Electrochemistry Communications 10 (2008) 10741077

Contents lists available at ScienceDirect

Electrochemistry Communications
journal homepage: www.elsevier.com/locate/elecom

Three-phase electrochemistry. Inuence of temperature on ion transfer

Elzbieta Bak, Mikolaj Donten, Zbigniew Stojek *
Department of Chemistry, Warsaw University, ul. Pasteura 1, PL-02-093 Warsaw, Poland

a r t i c l e i n f o a b s t r a c t

Article history: A study on the inuence of temperature and time on ion transfer across the liquid|liquid interface is
Received 19 April 2008 
reported. Three anions of different hydrophobic properties were selected for the examination: ClO4 ,
Received in revised form 2 May 2008 NO and Cl 
. It was found that while the current magnitudes of the oxidation of dmfc in NB in the pres-
3
Accepted 7 May 2008
ence of selected anions in water are very different at room temperature, they become comparable above
Available online 15 May 2008
45 C. This nding indicates that the ability of the examined anions to reach signicant concentrations in
the organic phase, at 45 C, becomes similar. For the anions of lower afnity to the organic phase (NO 3
Keywords:
and Cl), distinct current oscillations appeared in the recorded experimental voltammograms and
Three-phase boundary
Cylindrical microelectrode
amperograms.
Nitrobenzene 2008 Elsevier B.V. All rights reserved.
Voltammetry
Chronoamperometry
Decamethylferrocene

1. Introduction
Three-phase electrochemistry or three-phase electrodes are
employed in several elds of electrochemistry and electroanalysis
[16]. One of the interesting aspects of three-phase electrochemis-
try is the investigation of ion transfer across the phase boundary
and of the electrode reactions that require such a transfer of ions.
The electrochemical examination of ion transfer across the liquid|
liquid interface started with the four-electrode arrangement [7],
however, due to serious limitations of that setup, the four-elec-
trode system has been replaced with the three-electrode one.
Marken et al. [810] deposited an oily electroactive compound
in the form of several droplets on the graphite electrode surface
and in such a way a three-phase junction electrode/organic
phase/aqueous phase was formed. Scholz et al. [1113] replaced
the oxidation of pure organic phase with the oxidation of an elect-
roactive compound in different organic solvents. In their approach
a droplet of organic liquid with dissolved hydrophobic electroac-
tive compound was attached to the graphite electrode surface.
The electrode was immersed in an aqueous solution containing
various inorganic salts. Opallo and coworkers used electrodes
based on hydrophobic porous silicate matrix and homogeneously
dispersed electrode conductor (graphite particles) [1416]. Such
electrodes were easily impregnated with hydrophobic organic
non-polar solvents. Similarly as before, the transfer of ions was
related to their lipophilicity and thus could be monitored as a
function of electrode potential. Li et al. [17] have shown that
micro- and nanopipettes are very useful in studying the transport
of uncharged and charged species at the water/1,2-dichloroethane
interface with low concentrations of or without supporting electro-
lyte. We have proposed a new three-electrode setup [18] in which
the graphite rod was replaced with a very thin gold or platinum
wire which was placed in a mixture of two immiscible liquids in
such a way that the upper part of the wire remained in the aqueous
phase and its lower part was totally immersed in organic liquid.
The proposed conguration provides reproducible and well de-
ned and controllable length of three-phase boundary. The organic
phase can be also represented by a drop dispensed from a capillary
[19]. By applying the microcylindrical approach and an electropo-
lymerization process it was possible to examine how deeply the
particular anions can penetrate the organic phase [20].
Other recent developments in the three-phase electrochemistry
include: (1) the nding that for very hydrophilic anions the cat-
ionic product may be transferred from the organic phase to water
[21], and (2) for prolonged electrolysis times the precipitation of
the dmfc cation with the corresponding anion may take place [22].
Since a change in temperature and prolonged electrolysis time
may affect substantially the ion ability to exist in the organic phase
at higher concentrations, we demonstrate in this paper, how tem-
perature and time affects the process of ion transfer in three-phase
electrochemistry.
2. Experimental

* Corresponding author. Tel.: +48 22 220 211x336; fax: +48 22 822 4889. Voltammetric measurements were performed in the three-elec-
E-mail address: stojek@chem.uw.edu.pl (Z. Stojek). trode system using an AUTOLAB, model PGSTAT 30, potentiostat

1388-2481/$ - see front matter 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.elecom.2008.05.019
 E. Bak et al. / Electrochemistry Communications 10 (2008) 10741077 1075

(Eco-Chemie, Utrecht, The Netherlands). A special thermostated

voltammetric cell has been constructed for the measurements. A
A
15
platinum cylindrical electrode of 300 lm in diameter served as 20
a
b
the working electrode. A silver/silver chloride/saturated KCl (E =
0.199 V vs. NHE) was used as the reference electrode and a plati- c
10 a

i / A
num wire was the auxiliary electrode. The organic phase in form
of a drop was dispensed from a capillary with the help of a preci- 10 b
0
sion dispenser. The cell was enclosed in a Faraday cage to minimise c
the electrical noise. ClO4-
-10
Lithium perchlorate (LiClO4), sodium chloride (NaCl) and lith-
-0.4 -0.2 0.0 0.2

i / A
ium nitrate (LiNO3) were of high purity (pure p.a.) and were pur- E / V vs Ag/AgCl
5
chased from Fluka. Decamethylferrocene and nitrobenzene were
purchased from Aldrich.
A detailed description how a platinum microwire can be im-
mersed partially into a nitrobenzene drop and partially in the
aqueous phase can be found in [19]. 0

3. Results and discussion

In one of our previous papers [22] we have found that under -5 anion NO3-, T = 13 oC
some conditions decamethylferrocenium salts may precipitate in
the solution. Such precipitation must affect immediately local con- -0.4 -0.2 0.0 0.2 0.4 0.6
ductivity in the organic phase, so we thought it might trigger the E / V vs Ag/AgCl
appearance of oscillations in the measured current. Since, it is
known that solubility of chemical compounds and the mass-trans-
B
port rate are usually strongly temperature dependent, we have 50 40 a
b
decided to examine the dmfc electrode processes in a wider time
c
range and at different temperatures compared to earlier papers. 20
a
It appeared that in the case of anions of more positive transfer 40
i / A

potentials (more hydrophilic) and, consequently, of limited ability 0

to move into the organic phase, current oscillations indeed can be
seen in the measured responses. For NO 
3 and Cl these oscillations
30 ClO4-
-20
were found in both: cyclic voltammograms and chronoampero- -0.4 -0.2 0.0 0.2 b
metric transients.
i / A

E / V vs Ag/AgCl
In the case of NO 20
3 the oscillations were observed in cyclic vol- c
tammograms in the entire investigated temperature range, i.e.
from 13 to 55 C. Exemplary cyclic voltammograms of decamethyl-
ferrocene obtained in the presence of LiNO3 in the aqueous phase 10
at two selected temperatures (13 and 25 C) are shown in Figs.
Fig. 1A and B. At small scan rates (longer experimental times),
0
for both selected temperatures, the oscillations appear earlier
and usually have larger amplitudes (see curves a in Figs. 1A and
B). At higher scan rates the oscillations are not seen in the rst sev-
-10 anion NO3-, T = 25 oC
eral scans (see curves c) and appear eventually in consecutive

scans. Interestingly, when ClO4 , was present in the aqueous phase,
-0.4 -0.2 0.0 0.2 0.4 0.6
no current oscillations were observed at any temperature; see the
insets in Fig. 1A and B. In chronoamperometry, the appearance of E / V vs Ag/AgCl
the oscillations NO 3 is proceeded by a delay time of ca. 100 s. Fig. 1. Cyclic voltammetric curves for 0.05 M decamethylferrocene obtained with
Once the oscillations appear they maintain their frequency at a the 300-lm Pt microcylindrical electrode. Depth of penetration of nitrobenzene
rather constant level. For the explanation of the histeresis in the drop: 1500 lm. Aqueous phase 1 M LiNO3; scan rate: 2 (a), 5 (b) and 20 mV/s (c);
voltammograms see Ref. [22]. temperature: 13 C (A) and 25 C (B). Inset: aqueous phase 1 M LiClO4; all other
conditions as in A and B.
Cyclic voltammograms obtained in the presence of the chloride
ions in the aqueous phase also exhibit the oscillations. There are, to move into the organic phase. The examined temperature range
however, some differences: at lower scan rates the magnitude of could not be too wide, since at high temperatures the mutual sol-
oscillations is smaller compared to nitrates and oscillations disap- ubility of water and nitrobenzene signicantly increases. For
pear at temperatures higher than 45 C. example, a temperature increase from 20 to 60 C is related to an
At low dmfc concentrations the oscillations appeared for none increase in the solubility of water in nitrobenzene from 0.24% to
of the examined anions. 0.7% m/m, and of nitrobenzene in water from 0.19% to 0.4% m/m
[23]. In the temperature measurements the main emphasis was
3.1. Dependence of current on temperature put on more hydrophilic anions like NO 
3 and Cl , however; we

started with ClO4 which is the most hydrophobic among the exam-
The inuence of temperature on the magnitudes of the oxida- ined anions, enters well the organic phase [24] and therefore was
tion peak currents of decamethylferrocene was investigated in expected to give simple, linear Arrhenius-like (ln i vs. 1/T) depen-

the presence of three selected anions (ClO4 , NO 
3 , and Cl ) in the dencies. Reliable activation-energy data cannot be obtained from
aqueous phase. The selected anions represent different abilities these dependencies, since the rate-limiting process is unknown.
1076 E. Bak et al. / Electrochemistry Communications 10 (2008) 10741077

Cyclic voltammograms of decamethylferrocene obtained in the

presence of ClO4 in the aqueous phase, as expected, revealed that
the measured current increased steadily with potential and de-
creased with increasing scan rate. It also increased with increasing
temperature. The current magnitudes measured at 45 C were ca.
twice as large as those obtained at 18 C.
The next investigated anion was NO 3 . The magnitudes of the
oxidation currents of decamethylferrocene recorded in the pres-
ence of NO 3 at various temperatures are plotted vs. scan rate in
Fig. 2A. These plots were subsequently transformed into ln i vs.
1/T dependencies which are shown in Fig. 2B. Generally, due to
the existence of oscillations, the reading of the current magnitudes
recorded at different temperatures was difcult. For preparing the
A ion (DMFc+), in the presence of very hydrophilic anion in the aque-
140
120
100
80
i / A
plots, the current values were averaged over a range of data points.
For relatively high scan rates the linear plots were obtained in the
entire temperature range. Their slopes have rather big value
which may mean that the current increase is not just due to the
transport increase; it may be caused by a signicant increase of
solubility of decamethylferrocene nitrate with the increase in
temperature.

Temperature investigations similar to those done for the ClO4
 
and NO3 anions were performed for Cl . However, the obtained
ln i vs. 1/T plots differed signicantly from the dependencies
constructed for NO 3 . The striking difference is the absence of a rec-
tilinear part in the plots; they are parabolic. However, the low-
temperature fragment of the plot has again a high value of the
slope. The reason for the parabolic shape of the plot is not clear.
We suggest a preliminary explanation here: this shape of the plot
might be connected with the transfer of decamethylferrocene cat-
ous phase, from nitrobenzene to water. Such a transfer has been
suggested in papers [21,22,15].
Temperature dependencies of ln i for all investigated anions
at the highest applied scan rate of 100 mV/s are presented in
Fig. 3. At this scan rate no oscillations or transport limitations
were observed, since only a small part of the microcylinder
was appropriately polarised within a short experimental time-
scale and the concentration of the anions could not grow much.


As it was mentioned above, for ClO4 a linear dependence of the
60 logarithm of the current on the temperature reciprocal was
expected. Surprisingly, the ln i vs. 1/T plots consisted of two
linear segments of two distinctly different slopes: the ratio of
40 55 oC
the slope values is 6.31. This means that in two temperature
45 oC ranges seen two different phenomena control the current. An
20 35 oC important conclusion that can be drawn from Fig. 3 is that the
25 oC current magnitudes of decamethylferrocene obtained in the pres-
13 oC ence of various anions at temperatures higher than 45 C are
0 comparable.

0 20 40 60 80 100 120
-1
v / mV s

B 5.0 4.0
a
v = 100 mV/s

55 oC
4.5 3.5 45 oC
b 35 oC
25 oC
c

4.0 d
3.0
e 18 oC
ln (i / A)

f
ln (i / A)

13 oC
3.5 2.5
ClO4-

3.0 2.0
NO3-

2.5
1.5

Cl-
2.0
1.0
0.0030 0.0031 0.0032 0.0033 0.0034 0.0035 0.0030 0.0031 0.0032 0.0033 0.0034 0.0035 0.0036
1 / T (K-1) 1 / T (K )
-1

Fig. 2. (A) The dependence of the magnitude of the oxidation current of decam-
ethylferrocene measured at 0.41 V on the scan rate at various temperatures. Con-
centration of dmfc: 0.05 M, concentration of LiNO3 in the aqueous phase: 1 M. (B) ln
i vs. 1/T plots for the oxidation of DMFc and at various scan rates: 2 (a), 5 (b), 10 (c),
20 (d), 50 (e), and 100 mV/s (f).
Fig. 3. ln i vs. 1/T plots for the oxidation of DMFc in the presence of various anions
in the aqueous phase. Concentration of LiClO4 (0.25 V), LiNO3 (0.41 V) and NaCl
(0.46 V): 1 M; scan rate: 100 mV/s. The currents were measured at the same ove-
rpotential (400 mV) vs. the potential of ion transfer, i.e. at the potentials 0.25, 0.41
and 0.46 V vs. Ag/AgCl, respectively.
 E. Bak et al. / Electrochemistry Communications 10 (2008) 10741077
4. Conclusions
The inuence of temperature on the magnitudes of the oxida-
tion currents of decamethylferrocene was investigated in the pres-

ence of three selected anions (ClO4 , NO 
3 and Cl ) in the aqueous
phase. The selected anions represent different hydrophobic proper-
ties and therefore different abilities to enter the organic phase. The
oxidation current of decamethylferrocene distinctly depended on
temperature. It appeared that while at room temperature the cor-
responding currents differed much the current magnitudes of
decamethylferrocene obtained in the presence of various anions
at temperatures higher than 45 C are comparable. Therefore it
can be concluded that under such temperature conditions the abil-
ities of all examined anions to exist in the organic phase at a signif-
icant concentration (due to increased solubility of dmfc+ salts) are
high enough and are no longer the limiting factors for the mea-
sured responses. The kinetic effect of increased ion and molecule
transport rate cooperates here with the solubility effect. Measure-
ments at higher temperatures, due to signicant changes in the
solubility of the participating species, may also inuence the phe-
nomena connected with the formation of emulsions at the oil/
water interface [25].
Looking for the reason for the formation of the current oscilla-
tions in the chronoamperograms and the cyclic voltammograms
one should note that they are formed only in the presence of those
anions that hardly enter the organic phase and, consequently, their
concentration in that phase is rather low. Decamethylferrocenium
is the only available cation in the organic phase. We already know
that it may form a hardly soluble salt with one of the investigated
anions and precipitate while the electrode process lasts [22]. Since
any precipitation of the salt must be accompanied by a change in
local resistivity in the organic phase, and this must trigger the cor-
responding change in the ohmic potential drop, in consequence,
the measured current must drop. This phenomenon should be peri-
odical, since after the precipitation the concentration of the anion
in the organic phase may increase again while the electrode pro-
cess lasts. Then local oversaturation may be reached and nally
the next precipitation event occurs.
Looking for an alternate reason of oscillations we want to stress
that the precipitation may also affect the surface tension at the
nitrobenzene/water interface. The presence of the precipitate on
the electrode surface may change its wettability by each of two sol-
vents. A change in wettability of the electrode surface at the inter-
face may generate various movements of the nitrobenzene droplet.
Such movements (short lasting vibrations) were detected using the
1077
optical microscope. They were well correlated with the current
spikes. Current instabilities caused by unstable liquid/liquid inter-
phase, and convection caused by a redox reaction at the three-
phase boundary have been reported recently [26,27].
Acknowledgement
Support for this work by the Polish Ministry of Education and Sci-
ence under Grant No. N-N204 244534 is gratefully acknowledged.
References
[1] J.O. Besenhard (Ed.), Handbook of Battery Materials, Wiley-VCH, Weinheim,
1999.
[2] W. Vielstich, A. Lamm, H.A. Gasteiger (Eds.), Handbook of Fuel Cells:
Fundamentals, Technology and Applications, John Wiley & Sons, Ltd., New
York, 2003.
[3] A. Walcarius, Anal. Chim. Acta 384 (1999) 1.
[4] F. Ricci, G. Palleschi, Biosens. Bioelectron. 21 (2005) 389.
[5] F. Scholz, U. Schrder, R. Gulaboski, Electrochemistry of Immobilized Particles
and Droplets, Springer, Berlin, 2005.
[6] N. Katif, S.M. MacDonald, A.M. Kelly, E. Galbraith, T.D. James, A.T. Lubben, M.
Opallo, F. Marken, Electroanalysis 20 (2008) 469.
[7] H. Girault, D.J. Schiffrin, in: A.J. Bard (Ed.), Electroanalytical Chemistry, vol. 15,
Springer, Berlin, 1989, pp. 1141.
[8] F. Marken, R.D. Webster, S.D. Bull, S.G. Davies, J. Electroanal. Chem. 437 (1997)
209.
[9] F. Marken, R.G. Compton, Ch.H. Goeting, J.S. Foord, S.D. Bull, S.G. Davies,
Electroanalysis 10 (1998) 821.
[10] C.E. Banks, T.J. Davies, R.G. Evans, G. Hignett, A.J. Wain, N.S. Lawrence, J.D.
Wadhawan, F. Marken, R.G. Compton, Phys. Chem. Chem. Phys. 5 (2003) 4053.
[11] F. Scholz, . Komorsky-Lovric, M. Lovric, Electrochem. Commun. 2 (2000) 112.
[12] . Komorsky-Lovric, K. Riedl, R. Gulaboski, V. Mirceski, F. Scholz, Langmuir 18
(2002) 8000.
[13] R. Gulaboski, A. Galland, G. Bouchard, K. Caban, A. Kretschmer, P.A. Carrupt, Z.
Stojek, H.H. Girault, F. Scholz, J. Phys. Chem. B 108 (2004) 4565.
[14] G. Shul, M. Opallo, F. Marken, Electrochim. Acta 50 (2005) 2315.
[15] G. Shul, M. Opallo, Electrochem. Commun. 7 (2005) 194.
[16] E. Rozniecka, G. Shul, J. Sirieix-Plenet, L. Gaillon, M. Opallo, Electrochem.
Commun. 7 (2005) 299.
[17] F. Li, Y. Chen, M. Zhang, P. Jing, Z. Gao, Y. Shao, J. Electroanal. Chem. 579 (2005)
89.
[18] E. Bak, M. Donten, Z. Stojek, Electrochem. Commun. 7 (2005) 483.
[19] E. Bak, M. Donten, Z. Stojek, F. Scholz, Electrochem. Commun. 9 (2007) 386.
[20] E. Bak, M. Donten, M. Skompska, Z. Stojek, J. Phys. Chem. B 110 (2006) 25635.
[21] G. Shul, W. Nogala, I. Zakorchemna, J. Niedziolka, M. Opallo, J. Solid State
Electrochem., doi:10.1007/s10008-008-0512-8.
[22] M. Donten, E. Bak, M. Gniadek, Z. Stojek, Electrochim. Acta 53 (2008) 5608.
[23] A. Seidell, A Compilation of Quantitative Solubility Data from the Periodical
Literature, vol. II, D. Van Nostrand Company Inc., New York, 1941. pp. 136, 436.
[24] E. Bak, M.L. Donten, M. Donten, Z. Stojek, Electrochem. Commun. 7 (2005)
10981104.
[25] J.Y. Chen, M. Sato, J. Electroanal. Chem. 572 (2004) 153.
[26] T. Kakiuchi, N. Nishi, T. Kasahara, M. Chiba, ChemPhysChem 4 (2003) 179.
[27] M. Satoh, K. Aoki, J. Chen, Langmuir 24 (2008) 4364.