Quantitative in situ measurement of ion transport in polypyrrole/poly(styrenesulfonate) films using rotating ring-disk voltammetry

Article abstract of DOI:10.1021/ac990139w

An approach based on rotating ring-disk electrode (RRDE) voltammetry is described for the quantitative, in situ measurement of ion transport between solution and conducting polymer films. The specific composite film studied in this report is polypyrrole/poly(styrenesulfonate) (pPy⁺/pSS^-). Cation flux in and out of the polymer was obtained from the mass-transport-limited reduction current for the dopant cation(s) measured at the ring during redox cycling of the polymer. Crucial to this method is the use of a supporting electrolyte that is sterically inhibited from passing into the film and the use of dopant ions that adhere to specific electrochemical constraints. With this method it was possible to quantitatively account for all changes in charge compensation in the film by the specific cation(s) involved. Three different cations were explored alone and in paired combinations. Solutions containing mixtures of dopant cations were studied to determine whether the pPy⁺/pSS^- films exhibit preferential doping. Kinetic factors, likely due to steric differences in the dopant cations, were found to lead to significant preferential doping of the polymer.

Full text of DOI:10.1021/ac990139w

Anal. Chem. 1999, 71, 3677-3683
Quantitative in Situ Measurement of Ion Transport
in Polypyrrole/Poly(styrenesulfonate) Films Using
Rotating Ring-Disk Voltammetry
Corey A. Salzer and C. Michael Elliott*
Department of Chemistry, Colorado State University, Fort Collins, Colorado 80523
Susan M. Hendrickson
Department of Chemistry, Davidson College, Davidson, North Carolina 28036
at least in part to ion influx/ efflux, is then correlated to the change
in polymer oxidation state obtained from its voltammetry or
coulometry. The advantages of EQCM are its high sensitivity to
mass changes and in situ capabilities. However, it suffers from a
lack of selectivity in identifying specific dopants and in differentiat-
ing solvent molecules from ions.5,7 Each of the other approaches
listed above has its own advantages and limitations as well. For
instance, luminescence is selective and sensitive, but is limited
to ions that luminesce and, consequently, are typically rather large.
On the other hand, XPS can be used to examine a wide array of
ions, but lacks the ability to measure ion transport in situ. Both
luminescence and SECM suffer from the fact that the polymer
must be charged with the probe ion prior to voltammetric
scanning. Consequently, after the first scan, probe ion transport
is convoluted with supporting electrolyte cation and anion trans-
port.
An approach based on rotating ring-disk electrode (RRDE)
voltammetry is described for the quantitative, in situ
measurement of ion transport between solution and
conducting polymer films. The specific composite film
studied in this report is polypyrrole/ poly(styrenesulfonate)
(pP y+/ pSS-). Cation flux in and out of the polymer was
obtained from the mass-transport-limited reduction cur-
rent for the dopant cation(s) measured at the ring during
redox cycling of the polymer. Crucial to this method is
the use of a supporting electrolyte that is sterically
inhibited from passing into the film and the use of dopant
ions that adhere to specific electrochemical constraints.
With this method it was possible to quantitatively account
for all changes in charge compensation in the film by the
specific cation(s) involved. Three different cations were
explored alone and in paired combinations. Solutions
containing mixtures of dopant cations were studied to
determine whether the pP y+/ pSS- films exhibit prefer-
ential doping. Kinetic factors, likely due to steric differ-
ences in the dopant cations, were found to lead to
significant preferential doping of the polymer.
At least two prior attempts to use rotated ring-disk electrode
(RRDE) voltammetry to measure ion flux across a conducting
polymer/ solution interface have been reported. Aizawa and co-
workers13 monitored the oxidation of Br- at the ring electrode
while a polypyrrole film on the disk was redox-cycled in 0.10 M
NaBr/ water-acetonitrile (1:1). Despite efforts to quantitate these
data, these results are of only qualitative value because NaBr was
the sole electrolyte in solution. Migration of Br- was neglected
in their data treatment. Earlier yet, Pickup and Osteryoung14 used
RRDE voltammetry to monitor Cl- flux from polypyrrole in a very
A number of analytical techniques have been employed to
study ion transport in conducting polymers, including impedance
spectroscopy,1,2 X-ray photoelectron spectroscopy (XPS),1,3,4 lu-
minescence techniques,5,6 and scanning electrochemical micros-
copy (SECM),7 usually combined with electrochemical methods
such as cyclic voltammetry or coulometry. However, the most
commonly used approach is to monitor in situ mass changes of
the polymer by means of electrochemical quartz crystal micro-
gravimetry (EQCM).4,6,9-12 The change in mass of a polymer, due
-
slightly basic (i.e., a slight excess of Cl-) AlCl4 molten salt. In
contrast to the work of Aizawa and co-workers, these data were
quantitatively useful since [Cl-] , [AlCl4-]. Unfortunately, as
these authors were also able to show, Cl- was not the only dopant
ion in the polymer; consequently, it was not possible to determine
exactly which ions were participating in the charge compensation
process. As we will demonstrate below, the type of difficulties
encountered in these earlier studies can be avoided altogether
with the right combination of dopant ion(s) and background
electrolyte. Moreover, it is possible with this technique to
quantitatively account for all of the ions participating in polymer
doping level changes in situ and in real time.
(1) Ren, X.; Pickup, P. G. J. Phys. Chem. 1 9 9 3 , 97, 5356.
(2) Ren, X.; Pickup, P. G. Electrochim. Acta 1 9 9 6 , 41, 1877.
(3) Bach, C. M. G.; Reynolds, J. R. J. Phys. Chem. 1 9 9 4 , 98, 13636.
(4) Bose, C. S. C.; Basak, S.; Rajeshwar, K. J. Phys. Chem. 1 9 9 2 , 96, 9899.
(5) Krishna, V.; Ho, Y.-H.; Rajeshwar, K. J. Am. Chem. Soc. 1 9 9 1 , 113, 3325.
(6) Reynolds, J. R.; Pyo, M.; Qiu, Y.-J. Synth. Met. 1 9 9 3 , 55-57, 1388.
(7) Arca, M.; Mirkin, M. V.; Bard, A. J. J. Phys. Chem. 1 9 9 5 , 99, 5040.
(8) Naoi, K.; Lien, M.; Smyrl, W. H. J. Electrochem. Soc. 1 9 9 1 , 138, 440.
(9) Lien, M.; Smyrl, W. H.; Morita, M. J. Electroanal. Chem. 1 9 9 1 , 309, 333.
(10) Baker, C. K.; Qui, Y.-J.; Reynolds, J. R. J. Phys. Chem. 1 9 9 1 , 95, 4446.
(11) Lim, J. Y.; Paik, W.; Yeo, I.-H. Synth. Met. 1 9 9 5 , 69, 451.
(12) Li, Y.; Liu, Z. Synth. Met. 1 9 9 8 , 94, 131.
(13) Shinohara, H.; Kojima, J.; Aizawa, M. J. Electroanal. Chem. 1 9 8 9 , 266, 297.
(14) Pickup, P. G.; Osteryoung, R. A. J. Electroanal. Chem. 1 9 8 5 , 195, 271.
10.1021/ac990139w CCC: $18.00 © 1999 American Chemical Society
Published on Web 07/24/1999
Analytical Chemistry, Vol. 71, No. 17, September 1, 1999 3677

BACKGROUND
Before considering how RRDE voltammetry can be applied to
polymer doping studies, it is instructive to review how this
technique is employed to study so-called EC reactions in solution.15
Consider the following hypothetical EC reaction:
E: O + e- T R
C: R 9
k8 Z
Typically, the electrochemical reaction, E, is initiated at the disk.
The ring potential is adjusted to give a mass-transport-limited
reoxidation of R back to O. Because of the electrode rotation, the
R formed at the disk is hydrodynamically transported across the
insulating gap toward the ring where it can be reoxidized. Were
it not for the chemical step, which irreversibly consumes some
of the fraction of the R before it reaches the ring, the ring and
disk currents would be related by eq 1. The quantity N is a
Figure 1. Cyclic voltammograms of pPy+/pSS- films (grown under
identical conditions) cycled in acetonitrile solutions of 2 mM dopant
cation and 50 mM PPNTos: cobalticinium (Cc+, solid line); 1-methyl-
3-cyanopyridinium (CMP+, dash-dot); 1,3-dimethylpyridinium (DMP+,
dash only).
-(ir)/ (id) ) N
(1)
doping; otherwise, any effort to quantitatively correlate the disk
current with total ion transport would be pointless. Fortunately,
differences in ion size can be used effectively to discriminate
against doping by the supporting electrolyte ions. For example,
the supporting electrolyte cation employed in this study, bis-
(triphenylphosphoranylidene)ammonium (PPN+), is sufficiently
large such that it is completely excluded from entering the
polymer (vide infra). It is the use of this sterically excluded electrolyte
that circumvents the problems encountered in earlier attempts to
quantitate dopant ion flux by RRDE.13,14
As stated above, RRDE voltammetry can, in principle, be used
to monitor flux of cations, anions, or both. The present work
focuses only on cationic doping in polypyrrole/ poly(styrene-
sulfonate) (pPy+/ pSS-) composites. When pyrrole is polymerized
in the presence of pSS-, the pSS- polyanion is irreversibly
incorporated into the polymer structure.8-10 Since pSS- cannot
leave the polymer upon subsequent redox cycling, charge neutral-
ity is maintained predominantly by cation influx and efflux:
constant called the collection efficiency, and its value is determined
solely by the physical construction of the RRDE electrode (i.e.,
the disk area and the dimensions of the insulating gap and the
ring).15 When there is follow-up chemistry, as reaction C above,
some of the R is irreversibly consumed before reaching the ring;
therefore, the current ratio on the left side of eq 1 becomes less
than N. The disk-to-ring transit time, and thus the chemical
reaction time, is determined in a mathematically well-defined way
by the rotation rate of the electrode.15 Without going to unneces-
sary detail, the numerical value of the chemical rate constant, k,
can be obtained by considering how the ratio of the ring limiting
current to the disk limiting current, ir/ id, changes as a function
of rotation rate. Of relevance here is the fact that the voltammetry
of an electroactive polymer film on the disk of an RRDE is related
to the mass-transport-limited voltammetry of the dopant ion(s) at
the ring. This relationship is closely analogous to how the disk
and ring voltammetry of O and R, respectively, are related in this
hypothetical EC reaction in solution.
pPy+/ pSS- + e- + C+ / pPy/ pSS-C+
Prior to discussing how the currents are related, the unique
constraints imposed on the experiment by the polymer/ dopant
system need to be addressed. Two requirements are made of the
dopant redox chemistry by the fact that the voltammetry of the
dopant ion (or ions) is to be used to quantitate the polymer/
solution interfacial ion flux. First, the dopant ion must be
electrochemically inactive over the entire potential range applied
to the polymer, and second, it must give mass-transport-limited
voltammetry at the ring within the solvent window. Moreover, if
multiple types of dopant ions are to be considered, their voltam-
metry must also be sufficiently separated in potential to avoid
simultaneous reduction (vide infra). In general, either cationic or
anionic dopants (or both) could be probed as long as the above
criteria are satisfied.
In these experiments, the potential of a pPy+/ pSS--coated disk
is typically scanned between 0.00 and -1.1 V. The range is
sufficient to convert the polymer successively from fully reduced
(-1.1 V) to significantly oxidized (0.00 V). Voltammograms of
pPy+/ pSS- obtained in electrolyte solutions of typical composition
for a RRDE study are shown in Figure 1. The features of these
voltammograms are similar to those of voltammograms obtained
in more standard electrolyte solutions such as tetraalkylammo-
nium hexafluorophosphate. In the RRDE experiment, the ring
electrode potential is chosen to be well onto the mass-transport-
limited current plateau of the dopant cation, C+, reduction.16
Consequently, in the ideal case, a constant, time-independent
current, ir,0, should pass at the ring in the absence of any polymer
voltammetry at the disk.
In addition to the dopant ions, the solution must contain an
inert supporting electrolyte to eliminate migration effects. How-
ever, that supporting electrolyte cannot participate in polymer
(16) It is important to emphasize that the ring potential be chosen such that it
is well onto the limiting current plateau of the cation reduction so that small
variations in potential from iR drop do not effect the ring current.
(15) Bard, A. J.; Faulkner, L. R. Electrochemical Methods; Wiley: New York, 1980.
3678 Analytical Chemistry, Vol. 71, No. 17, September 1, 1999

fC+ ) -ir(t)/ (N(id(t)))
(4)
Here the doping experiment deviates from the classical EC
study. In the EC case, R is not present in bulk solution so there
is no current at the ring unless R is being produced at the disk.
By contrast, in the absence of any redox process at the disk, C+
EXPERIMENTAL SECTION
is present in bulk solution and produces a constant current (ir,0
)
Chemicals and Equipment. All chemicals except as noted
below were purchased from Aldrich and used as received. Pyrrole
was purified by distillation under nitrogen. Bis(triphenylphospho-
ranylidene)ammonium p-toluenesulfonate (PPNTos) was prepared
by combining aqueous solutions of PPNCl and sodium p-toluene-
sulfonate from which it precipitated. The PPNTos was filtered,
rinsed with distilled water, and dried under vacuum. The dried
product was subsequently recrystallized from methanol/ water and
from dichloromethane/ ethyl acetate. In each case, small white
crystals formed which were filtered from the solution and washed,
respectively, with either distilled water or ethyl acetate. The rinsed
solid was dried under vacuum overnight.
at the ring. When the polymer doping level changes, the
concentration of C+ reaching the ring changes, causing the ring
current to deviate from ir,0. It is the difference between ir,0 and
the total ring current, ir(t), which is quantitatively related to the
total disk current, id(t). Consider a positive potential scan of the
disk starting at a value where the pPy+/ pSS- film is fully reduced
(e.g., -1.0 V). As the polymer becomes progressively more
oxidized, electrons are being removed from the polymer and enter
the electrode. For charge neutrality to be maintained, either
cations must be ejected into solution or anions from solution must
enter the polymer. If it is assumed for the present that only cation
doping occurs in pPy+/ pSS-, every electron removed from the
polymer releases a cation into solution. From the point of view of
the ring electrode, the act of releasing C+ from the polymer is
indistinguishable from creating C+ electrochemically at the disk
(as was the origin of the R being formed at the disk in the EC
experiment). The difference in the two experiments is, again, that
in the EC studies the only source of R is electrochemistry at the
disk; in the doping experiment, C+ released from the polymer
adds to the constant flux of C+ from the bulk solution, thus
increasing ir(t) over ir,0. An identical argument can be made for
the reductive potential scan of the disk except that in that case
C+ enters the polymer, reducing its concentration at the ring and
causing ir(t) to drop below ir,0. If the corrected ring current, ir′(t),
is defined as
Tetraethylammonium poly(4-styrenesulfonate) (TEApSS) was
prepared via dialysis (Spectra/ Por membrane MWCO 6-8000
Spectrum Medical Industries, Inc., Los Angeles, CA) from tetra-
ethylammonium bromide (TEABr) and poly(sodium 4-styrene-
sulfonate) (NapSS) in distilled water. Aqueous solutions of TEABr
and NapSS were combined. The TEABr was in excess of the
NapSS (∼2.5 g) by 2-fold. The mixture was sealed in a dialysis
tube, which was then submerged in distilled water for 48 h. After
this time, the tube had swollen and become taut. External water
in the container was changed and to it several grams of TEABr
was added. The dialysis tube remained in the stirred solution for
48 h. This external solution was then replaced with pure distilled
water, which was replenished several times over a span of 48 h.
The tube was removed from the container and rinsed with distilled
water, and the contents were poured into a flask. The presence
of bromide in the solution was tested with aqueous silver nitrate.
The absence of a precipitate indicated the dialysis was complete.
The aqueous solution was taken to dryness on a rotary evaporator
leaving a tan film on the flask walls. This film was hard and difficult
to remove from the side of the flask. By dissolving the product in
a minimum of absolute ethanol and slowly re-evaporating the
ethanol on the rotary evaporator, a tan, bubbly solid coated the
flask walls, which could readily be removed. Once removed, the
material was further dried under vacuum. Finally, this procedure
is general for preparing acetonitrile-soluble R4N+pSS- salts when
R is hexyl or smaller (for R larger than hexyl, see ref 18).
1-Methyl-3-cyanopyridinium hexafluorophosphate (CMPPF6)
was prepared by dissolving 3-cyanopyridine in acetonitrile and
adding dropwise a stoichiometric amount of iodomethane. The
reaction mixture was refluxed for 4 h and then allowed to cool to
room temperature. The iodide salt (CMPI) formed as a yellow
solid upon cooling. It was filtered and recrystallized from aceto-
nitrile to yield yellow, needlelike crystals. The CMPI was then
dissolved in water and excess of ammonium hexafluorophosphate
added. The solution was cooled in a refrigerator until white
needlelike crystals formed. This mixture was filtered and the solid,
CMPPF6, was washed with cold water and then dried for 12 h
under vacuum. 1,3-Dimethylpyridinium hexafluorophosphate (DMP-
PF6) was synthesized in a similar manner starting with 3-picoline
in place of 3-cyanopyridine. The half-wave potentials for the
ir′(t) ) ir(t) - ir,0
(2)
a relation analogous to eq 1 exists which takes the form
-ir′(t)/ id(t) ) N
(3)
When the doping change is accomplished exclusively by C+, the
collection efficiency in eq 3, N, is the same as in eq 1. Just as
deviations of the current ratio from N signal a nonzero rate of
follow-up chemistry for an EC process, deviations in the current
ratio in eq 3 from N signal that C+ is not the sole ion participating
in changing the polymer doping level.17 The fraction of the doping
change that is due to C+, fC+, can be determined by dividing the
left side of eq 3 by the right and setting this equal to fC+, namely
(17) The magnitude of ir,0 in eq 2 is proportional to ω1/ 2, where ω is the electrode
rotation rate; the quantities id(t) and ir′(t) are, on the other hand, independent
of ω. Consequently, the signal-to-background ratio in these experiments
increases as
ω decreases. In all experiments reported here, ω was
maintained at 400 rpm, which is the lowest rotation rate available with the
PIR rotator. While they are independent of ω, the values of id(t) and ir′ do
depend on the polymer thickness and the potential scan rate, ν. At fast scan
rates, a collection of problems can arise from iR drop, from ion/ electron
transport rates within the polymer film, and from maintaining insufficient
flux of dopant ions from solution to the polymer interface (see text). A scan
rate of 50 mV/ s was determined to be an optimum compromise and was
used in all of the studies reported here unless otherwise noted.
(18) Elliott, C. M.; Kopelove, A. B.; Albery, W. J.; Chen, Z. J. Phys. Chem. 1 9 9 1 ,
95, 1743.
Analytical Chemistry, Vol. 71, No. 17, September 1, 1999 3679

reduction of CMP+ and DMP+ were determined from steady-state
voltammetry at the Pt ring of the RRDE. The values for CMP+
and DMP+ were calculated to be -1.13 and -1.70 V, respectively
vs a Ag/ Ag+ (0.10 M AgNO3, DMSO) reference electrode. The
measurements were made in acetonitrile solutions containing 2
mM electroactive cation and 50 mM PPNTos.
rigorously constant, significant errors would be introduced into
the data treatment. In actual experiments, ring current data are
collected over many cycles of the disk potential requiring between
5 and 10 min. Over such extended periods, changes to the ring
electrode surface often resulted, which were sufficient to cause
the value of ir,0 to slowly, and monotonically, decrease. This
decrease is typically small in absolute terms, but is large enough
to make eqs 3 and 4 unusable as previously defined. To deal with
this problem, the time dependence of the background current at
the ring was determined and the constant value of ir,0 in eq 2
replaced with the time-dependent function, ir,0(t) thus, redefining
ir′(t) in eqs 3 and 4:
The bis(cyclopentadienyl)cobalt(II) hexafluorophosphate
(CcPF6) was purchased from Aldrich and was used as received.
The half-wave potential was determined as above and calculated
to be -1.21 V vs a Ag/ Ag+ (0.10 M AgNO3, DMSO) reference
electrode.
An EG&G Princeton Applied Research model 173 potentiostat/
galvanostat with a model 175 digital coulometer was used for film
growth (vide infra). A Pine Instruments RDE4 bipotentiostat,
modified to provide a potential range of (5 V, was utilized for
the ring-disk electrochemistry. Electrode rotation was accom-
plished using a Pine Instruments PIR electrode rotator. Data
acquisition was accomplished using a program, written in ASYST,
provided by Professor Daniel Buttry of the Department of
Chemistry, University of Wyoming. The program was modified
locally. A BAS 100B Electrochemical Workstation was used for
the chronocoulometry experiments.
ir′(t) ) ir(t) - ir,0(t)
(2′)
To approximate the function ir,0(t), the points where id(t) crossed
the zero current axis were determined over the course of multiple
cycles of the disk potential. When id(t) ) 0.00 mA, no net change
in polymer doping can be occurring; thus, the concentration of
C+ at the ring should be exactly its bulk concentration. It follows
then that ir(t) at these points equal values of ir,0(t). The ring
current data for the points where id(t) ) 0.00 mA (as a function
of potential or time) were quite adequately fit to a simple quadratic
function. This function was then used to generate values of ir,0(t)
at the intervening potentials (times).
Cells and Electrodes. The electrochemical cells used in film
growth and in ring-disk experiments were each single-compart-
ment cells and were employed in three- and four-electrode
configurations. The working electrode was a platinum ring-
platinum disk electrode purchased from Pine Instruments. The
disk electrode had a 2.6-mm radius. The radius to the inside of
the ring was 4.0 mm, and the radius to the outside of the ring
was 4.3 mm. This gave a large ring-disk gap width of 1.4 mm.
The counter electrode was a platinum wire coil. A Ag/ Ag+ (0.10
M AgNO3, DMSO) reference electrode was employed for film
growth and for some of the ring-disk experiments; however,
more typically, a Ag pseudoreference electrode was utilized in
the RRDE experiments. All potentials are reported relative to the
Ag/ Ag+ (0.10 M AgNO3, DMSO) electrode.
RESULTS AND DISCUSSION
Cyclic Voltammetry. Films of pPy+/ pSS- give stable, repro-
ducible voltammetry with each of the three electroactive cation
dopants examined. The general shape of the voltammograms are
as expected; however, the total charge passed for a given film
during redox cycling is different for the different dopant cations.
Comparison of the voltammograms in Figure 1 shows that the
amount of charge passed follows the order DMP+ > CMP+
>
Cc+. The peak potentials of the voltammograms are also dopant
dependent; the less charge passed, the more positive the peak
potentials for a given film.
Film Growth. Composite polypyrrole/ poly(styrenesulfonate)
films were grown potentiostatically at +0.8 V. The acetonitrile
growth solution was 1.0 M in pyrrole and 0.10 M in TEApSS and
was stored at 0 °C under argon between uses. The extent of film
growth was determined from the coulombs passed during polym-
erization. All films were grown with the passage of 15 mC.
Following film polymerization, the working electrode was removed
from the growth solution and rinsed with copious amounts of
acetonitrile. The electrode and film were then placed into the
rotator and positioned in the RRDE cell. Shape and peak
magnitudes of the polymer’s cyclic voltammetry were used to
qualitatively verify film quality.
Single-Dopant Ion Transport Measurements. Figure 2
shows both the ir(t) and id(t) data for a pPy+/ pSS- film cycled
repeatedly between 0.0 and -1.1 V in a solution containing DMP+.
The ring potential is held at -1.9 V giving a mass-transport-limited
reduction of DMP+. The qualitative response for pPy+/ pSS- films
is similar to that shown in Figure 2 regardless of which electro-
active dopant is being studied. The current response of the
polymer is stable and centered about the zero current axis. The
value of ir,0 at the beginning of the experiment is represented in
the upper part of Figure 2 by a solid line. Over the course of this
experiment (∼4 min.), the background current at the ring decays
by roughly 10%. The dashed line labeled ir,0(t) is the time-
dependent background current obtained as described in the
Experimental Section. The corrected ring current response
calculated from eq 2′, ir′(t), is less than, and has the opposite sign
from, the disk current, id(t). From consideration of eq 4, if
Solutions. All solutions used for analysis of cation transport
in the composite films were prepared in the same manner. The
total concentration of electroactive cation(s) in solution was
maintained at 2.0 mM. The concentration of the PPNTos sup-
porting electrolyte was 50 mM. Acetonitrile was used as the
solvent in all electrochemical experiments.
+
fC ) 1, then ir′(t) should be smaller than id(t) by exactly the
Current Decay Correction. In the discussion leading up to
eqs 2-4, the results from the RRDE experiment were treated as
ideal. In practice, the peak-to-peak excursion of the (ir(t) - ir,0) is
typically only ∼10% of the value of ir,0. Therefore, if ir,0 were not
collection efficiency, N (i.e., 0.12 in the present case).
Typically, for each dopant ion studied, the value of fC+ obtained
from eq 4, within experimental error, is equal to unity. A visually
informative way of displaying the current data is in the form of a
3680 Analytical Chemistry, Vol. 71, No. 17, September 1, 1999

Figure 4. Cyclic voltammograms showing the change in the current
response of a pPy+/pSS- film as the solution is progressively changed
from pure CMP+ (inside solid curve) to pure DMP+ (outside solid
curve). Each solution was 50 mM in PPNTos and the total dopant
ion concentration ([CMP+] + [DMP+]) was kept at 2 mM. Electrode
rotation rate was 400 rpm.
Figure 2. Current vs time plot for a pPy+/pSS- film on the disk
electrode (id(t), dash-dot line) in an acetonitrile solution of 2 mM
DMPPF6 and 50 mM PPNTos cycled between 0.00 and -1.10 V and
the simultaneously recorded mass-transport-limited ring current for
the reduction of the DMP+ (ir(t), solid line). Electrode rotation rate
was 400 rpm.
Doping Competition Studies. Studies with each of the
individual electroactive dopants demonstrate that, at potentials
negative of +0.8 V, these pPy+/ pSS- composite films undergo
doping changes exclusively via cation transport. Having estab-
lished this fact, we were interested in how the various cations
compete in the doping process. Such competition studies would
be virtually impossible via techniques such as EQCM, but they
are straightforward employing the RRDE experiment.
We chose to consider two pairs of cations: DMP+ vs Cc+ and
DMP+ vs CMP+. The other possible combination, CMP+ vs Cc+,
was not considered because the reduction potentials of the cations
are too close to allow detection of each cation independently.
In the competition studies, the potential of the disk electrode
was cycled in a series of solutions in which the relative concentra-
tions of the two dopant cations was varied but their sum was held
constant. Figure 4 shows a progression of voltammograms in
solution containing 2.0 mM total electroactive cation ([CMP+] +
[DMP+]) as the relative fractional concentration of CMP+ is varied
from 1.0 to 0.0. Consistent with the voltammograms shown in
Figure 1, the amount of charge passed increases, the voltammo-
gram becomes more peak-shaped, and the peak shifts to more
negative potentials. Somewhat suprisingly, the shape of the disk
voltammogram proved to be mildly dependent upon the order in
which the polymer was exposed to the two dopant cations. If the
solution was changed progressively from one of pure DMP+ to
one of pure CMP+, the shape of the final voltammogram differed
moderately from that obtained from a film exposed only to CMP+.
Apparently, cycling the oxidation state of the polymer in a solution
containing DMP+ causes irreversible changes in the film. When
the reverse order of cation exposure is used, as depicted in Figure
4, the voltammograms at each extreme (i.e., DMP+ only and
CMP+ only) are the same as if the film had been exposed only to
the respective electrolyte. Similar results were observed in Cc+
and DMP+ competition studies. Consequently, for the majority
of the data reported here, the film was exposed to DMP+ last.
Irrespective of the film’s history and the order of electrolyte
exposure, the relative fraction of the doping level change due to
Figure 3. Plot of id (dashed line) and -Nir’(solid line) vs Edisk for a
pPy+/pSS- film (disk) cycled in an acetonitrile solution of 2 mM
DMPPF6 and 50 mM PPNTos. Ering was held at -1.9 V in the mass-
transport-limited reduction of DMP+. Electrode rotation rate was 400
rpm.
conventional cyclic voltammogram. When fC+ ) 1, a plot of -N(ir′)
+
should exactly overlay the disk voltammetry; when fC < 1, the
plot of -N(ir′) will be smaller than id by fC+. Figure 3 contains an
overlay plot of id and -N(ir′) for three multiple scans of the disk
potential. Once steady state is obtained, the two currents correlate,
within experimental error, over all but the most positive part of
the three scans. Only when the potential goes positive of ∼+0.8
V do the two plots deviate. Positive of +0.8 V on both the positive-
and negative-going scans id > -Nir′ indicating fC+ < 1. Recall that
the pPy+/ pSS- films were grown potentiostatically at +0.8 V.
Cycling the potential positive of +0.8 V therefore oxidizes the
polymer beyond the state in which it was grown. Consequently,
there is insufficient negative charge from the pSS- to compensate
the additional positive charge injected into the film; thus, additional
anions from solution must enter the film to maintain charge
balance.
Analytical Chemistry, Vol. 71, No. 17, September 1, 1999 3681

each dopant cation in a given mixture appears to depend only on
the ratio of the dopant cation concentrations in solution (vide
infra).
When the potential of the disk is cycled, the rate of change of
polymer doping (as reflected by the disk current) changes over
the course of the experiment. From eq 4, the fractional doping
change rate due to a given cation (or combination of cations) can
be calculated at every point in the voltammogram. When multiple
dopant cations are present in solution, there is no a priori reason
+
that fC for each cation remains constant at all doping rates
+
(currents). Nonetheless, under most circumstances, fC remains
constant within experimental error over the entire potential scan.
There are two situations where fC+ was observed to change during
a potential sweep: (1) when the potential is taken positive of where
the film was grown (as discussed previously) and (2) when the
flux of the lower concentration cation during the anodic sweep is
insufficient to provide the required degree of doping. To determine
whether sufficient cation flux exists, one need only compare the
ratio ir,0/ id to the electrode parameter â2/ 3. The value â, like the
collection efficiency, N, is specific for each RRDE and is deter-
mined by electrode geometry.13,19
Figure 5. Results of doping competition studies for CMP+ vs DMP+
(triangles) and Cc+ vs DMP+ (circles). The diagonal line represents
the case of no doping preference between the cations. The ordinate
For the reduction of the polymer at the disk not to be cation-
flux-limited, the following must be true:
+
is percent fC for the more easily reduced ion (i.e., either CMP+ or
Cc+). The abscissa is the relative fraction of the same ion in solution.
The horizontal line at the top of the plot indicates values having a
total fC+ ) 1(i.e., accounting for all doping occurring in film). Departure
of the data from the diagonal line indicates a preferential doping in
the film by DMP+.
ir,0/ id > â2/ 3
(5)
In experiments where the solution contained a single dopant cation
at 2 mM, ir,0/ id was always at least a factor of ∼2.5 greater than
When there is no preference for one cation over the other,
the fractional contribution to the doping, fC+, of each cation should
exactly correspond to its fractional solution concentration. All data
would then lie on the diagonal line in Figure 5. For both pairs of
cations represented in the Figure 5, all data points lie below the
diagonal line indicating a preference for DMP+.
â
2/ 3. In the mixed electrolyte experiments, the doping rate can
become flux-limited for the less concentrated component. In this
+
situation, fC for both cations was observed to vary during the
anodic scan. Fortunately, this situation only arises when one
component is present in a very small relative concentration.20
+
Since fC was determined to be effectively constant over the
Based on molecular models,21 the Cc+ is roughly cylindrical
in shape with a long-axis dimension (measured from a corner of
one cyclopentadiene through the cobalt to the far corner of the
other ring) of ∼5.6 Å. The length from center to center of the
two cyclopentadiene rings is 4.6 Å, and the diameter of each ring
is 4.2 Å. DMP+ is more nearly the shape of an oblong disk with
a long-axis dimension of ∼6.6 Å (measured from carbon to carbon
of the two methyl groups) and a short-axis dimension of ∼4.9 Å.
Given their physical dimensions, one might anticipate a preference
for DMP+ based simply on size. In fact, the data in Figure 5
demonstrate a strong preference for the smaller DMP+ cation over
Cc+. For example, when the relative solution concentration of Cc+
is 70%, it is responsible for less than 20% of the polymer doping
change!
While the Cc+/ DMP+ pair was studied because of difference
in size, the CMP+/ DMP+ pair was selected because of differences
in reduction potential (CMP+ is ∼700 mV easier to reduce than
DMP+). Assuming initially both ions to be the same size, we
presumed any doping preference in this pair would reflect
electronic differences in the ions (e.g., charge-transfer interactions
between the cation and the pyrrole sites). Consideration of the
entire polymer voltammogram, except as considered above, it has
a value characteristic of each set of concentrations in the binary
mixtures. Figure 5 is a plot of fC+ vs percent solution concentration
of the most easily reduced cation of each pair in (i.e., the solid
triangles for fCMP+ in the CMP+/ DMP+ pair and the solid circles
for the fCc+ in the Cc+/ DMP+ pair). The open triangles and circles
+
along the horizontal line at fC ) 1.0 represent the combined
doping by both cations obtained from eq 4 by potentiostating the
ring in the mass-transport-limited reduction of both dopant cations
of the pair. Within experimental error, the total doping change of
the polymer can be accounted for by the electroactive cations.
The different sizes of the data symbols delineate data obtained
from different, yet identically grown, pPy+/ pSS- films. Finally, for
none of the data presented in Figure 5 was the doping rate limited
by flux of either ion to the polymer surface.
(19) The quantity â2/ 3 is experimentally the ratio of the independently determined
ring and disk currents measured from the mass-transport-limited oxidation
or reduction of some arbitrary solution species. Since the limiting currents
at either electrode are proportional to the flux of species to the respective
surface, knowing the limiting flux at one electrode and â, the flux at the
other electrode can be calculated.
+
(20) While it would be interesting to quantitate the potential dependence of fC
under flux-limited conditions, the small absolute and relative concentrations
of the minor component make such studies virtually impossible because of
the poor overall signal-to-noise ratio. Nonetheless, the dependence is
qualitatively obvious.
(21) The Cc+ was modeled using CAChe Editor, Release 3.9. The CMP+ and
DMP+ measurements were done with Biograf by BioDesign, Inc., Version
2.2.
3682 Analytical Chemistry, Vol. 71, No. 17, September 1, 1999

to-noise ratio (S/ N) of the ring current also decreases. At scan
rates much below 50 mV/ s, the S/ N becomes so poor that
measurements of changes in doping level are unreliable.
While the chronocoulometry and slow-scan CV results dem-
onstrate that there is a major kinetic component to the preferential
doping at 50 mV/ s scan rates, these experiments do not rule out
the possibility of concomitant thermodynamic preferences. In an
effort to obtain some qualitative information on the equilibrium
doping state, a pPy+/ pSS- film was potentiostated at -1.10 V in
1:1 solutions of each doping ion pair (1 mM concentration of each
of the two ions) for 10 min. The potential of the disk was then
scanned in the positive direction. On this first anodic scan, the
+
+
values for fCMP and fCc were slightly greater than predicted by
the data in Figure 5. While there is no way to determine whether,
even after 10 min, the film has reached a true partition equilibrium,
these results do suggest that if there is a thermodynamic
preference for DMP+, it is smaller than the kinetic preference.
Given the absence of any other compelling explanations, it
seems reasonable to conclude that steric factors are responsible
for the kinetic preference of pPy+/ pSS- films to dope with DMP+
in both pairs of ions. The Cc+ unarguably occupies a larger volume
than does the DMP+ and it is slower to pass through the film.
For the CMP+/ DMP+ case, there is only a slight preference for
the later dopant, which is consistent with steric differences in the
two cations based on differential solvation.
Figure 6. Chronocoulometry for a pPy+/pSS- film (charge passed
by reduction of the polymer) on the disk of the RRDE with rotation at
400 rpm. The potential step is from 0.0 to -1.1 V. Each solution was
2 mM in electroactive cation and 50 mM in PPNTos.
data shown in Figure 5 (triangles) shows that there is indeed an
obvious preferential doping, but it is for the more electron rich
DMP+. Consideration of charge-transfer interaction predicts the
opposite. While the physical size of the bare cations does not
suggest a steric preference, allowing for differences in solvation
might. CMP+ incorporates a nitrile function (like the solvent) while
DMP+ is distinctly hydrophobic around its entire periphery.
Therefore, postulating differences in the “effective volume” of
these two cations is reasonable and may explain the preferential
partitioning of DMP+.
CONCLUSION
We have demonstrated the utility of rotating ring-disk volta-
mmetry in quantitating studies of ion transport in conducting
polymer films. In the past, pPy+/ pSS- has been assumed, without
real direct proof, to undergo only cation doping. With judicious
selection of the supporting electrolyte and the dopant cation(s),
we have been able to show by RRDE voltammetry that this
assumption is entirely correct under conditions similar to those
used in previous studies. Moreover, we have shown that RRDE
voltammetry can provide quantitative, in situ information on
competitive ion-doping processes. Finally, there is, in principle,
no reason why the technique is limited to cation dopants and
composite films. For example, we presently have preliminary data
on Cl- doping in conventional polypyrrole.
The major limitation of this approach lies in the fact that, for
the dopant ions to be directly monitored, they must be electro-
chemically active within the solvent window but inactive over the
potential range of the polymer. This is admittedly a nontrivial
constraint because it significantly reduces the number of dopant
ion candidates for study. That caveat notwithstanding, we have
found in preliminary work that, by using mixtures of electroactive
and nonelectroactive ions, it is often possible to gain considerable
in situ information on doping by electroinactive ions. This can be
accomplished by considering the difference between the total
doping requirement of the polymer and the part due to the
electroactive dopant.
The RRDE experiments, which generated the data in Figure
5, confirm preferential doping of pPy+/ pSS- by DMP+. What these
experiments do not address is whether the origin of the preference
is kinetic, thermodynamic, or some combination of both. Exami-
nation of the voltammograms taken in Cc+ and CMP+ electrolyte
included in Figure 1 show that the polymer continues to reduce
after the -1.1 V switching potential is reached. Even at a sweep
rate as low as 50 mV/ s, the polymer is not entirely at equilibrium
in these electrolytes, at least at the most reducing potentials. To
develop a more quantitative picture of the doping kinetics, potential
step chronocoulometry experiments were conducted for a pPy+/
pSS- film in solutions that were each 50 mM in PPNTos and 2
mM in one of the electroactive cations. In these experiments, the
potential of the polymer was stepped between 0.0 V, where the
film is partially oxidized, and -1.1 V, where, in the DMP+ case at
least, the polymer appears to be fully reduced. During each
experiment, the disk was rotated at 400 rpm to ensure that the
rate of reduction was not limited by cation flux to the polymer/
solution interface. The charge-time responses for the first 32 s
of a reductive potential step are shown in Figure 6 for each cation.
In none of the electrolytes has the polymer reached its equilibrium
doping level. In fact, it was found that the DMP+ charge response
attains a steady state at ∼53 s while the CMP+ takes ∼122 s and
Cc+ takes ∼137 s. The final value for the charge passed, however,
is the same for each of the cations. It is obvious from these data
that there is a significant difference in the rate of doping by the
three cations, this being fully consistent with the different shapes
of the cyclic voltammograms in Figure 1. Slowing the scan rate
to 5 mV/ s yields voltammograms that are more nearly the same
shape and size. Unfortunately, as scan rate decreases, the signal-
ACKNOWLEDGMENT
The authors acknowledge support of this work by the National
Science Foundation (Grant CHE-971408).
Received for review February 10, 1999. Accepted May 28,
1999.
AC990139W
Analytical Chemistry, Vol. 71, No. 17, September 1, 1999 3683