
recombination responses were observed in the presence of the
JA992215I