Sunteți pe pagina 1din 63

http://www.cp.umist.ac.uk/lecturenotes/Echem/index_main.

htm

Electrochemical Techniques
Contents List
This document is designed to serve as a summary of those parts of the Corrosion Basics and
Experimental Techniques courses that are taught by Dr Cottis. 

Introduction
Activation Polarization
Concentration Polarization
Resistance Polarization
Non-Perturbed Measurements
Linear Perturbed Measurements
Non-Linear Perturbed Measurements
Summary of Electrochemical Theory
Electrochemical Reactions
Faraday's Law
Electrochemical Half Cells
Reversibility of Electrochemical Reactions
Electrochemical Equilibria
Reference Electrodes
Electrochemical Kinetics
Electrochemical Double Layer
Kinetics of a Single Electrochemical Reaction
Activation Control
Mass Transport Control
Kinetics of Multiple Reactions
Rate Determining Step
Basics of Instrumentation
Electricity
Charge
Unit of charge the Coulomb, C (not to be confused with C for capacitance).
Charge on the electron (1.6 x 10-19 coulomb).
Current
Unit of current the Amp, A (= 1 coulomb/second).
Electronic and ionic current; conservation of charge.
Voltage (or Potential Difference)
Unit of voltage the Volt, V ( 1 joule required to move 1 coulomb through a potential difference of 1
volt).
Resistance
Ohm's Law (voltage = current x resistance, V = IR).
Unit of resistance the Ohm, W (that resistance which allows 1 amp to flow with a voltage of 1 volt).
Capacitance
Unit of capacitance the Farad, F ( 1 farad will absorb 1 coulomb when the voltage across it increases
by 1 volt).
Charging of a capacitor, C ( I = C dV/dt, Q = C V).
Inductance
Electrical Measurements
Measurement of current, effect of meter impedance (want Zmeter = 0).
Measurement of voltage, effect of meter impedance (want Zmeter = ).
Noise
Interference Noise
Electronic Noise
Frequency Response or Filtering
Operational Amplifiers
Non-Inverting Buffer
Properties
Cell Design for Electrochemistry
Working Electrode
Characterize material, chemical composition and metallurgical condition.
Mount specimen - avoid crevices; avoid cut edges.
Cast in epoxy resin (but problems with disbonding).
SternMakrides PTFE pressure seal.
Prepare surface - as received, ground or polished.
Clean surface.
Degrease.
Electrical connection - do not just twist wires together or wrap around specimen, solder, spot weld or
use bolted joint
Beware of exposure of second metal to the solution.
Counter Electrode (or Auxiliary Electrode)
Must be inert, typically Pt or graphite.
Large surface area compared to working electrode.
Reference Electrode
Commonly SCE
Ag/AgCl for small, cheap electrodes (but chloride sensitive).
Hg/Hg2SO4 to avoid chloride contamination.
Luggin Capillary
Shields potential measurement from potential drop in solution
Close to working electrode, but not so close as to shield part of the surface from the current flow
(ideal is about 3 x tip diameter).
Solution
Composition, including aeration, pH etc.
Temperature.
Flow,including rotating disk electrode, rotating cylinder electrode and flow channels.
Electrochemical Instrumentation
Potential Measurement
Digital voltmeter (input impedance 10 to 1000 MOhms).
pH meter or electrometer (input impedance ~ 1014 Ohms).
Buffer amplifier (input impedance 106 to 1014 Ohms).
Analogue meter (input impedance ~ 100 kOhms).
Chart recorder (input impedance ~ 10 MOhms).
Oscilloscope (input impedance ~ 10 MOhms).
Current Measurement
Digital multimeter (voltage drop up to 1 V)
Digital voltmeter / resistor (voltage drop ~ 10 mV).
Electrometer (voltage drop « 1 mV).
Zero resistance amplifier (voltage drop « 1 mV).
Analogue meter (voltage drop ~ 1-100 mV).
Chart recorder / resistor (voltage drop ~ 10 mV).
Oscilloscope / resistor (voltage drop ~ 10 mV).
Potentiostat
Connections
Auxiliary Electrode
Reference Electrode
Working Electrode
Earth
Reference Electrode Guard
Count Resistor
External Control Input
Overload Indicator
I R Compensation
Common Problems
Oscillation
Noise Pickup
Electronic Noise
Overloading
I R Compensation
Control of Current
Idealized potentiostat.
Real potentiostats - limitations and possible problems.
Design of potentiostats
The Galvanostat
Ideal galvanostat.
Real galvanostats.
Use of potentiostat as galvanostat.
SteadyState and Potential Sweep Methods
Rest Potential Measurements.
Potentiodynamic Measurements.
Polarization Resistance.
Transient Methods
Open Circuit Potential Decay.
Alternating Current Methods.
 
AC Impedance Measurement (Electrochemical Impedance Spectroscopy)
Definition
Application of AC Impedance Measurement to Electrochemistry
Measurement
 Wheatstone Bridge

 Phase and Amplitude Measurement

 Phase-Sensitive Detection

 Sinewave Correlation

 Fourier Analysis
 Multiple Frequency Methods
Cell Configurations
Validation of ac Impedance Data
Electrochemical Noise Methods
Theoretical Basis of Noise Generation Processes
Measurement of Electrochemical Noise
Interpretation of Electrochemical Noise
Relationship between Potential and Current Noise
Electrochemical Noise for the Detection of Localized Corrosion
Glossary
Copyright, Disclaimer and Acknowledgement
Introduction to Electrochemical Techniques

The following links provide rapid access to various sections of this document
Electrochemical Reactions
Polarization of Reactions
 Activation Polarization

 Concentration Polarization

 Resistance Polarization

Electrochemical Measurements
 Non-Perturbed Measurements

 Linear Perturbed Measurements

 Non-Linear Perturbed Measurements

Electrochemical Reactions
Electrochemical techniques have a wide range of application, but their use in corrosion and
electroplating tends to be concerned with trying to find out about properties of the metal-
solution interface: for example, the rate of reactions at the surface, the nature of films on
the surface or the morphology of the surface. The basic tools available to us are voltage and
current. The voltage across the interface can be changed and the current recorded or vice
versa. From these two parameters, we must attempt to deduce everything we can about
what is happening at the interface. When we immerse a metal in solution, there will be a
tendency for the metal to react with the solution, either with metal atoms dissolving as
cations or cations already in the solution depositing as metal atoms:

As a result of these reactions, the metal will tend to accumulate a negative or positive
charge. The build-up of this charge on the metal will change its potential in such a way as
to inhibit the reaction generating the charge until the potential reaches a value at which the
rates of the two reactions are equal and opposite. This is known as the equilibrium
potential, and is the potential the metal will adopt in the solution in the absence of any
other reactions.
It is very important to appreciate that when a piece of metal is sitting in a solution at its
equilibrium potential, this does not mean that the rates of the metal dissolution and
reprecipitation reactions are zero. Instead it implies that the rates of the two reactions are
equal. Since electrochemical reactions invariably involve a transfer of charge, we can
define their rates in terms of charge/unit area/unit time or current density. When the metal
dissolution and reprecipitation reactions are in equilibrium, we refer to the (equal and
opposite) rates of each of the two reactions as the exchange current density.
In corroding systems other anodic reactions are possible, the two most important being the
reduction of dissolved oxygen to hydroxyl ions and the reduction of hydrogen ions or water
molecules to hydrogen gas:

The balance between one or other of these cathodic reactions and the metal dissolution
reaction results in a rate of reaction given by the corrosion current density. One of the main
applications of electrochemical methods to the study of corrosion is the estimation of the
magnitude of the corrosion current density. Electrochemical techniques are also used to
study the mechanisms of corrosion processes.

Polarization of Reactions
An obvious assumption that we need to make is that the way in which the currents of
individual reactions change with potential is described by a reasonably simple theory. Such
a theory has been developed for electrochemical reactions, and provides a good description
of many reactions. In essence this theory proposes that the potential changes (polarization)
which arise as the result of current flowing across a metal-solution interface arise from
three basic processes:
Activation Polarization
This is the change in potential that is required just to make the reaction go faster. This phenomenon
can be analyzed in terms of the energy barrier between the reactant and the product states, and this
gives rise to a relationship between current and potential of the form:

A reaction for which activation polarization dominates is referred to as activation controlled.

Concentration Polarization
Many electrochemical reactions produce or consume species in the solution. The rates of reactions
involving dissolved species which participate as reactants in the rate-determining step (that stage in
the reaction which controls its rate) will be dependent on the concentration of the dissolved species.
As the dissolved species are consumed by the reaction, so a greater change in potential will be
required to maintain the current, and this is known as concentration polarization. A reaction for
which concentration polarization dominates is referred to as mass-transport or diffusion controlled.

Resistance Polarization
Solutions of electrolytes generally have a rather poor conductivity compared to metals, particularly
for dilute solutions. In corrosion systems this is often compounded by paints or other films of
insulating materials which can only conduct by way of traces of water dissolved in the coating. As
we have seen above, corrosion consists of two or more essentially independent reactions, and it is
quite common for these to occur at different sites on the metal surface. If the solution has a high
electrical resistance this will give rise to a potential difference between the anodic and cathodic sites,
which is known as resistance polarization. If resistance polarization dominates a reaction, it is
referred as resistance or IR controlled (the latter term arises from Ohm's Law, V = IR).
Electrochemical Measurements
In real systems all three types of polarization will be observed to a greater or lesser extent,
and much of the underlying theory of electrochemistry is concerned with relating the
polarization to fundamental physico-chemical properties such as activation energy,
temperature, diffusion coefficient and ionic mobilities. As electrochemists interested in
reactions involving metal surfaces, we take these theories, make any necessary simplifying
assumptions (usually very sweeping assumptions!) and then use the relationships between
experimental variables such as current, potential, temperature, and solution composition to
analyze and interpret our experimental results. The theories of electrochemistry are well-
established and reasonably simple - the real skill in their application to real problems lies in
making the optimum assumptions, and in recognizing when these assumptions are starting
to break down.
In order to investigate the values of the various parameters controlling the electrochemical
reaction it is often necessary to perturb the system in some way. There are a wide range of
electrochemical techniques which have been developed to study reactions at metal surfaces,
but by taking a rather broader view of the nature of an electrochemical experiment we can
classify all possible experiments into one of three general classes:
Non-perturbed measurements
This group incorporates all measurements that are made without any external perturbation of the
system. It includes potential-time and current-time monitoring and electrochemical noise
measurements. These techniques are very attractive because any information they are able to provide
relates to the actual system being studied, with little possibility of artefacts due to the measurement
technique.

Linear perturbed measurements


This group includes all techniques that are designed and analyzed on the assumption that the
electrochemical system can be treated as a combination of linear circuit elements (e.g. resistors,
capacitors etc.). This is almost invariably an approximation to reality, but by restricting ourselves to
small changes of potential (of the order of 10 mV or less), we generally obtain a reasonable
approximation to linear behaviour. It can be shown that all linear techniques sample the same basic
characteristics of the electrochemical system, although some methods provide a more thorough
sampling than others. Examples of linear techniques include linear polarization resistance
measurement, electrochemical impedance spectroscopy and some small amplitude transient
techniques.

Non-linear perturbed measurements


These techniques differ from the linear techniques in that the electrochemical system is not assumed
to respond in a linear fashion. Usually this implies that larger amplitude perturbations are used,
although this is not necessarily the case. Typical examples of non-linear techniques include the
measurement of polarization curves, larger amplitude transient methods and harmonic analysis
techniques.
Summary of Electrochemical Theory
Electrochemical Reactions
Faraday's Law
Electrochemical Half Cells
Reversibility of Electrochemical Reactions
Electrochemical Equilibria
Reference Electrodes
Electrochemical Kinetics
Electrochemical Double Layer
Kinetics of a Single Electrochemical Reaction
Activation Control
Mass Transport Control
Kinetics of Multiple Reactions
Rate Determining Step

Electrochemical Reactions
An electrochemical reaction is a reaction involving the transfer of charge as a part of a chemical reaction.
Typical electrochemical reactions in corrosion are metal dissolution and oxygen reduction:

In contrast a chemical reaction, such as the precipitation of a metal hydroxide, does not
involve a transfer of charge:

Faraday's Law
Faraday's Law relates the amount of charge involved in an electrochemical reaction with the number of moles
of reactant reacting and the number of electrons required for the reaction.

In addition to Faradaic processes that obey Faraday's Law, non-Faradaic processes may
also occur. Typically these are processes such as adsorption that do not involve a complete
transfer of charge from the solution to the metal.
Electrochemical Half Cells
A half cell is an electrochemical reaction that results in a net surplus or deficit of electrons, and it corresponds
to the smallest complete reaction sequence (while it may proceed as a sequence of simpler reactions, the
intermediate stages are not stable).
Oxidation or anodic reactions are those that result in a surplus of electrons, and for
corrosion these typically correspond to the various metal dissolution reactions, such as:
Reduction or cathodic reactions result in the consumption of electrons, and for corrosion
these typically correspond to the oxygen reduction or hydrogen evolution reactions:

Note that the above reactions have been shown going in one direction only. While the
reverse reactions are perfectly possible, they reverse of an anodic reaction is a cathodic
reaction and vice versa.
Reversibility of Electrochemical Reactions
A reaction is said to be reversible if it can proceed easily in either direction as conditions change (typically as
the electrochemical potential is changed). There are several aspects of reversibility.

Chemical reversibility relates to the chemical feasibility of the reaction, with a chemically
irreversible reaction being one in which the reverse reaction is prevented by the occurrence of
competing reactions (Example).
A thermodynamically reversible reaction is a chemically reversible reaction for which the reaction
will change direction as a result of an infinitesimal change in potential.

A practically reversible reaction is a thermodynamically reversible reaction that occurs at a


significant rate with small overpotentials.

Electrochemical Equilibria
Thermodynamically reversible reactions will adopt an equilibrium potential which is described by the Nernst
equation:

(example)
Reference Electrodes
Reference electrodes are needed to convert from the charge carriers in the metal (electrons) to the charge
carriers in solution (ions) in a reproducible fashion. They must be practically reversible.
The Normal Hydrogen Electrode (NHE) is used as the (arbitrary) standard. This consists of
hydrogen at unit activity (i.e. solution in equilibrium with hydrogen gas at 1 atmosphere) in
equilibrium with unit activity of hydrogen ions in solution (1.19 M HCl solution). The
equilibrium potential is detected with a platinum electrode that is coated with platinum
black (finely divided platinum) to enlarge the effective surface area.
The NHE is inconvenient for general-purpose use, and a range of secondary reference
electrodes have been developed (Table 1).

Table 1 Practical Reference Electrodes


Common V vs
Electrode
Name NHE
Hg/Hg2Cl2/sat.
SCE +0.241
KCl
Hg/Hg2Cl2/1M
Calomel +0.280
KCl
Mercurous Hg/Hg2SO4/sat.
+0.640
sulphate K2SO4
Hg/Hg2SO4/0.5M
+0.680
H2SO4
Mercurous Hg/HgO/1M
+0.098
oxide NaOH
Ag/AgCl/sat.
Silver chloride +0.197
KCl
Copper
Cu/sat. CuSO4 +0.316
sulphate
Zinc/Seawater Zn/seawater -0.8
Note that zinc in seawater is a useful practical reference electrode, although it has no
theoretical basis for the reference potential.

Electrochemical Kinetics
The Electrochemical Double Layer
There is a tendency for charged species to be attracted to or repelled from the metal-solution interface. This
gives rise to a separation of charge, and the layer of solution with different composition from the bulk solution
is known as the electrochemical double layer. There are a number of theoretical descriptions of the structure
of this layer, including the Helmholtz model, the GouyChapman model and the GouyChapmanStern model.
 
As a result of the variation of the charge separation with the applied potential, the electrochemical double
layer has an apparent capacitance (known as the double layer capacitance).

Kinetics of a Single Electrochemical Reaction


Activation control
An activation controlled reaction is one for which the rate of reaction is controlled solely by the rate of the
electrochemical charge transfer process, which is in turn an activation-controlled process. This gives rise to
kinetics that are described by the ButlerVolmer equation:
While the Butler-Volmer equation is valid over the full potential range, we can obtain
simpler approximate solutions over more restricted ranges of potential:
Large overpotentials - the Tafel equations are obtained (only applicable for irreversible
reactions):

or, more generally

Small overpotentials, assuming A = C = 0.5:


 

Hence

or, in terms of the anodic and cathodic Tafel slopes, ba and bc,

(note that bc above is taken to be positive, as this is commonly assumed; however, b c is


strictly negative, and the divisor in the Stern-Geary equation should be 2.303(ba - bc) RCT )
Mass transport control
This implies fast kinetics, hence the surface reaction is reversible and the potential is given by the Nernst
equation applied to the surface concentrations:
This analysis leads to a limiting current density:

The Kinetics of Multiple Reactions


When multiple reactions are possible, the resultant kinetics are described by the mixed potential theory. This
simply says that the total current in an external circuit is the sum of all of the currents due to the individual
reactions (with anodic currents being positive and cathodic current negative). In free corrosion conditions this
implies that the sum of the negative currents is equal to the sum of the positive currents. When there are only
two reactions of any significance (one anodic and one cathodic), the theory is analogous to that for a single
reaction.
The behaviour can be summarized diagrammatically with the Evans and E-log | i |
diagrams.

The mixed potential theory leads to the concepts of the corrosion potential, Ecorr, and the
corrosion current density, icorr. Ecorr is simply that potential at which the sum of the anodic
and cathodic currents is zero, and it is therefore that potential that the specimen will adopt
in free corrosion. icorr is, in some ways, a more important parameter, as it describes the rates
of the anodic and cathodic reactions. Providing all of the anodic reactions lead to the
oxidation of metal, then icorr is the corrosion rate of the metal (with the current being
related to the rate of loss of metal through Faraday's Law).
The Rate Determining Step
Real electrochemical reactions tend to occur as a sequence of very simple steps. For example, even a very
simple reaction such as hydrogen evolution occurs as two steps, with two alternatives for the second step:

The rate of the overall reaction is controlled by the rate of the slowest reaction, and this is
known as the rate controlling step. This may be an electrochemical reaction (such as Step 1
above) or a chemical reaction (such as Step 2a). Different rate controlling steps will
typically give a different Tafel slope for the reaction and a different reaction order
(dependence on concentration of reactants). Electrochemical measurements may be used to
help to determine the reaction mechanism and the rate-controlling step.
Basics of Instrumentation
The objective of this section is to present brief introductions to the principles underlying
electrochemical measurement techniques. Specific designs for a number of simple
instruments are also presented (in preparation).

Basic Electricity Measurement of Current


Charge Measurement of Small Currents
Current Operational Amplifiers
Potential Non-Inverting Buffer
Potential Difference (Voltage) Properties
Resistance Noise
Capacitance Interference Noise
Inductance Electronic Noise
Measurement of Voltage Frequency Response or Filtering
Measurement of Small Voltages  

Basic Electricity
Charge is the most fundamental concept in electricity. It derives from the properties of
elementary particles, with protons (and hence the nucleus of the atom) being positively
charged, and electrons negatively charged. The unit of charge is the coulomb (C). The
charge on the electron is 1.60219 x 10-19C (i.e. one coulomb corresponds to about 6 x 1018
electrons). An important quantity in electrochemistry, known as Faraday's Constant (or
often just the Faraday) and given the symbol F, is the charge associated with one mole of a
singly charged species such as H+ or Cl-. As one mole contains Avogadro's number (6.0228
x 1023) molecules, the Faraday is 6.0228 x 1023 x 1.60219 x 10-19C, or 96485 C.
Current is the rate of flow of charge along a conductor (note that this charge may be
electrons flowing in a metal or ions flowing in solution). 1 Amp (A) corresponds to a flow
of 1 C/s.
Potential is an indication of the potential energy of a unit charge at a particular point in a
circuit. Strictly it is the potential energy involved in moving a charge of 1 C to that point
from infinity (it is therefore quite difficult to measure!).
Potential difference or voltage is the difference in potential between two points. There is a
voltage of 1 volt between two points if 1 Joule is required to move 1 C from one point to
the other. As with current, potential may apply to charge in the form of ions or in the form
of electrons. However, for a valid potential difference, the charge must be the same at each
location.
Resistance is the tendency of a conductor to obstruct the flow of current. Ohm's Law states
that the voltage (V) across a resistor is proportional to the current (I) flowing through it:
V=IR
where R is the resistance, which has units of Ohm. Like many laws, this is an
approximation, and many conductors, including the metal-solution interface, have a non-
linear resistance.
Capacitance is the tendency of a device incorporating two conductors which are insulated
from each other to absorb charge when the voltage between the conductors is changed. The
charge, Q, is given by:

Q=C V

where C is the capacitance (which has units of Farads) and V the change in voltage. It is
unfortunate that the symbol C is used conventionally both for capacitance and coulombs.
We can see the effect of trying to pass a current through a capacitor by remembering that
current is charge per unit time. Hence:

Q=I t

Inductance
Due to the interrelation between electric currents and magnetic fields, there is a tendency for currents to flow
at a constant rate through a conductor. In most real conductors, this tendency is counteracted by the resistance
of the conductor, although in superconductors, which have no resistance, currents will flow essentially for
ever unless the current is caused to change by the application of a voltage. The inductance of a particular
conductor, L, is a measure of the voltage needed to cause the current to change at a particular rate.

The units of inductance are Henrys (1 Henry will give a rate of change of current of 1 A/s
with 1 V applied across it).

Impedance
Impedance is a general term used to describe the relationship between the voltage across a
component (or essentially any device capable of allowing at least some current to flow) and
the current flowing through that device. It is normally used in relation to alternating
currents with a sine waveform, but it is perfectly valid to refer to the impedance at zero
frequency (i.e. direct current).
As well as their amplitude, it is also important to know the relationship in time between the
current and voltage sine waves. This is described by the phase.
Measurement of Voltage
An ideal voltmeter will measure the voltage between its two input terminals, without in any way affecting that
voltage. The main problem with voltage measurements takes the form of a requirement for current to flow
through the voltmeter. Real voltmeters can be treated as an ideal voltmeter, together with a resistor between
the two input terminals which represents the requirements of the voltmeter for a current to flow in order to
measure the voltage.
Rm in the figure above is referred to in voltmeter specifications as the input resistance or
input impedance and for conventional digital voltmeters or multimeters it will commonly
be 10 MOhm, although it can be up to 1000 MOhm. Electrometers and pH meters are
designed to give a very high input impedance, typically in the region of 1014 Ohm .
The effect of these resistances is to allow current to flow, and if there are high resistances
associated with the voltages being measured (the source resistance), this may lead to
significant errors. For example, if potential of a painted specimen is being measured with a
voltmeter with a 10 MOhm input resistance, and the resistance of the paint film is 1 MOhm,
this will give an error of 10%. If, as is entirely possible, the resistance of the paint film is
108Ohm, the error will be 90%. For most situations, electrometers are unlikely to give
errors of any significance. With modern microelectronic devices, it is very easy to construct
an amplifier that will give an input impedance which is comparable to that of an
electrometer (see Operational Amplifiers).
Measurement of Small Voltages
With conventional instrumentation, the smallest voltage that can be resolved accurately is controlled by the
input offset voltage of the amplifier used, or the inbut bias current of the amplifier flowing through the source
resistance of the voltage being measured. The most stable amplifiers are automatically zeroed by switching
the input between the voltage to be measured and a short circuit. This produces devices with input offset
voltages of less than 1 µV. The best commercial digital voltmeters give resolutions down to 10 nV. Great care
must be taken to minimize noise pickup in such sensitive measurements, and thermal emfs associated with the
interconnection of different metals can lead to significant errors (of the order of 1 µV).
 
Measurement of Current
An ideal ammeter will measure the current flowing between its two terminals while at the same time behaving
as a perfect conductor, thus maintaining the two terminals at the same potential. Real ammeters will create a
potential difference between the terminals, and this may be represented as a resistor, known as the internal
resistance, in series with an ideal ammeter.
Typical digital multimeters measure current by measuring the voltage across a resistor, and
they will usually develop 10 to 100 mV across the resistor for a full scale current reading.
In many electrochemical experiments this is of no consequence because the potential drop
will have negligible effect on the current flowing. For example, in a potentiostatic
experiment the cell current may be measured in the lead between the potentiostat and the
counter electrode, and the potentiostat will provide the extra voltage needed with no
difficulty. The main application of current measurements where it is important to minimize
the potential drop across the meter is in the study of galvanic corrosion and the
measurement of electrochemical current noise. Currents can be measured with essentially
zero potential drop across the meter (< 1 mV) by the use of a current amplifier. This is
available incorporated into an ammeter, in which case it is known as a zero resistance
ammeter, or it can be constructed as an attachment for a conventional multimeter, or a
potentiostat can be configured to operate as a zero resistance ammeter.
Measurement of Small Currents
Current amplifiers provide the most sensitive method of measuring very small currents. In this case the lower
limit to the currents which can be measured is controlled by the input bias currents of the amplifier used, or by
the input offset voltage of the amplifier acting on the source impedance of the current source being measured.
At the time of writing, devices are available with maximum input bias currents of 75 fA (75 x 10 -15A).
Remembering that the charge on the electron is 1.6 x 10 -19C, this corresponds to only about 5 x 105 electrons
per second, or one electron every 2 microseconds. Similar performance can also be achieved with commercial
electrometers which provide very high quality zero resistance ammeters. In order to maintain the very low
leakage currents, it is essential to take very great care with connections to the amplifier input, as leakage
through dirty or poor quality insulators can far exceed the amplifier input leakage currents.
Noise
Noise may be defined as an unwanted signal superimposed on a signal of interest, although the term may also
be used to describe a signal consisting of apparently random fluctuations. The latter may be of interest, as in
the case of electrochemical noise. However, in this section we are concerned with noise as an unwanted
component of a signal, and the ways in which we can reduce noise or cope with it.
We shall typically find two types of noise; 'random' noise with a wide frequency content, which derives from
the properties of conductors and electronic devices, and noise derived from interference from man-made
sources, such as radio-frequency emissions, mains-frequency pick up, spikes due to fridges switching on or
off and so forth.
 
Interference Noise
As far as noise derived from interference is concerned, there are two basic rules to minimize the
amount of noise coupled into the measuring circuits.
 
Screen circuits
The principal of screening is to surround the signal circuits with conductors held at ground potential.
Then electromagnetic radiation will couple with the screening, where it will be harmlessly adsorbed,
rather than the signal circuits. The simplest form of screening is to use screened cables, in which the
signal cable is surrounded by braiding which is connected to ground. This is reasonably effective,
although it is difficult to extend the screening to cover all parts of the circuit, in particular the
electrochemical cell itself.
 
For sensitive measurements a more thorough approach is to surround the entire experiment by a
`Faraday cage'. Essentially this is a grounded conductive box which shields its contents from
electromagnetic radiation. Since the Faraday cage only shields against radiation coming from outside
the box, it is usually best to use only battery-powered instruments inside the Faraday cage, in order
to avoid introducing mains-frequency noise into the cage with mains-powered instruments. Outputs
from these devices should be carried through the wall of the cage by screened cables with the screen
connected to the cage.
 
Avoid signal or ground loops
One of the main `man-made' noise problems is the pick-up of noise at mains frequency. This is most
severe when a loop of wire is exposed to a mains-frequency electromagnetic field, forming what is,
in effect, a single turn transformer. If the loop is complete, large currents can flow through it,
developing significant potential differences between different parts of the loop.

The figgure above show an example of a typical source of problems due to a ground loop. Loops in
ground circuits can be particularly insidious, since instruments frequently have parts of their circuitry
connected to ground.
For example, the working electrode terminal of a potentiostat is commonly
connected to ground, as is the negative input of an oscilloscope. However, if both of
these connections remain in place while trying to monitor the cell potential, this will
establish a ground loop, which will increase the noise in the circuit due to
circulating ac currents in the ground loop. The remedy to this problem is to
disconnect all but one of the connections to ground. In doing this, one must, of
course, pay attention to safety requirements, since the ground connections serve to
protect against hazardous voltages as well as controlling noise.  
Electronic Noise
There are various sources of noise in electronic components. The most fundamental of these is due to
the random motions of electrons in a conductor. The random motion corresponds to a fluctuating
current, and this current will develop a voltage across the resistance of the conductor. For good
conductors, the amplitude of this noise is very small, but for large resistances it can become
significant. The phenomenon is known as Johnson noise, and the rms noise voltage is given by:
In much electrochemistry we are concerned with measurements from dc to about 10 Hz, hence we
have a 10 Hz band-width. If our source resistance is 1 MOhm, this will give an rms noise of
approximately 0.4 V. This would usually be tolerable, but we should be aware that noise of this
level is unavoidable. From the equation, it can be seen that Johnson noise will create the most severe
problems when we wish to measure over a wide frequency range (increasing b) with a very high
source impedance.
 
In 'real' electronic components other forms of noise are possible. When a current is flowing through a
circuit there will inevitably be a level of shot noise as a result of the quantized nature of electrical
charge. Additionally, semiconductors are subject to flicker noise, which gives particularly strong
noise at low frequencies.
 
Frequency Response and Filtering
In many electrochemical measurements, we are concerned with very low frequency measurements,
and there is a tendency to ignore the complications which result from the limited frequency response
of the instruments being used. However these may become significant in some circumstances.
 
Frequency determining components tend to take the form of resistor-capacitor combinations, such as
those shown below:

These two configurations are described as low and high-pass filters on the basis of those frequencies
which the filter allows to pass through it. Considering the low-pass filter, the impedance of the
resistor will be constant at R, while the impedance of the capacitor will be 1/2 fC. These two
impedance's will act as a potential divider, giving:
 

 
For frequencies very much less than ½ RC, Vout will be approximately equal to Vin, whereas for
frequencies very much greater than 1/2 RC, Vout will be given by Vin/2 RC. This may be seen by
plotted Vout/Vin on a logarithmic plot of Vout/Vin vs. frequency (see below). Vout/Vin may be described
as the gain of the filter and is frequently described in units of decibels (dB). These are slightly odd
logarithmic units which happen to be convenient in signal processing. One bel (which for some
reason unknown to us is never used as a unit in its own right) corresponds to a factor of ten increase
in the output power of a circuit as compared to the input circuit. Assuming a constant load resistance,
the power is proportional to the square of the voltage, hence a gain of 20 dB corresponds to the
output voltage being ten times the input voltage.
Then the gain of a low-pass filter is zero dB at low frequencies. At f = 1/2 RC, the gain will be ½
(since the impedance of the resistor and the capacitor are equal at this frequency), hence the gain will
be -6 dB (20 x log 0.5). At higher frequencies the gain will fall by a factor of 10 for each decade of
frequency, i.e. -20 dB/decade of frequency (also described as -6 dB/octave).
Operational Amplifiers
Modern microelectronic technology has made a wide range of extraordinarily sophisticated devices available
to the engineer. Fortunately it has also made some of the simpler devices very easy to use, and you do not
need a degree in electronic engineering to produce some very useful results. The basis of many useful devices
is a class of general-purpose amplifiers called operational amplifiers. These were originally developed as
building blocks for analogue computers, but they have since proved to be extremely versatile devices, and
will be found in most instruments. Because of their wide applicability, they are also available very cheaply,
with useful devices costing less than £1. A typical operational amplifier has five connections:

 
Two of these connections supply the power to the amplifier (shown here as the common values of +15 and
-15 V, although other voltages are also used); two give inputs to the amplifier, while one provides the
amplifier output. The operational amplifier is a voltage-controlled device. The output voltage (measured
relative to the mid-point of the power supply voltages) is a function of the difference between the two input
voltages according to the equation:

where A is known as the gain of the amplifier, and is very large. Ideally the gain is infinite; practical gains are
in the region of 105 to 107.
When drawing circuits using operational amplifiers, it is common to omit the power supply connections for
clarity, but these will always be required.
 
The Non-Inverting Buffer 
An amplifier with such a high gain is of no use as it stands, but it proves to be very easy to modify the
characteristics of the amplifier by connecting the output back to one or other of the inputs. The simplest
example is obtained by connecting the output directly to the non-inverting input:

Thus with this configuration, the output voltage will be almost exactly the same as the input voltage. Again
this may not seem terribly useful until we examine the properties of the operational amplifier in a little more
detail. For the ideal operational amplifier, the input impedance at both the inverting and the non-inverting
inputs is infinite (i.e. no current will flow into the input terminals). Conversely the output impedance is zero
(i.e. the voltage at the output will remain the same whatever current is drawn from the output). Now we can
see that this circuit can be very useful. For example, if we have a voltmeter with an input resistance of 10
MOhms, and we wish to measure the potential of a painted specimen, we have seen that we are liable to get
measurement errors due to the high resistance of the paint film. If we place our circuit between the lead from
the reference electrode and our voltmeter, we shall be able to make accurate measurements, since the circuit
will accurately reflect the reference electrode potential without allowing any current to flow.
 
So far we have been simplifying the discussion slightly by talking in terms of the properties of an ideal
amplifier. In reality, of course, we can only approach the ideal, but in many cases, developers of
microelectronic devices have got remarkably close to the ideal. The properties of typical devices are shown in
Table 2
 
Table 2 Ideal and typical operational amplifier properties
Ideal Typical device Best device
6
Gain 10 5 x 106
Input Offset Voltage 0 25 V 0.7 V
Input Offset Voltage
0 0.6 V/oC 0.01 V/oC
Temperature Drift
Input Bias Current 0 40 nA 75 fA
Input Bias Current
0 < 1 nA/oC x 2/10oC
Temperature Drift
Output Impedance 0 70 Ohms 0.1 Ohms
Maximum Output Current 17 mA 15A
Maximum Output Voltage 10 V 145 V
Unity Gain Band-width 8 MHz 5 GHz
Slew Rate 1.9 V/ s 500 V/ s
Notes to Table 2
The typical values are actually those for a low-cost device, the OP-27, which is designed for low noise and
low input offset errors. The best values are the best that we happened to find in the Burr-Brown and RS
components catalogues. The values are not, of course, all for the same device, and a device which is good for
one thing is liable to be worse for another. Specifications for modern devices are available on manufacturer's
Web sites:
Analog Devices Burr-Brown Texas Instruments
National Seminconductors Motorola Linear Technology
Maxim Harris
Operational Amplifier Properties
 
Gain
The ratio of the output voltage (relative to the mid-point of the power supply) to the difference
between the voltages at the non-inverting and the inverting inputs. The gain is specified for dc
voltages (it is typically constant below about 1 Hz), and the gain for ac signals will decrease as the
frequency increases.
Input offset voltage
The difference in voltage between the two input terminals when the output is actually zero (this can
often be adjusted to near zero if required).
Input offset voltage drift
The change in input offset voltage as a function of temperature.
Input bias current
The current that flows into either of the input terminals.
Input bias current drift
The variation of the input bias current with temperature. The nature of this will vary according to the
technology used for the device. Bipolar devices will typically provide a drift that is reasonably
constant over a moderate temperature range, and this will be specified as A/ oC. Devices based on
field-effect transistors will typically have an input bias current drift that doubles about every 10oC.
Output impedance
The effective impedance at the output (this may be thought of as a resistor between the output of an
ideal amplifier and the real amplifier). It is not usually very important as the feedback circuits tend to
compensate for any errors, although it will place one limit on the maximum current that the device
can deliver.
Maximum Output Current
Standard low-cost amplifiers will usually deliver at least 10 mA without problems. Larger current
can be obtained using power amplifiers. These will often require heat-sinks to prevent the device
overheating when delivering a large current.
Maximum Output Voltage
Standard low-cost amplifiers will usually provide an output voltage of +/- 10 V without problems
when operating from +/- 15 V power supplies. Higher output voltages require special-purpose
amplifiers or discrete transistor output stages.
Unity gain band-width
The gain of a typical operational amplifier is inversely proportional to frequency. The unity-gain
band-width is that frequency at which the gain has fallen to one, and indicates the frequency
response of the amplifier.  
Slew rate
This is the rate at which the output voltage will change when the input is instantaneously changed by
significant amount.
Operational Amplifier Circuits

Unity-gain non-inverting buffer

The ouput of this circuit will be the input voltage. The input impedance is very high (the
effective input impedance of the amplier is increased by the feed-back circuit), while the
output impedance is very low (the output impedance is similarly reduced by the feed-back
circuit). The main errors are associated with the input offset voltage of the amplifier,
together with the effects of the input bias current when using high source impedances.

Non-inverting voltage amplifier

By reducing the proportion of the output voltage fed back to the inverting input, the gain of
the amplifier can be increased. For the circuit shown, the gain will be R2/(R1+R2). Gains
up to about 1000 can be obtained reasonably easily, although the input offset voltage of the
amplifier is magnified by the gain, so this may become a significant voltage when large
gains are used. Note that the ground connection shown in the figure is the reference point
for the measurement of voltages.
Differential voltage amplifier

Note that the two resistors labelled R1 (and similarly the two labelled R2) have the same
value. The ouput voltage is given by the difference between the + and - inputs times the
gain, where the gain is R2/R1. One limitation of this circuit is the low input impedance
(approximately R1+R2), and if this presents a problem an instrumentation amplifier should
be used. Gains of up to about 1000 may be used without problems.

Instrumentation Amplifier
An instrumentation amplifier is essentially a differential voltage amplifer with a high input
impedance. It can be constructed using two or three operational amplifiers, but it is simpler
to use a ready-made device. An example is the Analog Devices
Current amplifier

In the current amplifier, the current at the -ve input terminal must be supplied from the
amplifier output via R1 (since the input to the amplifier has a very high impedance). Also,
the action of the amplifier will be to deliver sufficient current such that the voltage at the
inverting input is the same as that at the non-inverting input. Hence there will be no
difference in potential between the two inputs (except for the input offset voltage of the
amplifier), and the output voltage will be IinR1 (from Ohm's Law). The maximum
sensitivity of the current amplifier is limited by the input bias current of the amplifier or
(probably more commonly) the effect of the input offset voltage of the amplifier on the
source resistance of the current source.
Potentiostat

In the simple potentiostat circuit, the voltage at the counter electrode terminal will be
maintained by the amplifier such that the voltage at the reference electrode input is the
same as that at the non-inverting input of the amplifier. Hence, the potential of the
reference electrode relative to the working electrode will be the same as the control voltage.
Note that this circuit connects the working electrode to the ground of the system power
supply. Some care may therefore be needed to avoid ground loops or similar problems
during the measurement of cell potential and current.

Galvanostat

 In the galvanostat, the output current will flow through the external circuit and then
through the resistor R. The output voltage will be controlled such that the voltage across R
is equal to the control voltage, hence controlling the current. The main limitations of this
circuit are
 the total voltage available will be limited by the maximum output voltage of
the amplifier (this is not usually a problem for electrochemical work),
 the current is limited by the maximum output current of the amplifier, and
 the control of very small currents will become limited by the input bias
current of the amplifier and/or the action of the input offset voltage on the
resistor R.
Active filter
There are many forms of active filter, but in general the amplifier is used to compensate for
the effect of the impedance of passive frequency-determining elements. The simplest active
filter consists of a resistor-capacitor high- or low-pass filter followed by a unity-gain non-
inverting buffer. This gives what is known as a first order response, with the gain above or
below the cut-off frequency being proportional (or inversely proportional) to frequency.
Addition of a second resistor and capacitor gives a second order filter, with gain
proportional to frequency2. Note that changing from a high-pass to a low-pass filter
essentially just involves swapping the resistors and capacitors.
The frequency response for active filters of orders 1 to 4 are shown below. The response for
high pass filters will be essentially the same, but reflected about the y axis.

 
Cell Design for Electrochemistry
Working Electrode
Counter Electrode
Reference Electrode
Composition
Solution Flow
Rotating Disk Electrode
Rotating Cylinder Electrode
Flow Channels

A typical electrochemical cell consists of three electrodes (working, counter and reference) together with
features for the control of solution flow and state of aeration.
 

Diagramic electrochemical cell


 
The Working Electrode
The first stage in the preparation of the working electrode should be to characterize the material in terms of its
chemical composition and metallurgical condition. The specimen must then be mounted in some way to
expose it to the solution. This presents a number of difficult problems, including the presence of gas-solution
interfaces or crevices between the mounting system and the specimen.
Whatever mounting method is used it is important that there is a sound electrical
connection to the specimen, preferably achieved by spot-welding, soldering or bolting. The
point of connection and connecting wire must be isolated from the solution, either by way
of the specimen mount, or by keeping it out of the solution.
Crevices are particularly damaging for passive specimens, where a very large error in the
observed passive current density may be obtained. Typical methods that have been used to
avoid crevice problems include PTFE compression seals (originally proposed by Stern and
Makrides), epoxy resin mounting with careful pretreatment of the metal and the flushing
with distilled water of a filter paper filled crevice (the Avesta cell).
The surface treatment of the specimen is also important. This should be reproducible and
relevant to the intended application, although these two requirements are often
incompatible. Typical surface conditions include as received (this is often very relevant, but
not very reproducible), ground and polished. Cut edges will often behave differently from
the surface and should be avoided if possible (unless it is intended to study the behaviour of
the edge).
Electrochemical reactions are very sensitive to surface contamination, and it is important to
clean the specimen before testing, particularly to remove grease. However, care must also
be taken to avoid contamination of the surface with species extracted from mounting
materials (e.g. epoxy resins) by powerful solvents such as acetone.
The Counter Electrode (or Secondary or Auxiliary Electrode)
The counter electrode must be inert in the conditions to which it is exposed in order to avoid contamination of
the solution. Typically platinum or graphite are used. It is often specified that the counter electrode should
have a large area compared to the working electrode, but this may not be necessary in many cases. The
fundamental requirement is that it be able to pass sufficient current into the solution without needing an
excessive cell voltage or creating a non-uniform current distribution on the working electrode.
 
The Reference Electrode
The Saturated Calomel Electrode (SCE) is the most commonly used in corrosion studies. The Ag/AgCl
electrode is very convenient for small, cheap electrodes (but it is chloride sensitive). Many corrosion systems
are very sensitive to chloride, and the Hg/Hg 2SO4 electrode is often useful to avoid chloride contamination.
Recent legislation on the transport by air of objects containing metallic mecury has led to a move from the
SCE to Ag/AgCl.
A Luggin Capillary is used to shield the potential measurement from potential drop in
solution, such that the potential is effectively measured at the tip of the capillary. This
should therefore be close to the working electrode, but not so close as to shield part of the
surface from the current flow (the ideal is about 3 x the tip diameter of the capillary).

The Solution
Composition
The solution composition, including aeration and pH, must be well-defined and controlled. Similarly the
solution temperature should be controlled, or at least measured, as the rates of electrochemical reactions are
typically very temperature dependent.
 
Solution Flow
For systems that may be subject to mass transport control it is also important to control the flow of solution
past the electrode. Various methods are available to achieve this. For systems with only a moderate tendency
to mass transport control, it may be sufficient simply to stir the solution or to rely on agitation from the gas
bubbling used to control the oxygen concentration of the solution. These methods will generally give
reasonably good mass transport, but the exact conditions will be difficult to define and reproduce.
 
In order to provide well-defined and reproducible flow conditions a number of techniques are available:
The rotating disk electrode (RDE) consists of a disk on the end of an insulated shaft that is rotated
at a controlled angular velocity. Providing the flow is laminar over all of the disk, the mathematical
description of the flow is surprisingly simple, with the solution velocity towards the disk being a
function of the distance from the surface, but independent of the radial position. Consequently the
mass transport conditions are uniform over the surface of the disk, and the limiting current is given
by:

The rotating cylinder electrode provides more uniform behaviour over the electrode surface than
the RDE when turbulent flow conditions are used (the central region of the RDE is always laminar).
Flow channels have the advantage that the electrode remains stationary, and is therefore easily
examined during the experiment. The design of a flow channel to achieve well-defined flow
conditions requires considerable care. In particular the electrodes must be accurately aligned with the
wall of the channel, with no gaps or steps that could induce turbulence, and the flow pattern must be
well-established before reaching the electrodes, which requires a long entry region to allow the flow
to stabilize.
Electrochemical Instrumentation
Electrochemical techniques generally involve the relatively simple operations (in electronic
terms) of measurement or control of voltage or current, and in this chapter we shall discuss
the principles and practice of these operations.

Measurement of Voltage
Measurement of Current
Control of Potential
Connections
Auxiliary Electrode
Reference Electrode
Working Electrode
Earth
Reference Electrode Guard
Count Resistor
External Control Input
Overload Indicator
I R Compensation
Common Problems
Oscillation
Noise Pickup
Electronic Noise
Overloading
I R Compensation
Control of Current
Galvanostat

1. The measurement of electrochemical potentials


Our most common requirement for voltage measurement in electrochemistry is for the
determination of the electrode potential. For steady-state (or dc) measurements this really
only presents one problem, that due to the finite, and often quite significant, resistance of
the components present in the measurement circuit. This resistance results from several
parts of the circuit:
 charge transfer resistance of both the reference electrode and the working electrode,

 resistance of the solution path between the two electrodes (especially for dilute solutions or
when one or other of the electrodes is covered with a high resistance film), and

 occasionally, the resistance of poor electron conducting electrodes.

An ideal voltmeter will measure the voltage without drawing any current, hence any
resistance in the measuring circuit will be ignored (since no current is flowing, the voltage
drop across the resistance will be zero - see voltage measurement for a detailed discussion
of this). Typical modern voltmeters have an input resistance in the range 10 to 1000
MOhms. The bottom end of this range (which is by far the most common input resistance
for low-cost meters) will occasionally be too low for electrochemical measurements, while
the high end of the range will usually be adequate, except for very high resistance
electrodes such as painted samples and glass pH electrodes. For measurement with very
high source resistances, electrometers or pH meters (both of which have input resistances in
the region of 1014 Ohms) can be used. Alternatively low-cost buffer amplifiers can be
constructed to increase the input resistance of standard voltmeters.
2. The Measurement of Galvanic Currents
It is an assumption in the measurement of galvanic currents that there is a negligible potential difference
between the two metals. This is not always strictly correct, as the current will often produce a significant
potential drop across the solution resistance, but it is still desirable to minimize the potential drop across the
current measuring instrument. Conventional ammeters, especially the ammeter ranges on multimeters, are
rather poor as current measuring instruments, with a rather large potential drop across the meter (up to 100
mV) when measuring full scale current. A simple solution to this problem is to use a sensitive voltmeter or
chart recorder to measure the voltage across a resistor chosen to produce a potential drop of about 1 mV.
3. Control of Potential

Figure 3 Control of potential by connection to stable reference electrode


 
The simplest method of controlling the potential of an electrode (generally called the working electrode) in
solution is to connect it to a large secondary electrode which naturally adopts the required potential (Figure
3). Thus the potential of a piece of steel in seawater can be controlled at approximately -1050 mV with
respect to the saturated calomel electrode by connecting it to a large piece of zinc. This approach is limited by
the fact that the current that flows causes the potential of the secondary electrode to change as well as
modifying the potential of the working electrode. The available potentials are also limited by the availability
of suitable low impedance secondary electrodes. This limitation can be overcome by using an external voltage
source to introduce a potential difference between the working and secondary electrodes (Figure 4). This can
be a very useful approach for systems that require low currents to be passed through the cell, but for higher
currents the effect of the current on the potential of the secondary electrode remains a problem.
Figure 5 Manual three electrode potential control circuit
 
In order to overcome this problem, separate electrodes must be used: the counter electrode
(also known as the auxiliary electrode or the secondary electrode) to supply the current,
and the reference electrode to measure the potential of the working electrode (the two-
electrode potential control technique of Figure 4 is often described as the use of combined
reference and counter electrodes). The current can be controlled with a simple manual
circuit (such as that of Figure 5), and this will allow reasonably stable control of potential
for some electrochemical systems.

Figure 6 Control of potential for a polarization curve showing an active-passive transition.


However there are systems for which it is nearly impossible to control the potential in a
stable fashion by using this approach. A typical example of this is the classical active-
passive transition shown in Figure 6.
The manually controlled current source will behave like a voltage source in series with a
resistor, and this will give rise to a load line similar to that shown in Figure 6. As the
resistance is decreased and the current increases, the potential will be defined by the
intersection of the load line and the active part of the polarization curve. In this region, the
potential control will be reasonably stable. However as the potential passes the active peak
it becomes impossible to control the potential, which will jump to a much higher value.
Looking at Figure 6, it can be seen that the ideal load line to permit stable control of
potential will be essentially horizontal (i.e. the potential will be the same, irrespective of the
current flowing), and this is the objective of the potentiostat. At a minimum, a potentiostat
has three connections for the working electrode, the reference electrode and the counter
electrode. The potential between the working and reference electrode is monitored and if it
is too small, the voltage between the counter electrode and working electrode is increased,
whereas if it is too large, the voltage is decreased. The overall result of this is to maintain a
constant voltage between the working and reference electrodes. Practical potentiostats will
typically have several additional connections and controls. Figure 7 shows the front panel
of a typical potentiostat.

Figure 7 Hypothetical potentiostat front panel


Connections
Auxiliary Electrode
Also known as counter electrode or secondary electrode. This supplies current to the cell and is
typically connected to a platinum electrode.
 
Reference Electrode
Connects to the reference electrode to monitor cell potential. This connection should have a very
high input impedance, and some care should be taken to avoid compromising this by leakage through
surface contaminants on connectors, etc.
 
Working Electrode
Connects to the working electrode. There are often two working electrode connections, in which case
one is used to supply current to the working electrode while the other is used to monitor its potential.
By taking separate leads to the working electrode, the effect of potential drops due to current flow
through leads and contacts can be eliminated.
 
Ground or Earth
Typical potentiostat circuits result in the working electrode being connected to 0 V of the power
supply. This is often connected to ac mains ground for safety reasons, but this may be left
unconnected, with a separate ground connection on the front panel. It is best to make this connection,
unless other parts of the measuring circuit (e.g. chart recorder or oscilloscope inputs) are already
connected to ground (even if the ground connection on the other instrument is also connected to the
working electrode do not connect the latter to ground on the potentiostat, as this will lead to a
ground loop.
 
Reference Electrode Guard
Because the reference electrode may have a very high input impedance, leakage of current through
insulation in cables and connections can be significant. This problem can be minimized by
surrounding the cable with a screen at the same potential. This is the function of a guard connection.
This is a relatively rare facility on commercial potentiostats.
 
Count Resistor
The count resistor is a resistor placed somewhere in the current path such that the current can be
measured by monitoring the voltage across the resistor. In many potentiostats a range of resistors are
used to provide an indication of the current on an internal meter, and the voltage across the resistor
may be made available as an output. Alternatively there may be provision for a user selected resistor
to be placed in the current path by connecting it between two terminals on the front panel. This will
often take the form of a "dummy" counter electrode terminal, with no internal connection (except for
a possible connection to the current measuring circuit), such that a resistor can be connected from the
real output to the dummy terminal which is then connected to the cell. Whatever method of current
monitoring is used, you are advised to check the accuracy of measurement occasionally with an
ammeter in the lead to the counter electrode, as it will often be found that the measuring resistors
have been damaged by the inadvertent passage of high currents.
 
External Control Input
In addition to the manual adjustment normally available with the front panel control, this input
makes it possible to control the potential with an external voltage input, such that potential ramps or
transients can be applied to the specimen.
 
Overload Indicator
This indicator (which is often omitted from cheaper potentiostats) will light when the potentiostat is
overloaded, and will not be controlling the potential correctly (see below).
 
IR Compensation
Some of the more sophisticated potentiostats may have facilities for compensation for potential drops
due to the resistance of the solution between the point at which the potential is measured and the
surface of the specimen. The methods available are discussed further below.
Common problems
Modern potentiostats should be very reliable devices if well-designed, but they are
susceptible to some inherent problems, and it is all too common to find a perfectly good
potentiostat stacked in the corner labeled DEAD because a user was unable to recognize
one of these problems.
Oscillation
The potentiostat relies on negative feedback of the reference electrode voltage to produce the counter
electrode voltage, i.e. if the reference electrode voltage goes down by a small amount, the counter
electrode voltage will go up by a large amount. However, if there is a delay in the transmission of the
counter electrode voltage to the reference electrode, there will exist a frequency at which the voltage
from the counter will produce a signal at the reference which reinforces the oscillation. Such delays
inevitably exist due to the time taken to charge stray capacitances. Under normal circumstances, the
necessary frequency is so high that it is not amplified by the amplifiers in the potentiostat, but
sometimes this is not the case, and the potentiostat can burst into oscillation. This can be detected by
monitoring the cell voltage (the potential between the counter electrode and the working electrode)
or the cell current with an oscilloscope or an ac voltmeter (do not monitor the potential between the
reference and working electrodes, as the change in impedance at the reference input caused by the
connection of the oscilloscope can modify the behaviour, and possibly stop the oscillation). The
problem is often associated with long connecting wires, particularly to the reference electrode, and
with a high resistance in the reference electrode connection. A crude but effective solution, if you are
not concerned about high frequency measurements, is to connect a capacitor (not electrolytic) of the
order of 1 µF between the counter electrode and the reference electrode terminals. This produces a
strong negative feedback at high frequencies and damps out the oscillation. If you need to maintain a
high frequency response in a system which seems prone to oscillation, you may find it useful to
switch to galvanostatic operation, as this is less liable to this problem.
 
Noise Pickup
A problem which often appears to be similar to oscillation is the pickup of ac noise, especially at the
high impedance reference electrode input. Most problems are experienced with electromagnetic
noise due to interference from ac mains-powered equipment, but local radio frequency heating units
may also cause problems. The problem can be distinguished from oscillation by the frequency
(which is usually 50 or 60 Hz for mains frequency interference, rather than tens or hundreds of
kilohertz for oscillation), and by the fact that the frequency will be fixed, even if the cell
characteristics are changed (e.g. by moving the Luggin probe further from the working electrode).
Owing to the high rejection of mains frequency noise by modern electronic meters, it is easy to
overlook the presence of quite high amplitudes of ac interference, and it is probable that very many
supposedly potentiostatic experiments are actually made with significant superimposed currents at
mains frequency. Again you should observe the counter electrode to working electrode voltage or the
cell current with an oscilloscope to check for this possibility (although the noise pickup is
predominantly at the reference electrode input, the amplitude of any potential fluctuations will
appear small because the potentiostat will superimpose an ac component on the cell current to try to
remove the noise).
 
Electronic Noise
Where very sensitive measurements are being made it is possible that the internally generated noise
produced by the electronic components within the potentiostat may produce a significant
interference. Modern instruments should have noise characteristics that approach the theoretical
limits, but this may not be the case for poor quality or older instruments. See Noise.
 
Overloading
The potentiostat is an electronic amplifier, and as such there are limits on the voltage and current that
can be supplied to the cell. The values available will depend on the particular devices but you should
always monitor the actual cell potential rather than assuming that it will inevitably adopt the set
value. Overloading is typically revealed as a constant potentiostat output current. As electrochemical
systems rarely give exactly constant current, you should always suspect overloading if the cell
current remains completely constant. Limited voltage output capability is most likely to be a problem
in corrosion investigations, particularly in high resistance solutions, while limited current capability
can be a problem with electroplating studies. In either case reducing the area of the working
electrode will help to improve matters, while reducing counter electrode to working electrode
separation will help to reduce the cell voltage required.
 
IR Compensation
The IR drop between the working electrode and the tip of the Luggin probe is not strictly a
potentiostat problem, but it is convenient to deal with it here as it is a source of significant error in
many experiments. One approach to IR compensation is to ignore it at the time of taking the
measurements, but to correct the measured potentials to take account of the measured or calculated
IR drop in the solution. This is a perfectly valid approach, but it does have some limitations. The
effect of the solution resistance is to produce an error in the potential that is proportional to the
current flowing, and this gives an effective load line such as that shown in Figure 8. This results in a
similar problem to that which we observed with attempts to measure the polarization curve with a
power supply and resistor, namely that it is impossible to access the region of the curve immediately
above the active-passivation transition. A more subtle problem arises from the fact that the changing
IR drop results in a changing sweep rate, so that a polarization curve which is nominally measured
under constant sweep rate conditions will actually be measured with a varying sweep rate.
 

Figure 8 Load line for an IR compensating potentiostat


The basic approach by which you can compensate for IR drops while a measurement is
being performed is to modify the set potential to allow for the solution resistance. There are
two common methods used to make this compensation. The simplest is to pass the cell
current through a resistor of the same magnitude as the solution resistance and add the
resultant potential on to the set potential. There are various ways in which the appropriate
value of resistor can be estimated. These include calculation from the cell geometry and the
resistivity of the solution or measurement with a high frequency ac signal. However,
probably the most common method is to use the fact that the IR compensation is a form of
positive feedback. If the value of the compensating resistor is significantly greater than the
IR drop, the positive feedback will cause the potentiostat to oscillate. Thus the resistance is
adjusted to that value which just fails to cause oscillation, and this is taken to be a
reasonable approximation to the correct value (a typical estimate is that this technique will
give compensation to within 10 to 20% of the correct value).
The alternative method of compensating for IR drops is based on the premise that if the
current flow is switched off, the potential due to the IR drop, being purely resistive, will
disappear immediately, whereas the interfacial potential will remain steady for a reasonably
long time, largely due to the charge stored in the interfacial capacitance. Thus, if the current
flow is briefly interrupted (typically for periods of the order of 1 ms) the IR drop may be
determined and added to the control potential. This is known as the interrupter method of
IR compensation, and should be able to compensate for IR drop to within a few percent,
although the design of the electronics presents significant problems.

Design of Potentiostats
The basic potentiostat is a relatively simple device, and thanks to modern microelectronics,
one can be constructed for a few pounds.
The essential principle of the potentiostat is to produce a voltage at the counter electrode
terminal that is given by some control voltage, minus the voltage at the reference electrode
input, multiplied by a large factor. The operational amplifier has just this capability, and
this leads to the simple circuit shown in Figure 9.
 

  Figure 9 Simple Operational Amplifier Potentiostat


Standard operational amplifiers can deliver about ± 10 mA, and can be powered by a pair
of 9 V batteries. Higher power devices can provide several amps, but these require much
more attention to power supply design, heatsinking, etc.
The simple potentiostat circuit does suffer from some limitations: it is not possible to
provide a separate potential sensing input for the working electrode; only one control
potential input is available; it is difficult to control the frequency response of the
potentiostat to limit oscillation problems; it does not provide a low impedance output of the
potential between the working and reference electrodes and high output current operational
amplifiers tend to have rather poor input characteristics (input offset voltage, bias current
and impedance). There are as many ways of circumventing these limitations as there are
designers, but a possible design is shown in Figure 10.

From John Gill


Figure 10 Typical Practical Potentiostat Design
4. Control of Current
The control of current is similar to the control of potential, although it tends to be slightly easier to achieve. In
order to control current it is inevitably necessary to make a second connection to the solution by way of a
counter electrode. In order to drive current between the working and counter electrodes it is necessary to
provide an external power supply, and the simplest circuit is as shown in Figure 11. The current flowing is
controlled by the value of the resistor, R, and the voltage provided by the power supply less that dropped
across the cell. By keeping the supply voltage high, the effect of the cell voltage can be minimized, and good
current control can be achieved. Typical changes in cell voltage will be of the order of a few tens or possibly
hundreds of millivolts, and reasonably stable current control can be achieved with a ten volt power supply,
without requiring unduly high power dissipation in the resistor R. If the power supply voltage is increased to
give better current control, the power dissipation in the resistor will increase in proportion to the voltage
(assuming the current remains the same), and this may become a problem. Additionally the use of high
voltage supplies (greater than about 100 V) presents a significant safety hazard.

Figure 11 Simple current control circuit


 
The Galvanostat
The galvanostat is an electronic instrument that overcomes the limitations of the simple power supply plus
resistor approach to current control. It is inherently simpler than the potentiostat (at least from the user's point
of view), as it only has two connections, and will drive the set current through the external circuit connected
to these terminals (providing the required voltage is within the capabilities of the instrument).
Standard laboratory power supplies will often have a constant current mode of operation,
and these provide a low cost galvanostat. Note, however, that for many power supplies the
current control is of secondary importance to the control of voltage and is not very stable;
you should check the specifications carefully before buying a power supply for use as a
galvanostat.
You should also be aware that some constant current power supplies are able to deliver a
very high voltage. These can be extremely dangerous, since the full output voltage will
appear at the output terminals if the load is disconnected, and they can deliver a lethal
shock.

Figure 12 Connection of potentiostat for Current Control


We can also produce a constant current using a potentiostat with the circuit of Figure 12.
The potentiostat controls the voltage, V, across the resistor R, and by Ohm's Law this
implies that a current of R/V will be flowing from the counter electrode terminal, through
the external circuit and then through R.
This configuration also forms the basis of the simple operational amplifier galvanostat.
Steady-State and Potential Sweep Methods
Generation of Potential Ramps
Motorized Petentiometers
Electronic Waveform Generators
Analog
Digital
Hybrid Digital/Analog Sweep Generators
Measurement of Polarization Curves
Potentiodynamic Sweep Method
Potential Step Method
Multiple Specimen Method
Measurement of Linear Polarization Resistance

5. Generation of Potential Ramps


Several electrochemical experiments require a slow potential or current sweep, and these are typically
obtained by using a voltage sweep to control a potentiostat via the external control input (in some
potentiostats a potential sweep generator may be incorporated into the instrument, but the principle is
essentially the same). There are a number of methods which are commonly used to generate potential ramps
for this purpose:
 
Motorized potentiometers
By applying a fixed voltage across the end terminals of a potentiometer and moving the slider from
one end of the potentiometer to the other, the slider voltage will vary from one potential to the other,
with the potential being a linear function of the slider position. By driving the slider with a motor of
the appropriate speed a smooth ramp can be obtained. Additional controls may be incorporated to
cause the motor to reverse direction at the end of the sweep, and by manually moving the slider to
the desired position the sweep can be started at any voltage desired. This approach has a limited
maximum sweep speed (typically tens of seconds for a full sweep), and the motorized potentiometer
tends to be have a limited life if heavily used. However it does provide a relatively smooth sweep
(there will typically be very small steps as the slider moves from one turn of a wire-wound
potentiometer to the next), and can operate down to indefinitely low sweep speeds (with suitable
motor gearing).
 
Analog Electronic Waveform Generators
These instruments typically generate a sweep by charging a capacitor with a constant current, and by
modifying the current the sweep rate can be changed. Additional circuit elements allow for reversal
of the direction of sweep at the required voltage, and (in some instruments) for starting at an
appropriate potential. This approach can provide very rapid sweep rates (of the order of µs for a full
scale sweep), with a very smooth sweep. However drift problems make it difficult to achieve low
sweep rates accurately (a typical lower limit is of the order of an hour for a full scale sweep), and it
is difficult to hold a fixed potential.
 
Digital Electronic Waveform Generators
The simplest digital waveform generator consists of a counter, which is set to count up and then
down, followed by an analog-to digital convertor, which converts the number produced by the
counter into a voltage. This can be made very much more versatile by interposing a memory between
the counter and the convertor, or by using a computer or dedicated microprocessor to generate
numbers for the convertor. Both of these approaches allow arbitrary waveforms to be generated. This
approach has the advantage of a wide frequency range, with the output being stable from dc (i.e. the
potential can be held at a specified value indefinitely) up to hundreds of hertz (for computer control)
to tens of kilohertz or greater (for waveform generation via a dedicated memory). As the analog-to-
digital convertor produces only a discrete set of numbers, there will inevitably be small steps in the
voltage generated. The size of these will depend on the quality of the analog-to-digital convertor,
partly because a better-quality convertor will have a larger number of discrete output values possible,
and partly because of convertor errors which can result in large steps for changes between particular
numbers. The size of the steps will also be larger at faster sweep rates due to limitations in the rate at
which new voltages can be generated, especially with computer-controlled systems.
 
Hybrid Digital/Analog Sweep Generators
The optimum sweep generator properties (smooth sweeps, a wide frequency range, stability at low
frequencies, the ability to maintain a steady voltage and flexible control of sweep rate) can be
achieved by using a combination of analog (to generate the smooth sweep) and digital (to provide the
control and the stability at low frequencies) techniques, although at the time of writing few
instruments are available commercially with these capabilities.
The Measurement of Polarization Curves
A polarization curve is the quasi steady-state plot of current as a function of potential,
typically plotted as log |i | vs E (where |i| means modulus of current, i.e. the amplitude of
the current with a positive sign, so that the logarithm can be taken). By quasi steady-state
we imply that the curve is measured sufficiently slowly that the current flowing at each
potential has reached an essentially constant value. There are three methods to achieve this:

Figure 13 Measurement of Polarization Curve with swept Potential


The potentiodynamic sweep method
In this method the potential is swept slowly from one potential to another. Typically sweep rates are
of the order of 1 mV/sec. For this method a sweep generator is used in conjunction with a
potentiostat and recorder, as shown in Figure 12.
 
The potential step method
This is similar to the potential sweep method, except that the potential is changed in a series of steps,
typically 5-50 mV per step, with a hold at each step to allow the current to reach steady-state
conditions. This has the advantage of requiring less instrumentation (the steps can be obtained by
manual adjustment of the potential), but the results are in the form of the series of discrete points,
rather than a continuous curve.
 
The multiple specimen method
A problem of the preceding methods is that the results obtained at a particular potential tend to be
affected by the previous measurements. This can be avoided by using a freshly-prepared specimen
for each potential. This makes it a very time-consuming method, but it can be valuable in some
situations.
The fact that the same specimen is being used in the first two methods above for
measurements under a range of conditions presents the problem that the results obtained are
a function of the order in which measurements are made. In general the best results will be
obtained if those potentials liable to change the specimen surface least are measured first.
Typically this implies that the cathodic branch should be measured first, followed by the
anodic branch, and this can often be achieved by a single sweep from the most negative
potential up to the most positive. However this will lead to some errors due to changes in
the specimen surface at the most negative potential (e.g. reduction of oxide films and the
dissolution of cathodic hydrogen in the metal), and it may be best in some circumstances to
use two separate samples, one to measure the cathodic branch and the other to measure the
anodic branch, starting both measurements from the corrosion potential.
Y-t recorders are commonly used to record the variation of current with time, particularly
for potentiodynamic sweep measurements. This is a useful technique, but it is important to
recognise that it is usually required to measure current over several decades, and in order to
make accurate measurements it is important to adjust the range settings on the recorder to
match the magnitude of the voltage being measured. It is also useful to mark the chart
record with the potential at regular intervals to ensure that the time axis can be accurately
converted to potential.
Translating current data from a Y-t chart recorder output to a log |i| vs E graph is relatively inefficient and
error prone. An alternative approach which avoids this problem is to use an X-Y recorder with a log converter
on the current axis. Computerized data acquisition systems provide another method of automating the analysis
and plotting procedures, although care is necessary to obtain adequate resolution of the current over the
necessary wide range.
See Measurement of current
The Measurement of Linear Polarization Resistance
The measurement of polarization resistance has very similar requirements to the
measurement of full polarization curves. There are essentially four different methods of
making the measurement according to whether the current or the potential is controlled and
whether the current (or potential) is swept smoothly from one value to another, or simply
switched between two values. In addition the measurement may be made between two
nominally identical electrodes (a two-electrode system), or a conventional three-electrode
system (working, reference and counter) may be used.
Measurement with current control has a number of advantages; it permits the specimen
potential to find its own level, hence the measurement is automatically balanced around
zero current and the instrumentation required can be very simple, especially if positive and
negative current steps are used. Its main disadvantage is that some care is needed to ensure
that the amplitude of the potential fluctuation is sufficiently large to measure accurately, but
not too large. It is therefore inappropriate for systems in which the corrosion rate is varying
by a large amount.

Figure 14 Techniques for the Measurement of Linear Polarization Resistance


Typical configurations for the four possible measurement methods are shown in Figure 14.
A fundamental problem in the performance and interpretation of linear polarization
resistance is deciding what cycle period to use. This is related to the problem of
interpretation of ac impedance data, since, to a first approximation, the polarization
resistance that is measured with a particular cycle period will correspond to the impedance
for a sine wave of the same frequency.
Figure 15 Analysis of Linear Polarisation Resistance Measurements
The interpretation of linear polarization experiments may range from the very simple to the
complex. The simple analyses just calculate a resistance from the current-time or current-
potential records (assuming controlled potential measurements) as shown in Figure 15.
Note that even for this simple resistance measurement, several different values can be
taken. Thus for the current sweep measurement, one can take the average slope of the E-i
curve, the slope at zero current, or the limiting slope at the end of the sweep, and each of
these approaches will give different results since they give different weight to different
frequency components of the input triangle wave. In this respect, if one takes the view that
Rp is the limiting value of dE/dI as the frequency tends to zero, the 'best' value to take will
be the slope at the end of the sweep, since this gives most weight to the lowest frequencies.
On the other hand it also implies that the slope is being taken at the potential which is
furthest from the rest potential, and this may introduce errors due to non-linearity. Fewer
options are available for measurements using stepped potential or current, where R p is
generally calculated as E / I as indicated in Figure 15.
Much more complex analyses of linear polarization resistance measurements are possible
and, at least in theory, it is possible to estimate the components of an appropriate equivalent
circuit.
Transient Techniques
Measurement of Transient Behaviour
Analysis of Transient Behaviour
Time Based Analysis
Frequency Based Analysis

Measurement of Transient Behaviour


The procedures for the measurement of electrochemical transient behaviour are essentially the same as those
for the measurement of linear polarization resistance and polarization curves, except that the controlled
potential or current will be stepped from one value to another in order to examine the response of the system.
The time scales of interest may vary from milliseconds up to hours. The response to the transient may, in
principal, be recorded on a chart recorder (or a UV recorder for high speed transients), but computer data
acquisition would normally be used to permit the use of numerical analysis procedures.
 
Analysis of Transient Behaviour
Time Based Analysis
In the time-based analysis, the transient response is taken to consist of the sum of a series of simple
transients. The largest transient is analyzed in terms of an appropriate equation which is fitted to the
measured data over an appropriate range. Then the calculated curve is subtracted from the measured
data, and this will usually show a second transient which can then be analyzed as before (see Sykes).
This approach is simplest for very long period measurements, such that the system can be taken to be
at equilibrium prior to the commencement of the transient. This approach is most useful when the
expected transient response is well understood, such that good theoretical predictions are available
for the expected behaviour. It may be used to analyze both large and small amplitude transients (as it
does not require that linear behaviour is assumed).
 

Figure 16 Fourier Transform Analysis of Transient Measurements


Frequency Based Analysis
A somewhat more general approach for the analysis of small-amplitude transients is to analyze the
results in terms of frequency. Fourier showed that any repetitive signal can be expressed as a series
of harmonically-related sine waves, and we can compute the phase and amplitude of these sine
waves by means of the Fourier transform. If we apply this procedure to the voltage and current
records from a polarization resistance measurement we obtain phase and amplitude values for two
series of sine waves. By dividing the voltage by current at each frequency, using complex arithmetic
to retain the phase information, we derive the impedance of the metal-solution interface over a range
of frequencies (Figure 16). While it is useful to appreciate this relationship between a polarization
resistance measurement and the ac impedance spectrum, it is probably not a technique that is very
useful in practice, as it suffers from a number of weaknesses:
 the power present at a given frequency is inversely proportional to the frequency, hence the power
present at the higher frequencies is low, and we can only cover one or two decades of frequency with
a given measurement period.

 harmonic information tends to be masked by the wide range of frequencies being applied, and the
fact that they are harmonically-related.
Measurement of Electrochemical Impedance Spectra
Definition
Application of AC Impedance Measurement to Electrochemistry
Measurement Methods
Wheatstone Bridge
Phase and Amplitude Measurement
Phase-Sensitive Detection
Sinewave Correlation
Fourier Analysis
Multiple Frequency Methods
Cell Configurations
Validation of ac Impedance Data
Linearity
Stationarity
Causality

There are a wide range of techniques available for the measurement of ac impedance
spectra. One of the major distinctions is between those techniques which measure a single
frequency at a time, and those which measure several frequencies simultaneously. The
latter techniques are relatively little used in corrosion studies, although in principal they are
quite attractive. The frequencies of interest in corrosion tend to be rather low (down to 10 -
3
Hz or less) and the time taken for a measurement is long. Consequently the specimen
properties can often drift during the measurement of the low frequencies, invalidating the
basic requirement for stationarity.
The Wheatstone Bridge

The earliest method of measuring the impedance of a cell used an ac Wheatstone bridge
(Figure 16). The values of R3 and C1 are adjusted to minimize the real and imaginary
components of the bridge voltage, as determined on the oscilloscope. This method is no
longer in widespread use because it has several disadvantages.
 it is labour intensive and very slow

 it is difficult to use for frequencies below about 1Hz.

 it is not possible to use a three electrode cell.

Phase and Amplitude Measurement


In this method the cell potential is perturbed with a sine wave signal and the amplitude and
phase of the resultant current is measured (Figure 18). Then the amplitude of the
impedance is simply the amplitude of the potential divided by the amplitude of the current,
and the phase of the impedance is the phase of the voltage with respect to the current (i.e.
minus the measured phase).
This technique is rarely used because of a number of disadvantages:
 the amplitude measurement is sensitive to noise pick-up at frequencies other than that being
measured (especially mains frequency),
 standard amplitude and phase meters are not designed to operate at the low frequencies that are of
interest in corrosion.
 
Phase-Sensitive Detection

One way round the problems of sensitivity to extraneous noise and low frequency operation
is to use the technique of phase-sensitive detection (the instruments used for this purpose
may also be called lock-in amplifiers).
This method uses the signal at the energizing frequency to control the rectification of the ac
signal (?). The result is equivalent to multiplying the input signal by a square wave of the
same frequency as the input signal and the appropriate phase. This is an effective
measurement technique, although it is susceptible to errors due to harmonics of the
measurement frequency (this is because the square wave includes all harmonics of the
fundamental frequency, and if harmonics are generated in the electrochemical cell as a
result of non-linearity of the response, this will produce an error in the measurement).
Sine Wave Correlation

The susceptibility of the phase-sensitive detection technique to errors due to harmonics can
be overcome by multiplying the signal to be measured by a sine wave instead of a square
wave. This is the principle of the most accurate method available for measuring ac
impedance. The process of multiplying one signal by another is known as correlation,
hence this technique can be described as sine wave correlation, although it is probably
better known as frequency response analysis, after the name used for the range of
instruments produced by Solartron which dominate this field.
In principle, the multiplication can be performed by analog computing techniques, but in
practice it is difficult to achieve a sufficiently wide dynamic range with this approach, and
these measurements generally use digital techniques.
Fourier Analysis
The process of sine wave correlation is a general method for determining the content of a sine wave of a
particular frequency and phase in an input signal. This procedure forms the basis of the Fourier transform, and
the computational techniques which have been developed for general signal analysis can be applied to
determine the components of the signal produced as the result of the application of a sine wave to an
electrochemical cell. The procedures are somewhat more complex than the single frequency sine wave
correlation, but the algorithms are well-known. This approach has the advantage that the harmonic content of
the input signal is determined at the same time as the determination of the phase and amplitude of the
fundamental.
This technique is used in a low cost instrument produced by Voltech, and it is likely to be a
common approach in the future.
Multiple Frequency Methods
The Fourier transform can in principle extract all of the frequencies present in a repetitive signal, therefore it
is possible to apply signals consisting of several sine waves superimposed on each other, using the Fourier
transforms of the voltage and current signals to extract the necessary amplitude and phase information. In
principle, this has the major advantage of reducing the time required to make a measurement. In practice,
there are some limitations to this approach.
The extreme example of this approach is to apply the first n harmonics of the fundamental
frequency, where n is a large number. This can be achieved by the use of pseudo-random
noise which can be generated very simply. This method is used by some commercial
spectrum analyzers (since the Fourier transform procedures are already available for the
spectral analysis, it is very easy to add the random noise source to the instrument). The
major theoretical limitation of this approach is that, since all harmonics are present in the
input signal, it is impossible to separate harmonic effects. The method also gives
uniformly-spaced output frequencies, which gives an unnecessary number of readings at the
higher frequencies, since we are generally concerned with analysis of the results versus log
(freq).
To reduce problems of harmonics, it is possible to use several non-harmonically related
sine waves. This may be a good compromise, although we have a suspicion that as the
number of frequencies measured together increases, so does the time needed to get results
with a given margin of error.
Cell Configurations
Various cell configurations can be used in the measurement of ac impedance, but we would recommend that
the simplest possible system should always be preferred. The frequencies are often relatively high (10 kHz
and above), and conventional electrochemical instrumentation often has a poor response at such frequencies.
To take a simple example, consider the problem of using a reference electrode to measure the potential. It is
not unreasonable for the reference electrode plus the solution in the Luggin probe to have a resistance of 100
kOhm (105Ohms), and this would give insignificant errors at dc.

Figure21 Cell Configuration for Impedance Measurement


However the capacitance of the reference electrode input to a potentiostat, together with a
screened cable, will be of the order of 100 pF (10 -7F), giving a time constant of 105 x 10-7 =
10-2 seconds. Thus the reference electrode will start to modify the measured potential at
frequencies above 100 Hz! There are ways of getting round this problem, but much the
simplest approach is to eliminate it altogether by measuring the impedance of two identical
electrodes.
The general principle of all configurations is to perturb the cell and to measure the effect on
the cell in terms of the cell or specimen potential, and to measure the current flowing by
passing it through a reference impedance (i.e. one with known characteristics). Typical
configurations are shown in Figure 21.
In the vast majority of cases, the reference impedance used is a pure resistor, but this is not
the ideal approach. While the impedance of a resistor is independent of frequency, the
resistance of a typical electrochemical cell can change by two or more orders of magnitude.
This leads to an imbalance between the impedance of the resistor and that of the cell which
has various implications, according to the measurement configuration used.
For systems that show extreme variation in cell impedance, it might be found useful to use
slightly more complex reference impedances including an element of capacitance such that
ZR has a similar impedance to that of the cell over the whole frequency range. For similar
reasons, it will generally be found best to use voltage or potential-controlled measurement
techniques, such that the measured voltages remain reasonably constant. Note, however,
that in configuration 6, the current will be controlled such that the voltage across Z R is
constant, hence if ZR matches the cell impedance, the voltage across the cell will be
reasonably constant.
Note that some of the configurations in ? provide a buffer amplifier before the input to the
impedance measuring system which usually has a relatively low impedance (around 1
Mohm), and will otherwise give problems with measurements on high impedance systems
such as painted metals.
Validation of ac impedance data
There a several requirements that must be met to obtain a valid impedance spectrum:
 
Linearity
The response of the system must be a linear function of the perturbation applied. This is hardly ever strictly
true for corroding systems, but we attempt to approximate to it by using small amplitude perturbations.
 
Stationarity
The system must not be changing while we are making the impedance measurement. Again this is rarely valid
for corroding systems, especially if we attempt to make measurements during the early stages of the corrosion
process. There is little that can be done to improve the behaviour of real systems, where a lack of stationarity
will be revealed most strongly in the low frequency regime, where individual measurements take a longer
time.
 
Causality
The output from the system must be exclusively a result of the perturbing input. This is rarely a serious
problem for corroding systems.
Unfortunately, the absence of any of the above requirements does not necessarily prevent
the measurement being made, and it is desirable to test the results for validity, especially
when a new system is being studied.
For any linear system, the phase and amplitude of the impedance are intimately linked
together. Thus, if we know the phase as a function of frequency we can deduce the way in
which the amplitude of the impedance is changing. As a result of this property we can
calculate the amplitude of the impedance from the phase and vice versa. This is achieved by
the Kramers-Kronig (K-K) transforms, and MacDonald has proposed the comparison of the
results of the Kramers-Kronig transform with the observed phase and impedance as a check
on the validity of impedance data. This is a useful check on the validity of a measured
impedance spectrum, but it does have some limitations; in particular the impedance
spectrum must have been measured of the full range for which the properties are changing,
otherwise the K-K transform cannot be computed. Furthermore it is important to appreciate
that while an invalid K-K transform indicates that there were some problems with the
impedance spectrum, a valid K-K transform does not necessarily indicate that the
impedance data are correct.
Electrochemical Noise
 
Theoretical Basis of Noise Generation Processes
Sources of Noise
 Uniform Corrosion

 Pitting Corrosion

 Mass Transport Fluctuations

 Bubble Nucleation, Growth and Detachment


Measurment of Electrochemical Noise
 Cell Configurations for Electrochemical Noise Measurement
Interpretation of Electrochemical Noise
 Visual Examination of Time Record

 Transient Analysis

 Statistical Analysis

 Fourier Transform (FFT)

 Maxinum Entropy Method(MEM)


Relationship between Potential and Current Noise
Electrochemical Noise for the Detection of Localized Corrosion
 

Electrochemical noise is a general term for the `random' fluctuations in current or


potential which occur as an electrochemical process proceeds. The theoretical
treatment of the phenomenon is incomplete, but there have been many useful
applications, both in scientific studies and in corrosion monitoring applications. The
use of electrochemical noise measurements has significant advantages: the
measurement does not involve any external perturbation of the corroding system,
and it can therefore be applied to real structures; the instruments required to make
the measurements are reasonably simple, particularly with modern computer-based
data acquisition techniques, and localized corrosion processes, which can be
difficult to monitor with other techniques, tend to give particularly strong
electrochemical noise signals.
Theoretical Basis of Noise Generation Processes
The starting point for the development of a theory of electrochemical noise is a
theoretical analysis of the noise associated with a randomly-occurring, brief pulse of
charge, with the occurrence of each event being independent of any other event.
This is known as a Poisson process, and the simplest example is the flow of
electronic current, in which case each event is the passage of an individual electron
through the measuring circuit. If we define the noise current, In, as the instantaneous
current minus the mean current, it can be shown that the noise current is given by

The result of this process is known as shot noise, and is an unavoidable minimum
noise current associated with the flow of current.
Considering an electrochemical reaction, providing we can treat the dissolution
process as a series of brief events, we can use a similar analysis to predict the noise
current:

If the dissolution event has a significant duration, the noise at high frequencies
(where the period becomes less than the duration of the event) will fall to that due to
the individual reactions (i.e. q will become the charge on the electron times the
number of electrons involved in the reaction). The slope of the power spectrum
between the low and high frequency limits will be a function of the shape of the
current transients associated with the individual events, although the slope will only
be clearly distinguishable if q corresponds to a large number of electrons.
Note that this analysis is based on the assumption that the noise is due to the
electrochemical reaction occurring as discrete bursts of charge. This may not always
be the case, and further work is needed to clarify the situation.
Sources of Noise
Uniform Corrosion
Uniform corrosion might be expected to be free of noise, with atoms leaving the metal surface at a
uniform rate. However, even a perfectly homogenous process will give some fluctuations in rate,
akin to Brownian motion. Furthermore, there are a number of mechanisms by which it might be
expected that even a uniform dissolution process will occur as a series of bursts. For uniform
dissolution processes it is expected that the value of q will correspond to the charge liberated by 10 2
to 106 atoms.
 
Pitting Corrosion
The pitting initiation process is often found to result in metastable pit nucleation and propagation,
giving rise to current transients lasting for a time of the order of 1 second, and involving a charge of
the order of 10-6 C (corresponding to around 10 12 atoms). Thus the noise associated with pitting
corrosion is much larger than that observed for general corrosion.
 
Mass Transport Fluctuations
Fluctuations in the boundary layer thickness for a mass transport process will give rise to fluctuations
in the current. This will be observed most strongly in turbulent conditions, although even in
nominally laminar conditions it is expected that some fluctuation will occur. It is not clear that this
source of noise can be analyzed as a Poisson process, and further work is needed to understand the
expected relationships between the measured noise and the underlying processes.
 
Bubble Nucleation, Growth and Detachment
When hydrogen evolution is the predominant cathodic reaction, it is clear that the growth and
detachment of hydrogen bubbles will tend to give rise to a fluctuation the currents flowing. This
source of noise is unlikely to behave according to the treatment discussed above.
Measurement of Electrochemical Noise
Cell Configurations for Electrochemical Noise Measurement
There are two related, but distinctly different, approaches to the measurement of electrochemical
noise:
 The potential noise of a single sample can be monitored relative to a low noise
reference electrode.

 The current noise between a pair of identical electrodes can be measured. A similar
measurement can also be made on a single electrode by controlling its potential
relative to a low-noise reference electrode.

It is also possible to combine both of these techniques by measuring the current


noise between two identical electrodes while at the same time monitoring the
potential noise of the coupled electrodes relative to a low-noise reference electrode
or to a third identical electrode. This technique offers interesting possibilities in the
analysis of the resultant data, and is probably the best technique for corrosion
monitoring.
Most electrochemically-generated noise occurs at relatively low frequencies (of the order of
1Hz and less) and the measurement procedures do not present major difficulties, except
where it is required to measure very low amplitude signals.
As a preliminary investigation of the nature of the electrochemical noise from a specific
system, the simplest method is to record the current or potential with an X-t recorder.
However this provides a limited resolution, and it is virtually impossible to perform further
processing such as spectral analysis. Hence for more sophisticated work the time record
should be recorded in a machine-readable form.
For many systems, the conversion speed offered by conventional digital multimeters is
sufficient. These also provide very good noise rejection, particularly for mains frequency
noise (since the voltmeters integrate the input voltage over one or more mains cycles). The
data can be transferred from the voltmeter to a computer for recording using a serial output
or the IEEE 488 bus.
If it is required to measure higher frequencies, commercial data acquisition cards can be
used. These can give sampling speeds in the kilohertz range, with typical resolutions of 12
bits (1 part in 4096). Some care is required in the signal conditioning to make the most use
of the available resolution. This approach is also susceptible to problems due to mains
frequency noise, and great care is necessary in the screening of the system in order to
minimize mains frequency pick-up. (Even when using digital multimeters or other
integrating converters it is wise to try to minimize mains frequency pick-up to reduce the
possibility of errors due to overloading of the input amplifier). The use of battery-powered
signal conditioning devices is usually necessary to avoid mains frequency pick-up through
the power supply.
Interpretation of Electrochemical Noise
While it is generally accepted that electrochemical noise contains useful information about
electrochemical processes, there is no consensus about the best approach to extracting the
information. To some extent this is because the best method varies according to the type of
process being studied, but this is also an area where further development is required.
Visual examination of time record.
This is a very effective method of detecting specific transients, such as those generated during pit initiation or
stress-corrosion cracking. It is also possible to see clear period signals, which may be generated during
crevice or pitting corrosion and you can estimate the standard deviation of the signal simply by observing the
`width' of the signal trace. It is recommended that a visual examination of the time record should always be
the first part of a data analysis procedure. See Figure 23.
 
Transient analysis.
When specific transients can be detected in the time record, their characteristics can be analyzed by curve
fitting procedures or by simple measurement from the time record. Thus Williams and his co-workers have
made extensive use of this approach to determine various properties of the transients associated with unstable
pit nucleation and growth on a statistical basis. This analysis can be performed manually (it is one of the few
procedures which can be applied to a chart recorder trace of the time record), although the analysis of
digitized time records can have advantages if suitable software is available. Thus approximate growth or
decay curves can be fitted to a transient with least-squares procedures, or the duration of a transient can be
determined by extrapolation. In the interpretation of transients it is important to be aware of the difference
between potential and current transients. An example of this which has caused considerable confusion is the
interpretation of transients associated with unstable pit propagation. As we now understand the phenomenon,
after the pit initiates it grows steadily, and as the surface area of the pit increases, so the current draw by the
pit increases. Eventually the pit abruptly repassivates, possibly because an oxide cap, which helps to maintain
the chemical conditions in the pit, ruptures. This gives a current transient similar to that shown in Figure 22.
However early workers monitoring potential noise observed transients with a sharp rise and a slow decay.
This was initially interpreted as a film rupture process, followed by a slow repassivation. It is now clear that
the apparently sharp rise in potential corresponds to the rising part of the current transient, as the current
drawn by the pit is supplied from the double layer capacitance of the rest of the specimen. The subsequent
slow decay corresponds to the recharging of the double layer capacitance through the passive film, and has no
direct connection with the pitting process.
 
Statistical Analysis.
Various statistical procedures have been used to analyze electrochemical noise time records. The simplest
approach, yet one which seems to be effective for corrosion monitoring, is to determine the standard deviation
of the time record. This can also be described as the rms (root mean square) voltage or current after having
subtracted the dc component and is a measure of the ac power in the signal. While many ac voltmeters will
offer `true rms' measurement of ac voltage, it is important to be aware of the frequency requirements of this
measurement, as they will usually be far lower than conventional meters will accommodate. Most of the
power in an electrochemical noise signal lies at frequencies below 1Hz, and the standard deviation
measurement should include frequencies down to about 10 -3 Hz. This can be achieved by analogue
techniques, but some of the filtering requirements are quite difficult.
Moving beyond the measurement of the overall ac power in the signal, more detailed
information may be obtained by calculating the power present as a function of frequency.
The plot of power versus frequency is known as a power spectrum, and the process of
estimating this plot is known as spectral estimation. There is no `correct' method of
determining a power spectrum for a given true record. The time record is a sample of the
complete time record which will almost inevitably be much longer than any sample we may
be able to analyze. The problem is therefore analogous to the problem of estimating the
mean and standard deviation of a population from the mean and standard deviation of a
small sample from that population.
There are various methods that have been used for spectral estimation, but two in particular
have been used in studies of electrochemical noise; the Fourier transform and the
Maximum Entropy Method.
Fourier transform (FFT)
See Fourier Transform. Effectively it provides an exact description of the frequency content of the
signal analyzed, if it is assumed that the signal is one cycle of a periodic waveform with all the
cycles before and after the measurement period being exactly the same. This assumption leads to a
number of complications. Firstly, if the mean value of a signal is drifting with time this is interpreted
as a saw-tooth waveform, and `creates' frequency components (the fact that the mean is changing
with time implies that the system is non-stationary, and the Fourier transform is not strictly
applicable in this case, but we cannot afford to be too fussy about problems like this!). Second, the
implied `joining-up' of the start and the end of the time record is liable to create sharp transitions
which are not representative of the signal. These two problems are usually overcome by pre-
processing of the signal before undertaking the Fourier transformation.
 
Firstly, any drift is removed by fitting a straight line to the time-record and replacing the time record
by the deviation from the line. This also serves to set the signal mean to zero. This may still leave
sharp transitions at the end of the signal, and these are eliminated by multiplying the signal by a
windowing function which is zero at the ends of the time record and one in the centre. A wide range
of windowing functions have been proposed, but the simple triangular Parzen window is probably as
good as any. (see Press et al. for further information). Having removed these sources of difficulty the
modified time record is transformed to the frequency domain, usually using the Fast Fourier
Transform (FFT). The resultant frequency spectrum must then be normalized to relate the power
present in the spectrum to that present in the signal. One might expect this to be uniquely defined,
but in fact there are several interpretations of the meaning of the power spectrum, and these result in
different normalization procedures. The common interpretation in a power spectrum (but one which
differs from that given in the generally excellent book due to Press et al ?) is that the spectrum is
plotted as power spectral density, power per cycle (e.g. a value of 1 V 2/Hz at 10 Hz implies that there
will be 1V2 of power in the band between 9.5 and 10.5 Hertz. With this interpretation the power at
each frequency is given by:
 
Maximum Entropy Method (MEM)
Whereas the Fourier transform directly calculates the frequencies actually present in the time record
being analyzed, the maximum entropy method (also known as the all-poles or autoregressive model)
may be thought of as a method of calculating the coefficients of a series of filters that would have to
be applied to white noise to obtain the observed time record. An advantage of the MEM is that it
gives smooth curves, rather than the very noisy spectra obtained with the Fourier transform. The
details of the calculation are beyond the scope of this book; a discussion of the underlying theory and
computer programs are given by Press et al.
 
Because the MEM gives as its result a set of coefficients, it is possible to calculate the power spectral
density for any frequency. This has led to the publication of some fictitious data where the range of
validity of the analysis have not been fully appreciated. Given a time record of duration t, we cannot
know anything about signals with a period >t, i.e. frequencies < 1/t; this may be expressed in terms
of the band width of the analysis being 1/t (or greater if trend removal or windowing are used). At
the high frequency limit one cannot know anything about signals for which there are less than 2
samples in each cycle. If the sample period is t, the frequency of 1/2 t is known as the Nyquist
frequency, and is the upper limit of validity of the analysis. Unfortunately, while we cannot know
anything about frequencies above the Nyquist frequency, this does not mean that they don't show up
in the results. If we take regular samples of a signal above the Nyquist frequency, the results are
exactly the same as if we had taken samples from a signal of the same amplitude, but a frequency
below the Nyquist frequency. This phenomenon is known as aliasing. Ideally it should be prevented
by using filters to remove these high frequency signals, but these are frequently omitted in
electrochemical noise measurements. This is not usually too much of a problem, as electrochemical
noise spectra typically show a power spectral density that falls with increasing frequency, and
aliasing is revealed as a flattening of the high frequency and of the spectrum.
 
Power spectra obtained by the Fourier transform and the MEM are shown in Figure 25.
 
Having obtained a power spectrum, the real problem lies in deciding what it means. The area under
the curve is the total power in the signal, and is identical to the standard deviation calculated from
the time record. Thus as the frequency spectrum moves to higher power spectral densities, so the rate
of reaction may be expected to increase.
 
Periodic signals will clearly give rise to a peak in the power spectrum, and these may be related to
crevice or pitting corrosion. However this is fairly rare, and is usually readily apparent from
inspection of the time record. More commonly the spectrum has an appearance similar to that shown
in Figure 25, with a power spectral density that falls with increasing frequency, giving a straight line
on a log-log plot, implying a relationship of the form:
Power spectral density 1/fn
where n is a constant
In addition to the sloping 1/fn region of the spectrum, there are often indications of plateaus at the
high and/or low frequency ends of the spectrum. You should be very careful to ensure that these are
real before attempting to place any theoretical interpretation on them. Apparent plateaus at low
frequencies can arise from several sources: 
 If the MEM is extrapolated below the band-width of the analysis, it will
necessarily give a constant power spectral density.

 Trend removal and windowing will tend to increase the band-with of the analysis,
especially if higher order regression lines are used for trend removal.
 In general it is suggested that the lower valid frequency should be taken as something like 3/ time
record period, and only if the plateau is clearly evident at higher frequencies than this should it be
regarded as genuine. If in doubt take a longer time record (taking samples less frequently if
necessary) in order to test that a plateau remains the same.
 
Similarly, there are several ways in which an artificial high frequency plateau may be obtained:
 
 Aliasing of higher frequencies

 Instrumentation noise
As with low frequency plateaux, the validity of a high frequency plateau can be checked by taking
samples more frequently, reducing the length of the time record if necessary. Noise due to
instrumentation is more likely at higher frequencies (and correspondingly lower amplitudes) and you
should check for instrumental noise by replacing the electrochemical cell by a dummy cell of
comparable impedance to that of the electrochemical cell.
The Relationship Between Potential and Current Noise
The fundamental processes that generate electrochemical noise are concerned with
fluctuations in the rate of the electrochemical reaction. This is essentially a process which
creates current noise. As an unstable pit nucleates and propagates it creates a current
transient; as a metal corrodes in acid, bursts of dissolution from a favourably oriented grain
or perhaps screw dislocation give rise to pulses of current. Thus current noise bears a fairly
direct relationship to the underlying physical processes, and in general as rates of reaction
increase, so the current noise will increase. In contrast, the potential noise is a secondary
manifestation of the current noise. As the current created within the pit, or at an intense
local anode, returns to the metal at a cathodic site it must pass through the electrochemical
impedance at that site. The foregoing description assumes that the anodic process is
responsible for the noise current - clearly there is no requirement that this should be the
case, and in practice the situation is rather complex. However, the simplified example given
forms a good basis for understanding the way in which potential noise changes.
However, if we ignore solution resistance effects, the impedance of the metal-solution
interface at low frequencies is given by the polarization resistance, Rp, hence

Thus we can estimate Rp and icorr by dividing the potential noise by the current noise. The
resultant parameter is known as the electrochemical noise resistance, and several studies
have shown that it gives a good indication of the corrosion rate, providing the solution
resistance is small compared to the charge transfer resistance.
We can take this analysis further and calculate the variation of impedance with frequency.
To do this we calculate the current and potential noise spectra, then the impedance at a
specified frequency is the potential noise at that frequency, divided by the current noise. In
principle, if we have simultaneous measurement of the potential and current time records
we can calculate phase as well as amplitude information by using the Fourier transform. In
practice the results of this analysis are highly scattered, and a more practical approach is to
use the MEM transform to calculate the amplitudes of the impedance.
Electrochemical Noise for the Detection of Localized Corrosion
Early studies suggested that localized corrosion processes give a particularly strong
electrochemical noise response, and various criteria have been developed to assess whether
or not localized corrosion is occurring. One of the earliest methods was to determine the
coefficient of variation for the current noise signal. The coefficient of variation is the
standard deviation divided by the mean, and this presents the problem that the mean current
has an expected value of zero, and it is only by chance that the mean current is generally
non-zero. More recently this criterion has been revised to the localization parameter or
pitting index:

While this still has a defined value when the mean is zero, it is also necessarily 1 in this
situation.
It can be argued that the problem with both the coefficient of variation and the localization
parameter is that they are based on the mean net current, whereas the variable giving rise to
the noise is the one of the two partial currents (while it is possible that both the anodic and
cathodic partial reactions will give rise to comparable noise, this seems unlikely). Thus we
should replace the mean net current by the corrosion current Icorr, and a suitable parameter

to characterize localized may therefore be . However, we can calculate Icorr from


the noise resistance:

Hence the localization parameter becomes:

However, if we consider the fundamental processes underlying the generation of


electrochemical noise, a parameter that may be more appropriate is the amount of charge in
each transient. We can compute this from the shot noise equation:

It remains to be seen whether this provides a valid parameter for the practical identification
of localized corrosion.
Glossary
AC Impedance Spectrum
Impedance is the response, in terms of the resultant ac voltage, of a circuit element to the application
of an ac current. It is analagous to resistance for a dc system. As any frequency may be applied, the
impedance will be a function of the frequency, and an impedance spectrum is a representation of the
impedance of a circuit element as a function of frequency.
 
Band-width
The difference between the two frequencies for which the voltage amplitude ratio falls to -3dB of its
maxinum value is called the 3dB Band-width.
 
Double Layer Capacitance
This term (Cdl) is voltage dependant and thus is intrinsically nonlinear. In aqueous media this
capacitance is typically a few tens of microfarads per square centermeter, and the measured value
can be used to estimate the effective area of a corroding electrode. See Double Layer Capacitance.
 
Equivalent Circuit
This is a method, devised by Thérenin and Norton, of reducing the number of independent nodes and
loops in a network through the replacement of series and parallel arrangements of passive elements
by their appropiate equivalents. See Equivalent Circuit.
 
Gain

 
Non-linear Systems
In these systems the current is not proportional to the voltage i.e. V = IR does not hold and a graph of
voltage against current would not be a straight line.
 
Phase

 
Potential Divider

 
Charge Transfer Resistance
see Charge Transfer Resistance.
Shot Noise
When a current flows through a galvanic cell then a noise is generated described as the Shot Noise.
This noise is caused by current bursts taking place at the electrodes. In corroding systems the noise
observed is related principally to localized corrosion. Processes such as pitting or other forms of
localized attack change drastically the value of the anodic reaction resistance on a small area of the
electrode. See Electrochemical Noise Methods.
 
Solution Resistance

 
Using a Potentiostat to Control Current
Note that there should be no other connections to the potentiostat. E.g. if reference electrode is used
to monitor the potential of the working electrode, it must not be connected to the RE terminal of the
potentiostat, and the working electrode must not be connected to the working electrode terminal.
 
References
Stern and Makrides - J.Electrochem Soc. 107, 782 (1960).
Sykes -
Copyright
All documents refering to this page are copyright R.A.Cottis and A.M.Llewellyn. They
may be freely referenced in World Wide Web and other documents. Neither the documents,
nor parts of the documents (such as images or sections of text) may be copied, incorporated
in other works or used in any other way for commercial purpose without the express written
consent of the authors. Copies of the documents (either printed or electronic) may be made
for private study, or for caching purposes. Educational establishments wishing to use these
documents for teaching purposes may take temporary copies to reduce network usage, but
are requested to update or delete these copies after no more than four weeks to ensure that
they are up-to-date. Copies made for any purpose may not be modified in any way.

Acknowledgement
The production of the electronic versions of these documents was supported by the
UMIST ........

Disclaimer

S-ar putea să vă placă și