High power direct methanol fuel cell with a porous carbon nanofiber anode layer

Applied Energy 113 (2014) 946–954

Contents lists available at ScienceDirect

Applied Energy journal homepage: www.elsevier.com/locate/apenergy

High power direct methanol fuel cell with a porous carbon nanofiber anode layer A.M. Zainoodin a, S.K. Kamarudin a,b,⇑, M.S. Masdar b, W.R.W. Daud a,b, A.B. Mohamad b, J. Sahari a a b

Fuel Cell Institute, Universiti Kebangsaan Malaysia, 43600 UKM Bangi, Selangor, Malaysia Department of Chemical and Process Engineering, Faculty of Engineering and Built Environment, Universiti Kebangsaan Malaysia, 43600 UKM Bangi, Selangor, Malaysia

h i g h l i g h t s  This study demonstrates a novel porous carbon nanofiber anode (PNCF) layer.  PNFC anode layer DMFC presents power density of 23.0 mW cm

2

.

 This unit operates at room temperature and consumes low concentration of methanol.

a r t i c l e

i n f o

Article history: Received 16 May 2013 Received in revised form 29 July 2013 Accepted 31 July 2013

Keywords: Electrode Direct methanol fuel cell Carbon nanofiber

a b s t r a c t Three anode electrodes containing Pt–Ru Black as a catalyst were fabricated with a porous layer made with different carbon materials: carbon black (CB), carbon nanofiber (CNF) and a combination of both carbon materials (CB + CNF). The carbon-based porous layer was coated onto a carbon cloth with PTFE pre-treatment for delivering hydrophobic properties and applied in direct methanol fuel cells (DMFCs). Characterisation of electrochemical properties for three different anode electrodes was performed with cyclic voltammetry (CV), chronoamperometry (CA) and electrochemical impedance spectroscopy (EIS) at room temperature in a half-cell configuration. The evolution of the surface morphology of diffusion layer and electrodes was characterised by using variable-pressure scanning electron microscopy (VP-SEM). The electrochemical results indicate that electrode with CNF layer showed the highest current densities compared to CB and CB + CNF with the same catalyst loading. VP-SEM measurements show the network formation within the structure, which could facilitate the methanol mass transfer and improve the catalyst efficiency. The electrodes were applied to a single-cell DMFC, and the cell performance was experimentally investigated under passive operating mode and room temperature. A maximum power density of 23.0 mW cm 2 at a current density of 88.0 mA cm 2 with a 3 M dilute methanol solution was achieved. The results show that the electrodes with a CNF layer could improve the performance of DMFC as compared with commercially used CB and prove it’s potentially application in DMFC technology especially for portable power source applications due to several advantages as followings: operating at low concentration of methanol, operating at room temperature, low catalyst loading in anode and cathode, cheaper, less hazardous and no parasitic load. Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction The direct methanol fuel cell (DMFC) is a power-generating system that converts the energy of a chemical liquid fuel directly into electrical energy without auxiliary devices, making it a promising power source for many mobile and portable applications [1–4]. In the DMFC process, the methanol fuel diffuses to the anode gas diffusion layer and reaches the catalyst layer, where the methanol oxidation reaction takes place [5]. The generated protons diffuse ⇑ Corresponding author at: Fuel Cell Institute, Universiti Kebangsaan Malaysia, 43600 UKM Bangi, Selangor, Malaysia. Tel.: +60 389216422; fax: +60 389216148. E-mail address: [email protected] (S.K. Kamarudin). 0306-2619/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.apenergy.2013.07.066

through the polymer electrolyte membrane (PEM) to the cathode catalyst layer and combine with the electrons flowing from external circuit and the oxygen fed to generate water molecules. However, the performance of DMFCs is always restricted by the methanol crossover through the PEM [6] and the sluggish electrochemical reaction at catalyst sites, which results in serious catalyst poisoning and a reduction in the cell efficiency [7]. To overcome these problems and to reduce the catalyst costs, one of the effective approaches used is modification of the diffusion layer in an attempt to increase the mass transfer and catalytic activity. In general, a diffusion layer consist of carbon cloth or carbon paper substrate as backing layer and a layer of carbon-based

947

A.M. Zainoodin et al. / Applied Energy 113 (2014) 946–954

porous layer which provides electronic conductivity to the electrode and controlling the methanol transport through the anode. The CNF is placed between carbon cloth backing layer and the catalyst layer, thus a uniform flat porous layer may reduce the contact resistance between backing layer and the catalyst layer. A good interaction between the catalyst and diffusion layer not only improve the catalyst efficiency but also increase the charge transfer as well as the two-phase transport of methanol and oxygen. In addition, the high surface area of carbon-based porous layer also provide physical support for catalyst where the catalyst always dispersed in the form of nano-particles on the surface of porous layer [8]. It is shown that by properly designing the anode diffusion layer, the catalyst utilization is improved and the methanol crossover is reduced, which in turn enhances the overall cell performance. Basically, the anode porous layer should have a high electronic conductivity, a high surface area for better catalyst dispersion and utilisation, good reactant diffusion properties and good stability under DMFC operating conditions. Carbon materials, such as Vulcan XC-72R carbon black, are preferred materials for porous layer preparation in commercial DMFC application because of their high electronic conductivity, high surface area, easy availability, mechanical and chemical stability and low cost. However, the presence of deep micro-pores in carbon black porous layer which trap the catalyst nano-particles and making them unapproachable to reactants [9,10]. The smaller surface area of carbon black compared with other carbon nano-structure materials (e.g. CNT, CNF) also provides fewer accessible sites for catalyst deposition thus leading to reduced catalyst dispersion and catalyst activity in methanol oxidation. At present, there are many types of carbon nanostructure materials being studied as the catalyst support or applied as diffusion layer that have a higher specific surface area compared with the conventional carbon black [11,12]. These materials include single-walled carbon nano-tubes (SWCNTs), multi-walled carbon nano-tubes (MWCNTs), carbon nanofibers (CNFs), fullerenes and carbon nano-horns. Santasalo-Aarnio et al. found that by using Pt–Ru/graphitised CNFs, better stability was achieved, although the cell performance was poor compared with Pt–Ru/Vulcan [13]. In addition, the application of non-carbon materials as catalyst supports has also shown some improvement in the dispersion of catalyst particles in DMFC [14–16]. Gharibi et al. [17] use a thick layer of polyaniline-supported Pt with a low activation energy to improve the catalyst dispersion and utilisation. Among all these carbon nanostructure materials, carbon nanofibers offer several additional beneficial properties, such as a high electrical conductivity, high oxidation resistance, high purity, high thermal conductivity and good permeability to reactants. The meso-pore geometry of CNF porous layer can enhance the reactant mass transport and its high surface area advantageous for catalyst deposition, which is important for DMFC electrochemical reactions. Besides that, several researches like, Abdelkareem et al. [18,19] had presented porous carbon plate (PCP) to improve the performance of DMFC and obtained a power density of 24–30 mW cm 2. Later, Park et al. [20] presented multi-layer electrodes for DMFC and obtained a power density of 10.7 mW cm 2. While Yuan et al. [21] had invented integrated porous methanol barriers (PMFSF) and obtained a 5.2 mW cm 2 of power density. Apparently, it is observed that the PCP [18,19] gives the highest power density for DMFC by reducing the methanol crossover. However, a proper storage system is essential due to the toxicity of methanol to human being. There are also several other researcher adopted the vapor feed methanol to improve the performance of DMFC, however extra vaporizer component is needed to vaporize the liquid methanol [22]. This will affect the water needed in anode side for reaction that will lead to water management problem that would affect the power generation [23,24].

Due to that, this study will focus on the improvement of the performance of passive DMFC by focusing on low concentration of methanol operating at room temperature. In this paper, we report a simple technique to prepare the anode porous layer with different carbon structure materials (CB, CNF, CB + CNF) with the aim to compare the effects of different anode porous layer on the performance of DMFC. The carbon porous layer was applied on a sheet of macro-porous carbon cloth, which contains hydrophobic and hydrophilic properties. The electrodes performance was studied in a half cell and single cell DMFC. The morphologies of the diffusion layer were examined with scanning electron microscopy (SEM). The electrochemical characteristics of electrodes were investigated by cyclic voltammetry (CV), chronoamperometry (CA) and electrochemical impedance spectroscopy (EIS) measurements. MEA preparation with different carbon structure porous layer with similar catalyst (Pt–Ru Black) loading was tested in passive air-breathing DMFC single cell. Furthermore, various methanol concentrations (2 M, 3 M, 4 M, and 5 M) were used to investigate the effect of methanol concentration on the performance of CNF layer in DMFC. Finally, the experimental results of our study for single cell proves that the CNF layers gives a significant effect on the DMFC performance at low concentration of methanol operating at room condition with other several advantages.

2. Experimental 2.1. MEA preparation The anode porous layers were prepared by the simple precipitation method on a 5 wt% polytetrafluoroethylene (PTFE) pre-treatment carbon cloth (E-TEK, USA). The CNF (Platelet GNF) used in this study was supplied by Catalytic Materials, and CB (Vulcan XC-72R) was supplied by Cabot Corporation with the detail specification is stated in Table 1. NafionÒ Solution 5 wt% (Wako Pure Chemical Industries, Ltd.) was used as an ionomer. Pt–Ru Black (HiSPEC 6000, Alfa Easer, USA) acted as a catalyst for anode, while Pt Black (HiSPEC 1000, Alfa Easer, USA) for cathode and isopropyl alcohol (IPA) as a solvent. Preparation method for electrode is a determination step in a MEA and need to be optimized. The porous layer should be uniformly distributed on the carbon backing layer without blocking the macro-pores of carbon cloth for reactant transport. For the preparation of anode porous layer, firstly the carbon-based materials (CB or CNF or CB + CNF) is mixed with IPA and ionomer solution, secondly the carbon ink is deposited onto one side of carbon cloth pre-treated with PTFE suspension. Then the diffusion layer is heat-treated and results a smooth surface of porous layer (2 mg cm 2) on carbon cloth. A cathode diffusion layer was similarly prepared with the PTFE coated carbon cloth and the carbon black porous layer. The catalyst ink was prepared by the dispersion mixing of Pt–Ru Black with a 5.0 wt% Nafion solution, deionised water and isopropyl alcohol in an ultrasonic bath. After casting, the electrodes were Table 1 Properties of porous layer based material. Based Product material name

Specific surface area (m2 g 1)

Weight (%)

Electrical conductivity (S cm 1)

Particle size (nm)

Porosity (nm)

Carbon black Carbon nano fiber

254

>99

4

30

<2

10–300

>99.9

102–104

80

2–50

Vulcan XC-72R Platelet GNF

948

A.M. Zainoodin et al. / Applied Energy 113 (2014) 946–954

dried in an oven for the appropriate period of time. Loading of catalyst on both anode and cathode is 8 mg cm 2. The electrodes were bonded with Nafion 117 membrane by hot-pressed at 135 °C with 12.5 kgf cm 2 for 180 s. Subsequently, the MEA were sandwiched between two polycarbonate end plates with a stainless steel mesh as a current collector for DMFC performance testing. The detailed procedure for MEA fabrication is schematically represented in Fig. 1. In this study, we focus on the effect of structure of anode porous layer on DMFC performance, thereby the other design parameter including PTFE loading in backing layer, the carbon loading in porous layer, the ionomer content, catalyst loading, the membrane used, casting and the hot press procedures are kept constant during the experiment. All the parameters were optimized by one-factor-at-a-time technique based on preparation process reported by Okada et al. [25] and Ahmad et al. [4].

2.2. Structural and electrochemical characterisation Scanning electron microscopy (SEM) images for surface morphology of the diffusion layers and catalyst layers were obtained using a VP-SEM EVO (ZEISS) microscopy. Electrochemical measurements were performed using a conventional three-electrode electrochemical cell that was connected to a potentiostat–galvanostat (Autolab PGSTAT) at room temperature. A KCI-Saturated Ag/AgCI electrode was used as the reference, and a platinum rod electrode was used as the counter electrode while the sample (anode electrodes with different carbon porous layers but similar loading of catalyst) was employed as the working electrode. Cyclic voltammetry (CV) was performed in a N2-saturated 0.5 M H2SO4 electrolyte to ensure that the electrode preparation had succeeded and no significant an ion adsorption from the electrolyte to the catalyst surface occurred. The N2-saturated electrolyte was prepared by purging purified nitrogen gas into the solution for 30 min in order to remove oxygen. Then, the catalytic activities of electrode toward the methanol oxidation reaction were measured in N2-saturated 0.5 M H2SO4 + 1 M CH3OH solution between 0.08 and 0.84 V with a scan rate of 10 mV s 1. For further study of the catalytic poisoning, chronoamperometry (CA) measurements were obtained for 60 min at 0.4 V vs. RHE potential. Also the activities of the electrodes varying porous layers were studied by halfcell EIS measurements, which were carried out in a potential mode

at a 0.4 V potential over a frequency range of 100 kHz–0.01 Hz with an oscillating potential amplitude voltage of 10 mV. 2.3. Fuel cell performance The cell performance was evaluated in a passive single cell with an active area of 4 cm2. Air was supplied directly through the opening area of the cathode end plate while a methanol tank was built in the anode fixture. The passive DMFC was fuelled with 4 mL of dilute methanol and the testing was conducted at room temperature. The polarisation curves of the passive DMFC were obtained with the Potentiostat/Galvanostat (WonATech) instrument and the test ended when the methanol in reservoir was exhausted. Prior to the performance testing, the MEA were conditioned and activated overnight with 2 M methanol solution. The MEA with optimum anode porous layer is further tested with different methanol concentration to study the effect of methanol concentration on DMFC performance. The performance curve shown in our results represents the average of three repeated MEAs respectively, that the simple manual casting method shows the good reproducibility of the electrode manufacturing process. 3. Results and discussion 3.1. Structural characterisation Fig. 2 displays a typical micrograph of a carbon cloth; a properly arranged fiber structure causes the pores to be better organised on average. The carbon cloth has been applied as backing layer and mechanical support for the electrode due to its high porosity for methanol transport and carbon dioxide removal and also well performed in terms of durability. In this study, the carbon cloth will undergo a PTFE pre-treatment process to make the structure hydrophobic, where it is necessary to prevent water flooding and facilitate mass transport at the electrode. However, the PTFE loading of backing layer should be as low as possible to maintain an acceptable water-crossover rate without blocking the reactant transport. Increasing the PTFE loading in anode backing layer not only increase the mass transfer resistance of water but also increases the mass transfer resistance of methanol from reservoir to anode catalyst layer. This will cause an inadequate methanol concentration at anode catalyst layer, and the cell performance

Fig. 1. Schematic procedure for MEA fabrication.

A.M. Zainoodin et al. / Applied Energy 113 (2014) 946–954

949

Fig. 2. VP-SEM morphology of carbon cloth (a) untreated at 100 mag, (b) untreated at 400 mag, (c) treated with PTFE 5 wt% at 100 mag, (d) treated with PTFE 5 wt% at 400 mag, and (e) EDS of carbon cloth treated with PTFE 5 wt%.

decreases especially at high current density [26]. Fig. 2(a) and (b) show images of the untreated carbon cloth, and Fig. 2(c) and (d) show the carbon cloth that is already treated with PTFE solution. It can be seen that PTFE is coated evenly on the carbon cloth. Fig. 2(e) shown the EDS result of carbon cloth treated. Increasing PTFE loading will leads to the higher hydrophobicity, but at the same time it also will lower the electrical conductivity. If the content of PTFE is to low, insufficient water removal capability will occurs. After long-term operation it may also lead to poor water management which will degrades the MEA rapidly. Therefore, it shows that diffusion media (DM) plays an important role in the durability of MEA. In literatures, generally the suggested PTFE content were up to 45 wt% [27–29]. However, this study proved that 5 wt% PTFE loading was sufficient to obtain a hydrophobic surface to facilitate CO2 removal at anode and alleviate water flooding at cathode. Higher PTFE content could block the diffusion layer surface pores and this may significantly reduce the permeability of the carbon cloth. As a result, the effective diffusivity of liquid phase will be reduced as the gas flow path increased. However below 5% gave higher electrical conductivity but it is not hydrophobic and this will lead to problem in water management.

Therefore, an adequate PTFE loading for anode backing layer is essential in order to facilitate the gas transport and enhance the mass transport of methanol solution. While for cathode backing layer, wet-proofing of carbon cloth is important for revise diffusion of product water from cathode to anode side in order to provide a sufficient proton conductivity of membrane. Fig. 3 presents the in-plane electrical conductivity of the PTFEtreated carbon cloth as a function of the PTFE content. One of the major functions of is to transport electrons from catalyst layer to current collector; this implies that DM must have a good electrical conductivity. It is expected that the conductivity would decrease with an increase in the PTFE loading as the PTFE materials tends to cluster in the gaps between carbon fibers. The carbon cloth with 5 wt% PTFE shown a slightly drop in conductivity is most likely due to the electrically conductive carbon fibers of carbon cloth which remain structurally unchanged after treated with PTFE. SEM images of porous layer raw material have been presented in Fig. 4. It shows a significantly different morphology between CB, Vulcan XC-72R represented in Fig. 4(a) and CNF, Platelet GNF represented in Fig 4(b). The Vulcan carbon is composed of nanosized carbon particles extensively agglomerated with micro-pore

950

A.M. Zainoodin et al. / Applied Energy 113 (2014) 946–954

Fig. 3. In-plane electrical conductivity of carbon cloth backing layer with different PTFE loading.

while CNF is composed of nanosized carbon fiber dispersed more freely with meso-pore. Structures consisting of graphene sheets oriented perpendicular to the fiber growth axis in a configuration identical to that of a stack of cards. Fig. 5 presents the SEM images of the anode diffusion layer made with different porous layers (CB, CB + CNF and CNF). From the images, the differences in microstructures, the pore distribution and pore size within the three different porous layers can be seen. Fig. 5(a) and (b) represent CB, (c) and (d) represent the combination of CB + CNF, and (e) and (f) represent the CNF porous diffusion layer. More cracks can be seen in CB (a) compared with CB + CNF (c) and CNF (e), and the cracks appear to be denser, which induced by volume shrinkage of carbon/Nafion slurry during annealing. Such cracks are undesirable in the anode porous structure since they permit high fuel transport to reach the surface of the membrane. The almost crack-free anode porous layer prepared by CNF could significantly decrease the rate of water and methanol crossover and thus better cell performance is expected. Also, the CNF porous layer uniformly and fully covers the backing layer structure which might reduce the contact resistance with catalyst layer. Conversely, for CB and CB + CNF porous layer, the structure of the carbon cloth can still be seen. In Fig. 5(b), an accumulation of the carbon ball can be seen as compared with Fig. 5(f); where the fibers attempt to form a uniform, continuous network with more meso-pores. The pore size is large enough for the continuous supply of reactants and easier to coat the catalyst layer because of the roughnesses surface of CNF each direction. The porous layer

with combination of both carbon materials (CB + CNF) can be seen in Fig. 5(d); a network with extra pores is formed compared to CB porous layer, which might be helpful for methanol transport and electron conductivity. Yuan et al. [30] noted that the network will help disperse the catalyst ink well in contact with the support layer, thus lowering the contact resistance. The catalyst layer on different porous layers has also been examined by the SEM images as presented in Fig. 6. A similar catalyst loading is spread over the surface of different porous layer. The CNF porous layer is totally covered by the catalyst, where a dense catalyst layer is formed onto the diffusion layer as shown in Fig. 6(c). This is due to the high surface area of fibers structures for catalyst deposition. On the contrary, CB and CB + CNF porous layer in Fig. 6(a) and (b) were appear like permeable to catalyst where the catalyst is entrapped inside the pores and thus the structure of diffusion layer still can be seen. By comparing these three catalyst layer supported with different porous layer, it shows that thicker catalyst layers is formed on the CNF porous layer compared to the CB + CNF and CB porous layer with the same catalyst loading. Thicker catalyst layer generally would induce higher surface tension, thus causing more cracks. It is also stated that more catalyst near the porous layer/catalyst layer interfaces; probably have higher catalyst utilization and higher performance [31]. The thicker catalyst layer on CNF porous layer could increase the amount of reaction sites, and thus increased the reaction rate in anode. With the same catalyst loading, thinner catalyst layer formed on CB and CB + CNF as the catalyst deeply fall inside the porous layer, which mean part of the catalyst may not be utilized due to unachievable from methanol and no connection with electrolyte, leading to a low utilization of catalyst. However, it is important also to reduce the cracking in the catalyst layer in order to maximize the interfacial contact between electrode and membrane. 3.2. Electrochemical characterisation Fig. 7 shows the activity of Pt–Ru catalyst on different anode porous diffusion layers towards methanol oxidation. The CV indicates that the peak current value of CNF (119 mA cm 2) is considerably higher than that for CB (46 mA cm 2) and CB + CNF (80 mA cm 2). Three electrodes with different anode porous layer have a similar onset potential of 0.3 V. The higher current density is obtained for CNF porous layer-electrode due to the higher active surface area for methanol oxidation. The CNF porous layer featured with homogeneous porosity and high surface area, which leads to a uniform dispersion of catalyst and enhances the mass transport, thus improved the electrode performance. The stability of carbon porous layer as catalyst support is studied via CA test in electrochemical cells. Fig. 8 presents the CA

Fig. 4. VP-SEM morphology at 50 KX magnifications taken of the porous layer raw material (a) CB, Vulcan XC-72R and (b) CNF, platelet GNF.

A.M. Zainoodin et al. / Applied Energy 113 (2014) 946–954

951

Fig. 5. VP-SEM morphology at 100 and 10 KX magnification taken of the diffusion layers with different porous layers (a, b) CB, (c, d) CB + CNF and (e, f) CNF.

Fig. 6. VP-SEM morphology at 100 magnification taken of the anode electrode (catalyst layer) with different porous layer (a) CB, (b) CB + CNF, and (c) CNF.

curves of all the electrodes with different porous layer for a constant potential of 0.4 V and shows that the drop in current density due to the poisoning effect of methanol oxidation intermediates [16] and the loss of pathways in anode [10]. The CNF-porous layer-electrodes obtained a higher current peak and decreased 4% after 500 s compared with CB- and CB + CNF-porous layer-electrodes, which decreased by only 2%. Even though CNF showed a higher percentage decrease in current density compared with CB and CB + CNF, but the CNF still achieved higher current density at the end of 3600 s, is mainly due to its good dispersion and support for the catalyst, as well as the improved catalytic utilization. The EIS spectra of anode electrodes with various porous layers in half-cell configurations are shown in Fig. 9. All the anodes were prepared and tested under the same conditions except the carbon

nanostructure porous layer. There are two separated arcs is observed, the high frequency arc is due to charge transfer resistance that is related to electrode kinetics in catalyst layer while the low frequency arc is represented the mass transport through the diffusion layer. The CNF porous layer-electrode exhibited the lower total resistance compared with CB and CB + CNF porous layerelectrodes, indicating that the CNF porous layer could effectively controls the catalyst efficiency and methanol transport. While for CB porous layer, the size of plot grew largest. This suggested that the increase mass transfer resistance due to the micro-pores in CB which reduce the active catalyst sites thus inhibit the methanol chemisorptions to the catalyst sites. This was practically correlated with its low performances compared to CNF and CB + CNF as shown in Fig. 10.

952

A.M. Zainoodin et al. / Applied Energy 113 (2014) 946–954

3.3. Fuel cell performance

CB CB+CNF CNF

50

0

-50

-100 0.0

0.2

0.4

0.6

0.8

Potential voltage (V) Fig. 7. Cyclic voltammetry of methanol oxidation on (j) CB, (d) a mixture of CB + CNF with mass ratio of 1:1 and (N) CNF porous layer-electrodes. Measurements were performed in 0.5 M H2SO4 + 1.0 M CH3OH with a scan rate of 10 mV s 1.

14

CB CB+CNF CNF

-2

I (mA cm )

12

10

8

6

4 0

500

1000

1500

2000

2500

3000

3500

Time (sec.) Fig. 8. Chronoamperometry curves at 0.4 V potential. (j) CB, (d) a mixture of CB + CNF with mass ratio of 1:1 and (N) CNF porous layer-electrodes in a 0.5 M H2SO4 + 1.0 M CH3OH.

The electrode performance in a membrane electrode assembly (MEA) may differ from that in a half cell tests, thus the activity of electrodes with different anode porous layers have been further tested in a single-cell passive DMFC. The single-cell passive airbreathing DMFC performance results for different anode porous layers with 2 M methanol at room temperature are presented in Fig. 10. The highest peak power density of 18.3 mW cm 2 with a peak current density of 69 mA cm 2 was achieved for the DMFC with CNF anode porous layer. It also observed that ohmic polarization loss was less significant compared with CB and CB + CNF. This can be explained by the fact that the higher electrical conductivity and better catalyst utilization are provided by the CNF porous layer. The morphology of CNF provides a unique electrically conductive surface that enhances the reactivity and selectivity of supported metal catalyst particles to the maximum. This exposed higher surface area and provide unique properties of the CNF. The catalyst deposited uniformly on the CNF fiber as shown in Fig. 6(c) and become easily reachable by the reactant for methanol oxidation reaction. The peak power density of MEA occurs in the following order: CNF > CNF + CB > CB porous layer-anodic electrodes. This trend could be elucidated by understanding the role of CNF as porous layer to enhance the DMFC performance. In addition, the single cell with CNF porous layer exhibited the highest OCV among the three different porous layers due to the lowest reactant cross-over and organized electrode layer structure (almost crack free porous layer). For further comparison, DMFC performance tests were also performed at a high methanol concentration of 5 M at room temperature, and the results are shown in Fig. 11. The three single cells with different anode porous layer show a significant reduction in OCV indicates that using a higher methanol concentration feed could generate a higher rate of methanol crossover. Such a crossover effect is stronger than the electrode kinetics of CNF porous layer-electrode; therefore, a lower maximum power density was measured as 16.1 mW cm 2 at 5 M methanol compared with 18.3 mW cm 2 at 2 M methanol. However, the CNF porous layer still outperformed CB and CB + CNF even at higher methanol concentration, as presented in Fig. 11. It should also be mentioned that the improvement occurs in low current density (activation region) and high current density region (the mass transport region) which attributed to the improved catalyst activity and enhanced twophase transport of methanol and carbon dioxide in CNF-porous

0.7

25

CB CB+CNF CNF

0.6

18 16

Voltage (V)

0.5

2

-Z'' (ohm/cm )

20

20

CB CB+CNF CNF

2

2

Current density (mA/cm )

100

15

10

14 12

0.4

10 0.3

8 6

0.2

4

5

0.1

2

0.0 0

0 0

10

20

30

40

50

60

70

80

90

2

Power density (mW/cm )

150

10

20

30

40

50

60

70

80

90

2

Current density (mA/cm )

Z' (ohm/cm ) Fig. 9. Nyquist plot for EIS of (j) CB, (d) a mixture of CB + CNF with mass ratio of 1:1 and (N) CNF-porous layer electrodes in a 0.5 M H2SO4 + 1.0 M CH3OH solution.

Fig. 10. Polarization curves of three passive DMFCs with (d) CB, (j) a mixture of CB + CNF with mass ratio of 1:1 and (N) CNF anode porous layer using a 2 M CH3OH solution.

953

A.M. Zainoodin et al. / Applied Energy 113 (2014) 946–954

18

CB CB+CNF CNF

16 14

0.5

Voltage (V)

12 0.4

10

0.3

8 6

0.2 4 0.1

2

0.6

Power density (mW/cm )

0.7

2 0

0.0 0

10

20

30

40

50

60

70

80

90

2

Current density (mA/cm ) Fig. 11. Polarization curves of three passive DMFCs with (d) CB, (j) a mixture of CB + CNF with mass ratio of 1:1 and (N) CNF anode porous layer using a 5 M CH3OH solution.

Table 2 The OCV, maximum power density and the corresponding current density. Anode porous layer

2 M methanol OCV (V)

CB CB + CNF CNF

Pmax (mW cm

5 M methanol Imax ) (mA cm

2

0.622 12.6 0.612 13.2 0.640 18.3

54 60 69

OCV ) (V)

2

Pmax (mW cm

Imax ) (mA cm

2

0.545 11.5 0.570 10.6 0.550 16.1

)

58 53 60

0.7 0.6

20

0.5

15

0.4 10

0.3 0.2

5

2

2M 3M 4M 5M

Power density (mW/cm )

25

0.8

Voltage (V)

2

0.1 0.0 0

10

20

30

40

50

60

70

80

90

0 100 110

layer anode. Table 2 presents the OCV, maximum power density and corresponding current density of different anode porous layer MEAs. The experiment result reveals that the microstructure in the CNF porous layer can improve the fuel cell performance. For the electrode with CNF porous layer, the catalyst has been deposited on CNF layer evenly due to the high surface area of CNF. The uniformly disperse and close contact of the catalyst layer with membrane is practically beneficial to the fuel cell performance. In contrast, part of the catalyst has not been attached on the surface of CB porous layer or contacted with membrane directly and this reduces catalyst utilization and results in lower cell performance. Fig. 12 shows the performance of passive DMFC fabricated with CNF layer operated at various methanol concentrations, from 2 M to 5 M. When the methanol concentration increased from 2 M to 3 M, it can be observed that the limiting current density and the peak power density exhibit an obvious increase. This improvement is due to the improved methanol transport and reaction kinetic with higher methanol concentration. At 2 M methanol, the crackfree CNF porous layer increases the mass transfer resistance of methanol and reduces the methanol crossover, therefore performs higher OCV. By increasing the to methanol concentration to 3 M, the mass transfer resistance of methanol decreases and higher amount of methanol reached the catalyst site that will leads to a better performance. However, further increase the methanol concentration to 4 M and 5 M caused a decrease in cell performance. The methanol crossover from anode to cathode increases with further increasing methanol concentration and resulted a mixed potential occurs. An appropriate methanol concentration is important to achieve a high performance of DMFC cell. In our case, 3 M methanol is the most suitable concentration for DMFC with CNF porous layer. The maximum power density of 23.0 mW cm 2 at peak current density of 88 mA cm 2 is achieved (3 M methanol), which is highest among 2 M (18.3 mW cm 2, 69 mA cm 2), 4 M (21.0 mW cm 2, 86 mA cm 2), and 5 M (16.1 mW cm 2, 60 mA cm 2) methanol. Table 3 lists the performance results of passive DMFCs by different researchers under different operating conditions for comparison. It is observed that, this study presented the highest power density as compared to other researchers except for PCP with 22 M of methanol concentration however the advantages of this units are operating at lowest methanol concentration, lower catalyst loading that contributes to cheaper DMFC unit and will be more competitive for commercialization and less hazardous. Besides that, this DMFC is operated at room temperature without any elevated temperature and parasitic load for additional device.

4. Conclusion

2

Current density (mA/cm ) Fig. 12. Polarization curves of four passive DMFCs with CNF porous layer tested by different methanol concentrations (I) 2 M, (N) 3 M, (j) 4 M and (.) 5 M.

In summary, three anode electrodes with different porous layer has been prepared and tested in a passive air-breathing DMFC at room temperature. Compared with CB and CB + CNF, MEA with

Table 3 Comparison of the performance of passive DMFC with other researchers. Diffusion layer

Anode catalyst

Cathode catalyst

Methanol conc. (M)

Catalyst loading (mg cm2)

Peak power density (mW cm

2

PCP

Pt–Ru Black

Pt Black

16

10 (Anode, cathode)

24.0

[18]

PCP

Pt–Ru Black

Pt Black

22

12 (Anode, cathode)

30.0

[19]

CB

Pt–Ru

Pt

2

4 (Anode) 2 (Cathode)

3.0

[21]

CB (XC-72R)

Pt–Ru/C

Pt/C

4

3 (Anode) 2 (Cathode)

2.2

[32]

CNF

Pt–Ru Black

Pt Black

3

8 (Anode, cathode)

23.0

This study

)

Reference

954

A.M. Zainoodin et al. / Applied Energy 113 (2014) 946–954

the CNF layer yielded the highest maximum power density of 18.3 mW cm 2 with a 2 M methanol concentration. Porous layer made with different carbon structure materials has a substantial effect on the electrode performance, where the effect may relate to the porosity and surface area of the carbon structure. The use of highsurface-area CNFs as anode porous layer allowing more efficient catalyst dispersion and utilisation and better electrical conductivity with lower charge transfer resistance. This could effectively enhance the cell performance. DMFC cells with different anode porous layers were also tested with higher methanol concentrations (5 M). From the performance results, a higher maximum power density with a corresponding increasing current density is still achieved with CNF porous layer compared with CB and CB + CNF, indicating the feasibility of CNF layer in DMFC applications. Additionally, the performance of a passive DMFC was strongly depending on the methanol concentration used as indicated in our experimental results. Finally this study had proved that, the CNF layer-DMFC with 3 M methanol has shown the highest performance with a power density of 23.0 mW cm 2 with several advantages. It is concluded here that CNF is a promising candidate for anode layer fabrication for passive DMFCs. Acknowledgment The authors gratefully acknowledge the financial support given for this work by the Universiti Kebangsaan Malaysia under DIP2012-04 and National Science foundation (NSF) under Ministry of Science, Technology & Innovation, Malaysia (MOSTI). References [1] Karim NA, Kamarudin SK. An overview on non-platinum cathode catalysts for direct methanol fuel cell. Appl Energy 2012;103:212–20. [2] Kamarudin SK, Achmad F, Daud WRW. Overview on the application of direct methanol fuel cell (DMFC) for portable electronic devices. Int J Hydrogen Energy 2009;34:6902–16. [3] Thiam HS, Daud WRW, Kamarudin SK, Mohammad AB, Kadhum AAH, Loh KS, et al. Overview on nanostructured membrane in fuel cell applications. Int J Hydrogen Energy 2011;36:3187–205. [4] Achmad F, Kamarudin SK, Daud WRW, Majlan EH. Passive direct methanol fuel cells for portable electronic devices. Appl Energy 2011;88:1681–9. [5] Baldauf M, Preidel W. Status of the development of a direct methanol fuel cell. J Power Sources 1999;84:161–6. [6] Kamarudin SK, Daud WRW, Ho SL, Hasran UA. Overview on the challenges and developments of micro-direct methanol fuel cells (DMFC). J Power Sources 2007;163:743–54. [7] Seo SH, Lee CS. A study on the overall efficiency of direct methanol fuel cell by methanol crossover current. Appl Energy 2010;87:2597–604. [8] Salgado JRC, Paganin VA, Gonzalez ER, Montemor MF, Tacchini I, Ansón A, Salvador MA, Ferreira P, Figueiredo FML, Ferreira MGS. Characterization and performance evaluation of Pt–Ru electrocatalysts supported on different carbon materials for direct methanol fuel cells. Int J Hydrogen 2013;38:910–20. [9] Sharma S, Pollet BG. Support materials for PEMFC and DMFC electrocatalysts – a review. J Power Sources 2012;208:96–119. [10] Qi J, Jiang L, Tang Q, Zhu S, Wang S, Yi B, et al. Synthesis of graphitic mesoporous carbons with different surface areas and their use in direct methanol fuel cells. Carbon 2012;50:2824–31. [11] Zainoodin AM, Kamarudin SK, Daud WRW. Electrode in direct methanol fuel cells. Int J Hydrogen Energy 2010;35:c4606–21.

[12] Lo AY, Hung CT, Yu N, Kuo CT, Liu SB. Syntheses of carbon porous materials with varied pore sizes and their performances as catalyst supports during methanol oxidation reaction. Appl Energy 2012;100:66–74. [13] Santasalo-Aarnio A, Borghei M, Anoshkin IV, Nasibulin AG, Kauppinen EI, Ruiz V, et al. Durability of different carbon nanomaterial supports with PtRu catalyst in a direct methanol fuel cell. Int J Hydrogen Energy 2012;379:3415–24. [14] Nam K, Lim S, Kim S-K, Yoon S-H, Jung D-H. Application of silica as a catalyst support at high concentrations of methanol for direct methanol fuel cells. Int J Hydrogen Energy 2012;37:4619–26. [15] Antolini E, Gonzalez ER. Polymer supports for low temperature fuel cell catalysts. Appl Catal A 2009;365:1–19. [16] Antolini E, Gonzalez ER. Ceramic materials as supports for low-temperature fuel cell catalysts. Solid State Ion 2009;180:746–63. [17] Gharibi H, Kakaei K, Zhiani M, Taghiabadi MM. Effect of polyaniline-doped trifluoromethane sulfonic acid nanofiber composite film thickness on electrode for methanol oxidation. Int J Hydrogen Energy 2011;36:13301–9. [18] Abdelkareem MA, Nakagawa N. DMFC employing a porous plate for efficient operation at high methanol concentrations. J Power Sources 2006;162:114–23. [19] Abdelkareem MA, Morohashi M, Nakagawa N. Factors affecting methanol transport in a passive DMFC employing a porous carbon plate. J Power Sources 2007;172:659–65. [20] Park Y-C, Kim D-H, Lim S, Kim S-K, Peck D-H, Jung D-H. Design of a MEA with multi-layer electrodes for high concentration methanol DMFCs. Int J Hydrogen Energy 2012;37(2012):4717–27. [21] Yuan W, Tang Y, Yang X. High-concentration operation of a passive airbreathing direct methanol fuel cell integrated with a porous methanol barrier. Renew Energy 2013;50:741–6. [22] Shahbudin Masdar M, Tsujiguchi Takuya, Nakagawa Nobuyoshi. Mass spectroscopy for the anode gas layer in a semi-passive DMFC using porous carbon plate Part I: relationship between the gas composition and the current density. J Power Sources 2009;194(2009):610–7. [23] Eccariusa Steffen, Krause Falko, Beard Kevin, Agert Carsten. Passively operated vapor-fed direct methanol fuel cells for portable applications. J Power Sources 2008;182(2008):565–79. [24] Shahbudin Masdar M, Tsujiguchi Takuya, Nakagawa Nobuyoshi. Mass spectroscopy for the anode gas layer in a semi-passive DMFC using porous carbon plate Part II: relationship between the reaction products and the methanol and water vapor pressures. J Power Sources 2009;194(2009):618–24. [25] Okada Masataka, Konta Yuri, Nakagawa Nobuyoshi. Carbon nano-fiber interlayer that provides high catalyst utilization in direct methanol fuel cell. J Power Sources 2008;185(2008):711–6. [26] Wu QX, Zhao TS, Chen R, Yang WW. Effects of anode microporous layers made of carbon powder and nanotubes on water transport in direct methanol fuel cells. J Power Sources 2009;191(2009):304–11. [27] Park Gu-Gon, Sohn Young-Jun, Yang Tae-Hyun, Yoon Young-Gi, Lee Won-Yong, Kim Chang-Soo. Effect of PTFE contents in the gas diffusion media on the performance of PEMFC. J Power Sources 2004;131(2004):182–7. [28] Moreira J, Ocampo AL, Sebastiana PJ, Smit Mascha A, Salazar MD, del Angel P, et al. Infuence of the hydrophobic material content in the gas di(usion electrodes on the performance of a PEM fuel cell. Int J Hydrogen Energy 2003;28(2003):625–7. [29] Giorgi L, Antolini E, Pozio A, Passalacqua E. Infuence of the PTFE content in the diffusion layer of low-Pt loading electrodes for polymer electrolyte fuel cells. Electrochim Acta 1998;43(24):3675–80. [30] Yuan T, Zou Z, Chen M, Li Z, Xia B, Yang H. New anodic diffusive layer for passive micro-direct methanol fuel cell. J Power Sources 2009;192(2009):423–8. [31] Liu F, Wang C. Optimization of cathode catalyst layer for direct methanol fuel cells: Part I. Experimental investigation. Electrochim Acta 2006;52(2006):1417–25. [32] Hashim N, Kamarudin SK, Daud WRW. Design, fabrication and testing of a PMMA-based passive single-cell and a multi-cell stack micro-DMFC. Int J Hydrogen Energy 2009;34(2009):8263–9.