SUPPORTING CARBON MATERIAL FOR SOLID POLYMER FUEL CELL AND CATALYST METAL PARTICLE-SUPPORTING CARBON MATERIAL

Abstract

Provided are: a supporting carbon material for a solid polymer fuel cell, said supporting carbon material making it possible to produce a high-performance solid polymer fuel cell in which there is little decrease in power generation performance as a result of repeated battery load fluctuation that inevitably occurs during operation of the solid polymer fuel cell; and a catalyst metal particle-supporting carbon material. The present invention relates to: a supporting carbon material for a solid polymer fuel cell, said supporting carbon material being a porous carbon material in which the specific surface area of mesopores having a pore diameter of 2-50 nm according to nitrogen adsorption measurement is 600-1,600 m.sup.2/g, the relative intensity ratio (IG/IG) of the peak intensity (IG) of the G-band 2,650-2,700 cm.sup.1 range to the peak intensity (IG) of the G-band 1,550-1,650 cm.sup.1 range in the Raman spectrum is 0.8-2.2, and the peak position of the G-band is 2,660-2,670 cm.sup.1; and a catalyst metal particle-supporting carbon material.

Claims

1. A supporting carbon material for a solid polymer fuel cell comprised of a porous carbon material having a specific surface area S.sub.A of mesopores of a pore size of 2 to 50 nm, found by analyzing a nitrogen adsorption isotherm of an adsorption process by the Dollimore-Heal method, of 600 m.sup.2/g to 1600 m.sup.2/g, having a relative intensity ratio (IG/IG) of a peak intensity (IG) of a peak present in the G-band of 2650 to 2700 cm.sup.1 in range at a Raman spectrum and a peak intensity (IG) of a peak present in the G-band of 1550 to 1650 cm.sup.1 of 0.8 to 2.2, and having a peak position of the G-band of 2660 to 2670 cm.sup.1.

2. The supporting carbon material for a solid polymer fuel cell according to claim 2 wherein in said mesopores, mesopores with a pore size of 2 nm to less than 10 nm have a specific pore area S.sub.2-10 of 400 m.sup.2/g to 1100 m.sup.2/g, mesopores with a pore size of 2 nm to less than 10 nm have a specific pore volume V.sub.2-10 of 0.4 cc/g to 1.6 cc/g, and mesopores with a pore size of 10 nm to 50 nm have a specific pore area S.sub.10-50 of 20 m.sup.2/g to 150 m.sup.2/g, and micropores of a pore size of less than 2 nm, found by analyzing a nitrogen adsorption isotherm of an adsorption process by the Horvath-Kawazoe method, have a pore area S.sub.2 of 250 m.sup.2/g to 550 m.sup.2/g.

3. The supporting carbon material for a solid polymer fuel cell according to claim 2 wherein said specific pore area S.sub.2-10 is 400 m.sup.2/g to 1000 m.sup.2/g, said specific pore volume V.sub.2-10 is 0.4 cc/g to 1.4 cc/g, said specific pore area S.sub.10-50 is 30 m.sup.2/g to 100 m.sup.2/g, and said specific pore area S.sub.2 is 300 m.sup.2/g to 500 m.sup.2/g.

4. The supporting carbon material for a solid polymer fuel cell according to claim 1 wherein a specific surface area S.sub.BET by the BET method is 600 m.sup.2/g to 1500 m.sup.2/g, the DBP oil absorption X is 200 cm.sup.3/100 g to 650 cm.sup.3/100 g, a ratio (X/S.sub.BET) of said DBP oil absorption X and said specific surface area S.sub.BET is 3 nm to 5 nm, and a half width of a peak present in the range of 1550 to 1650 cm.sup.1 called the G-band is 30 cm.sup.1 to 75 cm.sup.1.

5. The supporting carbon material for a solid polymer fuel cell according to claim 4 wherein said specific surface area S.sub.BET is 800 m.sup.2/g to 1300 m.sup.2/g, said DBP oil absorption X is 300 cm.sup.3/100 g to 550 cm.sup.3/100 g, the ratio (X/S.sub.BET) of said DBP oil absorption X and said specific surface area S.sub.BET is 3 nm to 5 nm, and said half width is 50 cm.sup.1 to 70 cm.sup.1.

6. The supporting carbon material for a solid polymer fuel cell according to claim 1 wherein said specific surface area S.sub.A is 700 m.sup.2/g to 1400 m.sup.2/g and a relative intensity ratio (IG/IG) of the G-band and G-band is 1.0 to 2.0.

7. The supporting carbon material for a solid polymer fuel cell according to claim 1 wherein said specific surface area S.sub.A is 700 m.sup.2/g to 1400 m.sup.2/g and a peak position of the G-band is 2662 to 2668 cm.sup.1.

8. The supporting carbon material for a solid polymer fuel cell according to claim 1 wherein said specific surface area S.sub.A is 700 m.sup.2/g to 1400 m.sup.2/g, a relative intensity ratio (IG/IG) of the G-band and G-band is 1.0 to 2.0, and a peak position of the G-band is 2662 to 2668 cm.sup.1.

9. A catalyst metal particle-supporting carbon material for a solid polymer fuel cell comprised of a supporting carbon material for a solid polymer fuel cell according to claim 1 on which platinum alone or mainly platinum catalyst metal particles are supported.

10. The supporting carbon material for a solid polymer fuel cell according to claim 2 wherein a specific surface area S.sub.BET by the BET method is 600 m.sup.2/g to 1500 m.sup.2/g, the DBP oil absorption X is 200 cm.sup.3/100 g to 650 cm.sup.3/100 g, a ratio (X/S.sub.BET) of said DBP oil absorption X and said specific surface area S.sub.BET is 3 nm to 5 nm, and a half width of a peak present in the range of 1550 to 1650 cm.sup.1 called the G-band is 30 cm.sup.1 to 75 cm.sup.1.

11. The supporting carbon material for a solid polymer fuel cell according to claim 3 wherein a specific surface area S.sub.BET by the BET method is 600 m.sup.2/g to 1500 m.sup.2/g, the DBP oil absorption X is 200 cm.sup.3/100 g to 650 cm.sup.3/100 g, a ratio (X/S.sub.BET) of said DBP oil absorption X and said specific surface area S.sub.BET is 3 nm to 5 nm, and a half width of a peak present in the range of 1550 to 1650 cm.sup.1 called the G-band is 30 cm.sup.1 to 75 cm.sup.1.

12. The supporting carbon material for a solid polymer fuel cell according to claim 2 wherein said specific surface area S.sub.A is 700 m.sup.2/g to 1400 m.sup.2/g and a relative intensity ratio (IG/IG) of the G-band and G-band is 1.0 to 2.0.

13. The supporting carbon material for a solid polymer fuel cell according to claim 3 wherein said specific surface area S.sub.A is 700 m.sup.2/g to 1400 m.sup.2/g and a relative intensity ratio (IG/IG) of the G-band and G-band is 1.0 to 2.0.

14. The supporting carbon material for a solid polymer fuel cell according to claim 4 wherein said specific surface area S.sub.A is 700 m.sup.2/g to 1400 m.sup.2/g and a relative intensity ratio (IG/IG) of the G-band and G-band is 1.0 to 2.0.

15. The supporting carbon material for a solid polymer fuel cell according to claim 5 wherein said specific surface area S.sub.A is 700 m.sup.2/g to 1400 m.sup.2/g and a relative intensity ratio (IG/IG) of the G-band and G-band is 1.0 to 2.0.

16. The supporting carbon material for a solid polymer fuel cell according to claim 2 wherein said specific surface area S.sub.A is 700 m.sup.2/g to 1400 m.sup.2/g and a peak position of the G-band is 2662 to 2668 cm.sup.1.

17. The supporting carbon material for a solid polymer fuel cell according to claim 3 wherein said specific surface area S.sub.A is 700 m.sup.2/g to 1400 m.sup.2/g and a peak position of the G-band is 2662 to 2668 cm.sup.1.

18. The supporting carbon material for a solid polymer fuel cell according to claim 4 wherein said specific surface area S.sub.A is 700 m.sup.2/g to 1400 m.sup.2/g and a peak position of the G-band is 2662 to 2668 cm.sup.1.

19. The supporting carbon material for a solid polymer fuel cell according to claim 5 wherein said specific surface area S.sub.A is 700 m.sup.2/g to 1400 m.sup.2/g and a peak position of the G-band is 2662 to 2668 cm.sup.1.

20. The supporting carbon material for a solid polymer fuel cell according to claim 2 wherein said specific surface area S.sub.A is 700 m.sup.2/g to 1400 m.sup.2/g, a relative intensity ratio (IG/IG) of the G-band and G-band is 1.0 to 2.0, and a peak position of the G-band is 2662 to 2668 cm.sup.1.

Description

BRIEF DESCRIPTION OF DRAWINGS

[0079] FIG. 1 are explanatory views for explaining the concept of the outside surface area used as an indicator in the present invention in comparison with the surface area, wherein FIG. 1A shows the surface area while FIG. 1B shows the outside surface area.

[0080] FIG. 2 are schematic views for explaining the tree-branch shaped network in the supporting carbon material of the present invention, wherein FIG. 2A shows the spaces formed by single carbon particles themselves while FIG. 2B shows other spaces formed between carbon particles.

DESCRIPTION OF EMBODIMENTS

1. Supporting Carbon Material for Solid Polymer Fuel Cell

1-1. Regarding Conditions (a), (b), and (c)

[0081] In the supporting carbon material of the present invention, the specific surface area S.sub.A of the mesopores with a pore diameter, determined by measurement of the nitrogen adsorption, of 2 to 50 nm is 600 m.sup.2/g to 1600 m.sup.2/g, preferably 700 m.sup.2/g to 1400 m.sup.2/g. Micropores do not function to support nanosize catalyst metal particles, so what is important as a support is the area of the walls forming the mesopores with a pore diameter of 2 to 50 nm. Therefore, in the supporting carbon material of the present invention, the specific surface area S.sub.A of mesopores calculated by measurement of the adsorption/desorption of nitrogen gas is made the indicator. If the specific surface area S.sub.A of the mesopores is less than 600 m.sup.2/g, it is not possible to support the amount of catalyst metal particles required for fuel cell power generation and the power generation performance falls. Conversely, if over 1600 m.sup.2/g, the average pore size becomes smaller, the size of carbon graphene forming the pore walls becomes smaller, and, as a result, the edge carbon increases and the desired oxidative consumption resistance cannot be obtained.

[0082] Note that, in the present invention, the specific surface area S.sub.A of mesopores is calculated by analyzing the nitrogen adsorption isotherm of the adsorption process by the Dollimore-Heal method. In general, for analysis of the pore structure, an adsorption isotherm of nitrogen gas at the liquid nitrogen temperature is used (Maruzen, Feb. 25, 2001, Seiichi Kondo et al., The Science of Adsorption). For analysis of the pore structure of in particular mesopores (pores of diameter 2 to 50 nm) among the pore structures, the Dollimore-Heal method, BJH method, and CI method are known. For the porous carbon material using a metal acetylide as a starting material in the present invention, from the process of production of the pores, it is suitable to envision a cylindrical shape pore structure. Therefore, the inventors used the Dollimore-Heal method to analyze the adsorption isotherm. Here, the program used for the analysis was the software attached to the BELSORP mini made by Bel Japan.

[0083] Further, in the supporting carbon material of the present invention, the peak intensity (IG) in the range of the G-band of 2650 to 2700 cm.sup.1 in the Raman spectrum and the peak intensity (IG) in the range of the G-band of 1550 to 1650 cm.sup.1 is 0.8 to 2.2 and the peak position of the G-band is 2660 to 2670 cm.sup.1. Preferably, the relative intensity ratio (IG/IG) of the G-band and G-band is 1.0 to 2.0, and the peak position of the G-band is 2662 to 2668 cm.sup.1. Note that, the peak intensities (IG and IG) are amounts corresponding to the areas of the peaks. If the relative intensity ratio (IG/IG) of the G-band and G-band is smaller than 0.8, the edge carbon becomes too sparse, the support performance of the catalyst metal particles fall, the catalyst metal particles are desorbed, the particles agglomerate with each other etc., and as a result, the power generation performance and the durability fall thereby making this material unsuitable for practical use. Conversely, if the relative intensity ratio (IG/IG) is larger than 2.2, the edge carbon becomes too numerous and, while the support performance is excellent, the oxidative consumption becomes remarkable and the durability drops thereby also making this unsuitable for practical use. Further, regarding the peak position of the G-band, the case when smaller than 2650 cm.sup.1 corresponds to a single layer of carbon graphene. The pore walls forming the skeleton of the material are weak in strength and the pores easily collapse due to the pulverization step or other external action thereby making impossible to achieve the required specific surface area. Conversely, if larger than 2700 cm.sup.1, the number of stacked layers becomes 5 layers or more. In this case as well, the required specific surface area can no longer be achieved. Note that, in the Raman spectrum of the carbon material, the spectrum shape (peak position, relative strength) changes depending on the excitation light energy, but in the present invention, 532 nm excitation light was used.

[0084] Here, the size of the graphene forming the pore walls and the number of stacked layers of graphene are prescribed by the peak of the range of 2650 to 2700 cm.sup.1 called the G-band in the Raman spectrum and the peak of the range of 1550 to 1650 cm.sup.1 called the G-band. By experimental and theoretical studies, it is proven that the peak position of the G-band reflects the number of stacked layers of graphene (M. A. Pimenta et. al, Physical Chemistry Chemical Physics, vol. 9, 2007, pages 1276-1291). Further, in the present invention, the peak position of the G-band corresponding to the number of stacked layers of graphene is 2650 to 2700 cm.sup.1, preferably 2662 to 2668 cm.sup.1. Further, the relative intensity ratio (IG/IG) of the intensity of the G-band (IG) in the Raman spectrum and the intensity of the G-band (IG) is an indicator correlated with the amount of edge carbon. The G-band corresponds to the vibration mode of the graphene. The intensity is believed to correspond to the relative mass ratio of graphene contained in a carbon material. Further, in the G-band, the intensity is strongly correlated with the number of stacked layers. The smaller the number of stacked layers, the stronger the intensity. Further, the larger the number of stacked layers, the smaller the intensity. If the number of stacked layers is constant, the intensity of the G-band depends on the graphene size. If the graphene size is large, IG becomes larger, while if the graphene size is small, IG becomes smaller.

1-2. Regarding Conditions (d) and (e)

[0085] In the catalyst supporting carbon material of the present invention, the mesopores of a pore size of 2 to 50 nm are controlled. Specifically, the specific pore area S.sub.2-10 and specific pore volume V.sub.2-10 of pores with a pore size of 2 nm to less than 10 nm and the specific pore area S.sub.10-50 of pores with a pore size of 10 nm to 50 nm are in specific ranges. Furthermore, for micropores of a pore size of less than 2 nm, the particle size of the platinum generally used as a catalyst metal is 1.0 to 6.0 nm. These platinum particles have difficulty entering into the pores, so the magnitude of their amount does not have much of an effect from the viewpoint of supporting the platinum. However, if the amount of micropores is large, since micropores have the property of not easily discharging water once holding it, it becomes difficult to discharge the produced water from the catalyst layer at the time of large current discharge. This causes flooding. Therefore, the specific pore area S.sub.2 of the micropores with a pore size of less than 2 nm is also kept within a specific range.

[0086] Further, among the mesopores as well, it is believed that an efficient reduction reaction occurs on the catalyst metal particles supported in mesopores with a pore size of 2 nm to less than 10 nm even in fuel cell operating conditions at the time of large current discharge.

[0087] In the catalyst supporting carbon material of the present invention, the specific pore area S.sub.2-10 of pores with a pore size of 2 nm to less than 10 nm is 400 m.sup.2/g to 1100 m.sup.2/g, preferably 400 m.sup.2/g to 1000 m.sup.2/g. If the specific pore area S.sub.2-10 of the mesopores is 400 m.sup.2/g to 1100 m.sup.2/g, the output performance at the time of large current discharge becomes good. If the specific pore area S.sub.2-10 is less than 400 m.sup.2/g, due to the reasons shown below, it is guessed that the catalyst metal particles can no longer be supported in the mesopores at a high density. Here, a large current indicates a current density of at least 1000 mA/cm.sup.2 and, furthermore, a current density of 1500 mA/cm.sup.2 or more. In order for the platinum particles to efficiently play a role in the catalyst reaction, the reaction gas has to equally diffuse to the platinum particles in the pores. If the platinum particles are closely supported, the adjoining platinum particles will compete over the diffused reaction gas molecules. This is not preferable. On the other hand, improving the diffusion of gas by making the catalyst layer thinner, that is, raising the supporting density of platinum particles, is demanded. To obtain the minimum support density of platinum particles required for taking out a large current in this way, the lower limit of the essential mesopore specific pore area S.sub.2-10 should be 400 m.sup.2/g. Conversely, if this specific pore area S.sub.2-10 exceeds 1100 m.sup.2/g, since the inside of the support only has a limited volume, if the mesopore area becomes larger, the average size of the mesopores will correspondingly become smaller. In this regard, since platinum particles are supported in the pores, to enable gas to easily diffuse through the spaces between the platinum particles and the pores, the pore size is preferably large. The upper limit of balance of the pore size and area is believed to be 1100 m.sup.2/g.

[0088] Further, in the catalyst supporting carbon material of the present invention, the specific pore volume V.sub.2-10 of pores with a pore size of 2 nm to less than 10 nm is 0.4 cc/g to 1.6 cc/g, preferably 0.4 cc/g to 1.4 cc/g. If this specific pore volume V.sub.2-10 is less than 0.4 cc/g, the volume becomes small with respect to the pore area, so the average pore size becomes smaller. When supporting platinum particles in the pores, the spaces between the pores and the platinum particles become smaller, so gas diffusion falls and the large current characteristic ends up falling. Conversely, if the specific pore volume V.sub.2-10 exceeds 1.6 cc/g, the skeleton used as the carbon support ends up becoming thinner, the oxidative consumption resistance falls, the stirring required when preparing the catalyst layer ink for preparing the catalyst layer causes the skeleton of this carbon support to easily break, and the characteristics derived from the shape can no longer be exhibited.

[0089] Furthermore, in the catalyst supporting carbon material of the present invention, the specific pore area S.sub.10-50 of pores with a pore size of 10 nm to 50 nm is 20 m.sup.2/g to 150 m.sup.2/g pore size, preferably 40 m.sup.2/g to 100 m.sup.2/g. If this specific pore area S.sub.10-50 is less than 20 m.sup.2/g, the skeleton size of the carbon material (thickness and length of branches of tree-branch shaped structure) becomes too thick, the number of points of contact between the catalyst-supporting carbon material in the catalyst layer becomes smaller, a drop in the electron conduction efficiency is invited, and the fuel cell performance is made to decline. If this specific pore area S.sub.10-50 exceeds 150 m.sup.2/g, the skeleton size of the carbon material becomes too fine, the spaces between the catalyst particles become smaller, and the gas diffusion ability of the catalyst becomes poor.

[0090] Furthermore, in the catalyst supporting carbon material of the present invention, the specific pore area S.sub.2 of pores with a pore size of less than 2 nm is 250 m.sup.2/g to 550 m.sup.2/g, more preferably 300 m.sup.2/g to 500 m.sup.2/g. If the specific pore area S.sub.2 exceeds 550 m.sup.2/g, it becomes difficult to discharge the produced water from inside the catalyst layer at the time of large current discharge and flooding is caused. Further, if the specific pore area S.sub.2 is not less than 250 m.sup.2/g, the amount of moisture stored in the micropores decreases, so the water holding and water affinity of the carbon material drop sharply or is lost, so it becomes difficult to store the moisture required for operation of the fuel cell in the catalyst layer and the output voltage falls.

[0091] 1-3. Regarding Conditions (f), (g), and (h)

[0092] In the supporting carbon material of the present invention, the specific value of the DBP oil absorption is 200 cm.sup.3/100 g or more, preferably 300 cm.sup.3/100 g or more. If satisfying such a value, excellent electrode characteristics are exhibited. If this DBP oil absorption is less than 200 cm.sup.3/100 g, the spaces such as shown in FIG. 2 become insufficient, the paths for the gas become smaller, the speed of gas diffusion cannot keep up with the electrode reaction, and as a result the gas diffusion resistance causes the output voltage to fall. Further, if this DBP oil absorption exceeds 1000 cm.sup.3/100 g, the bulk density of the support becomes too low and the thickness of the catalyst layer formed for supporting the basis weight amount (g/m.sup.2) of catalyst required for imparting a predetermined power generation ability ends up becoming too great. That is, even if forming the same thicknesses of catalyst layers, the supported amount of catalyst metal particles forming the reaction sites decreases, so the catalyst layer has to be made thicker. Further, if the thickness of the catalyst layer becomes greater in this way, the diffusion distance of the gas ends up becoming longer. As a result, the gas diffusion resistance in the catalyst layer increases and the output voltage ends up falling.

[0093] Note that, in the supporting carbon material of the present invention, to make the later explained BET specific surface area and the half width G of the range of 1550 to 1650 cm.sup.1 called the G-band, obtained from the Raman spectrum, suitable ranges, the range of DBP oil absorption which can be substantially utilized is 200 cm.sup.3/100 g to 650 cm.sup.3/100 g.

[0094] On the other hand, in the case of a material with a large BET specific surface area such as the supporting carbon material of the present invention, since the pores running from the particle surfaces to the insides contribute to absorption of liquid, sometimes a DBP oil absorption somewhat larger than the DBP oil absorption corresponding to the holding volume between true particles is observed corresponding to the pore volume. Note that, here, the BET specific surface area is the value of the specific surface area found by the BET method from measurement of an adsorption isotherm at the liquid nitrogen temperature. If this value is large, it shows that a large number of nanometer (nm) pores are present in the carbon material. That is, the DBP oil absorption, as a primary approximation, becomes a general indicator with respect to the spaces between particles, but is observed to be greater than the amount of DBP supply corresponding to the holding volume between true particles. Therefore, in the present invention, in addition to this DBP oil absorption, the ratio (X/S.sub.BET) between the DBP oil absorption X (cm.sup.3/100 g) and the BET specific surface area S.sub.BET (m.sup.2/g) was found and prescribed.

[0095] That is, in a porous carbon material having a high tree-branch shaped network, the DBP oil absorption X becomes greater compared with the BET specific surface area S.sub.BET, while conversely in a material having a low tree-branch shaped network, the DBP oil absorption X becomes smaller compared with the BET specific surface area S.sub.BET. Therefore, by prescribing this ratio (X/S.sub.BET), it becomes possible to judge to what extent the effect of the BET specific surface area S.sub.BET is felt. Specifically, in a porous carbon material like in the present invention which has a relative large BET specific surface area S.sub.BET of 600 m.sup.2/g or more, the ratio (X/S.sub.BET) of the DBP oil absorption X (cm.sup.3/100 g) and BET specific surface area S.sub.BET (m.sup.2/g) is 3 nm to 5 nm. If this ratio (X/S.sub.BET) exceeds 5 nm, in the porous carbon material, the ratio of pores which cannot form passageways for gas becomes larger and the value of the apparent DBP oil absorption X tends to become higher. However, the spaces shown in FIG. 2 do not greatly change. Further, if less than 3 nm, formation of spaces between the branches cannot be expected, the gas diffusion paths become insufficient, and a sufficient gas diffusion ability cannot be secure.

[0096] Further, if the BET specific surface area is 600 m.sup.2/g or more, preferably 800 m.sup.2/g or more, catalyst metal particles of several nm size are supported in a well dispersed state, that is, a state where the distance between catalyst metal particles is held at a constant value or more and the particles can remain present in an independent state. Conversely, if this BET specific surface area is less than 600 m.sup.2/g, the distance between catalyst particles becomes shorter and locations arise where part of the catalyst metal particles are supported contacted in form. As a result, the effective area of the catalyst metal particles fall and the fuel cell characteristics greatly end up falling. Note that, there is no theoretical upper limit this BET specific surface area, but the BET specific surface area of the porous carbon material substantially able to be actually utilized in the present invention is 1500 m.sup.2/g or less. In this way, if using a porous carbon material with a BET specific surface area of 600 m.sup.2/g or more as a support and making this support the catalyst metal particles, the catalyst metal particles become supported on the supporting carbon material by a certain distance and the spatial districution of the catalyst metal particles in the catalyst layer becomes sparse. As a result, the speed of consumption of oxygen per unit volume in the catalyst layer becomes lower and it becomes possible to suppress a voltage drop due to gas diffusion.

[0097] Furthermore, to improve the durability in the environment of use of the fuel cell, the crystallinity of the supporting carbon material is raised. The crystallinity of the supporting carbon material can be evaluated by the half width (G) in the range of 1550 to 1650 cm.sup.1, called the G-band, obtained from the Raman spectrum. Specifically, the smaller this G-band half width (G), the higher the crystallinity of the supporting carbon material. Therefore, in the present invention, the G-band half width (G) is controlled to 75 cm.sup.1 or less. If the G-band half width (G) is 75 cm.sup.1 or less, the durability can be secured in the environment of fuel cell usage. In a carbon material with this G-band half width (G) exceeding 75 cm.sup.1, the crystallinity is low, there are many end faces of sheets forming starting points of oxidative consumption, and practical use cannot be withstood. Note that, the lower limit of the G value is not particularly set, but the G value of the carbon material which can be substantially utilized in the present invention is 30 cm.sup.1 or more.

2. Method of Production of Supporting Carbon Material for Solid Polymer Fuel Cell

[0098] For the method of producing the supporting carbon material for a solid polymer fuel cell of the present invention meeting the above-mentioned conditions (a) to (c), preferably the following method may be mentioned.

[0099] That is, this is a method comprising an acetylide producing step of blowing acetylene gas into a solution containing a metal or metal salt to cause the production of a metal acetylide, a first heat treatment step of heating the metal acetylide to 40 to 80 C. in temperature to prepare a metal particle-containing intermediate containing metal particles, a second heat treatment step of press-forming the metal particle-containing intermediate, heating the obtained shaped articles by a speed of temperature rise of 100 C. per minute or more until 400 C. or more to make the metal particle-containing intermediate eject metal particles to obtain the carbon material intermediate, a washing step of bringing the carbon material intermediate obtained by the second heat treatment step into contact with hot concentrated nitric acid or concentrated sulfuric acid to clean the carbon material intermediate, and a third heat treatment step of heating the carbon material integrident washed in the washing step in vacuo or in an inert gas at an atmosphere of air 1400 to 2100 C. to obtain a supporting carbon material.

[0100] In the method of production of this supporting carbon material, the acetylide generating step, the first heat treatment step, and the washing step may for example be similar to the steps in the method described in PLT 2.

[0101] As opposed to this, in the second heat treatment step of heating the metal particle-containing intermediate obtained at the first heat treatment step to make this metal particle-containing intermediate eject metal particles to obtain a carbon material intermediate, it is necessary to first press-form the metal particle-containing intermediate, then heat the obtained shaped articles by a speed of temperature rise of 100 C. per minute or more until 400 C. or more.

[0102] Here, when press-forming the metal particle-containing intermediate to form the shaped articles, it is preferable to use as high a pressure as possible for the treatment, but the pressing operation itself becomes a trigger causing an explosion, so usually the press-forming is performed at 0.1 kg/cm.sup.2 to 10 kg/cm.sup.2, preferably 0.5 kg/cm.sup.2 to 5 kg/cm.sup.2. Further, the size of the shaped articles (pellets) at the time of press-forming is also basically not particularly limited, but to avoid the danger of explosion, the intermediate should be press-formed into relatively small pellets. Usually, it is press-formed into pellets of a weight of 0.1 g to 10 g or so. If the weight of the pellets which are press-formed is less than 0.1 g, in the second heat treatment step, the heat energy when heating the metal particle-containing intermediate press-formed into pellets to make it eject metal particles is small. Due to the heat at this time, it becomes difficult to decrease the amount of edge carbon. This is not suited to the present invention.

[0103] Further, in the second heat treatment step, the shaped articles of the metal particle-containing intermediate obtained by press-forming are preferably heated by a 100 C./min or more, preferably 300 C./min or more, speed of temperature rise until 400 C. or more, preferably 400 C. to 600 C., in temperature and made to explode all at once. Due to this, it is possible to realize the desired growth of graphene and decrease of edge carbon in the obtained carbon material intermediate. If the speed of temperature rise at this time is 100 C./min or less or if the heating temperature is less than 400 C., the decrease of the edge carbon becomes insufficient. As a result, it no longer becomes possible to suppress the oxidative consumption of the supporting carbon material due to the fluctuation of potential at the time of fuel cell operation and the durability falls. Further, if the heating temperature exceeds 600 C., the 3D tree-branch shaped structure of the metal acetylide is destroyed, the gas diffusion ability of the catalyst layer due to the tree-branch shaped structure falls, and as a result the power generation output is liable to fall.

[0104] Further, in the third heat treatment step of heating the carbon material intermediate cleaned in the above washing step in a vacuum or in an inert gas atmosphere to obtain a supporting carbon material, the intermediate has to be heated at 1400 C. to 2100 C., preferably 1500 C. to 2100 C. in temperature. In this third heat treatment step, the edge carbon is decreased, but it is necessary that the changes in structure be ones where the pore structure is maintained, that is, where there is no large movement of graphene, increase in stacked layers, etc. If the above heating temperature is lower than 1400 C., the edge carbon decreases and the oxidative consumption resistance cannot be improved. Conversely, if over 2100 C., the texture of the graphene changes, the number of stacked layers increases, the pores are crushed, the amount of pores decrease, etc. As a result, the supporting performance of the catalyst metal particles is impaired, the support of the catalyst metal particles becomes uneven, the supported amount decreases, and the function as a catalyst itself falls. In this third heat treatment step, the object is to repair growth of the graphene, that is, merger of graphene, repair defects in the graphene, repair stacking defects of the graphene, etc. As the temperature region for avoiding remarkable growth of graphene and growth of the stacked structure, 1400 C. to 2100 C. was set.

[0105] Note that, to satisfy the conditions (d) and (e), in the above first heating step for removing the moisture, this heating temperature should be made 80 C. to 100 C., preferably 90 C. to 100 C. By this, adsorbed water partially remains in the obtained metal particle-containing intermediate and the ejection of metal particles by the explosion in the later second heat treatment step becomes uneven. It is possible to prevent as much as possible the distribution of pore size of the supporting carbon material obtained as a result from becoming broader exceeding the range of 2 nm to less than 10 nm and to obtain a supporting carbon material having as uniform a pore size as possible. If the heating temperature in this first heat treatment step is less than 80 C., it becomes difficult to obtain a supporting carbon material having a uniform pore size, while conversely if over 100 C., while there is no problem in the point of obtaining a supporting carbon material having a uniform pore size, there is no difference in the supporting carbon material obtained at 100 C. or more. From an economic perspective, there is the problem that the extra added heat becomes wasted.

[0106] Further, to satisfy the conditions (f), (g), and (h), in the acetylide producing step of blowing acetylene gas into the above solution containing a metal or metal salt to cause the production of a metal acetylide, it is possible to replace the acetylene gas blowing step with a step of adding a saturated acetylene solution to produce a metal acetylide. Furthermore, by having the saturated acetylene solution be an aqueous solution having acetylene dissolved in a saturated state (saturated acetylene aqueous solution) or a ketone-based solution having acetylene dissolved in a saturated state (saturated acetylene-ketone-based solution), greater effects can be achieved. With this method, it is possible to use the saturated acetylene aqueous solution or saturated acetylene-ketone-based solution to set the acetylene concentration at the time of the reaction high, so the reaction of silver acetylide is promoted and carbon material of properties different from PLT 2 can be formed. The details of the actual formation of the carbon material in this reaction step are unknown, but the inventors etc. believe that the reaction for formation of silver acetylide is so fast that the acetylene blown in is instantaneously consumed, so by increasing the concentration of acetylene in the solution, many nuclei of silver acetylide molecule-like crystals are formed in the reaction system. The amount of the nuclei present increases whereby the nuclei contact and merge frequently and repeatedly resulting, the inventors believe, in promotion of growth of the silver acetylide. As a result, the obtained silver acetylide becomes larger in thickness and length of branches, the extent of branching also becomes greater, and a 3D tree-branch shaped structure grows resulting in a tree-branch shaped crystal. A carbon material having physical properties different from PLT 2 etc., specifically a carbon material having a large BET specific surface area or DBP oil absorption, may be formed.

3. Catalyst Metal Particle-Supporting Carbon Material and Method of Production of Same

[0107] In the present invention, the above obtained carbon material for supporting a solid polymer fuel cell is used to produce a catalyst metal particle-supporting carbon material for forming a catalyst layer of an anode and/or cathode of a fuel cell. Here, the method of production of the catalyst metal particle-supporting carbon material is not particularly limited so long as a catalyst metal particle-supporting carbon material preferable for forming a catalyst layer of a fuel cell is obtained, but preferably the following method may be mentioned.

[0108] This method is the method of producing a catalyst metal particle-supporting carbon material by making the supporting carbon material disperse in a liquid dispersion medium, adding to the obtained dispersion a complex or salt of the metal forming the catalyst metal particles and a reducing agent, reducing the catalyst metal ions in the liquid phase to make the catalyst metal particles precipitate, and supporting the thus precipitated catalyst metal particles on the supporting carbon material. Preferably, it may comprise making the supporting carbon material disperse in advance in a liquid phase containing catalyst metal ions, adding the reducing agent in the state where there are sufficient catalyst metal ions present inside the pores, and thereby ensuring that the catalyst metal particles reduced inside the pores are reliably supported inside the pores.

[0109] Here, the catalyst metal particles forming the catalyst layers of the fuel cell are not particularly limited so long as having the function of promoting the chemical reaction required at the anode or cathode. As specific examples, platinum, palladium, ruthenium, gold, rhodium, osmium, iridium, tungsten, lead, iron, chrome, cobalt, nickel, manganese, vanadium, molybdenum, gallium, aluminum, or other metals or composites or alloys etc. comprised of two or more types of these metals may be mentioned, but in accordance with need, other catalyst metal ingredients may also be used.

[0110] The catalyst metal ingredient forming this catalyst metal particles is preferably platinum (Pt) or a metal mainly comprised of platinum (Pt). As the metal element to be added other than Pt, the elements Co, Ni, Fe, Pd, Au, Ru, Rh, Ir, etc. meant to improve the activity of catalyst metal particles may be mentioned. Further, the amount of these metal elements added is 50 at % or less in terms of percent of composition with respect to Pt. If the amount of addition of the metal element other than Pt exceeds 50 at %, the ratio of presence of the metal element other than Pt at the surface of the catalyst metal particles becomes greater, the element dissolves in operation of the fuel cell, and the power generation performance is liable to fall.

[0111] Further, as the liquid dispersion used for production of the catalyst metal particle-supporting carbon material, for example, water, alcohol, polyhydric alcohol, diphenyl ether, etc. are preferably used. Further, as the reducing agent added to the dispersion, for example, sodium borohydride, potassium borohydride, lithium borohydride, lithium aluminum chloride, an alcohol, formaldehyde, citric acid, formic acid, oxalic acid, etc. having carboxyl groups, etc. are suitably used.

[0112] Furthermore, the amount of the catalyst metal particles in the catalyst metal particle-supporting carbon material (rate of support of catalyst metal particles (mass %)) is 10 mass % to 60 mass % with respect to the total mass of the catalyst metal particles and supporting carbon material, preferably 20 mass % to 50 mass %. This is because when forming a catalyst metal particle-supporting carbon material as catalyst layers and generating power as a solid polymer fuel cell, diffusion of the hydrogen gas or oxygen gas used as fuel is essential. Diffusion of the hydrogen gas or oxygen gas in this catalyst layer becomes good under the above condition. Here, the rate of support of the catalyst metal particles is correlated with the distance between catalyst metal particles. If the amounts of the catalyst metal particles in the catalyst layer are equal, if the rate of support of the catalyst metal particles is low, the catalyst layer becomes thicker, the oxygen gas becomes harder to diffuse, and the fuel cell falls in power generation performance. Conversely, if the rate of support of the catalyst metal particles is high, the catalyst layer becomes thin and oxygen gas easily diffuses, but the problem arises that the distance between catalyst metal particles becomes shorter and they compete for oxygen molecules.

[0113] Note that, the supporting carbon material used in the method of production of the above catalyst metal particle-supporting carbon material is not particularly limited so long as satisfying the values of the physical properties prescribed in the present invention. However, a material which causes a chemical reaction other than the inherently sought reaction or a material which would elute a substance forming the carbon material by contact with condensed water would not be preferable. A chemically stable material is preferable.

[0114] Further, in the above explanation, mainly the application of a cathode electrode where the conditions sought are most stringent among applications of electrodes of solid polymer fuel cells was explained, but the invention may also be used for the application of an anode electrode where the conditions sought are easier than a cathode electrode.

EXAMPLES

[0115] Below, examples and comparative examples will be given to specifically explain the present invention, but the present invention is not limited to these examples and comparative examples.

Test 1

Examples 1 to 22 and Comparative Examples 1 to 20

1. Preparation of Supporting Carbon Material

(1) Acetylide Generating Step for Generating Silver Acetylide

[0116] 300 ml of a NH.sub.3 concentration 1.9 wt % silver nitrate-containing ammonia aqueous solution containing silver nitrate in a concentration of 1.1 mol % was taken in a flask. The air in this flask (residual oxygen) was replaced with argon or dry nitrogen or other inert gas to remove it, then, while stirring, acetylene gas was blown into 150 ml of this silver nitrate-containing ammonia aqueous solution by a flow rate of 25 ml/min over about 4 minutes. The generated silver acetylide was made to precipitate as a solid and was separated by filtration by a membrane filter to obtain silver acetylide.

(2) First Heat Treatment Step of Preparing Intermediate Containing Silver Particles

[0117] Next, the silver acetylide obtained in the above acetylide generating step was split into 50 mg amounts which were placed in Petri dishes. These were placed into vacuum heating containers. The gas generated during the heat treatment was sucked out to maintain a vacuum and the samples were heated at a 50 C. temperature for 3 hours. By this first heat treatment, a powder-shaped silver particle-containing intermediate with a moisture content decreased to 0.1 to 1.0 mass % or so was obtained.

(3) Second Heat Treatment Step of Obtaining Carbon Material Intermediate

[0118] Next, a pellet press-forming apparatus was used to form the powder state silver particle-containing intermediate obtained in the first heat treatment step by pressing at 0.5 kg/cm.sup.2 into pellets of a size of a diameter of 5 mmthickness of 0.2 mm and a weight of 0.5 g. The obtained pellets were placed on a heating plate of a reduced pressure heating apparatus then the system in the apparatus as a whole was held at a 1 Torr or less reduced pressure state. Next, the heating plate was used to heat the pellets by a speed of temperature rise shown in Table 1 to the heating temperature shown in Table 1 to cause an explosive reaction and prepare the carbon material intermediate.

[0119] Note that, in the case of Comparative Examples 15 to 20, in this second heat treatment step, pellets were not formed. The powder state silver particle-containing intermediate obtained at the first heat treatment step was heated as is at the heating temperature shown in Table 1 to cause an explosive reaction.

(4) Washing Step of Washing Carbon Material Intermediate

[0120] Next, the carbon material intermediate obtained in the second heat treatment step was washed by concentrated nitric acid to remove the silver particles and other unstable carbon compounds deposited on the surface of the carbon material intermediate to clean it.

(5) Third Heat Treatment Step of Obtaining Supporting Carbon Material

[0121] The cleaned carbon material intermediate obtained at the above second heat treatment step was heated using a Tammann furnace in an inert atmosphere using argon gas at a heating temperature shown in Table 1 for 2 hours to obtain the supporting carbon materials of Examples 1 to 22 and Comparative Examples 1 to 20.

[0122] Note that, in the obtained supporting carbon materials of Examples 1 to 22 and Comparative Examples 1 to 20, the supporting carbon materials of Examples 1 to 22 and Comparative Examples 1 to 14 are identified as A while the supporting carbon materials of Comparative Examples 15 to 20 are identified as B. Further, the values of the speed of temperature rise ( C./min) and heating temperature ( C.) at the second heat treatment step and the heating temperature ( C.) at the third heat treatment step without their units are used as identifiers. For example, the case of Example 1 where the second heat treatment is performed by a speed of temperature rise of 100 C./min and a heating temperature of 400 C. and, further, the third heat treatment is performed at 1400 C. can be expressed as A-100-400-1400.

TABLE-US-00001 TABLE 1 Second heat Second heat Third treatment treatment heat Rate of Rate of Third heat Rate of treat- temp. temp. treatment: temp. ment: rise rise heating rise Heating heating ( C./ ( C./ temp. Comp. ( C./ temp. temp. Ex. min) min) ( C.) ex. min) ( C.) ( C.) 1 100 400 1400 1 30 400 1800 2 100 400 1600 2 80 400 1800 3 100 400 1800 3 5 400 1800 4 100 400 2000 4 100 200 1800 5 100 400 2100 5 100 300 1800 6 100 450 1500 6 100 200 1000 7 100 500 1500 7 100 200 1300 8 100 600 1500 8 100 400 1000 9 150 500 1500 9 100 400 1300 10 200 500 1500 10 100 400 2200 11 300 500 1500 11 100 400 2400 12 500 500 1500 12 80 200 1800 13 300 500 1400 13 80 300 1800 14 300 500 1600 14 80 300 1200 15 300 500 1800 15 200 16 300 500 2000 16 200 1400 17 300 500 2100 17 200 1600 18 300 600 1800 18 200 1800 19 400 600 1900 19 200 2000 20 500 600 1900 20 200 2200 21 400 600 2000 22 500 600 2000

Comparative Examples 21 to 33

[0123] Further, as the commercially available carbon material, ketchen black (made by Lion, EC600JD, carbon material C), activated carbon (made by Kuraray Chemical, YP50F, carbon material D), and acetylene black not made porous (carbon black) (made by Denka, Denka Black, carbon material E) were used to prepare the supporting carbon materials of Comparative Examples 21 to 33. In Comparative Examples 21 to 26, the carbon material C was used as is without heat treatment (Comparative Example 21: C) or was used heat treated at 1400 C. (Comparative Example 22: C-1400), 1600 C. (Comparative Example 23: C-1600), 1800 C. (Comparative Example 24: C-1800), 2000 C. (Comparative Example 25: C-2000), or 2200 C. (Comparative Example 26: C-2200) to obtain supporting carbon materials. Further, in Comparative Examples 27 to 32, the carbon material D was pulverized to an average particle size of 1.2 m and either used as is without heat treatment (Comparative Example 27: D) or used heat treated at 1400 C. (Comparative Example 28: D-1400), 1600 C. (Comparative Example 29: D-1600), 1800 C. (Comparative Example 30: D-1800), 2000 C. (Comparative Example 31: D-2000), and 2200 C. (Comparative Example 32: D-2200) to obtain the supporting carbon materials. Furthermore, in Comparative Example 33, the carbon material E was used as is without heat treatment to obtain the supporting carbon material (E).

2. Evaluation of Physical Properties of Supporting Carbon Material

(a) Measurement of Specific Surface Area S.SUB.A .of Mesopores

[0124] For the supporting carbon materials of Examples 1 to 22 and Comparative Examples 1 to 33 prepared or obtained by the above-mentioned methods, each supporting carbon material was taken in about 50 mg, dried at 90 C. in vacuo, then measured for nitrogen gas adsorption/desorption characteristics using an automatic specific surface area measurement device (made by Bel Japan, BELSORP mini). At this time, the adsorption isotherm of the adsorption process was analyzed by the Dollimore-Heal method to calculate the specific surface area S.sub.A of the pore size 2 to 50 nm mesopores. The specific surface areas S.sub.A of the mesopores of the supporting carbon materials A of Examples 1 to 22 are shown in Table 2, while the specific surface areas S.sub.A of the mesopores of the supporting carbon materials B to E of Comparative Examples 1 to 33 are shown in Table 3.

(b) Measurement of Raman Spectrum

[0125] For the supporting carbon materials of Examples 1 to 22 and Comparative Examples 1 to 33 prepared or obtained by the above-mentioned methods, each supporting carbon material was taken in about 3 mg and was measured using a Raman spectroscopy apparatus (made by JASCO Corporation, NRS-7100) under conditions of an excitation laser: 532 nm, laser power: 100 mW (sample irradiation power: 0.1 mW), microscope arrangement: back scattering, slit: 100 m100 m, object lens: 100, spot diameter: 1 m, exposure time: 30 sec, observation wave number: 3200 to 750 cm.sup.1, and cumulative addition: 2 times. From the Raman spectrum obtained by measurement, the relative intensity ratio (IG/IG) of the peak intensity (IG) of the range of 2650 to 2700 cm.sup.1 called the G-band and the peak intensity (IG) of the range of 1550 to 1650 cm.sup.1 called the G-band was calculated and the peak position of the G-band was found. The measurement was conducted three times and the average value of the three times was made the measurement data. The relative intensity ratios (IG/IG) and G-band peak positions (G position) of the supporting carbon materials A of Examples 1 to 22 are shown in Table 2, while the relative intensity ratios (IG/IG) and G-band peak positions (G position) of the supporting carbon materials B to E of Comparative Examples 1 to 33 are shown in Table 3.

3. Preparation of Solid Polymer Fuel Cell and Evaluation of Power Generation Performance

(1) Preparation of Catalyst Metal Particle-Supporting Carbon Material

[0126] The supporting carbon materials prepared or obtained at Examples 1 to 22 and Comparative Examples 1 to 33 were used to prepare platinum-supporting carbon materials in the following way.

[0127] That is, first, water and ethanol were mixed in a volume ratio of a 1:1 ratio. To the obtained mixed solution (liquid dispersion medium): 150 ml, 0.4 g of chloroplatinic acid by platinum conversion was added and made to dissolve, then polyvinyl pyrrolidone (molecular weight 10000): 0.2 g was further made to dissolve. In the obtained solution, the supporting carbon material: 0.6 g was dispersed, then the dispersion was heat treated for 2 hours in the boiling state to simultaneously cause reduction and supporting. After the end of the reaction, the filtrate was dried in vacuo at 90 C. The obtained solid content was fired in argon gas containing hydrogen: 10 vol % at 300 C. for 1 hour to obtain the platinum particle-supporting carbon materials of Examples 1 to 22 and Comparative Examples 1 to 33 in which particle size 4 to 6 nm platinum particles were supported in a ratio of a supported amount of 40 mass %. Note that, the supported amount of the platinum particles was confirmed by measurement by ICP-AES. Further, the particle size of the platinum particles was evaluated by the half width of the main peak of platinum in powder X-ray diffraction.

(2) Preparation of Catalyst Layer

[0128] For each of the platinum particle-supporting carbon materials of the above Examples 1 to 22 and Comparative Examples 1 to 33, first, the platinum particle-supporting carbon material: 10 mass parts was taken in a container in an Ar atmosphere. To this, as an electrolyte material, Nafion made by Dupont: 8 mass parts (ratio of solid content) was added. This was lightly stirred, then ultrasonic waves were used to crush the platinum-supporting carbon material. Furthermore, while stirring, ethanol was added to adjust the solid content concentration of the platinum particle-supporting carbon material (Pt catalyst) and persulfonic acid-based ion exchange resin to 1 mass % and prepare a platinum catalyst ink comprised of a Pt catalyst and an electrolyte resin mixed together. After that, part of the obtained platinum catalyst ink was taken and ethanol was added to this platinum catalyst ink while stirring to prepare a platinum catalyst ink for coating use with a platinum concentration of 0.5 mass %.

[0129] Next, the mass of the platinum particles (catalyst metal particles) coated per unit area of the catalyst metal particles catalyst layer (below, platinum basis weight) was set to 0.2 mg/cm.sup.2 and the spray conditions were adjusted to give the targeted platinum basis weight. The above coating-use platinum catalyst ink was sprayed on a Teflon sheet under the adjusted conditions, then was treated to dry it in argon at 120 C. for 60 minutes to obtain the catalyst layers of the anode and cathode.

(3) Preparation of MEAs

[0130] The above prepared catalyst layers of Examples 1 to 22 and Comparative Examples 1 to 33 were used to prepare membrane electrode assemblies (MEA) by the following method.

[0131] From a Nafion membrane (made by Dupont, NR211), a square shaped electrolyte membrane having 6 cm sides was cut out. Further, the catalyst layers of the anode and cathode coated on Teflon sheets were cut out by cutter knives to square shapes of 2.5 cm sides. Between the catalyst layers of the thus cut out anode and cathode, this electrolyte membrane was sandwiched so that the catalyst layers straddled and contacted the center part of the electrolyte membrane without being offset from each other. These were pressed for 10 minutes while heating and pressing by 120 C. and 100 kg/cm.sup.2, then were cooled down to room temperature. After that, at both the anode and cathode, just the Teflon sheets were carefully peeled off to prepare a catalyst layer-electrolyte membrane assembly comprised of an electrolyte membrane at the two sides of which catalyst layers of the anode and cathode were fixed.

[0132] Next, as the gas diffusion layers, from carbon paper (made by SGL Carbon, 35BC), a pair of square shapes of carbon paper of sizes of 2.5 cm sides were cut out. Between these sheets of carbon paper, the above catalyst layer-electrolyte membrane assembly was inserted so that the catalyst layers of the anode and cathode matched without offset. The assembly was pressed by heating and pressing at 120 C. and 50 kg/cm.sup.2 for 10 minutes to prepare the MEA of each of Examples 1 to 22 and Comparative Examples 1 to 33.

[0133] Note that, the composition of the catalyst metal and the basis weights of the components of the carbon material and electrolyte material at the thus prepared MEAs were found by finding the mass of the catalyst layers fixed to the Nafion membrane electrolyte membrane from the difference between the mass of the Teflon sheets with the catalyst layers before pressing and the mass of the Teflon sheets after pressing and peeling (electrolyte membrane) and calculating using the mass ratio of the composition of the catalyst layer.

[0134] Further, for the anode, the carbon material A-100-400-1400 (Example 1) was used in common so as to extract only the performance of the cathode catalyst layer from the results of evaluation of the power generation characteristics.

(4) Evaluation of Power Generation Performance

[0135] The MEAs prepared in Examples 1 to 22 and Comparative Examples 1 to 33 were assembled into fuel cells. The cells were set in a fuel cell measurement device, then the following procedure was used to evaluate the power generation performance of the fuel cells by the following procedure.

[0136] For the gas, air was supplied to the cathode and, further, pure hydrogen was supplied to the anode under atmospheric pressure so that the rates of utilization became 40% and 70%. Further, the cell temperature was set to 80 C. The supplied gases for both the cathode and anode were bubbled through distilled water warmed to 65 C. in a humidifier to introduce water vapor corresponding to modified hydrogen for supply to the cell.

[0137] Under conditions of supplying gases to the cell under such settings, the load was gradually increased and the voltage across cell terminals at 1200 mA/cm.sup.2 was recorded as the output voltage for evaluation as the initial power generation performance of the fuel cell (characteristic before deterioration).

[0138] Further, for evaluating the durability of the fuel cells, a durability test was run using a load adjuster to hold a cell from the outside by first 0.6V for 6 seconds, then hold it at 1.3V for 6 seconds, then again hold it at 0.6V for 6 seconds and repeating this load fluctuation 500 times. After that, the same procedure was performed as the evaluation of the initial power generation performance (characteristic before deterioration) to record the voltage across cell terminals at 1200 mA/cm.sup.2 as the output voltage and evaluate the power generation performance after a durability test of the fuel cell (characteristics after deterioration).

[0139] The power generation performances of the obtained fuel cells were evaluated overall by the following criteria (overall evaluation):

(very good) level: Sample satisfying both output voltage before durability test at 1200 mA/cm.sup.2 of 0.60V or more and output voltage after durability test of of 0.50V or more.
(good) level: Sample satisfying both output voltage before durability test at 1200 mA/cm.sup.2 of 0.60V or more and output voltage after durability test of 0.45V to less than 0.50V.
x (poor) level: Sample not satisfying (good) level.

[0140] The above results are shown in Table 2 and Table 3.

TABLE-US-00002 TABLE 2 Supporting carbon material Meso- Power generation pore performance specific G Du- O- surface posi- ra- ver- Heat area IG/ tion Ini- bility all treatment type (m.sup.2/g) IG (com.sup.1) tial test eval. Ex. 1 A-100-400-1400 1580 0.82 2664 0.62 0.45 Ex. 2 A-100-400-1600 1350 0.96 2663 0.62 0.46 Ex. 3 A-100-400-1800 1120 1.16 2661 0.61 0.54 Ex. 4 A-100-400-2000 890 1.86 2669 0.62 0.56 Ex. 5 A-100-400-2100 740 1.93 2669 0.62 0.55 Ex. 6 A-100-450-1500 1410 0.89 2664 0.61 0.45 Ex. 7 A-100-500-1500 1270 2.08 2664 0.66 0.46 Ex. 8 A-100-600-1500 1320 2.16 2665 0.67 0.46 Ex. 9 A-150-500-1500 1350 2.11 2664 0.67 0.47 Ex. 10 A-200-500-1500 1390 2.03 2663 0.66 0.47 Ex. 11 A-300-500-1500 1360 1.96 2661 0.67 0.46 Ex. 12 A-500-500-1500 1290 1.55 2661 0.67 0.46 Ex. 13 A-300-500-1400 1260 1.32 2661 0.66 0.46 Ex. 14 A-300-500-1600 1210 1.23 2661 0.68 0.48 Ex. 15 A-300-500-1800 1160 1.63 2663 0.67 0.56 Ex. 16 A-300-500-2000 890 1.86 2668 0.67 0.57 Ex. 17 A-300-500-2100 730 1.93 2668 0.66 0.57 Ex. 18 A-300-600-1800 1090 1.66 2664 0.68 0.55 Ex. 19 A-400-600-1900 950 1.73 2666 0.69 0.57 Ex. 20 A-500-600-1900 960 1.77 2666 0.67 0.56 Ex. 21 A-400-600-2000 810 1.84 2667 0.65 0.58 Ex. 22 A-500-600-2000 820 1.89 2666 0.65 0.57

TABLE-US-00003 TABLE 3 Supporting carbon material Power Meso- generation pore performance specific G Du- O- Heat surface posi- ra- ver- treatment area IG/ tion Ini- bility all type (m.sup.2/g) IG (com.sup.1) tial test eval. Comp. A-30-400-1800 1280 2.35 2658 0.56 0.42 Ex. 1 Comp. A-80-400-1800 1210 2.27 2658 0.56 0.41 Ex. 2 Comp. A-5-400-1800 1310 2.61 2655 0.56 0.42 Ex. 3 Comp. A-100-200-1800 1260 2.52 2659 0.54 0.41 Ex. 4 Comp. A-100-300-1800 1230 2.44 2657 0.56 0.41 Ex. 5 Comp. A-100-200-1000 1710 0.55 2654 0.56 0.4 Ex. 6 Comp. A-100-200-1300 1460 0.61 2655 0.55 0.4 Ex. 7 Comp. A-100-400-1000 1690 0.59 2654 0.56 0.39 Ex. 8 Comp. A-100-400-1300 1580 0.66 2656 0.55 0.39 Ex. 9 Comp. A-100-400-2200 580 2.69 2673 0.53 0.42 Ex. 10 Comp. A-100-400-2400 420 2.26 2675 0.53 0.41 Ex. 11 Comp. A-80-200-1800 1210 2.31 2671 0.52 0.41 Ex. 12 Comp. A-80-300-1800 1190 2.41 2672 0.53 0.4 Ex. 13 Comp. A-80-300-1200 1630 0.71 2558 0.53 0.41 Ex. 14 Comp. B 1260 2.16 2673 0.51 0.32 Ex. 15 Comp. B-1400 1100 2.23 2675 0.52 0.38 Ex. 16 Comp. B-1600 750 2.44 2677 0.49 0.4 Ex. 17 Comp. B-1800 520 2.86 2677 0.46 0.42 Ex. 18 Comp. B-2000 430 3.11 2677 0.44 0.39 Ex. 19 Comp. B-2200 410 3.62 2678 0.41 0.32 Ex. 20 Comp. C 960 0.31 2651 0.41 0.32 Ex. 21 Comp. C-1400 920 0.42 2657 0.39 0.32 Ex. 22 Comp. C-1600 750 0.55 2663 0.39 0.33 Ex. 23 Comp. C-1800 640 0.61 2668 0.39 0.33 Ex. 24 Comp. C-2000 420 0.86 2672 0.38 0.35 Ex. 25 Comp. C-2200 360 1.02 2674 0.38 0.36 Ex. 26 Comp. D 1820 0.31 2651 0.41 0.32 Ex. 27 Comp. D-1400 1450 0.42 2657 0.39 0.32 Ex. 28 Comp. D-1600 1060 0.55 2663 0.39 0.33 Ex. 29 Comp. D-1800 680 0.61 2668 0.39 0.33 Ex. 30 Comp. D-2000 420 0.86 2672 0.38 0.35 Ex. 31 Comp. D-2200 360 1.02 2674 0.38 0.36 Ex. 32 Comp. E 82 1.36 2671 0.46 0.4 Ex. 33

[0141] The MEAs using the supporting carbon materials of Examples 1 to 22 of the present invention shown in Table 2 all exhibited excellent power generation performances of the (good) level or more. In particular, the MEAs using the supporting carbon materials of Examples 15 to 22 had values of the mesopore specific surface area, IG/IG, and G peak position satisfying the preferable ranges and exhibited better power generation performances. As opposed to this, none of the MEAs using the supporting carbon materials of the comparative examples shown in Table 3 exhibited the (good) level.

[0142] In each of Comparative Examples 1 to 3, the speed of temperature rise in the second treatment step at the time of production was low, the growth of graphene was not sufficient, the IG/IG and the peak position of the G band were outside the ranges of the present invention, and, in particular, the initial characteristics were low.

[0143] In each of Comparative Examples 4 and 5, the heating temperature at the second heat treatment step was low, in the same way as the case of Comparative Examples 1 to 3, the growth of the graphene structure was insufficient, further, even if the heating temperature at the third heat treatment step was sufficiently high, the skeleton structure was believed to be changed. The durability was improved, but the initial characteristics remarkably fell. In the supporting carbon materials of Comparative Examples 4 and 5, the value of IG/IG in the Raman spectrum and the peak position of the G band were outside the ranges of the present invention.

[0144] In each of Comparative Examples 6 and 7, the heating temperature at the second heat treatment step was low and, further, the heating temperature at the third heat treatment step was low, so the initial characteristics were relatively good, but there was little graphene growth, the crystallinity was low, and the durability was poor.

[0145] In each of Comparative Examples 8 to 11, at the second heat treatment step, the heating temperature was sufficient and a sufficient explosive energy was given, but the heating temperature at the third heat treatment step was low or too high. If the heating temperature is low (Comparative Examples 8 and 9), the initial characteristics become relatively good, but the durability is poor. Conversely, when the heating temperature is high (Comparative Examples 10 and 11), along with the growth of the crystal structure, the pore structure ends up being crushed. Due to this change in the pore structure, the initial characteristics greatly fall. The supporting carbon materials of Comparative Examples 8 and 9 had a small IG/IG, further had a small wave number even at the position of the G-band, and had a low developed graphene structure. Further, the supporting carbon materials of Comparative Examples 10 and 11 had a large IG/IG, further had a large wave number of the peak position of the G-band, had a highly developed crystallinity, and had too many graphene layers.

[0146] In Comparative Examples 12 to 14, both the speed of temperature rise and the heating temperature at the second heat treatment step were low, the explosive energy was small, and the growth of the graphene structure was weak. In Comparative Examples 12 and 13, the heating temperature in the third heat treatment step was suitable, but the initial characteristics and durability could not be balanced. If the explosive energy is not suitable in state, even if changing the heat treatment in the third heat treatment step, the graphene skeleton determined at the second heat treatment step is weak and easily changes by heat treatment, so the initial characteristics and durability cannot both be obtained. The supporting carbon materials of Comparative Examples 12 and 13 had a large IG/IG (too highly developed a graphene structure) and, further, a large wave number of the peak position of the G-band (too large a number of graphene layers). Further, in Comparative Example 14, heat treatment at the second heat treatment step the same as Comparative Example 13 was performed while the heating temperature at the third heat treatment step was set low. However, the supporting carbon material of Comparative Example 14 had a small IG/IG (weak graphene structure), further had a small wave number of the peak position of the G-band (too small a number of stacked layers of graphene), and did not have a good balance of the initial characteristics and the durability.

[0147] Comparative Examples 15 to 20 show the results of the carbon material B prepared by substantially the same method of production as PLT 2. If the explosive energy at the second heat treatment step is weak, the growth of the graphene structure is weak, and the heating temperature at the third heat treatment step is raised, the skeleton structure ends up changing and the initial characteristics (skeleton structure) and durability (growth of graphene structure) cannot both be obtained or balanced. In the third heat treatment step, if raising the heating temperature to 1800 C. or more, the pore structure is crushed, the basic function of the support (supporting performance) falls, and the power generation performance falls. Further, with 1600 C. or less heating temperature, the power generation performance is relatively good, but the durability is poor. The supporting carbon materials of Comparative Examples 15 to 20 had large wave numbers of the peak positions of the G-band and had basically too many stacked graphene layers, so the performance could not be balanced.

[0148] Comparative Examples 21 to 26 are cases using supporting carbon materials comprised of the conventional carbon material C. When performing heat treatment by the third heat treatment step, the specific surface area remarkably fell and the initial power generation performance and durability could not both be obtained. Further, Comparative Examples 27 to 32 are cases of using supporting carbon materials comprised of the conventional carbon material D. They failed to reach the (good) level even with heat treatment by the third heat treatment step. Furthermore, Comparative Example 33 is the case of using a supporting carbon material comprising a conventional carbon material E, having a mesopore specific surface area of a small 82 m.sup.2/g, further having a large wave number of the peak position of the G-band, and not reaching the (good) level.

Test 2

Examples 101 to 116 and Comparative Examples 101 to 104

1. Preparation of Supporting Carbon Material

[0149] In Test 1, the temperature 50 C. of the vacuum drying in the (2) first heat treatment step for preparing the silver particle-containing intermediate was changed to 40 to 100 C. as shown in Table 4. The treatment time was studied in the range of 0.5 hour to 10 hours. Further, the temperature of the second heat treatment, as shown in Table 4, was studied in the range of 300 C. to 600 C. The rest of the steps were made the same as Example 15 of Test 1. That is, the speed of temperature rise in the second heat treatment was made 300 C./min while the third heating temperature was made 1800 C.

TABLE-US-00004 TABLE 4 Second heating First heating temperature temperature Temp. ( C.) Time (hr) Temp. ( C.) Ex. 101 80 3.0 400 Ex. 102 80 5.0 400 Ex. 103 80 7.0 400 Ex. 104 80 10.0 400 Ex. 105 85 3.0 400 Ex. 106 85 5.0 400 Ex. 107 85 7.0 400 Ex. 108 85 10.0 400 Ex. 109 90 5.0 400 Ex. 110 100 3.0 400 Ex. 111 100 5.0 400 Ex. 112 100 7.0 400 Ex. 113 100 10.0 400 Ex. 114 100 10.0 500 Ex. 115 100 10.0 600 Ex. 116 75 1.0 400 Comp. Ex. 101 45 0.5 350 Comp. Ex. 102 40 1.5 350 Comp. Ex. 103 50 1.0 300 Comp. Ex. 104 45 1.5 300

2. Evaluation of Physical Properties of Supporting Carbon Material

[0150] In addition to the evaluation of physical properties of the Test 1, the same procedure was followed as with measurement of the specific surface area of the mesopores of Test 1 so as to measure the supporting carbon materials of Examples 101 to 116 and Comparative Examples 101 to 104 for nitrogen gas adsorption/desorption characteristics. The obtained nitrogen adsorption isotherms of the adsorption process were analyzed by the Dollimore-Heal method to find the values of the specific pore area S.sub.2-10 and specific pore volume V.sub.2-10 of pores of a pore size of 2.0 to 10.0 nm and the specific pore area S.sub.10-50 of pores of a pore size of 10.0 to 50.0 nm. The value of the specific pore area S.sub.2 of the pores with a pore size of less than 2.0 nm was found by analysis of a nitrogen adsorption isotherm of the adsorption process by the Horvath-Kawazoe method (HK method). The results of these specific pore area S.sub.2-10, specific pore volume V.sub.2-10, specific pore area S.sub.10-50, and specific pore area S.sub.2 are shown in Table 5. Note that, the analyses by the Dollimore-Heal method and HK method were performed using analysis software attached to an apparatus for measuring the nitrogen gas adsorption/desorption characteristics (Bel Mini made by Bel Japan).

3. Preparation of Solid Polymer Fuel Cell and Evaluation of Power Generation Performance

[0151] Next, the thus prepared supporting carbon materials of Examples 101 to 116 and Comparative Examples 101 to 104 were used to perform the same procedure as in Test 1 to prepare catalysts for solid polymer fuel cells, catalyst layers, and MEAs.

[0152] (1) Evaluation of Power Generation Performance

[0153] The prepared MEAs were assembled into cells. The cells were set in a fuel cell measurement device. As the evaluation of performance of the fuel cells, in addition to the evaluation of performance of Test 1, the current density at a cell voltage of 0.3V was measured.

[0154] The results of evaluation of the performances of the obtained fuel cells were evaluated overall by the following criteria (overall evaluation).

(excellent) level: Sample satisfying conditions of output voltage at 1200 mA/cm.sup.2 of 0.65V or more, output voltage after durability test of 0.55V or more, and current density at 0.3V cell voltage of 1700 mA/cm.sup.2 or more
(very good) level: Sample satisfying output voltage at 1200 mA/cm.sup.2 of 0.60V or more, output voltage after durability test of 0.50V or more, and and current density at 0.3V cell voltage of 1600 mA/cm.sup.2 or more
(good) level: Sample satisfying output voltage at 1200 mA/cm.sup.2 of 0.60V or more, output voltage after durability test of 0.45V or more, and current density at 0.3V cell voltage of 1600 mA/cm.sup.2
x (poor) level: Sample not satisfying (good) level The results are shown in Table 5.

TABLE-US-00005 TABLE 5 Power generation performance Supporting carbon material Output voltage at Current Mesopore 1200 mA/cm.sup.2 density at specific G Specific Specific Specific Specific After 0.3 V cell surface area position pore area pore volume pore area pore area Initial durability voltage Overall SA (m.sup.2/g) IG/IG (cm.sup.1) S.sub.2-10 (m.sup.2/g) V.sub.2-10 (m.sup.2/g) S.sub.10-50 (m.sup.2/g) S.sub.2 (m.sup.2/g) (V) test (V) (mA/cm.sup.2) eval. Ex. 101 1240 1.2 2665 1090 1.3 146 530 0.61 0.46 1610 Ex. 102 605 1.3 2666 455 0.52 145 310 0.62 0.46 1620 Ex. 103 1010 1.3 2667 890 1.56 120 420 0.62 0.47 1630 Ex. 104 740 1.4 2665 620 0.43 115 380 0.63 0.48 1610 Ex. 105 1180 1.4 2667 1030 1.3 149 510 0.62 0.46 1620 Ex. 106 1000 1.1 2665 980 1.4 22 290 0.61 0.46 1620 Ex. 107 960 1.2 2663 850 1.2 105 545 0.62 0.47 1630 Ex. 108 890 1.2 2664 760 1.0 125 255 0.63 0.47 1630 Ex. 109 960 1.3 2665 870 1.3 85 420 0.62 0.46 1700 Ex. 110 1070 1.6 2663 980 1.2 85 320 0.63 0.51 1710 Ex. 111 970 1.6 2667 880 1.3 90 460 0.63 0.52 1710 Ex. 112 1050 1.5 2666 950 1.2 95 330 0.63 0.52 1710 Ex. 113 995 1.3 2665 920 1.1 75 360 0.63 0.52 1710 Ex. 114 1040 1.2 2664 960 1.2 80 350 0.66 0.55 1720 Ex. 115 1065 1.3 2665 980 1.3 85 400 0.67 0.56 1760 Ex. 116 1310 1.0 2662 1150 1.5 160 450 0.62 0.46 1500 Comp. Ex. 101 1610 1.0 2663 1460 1.6 145 530 <0.30 <0.30 1450 Comp. Ex. 102 425 1.1 2661 410 0.31 15 255 <0.30 <0.30 1460 Comp. Ex. 103 1060 0.6 2661 850 1.55 210 420 <0.30 <0.30 1430 Comp. Ex. 104 1045 0.9 2598 900 1.4 145 620 <0.30 <0.30 1440

[0155] In the results of evaluation of the fuel cell performance shown in Table 5, Examples 101 to 115 had characteristics evaluated by Test 1 (specific surface area S.sub.A, IG/IG, and G peak position) in the ranges of the present invention and a specific pore area S.sub.2-10, specific pore volume V.sub.2-10, specific pore area S.sub.10-50, and specific pore area S.sub.2 in the ranges of the present invention. As a result, these samples satisfied the (good) level. In particular, Examples 110 to 115 had a specific pore area S.sub.2-10, specific pore volume V.sub.2-10, specific pore area S.sub.10-50, and specific pore area S.sub.2 in the preferable ranges, so satisfied the (very good) level or the (excellent) level.

[0156] Note that, Example 116 had the characteristics evaluated by Test 1 in the ranges of the present invention, so is a working example and satisfied the (good) level of Test 1. However, Examples 116 had a specific pore area S.sub.2-10 and specific pore area S.sub.10-50 outside the ranges, so had a low current density and could not satisfy the (good) level of Test 2.

[0157] Further, Comparative Examples 101 to 104 where at least one of the characteristics evaluated in Test 1 was outside the range of the present invention and at least one of the specific pore area S.sub.2-10, specific pore volume V.sub.2-10, specific pore area S.sub.10-50, and specific pore area S.sub.2 was outside the range of the present invention failed to satisfy the (good) level of Test 1 and the (good) level of Test 2.

Test 3

Examples 201 to 216 and Comparative Examples 201 to 204

1. Preparation of Supporting Carbon Material

[0158] In the Test 2, only (1) the step of preparation of the silver acetylide was changed to the following step. The other steps were the same as Example 105 of Test 2. That is, first, an ammonia aqueous solution containing silver nitrate in the concentration shown in Table 6 (concentration: 1.3 to 2.7 mass %) was taken in a flask. The air inside this flask was replaced with argon or dry nitrogen or other inert gas to remove the oxygen, then, while stirring, a tabletop ultrasonic generator (70 W) was used to fire ultrasonic waves while the amount of saturated acetylene aqueous solution or saturated acetylene-acetone solution shown in Table 1 was added to 500 ml of the above ammonia aqueous solution of silver nitrate to cause solids of silver acetylide to precipitate in the ammonia aqueous solution. Next, the precipitate was separated by filtration by the membrane filter, the obtained precipitate was washed with methanol, and further some ethanol was added after washing to impregnate this precipitate with methanol. By doing this, it is possible to prepare silver acetylide serving as a carbon material with a large DBP oil adsorption grown in a tree-branch shaped structure.

TABLE-US-00006 TABLE 6 Silver Amount of Amount of nitrate aqueous acetone concentration solution solution (mass %) (ml) (ml) Ex. 201 1.3 250 None Ex. 202 1.3 500 None Ex. 203 1.3 1000 None Ex. 204 1.5 500 None Ex. 205 1.5 1000 None Ex. 206 1.7 500 None Ex. 207 1.7 1000 None Ex. 208 1.9 500 None Ex. 209 1.9 1000 None Ex. 210 2.1 500 None Ex. 211 2.1 1000 None Ex. 212 1.3 None 90 Ex. 213 1.5 None 90 Ex. 214 1.7 None 90 Ex. 215 1.9 None 90 Ex. 216 2.1 None 90 Comp. Ex. 201 2.1 1000 130 Comp. Ex. 202 2.5 250 None Comp. Ex. 203 2.7 250 None Comp. Ex. 204 2.7 None 50 Comp. Ex. 205 2.7 None 70

2. Evaluation of Physical Properties of Supporting Carbon Material

[0159] In addition to the evaluations of properties of Tests 1 and 2, the following were evaluated:

(a) BET Specific Surface Area

[0160] The BET specific surface area S.sub.BET was measured for the supporting carbon materials of Examples 201 to 216 and Comparative Examples 201 to 204 using an Autosorb I-MP made by Quantachrome Instruments by the gas adsorption method using nitrogen gas and determining the specific surface area by the one-point method based on the BET method.

(b) DBP Oil Absorption

[0161] The DBP oil absorption was determined for the supporting carbon materials of Examples 201 to 216 and Comparative Examples 201 to 204 using an Absorptometer (made by Brabender) and converting the amount of addition of DBP at the time of 70% of the maximum torque to the DBP oil absorption per 100 g of sample.

(c) Raman Spectrum

[0162] The Raman spectrum was measured for the supporting carbon materials of Examples 201 to 216 and Comparative Examples 201 to 204 by measuring out about 3 mg of the supporting carbon materials and using a laser Raman spectrophotometer (made by JASCO Corporation, NRS-3100) under conditions of an excitation laser: 532 nm, laser power: 10 mW (sample irradiation power: 1.1 mW), microscope arrangement: back scattering, object lens: 100, spot diameter: 1 m, exposure time: 30 sec, observation wave number: 2000 to 300 cm.sup.1, and cumulative addition: 6 times. From the Raman spectrum obtained by measurement, the half width (G) of the G-band (range of 1550 to 1650 cm.sup.1) was calculated.

2. Preparation of Solid Polymer Fuel Cell and Evaluation of Power Generation Performance

[0163] Next, using the above prepared supporting carbon materials of Examples 201 to 216 and Comparative Examples 201 to 204, the same procedure was performed as Test 1 to prepare catalysts, catalyst layers, and MEAs for solid polymer fuel cells.

[0164] (1) Evaluation of Power Generation Performance

[0165] The prepared MEAs were assembled into cells and set in a fuel cell measurement device to evaluate the performance of the fuel cells. In addition to the evaluation of performance of Tests 1 and 2, a durability test different from the durability test of Test 1 was performed.

[0166] (2) Durability Test of Fuel Cells

As the durability test, the voltage across the cell terminals was held at 1.0V for 1 second, then the voltage across the cell terminals was made to rise to 1.5V and held there for 1 second, then the initial cell terminal voltage was returned to 1.0V. This cycle was repeated 4000 times. After that, the battery performance was measured in the same way as the case of the test for evaluation of the initial performance before the durability test. For the gas, air was fed to the cathode and pure hydrogen was fed to the anode to give respective rates of utilization of 40% and 70%. The respective gas pressures were adjusted to atmospheric pressure at back pressure valves provided downstream of the cell. The cell temperature was set to 80 C., but the fed air and pure hydrogen were respectively bubbled in distilled water warmed to 60 C. to moisten them. As the durability test, compared with the durability test of Test 1, the amount of fluctuation of voltage became larger. How much the output voltage at 1200 mA/cm.sup.2 after the durability test dropped compared with the output voltage before the durability test was found as the rate of drop.

[0167] The obtained results of the evaluation of performance of the fuel cell were evaluated overall (overall evaluation) by the following standards:

(excellent) level: Sample satisfying output voltage at 1200 mA/cm.sup.2 of 0.65V or more, output voltage after durability test of 0.55V or more, current density at 0.3V cell voltage of 1700 mA/cm.sup.2 or more, and rate of drop of less than 10%
(very good) level: Sample satisfying output voltage at 1200 mA/cm.sup.2 of 0.60V or more, output voltage after durability test of 0.50V or more, current density at 0.3V cell voltage of 1600 mA/cm.sup.2 or more, and rate of drop of less than 15%
(good) level: Sample satisfying output voltage at 1200 mA/cm.sup.2 of 0.60V or more, output voltage after durability test of 0.45V or more, current density at 0.3V cell voltage of 1600 mA/cm.sup.2 or more, and rate of drop of less than 20%
x (poor) level: Sample not satisfying the (good) level

[0168] The results are shown in Table 7.

TABLE-US-00007 TABLE 7 Supporting carbon material Power generation performance Mes- Spe- DBP BET Output Current opore Spe- cific Spe- Spe- oil spe- voltage at density Out- spe- cific pore cific cific absorp- cific G- 1200 mA/cm.sup.2 at 0.3 put cific G pore vol- pore pore tion surface band After V cell volt- O- surface posi- area ume area area (X) area X/ half Ini- dura- voltage age ver- area SA IG/ tion S.sub.2-10 V.sub.2-10 S.sub.10-50 S.sub.2 (cm.sup.3/ S.sub.BET S.sub.BET width tial bility (mA/ drop all (m.sup.2/g) IG (cm.sup.1) (m.sup.2/g) (m.sup.2/g) (m.sup.2/g) (m.sup.2/g) 100 g) (m.sup.2/g) (nm) (cm.sup.1) (V) test (V) cm.sup.2) (%) eval. Ex. 201 1155 0.9 2669 1010 1.45 145 265 640 1410 4.5 71 0.62 0.46 1620 19 Ex. 202 1206 2.1 2668 1080 1.5 125 285 450 1465 3.1 72 0.61 0.46 1615 18 Ex. 203 1196 2.15 2669 1080 1.45 115 290 670 1470 3.9 39 0.62 0.45 1610 19 Ex. 204 605 0.85 2661 465 1.55 135 255 300 620 4.8 43 0.63 0.47 1620 18 Ex. 205 1080 0.85 2661 1060 1.5 20 300 560 1380 4.1 47 0.63 0.46 1625 19 Ex. 206 1066 2.05 2662 1030 1.5 25 270 620 1420 4.4 37 0.61 0.47 1610 19 Ex. 207 1130 2.1 2689 1020 1.45 110 290 490 1410 3.5 73 0.62 0.47 1615 18 Ex. 208 1125 2.15 2689 1010 1.5 115 255 435 1380 3.2 32 0.63 0.46 1620 18 Ex. 209 925 2.0 2662 890 1.55 35 400 500 1320 3.8 72 0.62 0.45 1710 14 Ex. 210 1260 1.0 2662 1010 1.35 25 250 460 1290 3.5 65 0.61 0.46 1620 15 Ex. 211 610 1.0 620 1.55 90 480 510 1090 4.7 60 0.61 0.47 1760 13 Ex. 212 895 1.6 2661 860 0.85 35 255 360 1150 3.1 45 0.65 0.52 1770 13 Ex. 213 1225 2.1 2663 850 0.65 85 270 470 1490 3.2 52 0.66 0.62 1766 14 Ex. 214 615 1.5 2665 410 1.2 85 320 420 1135 3.7 55 0.69 0.57 1770 8 Ex. 215 610 1.6 2661 385 0.35 229 260 295 970 3.4 45 0.61 0.45 1630 19 Ex. 216 620 0.9 2661 495 0.65 125 360 285 980 2.9 69 0.62 0.45 1620 29 Comp. 706 0.6 2657 495 0.41 210 125 670 840 8.0 55 0.51 0.39 1480 32 Ex. 201 Comp. 1320 0.65 2658 1096 0.38 230 630 1550 4.1 63 0.52 0.36 14602 31 Ex. 202 Comp. 915 0.55 2658 750 0.36 165 165 220 1100 2.0 72 0.49 0.37 1490 35 Ex. 203 Comp. 1025 0.75 2659 800 0.39 225 160 370 1165 3.1 80 0.5 0.34 1500 30 Ex. 204 indicates data missing or illegible when filed

[0169] In the results of evaluation of fuel cell performance shown in Table 7, Examples 201 to 214 each had characteristics evaluated at Test 1 (S.sub.A, IG/IG, and G peak position) and characteristics evaluated at Test 2 (S.sub.2-10, V.sub.2-10, S.sub.10-50, and S.sub.2) inside the ranges of the present invention and had a BET specific surface area S.sub.BET, DBP oil absorption (X), X/S.sub.BET, and G-band half width (G) inside the ranges of the present invention. As a result, these samples satisfied the (good) level. In particular, Examples 212 to 214 each had a BET specific surface area S.sub.BET, DBP oil absorption (X), and G-band half width (G) inside the preferable ranges, so satisfied the (very good) level or (excellent) level.

[0170] Note that, Example 215 is an example where the characteristics evaluated in Test 1 were in the ranges of the present invention, so is a working example. However, Example 215 had an S.sub.2-10, V.sub.2-10, and S.sub.10-50 outside the ranges, but the BET specific surface area S.sub.BET, DBP oil absorption (X), X/S.sub.BET, and G-band half width (G) were in the ranges of the present invention. For this reason, the current density was improved, the rate of drop also fell, and the (good) level of Test 3 was satisfied.

[0171] Further, Example 216 had characteristics evaluated in Test 1 and Test 2 in the ranges of the present invention, so is a working example. Therefore, the (good) levels of Test 1 and Test 2 were satisfied. However, Example 216 had an X/S.sub.BET outside the range, so the rate of drop was large and the (good) level of Test 3 could not be satisfied.

[0172] Further, Comparative Examples 101 to 104 where, among the characteristics evaluated at Test 1 and Test 2, at least one was outside the range of the present invention and at least one of the BET specific surface area S.sub.BET, DBP oil absorption (X), X/S.sub.BET, and G-band half width (G) was outside the range of the present invention failed to satisfy the (good) levels of Tests 1 to 3.