Electrochemical oscillations in the methanol oxidation on Pt

Jaeyoung Lee *, Christian Eickes, Markus Eiswirth, Gerhard Ertl

Fritz-Haber-Institut der Max-Planck-Gesellschaft, Faradayweg 4-6, D-14195 Berlin, Germany

Received 22 June 2001

Abstract

Experimental observations of temporal dynamics in the electrocatalytic oxidation of methanol (CH3OH) on a polycrystalline

platinum electrode are reported. Hidden negative differential resistance (HNDR) and instabilities of the system were investigated by

means of an electrochemical impedance spectrum analysis and potential oscillations under galvanostatic control. This result could

be applied for a direct methanol fuel cell (DMFC) with higher energy efficiency. # 2002 Published by Elsevier Science Ltd.

Keywords: Methanol; Electro-oxidation; Hidden negative differential resistance; Galvanostatic potential oscillations

1. Introduction

There has been a resurgence of interest in the studies

of the oscillatory behaviour in electrochemical systems

like electrocatalysis, electrodeposition, and electro-dis-

solution [1�/15]. In electrocatalysis, the most important

system with regard to dynamic behaviour of small

organic molecules, is formic acid (HCOOH). Current

oscillations under potentiostatic conditions and poten-

tial oscillations under galvanostatic systems of HCOOH

on a Pt electrode were investigated [5�/7]. However, only

a few studies of potential oscillations during anodic

CH3OH oxidation were reported [8�/10]. Pavela re-

ported that diffusion of CH3OH might be the reason

for the potential oscillations [8]. In early 1960s, Buck

and Griffith [9] observed only small potential oscilla-

tions in the CH3OH oxidation. Krausa and Vielstich

studied the experimental conditions of potential oscilla-

tions during CH3OH oxidation on a Pt electrode and

they suggested a reaction mechanism resulting in

different types of oscillations [10]. The investigation of

the oxidation mechanism of CH3OH and HCOOH on a

Pt electrode has attracted attention, since the under-

standing of the oxidation of C1 fuels can be applied for

the development of clean energy techniques (fuel cell).

While the temporal as well as spatiotemporal behaviour

of HCOOH oxidation has been studied in detail bothexperimentally and theoretically, the mechanistic origin

of bistability and oscillations in CH3OH oxidation is not

well understood [7,11�/15].

In this work, we try to categorise the mechanistic

origin of the CH3OH oxidation by electrochemical

impedance spectroscopy (EIS) and the bifurcation

sequence under galvanostatic conditions. This study

could be also considered as a reference to understandthe electro-oxidation of C2 and C3 compounds [16,17]

and can give important clues to get higher power output

under oscillatory conditions.

2. Experimental

The schematic diagram of the principal three-elec-

trode arrangement is shown in Fig. 1. The electroche-

mical cell body consisted of a glass cylinder capped witha Teflon lid holding all electrodes. A smooth polycrys-

talline Pt ring of 34.5 mm inner diameter and 42.5 mm

outer diameter cut from a Pt plate of 0.1 mm thickness,

thus exhibiting a geometric surface area of 10 cm2, was

used as the working electrode (WE). All solutions were

prepared with ultrapure water (Millipore Milli-Q water,

18 MV �/cm). Prior to each experiment, the Pt ring WE

was first subjected to cleaning with C3H6O and purewater in an ultrasonic bath followed by chemical

cleaning in conc. H2SO4 (98%)�/30% H2O2 (1:1). Then,

a preliminary check of the voltammetric curve between* Corresponding author.

E-mail address: jlee@fhi-berlin.mpg.de (J. Lee).

Electrochimica Acta 47 (2002) 2297�/2301

www.elsevier.com/locate/electacta

0013-4686/02/$ - see front matter # 2002 Published by Elsevier Science Ltd.

PII: S 0 0 1 3 - 4 6 8 6 ( 0 2 ) 0 0 0 7 5 - 0

�/600 and �/700 mV of the WE with scan rate of 100

mV s�1 was performed in 0.5 M H2SO4 (suprapur,

MERCK) deaerated by N2 to verify initial cleanliness.

After this preliminary check of the Pt electrode, WE was

mounted in the Teflon lid of the cell which already held

both by Hg�/Hg2SO4, K2SO4 (saturated) referenceelectrode and by the platinized Pt counter electrode.

CH3OH (0.03 M, p.a. MERCK) in 0.1 M HClO4

(suprapur, MERCK) was used for CH3OH oxidation

and the electrolyte was extensively bubbled with N2

before each experiment. A nitrogen atmosphere was

maintained over the unstirred solutions during sweep

experiments.

Cyclic voltammogram, galvanostatic scan, and con-stant current method were executed with an in-house-

built potentiostat (Electronic Laboratory of Fritz-Ha-

ber-Institut) and a bi-potentiostat (EG&G, Model 366).

The impedance spectrum was measured by a frequency

response analyzer (S-5720, NF Electronic Instruments,

Yokohama). The frequency range from 10 kHz to 0.1

Hz was digitally recorded with 25 points per decade.

3. Results

Fig. 2 shows the typical I /U profile in the CH3OH

oxidation on the Pt electrode. On the anodic scan the

adsorption of CH3OH and dehydrogenation of metha-nol adsorbed result in the formation of Pt�/CO at low

overpotential that is marked with (a) in Fig. 2. We

observed a current peak around �/50 mV, in which

surface water might react with adsorbed CO on a Pt

electrode and produce CO2, proton (H�), electron (e�),

and vacant site (*) of Pt surface. With increasing anodic

potential, the current decreases due to the formation of

hydroxide on the Pt surface. On the cathodic scan, thedesorption of Pt�/OH results in a burst of anodic current

at �/150 mV and the oxidation of CH3OH follows. We

assume that the current plateau marked with (b) in Fig.

2 may be associated with galvanostatic potential oscilla-

tions, but the reaction mechanism is still unclear.

Fig. 3 shows the galvanostatic scan (I /U ) of CH3OH

oxidation on a Pt electrode. Scan rate is �/25 mA s�1. A

hard transition from oscillatory to stationary behaviour

is seen at high current, while a soft supercritical Hopf

bifurcation is discerned at low current. This measure-

ment reveals an onset potential of spontaneous sus-

tained oscillations between �/45 and �/65 mV. The

value of the anodic current of oscillation initiation is

about �/17 mA. According to previous mechanistic

classification [11,18], we assume that the origin of

potential oscillations in CH3OH oxidation belong to

class IV, i.e. hidden negative differential resistance

(HNDR).

To prove the mechanistic origin of CH3OH oxidation

on a Pt electrode classified as HNDR, impedance

spectra were measured for various potentiostatic opera-

tion points prior to the onset of periodic instabilities.

EIS is a common tool to investigate the kinetics of

electrochemical processes over a large range of frequen-

cies. This makes it an ideal predictive technique to probe

electrochemical systems for their capability to exhibit

Fig. 1. Experimental set-up (2D plot).

Fig. 2. Cyclic voltammogram at a Pt electrode with scan rate of 10 mV

s�1 in 0.03 M CH3OH�/0.1 M HClO4 solution. Current plateau

around �/60 mV on the positive going scan is obtained.

Fig. 3. Onset of sustained potential oscillations (indicating Hopf

bifurcation) in the galvanostatic operation mode (I /U plot) with scan

rate of 25 mA s�1 in 0.03 M CH3OH�/0.1 M HClO4.

J. Lee et al. / Electrochimica Acta 47 (2002) 2297�/23012298

dynamical instabilities. Nyquist plots at three different

applied outer potentials (U ) and the Bode plot at �/65

mV are shown in Fig. 4. At outer potential of �/45 mV

(Fig. 4a), the impedance plot Z (v) exhibits a clockwise

capacitive�/inductive loop indicating dynamically stable

stationary operating points. However, the impedance

plot is drastically changed at higher overpotential �/65

mV, as shown in Fig. 4b. Z (v ) wraps around the origin

anti-clockwise. The intersection of the impedance and

the negative real axis is obtained at a finite frequency,

v0:/1 Hz. Then Z (v ) again exhibits positive real

resistances for very low frequencies. The detailed shape

of the Nyquist plot bears information on the mechan-

istic origin of the oscillatory behaviour, which can be

used to predict the dynamical stability of oscillatory

systems.

Fig. 5 is the Bode plot at �/65 mV and it clearly

demonstrates that the phase angle jumps from �/1808 to

�/1808 at frequency v0:/1 Hz, i.e. this drastic phase

change results in an anti-clockwise impedance profile in

the Nyquist plot (Fig. 4b and c).

Fig. 6a and b show the potential oscillations at

constant anodic current of �/14.5 mA at different times.

It was measured without applying any forced convectionsuch as stirring or nitrogen bubbling. Note that the

value of v0 in impedance spectra (see Fig. 4b and c), the

intersection where Im Z�/0 with Re Z B/0, agrees with

the initial oscillation frequency of 1 Hz obtained in Fig.

6. As discussed in previous work [18,19], the critical

potential, where the Nyquist plot blows up such that

approaches 9/�, marks the galvanostatic transition to

periodic behaviour, i.e. Hopf bifurcation. The value ofv0 represents the intrinsic frequency of the oscillation in

the electrocatalytic oxidation of CH3OH on a pure

polycrystalline Pt.

4. Discussion

According to previous work [4], stepwise dehydro-

genation mechanism for the oxidation of methanol is asfollows:

CH3OH�Pt 0 Pt�CH2OH�H��e� (1)

Pt�CH2OH�Pt 0 Pt2�CHOH�H��e� (2)

Pt2�CHOH�Pt 0 Pt3�COH�H��e� (3)

Pt3�COH 0 Pt(3�x)�CO�xPt�H��e� (4)

H2O�Pt 0 Pt�OHad�H��e� (5)

Pt(3�x)�CO�Pt�OHad0(4�x)Pt�CO2��H�

�e� (6)

After applying an anodic current the surface will be

covered by CH3OH residues with different oxidation

states. If reaction rates of dehydrogenation from Eqs.(1)�/(4) are slow, only few vacant Pt sites remain and

thus further dehydrogenation or oxidation of the

CH3OH residues is slowed down further. Therefore,

Fig. 4. Electrochemical impedance behaviour of methanol oxidation at a Pt electrode measured at three outer potentials of: (a) �/45; (b) �/65; and (c)

�/85 mV.

Fig. 5. The Bode plot at �/65 mV shows the phase shift as a function

of frequency. Experimental condition is same as in Fig. 2. Closed and

open circles indicate log jZ j and phase angle, respectively.

J. Lee et al. / Electrochimica Acta 47 (2002) 2297�/2301 2299

the adsorption of H2O must be considered (Eq. (5)) for

the complete oxidation of CH3OH as shown in Eq. (6),

since H2O is the source of oxygen. However, this

adsorption process also needs free Pt sites. Thus, the

potential rises until greater amounts of Pt�/OH are

formed and more free Pt sites are present by the

autocatalytic reaction (see Eq. (6)) of the oxidation ofPt�/CO and the conversion of the residues can proceed

with a higher reaction rate. During this reaction the

potential will decrease slowly until larger areas of the

surface are empty, then the potential decreases rapidly

for about 200 mV in a negative direction.

Of late, it was [11] suggested in a simple theoretical

calculation that higher power output of direct C1 fuel

cell might be possible by applying autonomous and/orexternally driven potential oscillations. This result is due

to smaller dissipation (energy loss) in the average

dissipation in periodic condition compared to the

stationary system. The detail theoretical explanation is

as follows.

Dissipation (D ) and efficiency (h ) are defined as D�/

dS /dt (S , entropy) and h�/Pout/Pin, respectively and we

normalise the efficiency like h�/1�/D ? according toD�/Pin�/Pout. Dissipation is time-dependent term.

�D��v

2p g2p=v

0

D(t) dt (period�2p=v) (7)

DD��D��Dstationary (8)

We can calculate average value of dissipation (�D�)

when the period is 2p /v . To increase the power

efficiency of direct C1 fuel cell, negative DD is wanted

in time-periodic processes.

In linear thermodynamics, DD is always bigger than

zero. But in nonlinear systems, the phase between fluxand force can vary over a wide range. If it is p (or �/p),

DD has maximum negative value, that is maximum

efficiency can be obtained by applying a periodic

perturbation which results in a phase of 1808 phase

between flux (current) and force (potential).

DD�kDU 2 v

2p g2p=v

0

cos vt cos(vt�8 ) dt

�1

2kDU 2 cos 8 (9)

8:0 0 DD�1

2kDU 2 (10a)

8:p 0 DD��1

2kDU 2 (10b)

where 8 is the phase between I and U ; �D�, averagevalue of dissipation when period is 2p /v . The required

perturbation period can be directly seen from the

impedance spectra.

It should be noted that the above thermodynamic

considerations only compare the performance of a given

electrode under stationary and periodic operation.

Certainly a crucial point, not addressed in this work,

is an improvement of the electrode kinetics, such as Ru-modification of a Pt electrode, which can reduce the

amount of CO poisoning [20].

5. Conclusions

EIS and the galvanostatic scan of CH3OH oxidation

on Pt exhibits a HNDR and a Hopf bifurcation at low

overpotential. Therefore, this oscillator belong to class

IV according to an earlier mechanistic classification

[11,18], as did HCOOH oxidation [5,7]. Comparison of

stationary conditions with oscillatory systems, higher

efficiency of direct C1 fuel cell is theoretically possible.Thus, further study will be carried out to get higher

power output by applying an external pulse with correct

frequency of a 1 Hz obtained by EIS technique.

Fig. 6. Potential oscillations at constant current of �/14.5 mA show oscillations frequency of 1 Hz. It is in good agreement with the result obtained in

Fig. 4b and c. (a) At initial oscillations time; and (b) at hard transition time.

J. Lee et al. / Electrochimica Acta 47 (2002) 2297�/23012300

Acknowledgements

Partial financial support by the Deutsche Forschungs-

gemeinschaft (DFG) through Sfb 555 is gratefullyacknowledged.

References

[1] T. Matsuda, Y. MuLouyama, H. Hommura, S. Yae, Y. Nakato,

J. Electrochem. Soc. 144 (1997) 2996.

[2] E.W. Bohannan, L.-Y. Huang, F.S. Miller, M.G. Shumsky, J.A.

Switzer, Langmuir 15 (1999) 813.

[3] D.T. Bowlin, A. Scheeline, A. Pearlstein, Electrochim. Acta 43

(1998) 417.

[4] V.S. Bagotsky, Yu.B. Vassiliev, O.A. Khazova, J. Electroanal.

Chem. 81 (1977) 229.

[5] J. Lee, P. Strasser, M. Eiswirth, G. Ertl, Electrochim. Acta 47

(2001) 501.

[6] W. Vielstich, Fuel Cells, Wiley, New York 1970.

[7] J. Lee, J. Christoph, P. Strasser, M. Eiswirth, G. Ertl, J. Chem.

Phys. 115 (2001) 1485.

[8] T.O. Pavela, Suomen Kemistilehti 30 (1957) 138.

[9] R.P. Buck, L.R. Griffith, J. Electrochem. Soc. 109 (1962) 1005.

[10] M. Krausa, W. Vielstich, J. Electroanal. Chem. 399 (1995) 7.

[11] P. Strasser, Ph.D. thesis, Freie Universitat Berlin, Berlin, 1999.

[12] C. Lamy, J.M. Leger, J. Clavilier, R. Parsons, J. Electroanal.

Chem. 150 (1983) 71.

[13] R.R. Adzic, W.E. O’Grady, S. Srinivasan, J. Electrochem. Soc.

128 (1981) 1913.

[14] J. Clavilier, C. Lamy, J.M. Leger, J. Electroanal. Chem. 125

(1981) 249.

[15] S. Sriramulu, T.D. Jarvi, E.M. Stuve, J. Electroananl. Chem. 467

(1999) 132.

[16] S. Chen, M. Schell, J. Electroanal. Chem. 478 (1999) 108.

[17] J.-W. Kim, S.-M. Park, J. Electrochem. Soc. 146 (1999) 1075.

[18] P. Strasser, M. Eiswirth, M.T.M. Koper, J. Electroanal. Chem.

478 (1999) 50.

[19] M.T.M. Koper, Adv. Chem. Phys. 92 (1996) 161.

[20] J. Lipkowski, P.N. Ross, Electrocatalysis, Wiley-VCH, New York

1998.

J. Lee et al. / Electrochimica Acta 47 (2002) 2297�/2301 2301