Microporous layer based on SiC for high temperature ...

Journal of Power Sources 288 (2015) 288e295

Contents lists available at ScienceDirect

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

Microporous layer based on SiC for high temperature proton exchange membrane fuel cells
ctor Zamora*, Pablo Can ~ izares, Jorge Plaza, Manuel Andre
s Rodrigo Justo Lobato, He University of Castilla-La Mancha, Chemical Engineering Department, Building Enrique Costa Novella, Ave. Camilo Jos
e Cela, 12, 13071, Ciudad Real, Spain

h i g h l i g h t s

g r a p h i c a l a b s t r a c t


SiC can be used in MPL of electrodes of HT-PEMFC.
SiC based MPLs exhibit good thermal and electrochemical stability.
HT-PEMFC operation results with SiC based electrodes are shown for the first time.
In comparing SiC and carbonaceous MPLs, improvement of the mass transport is observed.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 21 January 2015 Received in revised form 25 March 2015 Accepted 16 April 2015 Available online

This work reports the evaluation of Silicon Carbide (SiC) for its application in microporous layers (MPL) of HT-PEMFC electrodes and compares results with those obtained using conventional MPL based on Vulcan XC72. Influence of the support load on the MPL prepared with SiC was evaluated, and the MPL were characterized by XRD, Hg porosimetry and cyclic voltammetries. In addition, a short lifetest was carried out to evaluate performance in accelerated stress conditions. Results demonstrate that SiC is a promising alternative to carbonaceous materials because of its higher electrochemical and thermal stability and the positive effect on mass transfer associated to its different pore size distribution. Ohmic resistance is the most significant challenge to be overcome in further studies. © 2015 Elsevier B.V. All rights reserved.

Keywords: Silicon carbide High temperature PEMFC Microporous layer PBI Electrodes

1. Introduction Nowadays, the inherent disadvantages of fossil fuels in terms of the negative environmental effects associated with the production, transportation and use of energy are becoming more and more severe. Consequently, different energy sources and alternative technologies for the transformation of energy are looked for and, in this context, fuel cells are receiving a great deal of attention. Among

* Corresponding author. E-mail addresses: [email protected], [email protected] (H. Zamora). http://dx.doi.org/10.1016/j.jpowsour.2015.04.102 0378-7753/© 2015 Elsevier B.V. All rights reserved.

the various types of fuel cells, the Proton Exchange Membrane Fuel Cell (PEMFC) has received much attention over the past decade, for its application in transportation and portable devices, due to its low operating temperature, high efficiency and power density, nearly zero emission, and silent working conditions [1,2]. Main drawbacks of this technology are the complex water management, the high demand of platinum and the easiness in the catalyst poisoning. In order to overcome these limitations, in the last years, extensive research activities have been carried out with a variety of PEMFC that operates with vapour water (at temperatures over 100 ), called High Temperature Proton Exchange Membrane Fuel

J. Lobato et al. / Journal of Power Sources 288 (2015) 288e295

Cells (HT-PEMFCs). This technology may exhibit numerous advantages as compared to conventional Low Temperature PEMFC (LTPEMFC), overcoming several of their limitations [3,4]. However, working at high temperatures requires new materials for membranes, different to the traditional polymers typically used in LTPEMFC (such as Nafion). Polybenzimidazol (PBI) e based membranes are one of the most important and seem to be the most suitable for both, automobile and stationary applications [5]. The key component of PEMFCs is the membrane-electrode assembly (MEA), which is composed of a polymer electrolyte membrane, catalyst layers for the anode and cathode, and the gas diffusion layers (GDLs). The GDL, situated between the catalyst layer and bipolar plate, is subdivided into the gas diffusion backing (GDB) and microporous layer (MPL). The MPL provides the proper pore structure for the diffusion of the reactant gases and liquids, minimizes the electric contact resistance between the catalyst layer and bipolar plates, and in LT-PEMFC manages the water balance among production, expulsion, supply and evaporation [1,3,5]. Usually, carbon blacks are used to compose MPL. The high surface area (250 m2/g for Vulcan XC-72), low cost and its easy availability reduce the manufacturing cost of fuel cell. However, carbon black also shows important limitations related to its thermochemical instability and its corrosion in acidic conditions. In addition, its pore size and distribution clearly affect the interaction between the ionomer and the catalyst particles. It is wellcharacterized the reduction in the catalyst activity associated to the sulphur impurities and the deep micropores distribution (which suitable size to trap catalyst nanoparticles) [6,7]. Thus, it is crucial the development of new materials that can address all the positive features that carbon materials present and, at the same time, minimize the negative characteristics. One of most promising alternatives is the silicon carbide (SiC). Silicon carbide is covalent 1:1 stoichiometric carbide, keeping diamond structure in spite of the different size of silicon (Si) and carbon (C). Attending to its crystalline structure, SiC is divided into different groups. The main groups are a-SiC and b-SiC [8e10]. The a-SiC is characterized by its greater occurrence and stability at high temperatures (above 1973 K). b-SiC is characterized by a higher surface area as compared to a-SiC, low band gap and high electron mobility. In addition, it is important to take into account that SiC is a hard, solid semiconductor and presents good properties under high temperature and voltage [8,9]. Regarding other properties of interest, it shows a high porosity, with values of pore volume about 0.7 mL/g, and large surface area [10]. The SiC has been extensively used as catalyst support for thermochemical reactions [9]. However, to the authors' knowledge only one paper has been recently published in the field of PEMFC and it was focussed on not on HT-PEMFC but on LT-PEMFCs [11]. Thus, this work studies, for the first time, the use of SiC in the electrodes for HT-PEMFCs. To meet this objective, physical, chemical and electrochemical characterization of electrodes with different amounts of SiC was performed. Fuel cell tests were also carried out in order to check the viability and main features in single cell operation, including a short preliminary lifetest. 2. Experimental procedure High purity SiC was gently provided by SICAT (Paris, France). Attending to the specs provided by the supplier, this b-SiC has 30 m2/g BET surface, particle size between 50 and 150 mm. Vulcan carbon was purchased from Cabot (Boston, USA) and has 250 m2/g BET surface. Different MPL inks were prepared with Vulcan Carbon XC72 or b-SiC. Vulcan carbon ink was prepared with a carbon load of 2 mg/cm2 and 10% PTFE content, using isopropanol as solvent, according to the results obtained in previous works [12]. SiC inks

289

were prepared using 10% PTFE and different amounts of raw material (2, 4 and 6 mg/cm2), in order to evaluate the influence of the SiC content on the performance of the GDL. Inks were deposited by air brushing onto a commercial GDL (Toray Graphite paper TGPH120, 0.35 mm). 2.1. Physicochemical characterization XRD measurements were performed on a Philips PW-1700 diffractometer with rotating anode applying Ka corresponding to the transition from copper radiation for different samples. The type and distribution of the pores of the GDL and the different MPLs were studied by the porosimetry injection of mercury method. Porosity measurements were carried out by CSIC (Madrid, Spain). Respect to the tortuosity, graphical points were calculated with the data obtained from the mercury intrusion porosimetry (MIP) process, according to Hager equation (1):

k¼

r 24t2 ð1 þ rVtot Þ

h¼rZc; max

h2 f vðhÞdh

(1)

h¼rc; min

where r is material density; Vtot is total pore volume; t is tortuosity which is defined as the ratio of actual distance travelled (le) to R h¼r max shortest distance (l) (t > 1), and h¼rc;c; min h2 f vðhÞdh is the pore volume distribution by pore size. Electrical conductivity of each electrode was calculated at 125, 150 and 175  C using a potentiostat/galvanostat AUTOLAB PGSTAT 30 equipped with a FRA (Frequency Response Analysis) module. The frequency range applied was 100e10,000 Hz at 0.00 V (amplitude 0.01 V). The study of thermal resistance was performed at 185  C for 8 h. The electrode was previously impregnated with 85% phosphoric acid (PA) with a load of 10 mg PA/cm2. Acid load that reaches the electrodes under the conditions of operation of the fuel cell is about 0.17 mg/cm2. After 8 h, the samples were weighed and analysed by XRD to evaluate changes in the crystallite size and surface of the sample. Finally, to evaluate the electrochemical corrosion, samples were subjected to cyclic voltammetries, which were performed in a reactor filled with 2M PA, with a reference electrode Ag/AgCl. The maximum and minimum voltages were 0.79 and 0.41 V vs Ag/AgCl, respectively. The scanning was performed at 20 mV/s. 350 cycles were performed on each sample in order to evaluate the electrochemical degradation of the different MPLs. 2.2. Preparation of the MEA The electrodes were deposited onto a gas diffusion media (Toray Carbon Paper e PTFE 10%, Fuel Cells Etc.) by air-spraying a microporous layer (MPL) consisting of 4 mg/cm2 b-SiC (SICAT) and 10% PTFE (Teflon™ Emulsion Solution, Electrochem Inc.), which prevents the migration of the catalytic layer into the carbon support. After the deposition of the MPL, a heat treatment was applied to the electrodes at 360  C for 30 min for PTFE sintering. A catalyst layer was deposited by spraying the catalyst ink over the electrodes. The catalyst ink consisted of a 40% Pt/C catalyst on Vulcan XC-72R Carbon Black (Fuel Cell Store), PBI ionomer (1.5 wt. % PBI in N,Ndimethylacetamide, DMAc, 1e20 PBI/support ratio), and DMAc as a dispersing solvent. In both cases, the Pt loading on the two electrodes was 0.6 mg Pt/cm2. After the deposition of the catalyst layer, the electrodes were dried at 190  C for 2 h, with the purpose of removing traces of DMAc. The electrodes were then wetted with a solution of 10% PA. Electrodes were left to absorb the acid for one day. For the preparation of the MEA, a thermally treated PBI membrane was provided by Danish Power Systems®. This

290

J. Lobato et al. / Journal of Power Sources 288 (2015) 288e295

membrane was doped in 85 wt. % PA for 5 days, in order to achieve good proton conductivity. The doping level acquired by the membrane was 9 molecules of acid per polymer repeating unit. The corresponding thickness of the membrane was 83.2 mm. The superficial acid on the membrane was thoroughly wiped off with filter paper and the membrane was used to prepare the MEA. In order to fabricate the MEA, the doped membrane was sandwiched between a couple of electrodes and the whole system was hotpressed at 130  C and 1 ton for 15 min. The completed MEA was inserted into the cell between bipolar plates of graphite (with a five serpentine channels frame in each plate). The geometrical area of each electrode was 25 cm2.

A

2.3. Fuel cell tests and preliminary stability test MEAs prepared with thermal cured PBI membranes were mounted and characterized in a commercially available Cell Compression Unit (CCU) provided by Baltic fuel cells GmbH (Germany). The break-in procedure consists of operation at 0.1 A/cm2 and l (H2/air ratio) of 1.5/2 for 48 h. Preliminary stability test was conducted by increasing the current density to 0.2 A/cm2 (160  C). For further characterization, a protocol test was carried out every 48 h since the final of the break-in procedure. This protocol test consists of the following routine: - Galvanostatic polarization curves. They were performed from the OCV to 0.4 V. First with air at l ¼ 1.5/2 and then with oxygen at l ¼ 1.5/9.5. - Electrochemical impedance spectroscopy (EIS) tests. This method was used to determine the different resistances involved in the fuel cell. The EIS were performed at different current densities (0.03, 0.10, and 0.20 A/cm2) with 10 mV AC perturbation amplitude and frequency range from 100 kHz to 100 MHz. This sequence of EIS tests was carried out with air as comburent and then the same procedure was repeated with oxygen. - Cyclic voltammetries (CV). The Electrochemical surface area (ECSA) of cathode was estimated with this technique. The cathode side was purged with nitrogen and hydrogen flowed through anode side with flows of 0.1/0.1 L/min N2/H2. The CV was carried out from 0.05 V to 1.0 V with a scan rate of 100 mV/s. - Linear sweep voltammetry (LSV). This technique was performed to find out any crossover of gas flow through MEA. The same gases of the CV were fed with flows of 0.3/0.3 L/min N2/H2. Thus, it could be considered this preliminary lifetest under accelerate stress conditions. 3. Results and discussion 3.1. Physical characterization SiC powders and MPLs were analysed by XRD. Diffractograms are shown in Fig. 1. In Part a, it can be observed that SiC presents characteristics peaks at 2q of 36 , 41, 60 and 74 , being 36 the main peak, which may be attributed to the face (111) [13]. In Part b, the XRD patterns of the electrodes prepared with MPLs with 2, 4 and 6 mg SiC/cm2 are compared. As it can be observed, one high intensity peak appears at 2q ¼ 25 and it decreases in size when amount of SiC increases. This peak may be attributed to carbon [14], and it can be associated with the carbonaceous structure of the gas diffusion layer which supports the microporous layer. As it was expected, as the SiC content increases the MPL thickness increases and hence the carbonaceous layer is more coated by the SiC based MPL (carbon content decreases).

B

Fig. 1. A) XRD patterns obtained for SiC powder. B) XRD patterns for different MPLs prepared with SiC.

Crystal size (Lc) and distance between planes (d002) was calculated for each sample using equations (2) and (3), respectively, where LC is the crystal size (nm), l corresponds to the Ka radiation of copper (l ¼ 0.15418 nm), B is the peak width of Imax/2 and q the angle corresponding to Imax.

LC ¼

0:89,l B,cosðqÞ

d002 ¼

l 2,senðqÞ

(2)

(3)

In the case of MPLs, Lc and d002 values are considered apparent, because of the addition of PTFE in the microporous layer, which causes agglomeration. Therefore, these values shall only be used for comparative purposes when making stability assessments, as described later on in this work. Table 1 shows the Lc and d002 values obtained for different MPLs before and after thermal and electrochemical tests (which will be discussed later on). First five rows compare the powder and the raw MPL properties. Respect to d002, it can be observed that values are comparable for the powder and the MPL samples. As expected, the SiC based materials are more crystalline as compared to the Vulcan carbon, with a lower d002 value, which means that its particles are more ordered. Fig. 2A shows the variation of cumulative pore volume for different SiC based MPL and a GDL. Porosimetry is an important parameter, which may be related to the diffusion coefficient of porous substrates [15], because it is directly related to the gas

J. Lobato et al. / Journal of Power Sources 288 (2015) 288e295 Table 1 Crystal size, apparent crystal size and d002 from all powder and samples analysed. Sample

Lc

SiC (powder) Vulcan XC72 (powder) SiC (MPL 2 mg/cm2) SiC (MPL 4 mg/cm2) SiC (MPL 6 mg/cm2) Vulcan XC72 (MPL) SiC (MPL 2 mg/cm2) after TT SiC (MPL 4 mg/cm2) after TT SiC (MPL 6 mg/cm2) after TT Vulcan XC72 (MPL) after TT SiC (MPL 2 mg/cm2) after CV SiC (MPL 4 mg/cm2) after CV SiC (MPL 6 mg/cm2) after CV Vulcan XC72 (MPL) after CV

15.9 2.3 14.0 14.1 13.8 18.5 14.1 13.6 13.3 20.9 14.3 14.1 14.0 23.3

SiC

(nm)

D002 (nm) 0.252 0.348 0.252 0.253 0.254 0.335 0.252 0.253 0.254 0.336 0.252 0.253 0.254 0.338

A

B

291

permeability of the electrode. In comparing results shown in Fig. 2, it can be clearly observed that the higher the SiC load in the MPL, the lower is the cumulative pore volume. This behaviour was expected because of the thicker SiC based layer and the change is more pronounced in the highest pore size range, which could be due to the agglomeration of the material. This process has been observed in previous works with carbonaceous materials [15] and suggests that the highest load of SiC (6 mg/cm2) should be avoided in the manufacturing of the MPL. Parts B and C of Fig. 2 compare the porosity and tortuosity of the different MPLs and GDL, respectively. As it is expected, the increasing of the SiC load decreases the porosity and increases the tortuosity. This observation can be explained in terms of the MPL intrinsic microporosity, which leads to a decrease in the overall porosity. On the other hand, higher SiC loads mean thicker MPL layer which can block part of the macropores of the GDL, leading to a decrease in the porosity and to an increase in the tortuosity [15]. Tortuosity can be defined as the reciprocal of the ratio between the actual trajectory covered by the fluid when moving through the medium between two points under rectilinear one. As expected, higher SiC contents lead to greater tortuosity for any fluid when this penetrates into the SiC based electrode. Nevertheless, the tortuosity values are similar (slightly higher) than others obtained in a previous work, where the influence of carbon load on the MPL was studied [12]. Fig. 3 shows the electrical conductivities at different temperatures, measured for the electrodes prepared with different MPLs. In comparing results, it can be observed that the higher the load of SiC, the lower is the conductivity. In addition, in the cases of the MPLs with 2 mg/cm2 and 4 mg/cm2 of SiC, the electrical conductivities are comparable to the conductivity reached by the standard electrode with the carbon material and even comparable to the one obtained using advanced carbonaceous materials [16]. This observation means that the conductivity values of SiC are acceptably enough to be used in a HT-PEMFC.

3.2. Electrochemical and thermal degradation As it has been pointed out previously, high stability of the components of the HT-PEMFC is a challenge, especially if it is taken into account the harsh operation conditions during HT-PEMFC performance, which combines acidic environment and high temperature. In order to assess the stability from the viewpoint of the high temperatures that undergo these systems, the MPL studied were subjected to a thermal treatment in hot acidic media. The samples were weighed to determine weight loses associated with the degradation of materials and no changes in the weight were

C

Fig. 2. A) Cumulative pore volume vs pore diameter for different samples. B) Evolution of the porosity. C) Evolution for tortuosity for different samples tested.

Fig. 3. Conductivity values vs Temperature for different MPLs tested.

292

J. Lobato et al. / Journal of Power Sources 288 (2015) 288e295

found. XRD were performed to the MPLs before and after the thermal treatment. Table 1 shows the results of this characterization (rows 6e9). In the case of the MPL with Vulcan Carbon, differences higher than 13% were found in the apparent crystal size before and after the thermal treatment. Conversely, variations lower than 5% were found in the SiC based MPLs. Nevertheless, in data not shown, it is important to take into account that in the case of MLP with 2 mg/ cm2, it was observed a higher degradation of the peak associated with GDL (carbon structure), which means that the MPL does not suffer degradation but this lower amount of SiC cannot cover and protect the whole GDL surface. This fact could also explain the higher conductivity reached by SiC MPL prepared with the lowest SiC content. Electrochemical resistance of the MPLs was evaluated by sequential cyclic voltammetry tests in PA media. Fig. 4 shows the changes observed in the voltammograms during the tests carried out tor MPLs prepared with of 2, 4 and 6 mg/cm2 SiC. Changes in the voltammograms are indicative of the degradation of the electrode [17]. However, and opposite to many other studies about degradation of carbonaceous electrodes, as SiC is a non-carbonaceous material, the changes of the quinoneehydroquinone (QeHQ) couple do not appear and, hence, they cannot be used as an indicator of the degradation of the surface [17e19]. As it can be observed in part A, the SiC based MPL prepared with the lowest load presents a high degradation rate during the first 100 cycles. After that, it presents a very high stability during the rest of the test and changes were very small. This behaviour could be attributed to the presence of impurities on the surface (deposited before the experiment) that are removed with the initial voltammetries, rather than to a low stability. Moreover, the cyclic voltammograms of the MPL prepared with 2 mg SiC/cm2 have a different form, probably due to the degradation of GDL. As explained before the formation of the Q-HQ peak at 0.4 V approx. (vs Ag/AgCl) was not observed in the anodic CV curve (as occurred in the case of the electrodes prepared with different carbonaceous materials and which is indicative of surface oxidation [16]). In the case of the MPL with 4 and 6 mg SiC/cm2 the CVs did not change during the tests, which means that these electrodes show a very good electrochemical stability. This fact can be confirmed with the XRD parameters obtained from the samples after the cyclic voltammetries, in Table 1 (rows 10e13). As it can be observed, the changes of apparent Lc in the case of SiC samples are very low as compared to the Lc variation in the case of Vulcan carbon. Not new peaks were detected which indicates the absence of oxides or another substances formation on the surface of the electrodes [17e19].

A

B

C

3.3. HT-PEMFC results Once the electrodes were characterized and the thermal and electrochemical degradation was evaluated, a short lifetest with the SiC based electrodes was performed in a single cell. To the authors' knowledge, this type of material has not been tested before in HTPEMFC, so this result becomes a first reference about performance of this novel material. At this point, it should be taken into account that it is not an optimized assessment but just a first proof of concept, in which harsh conditions (in terms of monitoring) are going to be applied. Fig. 5 shows the changes in the voltage and the polarization curves performed to the fuel cell during the lifetest. As it can be observed in part A, three characterization protocols were carried out throughout the essay, as detailed in the Section 2.3. The initial voltage of the fuel cell was close to 0.40 V and it decreases continuously during the lifetest. At the end of

Fig. 4. Evolution of voltammograms from SiC based MPLs. A) SiC MPL 2 mg/cm2. B) SiC MPL 4 mg/cm2. C) SiC MPL 6 mg/cm2.

experiment, the degradation rate of the fuel cell was around 470 mV/h. In comparing this value with other previously shown in the literature, it can be stated that this value is too high, although it must be pointed out that during this essay different electrochemical characterization tests (polarization curves with air and with oxygen, EIS at different current densities with air and with O2, CVs and LSVs) were performed which could contribute to accelerate the degradation of the fuel cell. Thus, for PBI-based membranes, degradation rates measured in previous studies were 14 mV/ h and 250 mV/h [20], 194.2 mV/h [21] and 25 mV/h [22].

J. Lobato et al. / Journal of Power Sources 288 (2015) 288e295

A

Test Protocol I

Test Protocol II

Cell Voltage (V)

0.7

Test Protocol III

0.6 0.5 0.4 0.3 0.2 0.1 0 0

10

20

30

40

50

60

70

80

90 100

Time (h)

B

Protocol 1

1 Voltage (V)

100

Protocol 2

80

Protocol 3

0.8

60

0.6 40

0.4

20

0.2 0 0

0 300

200

Current density [mA/cm2]

C

120

Protocol 2

1

100

Protocol 3

0.8

80

0.6

60

0.4

40

0.2

20

0 0

100

200

300

0 400

Power density mW/cm2)

Protocol 1

1.2

Voltage (V)

100

Power density (mW/cm2)

1.2

293

operating at 160  C and testing new membranes in a 50 cm2 single cell. No information about the catalyst loading was reported. During the preliminary lifetest the power density decreased 9.7% when oxygen was used and 5.5% with air. This means that under the applied operation conditions, the degradation was much higher with oxygen than with air. Moreover, during the lifetest the MEA showed very good open-circuit voltage values at 0.0 A/cm2, indicating that not short circuits, electrical losses or damages in the membrane occurred. As it was pointed out above, during the lifetest, impedance analyses were carried out at different operation times. Fig. 6 shows the Nyquist plots of the MEA with SiC based electrodes. For comparison purposes, it is also shown the Nyquist plots of a MEA with standard carbon based electrodes, measured under the same operation conditions and with the same equipment. In order to know the resistance values originate in the fuel cell, an equivalent circuit R(RQ) (RQ) was used to simulate the impedance data in this work [25e27]. The parameters obtained are shown in Table 2. The high frequency resistance is typically explained in terms of the sum of the ohmic resistances of the membrane electrolyte and the electric resistance of the electrode (GDL þ MPL, mainly). It can be observed that this high frequency resistance is higher in the case of the SiC based electrodes than in the standard carbon based electrodes. Ohmic resistance values around 1 U/cm2 were obtained for the SiC based electrodes, which are five times higher than the values obtained under the same operation condition with standard carbon based electrodes (around 0.2 U/cm2). Taking into account that in both cases the same PBI membrane and the same catalyst layer were used, this difference should be attributed to the different MPL. Thus, it seems that the SiC does not behave in a HT-PEMFC as a good electrical conductor, contrary to it was expected from the results showed in Fig. 3. Furthermore, the semicircle observed in the Nyquist plot is explained in terms of the charge transfer resistance (RCT) or of the oxygen reduction reaction (ORR) activation in the cathode catalyst layer. The electrodes prepared with the SiC based MPL show a higher RCT as compared to the carbonaceous supported MPL (0.6 U/ cm2 vs 0.45 U/cm2 for the standard one). These values are higher, although it has to be accounted that the novel SiC based electrodes are not optimized in terms of the overall composition, particle size, the ratio ionomer/SiC, the amount of phosphoric acid, etc. Nevertheless, it is remarkable that the change of the charge transfer resistance, when air was used, during the 100 h of operation was more noticeable in the case of the electrode with the carbon based MPL (28.0%) while in the case of the SiC based MPL the charge transfer increased 18.3% which means that the electrode suffered a

Current density (mA/cm2) Fig. 5. A) Evolution of Cell Voltage vs time for MEA prepared with SiC based electrodes. B) Polarization curve performed with Air. C) Polarization curve performed with Oxygen.

Results obtained in the different polarization curves carried out with air and oxygen are shown in parts B and C, respectively. The maximum power density values are about 100 mW/cm2 when it was operating with oxygen and 79 mW/cm2 when air was used as comburent. These values are much lower than values obtained by the same group (600 mW/cm2 [21]) with a carbonaceous support. However, this power density values are similar or even higher than others reported recently in the literature when new materials have been tested. Thus, Kurdakova [23] reported power densities around 75 mW/cm2 operating with air, at 150  C and 0.5 mg Pt/cm2 in both electrodes and testing new PBI based membranes in a 5 cm2 single cell. Ossiander [24] reported only values around 33 mW/cm2

Fig. 6. Nyquist plot for Vulcan carbon and SiC based electrodes at different times.

294

J. Lobato et al. / Journal of Power Sources 288 (2015) 288e295

Table 2 Ohmnic, charge transfer and mass transfer resistances obtained from MEA impedances carried out at 0.1 A/cm2 during the short lifetest. Values of ECSA at different times from CV are also shown. Air MPL

Time (h)

Rohm (mU/cm

SiC

Vulcan XC72

0 46 100 0 46 100

1047.5 1090.0 1192.5 184.0 194.0 203.5

ECSA (m2/g)

Oxygen RCT 2

)

(mU/cm

RMT 2

)

790.0 932.0 935.2 402.3 488.3 514.8

lower degradation process. When oxygen was used, the changes in the charge transfer resistance were similar in both electrodes, around 3%. On the other hand, at low frequency, another arc appears on the EIS that may be related to the mass transport limitations. It can be clearly observed that in the case of the SiC based electrodes this arc does not appear whereas in the case of the carbon based electrodes it clearly does. This means that the SiC based MPL improves the mass transfer and is more beneficial at high current conditions. Table 2 also shows also the ECSA (ElectroChemical Surface Area) values obtained at different times of the essay. At the beginning, a value of 15.04 m2/g is obtained. In comparing this value with other shown in the literature, it can be observed that it is similar to these values (above 19.4 m2/g) [22]. However, there is a significant decrease throughout the test obtaining a non-negligible percentage of degradation of 14.3%, as it has been stated previously. This value (14.3%), is lower than the one (19.0%) obtained by the electrode prepared with the standard MPL (Vulcan carbon) which means that the SiC based MPL contributed to a better stabilization of the catalyst layer. This observation is in agreement with the previous statement about the charge transfer resistance where a better stability of the electrodes prepared with the SiC based MPL was observed. It cannot only be related to the material but to the three protocol tests which may cause a degradation more pronounced than that reported by other authors who made a less aggressive characterization during the lifetest. This is not a good result, because stability was the main feature looked for in this research and although off-site tests were promising, lifetest yields less interesting results. For studying the effect of this novel material on mass transport polarization curves at constant flow corresponding to a current density of 200 mA/cm2 were performed to the fuel cells with SiC based electrodes and carbon based electrodes at different times, both with air as oxidant (Fig. 7). In comparing results, it can be observed that the fuel cell with the carbon based electrodes mass transfer limitations appear at high current density because the gas flow is fed below the stoichiometric ratio. A limiting current density about 270 mA/cm2 is clearly observed. However, for the same conditions, these limitations do not appear for the SiC based electrodes. These results confirm the previous data obtained from the EIS under different operation conditions: the SiC based MPL contributes to improve the performance of the fuel cell at high current. Nevertheless, much effort must be carried out to improve the electrical conductivity to decrease the ohmic resistance and create a more continuous interface between the catalyst layer and the MPL in order to decrease the RCT.

4. Conclusions For the first time, MPLs with three different SiC loads (2, 4 and

(mU/cm e e e 867.5 707.0 576.0

Rohm 2

)

(mU/cm 1087.5 1175.0 1270.0 182.0 198.3 199.8

RCT 2

)

(mU/cm2) 646.2 619.7 678.0 446.5 465.0 458.5

15.07 14.21 12.91 27.88 22.87 22.58

A

B

Fig. 7. Polarization curves at constant flow (j ¼ 200 mA/cm2) and different times. A) MEA with SiC based electrodes. B) MEA with Vulcan based electrodes.

6 mg/cm2) have been prepared and compared with a carbon based MPL. The following conclusions can be drawn from this work: 1) The electrical conductivity decreases with the SiC content in the MPL. The conductivity for the two lowest contents is comparable to the conductivities obtained with the carbon based MPL. 2) The SiC based MPLs with 4 and 6 mg SiC/cm2 showed much better thermal and electrochemical stability under the applied experimental conditions than carbonaceous based MPLs. 3) The electrode prepared with a SiC based MPL shows a very high ohmic resistance which contributes to lower fuel cell performance. Stability measured under the same operation conditions is better to that obtained with the standard electrodes. 4) The electrode prepared with the SiC based MPL did not show mass transport problems when undergo conditions in which electrodes prepared carbonaceous based MPL do.

J. Lobato et al. / Journal of Power Sources 288 (2015) 288e295

Thus, it can be confirmed that SiC is a promising material that can be used in HT-PEMFCs technology although much work must be carried out in order to improve both the fuel cell performance and durability. Acknowledgements The authors further thank the European Commission as this work was supported by the Seventh Framework Programme through the project CISTEM (FCH-JU Grant Agreement Number 325262). Moreover, Sicat Co is also gratefully acknowledged for the SiC supply. References [1] M. Ahn, Y.-H. Cho, Y.-H. Cho, J. Kim, N. Jung, Y.-E. Sung, Electrochim. Acta 56 (2011) 2450e2457. [2] L. Carrette, K. Friedrich, U. Stimming, Fuel Cells 1 (2001) 5e39. [3] Q. Li, R. He, J.O. Jensen, Fuel Cells 4 (2004) 147. ~ izares, M.A. Rodrigo, D. Úbeda, F.J. Pinar, J. Power Sources 196 [4] J. Lobato, P. Can (2011) 8265e8271. [5] A. Chandan, M. Hattenberger, A. El-kharouf, S. Du, A. Dhir, V. Self, B.G. Pollet, A. Ingram, W. Bujalski, J. Power Sources 231 (2013) 264e278. [6] B. Avasarala, R. Moore, P. Haldar, Electrochim. Acta 55 (2010) 4765e4771. [7] Y.-C. Park, K. Kakinuma, M. Uchida, H. Uchida, M. Watanabe, Electrochim. Acta 123 (2014) 84e92. [8] S. Gianella, D. Gaia, A. Ortona, Adv. Eng. Mater. 14 (2012) 1074e1081. [9] Y. Liu, B. de Tymowski, F. Vigneron, I. Florea, O. Ersen, C. Meny, P. Nguyen,

295

C. Pham, F. Luck, C. Pham-Huu, ACS Catal. 3 (2013) 393e404. [10] A. Walcarius, Chem. Soc. Rev. 42 (2013) 4098e4140. [11] R. Dhiman, E. Johnson, E.M. Skou, P. Morgen, S.M. Andersen, J. Mater. Chem. A 1 (2013) 6030e6036. ~ izares, M.A. Rodrigo, D. Úbeda, F.J. Pinar, J.J. Linares, Fuel Cells [12] J. Lobato, P. Can 10 (2010) 770e777. [13] P. Nguyen, C. Pham, Appl. Catal. A General 391 (2011) 443e454.
nez, A. Nieto-Ma
rquez, J.A. Díaz, R. Romero, P. Sa
nchez, J.L. Valverde, [14] V. Jime A. Romero, Ind. Eng. Chem. Res. 48 (2009) 8407e8417. ~ izares, M.A. Rodrigo, C. Ruíz-Lo
pez, J.J. Linares, J. Appl. Elec[15] J. Lobato, P. Can trochem. 38 (2008) 793e802. ~ izares, M.A. Rodrigo, J. Lobato, Fuel Cells 15 (2) (2015) [16] H. Zamora, P. Can 375e383. [17] B. Avasarala, R. Moore, P. Haldar, Electrochim. Acta 55 (2010) 4765e4771. [18] K.H. Kangasniemi, D.A. Condit, T.D. Jarvi, J. Electrochem. Soc. 151 (2004) E125. [19] E. Passalacqua, P.L. Antonucci, M. Vivaldi, A. Patti, V. Antonucci, N. Giordano, K. Kinoshita, Electrochim. Acta 37 (1997) 2725e2730. [20] S. Yu, L. Xiao, B.C. Benicewicz, Fuel Cells 3e4 (2008) 165. ~ izares, M.A. Rodrigo, D. Úbeda, J.J. Linares, J. Power Sources 196 [21] J. Lobato, P. Can (2011) 8265e8271. [22] A.D. Modestov, M.R. Tarasevich, V.Y. Filimonov, N.M. Zagudaeva, Electrochim. Acta 54 (2009) 7121e7127. [23] V. Kurdakova, E. Quarterone, P. Mustarelli, A. Magistris, E. Caponetti, M.L. Saladino, J. Power Sources 195 (2010) 7765e7769. €nberger, P. Vo € lk, M. Welsch, C. Scheu, [24] T. Ossiander, C. Heinzl, S. Glech, F. Scho J. Membr. Sci. 454 (2014) 12e19. ~ izares, M.A. Rodrigo, J.J. Linares, D. Úbeda, F.J. Pinar, Fuel Cells [25] J. Lobato, P. Can 10 (2010) 312e319. [26] J. Zhang, Y. Tang, C. Song, J. Zhang, J. Power Sources 172 (2007) 163e171. [27] X. Yuan, H. Wang, J.C. Sun, J. Zhang, Int. J. Hydrog. Energy 32 (2007) 4365e4380.