Pore structure and effective diffusion coefficient of

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 0

Available online at www.sciencedirect.com

ScienceDirect journal homepage: www.elsevier.com/locate/he

Pore structure and effective diffusion coefficient of catalyzed electrodes in polymer electrolyte membrane fuel cells Jian Zhao, Samaneh Shahgaldi, Ibrahim Alaefour, Song Yang, Xianguo Li* 20/20 Laboratory for Fuel Cell and Green Energy RD & D, Department of Mechanical and Mechatronics Engineering, University of Waterloo, Waterloo, ON, N2L 3G1, Canada

article info

abstract

Article history:

For polymer electrolyte membrane (PEM) fuel cells, the pore structure and small effective

Received 21 November 2017

diffusion coefficient (EDC) of the catalyst layers have significant impact on the cell per-

Received in revised form

formance. In this study, both the pore structure and EDC of the catalyst layers are inves-

20 December 2017

tigated experimentally; the pore structure of the catalyst layer is characterized by the

Accepted 7 January 2018

method of standard porosimetry, and the EDC is measured by a modified Loschmidt cell for

Available online xxx

oxygen-nitrogen mixture through the catalyzed electrodes. It is found that Pt loading has a direct impact on the pore structure and consequently the EDC of the catalyzed electrodes.

Keywords:

As the Pt loading is increased, the porosity and mean pore size of the catalyzed electrode

PEM fuel cell

decrease, and the EDC decreases accordingly, however, it is increased by 15e25% by

Pore structure

increasing the temperature from 25 C to 75 C. The EDC of the catalyst layer is about

Effective diffusion coefficient

4.6 107 m2 s1 at 75 C, compared with 25.0 107 m2 s1 for the uncatalyzed electrode at

Catalyzed electrode

the same temperature.

Pt loading

© 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.

Introduction Gas diffusion in the electrode of polymer electrolyte membrane (PEM) fuel cell is one of the dominant mechanisms of transporting the reactants from the flow channels to the reaction sites [1e3]. The gas diffusion in the porous electrode is commonly represented by the effective diffusion coefficient (EDC) [4e8], and higher values of EDCs indicate better capabilities of gases to penetrate through the porous electrode, which is beneficial for fuel cell performance [9,10]. In a single PEM fuel cell, the reactant diffusion occurs in both the gas diffusion layer (GDL) and catalyst layer (CL), which constitute the electrode. The EDC, in turn, is governed by the pore

structures of these porous layers at the cell operating condition, and is instrumental to fuel cell operation at high current densities, which is essential for practical applications [7]. Great efforts have been made to reconstruct the micromorphology of electrode materials and to optimize the microstructure of the electrodes through three-dimensional pore-scale simulation [11e14]. The pore structure of the electrode is complex due to the presence of various materials in different layers [15e18]. The catalyst layer, adjacent to the membrane, is made of carbon-supported platinum (Pt/C) particles and ionomer. The solid structure provides electrochemical reaction sites as well as electronic and protonic pathways, while the remaining pores for the reactant diffusion and water management. The catalyst layer is protected

* Corresponding author. E-mail address: [email protected] (X. Li). https://doi.org/10.1016/j.ijhydene.2018.01.019 0360-3199/© 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved. Please cite this article in press as: Zhao J, et al., Pore structure and effective diffusion coefficient of catalyzed electrodes in polymer electrolyte membrane fuel cells, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.01.019

2


and supported by the GDL, which is made of a carbon paper or cloth and a micro-porous layer (MPL). The carbon paper/cloth is often treated by Polytetrafluoroethylene (PTFE) in order to improve its hydrophobicity, while the MPL, made of carbon particles with PTFE binders, provides a good connection between the carbon paper/cloth and catalyst layer, reducing the interfacial resistance and enhancing the performance of the PEM fuel cells [1,19,20]. The pore structure of the GDLs has been extensively investigated in terms of fiber diameter, fiber direction, layers of fibers, PTFE treatment, as well as MPL coating [21e24]; however, the factors affecting the pore structure of the CLs including the types of catalyst and ionomer, CL composition, as well as Pt loading have not been investigated adequately, although pore structure of the catalyst layer has been shown to have significant impact on the cell performance [15,25]. Various techniques for the characterization of the pore structure in terms of the porosity and pore size distribution (PSD) have been employed including mercury porosimetry [26,27], gas adsorption [28], capillary condensation [29,30], small-angle X-ray scattering [31], displacement method [32], and optical and electronic microscopy [33]. However, each of these methods has its own limitations on testing the GDL and CL [26e34]. The method of mercury porosimetry might deform the target pore structure due to the high pressure of mercury [35e37]. The methods of gas adsorption, capillary condensation, and small-angle X-ray scattering are suitable for small pores (up to 50 nm), while the displacement method works well only for large pores (10e200 mm) [28e35]. The PSD obtained from the optical microscopy and electronic microscopy depends on the data-processing algorithm, and the pore size is calculated mostly from surface images. Volfkovich et al. [35e37] developed the method of standard porosimetry (MSP) based on capillary equilibrium, which is able to measure the GDLs and CLs in a wide range of pore sizes (0.3 nm - 300 mm) under room conditions. Theoretically, the EDC in the porous media is related to the volume fraction of the void region (i.e., porosity F), the length fraction of the tortuous flow path to the straight line length (i.e., tortuosity t), and the bulk diffusion coefficient [38]. The EDC of porous media can be calculated using the following equation [38]: Deff ¼

∅Dbulk t

(1)

However, the tortuosity, t, is unknown for the porous electrode in PEM fuel cells. For simplicity, many models have been developed for the EDC of porous materials as a function of porosity only [39e45], as summarized in Table 1. However, these models are built based on the assumption that the materials are made of either fibers or particles; therefore, these models may not be suitable for the PEM fuel cell electrodes because their pore structure is more complex with agglomerates, binders, and ionomers. Various methods have been developed to experimentally measure the EDCs of the porous layers in PEM fuel cells. Baker et al. [46] determined the EDC of the GDLs using water vapor in air mixture through the method of limiting current. LaManna and Kandlikar [47] investigated the water vapor diffusion coefficients of various GDLs with and without MPLs using a dynamic diffusion test cell. Casalegno et al. [48] utilized a single fuel cell with humidified and dry air flow to measure water vapor concentrations for GDLs. Kramer et al. [49] and Flu¨ckiger et al. [50] employed the electrochemical impedance spectroscopy to determine the effective relative diffusivity for GDLs under compression conditions. Shen et al. [8] and Chan et al. [51] modified the Loschmidt cell (also referred to as the closed-tube method) to investigate the EDCs of the nitrogenoxygen gas pair for the GDLs and Al2O3-membrane-supported catalyst layers, respectively. Of all the above methods, the Loschmidt cell is one of the most commonly used methods to determine the diffusion coefficients of the gas pairs because of its feasibility of modification for different porous media, high accuracy, short experimental execution time, ease of operation, and simple experimental configuration [8,38,52]. Further, the measured EDC of a carbon paper was found to be in good agreement with the three-dimensional (3D) simulation result of gas diffusion in the sample [44], indicating the advantages of the Loschmidt cell in studying the porous layers in PEM fuel cells. Despite the fact that the pore structure and EDCs of the GDLs have been widely investigated, few studies have been directed to the practical catalyst layers or the GDL-supported catalyst layers which are the complete paths for the gas transport in PEM fuel cells. Therefore, the objective of the present study is to experimentally investigate the effect of the catalyst layers, with a focus on Pt loadings, on the pore structure and EDC of the catalyzed electrodes. In this study, the catalyzed electrodes with the Pt loadings of 0.1, 0.2, 0.3,

Table 1 e Various models for the effective diffusion coefficient (EDC) of porous materials. Model Bruggeman (1937) Neale and Nader (1973) Tomadakis and Sotirchos (1993)

Effective diffusion coefficient Deff ¼ ∅ Dbulk Deff ¼ ½2∅=ð3 ∅ÞDbulk 1:5

Deff ¼ ∅½ð∅ 0:037Þ=0:9630:661 Dbulk

Zamel et al. (2009)

Deff ¼ ½1 ð1 ∅Þ0:46 Dbulk ð0 ∅ 0:65Þ Dbulk ð0:33 ∅ 1Þ Deff ¼ 1 2:76∅coshð3∅ 1:92Þ 3ð1∅Þ 3∅

Das et al. (2010)

Deff ¼ Dbulk

Mezedur et al. (2002)

3ð1∅ÞDbulk 3Dbulk ∅ 2f D Dbulk m m 3fm

Note

Source

Spherical particles

[39,40]

Spherical particles Fibers

[41] [42]

Tetragonal network

[43]

Fibers

[44]

Catalyst layers

[45]

Note: Dm is the diffusivity in ionomer and fm is the volume fraction of ionomer.

Please cite this article in press as: Zhao J, et al., Pore structure and effective diffusion coefficient of catalyzed electrodes in polymer electrolyte membrane fuel cells, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.01.019


and 0.4 mg cm2 are examined in comparison with the reference GDL (without the catalyst layer, or uncatalyzed). For today's PEM fuel cell technology, the Pt loading is typically around 0.4 mg cm2, and the target low Pt loading is about 0.1 mg cm2, mostly based on the US Department of Energy (DOE) target for total Pt loading of 0.125 mg cm2 to be achieved by 2020 [53,54]. As a result, for the present study, the Pt loadings of 0.1, 0.2, 0.3, and 0.4 mg cm2 are selected as representatives of the typical Pt loadings [55,56]. The pore structure is measured by the method of standard porosimetry based on the principle of capillary equilibrium, and the EDC of the electrodes for the gas pairs of oxygen and nitrogen is determined by the modified Loschmidt cell at the temperatures of 25 C and 75 C, which are representative of the startup and design-point operation temperatures for practical PEM fuel cells [57]. This is because the PEM fuel cell is usually operated at the temperature of 70e80 C as an optimal balance between the performance and durability [19]. The dependence of the EDC upon the pore structure including porosity and mean pore size is also explored based on the experimental results. For the present study, optimal ionomer content is used for the CLs [58] such that the effect of ionomer content in the CLs is not considered. However, water saturation in the pore structure affects the pore region available for gas transport, thus influencing the cell performance. It is known that gas diffusion (or transport) through liquid water is very small, in fact, negligibly small, in comparison with the gas diffusion (or transport) through the pore region without the liquid water presence; therefore, liquid water saturation can be approximated as a region with negligible gas diffusion (or transport) [59,60]. As a result, the effect of water saturation level is not investigated in the present study.

3

Pore structure characterization The pore structure is of great importance to determine the transport properties of the porous media in PEM fuel cells. Specifically, the porosity [39,40,41,42,43,44,45] and mean pore size [8,61] are the two most important parameters. Generally, a larger porosity will lead to a bigger effective diffusion coefficient when the mean pore size is kept constant. For the same porosity, a smaller mean pore size, when it is below the continuum regime, indicates that the Knudsen diffusion cannot be ignored, which means the gas molecules collide with the pore walls more frequently. Therefore, the accurate measurement of the porosity and mean pore size is significant. The porosity can be measured directly, while the mean pore size is determined from the specific surface area and pore volume, which can be obtained from a pore size distribution curve. All the required parameters are described and defined below. The pore structure is characterized by the method of standard porosimetry (MSP), which is developed based on the law of capillary equilibrium [35,36,62] and implemented by Standard Porosimetry 3.1 (POROTECH Ltd.). The electrode samples are prepared in a disk-like shape with the diameter of 2.3 cm using a die. The tested samples and standards samples are then immersed in the N-Octane (Sigma-Aldrich Co. LLC.) under the vacuum condition so that all the pores are filled with octane. A digital balance (Sartorius ALC-210.4) is employed to record the mass changes of the samples, and a hot plate is used to accelerate the evaporation process. The weighing and drying processes are automatically repeated and accomplished. Fig. 1 represents the process of how the pore size distribution is achieved by the MSP. On the left side, curve 2 shows the relationship between the pore volume of the standard

Experimental Electrode preparation The fuel cell electrodes are prepared by spraying the catalyst ink with various ingredients and compositions on the GDLs (AvCarb GDS3250). This commercial GDL composed of one layer of carbon fibers with PTFE treatment and one microporous layer (MPL) made of carbon particles. The GDLs have the thickness of 221.6 ± 2.1 mm and porosity of 74.7 ± 1.0%. The catalyst ink is prepared by mixing proper amounts of ionomer, catalyst, and solvent. Ionomer from Nafion solution (5 wt% Nafion® PFSA) is used to bind and stabilize the catalyst particles and to provide the pathway for protons in catalyst layers. The catalyst particles (TEC10E60TPM) with a Pt/C ratio of 58.3% (nominally referred to as 60% Pt/C) are used to promote the electrochemical reactions. The solvent is 2-propanol (99.9% purity) which helps uniformly disperse the particles and ionomer. The weight ratio of the catalyst particles to ionomer is kept at 3:1, which is the optimal ionomer content that yields optimal cell performance [58]. The prepared mixture is then subjected to ultrasonic treatment for 1 h in order to make solution uniform. The prepared ink mixture is sprayed on the MPL side of the GDLs until the desired Pt loading of 0.1, 0.2, 0.3, and 0.4 mg cm2 is reached.

Fig. 1 e Principles of the method of standard porosimetry: (1) pore size distribution of the standard sample; (2) pore volume of the standard sample Vs vs. the pore volume of the test sample Vt; and (3) pore size distribution of the tested sample [35,36].


4


sample (Vs) and that in the tested sample (Vt), whereas the pore volume of the standard sample is presented as a function of log (rm) on the right-hand side. Curve 1 reveals the pore size distribution curve for the standard sample given by the manufacturer. Curve 3 indicates the cumulative pore size distribution of the tested samples. For a particular case of capillary equilibrium, the Vs and Vt determine point B on curve 2. By referring to the standard curve, point A can be found and the maximum size of pores filled with the liquid in the standard sample can be determined. Due to capillary equilibrium, the maximum pore size of the tested samples remains the same as the standards samples. Therefore, the relationship between Vt and rm in the tested samples can be established. By repeating the evaporating and weighing procedures, more points can be generated. The overall distribution curve 3 for the test sample is thus determined. The cross-sectional area, Ac, of each type of samples being tested is Ac ¼

pd2 4

(2)

Fig. 2 e Schematic of a Loschmidt cell: (1) oxygen inlet; (2) nitrogen inlet; (3) oxygen inlet; (4) gas outlet.

where d is the diameter of the disk-like sample. The bulk volume, Vbulk, can be calculated as Vbulk ¼ Ac dN

(3)

where d is the thickness of the sample, and N is the number of the samples being tested together. In the present study, d is 2.3 cm, and N equals 2. The porosity, ∅, is defined as: ∅¼

Vpore Vbulk

(4)

The pore surface area, Sp, can be calculated from the integral pore radius distribution curve by using [23,35]. Zrmax Sp ¼ 2

1 dVt dr r dr

(5)

rmin

The specific surface area, SSA, is defined by the following equation in order to make a good comparison with the different samples. SSA ¼

Sp NAc

(6)

The mean pore size, MPS, is defined as [63]. MPS ¼

4Vpore Sp

(7)

Effective diffusion coefficient The Loschmidt cell, which is used to determine the diffusion coefficient in the present study, is established based on Fick's law of diffusion [8,51,52]. Fig. 2 represents a modified Loschmidt cell which consists of two chambers separated by a sliding gate, based on [8]. The top and bottom chambers with an interior length and diameter of 42.5 cm and 3.8 cm are used to store nitrogen and oxygen gases, respectively. The nitrogen and oxygen gases can be separated or connected by the sliding gate made of a non-porous metal. Two mass flow controllers (Omega, Model FMA-5508) with a flow capacity of 0e500

(mL$min1) are used to control the flow rate of N2 and O2 during the calibration and experimental processes. An oxygen sensor (Ocean Optics FOXY-AL300) is employed to measure the oxygen concentration over time. The typical experimental procedures can be divided into four major steps as shown in Fig. 3. First, the oxygen probe is calibrated under 50% oxygen volumetric fraction condition (see Fig. 3a). O2 and N2 gases are supplied with a constant flow rate of 500 mL min1 for a prolonged period of time through valve #1 and #2, while valve #4 is open to expel the existed gas in the chambers (50% O2 point calibrated). Second, the oxygen probe is calibrated under 0% oxygen concentration condition (see Fig. 3b). Valve #1 and #3 are closed, and N2 gas is supplied through valve #2 and expelled through valve #4 for a prolonged period of time (0% O2 point calibrated). Third, the oxygen and nitrogen gases are filled in separated chambers (see Fig. 3c). In this case, the sliding gate is closed to separate the O2 and N2 gases in the bottom and top chambers, respectively. Therefore, the chambers are filled with pure O2 and N2, respectively. Finally, the sliding gate is opened and the measurement process starts. The experimental data for the oxygen concentration change over time are obtained through the oxygen sensor. More detailed experimental setup and procedure are available elsewhere [38,64]. The EDC of the electrode, Deff, can be calculated based on the resistance network theory [8,38] such that. Deff ¼

z Deq

d Dzd bulk

(8)

where Deq is the equivalent diffusion coefficient, z is the diffusion distance in the diffusion direction, and Dbulk is the bulk diffusion coefficient. The equivalent diffusion coefficient, Deq, in the combination of the bulk region and porous media is obtained by curve fitting of the experimental data to Crank's equation [8,38,65],


5


Fig. 3 e Experimental procedures for the modified Loschmidt cell.

Cðz; tÞ ¼

Cb z erfc pffiffiffiffiffiffiffiffiffiffi 2 2 Deq t

! (9)

where t is the experimental time, C means the oxygen concentration, and the subscript b denotes the bottom chamber. According to Marrero and Mason's formula, the bulk diffusion coefficients for O2eN2 gas pair can be calculated through the following formula with an accuracy of 3% [51,52], Dbulk ¼ 1:13 109

T1:724 P

(10)

where Dbulk is the bulk diffusion coefficient in m2 s1, T is the gas temperature in K, and P is the pressure in atm. The diffusion resistivity, Rd, which is defined as the diffusion resistance per unit area of the cross-section, is calculated according to Rd ¼

d Deff

(11)

The EDC tests are repeated five times under each condition, and the standard deviations are all within 4% for the GDLs with and without the CLs, and still within 8.3% for the worst case when the EDCs of the thin catalyst layers are determined.

distribution (PSD), pore surface area distribution, porosity, specific surface area (SSA), and mean pore size (MPS). The effect of the catalyst layer on the pore structure of the entire electrode (or referred to later as catalyzed GDL) is studied under a range of Pt loadings from 0.1 to 0.4 mg cm2. The pore structure of the catalyzed GDLs is compared to the uncatalyzed GDL, which is labeled as 0.0 mg cm2. Fig. 4 (a) represents the cumulative pore size distribution of the porous electrodes with five different Pt loadings of 0.0, 0.1, 0.2, 0.3, and 0.4 mg cm2. It can be observed that the pore diameter can be as large as 20 mm, and the Pt loadings have a significant influence on the small pores typically within the range of 10e200 nm. These results are consistent with the previous studies in Ref. [66]. Clearly, the high-Pt-loading electrodes have more small pores in this range. The increase of the meso (2e50 nm diameter) and macro (>50 nm diameter) pores, which are defined by the International Union of Pure and Applied Chemistry (IUPAC) [67], is likely contributed to by the coated catalyst layer, which shows consistency with Shen et al.’s work [8]. However, it should be noted that the volume of larger pores (>1 mm) decrease when a small amount of catalysts (e.g., 0.1 and 0.2 mg cm2 Pt loading) are deposited on GDLs because the catalysts can more easily occupy the large pore in GDLs. As the Pt loading is increased to 0.3 and 0.4 mg cm2, more catalysts are sprayed and accumulate on

Results and discussion Validation of Loschmidt cell Table 2 compares the measured bulk diffusion coefficient of oxygen-nitrogen gases at different temperatures against the theoretical bulk diffusion coefficient obtained by Marrero and Mason's formula (i.e., Eq. (10)). It can be seen that the relative errors between the measured and theoretical values can be as small as 3.6%. This suggests that the experimental results are at an acceptable level of accuracy.

Pore structure characterization The pore structure of the electrode is investigated by using the method of standard porosimetry (MSP) in terms of pore size

Table 2 e Bulk diffusion coefficients of oxygen gas in nitrogen gas at different temperatures measured in the present study. Temperature ( C)

20 25 30 35 45 50 55 60

Measured diffusion coefficient [105 m2 s1] 2.13 2.17 2.20 2.26 2.38 2.44 2.51 2.62

± 0.06 ± 0.06 ± 0.07 ± 0.07 ± 0.08 ± 0.08 ± 0.09 ± 0.13

Marrero and Mason's formula [52] [105 m2 s1]

Relative error (%)

2.10 2.14 2.16 2.21 2.31 2.36 2.42 2.53

1.4 1.4 1.9 2.2 3.0 3.4 3.5 3.6


6


Fig. 4 (b) presents the bulk volume, pore volume, and porosity of the five prepared samples. As the Pt loading is increased from 0.0 to 0.4 mg cm2, the bulk volume increases due to the catalyst material deposition and the resulted thicker layer. Further, increasing the Pt loadings has a minor impact on the pore volume, hence, the porosity decreases from 74.7% to 68.8% as shown in Fig. 4 (b). Furthermore, Fig. 4 (c) exhibits the cumulative surface area distribution of the pores from the maximum to minimum sizes for these five prepared electrodes. As can be seen, the surface area is greatly influenced by the small pores, e.g., 95.0%e96.5% for the pores with a diameter smaller than 200 nm. Additionally, the low Pt loading of 0.1 mg cm2 can lead to a 1.2% increase in the specific surface area, while such an increase can be magnified by 24.0% with a high Pt loading of 0.4 mg cm2. The increase in the pore surface area is mainly contributed by the presence of the small Pt/C particles, resulting in more sites for electrochemical reaction to occur. It can also be observed that even though the overall pore surface area shows a clearly increasing trend with the Pt loading, the pore surface area contributed by a certain range of pore sizes may vary greatly due to the different degrees of particle penetration into the GDL pores when various Pt loadings are applied. As the catalyst layer thickness increases, the mean pore size of the entire electrode is reduced accordingly as shown in Table 3. Since the mean pore size is defined as 4 times the ratio of pore volume to pore surface area, a higher surface area leads to a smaller mean pore size when the pore volume is almost the same as the Pt loading is increased.

Effective diffusion coefficient measurement

Fig. 4 e Pore structure of the porous electrodes with various Pt loadings: (a) Cumulative pore size distribution, (b) Porosity, and (c) Cumulative pore surface area distribution.

GDLs. It can be observed that the amount of large pores (>1 mm) increases because the catalyst layer itself also contributes to the increase of pores when the Pt loading is large enough. In other words, when catalyst inks are sprayed onto the GDLs, the pore structure can be changed by two ways: pore volume decrease due to catalyst particle penetration into the GDLs and pore volume increase due to the new pores introduced by the deposited catalyst layers. The importance of these two effects depends on the Pt loading applied and the GDL pore structures.

Fig. 5 (a) exhibits the EDC of the electrodes with a range of Pt loadings from 0.0 to 0.4 mg cm2 at the temperature of 25 and 75 C, respectively. As discussed earlier, a higher-Pt-loading can lead to a thicker catalyst layer, lower porosity, and smaller mean pore size. Thus, the EDC of the electrode with a higher-Pt-loading becomes smaller as shown in Fig. 5 (a). The corresponding diffusion resistivity equals the thickness divided by EDC. Due to the thicker electrode and its associated lower diffusion coefficient, the diffusion resistivity of the higher Pt-loading electrode becomes larger as indicated in Fig. 5 (b). It should be noted that a higher operating temperature will promote the diffusion of oxygen through the porous electrode, resulting in a lower mass transport resistance. In order to determine the EDCs of the CLs alone, the catalyzed and uncatalyzed GDLs are measured separately following the detailed experimental procedures in Section 2.3. From the EDCs of the catalyzed and uncatalyzed GDLs shown

Table 3 e Parameters of the pore structure of the uncatalyzed and catalyzed electrodes based on porosimetry results. Pt loading [mg cm2]

0.0

0.1

0.2

0.3

0.4

Electrode thickness [mm] 221.6 224.6 226.4 229.2 231.0 Specific surface area [cm2 cm2] 5394 5455 5589 6375 6701 Mean pore size [nm] 122 119 115 99 95


7


Table 4 e Effective diffusion coefficient of the catalyst layers measured in the present study. Pt loading [mg cm2]

25 C 75 C

Effective diffusion coefficient of catalyst layers [107 m2 s1] 0.1

0.2

0.3

0.4

4.9 ± 0.3 5.1 ± 0.4

4.6 ± 0.1 4.8 ± 0.4

4.4 ± 0.3 4.5 ± 0.1

2.8 ± 0.1 3.9 ± 0.1

The effective diffusion coefficients of the uncatalyzed GDL measured in this study is 20.8 107 and 25.0 107 m2 s1 at 25 and 75 C, respectively.

which means the results are more related to the practical PEM fuel cell operating conditions. It might be mentioned that the present study uses 60% Pt/C catalyst. If lower Pt/C catalyst is used, and for the same Pt loading for the electrode, more carbon and ionomer content will be present, resulting into a thicker CL [57], and also since the CL has smaller pore sizes and porosity than the uncatalyzed electrode [57], the entire catalyzed electrode will have lower porosity and lower EDC as a result, which is undesirable for the mass transport. On the other hand, the effective catalyst surface area is increased for lower Pt/C catalyst with the same Pt loading. Therefore, the final cell performance will be a balance between the reduced mass transport and increased reactive surface area [57].

Relation between pore structure and effective diffusion coefficient

Fig. 5 e Diffusion properties of the porous electrodes with different Pt loadings of 0.0, 0.1, 0.2, 0.3, and 0.4 mg cm¡2 at the temperature of 25 and 75 C: (a) Effective diffusion coefficient and (b) Diffusion resistivity.

in Fig. 5, the EDCs of the CLs can be calculated based on the resistance network theory [2] as follows: eff DCL ¼ dGDL

CL

dGDL

dGDL

CL

DGDL

CL

dGDL DGDL

1 (12)

where d is the thickness of the corresponding component. The results determined are given in Table 4. The repeatability is within 8.3% for all cases. It is seen that the EDCs of the CLs are about one order of magnitude smaller than that of the uncatalyzed GDL measured in this study, and this observation is also consistent with the previous results as shown in Shen et al.’s work using Al2O3 as the substrate [8]. The average EDCs of the catalyst layers are (4.2 ± 0.9) 107 and (4.6 ± 0.5) 107 m2$s1at 25 and 75 C, respectively. The current experimental data are slightly larger than those in Shen et al.’s work due to the different materials, composition, and fabrication methods employed, and the substrate selection might also contribute to the difference. It should be noted that this study utilizes a commercial GDL as the substrate,

Fig. 6 (a) and (b) illustrate the effect of porosity and mean pore size on the EDCs, respectively. Experimental results indicate that the EDC is directly related to the porosity. Higher porosity has a positive impact on enhancing the EDC. This trend is found to be in good agreement with the empirical correlation models shown in Table 1. Results suggest that mean pore size can also have a significant effect on the EDCs as shown in Fig. 6 (b). The EDCs of the catalyzed electrodes present a clear increasing trend with the mean pore size. When the pore size is small, the gas molecules collide more frequently with the pore surface, resulting in a larger diffusion resistance (also known as Knudsen effect). The Knudsen diffusion coefficient, which is proportional to the pore diameter [8,61], indicates that the sample with a large mean pore size has a higher capability for diffusion. The Knudsen diffusion becomes significant when the pore size is less than 1 mm. As discussed earlier, the electrode with higher Pt loadings possesses a higher pore surface area and smaller mean pore size, thus leading to a lower EDC. Fig. 7 represents the comparison of the experimental data with the different models as a function of porosity. The description of these models is given in Table 1. It can be observed that the experimental data are much smaller than the predicted values. Among these models, Zamel et al.’s model is the closest to predict the EDC of the porous electrode in PEM fuel cells. However, this model still over-estimates the EDC of the electrode by almost 53% as shown in Fig. 7. This over-prediction is likely due to the presence of the catalyst layer and microporous layer which results in a smaller mean


8


catalyst layers and catalyzed electrodes. It should be noted that Das et al.’ model is not included since the porosity of the catalyst layer only is required. This parameter cannot be measured in this study because the catalyst layers are too thin and too delicate to be separated from the GDLs.

Conclusions

Fig. 6 e Relation between the diffusion coefficient of the porous electrodes and the pore structure in terms of porosity (a) and mean pore size (b).

In this study, an experimental investigation has been carried out to determine the pore structure and effective diffusion coefficient (EDC) of the porous electrode in polymer electrolyte membrane (PEM) fuel cells. Catalyst layers made of 60% Pt/C particles with four different Pt loadings (i.e., 0.1, 0.2, 0.3, and 0.4 mg cm2) are investigated in the study. The pore structure is characterized by the method of standard porosimetry (MSP) in terms of pore size distribution (PSD), porosity, specific surface area (SSA), and mean pore size (MPS), while the EDC is measured by the modified Loschmidt cell. The results show that the presence of the catalyst layer significantly impacts the pore structure and EDC of the PEM fuel cell electrode. As the Pt loading is increased, the porosity of the electrode is decreased due to a thicker catalyst layer with much smaller porosity. The electrodes with a high Pt loading have more pores with a radius smaller than 100 nm than those with a low Pt loading, and these small pores (0e100 nm) contribute 95.0e96.5% to the overall pore surface area. Thus the pore surface area increases with the Pt loading due to the presence of more catalyst particles. The EDC decreases due to the smaller porosity and mean pore size for the higher-Pt-loading electrodes. In addition, a higher temperature enhances the EDC by 15e25% from 25 C to 75 C, which are the typical operating temperatures of PEM fuel cells. Finally, the EDC of the catalyst layer is about 4.6 107 m2 s1 at 75 C. These results highlight the importance of catalyst layers, with a focus on the impact of the Pt loading, in determining the overall mass transport in the porous electrode of PEM fuel cells.

Acknowledgement This work is financially supported by Ontario-China Research and Innovation Fund (OCRIF Round 3), the Natural Sciences and Engineering Research Council of Canada (NSERC) via a Discovery Grant, and was conducted as part of the Catalysis Research for Polymer Electrolyte Fuel Cells (CaRPE-FC) Network administered from Simon Fraser University and supported by Automotive Partnership Canada (APC) Grant No. APCPJ 417858-11 through the Natural Sciences and Engineering Research Council of Canada (NSERC).

references Fig. 7 e Comparison of the experimental data with the models shown in Table 1. pore size of the catalyzed electrode. The gas species experience higher diffusion resistances in smaller pores, and thus the EDC is lower for the porous media with smaller pore sizes. This indicates that Knudsen effect has to be considered for

[1] Li X. Principles of fuel cells. New York: Taylor & Francis Group; 2006. [2] Pant LM, Mitra SK, Secanell M. Absolute permeability and Knudsen diffusivity measurements in PEMFC gas diffusion layers and micro porous layers. J Power Sources 2012;206:153e60.



[3] Zhao J, Shahgaldi S, Alaefour I, Xu Q, Li X. Gas permeability of catalyzed electrodes in polymer electrolyte membrane fuel cells. Appl Energy 2018;209:203e10. https://doi.org/10.1016/ j.apenergy.2017.10.087. [4] Zhang X, Gao Y, Ostadi H, Jiang K, Chen R. Modelling water intrusion and oxygen diffusion in a reconstructed microporous layer of PEM fuel cells. Int J Hydrogen Energy 2014;39:17222e30. https://doi.org/10.1016/ j.ijhydene.2014.08.027. [5] Andisheh-Tadbir M, El Hannach M, Kjeang E, Bahrami M. An analytical relationship for calculating the effective diffusivity of micro-porous layers. Int J Hydrogen Energy 2015;40:10242e50. https://doi.org/10.1016/ j.ijhydene.2015.06.067. [6] Ismail MS, Ingham DB, Hughes KJ, Ma L, Pourkashanian M. Effective diffusivity of polymer electrolyte fuel cell gas diffusion layers: an overview and numerical study. Int J Hydrogen Energy 2015;40:10994e1010. https://doi.org/ 10.1016/j.ijhydene.2015.06.073. [7] Zamel N, Li X. Effective transport properties for polymer electrolyte membrane fuel cells e with a focus on the gas diffusion layer. Prog Energy Combust Sci 2013;39:111e46. [8] Shen J, Zhou J, Astrath NGC, Navessin T, Liu ZS, Lei C, et al. Measurement of effective gas diffusion coefficients of catalyst layers of PEM fuel cells with a Loschmidt diffusion cell. J Power Sources 2011;196:674e8. [9] Gostick JT, Fowler MW, Pritzker MD, Ioannidis MA, Behra LM. In-plane and through-plane gas permeability of carbon fiber electrode backing layers. J Power Sources 2006;162:228e38. [10] Sohn Y-J, Yim S-D, Park G-G, Kim M, Cha S-W, Kim K. PEMFC modeling based on characterization of effective diffusivity in simulated cathode catalyst layer. Int J Hydrogen Energy 2017;42:13226e33. https://doi.org/10.1016/ j.ijhydene.2017.04.036. [11] Salomov UR, Chiavazzo E, Fasano M, Asinari P. Pore- and macro-scale simulations of high temperature proton exchange fuel cells e HTPEMFC e and possible strategies for enhancing durability. Int J Hydrogen Energy 2017;42:26730e43. https://doi.org/10.1016/ j.ijhydene.2017.09.011. [12] Salomov UR, Chiavazzo E, Asinari P. Pore-scale modeling of fluid flow through gas diffusion and catalyst layers for high temperature proton exchange membrane (HT-PEM) fuel cells. Comput Math Appl 2014;67:393e411. https://doi.org/ 10.1016/j.camwa.2013.08.006. [13] Shojaeefard MH, Molaeimanesh GR, Nazemian M, Moqaddari MR. A review on microstructure reconstruction of PEM fuel cells porous electrodes for pore scale simulation. Int J Hydrogen Energy 2016;41:20276e93. https://doi.org/10.1016/ j.ijhydene.2016.08.179. [14] Molaeimanesh GR, Saeidi Googarchin H, Qasemian Moqaddam A. Lattice Boltzmann simulation of proton exchange membrane fuel cells e a review on opportunities and challenges. Int J Hydrogen Energy 2016;41:22221e45. https://doi.org/10.1016/j.ijhydene.2016.09.211. [15] Ye L, Gao Y, Zhu S, Zheng JJP, Li P, Zheng JJP. A Pt content and pore structure gradient distributed catalyst layer to improve the PEMFC performance. Int J Hydrogen Energy 2017;42:7241e5. https://doi.org/10.1016/ j.ijhydene.2016.11.002. cio RN, Neto AO, Linardi M. Influence of the relative [16] Bonifa volumes between catalyst and Nafion ionomer in the catalyst layer efficiency. Int J Hydrogen Energy 2014;39:14680e9. https://doi.org/10.1016/ j.ijhydene.2014.07.004. [17] Suzuki T, Tanaka H, Hayase M, Tsushima S, Hirai S. Investigation of porous structure formation of catalyst layers

[18]

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26] [27] [28]

[29] [30]

[31] [32]

[33]

[34]

[35]

[36]

[37]

9

for proton exchange membrane fuel cells and their effect on cell performance. Int J Hydrogen Energy 2016;41:20326e35. Inoue G, Kawase M. Understanding formation mechanism of heterogeneous porous structure of catalyst layer in polymer electrolyte fuel cell. Int J Hydrogen Energy 2016;41:21352e65. https://doi.org/10.1016/j.ijhydene.2016.08.029. Wu J, Yuan XZ, Martin JJ, Wang H, Zhang J, Shen J, et al. A review of PEM fuel cell durability: degradation mechanisms and mitigation strategies. J Power Sources 2008;184:104e19. https://doi.org/10.1016/j.jpowsour.2008.06.006. Suzuki A, Hattori T, Miura R, Tsuboi H, Hatakeyama N, Takaba H, et al. Porosity and Pt content in the catalyst layer of PEMFC: effects on diffusion and polarization characteristics. Int J Electrochem Sci 2010;5:1948e61. Vol’fkovich YM, Sosenkin VE, Nikol’skaya NF, Kulova TL. Porous structure and hydrophilic-hydrophobic properties of gas diffusion layers of the electrodes in proton-exchange membrane fuel cells. Russ J Electrochem 2008;44:278e85. https://doi.org/10.1134/S102319350803004X. Roohparvarzadeh S. Experimental characterization of the compressive behaviour of gas diffusion layers in PEM fuel cells. 2014. Volfkovich YM, Sosenkin VE, Bagotsky VS. Structural and wetting properties of fuel cell components. J Power Sources 2010;195:5429e41. https://doi.org/10.1016/ j.jpowsour.2010.03.002. Oh H, Park J, Min K, Lee E, Jyoung JY. Effects of pore size gradient in the substrate of a gas diffusion layer on the performance of a proton exchange membrane fuel cell. Appl Energy 2015;149:186e93. https://doi.org/10.1016/ j.apenergy.2015.03.072. Yoon YG, Park GG, Yang TH, Han JN, Lee WY, Kim CS. Effect of pore structure of catalyst layer in a PEMFC on its performance. Int J Hydrogen Energy 2003;28:657e62. https:// doi.org/10.1016/S0360-3199(02)00156-8. Drake LC. Pore-size distribution in porous materials. Ind Eng Chem 1949;41:780e5. Giesche H. Mercury porosimetry: a general (practical) overview. Part Part Syst Char 2006;23:9e19. Yan Z, Chen C, Fan P, Wang M. Pore structure characterization of ten typical rocks in China. Electron J Geotech Eng 2015;20:479e94. Gregg SJ, Sing KSW, Salzberg HW. Adsorption surface area and porosity. J Electrochem Soc 1967;114. 279Ce279C. Neimark AV, Ravikovitch PI. Capillary condensation in MMS and pore structure characterization. Microporous Mesoporous Mater 2001;44:697e707. Dubinin MM, Plavnik GM. Microporous structures of carbonaceous adsorbents. Carbon N Y 1968;6:183e92. Lee Y, Jeong J, Youn IJ, Lee WH. Modified liquid displacement method for determination of pore size distribution in porous membranes. J Memb Sci 1997;130:149e56. Abell A, Willis K, Lange D. Mercury intrusion porosimetry and image analysis of cement-based materials. J Colloid Interface Sci 1999;211:39e44. Miller B, Tyomkin I. An extended range liquid extrusion method for determining pore size distributions. Textil Res J 1986;56:35e40. Volfkovich YM, Sakars AV, Volinsky AA. Application of the standard porosimetry method for nanomaterials. Int J Nanotechnol 2005;2:292e302. Volfkovich YM, Bagotzky VS. The method of standard porosimetry. 1. Principles and possibilities. J Power Sources 1994;48:327e38. Volfkovich YM, Bagotzky VS. The method of standard porosimetry 2. Investigation of the formation of porous structures. J Power Sources 1994;48:339e48.


10


[38] Unsworth G, Dong L, Li X. Improved experimental method for measuring gas diffusivity through thin porous media. Am Inst Chem Eng J 2013;59:1409e19. [39] Bruggeman DAG. Berechnung verschiedener physikalischer Konstanten von heterogenen Substanzen. III. Die elastischen Konstanten der quasiisotropen Mischkorper aus isotropen substanzen. Ann Phys 1937;421:160e78. [40] Mintani M. Geometric factor for diffusion in porous media. J Chem Eng Jpn 1984;17:441e3. [41] Neale GH, Nader WK. Prediction of transport processes within porous media: diffusive flow processes within an homogeneous swarm of spherical particles. Am Inst Chem Eng J 1973;19:112e9. [42] Tomadakis MM, Sotirchos SV. Ordinary and transition regime diffusion in random fiber structures. Am Inst Chem Eng J 1993;39:397e412. [43] Mezedur MM, Kaviany M, Moore W. Effect of pore structure, randomness and size on effective mass diffusivity. Am Inst Chem Eng J 2002;48:15e24. [44] Zamel N, Li X, Shen J. Correlation for the effective gas diffusion coefficient in carbon paper diffusion media. Energy and Fuels 2009;23:6070e8. [45] Das PK, Li X, Liu ZS. Effective transport coefficients in PEM fuel cell catalyst and gas diffusion layers: beyond Bruggeman approximation. Appl Energy 2010;87:2785e96. [46] Baker DR, Wieser C, Neyerlin KC, Murphy MW. The use of limiting current to determine transport resistance in PEM fuel cells. ECS Trans 2006;3:989e99. [47] Lamanna JM, Kandlikar SG. Determination of effective water vapor diffusion coefficient in pemfc gas diffusion layers. Int J Hydrogen Energy 2011;36:5021e9. https://doi.org/10.1016/ j.ijhydene.2011.01.036. [48] Casalegno A, Colombo L, Galbiati S, Marchesi R. Quantitative characterization of water transport and flooding in the diffusion layers of polymer electrolyte fuel cells. J Power Sources 2010;195:4143e8. https://doi.org/10.1016/ j.jpowsour.2010.01.032. [49] Kramer D, Freunberger SA, Flu¨ckiger R, Schneider IA, Wokaun A, Bu¨chi FN, et al. Electrochemical diffusimetry of fuel cell gas diffusion layers. J Electroanal Chem 2008;612:63e77. https://doi.org/10.1016/ j.jelechem.2007.09.014. [50] Flu¨ckiger R, Freunberger SA, Kramer D, Wokaun A, Scherer GG, Bu¨chi FN. Anisotropic, effective diffusivity of porous gas diffusion layer materials for PEFC. Electrochim Acta 2008;54:551e9. https://doi.org/10.1016/ j.electacta.2008.07.034. [51] Chan C. Experimental measurement of effective diffusion coefficient of gas diffusion layer/microporous layer in PEM fuel cells. Electrochim Acta 2011;65:13e21. [52] Marrero TR, Mason EA. Gaseous diffusion coefficients. J Phys Chem Ref Data 1972;1. 3e118.

[53] Huan TN, Jane RT, Benayad A, Guetaz L, Tran PD, Artero V. Bio-inspired noble metal-free nanomaterials approaching platinum performances for H 2 evolution and uptake. Energy Environ Sci 2016;9:940e7. https://doi.org/10.1039/C5EE02739J. [54] Najafabadi AT, Leeuwner MJ, Wilkinson DP, Gyenge EL. Electrochemically produced graphene for microporous layers in fuel cells. ChemSusChem 2016;9:1689e97. https://doi.org/ 10.1002/cssc.201600351. [55] He C, Desai S, Brown G, Bollepalli S. PEM fuel cell catalysts: cost, performance, and durability. Electrochem Soc Interface 2005:41e4. [56] Wee J-H, Lee K-Y, Kim SH. Fabrication methods for low-Ptloading electrocatalysts in proton exchange membrane fuel cell systems. J Power Sources 2007;165:667e77. https:// doi.org/10.1016/j.jpowsour.2006.12.051. [57] Shahgaldi S, Zhao J, Alaefour I, Li X. Investigation of catalytic vs reactant transport effect of catalyst layers on proton exchange membrane fuel cell performance. Fuel 2017;208:321e8. https://doi.org/10.1016/j.fuel.2017.07.035. [58] Shahgaldi S, Alaefour I, Unsworth G, Li X. Development of a low temperature decal transfer method for the fabrication of proton exchange membrane fuel cells. Int J Hydrogen Energy 2017;42:11813e22. https://doi.org/10.1016/ j.ijhydene.2017.02.127. [59] Baschuk J, Li X. Modelling of polymer electrolyte membrane fuel cells with variable degrees of water flooding. J Power Sources 2000;86:181e96. https://doi.org/10.1016/S03787753(99)00426-7. [60] Zamel N, Li X, Becker J, Wiegmann A. Effect of liquid water on transport properties of the gas diffusion layer of polymer electrolyte membrane fuel cells. Int J Hydrogen Energy 2011;36:5466e78. https://doi.org/10.1016/ j.ijhydene.2011.01.146. [61] Cunningham RE, Williams RJJ. Diffusion in gases and porous media. 1980. https://doi.org/10.1007/978-1-4757-4983-0. [62] Morrow NR, Harris CC. Capillary equilibrium in porous materials. Soc Petrol Eng J 1965;5:15e24. [63] Dullien FAL. Porous media: fluid transport and pore structure. New York: Academic Press; 1979. [64] Dong L. Accuracy improvement for measurement of gas diffusivity through thin porous media. 2012. [65] Crank J. The mathematics of diffusion. New York: Oxford University Press; 1979. [66] Lamanna JM, Bothe JV, Zhang FY, Mench MM. Measurement of capillary pressure in fuel cell diffusion media, microporous layers, catalyst layers, and interfaces. J Power Sources 2014;271:180e6. https://doi.org/10.1016/ j.jpowsour.2014.07.163. [67] Rouquerol J, Avnir D, Fairbridge CW, Everett DH, Haynes JH, Pernicone N, et al. Recommendations for the characterization of porous solid. Pure Appl Chem 1994;66:1739e58.
