Introduction to EIS and

Conversion of CPE into C

Dr. Zack Qin

The Blind Men and the Elephant

… Is very like a wall!

… Is very like a spear!
… Is very like a snake!
… Is very like a tree!
… Is very like a fan!
… Is very like a rope!
…….
Not one of them has seen!

J. G. Saxe (1816-1887)
1
 Z is a transfer function

Excitation Transfer Response

E(t) Function I(t)

Fourier transform

E(jω) = Z(jω) x I(jω)

™ Z is defined in the frequency domain that relates to the time

domain through Fourier/Laplace Transform.
™ The equation can be considered as a generalized Ohm’s law.
2
Z is a complex function

Z (ω ) = Z ' (ω ) + jZ " (ω ) = Z (ω ) e − jθ (ω )

Z ' (ω ) = Re[ Z ( jω )]
Z " (ω ) = Im[ Z ( jω )] Z”

Z (ω ) = Z ' (ω ) 2 + Z " (ω ) 2
⎛ Z " (ω ) ⎞
θ (ω ) = arctan ⎜ − ' ⎟ Z’
⎝ Z (ω ) ⎠

™ Z is described by pair-functions, Z’ and Z”, or |Z| and θ.

™ Z’(ω), Z”(ω), |Z(ω)|, and θ(ω) are real functions.
™ Other related immittance functions:
Admittance Modulus Dielectric constant
Y = Z-1 M = jωC0Z ε = (jωC0Z)-1
3
Element impedances
Resistance (Ω)
R=E/I ⇒ ZR = R
™ Dissipation of energy (Ohm’s law): I in phase with E
™ For a uniform area resistor: R = ρ L/A

Capacitance (F) dQ I 1
C= = ⇒ ZC =
dE ( dE dt ) jω C
™ Storage of charge (Coulomb’s law): I leads E.
™ For a parallel-plate capacitor: C = εε0 A/d

Inductance (H)
dΦ E
L= = ⇒ Z L = jω L
dI ( dI dt )
™ Storage of magnetic energy (Faraday’s law): I lags E.
™ Inductive behaviours are not normally observed in electrochemical
systems, but can be attributed to adsorption phenomenon.
4
 Kirchhoff’s rules
™ Conservation of energy -- the sum of potential differences across
each element around any closed circuit loop must be zero.
™ Conservation of charge -- the sum of the currents entering any
junction must equal the sum of the currents leaving that
junction.

Impedances in series:

E1 E2 E1 + E2
Z1
I
Z2
I Z= = Z1 + Z 2
I

Impedances in parallel:

I1
E Z1 1 I1 + I 2 1 1
= = +
I2 Z E Z1 Z 2
Z2
5
 Randles circuit

Cdl

Rs

Rct

Rct
Z ' = Rs +
1 + (ωCdl Rct ) 2 Rs = lim Z
ω →∞
ωCdl R 2
Rct = lim Z − lim Z
Z"= − ct
1 + (ωCdl Rct ) 2 ω →0 ω →∞

⎛ Z '− R − Rct
2
⎞ + ( Z ") 2 = ⎛ Rct ⎞
2 log Cdl = − log Z (ω = 1)
⎜ 2 ⎟⎠ ⎜ 2⎟
⎝ ⎝ ⎠
s

6
 Time constant
™ The time constant (relaxation time), τ, characterizes the response
of a first order linear time-invariant system. It represents the
time required to reach (1-1/e) of its asymptotic value in a step
response, or 1/e of its initial value in a impulse response.
™ The time constant will remain the same for the same system
regardless of the starting conditions.
™ RC circuit: τ = RC, and RL circuit: τ = L/R, regardless of series
circuits or parallel circuits.

Vin − VC dVC Vin ( t = 0) = 0, Vin ( t > 0) =V0

−t
=C ⎯⎯⎯⎯⎯⎯⎯⎯
→ VC (t ) = V0 (1 − e RC
)
R dt

7
Characteristic frequency
1
ω c = 1τ =
RC
™ ωc is the physical property of a RC circuit.
™ ωc is defined in the frequency domain, while τ is defined in the
time domain.
™ ωc is the resonance frequency at which –Z” is maximum.
™ ωc is the cutoff frequency at which energy entering the system
begins to be attenuated or reflected instead of transmitted.

8
 Other frequencies
™ The inflection frequency, ωθ, is the frequency at which the absolute
value of the phase angle θ reaches maximum.
™ The breakpoint frequencies, ω1 and ω2, are frequencies at θ=-45°.
™ While ωc is the property of a RC circuit, ωθ and ω1,2 are the
properties of Randles circuit.

Relationships to ωc

Rct
ωθ = ωc 1 +
Rs

ωc ⎛ Rs Rs2 ⎞
ω1,2 = ⎜1 ± 1 − 4 −4 2 ⎟
⎜
2 ⎝ Rct Rct ⎟⎠
ωθ

9
Constant phase element

Z CPE = [Y0 ⋅ ( jω )α ]−1 (θ = απ 2 )

= p/2 - θ

™ CPE is a phenomenological term first used by Brug in 1984. It

is an empirical impedance that its phase angle is independent
of frequency.
™ CPE is often related to frequency dispersion attributed to
surface inhomogeneities and distributed time constants.
™ CPE obeys Kramers-Kronig relations provided that |α| ≤ 1.
™ The physical justification of CPE is not obvious except on the
following circumstances:
α 1 0 -1 0.5
Y0 C 1/R 1/L 1/√2σ
10
Conversion of CPE into C
™ While the use of CPE usually increases the “goodness of the
fit”, the physical meaning of the CPE should be discussed.
™ Avoid using CPE. If this is inevitable, the causes of the non-
idealism should be identified.
™ CPE is often used to describe non-ideal capacitive behaviour.
However, the amplitude Y0 is not a capacitance.
™ The dimension of Y0 is secα/Ω, while that of C is F or sec/Ω.
™ Small deviations of the exponent α from 1 can lead to large
computational errors of capacitance.
™ To obtain a trustworthy value of α, the testing frequencies
should be at least a decade lower than ωc.
™ For conversion into C, the exponent α must be in the range of
0.8-1. Otherwise, CPE won’t represent a capacitor.
™ The conversion of Y0 into C has been controversial. At least
three distinct equations have been proposed.

CPE is not a capacitor, and Y0 is not capacitance

11
 Brug conversion
1α
⎡ ⎛ 1 1 ⎞ ⎤ α −1

C = ⎢Y0 ⎜ + ⎟ ⎥
⎢⎣ ⎝ Rs Rct ⎠ ⎥⎦
™ It applies to Randles circuit with no more than one τ.
™ It depends on the solution conductivity. This dispersion is largest
when Rs dominates over Rct.
™ It obtained by comparing impedance (admittance) with relaxation
analysis.

1⎡ 1 ⎤
Y= −
⎢ 1 + R R + R Y ( jω )α ⎥
1
⎣
Rs s ct s 0 ⎦
∞
1 ⎡ 1 ⎤
Y= 1 −
Rs ⎢⎣ −∞ ∫ 1 + Rs Rct + jωτ 0 exp(s) F ( s ) ds ⎥
⎦

G.J. Brug, et al, J Electroanal. Chem. 176 (1984), 275-295. 12

 Mansfeld conversion
α −1
C = Y0ωc
™ It applies to an individual R||C, regardless how many branches
are nested in the circuit.
™ No dependency on the solution conductivity.
™ The derivation was based on
the fact that Z’ is independent 12000 -6000

of α at ωc, which itself is also a=0.8.z

a=0.9.z

independent of α. Therefore,
10000 -5000
a=1.z

8000 -4000

1 1

Z''
Z'
6000 -3000

=
Y0 ( jωc )α jωc C 4000 -2000

2000 -1000

1 1
Z = = 2 2
2
2α
Y ω ωc C
0 0
2 10-3 10-2 10-1 100 101 102 103 104

0 c Frequency (Hz)

C.H. Hsu and F. Mansfeld, Corrosion, 57 (2001), 747-748. 13

Extension of Mansfeld

With Warburg

Nested R||C

14
 Westing/Mertens conversion

Y0ωθα −1
C=
sin(απ 2)
™ First described in the PhD thesis of Westing (1992), and later
modified by Mertens (1997).
™ The derivation assumed that ZCPE(ωθ) = ZC(ωθ), where ωθ is the
inflection frequency.

1 1
= (ω = ωθ )
Y0 ( jω )α jωC
−α ⎛ απ ⎞ ⎛ απ ⎞
j = cos ⎜ − ⎟ + j sin ⎜ − ⎟
⎝ 2 ⎠ ⎝ 2 ⎠

™ Ilevbare and Scully proposed the use of the equation, but Hsu
and Mansfeld claimed the equation is incorrect.
S.F. Mertens, et al, Corrosion, 53 (1997), 381-388. 15
CPE2C program

™ CPE2C is a lab-built code for the conversion of CPE into C, based

on the Mansfeld or Brug equations.
™ It is available to members of Shoesmith and Wren groups.
™ Feedback and suggestions are welcome.
16
Use of CPE2C
™ Fit EIS spectrum with an appropriate equivalent circuit according
the spectrum and the system under investigation.
™ Identify the origin of the CPE (capacitive, diffusive, porous, a
combination of them, or even a wrong circuit).
™ For non-ideal capacitive behaviour, select the conversion method:
Mansfeld
• The ωc is the characteristic frequency associated with that
R||C to be converted. Note the angular frequency ω = 2πf.
• The calculated C must satisfy the condition when comparing
with the measured ωc,
1/Tolerance < RCωc < Tolerance
• Independent of solution conductivity at all.
Brug
• Apply to a Randles circuit, a blocking electrode, and
geometry-induced frequency dispersion.
• Conversion depends on solution conductivity.
™ The exponent α must be in the range of 0.8-1.

17
Example

Mansfeld R (Ω) ωc (rad/s) Y0 α C (F)

CPE-1 100 5384.4 10-5 0.8 1.794e-6
CPE-2 250 4.6572 0.001 0.9 8.574e-4
CPE2C Brug Rs (Ω) R1 Y0 α C (F)
A 10 100 10-5 0.8 9.764e-7
B 1 100 10-5 0.8 5.609e-7
18