Table of Contents 1

INVESTIGATIONS OF THE REDOX KINETICS OF

CONDUCTING POLYMERS

Dissertation
zur
Erlangung des Doktorgrades
der Mathematisch-Naturwissenschaftlichen Fakultät
der Christian-Albrechts-Universität
zu Kiel

vorgelegt von

Evgenij Barsoukov

Kiel 1996
Table of Contents 2

Table of Contents

1 Introduction ___________________________________________ 4

2 Experimental__________________________________________ 10

2.1 Electronic equipment 10

2.1.1 Conventional electrochemical techniques 11

2.1.2 Impedance measurements 11
2.1.3 Impedance measurements on the twin electrode 12
2.2 Cells and Electrodes 14

2.2.1 Three-electrodes-cell 14
2.2.2 The twin electrode 15
2.3 Chemicals and Electrolytes 16

3 Basic experimental data analysis__________________________ 17

3.1 The analysis of the impedance spectra of conducting polymer layers 17

3.1.1 Monolayer of an electroactive species 17

3.1.2 Homogenous nonporous electroactive layers 19
3.1.3 Porous electroactive layers 21
3.2 Field considerations for measurements on the twin electrode 25

3.3 Complex nonlinear least squares fitting 28

4 Results _______________________________________________ 30

4.1 Impedance measurements during oxidation and reduction of

conducting polymers 30

4.1.1 Simulation procedure 30

4.1.2 Measurements and simulation results 31
4.2 Impedance measurements on twin working electrodes bridged with
polybithiophene layers 37
Table of Contents 3

4.2.1 Verification of the measurement set-up 37

4.2.2 Impedance spectra 39
4.2.3 Estimations of conductivity and thickness of the polymer layer 46
4.3 The thickness dependence of the impedance spectra of polybithiophene
layers measured under galvanostatic condition 48

4.3.1 Measurement results 48

4.3.2 Data processing 52
4.4 Impedance investigation of conducting polymer layers considering their
inhomogeneity. 57

4.4.1 Model of an inhomogeneous porous electroactive layer 57

4.4.2 Experimental results and validation of the model 63
5 Discussion ____________________________________________ 71

5.1 The contribution of faradaic currents in the oxidation/reduction of

polybithiophene and polypyrrole during cyclic voltammetry experements. 71

5.1.1 Polybithiophene. 72
5.1.2 Polypyrrole. 74
5.2 Contribution of electronic and ionic resistance to the resistive
hindrance of the recharge processes of polybithiophene 74

5.3 Evaluation of kinetic relevant parameters of conducting polymers by

analysis of layers with different thickness. 76

5.4 Influence of inhomogeneity on the electrochemical behavior of

conducting polymer layers. 79

6 Conclusion ____________________________________________ 83

7 Acknowledgments ______________________________________ 85

8 References ____________________________________________ 89
1 Introduction 4

1 Introduction
A new exiting branch of electrochemistry was born in 1977, as Heeger & McDiarmid
[1] discovered the ability of polyaromatics with conjugated double bonds to undergo
chemical and electrochemical redox transitions to yield polymers with high electronic
conductivity. The simple electrochemical technique of producing conducting polymer
layers was first reported by Diaz et al. [2] in 1979. After that, conducting polymers have
become a beloved object of investigation by electrochemists. Many proposals for
possible applications of these materials have been put forward in the last few years
because of their unique electronic, electrochemical and optical properties. Some of
these include their use in storage batteries [3], electrochromic displays [4] and sensors
[5]. However, no extensive commercialization of any of these systems has been
observed up to the present. This can be partially accounted for their very complex
chemical and physical behaviors. The price paid for the simplicity of their
electrochemical preparation is the high complexity of the structure and properties of the
obtained polymer layers. For such applications as supercapacitors or electrochromic
devices, thin films of polymer on the electrode are required, however preparation of
such films by use of chemically synthesized polymers would be connected with
unreasonable expenses. Hence, the understanding of the electrochemical kinetics in
view of polymer layer properties is decisive for the future use of conducting polymers.

A large number of different processes influence the kinetics of the recharge of

electrochemically synthesized conducting polymers, however a couple of processes is
common to most of them. These are: charge transfer through the polymer and the
electrolyte, included in the polymer layer; transport of counterions into the polymer
matrix; formation of double layer on the polymer interface; establishment of chemical
and membrane equilibrium and the associated structural changes. Practically all kinds of
existing physical and chemical methods have been used to investigate the properties of
conducting polymers. Investigations have been performed under potentiostatic control
by cyclic voltammetry along with other measurement techniques, such as
electrochemical impedance spectroscopy (EIS) [6- 9], conductivity measurements
1 Introduction 5

[10- 12], ESR [12], quartz microbalance [13], uv-vis spectroscopy [14] etc. Results of
current pulse experiments under galvanostatic control have also been reported [7,15,16].
However, due to complexity of the electrochemical behavior of conducting polymers,
much questions remain still open.

• Due to the manifold of processes, determining the behavior of conducting polymers,

the cyclic voltammetry is not able to give their physical and chemical parameters
directly. One of the commonly admitted features of the cyclic voltammograms (CVs) of
conducting polymers is its strong capacitive character in potential range, where polymer
is in an oxidized form [17-20]. Several attempts have been made to describe such a
behavior of coated with conducting polymer electrodes [9,21-26]. Feldberg [21]
assumed that the capacitance is proportional to the amount of oxidized polymer but is
independent of the electrode potential once the polymer film has been converted into its
oxidized form. White and coworkers [22,23] extended the approach of Feldberg, taking
into account the porosity of the polymer films. Kaplin and Qutubuddin [24] simulated
charge transfer during switching of conducting polymers for both the reversible and the
irreversible cases, but neglected the capacitive effects.

The combined analysis of the cyclic voltammograms and data obtained by

measurements from a different source, can allow to exclude one of the processes,
effecting the current response during the potential sweep and therefore to simplify the
investigation significantly. Tanguy and coworkers measured the impedance of
chemically synthesized polypyrrole [25] and of electrodes covered with 3-
methylthiophene [9] over a large frequency range for different values of the electrode
potential. They measured electrochemical impedance spectra in steady state and
calculated the capacitive part of the current as Ic = C(U)(dU/dt) where: C(U) is an
expression for the voltage dependence of the capacitance obtained by fitting of the low
frequency part of experimental impedance data and dV/dt is the sweep rate. Use of
steady state impedance data for predicting CVs requires consideration of
pseudoequlibrium conditions during the potential cycle. Cyclic voltammograms with
different scan rates performed on polybithiophene films do not prove this assumption.
Impedance spectra measured at a particular potential are dependent on the rate of
potential scan. Therefore, the attempt has been made, to predict CVs using the
1 Introduction 6

resistance and capacity values evaluated from impedance spectra, which were measured
during potential sweep with a respective scan rate (chapter 4.1). Current response was
calculated by a numerical solution of a differential equation considering both C and R
values dependent on the applied electrode potential.

• The model of serial connected capacitance and resistance is only usable to describe
the slow recharge of conducting polymers. In the intermediate frequency range,
corresponding to fast processes in the time domain, the porous nature of
electrochemically deposited polymer layer has to be taken into account. Due to porosity,
the polymer-electrolyte interface resides not only on the geometrically edge of the layer
but also inside of it. Different inner interfaces become nonequipotential during the
current flow through the layer. The analysis of alternating current (ac) response allows
us to evaluate the resistance of polymer layer, however, the path of current flow by
conventionally impedance measurements travels through the pores, filled with
electrolyte, as well as through the compact polymer aggregates. Hence, the electronic
conductivity of polymer becomes indistinguishable from the ionic conductivity of
electrolyte. An analysis of experimental impedance spectra of polymers in conducting
state shows, that the contribution of one of the two possible resistances can be neglected
and does not have considerable effect [27]. The question remains open though, as to
which one is it. Measurements, where the charge transfer does not cross the
polymer-electrolyte interface can allow independent evaluation of one of the
resistances. Several experimental designs for performing such experiments have been
reported [5,11, 28-31]. Paul et al. [5] have investigated the conductivity properties of
polyaniline, electrochemically deposited onto an Au microelectrode array. The
resistance was measured between two microelectrodes, spaced 1.7 µm apart. Schiavon
et al. [28], Holze and Lippe [11], and Kankare and Kupila [29] have used a simpler,
twin electrode configuration with two metal stripes separated by a gap a few
micrometers wide, exposed to the electrolyte. The gap was bridged with a polymer layer
through electrochemical deposition. The resistance measurement has been performed by
applying a small dc voltage [11,28] or a 130 Hz ac voltage [29] to the electrodes across
the gap. The 4-probe technique has also been successfully applied to in situ
conductivity measurements on polymer covered electrodes by Olmedo et al. [31].
1 Introduction 7

Measurement with a polymer layer, which is only contacted at the polymer or

electrolyte side, could give resistance data along with other information, relevant to
recharge kinetic, if performed within a wide frequency range. A new technique for
multifrequency EIS measurements in the surface layer, nearly described in chapter 4.2,
has been developed for this purpose. The technique has been applied to a twin-working
electrode, bridged with electrochemically deposited polybithiophene. Measurements,
both time-resolved in situ EIS during electrochemical deposition of polybithiophene and
potential-resolved during oxidation/reduction potential sweeps, have been performed.
The analysis of the electric field taking place in the electrode configuration used has
allowed to make an estimation for the specific electronic resistance and the polymer
layer thickness.

• As mentioned above, the structure of electrochemically deposited polymer layer

highly depends on the condition of preparation and considerably influences the recharge
behavior of the conducting polymer. The deposition of conducting polymers onto metal
electrodes during electrochemical polymerization from monomer containing
electrolytes is a rather complex process involving diffusion of monomer to the
electrode surface and electrooxidation, subsequent back diffusion, oligomerization,
and precipitation onto the electrode. Hillman and coworkers [32,14] suggested a
deposition mechanism including instantaneous nucleation, followed by 3-D growth of
the nuclei until they overlap leading to formation of a plain polymer layer, but Li and
Albery [33] have pointed out that the alternative 2-D growth mechanism is also
possible. Investigations of the electrochemical polymerization of thiophene and its
derivatives have been performed by potential pulse [32,33] and potential sweep [34-36]
techniques combined with other measurement techniques, such as ellipsometry [36-38],
conductivity measurements [29,39,40], quartz microbalance [41,42], uv-vis
spectroscopy [32], EIS [43], etc. Measurements upon galvanostatic conditions can be
very useful for the study of the early stages of polymer deposition, where formation
rate of polymer layer can be controlled according to the requirements of an impedance
experiment.

Investigation of thickness dependence of impedance spectra can be easily performed

in situ, without removing the polymer coated electrode from the monomer containing
1 Introduction 8

electrolyte, by setting a zero current after deposition of every new layer. While
galvanostatic methods were used earlier predominantly for preparation of polymer
films and for proving their charge/discharge cyclability performance [44-47], a new
technique of EIS measurements, using the advantages of the galvanostatical condition,
has been developed. EIS measurements, aimed to investigate the process of
electrochemical growth of polybithiophene and the dependence of impedance spectra on
the layer thickness, have been performed as closely described in chapter 4.3. Time
resolved impedance spectra have been measured in situ during current controlled
polymerization. Thus the polymer film formation could be traced out for a wide range
of electric charge passed, i.e. from the onset of polymer deposition until a relatively
thick layer was grown. The impedance spectra obtained has been analyzed using models
of Ho et al. [48] and Paasch et al. [27], and the suitability of both models for analysis of
layers with different thickness has been discussed.

• The simple model [6], based on the work of Ho and coworkers [48], considers a
homogeneous nonporous polymer layer and can thus only be used for very thin layers.
Impedance measurements on polymer layers of different thicknesses, made by Tanguy
[9] as well as measurements made in this work let conclude, that transport of dopants
during recharge is not hindered in the bulk of polymer layer but in the polymer
aggregates, which are relatively small in comparison to the layer thickness. A polymer
layer can thus be imagined as a 2-phase medium, where one phase consist of the
polymer aggregates and another the electrolyte, which fills the cavities between them. A
mathematical model describing electrochemical behavior of such polymer layer is
developed by Paasch et al. [27]. A good correspondence of experimental results to this
model was observed during the investigation of oxidized polybithiophene layers with
different thicknesses. However, the assumption of macrohomoheneity in direction
perpendicular to the electrode interface was made, which is not strictly fulfilled by most
of electrochemically prepared polymer layers. Discrepancies between experimental and
simulated spectra were not considerable for polymer in oxidized state.
Macroinhomogeneity of specific resistance does not affect the spectra because of the
small value, which resistance of an oxidized polymer exhibits nonetheless. The situation
change in case of reduced polymer, where experimental data can not be described by
this model any further. To analyze experimental data measured on polymer layers in
1 Introduction 9

reduced state, the model of Paasch has been extended for the case of inhomogenenous
polymer layers (chapter 4.4). EIS measurements on polybithiophene layers were
performed steady state at different electrode potentials and the experimental result were
analyzed using the developed model.
2 Experimental 10

2 Experimental

2.1 Electronic equipment

All measurements described in this thesis have been performed using a computer
operated fast potentiostat/galvanostat, developed in our working group [xlix]. The
normally used measurement design is shown below.

Fig. 1 Electronic design used for electric measurements, dashed lines - output signal
routing, normal lines- input signal routing.
2 Experimental 11

The characteristics of electronic devices, presented in Fig. 1, are listed in detail in the
following table:

Fast potentiostat, Inst. f. Phys. Chem. d.c. - 100 kHz

University of Kiel
Signal generator, Inst. f. Phys. Chem. • Perturbation signal burned in EPROM
University of Kiel • DAC - 16 bit, synchronized with the
transient recorder
• 2×low pass filters, Kemo LTD
Beckenham Kent, type VBF8
Transient recorder, Inst. f. Phys. Chem. • 2×differential amplifiers
University of Kiel • 2×low pass filters, Stanford Research
Systems, inc. model SR640
• 2×ADC - 12 bit, 3 µs
• 2×256 Kbytes record length
ADC, DAC-1, DAC-2, Inst. f. Phys. Chem. 12 bit, 3 µs
University of Kiel
AT-486 16 Mbytes RAM, 256 Kbytes cache,
XENIX

2.1.1 Conventional electrochemical techniques

CV, coulometry, potential- and current pulse experiments were carried out using the
simplified electronic design shown in Fig. 1. Potentiostat/galvanostat was operated by
signal of DAC-1, which was controlled by the computer. Output signal was filtered by a
low pass filter, digitized by ADC and recorded. Software running under DOS and
XENIX were used to operate and record the measurements. For especially exact low
current measurements an extern DC-standard (Knick Berlin37 type S13) was used for
operating of the galvanostat.

2.1.2 Impedance measurements

In situ FFT- impedance measurements were performed using the electronic equipment
[xlix], shown in Fig. 1. In case of potentiostatic EIS measurements, the bias potential
given by DAC-1 was overlaid by an alternating voltage signal, produced by the signal
generator. A sum of 52 sinus waves, were burned into EPROM and used for generation
of the perturbation signal in the frequency ranges shown below.
2 Experimental 12

• 4 Hz- 30 kHz (τ=10 µs).

• 0.15 Hz- 5 kHz (τ=100 µs).

• 0.015 Hz - 450 Hz (τ=1000 µs).

Peak signal amplitude was selected specially for each particular measurement and never
exceed 15 mV, so linear conditions for EIS could be provided. Alternating potential and
current output signals were subtracted from the d.c. - bias component by differential
amplifiers, filtered by a low pass filter to prevent aliasing effects, digitized and recorded
by a two channel transient recorder. The perturbation signal was applied twice and only
the second half of the signal was recorded to prevent transient effects. Information
interchange between transient recorder and computer was initiated after each
measurement. Recorded signal in time domain was transformed into frequency domain
by means of fast Fourier transformation (FFT) immediately after each measurement or
stored in RAM and analyzed after finishing the entire experiment. In some cases the
„window“ method [l], using Hanning weighing function, was applied to additionally
improve the impedance spectra quality. Amplitude spectra of potential and current
signals were analyzed to prove the validity of measurements.

Measurement under galvanostatical conditions was performed analogously with

the difference, that the operating signals were given on the input of galvanostat and
resulting potential signals were recorded. The measurement equipment and data
analysis are detailed described in [43].

2.1.3 Impedance measurements on the twin electrode

Measurement set-up used for EIS- measurements on the twin electrode was described in
detail earlier [li]. It consisted of an FFT impedance spectrometer and a fast potentiostat
connected to the electrochemical cell as shown in Fig. 2.
2 Experimental 13

Fig. 2 Electronic circuit diagram of impedance measurement on twin electrode

An equal bias potential Ubias, measured against the reference electrode, was applied to
~
both W1 end W2 electrodes. Small perturbation voltage Upert with peak-to-peak
~
amplitude 15 mV was applied to the electrode W1. The perturbation voltage Upert and

the resulting current between W1 and W2 was used to calculate the impedance of the
twin-electrode at the preset bias electrode potential Ubias.
2 Experimental 14

2.2 Cells and Electrodes

2.2.1 Three-electrodes-cell
A multifunctional three electrode cell, presented below, was used for all measurements.

Fig. 3 The multifunctional 3 electrode cell used for electrochemical measurements

It was made out of Jena glass and rid up with 4 ground glass stopper sockets A,B,C and
D for attachment of different electrodes. For most of measurements two such cells were
used. One was filled with electrolyte, containing monomer and used for preparation of
polymer films as well as for investigation of the polymerization process. After the
preparation of polymer layer the stopper with the working electrode was removed from
the socket B, rinsed in clean acetonitrile for 5 minutes and then placed in the socket B
of the second analogous cell, filled with free of monomer electrolyte.

For all measurement a nonaqueous reference electrode was used. As reference, the
couple Ag/0.01 M AgNO3 +0.1 M TBAB was chosen. It is reversible and rapid in
acetonitrile and has a potential +0.291 V vs. SCE as reported in [lii]. The reference
electrode was mounted together with a Lugging capillary into a polyethylene stopper, as
shown in Fig. 4, at short distance to the working electrode to prevent for possible
disturbances.
2 Experimental 15

Fig. 4 The reference electrode used for

electrochemical measurements in the 3-electrode
cell.
1. Silver wire
2. Syringe needle for filling the Luggin capillary
tube with the electrolyte
3. Ag/0.01M AgNO3 solution in acetonitrile
4. Electrolyte solution hoisted from the cell
5. Frit membrane
6. Polyethylene stopper
7. Polyethylene Luggin capillary tube

A Platinumum wire melted into a glass tube and

cut off near the glass range served as the working
electrode. The cut was grounded with emery paper.
The diameter of the used Platinum wire determined
thus the diameter of resulting disc electrode and
was chosen specially for every particular
experiment. The glass tube with the Platinum wire
was fixed in a polyethylene stopper as shown in Fig. 3. Platinumum wire melted in a
glass tube and fixed analogously in a polyethylene stopper was used as the counter
electrode. All measurements were performed under Ar atmosphere.

2.2.2 The twin electrode

The twin-working electrode used for impedance measurements with one-side contacted
polymer layer was constructed analogous to one described in [31]. It possessed a
sandwich configuration consisting of two sheets of Pt with an insulation foil in between,
pressed together with a very little amount of epoxy seal and stored under pressure till
hardening (24 hours). The so prepared „sandwich“ was inserted into a glass stopper and
sealed in epoxy resin, high resistive against organic solvents, which was hardened by
120° for 20 minutes. The surface directed to the socket was grounded with series of
different emery papers with dimension of abrasive particles from 30 to 0.05 µm. The
2 Experimental 16

two Pt stripes, 5 mm × 0.5 mm each, separated by a 20 µm gap appeared after the

preparation as shown below.

Fig. 5 Cross section of the twin-working electrode. Platinum strips: 5 mm × 0.5 mm

each, gap: 20 µm.
The glass stopper with incorporated twin electrode was inserted into the socket D in
Fig. 3 and the bands of the twin electrode was then covered with polymer layer by
electrochemical polymerization in solution containing monomer. Dependent on the
particular experiment it was then rinsed with clean acetonitrile and placed in socked D
of an another cell with free from monomer electrolyte for further measurements.

2.3 Chemicals and Electrolytes

The acetonitrile used for preparing of all electrolyte solutions was „acetonitrile extra
pure“ from Merck, Art. 15500, which was stored for 30 minutes with heated Na2SO4,
then distilled from KMnO4, processed with small amount of concentrated H2SO4 and
then finally distilled from P2O5 as described in [liii].

TBAB (tetrabutilamoniumborfluorid) was used as the conducting salt. It was supplied

by Fluka, cleaned by recrystallization from double distilled water and dried by 80°
under vacuum. The background current measured to prove the absence of electroactive
species in the electrolyte on Pt electrode at 1.2 V vs. SCE has been found to be 7
µA.cm-1.

Bithiophene (Merk) was used as supplied, Pyrrole (Merk) was distilled under vacuum in
Ar atmosphere and stored in a refrigerator by -5°.

Electrolytes containing 0.05 M of monomer were used for electropolymerization in all

experiments.
3 Basic experimental data analysis 17

3 Basic experimental data analysis

3.1 The analysis of the impedance spectra of conducting polymer

layers
The EIS spectra of conducting polymers are quite complex and still not well understood
because of numerous processes taking place simultaneously during recharge of the
polymer layer. However, a number of models, describing the impedance spectra for
some particular kind of polymer layers, has been proposed in the literature [6,27,liv,62-
66]. In this work only the models, describing the capacitive effects in the low
frequency range through chemical processes, as first proposed by Hunter et al.[6], have
been applied. This well known from the literature approach emphases the restricted
amount of the electroactive species on the interface of the electrode as the reason for the
capacitive behavior of the polymer layer in the low frequency region. The mechanism
of this kinetical effect can be demonstrated on the example of the simplest
representative of such systems - the monomolecular layer of electroactive species fixed
on the electrode. The probability of diffusion of the dopping ions from the electrode
interface into a polymer layer with a finite thickness [48] and the influence of the
porosity of the electroactive layer [27] are the complications solved by two models
presented further in this chapter.

3.1.1 Monolayer of an electroactive species

Monomolecular layer of electroactive species, fixed on the interface, is the simplest
representative of the electroactive coatings and offers thus a favorable object for
illustrating the common properties, shared by more complicated systems. An
electrochemical behavior of such a system at higher frequency range is analogous to
one, characteristic for species dissolved in the electrolyte. The charge transfer
resistance Rct in parallel with the double layer capacitance Cdl results in a semi-circle in
the impedance spectrum, if presented as a Nuquist plot. The change of the concentration
of the electroactive specie during the harmonical perturbation of potential becomes
significant at lower frequencies and leads to more complicated impedance behaviour.
3 Basic experimental data analysis 18

To evaluate the quantitative correlation between the kinetic characteristics of this

system and its impedance spectrum it can be helpful to make use of the equivalent
circuit presented below, known as the Randles circuit.

Fig. 6 Randles equivalent circuit for the ac response of an activated charge transfer
and concentration hindrance of reaction. Rs is the serial resistance of electrolyte, Cdl,
the double layer capacity, Rct, the charge transfer resistance and Zlim is the
concentration hindrance impedance.
This circuit describes a behavior of all non distributed electrochemical systems, if only
one redox couple is present and the Zlim is considered to be any kind of a concentration
hindrance, i.g. Warburg impedance for the case of an active specie, dissolved in
electrolyte. The evaluation of the impedance of electroactive monolayer, fixed on the
electrode is thus simplified to the determination of the expression for Zlim. Evaluation of
this impedance means finding the relation between the small perturbation voltage δE,
and the response current δi under the assumption, that no charge transfer resistance or
double layer capacity is present. In this case the flow of charge δQ is linearly connected
with the change of the concentration of the oxidized species by the Faradaic law.
Assuming further, that δE/δt = dE/dt, δE/δQ = dE/dQ for small amplitude of
perturbation potential, the definition of the current i=δQ/δt can be rearranged, as

dE SFz
i= eqn. 1
d t dE dc

Here c is the interfacial concentration of the oxidized species, S, the interface of the
electrode, F, the faradaic constant and z, the number of electrons involved in the
reaction. This expression is analogous to the equation describing the current response of
3 Basic experimental data analysis 19

capacitor. Hence it is justified to denote the time independent multiplier as

„pseudocapacity“, Cps.

dE SFz eqn. 2
i= C ps C ps =
dt dE d c

If the systhem exhibits an ideal behaviour, the dependence dE/dc can be determined by
differentiating the Nernst's equation. The corresponding equation for Cps for such a
system is then given as:

 zF 
2 exp  ( E − E 0 )
F c0 z  RT 
Cps = 2 eqn. 3
RT   zF  
exp  ( E − E 0 ) + 1
  RT  

Here c0 is the sum of interfacial concentrations of the oxidized and reduced species, E is
the bias electrode potential, at which the small alternating voltage δE is applied and E0
is the formal potential of the redox couple. It can be seen, that the observed system
exhibits an impedance behavior of a capacitor, which capacitance varies with the bias
electrode potential. The peak of the capacitance value is reached at the formal potential
of the particular redox couple.

3.1.2 Homogenous nonporous electroactive layers

If thin, but not monomolecular layers of the electroactive substance has to be
considered, as normally the case with conducting polymers, a diffusion of dopping ions
into the inner parts of the layer has to be taken into account. The analysis of the ac
current response of such a system is given by Ho et al. [48]. Analogously to the
approach used in chapter 3.1.1, the Randles equivalent circuit can be used. The Zlim has
to be evaluated considering the gradient of concentration at the interface, described by
the Fick's law. The solution of the Fick's differential equation

∂ [δc( x, t )] ~ ∂ 2 [δc]
=D eqn. 4
∂t ∂x 2

with the boundary conditions dc(0,t)=0 and dc(l,t), as given by eqn. 2, allows to
determine the impedance of the concentration hindrance as:
3 Basic experimental data analysis 20

dE 1 coth( l iω D )
ZrfW = eqn. 5
dc zF iω D

Here ZrfW is the "restricted finite Warburg" impedance and l is the thickness of the
layer. The impedance of the equivalent circuit shown on Fig. 6, with the ZrfW
substituting Zlim, is given by:

.−1
 1  
ZHo =  + iω Cdl  S eqn. 6
 ZrfW + Rct  

An example of an impedance spectra of such a system is shown as Nuquist plot below.

Fig. 7 Computer simulation of the impedance of homogenous nonporous layer of

electroactive species using the equivalent circuit shown in Fig. 6 with restricted finite
Warburg impedance as concentration hindrance.
It can be seen, that the kinetics changes from charge transfer control by high
frequencies to diffusion control by intermediate frequencies. At very low frequencies
range, the effect of the finite layer thickness becomes dominant and the restricted finite
Warburg impedance approaches the impedance of a capacitor Cps, connected in series
with a resistance RD. The values of this elements can be expressed as:

SFz dE dc
C ps = l RD = l eqn. 7
dE dc 3DSFz
3 Basic experimental data analysis 21

It can be seen, that the pseudocapacity for the case of a multimolecular layer is
expressed analogously to one of the monomolecular layer with the difference, that here
c is the volume concentration of the electroactive species. dE/dc can be determined by
differentiating of the Nernst's equation (cf. eqn. 3), if the particular redox couple
exhibits ideal behavior.

3.1.3 Porous electroactive layers

Thick polymer layers exhibit impedance behavior, which can not be described by the
simple model, presented in the previous chapter. The semicircles observed in the higher
frequency range in the complex impedance plots are depressed, which can be
accounted for the distribution of the time constants. The possible reason for such a
distribution for the case of conducting polymer layers could be their high porosity.
Paasch et al. [27] have proposed a theoretical model, which allows an analysis of
systems exhibiting such behavior. In this theory the polymer layer is assumed to be a
macrohomogenous two-phase system consisting of small polymer aggregates and
electrolyte. The diffusion of the dopping ions occurs inside of this aggregates, therefore
the characteristic diffusion length can be assumed to be independent of the polymer
thickness, corresponding to results of experimental measurements of Tanguy et al. [9].
To find the total impedance of such a system, the entire polymer layer is divided into N
elementary volume elements with cross-sections A. The polymer is assumed to be
contacted to polymer and the electrolyte to electrolyte at each cross-section.
3 Basic experimental data analysis 22

Fig. 8 An elementary volume of the porous polymer layer and the corresponding
equivalent circuit as assumed by model of Paasch et al.
The charge transfer through the polymer-electrolyte boundary at each cross-section is
determined as in the model of Ho et al. under consideration, that the diffusion of the
dopping ions take place inside of the elementary agglomerates of the polymer, having a
characteristic dimension l, and not through the whole polymer layer. This interfacial
impedance can therefore be expressed as

.−1
 1  Sf 
Z Ho =  ∗
'
+ iω Cdl  ⋅  eqn. 8
 ZrfW + Rct  d 

The effective interface of the polymer agglomerates inside one elementary volume is
given here by the expression Sf/d, where f is the effective interface coefficient and d the
thickness of the polymer layer. The ZrfW is calculated analogously to eqn. 5 with the
difference, that the diffusion length l is treated as the average dimension of polymer
aggregates independent on the layer thickness. The current flow through each cross-
section is determined by an electronic resistance of the polymer ρ1, interfacial
impedance Z'Ho and the ionic resistance of the electrolyte ρ2. The impedance the entire
3 Basic experimental data analysis 23

polymer layer is then analogous to one of a transmission line consisting of chain

connected elements shown in Fig. 8.

Fig. 9 Generalized transmission line representing the impedance of porous

macrohomogenous electrode
The solution for the input impedance of this circuit under consideration of infinite N is
given by Paasch et al. However, for the case of oxidized or slightly reduced conducting
polymers, the expression for the total impedance of presented generalized transmission
line can be considerable simplificated by an assumption, that either electronic or ionic
resistance of the polymer layer is neglectable small. The impedance of a cross-
connected transmission line will then be equal to the impedance of normal transmission
line, terminated by an impedance Zt.

Fig. 10 Generalized transmission line representing the impedance of porous

macrohomogenous electrode neglecting the specific resistance of the polymer or of the
electrolyte.
The terminating impedance Zt can be either infinity for a normal impedance
measurement or zero for measurement on polymer layers contacted only on the
electrolyte side or on the polymer side. The expression for the total impedance of this
3 Basic experimental data analysis 24

circuit under consideration of an infinite small distance δx and Zt equal to infinity is

derived by Fletcher [liv]:

 ρ S
Z = Z' Ho ρ S ⋅ coth d ⋅ '  + Rs eqn. 9
 Z Ho 

The serial distributed resistance per unit length ρ´1 is replaced here by ρ/S, where ρ is
the specific resistance of the polymer. A computer simulation results for the same
parameter using eqn. 6 and eqn. 9 are presented below.

Fig. 11 Computer simulation of impedance spectra for the same parameters using Ho
et al. model (eqn. 6)- solid line, and simplified Paasch et al. model (eqn. 9)- dashed
line. Specific resistance ρ = 4.8 E+5 Ω.cm was used in eqn. 9.
In the low frequency region the impedance spectra predicted by the model is practically
analogous to one of serial connected resistance R and capacity C. The interpretation of
this quantities is similar to one given in eqn. 7. The difference is in the additional serial
resistance Rl, which accounts for the resistance of polymer and is related to the specific
resistance as

Rl = ρ d 3S eqn. 10
3 Basic experimental data analysis 25

3.2 Field considerations for measurements on the twin electrode

The specific conductivity of an electrochemically deposited polymer layer is an
essential electrochemical characteristic. However, the conventional impedance
measurements, where the polymer layer is contacted on both polymer and electrolyte
sides, do not allow to distinguish between the electronic conductivity of polymer itself
and the ionic conductivity of the electrolyte in pores. Conductivity measurements
between the bands of a twin-electrode, as shown in the picture below, covered with a
polymer layer, was proposed by several authors for this purpose [11,12,29].

Fig. 12 Cross-section of a twin electrode covered with a polymer layer of thickness g.

The polymer-electrolyte interface, which is partly present also inside of the polymer
layer because of its porosity, has highly capacitive character. Therefore, high frequency
ac current flows between the electrodes not only trough the polymer but can also across
the walls of cavities and uses the electrolyte as an alternating path. This effect is
diminished by very low frequencies, however, where the pure electronic resistance can
be measured.

Measurements between the bands of the twin-electrode can be applied to obtain the
value of conductance, which can be used to compare polymer layers with different
oxidation level, but depends on the particular electrode configuration. To evaluate the
specific conductance of polymer layer, an analysis of electrical field between the bands
of the twin electrode is required.

The theory of conformal mapping was applied by Kankare et al. [29] to find the solution
of the current flow field for the electrode configuration analogous to this used in our
work. According to this approach, the electrode configuration depicted into complex
3 Basic experimental data analysis 26

coordinate plane is transformed with preserving the angles between the boundaries into
another complex plane, where the solution of the field can be easily found. For the
transformation of the polygonal boundaries, the Schwarz-Christoffel equation is usually
used to calculate the transformation function of the coordinates. The authors have
shown, that the configuration in Fig. 12 can be transformed to a plate-capacitor-like
configuration:

Fig. 13 Space obtaining by two-stage conformal mapping of the configuration shown in

Fig. 12 with the polymer layer mapped into the interior of a rectangle and electrodes
into the opposite sides of the rectangle.
The virtual width of the electrodes, φ0, in this new complex coordinate system can be
expressed through the coordinates of electrodes in the original plane as [lv]

eqn. 11
3 Basic experimental data analysis 27

1 a
tanh( π )
K ′( k ) 2 g
φ0 = k=
2K ( k ) 1 b
tanh( π )
2 g

Here, a and b are dimensions of the twin-electrode and g the thickness of the layer, as in
Fig. 12; K(k) and K'(k) are complete elliptic integrals modulus k of first and second
kind respectively, as in

1
1 k
1 1
K(k ) = ∫ dx K ′( k ) = ∫i dx
0 (1 − x )(1 − k x )
2 2 2
1 (
(1 − x ) 1 − k 2 x 2
2
)
eqn. 12
The configuration shown in Fig. 13 is one of a plate capacitor thus the conductance
between the two plates is given by:

G = σ l φ0 eqn. 15

Kankare et al have shown, that the expression for conductance G can be simplified for
the case of an intermediate layer thickness (a < g < b) as in eqn. 13.

σl 8 
G≈  ln + ln g eqn. 13
π  πa 

In the region g > b the resistance between the bands of twin electrode becomes
independent on the of the layer thickness:

σl eqn. 14
G≈ ln( 4b a )
π
3 Basic experimental data analysis 28

Using these two expressions, thickness and specific resistance of the polymer layer can
be determined, if the dependence of conductivity between the bands of the twin
electrode on polymerization charge is known. Assuming, that the thickness of deposited
polymer layer depends linearly on the charge used for polymerization per unit area (g =
kQ) , eqn. 13 can be rearranged as follows:

σ l  8k 
G≈  ln + ln Q eqn. 16
π  πa 

Accordingly, the specific conductivity can be calculated from the slope of G vs ln Q.

The intercept gives the thickness growth coefficient k.

π πa
σ= tan(α ) k= exp( G ( 0)π σ l ) eqn. 17
l 8

An example of a dependence of G on Q, simulated for a dummy twin electrode is shown

in Fig. 14.

Fig. 14 Conductance between the bands of the twin-electrode vs. ln(Q) exact simulated
using eqn. 15 (a) and approximately simulated using eqn. 13 (b) and eqn. 14 (c). Here,
a=1, b=20 and l=k=σ=1 were used in simulation.

3.3 Complex nonlinear least squares fitting

3 Basic experimental data analysis 29

EIS spectra are often complex and can give only qualitative information about
properties of the examined electrochemical system. Quantitative informations can be
obtained by fit of experimental data to a complex function resulting from the used
physical model. For this purpose, the complex nonlinear least squares (CNLS)
immittance fitting program LEVM 6.1 by J.R.Macdonald [lvi] was used and
modificated in this work.

The least squares fitting procedure optimizes parameters of the used physical
model by minimizing the sum:

[
S ( P ) = ∑ w j Z j − Z (ω j , P ) j ]
2
eqn. 18
j

Here, Z is the vector of experimental impedance values for each frequency, Z(P) is the
functional dependence of impedance on frequency ω and vector of parameters P,
resulting from a particular physical model. The sum is calculated in range j = 1-N,
where N is the total number of experimental data points. W is the vector of weights for
each point. Weighting factor wj is calculated from the uncertainty of the jth data point,
βj, as follows:

( )
2
wj =1 β j eqn. 19

Levenberg-Marquardt algorithm [lvii] is used to minimize the sum in eqn. 18.

Relative and absolute standard deviations for each parameter fitted are calculated during
the minimization. These values were used to check the goodness of the entire fit as well
as to reveal the effect of any parameter on the function value in view of a particular
parameter set. The value of relative deviation for every parameter is conversely
proportional to its average partial derivative over the frequency range where the fit is
performed. The parameters having insignificant effect on the functional value in
considered frequency range would therefore be calculated with very large relative
deviations. This makes the error variance analysis important for validation of the
parameter reliability.
4 Results 30

4 Results

4.1 Impedance measurements during oxidation and reduction of

conducting polymers

4.1.1 Simulation procedure

As well known from measurements performed under potential control, the low
frequency impedance of polymer covered electrodes can be approximated by a
capacitance and a series resistance [6,lviii-lx]. The nature of the capacitance is still not
well understood. It is often referred to as "limiting", that reflects its connection to the
finite diffusion at limiting low frequencies, or as "redox" to point out its relation to
oxidation/reduction processes of the polymer film.

In order to simulate the capacitive part of the current-voltage curve we assumed that the
equivalent circuit of the cell can be represented as a series combination of a resistor R
and a capacitor C, both dependent on the applied electrode potential U. Then the
potential drop Uc across the capacitor when current I flows is

U c = U − IR eqn. 20

and the charge Q = Uc . C , stored in the capacitor is

Q = (U − IR)C eqn. 21

Taking into account the dependence of R and C on the potential and assuming a
constant sweep rate υ = dU/dt in cyclic voltammetry, differentiation of eqn. 21 gives

dQ  dR dI  dC
= υ − Iυ − R  C + υ (U − IR) eqn. 22
dt  dU dt  dU

which can be rearranged to give the differential equation for the current as

dI  1 1 dR 1 dC  1 U dC
+ + +  I= + eqn. 23
dU  υRC R dU C dU  R RC dU

The solution of eqn. 23 for I = I(U) has been obtained numerically using the Runge -
Kutta method. The data for R(U) and C(U) used in the calculations have been obtained
4 Results 31

from low frequency part of EIS measurements. The anodic and the cathodic sweeps
have been simulated separately with the initial conditions I+,start = 0 at the start of the
anodic sweep and I-,start = I+,end , where I-,start is the current at the start of the cathodic
sweep and I+,end is the current at the end of the anodic sweep. For the sake of
comparison, simulation of cyclic voltammograms (CVs) using obtained C(U) data has
been also performed using

I c = C(U )υ , eqn. 24

neglecting the potential dependencies of C and R, as previously reported by Tangui at

al. [9].

4.1.2 Measurements and simulation results

The measurements have been performed on a polished polycrystalline Pt electrode (0.02
cm2) coated with PBt or PPy layers, obtained galvanostatycally at 2 mA. cm-2 for 40 s.
The impedance spectra measured during anodic and cathodic potential sweeps at a scan
rate of 20 mV. s-1 in the ranges - 0.3 to 1.0 V for PBt and 0.4 to 0.7 V for PPy are
shown in Fig. 15 and Fig. 16 respectively.
4 Results 32

Fig. 15 Complex plane impedance

plots for a PBt-coated electrode
measured at different electrode
potentials: (a),(b) anodic sweep;(c)-
cathodic sweep.
4 Results 33

Fig. 16 Complex plane impedance plots for a PPy-coated electrode measured at

different electrode potentials; (a) - anodic sweep; (b) cathodic sweep.
It was previously observed, that condition of stady state in polymer layer is reached
much faster at potentials, where the polymer is oxidized, then at reducting potentials.
Thus, in order to reach reproducible initial conditions for carrying out CV and EIS
measurements, the voltammetric cycles were started with the cathodic potential sweep.
The potential limits were chosen such that (i) the polymer layer was still slightly
oxidized at the end of reduction scan and (ii) high anodic potentials were avoided to
prevent the layer from damage due to over-oxidation. The potential dependencies R(U)
and C(U) of the resistance and capacitance of the samples, derived from the low
frequency parts of the measured impedance spectra, are shown in Fig. 17 and Fig. 18 for
PBt and PPy respectively.
4 Results 34

Fig. 17 Capacitance C and

resistance R derived for the
PBT-coated electrode from
impedance spectra measu-
rements.

Fig. 18 Capacitance C and

resistance R derived for the
PPy-coated electrode from
impedance spectra measu-
rements.
4 Results 35

Results from numerical integration of eqn. 23 and eqn. 24 for the case of PBt are shown
in Fig. 19 for the anodic and the cathodic sweeps.

Fig. 19 CV of the PBt-coated electrode. The capacitive current was obtained by

numerical integration of (a) eqn. 23 and (b) eqn. 24. The CV recorded experimentally at
a scan rate of 8.5 mV. s-1 is shown for comparison as curve (c). Curve (d) is the
difference between curves (c) and (a).
As mentioned above, the C and R data were derived from a measurement in which the
nominal potential sweep rate was 20 mV. s-1. However, the potential sweep was
interrupted during the perturbation and data transmission to the computer, so that the
effective scan rate was 8.5 mV. s-1. This value was used in the integration and thus a CV
recorded at 8.5 mV. s-1 is presented for comparison, in Fig. 18. The calculated capacitive
current and the measured current-voltage curve for the PPy- covered electrode are
shown in Fig 20.
4 Results 36

Fig 20 CV of the PPy-coated electrode. The capacitive current was obtained by

numerical integration of (a) eqn. 23 and (b) eqn. 24. The CV recorded experimentally at
a scan rate of 8.5 mV. s-1 is shown for comparison as curve (c). Curve (d) is the
difference between curves (c) and (a).
4.2 Impedance measurements on twin working electrodes bridged with
polybithiophene layers

4.2.1 Verification of the measurement set-up

~
As shown in 2.1.3, an application of a perturbation voltage U pert accross the bands of
~
the twin electrode results in a flow of response current Iresp , which can be measured and

used for calculating the impedance spectra of the part of the polymer layer, bridging the
gap. However, it must be established that the obtained impedance spectra characterize
only the polymer layer between the two metal stripes, i.e., that the ac response current,
~
Iresp , has no components due to currents crossing the polymer/electrolyte interface.

The twin-working electrode, with the gap bridged by conducting polymer can be
represented by a simplified equivalent circuit as depicted in Fig. 21.

Fig. 21 Simplified equivalent circuit of

a twin-working electrode, covered with
conducting polymer.
4 Results 38

The impedance elements Z1 and Z2 represent the metal/polymer/electrolyte branches for

the two twin working electrodes, and Z1,2 is used for the metal/polymer/metal branch,
i.e. for the polymer layer bridging the gap. Note that U1 and U2 are given with respect to
~
system ground. When a small perturbation voltage, U pert , is applied to W1, two currents,
~ ~ ~
I1 and I1,2 , flow through Z1 and Z1,2, respectively. The response current, Iresp , measured
~ ~ ~
at W2 should be equal to the sum of I1, 2 and I 2 , where I 2 is the a.c. current through Z2,
~
due to the perturbation. One can consider I 2 as this part of the response current that
crosses the interface polymer/electrolyte at W1, passes through the electrolyte and
enters W2 crossing again the polymer/electrolyte interface. On the other hand the bias
voltage, Ubias, used to adjust the working potential in an EIS experiment is generally set
to be a d.c. voltage, and the fast potentiostat used provides that the potential difference
measured between the reference electrode and the working electrode is exactly equal to
the preset bias voltage. The working electrode W2 is connected to virtual ground and it
appears as a true working electrode to the potentiostat. Thus, the current across Z2
~ ~
depends only on Ubias, and does not depend on U pert , i.e. its a.c. component I 2 = 0, and
~ ~
consequently Iresp = I1,2 .

Nevertheless, an experiment has been curried out to additionally assure the prevention
of cross-interface currents by the measurement set-up. The EIS spectra, resulting from
the test measurements, are shown in the figure below:

Fig. 22 Impedance spectra measured on the dummy cell shown on Fig. 21. The curve
(a) marked with squares - switch A OFF, potentiostat OFF. Curve (b) marked with
4 Results 39

triangles- switch A ON, potentiostat OFF. Curve (c) marked with circles- measurement
with switch A in position ON, potentiostat control ON.
Curve (a) present a measurement made with a circuit representing interfacial impedance
switched OFF. This provide an ideal case, where EIS of the circuit modeling the "metal-
polymer-metal" bridge is measured only. In case (b), the normal two-electrode
impedance between W1 and W2 is measured. Potentiostat is switched OFF, therefore no
compensation of cross-interface currents is performed. The impedance spectra
corresponds to the electrical circuit W1-A-W2 and differ from (a). A reliable
measurement of the impedance of the polymer bridge over the gap between the bands of
twin electrode is not possible in this configuration. Case (c) occurs in experiments made
in a real electrochemical cell. The circuit representing interfacial impedance is switched
to ON but the cross-interface current is compensated by the potentiostat, which
establishes a constant potential between the pair of working electrodes and the reference
electrode. It can be seen, that curves (a) and (c) are almost identical. Thus, the
performed test measurements prove experimentally, that the used measurement setup
effectively prevents flowing of cross-interface currents.

4.2.2 Impedance spectra

As shown in the previous section, in situ EIS measurements, free of distortion due to
current flow through electrolyte, can be performed on twin electrode bridged with
polymer layer and can give information about the electrochemical properties of the
polymer. Impedance measurements on polymer layers in different oxidation levels were
carried out.

The twin-electrode was covered by polybithiophene layer by means of potentiostatical

electropolymerization at 0.8 V in monomer solution until the total charge of 800
mC.cm-2 passed. The obtained twin-electrode, covered with polymer, was rinsed with
acetonitrile and transferred into an electrochemical cell with solution free of monomer.
By means of potential sweeps in range 0.3 V - 0.8 V the polymer layer could be reduced
and oxidized reversibly. The impedance spectra resulting from the series of EIS
measurements, performed for electrode potentials ranging from 0.36 V to 0.8 V are
shown in a complex plane diagram.
4 Results 40
4 Results 41

Fig. 23 Series of impedance spectra in the complex plane as obtained during anodic
sweep in monomer-free solution in potential range from 0.36 to 0.6 V (a) and from 0.73
to 0.8 V (b).
4 Results 42

The results can also be presented as Bode-plots, which allow better comparison of the
spectra measured for high and low potentials.

Fig. 24 Series of impedance spectra measured on a polybithiophene covered twin-

working electrode during anodic sweep in potential range from 0.36 to 0.8 V presented
as Bode-plots.
The values of the resistance between the bands of the twin electrode, extrapolated from
the low frequency region of impedance spectra, are presented versus the applied
electrode potential:
4 Results 43

Fig. 25 Resistance data, obtained from the impedance spectra (Fig. 24), measured
between the bands of the twin-electrode covered with polybithiophene layers at different
electrode potentials.
For the sake of comparison, the two twin- working electrodes were externally short
circuited and an impedance spectrum of the polymer-electrolyte interface was measured
using the conventional three-electrode cell configuration. Fig. 26 shows the impedance
spectrum, as obtained with anodic polarization of 0.8 V, i.e., with the polymer in its
oxidized state.

Fig. 26 Impedance spectrum of the

twin-electrode, covered with
polybithiophene, measured at 0.8 V
with a conventional EIS technique with
the bands of the twin-electrode
connected externally.
4 Results 44

As shown in Chapter 3.2, the specific conductivity of polymer layer, deposited

on the twin electrode, can be determined by analyzing the dependence of the
conductivity between the bands of twin electrode on the charge used for polymerization.
The required conductivity data can be obtained from the low frequency parts of
impedance spectra, measured between the bands. Aiming to evaluate the
conductivity/polymerization charge dependence for a wide range of thicknesses, a series
of impedance measurements during a deposition of polymer layer was performed. With
this purpose a polished twin-working electrode was subjected to an anodic potential of
0.8 V in a bithiophene solution. The metal stripes were covered with polymer and the
growing polymer layer bridged the gap. The charge, Q, passed through the twin-
electrode was measured simultaneously by a current integrator and is presented in the
figure below.

Fig. 27 Total charge, Q, passed through the cell during polymer deposition on the two
bands of the twin-electrode. EIS measurements were performed at 40 s intervals as
marked with the numbered points.
A series of 20 EIS measurements was performed at time intervals of 40 s after the
beginning of polymerization. The impedance spectra obtained are shown in a complex
plane diagram, Fig. 28a, and as Bode-plots, Fig. 28b and c.
4 Results 45

Fig. 28 Series of impedance spectra in

the complex plane (a), and as Bode
plots (b, c), as obtained between the
two twin-working electrodes during
polymerization at 0.8 V in monomer
solution. The measurements were
performed at 40 s intervals, and the
numbers correspond to the respective
points in Fig. 27.

These results were used to extract data for the dependence of the polymer resistance on
polymerization charge as depicted in the picture below.
4 Results 46

Fig. 29 Resistance data, extracted from the low frequency region of the impedance
spectra, presented in Fig. 28.

4.2.3 Estimations of conductivity and thickness of the polymer layer

The resistance data, obtained from impedance measurements between the bands of the
twin electrode during electropolymerization, can be used for estimation of specific
resistance of the deposited polymer layer as shown in chapter 3.2. For this purpose the
resistance data from Fig. 29 has been converted into conductance and presented versus
logarithm of charge, used for deposition of the particular layer as follows:
4 Results 47

Fig. 30 (a) Conductivity data obtained from impedance spectra measured during the
electrodeposition of polybithiophene vs. logarithm of charge, used for deposition; (b) fit
of experimental data by eqn. 13 using dimensions of the twin electrode as in Fig. 5 and
σ=0.273 S.cm-1, k= 0.011 µm mC-1.
It can be seen, that at the beginning of deposition no linear dependence of G on
logarithmic charge is observed. That can be explained through an absence of an even
bridging of a gap between the bands of the twin-electrode by the growing polymer
layer. For thicker layers the conductance becomes linear dependent on the logarithm of
charge according to eqn. 13. Hence, the values of the specific conductance of polymer
layer σ and the thickness growth coefficient, k, related to 1.cm2 electrode interface,
could be estimated as 0.27 S.cm-1 and 0.011 µm mC-1 respectively. The thickness of the
layer deposited upon a polymerization charge of 800 mC.cm-2 has been found to be 8.8
µm.
4 Results 48

4.3 The thickness dependence of the impedance spectra of

polybithiophene layers measured under galvanostatic condition
The electrochemical properties of conducting polymer layers are strongly dependent on
their structure. The analysis of impedance spectra for polymer layers of different
thickness in view of a particular physical model allows to prove or to refute the
structural deliberations, laying in its basis. If the impedance dependence on thickness,
predicted by the model (cf. chapter 3.1.2), take place, an estimation of the kinetic and
thermodynamic parameters, common for the layers of different thickness, becomes
possible.

As known from literature [32-36], the electrosynthesis of polybithiophen starts with

initial formation of a dense polymer layer on the electrode substrate, but thicker layers
have porous structure. Under the assumption, that the thermodynamic properties of
polybithiophen do not depend on the layer structure, a combined analysis of thin and
thick polymer layers is possible. Thereby, it was interesting to investigate very thin
polymer layers, for which the simplified model (chapter 3.1.2) can be used and then to
apply the evaluated parameters to analysis of thicker layers, for which the porosity has
to be considered in addition. The EIS spectra, required for such investigation, were
obtained using electrodes covered with both thin and thick polymer layers performing
the impedance measurements under galvanostatical conditions.

4.3.1 Measurement results

Two series of EIS measurements were performed during galvanostatic polymerization
of bithiophene. In the first one, the polymerization was curried out at relatively low
current density of 0.15 mA.cm-2 in order to slow down the deposition rate and thus
allow impedance spectra measurements at early stages of polymer growth. The
recorded electrode potential vs. time dependence is shown in Fig. 31.
4 Results 49

Fig. 31 Potential vs. time curve, obtained during electropolymerization of bithiophene

at 0.15 mA.cm-2. Impedance spectra measured as marked with dots.
The potential of the working electrode increases abruptly to a value of about 0.79 V
and does not change considerably even upon prolonged polymerization. The EIS
measurements were performed in 5 s intervals as marked in the figure. The obtained
impedance spectra are shown as Bode-plots in Fig. 32 a, b. and the 5th spectrum is
presented as a Nuquist plot in Fig. 32, c.
4 Results 50

Fig. 32 A series of impedance spectra

obtained during galvanostatically
electropolymerization of bithiophene at
0.15 mA.cm-2 and the fits made using
Ho et al. model, presented as Bode
plots (a,b) and the 5th spectra as a
Nuquist plot (c).
4 Results 51

In a second experiment the polymer layer was deposited at a higher rate with
current density of 2 mA.cm-2. The EIS measurements were performed at intervals of 15
s. In order to avoid polymer deposition during the impedance measurements, the cell
current was switched to 0.0 mA.cm-2 for the duration of the perturbation pulses and
respective data transmission to computer. The impedance spectra obtained are shown
below as Bode plots. The total polymerization charge reached during this experiment
was 450 mC. cm-2.

Fig. 33 A series of impedance spectra, measured during the polybithiophene layer

deposition by polymerization current of 2 mA.cm-2. Solid lines represent the fit of the
experimental data using the Paasch et al. model.
4 Results 52

4.3.2 Data processing

The impedance spectra measured during polymerization at 0.15 mA.cm-2 were analyzed
using the model of Ho et al. [48]. A second charge transfer resistance, Rp, has been
introduced to account for the polymerization reaction. The eqn. 6 thus can be rewritten
to give:
.−1
 1 1 
Z1 =  + iω Cdl +  ⋅ S + Rs eqn. 25
 ZrfW + Rct Rp  

where Rct is the oxidation/reduction charge transfer resistance, Cdl is the double layer
capacitance, S is the surface area of the electrode and Rs is the serial resistance. CNLS
fitting of experimental data by use of eqn. 25 yielded the impedance spectra shown in
Fig. 32. The thickness of the polymer layer obtained after passing 30 mC.cm-2
polymerization charge was estimated to be 80 nm using the value for polymer density ρ
= 1.5 g.cm-3, as reported by Koßmel et al [30]. The value of dE/dc = 299.9 V.cm3 (cf.
eqn. 5), determined from the fit of EIS data obtained for 30 mC.cm-2 layer, was assumed
to be constant during polymerization and has been used further to estimate the polymer
layer thickness for different values of polymerization charge passed as shown in Fig.
34a.
4 Results 53

Fig. 34 Thickness of polymer layer and apparent diffusion constant, vs. polymerization
charge passed during the polymer deposition at current density 0.15 mA.cm-2.

The estimated values for the apparent diffusion constant of counterions, D, are shown in
Fig. 34b. In the region, where the apparent diffusion constant becomes independent of
the layer thickness and therefore Ho et al. model is valid, the diffusion constant D was
determined as 1.8.10-8 cm2 .s-1. The values for Rp, Cdl and Rct obtained from the fits are
presented in Fig. 35.
4 Results 54

Fig. 35 Charge transfer resistance of polymerization reaction (a), double layer

capacitance (b) and charge transfer resistance of oxidation/reduction of
polybithiophene layer (c) vs. polymerization charge passed during the polymer
deposition at current density of 0.15 mA.cm-2.

The treatment of the impedance data measured on polymer layers consecutively

deposited at a polymerization current of 2 mA. cm-2 was performed using the simplified
model of Paasch et al (eqn. 9). The values of dE/dc, Rct and the diffusion constant of
counter ions D were assumed to be invariable and equal to the values estimated in the
experiment with thin polymer layers. The value of diffusion length l of the counter ions
could thus be evaluated from the fits as in Fig. 36a.
4 Results 55

The assumption of a constant Rct value allowed to estimate the effective interface of
polymer, accessible for electron transfer. The effective interface coefficients obtained
by fit of experimental data are shown on Fig. 36b.

Fig. 36 Specific diffusion length (a) and effective interface coefficients vs.
polymerization charge passed during the polymer deposition by current 2 mA.cm-2.
4 Results 56

The value of the effective diffusion length increases with the thickness of the polymer
and saturates for polymerization charges > 200 mC. cm-2 . Thus the region of validity of
the model of Paasch et al. could be determined (Q > 200 mC. cm-2) and the thickness of
polymer layer, the double layer capacity and the specific resistance could be evaluated,
applying this model to the experimental EIS data obtained.

Fig. 37 The thickness of the polymer layer (a), double layer capacity (b) and sum of
ionic and electronic conductivities of polybithiophene layer (c) vs. polymerization
charge passed during the polymer deposition at current 2 mA.cm-2.
4 Results 57

4.4 Impedance investigation of conducting polymer layers considering

their inhomogeneity.

4.4.1 Model of an inhomogeneous porous electroactive layer

Experiments described in this thesis as well as data from the literature [9,25] show, that
electrodeposited polymer layers are porous and behave as a two-phase system. The
model of Paasch et al. is based on the assumption of macrohomogeneity of the polymer
layers and allows good description of their impedance spectra in the oxidized state.
However, the observed impedance spectra of polymers in reduced state, presented in the
complex plane, exhibit a slope between 50 and 70 degree in the lower frequency range
(cf. Fig. 15), which can not be descibed by the model of Paasch at al. One of the
possible reasons for such a behavior could be inhomogeneity of the polymer layer in the
direction perpendicular to the electrode interface. The analysis of the polymer layers of
different thickness also prove the reliability of this assumption. Thus, a model taking
into account both porosity and macroinhomogeneity was developed as an improvement
of the Paasch et al. model.

We consider a polymer layer with thickness d to be divided by cross-sections

with surface S into N sublayers, each with thickness ∆d = d/N, as shown in the scheme
below.
4 Results 58

Fig. 38 Scheme presenting

the dividing of macroscopic
polymer layer into
elementary volumes and the
network consisting of chain
connected transmission
lines, describing the
electrical behavior of such
a system. l(x) is the
characteristic diffusion
length and S(x) the effective
interface of the elementary
volume with recession x
from electrode.
4 Results 59

The resulting polymer sublayers are assumed to be described by parameters.

functionally depending on the distance from the electrode. The total number of the
cross-sections has to be high enough, so that additional division does not effect the
value of input impedance considerably.

The electrical response of every layer is presented by a transmission line as proposed by

Paasch et al (cf. Fig. 9). It is assumed, that the polymer layers, which consist of polymer
and electrolyte, included in pores, are connected polymer to polymer and electrolyte to
electrolyte. Thus, the total impedance of such a structure corresponds to a cross-
contacted network of chain connected transmission lines. An important simplification
for the calculation of the total impedance of the considered equivalent circuit can be
reached, if the electronic or ionic conductances can be neglected, yelding an infinitely
conducting line on the corresponding side of the circuit (Fig. 10). The above made
assumption about the connection between the elementary layers as polymer-to-polymer
and electrolyte-to-electrolyte has a consequence, that the infinitely conducting line of
each element is connected to the infinitely conducting line of the next element. In this
case the measurements with cross-contacted and direct-contacted network give equal
total impedance and therefore the circuit shown in Fig. 38 can be rearranged as follows:

Fig. 39 The equivalent circuit describing an inhomogeneous in the x-direction polymer

layer, where either electronic or ionic resistance can be neglected. Zt = ∞ for the case
of conventional impedance measurement and Zt = 0 for measurement of a layer,
contacted on the one side.
The terminating impedance Zt can be either infinite to describe the case of conventional
impedance measurement or zero to represent the impedance of electrode-polymer-
electrode or electrode-electrolyte-electrode connected polymer layer. An expression for
the impedance of such a network can be derived by technique, common in electronic
engineering.
4 Results 60

The required input impedance on the left gate of a homogenous transmission line
is determined by the terminating impedance on it right gate and the characteristics of the
line. Now, we have a chain of transmission lines with different parameters. The input
impedance of the first one Zi(1) is the required input impedance of the whole network.
Every elementary transmission line from left to right has as terminating impedance,
Zt(n), the input impedance of the next one, Zi(n+1), which is unknown. We know only
the impedance Zt terminating the last of them. Therefore the solution for the first input
impedance can be found by "transmission" of the terminating impedance at the end of
the network through all of the homogenous transmission lines.

It can be shown [61], that a dependence of the input impedance Zi of a

homogenous transmission line of length ∆d on the terminating impedance Zt is given by

Zt Z 0 − 1 −2 γ ∆d
1+ e
Zt Z0 + 1 eqn. 26
Zi = Z 0
Zt Z 0 − 1 −2 γ ∆d
1− e
Zt Z0 + 1

ρ′
Here, γ ≡ is commonly called the propagation coefficient and
′
Z Ho

Z 0 ≡ ρ ′ Z Ho is denoted as characteristic impedance of the transmission line.

To simplify further treatment, it is expedient to introduce the so called reflection factor

on the beginning of the transmission line, ri, and the terminating reflection factor on the
end of transmission line rt.

Z i Z0 − 1 Z t Z0 − 1 eqn. 27
ri = rt =
Z i Z0 + 1 Z t Z0 + 1

The eqn. 26 can be then rewritten as

1 + rt e −2γ ∆d eqn. 28
Zi = Z 0
1 − rt e −2γ ∆d

Comparing eqn. 27 and eqn. 28 can be seen, that the input reflection factor is related to
the terminating reflection factor by the simple expression

ri = rt e −2 γ l eqn. 29
4 Results 61

Now we have to evaluate the terminating reflection factor of the nth line, through the
terminating reflection factor of the (n+1)th line. As mentioned above, Zt(n)=Zi(n+1).
Thus, if we substitute the Zi(n+1), expressed through the terminating reflection factor of
the (n+1)th line, as in eqn. 28, into eqn. 27 we get

( Z 0 ( n + 1) − Z 0 ( n )) + rt ( n + 1)e −2 γ ( n +1) ∆ d ( Z 0 ( n + 1) + Z 0 ( n )) eqn. 30

rt ( n ) =
( Z 0 ( n + 1) + Z 0 ( n )) + rt ( n + 1)e −2 γ ( n +1) ∆d ( Z 0 ( n + 1) − Z 0 ( n ))

Now through consecutive calculating rt(n) beginning from the last chain rt(N) we can
evaluate the reflection factor for the first chain and therefore the required input
impedance of the whole network. The terminating reflection factor of the last chain,
rt(N), can be directly calculated by substituting the value of terminating impedance Zt
into eqn. 27. As mentioned above, Zt= ∞ and Zt = 0 are the cases interesting for
impedance measurements on conducting polymers. The terminating reflection factors
for the last chain are correspondingly rt ∞(N) = 1 and rt 0(N) = -1

A close mathematical expression for the input impedance of the considered network can
be found for the case of small inhomogeneity over the distance ∆d. In this case an
equality of the characteristic impedances Z0 of the adjoining transmission lines can be
assumed. A substitution Z0(n) = Z0(n+1) into eqn. 30 gives:

rt ( n ) = rt ( n + 1)e −2 γ ( n +1) ∆d eqn. 31

The consecutive substitution of every rt(n+1) until end of the line gives the terminating
reflection factor of the first chain for the cases of the open-circuit and the case of the
short-circuit respectively as:

N  N 
rt ∞ (1) = exp ∑ −2 γ ( n )∆d ) rt 0(1) = − exp ∑ −2 γ ( n )∆d ) eqn. 32
 n=0   n=0 

Substitution of these reflection factors into eqn. 27 gives expressions for the input
impedance of the first chain for both cases
4 Results 62

N  N 
1 + exp ∑ −2γ ( n )∆d  1 − exp ∑ −2γ ( n )∆d 
Zi ∞ = Z 0 ( 0)  n= 0  Z = Z ( 0)  n= 0  eqn. 33
i0 0
 N
  N

1 − exp ∑ −2γ ( n )∆d  1 + exp ∑ −2γ ( n )∆d 
 n= 0   n= 0 

This expressions can be rewritten using hyperbolic functions as

N  N 
Zi ∞ = Z 0 ( 0) coth ∑ −2γ ( n )∆d  Zi ∞ = Z 0 ( 0) tanh ∑ −2γ ( n )∆d  eqn. 34
 n= 0   n= 0 

It can be seen, that this expressions go over into the well known expressions for the
impedance of the homogeneous transmission line (cf. eqn. 9) if all the γ(n) are equal.
The ability of this approximate solution to describe the behavior of real conducting
polymer layers will be discussed further in this work.

The proposed mathematical solution can correctly describe a network with any
arbitrarily set parameters for all sublayers. However, for a physically relevant model of
conducting polymer layers, only smooth dependence of the parameters on the distance
to the electrode can be considered. Using a function describing the expected
dependence and fitting its parameters instead of the parameters of each sublayer can
greatly improve the efficiency of the fitting procedure. The choice of such function can
be different for every special case. A function, further refered as inhomogeneity
function, can be used to describe the uniform decrease of sublayer parameters with the
distance from electrode.

1 − ek(d−x)I
p( x ) = p 0 eqn. 35
1− ek d

Here x is the distance from the electrode and d the thickness of the polymer layer. The
parameter value change from p0 (scale factor) to zero over the interval d. The product
k.d can be treated as a form factor, which controls the decrease of the parameter. Its
change from -∞ to ∞ causes the change of the parameter dependence on the distance as
shown in Fig. 40.
4 Results 63

Fig. 40 Parameter dependence on distance x from the electrode as described by the

inhomogeneity function for different values of the form factor. Values of p0=1 and d=10
were used to simulate p(x) using eqn. 35
It should be mentioned, that performing numerical calculations of reflection factor for a
value of parameter equal to zero is unreasonable and hence the calculation for the Nth
layer was not carried out. The practical consequence of this approach is the absence of
inhomogeneity for parameters having a form factor < -1000.

4.4.2 Experimental results and validation of the model

The model described above was applied to impedance spectra measured on
polybithiophe layers in different oxidation state, aiming to investigate the character of
their inhomogeneity. The impedance measurements were performed in steady state
conditions on polybithiophene layers, electrodeposited by total charge flow of 80
mC.cm-2, in an electrolyte solution, free from monomer. The resulting impedance
spectra are presented below as Bode plots.
4 Results 64

Fig. 41 A series of impedance spectra of polybithiophene layer at different potentials in

steady state (a) real part, (b) imaginary part vs. frequency. Solid lines corresponds to
CNLS-fit of experimental data by use of eqn. 30 with N=50
The critical number of cross-sections needed to provide the condition, mentioned above,
that the simulated impedance spectra should not further change with increasing N, must
be estimated experimentally for every particular set of parameters. In general, the strong
inhomogeneity necessitates a large number of cross-sections. In our case this conditions
was fulfilled by N = 50 even for cases of strong inhomogeneity. As can be seen in Fig.
42 below, the difference between the spectra calculated with N more than 20 becomes
negligibly small for the set of parameters describing the layer of a reduced polymer,
where the inhomogeneity of the resistance reaches its maximal value.
4 Results 65

Fig. 42 Impedance spectra simulated by use of eqn. 30 for different numbers of cross-
sections, N. The set of parameters, optimized to describe the experimental spectra
measured by 0.5 V has been used for the simulation.
Using the approximate solution for the impedance of inhomogeneous polymer layer,
given by eqn. 34, could considerably simplify the numerical optimization, carried out to
fit the experimental data. However, the justification for use of this equation is based on
the assumption of small inhomogeneity over the length of one elementary section and is
dependent on the particular set of parameters. A comparison of impedance spectra
simulated with both the strict (eqn. 30) and approximate (eqn. 34) solution using same
parameters calculated from the experemental data is shown in Fig. 43.
4 Results 66

Fig. 43 Simulated impedance spectra using the parameters, optimized to fit to

experimental data measured at electrode potential 0.9, 0.65 and 0.5 V. Lines -
simulation by use of the approximate solution (eqn. 8) and markers - simulation by use
of strict solution (eqn.5)
The set of parameters were optimized to describe the experimental data at different
electrode potentials by use of the strict solution (eqn. 30). The obtained set of
parameters was then used to compute the spectra by the approximate solution (eqn. 34).
It can be seen, that the obtained spectra are in a good agreement with the exact
calculated spectra only in the case of well oxidized polymer (0.9 V), where the
inhomogeneity, required to describe the experimental data, is small. Therefore, the
analysis of the experimental data was further performed by use of the strict solution for
the impedance at all electrode potentials.
4 Results 67

The number of the parameters to be optimized could be reduced by use of the values for
dE/dc, D and Rct estimated in 4.3.2 for thin polymer layers, which can be considered as
non porous and homogeneous. These values were used as fixed parameters to fit the
impedance spectrum measured on oxidized layer (0.9 V) and allowed evaluation of the
polymer layer thickness as well as of the scale factors for the effective surface
coefficient f0 and characteristic diffusion length l0 by fixed form factors kfd and kld.

The inhomogeneity of l and f parameters is characteristic for the structure of any

particular polymer layer and therefore should be independent on the potential. This
consideration allows to reduce the number of optimized parameters by assuming
constant values for the form factors for l and f for all oxidation levels. The
inhomogeneity functions satisfying this criterion were estimated by fit of series of
experimental impedance spectra, as shown in Fig. 41. The curves representing the
dependence of the specific diffusion length l and efective surface coefficient f on the
distance from electrode are presented below.
4 Results 68

Fig. 44 Inhomogeneity profiles of characteristic diffusion length l (a) and effective

surface coefficient f (b), which allow good fits of impedance spectra of polybithiophene
layers for all oxidation potentials.
In contrast to the structural parameters l and f, the inhomogeneity of the specific
conductivity can increase during prolonged reduction due to slow relaxation processes
in polymer. Thus, the form factor of this parameter was set free in the fit of the
experimental impedance spectra at different potentials. The resulting inhomogeneity
profiles are shown in Fig. 45
4 Results 69

Fig. 45 Inhomogeneity profiles of the specific conductivity of polybithiophene layer at

different electrode potentials.
The double layer capacity and charge transfer resistance, related to the effective
interface of polymer, are presented in Fig. 46a and 46b, respectively. The change of
dE/dc with electrode potential can be seen in Fig. 46c.
Fig. 46 Charge transfer
resistance, Rct, (a), double
layer capacity, Cdl, (b) and
dE/dc (c), evaluated by fit
of series of impedance
spectra, measured at
different potentials, using
eqn. 30.
4 Results 70

The form factor of the inhomogeneity function of the resistance was found to change
much more significantly with the reduction of polymer then the scale factor, in this case
the resistance of the first layer. Thus, aiming to improve the efficiency of the fitting
procedure, the resistance of the first layer was assumed to be constant during the first
stages of reduction and was fixed at 2.4.106 Ohm.cm-1. The apparent specific resistance
of the whole layer, estimated by integration of the inhomogeneous resistance and
dividing by layer thickness, is shown in Fig. 47a vs. electrode potential. The apparent
diffusion constant of dopant in the polymer particles is presented in Fig. 47b.
Fig. 47 Apparent specific
resistance of the polymer
layer, ρ, (a), and the
apparent diffusion
coefficients of the dopant
ions, D, (b), vs. applied
electrode potential
5 Discussion 71

5 Discussion

5.1 The contribution of faradaic currents in the oxidation/reduction of

polybithiophene and polypyrrole during cyclic voltammetry
experements.
The combined analysis of CVs and impedance spectra, performed by Tanguy et al
[9,25], has demonstrated the ability to distinguish between different processes,
determining the redox behavior of conducting polymers. Their results are based on
impedance measurements, performed in steady state. The results of our measurements
have shown, however, that the impedance spectra are dependent not only on potential
but also on time, due to the slow relaxation processes observed during oxidation and
reduction. Thus, scan rate dependence should be expected in cyclic voltammetry
measurements. The application of FFT-impedance spectroscopy allows impedance
measurements during the potential sweep, so the measured impedance spectra
correspond to the actual state of polymer layer and can be used for simulation of CVs.

Several recent reports have dealt with interpretation of the impedance spectra of
polymer films deposited on metal electrodes [6,27,liv,62-66]. Some authors [6,62,63]
apply the model of Ho et al. [48], assuming that at high frequencies charge transfer
dominates, leading to a semicircle in the complex impedance diagram, at intermediate
frequencies the diffusion of ions determines the impedance leading to a Warburg-type
behavior and at low frequencies to a highly capacitive behavior, due to the finite
thickness of the polymer layer. Others [27,liv,64-66] have applied a transmission line
model to take account for the porosity of the polymer layer.

The high frequency impedance behavior of an electrochemical cell does not influence
its current-voltage characteristic at moderate scan rates and in the case of slow CVs,
only the low frequency range must be considered. Thus, a simple equivalent circuit,
represented by a resistor, R, and a series capacitor, C, can be justified experimentally by
the observed vertical segment in the impedance spectra of oxidized polymer electrodes
- cf. Fig. 15 and Fig. 16. This simplification results also from the model of Ho et al (see
5 Discussion 72

eqn. 7) and the transmission line based model of Paasch et al (eqn. 10). Here, the
quantities R and C have been only used for prediction of CVs using the results obtained
from in situ EIS measurements on PPy and PBt covered electrodes. The actual physical
parameters included in this apparent impedance elements have been analyzed by fit of
the impedance spectra in the entire accessible frequency range and will be discussed in
following sections.

The important question how to distinguish between faradaic and capacitive processes
during a potential sweep experiment has been repeatedly discussed in the literature [21,
70-69]. Most of the authors emphasize the inseparability of the faradaic and the
capacitive current components by electrical measurements. However, the definition of
the notion "capacitive" is often not well determined, which additionally complicates the
discussion. In this work the capacitive part of the current is defined formally as one of a
serial capacitor as revealed from the impedance spectra. The actual meaning of C can be
derived only from a physical model, for instance the one proposed by Ho et al (eqn. 7).
With regard to this model, by subtraction of the current resulting from this capacity we
actually exclude processes connected with the establishment of a chemical equilibrium
between different redox species of conducting polymer and the recharge of the double
layer. The remaining current, referred further on as faradaic current, can be due to
overoxidation of polymer, solid state polymerization or reducing of some species,
present in the electrolyte in small amount.

5.1.1 Polybithiophene.
The impedance spectra measured in situ for PBt covered electrode can be, to a good
approximation, described by the assumed series connection of R and C in the potential
ranges 0.6 V to 1.0 V in the anodic sweep and 1.0 V to 0.3 V in the cathodic sweep - cf.
Fig. 15. The slope of the low frequency spectra segments is between 75o and 85o which
supports the suggested simple equivalent circuit, applied for low frequencies. The
impedance spectra obtained during an anodic sweep in the potential range 0.3 - 0.5 V
can not be approximated by an R-C circuit. The analysis of such spectra can only be
correctly performed taking into account the inhomogeneity of a polymer layer, as
discussed in chapter 5.4. For the purpose of simulating CVs, however, the exact
5 Discussion 73

determination of R and C in this region of potentials is not a matter of concern. The

current in this potential range is low and hardly measurable, therefore no reliable
comparison with the experimental data can be performed anyway.

As shown in Fig. 17, R changes significantly with the onset of oxidation during the
anodic sweep, however, it changes only little during the cathodic sweep, even at
reductive potentials. Similarly, the build-up of C can not be completed during the
anodic sweep and continues to increase further during the reverse sweep, until the
reduction becomes dominant at lower potentials (cf. Fig. 17). Such an effect has been
observed even for steady state impedance measurements by Tanguy et al. [25]. They
ascribe this phenomenon to a profound structural change of the polymer during
oxidation. According to Heinze et al. [70] such structural relaxation processes should
accompany the oxidation and reduction of conjugated materials, such as PPy and
polythiophene. The delayed increase of the capacitance depends further on the sweep
rate and this underlines once more the necessity of in situ impedance measurements
during potential sweeps for the purpose of simulating cyclic voltammograms.
According to eqn. 24, usually applied for calculation of the capacitive current during a
potential sweep, it should be expected that the most significant capacitive effects on the
current-voltage characteristics of a polymer film occur in the potential range where the
capacitance has its highest values. A curve calculated using eqn. 24 is presented in Fig.
19b for comparison. However, as can be seen from the more general eqn. 23, the
capacitive current should strongly depend on the derivative dC/dU and not only on the
value of C(U). In cases where the capacitance of the sample significantly changes with
the electrode potential (cf. Fig. 17), the capacitive part of the current can not be
calculated using the simple eqn. 23. At a first look, the numerically computed capacitive
current as depicted in Fig. 19a, seems to have a very strange potential dependence. High
capacitive current flows at the onset of the oxidation of the polymer, an effect that is
related to the increase of the sample capacitance in this potential range. It seems that the
main contribution to the total current flow during the first stages of the oxidation and
reduction is of capacitive nature as can be seen from the comparison of curves (a) and
(c) in Fig. 17. At higher anodic potentials, 0.6 - 0.8 V, the capacitive current decreases,
approaching Ia=C(U).υ, because the capacitance of the polymer does not change
5 Discussion 74

significantly in the range 0.9 - 1.0 V (cf. Fig. 17). After the sweep reversal, however,
the capacitive current cannot approach the values predicted by eqn. 23, because of the
further increase of the capacitance during the first half of the sweep, but remains low
and even decreases further.

5.1.2 Polypyrrole.
EIS measurements of PPy covered electrodes were performed in situ in the potential
range -0.4 V to 0.7 V in which the obtained impedance spectra can be described by the
assumed series connection of R and C for low frequencies. The potential dependence of
C and R is less pronounced, as compared with the PBt results and the capacitive current
calculated from eqn. 23 shows a similar pattern as the curve obtained by using eqn. 24
instead, as can be seen in Fig.6. The faradaic current, obtained by subtraction of the
calculated capacitive part, exhibits evidently different behavior in the two potential
regions. It remains constant in the area under 0.4 V and begins to increase significantly
above this potential. The current in the first potential region could be accounted for the
solid state polymerization, which is imaginable to be connected with significant spatial
difficulties and therefore not limited by the charge transfer (and potential). The faradaic
current has an onset at 0.4 V and can either result from oxidation of some species
present in the electrolyte or from overoxidation of the polymer.

5.2 Contribution of electronic and ionic resistance to the resistive

hindrance of the recharge processes of polybithiophene
The impedance measurements with the twin electrode, presented in Fig. 12, were
performed aiming to obtain the value of electronic conductivity of polymer layer
independent on its ionic conductivity. The measurements, performed during anodic
potential sweep in monomer-free electrolyte enabled the determination of the polymer
layer impedance for different stages of oxidation. It was observed (cf. Fig. 23), that
even in its highly conductive state the polymer layer exhibits complex impedance
behavior. The impedance has however no imaginary components in the frequency range
below 10 Hz and presents therefore pure electronic resistance of polymer. As it could be
5 Discussion 75

expected, it decreases strongly during the process of oxidation, exhibiting a change of

more than two orders of magnitude, as can be seen in Fig. 25.

The impedance spectrum shown in Fig. 26, obtained in a conventional three-electrode

EIS measurement is typical for conducting polymers in the oxidized state and is well
known from literature. The measurement was performed on the same sample and with
no changes in electrochemical cell configuration, except that the two twin-working
electrodes were externally connected with a wire. It can be seen that the extreme
capacitive and blocking behavior of the polymer/electrolyte interface determines almost
entirely the impedance spectrum of the sample. The influence of the polymer resistance
can be seen on a slightly depressed character of the semicircle in the high frequency
range, but is hardly to determine quantitatively. The measurements performed with a
twin-working electrode allows, on the contrary, an exact evaluation of the electronic
conductivity of polymer along with the information about the processes on the polymer
interface.

During polymerization at a controlled anodic potential, the polybithiophene

layer grows in its oxidized and conducting state [71,72]. The total electric charge passed
through the cell is related to the number of monomer units oxidized and deposited as
polymer layer on the electrodes and to the charge used in the process of the polymer
oxidation. Thus, although indirectly, the curve presented in Fig. 27 characterizes the
formation of an oxidized polymer layer during polymerization. The growth rate
evidently increases with time due to the polymer layer porosity and the increase in
active surface area.

In situ EIS measurements between the bands of the twin electrode has been performed
during polymerization as marked with the numbered points. It can be seen from the
obtained impedance spectra (cf. Fig. 28) that the gap of the twin- working electrode was
already bridged with polymer before the first impedance spectrum has been measured,
i.e. in less than 40 s polymerization time. A complete interpretation of the obtained
impedance spectra can be given using an approach described in the chapter 4.4.1. and
the knowledge of the field distribution between the bands of the twin-electrode.
However, such a distribution for the case of the inhomogenous polymer layer is rather
complicated and still remain a problem to be solved.
5 Discussion 76

Some general features of the impedance spectra are evident, though. The Ohmic
behavior for low frequencies, as seen in Fig. 28, can be attributed to electronic
conduction through the polymer aggregates. Low frequency resistance data, obtained
from Fig. 28b, are compiled in Fig. 29 to represent the dependence of the d.c. resistance
on the polymerization time. Evaluation of specific resistance values, however, depends
on particular electrode configuration. Model calculations, based on the electric field
distribution between the two metal stripes of the twin-working electrode, were
performed to gain insight into the dependence of the specific conductance between the
bands of the twin electrode on the layer thickness. The assumption of a linear
dependence of the layer thickness on the total charge used for polymerization had to be
used to evaluate the values of the specific conductivity and thickness of the layer. The
value of the specific electronic conductivity of the polymer has been found to be much
higher as the conductivity determined from conventional impedance measurements,
where ionic resistance has a significant effect. The analysis of the inhomogeneity of
even much thinner polymer layers, presented in chapter 4.4.2, let imagine, however, that
the assumption of the homogenous polymer layer, used in the the calculation of the
specific resistance, does not satisfactorily match the real polymer structure. A field
calculation taking into account the inhomogeneity of the polymer layer can be proposed
as further step in the development of this investigation method.

5.3 Evaluation of kinetic relevant parameters of conducting polymers

by analysis of layers with different thickness.
The formation of conducting polymer layers during the electropolymerization has been
intensively investigated in the last years [14,32-38]. However, the potentiostatic
measurements, commonly used for this purpose, do not allow to observe closely the first
stages of the layer growth, due to the nonlinear dependence of the formation rate on
potential at the onset of deposition. Thus, impedance measurements were performed
during galvanostatic polymerization of bithiophene at low enough current density of
0.15 mA.cm-2. The absence of distortions of the potential response, due to changes in
the polymer layer during the EIS- measurement were checked by analysis of obtained
power spectra using an approach, described in [l]. Separation of two stages of layer
5 Discussion 77

growth was possible in the time scale of the polymerization experiment using this
current density, as obtained data have shown.

The use of the Macdonald's CNLS program for fitting the experimental data allowed
comprehensive analysis of the fit quality. The relative standard deviations of parameters
supplied by the optimization program are representative for the goodness of the entir fit
as well as for the influence of each parameter on the function value. All of the above
presented parameters had average relative standard deviations bellow 10%, whereas a
random distribution of the deviations for each frequency point has been detected. This is
a confirmation for a correct separation of parameters during the fit of the particular
experimental data by the used model. The comparison of the fit curves in Nuquist and
Bode presentation with the experimental data also verifies the reliable choice of fitting
functions.

The application of the model of Ho et al. was performed, assuming thin polymer layers
nonporous and homogeneous. However, the apparent diffusion constant of the
counterions depends on the film thickness at the initial stages of polymerization (Fig.
34b), what can correspond to 3-D growth of polymer nuclei rather than existence of
homogenous polymer film, assumed by the model. For polymerization charge > 20
mC.cm-2, the apparent diffusion constant becomes practically invariable and so the
requirements for the use of Ho et al. model become fulfilled. That can be a consequence
of a crossover to 2-D growth with a formation of homogenous polymer layer. The
estimated diffusion constant for this region amounts to 1.8 . 10-8 s-1 . cm2 . The charge
transfer resistance (Fig. 34a) does not change significantly during the polymerization,
indicating a constancy of the interface available for oxidation/reduction reaction of the
polymer. The exchange current density corresponding to the Rct value of the thickest
film was estimated to be 4 mA.cm-2 . The capacitance of the double layer obtained from
the fits increases linearly with the film growth (Fig. 35b). This result shows, that the
inner regions of polymer film are possibly also involved in the formation of double
layer. The approximation of the obtained linear dependence to 0 polymerization charge,
where the interface can be assumed equal to the geometrical, yields for the specific
double layer capacitance the value of 2.4 µF. cm-2 .
5 Discussion 78

Increasing of the electrode interface covered with polymer or a lower activation energy
for electron transfer from the monomer result in a decrease of the charge transfer
resistance during the first stages of film growth (Fig. 35c). The latter can be due to the
fact that the already deposited film is better substrate for polymer precipitation than
Platinum. After a formation of a compact polymer layer the charge transfer resistance
becomes nearly constant.

The use of the impedance equations proposed by Ho et al. and Paasch et al. in
the form, including the structure independent layer parameter dE/dc, allows to perform
a combined analysis of both thin and thick polymer layers. Independence of this
parameter on the layer thickness requires only constancy of the polymer density inside
of elementary polymer aggregates during the polymer growth. Such an assumption can
be justified in view of the results of the polymer density measurements made by
flotation method [73].

During the first stages of polymerization the characteristic diffusion length,

resulting from fit of experimental data (Fig. 36a), changes with the film thickness,
which does not prove the assumptions made in the model of Paasch et al. After
deposition of polymer of up to 180 mC. cm-2 polymerization charge the characteristic
diffusion length becomes constant. Similar behavior was reported by Tanguy et al. for
methylthiophene [9]. The estimated value for the characteristic diffusion length (300
nm) let us conclude, that fibriles of polymer, for which similar dimensions are estimated
by SEM measurements [74], are the regions where diffusion becomes the limiting step
of charge transfer.

The effective surface coefficient increases with layer thickness until the
polymerisation charge reaches the value of 250 mC.cm-2 . The growth of the interface
due to deposition of more polymer becomes then compensated by blocking of already
accessable interface. This effect should result in macroinhomogeneities, that should be
taken into account in the analysis of thicker polymer layers. The possible model
approach for this case is discussed in the next chapter.

The estimated maximal effective surface is 4.7 times larger then the geometric
surface of electrode. This value is much smaller than values estimated by adsorption
experiments [75]. The observed difference could be explained by the assumption, that
5 Discussion 79

organic molecules can adsorb inside of the polymer conducting areas whereas the
electron transfer can occur only on their interface. The volume occupied by polymer
aggregates can be estimated by multiplication of the effective interface and
characteristic diffusion length. This volume amounts to 0.7 times the value of
geometrical volume of the film, calculated from film thickness obtained by the fit (Fig.
37a) and can serve as a characteristic for porosity of the polymer layer. The
dependence of the double layer capacity on the polymerization charge (Fig. 37b) is
nearly linear and similar to one, estimated for thin polymer films. The value of the
estimated conductivity of polymer layer (Fig. 37c), is between 0.5 and 2 . 10-5 S . cm-1.
This value is much smaller than the values of the electronic conductance characteristic
for polybithiophene and can be accounted for ionic conductance of the polymer layer.
However, the effect of the inhomogeneity was not considered in the calculation of this
value, and it should be thus interpreted as an average through the entire film thickness.
Its growth can therefore be interpreted as a sign for growing porosity of the film.

5.4 Influence of inhomogeneity on the electrochemical behavior of

conducting polymer layers.
Many authors [25,64 -80] have pointed out the necessity of accounting the
inhomogeneity effects by analysis of the impedance spectra of conducting polymers.
However, only approximate models, like one with constant phase elements (CPE)
[lx,81] or with diffusion hindrance, calculated with different from 1/2 value of α
[78,79], not allowing quantitativ characterization of inhomogeneity, were proposed up
to date. The application of the approach of reflection factors, common in electronic
ingineering, allowed to simplify the mathematical treatment considerably and to
evaluate an exact solution for inhomogeneous transmission line, which takes into
account the dependence of all parameters on the distance from the electrode.

The proposed solution is valid only for cases, for which either the ionic or the electronic
resistance is negligible. This condition is fulfilled for the electrodeposited layers of
polybithiophene, as can be seen comparing impedance spectra measured on the blank Pt
electrode and on the same electrode after the deposition of a polymer layer. The real
part of the impedance high frequency limit does not change after the polymer deposition
5 Discussion 80

and therefore represent only the resistance between the tip of the Luggin capillary tube
and the polymere interface, not including the resistance of the polymer. Test
calculations using the expression for the impedance of a transmission line under
consideration of both ionic and electronic resistances [27], having comparable values,
show however, that the real part of the impedance of such system does not become zero
at high frequencies end and therefore should contribute to the entire high frequency
resistance. The absence of such contribution justifies the assumed simplification at
least for the case of slightly reduced polymer layers, analyzed in this work.

The choice of a function to describe, how particular parameters depend on the distance
from electrode should be made in accordance to the expected character of
inhomogeneity. For the case of galvanostatically deposited polybithiophene, the ESM
images [30] show a macroscopic structure consisting of conic agglomerates with the
base on the electrode and the tip directed into the electrolyte solution, as shown in Fig.
38 in diagram form. The specific effective surface, characteristic diffusion length and
specific conductivity should therefore decrease with the distance from the electrode and
become zero at the distance equal to the layer thickness. The proposed inhomogeneity
function (eqn. 35) exhibits the required behavior over the distance range from x = 0 to
x = d. Its form undergoes a wide range of possible alterations by change of only two
parameters, having a physical meaning as the scale factor (the value of the function at
x=0) and the form factor, controlling the character of decreasing, as in Fig. 40. The use
of this function allowed good description of experimental data, as can be seen in Fig.
41.

The number of cross-sections, N, required to provide a quasi-infinite dividing of

polymer layer has been estimated as 50 for the case of polybithiophene in slightly
reduced state, cf. Fig. 42. The computing time for optimization of parameters describing
the impedance spectra are direct proportional to the number of the layers but remain still
acceptable for N = 50, when using an optimized for speed CNLS fitting program
(LEVM) on the IBM RISC 6000 workstation.

It has been found that the possible simplification of calculations using approximate
solution (eqn. 34) is unusable for fitting the impedance spectra for inhomogeneities,
taking place in reduced polybithiophene layers (cf. Fig. 43). It should be mentioned,
5 Discussion 81

that some approximate expressions for the impedance of inhomogeneous conducting

polymers, described in the literature [76,79], are based on similar assumptions of low
inhomogeneity and so their applicability to polymer layers, prepared by common
electrodeposition techniques are questionable and should be carefully considered in
view of the results obtained using the exact solution.

The proposed mathematical model can correctly describe an inhomogeneous system

with all parameters depending on the distance. However, the number of parameters,
which can be considered to be inhomogeneous in a polymer layer, is restricted by the
applied physical model. For instance, if the effective surface coefficient f is considered
inhomogeneous, the interface related parameters, such as specific capacitance Cdl and
charge transfer resistance Rct, have to be independent on the distance from the electrode.
Analogous, the apparent diffusion coefficient D and the thermodynamic parameter
dE/dc have to be independent on distance in case of an assumption of inhomogeneous
characteristic diffusion length l. Hence, only the inhomogeneities of the effective
interface coefficient f, the characteristic diffusion length l and the specific polymer
resistance ρ have to be analyzed.

The inhomogeneity profiles of the specific effective interface coefficient and the
characteristic diffusion length, which allow good fit of the experimental data at all
electrode potentials, are presented in Fig. 44. The observed character of the distance
dependencies are in a good aggreement with the assumed granular macrostructure of the
polymer layer. The decrease of the characteristic diffusion length is followed by slower
decrease of the specific effective surface. The possible interpretation of this
phenomenon is that polymer agglomerates become smaller with the increasing distance
from the electrode and simultaneously diminish its volume fraction, which results the
corresponding reducing of the specific interface as the summary effect.

The nature of the resistance, determining the charge/discharge kinetics of the polymer
layer, is still not well understood. However, the character of the dependence of this
resistance on the distance to electrode (Fig. 45) let assume, that the electronic resistance
of polymer agglomerates is the reason for distribution of relaxation times. As can be
expected, increasing with the distance porosity and decreasing volume fraction of
polymer agglomerates yield increasing resistance of the layer. If the ionic resistance
5 Discussion 82

would be the reason for the distribution, it should be decreasing with the distance from
the electrode.

The charge transfer resistance of polybithiophene, evaluated from the fit, almost
linearly decreases with the electrode potential in the range between 0.9- 0.45 V, as in
Fig. 46a. The potential dependence of this quantity is evidently non Nernstian, which
indicates the presence of a complex electrochemical equilibrium on the interface of
polybithiophene. The potential dependence of the partial derivative (dE/dc)E, denoted in
Fig. 46c, is non Nernstian either, in accordance with the results of coulometric
investigations reported in [82,83]. Accounting of the Donnan equilibrium and the
interaction between the ions, incorporated into polymer matrix, could possibly allow to
describe the observed potential dependence.

The double layer capacity decreases slightly with the potential (Fig. 46b). The decrease
of the concentration of doping ions on the interface with reducing the polymer could be
the possible reason for this effect. The apparent diffusion coefficient D also exhibits a
slight decrease with the electrode potential, as in Fig. 47b. This fact confirms the
observed by many authors [37,38] decrease of the volume of the polymer during
reduction, which respectively yields more compact polymer agglomerates in reduced
state.
6 Conclusion 83

6 Conclusion
In this work the processes involved in recharge of conducting polymers have been
investigated by means of various electrical measurement techniques. Conventional
methods (cyclic voltammetry, coulometry, potentiometry, potential and current pulse,
EIS), EIS on twin electrode and EIS under galvanostatical conditions were applied. The
two latter investigation methods were developed in our working group and their
application and analysis were performed for the first time during this work.

Two recently published models, connecting the kinetic and thermodynamic parameters
relevant to the recharge processes of the polymer layer with its electrical properties,
have been used to analyze the experimental data and their applicability for different
kind of polymer layers has been discussed.

To account for the inhomogeneity of thick polymer layers, which has been established
in some cases, the model of Paasch has been generalized using an approach of an
inhomogenous transmission line and was applied to analyze the impedance spectra of
polymer layers at different oxidation potentials.

• The impedance measurements during the cyclic recharging of polybithiophene and

polypyrrole layers show considerable hysteresis in the potential dependence of the
pseudocapacity and the resistance. The comparison of calculated CVs with
experimentally measured ones allowed the estimation of the faradaic current for high
anodic potentials, which was ascribed for solid state polymerisation and/or
overoxidation of polymer.

• In order to estimate the electronic resistance of the polymer separately, impedance

measurements on a polymer layer, contacted only on the polymer side, have been
carried out. Using an analytical solution for the electrical field in the used
twin-electrode configuration, specific electronic conductivity and thickness of the layer
could be estimated.

• An analysis of the impedance spectra obtained during the first stages of

polybithiophene deposition allowed to observe a transition from a granular 3-D to the
6 Conclusion 84

2-D growth of the polymer film. An application of the Ho et al. model for layers of
intermediate thickness, prepared at low current densities, could be justified. Polymer
layers produced by higher current density has been found to be porous and therefore
well matching the considerations of the 2-phase model proposed by Paasch at el. A
combined analysis of thick and thin polymer layers assuming the constancy of their
thermodynamic properties allowed to estimate the effective interface and the average
dimensions of the polymer agglomerates. An increasing inhomogeneity has been
observed for polymer layers with growing thickness.

• A model has been developed in order to describe macroinhomogenous polymer

layers. The assumption of the polymer as a 2-phase medium with structural parameters,
differing in the direction perpendicular to the electrode interface, has been realized
using the approach of an inhomogeneous transmission line. Good correspondence of the
proposed model to the impedance spectra of polybithiophen, measured at different
electrode potentials, has been observed. CNLS fit of experimental data with the model
function allowed to estimate thermodynamic and kinetic parameters of the polymer,
along with the information about the structure of polymer layer.
7 Acknowledgments 85

7 Acknowledgments
This doctoral thesis is a product of three years I spent inside of a lively and creative
working group, which made a considerable contribution to my development as a
scientist as well as to my acclimatizing in unusual ethnic surroundings. This group, as I
have found it, is a result of long scientific and organizing activity of Prof.
R.N.Schindler, whom I am indebted for his credit of trust to invite me for carrying out
my Ph.D. research and for an effective help in every scientific and private affairs, which
always came in time.

Helpless in the foreign country, I stepped into the office of Dr. Popkirov one
good day of 1990, and to my great surprise and relieve conceived the words of my
mother language. Since that, George steadily assisted me in all this small practical and
theoretical problems, usual in the scientific work, actually forming my attitude to the
electrochemistry as to an exact science. I also spent uncountable hours with the unique
impedance experimental equipment, which he created. It always allowed to realize even
my most wild ideas. And this long discussions with him, where we always came to the
consensus....I am greatly thankful to him for all this!

Always there with his good advise was Dr. Alexander Kukui, who considerably
influenced the formation of my taste to the mathematical analysis and to scientific work
in generally. Each time, he found my great mathematical problems "quite simple".

To all my colleagues, who created this indescribable atmosphere of the helpful

cheerfulness, I would like to extend my warm thanks. Axel Müller, Mathias Burmeister,
Martin Olbrich-Stock, Thorsten Benter, Volker Sauerland, Martin Liesner, Ulf Kirchner
- all you made me feel at home at the Institute and helped me to find my way through all
imaginable small and big problems. Uwe Eggers and Michael Karstens were
responsible for solving the technical ones, which freighted me most. Mr G.Hammerich
was the creator of the collection of beautiful glassy measurement equipment, which
includes parts I was sure were impossible to prepare. Mr A.Kracht made his
contrubution to the electronical equipment I could not dispence in my work.
7 Acknowledgments 86

The hours I did not spent in the Institute during this years was blessed by the
sympathy of Rex and Leslie Slate, Lourdes Lemos, the family of Kukui and Sudesha,
who made my love to Kiel even deeper. The love of my parents accompanied me since I
know me, along with the effective support, which was always as understoodly as the air
I breath. C¯ á¨¡®!

Last but not least I would like to speak out my deep thankfulness to the Gottlieb-
Daimler-und-Karl-Benz Stiftung and its staff not only for making possible my research
stay in Germany, but also for the feeling to be a part of a big family they gave each
participant.
7 Acknowledgments 87

Referent: ...Dr. Georgy Popkirov..................................................................................

Korreferent: ...Prof. Schindler.............................................................................

Tag der mündlichen Prüfung: ....6/7/1996..................................................

Zum Druck genehmigt: Kiel, den ..10/7/1996..............................................

.................................................

Dekan
7 Acknowledgments 88

Diese Dissertation wurde mit der freundlichen Erlaubnis des Dekans der
Mathematisch- Naturwissenschaftlichen Fakultät, Prof. Dr. R. Haensel, in englischer
Sprache angefärtigt. Dies wurde nicht durch meine mangelnde Zuneigung zur deutscher
Sprache verursacht, sondern vielmehr durch meinen Wunsch, die Ergebnisse meiner
Arbeit in Kiel auch meinen russischen und ukrainischen Kollegen zur Verfügung stellen
zu können.
8 References 89

8 References

1. H.Shirakawa, E.J. Louis, A.G. MacDiarmid, C.K. Chiang, A.J. Heeger

J.Chem.Soc.,Chem.Commum., (1977) 578

2. A. F. Diaz, K. Kanazawa, G. Gardini, J. Chem. Soc. Chem. Commun., 14, (1979)

635

3. A. G. MacDiarmid, S.-L. Mu., N. L. Somasiri, and W. Wu, Mol. Cryst. Liq. Cryst.,

121, (1985) 187

4. T. Kobayashi, H. Yoneyama, and H. Tamura, J. Electroanal. Chem. 161, (1984), 187

5. E. W. Paul, A. J. Ricco, and M. S. Wrighton, J. Phys. Chem., 89, (1985) 1441

6. T.B.Hunter, P.S.Tyler, W.H.Smyrl and H.S.White, J.Electrochem.Soc., 134 (1987)

2198.

7. F.Beck and P.Hulser, J.Electroanal.Chem., 280 (1990) 159.

8. M.M.Musiani, Electrochim.Acta, 35 (1990) 1665.

9. J.Tanguy, J.L.Baudoin, F.Chao and M.Costa, Electrochim.Acta, 37 (1992) 1417.

10. R.John and G.G.Wallace, J.Electroanal.Chem., 354 (1993) 145.

11. R.Holze and J.Lippe, Synth.Met., 38 (1990) 99.

12. J.Kruszka, M.Nechtschein and C.Santier, Rev.Sci.Instrum., 62 (1991) 695.

13. R.Bilger and J.Heinze, Synth.Metals, 43 (1991) 2893.

14. A.R.Hillman and M.J.Swann, Electrochim.Acta, 33 (1988) 1303

15. C.Arbizzani and M.Mastragostino, Electrochim.Acta, 35 (1990) 251.

16. G.Morea, C.Malitesta, L.Sabbatini and P.G.Zambonin, J.Chem. Soc.Faraday Trans.,

86 (1990) 3769.
8 References 90

17. G.G. Pickup and R.A. Osteryoung, J. Electroanal. Chem., 195 (1985) 271

18. N. Mermiliod, J. Tanguy and F. Petiot, J. Electrochem. Soc.,133 (1986) 1073

19. F. Beck and M. Oberst, Makromol. Chem. Makromol. Symp., 8 (1987) 97

20. S. Panero, P. Prosperi, S. Passerini and B. Scrosati, J. Electrochem. Soc., 136 (1989)
3729

21. S.W.Feldberg, J.Am.Chem.Soc., 106 (1984) 4671

22. T. Yeu, T.V. Nguyen and R.E. White, J.Electrochem.Soc., 135 (1988) 1971

23. T. Yeu, K.M. Yin, J. Carbajal and R.E. White, J. Electrochem.Soc., 135 (1991)
2869

24. D.A. Kaplin and S. Qutubuddin, J. Electrochem. Soc., 140 (1993) 3185

25. J. Tanguy, N. Mermilliod and M. Hocklet, J.Electrochem.Soc., 134 (1987) 795

26. M.A.Vorotyntsev and J.P.Badiali, Electrochim.Acta, 39 (1994) 289

27. G.Paasch, K.Micka and P.Gersdorf, Electrochim.Acta, 38 (1993), 2653

28. G.Schiavon, S.Siran and G. Zotti, Synth. Met., 32, (1989) 209

29. J. Kankare and E.-L. Kupila, J. electroanal. Chem., 322, (1992) 167

30. G.Kossmehl, D.Fechler, W.Plieth, W.-F. Zhang and J.Zerbino, in Werkstoffe in der
electrochemnie, Dechema-Monographien- Band 121, 279, VCH Verlagsgeselschaft-
Weinheim, Germany (1990)

31. L.Olmedo, I. Chanteloube, A. Germain, M.Petit and E.M.Genies, Synth.Met., 30,

(1989) 159

32. A.R. Hillman and E. F. Mallen, J. Electroanal. Chem. 220 (1987) 351.

33. F.Li and W.J.Albery, Electrochim.Acta, 37 (1992) 393.

8 References 91

34. J. Heinze, K. Hinkelmann, M. Dietrich and J. Mortensen, Dechema Monographien,

102 (1986) 209.

35. P. Tschuncky, J. Heinze , Voltammetric Studies on the Electropolymerization

Mechanism of Methoxythiophenes , Synth. Met. 55 ( 1993 ) 1603.

36. G. Koßmehl, D. Fechler, W. Plieth, W.-F. Zhang, J. Zerbino , Untersuchungen

zur Struktur und Morphologie elektrisch leitenderPolymerschichten auf Polythiophen-
Basis , DECHEMA-Monogr. 121 ( 1990 ) 279

37. A. Hamnett, A. R. Hillman, J. Electrochem. Soc., 135 (1988) 2517.

38. F. Chao, M. Costa, C. Tian , Synth. Met. 53 ( 1993 ) 127.

39. G. Tourillon and F. Garnier, J. Phys. Chem., 87 (1983) 2289.

40. G. Schiavon, S. Sitran and G. Zotti, Synth. Metals, 32 (1989) 209.

41. R.Bilger and J.Heinze, Synth.Metals, 43 (1991) 2893.

42. C.K. Baker, J.R. Reynolds , Am. Chem. Soc., Div. Polym. Chem., 28 ( 1987 )
284.

43. G.S. Popkirov aund R.N. Schindler, Ber. Bunsenges.phys.Chem, 97 (1993) 479.

44. J. Roncali, F. Garnier, M. Lemaire and R. Garreau, Synth. Metals, 15 (1986) 323.

45. J. Roncali, M. Lemaire, R. Garreau and F. Garnier, Synth. Metals 18 (1987) 139.

46. B. Zinger, Y. Greenwald, I. Rubinstein, Synth. Metals 41 ( 1991 ) 583

47. F. Sundholm, G. Sundholm, M. Törrönen , Synth. Met. 53 (1993) 109.

48. C.Ho, I.D.Raistrick and R.A.Huggins, J.Electrochem.Soc., 127 (1980) 343.

xlix. G.S.Popkirov and R.N. Schindler, Rev. Sci. Instrum. 63/11 (1992) 5366

l. G.S.Popkirov and R.N. Schindler, Electrochim. Acta, 39 (1994) 2025

li. G.S.Popkirov, E.Barsoukov and R.N.Schindler, Electrochimica Acta, 40 (1995) 1857

8 References 92

lii. R.C.Larson, R.T.Iwamoto, and R.N.Adams, Anal. Chim. Acta, 25 (1961) 371

liii. J. E.McClure, D.L.Maricle, Anal. Chem., 39 (1967) 236

liv. S. Fletcher, J.Electroanal.Chem., 337 (1992) 127.

lv. K.J.Binns, P.J. Lawrenson, C.W.Trowbridge "The Analytical and Numerical

Solution of Electric and Magnetic Fields", John Wiley & Sons, 1992

lvi J. R. Macdonald, Electrochimica Acta, 35 (1990) 1483.

lvii. J.J.More in G.A.Watson (ed.), Lecture Notes in Mathemiatics: v. 630, Numerical

Analysis, Springer-Verlag, Berlin, 1978, p. 105

lviii. O.Genz, M.M.Lohrengel and J.W.Schultze, Electrochim.Acta, 39 (1994) 179.

lix. G.S.Popkirov and E.Barsoukov, J.Electroanal.Chem., 383 (1995) 155.

lx. J.Tanguy, N.Mermilliod and M.Hoclet, J.Electrochem.Soc., 134 (1987) 795.

61. O. Zinke, A.Vlcek, "Lehrbuch der Hochfrequenztechnik", Bd.1, 3., neubearb. u.

erw. Aufl. (1987), Springer-Verlag New York Heidelberg Berlin, pp. 60- 64

62. K.Naoi, K.Ueyama, T.Osaka and W.H.Smyrl, J.Electrochem.Soc. 137 (1990) 494.

63. R.M.Penner and C.R.Martin, J.Phys.Chem. 93 (1989) 984. 21. M.M.Musiani,

Electrochim.Acta, 35 (1990) 1665.

64. X.Ren and P.G.Pickup, J.Chem.Soc. Faraday Trans., 89 (1993) 321.

65. W.J.Albery, C.M.Elliott and A.R.Mount, J.Electroanal.Chem., 288 (1990) 15.

66. W.J.Albery and A.R.Mount, J.Electroanal.Chem., 288 (1990) 15.

67. J.Heinze, M.Strzbach and J.Mortensen, Ber.Bunsenges. Phys. Chem., 91 (1987)

960.

68. I.Rubinstein, E.Sabatani and J.Rishpon, J.Electrochem.Soc., 134 (1987) 3078.

69. R.S.Hutton, M.Kalaji and L.M.Peter, J.Electroanal.Chem., 270 (1989) 429.

8 References 93

70. J.Heinze, R.Bilger and K.Meerholz, Ber.Bunsenges. Phys. Chem., 92 (1988) 1266.

71. J.Heinze, "Electronic conducting polymers", in "Topics in current chemistry",

v.152, Springer-Verlag, Berlin 1990

72. G.P.Evans, "The electrochemistry of conducting polymers", in "Advances in

Electrochemical Science and Engineering",Ed.:H.Gerischer and C.W.Tobias, VCH
Verlagsgeselschaft GmbH, Weinheim (Germany) 1990.

73. M.Salmon, A.F Diaz., A.J.Logan, M. Krounbi, J. Bargon, Mol.Cryst. Liq. Cryst., 83
(1983) 1297

74. G. Tourillon, Handbook of conducting polymers, Ed. T.A.Skotheim, Marcel

Dekker, Inc., 1986, p. 293

75. D. L. Miller and J. O’M. Bockris, J. Electrochem. Soc. 139 (1992) 967.

76. G. La'ng and G. Inzelt, Electrochim. Acta, 36 (1991) 847

77. C. Gabrielli, H. Takenouti, O. Haas and A. Tsukada, J. Electroanal. Chem., 302

(1991) 59.

78. G. Inzelt, G. La'ng, J. Electroanal. Chem. 378 (1994) 39-49.

79. Z. Stoynov, Electrochim. Acta, 35 (1990) 1493

80. M. Edeling, K. Mund, W.Naschwitz, Siemens Forsch. u. Entwickl. Ber. 12 (1983),

85

81. G. La'ng, J. Bacskai and G. Inzelt, Electrochim. Acta, 38 (1993) 773

82. S. Porter, Mat. Sci. Forum. 42 (1987) 43

83. J. H. Kaufman, J. W. Kaufer, A. J. Heeger et al., Phys. Rev. B., 26 (1982) 2327