Annals"of BiomedicalEngineering, Vol. 20, pp. 363-383, 1992 0090-6964/92 $5.00 + .

00
Printed in the USA. All rights reserved. Copyright 9 1992Pergamon Press Ltd.

Impedances of Thin and Layered Systems: Cells with

Even or Odd Numbers of Interfaces

Richard P. Buck
Department of Chemistry
University of North Carolina
Chapel Hill, NC

(Received 4/18/91; Revised 10/22/91)

Impedance data, e.g., system responses, from perturbing small amplitude applied
sinusoid signals of near DC to high kilohertz frequencies, give chemical information.
Analysis of frequency-dependent imaginary and real impedance proceeds from equiv-
alent analog circuit elements to chemical and physical significance determined from
many model systems. Already, it is possible to interpret bulk transport processes, sur-
face kinetic effects, diffusion phenomena, and dependencies on the type of contacts:
symmetric ion contact, symmetric metal contact or asymmetric metal-ion interfaces,
and cell design; even (battery or sensor) and odd numbered (constrained junction or
immiscible liquid) interfaces in a system. These analyses cover the chemical origins,
locations and meanings of the lumped resistances, capacitances and transmission lines
that are introduced by engineers in their strict analog interpretations of the imped-
ance data.
Examples cover simple ohmic, simple diffusive behavior, complex behavior with
surface interracial kinetics or surface resistances, and with finite (nonblocking) or in-
finite (blocking) DC impedance. High and low frequency responses may show so-
called constant phase element character that suggests fractal behavior. Low frequency
data occasionally appear in the second quadrant of impedance plane plots. These re-
sults are caused by negative capacitances and resistances.
In this paper, chemical interpretations of analog circuit elements are mainly based
on theory and observations of thin cells of electrolytes and solid and liquid films
(membranes) that are ionic or mixed ionic~electronic conductors. The information
should carry over into thickened, gelled, and tissue electrolyte phases and serve as a
basis for medically-oriented, perhaps diagnostic impedance measurement applications
already pioneered by Herman Schwan.
Keywords- Impedances, Time constants, Capacitances, Resistances, Donnan failure,
I- V curves, Fermi levels, Ion exchange.

INTRODUCTION
I n all o f these film systems it is the k i n d , l o c a t i o n , q u a n t i t i e s , a n d m o b i l i t i e s o f
c h a r g e d species in b o t h the film a n d c o n t a c t i n g p h a s e s t h a t u l t i m a t e l y d e t e r m i n e the
i m p e d a n c e s (11,18). T h e o p t i m a l e q u i v a l e n t circuit has a m i n i m u m n u m b e r o f ele-

Address correspondence to Richard P. Buck, Department of Chemistry, University of North Carolina,

Chapel Hill, NC 27599-3290.
363
364 R.P. Buck

ments. These will be frequency-independent R's and C's, but may include one or more
transmission line ladder network elements. The chemistry determines the origin of
these elements and they are not arbitrarily selected, except perhaps, during the ini-
tial curve-fitting analysis.

Capacitors and Capacitances of Simple,

Relatively Nondistributed Systems
It is the distribution of charge that gives rise to the fields and potential profiles
(9,10,12). The phases, initially electrically neutral, develop space charge especially at
interfaces, but the space charge spreads into the phases in decreasing amounts. It is
the fact that energies of permeable charged species at contacting electrolyte/film, elec-
trolyte/metal, or film/metal are not the same in each phase. These energy differences
lead to generation of space charge double layers. The width of these inhomogeneous
regions is variable but under 100 nm unless current is flowing. Width depends on
charged species activities and standard free energies in each phase (reckoned with re-
spect to standard cells or to vacuum). Space charge, curvature of activity profiles, and
electron energy band bending are synonymous features.
At the simplest interfaces with uniform distribution of charges everywhere except
at an interface, the capacitance is not a constant value, independent of potential dif-
ference (pd). Unfortunately, in most chemical systems, the spacing of the interfacial
capacitor "plates" is variable and depends on chemical composition and applied volt-
age. Consequently, a single constant capacitor in an equivalent circuit is only appro-
priate when the system is being described at the natural, zero-current or open circuit
voltage (or electromotive force, emf), and when the impedance is measured (as it usu-
ally is) with a small AC perturbation voltage around the DC voltage of zero current.
However, impedances need not be measured at zero current, and the impedance mea-
surements at several different DC currents can add essential information to the anal-
ysis of an unknown system. The capacitors may well change their values in each
experiment!
A further complication is another capacitor in series with the space charge capac-
itor, arising in fluid systems, by adsorption of dipolar solvent, dipolar electrolyte, or
individual ions of one sign from an electrolyte.

Resistors and Resistances in Simple Systems

Constant resistors in equivalent circuits are associated with charge transport
through regions with a uniform concentration of charge carriers. Charge motion in
a uniform electrolyte is field driven by the process "migration" or "drift" in the re-
spective electrochemical and physics literature. As long as there is no possibility for
depletion of charge carriers and, therefore, no concentration gradients, e.g., electrons
in metals and ions in fixed-site ion exchangers, then transport shows a constant re-
sistance. However in fluid ionic systems and in electron semiconductors, depletion or
accumulation regions (layers) are easily formed, especially at interfaces where charge
carriers move across an interface. Then the current flow is non-ohmic and time de-
pendent. Of course, the usual situation is one in which both diffusion down a con-
centration gradient and migration determine the flux of charge carriers to an
interface.
Impedances of Thin and Layered Systems 365

Diffusion~Migration (Warburg) Impedance

When a potential difference is applied at a single interface and a depletion or ac-
cumulation layer is formed, the flux is determined by time-dependent solutions of the
appropriate transport equation. For a common case of a high supporting electrolyte
and a low concentration of the ion passing the interface, then transport is solely by
diffusion. The current under a voltage step will be large initially and fall as t -1/2.
The circuit element is an infinite transmission line with equal R and C per unit length.
Both terminals are, of course, at the same side of the line. The risers and rungs of the
equivalent ladder network may be interchangeably R or C since no DC current is pos-
sible in either arrangement. For real, finite thickness membranes and films, the trans-
mission line is finite, terminated with a resistance, and passes current at DC (24,37).

Simple Layered Conductors

A stack of different, homogeneous, conducting phases in contact, is a series of re-
sistances and interfacial capacitances. This statement is true of a series of different
metal wires, e.g., copper/silver/gold/copper, where normally, the resistance would
be calculated simply enough. But, there are also interfacial capacitances at each phase
contact. These quantities are not often calculated. When two or more electronic semi-
conductors (of different kind or different doping levels of the same kind) are con-
tacted in series, or ion exchangers of different charge carrier concentrations are
contacted, then the system behaves as a series of parallel RC elements. At the high-
est frequencies the impedance is dominated by the least resistive layer, coupled to the
total exterior capacitance. The latter gives a single semicircle (or part of one) in the
complex plane. The next most resistive layer can pass charge only at lower frequen-
cies and so produces a second semicircle with nearly the same capacitance. When the
layers are different enough (i.e., their time constants differ by about 100, there can
be a series of semicircles. However, when the time constants approach similar values
the semicircles are merged to give a distribution of overlapping semicircles.

APPLICATION OF ELECTROCHEMICAL POTENTIALS

AND FERMI LEVELS
Local Equilibrium and Potential Differences at Interfaces
with Reversible Ion and Electron Exchange
Ion exchange and electron exchange are potential generating processes. When
phases are contacted, mobile ion electrochemical potentials equilibrate locally. Like-
wise, Fermi levels balance locally. Mass transport is then set in motion and potential
differences (pds) often depend on mobilities as well as on intrinsic energy differences
(17). The concept of local vs. global equilibrium is essential for understanding charge
exchange phase behavior. When phases, containing mobile charged species, are con-
tacted, local equilibrium allows exchange of charged species at the interface, forma-
tion of space charge, an electric field and a potential distribution with a net potential
difference (pd) reckoned from the inner homogeneous regions of each phase. This
local equilibrium may also be total or global equilibrium because any further charge
motion would violate electroneutrality of the phases. On the other hand, the system
may continue to pass charge at a very low rate (because of low species mobilities), and
366 R.P. Buck

so the system very slowly passes through many local equilibrium states. Experience
with electrochemical systems and materials allows quick recognition o f systems that
reach global equilibrium quickly (in a few Debye relaxation times, e.g., micro- to mil-
liseconds) and those that may move to global equilibrium in minutes, hours or days.
The latter cases, such as glass membranes, exist in a long-lived steady state for years!
Relaxation of unbalanced space charge species across interfaces occurs in the
Debye relaxation time and is driven by the energy difference of the species in each
phase. Of course, as mentioned above, formation of space charge is not the only po-
tential-determining process since reorientation of solvent, and perhaps other ions,
must also occur. For electrons in each phase, the Fermi levels are:

ire = e + ( R T / n ) ln(area/aox) - Feb , (1)

n = Zo• - zr~a 9 (2)

For electrons in metals:

~ --- e - F~b . (3)

For ions, the equivalent electrochemical potentials are:

~i = t~~ + R T l n ( a i ) + ziF~ . (4)

Upon equating Fermi levels, differences in "standard" energies, e's, are linearly re-
lated to the negatives of the standard electrode potentials.

Potential Differences A c r o s s the Interior o f Phases

Normally, it is thought that p d s exist across homogeneous bulk phases only when
current (ionic or electronic) passes through the phase (3). Generally a voltage is ap-
plied to produce the current, and one accounts for the applied voltage in terms of
both interfacial p d s and the I R drop in the bulk phase, where R is conventional re-
sistance. The interfacial p d s are the net potential differences across the few tens of
nanometers of space charge that separate electrically neutral bulk regions in each
phase. The so-called I R drop is the residual voltage in the direction of uniform cur-
rent flow through a region of constant resistivity (constant concentration of the main
resistance-determining species). Generally, the region where the IR voltage develops
is much larger than the non-uniform concentration, space-charge regions. Small vari-
ations of local concentrations, by virtue of space charge, are unimportant in calcu-
lating R. To establish the interfacial p d components, local equilibrium equations, e.g.,
Nernst, Nernst-Donnan, etc. (23), are used when the net current is much less than the
exchange current. Alternatively, kinetic equations (3) can be used to establish the in-
terfacial parts. The Nernst-Planck equation is used to determine the bulk resistance
and the bulk p d (13). When the analysis is correctly performed the three calculated
quantities add up to the known applied voltage.
However, it is also important to identify common, non-homogeneous conditions
under which a net p d occurs, even at zero current. The net p d includes interfacial and
bulk components that arise from asymmetric bathing (more generally the external
Impedances o f Thin and Layered Systems 367

contacting phases), not from application of an external voltage! These are: 1) one or
more concentration gradients of charge species exists when films are bathed asymmet-
rically with more than one permeable species; 2) individual fluxes of ions during
counter-current flow are non-zero, and 3) mobilities are different for uni- or counter-
directional flows; 4) membranes are contacted with electrolyte on one side and a
reversible ion-exchanging, or electron-exchanging metal on the other. Pressure dif-
ferentials also can produce voltages and currents, but these effects are not pursued
here. One deduces the transport properties from the flux equations: For electrons (in
the absence of band structures)

')re = -- Ue Cox Cred O~e / OX . (5)

For ions,

9li = - u i Ci O~i/Ox , (6)

and from the continuity equation. Occasionally Poisson's equation is required, espe-
cially for films on the order of, or thinner than the Debye Screening Length. It is an
important fact that the transport equations for any ion blocked at an interface gen-
erate a Boltzmann distribution that is the basis of the Gouy-Chapman theory and the
origin of the interfacial capacitance.

CELLS W I T H EVEN A N D O D D N U M B E R S OF C O N T A C T S
To study electrochemical properties o f membranes or films, two active interface
liquid, semi-solid and glassy systems of ion conductors and mixed electron/ion con-
ductors are frequently involved. Even numbers of interfaces within cells require pairs
of ion contacts, or electronic contacts to be used. When one of each type of contact-
ing phase is used, the cell contains an odd number of interfaces. The impedance spec-
tra show properties of this symmetry (evenness) or asymmetry (unevenness) of the
number of contacts (interfaces). Equivalent effects in voltage step, voltage scan or ac-
tivity step experiments produce an interesting correlation with impedances. These will
be discussed later.
The common impedance features: bulk geometric, surface kinetic, finite Warburg
with DC resistance, and finite Warburg with pure DC capacitance are considered here.
Membranes and films are considered to be "unsupported" systems, i.e., there are no
inert, extraneous salts present that cause the interfacial double layer space charge re-
gions to have minimal thickness with corresponding collapse of the bulk electric field
to minimal value. Impedances are then related to electrochemical characteristics, es-
pecially described by coupled ion (Nernst-Planck) or coupled ion/electron hopping
in the films or membranes and the contact arrangement. Odd number interface cells
are used to determine apparent interracial ion kinetic sources and their effects. Liq-
uid-liquid interface cells are described later together with single interface impedance
measurements to isolate kinetic processes at one interface, that would otherwise be
coupled with processes at the second interface o f normal two-sided films and
membranes.
A feature of all cell designs for impedance measurements is the conscious attempt
to make impedances of the interface or interfaces (say, of a film) the largest item in
368 R.P. Buck

the series circuit of the cell. This means making low impedance contacts and refer-
ence electrodes.

Typical Membrane (Film) Materials

Passive Ionic Membranes (9,10,12).

I. Fixed-site (covalently bonded, immobile charged groups)

A. Homogeneous: glass, ion conducting crystals, sulfonated polystyrene
B. Heterogeneous: pressed pellets, zeolites in binders, plasticized charged-site,
poly(vinyl chloride) membranes with selective reagents

II. Mobile-site (hydrophobic, trapped, organic ions)

A. Homogeneous: plasticized extractants in poly(vinyl chloride)
B. Heterogeneous: plasticized liquid extractants in pores of inert supports

III. Thin Layer Electrolyte Cells: silver nitrate electrolyte between silver metal
electrodes

IV. Initially Site-Free Membranes or Constrained Regions

A. Hydrophobic Ion Pairs Trapped in Plasticized Membranes
B. Constrained Liquid Junctions
Passive Mixed Conductor Membranes (14-16).

I. Fixed-site Redox Polymer Electron/Ion Exchangers

A. Polymer Radical Ion Sites: pyridinium-based, pyrrole-based, quinone-
based, viologen-based
B. Redox Ion Sites: ferrocene-based, Fe(II/III), Ru-based, Os-based, Co-
based, prussian-blue-based
C. Inert, Non-redox Site: conventional ion exchangers, e.g., Nafion redox
counterions
D. AgBr with ionic electrolyte contacts, or with non-Ag metal contacts

II. Mobile-site Redox Electron/Ion Exchangers

A. Trapped Cation Sites: Aliquat ferri-ferrocyanide
B. AgBr with Ag metal contacts
EXPERIMENTAL CELL ARRANGEMENTS
To illustrate the kinds of cells for impedance studies of films and membranes, the
following examples of even, symmetric electrode cells, and asymmetric, odd number
electrode cells are offered.

Ionic Contact Cells

Ag/AgCl MX (aqueous) membrane MX (aqueous) Ag/AgCl

low impedance membrane film membrane low impedance (I)
ref.l contact solution contact ref.r
x=--d x=d
Impedances of Thin and Layered Systems 369

Ionic~Electronic Contact Cells

Ag/AgC1 MX (aqueous) membrane] inert

low impedance membrane film metal (IIA)
ref.~ contact solution contact
x=--d x=d

Redox Half Cell Contacts

reactive membrane reactive

metal film non-metal (IIB)
contact solution contact
x=-d x=d

Electronic Contact Cells

inert membrane inert

metal film metal (III)
contact solution contact
x=--d x=d

Cells I and III are symmetric. Cell I is used to determine ionic properties since any
electron transport across interfaces is blocked, unless one specifically incorporates
pairs of redox ions in all phases. Cells of type I were used only in the last 25 years
to isolate ion conductivity properties from electronic (8,54). Similarly, Cell III iso-
lates electronic transport since ions are blocked (31,60).
Cells IIA and lib are asymmetric and necessarily involve transport of ions and
electrons. The silver wires and AgC1 (or Hg/calomel) coating should have large ar-
eas so that the interfacial resistance is low and capacitance is high compared with the
parameters of the membrane, film, or solution. Stirred solution resistances should be
less than the membrane resistance. All cells use copper contacts for terminal connec-
tions to avoid contact potential differences. In most cases, circuit impedances in se-
ries with film impedances are small (102 to 104 fl in comparison with 105 to 107 fl).
When film impedances are comparable with cell impedances without the film, anal-
ysis of impedance plots becomes more difficult.
Membrane bathing solutions contain salt with an ion or ions in common with the
membrane interior. Vertical lines represent interfaces with major pds. Dashed lines
show location of junction potentials that are usually small and constant. Reading
across any cell, one encounters interfaces between different phases and bulks of
phases. Fields and pds can develop spontaneously or be produced at interfaces and
in bulk by application of a voltage. Interfaces that do not permit passage of a given
species are "blocked" for that species. Ideally the charged species internal to the mem-
brane: fixed sites or trapped hydrophobic sites, are blocked, and they confer the ref-
erence species and sign. Ions that move in and through phases are grouped as
"counterions," and have signs opposite to the sites. Ions of the same sign as sites are
thermodynamically rejected (but not blocked) and are "coions."
370 R.P. Buck

RECOMMENDATIONS FOR IMPEDANCE DATA PLOTS

Impedance or admittance data can be plotted in several ways: absolute magnitude
of impedance vs. frequency and phase angle vs. frequency; real and imaginary im-
pedance components vs. frequency (usually on logarithmic scales), or impedance
plane plots, e.g., negative imaginary impedance vs. real impedance with frequency
as the implicit variable. Modulus plots, e.g., a dimensionless impedance function ob-
tained by multiplying by jo~Co with Co taken as the empty cell capacitance, can be
substituted in the various plots. They have the advantage of spreading out extreme
frequency data. The impedance plane plot is preferred because individual processes
are emphasized and they appear as semicircles with center on the real axis, in the sim-
ple RC cases. These processes have characteristic traces in the Bode-type of plots, but
when processes become "overlapping" by virtue of similar time constants, the imped-
ance plane plots are more easily recognized. A selection of impedance plane plots are
shown in Fig. 1.

GENERAL IMPEDANCE PLANE PLOT CHARACTERISTICS

Macdonald and coworkers have worked out impedance behavior of many homo-
geneous, one-dimensional phases with plane, parallel contacting phases or electrodes

CB CR CA CG(r Cw( 0U)

I 11 ii
I Absorption ~ ]
ii
G/R [
i~ O
~ n

RB RR RA RG(m) Rwlwl
(a)

-I~(ZT)

i ~-" GIR "~

t) Roo RD
Re (ZT)
(b)

--Irn (ZT) ]
/

/ G / R ~\ D

0 Rco
Re(ZT)
(c)

FIGURE 1. Passive network of parallel RC elements for electrochemical impedances and correspond-
ing impedance plane plots: (a) series of parallel element model for processes identified in the text;
(b) manifestation in the impedance plane plot for corresponding processes with widely separated time-
constants (Cells of types I and III); (c) manifestation in the impedance plane plot for corresponding
processes with widely separated time-constant (Cells of type IIA and liB).
Impedances of Thin and Layered Systems 371

(26,39-45,47). In a well-behaved system, one expects perhaps four or five features in

the impedance plane plots provided there are met the requirements: 1) chemical ho-
mogeneity on a microscopic level; 2) physical uniformity and freedom from cracks
or grain boundaries; 3) a fixed number and kinds of charge carriers; 4) fixed, con-
centration-independent mobilities of each charge carrier; 5) measurable rate of trans-
fer of carriers at an interface (Butler-Volmer or other kinetics); 6) a rate controlling
adsorption process or adsorption-reaction step; 7) measurable rate of generation and
recombination of charge carriers including ion pairing and complex formation; 8)
conventional space charge capacitances including effects of specific adsorption and
surface states.
It is not necessary to solve the equations of transport under a pulse or step to ob-
tain the impedance function, although analytical solutions are satisfying. Digital sim-
ulations of the Nernst-Planck equation, or the Nernst-Planck-Poisson equations, or
sets of parallel equations for several ions, can be achieved. Most examples use con-
stant current, but constant voltage perturbations are possible as well for thin cells,
with one dimensional transport. From the response calculations, we generated imped-
ance plane plots using fast Fourier Transform (5) and applied the method to trans-
mission line equivalents (7), kinetics (55) and to Donnan Failure (6).
The representation of the theoretical processes can be done as a circuit with par-
allel paths containing series elements. However, these circuits are harder to interpret
intuitively than the equivalent series circuits with parallel R C elements. In membrane
systems the resistors and capacitors are usually in a series arrangement in space, so
the latter circuit type, shown in Fig. la has good physical meaning.
The middle panel, Fig. lb shows a number of semicircles or skewed semicircles.
These are theoretically analyzed processes with selected time constants such that each
process provides an isolated feature. They are particularly appropriate for solid sam-
ples, but include very high, and very low frequency features that also occur for im-
miscible liquid/gelled or plasticized liquid membrane phases. All of the features do
not necessarily occur in the impedance spectrum for any given sample. This figure
should be considered a composite of possible behavior. There is included:

1. The high frequency semicircle of a homogeneous system. The resistance is the

high frequency value with no depletion regions. The capacitance is geometric
since the charge buildup is on the contacting electrodes (or in the external space
charge regions in the bathing electrolytes) external to the film itself.
2. The next highest frequency semicircle that is associated with a layer of higher
resistance and, often, higher capacitance than the first. If this phase resistance
is not more than about 10% of the first, the second semicircle is difficult to see
on the side of the first. This higher resistance layer can be a surface or an ac-
tivation (kinetic) resistance.
3. The surface-adsorbed state that is occasionally implicated in the ion- or elec-
tron-transfer step. A possible third semicircle may appear at lower frequencies.
4. In semiconductors and in defect ionic crystals, there are characteristic times for
generation and recombination of charge carriers of opposite sign. Conceivably
some processes, such as ion pairing or trimer formation, may involve a change
in the number of charge carriers, and another semicircle might appear in the im-
pedance plane plot.
5. The low frequency feature that has been extensively studied and analyzed is the
372 R.P. Buck

diffusion-migration controlled Warburg impedance. There are three manifes-

tations of this ladder network. The first is an equal magnitude real and imagi-
nary impedance appearing as a 45 o line in the impedance plane plot. This line
signals a semi-infinite diffusion process corresponding to the absence of a steady
state current. The "finite Warburg" begins on the high-frequency side with the
same 45 ~ line, but the imaginary part passes through a maximum and the real
part terminates at the DC resistance of the membrane (Fig. lb). In contrast with
the continual expansion of the depletion layer during semi-infinite diffusion, the
depletion layer eventually reaches across a film phase to produce a finite deple-
tion layer. The third situation is the electroactive film phase that is oxidized or
reduced to a specified level of completion set by the applied DC voltage. The
imaginary impedance increases without limit, while the real part terminates at
!3 the finite-thickness value (Fig. lc).
. "Sunken" semicircles are those semicircle impedance features whose centers are
below the real axis. These semicircles seem to obey an equation in which the de-
nominator shows a non-integral exponent:

R
Z(j~) - (7)
1 + (jo~z) ~

in which ~ lies between 0 and 1; the latter corresponds to a normal R C circuit.

The equation can be readily analyzed and written in terms of frequency-depen-
dent R and C. This equation corresponds to a unique distribution of time con-
stants (11). However, the equation cannot be correct because it predicts that the
capacitance diverges at extreme frequencies. This means that real data that seem
to obey Eq. 7 must really intersect the real axis with an infinite slope (e.g., ver-
tically). The origin o f the distribution o f time constants is the distribution o f
capacitance across any finite membrane surface that is clamped in an insulat-
ing holder. This effect is also possible in some pressed powder membrane films;
but there is also a possible distribution of resistive paths normal to the surface.
Both could combine to give a nearly random distribution (bell-shaped) of time
constants around the most prevalent value. Some so-called "constant phase el-
ements" are invoked to represent these frequency dependent processes. But
these are formal solutions to a real physical effect. Self-similar or self-affine
surface roughness (fractal behavior) is a popular interpretation of the dis-
tribution of capacitance. This topic is addressed in one of the Symposium
contributions.

Curve fitting of real data, corresponding to similar time constant, overlapping fea-
tures, can be done by non-linear least square fitting routines. This topic and refer-
ences occur in another contribution to this Symposium.

F R E Q U E N T L Y OBSERVED C H A R A C T E R I S T I C I M P E D A N C E S
C O R R E S P O N D I N G TO I-V CURVES
There are basically four types of impedance plane plots that characterize all of the
well-behaved, reversible, homogeneous systems:
Impedances of Thin and Layered Systems 373

O.e

LO.O

-0. 2 i i

-0. 2 0.0 0.2

Pot@ntiol
(e)

0.5

-Z z

0, O i i i

O.O 0.5 1.0

Z,
(b)

FIGURE 2. (a) Steady state ohmic response current-voltage curve (fixed-site, single charge carrier,
uniformly distributed resistance). (b) Impedance plane plot for corresponding I-V curve, typical geo-
metric impedance with external space charge capacitance (frequencies increase counter clockwise).

A Single Geometric Semicircle in the Impedance Plane Plot

These systems are entirely ohmic over all frequencies and polarizing DC biases,
i.e., the I-V curve is a straight line as shown in Fig. 2. The width o f the semicircle,
usually called R~o, is also the DC resistance RDC. Fixed-site membranes with a sin-
gle permeable ion, ionic crystals with a single mobile defect, and ohmic contact semi-
conductors belong to this group (20,48,52). When the film resistance is lower than the
bathing solution resistance, one can observe a "high frequency" Warburg slope on the
high frequency side of the semicircle (30). A lowered semicircle usually means distrib-
uted capacitance and surface roughness or so-called fractal behavior (49), while a flat-
tened semicircle implies merger of two comparable time-constant semicircles caused
by a resistive surface film, or true surface kinetics.
Most impedance measurements are made with a cell voltage bias to achieve zero
current through the cell. The reason is that the classical theories of network responses
374 R.P. Buck

presume that the current voltage (I-V) curve crosses the zero current at the origin. The
bias simply shifts the voltage scale by compensating for the usually non-zero voltage
of arbitrary chemical cells. Most I-V data are collected using stirred bathing solutions.
This requirement comes about because of the possibility that the membrane may have
lower resistance than the solutions! Then, a voltage scan will cause depletion of charge
carriers in the solution where ions enter the membrane. This effect is also called "con-
centration polarization" in electrochemical literature. The I-V curve will not be lin-
ear and the results can be mistaken for a liquid membrane subject to interior carrier
depletion. In addition to the single semicircle in the impedance plane plot, the deple-
tion will cause distortion of the log frequency data and appearance of a Warburg
impedance.
Existence of the high frequency semicircle in the impedance plane is not restricted
to ohmic systems. It will also appear in liquid membrane, mobile charge carrier sys-
tems. However, there will also be a characteristic finite Warburg impedance skewed
semicircle at lower frequencies.

Finite Warburg Behavior with DC Resistance (14,29)

Appearance of a finite Warburg element with DC resistance is characteristic of
symmetric electrode cells I and III (Fig. 3). The classical unsupported electrolyte ex-
amples are symmetric thin layer cells such as two parallel plane Ag electrodes con-
taining a solution of aqueous AgNO3; but, also include ceils containing mobile,
trapped site membranes with ionic contacts, and fixed redox site membranes with elec-
tronic and ionic contacts together on both sides. In the steady state, the concentra-
tion profile of charge carriers is tipped in space: high at the interface where charge
carriers enter and depleted at the exiting interface. There is a steady state DC current,
and the width of the finite Warburg arc is determined by the concentrations, charge
and mobility of that species that actually passes through the cell. This contrasts with
Roo that contains contributions from all of the mobile charged species, essentially
just oscillating in space.
Point-by-point current data collection for a very slow, steady state scan of a liq-
uid ion exchanger membrane or Ag cell (above), produces the I-V curve in Fig. 3.
Unique I-V data on these systems cannot be collected simply by applying a voltage
ramp with arbitrary sweep rate while observing the resulting instantaneous currents.
The reason is that the magnitude of the current depends on the sweep rate for most
membranes with slow moving ions. This well-known effect can be understood by
imagining the voltage sweep to be a series of voltage steps. For these mobile, non-
ohmic systems, the measured current depends on the sampling time after voltage ap-
plication, because charge carrier concentration continuously changes with time until
the steady state current is reached. It is the built-in time-dependence of the I-V curves
that corresponds to the frequency-dependence of the Warburg impedance element.

Finite Warburg Behavior With Pure Capacitive Character at DC (14)

Pure capacitive character at DC is characteristic of membranes and films that ab-
sorb charge and can be completely oxidized or reduced during measured external cur-
rent flow. This behavior is found for asymmetric thin layer cells (different electrodes
on each side such as Cell IIA" redox site membranes with electronic contact on one
side and electrolyte on the other, and reactive electrode Cell IIB: Li/LiI/I2 (Fig. 4).
Impedances of Thin and Layered Systems 375

1.0

"d
t.
t.
0.0
(.a

-i. 0
-0. 2 0.0 0.2
Potential
(a)

0.5
-Z I

0.0
0.0
/ i

0.5
j i

1.0
ZR
Ib)
FIGURE 3. (a) Steady state current voltage curve for a symmetric cell with concentration polariza-
ble sites or carriers, e.g., trapped mobile site systems. A typical S-shape curve obtained at fixed times
after a voltage step in a static electrolyte, or in the steady state of a stirred system. (b) Classical fi-
nite Warburg impedance plane plot, normalized to unity at DC, corresponding to the |-V curve.

These systems have no steady state current because the cell electrolyte composition
varies until any applied voltage becomes the new equilibrium cell voltage and current
necessarily ceases to flow. Ceils of the liB class give "bell-shaped" I-V curves when
voltage is scanned from large values (Iz positive, Li negative), e.g., no initial current,
to lesser voltages. As the applied voltage is diminished very slowly, to allow time for
the cell to remain continuously in equilibrium, passage o f DC current changes the
electrolyte composition by formation of LiI; current increases to a maximum value
in the vicinity of the standard cell voltage where concentration gradients are great-
est; and eventually current drops to zero as the electrodes are dissolved. There is, at
the beginning, a characteristic infinite frequency bulk resistance; at the end of the ex-
periment there is a new high frequency resistance that must be probed by applying
a new high frequency voltage.
Both cases of Warburg character are anticipated in Macdonald's work (27,28).
However, Ho, Raistrick and Huggins (33) developed the theory in a somewhat dif-
376 R.P. Buck

~ 0.2

0.0
-0.2 0.0 0.2

Pot@ntiol
(a)

0.5
-Zj

0.0
0.0 0.5 1.0
Z,
(bl
FIGURE 4. (a) Non-steady-state current-voltage curve for an asymmetric cell. Typical "bell" curve
dependent on voltage scan rate. |b) Impedance plane plot, Warburg impedance capacitive at DC, with
same parameters as Fig. 3b.

ferent form from (14) since the latter does not consider interfacial kinetics. Experi-
mental verifications for conventional asymmetric cells (33) and asymmetrically bathed
redox films on electronic conductors (1,2,35,53) leave little to be desired.

Negative Capacitance~Resistance Coupled with Warburg Behavior

Apparent inductive character at low frequencies, shown by fixed site membranes

with two (or more) permeable counterions or mobile site membrane under Donnan
exclusion failure is caused by too few sites at one interface during DC bias (Fig. 5).
This type of behavior is found extensively in the electrophysiology and biophysical
chemistry literature (25). However, the examples are often complicated by the living
"active" transport character. Nevertheless, for passive membranes, conditions can be
easily found for apparent negative resistance, negative capacitance, and oscillatory
character from which apparent "inductive" loop impedance plots are the correspond-
ing manifestation. Franceschetti and Macdonald (26) have pointed out that imped-
ance arcs in the second quadrant correspond to combined negative resistances and
capacitances, just as positive values lead to arcs in the normally-encountered first
Impedances of Thin and Layered Systems 377

(a) 12 I I |

I1.
~-6

4 g

I I I
O0 5 10 15
-FA~IRT

(b) wc
Rdc

i'
N

WL
, , , i I i , i , I , , 9 , I = i I I [
O
30 35 40 45 50
Z REAL

FIGURE 5. Upper: (a) Steady state I-V curve for ideal Donnan exclusion; (b) steady state curve
(trapped, mobile site membrane) with onset of Donnan exclusion failure; (c) complete Donnan fail-
ure; (d) h y p o t h e t i c a l o h m i c r e s p o n s e . Lower: Result of n e g a t i v e c a p a c i t a n c e on an i m p e d a n c e plane
plot, corresponding to case (b) upper.

quadrant. They also assert that circuits with inductors can most always be replaced
by identical frequency response circuits with negative differential resistances and ca-
pacitances. It is easy to find negative differential resistances in the Donnan failure ex-
amples, and with some imagination, negative capacitances. It is not easy to reconcile
these systems with real inductors in the circuits. However, examples of parallel pro-
cesses: slow direct surface reaction coupled to reactions from the surface-adsorbed
states (26) may be better represented by inductances, even though the processes have
nothing to do with magnetic field induced currents.
When films contain sites, there are counterions present in maintain electroneutral-
ity. There are also some, albeit small numbers, of coions present as well. Electroneu-
trality of film phases requires zs (sites) + Zoo (coions) = -Zcou,ter (counter) where
3 78 R.P. Buck

parentheses denote concentrations. However, when one side of a film is bathed in very
high activities of salt, the whole salt can extract in addition to the nearly site-equiv-
alent number of counterions. Normally, when site concentrations are made intention-
ally large and bathing solutions are dilute, the system obeys "Donnan exclusion" (of
coions). When salt is admitted, one refers to "Donnan exclusion failure." The theory
is thermodynamic and has been worked out for fixed-site systems (32), and mobile,
trapped site systems (21,22,56).
Donnan exclusion can also fail simply by passing current (biasing) the film. On the
exiting side on a film, the site concentration is very much lower than on the entering
side. The result is that salts enter the membrane and the membrane resistance actu-
ally d e c r e a s e s with i n c r e a s i n g applied voltage. This phenomenon in transform space
gives the "inductive" behavior (57). The manifestation of this negative resistance/ca-
pacitance "pseudo inductive" responses are found for systems with I-V curves that are
typical liquid membrane curves at the zero crossing and on to the limiting current
range (Fig. 5). But at high biases, increased current flows, i.e., the resistance decreases
at high applied voltages. Even selective glass fixed site membranes bathed asymmet-
rically show, by digital simulations, inductive loops at low frequencies. More com-
monly, mobile-site membranes with DC biases such that site concentrations (and
counterion concentrations) are low at one interface, undergo Donnan failure and co-
ions are admitted into the membrane. This decreasing resistance at larger biases cor-
responds to an "inductive loop" in the impedance plane plots.

C H A R A C T E R I S T I C TIME CONSTANTS
OF M E M B R A N E SYSTEM
When there are well-resolved features in the impedance plane plots, the time con-
stant of the process is found by noting the frequency of the imaginary impedance
maximum. For the usual R C combinations, this frequency is related to the time
constant:

r x ~- 1/7 = 1 / R C . (8)
The real resistance is simply read from the width of the semicircle and C is computed.
The chemical significance of the R's and C's has been touched on earlier in this pa-
per. For the high frequency geometric R and C process, C is given by

C = eoeA/2d (9)

for the example cells with films or membranes of thickness 2d, cross-sectional area
A, conventional dielectric constant e, and permittivity Co. These capacitors are of the
order of picofarads to nanofarads, roughly. For surface films, potential-dependent
surface rate processes, and second and subsequent more resistive films, the capaci-
tance is much larger, of the order of microfarads. For a single interface it is typically
the Gouy-Chapman value corrected for voltage and compact layer capacitance. The
Gouy-Chapman base value, per unit area, is

C = ~e0e , (10)

where r is the reciprocal thickness of the diffuse double layer. For any electro-
lyte/plasticized membrane with different concentrations of ionic species and differ-
Impedances of Thin and Layered Systems 379

ent dielectric constants, there are two back-to-back diffuse layers in series, and so two
Gouy-Chapman capacitances in series. If the membrane is symmetrically bathed in
electrolytes, there are four capacitances added in series from the two interfaces. The
time constants, resistances and capacitances for adsorption/reaction and for gener-
ation/recombination of charge carriers (e.g., ion-pairing) are sufficiently complicated
that the original papers should be consulted (46,47). These values are even larger than
the interfacial diffuse capacitances, and it is this reason that time constants tend to
be longer and the frequency range lower.
There is no time constant for semi-infinite diffusion and not a very characteristic
one for DC-blocked interfaces, although the capacitances and resistances have sig-
nificant meaning and values. For the finite Warburg impedance (Fig. 3) there is a
maximum imaginary impedance occurring at r = 5.069, where r is defined

r = (thickness)2/2D . (II)

The resistance is determined by transport of the ion that crosses the interfaces, and
is often a diffusion-migration process down a concentration and local voltage gradi-
ent. The capacitance is not the usual interfacial space charge variety, but is called a
"pseudo" capacitance because it involves transport parameters that reflect the distri-
bution of space charge across the whole bulk of the film. Normally when one ion car-
ries current in a supporting electrolyte, the concentration of all ions remains constant
in space. The resistance and local resistivity is everywhere constant except possibly at
the electrode surface where the space charge resides. Then, the field is constant and
the potential profile is linear in the bulk. This is the usual situation with semi-infinite
cells. The resulting IR drop (the voltage across the uniform part o f the electrolyte),
is an intercept in the impedance plane plot at the high frequency limit of the geomet-
ric semicircle.
In membrane electrochemistry where all system are unsupported (e.g., free from
mobile supporting electrolyte), there are many situations when the charge carrier pro-
files (and the mobile site profile, as well) are curved during the transient, but approach
a linear, usually tipped profile in the steady state. The potential profile may not be
linear in space. In almost all fixed site systems with more than one mobile species, the
concentration and potential profiles are not linear in space even in the steady state
(17). In these cases there is clearly some distributed space charge in the bulk mem-
brane associated with the transport process itself.
For diffusion with conversion of the entire membrane to an oxidized or reduced
state (Fig. 4), there is no time constant because there is no maximum imaginary part.
The resistance approaches ~ of the DC film value, while the capacitance is entirely
"pseudo." The capacitance is the total charge storing capacity, regardless of the fact
that the stored charge is not space charge. For a redox process involving n electrons
transferred, the capacitance is

C = ( n F ) 2 ( f o x + ere ) ~Kfilm t h i c k n e s s / R T (12)

A SINGLE I N T E R F A C E - A P P L I C A T I O N S TO
I N T E R F A C I A L KINETICS

A pervasive problem with impedance spectroscopy of two interface systems is the

subdivision of events at one interface from those of the other. This problem has been
380 R.P. Buck

pursued by studying single interface impedances (actually three, but two are intention-
ally low impedance interfaces in series). Although we have not succeeded in insert-
ing low impedance reference probes or other reversible contacts into membranes
(liquid or solid), it is possible to simulate a single interface of a liquid electrolyte/semi-
solid (plasticized) membrane by using the immiscible liquid/liquid interface. These
measurements have been crucial in establishing validity of the Verwey-Niessen (back-
to-back diffuse double layer capacitance) theory (50,51), and in discovering measur-
ably small interfacial kinetics (19,61). These modern "single" interface impedance
measurements are related to the classical metal/redox electrolyte impedances that are
based on semi-infinite transport with or without coupled interfacial electron trans-
fer kinetics. However, in the modern examples, there is diffusion-migration mass
transport in each of two phases on each side of the liquid/liquid interface.
The measurement of single interface kinetics has been pursued using cells of the
type (59),

Ag/AgCl MX (aqueous) membrane Ag/AgTPB

low impedance membrane film, immisc. low impedance (IV)
ref.l contact solution ref.r
X=0

in which TPB- (tetraphenylborate) is one of several anions soluble in many mem-

brane phases and in non-aqueous immiscible electrolytes, e.g., nitrobenzene or 1,2-
dichloroethane. Cells of this kind using low impedance references, permit study of
the interface electrochemistry between two phases (19,50,51,61). Berlouis et aL (4)
were the first to measure impedances, although the technique and systems were be-
ing investigated by several groups, simultaneously, to study not only kinetics, but dou-
ble layer capacitances. For the aqueous electrolyte/nitrobenzene system, impedance
plane plots show, typically, kinetic and Warburg character (19), on which a kinetic
semicircle can be fit. The exchange current density is calculated from the activation
resistance classically and rate constants for forward and backward steps are estimated.
More recent and precise analyses were described by Samec et al. (61).
It is a moot point whether the surface resistance is activation controlled, or resis-
tive layer controlled. It is possible that the interfacial region is a mixture of solvents,
running from water to nitrobenzene because the solubility of nitrobenzene in water
(and vice versa) is not negligible. Mobilities of ions may vary from point-to-point in
the mixture region, and desolvation-resolvation kinetics may be involved.
The situation of the ion transfer kinetics at water/dioctylsebacate (plasticizer) in-
terfaces (34,36,58) and water/silicone rubber interfaces (38), used to construct many
ion sensitive devices, is equally unclear. The bulk of evidence from impedances and
surface clean-up techniques demonstrates the presence of resistive external surface
films (from impurities in plasticizers and suspending agents used for polymerization)
and resistive inhomogeneous regions with lower-than-average site densities, just be-
low membrane surfaces.

REFERENCES

1. Armstrong,R.D. Impedanceplane displayfor an electrodewith diffusionrestrictedto a thin layer.

J. Electroanal.Chem. 198:177-180; 1986.
Impedances of Thin and Layered Systems 381

2. Armstrong, R.D.; Lindholm, B.; Sharp, M. Impedance characteristics of a modified electrode. J. Elec-
troanal. Chem. 202:69-74; 1986.
3. Bard, A.J.; Faulkner, L.R. Electrochemical methods. New York: John Wiley & Sons; 1980.
4. Berlouis, L.E.A.; Chatham, J.; Girault, H.H.J.; Schiffrin, D.J. Transport properties of phospholipid
monolayer membrane models. Ext. Abstr. Electrochem. Soc. 83-1(620):935-936; 1983.
5. Brumleve, T.R.; Buck, R.P. Numerical solution of Nernst-Planck and Poisson equations systems: Ap-
plications to membrane electrochemistry and solid state physics. J. Electroanal. Chem. 90:1-31; 1978.
6. Brumleve, T.R.; Buck, R.P. Potential reversals across site-free, passive m e m b r a n e s - A simulation anal-
ysis. J. Electroanal. Chem. 126:55-72; 1981.
7. Brumleve, T.R.; Buck, R.P. Transmission line equivalent circuit models for electrochemical impedances.
J. Electroanal. Chem. 126:73-104; 1981.
8. Buck, R.P. Transient electrical behavior of glass membranes. J. Electroanal. Chem. 18:363-399; 1968.
9. Buck, R.P. Potential-generating processes at interfaces: From electrolyte/metal and electrolyte/
membrane to electrolyte/semiconductor. In: Cheung, P.W.; Fleming, D.G.; Neuman, M.R.; Ko, W.H.,
eds. Theory, design, and biomedical applications of solid state chemical sensors. W. Palm Beach: CRC
Press, Inc.; 1978: pp. 1-41.
10. Buck, R.P. Transport properties of ionic conductors. Bergveld, P.; Zeme, J.; Middelhoek, S., eds; In:
Chemically sensitive electronic devices. Amsterdam, Netherlands: Elsevier Pub. Co.; 1981: pp. 137-260.
11. Buck, R.P. The impedance methods appfied to the investigation of ion-selective electrodes. Ion-Selective
Electrode Rev. 4:3-74; 1982.
12. Buck, R.P. Electrochemistry of ion-selective electrodes. In: White, R.E.; Bockris, J.O.; Conway, B.E.;
Yeager, E., eds. Comprehensive treatise of electrochemistry. New York: Plenum Pub. Co.; 1984:
pp. 137-248.
13. Buck, R.P. Kinetics of bulk and interfacial ionic motion: Microscopic bases and limits for the Nernst-
Planck equation applied to membrane systems. J. Membr. Sci. 17:1-62; 1984.
14. Buck, R.P. Diffusion-migration impedances for finite, one-dimensional transport in thin layer and
membrane cells: An analysis of derived electrical quantities and equivalent circuits. J. Electroanal.
Chem. 210:1-19; 1986.
15. Buck, R.P. Diffusion-migration impedances for finite, one-dimensional transport in thin-layer and
membrane cells, mixed conduction cases: Os(III)/Os(II)C104. J. Electroanal. Chem. 219:23-48; 1987.
16. Buck, R.P. Electron hopping in one dimension: Mixed conductor membranes. J. Phys. Chem.
92:4196-4200; 1988.
17. Buck, R.P. Steady state diffusion-migration potential differences in mixed conductor polymer films
and thin layer cells. J. Electroanal. Chem. 271:1-14; 1989.
18. Buck, R.P. Impedances of membrane systems with metal and/or ionic contacts. Electrochim. Acta
35:1609-1617; 1990.
19. Buck, R.P.; Bronner, W.E. Prediction of salt effects on rates of single-ion crossings in ITIES exper-
iments. J. Electroanal. Chem. 197:179-188, 1986.
20, Buck, R.P.; Mathis, D.E.; Rhodes, R.K. Impedance measurements on purified silver chloride crys-
tals using ionic vs electronic contacts. J. Electroanal. Chem. 80:245-257; 1977.
21. Buck, R.P.; Stover, F.S.; Mathis, D.E. Site concentration determination in liquid ion exchange mem-
branes: Theory and techniques. J. Electroanal. Chem. 82:345-360; 1977.
22. Buck, R.P.; Stover, F.S.; Mathis, D.E. Site concentration determination in liquid ion exchange mem-
branes: Experimental. J. Electroanal. Chem. 100:63-70; 1979.
23. Buck, R.P.; Vanysck, P. Interfacial potential differences at mixed conductor interfaces: Nernst, Nernst-
Donnan, Nernst Distribution and generalizations. J. Electroanal. Chem. 292:73-91; 1990.
24. Cole, K.S. Membranes, ions, and impulses. Berkeley, CA: University of California Press; 1972; pp.
167-203.
25. Cole, K.S. Membranes, ions, and impulses. Berkeley, CA: University of California Press; 1972; p. 303.
26. Franceschetti, D.R.; Macdonald, J.R. Electrode kinetics, equivalent circuits, and system characteriza-
tion: Small-signal conditions. J. Electroanal. Chem. 82:271-301; 1977.
27. Franceschetti, D.R.; Macdonald, J.R. Diffusion of neutral and charged species under small-signal a.c.
conditions. J, Electroanal. Chem. 101:307-316; 1979.
28. Franceschetti, D.R.; Macdonald, J.R. Small-signal response theory for electrochomic thin films. J. Elec-
troanal. Chem. 120:1754-1756; 1982.
29. Franceschetti, D.R.; Macdonald, J.R.; Buck, R.P. Interpretation of finite-length-Warburg-type im-
pedances in supported and unsupported electrochemical cells with kinetically reversible electrodes. J.
Electrochem. Soc. 138:1368-1371; 1991.
382 R.P. Buck

30. Gratzl, M.; Pungor, E.; Buck, R.P. Impedance measurements for pressed-pellet electrode membranes
based on silver iodide and silver iodide/silver sulfide with solution contacts. Anal. Chim. Acta 189:
217-228; 1986.
31. Hebb, M.B. Electrical conductivity of silver sulfide. J. Chem. Phys. 20:185-190; 1952.
32. Hills, G.J.; Jacobs, P.W.M.; Lakshminarayanaiah, N. Membrane potentials: Part I The theory of the
e.m.f, of cells containing ion-exchanging membranes. Part II The measurement of the e.m.f, of cells
containing the cation-exchange membrane, cross-linked polymethacrylic acid. Proc. Roy. Soc. A262:
246-279; 1961.
33. Ho, C.; Raistrick, I.D.; Haggins, R.A. Application of a-c techniques to the study of lithium diffu-
sion in tungsten trioxide thin films. J. Electrochem. Soc. 127:343-350; 1980.
34. Horval, G.; Graf, E.; Toth, K.; Pungor, E.; Buck, R.P. Plasticized poly(vinyl chloride) properties and
characteristics of valinomycin electrodes: High-frequency resistances and dielectric properties. Anal,
Chem. 58:2735-2740; 1986.
35. Hunter, T.B.; Tyler, P.S.; Smyrl, W.H.; White, H.S. Impedance analysis of poly(vinylferrocene) films.
J. Electrochem. Soc. 134:2198-2204; 1987.
36. Iglehart, M.L.; Buck, R.P. Ion transport properties of cyclic and acyclic neutral carrier-containing mem-
branes. Talanta. 36:89-98; 1989.
37. Labes, R.; Lullies, H. Analyses der nerveneigenschaften durch wechselstrommessungen mit hilfe der
membrankernleitertheorie. Arch. Ges. Physiol. 230:738-770; 1932.
38. Lindner, E.; Niegreisz, Zs.; Toth, K.; Pungor, E.; Berube, T.R.; Buck, R.P. Electrical and dynamic
properties of non-plasticized potassium selective membranes. J. Electroanal. Chem. 259:67-80; 1989.
39. Macdonald, J.R. Theory of space-charge polarization and electrode-discharge effects. J. Chem. Phys.
58:4982-5001; 1973.
40. Macdonald, J.R. Simplified impedance/frequency-response results for intrinsically conducting solids
and liquids. J. Chem. Phys. 61:3977~3996; 1974.
41. Macdonald, J.R. Binary electrolyte small-signal frequency response. J. Electroanal. Chem. 53:1-55;
1974.
42. Macdonald, J.R. Some a.c. response results for solids with recombining space charge. J. Phys.
C7:L327-L331; 1974.
43. Macdonald, J.R. Complex rate constant for an electrochemical system involving an adsorbed inter-
mediate. J. Electroanal. Chem. 70:17-26; 1976.
44. Macdonald, J.R. Interpretation of ac impedance measurements in solids. In: Mahan, G.D.; Roth, W.L,
eds. Superionic conductors. New York: Plenum Press; 1976: pp. 81-97.
45. Macdonald, J.R. Space charge polarisation. In: Kleitz, M.; Dupuy, J., eds. Electrode processes in solid
state ionics. Dordrecht: Reidel; 1976: pp. 149-180.
46. Macdonald, J.R. Impedance spectroscopy--Emphasizing solid materials and systems. New York:
Wiley-Interscience Pubs.; 1987.
47. Macdonald, J.R.; Franceschetti, D.R. Theory of small-signal a.c. response of solids and liquids with
recombining mobile charges. J. Chem. Phys. 68:1614-1637; 1978.
48. Mathis, D.E.; Buck, R.P. Ion transport in free and supported nitrobenzene aliquat nitrate liquid mem-
brane ion-selective electrodes I: Bulk electrical properties including ion association and dielectric con-
stant. If: Interfacial kinetics and time-dependent phenomena. J. Membr. Sci. 4:379-414; 1979.
49. Nyikos, L.; Pajkossy, T. Diffusion to fractal surfaces parts I, II, and IIl. Electrochim. Acta 31:1347,
1986; 34:171-186; 1989.
50. Reid, J.D.; Vanysck, P.; Buck, R.P. Potential dependence of capacitance at a polarizable (blocked)
liquid/liquid interface. J. Electroanal. Chem. 161:1-15; 1984.
51. Reid, J.D.; Vanysck, P.; Buck, R.P. Potential dependence of capacitance at a liquid/liquid interface,
unblocked interface. J. Electroanal. Chem. 170:109-125; 1984.
52. Rhodes, R.K.; Buck, R.P. Impedance measurements using ionic contacts on purified and doped sil-
ver bromide crystals. J. Electroanal. Chem. 86:349-360; 1978.
53. Rubinstein, I.; Rishpon, J.; Gottesfeld, S. An a.c.-impedance study of electrochemical processes at
Nation-coated electrodes. J. Electroehem. Soc. 133:729-734; 1986.
54. Sandifer, J.R.; Buck, R.P. Impedance characteristics of ion selective glass electrodes. J. Electroanal.
Chem. 56:385-398; 1974.
55. Stover, ES.; Brumleve, T.R.; Buck, R.P. Comparison of time constants for liquid ion-exchange mem-
brane electrode responses determined by an impedance method and an activity step method. Anal.
Chim. Acta 109:259-278; 1979.
Impedances of Thin and Layered Systems 383

56. Stover, ES.; Buck, R.P. Site concentration determination in liquid ion exchange membranes. J. Elec-
troanal. Chem. 94:59-66; 1978.
57. Stover, ES.; Buck, R.P. Electrical properties of mobile site ion exchange membranes with interfacial
permselectivity breakdown: Non-zero current properties. J. Electroanal. Chem. 107:165-175; 1980.
58. Toth, K.; Graf, E.; Horvai, G.; Pungor, E.; Buck, R.P. Plasticized poly(vinylchloride) properties and
characteristics of valinomyvcin electrodes. Part 2: Low-frequency, surface-rate and Warburg impedance
characteristics. Anal. Chem. 58:2741-2744; 1986.
59. Vanysck, P. Electrochemistry on liquid/liquid interfaces: Lecture notes in chemistry No. 39. Berlin:
Springer-Verlag; 1985.
60. Wagner, C. Galvanic cells with solid electrolytes involving ionic and electronic conduction. In: Pro-
ceedings of the 7th Meeting of the International Commission of Electrochemical Thermodynamics and
Kinetics, London: Butterworths; 1957: pp. 361-377.
61. Wandlowski, T.; Marecek, V.; Samec, Z. Kinetic analysis of the picrate ion transfer across the inter-
face between two immiscible electrolyte solutions from impedance measurements at the equilibrium
potential. J. Electroanal. Chem. 242:291-302; 1988.

NOMENCLATURE
A active area
ai = ion activity
ared = reduced species activity
aox = oxidized species activity
C = capacitance or capacitance/unit area
c~ = ion concentration
Cre d = reduced species concentration
Cox = oxidized species concentration
D = diffusion coefficient
d = half film (membrane) thickness
F = Faraday constant
jo = electron flux
J~ = ion flux
n = number of electrons in half cell reaction
R = resistance
R = gas constant appearing with T
T = absolute temperature
/d e electron mobility
ion mobility
z(joJ) = impedance
Zi = ion charge
Zred reduced species charge
go x oxidized species charge
OL = parameter of Cole-Cole plot
dielectric constant