Electron transfer at liquid/liquid interfaces. The effects of ionic adsorption, electrolyte concentration, and spacer length on the reaction rate

Article abstract of DOI:10.1021/jp013543j

Several factors affecting the rate of heterogeneous electron transfer (ET) reactions at the nonpolarized interface between two immiscible electrolyte solutions (ITIES) were investigated by scanning electrochemical microscopy (SECM). The reactions between zinc porphyrin in organic phase and aqueous Ru(CN)₆^3/4-or Fe(CN)₆^3/4- redox species, which have previously served as model systems for probing the potential dependence of the ET rate constant, were used to investigate the effects of adsorption of a reactant and a potential-determining ion on ET rate at the interfaces between water and different organic solvents. It was demonstrated that the rate constant of an interfacial ET can be either potential-independent or potential-dependent under different experimental conditions. The same model experimental systems were used to probe the kinetics of long-range ET across a monolayer of phospholipid adsorbed at the ITIES. The results obtained with a family of phosphatidylserine lipids are compared to those previously obtained with synthetic phosphatidylcholine lipids. The extent of blocking effect of adsorbed lipids on interfacial ET was investigated at different pHs and also in the presence of calcium ions in the aqueous phase. The micrometer-sized domain formation in the monolayer was observed in the presence of divalent cations.

Full text of DOI:10.1021/jp013543j

J. Phys. Chem. B 2002, 106, 3933-3940
3933
Electron Transfer at Liquid/Liquid Interfaces. The Effects of Ionic Adsorption, Electrolyte
Concentration, and Spacer Length on the Reaction Rate
Biao Liu and Michael V. Mirkin*
Department of Chemistry and Biochemistry, Queens CollegesCUNY, Flushing, New York 11367
ReceiVed: September 17, 2001; In Final Form: January 14, 2002
Several factors affecting the rate of heterogeneous electron transfer (ET) reactions at the nonpolarized interface
between two immiscible electrolyte solutions (ITIES) were investigated by scanning electrochemical microscopy
3/4-
3/4-
(SECM). The reactions between zinc porphyrin in organic phase and aqueous Ru(CN)6
or Fe(CN)6
redox species, which have previously served as model systems for probing the potential dependence of the
ET rate constant, were used to investigate the effects of adsorption of a reactant and a potential-determining
ion on ET rate at the interfaces between water and different organic solvents. It was demonstrated that the
rate constant of an interfacial ET can be either potential-independent or potential-dependent under different
experimental conditions. The same model experimental systems were used to probe the kinetics of long-
range ET across a monolayer of phospholipid adsorbed at the ITIES. The results obtained with a family of
phosphatidylserine lipids are compared to those previously obtained with synthetic phosphatidylcholine lipids.
The extent of blocking effect of adsorbed lipids on interfacial ET was investigated at different pHs and also
in the presence of calcium ions in the aqueous phase. The micrometer-sized domain formation in the monolayer
was observed in the presence of divalent cations.
Introduction
organic redox species is small.14 At the same time, the diffuse
layer effect, similar to the Frumkin effect at metal electrodes,
There has been a surge of activity in studies of electron
transfer (ET) processes at the interface between two immiscible
electrolyte solutions (ITIES) during the last several years. The
kinetics of ET reactions between different aqueous and organic
redox species have been measured.1 Several research groups
employed scanning electrochemical microscopy (SECM) to
probe ET at the nonpolarizable2-8 and polarizable9 ITIES. Those
studies identified the experimental conditions under which
interfacial electron transfer can be probed and demonstrated that
ET can be quantitatively separated from ion transfer processes.
As expected from theory,10 the studied reactions were found to
be of the first order in both organic and aqueous redox species
and the effective heterogeneous rate constants were strongly
dependent on the difference between standard potentials of two
redox couples (∆E°). Perhaps the most controversial point in
this area is the contribution of the interfacial potential drop to
the driving force for ET reactions.
may result in either accumulation or depletion of a reactant
species in the proximity of the phase boundary (mostly on the
organic side14). Thus, the apparent rate constant is expected to
depend on ∆owæ if the organic reactant is charged and to be
essentially independent of ∆owæ when neutral organic species
participate in the ET reaction.14c
These theoretical predictions have been corroborated only for
one model experimental system, the reaction between zinc
porphyrin (ZnPor) in organic phase and hexacyanorutinate in
water. Both the forward (reaction 1b3b,c
)
ZnPor - e f ZnPor+
(tip in organic phase) (1a)
kf
ZnPor+(o) + Ru(CN)64-(w) 8 ZnPor(o) + Ru(CN)63-(w)
9
(ITIES) (1b)
and reverse (reaction 2b7)
Most experimental data obtained at polarized9,11,12 and
nonpolarized ITIES3,6b,d,8 supported the assumption that ET at
the ITIES is driven by the potential drop across the interfacial
boundary [for the ITIES this is the difference of Galvani
potentials of the organic and aqueous phases (∆owæ)1]. Ac-
cordingly, the dependence of the ET rate constant on ∆owæ is
expected to be exponential (Butler-Volmer equation) at lower
overpotentials and to level off at much higher overpotentials.10
However, the Butler-Volmer model (and Marcus theory) is
applicable only if most of the interfacial voltage drops between
the reacting redox moieties. This assumption is not highly
plausible because the interfacial voltage drops mostly within
the diffuse layer in organic phase.13 According to a widely
accepted three-layer model of the ITIES, the ET rate constant
should be essentially potential-independent because the potential
drop across the mixed-solvent layer separating aqueous and
3-
Ru(CN)64- - e f Ru(CN)6
(tip in water) (2a)
kb
Ru(CN)63-(w) + ZnPor(o)
9
8 Ru(CN)64-(w) + ZnPor+(o)
(ITIES) (2b)
ET processes were investigated by SECM. In those experiments,
the ultramicroelectrode (UME) tip was positioned in the top
phase, which was either organic (Figure 1A) or aqueous (Figure
1B). The redox mediator contained in the top phase (either
ZnPor or Ru(CN)64- ) was oxidized at the UME surface. When
the tip approached the bottom liquid layer, the mediator was
regenerated via interfacial bimolecular ET (either reaction 1b
or 2b). The rate constants of reactions 1b and 2b were evaluated
from the SECM tip current vs distance curves. To probe the
10.1021/jp013543j CCC: $22.00 © 2002 American Chemical Society
Published on Web 03/23/2002

3934 J. Phys. Chem. B, Vol. 106, No. 15, 2002
Liu and Mirkin
It was suggested later,7 that the observed dependence of the
ET rate on ∆owæ can also be explained by Schmickler’s
model.14c Within the frame of that model, the increase in the kf
at more positive ∆owæ is attributed to accumulation of ZnPor+
at the interface rather than to increasing driving force for ET.
Unlike forward reaction 1b, the electron acceptor in the reverse
reaction 2b is anionic and the electron donor is a neutral species.
The concentration of the neutral ZnPor at the interface is
potential-independent. Since only a small fraction of the
interfacial voltage drops within the aqueous diffuse layer, the
effect of ∆o æ on surface concentration of Ru(CN)63- is small.
w
Accordingly, the kb remained essentially constant within the limit
of experimental error when the ∆owæ was changed by about
120 mV (inset in Figure 1B). Similar results were obtained for
three different organic solvents, i.e., BZ, NB, and DCE.7
One should notice, however, that no other potential-
independent rate constants have been measured at either
polarizable or nonpolarizable ITIES (the only exception is ref
15, where the kinetic analysis was complicated by the transfer
of ferrocenium ion, which could occur simultaneously with ET).
The rates of ET reactions involving such neutral organic
reactants as 7,7,8,8-tetracyanoquinodimethane (TCNQ)6b,8 or
decamethylferrocene (DMFc)6c,7 at nonpolarizable ITIES de-
pended significantly on the ∆owæ (though in the latter case the
potential dependence was at variance with both Butler-Volmer
and Marcus models). At a polarizable ITIES, the rate of ET
reactions between neutral phthalocyanine complexes of lutetium
or tin in organic phase and charged aqueous species also
appeared potential-dependent, and the Butler-Volmer model
was used for kinetic analysis.11a,b Recent spectroelectrochemical
studies using neutral organic reactants (e.g., ferrocene)13,16 have
also showed some potential dependence of the interfacial ET
rate constant. Here, we report more experiments aimed to clarify
the origins of potential-independent kinetics of reaction 2b.
Another objective of this work is to compare the kinetics of
long-range ET across monolayers of different phospholipids
adsorbed at the ITIES. Our interest in studying ET across
phospholipid films at the ITIES is 2-fold. First, this process is
of fundamental importance for a number of biological systems.17
On the other hand, by separating the electron donor and acceptor
with a well-defined spacer one can slow the ET rate and carry
out measurements at high overpotentials in order to check the
ET theory.18 ET can also be used as a yardstick to probe the
properties of the interfacial film.19 In this way, Delville et al.3d
compared the rates of ET across the monolayers of saturated
dipalmytoylphosphocholine and polyconjugated 2(3-(diphenyl-
hexatrienyl)propanoyl)-1-hexadecanoyl-sn-glycero-3-phospho-
choline lipids adsorbed at the water/benzene interface and
showed that polyconjugated chains added to dipalmytoyl film
behave as conducting wires increasing the overall ET rate.
SECM imaging of mixed lipid films at the ITIES revealed the
formation of micrometer-size domains.3d Separating ET reactants
with a long-chain spacer can also increase the fraction of ∆owæ
dropping between them and in this way increase the potential
dependence of the rate constant.
Figure 1. Schematic diagram of SECM measurements of the kinetics
of ET between (a) ZnPor+ in organic phase and Ru(CN)64- in aqueous
phase and (b) ZnPor in organic phase and Ru(CN)63- in aqueous phase.
The insets show the corresponding potential dependences of the rate
constant from ref 3b (A) and 7 (B).
potential dependence of the rate constant, ∆owæ was varied by
changing the concentration of the common ion in the aqueous
-
phase ([ClO4
]
w
3b,7). The potential drop across a nonpolarizable
ITIES was shown to follow the Nernst-Donnan equation
-
∆owæ ) const - S log[ClO4 ]w
(3)
and the measured slope, S, was 56 mV/decade for nitrobenzene
(NB) and 60 mV/decade for 1,2-dichloroethane (DCE) and
benzene (BZ).3b,7 For reaction 1b, the log kf vs ∆o æ depen-
An earlier SECM study3c showed that adsorption of sym-
metric saturated lipids (1,2-dicacyl-sn-glycero-3-phospho-
choline) with a different number of methylene groups in a
hydrocarbon chain (n) on the water/BZ interface results in a
formation of a compact monolayer. The defect density in the
film was too low to produce detectable feedback current.
w
dence was linear with a transfer coefficient, R ) 0.5 (see inset
in Figure 1A). Thus, it was proposed that the ET kinetics can
be described by the Butler-Volmer model, and the driving force
for the interfacial ET reaction is the sum of the ∆owæ term and
the difference of standard potentials of the organic and aqueous
redox mediators:3b,c
Although the redox reactants (i.e., ZnPor+ in BZ and Fe(CN)6
4-
in water) were separated by the lipid film and the rate of ET
decreased markedly with the increasing length of the spacer,
∆G° ) -F(∆E° + ∆owæ)
(4)

Electron Transfer at Liquid/Liquid Interfaces
J. Phys. Chem. B, Vol. 106, No. 15, 2002 3935
the analysis of distance dependence of the rate constant
suggested partial penetration of ZnPor+ into the lipid layer. This
prevented quantitative interpretation of the kf vs n results.
Here we explore the possibility of improving the blocking
properties of a spacer by using another family of saturated
symmetric phospholipids, i.e., 1,2-diacyl-sn-glycero-3-phospho-
L-serines. An interesting feature of phosphatidylserine films is
that their structure can be easily and predictably changed. The
liquid-expanded monolayers formed by these lipids at the ITIES
at room temperature20 undergo a sharp phase transition and
become solid upon addition of a low concentration of a divalent
cation (e.g., Ca2+ or Mg2+).20b Such films are supposed to be
tightly packed and may not be permeable to large porphyrin
molecules.
Experimental Section
Chemicals. 5,10,15,20-Tetraphenyl-21H,23H-porphine zinc
(ZnPor), tetrabutylammonium tetraphenylborate (TBATPB),
NaClO4, and NaCl from Aldrich (Milwaukee, WI), and Na4-
Fe(CN)6 (Fisher Scientific, Fair Lawn, NJ) were used as
received. Na4Ru(CN)6 was synthesized as reported previously,3b,21
recrystallized several times from methanol/water and dried under
vacuum overnight at 50 °C. Tetrahexylammonium perchlorate
(THAClO4; Johnson Matthey, Ward Hill, MA) was recrystal-
lized twice from an ethyl acetate:ether (9:1) mixture and dried
under vacuum overnight at room temperature. Tetrabutyl-
ammonium chloride (TBACl), tetrabutylammonium hexafluo-
rophosphate (TBAPF6), and nitrobenzene (>99%) were from
Fluka Chemika, Switzerland. Benzene (spectrophotometric
grade) and 1,2-dichloroethane (99.8% HPLC grade) were from
Aldrich. All organic solutions were washed with a larger volume
of Milli-Q water several times before measurements to remove
impurities from the organic phase. All other chemicals were
ACS reagent grade.
Chloroform solutions of symmetric saturated synthetic lipids
(1,2-dicacyl-sn-glycero-3-phosphocholine and 1,2-diacyl-sn-
glycero-3-[phospho-L-serine]) with different numbers of meth-
ylene groups in a hydrocarbon chain (n ) 10, 12, 16; the
abbreviations C-10, C-12, etc. will be used to designate different
phospholipids) from Avanti Polar Lipids, Inc. (Alabaster, AL)
were stored at -20 °C. Before an experiment, a flow of nitrogen
was passed above a small volume (∼40 µL) of phospholipid
solution in chloroform for complete evaporation of solvent. The
precipitated lipid was then redissolved in either BZ or DCE or
NB to prepare a 500 µM stock solution, which also contained
ZnPor and supporting electrolyte. The mixture of the lipid stock
solution and BZ solution containing the same concentrations
of ZnPor and THAClO4 served as the organic phase.
Electrodes and Electrochemical Cells. A three-electrode
setup was employed with a 12.5-µm-radius Pt ultramicroelec-
trode tip, a Pt wire as the counter electrode, and a Ag/AgCl
reference electrode; all three electrodes were placed in the top
liquid phase. The UME tip was rinsed with ethanol and water
and then polished and dried before each measurement. When
the top phase was organic, an ionic bridge containing a solution
of NaCl and NaClO4 (or TBACl) was placed between the Ag/
AgCl reference electrode and organic solution.3b
To measure the rate of reaction 1b, the top phase contained
a solution of 0.25 M THAClO4, 0.5 mM ZnPor, and 0-100
µM lipid in BZ or NB or DCE, and the bottom phase contained
a solution of 7 mM of Na4Ru(CN)6 or Na4Fe(CN)6, 0.1 M NaCl,
and 0.01-2 M NaClO4 or 5-100 mM TBACl. The organic
solution was deaerated before each experiment. To measure the
rate of reaction 2b, the top phase contained a 0.1 mM solution
Figure 2. Steady-state voltammograms of 2 mM ZnPor at a 25-µm
diameter Pt disk electrode in (A) NB and (B) DCE containing 25 mM
TBAPF6. Scan rate was 5 mV/s.
of Na4Ru(CN)6, 0.1 M NaCl, and 0.01-2 M NaClO4 or 5-100
mM TBACl; and the bottom (organic) phase contained a
solution of 2 mM ZnPor, 1-100 mM THAClO4, or 10-200
mM TBAPF6 in NB or DCE. The steady-state voltammograms
of ZnPor in solutions containing TBA+ (Figure 2) were very
similar to those obtained previously in perchlorate solutions.3b,7
The electrochemical cell designed to keep a lower density liquid
phase (e.g., water) under a heavier liquid (e.g., DCE) was
described previously.22
SECM Apparatus and Procedure. The SECM instrument
was described previously.22,23 Before SECM measurements, the
tip electrode was positioned in the top phase. The tip was biased
either at a potential corresponding to the plateau current of
4-
Ru(CN)6 oxidation in water or to the plateau current of the
first oxidation wave of ZnPor in organic phase.
The approach curves were obtained by moving the tip toward
the liquid/liquid interface and recording the tip current (iT) as a
function of the tip/ITIES separation distance (d). The coordinate
of the liquid/liquid interface (d ) 0) was determined from the
sharp change in iT that occurred when the tip touched the ITIES.
The concentration of ZnPor in organic phase (e5 mM) was
limited by its solubility. When an organic solvent was used as
a bottom phase, the concentration of aqueous redox species was
not higher than [ZnPor]o/20 to avoid diffusion limitations in
the bottom layer.3,7
Potential Drop across the ITIES. In SECM measurements,
a nonpolarizable ITIES is poised by the concentrations of the
common ion providing a controllable driving force for the ET
reaction.3 The interfacial potential drop (∆wo æ) is governed by
-
the ratio of common ion (i.e., ClO4 or TBA+) concentration
-
in the organic and aqueous phases (i.e., [ClO4-]o/[ClO4
] or
w
[TBA+]o/[TBA+]w). The ∆owæ value was varied either by
changing the concentration of the common ion in water (i.e.,
or [TBA+]w) at a constant concentration of that ion
-
[ClO4
]
w
in organic phase (typically, 10-100 mM), or vice versa, the
common ion concentration was changed in the organic phase
and kept constant in the aqueous phase. The relative values of
∆owæ were obtained from cyclic voltammograms of either

3936 J. Phys. Chem. B, Vol. 106, No. 15, 2002
TABLE 1: Slope Values (S) for the Concentration
Liu and Mirkin
Dependencies of ∆owæ Measured at the DCE/Water and
NB/Water Interfaces
concentration varied organic solvent S, mV/decade
reference
-
[ClO4
[ClO4
[ClO4
[ClO4
]
]
]
]
DCE
DCE
NB
-59, -60
59, 65
-56
6c, 7
6c, this work
7
this work
6c, this work
this work
this work
this work
w
o
-
-
-
w
o
NB
62
[TBA+]w
[TBA+]o
[TBA+]w
[TBA+]o
DCE
DCE
NB
59, 66
-61
60
NB
-59
ZnPor (Figure 2) or TCNQ at a 12.5-µm-radius microdisk
electrode in either NB or DCE as described earlier.3,7 The
driving force for ET reaction was expressed as ∆E° + ∆owæ.
Although ∆owæ could not be found without an extrathermody-
namic assumption, the difference of two reversible half-wave
potentials obtained from the steady-state voltammograms of
organic and aqueous redox species (with respect to the same
Ag/AgCl reference electrode) is approximately equal to ∆E°
+ ∆owæ.3 The use of the same potential scale allowed direct
comparison of our results to those in refs 3b,c and 7.
Figure 3. Current-distance curves for a 12.5-µm-radius Pt tip
approaching the water/DCE interface. (A) Aqueous phase contained
0.1 mM Na4Ru(CN)6, 0.1 M NaCl, and 3.1 (1), 12.5 (2), and 50 (3)
mM TBACl. DCE was 2 mM in ZnPor and 50 mM in TBAPF6.
Symbols, experimental data; solid lines, theory.3a (B) Current oscilla-
tions caused by the film formation at [TBACl] ) 100 mM.
Results and Discussion
Concentration Dependencies of ∆owæ. Depending on the
phase in which the concentration of the common ion was varied,
the relative ∆owæ value is given either by eq 5a
not be seen by the naked eye, the appearance of abnormal peaks
on the current-distance curves (Figure 3B) points to the
multilayer film formation at the ITIES.3b In contrast, perchlorate
is not strongly adsorbed at the ITIES. Zhang et al. showed
∆owæ ) const + S log[common ion]w
(5a)
or by eq 5b
-
recently that the adsorption of ClO4 (if any) does not
significantly affect the kinetics of interfacial ET.6c
∆owæ ) const + S log[common ion]o
(5b)
The SECM approach curves obtained at different concentra-
tions of TBA+ in the aqueous phase (Figure 3A) show the lower
tip current at higher [TBA+]w. The apparent heterogeneous rate
constants (kb) of reaction 2b at the DCE/water interface extracted
from the approach curves using previously developed SECM
theory3a exhibit strong dependence on [TBA+]w (Figure 4A).
The apparent transfer coefficient (R) extracted from this plot is
0.56. The TBA+ adsorption at the NB/water interface was less
pronounced at [TBA+]w < 50 mM (the interfacial shape did
not change greatly when the TBA+ concentration was <50 mM).
Accordingly, the log kb vs [TBA+]w dependence in this
concentration range was relatively weak (R ) 0.1). However,
the kb value decreased markedly at higher [TBA+]w (Figure 4B).
These findings are in a sharp contrast with the results of our
where S is the slope. The experimentally determined slope values
are summarized in Table 1. In all cases, the ∆owæ vs
log[common ion] dependences were linear with high correlation
coefficients (>0.99). As discussed previously,6c ZnPor is not
suitable for obtaining potential dependences (i.e., ∆owæ vs
log[common ion]o) when ClO4- is used as a common ion. Very
small S values are obtained in such experiments because of
-
extensive ion pairing between ClO4 and a cationic form of
redox species. In contrast, using TCNQ as a redox probe, the
slope values close to 59 mV/decade were obtained for both DCE
and NB (Table 1). The almost Nernstian slope values obtained
by varying the common ion concentration in either aqueous or
organic solution indicate that ion pairing of supporting elec-
trolyte in either phase does not affect the interfacial potential
drop.
Effect of the Common Ion Adsorption. The adsorption of
ZnPor at the ITIES recently observed by Girault’s group24 can
be the reason for the potential-independent rate of ZnPor
oxidation. To check this hypothesis, we probed the kinetics of
reaction 2 with TBA+ used as a surface-active common ion
(Figure 3). Reid et al.25 observed rapid decrease in surface
tension at the liquid/liquid interface with increasing concentra-
tion of TBA+ in aqueous phase. More recently, the adsorption
of TBA+ at the water/NB interface from aqueous bromide and
chloride solutions was reported by Uchiyama et al.26 In our
experiments, at lower concentrations of TBA+ in water (e.g.,
[TBA+]w < 25 mM), the interface formed a meniscus in the
glass cell (i.e., the organic phase was convex), but at higher
concentrations, the interface became flat. Moreover, at high
[TBA+]w (i.e., >0.1 M for DCE/water and >0.2 M for NB/
water), a film forms at the interface. Although the film could
previous experiments, in which the kb of reaction 2 was
-
independent of [ClO4
]
when perchlorate was used as a
w
common ion (see inset in Figure 1B). Since neither perchlorate
nor TBA+ is expected to strongly interact with redox reactants,
the major difference between kb vs ∆o æ dependences observed
w
with those common ions is likely to be due to adsorption of
TBA+ at the interface. The adsorbed TBA+ species can displace
ZnPor from the interfacial region, thus increasing the distance
between the redox reactants and decreasing the ET rate. The
specific adsorption of the bulky TBA+ (3.85 Å radius) at the
ITIES should result in a thicker mixed solvent layer.27 This effect
may also contribute to the increase in the distance between the
aqueous and organic reactants. The larger separation distance
in turn increases the fraction of the interfacial voltage dropping
between the aqueous and organic reactants, so that the depen-
dence of the kb on ∆owæ should appear when TBA+ is used as
a common ion. Thus, the decrease in the kb with increasing
[TBA+]w can be caused by two concomitant and interrelated

Electron Transfer at Liquid/Liquid Interfaces
J. Phys. Chem. B, Vol. 106, No. 15, 2002 3937
Figure 6. Dependence of the rate of ET between ZnPor and Ru(CN)63-
on THAClO4 concentration in organic phase. The organic phase was
DCE (A) or NB (B) containing 2 mM ZnPor. The aqueous phase
contained 0.1 mM Na4Ru(CN)6, 0.1 M NaCl, and 100 mM NaClO4.
Figure 4. Dependence of the rate of ET between ZnPor and Ru(CN)63-
on TBACl concentration in the aqueous phase. In addition to TBACl,
aqueous solution contained 0.1 mM Na4Ru(CN)6 and 0.1 M NaCl. The
organic phase was DCE (A) or NB (B) containing 2 mM ZnPor and
50 mM TBAPF6. Here and in other figures horizontal dashed lines show
the diffusion limit for the ET rate measurements by SECM under
corresponding experimental conditions.
0.03 at [TBA+]w ) 10 mM (curve 1) to 0.34 at [TBA+]w ) 50
mM (curve 2) and to 0.58 at [TBA+]w ) 100 mM (curve 3).
Kakiuchi’s treatment of ionic adsorption at the ITIES28 is
useful for understanding of these results. According to that
model, at any given ∆owæ the adsorption of a common ion
occurs either from the organic or from the aqueous side of the
interface. Because the interfacial voltage was varied within a
narrow range (∼100 mV) and the experimental results point to
the adsorption of TBA+ from the aqueous phase, one can expect
no significant adsorption of TBA+ from organic phase under
our experimental conditions. (This conclusion is also supported
by essentially the same shape of the interfacial meniscus
observed for a wide range of [TBA+]o, between 10 and 200
mM, and by the absence of a multilayer interfacial film at high
[TBA+]o.) Thus, at low concentrations of TBA+ in water, its
surface concentration is also low, and ZnPor remains adsorbed
on the ITIES. Under these conditions, the ET rate constant is
essentially potential-independent (curve 1 in Figure 5A,B). At
higher [TBA+]w, ZnPor gets displaced from the interface, and
the rate constant of reaction 2b depends strongly on [TBA+]o
(curves 2 and 3 in Figure 5).
One should notice that the kb values in Figure 5 increase with
increasing [TBA+]o. This cannot be attributed to displacement
of ZnPor from the phase boundary. Unlike the kb vs [TBA+]w
dependences in Figure 4, which represent the combination of
the effects of ZnPor desorption and potential dependence of
the ET rate, the kb vs [TBA+]o dependences reflect only the
effect of ∆owæ on the ET rate. The increase in the apparent R
value with [TBA+]w can be explained by increasing the fraction
of the overall interfacial voltage dropping between ZnPor and
Figure 5. Dependence of the rate of ET between ZnPor and Ru(CN)63-
on TBAPF6 concentration in organic phase. The organic phase was
DCE (A) or NB (B) containing 2 mM ZnPor. The aqueous phase
contained 0.1 mM Na4Ru(CN)6, 0.1 M NaCl, and 10 mM (curve 1),
50 mM (curve 2), or 100 mM (curve 3) TBACl.
effectssthe increasing separation between redox species and
the potential dependence of the ET rate. These two factors can
be separated by probing the dependence of the rate constant on
[TBA+]o.
3-
Ru(CN)6 species. The diffuse double layer effect in water
cannot be responsible for the observed increase in ET rate with
[TBA+]o because the increase in [TBA+]o shifts ∆owæ to more
negative values and causes the depletion (rather than accumula-
Effect of Common Ion Concentration in Organic Phase.
In previous SECM experiments,3,6-8 the concentration of the
common ion in organic solvent was kept constant to avoid
complications, which might be caused by changes in the diffuse
layer thickness and the extent of ion pairing. As shown in Table
1, ∆owæ dependences on [TBA+]o and [ClO4-]o are essentially
Nernstian for both DCE/water and NB/water interfaces. Two
log kb vs log[TBA+]o dependencies obtained for reaction 2b at
the water/DCE interface at [TBA+]w ) 10 mM (curve 1) and
[TBA+]w ) 50 mM (curve 2) are shown in Figure 5A. While
at a lower [TBA+]w value the ET rate is independent of [TBA+]o
(R ) 0.0), at [TBA+]w ) 50 mM the linear log kb vs log[TBA+]o
dependence can be seen (R ) 0.21). At NB/water interface, the
picture is even more clearsthe apparent R value changes from
3-
tion) of negatively charged Ru(CN)6 at the interface.
Changing the electrolyte concentration in organic solution
could affect the interfacial ET rate in different ways. The ion
pairing effect discussed recently by Quinn and Kontturi29 is less
important in our system because the organic reactant (ZnPor)
is a neutral species. To check the possibility that the observed
potential dependence of ET rate is caused by diffuse layer effect
in organic phase,30 the log kb vs log[ClO4-]o dependences were
obtained by using perchlorate as a common ion (Figure 6). Very
small a values (i.e., 0.05 at the DCE/water and 0.07 at the NB
water interface) extracted from Figure 6 indicate the organic
diffuse layer effect on the rate of reaction 2b is negligibly small.

3938 J. Phys. Chem. B, Vol. 106, No. 15, 2002
Liu and Mirkin
TABLE 2: Rate Constants of Long-Range ET between
4-
ZnPor+ and Fe(CN)6 across Phosphatidylserine (PS) and
Phosphatidylcholine (PC) Monolayers Adsorbed at the
BZ/Water Interfacea
PS
kf x 103, cm/s
PC
kf x 103, cm/s (from ref 3c)
C-10
C-12
C-16
1.9
1.4
0.3
C-10
C-12
C-16
12
4.8
1.6-6.5
a BZ contained 0.5 mM ZnPor, 0.25 M THAClO4, and 100 µM lipid.
The bottom (aqueous) phase contained 7 mM Na4Fe(CN)6, 0.1 M NaCl,
and 0.1 M NaClO4.
Figure 7. Dependence of the rate constant of ET between ZnPor+
and Ru(CN)64- (filled symbols) or Fe(CN)64- (open symbols) on lipid
concentration in benzene. The number of methylene groups in the lipid
hydrocarbon chain was 10 (triangles), 12 (squares), and 16 (circles).
The organic phase contained 0.25 M THAClO4, 0.5 mM ZnPor, and
lipid. The water phase contained 0.1 M NaCl, 0.1 M NaClO4, and 7
mM Ru(CN)64- or Fe(CN)64-. Rate constants were obtained by fitting
experimental approach curves to the theory.3a
interface. Thus, similar blocking of interfacial ET by both types
of lipid films can be expected. However, the rate constants of
reaction 1b across different PS monolayers turned out to be
significantly lower than the rate constants across PC spacers
(with the same number of methylene groups) measured under
the same experimental conditions (Table 2). This cannot be
explained by smaller penetration of ZnPor+ into the lipid phase.
If PS films were less permeable to ZnPor+ than PC films, a
strong distance dependence of the ET rate would be observed.
On the contrary, the ET rates across a C-10 and a C-12
monolayers were almost identical (Figure 7), and the rate
constant for C-16 was only ∼5 times lower than that for C-10.
Unlike phosphatidylcholines, which are neutral, phosphati-
dylserine lipids are negatively charged at pH 7.32 The differences
in ET rates across PS and PC films can be related to the
Long-Distance ET across Phosphatidylserine Monolayers.
Similarly to the previous experiments with phosphatidylcholine
(PC) lipids,3c the formation of a molecular monolayer of a long-
chain phosphatidyl serine (PS) lipid impeded ET across the
liquid/liquid interface. After the addition of a lipid stock solution
to the organic phase, the measured rate constant decreased
rapidly and within 15-20 min, reached the limiting value
(corresponding to a specific concentration of lipid in organic
phase). The diffusion-controlled adsorption of PS is slightly
faster than PC adsorption, which required 30-40 min to come
to equilibrium.3c
4-
repulsion of anionic Fe(CN)6 species by negatively charged
Several dependences of kb vs lipid concentration obtained for
heads of the PS lipid surfactants. Such a repulsion may result
in both the decreased interfacial concentration of Fe(CN)64- and
the larger separation of the redox reactants. The PS films can
be rendered neutral by decreasing the aqueous phase pH to <5.
This should produce two opposite effects. A neutral PS film is
expected to be more compact than a charged film because of
lower electrostatic repulsion between lipid molecules and thus
should block ET stronger. Apparently, this effect is counterbal-
4-
two different aqueous redox species (Ru(CN)64- and Fe(CN)6
,
filled and open symbols, respectively) and three different lipids
[C-10 (∆), C-12 (0), and C-16 (O)] are shown in Figure 7. As
discussed earlier,3c the decrease in ET rate with increasing lipid
concentration points to the increasing fraction of the interfacial
area covered with lipid. At higher concentrations of any lipid
(i.e., g50 µM), the ET rate between Ru(CN)6 and ZnPor+
4-
across a complete monolayer of any lipid with a number of
methylene groups, n g 10, is immeasurably slow (filled symbols
in Figure 7). This shows that the lipid films are sufficiently
compact to make the current through defects negligibly small
under our experimental conditions. For the Ru(CN)64-/ZnPor+
reaction, the effective heterogeneous rate constant can be related
to the fraction of the ITIES area covered with lipid (θ):31,3c
4-
anced by the elimination of Fe(CN)6 repulsion discussed
above, and the net effect of the pH change on the ET rate is
small (e.g., kb ) 1.8 x 10-3 cm/s at pH 4 vs 2.2 x 10-3 cm/s at
pH 6.5 for a complete C-10 monolayer under the same
conditions). Although this explanation is tentative, quantitative
modeling of the pH effects on ET rate is problematic because
the changes in film structure can also affect the extent of ZnPor
penetration into the lipid layer.
kf(θ) ) (1 - θ)kf(0)
(6)
The ZnPor+ penetration into lipid phase prevents any
quantitative analysis of the kf vs n results in terms of a distance
dependence of ET. An attempt was made to improve the
blocking properties of the spacer by solidifying liquid lipid
monolayers. Liquid-expanded PS monolayers formed at the
ITIES at room temperature can be converted to the liquid-
condensed state or even made solid by addition of a low
concentration of a divalent cation (e.g., Ca2+ or Mg2+).20b,33
Three approach curves for the Fe(CN)64-/ZnPor ET across a
C-12 monolayer in the absence (curve 1) and presence (curves
2, 3) of Ca2+ in the aqueous phase are presented in Figure 8.
While the lower feedback current in the presence of calcium is
expected (curve 2), the increase in iT upon addition of calcium
to the aqueous phase (curve 3) is surprising. The higher value
of the effective rate constant (i.e., 0.0042 cm/s from curve 3 vs
0.0014 cm/s from curve 1 in Figure 8) points to a weaker
blocking effect exerted by a solidified lipid film. This observa-
tion suggests that the addition of calcium ion results in a
significantly higher defect density in the lipid film. This
observation is not in conflict with low values of differential
where kf(0) is the rate constant at the lipid-free ITIES. Thus,
Figure 7 essentially represents the θ vs [lipid] dependence,
which can be fit to the Langmuir isotherm, as it was shown for
PC adsorption in ref 3c.
When Ru(CN)64- is replaced with Fe(CN)64- (for which ∆E°
is about 0.5 V larger), the kf decreases markedly with increasing
concentration of lipid but does not vanish at higher concentra-
tions (open symbols in Figure 7). Instead, it reaches a limiting
value at about 50 µM and does not change at higher lipid
concentrations. This saturation points to the formation of a
complete phospholipid monolayer at the ITIES. In the same
concentration range (i.e., [lipid] J 20 µM), the formation of
compact PS monolayers was observed previously at the water/
nitrobenzene interface.20
The ET between Fe(CN)64- and ZnPor+ occurs at a measur-
able rate via tunneling through the monolayer. Kakiuchi et al.20b
reported that the differential capacity of the ITIES covered with
a PS monolayer was similar to that obtained with an analogous
PC lipid. They concluded that at room temperature both families
of lipids form liquid-expanded monolayers at the NB/water
capacity of PS monolayers measured in the presence of Ca2+ 20b
.

Electron Transfer at Liquid/Liquid Interfaces
J. Phys. Chem. B, Vol. 106, No. 15, 2002 3939
Figure 8. Approach curves for a full monolayer (100 µM lipid in
benzene) of C-12 in the absence (1) and presence (2, 3) of Ca2+ in the
aqueous phase. Curves 2 and 3 were obtained with a tip approaching
different areas of the interface. The effective rate constant values found
from the theoretical fit (solid lines) to the experimental points (symbols)
were (1) 0.0014 cm/s, (2) 0.00038 cm/s, and (3) 0.0042 cm/s. The tip
approached at 1 µm/s. Aqueous phase contained 7 mM Na4Fe(CN)6,
0.1 M NaCl, 0.1 M NaClO4, and 20 µM CaCl2 (2, 3). BZ was 0.5 mM
in ZnPor and 0.25 M in THAClO4.
Figure 10. An SECM tip was scanned in the x-direction above a C-10
monolayer adsorbed at the BZ/water interface 1 h after the addition of
20 µM Ca2+ to the aqueous phase. Other conditions are the same as
for curve 2 in Figure 9.
larger kf value) as compared to curve 2. We take the former to
correspond to domains of solidified lipid monolayer that
produces a stronger blocking effect on the ET reaction. Clearly,
a micrometer-sized zone, which produces a higher feedback
current, cannot be completely free of lipid (if this were the case,
the Fe(CN)64-/ZnPor ET reaction at such a domain would be
diffusion-controlled, and the measured feedback current would
be much higher). A more plausible explanation is that the
pinhole defects in the monolayer are clustered within such
domains. The highest feedback current in Figure 8 (curve 3)
apparently corresponds to such a domain.
The stability of the domains on the phase boundary was
checked by using forward (Figure 9, curve 1) and reverse (curve
3) scans at the same spot of the interface. Although the forward
and reverse scans are not completely retraceable and some
interfacial features change on the experimental time scale (∼2
min), the essential features of the imaged domains are reproduc-
ible. Interestingly, the distribution of defects in the monolayer
depend greatly on the way those defects were nucleated. Figure
10 shows the data obtained at a PS (C-10) monolayer, which
was adsorbed at the ITIES with no divalent cation present in
the aqueous solution. The feedback current increased signifi-
cantly after the addition of 20 µM Ca2+ to the water phase (cf.
curve 2 in Figure 9). However, no domain structure was detected
when the tip was scanned horizontally (in the x-y plane) above
the interface (Figure 10). Thus, the exposure of a preformed
PS monolayer to calcium results in the formation of pinhole
defects, which are uniformly distributed on the interface, and a
consistently increased ET rate across the film.
Figure 9. Images of the lipid films adsorbed at the BZ/water interface
obtained by scanning a 12.5-µm-radius tip along one axis parallel to
the phase boundary. The ITIES was covered with a complete monolayer
of PS lipid (C-10). Aqueous phase contained 10 mM Na4Fe(CN)6, 0.1
M NaCl, and 0.1 M NaClO4. BZ was 0.5 mM in ZnPor and 0.25 M in
THAClO4. The lipid concentration in BZ was 100 µM. The concentra-
tion of Ca2+ in the aqueous phase was (1, 3) 20 µM and (2) 0 µM. The
tip/ITIES distance was ∼4 µm. Curves 1 (forward) and 3 (reverse)
were obtained by scanning a SECM tip above the same area of the
lipid film formed in the presence of calcium ion in the water phase.
Conclusions
Even a small number of pinholes greatly increases the ET rate
across a blocking monolayer film, while the combined area of
those defects (which affects the interfacial capacitance) can be
very small.18b,d,31
The essentially potential-independent rate of heterogeneous
ET between organic ZnPor and aqueous Fe(CN)6 can be
3-
attributed to the adsorption of ZnPor at the interface. When a
surface-active common ion (e.g., TBA+) is added to the system,
it can displace ZnPor from the interface. Under these conditions,
the ET rate exhibits strong dependence on the interfacial
potential (∆owæ). The desorption of ZnPor apparently increases
the fraction of the interfacial voltage dropping between the
reactants and in this way increases the potential dependence of
To investigate the film uniformity, the tip was scanned above
the monolayer in the x-y plane parallel to the interface. (In
this way micrometer-sized domains of different reactivity have
previously been detected in monolayers consisting of two
different lipids.3d) When the tip was scanned above the interface
covered with C-10 lipid and no calcium was added to the
solution, the iT was practically constant and independent of the
(x,y) position of the tip (curve 2 in Figure 9). The interface
appears to be flat and smooth. Major fluctuations of the current
appeared after the addition of 20 µM of Ca2+ to the aqueous
layer (curve 1 in Figure 9). Two different kinds of micrometer-
sized domains can be seen in curve 1 corresponding to a lower
iT (i.e., smaller kf value) and to a higher current response (i.e.,
the ET rate. A strong dependence of kb on ∆o æ observed for
w
ET involving neutral redox species dissolved in the organic
phase suggests that the potential drop across a mixed-solvent
layer may be larger than the few millivolt value suggested
previously.14
Long-range ET was probed across 1,2-diacyl-sn-glycero-3-
phospho-L-serine monolayers. The blocking properties of these

3940 J. Phys. Chem. B, Vol. 106, No. 15, 2002
Liu and Mirkin
(11) (a) Geblewicz, G.; Schiffrin, D. J. J. Electroanal. Chem. 1988, 244,
27. (b) Cunnane, V. J.; Schiffrin, D. J.; Beltran, C.; Geblewicz, G.; Solomon,
T. J. Electroanal. Chem. 1988, 247, 203. (c) Cheng, Y.; Schiffrin, D. J. J.
Electroanal. Chem. 1991, 314, 153.
(12) Cai, C.; Mirkin, M. V., unpublished results.
(13) Fermin, D. J.; Ding, Z.; Duong, H. D.; Brevet, P.-F.; Girault, H.
films turned out to be similar to those previously reported for
phosphatidylcholine monolayers.3c A relatively weak distance
dependence of ET rate points to partial penetration of ZnPor+
into the lipid monolayer. An attempt to solidify phosphatid-
ylserine films by adding calcium ions to the aqueous phase
resulted in a formation of two types of domains in the
monolayer. Although micrometer-sized domains in lipid mono-
layers often form under the increased surface pressure and also
in the mixed films, the observation of a domain structure in a
monolayer formed from a single lipid at the ITIES is not
common. The properties of the film depend dramatically on the
adsorption conditions, thus indicating the importance of nucle-
ation phenomena for the formation of the domains.
H. J. Phys. Chem. B 1998, 102, 10334.
(14) (a) Girault, H. H.; Schiffrin, D. J. J. Electroanal. Chem. 1988, 244,
15. (b) Katano, H.; Maeda, K.; Senda, M. J. Electroanal. Chem. 1995, 396,
391. (c) Schmickler, W. J. Electroanal. Chem. 1997, 428, 123.
(15) Hanzlik, J.; Hovorka, J.; Samec, Z.; Toma, S. Collect. Czech. Chem.
Commun. 1988, 53, 903.
(16) Ding, Z.; Fermin, D. J.; Brevet, P.-F.; Girault, H. H. J. Electroanal.
Chem. 1998, 458, 139.
(17) (a) Long-Range Electron-Transfer in Biology; Palmer, G. A., Ed;
Structure and Bonding 75; Springer-Verlag: Berlin, 1991. (b) Lahiri, J.;
Fate, G. D.; Ungashe, S. B.; Groves, J. T. J. Am. Chem. Soc. 1996, 118,
2347.
(18) (a) Chidsey, C. E. D. Science 1991, 251, 919. (b) Finklea, H. O. In
Electroanalytical Chemistry; Bard, A. J., Ed.; Marcel Dekker: New York,
1996; Vol. 19, p 109. (c) Sachs, S. B.; Dudek, S. P.; Hsung, R. P.; Sita, L.
R.; Smalley, J. F.; Newton, M. D.; Feldberg, S. W.; Chidsey, C. E. D. J.
Am. Chem. Soc. 1997, 119, 10563. (d) Miller, C. J. In Physical Electro-
chemistry: Principles, Methods, and Applications; Rubinstein, I., Ed.;
Marcel Dekker: New York, 1995; p 27.
Acknowledgment. The support by the donors of the
Petroleum Research Fund administrated by the American
Chemical Society, NSF Division of International Programs, and
a grant from PSC-CUNY are gratefully acknowledged. The
software for SECM measurements was generously provided by
Prof. D. O. Wipf (Mississippi State University).
(19) (a) Finklea, H. O.; Liu, L.; Ravenscroft, M. S.; Punturi, S. J. Phys.
Chem. 1996, 100, 18 852. (b) Slowinski, K.; Chamberlain, R. V.; Miller,
C. J.; Majda, M. J. Am. Chem. Soc. 1997, 119, 11 910. (c) Slowinski, K.;
Fong, H. K. Y.; Majda, M. J. Am. Chem. Soc. 1999, 121, 7257.
(20) (a) Kakiuchi, T. In Liquid-Liquid Interfaces. Theory and Methods;
Volkov, A. G.; Deamer, D. W., Eds; CRC Press: Boca Raton, FL, 1996;
p 317. (b) Kakiuchi, T.; Kondo, T.; Senda, M. Bull. Chem. Soc. Jpn. 1990,
63, 3270.
(21) Krause, R. A.; Violette, C. Inorg. Chim. Acta 1986, 113, 161.
(22) Shao, Y.; Mirkin, M. V.; Rusling, J. F. J. Phys. Chem. B 1997,
101, 3202.
(23) Wipf, D. O.; Bard, A. J. J. Electrochem. Soc. 1991, 138, 489.
(24) Girault, H. H., unpublished results.
References and Notes
(1) Liquid Interfaces in Chemical, Biological, and Pharmaceutical
Applications; Volkov, A. G., Ed.; Marcel Dekker: New York, 2001.
(2) (a) Amemiya, S.; Ding, Z.; Zhou, J.; Bard A. J. Electroanal. Chem.
2000, 483, 7. (b) Barker, A. L.; Gonsalves, M.; Unwin, P. R. Anal. Chim.
Acta 1999, 385, 223. (c) Mirkin, M. V.; Tsionsky, M. In Scanning
Electrochemical Microscopy; Bard, A. J., Mirkin, M. V., Eds.; Marcel
Dekker: New York, 2001; p 299.
(3) (a) Wei, C.; Bard, A. J.; Mirkin, M. V. J. Phys. Chem. 1995, 99,
16 033. (b) Tsionsky, M.; Bard, A. J.; Mirkin, M. V. J. Phys. Chem. 1996,
100, 17 881. (c) Tsionsky, M.; Bard, A. J.; Mirkin, M. V. J. Am. Chem.
Soc. 1997, 119, 10 785. (d) Delville, M.-H.; Tsionsky, M.; Bard, A. J.
Langmuir 1998, 14, 2774.
(4) Solomon, T.; Bard, A. J. Anal. Chem. 1995, 67, 2787.
(5) Selzer, Y.; Mandler, D. J. Electroanal. Chem. 1996, 409, 15.
(6) (a) Barker, A. L.; Macpherson, J. V.; Unwin, P. R. J. Phys. Chem.
B 1998, 102, 1586. (b) Zhang, J.; Unwin, P. R. J. Phys. Chem. B 2000,
104, 2341. (c) Zhang, J.; Barker, A. L.; Unwin, P. R. J. Electroanal. Chem.
2000, 483, 95. (d) Barker, A. L.; Unwin, P. R.; Bard, A. J. J. Phys. Chem.
B 1999, 103, 7260.
(7) Liu, B.; Mirkin, M. V. J. Am. Chem. Soc. 1999, 121, 8352.
(8) Ding, Z.; Quinn, B. M.; Bard, A. J. J. Phys. Chem. B 2001, 105,
6367.
(9) Zhang, Z.; Yuan, Y.; Sun, P.; Su, B.; Guo, J.; Shao, Y.; Girault,
H. H. J. Phys. Chem. B, in press.
(10) (a) Marcus, R. A. J. Phys. Chem. 1990, 94, 1050. (b) Marcus, R.
A. J. Phys. Chem. 1990, 94, 4152. (c) Marcus, R. A. J. Phys. Chem. 1991,
95, 2010.
(25) Reid, J. D.; Melroy, O. R.; Buck, R. P. J. Electroanal. Chem. 1983,
147, 71.
(26) Uchiyama, Y.; Kitamori, T.; Sawada, T. Langmuir 2000, 16, 6597.
(27) Girault, H. H. Electrochim. Acta 1987, 32, 383.
(28) Kakiuchi, T. J. Electroanal. Chem. 2001, 496, 137.
(29) Quinn, B.; Kontturi, K. J. Electroanal. Chem. 2000, 483, 124.
(30) Pereira, C. M.; Silva, F.; Sousa, M. J.; Kontturi, K.; Murtoma¨ki,
L. J. Electroanal. Chem. 2001, 509, 148.
(31) (a) Amatore, C.; Save´ant, J. M.; Tessier, D.J. Electroanal. Chem.
1983, 147, 39. (b) Forouzan, F.; Bard, A. J.; Mirkin, M. V. Isr. J. Chem.
1997, 37, 155.
(32) Seimiya, T.; Ohki, S. Biochim. Biophys. Acta 1973, 298, 546.
(33) Ohnishi, S.; Ito, T. Biochemistry 1974, 13, 881.