PII:

Electrochimica Acta, Vol. 43, No. 24, pp. 36753680, 1998 # 1998 Published by Elsevier Science Ltd. All rights reserved Printed in Great Britain S0013-4686(98)00125-X 00134686/98 $19.00 + 0.00

Inuence of the PTFE content in the diusion layer of low-Pt loading electrodes for polymer electrolyte fuel cells
L. Giorgi,a* E. Antolini,a A. Pozioa and E. Passalacquab
a

ENEA, C.R. Casaccia, ERG-TEA-ECHI, Via Anguillarese 301, 00060 S. Maria di Galeria, Roma, Italy b Institute CNR-TAE, Via Salita S. Lucia sopra Contesse 39, 98126 Messina, Italy (Received 7 February 1998; in revised form 7 February 1998)

AbstractThe inuence of the structure and composition of the diusion layer on polymer electrolyte fuel cell (PEFC) cathode performance was investigated. Electrodes were prepared with dierent poly-tetrauoroethylene (PTFE) content in the diusion layer and maintaining a constant composition for the catalytic layer with a low-Pt loading (0.11 mg cm2). Electrodes were characterized by Hg-intrusion porosimetry, scanning electron microscopy and electrochemical techniques (cyclic voltammetry, galvanostatic polarization and ac-impedance spectroscopy). # 1998 Elsevier Science Ltd. All rights reserved Key words: polymer electrolyte fuel cell, gas diusion electrode, PTFE loading, impedance, oxygen transport. INTRODUCTION Electrode structures in polymer electrolyte fuel cell (PEFC) have evolved from two layers to three layers structure [1]. Dual-layer electrodes are constituted by a hydrophobic support layer and a porous electrocatalyst layer formed by Pt on carbon (Pt/C) as a catalyst agent, polytetrauoroethylene (PTFE) and polyperuorosulphonic acid (Naon). Threelayer electrodes are made by a porous support layer, a diusion layer consisting of carbon particles and PTFE and a catalyst layer formed by Pt/C and Naon. The performances of the electrodes depend on many parameters [25]: (1) type of support (carbon paper or carbon cloth) and its characteristics (porosity and thickness); (2) type of catalyst (Pt alone, Pt with other metals, macrocycles, Pt amount, Pt particles size, type of carbon as the support); (3) PTFE amount; (4) Naon amount; (5) thermal treatment; (6) thickness of diusion and catalytic layers; 7) fabrication process. The fabrication process of both diusion and catalyst layer can be carry out by dierent techniques as screen-printing, rolling, brushing, ltering and spraying. The incorporation of PTFE into the diusion or catalyst layer serves two functions: binding the high surface carbon particles into a cohesive layer, and imparting some hydrophobic character to the layer.
*Author to whom correspondence should be addressed. Tel.: +39 6 30486358; Fax: +39 6 30486357; E-mail: giorgil@casaccia.enea.it

Recently, Paganin et al. [6] investigated the inuence of PTFE loading in the diusion layer for a three-layer structure, in the case of 20 wt% Pt/C and 0.4 mg Pt/cm2; they found a small increase in the performance when the PTFE content is diminished from 40 to 15 wt%. Little has been published [7] hitherto about the detailed micromorphology of three layers low Ptloading electrodes and the role of electrode porosity on electrode performance. It is well known, that the porosity of thicker fuel cell electrodes (e.g. of phosphoric acid and alkaline fuel cells) is decisive for their performance. Therefore, the aim of this paper was to investigate the inuence of the diusion layer porosity on the eective catalyst activity for the cathodic oxygen reduction. The inuence of the composition (PTFE content) and structure of diusional layer on the performance of low Pt-loading electrodes for PEFC was investigated. Gas diusion electrodes with a three-layer structure and a low catalyst loading were prepared. A spray technique for the deposition of both diusion and catalyst layers was used. EXPERIMENTAL The electrodes were prepared by a spray technique using platinum-on-carbon (20% Pt/C) catalyst (Electrochem.), grounded carbon powder (Vulcan XC-72, Cabot), carbon paper substrate (Toray TGPH-090), polytetrauoroethylene (PTFE)

3675

3676

L. Giorgi et al.

dispersion (35 wt% PTFE aqueous dispersion, Hostaon 5033, Hoechst), Naon solution (5 wt% Naon 1100 EW solution, Solution Technology) and glycerol. The carbon loading in the diusion layer was 2 mg cm2; the platinum loading in the catalyst layer was 0.11 mg cm2. For the preparation of the diusion layer, a homogeneous suspension was prepared by mixing and stirring in an ultrasonic bath at room temperature for 25 min the carbon with an appropriate amount of PTFE dispersion. The suspension was spray deposited onto a porous support. The layer was dried in air at 1208C for 1 h, followed by a thermal treatment at 2808C for 30 min to remove the dispersion agent contained in PTFE, and nally sintered at 3508C for 30 min. To prepare the catalyst layer, a homogeneous suspension, formed from the desired amounts of the Pt/C catalyst, the glycerol and the Naon solution with ethanol as solvent, was spray deposited onto the diusion layer. Porosimetry measurements were performed using a Carlo Erba Mercury Porosimeter mod. 2000. Micrographs were obtained with a Cambridge Stereoscan 250 MK3 scanning electron microscope. A pyrex cell with a platinum foil as counter electrode and a reversible saturated calomel reference electrode was used for all electrochemical measurements. The working electrode (1 cm2 of geometric area) was inserted in a PTFE specimen holder, which allowed to expose the electrode active reaction layer to the electrolyte (H2SO4 1 M at 258C) and the backside to the owing gas. Cyclic voltammetry and galvanostatic polarization were carried out using a potentiostatgalvanostat EG&G PAR 273A interfaced via GPIB to computer HewlettPackard Vectra 5/75. Characteristic curves E vs i were recorded in oxygen at a owing rate of 100 mL min1. To determine the charge associated to active area, cyclic voltammetries (CV) were recorded after cycling the electrodes until a stationary response was obtained; the electrolyte was purged with argon and CV proles were recorded at a scan rate of 100 mV s1 between the potential limits for hydrogen and oxygen evolution. The instrumentation for the electrochemical impedance spectroscopy (EIS) consisted in an electrochemical interface Solartron Instruments mod. 1287 and an impedance phase analyzer Solartron Instruments

Fig. 1. Galvanostatic steady-state polarization curves for the oxygen reduction at 258C in H2SO4 1 M on gas diusion electrodes, with dierent PTFE loading in the diusion layer (0.11 mg Pt cm2 in the catalytic layer).

mod. 1260, both interfaced via GPIB to a computer Hewlett-Packard Vectra 5/75. The EIS measurements were carried out at various cathodic potentials in the range from 500 to 900 mV vs NHE. An alternating sinusoidal signal of 10 mV peak-to-peak was superimposed on the dc-potential. The impedance spectra were collected in the frequencies range between 150 kHz and 50 mHz. RESULTS AND DISCUSSION Analysis of the experimental polarization data was made using the semi-empirical equation proposed by Srinivasan et al. [8] E E 0 b log i Ri where E 0 Er b log i0 X 2 Er is the reversible potential for the electrode, b the Tafel slope, i0 the exchange current density for the oxygen reduction reaction (ORR) in the Pt/C catalyst determined by extrapolation at OCV, and R represents the total contribution of the polarization components, which includes the charge transfer resistance for ORR, the electrolyte and the mass transfer resistance. The tting of the above equation to the experimental data was made by a nonlinear least squares method. In Fig. 1, the plots of the electrode potential vs current density, for electrodes with dierent PTFE contents in the diusion layer are shown. In the insert the magnication of the low overvoltage 1

Table 1. Kinetic parameters obtained from the tting of polarization data for the electrodes with dierent PTFE loading in the diusion layer Parameter 10 b (mV dec1) i0 (A cm2105) R (O) E0 (mV) Er (mV) 6722 2.320.2 1.2320.02 56926 88221 20 672 2 1.62 0.2 1.552 0.02 5532 5 8762 1 PTFE loading (wt%) 40 66 22 0.5 20.1 1.91 20.04 561 27 915 21 60 69 23 2.520.4 2.11 20.04 559 28 891 21

Inuence of the PTFE content in the diusion layer of low-Pt

3677

Fig. 2. Total porosity of the diusion layer as a function of the PTFE content.

Fig. 4. Dependence of the electrode potential, during the oxygen reduction at 5 mA cm2, as a function of porosity in the range from 2 to 50 mm.

region of the polarization curves is reported. At high current densities, the electrode performance decreases with increasing PTFE content in the diusion layer, while at low current densities the electrode performance goes through a minimum, for 20 wt% PTFE. As shown in Table 1, some electrochemical parameters depends on the PTFE loading. In particular the resistance R greatly increases with PTFE content (two times when PTFE content changed from 10 to 60 wt%). These results indicate that for this type of electrodes, the PTFE loading has an important eect on the performance. As just explained, the term R is the sum of the polarization resistance Rp (charge and mass transfer), the electrolyte (Rel) and the electrode (Re) resistance. We can consider the last two terms to be constant, as drawn out from EIS measurements. So the change in the slope of E vs i plot, is related to diusion and/or charge transfer phenomena. The diusion layer has two distinctive pore distributions with boundary at about 0.35 mm. To identify these dierent regions, we call micropores the pores with size lower than 0.35 mm and macropores the pores having size higher than 0.35 mm. Figure 2 shows the dependence of total porosity on PTFE content of the diusion layer: total porosity decreases with increasing PTFE loading. As it can be observed, the electrode performance at high current density shows the same trend of the total porosity data. Diusion layer porosity inuences the electrode performance as the rate of mass transport

Fig. 3. Porosity in the range from 2 to 50 mm and macroporosity of the diusion layer as a function of PTFE content.

of reactant gas on catalytic sites increases with total porosity of the diusion layer. In the high current density region, the electrode is under diusion control. So in this case, the contribute of total porosity, related to a larger amount of reactant gas on catalytic sites, is predominant on electrode performance. In the low current density range (15 mA cm2), where the electrode is under the control of charge transfer reaction (insert in Fig. 1), the electrode performance showed a dierent trend than that regarding high current density region. To explain this behavior, it is necessary to take into account the inuence of macroporosity on the electrode performance. In Fig. 3 the diusion layer macroporosity and the porosity in a pore size range between 2 and 50 mm, are reported as a function of PTFE loading. A minimum appears for a PTFE loading between 20 and 40 wt%. By plotting the electrode potential vs the macroporosity, in the range from 2 to 50 mm, we obtain a straight line, as shown in Fig. 4. This behavior can be explained considering that a diusion layer with a high macroporosity allows an higher eective catalytic area, because spray deposited catalyst can penetrate into the larger pore of the diusion layer and the triple contact zone Pt/Naon/gas is increased. Exchange current density reported in Table 1 shows a minimum value at 40 wt% PTFE, conrming the change in the electrode eective area. SEM micrographs of the diusion layer on the carbon substrate, before the deposition of catalytic layer, for dierent PTFE loading electrode seem to conrm these conclusions. In Fig. 5(a) both macropores and the bers of the carbon paper are visible. At higher PTFE contents (Fig. 5(bd)), it seems that the polymer covers all the surface of the layer, but the presence of cracks, increasing with PTFE content, is also revealed. The change of eective electrode area with diusion layer macroporosity is evident only in a limited electrode potential vs current density region, while the predominant eect of PTFE loading increase appears to be related to the decrease of oxygen amount on the catalytic sites. Such an anomalous eect denoted at low current density for electrode with dierent PTFE loading has been conrmed from electrochemical impedance spectroscopy data.

3678

L. Giorgi et al. tial of 782 mV vs NHE. It mainly consists of one semicircle characteristic of an electrochemical interface that can be simulated, by a non linear re-

Impedance spectra have been obtained for the gas diusion electrodes at various potentials. Figure 6 shows the Nyquist plot obtained at electrode poten-

Fig. 5(a) and (b)caption opposite

Inuence of the PTFE content in the diusion layer of low-Pt gression, with an equivalent electrical circuit RO(RpCdl), where RO is the sum of electrode and electrolyte resistance, Rp is the polarization resist-

3679

ance and Cdl is the double layer capacity. The polarization resistance is the sum of two terms: charge transfer and diusion resistance. The highest

Fig. 5. Scanning electron micrographs of the surface of diusion layer-carbon paper system with (a) 10, (b) 20, (c) 40 and (d) 60 wt% PTFE (magnication 50).

3680

L. Giorgi et al.

Fig. 6. Nyquist plots of electrodes (0.11 mg Pt cm2) at 782 mV vs NHE, as a function of the PTFE content in the diusion layer.

Fig. 8. Double layer capacitance from impedance and hydrogen adsorption charge from cyclic voltammetry, as a function of PTFE content in the diusion layer.

impedance has been obtained for electrode with 20 wt% of PTFE. As the electrode overpotential increases, the charge transfer resistance and the relaxation time of the charge transfer process decrease, whereas diusion resistance increases and the RO resistance is constant. The highest impedance has been measured for 60 wt% PTFE electrode. Figure 7 shows Rp for dierent PTFE loading electrodes at 782 mV vs NHE and 542 mV vs NHE. The Rp trend is similar to that obtained from electrode potential vs current density curves. These data show a dierent electrode performance at high and low overpotential. Polarization resistance and double layer capacity are related to maximum frequency by the following equation: 1 o max X 3 Rp Cdl From the omax value and assuming that the semicircle diameter like the polarization resistance, we can calculate double layer capacity. Double layer capacity is related to the ratio between eective electrode area and double layer thickness. In this case the catalytic layer for dierent PTFE loading electrodes, is the same. So we can attribute the change of Cdl values to the diusion layer morphology. Figure 8 shows the double layer capacity obtained from impedance spectra at 782 mV vs NHE; the curve is similar to that obtained from macroporosity data (Fig. 3) related to the diusion layer, showing a minimum value for a PTFE loading of 40 wt%. The same trend as capacity, has

been obtained for adsorbed hydrogen charge from cyclic voltammetry data (Fig. 8). CONCLUSIONS The PTFE content in the diusion layer of PEFC cathodes changes the electrode microstructure in two ways. First, total porosity of the diusion layer decreases by increasing PTFE content; second, the catalytic surface area depends on the macroporosity of the diusion layer, which varies with PTFE loading. These changes aect the electrochemical properties of the electrode. On this basis the eect of PTFE loading on cathode performance depends on the current density region: * at high current density (diusion layer control), as a consequence of total porosity increase, the mass transport rate, as well as the performance, increases by decreasing PTFE loading; * at low current density (catalytic layer control), as a consequence of macroporosity change, the performance goes through a minimum, at about 20 wt% PTFE. Thus the global best performances are obtained at the lowest PTFE loading, but it is not possible to reduce to zero the PTFE content to avoid the electrode ooding for lack of hydrophobicity. Furthermore a minimum quantity of PTFE in the diusion layer is necessary to bond together the carbon particles. REFERENCES
1. Proceedings of 1996 Fuel Cell Seminar, November 17 20, 1996, Orlando. 2. Proceedings of 2nd International Fuel Cell Conference, February 58, 1996, Kobe. 3. T. Maoka, Electrochim. Acta 33, 379 (1988). 4. M. Schulze, N. Wagner, G. Steinhilber, E. Gulzow, M. Wohr and K. Bolwin, Proceedings of 1996 Fuel Cell Seminar, November 1720, 1996, Orlando. 5. A. S. Arico, V. Antonucci, V. Alderucci, E. Modica and N. Giordano, J. Appl. Electrochem. 23, 1107 (1993). 6. V. A. Paganin, E. A. Ticianelli and E. R. Gonzalez, J. Appl. Electrochem. 26, 297 (1996). 7. A. Fischer and H. Wendt, Proceedings of 1996 Fuel Cell Seminar, November 1720, 1996, Orlando. 8. E. A. Ticianelli, C. R. Derouin, A. Redondo and S. Srinivasan, J. Electrochem. Soc. 135, 2209 (1988).

Fig. 7. Dependence of total polarization resistance on PTFE content in the diusion layer, at low (782 mV vs NHE) and high (542 mV vs NHE) overvoltage for oxygen reduction.