EC-Lab - Application Note #51

04/28/2014

Supercapacitor characterization by galvanostatic polarization method

I – INTRODUCTION Each supercapacitor was cycled at ambient

temperature (~ 19°C) between its rated volt-
Supercapacitors known as Electric Double
age and 0 V using an SP-300 potentiostat/
Layer Capacitors (EDLCs) are used as energy
galvanostat with a 10 A booster.
storage devices for applications requiring
Supercapacitors were charged up to the rat-
power for short periods of time. Typically an
ed voltage and then discharged to 0 V. Nu-
EDLC consists of two porous carbon-based
merous charge/discharge rates were tested.
electrodes electrically isolated by a porous
The data were acquired and processed using
separator. The separator and the electrodes
EC-Lab® software.
are impregnated with an electrolyte, which
allows the ionic current to flow between the
The galvanostatic polarization protocol used
electrodes while preventing electronic cur- in this note is the GCPL technique. Alterna-
rent from discharging the supercapacitor [1]. tively the constant current (CstC) technique
located in the supercapacitor section of EC-
Compared to batteries, the supercapacitors
Lab® can also be used for supercapacitors
are characterized by a quick charge and dis-
testing.
charge regime, high power density and long
The figure below shows the setup used for
cycle life. These devices are rated to support
the charge and discharge of the SC400.
over one million cycles.

Many different methods are used to measure

supercapacitor parameters: cyclic voltam-
mtry, impedance spectroscopy, potentiostat-
ic and galvanostatic methods. The first meth-
ods were explained in the application notes
(#33) and (#34) [2-3].

The present note describes how to evaluate

the characteristics of a supercapacitor (ca-
pacitance, capacity, energy, internal re-
sistance and coulombic efficiency) using gal-
vanostatic polarization method and the “Ca-
pacity & Energy per Cycle or Sequence” anal-
ysis tool available in EC-Lab® software.

II – EXPERIMENTAL CONDITIONS
The tested commercial supercapacitors are
manufactured by two different companies: a
PANASONIC gold supercapacitor (HW series)
with a rated voltage of 2.3 V and a nominal
capacitance of 22 F and an IOXUS superca-
pacitor with a nominal capacitance of 400 F
and a rated voltage of 2.7 V, referred to as
SC22 and SC400, respectively. Fig. 1: Example of setup for SC400 testing.

Bio-Logic Science Instruments, 1 Rue de l'Europe, 38640 Claix, FRANCE 1

Tel: +33 476 98 68 31 – Fax: +33 476 98 69 09 www.bio-logic.info
 EC-Lab - Application Note #51
04/28/2014

III – RESULTS in the supercapacitor section of Analysis

III-1 E vs. TIME CURVES tools menu of EC-Lab®.
Fig. 2 shows the obtained voltage profiles for
the two tested supercapacitors. The charge
and discharge were performed at the con-
stant current of +/- 1 A for the SC22 and +/-8
A for the SC400.

Fig. 3: “Capacity & Energy per Cycle or Sequence”

window.

As shown in Fig. 3, the “Capacity & Energy

per Cycle or Sequence” tool allows an EC-
Fig. 2a: Voltage profiles for SC22.
Lab® user to calculate either for charge and
discharge processes the capacity Q/mA.h, the
capacitance C/F and the coulombic efficien-
cy/%. These parameters are calculated as
follow:

- Capacity: Q= I.Δt where I is the applied

charge/discharge current and Δt is the
charge/discharge duration.
- Capacitance C is calculated using the for-
mula C =Qch/disch/ΔEwe, where Qch/disch is
the total charge stored or released by the
Fig. 2b: Voltage profiles for SC400. supercapacitor. ΔEwe is the voltage differ-
ence between the initial and final poten-
The beginning of each charge and discharge tial either on charging or discharging pro-
sequence on the SC22 voltage profile (Fig. 2a) cess.
displays a voltage drop (iR drop). This voltage - Coulombic efficiency is the ratio
drop corresponds to an important internal (Qdisch/Qch).100. As the charge and dis-
resistance. However no significant IR drop charge were performed at the same cur-
was observed in the Fig. 2b indicating a low rent rate the coulombic efficiency is equal
internal resistance for the SC400 to (Δtdisch/Δtch).100 [4, 5] where Δtdisch and
supercapacitor. The internal resistance is Δtch are the discharge and charge time, re-
discussed in the paragraph III-5. spectively.
The supercapacitor characteristics were cal- Once the charge-discharge data file is pro-
culated by processing the charge-discharge cessed, the user can have access to the value
data using “Capacity & Energy per Cycle or of all the parameters mentioned above. In
Sequence” analysis tool. This tool is available case of a charge-discharge cycling test, the

Bio-Logic Science Instruments, 1 Rue de l'Europe, 38640 Claix, FRANCE 2

Tel: +33 476 98 68 31 – Fax: +33 476 98 69 09 www.bio-logic.info
 EC-Lab - Application Note #51
04/28/2014

user can also determine the parameters evo-

lution vs. time (or vs. cycle number, volt-
age,...) using the following file selection win-
dow (Fig. 4).

Fig. 5b: Capacity vs. cycle number for the SC22.

The measured capacities at the cycle 1 are

around 239.0 mAh and 12.1 mAh respectively
for the SC400 & SC22. These values decrease
slightly with growing cycle number.

Fig. 4: File selection window. III – 3 CAPACITANCE MEASUREMENT

Capacitance is the main characteristic of a
III – 2 CAPACITY MEASUREMENT supercapacitor. It expresses the ability of a
The capacity expresses the total charge supercapacitor to store electrical energy.
stored or released by the capacitor during a The values of capacitance obtained for the
full charge-discharge process. It can be dis- cycle 1 and for SC400 and SC22 are 392 F and
played for each charge and discharge se- 21 F respectively. These results are in
quence and also for each cycle (1 charge+1 agreement with the nominal capacitance
discharge). values provided by manufacturers which are
Fig. 5a and 5b show the evolution (vs. cycle 400 F + 10%/- 5% and 22 F +/- 30%.
number) of the measured capacity Qch & The two supercapacitors have been cycled
Qdisch during charge and discharge of the two over 100 times at charge/discharge rates of 8
tested supercapacitors cycled 100 times us- A and 2 A respectively. The evolutions of the
ing a galvanostatic charge-discharge protocol capacitances vs. cycle number are presented
with current rates of 8 A and 2 A. in the Fig. 6a and 6b below.

Fig. 6a: Charge & discharge Capacitance vs. cycle

Fig. 5a: Capacity vs. cycle number for the SC400.
number for the SC400 supercapacitor.

Bio-Logic Science Instruments, 1 Rue de l'Europe, 38640 Claix, FRANCE 3

Tel: +33 476 98 68 31 – Fax: +33 476 98 69 09 www.bio-logic.info
 EC-Lab - Application Note #51
04/28/2014

curve. The obtained values of capacitance are

387 F and 21 F for SC400 and SC22
respectively.

Contrary to the classical method which leads

to large errors for low rates of
charge/discharge where the Ewe vs. t curve is
not perfectly linear, the “Capacity and Energy
per Cycle or Sequence” analysis tool provide
Fig. 6b: Charge & discharge Capacitance vs. cycle an accurate capacitance value for low
number for the SC22 supercapacitor. charge/discharge rates.
The SC400 was cycled over 100 times at
The Fig. 6a and 6b show a shallow, constant charge/discharge rate of 8 A, 4 A and 2 A.
decrease of capacitance upon cycling. The The Fig. 8 shows a decrease of the measured
measured capacitance decrease reached capacitance with the increase of the cycling
around 2% of the initial measured value after current. The capacitance decrease with in-
100 cycles. creasing charge-discharge current was al-
In order to check the evolution of the meas- ready observed in numerous studies [11, 12].
ured capacitance for a long time a long-term The obtained values of capacitance are more
cycling test (10,000 cycles) was performed on accurate and close to the nominal
the two supercapacitors. The result shows a capacitance for the low current rates.
capacitance decrease during the initial hun-
dred cycles. The decrease becomes slightly
smaller with growing cycle number. The Fig.
7 shows the obtained capacitance vs. cycle
plot for the SC22.

Fig. 8: Capacitance vs. cycle number for the SC400 with

three current rates (green 2 A, red 4 A, blue 8 A).

Fig. 7: Capacitance vs. cycle number for a long-term

cycling test of the SC22. III – 4 EFFICIENCY MEASUREMENTS
Once the charge-discharge data is processed
The capacitance was also calculated for the by the analysis tool the coulombic efficiency
same charge-discharge rates using the vs. cycle number can be automatically
classical equation below: plotted. The Fig. 9a and 9b show the
obtained graphs.

Where Ich/disch is the charge or discharge

current and abs (slope Ewe vs. t) is the
absolute value of the slope of Ewe vs. time

Bio-Logic Science Instruments, 1 Rue de l'Europe, 38640 Claix, FRANCE 4

Tel: +33 476 98 68 31 – Fax: +33 476 98 69 09 www.bio-logic.info
 EC-Lab - Application Note #51
04/28/2014

the frequency range 200 kHz-100 mHz with a

voltage amplitude of 10 mV.
Fig. 10a and 10b present the Nyquist dia-
grams for the two capacitors.

Fig. 9a: Coulombic efficiency vs. cycle number for

SC400.

Fig. 10a: Nyquist diagram of SC400 supercapacitor.

Fig. 9b: Coulombic efficiency vs. cycle number for

SC22.

As shown in the Fig. 9a and 9b, both tested

supercapacitors display a coulombic
efficiency higher than 99%.

III – 5 ESR MEASUREMENT

The internal resistance or Equivalent Series
Resistance of a supercapacitor is classically
determined from the charge-discharge
curves using the voltage drop at the begin-
ning of the discharge curve. The internal re- Fig. 10b: Nyquist diagram of SC22 supercapacitor.
sistance which is also referred to as the
Equivalent Series Resistance (ESR) [6] is due In the high-frequency range, the Nyquist dia-
to resistance of terminal leads, electrolyte, grams show inductive behavior and then,
separator and current collectors. intersect the real axis at approximately
Two different methods have been used to Re(Z)= 2.7 mΩ and 31.0 mΩ for SC400 and
determine the internal resistance of each SC22, respectively. These values, also ob-
supercapacitor: EIS method and galvanostatic tained by fitting the two diagrams, corre-
polarization method. spond to internal resistance ESR [6-8] of the
a- EIS method two supercapacitors.
Impedance measurements at rated voltage
were performed on each supercapacitor in Fig. 10c shows the fitted Nyquist diagram for
the SC22.

Bio-Logic Science Instruments, 1 Rue de l'Europe, 38640 Claix, FRANCE 5

Tel: +33 476 98 68 31 – Fax: +33 476 98 69 09 www.bio-logic.info
 EC-Lab - Application Note #51
04/28/2014

Where Rd is the diffusion resistance, ω is the

angular frequency, τ is the time constant and
α a real parameter equal to 0.94.

The values of the ESR obtained from the fit of

the Nyquist diagrams are close to those pro-
vided by the supercapacitors manufacturers
(respectively ESR (SC400) = 2.0 mΩ & ESR
(SC22) < 100 mΩ).

As the internal resistance values of superca-

pacitors are very low the use of the four ter-
minal connection points is strongly recom-
mended to minimize the effect of the wiring
on the value of ESR.

Fig. 10c: Fitting of the Nyquist diagram of SC22.

The equivalent circuit used to model the

Nyquist diagram of the SC22 is shown below
Fig. 12: Four terminal connection points performed
with SP-300/10A system.

b- Galvanostatic polarization method

Fig. 11: Equivalent circuit used for Nyquist diagram of The internal resistances of the two superca-
SC22 supercapacitor.
pacitors were also estimated from the volt-
age drop (iRdrop) divided by the total change
Where R1 is the internal resistance (ESR) of
in current applied (2. Ich/disch) using the fol-
SC22, L1 inductance due to the wiring setup.
lowing equation [10]:
R2 and Q2 were used to fit the semicircle at
high frequency. R2 and Q2 are the charge
ESR =iRdrop/2.Ich/disch
transfer resistance and the CPE correspond-
ing to the charge transfer resistance in paral-
The Fig. 13 shows one charge discharge cycle
lel with the double layer capacitance. Ma3 is
performed on SC400 supercapacitor with a
the Modified restricted diffusion element.
zoom on the voltage drop.
Indeed at low frequencies (lower than 1 Hz)
Fig 10c shows a typical modified restricted
diffusion behavior. Thus the “Modified re-
stricted diffusion element” Ma element
available in Z Fit library of EC-Lab® was used
for modeling this kind of diffusion. The im-
pedance expression of Ma is as follows [9].

Bio-Logic Science Instruments, 1 Rue de l'Europe, 38640 Claix, FRANCE 6

Tel: +33 476 98 68 31 – Fax: +33 476 98 69 09 www.bio-logic.info
 EC-Lab - Application Note #51
04/28/2014

REFERENCES
[1] M. D. Stoller, R.S. Ruoff, Energy Environ.
Sci., 3 (2010) 1294.
[2] Application Note #33 “Supercapacitors
investigations – Part I: Charge/discharge
cycling” http://www.bio-
logic.info/potentiostat-electrochemistry-ec-
lab/apps-literature.
[3] Application Note #34 ”Supercapacitors
investigations – Part II: Time constant
Fig. 13: IR drop at the initiation of a 8A discharge
curve of SC400 supercapacitor. determination”http://www.bio-
logic.info/potentiostat-electrochemistry-ec-
As can be seen from Fig. 13, the voltage drop lab/apps-liter-ature.
observed during switching from charge to
discharge regime is ΔEwe= 43 mV. This voltage [4] C. T. Hsieh, Y.W. Chou, W.Y. Chen, J. Solid
was divided by twice the applied current. The State Chem., 12 (2008) 663.
obtained ESR value is ESR = 2.7 mΩ. [5] H. Yu, Q. Tang, J. Wu, Y. Lin, L. Fan, M.
Huang, J. Lin, Y. Li, F. Yu, J. Power Sources
IV – CONCLUSIONS 206 (2012) 463.
Galvanostatic polarization method is a fast [6] K.H. An, K.K. Jeon, J.K. Heo, S.C. Lim, D.J.
and easy method for the determination of Bae, Y.H. Leez, J. Electrochem. Soc., 149
supercapacitor characteristics. The “Capacity (2002) A1062.
and Energy per Cycle or Sequence“ is a pow- [7] X. Liu, P.G. Pickup, J. Power Sources 176
erful tool for the calculation of the superca- (2008) 410.
pacitor parameters (capacity, energy, capaci-
tance & efficiency). These parameters are [8] R. Kotz, M. Hahna, R. Gallay, J. Power
quickly determined by processing the gal- Sources 154 (2006) 550.
vanostatic charge-discharge data using the [9] T. Hang, D. Mukoyama, H. Nara, T. Yoko-
analysis tool. The obtained parameters for shima, T. Momma, M. Li, T. Osaka, J. Power
the two supercapacitors clearly showed a Sources 256 (2014) 226.
good agreement with the characteristics val-
[10] G. Ma, J. Li, K. Sun, H. Peng, J. Mu, Z. Lei,
ues provided by the manufacturers.
J. Power Sources 256 (2014) 281.
Based on an analytical approach, the “Capac- [11] Y. Fu, J. Song, Y. Zhu, C. Cao, J. Power
ity & Energy per Cycle or Sequence“ provides Sources 262 (2014) 344.
accurate values of supercapacitors parame- [12] Y. Bai, R.B. Rakhi, W. Chen, H.N.
ters. Alshareef, J. Power Sources 233 (2013) 313.

Bio-Logic Science Instruments, 1 Rue de l'Europe, 38640 Claix, FRANCE 7

Tel: +33 476 98 68 31 – Fax: +33 476 98 69 09 www.bio-logic.info