Optimized CeO2 content of the carbon nanofiber support of PtRu catalyst for direct methanol fuel cells

Optimized CeO2 content of the carbon nanofiber support of PtRu catalyst for direct methanol fuel cells

Journal of Power Sources 297 (2015) 400e407 Contents lists available at ScienceDirect Journal of Power Sources journal homepage: www.elsevier.com/lo...

2MB Sizes 2 Downloads 43 Views

Journal of Power Sources 297 (2015) 400e407

Contents lists available at ScienceDirect

Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour

Optimized CeO2 content of the carbon nanofiber support of PtRu catalyst for direct methanol fuel cells Hikari Kunitomo a, Hirokazu Ishitobi b, Nobuyoshi Nakagawa b, * a

Department of Environmental Engineering Science, Graduate School of Science and Technology, Gunma University, 1-5-1, Tenjincho, Kiryu, Gunma 376-8515, Japan b Division of Environmental Engineering Science, Gunma University, 1-5-1, Tenjincho, Kiryu, Gunma 376-8515, Japan

h i g h l i g h t s  The  The  The  The

catalytic activity of the PtRu/CECNF was maximized at Ce/C ¼ 0.4. increased PtRu mass activity was 2-folds of the commercial catalyst. DMFC with PtRu/CECNF generated a 2.5-folds power with half the PtRu loading. XPS spectra suggested an interaction between the PtRu and CeO2 on the CNF.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 17 March 2015 Received in revised form 10 June 2015 Accepted 2 July 2015 Available online 25 August 2015

A series of CeO2 embedded carbon nanofibers, CECNFs, with different CeO2 contents was prepared by an electrospinning technique. About 15 wt% PtRu nanoparticles were deposited on the fibers, and the effect of the CeO2 content on the methanol oxidation activity of the catalyst, PtRu/CECNF, was investigated. Cyclic voltammetry (CV), chronoamperometry (CA) and CO stripping electrochemical measurements and physical characterization along with X-ray diffraction (XRD) analysis, energy dispersive X-ray (EDX) analysis, scanning electron microscopy (SEM), transmission electron microscopy (TEM) and X-ray photoelectron spectroscopy (XPS) were carried out on the prepared catalysts. The mass activity of the PtRu was significantly increased by the CeO2 addition up to Ce/C ¼ 0.4, and the maximized activity was 2 times higher than that without CeO2. The increased activity was attributed to the strong interaction between the metal and oxide in the embedded nanofiber structure. A DMFC with the PtRu/CECNF exhibited more than 2.5 times high power density with one half the PtRu loading compared to that of the commercial catalyst, PtRu/Ccom. © 2015 Elsevier B.V. All rights reserved.

Keywords: Direct methanol fuel cell CeO2 embedded carbon nanofiber support PtRu nanoparticle catalyst Methanol oxidation reaction Strong metal support interaction Catalyst layer structure

1. Introduction A direct methanol fuel cells (DMFC) has been expected as attractive power source for mobile and portable applications due to its advantages including a high theoretical energy density and simple system structure. However, performance of the current DMFC is still far from ideal. One of the main reasons is attributed to the sluggish electrode reactions at both the anode and cathode. Many studies have already been devoted to finding an active catalyst for the electrode reactions. At the anode of the DMFC, it is known that the oxidation of CO,

* Corresponding author. E-mail address: [email protected] (N. Nakagawa). http://dx.doi.org/10.1016/j.jpowsour.2015.07.002 0378-7753/© 2015 Elsevier B.V. All rights reserved.

an intermediate of the methanol oxidation reaction (MOR) and strongly adsorbed on Pt, is the rate limiting reaction. In order to improve the reaction rate, secondary metal additions to Pt have been studied. PtRu has been widely used [1,2] as one of the most active catalysts for the MOR. The PtRu is typically an alloy with the atomic ratio of Pt:Ru is 1:1. The effect of Ru on the activity of the alloy catalyst has been expressed by the binary function mechanism [1,2] and/or the ligand effect [3]. It is usually used as nanoparticles supported on carbon, PtRu/C, or black catalyst. DMFC usually requires a relatively high loading, 3 mg cm2 or higher, of PtRu for the anode. However, the power density of the DMFC is still low compared to that of a hydrogen fuel cell. Since Pt and Ru are precious metals, it is important to reduce the loading of these elements for commercialization. Recently, there have been many reports that described the

H. Kunitomo et al. / Journal of Power Sources 297 (2015) 400e407

enhanced MOR activities of the Pt-based catalysts on the metal oxide supports [4e8] like TiO2 [6,7], CeO2 [8], V2O5 [5,9], WO3 [10], etc. Among them, CeO2 is one of the most attractive materials that have been under intense scrutiny as catalysts and as structural and electronic promoters of heterogeneous catalytic reactions [11]. The improved activity was sometimes interpreted by the bifunctional mechanism [6] and the high oxygen storage capacity of CeO2 [12]. It was also explained by the strong interaction between Pt and CeO2 [13e15]. Bera et al. considered that oxide ion vacancies lead to a strong Pt2þeCeO2 interaction and are responsible for the higher catalytic activity [15], although the interaction between Pt and CeO2 has not been ascertained. It is important to clarify the mechanism of the increased catalytic activity for the Pt-based catalyst supported on CeO2. On the other hand, the CeO2 support causes a high ohmic resistance which is a disadvantage in a fuel cell application because of its very low electrical conductivity. To overcome the problem, mixed catalysts of Pt/C with CeO2 [13,14] and the composite support of CeO2 and carbon [8,16,17] were attempted. Gu et al. have optimized the volume ratio of CeO2 in Pt/C þ CeO2 [18]. Yu et al. optimized the particle size of CeO2 in Pt/C þ CeO2 [19]. Zhang et al. investigated the optimum lattice plane of CeO2 that contacts the Pt [20]. One of the authors proposed PtRu nanoparticles supported on composite nanofibers in which CeO2 nanoparticles were embedded in the carbon nanofibers, PtRu/(CeO2eC)NF, as the anode catalyst for the DMFC [8]. It showed a significantly improved mass activity of PtRu compared to that of the commercial one. Another report has demonstrated the significantly increased mass activity of PtRu nanoparticles deposited on TiO2-embedded carbon nanofibers. The activity was strongly dependent on the TiO2 content and showed a maximum at Ti/C ¼ 1 [7]. The maximized activity was attributed to the strong interaction between the PtRu and metal oxide that was enhanced by the improved electric conductivity by the embedding structure. For the CeO2 embedded carbon nanofiber (CECNF) support, the effect of the CeO2 content in the support on the catalytic activity of PtRu has not yet been investigated. Clarifying how much the activity is improved by optimizing the CeO2 content is the purpose of the increasing mass activity of the precious metal catalyst of the DMFC. Furthermore, it is important to clarify the mechanism of the enhanced mass activity of the composite nanofiber catalyst. From this viewpoint, the surface composition rather than the bulk composition is more important related to the reaction site at the surface. Also, not only the particle sizes of the PtRu and CeO2, but also the dispersion of these particles at the surface are more important for the catalysts activity related to the circumstances of the reaction site. However, these have not yet been measured and discussed. In this study, the effect of the CeO2 content in the CECNF support on the catalytic activity of PtRu/CECNF for the MOR was investigated. According to a previous study, the composite carbon nanofibers were prepared by an electrospinning technique. The catalytic activity was evaluated by cyclic voltammetry (CV), chronoamperometry (CA) and CO-stripping voltammetry. The characterization of the prepared catalysts was done by XRD, EDX, SEM, TEM, XPS, and BET surface area measurements. By comparing both compositions obtained from the EDX and XPS, the surface composition and particle dispersion of PtRu on the nanofiber surface were discussed. The enhanced activity of the catalyst was also discussed based on the XPS results. 2. Experimental 2.1. Preparation of PtRu/CECNF catalysts CECNFs with different CeO2 contents were prepared by the

401

electrospinning technique based on a previous study [21]. The CeO2 nanoparticles (Nano Tek®, average particle size 14 nm) were dispersed in dimethylformamide (DMF, 99.5% Wako Pure Chemicals Ind., Ltd.) by ultrasonic agitation for 30 min. Polyacrylonitrile (PAN, SigmaeAldrich Co., Ltd.) was added to the mixed solution and stirred overnight at 333 K. The ratio of DMF to the sum of the PAN and CeO2 was optimized to make the nanofiber as thin as possible. The mixed solution was placed in a glass syringe attached to a stainless needle. Electrospinning was performed by applying 17 kV between the needle and a metallic collector covered with aluminum foil at a distance of 15 cm, with the solution flowing at 0.025 mL min1. The obtained felt sheet was stabilized at 513 K in air for 3 h, then carbonized at 1173 K in N2 for 1 h. About 15 wt% PtRu (at atomic ratio of 1:1) was deposited on the CECNFs by the microwave polyol method [22]. A certain amount of the nanofiber was dispersed in ethylene glycol under ultrasonic agitation for 30 min. A certain amount of H2PtCl6$6H2O (98.5% Wako Pure Chemicals Ind., Ltd.) and RuCl3$nH2O (99.9% Wako Pure Chemicals Ind., Ltd.) were dissolved in the solution by stirring for 3 h. A 100 mL portion of the mixture in a beaker was heated at 600 W for 4 min in a microwave oven, and the solution was stirred overnight. The slurry was then filtered and washed with a sufficient amount of distilled water and methanol. Finally, the catalyst powder, PtRu/CECNF, was dried at 343 K under vacuum for 24 h. 2.2. Characterization of the catalyst The chemical composition of the catalyst was analyzed by an energy dispersive X-ray spectroscope (EDX, EX-200k, Horiba, Ltd.). An X-ray diffraction (XRD) measurement of the samples was conducted to determine the crystal structure using an X-ray diffractometer (Rint2100/PC Rigaku Corp., Ltd.) with a Ni filtered Cu-Ka radiation source at 32 kV and 20 mA. The surface morphology of the prepared catalysts was studied using a field emission scanning electron microscope (FE-SEM, JSM-6330F, JEOL, Ltd.) operated at 15 kV. A transmission electron microscope (TEM, JEM 2010, JEOL, Ltd) was also used for some of the catalysts. A measurement by an X-ray photoelectron spectroscope (XPS, AXIS-NOVA, KRATOS Co., Ltd.) with Al monochromatic X-rays at 15 kV and 10 mA was also conducted to investigate the electronic states of the related elements and the elemental composition of the surface. The specific surface area of the catalysts was analyzed by the single point BrunauereEmmetteTeller (BET) method using a chemisorption analyzer (AutoChemII, Micromeritics Co., Ltd.). 2.3. Electrochemical measurements using a half-cell The catalytic activity was evaluated using a three-electrode cell with a rotating disk electrode (RDE-1, NIKKO KEISOKU Co.) of glassy carbon. The catalyst ink was prepared by dispersing 5 mg of the catalyst in a mixed solution containing 80 ml of distilled water, 80 ml of ethanol and 25 ml of a 5 wt% Nafion solution (Wako Pure Chemicals Ind., Ltd.) and agitated for 1 h. A 7 ml sample of the catalyst ink was deposited on the disk electrode (5 mm diameter) and dried at 373 K for 30 min. A Pt mesh and Ag/AgCl electrode were used as the counter and reference electrodes, respectively. The cyclic voltammetry (CV) was carried out in the aqueous solution of 0.5 M methanol and 0.5 M H2SO4 at 20 mVs1 in the potential range of 0.02e1.22 V vs. RHE. The rotation speed of the disk electrode was 1600 rpm. Before the measurement, the solution was bubbled with N2 at 200 mL min1 for 30 min. The measurement was conducted with N2 bubbling at 50 mL min1. After the CV measurement, the chronoamperometry (CA) was conducted at 0.72 V vs. RHE for 1 h in the same solution.

402

H. Kunitomo et al. / Journal of Power Sources 297 (2015) 400e407

Table 1 Properties of the prepared catalyst with the different Ce/C value measured by XRD and EDX.

PtRu/CECNF[0] PtRu/CECNF[0.1] PtRu/CECNF[0.3] PtRu/CECNF[0.4] PtRu/CECNF[0.6] PtRu/CECNF[0.8] PtRu/CECNF[1.4] PtRu/Ccom

Composition [wt.%] Pt

Ru

Ce

O

C

8.5 7.0 11.2 11.2 9.2 9.6 11.6 30.3

4.3 2.5 3.9 3.7 2.9 4.1 3.6 23.3

0.0 5.1 17.7 23.1 27.6 32.8 42.7 e

9.0 12.3 10.1 8.3 12.0 10.8 12.5 e

78.2 73.2 57.2 53.7 48.4 42.7 29.6 46.7

CO stripping voltammetry was also carried out using the catalyst to evaluate its electrochemically active surface area (ECSA). It was conducted in the following three steps: (1) CO adsorption by flowing 99.95% CO at 0.22 V vs. RHE for 30 min, (2) removal of the dissolved CO by bubbling N2 while maintaining the potential, and (3) CO stripping voltammetry, started from the constant potential between 0.02 V and 0.82 V vs. RHE at 0.01 V s1. The ECSA was calculated using the following equation.

 2 m ¼ ECSA g

Q ðCÞ  co  100: mC 420 cm2  WPtRu ðgÞ

(1)

where Qco is the measured charge for the CO stripping and WPtRu is the mass of Pt and Ru. The above measurements were carried out using an electrochemical measurement system (HZ-5000, Hokuto Denko, Co., Ltd.). The same measurements were also conducted using the commercial catalyst, PtRu/Ccom (TEC61E54, Tanaka Kikinzoku Kogyo K.K., Pt 30.0 wt%, Ru 23.3 wt%), for comparison. 2.4. MEA preparation and DMFC operation The DMFC performance of the membrane electrode assembly (MEA) with the prepared anode catalyst (PtRu/CECNF) was evaluated and compared to that with the conventional catalyst, PtRu/ Ccom. The catalysts dispersed in the mixed solution of ethanol, distilled water and 5 wt% Nafion solution were deposited on carbon paper (2.2  2.2 cm2, TGP-H-090, Toray) then dried. The carbon paper with the catalyst (0.81 mg-PtRu cm2) was hot pressed on one side of the electrolyte membrane, Nafion 117 (Du pont), with another carbon paper (Pt: 1.0 mg cm2, EC-20-10, ElectroChem, Inc.) for the cathode on the other side at 410 K and 5 MPa for 3 min. The MEA was placed in the holder with serpentine flow channels in both the anode and cathode (FC05-01SP, ElectroChem, Inc.). The currentevoltage curves were measured at the scan rate of 2 mV s1 by pumping a 2 M aqueous methanol solution at 1.5 mL min1 to the anode and oxygen at 1 L min1 to the cathode at 353 K. This measurement was also conducted using the same electrochemical measurement system. In a similar way, an MEA with the PtRu/Ccom (1.90 mg-PtRu cm2) was also prepared, measured and used for comparison. 2.5. Measurement of electric resistance of the catalyst layer The electric resistance of the nanofiber support was measured using a cylinder cell. A certain amount of the nanofiber support was homogeneously put into a column (7.0 mm in diam.) of which inner wall was made of DURACON (Polyplactics Co., Ltd.). The nanofiber layer was pressed in the axial direction by two cylinders with a flat head (7.0 mm in diam.) made of SUS304 stainless steel, and the

Pt/Ru [at./at.]

Ce/C [wt./wt.]

PtRu size [nm]

CeO2 size [nm]

1.0 1.4 1.5 1.5 1.6 1.2 1.6 1.3

0 0.070 0.309 0.430 0.570 0.768 1.44 e

2.9 3.9 5.1 5.0 5.2 6.3 6.4 3.8

e 25 27 24 27 32 25 e

electric resistance between the cylinders was measured using a 10 kHz impedance meter. 3. Results and discussion 3.1. Characterization of the catalyst Table 1 shows the catalyst composition measured by EDX for the prepared catalyst. It was confirmed that the CECNFs with different CeO2 contents, i.e., different weight ratios of Ce/C, were successfully prepared. Hereafter, the CeO2 content of the CECNF will be expressed by the weight ratio of Ce/C and noted in brackets as PtRu/ CECNF[Ce/C]. The PtRu loading was about 15 wt% and almost the same among the prepared catalysts except for PtRu/CECNF[0.1]. The atomic ratio of Pt:Ru was about 1.5:1 and almost the same except for PtRu/CECNF[0]. Fig. 1 shows the XRD profiles of the PtRu/CECNFs. The peaks of C, Pt and CeO2 were present. The broad peaks of Pt suggested a facecentered cubic structure. The Pt (111) at 2q ¼ 40.0 was slightly shifted from the peak position of pure Pt (2q ¼ 39.76 ) suggesting that the deposited Pt was alloyed with Ru. The crystallite diameters of the Pt and CeO2 were calculated based on the width of the peak at half height for the Pt (111) and CeO2 (111), respectively, using Scherrer's equation and listed in Table 1. The CeO2 crystal size was about 25 nm and independent of the CeO2 content in the support. However, the Pt crystal size tended to increase with the increasing CeO2 content. This would be due to the different surface properties of the carbon and CeO2 on the composite fiber as discussed below. Table 2 shows the composition measured by XPS for the CECNFs and PtRu/CECNFs with the different Ce/C weight ratios of 0.4 and

CeO2

Ce/C=1.4

Intensity [a. u.]

Catalyst

0.8 0.6 0.4 0.3 0.1 0 10

20

30

40

50

60

70

80

90

2 Fig. 1. XRD patterns for the prepared catalysts with the different CeO2 contents.

H. Kunitomo et al. / Journal of Power Sources 297 (2015) 400e407 Table 2 Surface compositions and bulk compositions calculated from the XPS and EDX results, respectively. Sample

Method

Composition [wt%] Pt

Ru

Ce

Ce/C [wt./wt.] O

C

CECNF[0.4] PtRu/CECNF[0.4]

XPS EDX XPS

0.0 11.2 38.4

0.0 3.7 13.4

19.9 23.1 3.4

8.2 8.3 7.3

71.9 53.7 37.6

0.28 0.43 0.09

CECNF[0.6] PtRu/CECNF[0.6]

XPS EDX XPS

0.0 9.2 40.0

0.0 2.9 9.6

28.8 27.6 6.4

9.5 12.0 6.8

61.7 48.4 37.3

0.47 0.57 0.17

0.6. It was based on the XPS spectra, some of them are shown in Fig. 7, that exhibited predominant core level peaks due to Ce, O, C, Ru and Pt. The composition was calculated by separating the peaks of Ce(3d), O(1s), C(1s), Ru(3p), and Pt(4f) using a curve fitting technique. The compositions obtained by the EDX for the catalysts were again listed in the table for comparison. The XPS analysis revealed the composition only of the surface within a several nanometer depth. Whereas, the EDX covers the whole catalyst since the nanofiber diameter, 200e300 nm, was in the range of the X-ray penetration depth. A comparison of these compositions clarified the features of the surface structure of the composite nanofiber. First, the Ce content at the surface, i.e., the XPS results for the CECNFs, was similar to that for the whole fiber based on the EDX

403

results for the PtRu/CECNFs, taking the contents of Pt and Ru for the latter into consideration. This suggested that the CeO2 particles were homogeneously dispersed in the fibers. The higher Pt and Ru contents at the surface based on the XPS results for the PtRu/ CECNFs, compared to their contents for the whole fiber based on the EDX results for the PtRu/CECNFs, confirmed that the PtRu particles had been deposited on the surface of the fibers. This was supported by the BET surface areas of 61, 73 and 50 m2 g1 for PtRu/ CECNF[0], PtRu/CECNF[0.4] and PtRu/CECNF[0.6], respectively, which are not high enough for porous fibers. It should be noted that the largely decreased Ce content at the surface compared to that for the whole fiber strongly suggested that the PtRu nanoparticles were predominantly deposited on the CeO2 part at the fiber surface. This may be related to the increased Pt crystal size with the increase in the CeO2 content shown in Table 1. It is usually difficult to deposit particles by a reduction method in a liquid solution on the surface of graphite and single and multiwalled carbon nanotubes due to their inert nature with a small number of defects and surface oxygen groups that contribute to nucleation for the crystallization, whereas, the CeO2 has many oxygen vacancies on the surface [11]. The PtRu particles would then predominantly deposit on the CeO2 part of the nanofiber surface. The concentrated PtRu deposition on CeO2 would be preferred in order to enhance the interaction, between PtRu and CeO2, which is considered as the main reason of the increased catalytic activity by the CeO2 addition in this system. Fig. 2 displays FE-SEM and TEM pictures for the PtRu/CECNFs

Fig. 2. FE-SEM (AeF) and TEM (G) pictures of the PtRu/CECNFs with the different CeO2 contents, A and D; Ce/C ¼ 0, B and E; Ce/C ¼ 0.4, C and F; Ce/C ¼ 0.8. D, E and F are the catalyst layers prepared on the glassy carbon electrode for the measurements. G is a TEM picture for Ce/C ¼ 0.4.

400

Current density [mA mg-PtRu-1]

H. Kunitomo et al. / Journal of Power Sources 297 (2015) 400e407

Current density [mA mg-PtRu-1]

404

(a)

PtRu/Ccom Ce/C=0 0.1 0.3

300 200 100 0.00 -100

0.2

0.4

0.6

0.8

1

1.2

Potential [V vs. RHE]

400

(b)

Ce/C=0.4 0.6 0.8 1.4

300 200 100 0.00 -100

0.2

0.4

0.6

0.8

1

1.2

Potential [V vs. RHE]

Fig. 3. CV curves for the PtRu/CECNF at the different Ce/C ratios measured in 0.5 M methanol and 0.5 M H2SO4 solution at the scan rate of 20 mV1, (a) Ce/C ¼ 0e0.3 and PtRu/Ccom for comparison, (b) Ce/C ¼ 0.4e1.4.

with the different CeO2 contents. The PtRu/CECNF[0], Fig. 2A, was relatively thick with long fibers, and the diameter and length of the fiber decreased with the increase in the CeO2 content as shown in Fig. 2B and C for PtRu/CECNF[0.4] and PtRu/CECNF[0.8], respectively. The average diameter was calculated based on 50 sampling points to be 300, 270 and 200 nm for PtRu/CECNF[0], PtRu/CECNF [0.4] and PtRu/CECNF[0.8], respectively. The decreased diameter at the higher CeO2 contents was attributed to the higher shrinkage of the carbon source, i.e., the polymer component, during the carbonization because of the higher DMF solvent contents in the spun solution. Fig. 2DeF were taken of the catalyst layer on the glassy carbon used for the electrochemical measurements. The catalyst loading on the electrode was 1 mg cm2 and may be relatively high compared to other reports. This loading was chosen as it shows a proportional activity to the loading for the nanofiber catalyst [8]. A decreased loading showed a lower mass activity that was not proportional to the loading. This would be due to the nanofiber structure that is different from the usual particulate catalysts. Hence, the activity evaluated in this study reflected the effect of the catalyst layer structure. In Fig. 2DeF, one can see the decreased void size in the catalyst layer with the increasing Ce/C

1400

3 2.5

1000

2

800 1.5 600 1

400

80.0

0

0.5 1 Ce/C [wt./wt.]

0 1.5

60.0 PtRu/Ccom 40.0 20.0

0

0.5

Fig. 3 shows the CV curves for the PtRu/CECNFs with the different CeO2 contents measured in a 0.5 M methanol with 0.5 M H2SO4 solution. The current density was based on the mass of PtRu showing a mass activity. The curve for the commercial catalyst, PtRu/Ccom, was also included in the figure for comparison. In the forward scan, the onset potentials for the methanol oxidation were almost the same at around 0.5 V vs. RHE for all of the catalysts. The peak potential around 0.9 V vs. RHE was almost the same among the catalysts except for the commercial one that showed the

0.5

200 0

3.2. Mass activity and stability for methanol oxidation

1

1.5

Ce/C [wt./wt.] Fig. 4. Relationship between the Ce/C ratio and mass activity; current density at 0.68 V vs. RHE of the CV curve shown in Fig. 3. The inset is relationship between the Ce/C and the electric resistance of the catalyst layer (40 mg-cat., column with 7.0 in diameter, pressed at 2 MPa and 4 MPa). Thickness of the catalyst layer was also shown.

Current density [mA mg-PtRu-1]

Resistance [ ]

100

[mA mg-PtRu-1]

Current density at 0.7 V vs. RHE

120

2 MPa 4 MPa

Thickness [mm]

40 mg-cat. 7.0 mm

1200

due to the decreased length of the nanofibers. This means that the catalyst layer porosity decreased with the increasing CeO2 content. At the relatively low CeO2 contents, a bulky catalyst layer was obtained due to the relatively long fibers. On the other hand, at the high CeO2 contents, a dense catalyst layer was formed by the short catalysts. It should be noted that the catalytic activity in this study includes the effect of the catalyst layer structure. This was similar to the case of the TiO2 embedded carbon nanofibers described in a previous study [7]. Fig. 2G shows a TEM picture for PtRu/CECNF[0.4]. PtRu nanoparticles around 3 nm in diameter were observed on the nanofiber. Some of them formed agglomerates here and there as shown in the picture. However, relating to the discussion of the surface composition in Table 2, it was not clear whether the agglomerates were predominantly deposited on CeO2 particles or not.

200 0.68 V vs. RHE 150

Ce/C=0 0.1 0.4 0.6 0.8 1.4

100

50.0

0.00

0

900

1800

2700

3600

Time [s] Fig. 5. CA curves for the PtRu/CECNFs with the different CeO2 content measured in 0.5 M methanol and 0.5 M H2SO4 at the constant potential of 0.68 V vs. RHE.

30.0

(a)

Ce/C=0 0.1 0.4

20.0 10.0 0.00 -10.0 -20.0 -30.0 -40.0

0

Current density [mA mg-PtRu-1]

Current density [mA mg-PtRu-1]

H. Kunitomo et al. / Journal of Power Sources 297 (2015) 400e407

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

30.0

(b)

Ce/C=0.6 0.8 1.4

20.0 10.0 0.00 -10.0 -20.0 -30.0 -40.0

0

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Potential [V vs. RHE]

Current density [mA mg-PtRu-1]

405

Potential [V vs. RHE]

30.0

(c)

20.0

PtRu/Ccom

10.0 0.00 -10.0 -20.0 -30.0 -40.0 0

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Potential [V vs. RHE] Fig. 6. CO stripping voltammograms for the different catalysts.

peak at a 0.1 V higher potential. The current densities in the range over the onset potential strongly depended on the CeO2 content showing the improved current densities compared to that without CeO2 addition, i.e., PtRu/CECNF[0] and PtRu/Ccom. The mass activity at 0.7 V vs. RHE, that is the typical anode potential of the DMFC, was plotted versus the Ce/C ratio for the PtRu/CECNFs with the different Ce/C ratios in Fig. 4. The dashed line shows the level of the PtRu/ Ccom. The mass activity increased with the increase in the Ce/C ratio up to 0.4 and then decreased. The maximum activity at Ce/C ¼ 0.4 was more than double the activities of the catalyst without CeO2, PtRu/CECNF[0] and PtRu/Ccom. The improved activity would be attributed to the strong interaction between PtRu and CeO2. The decreased activity in the Ce/C range over 0.4 can be explained by the reduced electron conductivity of the catalysts. The inset of Fig. 4 shows the electric resistance of the CECNF layers with the different ratio of Ce/C. The electric resistance of the layer drastically increased with the increase of the Ce/C in the rage over 0.4. Also, the dense structure of the catalyst layer containing the shorter fiber structure, as shown in Fig. 2F, would be related to the decrease in the activity. Fig. 5 shows the CA curves for the PtRu/CECNFs with the different Ce/C ratios measured at the constant potential of 0.68 V vs. RHE. The mass activity decreased with time in all cases. However, the activity for PtRu/CECNF[0.4] was the best during the 1 h measurement. As shown above, the mass activity of PtRu/CECNF was significantly improved by the CeO2 addition, and the CeO2 content was optimized at Ce/C ¼ 0.4. The effect of CeO2 will be discussed in the next section. Fig. 6 shows the CO stripping voltammograms for the different catalysts. The onset potential and peak potential for the CO oxidation and ECSA for the catalysts are listed in Table 3. As shown in the

Table 3 The results of CO stripping with the prepared catalysts. Catalyst

Onset potential [eV]

Peak potential [eV]

ECSA [m2 g1 -PtRu]

PtRu/CECNF[0] PtRu/CECNF[0.1] PtRu/CECNF[0.4] PtRu/CECNF[0.6] PtRu/CECNF[0.8] PtRu/CECNF[1.4] PtRu/Ccom

0.50 0.45 0.43 0.46 0.46 0.46 0.39

0.59 0.55 0.58 0.57 0.55 0.56 0.52

25.7 66.0 64.2 41.6 69.4 19.4 45.0

table, ECSA did not show a clear correlation with the activity shown in Fig. 4, although the ECSA was improved by the addition of CeO2 from 26 to around 70 m2 g1 -PtRu. On the other hand, the onset potential was in accordance with the mass activity showing the lowest potential, 0.43 V vs. RHE, for the PtRu/CECNF[0.4], although it was 0.04 V higher than that of the commercial catalyst, PtRu/Ccom. This suggested that the increased activity by the CeO2 addition was not mainly related to ECSA, as a quantity of the reaction site, but to the onset potential as a quality of the reaction site. The difference of the onset potential between the PtRu/CECNFs and PtRu/Ccom would be due to the different support materials and the different PtRu loadings as shown in Table 1. The reduced onset potentials for the PtRu/CECNFs would be attributed to the interaction between the metal particles and the support. This is supported by the following XPS analysis. The property of the catalyst surface is sometimes discussed based on the XPS spectra. Fig. 7 shows the XPS spectra for the Pt(4f) and Ce(3d) core level regions of the PtRu/CECNF with different CeO2 contents. The two major peaks were isolated into the four Pt(4f) and two Ru(4s) components by a curve fitting as shown in Fig. 7a and b. The Ce(3d) core level region was isolated into the twelve

406

H. Kunitomo et al. / Journal of Power Sources 297 (2015) 400e407

Ce/C=0.8

Intensity

Ce/C=0.4

Intensity

(b)

(a)

Pt 4f

Ce/C=0.6

Ce/C=0

84

82

80

78

76

74

72

70

84

68

82

80

(c)

70

68

Intensity

Intensity

Intensity

72

(e)

Ce/C=0.8

Ce/C=0.6

910 900 890 Binding energy [eV]

74

(d)

Ce/C=0.4 920

76

Binding energy [eV]

Binding energy [eV] Ce 3d

78

880

920

910 900 890 Binding energy [eV]

880

920

910 900 890 Binding energy [eV]

880

Fig. 7. XPS spectra for Pt(4f), a and b, and Ce(3d), c, e and f, core level regions of the PtRu/CECNF with the different CeO2 contents. Dot; measured, line; calculated (Assigned binding energy [eV]; 74.3 and 71.1 for Pt(4f)0, 75.1 and 71.8 for Pt(4f)2þ, 882.0, 885.9, 900.4 and 904.3 for Ce(3d)3þ, 882.5, 885.2, 889.1, 898.4, 900.9, 903.7, 907.6 and 916.7 for Ce(3d)4þ.).

components according to a previous report [23] as shown in Fig. 7cee. Table 4 shows the existence percentage of the Pt metal and Pt2þ and that of Ce3þ and Ce4þ based on the isolated peak areas. It is noted that percentage of Pt metal increased by the addition of CeO2 into the nanofibers. The higher presence of Pt metal reflected the higher utilization of the Pt metal. It has been considered that the electron density of Pt increased by the electronic interaction between CeO2 and Pt, i.e., an electron donation from C of the adsorbed CO to Pt decreased. This would result in the easy removal of CO from the Pt surface. The increased percentage of Pt metal may be caused by the effect of the interaction. It is also noted that the percentage of Ce3þ is quite high for PtRu/CECNF[0.4], that showed the highest activity, and the percentage decreased with the increase of CeO2 content. The increased percentage of Ce3þ shows a more reduced surface of CeO2 suggesting a stronger interaction between the Pt and CeO2. These XPS analyses suggested an electrical interaction between Pt and CeO2 for improvement of the catalytic activity of the catalyst with CeO2, although the measured data did not clearly explain the difference in the CeO2 contents. This may be due to the different

structure of the catalyst and the layer as shown in Fig. 2. At the higher CeO2 contents, the short fibers with the low carbon content reduce the electric conductivity of the catalyst layer and the dense catalyst layer would cause a high mass transport resistance of the catalyst layer. Noted again, these effects of the different structures of the catalyst and its layer would be included in the results of the mass activity of the PtRu/CECNFs with the different CeO2 contents. 3.3. DMFC performance Fig. 8 shows the comparison of the DMFC performances between the MEA with the PtRu/CECNF[0.3] and that with the commercial one, PtRu/Ccom. The PtRu loading was 0.81 and

Table 4 The results of isolated Pt(4f) and Ce(3d) spectra for PtRu/CECNF with different Ce/C. Element

Existence percentage [%]

Valency

0







PtRu/CECNF[0] PtRu/CECNF[0.4] PtRu/CECNF[0.6] PtRu/CECNF[0.8]

42.6 44.5 50.9 46.1

57.4 55.5 49.1 53.9

e 0.56 0.18 0.10

e 0.44 0.82 0.90

Pt

Ce

Fig. 8. Power generation characteristics of the DMFC with PtRu/CECNF[0.3] and that with PtRu/Ccom.

H. Kunitomo et al. / Journal of Power Sources 297 (2015) 400e407

1.90 mg cm2 for the PtRu/CECNF and PtRu/Ccom, respectively. In spite of the half loading of PtRu, the DMFC with the PtRu/CECNF delivered 71 mW cm2 that was about 2.5 times higher than that with PtRu/Ccom at 28 mW cm2. The power density of the DMFC with the commercial one was reasonably compared to that with the different catalyst loadings measured for the same system under the same conditions [7]. The significantly high improvement of the PtRu utilization in the DMFC with PtRu/CECNF should be noted. The improved power density can be expressed by the improved mass activity of the catalyst by the interaction between Pt and CeO2 and also of the bulky nanofiber catalyst layer that would be preferred for the mass transport of the related compounds. The effect of the catalyst layer structure of the nanofiber catalyst depending on the aspect ratio of the catalyst on the catalyst layer activity, i.e., potential loss, should be investigated. This will be one of our future subjects. Such a significant improvement in the mass activity of PtRu and Pt has been also reported for the TiO2 embedded nanofiber support [7]. This report revealed that CeO2 was another effective oxide material for enhancing the mass activity of the metals. We strongly suggest that the embedding architecture of such metal oxide particles in carbon nanofibers can effectively improve the mass activity of the PtRu nanoparticle catalysts and then the DMFC performance. 4. Conclusion PtRu/CECNFs with different CeO2 contents were prepared by the electrospinning technique and the microwave polyol method. The effect of the CeO2 content on the mass activity of the PtRu support was investigated both in a half cell and in an actual DMFC. The addition of CeO2 significantly improved the activity up to the Ce/C of 0.4, then the activity decreased. The decrease of the activity in the range Ce/C > 0.4 was explained by the increase of the electric resistance of the catalyst layer. The maximum mass activity with PtRu/CECNF[0.4] was about 2 times higher than that of PtRu/Ccom. The improved activity of PtRu by the CeO2 addition was due to the interaction between Pt and CeO2 based on the CO stripping data and the XPS analysis. The relatively bulky structure of the catalyst layer of the PtRu/CECNF with the lower CeO2 contents of less than 0.4 would also be effective to improve the activity. The DMFC with

407

PtRu/CECNF generated a 2.5 times higher power density compared to that with PtRu/Ccom in spite of about a half PtRu loading. This demonstrated the reduction of the PtRu loading on the DMFC with the PtRu/CECNF. PtRu/CECNF is a promising anode catalyst for DMFCs. Acknowledgments Part of this study was supported by the Element Innovation Project, Ministry of Education, Japan, and by JSPS KAKENHI (26289300). References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23]

M. Watanabe, S. Motoo, J. Electroanal. Chem. 60 (1975) 267e273. M. Watanabe, S. Motoo, J. Electroanal. Chem. 60 (1975) 275e283. H. Igarashi, et al., Phys. Chem. Chem. Phys. 3 (2001) 306e314. H.L. Pang, X.H. Zang, X.X. Zhong, B. Liu, X.G. Wei, Y.F. Kuang, J.H. Chen, J. Colloid Interface Sci. 319 (2008) 193e198. S. Jayaraman, T.F. Jaramillo, S.-H. Baeck, E.W. McFarland, J. Phys. Chem. B 109 (2005) 22958e22966. P. Kolla, A. Smirnova, Int. J. Hydrogen Energy 38 (2013) 15152e15159. Y. Ito, et al., J. Power Sources 242 (2013) 280e288. C. Feng, M.A. Abdelikareem, T. Tsujiguchi, N. Nakagawa, J. Power Sources 242 (2013) 57e64. P. Justin, G.R. Rao, Catal. Today 141 (2009) 138e143. B. Rajesh, et al., Fuel 81 (2002) 2177e2190. A. Trovarelli, Catalysis reviews, Sci. Eng. 38 (1996) 439e520. J.W. Guo, T.S. Zhao, J. Prabhuram, R. Chen, C.W. Wong, J. Power Sources 156 (2006) 345e354. D.R. Ou, T. Mori, H. Togasaki, M. Takahashi, F. Ye, J. Drennan, Langmuir 27 (2011) 3859. A. Bruix, et al., J. Am. Chem. Soc. 134 (2012) 8968e8974. P. Bera, et al., J. Phys. Chem. B 107 (2003) 6122e6130. M.A. Schibioh, S.K. Kim, E.A. Cho, T.H. Lim, S.A. Hong, H.Y. Ha, Appl. Catal. B Environ. 84 (2008) 773e782. M. Wu, M. Han, M. Li, Y. Li, J. Zeng, S. Liao, Electrochim. Acta 139 (2014) 308e314. D.M. Gu, Y.Y. Chu, Z.B. Wang, Z.Z. Jiang, G.P. Yin, Y. Liu, Appl. Catal. B Environ. 102 (2011) 9e18. L. Yu, J. Xi, Int. J. Hydrogen Energy 37 (2012) 15938e15947. D. Zhang, C. Zhang, Y. Chen, Q. Wang, L. Bian, J. Miao, Electrochim. Acta 139 (2014) 42e47. M.A. Abdelkareem, Y. Ito, T. Tsujiguchi, N. Nakagawa, ECS Trans. 50 (2012) 1959e1967. Y. Cheng, S.P. Jiang, Electrochim. Acta 99 (2013) 124e132. M.N. Revoy, R.W.J. Scott, A.P. Grosvenor, J. Phys. Chem. C 117 (2013) 10095e10105.