NANOSTRUCTURED CORE-SHELL ELECTROCATALYSTS FOR FUEL CELLS

-

Catalytic layers for use in the electrodes of fuel cells including a non-noble metal substrate layer coated with one or a few monolayers of noble metal, such as Pt. These thin, highly porous structures with large catalytically active surface areas, should exhibit a significantly higher power output per mg of Pt and per cm2 of the membrane than the current Polymer Electrolyte Fuel Cells catalytic layers.

Skip to: Description  ·  Claims  · Patent History  ·  Patent History
Description
INCORPORATION BY REFERENCE

This divisional application claims the priority benefit of U.S. application Ser. No. 11/267,737, filed Nov. 5, 2005, entitled “Nanostructured Core Shell Electrocatalysts For Fuel Cells”, by Yuriy Viacheslavovich Tolmachev, the disclosure of which is hereby incorporated by reference in its entirety.

BACKGROUND

The present exemplary embodiments relate to nanostructured catalyst layers. It finds particular application in conjunction with Polymer Electrolyte Fuel Cells (PEFCs), and will be described with particular reference thereto. However, it is to be appreciated that the present exemplary embodiment is also amenable to other like applications.

The last 5-10 years witnessed a dramatic increase in interest, both from industry and governments, in fuel cells as power sources for automotive, portable and stationary applications. This was prompted by the problems of limited and unstable supplies of fossil fuel resources, growing atmospheric pollution and greenhouse gas emissions, aggravated by a projected increase in the global power demand, as well as by several recent ground-breaking inventions, which lead to a significant improvement in the performance of fuel cells.

Polymer Electrolyte Fuel Cells (PEFCs) are usually fueled with hydrogen or methanol. The hydrogen-air fuel cell (HAFC) is currently considered as the power source of choice for vehicular applications thanks to its superb power density, lifetime and short start-up time. Due to problems with hydrogen storage, another type of PEFC, Direct Methanol Fuel Cell (DMFC), is also investigated for portable and vehicular applications. DMFCs have already entered the market for portable applications, such as military radio stations and laptop computers.

Since PEFCs typically operate at temperatures below 100° C., they require catalysts to achieve practically meaningful power density, particularly in the case of oxygen reduction (ORR) and methanol oxidation (MOR) reactions. A DOE report in 1982 concluded that the prospects of non-precious metal catalysts in low temperature fuel cells were slim either due to their poor activity or stability. P. N. Ross, F. T. Watner. Prospects for the development of non-noble metal catalysts for hydrogen air fuel cells. Lawrence Berkeley Lab. Report No. LBL-14192 (Mar. 1, 1982) 57. Even though some new materials have been discovered since then, this conclusion has not been successfully challenged.

Currently, all catalysts used in PEFCs are platinum-based. The current state-of-the-art PEFCs use 0.2-0.4 mg/cm2 of Pt for air cathode, 0.05 mg/cm2 of Pt for hydrogen anode, 0.2 mg/cm2 of Pt+Ru for reformate gas (H2+100 ppm CO) anode, and 4.0 mg/cm2 of Pt+Ru for methanol anode,6 and the reported performances at 80-100° C. with 1 atm of air are ca. 0.35-1 g/kW for pure hydrogen-fed HAFC, and 20-80 g/kW for DMFC. Published analyses for different types of automotive HAFCs estimate the cost of catalyst as 50-250 $/kW or 20-35% of the fuel cell stack cost. These numbers alone are higher than the 2015 DOE's goals for the complete fuel cell power plant including hydrogen storage, i.e. $30/kW, which is cost competitive the internal combustion engine technology. The situation looks even less promising in the case of automotive DMFCs.

The high cost of PEFCs is not the only problem that impedes their broad use; the scarcity of Pt metal is even a larger concern. Current fuel cell powered automobiles require ca. 60 g of Pt compare to 2-5 g of Pt for internal combustion engine vehicles. According to various estimates, Pt production will be able to meet the demand of projected FCA market growth, only if the required amount of Pt is reduced 4-5 fold, to 15 g/vehicle or to 0.2 g/kW by 2015 and to lower values later. An even more drastic requirement, a 10-fold reduction of Pt loading, was formulated as a necessary condition for a large-scale market penetration of automotive fuel cells by a Nissan researcher. M. Arita; Technical Issues of Fuel Cell Systems for Automotive Applications. Fuel Cells 2 (2002) 10-14.

The commercial viability of Polymer Electrolyte Fuel Cells (PEFCs) for large-market applications requires, among other factors, a substantial reduction of platinum metals loading. In this respect, according to a GM study, improvements within the current state-of-the-art PEFC catalytic layer design cannot reduce Pt loading to the acceptable level, unless new types of more active catalysts are discovered.

The state-of-the-art design of the catalytic layer in PEFCs based on carbon-supported ca. 3 nm Pt nanoparticles suffers from three main drawbacks. The first is a poor utilization of Pt nanoparticles, since 25-50% of catalyst particles are buried in carbon agglomerates and, therefore, are not accessible to reactants or protons. The second drawback is a large fraction (50-60%) of subsurface Pt atoms, which do not participate in electrocatalysis. The third is the significant thickness of the catalytic layer (ca. 10 μm), which prevents efficient mass-transport throughout the layer.

Proposed herein is a conceptually new design of the catalytic layer based on recently developed nanofabrication principles.

BRIEF DESCRIPTION

In accordance with one aspect of the present exemplary embodiments, there is provided a catalyst layer for use in fuel cells, the catalyst layer including a nanoporous substrate and a surface layer of a different material.

In accordance with a second aspect, there is provided a fuel cell including an ion exchange membrane positioned between first and second porous catalyst layers and first and second electrodes, each of the catalyst layers including a nanoporous substrate and a surface layer of a different material.

In a third aspect, there is provided a membrane-electrode assembly including a proton conducting polymer membrane separating two porous electrodes, and positioned between two porous catalyst layers, each of the catalyst layers including a nanoporous substrate and a surface layer of a different material.

In a fourth aspect, there is provided a method for preparing a fuel cell including the steps of: providing first and second electrodes; positioning an ion conducting membrane between the electrodes; and disposing first and second porous catalyst layers on opposing surfaces of the membrane, each of the catalyst layers including a nanoporous substrate and a surface layer of a different material.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic diagram of a single stack HAFC cross-section. The zoomed view shows the MEA.

FIG. 2 is a microscopic single pore model of the catalytic layer in a fuel cell.

FIG. 3 are three historical types of PEFC electrodes.

FIG. 4 is a honeycomb catalytic layer structure.

FIG. 5 are a few possible structures of a catalyst layer. Dark represents the core-shell skeleton, white represents the voids. Typical feature size is 5-50 nm.

FIG. 6 is a graph of the current transient during Ni plating into an ion track etched membrane. Membrane pore size 15 nm, pore density 6·108 cm−2, thickness 6 μm, diameter 47 mm. Plating bath: Ni(H2NSO3)2×4H2O (515 g/l)+NiCl2×6H2O (20 g/l)+H3BO3 (20 g/l). 22° C. Potential step from −0.350 to −1.000 V vs Ag/AgCl.

FIG. 7 is a schematic representation of an ALD system (not to scale).

FIG. 8 is an SEM image of Ni nanowires produced by electroplating into ion-track etched polycarbonate membrane.

 DETAILED DESCRIPTION

In order to give the reader a better understanding of the proposed design, we will review briefly the historical development and the current designs of the PEFC catalytic layers.

FIG. 1 shows a schematic diagram of a single stack of a conventional HAFC 10. The heart of the PEFC is a Membrane-Electrode Assembly (MEA) 12, which includes an ion-conducting layer 14 separating an anode and a cathode 16, 18 coated with catalytic layers 20, 22, and sandwiched between two gas diffusion layers 24, 26. Although not intended to be limiting as other electrolytes are also useful in the present embodiments, an exemplary ion-conducting membrane for use herein is Nafion®. Nafion® is a perfluorinated polymer available from DuPont containing a certain proportion of sulfonic acid functional groups and having the general structure below, with H+ being the counter ion:

Structural optimization of all MEA components is important for maximizing the power density of a fuel cell, but, since precious metals are used only in the catalytic layer, this component is the most relevant for the present discussion. Since the structures of air cathode, hydrogen, and liquid methanol anodes are in many ways similar, we will focus mostly on the former, since it is common to all types of PEFC and because it consumes 80% or more of the total Pt in HAFC.

Electrochemical reactions in PEFCs require the presence of a catalyst, electrons, protons and reactants. As a result, they occur only in a three-phase zone where catalyst, electron conductor, polymer electrolyte, and gas channels meet. This point is illustrated in FIG. 2, which shows a single pore with catalytic walls 30, 32 partially filled with electrolyte 34 (in this case Nafion®). In this figure, the Pt wall surface, which is in contact with air does not participate in catalysis, because there is no electrolyte to transportions to and from region I (More recent studies show that protons can be transported via surface diffusion on Pt, yet not fast enough to sustain large current densities over 0.1 μm away from the interface). Deep inside the Nafion-filled pore (region II) the current density is also small, since the supply of the reactant via diffusion through the electrolyte is slow.

As can be seen from FIG. 2, there are two ways to improve the utilization of the catalyst: by localizing the catalyst only in a thin region (III) of the pores at the interface between electrolyte and gas, and by reducing the amount of catalyst buried under the first surface layer. Historically, the advancements in the reduction of precious catalyst loading in PEFCs targeted both approaches, yet neither of them has been pushed to the limit.

The PEFC electrodes developed by General Electric for NASA in the 1960's as shown in FIG. 3 (left) were prepared by random mixing of about 10-20 nm Pt catalyst particles (which also provide electronic conductivity) and Teflon (which forms pores filled with gaseous reactant) in contact with Nation membrane (which provides ionic conductivity). Typical Pt loading was 4 mg/cm2, and the maximal power, when operated on H2 and air, was 0.1 W/cm2, yielding 40 g Pt/kW value.

The first significant improvement in reduction of Pt loading (down to 0.4 mg/cm2 and 4 g Pt/kW) came from Los Alamos National Laboratory (LANL) in the 1980's when both the use of Pt supported on carbon particles and impregnation of the catalyst layer with Nafion were proposed (FIG. 3, middle). The carbon support allowed the use of smaller Pt particles (3 nm) with a larger surface-to-volume ratio. The impregnation with Nafion allowed the extension of the three-phase zone over larger thickness. Yet, such electrodes still suffered from (i) poor Pt utilization (20-25% by cyclic voltammetry, CV), since many catalytic particles were not in contact with electrolyte, and (ii) from the large thickness (100, and later 50 μm), which prevented fast reactant transport throughout the catalytic layer.

The second major improvement also came from LANL in the early 1990s (U.S. Pat. No. 5,234,777), when thin-layer electrodes consisting of carbon supported Pt particles mixed with Nafion were proposed (FIG. 3, right). These electrodes do not contain Teflon, and, thus, have a smaller volume fraction of gas-filled channels (recently estimated as 4% based on the partially flooded agglomerate model for the cathode) that surround unutilized Pt particles. This allowed for an increase in Pt CV utilization to 50-75%. Since these electrodes are significantly thinner (5-10 μm), the mass transfer of the gas reactants is more efficient and the ohmic resistance is reduced. Numerous advancements of this design have been reported by other groups with the resulting state-of-the-art performance 0.7 W/cm2 or 0.35 g Pt/kW with 0.05 mg/cm2 of Pt on the hydrogen electrode and 0.2 mg/cm2 on the air electrode.

Recently several groups reported further reduction in the Pt loading using sputter or pulsed laser deposited catalyst. These studies showed that such electrodes with 0.017 mg/cm2 Pt loading perform as well as the conventional electrodes with 0.4 mg/cm2 Pt loading in the case of H2 anodes, and that methanol anode performance can reach 0.5 g/kW. The cathode Pt loading can reportedly be reduced from 0.4 to 0.1 and even to 0.04 mg/cm2 without a significant sacrifice in performance. The conclusions that emerged from these results are: (i) that it is possible to decrease the catalytic layer thickness down to 10 nm without performance degradation, and (ii) that the introduction of Nafion is not necessary for such thin continuous Pt layers. The first observation can be explained by the fact that the reaction layer in porous electrodes at practical current densities extends only to few nm (see below). The second observation can be attributed to the surface proton diffusion on Pt.

Sputter-deposited electrodes, despite their impressive performance, are far from the ideal. They still have a significant fraction of Pt atoms buried under the surface, and their preparation route does not allow for much flexibility in making optimized 3D porous nanostructures. The present embodiments propose to solve these issues using recently developed nanofabrication methods.

The present designs use, for example, a non-noble metal rather than carbon as a support or substrate for Pt. While non-noble metals will be used in the discussion of the substrate herein, it is possible that other, non-metal materials may be used to form the substrate as well.

The morphology and thickness of the substrate skeleton can be designed in order to optimize the mass-transport properties. The nature of the support material is selected on the basis of cost, ease of preparation of nanostructures, corrosion resistance, adhesion to the catalytic shell, and promoting activity in electrocatalysis. Such thin, highly porous structures with large catalytically active surface areas, according to a presented prophetic example, exhibit a significantly higher power output per mg of Pt and per cm2 of the membrane than the current PEFC catalytic layers.

The present embodiments suggest a replacement in fuel cell electrodes of carbon-supported Pt nanoparticles for nanostructured non-noble metal supports coated with one or a few monolayers of a catalytic noble metal. The possible non-noble metal substrates are chosen based on their cost, corrosion stability, promoting effect in electrocatalysis, the ease of fabrication of nanostructures from these materials and adhesion energy of the catalytic shell material. Candidates may include but not limited to Fe, Co, Ni, Cr, Ti and other first row transition metals, valve metals or coinage metals or a combination thereof. The noble shell material may be platinoid, and is preferably Pt. The embodiments will be described using Pt as an exemplary metal, but it should be noted that the discussion and embodiments may apply to other platinoid metals as well, or possibly to other metal and non-metal materials. The nanostructures include nanoparticles, nanowires, honeycombs, and inverse opals. The thickness of the catalytic layer is preferably 0.1-10 μm, and the characteristic pore size is preferably 3-300 nm, more preferably 15-150 nm.

Useful in the practice of the present embodiments is the method of preparation of conformal coating of a noble metal on a non-noble substrate. One method proposed for this purpose is Atomic Layer Deposition (ALD).

The impact of the present embodiments is the development of a technology allowing for a 5-10 fold reduction of Pt loading per kW in PEFCs, and, potentially, reaching the ultimately best design of the catalytic layer for the three most important types of PEFC electrodes: air cathode, hydrogen anode, and liquid-fed methanol anode. It is believed that the present concept has the potential of meeting the 2015 DOE goal for Pt load reduction in automotive fuel cells.

The elimination of Pt atoms in the bulk of catalyst particles is mandatory for the next generation of electrocatalysts. So far, the research in this direction has been aimed at decreasing the particle size. Yet, such approaches face limitations for particles smaller than 1-3 nm (where there is still as much as 60% of the Pt atoms wasted in the bulk) due to a decrease in the surface specific catalytic activity and due to the problems with preparation and stability of small Pt particles even on carbon supports. The use of metal supports allows one to prepare monolayers of Pt without agglomeration, i.e., to form core-shell catalysts. In addition to reducing the amount of bulk Pt, the use of metal supports may promote the activity of surface Pt, as in the case of Pt—Ni catalysts in oxygen reduction reactions, methanol oxidation reactions, and reform gas (H2+CO) oxidation. The elimination of the carbon support has three more advantages: (i) it precludes the burying the catalysts particles inside agglomerates of carbon particles, (ii) it results in a larger active surface area at a given thickness of the catalytic layer, and (iii) it creates the possibility of proton transfer by surface diffusion on Pt by extending the three-phase zone into the gas-catalyst interface.

All the previously described gains achieved by a core-shell atomic design may be lost at high current densities if only a very small fraction of catalyst has a sufficiently fast supply of reactants and protons. This calls for a proper 3D design of the catalytic layer at the nanoscale level, which minimizes electric and mass-transport resistances. A simple assembly of core-shell nanospheres in the catalytic layer may not provide suitable results because (i) some catalytic particles may end up outside of the percolating cluster, (ii) because the electronic resistance of the cluster may be high due to numerous particle-particle contacts, and (iii) because the reactant and proton transfer in such low porosity and high tortuosity structure is inefficient. Structures with highly porous interpercolating continuous networks of catalyst and electrolyte/gas pores present better opportunities. We will start our analysis with the core-shell honeycomb structure shown in FIG. 4.

PROPHETIC EXAMPLES

We will start our analysis with the core-shell honeycomb structure shown in FIG. 4. Table 1 below compares estimated properties of the proposed honeycomb catalytic layer, having all pores oriented perpendicular to the membrane plane, with experimental data for a current state-of-the-art design. We used this type of electrode, since the calculations are the most simple for this geometry. The calculations were performed for air-fed cathodes with the pore diameter of 30 nm, wall thickness of 5 nm and three different layer thicknesses: 0.1, 1 and 10 μm. The effects of water and proton transport were ignored in the calculations. The HAFC powers at 0.2 A/cm2 and at 0.6 V are used in this comparison.

As can be seen from Table 1, the ratio of surface-to-total Pt atoms is 3-5 times higher for the proposed design due to a better surface dispersion and utilization of Pt. This number means 3-5 times higher current density at high operating voltages. These calculation do not take into account the promoting enhancement by the substrate (up to 4 times), which may be degraded in practice by the need to deposit multilayer Pt shell. When these factors are considered, the estimates for the proposed design yield the potential at 0.2 N cm2 for 0.051 mg/cm2 Pt loading about 80 mV higher than that for the current electrodes with 0.2-0.4 mg/cm2 of Pt, which translates into 4-8 fold reduction of Pt loading per kW.

We used the Thiele model of reaction kinetics (as described in Chemical Engineering Kinetics, McGraw-Hill (1970), pp 612) in a Nafion-filled cylindrical pore (without considering the top flat part of the catalytic layer) to estimate the geometric current density at 0.6 V for an air-fed electrode. Since the reaction layer is confined to 4.2 nm depth, electrodes of different thicknesses yield the same current of 2.2 A/cm2.

Even though this current density is approximately 5 times higher than for the state-of-the-art HAFC anodes at the same potential, the utilization of the catalyst is rather poor, as the reaction layer does not go deep into the pore. Fortunately, our catalytic layers may require only partial pore filling with Nafion® or other ion exchange resin and still have 100% surface utilization. This is due to the fact that reactive species can be transported via surface diffusion on Pt over more than 0.1 μm. Such effect does not occur with conventional electrodes, where Pt nanoparticles are separated by carbon support.

We calculated, using the Thiele model again, the geometric current densities for the extreme case of completely gas filled pores, taking roughly into account binary and Knudsen diffusion effects for O2 transport in air-filled nanopores. The estimate shows that the reaction layer in this case extends to 650 nm and that mass-transport limitations in the catalytic layer do not occur under such conditions. Note that even if we assume that for a half-filled pore the current is half of the one calculated for a completely gas-filled pore, and the ionic resistance is twice that of the all-Nafion® pore, the ohmic losses in the 0.1 μm layer are only 30 mV.

In practice, the membrane resistance and gas-diffusion layer flooding are likely to limit the performance at current densities over 1-2 A/cm2. For a 0.1 μm layer, 2 A/cm2 will be reached at 768 mV. If we subtract from this number a 0.1V loss due to the resistance of 50 μm Nafion® membrane, we will obtain a power of 1.34 W/cm2 or loading of 0.07 g of Pt/kW, if both HAFC electrodes are 0.1 μm thick and anode losses are negligible.

Thus, it appears that the present catalytic layer design should allow for 5-10 fold reduction of the amount of Pt per kW power for HAFC at both low and high currents, as well as for Pt+Ru DMFC anodes. At low currents this reduction is due to a better dispersion and utilization of the three-phase zone, and at high currents it is due to an improved mass-transport of reactant gases and protons in the thin porous layer. We also have discovered that 0.1-1 μm thick electrodes can provide the best Pt/kW savings, and that the pore-filling with electrolyte may not be necessary at such thicknesses.

TABLE 1 Calculated parameters for the proposed catalytic layer design (FIG., 4, with 30 nm pore diameter and 5 nm pore wall) of various thicknesses and experimental parameters for the current state-of-the-art PEFC cathodes. 0.1 μm 1 μm 10 μm Current Parameter thick thick thick electrodes Roughness factor 12.1 119 1185 200-400 Pt loading, mg/cm2 0.051 0.503 5.02 0.2-0.4 Surface/total Pt atoms 100% 100% 100% 41% CV utilization of surface Pt 100% 100% 100% 50-75% Active Pt/total Pta 100% 100% 100% 20-30% g of active Pt/cm2 0.051 0.503 5.02 0.04-0.12 air-fed anode potential at 843 mV 903 mV 963 mV 765 mV 0.2 A/cm2 Power density at 0.2 A/cm2, 0.169 0.181 0.193 0.155 W/cm2 g of Pt/kW at 0.2 A/cm2b 0.30 2.8 26 1.3-2.6 Geometric current density 2.2c 2.2c 2.2c 0.43d at 0.6 V for Nafion-filled pores, A/cm2 Geometric current density 243e 1,464e 1,606e at 0.6 V for air-filled pores, A/cm2 Ionic resistance, mΩ/cm2 0.11 1.1 11 20f 100-200g g of Pt per kW at maximal 0.035h 0.26-0.56 powerb aSurface/total Pt atoms × CV utilization of surface Pt bthis number includes only cathode Pt loading. Anode Pt should increase this value by 10-30%. cCalculated using Thiele model for an air feed with extrapolated kinetic current 21 A/cm2, and O2 solubility and diffusion coefficients from K. Lee, A. Ishihara, S. Mitsushima, N. Kamiya, K. Ota., Effect of recast temperature on diffusion and dissolution of oxygen and morphological properties in recast Nafion. J. Electrochem. Soc. 151 (2004) A639-A645. dState-of-art data from S. Litster, G. McLean. PEM fuel cell electrodes. J. Power Sources 130 (2004) 61-76. with air cathode corrected for membrane resistance eCalculated using Thiele model with binary (0.21 O2 + 0.79 N2) diffusion and Knudsen corrections for effective diffusion coefficient but without convective effect (see text for details). ftheoretical resistance for 10 μm Nafion layer with conductivity 0.1 Sm/cm and 0.5 porosity. gexperimental resistance including particle-to-particle contact and O2 transport at ca. 0.6 V with pure oxygen feed. hestimated for 2 A/cm2, as explained in the text.

Proposed Designs of the Catalytic Layers

While sufficient, the honeycomb structure discussed above has a relatively low surface-to-volume ratio and does not have an interconnected pore network, which may cause an inefficient use of catalysts if some pores are clogged, as in the case of electrode flooding. Several other structures, like those shown in FIG. 5, may offer better performance.

Structure A is composed of randomly arranged core-shell nanowires. Its biggest advantage is the ease of fabrication and of porosity variation. The disadvantage is the difficulty in controlling the exact structure, which makes it less suitable for optimization and model studies. Ordered structure B, with rigidly connected hexagonally packed parallel nanowires, is free of these problems, but it can only be produced, using known nanofabrication procedures, in a small range of dimensions and with a low porosity. Structure C, an inverse opal, which can be prepared by depositing the core metal into an ordered assembly of silica or polymer nanospheres, provides the greatest dimensional flexibility, as well as superior mass-transfer properties.

An increase in the pore size at a constant thickness of the catalytic layer is beneficial for the mass-transport, yet at the same time the catalytically active surface area (Pt roughness factor) decreases. Our estimates show that the optimal pore size for 0.1-1 μm thick catalytic layers lies between 5 and 50 nm, depending on the assumptions about proton and reactant transport.

In the following section we present the proposed routes to fabricate the structure in FIG. 5A and methods of its characterization.

Preparation of Ni Core Nanowires

We selected Ni as the core substrate for this work since it shows promotion in the catalytic activity of Pt in oxygen reduction, methanol and for reformate gas (H2+CO) oxidation, and because preparation techniques for Ni nanostructures are well established. Nevertheless, other metals may be used as discussed above.

One can obtain or fabricate 15-60 nm diameter×5-15 μm length nanowire cores (FIG. 5A) by electroplating Ni into nanoporous alumina or polymer membrane templates. The porous alumina substrate allows for fabrication of a larger number wires in one step and the lengths of the wires are larger.

This example deal specifically with a deposition into an ion-track etched membrane, which may be obtained commercially. In order to fill the pores with Ni, a 150-200 nm Au layer is sputtered on one side of an ion track etched membrane. The membrane is then placed on a sheet of gold, with the sputtered layer facing the sheet, and clamped between a glass filter cup top with an aluminum bottom piece. The nickel plating solution is introduced into the filter cup and Ni is plated into the membrane at constant potential. A typical current transient is shown in FIG. 6. Three regions can be distinguished. The fast current decay within the first few seconds is related to double layer charging after the potential step. The following constant current region corresponds to the growth of Ni nanowires in the pores. The current increase in the third region starts when the pores are completely filled and a Ni deposit is formed on the outer surface of the membrane. The plating is usually stopped just before the end of the constant current region (about 300 s in FIG. 6). It was found that placing a large Pt gauze counter electrode parallel to the membrane solves the problem with non-uniform Ni growth rate across the membrane, which is caused, apparently, by the large ohmic drop in solution.

Most of the sputtered gold layer can be peeled off with a scotch tape, and the rest can be readily dissolved in a cyanide solution or in mercury, if needed. The Ni nanowires are released after the template dissolution. They are collected by centrifugation, rinsed and stored in ethanol. One run like the one shown in FIG. 6 results in 1.5 mg of 35 nm diameter×6 μm long non-bundled Ni nanowires from one ion-track etched polycarbonate membrane. The SEM image of the wires in shown in FIG. 8. Our procedure combines methods for electroplating of nanowires developed by Searson et al. (T. M. Whitney, J. S. Jiang, P. C. Searson, C. L. Chien; Fabrication and Magnetic-Properties of Arrays of Metallic Nanowires; Science 261 (1993)1316-1319) and Mallouk et al. (M. L. Tian, J. U. Wang, J. Kurtz, T. E. Mallouk, M. H. W. Chan; Electrochemical growth of single-crystal metal nanowires via a two-dimensional nucleation and growth mechanism; Nano Lett. 3 (2003) 919-923).

Deposition of Pt Shell

Preparation of (sub)monolayers of noble metals on noble substrates (like Pt on Ru or vice versa) through spontaneous deposition or underpotential deposition exchange techniques is well known in the electrocatalysis community. However, these reactions are not suitable in the case of non-noble substrates because of the spontaneous cementation reaction, i.e. deposition of the noble metal and dissolution of the non-noble core. Since this is an electrochemical process, the deposition and dissolution can occur on spatially different areas of the conductive substrate, which results in destruction of the substrate shape. Recently a self-limiting decomposition reaction in a non-polar solvent was proposed to prepare a submonolayer of Ru on Pt. D. X. Cao, S. H. Bergens; An organometallic deposition of ruthenium adatoms on platinum that self poisons at a specific surface composition. A direct methanol fuel cell using a platinum-ruthenium adatom anode catalyst. J. Electroanal. Chem. 533 (2002) 91-100. Even though such conditions may suppress the cementation, it is not clear whether good quality mono- and multilayers can be prepared via this route.

Gas-phase deposition offers definite advantage in terms of the cementation problem. In this case it is also possible to control layer-by-layer growth using an Atomic Layer Deposition (ALD) technique. The ALD process consists of alternating cycles of precursor adsorption and decomposition. To generate a thin film, first a dose of the precursor vapor is brought to the surface of a substrate. Then, any excess non-adsorbed precursor is purged and pumped away. Next, a vapor dose of another reactant is used to decompose the adsorbed precursor, and the excess is purged and pumped away. One cycle normally deposits a monolayer or less of the material. ALD is reputed to generate reproducibly conformal coatings even in the case of high aspect ratio nanosubstrates.

The only published successful procedure for Pt ALD involves adsorption of (η5-methylcyclopentadienyl)trimethylplatinum, (MeCp)PtMe3, followed by the ligand oxidation with oxygen at 300° C. This procedure is not suitable for the present purpose since the Ni substrate will be oxidized under these conditions. Decomposition of (MeCp)PtMe3 in the presence of H2 occurs at a lower temperature, i.e. 90-250° C., which has been used in a conventional MOCVD of Pt. Considering the fact that the thermal decomposition of this precursor occurs only above 300° C., the use of H2 as the reactant for the decomposition of (MeCp)PtMe3 in Pt ALD seems well justified. During the MOCVD process, this precursor reacts cleanly with H2 producing corresponding hydrocarbon ligands according to the following:


(MeCp)PtMe3+4H2→Pt+3CH4+MeC5H9

It has been reported that this reaction requires long induction periods with Teflon and clean glass substrates. This type of complication is not expected when H2-treated Ni is used for the adsorption of the precursor, since Ni is known to promote H2 dissociation. On the other hand, the induction period could be beneficial, as it can prevent the Pt deposition on the walls of the reactor.

Other advantages of (MeCp)PtMe3 are its commercial availability, the highest purity of Pt deposits produced compare to other precursors, and a high vapor pressure even at room temperature (0.053 Torr, mp 30° C.). A disadvantage is its sensitivity to air. In lieu of this the loading of the precursor into the source bottle may be performed in a glove box. It has been claimed that (EtCp)PtMe3 is less sensitive to air and it is a liquid at room temperature and this may also be suitable for use in the present embodiments. Due to its higher cost though, it may not be preferred. Other contemplated precursors include (η5-ethylcyclopentadienyl) trimethylplatinum, and (η1,5-cyclooctadiene)dimethylplatinum.

The Pt ALD may be performed in the proposed all glass reactor shown in FIG. 7. This system consists of two nitrogen tanks (N2), a H2 tank (H2) equipped with low pressure regulators, the platinum precursor container in a thermostat (14° C.), precursor delivery line wrapped with a heating tape (60° C.), an ALD reactor in an oven (90-180° C.), a pressure transducer (P), two cold traps (one used during adsorption step for the Pt precursor recycling, the other is used during the reduction step) and a rotary vane pump. The set up uses four vacuum valves may be controlled by commercial software via actuators, and several manual valves (not shown). Typical operating pressure is about 10 mbar.

The atomic layer deposition is preferably performed in a cold-wall glass or Teflon reactor, with the substrate temperature 25 to 300° C., preferably at 175° C. The purging times in the cycle can be 1s to 30 min, preferably 5 min for each step. Flow rates may vary between 0.1 to 100 cm/s, preferably 1 cm/s. The number of cycles can vary from one to over one hundred, preferably 5.

Characterization of Core-Shell Nanostructures

The shape, size and the preferred orientation of the nanowires may be determined before and after Pt coating using Scanning Electron Microscopy, wide- and small-angle. TEM measurements may also be performed.

The surface area, pore size and pore size distribution of the catalysts may be evaluated from N2 adsorption isotherms at 77K. The Pt content in the sample may be determined primarily by Energy Dispersive X-ray Fluorescence Spectroscopy (EDXRFS). This is a non-destructive technique, it requires a minimal sample preparation, and it has enough sensitivity to measure even 2% of monolayer of Pt in 1.5 mg of 30 nm thick Pt-shell Ni-core nanowires. With such thin dilute samples the matrix effect is constant and 2% accuracy can be obtained with a proper choice of calibration standards. For this purpose one may use filter papers impregnated with known amounts of standard Pt and Ni solutions. The results can be validated by EDXRFS on dissolved samples and by ICP-AES. The data on surface area and Pt content can be used to calculate the average thickness of the Pt shell.

The atomic-level structure of the shell can be determined by solid-state 195Pt NMR and by Pt L-edge EXAFS.

For the initial electrochemical characterization of the catalysts the nanowires may be placed on a glassy carbon electrode using the drop-and-dry technique and coated with Nafion. The presence of uncovered Ni patches, corrosion stability and the Pt surface area can be determined by cyclic voltammetry (CV). The electrocatalytic activity of the new catalysts in O2 reduction, CO and methanol oxidation can be studied with a thin-layer rotating disk electrode up to 60° C. Pt shells with various thicknesses can be studied in order to see the spatial extent of the Ni promoting effect. If such effect is found beyond monolayer thick coating, it would be a strong evidence in favor of the electronic mechanism of promoting. It has been reported that Pt—Ni alloy nanoparticles do not corrode under practical electrocatalysis conditions, yet it is likely that those catalysts had a thick Pt skin. If Ni corrosion turns out to be a serious problem with the present catalysts, one can limit electrocatalytic studies to samples with multilayer Pt shells.

Membrane-Electrode Assembly and Fuel Cell Testing

Rotating disk electrode measurements are adequate for electrocatalytic studies only at low current densities. At higher currents, when mass-transport effects become significant, fuel cell measurements are necessary. For these experiments, the catalysts will be deposited by drop-and-dry or by immersion techniques onto ion exchange membranes (such as Nafion®) or commercial gas-diffusion layers (E-TEK and SGL). The membrane-electrode assembly can be prepared by hot-pressing and clamped between two 5×5 cm2 gold-coated gas-flow fields with two dynamic hydrogen reference electrodes attached as described by W. S. He, T. Van Nguyen; Edge effects on reference electrode measurements in PEM fuel cells; J. Electrochem. Soc. 151 (2004) A185-A195. The fuel cell performance can be measured using a 250 W fuel cell testing station, which is capable of working up to 20 atm backpressure and 180° C. temperature. All data for the new electrodes can be compared with the performance of a conventional membrane-electrode assembly (E-TEK) under the same conditions.

More specifically, one can determine the following parameters: the electrochemically active surface area of Pt by cyclic voltammetry; the kinetics of oxygen reduction, hydrogen and CO oxidation by steady state-polarization curves; the effects of O2 and H2 dilution with an inert gas (N2 or He) and the effect of the flow rate on the polarization curves. The later data allows one to obtain the relative contributions of bulk, Knudsen and thin-film diffusion into mass-transport.

How the structure of the catalytic layers affects the performance of the fuel cell can be investigated. The thickness, porosity and Nafion content of the nanowire assembly will be varied by changing preparation conditions: the rate of emersion, the rate of suspension evaporation, the concentration of Nafion and a pore-forming agent (HF-soluble colloidal silica) in the suspension, the viscosity and the surface tension of the suspension. Elemental analysis, cyclic voltammetry, lateral and cross-sectional SEM imaging can be used to determine the produced structures. As explained above, we expect to see that the power per geometric area will increase up to 10 μm thick layer and remains constant at higher thicknesses. Also, the addition of Nafion may not be necessary for layers thinner than 1 μm. The pore forming agent may improve performance for the thickest layers only. Also, it is believed that catalytic layers deposited on to the gas-diffusion layer will show a better performance than those deposited on to the membrane, similarly to the case of sputtered electrodes. This effect likely originates from a rougher surface of the gas-diffusion layer. We also expect that our thin catalytic layers in contact with the hydrophobic side of the gas-diffusion layer will be less susceptible to flooding.

The applicants are aware that the proton transport resistance of the membrane and flooding of the gas diffusion layer are likely to limit the performance of our fuel cells at high currents. While we will attempt to minimize these effects by increasing the back pressure and the flow rates in order to assess the intrinsic performance of the thick layers, the present embodiments are primarily concerned with showing that the current state-of-the-art power output can be achieved with 5-10 times smaller Pt loading using thin catalytic layers.

The present concept demonstrates the feasibility of Pt loading reduction in fuel cell using nanostructured catalysts. Further investigation will verify quantitative predictions of the present analysis, identify the factors limiting the performance, and measure the experimental parameters needed for more detailed modeling.

The novelty of the present approach is the use of the nanofabrication techniques, which were developed within the last decade, to address a crucial issue in a socially important technology. The proposed study promises to open a new paradigm in the development of PEFCs with ultra-low Pt loading, which may reach the projected 2015 DOE goal of 15 g of Pt per car. It is not limited to one type of electrode, but encompasses all three main types of electrodes: oxygen (air) cathode, hydrogen, liquid- and vapor-fed methanol anodes, and, thus, is applicable to portable and stationary fuel cells as well.

The proposed design of the catalytic layer presents an integrated approach over three length scales: atomic (shell), nano-(core skeleton) and micro-(layer thickness). Thus, it is an example of the emerging field on nanocatalysis, which may be defined as an explicit design of a catalyst over length scales larger than a single active site. Successful demonstration of this approach is expected to stimulate the adoption of similar strategies not just in electrocatalysis but in other technological areas.

The procedures for fabrication of Ni nanowires have been already developed by other groups and proven reliable. On the other hand, atomic layer deposition of Pt on Ni has never been published, and such procedure, especially for nanostructures, was developed in our laboratory. Another difficulty arises in preparation of 0.1-1 μm thick layers of nanoporous catalysts. The drop-and-dry technique may not produce the desired uniform thickness. As an alternative we propose immersion coating. By varying the viscosity, surface tension of the suspension and the immersion rate of the substrate (membrane or gas-diffusion layer) one will be able to prepare catalytic layers with different thicknesses and porosities.

The present embodiments will be used by us and others to develop better performing layers through modeling and synthetic advancements. Specifically, we intend to investigate the performance of the inverse opal structures (FIG. 5C), as well as of Cr and Ti supports, which also show a promoting activity in O2 reduction but are more stable in aqueous media due to the formation of passive oxide films.

The exemplary embodiment has been described with reference to the preferred embodiments. Obviously, modifications and alterations will occur to others upon reading and understanding the preceding detailed description. It is intended that the exemplary embodiment be construed as including all such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.

Claims

1. A fuel cell comprising an ion exchange membrane positioned between first and second porous catalytic layers and first and second electrodes, each of said catalytic layers comprising a carbon-free non-noble porous metal nanostructured substrate and a coating layer comprising a noble metal.

2. A fuel cell according to claim 1, wherein said noble metal is platinum.

3. A fuel cell according to claim 1, wherein said nanostructure comprises at least one of nanoparticles, nanowires, honeycombs, and inverse opals.

4. A fuel cell according to claim 1 wherein said first electrode is an oxygen or air cathode.

5. A fuel cell according to claim 4 wherein said second electrode is a hydrogen anode.

6. A fuel cell according to claim 5, wherein a platinum loading in said catalyst layers is from 0.2 to 0.017 mg/cm2.

7. A fuel cell according to claim 4, wherein said second electrode is a methanol anode.

8. A fuel cell according to 6, wherein a platinum loading in said catalyst layers is from 0.1 to 0.04 mg/cm2.

9. A fuel cell according to claim 2, wherein the thickness of the nanoporous catalytic layer is from 0.1 to 10 μm.

10. A fuel cell according to claim 1, wherein a pore diameter of the nanostructure is from 3 to 200 nm.

11. A fuel cell according to claim 10 wherein a pore diameter is from 5 to 50 nm.

12. A fuel cell according to claim 1, wherein said ion exchange membrane comprises a proton-conducting polymer and it fills at least part of the pores of said substrate.

13. A fuel cell according to claim 1, wherein said non-noble metal comprises a first-row transition element, valve metal or coinage metal or a combination thereof.

14. A fuel cell according to claim 1, wherein said fuel cell is capable of achieving an energy output of 0.02 g Pt/kW.

15. A membrane-electrode assembly comprising a proton conducting membrane separating two porous electrodes, and positioned between two porous catalytic layers, each of said catalytic layers comprising a carbon-free non-noble metal porous nanostructured substrate and a noble metal surface layer.

16. A membrane-electrode assembly according to claim 15, wherein said nanostructure comprises at least one of nanoparticles, nanowires, honeycombs, and inverse opals.

17. A membrane-electrode assembly according to claim 15, wherein said first electrode is an oxygen or air cathode.

18. A membrane-electrode assembly according to claim 17 wherein said second electrode is a hydrogen anode.

19. A membrane-electrode assembly according to claim 18, wherein a noble metal loading in said catalyst layers is from 0.2 to 0.017 mg/cm2.

20. A membrane-electrode assembly according to claim 17 wherein said second electrode is a methanol anode.

21. A membrane-electrode assembly according to 20, wherein a noble metal loading in said catalyst layers is from 0.1 to 0.04 mg/cm2.

22. A membrane-electrode assembly according to claim 15, wherein the nanoporous catalytic layer thickness is between 0.1 and 10 (im.

23. A membrane-electrode assembly according to claim 15, wherein a pore diameter of the nanostructure is from 3 to 200 nm.

24. A membrane-electrode assembly according to claim 15, wherein said ion exchange membrane comprises a proton-conducting polymer and it fills at least part of the pores of said substrate.

25. A membrane-electrode assembly according to claim 15, wherein said non-noble metal comprises a first-row transition element, valve metal or coinage metal or a combination thereof.

26. A membrane-electrode assembly according to claim 15, further comprising first and second gas diffusion layers in contact with said catalytic layers.

27. A method for preparing a fuel cell comprising the steps of: providing first and second electrodes; positioning an ion conducting membrane between said electrodes; and disposing first and second porous catalytic layers on opposing surfaces of said membrane, each of said catalytic layers comprising a carbon-free non-noble metal porous nanostructured substrate and a noble metal surface layer.

28. A method according to claim 27, wherein said non-noble metal comprises a first-row transition element, valve metal or coinage metal or a combination thereof.

29. A method according to claim 27, wherein said noble metal comprises a platinoid.

Patent History
Publication number: 20100209815
Type: Application
Filed: Apr 28, 2010
Publication Date: Aug 19, 2010
Applicant:
Inventor: Yuriy Viacheslavovich Tolmachev (Kent, OH)
Application Number: 12/768,810
Classifications
Current U.S. Class: Membrane Electrode Assembly (mea) (429/483); Electrode Structure Or Composition (429/523); Including Platinum Catalyst (429/524); Making Catalytic Electrode, Process Only (502/101); Nanostructure (977/700)
International Classification: H01M 8/10 (20060101); H01M 4/36 (20060101); H01M 4/92 (20060101); H01M 4/88 (20060101);