J. CHEM. SOC. FARADAY TRANS.

, 1991, 87(6), 853-860 853

Optical and Electrical A.C. Response of Polyaniline Films

Maher Kalaji and Laurence M. Peter*
Department of Chemistry, The University of Southampton , Southampton SO9 5NH, UK

The small-amplitude periodic optical transmittance response of redox polymer films is derived and related to the
corresponding electrical behaviour. The combination of optical and electrical measurements allows Faradaic
and non-Faradaic charging processes to be distinguished. The analysis of the modulated transmission data
obtained for polyaniline films on indium-doped tin oxide coated transparent electrodes in 1.0 mol dm-3 H,SO,
indicates that the capacitance measured in a.c. experiments is associated with redox transformations in the
PAN1 film. The a.c. capacitance was found to be lower than the pseudo-capacitance derived from cyclic voltam-
mograms and possible reasons for this discrepancy are discussed. The absorption coefficient of the polaron has
been derived from the relationship between the modulated transmittance and the film capacitance which also
shows that bipolaron states are formed at potentials more positive than 0.15 V vs. SCE.

The electrical and optical properties of polyaniline (PANI)
continue to attract attention. The material is particularly
interesting since its conductivity depends not only on its
degree of oxidation but also on protonation equilibria. l P 3
Oxidation of the insulating leucoemeraldine form of PANI in
acid solutions results in the formation of a conducting phase
whereas upon further oxidation the polymer becomes insulat-
ing again.
trical double layer at the polymer/solution interface. It was
suggested by Feldberg7 that the double-layer capacitance
associated with this process should depend linearly on the
fraction of the polymer converted to its conducting (oxidized)
form. The experimental impedance behaviour of electro-
chemically synthesized PANI films has been discussed by
several and it is clear that no Warburg behav-
iour is observed. It has been proposed by Albery et ~ 1 . ~ ~ 7 ~

Several aspects of the behaviour of PANI are not well
understood. It is not clear, for example, why the pair of volt-
ammetric peaks associated with the insulator-conductor
transition exhibit hysteresis which is independent of sweep
rate.4,5At the same time, the origin of the apparent ‘charging
current’ in the cyclic voltammogram for the oxidised
(conducting) form of PANI remains c o n t r ~ v e r s i a l .There
~ * ~ is
also disagreement about the interpretation of EPR data for
oxidised PANI in terms of localised Curie spins as against
Pauli susceptibility of metallic islands.8-’ Several of these
key problems were highlighted during the recent Faraday
Discussion on Charge Transfer in Polymeric Systems.’ Here
we address the problem of distinguishing between Faradaic
charge-transfer processes and charging of the electrical
double layer around polymer strands.
PANI films can be electrosynthesized on optically trans-
parent electrodes (OTEs) constructed from indium-doped tin
oxide (ITO) coated glass, and changes in the absorption spec-
trum associated with the subsequent oxidation and reduction
of the polymer have been widely In this paper,
we are concerned with the small-amplitude non-steady-state
optical response of PANI. This contains additional informa-
tion since the time-dependent absorbance changes can be
related to the Faradaic component of the total current. The
study has been restricted to the first insulator-conductor
transition, since further oxidation leads to rapid film degra-
dation.
Recently, we have shown that the transient response of
PANI-coated macroelectrodes is limited by the solution and
electrode resistances’ (the latter is particularly important for
ITO-coated OTEs). The kinetic processes involved in the
oxidation-reduction mechanism are extraordinarily rapid
and the transformation process can be resolved only if iR
errors are minimized by using ultramicroelectrodes with radii
as small as 10 pm. There are, however, two further processes
which might be expected to influence the transient or periodic
electrical response of PANI films. The first of these is the
transport of counterions which is required to maintain elec-
troneutrality during oxidation and reduction. If ion transport
is diffusion controlled, an additional Warburg impedance
should be observed in the ax. response. The second process
that could be important is non-Faradaic charging of the elec-
that the absence of a Warburg response in the complex plane
plots of the a.c. impedance of PANI can be explained by a
dual rail transmission line model in which the distributed
resistance elements in the polymer and aqueous phases are
equal, but this seems physically implausible (for a critical dis-
cussion of this model see ref. 13). There is also no compelling
evidence for the existence of a double-layer capacitance corre-
sponding to ‘metallic’ behaviour (i.e. arising from an excess
of surface density of mobile electronic charge carriers and an
ionic double layer in solution) of PANI in its oxidized state.
However, it is not possible to distinguish between Faradaic
pseudo-capacitance and double-layer capacitance on the
basis of electrical measurements alone.
In a recent c o m m u n i c a t i ~ nwe
, ~ ~reported that an exact
correspondence exists between the small-amplitude periodic
electrical and optical responses of PANI films on OTEs. In
particular, the characteristic relaxation frequencies of the two
responses are identical and appear to be determined by the
RC combination consisting of the Faradaic pseudo-
capacitance of the PANI film in series with the electrode and
solution resistances. The modulated optical response was
found to persist even when the film was converted to its con-
ducting state, showing clearly that the apparent ‘double-layer
charging current’ observed in the plateau region of the cyclic
voltammogram is associated with a Faradaic process. In spite
of its power, there have been few applications of modulated
transmittance. Kankare and ParenZ5 have used a similar
method to characterize poly(3-methylthiophene) and Ked-
dam et aLZ6 have described modulated reflectance measure-
ments on PANI films. In the present study, we consider the
relationship between frequency-resolved periodic optical
(transmission) and electrical (impedance) responses of PANI
films grown electrochemically on I T 0 electrodes. Impedance
and modulated reflectance measurements of PANI-coated
platinum electrodes will be discussed el~ewhere.~
Theory
Relationship between Electrical and Optical A.C. Response
The optical transmission of a film of thickness, d, is given by
= exp( - ad) = exp( - 4nkd/R)
Itrans/ZO
854
where I , is the incident light intensity (corrected for
reflection) and Itran,
is the transmitted light intensity. Here a
is the absorption coefficient and k is the imaginary com-
ponent of the complex refractive index at the wavelength, 1.
For the present purposes, it is convenient to eliminate the
film thickness by rewriting eqn. (1) in terms of the surface
concentrations r, (mol mP2) of absorbing species and the
corresponding molar absorption cross-sections, on (m2
mol - I ) .
Itransl'o ( Ern an)
= ~ X P-
For a potential perturbation between two values a and b, the
normalized change in transmission is related to the corre-
sponding changes, Ar, in surface concentration
For small optical perturbations, eqn. (3) can be written in the
linearized form
ATIT = -mr,o,
The changes in surface concentration, Arn,are related to the
Faradaic component of the current density (IF)and hence to
the corresponding component of the charge density AQF . If,
for example, we consider two species that are involved in the
redox reaction R = 0 + ne, then ArB= -ArR= AQF/nF.
For an a.c. perturbation, eqn. (4)then takes the form
AT/T = [(OR - ~o)/nF]
s Z F ( W ~ )dt
We can now relate the optical and electrical responses of the
film by noting that the total a.c. current density, I,,, (which is
the sum of the Faradaic and non-Faradaic components), is
related to the perturbing voltage, V , by
I,,, = V/Z = VY
where Z is the film impedance and Y is the film admittance.
The total a.c. charge density Qtotis related to the perturbing
voltage by
Q,,, =
i I,,, dt = CV
where C is the complex capacitance, defined by C = Y dt. It
follows from this discussion that AT/T should be linearly
related to the Faradaic complex capacitance and it is there-
fore possible to define the optical complex capacitance, Copt,
as
Copt = nF(AT/T)/[(oR -
Alternatively, we can consider the time derivative of the nor-
malized transmission
d(AT/T)/dt = [(OR - ~o)/nF]Zdwt)
in order to derive the optical admittance, Y,,,
x p t = {nF/[(oR - d(AT/T)/dt
Finally, it may be of interest to obtain the optical impedance,
Zopt,which can be defined as Zopt= Yoi: .
It is important to note that the electrical terms C , Y , and Z
become identical with their optical analogues Copt, and
Zopt,only in the case where non-Faradaic processes make no
contribution to the total current, i.e. It,, = I , . Provided that
interference effects are negligible, the foregoing analysis can
be extended to approximate the reflectance behaviour of thin
films on a reflecting substrate by taking the increased optical
pathlength into account. However, for thicker films it is
necessary to use the three-layer
J. CHEM. SOC. FARADAY TRANS., 1991, VOL. 87
The relationship between AT/T and the complex capac-
itance will no longer be linear if more than two species are
involved in the redox equilibria. There is evidence from EPR
measurements1'>12for the existence in the conducting state of
PANI of unpaired spin radical cation states (polarons) and
paired spin states (bipolaron or quinone imine). Equilibrium
between the polaron (P) and bipolaron (B) states is expected
to be established by the reaction
P+P=B+R (1)
(2)
where R is the corresponding reduced species. Eqn. (4)then
takes the form
AT/T = -(ArR OR + A r p o p + ArB oB) (11)
and although the relationship between AT/T and AQ will
still be linear for small perturbations, the optical complex
capacitance is less easy to define and AT/T will in general no
longer be a linear function of the complex capacitance. This
situation is considered in more detail in the discussion
section.
(4)
The Effect of the Distributed Resistance of the I T 0 Electrodes
When considering the impedance of PANI films on I T 0 sub-
strate electrodes, it is necessary to take into account the dis-
tributed sheet resistance of the transparent oxide coating. In
our experiments, the resistance measured between the ends of
the I T 0 electrodes (1 cm square) was typically 20-30 Q.
(5)
Some of this resistance is associated with the part of the I T 0
that is covered with insulating epoxy, and so it can be treated
simply as a series resistance. On the other hand, the exposed
part of the I T 0 electrode can be represented by a distributed
resistive network28 which is connected via the impedance of
the PANI film to the distributed resistance of the electrolyte
solution. The equivalent circuit of the PANI-coated I T 0
electrode is shown in Fig. 1.
(6)
The total impedance of the circuit shown in Fig. 1 is given
by the sum of the contact resistance R,,,, and the distributed
impedance Zdist,28 where
Zdist= (ZpRito)1/2cotanh[IRjto/(ZpRj3'/2] (12)
5
(7) Here I is the length of the IT0 electrode, 2, is the distributed
impedance made up of the capacitance of the PANI film and
the electrolyte resistance, and Rito is the distributed sheet
resistance of the I T 0 electrode.
(8)
(9)
Ir I
Rc o n t a c t
Rsneet
(10)
x,,
Cpolymer Rsolutlo" ;
I
Fig. 1. Equivalent circuit of a PANI-coated I T 0 electrode. The
impedance of the PANI film has been represented as a pseudo-
capacitance (valid for thin films)
J. CHEM. SOC. FARADAY TRANS., 1991, VOL. 87 855

real( impedance/l 0 n)
real (admittance/l 0 - 2n-’)
real(capacitance/l 0-4 F )

0
) ? 0

log [frequency/Hz]
Fig. 2 (a) Complex plane plots of the impedance (+), admittance (a)and capacitance (0) corresponding to the equivalent circuit shown in
Fig. 1 , where RcOntact
= 10 R, Rsheet = 20 R cm-’, Cpolymer = 0.5 mF cm-’ and Rsolution= 0.2 R cm2; ( b ) Bode plot of the impedance of the
equivalent circuit shown in Fig. 1 calculated using the same values as those given in Fig. 2(a); ( 0 )magnitude (0) phase angle

The effects of the distributed and series resistance on the
impedance, admittance and complex capacitance are illus-
trated by Fig. 2(a) and (b). It can be seen that at high fre-
quencies the impedance plot deviates from the vertical line
corresponding to the simple series RC circuit, and the corre-
sponding response in the admittance is manifest as a second
high-frequency semicircle. By contrast, the complex capac-
itance plot is much less sensitive to the effects of the distrib-
uted sheet resistance, since it compresses the high-frequency
response and emphasizes the low-frequency semicircle. This is
demonstrated by Fig. 2(a) which shows that the complex
capacitance plot forms an almost perfect semicircle. The Bode
plot [Fig. 2(b)] gives the most balanced representation of the
impedance. The low-frequency behaviour tends towards
purely capacitive behaviour as expected, but the transition at
high frequencies to near resistive behaviour proceeds via an
intermediate region where the distributed impedance is
important. This is evident in the change of slope in the
log[Z]/logf plot from - 1 towards -0.5 and the corre-
sponding flattening off in the phase-angle plot.
Experimental
PANI films were grown p~tentiodynamically~~ on I T 0 elec-
trodes by cycling the potential between -0.2 and 0.75 V us.
SCE in a solution containing 0.1 mol d m - 3 of freshly dis-
tilled aniline (Aldrich ‘Gold Star’ grade) in 1.0 mol dm-3
H,SO, (BDH ‘Aristar’ grade). The potential was extended to
0.9 V for the first cycle to promote nucleation of the film.
OTEs were constructed from 1 cm2 pieces of indium-doped
tin oxide (ITO) coated glass (Hoya). Contacts were made
with conducting silver paint and the contact area was insu-
lated by a layer of epoxy resin. The PANI film thicknesses
were calculated from the relationship between thickness and
switching charge in the cyclic voltammogram established
independently by e l l i p ~ o m e t r y . ~ ~
Electrochemical measurements were carried out in a three-
electrode cell controlled by a potentiostat constructed using
high-bandwidth operational amplifiers. Frequency-resolved
transmission measurements were performed at fixed d.c.
potentials and wavelengths in a single-beam arrangement in
which monochromatic light, provided by a tungsten lamp
and a grating monochromator, was detected by a suitably
positioned silicon photodiode. The output of the photodiode
was measured with a wideband current amplifier with a d.c.
offset facility and was fed to a Solartron 1250 frequency-
response analyser, which in turn provided the small-
amplitude a.c. modulation to the potentiostat. Impedance
measurements were performed at constant d.c. potential
under identical conditions using the Solartron frequency-
response analyser. Data acquisition, treatment and presen-
tation were carried out under microcomputer control.
856 J. CHEM. SOC. FARADAY TRANS., 1991, VOL. 87

obtained on the first cycle after holding the potential at the

negative limit sufficiently long to ensure complete reduction
of the PANI film (see ref. 17 for a discussion of the effect of
multicycling on the anodic peak). The voltammogram
obtained with the I T 0 electrode exhibits the familiar shape
observed for PANI-coated platinum, although the back-
ground current at the negative limit is much smaller owing to
the much higher overpotential for hydrogen evolution on
ITO. The hysteresis observed in the voltammogram, which is
typical of many conducting polymer systems, does not appear
to arise from slow electron-transfer kinetics since the shape of
the voltammogram remains unchanged at sweep rates of up
to 1000 V s-'.I7 It seems more likely that a structural trans-
formation of the type discussed by Heinze et aL4 and by
Feldberg and Rubinstein' is responsible. However, more evi-
dence is needed to support the proposed mechanisms or any
other possibilities, such as film swelling during the influx of
counter ion^.^^
It is significant that the optical absorbance in Fig. 3(a)
increases over the entire range of the forward potential scan,
indicating that oxidation still takes place in the plateau
region following the anodic peak. Fig. 3(b) shows that the
hysteresis, exhibited in the absorbance plot of the forward
and reverse potential scans, disappears in a plot of absorb-
ance us. charge indicating that the composition of the film is
determined by the charge rather than the potential (we note,
however, that some hysteresis was observed in the
1 1 1 1 1 1 1 1 1

-0.2 0 0.2 0.4 0.6 absorbance-charge plots measured at 450 nm). Fig. 3(b) also
shows that the absorbance does not increase linearly with
potentialp vs. SCE
charge, which indicates that more than two species are
involved in the redox reactions. Zotti and Schiavon14 have
analysed the non-linear charge dependence of the absorbance
at different wavelengths in terms of the polaron-bipolaron
equilibrium, and have obtained the polaron density as a func-
tion of oxidation charge and the corresponding association
constan t.
40 Impedance measurements were carried out on a range of
PANI films on I T 0 electrodes. Fig. 4 contrasts the complex-
h
plane impedance, admittance and capacitance plots for a 26
$
v
nm thick PANI film on an I T 0 electrode at 500 mV us. SCE
Q,

m
C
e
v)
n
20

0 1 2
charge/mC cm-*
3

'9
0
0
0

iw
Fig. 3 (a) Cyclic voltammagram for an 80 nm PANI film on ITO- 0100 Hz

coated glass (1 cm') at 20 mV s - * . The figure also shows the simul- * O

taneously recorded change in the optical absorption of the film at

620 nm. The absorbance increases over the whole range of potential.
(b) Plot of absorbance us. charge constructed from the data in (a).
Note the absence of hysteresis and the deviation from linearity; (a) reaI(,ZJn)

.
5 - p _ _ _____*
~
anodic sweep; (())cathodic sweep
---o real(Y/Q-') ~ -012
0 4---- - rea1(C/1o - F)
~ ___- 10
Results and Discussion Fig. 4 Complex-plane impedance (+), admittance (a)and capac-
itance (0)plots of a 26 nm PANI film on an ITO-coated glass elec-
Fig. 3(a) illustrates the relationship between the cyclic volt- trode (1 cm2) measured at 500 rnV us. SCE. Note the similarity
ammetric behaviour of a PANI film and the corresponding between the experimental electrode response and the calculated plots
changes in optical transmittance at 620 nm. The data were shown in Fig. 2(a)
J. CHEM. SOC. FARADAY TRANS., 1991, VOL. 87 857

5
admittance plot which shows the predicted high-frequency
h
LL
semicircle. By contrast, the complex capacitance appears to
d form an almost perfect semicircle since the high-frequency
response is compressed. The presentation of the data in the
form of the complex capacitance is particulary convenient
since it allows the film capacitance to be obtained by
extrapolation of the plot to zero frequency.
The effect of potential on the a.c. response of the PANI-
coated electrode is illustrated by the complex capacitance
plots in Fig. 5(a) and (b) which were recorded for successively
0 0.5 1 less positive values of the electrode potential in order to
avoid the 'memory effects' associated with the reduced form
real (C/m F)
of PANI. The extrapolation of the complex capacitance plots
to zero frequency showed that the capacitance goes through a
rather small maximum at around 150 mV on the reverse
- I I scan. Measurements for film thicknesses in the range 20-300
nm showed that the film capacitance increases and a,,,
decreases approximately linearly with film thickness.
The transition in the electrical behaviour of the film from
capacitive at low frequencies to resistive at high frequencies is
best illustrated in the Bode plot shown in Fig. 6. The resem-
blance between Fig. 6 and Fig. 2(b)further confirms the valid-
ity of the equivalent circuit of the PANI-coated I T 0
0 0.5 electrode shown in Fig. 1. Note that the PANI films con-

.,
real (C/m F)
Fig. 5 Complex plane plots of the complex capacitance of a 26 nm
PANI film on an ITO-coated glass electrode (1 cm2) at different
potentials us. SCE: (a) 0, 500 mV; 400 mV; 0 , 200 mV; 0, 150
mV; (b) +, 100 mV; 0 , 50 mV; 0, 0 mV. The data were obtained
for d.c. potentials decreasing in steps from 0.5 V
[i.e. at a potential corresponding to the anodic limit of the
cyclic voltammogram in Fig. 3(a)].Here, as in the simulations
in Fig. 2, the impedance plot deviates at high frequencies
from the vertical line corresponding to the series RC circuit.
The remarkably good fit to the theory is manifest in the
sidered here are sufficiently thin that no additional high-
frequency semicircle appears in the complex plane impedance
plot.20,21
Fig. 7 compares capacitance values derived from the
impedance measurements with the apparent capacitance
derived from the cyclic voltammogram by assuming that the
measured current is given by the product of the sweep rate
and the pseudo-capacitance. It is immediately clear that the
capacitance measured using a small-amplitude perturbation
at constant d.c. potential is considerably lower than the
apparent capacitance derived from the cyclic voltammogram.
Furthermore, the positions of the capacitance maxima do not
coincide. This kind of behaviour has been noted for other

- 90
_ .
-1.0 0 1.0 2.0 3.0 4.0
log (frequency/Hz)
Fig. 6 Bode plot of the impedance of the same PANI film (26 nm) as in Fig. 5 at -200 mV ( 0 , magnitude; 0,phase) and 500 mV (m,
magnitude; 0,phase)
858 J. CHEM. SOC. FARADAY TRANS., 1991, VOL. 87

I 4 rnF c m - 2

-0.5 -1 .o -1.5
real(AT/T)/l 0-4

real(AT/T)/l 0 -
Fig. 8 Complex-plane plots of the modulated normalized transmit-
tance AT/T at 630 nm for the same PANI film as in Fig. 5 for
500 mV;W, 400 mV; 0 , 200
decreasing potentials from 0.5 V; (a) 0,
- 0.2 0 0.2 0.4
potentialp vs. SCE
mV; ( b ) 0,150 mV, +, 100 mV; .,50 mV

Fig. 7 Comparison of the capacitance of a 66 nm PANI film deter-

mined from the impedance (0) with the pseudo-capacitance derived The dependence of AT/T on oxidation charge is compared
from the cyclic voltammogram in Fig. 9 with the corresponding variation of the complex
capacitance. For simplicity, the comparison has been
restricted to the reverse sweep from 0.5 V. The first point to
note is that no maximum is observed in the AT/T plot, in
polymer system^.^'.^^ In the case of PANI at least, the differ-
spite of the fact that the complex capacitance passes through
ences cannot be attributed to slow kinetics since cyclic volt-
a maximum at 0.15 V. In fact, the largest AT/T response is
ammetry at microelectrodes has shown that the shape of the
obtained at the anodic limit of 0.5 V, which is in the region
cyclic voltammogram remains unchanged even at sweep rates
where the plateau current in the cyclic voltammogram might
of up to loo0 V s-'. It must be concluded that small-
be attributed to double-layer charging of the fully oxidized
amplitude periodic perturbation results in the reversible
polymer. The optical response provides convincing proof that
transfer of a smaller quantity of charge than is observed
during a large potential excursion. The reasons for this are
not clear at present, but it has been suggested that rapid elec-
tron transfer is accompanied by significant conformational 0.C
changes in the polymer which occur on a longer time~cale.~
We note that the capacitance values obtained from the
impedance measurements for increasing and decreasing d.c.
potentials still show significant hysteresis, in spite of the fact
that each measurement was made after waiting for three
minutes. The first-order rate constants for the structural 0.C
transformation obtained by Albery et from the analysis
NI
of spin-density transients are sufficiently high that we should
expect the most stable conformation to be achieved on the L L 5
timescale of our measurements. If this were indeed the case, E
the capacitance data should not exhibit hysteresis. At present 0.4
we have no evidence that the hysteresis would disappear on
longer timescales.
The correlation between the complex electrical and optical
capacitances can be seen by comparing Fig. 5 with Fig. 8 ; the
latter shows the normalized modulated transmittance at 630 0.;
nm of the 26 nm PANI film in Fig. 5. The correspondence
between the electrical and optical responses is clearly demon-
strated by the fact that their characteristic relaxation fre- I 1 I I I I I I
quencies are identical. This confirms our previous conclusion 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
that the 'capacitive current' is associated with charge trans- QlQ,,,
fer.24 As expected, the value of AT/T at zero frequency Fig. 9 Plots of complex capacitance (0) and modulated transmit-
increases whereas w,,, decreases linearly with film thickness, tance).( as a function of oxidation charge normalized to charge at
following exactlv the behaviour of comdex caDacitance. 0.5 V. All data were taken for the reverse sweeD
J. CHEM. SOC. FARADAY TRANS., 1991, VOL. 87
the charging in this potential region is still due to a redox
process in the film.
Eqn. (8) predicts that a plot of AT/T against the complex
capacitance should be a linear function provided that the
redox equilibrium is dominated by only two species and non-
Faradaic charging is negligible. The EPR data” suggest that
the initial oxidation leads to a radical cation or polaron state,
whereas the bipolaron is stable only at more positive poten-
tials. Fig. 10 shows that the initial part of the plot of AT/T
against film capacitance is indeed linear, but as the degree of
oxidation exceeds the value where bipolarons are presumed
to form, the plot folds back as AT/T continues to rise but the
capacitance falls again. Analysis of this part of the curve is
difficult because a realistic model describing the polaron-
bipolaron equilibrium must include interaction parameters to
explain the broadening of the voltammogram. Here we
restrict attention to the first part of the plot which we analyse
assuming that only polaron species are present in addition to
the reduced form. According to eqn. (8), the plot of AT/T us.
C should have a slope which is equal to (oR- ao)V/nF (in the
case of the polaron, n = 1). Since the reduced form of the
polymer is transparent at 630 nm,29it is possible to obtain an
approximate value of the optical absorption cross-section of
the polaron; the value estimated from the linear portion of
the plot in Fig. 10 is 1.4 x lo7 cm2 mol- ’.
The absorption coefficient of the PANI films has been
determined as a function of wavelength and potential by
ellip~ometry~’ at 633 nm; the value of k at 0.15 V on the
reverse sweep is 0.12 corresponding to tl = 2.3 x lo4 cm-’.
The molar volume V, of the polaron species at the upper
limit of the linear plot can now be obtained since V, = op/tl.
We find in this way that V, = 600 cm3 mol-’. If we assume
that the polaron repeat unit consists of four benzene units
and take the density of PAN132 to be 1.2 g cm-3, the mole
fraction of polarons at the point where bipolarons begin to
form is ca. 0.6, which is in reasonable agreement with the
estimate obtained by Zotti and S ~ h i a v o n . ’ ~
Conclusions
The simultaneous analysis of the periodic electrical and
optical responses of PANI films clearly offers a powerful new
12
10 d
(0.9 2
LD
zl a 17 M. Kalaji, L. M. Peter, L. M. Abrantes and J. C. Mesquita, J .
2 18 S . H. Glarum and J. H. Marshall, J. Electrochem. Soc., 1987, 134,
2 142.
9 19 I. Rubinstein, E. Sabatani and J. Rishpon, J . Electrochem. Soc.,
4
2 22 W. J. Albery, Z . Chen, B. R. Horrocks, A. R. Mount, P. J.
-
0 ~~
0.6 0.7 0.8
C/mF c m - 2
Fig. 10 Plot of A T I T us. capacitance [see eqn. (8)1 constructed from
data in Fig. 9. Note that the plot is linear at low-oxidation charges
where the polaron dominates, but folds round as the bipolaron is
formed at higher charges. The linear portion of the plot can be used
to derive the absorption cross-section of the polaron (see text)
859
approach to the characterization of this complex material,
and the method should also find application to other modi-
fied electrode systems. The main conclusion of the study is
that the a.c. response of PANI is dominated by Faradaic pro-
cesses, but the problem of the discrepancy between the small-
amplitude (a.c.) and large-amplitude (cyclic voltammetry)
capacitance values remains unsolved. It seems unlikely that
analysis of a.c. impedance data alone will offer any new
insights into the mechanisms of conducting-polymer oxida-
tion and reduction, and approaches based on complex equiv-
alent c i r c ~ i t s ~ ~or
, ~ ’ on invalid models involving
non-Faradaic processes” have little to recommend them. A
complete understanding of these complex systems will require
simultaneous application of even more than the two tech-
niques discussed in this paper. In this context, the intro-
duction of the a.c. modulated mirobalance method34 may
prove to be an important step forward since this method can,
in principle, be used to deconvolute anion and cation fluxes
during the cycling of conducting-polymer films.
This work was supported by the SERC and the Ministry of
Defence. The authors wish to thank the Hoya Corporation
(UK) for the donation of I T 0 materials.
References
1 A. G . MacDiarmid and A. J. Epstein, Faraday Discuss. Chem.
SOC.,1989,88, 3 17.
2 P. M. McManus, R. J. Cushman and S. C. Yang, J . Phys. Chem.,
1987,91, 744.
3 H. Reiss, Synth. Met., 1989,30, 257.
4 J. Heinze, M. Storzbach and J. Mortensen, Ber. Bunsenges. Phys.
Chem., 1987,91,960.
5 S. W. Feldberg and I. Rubinstein, J . Electroanal. Chem., 1988,
240, 1.
6 J. Heinze, R. Bilger and K. Meerholz, Ber. Bunsenges. Phys.
Chem., 1988,92, 1266.
7 S. W. Feldberg, J . Am. Chem. Soc., 1984,106,4671.
8 M. E. Josefowicz, R. Laversanne, H. H. Javadi, A. J. Epstein,
J. P. Pouget, X. Tang and A. G. MacDiarmid, Phys. Rev. B,
1989,39, 12958.
9 B. Villeret and M. Nechtschein, Phys. Reo. Lett., 1989,63, 1285.
10 S. H. Glarum and J. H. Marshall, J. Electrochem. Soc., 1987, 134,
2 160.
11 E. M. Genies and M. Lapkowski, J . Electroanal. Chem., 1987,
236,199.
12 M. Lapkowski and E. M. Genies, J. Electroanal. Chem., 1990,
279, 157.
13 Faraday Discuss. Chem. Soc., 1989,88.
14 G. Zotti and G. Schiavon, Synth. M e t . , 1989,30, 151.
15 D. E. Stilwell and S-M. Park, J. Electrochem. Soc., 1989, 136,
427.
16 Y. Li, B. Yan, J. Yang, Y. Cao and R. Qian, Synth. Met., 1988,
25, 79.
Electroanal. Chem., 1989,274, 289.
1987,134,3078.
20 M. Lyons, H. Fay, T. McCabe, J. Corish, J. Vos and A. Kelly, J .
Chem. Soc., Faraday Trans., 1990,86,2905.
21 T. Boschi, G. Montesperelli, P. Nunziante and G. Pistoia, Solid
State lonics, 1989,31, 281.
Wilson, D. Bloor, A. T. Monkman and M. C. Elliott, Faraday
Discuss. Chem. Soc., 1989,88, 247.
23 W. J. Albery, C. Elliot and A. Mount, J . Electroanal. Chem.,
1990,288, 15.
24 R. Hutton, M. Kalaji and L. M. Peter, J . Electroanal. Chem.,
1989,270,429.
25 J. Kankare and A. Paren, Extended Abstracts First Int. Symp. on
Electrochemical lmpedunce Spectroscopy, 1989, C6.1.
26 M. Keddam, M. Musiani and H. Takenouti, Extended Abstracts
40th ISE Meeting, 1989, D2 21-13-14-P.
860
27 M. Kalaji and L. M. Peter, in preparation.
28 R. De Levie, Adv. Electrochem. Electrochem. Eng., 1967,6,329.
29 R. Greef, M. Kalaji and L. M. Peter, Faraday Discuss. Chem.
Soc., 1989,843,277.
30 B. Lindholm, in ref. 13, p. 297.
31 J. Tanguy, M. Salma, M. Hoclet and J. Baudouin, Synth. Met.,
1989,28, C145.
J. CHEM. SOC. FARADAY TRANS., 1991, VOL. 87
32 D. Orata and D. Buttry, J. Am. Chem. Soc., 1987,109,3574.
33 H. Mao, J. Ochmanska, C. Paulse and P. Pickup, Faraday
Discuss. Chem. Soc., 1989,843, 156.
34 C. Gabrielli, M. Keddam and R. Torresi, Extended Abstracts
41st ISE Meeting, 1990, Fr-169.
Paper 0/04142D;Received 1lth September, 1990