JP7527740B2 - Membrane Electrode Assembly - Google Patents

Description

The present invention relates to a membrane electrode assembly, and more specifically, to a membrane electrode assembly that has excellent durability against power generation in a sub-freezing environment (sub-freezing start-up).

The basic unit of a polymer electrolyte fuel cell is a membrane electrode assembly (MEA) in which electrodes (catalyst layers) are bonded to both sides of a solid polymer electrolyte membrane. In a polymer electrolyte fuel cell, a gas diffusion layer is generally placed on the outside of the catalyst layer. The gas diffusion layer is used to supply reactive gas and electrons to the catalyst layer, and is made of carbon paper, carbon cloth, or the like. The catalyst layer is the reaction field for the electrode reaction, and is generally made of a composite of carbon carrying an electrode catalyst such as platinum and a solid polymer electrolyte (catalyst layer ionomer). Furthermore, a separator with a gas flow path is placed on the outside of the gas diffusion layer.

The types and contents of materials constituting an MEA affect the performance of the MEA, and therefore, various proposals have been made regarding the relationship between the materials constituting an MEA and the performance of the MEA.
For example, Patent Document 1 discloses a fuel cell in which the ratio A/B of the mass A of a proton conductor (a fluoropolymer) contained in a catalyst layer to the mass B of a conductive support contained in the catalyst layer is controlled within a predetermined range.

The same document states:
(A) When the water remaining in the cell freezes, the catalyst layer may be damaged by stress caused by ice crystal growth. To prevent this, it is necessary to increase the mechanical strength of the catalyst layer.
(B) Increasing the mass ratio (I/C) of ionomer to carbon in the catalyst layer increases the strength of the catalyst layer, but if the I/C is increased more than necessary, the performance of the fuel cell may decrease; and
(C) By determining the ratio A/B so that the strength of the catalyst layer is equal to or greater than a predetermined value, and by determining the ratio A/B so that the voltage value is equal to or greater than a predetermined value, it is possible to improve both durability and performance.
is stated.

In Patent Document 2,
(a) forming a catalyst layer on the surface of a diffusion layer by applying a slurry containing a catalyst and an ion-conductive polymer to the surface of the diffusion layer;
(b) placing the coated surface of the slurry (i.e., the catalyst layer) on the electrolyte membrane;
(c) A method for manufacturing a fuel cell electrode assembly is disclosed in which a laminate of an electrolyte membrane, a catalyst layer and a diffusion layer is heated under pressure.

The same document states:
It is described that (A) this method makes it possible to obtain an electrode structure for a fuel cell in which a portion of the catalyst particles penetrates a predetermined distance into the electrolyte membrane, and (B) the electrode structure thus obtained has a composition that changes continuously at the interface between the electrolyte membrane and the catalyst layer, so that peeling does not occur at the interface between the electrolyte membrane and the catalyst layer and durability against temperature cycles is improved.

It is known that the performance of fuel cells deteriorates when they are repeatedly started below freezing (see Patent Documents 1 and 2). The deterioration in performance during start-up below freezing is believed to be mainly due to peeling of the cathode catalyst layer near the electrolyte membrane. Therefore, in order to improve durability during start-up below freezing, it is important to improve the strength of the interface between the electrolyte membrane and the catalyst layer (hereinafter, this is also referred to as "interface adhesion strength").

In this regard, Patent Document 2 describes that by allowing part of the catalyst particles to penetrate into the electrolyte membrane, durability against temperature cycles is improved. However, the fracture strength that is believed to occur in the frost heave phenomenon (the phenomenon in which ice is generated and grows) is very large, and durability cannot be sufficiently improved simply by increasing the interfacial strength by allowing part of the catalyst particles (and carbon support) to penetrate into the electrolyte membrane.

JP 2009-259664 A JP 2002-075382 A

The problem that this invention aims to solve is to provide a membrane electrode assembly that has excellent durability against repeated subzero start-up.

In order to solve the above problems, the membrane electrode assembly according to the present invention has the following configuration.
(1) The membrane electrode assembly is
An electrolyte membrane made of a resin (A) having proton conductivity;
a cathode catalyst layer bonded to one surface of the electrolyte membrane;
and an anode catalyst layer bonded to the other surface of the electrolyte membrane.
(2) The cathode catalyst layer comprises:
An electrode catalyst (B) in which catalyst particles (B) are supported on a conductive support (B);
and a resin (B) having proton conductivity.
(3) The membrane electrode assembly has a catalyst layer resistance change rate of 0.33 or less per cycle.
(4) The cathode catalyst layer has a void volume per unit area of 0.15 μL/cm ² or more and 0.75 μL/cm ² or less.

The cathode catalyst layer is made of a composite of the electrode catalyst (B) and the resin (B), and voids exist in the cathode catalyst layer. If the amount of voids is relatively small, a large force is applied to the cathode catalyst layer when ice is formed in the cathode catalyst layer, causing frost heave destruction of the cathode catalyst layer.
On the other hand, the amount of voids in the cathode catalyst layer mainly depends on the structure of the conductive carrier (B) and the manufacturing conditions of the cathode catalyst layer. By optimizing these, a relatively large amount of voids can be introduced into the cathode catalyst layer. As a result, even if ice forms in the cathode catalyst layer, the force applied to the cathode catalyst layer is reduced, and frost heave damage of the cathode catalyst layer is suppressed.

1A is a schematic diagram of a catalyst layer with a small amount of voids, and FIG 1B is a schematic diagram of a catalyst layer with a large amount of voids. Fig. 2(A) is an image of a cross section of a catalyst layer. Fig. 2(B) is an image in which the threshold gradation of the image of Fig. 2(A) is increased until the areas considered to be voids turn black. Fig. 2(C) is an image in which the threshold gradation of the image of Fig. 2(B) is decreased to a value just before the areas considered to be carbon cross sections turn black. Fig. 2(D) is a binarized image in which the white areas remaining in the voids of the image of Fig. 2(C) are filled in black. FIG. 13 is a graph showing the relationship between the void volume and the rate of change in resistance of a catalyst layer. FIG. 1 is a diagram showing the relationship between tap density and void volume.

An embodiment of the present invention will be described in detail below.
[1. Membrane electrode assembly]
The membrane electrode assembly according to the present invention has the following configuration.
(1) The membrane electrode assembly is
An electrolyte membrane made of a resin (A) having proton conductivity;
a cathode catalyst layer bonded to one surface of the electrolyte membrane;
and an anode catalyst layer bonded to the other surface of the electrolyte membrane.
(2) The cathode catalyst layer comprises:
An electrode catalyst (B) in which catalyst particles (B) are supported on a conductive support (B);
and a resin (B) having proton conductivity.
(3) The membrane electrode assembly has a catalyst layer resistance change rate of 0.33 or less per cycle.
(4) The cathode catalyst layer has a void volume per unit area of 0.15 μL/cm ² or more and 0.75 μL/cm ² or less.

[1.1. Electrolyte membrane]
The electrolyte membrane is made of a proton-conductive resin (A). The material of the resin (A) is not particularly limited, and an optimum material can be selected depending on the purpose.
Examples of the resin (A) include:
(a) perfluorocarbon sulfonic acid polymers such as Nafion®, Flemion®, Aciplex®, and Aquivion®;
(b) Hydrocarbon polymers such as sulfonated polyether ketones, sulfonated polyether sulfones, sulfonated polyether ether sulfones, sulfonated polysulfides, and sulfonated polyphenylenes;
(c) a highly oxygen-permeable ionomer, as described below;
and so on.

[1.2. Cathode catalyst layer]
A cathode catalyst layer is bonded to one surface of the electrolyte membrane. In the present invention, the cathode catalyst layer is
An electrode catalyst (B) in which catalyst particles (B) are supported on a conductive support (B);
and a resin (B) having proton conductivity.

[1.2.1. Electrode catalyst (B)]
The electrode catalyst (B) is composed of a conductive support (B) carrying catalyst particles (B).

[A. Conductive support (B)]
A.1. Materials
In the present invention, the material of the conductive support (B) is not particularly limited.
Examples of the conductive support (B) include
(a) Carbon supports such as carbon black, carbon nanotubes, carbon nanohorns, activated carbon, natural graphite, mesocarbon microbeads, and glassy carbon powder;
(b) a metal support made of a pure metal or alloy;
(c) an oxide support made of a metal oxide or composite metal oxide having electrical conductivity;
and so on.
The shape of the conductive support (B) is not particularly limited, and an optimal shape can be selected depending on the purpose.

[A.2. Tap density]
The "tap density" refers to a value measured in accordance with JIS Z 2512.
The conductive support (B) usually has a primary particle diameter of 10 nm to 300 nm. This is to support the catalyst particles (B) on the surface of the conductive support (B) in a highly dispersed manner. In the cathode catalyst layer, these fine primary particles usually do not exist alone, but exist as secondary particles formed by aggregation of the primary particles. The degree of aggregation of the primary particles affects the void volume of the cathode catalyst layer.

The degree of aggregation of the primary particles can be expressed by tap density. In general, the smaller the tap density of the conductive support (B), the more voids can be introduced into the cathode catalyst layer. In order to ensure the amount of voids described later, the conductive support (B) preferably has a tap density of 0.2 g/ ^cm3 or less. The tap density is preferably 0.15 g/ ^cm3 or less.

However, if the tap density is too small, the thickness of the cathode catalyst layer required to make the weight of the catalyst particles (B) a certain value or more becomes large, and the proton resistance of the cathode catalyst layer increases. Therefore, the tap density is preferably 0.03 g/cm ³ or more. The tap density is preferably 0.05 g/cm ³ or more, and more preferably 0.08 g/cm ³ or more.

[B. Catalyst particles (B)]
In the present invention, the material of the catalyst particles (B) is not particularly limited.
Examples of the material for the catalyst particles (B) include:
(a) Precious metals (Pt, Au, Ag, Pd, Rh, Ir, Ru, Os),
(b) an alloy containing two or more precious metal elements;
(c) Alloys containing one or more precious metal elements and one or more base metal elements (e.g., Fe, Co, Ni, Cr, V, Ti, etc.);
and so on.

Among these, the catalyst particles (B) are preferably made of Pt or a Pt alloy, since they have high activity in the electrode reaction of the fuel cell.
Examples of Pt alloys include Pt--Fe alloys, Pt--Co alloys, Pt--Ni alloys, Pt--Pd alloys, Pt--Cr alloys, Pt--V alloys, Pt--Ti alloys, Pt--Ru alloys, and Pt--Ir alloys.

The particle size of the catalyst particles (B) is not particularly limited, and an optimal particle size can be selected according to the purpose. In general, if the particle size of the catalyst particles (B) is too small, the catalyst particles (B) tend to dissolve. Therefore, the particle size of the catalyst particles (B) is preferably 1 nm or more.
On the other hand, if the particle size of the catalyst particles (B) is too large, the mass activity decreases. Therefore, the particle size of the catalyst particles (B) is preferably 20 nm or less. The particle size of the catalyst particles (B) is preferably 10 nm or less, and more preferably 5 nm or less.

[1.2.2. Catalyst layer ionomer (B)]
The catalyst layer ionomer (B) contained in the cathode catalyst layer is made of a resin (B) having proton conductivity.

A. Materials
In the present invention, the material of the resin (B) is not particularly limited. The resin (B) may be any of a perfluorocarbon sulfonic acid polymer, a hydrocarbon polymer, and a highly oxygen-permeable ionomer. In order to reduce the oxygen transfer resistance in the cathode catalyst layer, the resin (B) preferably contains a highly oxygen-permeable ionomer.

The resin (B) may be composed solely of a highly oxygen-permeable ionomer, which has the advantage that oxygen can be easily transferred to the catalyst particles (B), facilitating the catalytic reaction, thereby improving performance.
Alternatively, the resin (B) may contain, in addition to the high oxygen permeability ionomer, another ionomer (e.g., a perfluorocarbon sulfonic acid polymer), which has the advantage that cracks are less likely to occur in the catalyst layer, thereby increasing the productivity of the catalyst layer.
However, if the content of the high oxygen-permeable ionomer is too small, the performance may be deteriorated. Therefore, the mass ratio of the high oxygen-permeable ionomer in the resin (B) is preferably 70 mass% or more. The mass ratio of the high oxygen-permeable ionomer is preferably 80 mass% or more, and more preferably 90 mass% or more.

Here, the term "highly oxygen-permeable ionomer" refers to a polymeric compound that contains an acid group and a cyclic structure in its molecular structure. The highly oxygen-permeable ionomer has a high oxygen permeability coefficient because it contains a cyclic structure in its molecular structure. Therefore, when this ionomer is used as an ionomer, the oxygen transfer resistance at the interface with the catalyst becomes relatively small.
In other words, the "highly oxygen-permeable ionomer" refers to an ionomer having an oxygen permeability coefficient higher than that of perfluorocarbon sulfonic acid polymers such as Nafion (registered trademark).

Generally, the performance of a fuel cell is determined by the diffusion of oxygen to the catalyst surface. By contrast, coating the surface of an electrode catalyst with a highly oxygen-permeable ionomer improves the oxygen permeability of the catalyst layer, thereby improving the performance of the fuel cell.
The molecular structure of the highly oxygen-permeable ionomer is not particularly limited as long as it exhibits a relatively small oxygen transfer resistance. In particular, an ionomer containing a cyclic structure (aliphatic cyclic structure) in its molecular structure has a smaller oxygen transfer resistance than an ionomer not containing a cyclic structure, and is therefore suitable as an ionomer for coating the surface of an electrode catalyst.

Examples of highly oxygen-permeable ionomers include:
(a) an electrolyte polymer including a perfluorocarbon unit having an aliphatic ring structure and an acid group unit having a perfluorosulfonic acid in a side chain;
(b) an electrolyte polymer including a perfluorocarbon unit having an aliphatic ring structure and an acid group unit having a perfluoroimide in a side chain;
(c) an electrolyte polymer containing a unit in which perfluorosulfonic acid is directly bonded to a perfluorocarbon having an aliphatic ring structure;
(See References 1 to 4).
[Reference 1] JP 2003-036856 A [Reference 2] WO 2012/088166 A [Reference 3] JP 2013-216811 A [Reference 4] JP 2006-152249 A

[B.E.W.]
The EW of the resin (B) affects the interfacial adhesion strength, durability (rate of change in resistance of the catalyst layer), proton conductivity and/or gas diffusivity of the cathode catalyst layer, and the like.
In general, if the EW of the resin (B) is too low, the water content becomes high and the resin (B) becomes easily dissolved in water. Therefore, the EW is preferably 588 g/eq or more.
On the other hand, if the EW of the resin (B) is too high, the proton conductivity of the cathode catalyst layer decreases, so the EW is preferably 830 g/eq or less.

C. Gelation of the ionomer
Ionomer molecules generally have a main chain and side chains attached to the main chain, with acid groups attached to the ends of the side chains. When such ionomer molecules are dispersed in water, they form a structure (micelle structure) in which the outside is mainly composed of hydrophilic side chains and the inside is mainly composed of hydrophobic main chains.

When an ionomer with such a micellar structure is used as is to prepare a dispersion, the micellar structure remains in the dispersion. Dispersions containing micellar structures generally have low viscosity. When a cathode catalyst layer is prepared using such a catalyst ionomer, a relatively high interfacial adhesion strength cannot be obtained. This is thought to be because there is little molecular entanglement at the interface between the electrolyte membrane and the cathode catalyst layer.

On the other hand, when preparing a dispersion using an ionomer with a micellar structure, if the ionomer is subjected to appropriate treatment, the ionomer will not form a micellar structure in the dispersion, but will be uniformly dispersed (this state is also called "gelation").

A dispersion containing a gelled ionomer generally has a high viscosity. When a cathode catalyst layer is produced using such a catalyst layer ionomer, a relatively high interfacial adhesion strength is obtained. This is considered to be due to increased molecular entanglement at the interface between the electrolyte membrane and the cathode catalyst layer. In order to obtain a high interfacial adhesion strength, the cathode catalyst layer is preferably produced using the gelled resin (B).
The degree of gelation can be expressed by the storage modulus. In order to obtain a catalyst layer having high interfacial adhesion strength, the resin (B) preferably has a storage modulus of 40 [Pa] or more and 140 [Pa] or less when a vibration having a frequency of 1 [Hz] is applied and the distortion amount is 1 [%].

1.2.3. Average I/C
The "average I/C value" of the cathode catalyst layer refers to the ratio of the total mass of the resin (B) to the total mass of the conductive carrier (B) contained in the cathode catalyst layer.
The average value of I/C of the cathode catalyst layer affects the interface adhesion strength, durability (rate of change in catalyst layer resistance), proton conductivity and/or gas diffusivity of the cathode catalyst layer.
Generally, if the average value of I/C of the cathode catalyst layer is too low, the proton conductivity of the cathode catalyst layer decreases. Therefore, the average value of I/C is preferably 0.80 or more. The average value of I/C is more preferably 0.95 or more.
On the other hand, if the average value of I/C of the cathode catalyst layer becomes too high, the amount of voids in the cathode catalyst layer decreases, and the gas diffusion property of the cathode catalyst layer decreases. Therefore, the average value of I/C is preferably 1.8 or less. I/C is preferably 1.5 or less.

[1.2.4. Amount of voids]
In the present invention, the "void volume per unit area (P)" refers to a value represented by the following formula (1).
P (μL/cm ² )=S×t/S ₀ …(1)
however,
S ₀ is the area of the field of view when the cross section of the cathode catalyst layer is observed under a microscope,
S is the area of the void included in the field of view,
t is the thickness of the cathode catalyst layer.

The void volume per unit area of the cathode catalyst layer (hereinafter, simply referred to as "void volume") affects durability against sub-freezing start-up. If the void volume is too small, a large force is applied to the cathode catalyst layer when ice is generated in the cathode catalyst layer, which may cause frost heave destruction. Therefore, the void volume must be 0.15 μL/ ^cm2 or more. The void volume is preferably 0.20 μL/ ^cm2 or more, and more preferably 0.25 μL/ ^cm2 or more.
On the other hand, if the void volume is excessive, the proton conductivity of the catalyst layer may decrease. Therefore, the void volume needs to be 0.75 μL/ ^cm2 or less. The void volume is preferably 0.70 μL/ ^cm2 or less, and more preferably 0.50 μL/ ^cm2 or less.

The void volume depends not only on the microstructure of the conductive support (B) but also on the manufacturing conditions of the cathode catalyst layer. For example, when manufacturing a catalyst ink by dispersing an electrode catalyst (B) having a sufficient void volume in a solvent, if an excessive dispersion process is performed, the microstructure of the electrode catalyst (B) may be destroyed during the dispersion process. If such a catalyst ink is used to form a cathode catalyst layer, the required void volume may not be obtained.
Therefore, when producing the cathode catalyst layer, it is preferable to select the material of the conductive support (B) and the production conditions of the cathode catalyst layer so that the necessary void volume can be secured.

[1.3. Anode catalyst layer]
An anode catalyst layer is bonded to the other surface of the electrolyte membrane.
An electrode catalyst (C) in which catalyst particles (C) are supported on a conductive support (C);
and a resin (C) having proton conductivity.

[1.3.1. Electrode catalyst (C)]
The electrode catalyst (C) is composed of a conductive support (C) and catalyst particles (C) supported thereon. In the present invention, the electrode catalyst (C) is not particularly limited, and may be selected according to the purpose. The conductive carrier (C) contained in the anode catalyst layer may be the same as the conductive carrier (B) contained in the cathode catalyst layer, or may be different. Similarly, the catalyst particles (C) contained in the anode catalyst layer may be the same as or different from the catalyst particles (B) contained in the cathode catalyst layer.
Other points regarding the conductive support (C) and the catalyst particles (C) are similar to those of the conductive support (B) and the catalyst particles (B), respectively, and therefore description thereof will be omitted.

[1.3.2. Catalyst layer ionomer (C)]
The catalyst layer ionomer (C) contained in the anode catalyst layer is made of a resin (C) having proton conductivity. In the present invention, the catalyst layer ionomer (C) is not particularly limited, and an optimal one can be selected depending on the purpose. The resin (C) contained in the anode catalyst layer may be the same as or different from the resin (A) constituting the electrolyte membrane or the resin (B) contained in the cathode catalyst layer.
Other points regarding resin (C) are similar to those of resin (A) or resin (B), and therefore description thereof will be omitted.

1.3.3. Average I/C
The "average I/C value" of the anode catalyst layer refers to the ratio of the total mass of the resin (C) to the total mass of the conductive carrier (C) contained in the anode catalyst layer.
The average value of I/C of the anode catalyst layer affects the performance of the membrane electrode assembly. In general, if the average value of I/C is too small, the proton conductivity of the anode catalyst layer decreases. Therefore, the average value of I/C is preferably 0.95 or more.
On the other hand, if the average value of I/C becomes too large, the porosity of the anode catalyst layer decreases, and the gas diffusivity of the anode catalyst layer decreases. Therefore, the average value of I/C is preferably 1.8 or less. The average value of I/C is preferably 1.5 or less.

[1.4. Catalyst layer resistance change rate]
The "catalyst layer resistance change rate" refers to a value expressed by the following formula (1).
Catalyst layer resistance change rate (/time)=(R−R ₀ )/(n×R ₀ ) (1)
however,
n is the number of cycles of the sub-zero start-up acceleration durability test,
R ₀ is the catalyst layer resistance before the sub-zero start-up acceleration durability test,
R is the catalyst layer resistance after n cycles of sub-zero accelerated starting durability testing.

In the present invention, the "subzero start-up acceleration durability test" refers to
(a) humidifying both the cathode/anode electrodes of a cell containing an MEA at 82° C. with a dew point of 40° C.;
(b) After humidification, the cell is sealed and cooled to −20° C.;
(c) supplying unhumidified hydrogen gas/air to the frozen cell and generating electricity at 0.3 A/ ^cm2 for 1000 s; and
(d) After power generation is completed, the cell is sealed again and the temperature and humidity are increased, which constitutes one cycle. This test is repeated a total of n cycles (n=12 in the present invention).

When the void volume of the cathode catalyst layer, the tap density of the conductive carrier (B), the EW of the catalyst layer ionomer (B), the average I/C value of the cathode catalyst layer, etc. are optimized, the catalyst layer resistance change rate becomes 0.33/cycle or less. Even when a highly oxygen-permeable ionomer is used as the catalyst layer ionomer (B), when the void volume of the cathode catalyst layer is optimized, the catalyst layer resistance change rate becomes 0.33/cycle or less. This is thought to be because the introduction of an appropriate amount of voids into the cathode catalyst layer alleviates the volume expansion associated with freezing due to the voids.
The "catalyst layer resistance" used in calculating the catalyst layer resistance change rate refers to the catalyst layer proton resistance calculated from the real part of the catalyst layer impedance measured (80° C., 30% RH).

[3. Action]
Fig. 1(A) shows a schematic diagram of a catalyst layer with a small amount of voids. Fig. 1(B) shows a schematic diagram of a catalyst layer with a large amount of voids. The cathode catalyst layer is made of a composite of an electrode catalyst (B) and a resin (B), and voids exist in the cathode catalyst layer. If the amount of voids is relatively small, as shown in Fig. 1(A), when ice is formed in the cathode catalyst layer, a large force is applied to the cathode catalyst layer, causing frost heave destruction of the cathode catalyst layer.
In contrast, the amount of voids in the cathode catalyst layer mainly depends on the structure of the conductive support (B) and the manufacturing conditions of the cathode catalyst layer. By optimizing these, a relatively large amount of voids can be introduced into the cathode catalyst layer, as shown in Figure 1 (B). As a result, even if ice forms in the cathode catalyst layer, the force applied to the cathode catalyst layer is reduced, and frost heave damage to the cathode catalyst layer is suppressed.

(Examples 1 to 7, Comparative Examples 1 to 2)
1. Preparation of Samples
[1.1. Preparation of cathode catalyst layer]
Six types of Pt/C with different carbon supports were used as electrode catalysts. Table 1 shows the specifications of the carbon supports used.
As the ionomer, a highly oxygen-permeable ionomer having an EW of 650 g/eq was used.

Water was added to 1 g of Pt/C and stirred with a spatula. An ionomer solution was added to the water and dispersed for 5 minutes with an ultrasonic homogenizer. Ethanol was further added to the dispersion and dispersed for 5 to 15 minutes with an ultrasonic homogenizer.
The mixture was then centrifuged for 3 minutes to remove bubbles, yielding a catalyst ink. The resulting catalyst ink was applied to a polytetrafluoroethylene (PTFE) sheet with an applicator so that the Pt basis weight was 0.2 mg/ ^cm2 , and the sheet was dried to obtain a catalyst layer.

[1.2. Preparation of fuel cell]
The above-mentioned cathode catalyst layer was transferred onto one surface of the electrolyte membrane, and the anode catalyst layer was transferred onto the other surface. The electrode area was ¹ cm2.
The anode catalyst layer was a composite of Pt/C (Pt loading: 30 mass%) and Nafion (registered trademark). The I/C of the anode catalyst layer was 1.0 in all cases. The Pt weight was 0.1 mg/ ^cm2 in all cases.

2. Test Method
2.1. Porosity
Each catalyst layer was subjected to cross-sectional processing using an Ar ion beam, and then observed under a SEM with an accelerating voltage of 2 kV and a working distance of about 3 mm, and images were obtained from 3 to 5 fields of view at magnifications of 30,000 to 50,000 times.
The acquired image was binarized using Image J. First, the acquired image was converted to a monochrome image with 256 gradations, and the brightness/contrast was automatically adjusted using the automatic brightness/contrast adjustment (Auto) function of Image J. Next, noise was removed from the image using a function called Non-local Means Denoising, which was added as a plugin function. The image that had been subjected to these image processes was binarized.

Figure 2(A) shows an image of the cross section of the catalyst layer. Figure 2(B) shows an image in which the threshold gradation of the image in Figure 2(A) has been increased until the areas that can be considered to be voids turn black. Figure 2(C) shows an image in which the threshold gradation of the image in Figure 2(B) has been lowered to the value just before the areas that can be considered to be carbon cross sections turn black. Figure 2(D) shows a binarized image in which the white areas remaining in the voids of the image in Figure 2(C) have been filled in black.

The threshold for binarization was determined as follows, using the threshold function of Image J. That is, for the image shown in Figure 2(A), when the threshold gradation was increased until the areas considered to be voids turned black, a part of the area considered to be the carbon cross section also turned black, as shown in Figure 2(B). Therefore, the gradation just before the areas considered to be the carbon cross section turned black was set as the threshold (Figure 2(C)). After that, the white areas remaining in the voids were painted black using the Paintbrush tool of Image J. The resulting image (Figure 2(D)) was used as the binarized image.

From the obtained binarized image, the area ratio of the void portion (ratio of void area (S) to field area ( _S0) ) was calculated. In addition, the thickness (t) of the catalyst layer was measured from the SEM image of the catalyst layer cross section. Furthermore, the void volume was calculated using formula (1).

[2.2. Sub-zero starting acceleration durability test]
Both the cathode and anode electrodes of the fuel cell were humidified at 82°C with a dew point of 40°C. The small cell was sealed and then cooled to -20°C. Then, while feeding unhumidified hydrogen gas/air to the fuel cell, electricity was generated at 0.3 A/ ^cm2 for 1000 s. After the power generation was completed, the fuel cell was sealed again and the temperature was raised and humidified. The cycle up to this point was counted as one cycle, and a total of 12 cycles were repeated. The power generation performance was measured before the durability test and at the 12th cycle. The catalyst layer impedance was measured at 80°C and 30% RH, and the catalyst layer resistance change rate (the rate of change in catalyst layer resistance per one durability cycle) was calculated from the real part of the catalyst layer impedance.

3. Results
The results are shown in Table 2. The specifications of the cathode catalyst layer are also shown in Table 2. Figure 3 shows the relationship between the void volume and the catalyst layer resistance change rate. Figure 4 shows the relationship between the tap density and the void volume. The following can be seen from Table 2 and Figures 3 and 4.

(1) It was found that the greater the void volume, the higher the durability against sub-zero startability. It was also found that in order to achieve a catalyst layer resistance change rate of 0.33 (/cycle) or less, the void volume needs to be 0.15 μL/ ^cm2 or more.
(2) A correlation was found between tap density and void volume, and it was found that the lower the tap density of a carbon support, the greater the void volume.
(3) In Comparative Example 2, the void volume was smaller and the catalyst layer resistance change rate was larger than in Example 6. This is considered to be because the total dispersion time in Comparative Example 2 was too long, causing the aggregate structure of Carbon F to break during dispersion, and further causing the carbon particles to be crushed.

Although the embodiment of the present invention has been described in detail above, the present invention is not limited to the above embodiment, and various modifications are possible without departing from the gist of the present invention.

The membrane electrode assembly of the present invention can be used in fuel cells used in vehicle power sources, small stationary power generators, etc.

Claims (5)

A membrane electrode assembly having the following configuration.
(1) The membrane electrode assembly is
An electrolyte membrane made of a resin (A) having proton conductivity;
a cathode catalyst layer bonded to one surface of the electrolyte membrane;
and an anode catalyst layer bonded to the other surface of the electrolyte membrane.
(2) The cathode catalyst layer comprises:
An electrode catalyst (B) in which catalyst particles (B) are supported on a conductive support (B);
and a resin (B) having proton conductivity.
(3) The membrane electrode assembly has a catalyst layer resistance change rate of 0.33 or less per cycle.
(4) The cathode catalyst layer has a void volume per unit area of 0.15 μL/cm ² or more and 0.75 μL/cm ² or less. The membrane electrode assembly according to claim 1, wherein the resin (B) contains a highly oxygen-permeable ionomer. 3. The membrane/electrode assembly according to claim 1, wherein the conductive support (B) has a tap density of 0.2 g/cm ^<3> or less. The membrane electrode assembly according to any one of claims 1 to 3, wherein the resin (B) has an EW of 588 g/eq or more and 830 g/eq or less. The membrane electrode assembly according to any one of claims 1 to 4, wherein the cathode catalyst layer has an average I/C value of 0.80 or more and 1.8 or less.

Images