Pore size effect of the DMFC catalyst supported on porous materials

Pore size effect of the DMFC catalyst supported on porous materials

Available online at www.sciencedirect.com International Journal of Hydrogen Energy 28 (2003) 645 – 650 www.elsevier.com/locate/ijhydene Pore size e'...

357KB Sizes 1 Downloads 38 Views

Available online at www.sciencedirect.com

International Journal of Hydrogen Energy 28 (2003) 645 – 650 www.elsevier.com/locate/ijhydene

Pore size e'ect of the DMFC catalyst supported on porous materials Gu-Gon Park∗ , Tae-Hyun Yang, Young-Gi Yoon, Won-Yong Lee, Chang-Soo Kim Fuel Cell Research Center, Korea Institute of Energy Research, P.O. Box 103, Yusung-Ku, Daejeon 305-343, South Korea

Abstract Activated carbons were employed for the support material of catalyst in the direct methanol fuel cell (DMFC). Until now, most of the papers related to the catalyst of fuel cells reported carbon blacks as catalyst support materials. In the present study, an activated carbon was re-activated by chemical activation method with NaOH at the various temperatures for the development of meso or/and macro pores. By using these pretreated activated carbons, Pt–Ru catalysts on the activated carbons (Pt–Ru/AC) which have various surface areas and porosities were prepared for the anode catalyst of DMFC. Surface areas and the crystal sizes of active metals were measured by N2 adsorption and XRD, respectively. The performance of prepared Pt–Ru/AC catalysts has been evaluated by using the typical I –V curve of air breathing-type DMFC single cell. By this study, the optimum conditions of catalyst were suggested by correlating the catalyst reactivity with surface areas, pore sizes, metal sizes and distances between active metals. ? 2002 International Association for Hydrogen Energy. Published by Elsevier Science Ltd. All rights reserved. Keywords: Activated carbon; NaOH; Chemical activation; Pore size distribution; DMFC

1. Introduction Nowadays, some types of fuel cells are approaching commercialization. Of them, most expected types are direct methanol fuel cell (DMFC) and polymer electrolyte membrane fuel cell (PEMFC). Unlike other types of fuel cells, DMFC needs no fuel processor to reform hydrocarbon fuels. This fact makes DMFC advantageous to apply small and portable devices. In this aspect, it is thought that the air breathing-type DMFC, which operates at room temperature and atmospheric pressure without any forced Aowing of air, is the most feasible application. But its intrinsic low power density is still a big problem which must be overcome. Suggested main topics, to increase the performance of DMFC, are electrolyte membrane and catalyst. In the aspect of DMFC catalyst, most papers focused on the active metals such as binary, ternary and even quaternary metal alloyed catalysts. The typical candidates for active metals are Pt, Ru, Ir, Os, Pd and Rh and the ∗

Corresponding author. E-mail address: [email protected] (G.-G. Park).

supporting material is carbon black [1–3]. Because of its corrosion resistance and electric conductivity, carbon black worked successfully as support material of catalyst in the fuel cells. With regard to the support materials, it was reported that the pretreatment of carbon blacks before preparing the catalysts can a'ect the catalytic performance [4,5]. In the recent study [4], the catalytic performance was related to the surface area of the supports as well as to the size of active metal. But restricted intrinsic surface area of carbon blacks hindered more speciGc study. In the present study, activated carbons were employed for the support material of catalyst. In general, activated carbons have high surface areas and micropores. When it is considered that the sizes of active metals are between 1.5 and 3:0 nm, the desirable morphology of activated carbons should have suHcient surface areas and large mesoporosity, simultaneously. To get such samples, the activated carbons were re-activated by chemical activation method with NaOH [6,7]. With these supports, catalysts which have various surface areas and pore sizes were prepared and the optimum morphology of the DMFC catalyst was suggested to get good catalytic performance. The optimum conditions

0360-3199/03/$ 30.00 ? 2002 International Association for Hydrogen Energy. Published by Elsevier Science Ltd. All rights reserved. PII: S 0 3 6 0 - 3 1 9 9 ( 0 2 ) 0 0 1 4 0 - 4

646

G.-G. Park et al. / International Journal of Hydrogen Energy 28 (2003) 645 – 650

of catalysts were correlated to the catalytic reactivity with surface areas, pore sizes, metal sizes and distances between active metals. 2. Experimental 2.1. Preparation of samples A commercial activated carbon was obtained from Seijung Carbon Technical Corporation (Korea) for the support of DMFC anode catalyst. To get various surface areas and pore sizes, this carbon was re-activated with NaOH which is known as an e'ective activating reagent that develops the porosity of the raw materials [6,7]. Screened (120/170 mesh) activated carbon was mixed with NaOH and water. This well-mixed slurry was dried in rota-evaporator to prepare the impregnated sample. The impregnation ratio of NaOH-to-activated carbon was 2. The re-activation was carried out in a horizontal furnace and the samples were heated (5◦ C=min) from room temperature to the Gnal re-activation temperatures (700◦ C; 800◦ C; 900◦ C and 1000◦ C) in nitrogen Aow (30 ml=min). Samples were kept at the Gnal activation temperature for 2 h. After chemical activation, samples were washed with hot distilled water followed by 0:1 N HC1 solution. Finally, cold distilled water was repeatedly used to remove residual chemicals until the pH of Gltered water was below 7.5. The washed samples were dried at 120◦ C for 12 h. The prepared samples were designated as AC, AC700, AC800, AC900 and AC1000 followed by their treated temperatures. The 60 wt% Pt–Ru catalysts supported on the activated carbons, Pt–Ru/AC, with a Pt:Ru ratio of 1:1 atomic were prepared using the pretreated activated carbons. The precursors of active metals were H2 PtCl6 · H2 O and RuCl3 · xH2 O from Alfa Aesar. Re-activated activated carbons were suspended and stirred in distilled water for 2 h, and the RuCl3 · xH2 O aqueous solution was added slowly to the activated carbon slurry followed by H2 PtCl6 · H2 O aqueous solution. These mixtures were reduced by modiGed borohydride method [4], and then washed repeatedly with distilled water till the pH of the Gltered water became lower than 7.5. Finally, dried at 120◦ C for 12 h in the air condition. 2.2. Characterization of samples SpeciGc surface areas and pore volumes of the prepared activated carbons were determined by nitrogen gas absorption–desorption isotherm. Micromeritics ASAP 2010 was employed for these measurements. Adsorption of N2 , as probe species, was performed at −196◦ C. Surface areas and micropore volumes of samples were determined from the BET and Dubinin–Astakhov (D–A) equations, respectively. BET equation was applied in a relative pressure range from 0.06 to 0.3. The total volume (Vtot ) were estimated to be the liquid volumes of adsorbate (N2 ) at a

relative pressure of 0.98. The mesopore volume was obtained by subtracting the micropore volume (from D–A equation) from the total volume. A Otsuka ELS-8000 zeta-potential analyzer was used to study the change of dielectric property at the surface of activated carbons. The pH of samples was controlled by HC1 and ammonia solutions and the used dilution solution was 0:001 M NaC1. The X-ray di'raction (XRD) data were acquired to calculate the particle size and the inter-particle distance of active metals. The range of 2 from 10◦ to 100◦ was scanned at the rate of 6◦ =min. For target copper K∝( = 1:5406 nm) line was used and to calculate the particle size of Pt crystal Scherrer’s equation was employed. In this case, the (2 2 0) face around the 2 of 68◦ was selected to avoid overlapping with Ru peak. More precise results were obtained by small angle X-ray scattering (SAS) technique with a scan rate of 2 1◦ =min from 63◦ to 73◦ . The morphology of catalysts and its metal size and inter-metal distance were investigated by transmission electronic microscopy (Philips F20). 2.3. Performance test of catalysts The electrode slurries for anode were prepared by mixing together the home-made catalysts (60 wt% Pt–Ru/AC catalysts) and an appropriate amount of 5 wt% NaGon solution (Du Pont) including some kinds of dispersant. Then, these well mixed slurries were coated on the waterproof carbon paper until the amount of Pt became 3 mg Pt=cm2 . Cathode electrodes were prepared by the same method as anode with 8 mg Pt=cm2 of Pt black catalyst. The prepared electrodes were dried in the vacuum oven to remove the residual dispersants. NaGon 115 was used to fabricate membrane electrode assembly (MEA). Before being applied to MEA, membranes were pretreated in three steps to remove trace organic and inorganic contaminants. First, membranes were boiled in 3 wt% H2 O2 solution followed by washing in a distilled water. Next, aqueous 0:5 M H2 SO4 solution was used as a boiling medium. Finally, the membranes were boiled again in the distilled water. Each step took over 4 h. The fabricated MEA was tested by the air breathing-type DMFC, which operates at room temperature and atmospheric pressure without any forced Aowing of air at the cathode side. In this non-Aow type tester, of which active electrode area is 9:6 cm2 , the fuel was just stu'ed on the anode side with 2:5 M aqueous methanol solution. During single cell performance testing of all samples, the temperature was kept in the range of 23–25◦ C. 3. Results and discussion Activated carbon was re-activated to get various surface areas and pore sizes. For the samples, nitrogen adsorption

G.-G. Park et al. / International Journal of Hydrogen Energy 28 (2003) 645 – 650 2.0

1000 800 600 400

1.5

1.0

0.5

0.0

200 0 0.0

Micropore Mesopore

3

1200

Pore Volume[cm /g]

AC AC700 AC800 AC900 AC1000

3

Volume Absorbed[cm /g STP]

1400

647

0.2

0.4

0.6

0.8

AC

AC700

AC800

AC900

AC1000

Fig. 2. InAuences of chemical activation temperature on the pore volume of activated carbons.

1.0

Relative Pressure[P/Po] Fig. 1. Nitrogen adsorption isotherms of activated carbons.

10 AC AC700 AC800 AC900 AC1000

5

Zeta Potential[mV]

and desorption isotherms were obtained for the samples re-activated at temperatures of 700◦ C; 800◦ C; 900◦ C and 1000◦ C with NaOH. In Fig. 1, the steep isotherms at the low relative pressure indicate that all samples except AC1000 have plenty of micropores with a narrow pore size distributions [8]. And as the re-activating temperature increases, the adsorbed volume also increases till 800◦ C. But at 900◦ C and 1000◦ C, the adsorbed volume decreases. Furthermore development of mesoporosity is indicated by the desorption hysteresis loops of samples. The e'ects of re-activation on the surface areas and pore volumes are shown in Table 1. The raw material, AC, has relatively high micropore surface area than other samples. But as the treating temperature increased, the ratio of Smi =SBET decreased gradually. In the case of AC800, microand meso surface areas as well as pore volumes were well developed, simultaneously. For the samples of AC900 and AC1000, the rapid declines of surface areas and pore volumes were observed, especially, in the micropore part. It can be thought that the pore volumes of the activated carbons increase with activation temperatures up to about 800◦ C. This indicates that the pores enlarge up to this temperature. Above 800◦ C, the excess enlargement causes the combination of pores, resulting in the increase of mesopores

0 -5 -10 -15 -20 -25 1

2

3

4

5

6

7

8

9

10

pH

Fig. 3. Zeta-potentials of activated carbons.

and the decrease of micropore volume and the surface area as well. The other reason is the graphitization of samples over 900◦ C. In Fig. 6, the evidence of graphitization can be seen in the XRD peak shown around the 2 of 26◦ . The inAuences of activation temperature on the pore volume of activated carbons are summarized in Fig. 2. This shows that the prepared activated carbons have various porosities which were intended at the start of the present study. Prior to making catalysts, the zeta-potentials of prepared activated carbons were measured to Gnd out the most

Table 1 Characteristics of activated carbons by N2 adsorption/desorption Sample

SBET (m2 =g)

Smi (m2 =g)

Sme (m2 =g)

Vmi (m3 =g)

Vtot (m3 =g)

Vme (m3 =g)

D (nm)

Vme =Vtot (%)

AC AC700 AC800 AC900 AC1000

1313 1720 2204 1121 251

1225 1495 1661 795 197

88 225 543 326 54

0.537 0.634 0.700 0.334 0.081

1.247 1.342 1.954 1.264 0.459

0.710 0.708 1.254 0.930 0.378

4.66 3.57 3.69 4.46 8.42

56 52 64 73 82

SBET : BET surface area; Smi : micropore surface area; Sme : mesopore surface area; Vmi : micropore volume; Vtot : total volume; Vme : mesopore volume; D: pore diameter.

648

G.-G. Park et al. / International Journal of Hydrogen Energy 28 (2003) 645 – 650

Table 2 Characteristics of catalysts by N2 adsorption/desorption Sample

SBET (m2 =g)

Smi (m2 =g)

Sme (m2 =g)

Vmi (m3 =g)

Vtot (m3 =g)

Vme (m3 =g)

D (nm)

Vme =Vtot (%)

Pt–Ru/AC Pt–Ru/AC700 Pt–Ru/AC800 Pt–Ru/AC900 Pt–Ru/AC1000

358 447 699 449 90

160 170 175 53 14

198 227 524 396 76

0.068 0.070 0.071 0.020 0.006

0.316 0.322 0.567 0.477 0.174

0.248 0.251 0.495 0.457 0.168

3.08 2.52 2.82 4.12 9.39

78 78 87 95 96

Micropore Mesopore

0.5

3

Pore Volume[cm /g]

0.6

0.4 0.3 0.2 0.1 0.0

AC

AC700

AC800

AC900

AC1000

Fig. 4. The inAuences of metal loading on the pore volume of catalysts.

desirable conditions of active metal adsorption [4]. Because the precursors of active metals used in this study exist in the form of anion in the slurry, it could be expected that the favorable surface states of the support material should be in the state of positive charge. In Fig. 3, it was shown that the isoelectric points of activated carbons were under the pH of 3. Referred to these data, all the catalysts were prepared under the pH of 2. Surface areas and pore volumes of the prepared catalysts are listed in Table 2.Compared to the support materials, surface areas and pore volumes of catalysts were much reduced. Peculiarly, the micropore surface areas and pore volumes were. This result shows that the active metals might block the micropores smaller than the particle size of metals. The micropore blocking was ascertained in Fig. 4. At the same time mesopore volumes were reduced offering the sitting site of active metals. In spite of these pore blockings, the average pore diameters were maintained larger than 2:5 nm indicating that there might be not so much hindrance to access for reactants to the active sites. The sizes and distributions of active metals were shown by TEM and XRD analysis in Figs. 5 and 6, respectively. With the TEM images only, it was diHcult to measure the exact particle sizes. But the di'erences in the sizes and dis-

tributions of them were clearly visualized. Combined with XRD data, the particle sizes and the inter-particle distances were calculated mathematically in Table 3 and showed a good correlation. While the diameter of Pt particles, d, had a narrow size distribution from 1.4 to 2:1 nm, the inter-particle distance, X , showed more broad values from 1.3 to 6:9 nm. These results were correlated with the performance of air breathing-type DMFC single cell to Gnd out the optimum conditions of catalysts. The reactivity of prepared catalysts was evaluated by air breathing-type DMFC single cell test. The I –V curves of DMFC are shown in Fig. 7. When compared with commercial catalyst (from E-TEK), home-made Pt–Ru catalysts supported on the activated carbons had the similar performance. Among them, Pt–Ru/AC700 catalyst showed the highest performance in power density of 18 mW=cm2 . Although Pt–Ru/AC800 and Pt– Ru/AC900 catalysts showed lower power density than other catalysts, the di'erences were as little as about 4 mW=cm2 . Furthermore, it can be seen that the catalytic performance is directly related to the inter-particle distance. Though, for instants, Pt–Ru/AC800 and Pt–Ru/AC1000 catalysts have the same average metal size of 2:08 nm, they showed different catalytic activities. That is, closer inter-Pt distanced catalysts have better catalytic activity, as is similar to the case between Pt–Ru/AC700 and Pt–Ru/AC900 catalysts. This correlation shows that the control of inter-particle distance is more important than that of particle size on the catalyst as long as the particle size is smaller than 2:1 nm. 4. Conclusion Activated carbons were successfully employed for the support material of catalyst in the DMFC. By re-activation process, activated carbons, which have various surface areas and porosities, could be obtained. In the present study, Pt–Ru/AC700 showed the best catalytic activity of 18 mW=cm2 in the air breathing-type DMFC single cell test. By combination of I –V curves and XRD analysis, it was revealed that more closely Pt dispersed catalysts have

G.-G. Park et al. / International Journal of Hydrogen Energy 28 (2003) 645 – 650

649

Fig. 5. TEM images of catalysts: (A) Pt–Ru/AC, (B) Pt–Ru/AC700, (C) Pt–Ru/AC800, and (D) Pt–Ru/AC900.

AC AC700 AC800 AC900 AC1000

A.U.

(220)

(A) (B) (C) (D) (E)

10

20

30

40

50

60

70

80

90

100



Fig. 6. XRD patterns of catalysts.

better catalytic activity. But as the inner-metal distances became closer the peak activity point occurred. The optimum catalyst condition obtained in this study is suggested as the Pt particle size of 1:68 nm with inter-particle is 2:59 nm.

Table 3 Platinum particle size and the inter-particle distance of catalyst by XRD Sample

d (nm)

X (nm)

Pt–Ru/AC Pt–Ru/AC700 Pt–Ru/AC800 Pt–Ru/AC900 Pt–Ru/AC1000

1.45 1.68 2.08 1.77 2.08

1.75 2.59 6.87 4.33 1.32

d = particle size calculated by Scherrer’s equation. X = 21:4d3 (100 − m)Sme − d, where Sme is the mesopore speciGc (m × 6000) area (m2 =g) of the catalysts and m the mass fraction of Pt in the catalytic powder [9].

Finally, for the porous support materials which have high surface areas, the inter-particle distance is a more important factor than particle size with regard to the catalytic activity.

G.-G. Park et al. / International Journal of Hydrogen Energy 28 (2003) 645 – 650 20 18

0.6

16 14

0.5 Pt-Ru/AC Pt-Ru/AC700 Pt-Ru/AC800 Pt-Ru/AC900 Pt-Ru/AC1000 E-TEK

0.4 0.3 0.2

12 10 8 6 4

0.1 0.0

2 0

20

40

60

80

100

120

140

References 2

0.7

Power Density[mW/cm ]

Cell Voltage[V]

650

0 160

2

Current density[mA/cm ]

Fig. 7. I –V characteristics and output power of air breathing-type DMFC single cell at room temperature.

Acknowledgement This work was supported by the Ministry of Science and Technology as a part of the National Research Laboratory Program and by the Ministry of Commerce, Industry and Energy.

[1] Schmidt TJ, Noeske M, Gasteiger HA, Behm RJ. J Electrochem Soc 1998;145:925. [2] Kua J, Goddard III WA. J Am Chem Soc 1999;121:10928. [3] Reddington E, Sapienza A, Gurau B, Viswanathan R, Sarangpani S, Smokin ES, Mallouk TE. Science 1998;280:1735. [4] Park GG, Jung JH, Jung DH, Shin DR, Kim CS. 2000 Fuel Cell Seminar Abstracts, October 30 –November 2, Portland, OR, 2000, p. 167. [5] Goodenough JB. J Electroanal Chem 1988;240:133. [6] Lillo-Rodenas MA, Lozano-Castello D, Cazorla-Amoros D, Linares-Solano A. Carbon 2001;39:751. [7] Hayashi J, Kazehaya A, Muroyama K, Watkinson AP. Carbon 2000;38:1873. [8] Hu Z, Srinivasan MP. Microporous Mesoporous Mater 2001;43:267. [9] Gamez A, Richard D, Gallezot P. Electrochim Acta 1995;41:307.