Method for producing supported metal catalyst and catalyst obtained thereby

Abstract

The present invention relates to a supported catalyst, in particular a bimetallic catalyst, and to such a catalyst thus obtained, which combines high stability with unique electronic and structural features.

Description

Technical Field

[0001] The present invention relates to a method for producing a supported metal catalyst, particularly a supported bimetal catalyst, and to the resulting catalyst which combines high stability with unique electronic and structural properties.

Background Art

[0002] Bimetal catalysts, and more generally alloy catalysts, are characterized by unique electronic and structural properties, which are responsible for the improved overall catalytic performance often observed when compared to their monometal counterparts. For example, alloying of Pt and Ru is beneficial for use in direct methanol fuel cells (DMFCs) and polymer electrolyte membrane fuel cells (PEMFCs) where CO-containing hydrogen feeds are used because the resistance of the catalyst to poisoning by CO is greatly improved. Similarly, PtNi- and PtCo-based systems are very attractive catalysts for the oxygen reduction reaction (ORR) because they feature a specific activity more than twice that of Pt-only-based systems. Therefore, these are state-of-the-art catalysts for the cathode compartment of PEMFCs. Au-Pd catalysts, instead, have been shown to exhibit a novel catalytic effect by synergistically enhancing redox reactions (e.g., oxidative dehydrogenation of hydroxymethylfurfural) through the separation of Pd and Au from the Au-Pd assembly. The most typical example of a synergistic effect is represented by the interaction between copper and zinc during the catalytic hydrogenation of CO2 to methanol, where the formation of a Cu-Zn alloy and subsequent oxidation results in a very rich Cu-ZnO interface that is beneficial for the reaction to occur. However, many other similar cases can be found in the literature.

[0003] The means used to synthesize supported monometal metal catalysts (e.g., wet and dry impregnation, co-precipitation, and colloidal deposition) are also used to prepare bimetal formulations but require multiple metal precursors. When using the wet impregnation method, the metal precursor and the support are dispersed in a solvent and then the suspension is filtered and dried. In this method, the attraction is mainly due to the adsorption of the metal precursor to the support (pre-catalyst), and therefore, control over the metal loading is limited. However, bypassing filtration and using drying directly to remove the solvent allows the maximum metal loading to be limited, but typically a broad particle size distribution is obtained.

[0004] An alternative to impregnation and drying is dry impregnation, in which case the metal precursor is deposited, and thus dispersed, on the support using very small amounts of solvent or no solvent at all to form the pre-catalyst. When using a porous support, incipient wetness impregnation is usually the optimal method, which is most suitable for maximizing the penetration of the metal precursor(s) into the pore system. Briefly, an amount of liquid metal precursor (solution or melt) equal to the pore volume of the support is impregnated via capillary forces. Unfortunately, such means often cannot achieve high loadings and may be difficult to scale up. Colloidal synthesis of alloyed nanoparticles has been mainly used to produce active metal catalysts with controlled size, shape, and composition. However, this method has little scalability and is uneconomical due to the need for high dilution and surfactant control. Co-precipitation is the most established synthetic technique for large amounts of supported alloyed catalysts. Therefore, usually, the metal precursor and the support are first dispersed in a solvent. Then, precipitation is induced by changing the pH or temperature or by adding a reducing agent. Large amounts of solvent (often organic-based) and harmful chemicals (e.g., formaldehyde, metal borates) are the drawbacks of this method. As is typical in the case of the most traditional wet chemical methods for supported bimetallic catalysts, filtration and drying typically follow. Finally, to obtain the final and most performant form of the catalyst, calcination, reduction, or other treatments are often required, independent of the method used to prepare the pre-catalyst. This accounts for a significant proportion of the cost of the catalyst and is the reason why simpler and preferably solvent-free synthesis methods are being actively developed.

[0005] Methods for preparing catalysts for various applications, including for fuel cells, are known in the prior art. For example, EP3614473A1 (Patent Document 1) discloses a method for manufacturing a catalyst for a solid polymer fuel cell using a heat-treated platinum catalyst.

[0006] Similarly, WO2021 / 156644A1 (Patent Document 2) discloses a method for preparing a catalyst for a fuel cell, and the final catalyst is represented as Pt-Fe-N-C(NH3).

[0007] EP3843186A1 (Patent Document 3) discloses a method for preparing a catalyst for a fuel cell, which has a Pt alloy catalyst having a Pt alloy (Pt-M1 alloy or Pt-M1-M2 alloy) as catalyst particles.

[0008] Furthermore, US2006 / 099483A1 (Patent Document 4) discloses a method for preparing a catalyst for a fuel cell, which includes a mixed metal catalyst containing at least one element selected from the group consisting of silicon, aluminum, and titanium and a catalyst metal precursor.

Prior Art Documents

Patent Documents

[0009]

Patent Document 1

Patent Document 2

Patent Document 3

Patent Document 4

Summary of the Invention

Problems to be Solved by the Invention

[0010] For a long time, many alternative synthetic strategies have been available to simplify procedures and thus avoid the difficulties associated with solvent-based methods. Among them, the mechanochemical method by ball milling is becoming increasingly popular as an alternative approach for the solid-phase synthesis of a wide range of materials. In fact, ball milling has already been successfully used to synthesize molecular complex compounds, coordination polymers, extended inorganic solids directly from elements, nanocomposites, and functional materials. Most importantly, mechanochemistry has also been recognized as a convenient means for the nanostructuring of solid materials for heterogeneous catalysts, including metal and metal oxide nanoparticles, porous materials, supported metal catalysts, and hybrid inorganic-organic materials. Mechanochemistry is particularly advantageous in many respects for the synthesis of solid catalysts from an industrial perspective, either in one step of its preparation or as the main stage of its preparation. In fact, ball milling is one of the simplest and cheapest techniques, since most mechanochemically induced reactions are typically carried out rapidly and efficiently using only a minimal amount of solvent or other substances as process control agents under conditions close to ambient, which also means limited ecologically harmful waste. In addition, ball milling enables the synthesis of solid catalysts with improved performance (e.g., activity and stability) in some cases due to the enhanced reactivity of the solid phase under mechanochemical conditions. Otherwise, at least a temperature decrease and simple interactions are observed during further processing (e.g., calcination, reduction, and annealing). However, these processes are difficult for the synthesis of metal nanoparticles or bi- or multi-metal nanoparticles supported on various carbons due to the mechanical properties of carbon.

Means for Solving the Problems

[0011] Here, the inventors developed a simple and potentially scalable mechanochemically assisted two-step method for synthesizing supported metal or bimetallic catalysts, such as PtNi, PtCo, PtSn, PtRh, PtIr, PtRu, and Pt-based catalysts, supported on various high surface area carriers, particularly carbon carriers. In the first step, at least one metal compound, such as metal acetylacetonate, is uniformly dispersed on the target carrier via mild ball milling. In the second step, the resulting material, i.e., the pre-catalyst, is heat-treated via follow-up by reduction with hydrogen and optionally annealing under an argon flow in a tubular furnace. The effects of various annealing temperatures and reduction on the properties of the final material were verified, and optimal conditions were found.

[0012] Therefore, the present invention relates to a method for preparing a supported metal catalyst supported on a conductive carrier, comprising, in a first step, subjecting at least one metal compound to ball milling together with an electrically conductive (conductive) target carrier selected from carbon, carbon black, titanium carbide, titanium nitride, boron nitride, niobium nitride, indium tin oxide, and antimony tin oxide, preferably from carbon, titanium nitride, boron nitride, more preferably from carbon, and, in a second step, hydrogenating the product of the first step at a temperature from 100 °C to 500 °C for a period from 1 to 3 hours under an H2 pressure from 1 to 10 atm, and, in an optional third step, annealing the product of the second step in a protective gas atmosphere at a temperature range from 400 °C to 1200 °C for a period from 2 to 12 hours.

[0013] The at least one metal compound is preferably selected from metal halides, metal nitrates, metal sulfates, metal carbonates, metal carboxylates of monocarboxylic or dicarboxylic acids having 1 to 12 carbon atoms, metal β-diketonates, metal carbonyls, metallocenes, or mixed complexes including cyclopentadienyl metal complexes, more preferably metal carboxylates of monocarboxylic or dicarboxylic acids having 1 to 12 carbon atoms, metal halides, metal nitrates, metal carboxylates, or most preferably metal carboxylates of monocarboxylic or dicarboxylic acids having 1 to 12, preferably 1 to 6 carbon atoms, or metal β-diketonates, such as acetylacetonate.

[0014] The at least one metal of the at least one metal compound is preferably selected from V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Zr, Mo, Re, Ru, Rh, Pd, Ag, Cd, Sn, Ir, Pt, Au, or mixtures thereof, preferably at least one catalytically active metal selected from Fe, Co, Ni, Ru, Pd, Ag, Rh, Ir, Sn, and Pt or mixtures thereof, more preferably Fe, Co, Ni, Ru, Rh, Ir, Sn, and Pt, or mixtures thereof.

[0015] According to the present invention, the target carrier is preferably an electrically conductive carbon material selected from carbon, carbon black, graphite, titanium carbide, or titanium nitride.

[0016] Ball milling is usually carried out in any mill, such as an attritor mill, a Simoloyer type ball mill, and a planetary ball mill, in a rotational speed range from 200 to 1400 rpm, preferably using balls in a size range from 2 to 20 mm, for a period from 15 minutes to 720 minutes, and the mill grinding container and the balls are preferably made of steel, carbide, oxide, nitride or polymer material. Ball milling is preferably carried out in a planetary mill, in a rotational speed range from 400 to 600 rpm, using balls in a size range from 5 to 10 mm, for a period from 1 to 5 hours, and the mill grinding container and the balls are made of silicon nitride.

[0017] Hydrogenation can generally be carried out in an inert gas under a hydrogen stream, using a hydrogen ratio from 5 to 50% by volume, preferably from 10 to 30% by volume.

[0018] Normally, the supported metal catalyst is composed of metal nanoparticles with an average size range from 0.1 to 20 nm, preferably from 0.5 to 10 nm, deposited on a support.

[0019] In one preferred form of the method of the present invention, a first metal acetylacetonate selected from Pd or Pt acetylacetonate, and a second metal acetylacetonate selected from Ni or Co are subjected to ball milling together with carbon, the resulting product is subjected to hydrogenation, and finally annealed. In the above aspect, the molar ratio of the first metal acetylacetonate to the second metal acetylacetonate is in the range from 1:1 to 5:1, preferably from 1:1 to 3:1. The loading amount of a metal compound such as metal acetylacetonate with respect to the support, calculated excluding the ligand with respect to the weight of the support, is preferably in the range from 0.5 to 50% by weight, more preferably from 5 to 40% by weight, and even more preferably from 5 to 15% by weight. This means that since the organic residue is removed during hydrogenation and / or annealing, the metal loading amount is the value with respect to the support of the final product.

[0020] The supported metal catalyst usually has a pore structure similar to that of the support before loading the metal nanoparticles, depending on the relative metal loading and the treatment applied, and the supported metal catalyst provides highly dispersed metal nanoparticles on the surface area of the support.

[0021] The supported metal nanoparticles exhibit the above-mentioned uniform metal distribution on the support, and when multiple metals are used in the synthesis method described in the claims, the dispersion of both metals is substantially the same.

[0022] The supported metal nanoparticles generally have a metal particle size that can be adjusted to the intended size by varying the annealing temperature in the range from 400 °C to 1200 °C, preferably in the range from 500 °C to 1000 °C.

[0023] Therefore, the present invention also relates to a supported metal catalyst that can be obtained according to any form of the method of the present invention, and preferably as an electrocatalyst, more preferably as an electrocatalyst for the cathodic oxygen reduction reaction (ORR) in a PEM fuel cell, and to the use of such a supported metal catalyst.

[0024] The success of the synthesis method reported here lies in the ability of ball milling to achieve a uniform solid mixture and thus a high dispersion level for the active components.

[0025] By applying preferred mixing conditions, when hard and soft, brittle and plastic samples are processed together, the relatively hard component can function as a grinding body with respect to the soft and plastic components, and thus it was found by the inventors that by subjecting it to activation by deformation, it is possible to coat the latter on the former. This evidence is probably the phenomenological reason for the success of the deposition of mechanochemically induced metal nanoparticles (soft) on a support such as a carbon or oxide support (hard).

[0026] Acetylacetonate has many advantageous physical properties. It is a soft molecular material characterized by a low melting point, which works favorably for the dispersion of these in combination with relatively hard components and improves the rheological properties of the solid mixture. Additionally, most metal acetylacetonates tend to decompose into metals when heated at a sufficiently high temperature (typically, a temperature exceeding the melting point) (auto-reduction). In this sense, materials composed of highly dispersed carbon-supported metal acetylacetonates would be a promising starting point (pre-catalyst) for the preparation of carbon-supported metal nanoparticles upon heating. Along these lines, the bimetal formulation of these materials can be obtained simply by using different metal acetylacetonates simultaneously.

[0027] The present invention will be further described with reference to the accompanying drawings and the following experimental section.

Brief Description of the Drawings

[0028]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

[0029] For the synthesis of the materials of the present invention, jars and balls made of silicon nitride were used exemplarily. However, the use of any other grinding material for the jars and balls is generally possible. In particular, in order to improve mixing, a 45 mL volume jar and a number of small grinding balls were used, taking into account the previously reported guidelines on how to maximize the mixing (impact) rate in a planetary ball mill. Exemplarily, Pt(acac)2, Ni(acac)2, and Co(acac)3 can be used as metal precursors, case by case, individually or in combination. Co(acac)3 was preferred over Co(acac)2 because it is known not to release metal-containing fragments upon heating. Carbon black was selected for the role of supporting the materials in all cases. As expected, various relative metal:metal ratios were verified when investigating the bimetal formulation. However, the total metal loading of interest is usually maintained per a fixed value (10 wt% relative to carbon), and the preparation scale is the same (450 mg carbon carrier and metal acetylacetonate).

[0030] Typically, an exemplary solid mixture comprising a metal compound such as metal acetylacetonate and a carbon support was milled at 600 rpm for 3 hours, a time selected to ensure maximum dispersion of the metal precursor on the carbon support. Depending on the system to be synthesized, other milling conditions may be desirable. In fact, in the case described, none of the typical XRD reflections of the acetylacetonate precursor were observed after 3 hours of treatment (Figure 1a). This suggests that the metal precursors were successfully dispersed on the carbon support at a scale low enough to prevent their detection via XRD. This insight was further supported by directly observing the distribution of metal species on the carbon support with an electron microscope, as reported for representative cases of (Pt,Co)-acac / CB (Figure 1b) and (Pt,Ni)-acac / CB (Figure 1c). Presumably, a shorter milling time may be sufficient to achieve similar results, but this option was not verified in detail.

[0031] Interestingly, the thermal analysis (TG-DSC) of the samples showed a reproducible decomposition pattern for all samples (Figure 2). Generally, carbon-supported metal acetylacetonates seem to decompose in one step when heated above 150 °C under Ar flow (TG). This suggests that the metal acetylacetonates are intimately mixed throughout the material. This is a promising result, since both the uniform mixing and distribution of the metal precursors will likely act favorably towards the bottom-up synthesis of (supported) alloy nanoparticles via subsequent heat treatment. On the other hand, the weight loss measured at 900 °C is advantageously different from what would be expected for just the pyrolysis of the metal complex (ICP-OES analysis of the samples as reference (Table 1)) in most cases. Presumably, the high-temperature gasification of the carbon support by the action of the metal component, which is a well-known gasification catalyst for all, is responsible for the additional weight loss observed for all samples. In line with these considerations, an increase in the total loading should be observed upon heating at high temperature.

[0032] Specifically, annealing at approximately 950 °C for 7 hours while flowing Ar induces the formation of alloy species, as clearly shown from the XRD pattern (Figure 1a). The XRD reflections typical of these metal species are now very distinct in clear contrast to the corresponding precursors before heat treatment. Most importantly, a systematic shift to larger 2Θ angles from Pt was observed for most of the reflections associated with bimetal formation, depending on the nature of the second metal, the relative amount with respect to Pt, and the ordering within the alloy. This is particularly evident for the reflections determined by diffraction from the (1 1 1) plane of the metal species, which is the strongest reflection in all cases. However, due to the formation of ordered intermetallic Pt x Ni y and Pt x Co y phases, additional reflections were also observed. As an addendum, typically, a regular-irregular transition is observed when the system is first heated above the transition temperature (e.g., 620 °C for bulk PtNi alloy, 825 °C for PtCo), then quenched from the disordered state (below the transition temperature), annealed for a long time, and finally cooled to room temperature at an appropriate cooling rate. Both the annealing and cooling steps are extremely important to obtain the ordered intermetallic phase. Perhaps a fortunate combination of annealing temperature and cooling rate was responsible for the accidental formation of the PtNi and PtCo ordered phases in our case. These findings, although preliminary, suggest that the composition of the alloy species can be fine-tuned via the preparation method described herein. Similarly, ordering can potentially be induced by appropriate annealing and cooling procedures. As expected, a moderate increase in the metal loading was observed due to the high-temperature heat treatment.

[0033] The formation of nano-sized alloy species was confirmed by direct observation of the nanoparticles with an electron microscope. In particular, the effect of heat treatment on the overall metal distribution is shown in Figures 1b-c and 1f-g using the examples of PtNi / CB-950 and Pt-Co / CB-950, respectively. In fact, high-temperature heating resulted in the formation of alloy nanoparticles in which Pt and Ni or Co were densely mixed as shown in the EDX elemental maps (Figures 1d-e and 1h-i). Furthermore, despite the use of elevated annealing temperatures, particles with an average size (diameter) between 4 nm and 5 nm and a standard deviation of less than 1.5 nm were obtained in all cases. More precisely, with the exception of PtCo / CB-950, all systems showed a reproducible metal particle size distribution regardless of the specific combination of the metals used and their relative amounts in the alloy. In fact, in the PtCo system, there seems to be a higher probability of forming particles with a size of approximately 3 nm, which resulted in an average particle size of about 4 nm, slightly smaller than in other cases.

[0034] It is important to note that the synthesis procedure, including mechanochemical mixing and subsequent heat treatment, did not seem to affect the characteristics of the carbon support established by verifying the N2 physisorption behavior of the samples. Generally, the N2 physisorption isotherms are indeed similar regardless of the treatment to which the material was subjected (Figure 3).

[0035] Table 1 shows an overview of the characterization results for the 10 wt% Pt of the materials obtained from the (xPt,yM)-acac / CB precursors (M = 0, Ni, Co) after annealing at 950 °C x M y / CB-R950 series. These results include the elemental composition via ICP-OES and SEM-EDX analysis, the texture characteristics (surface area - SA, external surface area - SA ext , pore volume - V p,micro , total pore volume - V p,tot ) determined from N2 physisorption experiments, the average metal particle size (d M ) and standard deviation (σ dIt includes (xPt,yM)-acac / CB. For (xPt,yM)-acac / CB, the metal content determined via ICP-OES analysis is reported as the confirmed value and as the corrected value (in parentheses) for the presence of organic ligands.

[0036]

Table 1

[0037] The above synthesis method was repeated using a stainless-steel grinding jar and balls under similar conditions. However, ball milling of metal acetylacetonates with carbon black functioned as expected by achieving a high level of dispersion of the metal precursors, and subsequent heat treatment led to the formation of alloy species intermixed with Fe resulting from wear of the grinding jar and balls in all cases, as clearly indicated by the powder XRD patterns of the resulting materials. Therefore, limited control should be exerted over the entire composition of the alloyed species, as various amounts of Fe were found instead in the samples (1.6 to 2.3 wt%).

[0038] In any case, by using a silicon nitride jar, a material (pre-catalyst) free from contamination from additional catalytically active components (Si from 0.6 to 0.9 wt% or Si3N4 from 1.0 to 1.5 wt% as determined by SEM-EDX analysis) can be easily obtained. Subsequently, an appropriate heat treatment can be carried out to yield the final catalyst material. As described above, a long-term anneal at 900 °C to 1000 °C, particularly at 950 °C, under an inert gas atmosphere, which was the condition initially selected to ensure alloying of the metal species during heating, did indeed result in Pt, Pt x Ni y 、and Pt x Co yresulted in the formation of nanoparticles. As expected, an overall metal loading of about 10 wt% was used as a typical example, although higher metal loadings were also investigated. In particular, PtNi-based catalysts were successfully obtained at higher metal loadings, namely 20 and 40 wt%, as pre-demonstrated from the XRD characterization of the samples (Figure 4). However, despite the nominal composition and promising material characteristics, the catalyst was generally recognized to be poorly active towards ORR in an RDE setup. More appropriate heat treatment is required to reach the most performant state for the catalyst material.

[0039] For this purpose, a reduction step at 220 °C in hydrogen was first introduced. Subsequently, the influence of the annealing temperature (7 hours after reduction) was verified for the Pt-Ni system in the range from 600 °C to 850 °C. Specifically, as seen from the XRD characterization of the material (Figure 5), the alloying of Pt and Ni is not complete except at relatively high temperatures within a given time. In fact, approximately symmetric reflections (supporting the statistical distribution of the Pt-Ni mixing pattern per nominal composition) are only observed for the materials annealed at 750 °C and 850 °C for 7 hours. Nevertheless, the two materials are not the same, with one being annealed at a higher temperature, featuring on average larger metal nanoparticles, as indicated by the sharper XRD reflections of the PtNi phase. In contrast, annealing at temperatures from 600 to 700 °C was clearly insufficient to reach a steady state for the alloying of Pt and Ni, as the contributions from Pt and Ni could be easily detected in the XRD pattern. As an addendum, post-reduction annealing at 750 °C and 850 °C resulted in the formation of a non-regularized PtNi phase (PtNi fcc solid solution) in both cases and did not result in the regularized intermetallic phase as observed previously.

[0040] For a deeper insight into the influence of heat treatment on the characteristics of the materials, a set of nine materials was constructed considering several aspects for the Pt-Ni system and thoroughly characterized. Specifically, the following series of materials were examined: (a) Pt(M) / CB-R obtained from the reduction of (1Pt,1M)-acac / CB (M = 0, Ni, Co) at 220 °C using hydrogen; (b) Pt(M) / CB-R750 obtained after reduction at 220 °C in hydrogen and annealing at 750 °C in argon; (c) Pt(M) / CB-R850 obtained after reduction at 220 °C in hydrogen and annealing at 850 °C in argon. The results from the characterization of the materials are summarized in Table 2.

[0041] In Table 2, an overview of the characterization results for 10 wt% PtM / CB-RXXX of the materials obtained from (1Pt,1M)-acac / CB (M = 0, Ni, Co) after reduction at 220 °C (R), reduction and annealing at 750 °C (R750), or reduction and annealing at 850 °C (R850) is shown. These results include the elemental composition via ICP-OES analysis, the texture properties (surface area - SA, external surface area - SA ext , pore volume - V p,micro , total pore volume - V p,tot ) determined from N2 physisorption experiments, the number-average metal particle size (d M ) and standard deviation (σ d ) from electron microscopy verification.

[0042]

Table 2

[0043] As shown in Figs. 6a, 6e, and 6o, the XRD patterns of the materials obtained after reduction at 220 °C in hydrogen are mostly dominated by broad reflections from the carbon support. However, in contrast to the as-ground powder (Fig. 1a), additional features can be easily detected at positions typical for Pt and the corresponding alloys, indicating the presence of metal species represented by small coherent scattering domains. Furthermore, specifically in the case of Pt / CB-R, weak but clearly sharp reflections are seen at positions consistent with Pt metal, which would imply a heterogeneous metal particle size distribution containing relatively large particles with low occupancy. The previous observations were corroborated by direct observation of the metal distribution under the electron microscope (Figs. 6b, 6f, and 6p). Generally, particles of about 1 nm were mainly observed, although large particles up to 10 nm were occasionally seen. Due to the high dispersion of the metal species, the strong interaction between the electron beam and the metal nanoparticles hindered the recording of significant EDX maps for PtNi / CB-R and PtCo / CB-R. Based on these premises, the EDX line scans measured for the same samples are considered to be more reliable. In any case, these data are consistent and, on a case-by-case basis, indicate the presence of highly dispersed Pt species and Ni or Pt species and Co species.

[0044] Instead, after annealing, the powder XRD patterns show the characteristic reflections of Pt, PtNi (non-ordered), and PtCo (ordered) as expected (Fig. 6e). After annealing at 750 °C and 850 °C, no other metal phases could be easily detected besides the target species. As before, the materials annealed at 750 °C and 850 °C (after reduction) were not identical. Narrower reflections of the metal components were confirmed with increasing temperature, which is consistent with the sintering of the metal nanoparticles. In fact, regardless of the nominal composition, an average particle size of about 3 nm was measured for Pt(M) / CB-R750 and about 4 nm for Pr(M) / CB-R850. Most importantly, according to the EDX elemental maps, Pt and Ni or Pt and Co were found to be in the same spatial positions instead, which supports the formation of alloy species (Figs. 6i-k and 6s-u).

[0045] To characterize the as-synthesized materials, powder X-ray diffraction, electron microscopy, nitrogen sorption, and thermal analysis were used. Using this method, several PtNi- and PtCo-based catalysts for ORR were synthesized using carbon black as the support. In addition, for comparison purposes, materials containing only Pt were also synthesized. The method was found to be flexible with respect to the total metal loading (e.g., 10, 20, and 40 wt%) and the metal:metal relative composition (e.g., 1 and 3 of Pt:M (M = Ni or Co)), enabling the synthesis of a wider set of materials. An important feature of the method is the extraordinary level of dispersion achieved during ball milling of the metal precursors, which results in a relatively narrow size distribution of the supported alloy nanoparticles (average size less than 5 nm). Most importantly, the materials were tested for ORR on a rotating disk electrode (RDE), and in the case of PtNi supported on carbon black, a specific activity (SA) of 1.2 mA cm -2 Pt and a mass activity of 0.9 A mg -1 Pt were demonstrated to be highly active.

[0046] All materials were also verified by N2 physisorption. As described previously, the porosity of the carbon support was not affected by either ball milling or subsequent heat treatment, and as a result, similar N2 physisorption isotherms were obtained independent of the specific heat treatment applied to the materials.

[0047] As shown in Figure 7a, after reduction at 220 °C, all materials had an approximately 150 m 2 g -1 PtIt is characterized by a high value of specific surface area (relative value with respect to the amount of Pt). These results are consistent with an average metal particle size of approximately 1 nm confirmed for all materials, regardless of the nominal composition, and further support the high degree of metal dispersion achieved during ball milling. However, correspondingly, the activity of all catalysts in the ORR is relatively weak, which is likely due to the small metal particle size and possibly incomplete alloying. The effect of metal particle size on ORR activity has been a long-standing issue in the literature. Understanding it has been mainly prevented by differences in the conditions under which electrochemical analysis was performed, the average interparticle distance, the shape of individual particles, or other reasons. Nevertheless, there is strong evidence that a particle size of less than 2 nm would lead to a decreased ORR activity in the case of Pt. For example, experimental findings supported by DFT calculations showed that the specific activity decreased by a factor of 4 when the average particle size decreased from 2.2 nm to 1.3 nm. This effect was attributed to an increase in the expected oxygen binding strength as the Pt particle size decreased. Nevertheless, regarding stability, there is a clear consensus that the stability of Pt particles less than 2 nm is insufficient for use in PEMFCs and has a higher tendency to dissolve under typical reaction conditions.

[0048] After annealing at 750 °C, the average particle size is approximately 70 m 2 g -1 Pt (as reflected by the ECSA of the material in (Fig. 7b)) increases to approximately 3 nm. Therefore, for Pt- and PtNi-based catalysts, high specific activities were measured. That is, in the case of Pt, it is approximately 0.49 mA cm -2 Pt and in the case of PtNi, it is 1.2 mA cm -2 Pt On the other hand, although PtCo / CB-R750 had similar ECSA values determined, it had a relatively low ORR activity, i.e., 0.64 mA cm -2 Ptshowed ORR activity. This would suggest a partial alloying of Pt with Co and thus an incomplete formation of the most active PtCO phase. Along these lines of thought, in addition to the contribution from the crystal size effect, which was initially assumed to be dominant, the high dispersion of the Pt x Co y species could have been responsible for the broad diffraction lines (Figure 6o) observed for each Pt x Co y phase. This idea is supported by the EDX maps and line scans measured across the samples, which report strongly fluctuating levels of Pt and Co.

[0049] Presumably, an annealing temperature as high as 850 °C is more suitable for the formation of the PtCo intermetallic phase, which is consistent with the Pt-Co phase diagram. Indeed, the order-disorder transition in the case of the PtCo bulk alloy is expected at a temperature as high as 825 °C, as described previously. Therefore, enhanced ORR activity was observed in the materials annealed at 850 °C (Figure 7c). In particular, for PtCO / CB-R850, although the ECSA was within the same range as that of PtCo / CB-R750 (Table 3), a relatively high specific ORR activity of about 1.1 mA cm -2 Pt was measured. In contrast, PtNi / CB-R850 and Pt / CB-R850 seemed to be less active than their corresponding counterparts annealed at 750 °C. However, in the case of the Pt-Ni system, the decreased activity could be directly attributed to an increase in the average metal grain size. In the case of Pt / CB-R850, this same evidence cannot be explained by just the sintering of the metal nanoparticles, since the average Pt grain size and the ECSA value do not match, i.e., the average grain size of about 3.8 nm in the case of Pt / CB-R850 does not match the relatively low ECSA of 27 m 2 g -1 Pt 2. Presumably, the carbon restructuring at such high temperatures implies the formation of a carbon coating layer around the metal nanoparticles, greatly reducing the surface area available for ORR.

[0050] Table 3 shows the ORR activity (activity as specific activity - SA and mass activity - MA) and the electrochemically active surface area (ECSA) of 10 wt% Pt(M) / CB-950 (M = 0, Ni, Co) after activation.

[0051]

Table 3

[0052] In summary, these highly active systems were obtained via a mechanochemically assisted synthesis procedure by simply varying the annealing temperature after reduction. In particular, Pt / CB-R750, PtNi / CB-R750, and PtCO / CB-R850 showed activity on the high side of the spectrum of values reported in the literature. For example, the 1.2 mA cm -2 Pt SA and 0.9 A mg -1 Pt MA measured for PtNi / CB-R750 are rarely exceeded, except for systems with shape or morphology control where stability is often an issue. In contrast, the catalysts presented here do not have a controlled morphology and thus are similar to industrially established PEMFC catalysts, and as a result, similar stability can be expected. In fact, exemplary stability was observed for PtCo / CB-R850, which supports the expected resistance to degradation in the normal potential range (Figure 8). To confirm the high activity of PtNi / CB-R750, further, a complete resynthesis and all characterization techniques, including electrochemical evaluation, were performed (Figure 9). As expected, the differences between these two materials were negligible, and reproducible preparation was demonstrated. Finally, although Pt(M) / CB-R (M = 0, Ni, Co) shows low activity and stability in PEMFC use, the existence of such a simple bottom-up approach for the synthesis of nanoparticles with a size of about 1 nm is important for other catalyst applications.

[0053] The solid-state based method reported here, which involves mechanochemical mixing and subsequent heat treatment, enables the dry synthesis of supported bimetallic catalysts over a wide range of compositions. An important feature of the method is the extraordinary level of dispersion achieved during ball milling of the metal precursor(s), which ultimately led to the formation of a relatively narrow size distribution of the supported alloy nanoparticles during subsequent heat treatment. Most importantly, the resulting metal nanoparticles are remarkably pure in their metal composition, because unalloyed species could not be detected by either XRD or TEM-EDX verification. For each desired composition, i.e., PtM / CB (M = Ni, Co), highly active ORR catalysts were obtained simply by varying the annealing temperature after reduction.

[0054] Because of its simplicity and flexibility, the inventors expect that the synthesis method will be used beyond the field of ORR catalysts in a short period of time. In fact, the inventors see no reason why the co-preparation method should not be applied to the synthesis of supported nanoparticles including other suitable combinations of metals and supports and used for other different applications.

Examples

[0055] Characterization method Powder X-ray diffraction X-ray diffraction (XRD) data were recorded on a STOE STADI P diffractometer operating in Bragg-Brentano geometry (CuKα 1,2 : 1.541862 Å) with a secondary graphite monochromator. Samples were placed on a silicon background-free sample holder, and data were collected continuously in the 15 to 75° 2Θ range at a scan rate of 0.6° per minute -1 (step size, 0.1°) using a proportional gas point detector. The measured patterns were qualitatively evaluated by comparison with entries in the ICDD PDF-2 database.

[0056] Thermal analysis Simultaneous thermal analysis (thermogravimetric analysis TG and differential scanning calorimetry DSC) coupled with mass spectrometry (MS) was performed using a NETZSCH STA449F3 Jupiter thermal analyzer connected to a NETZSCH QMS403D Aeros mass spectrometer. A sample (about 10 mg) was heated from 40 to 900 °C in a continuous flow of Ar (40 mL min -1 ) (heating rate, 10 °C min -1 ).

[0057] Nitrogen physical adsorption Nitrogen adsorption isotherms were recorded using a Micromeritics 3Flex instrument. The apparent specific surface area was determined using the Brunauer–Emmett–Teller (BET) algorithm. The t-plot method was used to determine the pore volume and external surface area. Carbon black STSA was used as the thickness curve. The total pore volume was determined using Gurvich's rule at 0.95 p / p0. All methods used were comprehensively reviewed in the literature.

[0058] Electron microscopy High-angle annular dark-field scanning transmission electron (HAADF-STEM) microscopy and energy-dispersive X-ray spectroscopy (EDX) elemental maps were acquired using a spherical aberration-corrected Hitachi HD-2700 microscope equipped with a cold field emission gun and two EDAX Octane T Ultra W EDX detectors at an acceleration voltage of 200 kV. In addition, conventional high-resolution transmission electron (HR-TEM) micrographs were collected using the same machine. Finally, scanning mode imaging was also performed. Samples were typically prepared by spreading a dried specimen on a TEM grid. In all cases, the particle size distribution was determined by evaluating the diameters of at least 200 particles from multiple images of the same sample. The elemental composition was determined via energy-dispersive X-ray (EDX) bulk analysis performed using a Hitachi TM3030PLUS table-top scanning electron microscope (SEM) equipped with an Oxford Instruments Explorer Compact 30 detector and operated at an acceleration voltage of 15 kV. All samples for SEM-EDX measurements were prepared by spreading a dried specimen on a C film.

[0059] Elemental analysis (ICP-OES) Inductively coupled plasma optical emission (ICP-OES) analysis was performed on a Spectro Green FXM46 using a UVPlus Optic ORCA (Optimized Rowland Circle Alignment) optical system. The emission wavelengths for each metal used are summarized in Table 4. The loading amount was determined by averaging the concentrations evaluated using different wavelengths.

[0060] Table 4: Summary of wavelengths used to evaluate metal concentrations during ICP-OES measurement

[0061] [Table 4]

[0062] To prepare the liquid sample for measurement, approximately 10 mg of the catalyst was leached in a mixture of 7.5 mL of HCl, 2.5 mL of HNO3, and 0.8 mL of H2O2. This suspension was heated in an Anton Paar Multiwave 5000 microwave oven at 180 °C for 5 minutes and then allowed to cool naturally to ambient temperature. It was then diluted to 50 mL with mQ water and stirred overnight before analysis to evaluate the metal concentration. Throughout the procedure, the sample weight was measured using a METTLER TOLEDO XA205 dual range and a Sartorius Entris II.

[0063] Materials Platinum(II) acetylacetonate and cobalt(III) acetylacetonate were purchased from Alfa Aesar. Nickel(II) acetylacetonate and acetylacetone were purchased from Sigma-Aldrich. Vulcan XC-72R was purchased from Cabot. All chemicals were used as received without further purification.

[0064] Methods Synthesis of (Pt,M)-acac / C (M = 0, Ni, Co) All preparations were carried out using a Pulverisette P7 (Classic Line) planetary ball mill from Fritsch. Carbon black (CB), Pt(acac)2, and optionally a second metal acetylacetonate precursor, either Ni(acac)2 or Co(acac)3, were charged into a 45 mL silicon nitride jar. Typically, the synthesis aimed for a final total metal loading of 10 wt% (excluding the organic ligand). Therefore, 450 mg of CB was used and combined with the exact amount of metal acetylacetonate. When Pt(acac)2 was combined with Ni(acac)2 or Co(acac)3, Pt:M molar ratios of 1 and 3 were alternatively used in each case. The solid mixture was then milled at 600 rpm for 3 h using 80 silicon nitride grinding balls with a diameter of 5 mm (total weight 17 g). Several preparations were repeated under similar conditions, including the number and size of the grinding balls (80 stainless steel balls with a diameter of 5 mm and a total weight of 35 g), the rotation speed (600 rpm), and the milling time (3 h), using a 45 mL stainless steel jar. For syntheses where a total metal loading of 20 or 40 wt% (excluding the organic ligand) was alternatively targeted, the same amount of grinding balls (80 silicon nitride balls with a total weight of 17 g) was used and milling was carried out at 500 rpm for 12 h. In all cases, the milling program periodically included the same two steps, i.e., first milling for 15 min at the selected rotation speed and then a 5 min pause, repeated until the desired milling time had accumulated. The rotation was reversed when moving from one repetition to the next to improve the uniformity of the treatment. At the end of this milling program, the material was scraped from the milling jar and thus recovered in a nearly quantitative yield. Blank experiments were also carried out with or without acetylacetone (acacH, 50 μL) to simulate the presence of the metal acetylacetonate in order to verify the effect of milling on the bare support (500 mg of CB). All preparations were repeated at least twice to evaluate reproducibility.

[0065] Synthesis of PtM / C (M = 0, Ni, Co) (Pt,M)-acac / C (M = 0, Ni, Co) was transferred to a suitable crucible and subjected to heat treatment in a tubular furnace. Typically, this heat treatment included the following: (a) reduction in a hydrogen stream (20 vol% H2 in Ar at 200 mL / min -1 ) for 90 minutes at 220 °C (heating rate, 3 °C / min -1 ), and (b) annealing in an argon stream (Ar at 100 mL / min -1 ) for 7 hours at 750 °C or 850 °C (heating rate, 5 °C / min -1 ). However, occasionally, this verification alternatively required using only reduction or thermal annealing to prepare the samples. Additionally, alternative annealing temperatures such as 600 °C, 650 °C, 700 °C, and 950 °C were investigated while keeping all other parameters the same. In each case, at the end of the treatment, the system was allowed to cool naturally to room temperature. Finally, before recovery, all materials were passivated with a mixture of increasingly higher concentrations of synthetic air in argon as follows: (a) 1 vol% synthetic air in Ar for 30 minutes at 100 mL / min -1 ; (b) 5 vol% synthetic air in Ar for 30 minutes at 100 mL / min -1 . All preparations were repeated at least twice to evaluate reproducibility. Considering the different conditions applied to the synthesis of the materials via thermal annealing, including combinations of steps and annealing temperatures, the materials were designated using the symbol PtM / CB-RXXX (M = 0, Ni, Co), where R represents whether they were reduced (or not reduced if R is absent), and XXX represents the annealing temperature. For example, PtNi / CB-R750 specifies PtNi nanoparticles supported on carbon black that were reduced as described above and then annealed at 750 °C for 7 hours. Similarly, PtNi / CB-950 was not reduced and was directly annealed in argon at 950 °C for 7 hours.

[0066] Synthesis of PtRh / CB-R1000 For the synthesis of Pt / CB-R1000, Pt(acac)2 and Rh(acac)3 were ball-milled at 600 rpm for 3 hours using a silicon nitride jar and balls in equimolar amounts (10 wt% Pt and Rh excluding ligands). The milling was carried out under the same conditions to obtain (Pt,Ni)-acac / CB. Next, this milled powder was reduced with hydrogen (20% H2 in Ar at 200 mL / min; 220 °C, 3 °C / min, 90 min), annealed in an argon stream (100 mL / min; 1000 °C, 5 °C / min, 10 h), and then passivated in a tubular furnace. According to the XRD pattern shown in Figure 10, this process probably results in Pt-Rh alloyed species with a relative composition close to the target value, supported on CB.

[0067] Synthesis of PtIr / CB-R850 For the synthesis of PtIr / CB-R850, milling was carried out under the same conditions to obtain PtRh / CB-R1000, except that [Ir(cod)(acac)] was used as the iridium precursor. Next, this milled powder was reduced with hydrogen (20% H2 in Ar at 200 mL / min; 220 °C, 3 °C / min, 90 min), annealed in argon (100 mL / min; 850 °C, 5 °C / min, 12 h), and then passivated in a tubular furnace. According to the XRD pattern, this process probably results in Pt-Ir alloyed species with a relative composition close to the target value, supported on CB.

[0068] Synthesis of PtM / CB-R1000 (M = Sn, Ru) For the synthesis of PtM / CB-R1000 (M = Sn, Ru), milling was carried out under the same conditions to obtain PtRh / CB-R1000. According to this case, the precursors of the second metal were Sn(OAc)2 (representing tin(II) acetate) and Ru(acac)3. Pt(acac)2 and the second metal precursor were mixed in equimolar amounts (10 wt% Pt + M excluding ligands). After milling, this material was reduced with hydrogen (20% H2 in Ar at 200 mL / min; 220 °C, 3 °C / min, 90 min), annealed in argon (100 mL / min; 1000 °C, 5 °C / min, 10 h), and then passivated in a tubular furnace. According to the XRD pattern shown in Figure 11, this process results in the formation of alloyed species (main phase) with a relative composition close to the target values in the cases of Pt-Sn and Pt-Ru. However, monometallic species, especially Pt in the case of Pt-Sn and Pt and Ru in the case of Pt-Ru, were also detected.

[0069] Synthesis of PtNi / TiC-R750 and 5 wt% PtNi / TiN-R750 For the synthesis of 10 wt% PtNi / TiC-R750 and 5 wt% PtNi / TiN-R750, milling was carried out as described above using a silicon nitride jar and balls. Pt(acac)2 and Ni(acac)2 were mixed in equimolar amounts with TiC and TiN (5 - 10 wt% Pt + Ni excluding ligands). After milling, this material was reduced with hydrogen (20% H2 in Ar at 200 mL / min; 220 °C, 3 °C / min, 90 min), annealed in argon (100 mL / min; 750 °C, 5 °C / min, 7 h), and then passivated in a tubular furnace. According to the XRD pattern shown in Figure 12, this process results in the formation of highly dispersed Pt and Ni species, although some alloying was observed. Elemental mapping via electron microscopy showed that Pt and Ni form separated supported nanoparticles.

[0070] Electrochemical measurements The specific activity (SA) and mass activity (MA) for ORR, as well as the electrochemical surface area (ECSA), were determined from electrochemical half-cell measurements. These measurements were performed in a three-electrode compartment Teflon cell using a rotating disk electrode (RDE) setup equipped with either a Radiometer Analytical rotating controller or a Radiometer from Orion Research, Inc., and a Gamry Reference 600 potentiostat. A graphite rod was used as the counter electrode, and a double-junction Ag / AgCl (3M KCl, Metrohm) electrode was used as the reference electrode. The single cell compartment of the reference electrode was additionally separated using a Nafion membrane to avoid chloride ion leaching during activity and stability measurements. All potentials in this work are values relative to the reversible hydrogen electrode (RHE). To ensure accurate correction of the Ag / AgCl electrode potential, the potential was pre-measured with the help of a Pt electrode in a H2-saturated electrolyte. This procedure was repeated for all activity and stability measurements using the Pt-based ORR catalyst under test as the RHE.

[0071] The catalyst film was first prepared by dispersing the catalyst powder in ultrapure water (VWR) using an ultrasonic vial processor. Then, after an initial treatment of 30 minutes (10 minutes if the dispersion is reused), 20 μL of the ink droplet was pipetted onto a freshly polished glassy carbon disk electrode with a diameter of 5 mm (geometric surface area of 0.196 cm 2 ), embedded in a Teflon shell, and finally dried in air. The amount of catalyst on the electrode was in the range of 5 - 20 μg cm -2 as the platinum loading on the carbon support and the dispersion (see Table 5). Thereby, a reproducible film thickness was obtained, ensuring suitable analysis and comparable results. For both activity and stability measurements, 0.1 M HClO4 (prepared by diluting 70% HClO4 (ultrapure from Carl Roth) with ultrapure water) was used as the electrolyte. The resistance drop was corrected via positive feedback to have a residual uncompensated resistance of less than 3 Ω.

[0072] Before each activity and stability measurement, the catalyst was subjected to 200 - 300 cleaning cycles (0.05 - 1.0 V or 0.05 - 1.4 V) in N2 - saturated electrolyte until a stable cyclic voltammogram (CV) was recorded. During this step, carbon impurities are removed from the surface. For more details on activation, see Table 5. The ORR activity was calculated from the anodic scan of the CV recorded in O2 - saturated electrolyte at a scan rate of 50 mV s -1 −1 and a rotation rate of 0.9 V / RHE and 1600 rpm. The CV was corrected for capacitive processes by subtracting the CV recorded in argon to isolate the current related to oxygen reduction. The ORR activity measurements were carried out according to the guidelines reported in the literature. The specific activity (SA) is shown as the kinetic current density normalized with respect to the actual surface area, derived from the electrochemical CO oxidation charge. For this, the peak of the CO stripping CV was integrated after subtracting the CV recorded in N2 - saturated electrolyte to correct for non - faradaic contributions. The electro - chemical surface area (ECSA) was calculated assuming a surface charge density of 380 μC cm -2 Pt −2. The mass activity (MA) was derived from SA and ECSA. The stability measurements consisted of an accelerated degradation protocol based on voltage cycles of 10800 CV in total at 1 V s -1 −1 between 0.4 V and 1.0 V / RHE. The degradation tests were carried out in N2 - saturated electrolyte without rotation, and the ECSA was determined via CO stripping by comparing the results before and after each of 360, 1080, 2160, 3600, 5400, 7200, and 10800 load cycles to monitor the change in charge at the surface.

[0073] Table 5: Catalyst loading on the electrode as a function of the average metal particle size (d M p) and Pt loading (ICP - OES). For activation via CV, the lower limit of the potential was 0.05 V / RHE. The upper limits of the potential are shown respectively.

[0074]

Table 5

[0075] As shown above, the method of the present invention for preparing supported metal catalysts, particularly bimetallic supported catalysts, results in the catalysts of the present invention having high stability and unique electronic and structural properties. The inventors propose a dry, potentially scalable and two-step method assisted by mechanochemistry as an alternative synthetic route to supported metal catalysts. First, the metal salts are dispersed on the support in a planetary ball mill. Next, this mixture is reduced and annealed to produce supported alloyed nanoparticles. By applying this procedure to the synthesis of PtM / C (M = Ni, Co), which are typical catalysts for the oxygen reduction reaction, the inventors demonstrate that the particle size, composition and loading can be finely tuned, ultimately resulting in very high performance catalysts.

Claims

1. A method for producing a supported metal catalyst, comprising: firstly, subjecting at least one metal compound to ball mill grinding with an electrically conductive support selected from an electrically conductive carbon material such as carbon, graphite, carbon black, or titanium carbide, or from titanium nitride, boron nitride, niobium nitride, indium tin oxide, and antimony tin oxide, preferably from carbon, titanium nitride, boron nitride, and more preferably from carbon; and secondly, heating the product from the first step to H1 to 10 atm. ２ The method comprising hydrogenating under pressure at a temperature range of 100°C to 500°C for a period of 1 to 3 hours, and in an optional third step, annealing the product of the second step at a temperature range of 400°C to 1200°C for a period of 2 to 12 hours under a protective gas atmosphere.

2. The method according to claim 1, wherein the at least one metal compound is selected from metal halides, metal nitrates, metal sulfates, metal carbonates, metal carboxylates of carboxylic acids having 1 to 12 carbon atoms, metal β-diketonates, metal carbonyls, metallocenes, or cyclopentadienyl metal complexes, preferably from metal carboxylates of monocarboxylic acids or dicarboxylic acids having 1 to 12 carbon atoms, particularly 1 to 6 carbon atoms, metal halides, metal nitrates, metal carboxylates, or more preferably from metal carboxylates of monocarboxylic acids or dicarboxylic acids having 1 to 12 carbon atoms, particularly 1 to 6 carbon atoms, or metal β-diketonates, such as acetylacetonate.

3. The method according to claim 1 or claim 2, wherein the metal of the at least one metal compound is selected from V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Zr, Mo, Re, Ru, Rh, Pd, Ag, Cd, Sn, Ir, Pt, Au, or mixtures thereof, preferably from Fe, Co, Ni, Ru, Pd, Ag, Rh, Ir, Sn, and Pt or mixtures thereof, and more preferably from Fe, Co, Ni, Ru, Rh, Ir, Sn, and Pt, and from at least one catalytically active metal.

4. The method according to claim 1 or 2, wherein the loading rate of the metal compound on the carrier is in the range of 0.5 to 50% by weight, preferably 5 to 40% by weight, and more preferably 5 to 15% by weight, calculated relative to the carrier and excluding the ligand.

5. The method according to claim 1 or 2, wherein the electrically conductive carrier is an electrically conductive carbon material selected from carbon, graphite, carbon black, titanium carbide, or titanium nitride.

6. The method according to claim 1 or 2, wherein the ball milling is performed in any ball mill machine from among an attritor mill, a Simoroyer type ball mill, and a planetary ball mill, at a rotational speed range of 200 to 1400 rpm, using balls in a size range of 2 to 20 mm, for a period of 15 minutes to 720 minutes, and the ball mill and balls are preferably made of steel, carbide, oxide, nitride, or polymer material.

7. The method according to claim 1 or 2, wherein the hydrogenation is carried out under a hydrogen stream in an inert gas having a hydrogen ratio of 5 to 50 volume%, preferably 10 to 30 volume%,.

8. The method according to claim 1 or 2, wherein a metal acetylacetonate selected from Pd or Pt acetylacetonate is subjected to ball milling with carbon, the resulting product is subjected to hydrogenation, and finally to annealing.

9. A supported metal catalyst that can be obtained according to the method of claim 1 or 2.

10. Use of the supported metal catalyst according to claim 9 as a catalyst, preferably as an electrocatalyst, and most preferably for the cathode oxygen reduction reaction (ORR) in a PEM fuel cell.

Images