Metal alloy catalysts for fuel cell anodes

Abstract

A catalyst for a fuel cell anode comprises an alloy of Pd and at least two other transition metals, at least one of which which binds to hydrogen and/or carbon monoxide at least as strongly as Pd does. Suitable transition metals which bind more strongly are Co, W, Ti, V, Cr, Fe, Mo, Nb, Hf, Ta, Zr and Re. PdCoW is the most preferred alloy. The catalyst is used on the anode of a hydrogen oxidising fuel cell, such as a PEMFC to catalyse the hydrogen oxidation reaction.

Claims

1. A fuel cell comprising: a hydrogen oxidising fuel cell anode comprising a substrate and a catalyst, the catalyst comprising an alloy of Pd and at least two other transition metals, at least one of which binds to hydrogen and/or carbon monoxide at least as strongly as Pd does, wherein the catalyst comprises a PdCoW alloy; a cathode; an electrolyte; a fuel supply; and a supply of oxidant; wherein the PdCoW alloy exhibits reduced absorption of hydrogen relative to unalloyed palladium.

2. The fuel cell of claim 1, wherein two of the at least two other transition metals bind to hydrogen and/or carbon monoxide at least as strongly as Pd does.

3. The fuel cell of claim 1, wherein at least one of the at least two other transition metals binds to hydrogen and/or carbon monoxide more strongly than Pd does.

4. The fuel cell of claim 1, wherein two of the at least two other transition metals bind to hydrogen and/or carbon monoxide more strongly than Pd does.

5. The fuel cell of claim 1, wherein the catalyst consists essentially of Pd, Co and W.

6. The fuel cell of claim 1, wherein the substrate is non-metallic.

7. The fuel cell of claim 6, wherein the substrate is carbon.

8. A fuel cell according to claim 1 wherein the fuel supply supplies hydrogen gas.

9. A fuel cell according to claim 1 wherein the electrolyte comprises a proton exchange membrane.

10. A method of oxidising fuel in a fuel cell comprising supplying fuel to an anode of the fuel cell, wherein the anode is as claimed in claim 1.

11. A method according to claim 10 wherein the fuel is hydrogen.

12. A method of generating electricity in a fuel cell according to claim 1 comprising supplying fuel to the anode and an oxidant to the cathode.

Description

(1) The invention will now be described in detail with reference to the figures, in which:

(2) FIG. 1 shows cyclic voltammograms (CVs) between approximately −0.025 V and 0.5-0.6 V vs. RHE for glassy carbon RDE electrodes with inks of: 70 wt. % Pt on C, 40 wt. % PdCoW on C and Pd black in Ar purged 0.1 M HClO.sub.4 at 20 mV s.sup.−1, all potentials are vs. RHE.

(3) FIG. 1a shows the same plots of FIG. 1, but just showing 70 wt. % Pt on C and 40 wt. % PdCoW on C;

(4) FIG. 2 shows the positive going sweeps of the CVs in the same region as FIG. 1 in Ar purged 0.1 M HClO.sub.4 at 20 mV s.sup.−1 for glassy carbon RDE electrodes with inks of each of the catalysts, the sweeps are compared with and without rotation at 400 rpm;

(5) FIG. 3 shows the positive going sweep of cyclic voltammetry experiments for glassy carbon RDE electrodes with inks of 70 wt. % Pt on C, 40 wt. % PdCoW on C and Pd black in H.sub.2 saturated 0.1 M HClO.sub.4 at 20 mV s.sup.−1 at 400 rpm;

(6) FIG. 3a is a zoomed view of the region in FIG. 3 where the currents first become positive (i.e. as hydrogen oxidation begins to occur); and

(7) FIG. 4 shows cyclic voltammograms between approximately 0 and 1 V vs. RHE for electrodes with inks of 40 wt. % PdCoW on C and Pd black in Ar purged 0.1 M HClO.sub.4 at 100 mV s.sup.−1 before and after hydrogen oxidation experiments have been carried out.

(8) FIG. 5 shows polarisation plots obtained in H.sub.2—O.sub.2 for Membrane Electrode Assemblies (MEAs) at 80° C. at various pressures (pressures shown in legend) for the 40 wt. % PdCoW catalyst at the anode (with a loading of 0.31 mg of PdCoW alloy cm.sup.−2) and a standard Pt catalyst at the cathode. Also shown for comparison is a polarisation curve (iR corrected) for an MEA containing an anode catalyst of Pt/C with an anode Pt loading of 0.3 mg cm.sup.−2, with a back pressure of 9 psi.

(9) The invention and its efficacy are further illustrated in the following Example.

(10) Three different catalysts have been investigated and compared in this study. These are a first comparative example of Pd black powder, a second comparative example of 70 wt. % (weight per cent) Pt on carbon powder and an example of the invention consisting of 40 wt. % PdCoW on carbon powder. The PdCoW catalyst has a composition of 50 at. % (atomic per cent) Pd, 43 at. % Co and 7 at. % W.

(11) The Pd black powder was obtained from Aldrich and the platinum on carbon was obtained from Alfa Aesar. PdCoW on carbon can be synthesised according to several well-known techniques, so the details are not provide here. WO2007042841; Fernández, J. L., D. A. Walsh, and A. J. Bard, Journal of the American Chemical Society, 2004. 127(1): p. 357-365; and Raghuveer, V., A. Manthiram, and A. J. Bard, The Journal of Physical Chemistry B, 2005. 109(48): p. 22909-22912 all of which are incorporated by reference, provide instruction on production of such catalysts.

(12) A small amount of each catalyst (40 mg) was made into an ink by mixing with 7 ml of deionised water and 80 μl of 5% nafion solution. The mixture was sonicated in an ultrasonic bath for 2 hours. 5 μl of an ink was pipetted onto a freshly polished glassy carbon rotating disc electrode (RDE), with a diameter of 0.5 cm. The electrode was polished prior to deposition of the ink using polishing pads and two different grades of alumina suspension (0.3 and 0.05 μm diameter alumina particles). Three minutes of polishing was carried out initially on 0.3 μm alumina to remove any previous material, then another 3 minutes on another pad with 0.3 μm alumina and finally 3 minutes on 0.05 μm alumina. The inks were left to dry overnight and the electrode was then attached to a rotating shaft and lowered into a three compartment electrochemical cell for analysis. A platinum gauze electrode was used as a counter electrode and a commercially available Mercury-Mercury Sulphate reference electrode was used. All potentials have been converted and will be quoted against the Reversible Hydrogen Electrode (RHE). All experiments were carried out in 0.1 M HClO.sub.4 (aq). Cyclic voltammetry experiments were carried out with and without rotation of the electrode using a potentiostat built in-house.

(13) The electrochemical screening procedure carried out on each sample is shown in Table 1 below.

(14) TABLE-US-00001 TABLE 1 Experimental procedure applied to the catalyst inks deposited on glassy carbon RDE electrodes. Potential limits/ Sweep rate/ Rotation Experiment Gas V vs. RHE mV s.sup.−1 rate/rpm Bubbling Ar 20 min 3 CVs in Ar above   0.025-1.00 100 0 deoxygenated solution solution Bubbling Ar −0.002 10 min 3 CVs in Ar above −0.025-0.5 3 0 deoxygenated solution solution Bubbling Ar −0.002 10 min 3 CVs in Ar above −0.025-0.5 3 400 deoxygenated solution solution Bubbling Ar −0.002 10 min 3 CVs in Ar above −0.025-0.5 3 900 deoxygenated solution solution Bubbling Ar −0.002 10 min 3 CVs in Ar above −0.025-0.5 3 1600 deoxygenated solution solution Bubbling H.sub.2 10 min 3 CVs in H.sub.2 H.sub.2 above −0.025-mass transport 3 0 saturated solution solution limited region (between 0.4 and 0.6 V) Bubbling H.sub.2 10 min 3 CVs in H.sub.2 H.sub.2 above −0.025-mass transport 3 400 saturated solution solution limited region (between 0.4 and 0.6 V) Bubbling H.sub.2 10 min 3 CVs in H.sub.2 H.sub.2 above −0.025-mass transport 3 900 saturated solution solution limited region (between 0.4 and 0.6 V) Bubbling H.sub.2 10 min 3 CVs in H.sub.2 H.sub.2 above −0.025-mass transport 3 1600 saturated solution solution limited region (between 0.4 and 0.6 V) Bubbling Ar 20 min 3 CVs in Ar above   0.025-1.00 100 0 deoxygenated solution solution 3 CVs in Ar above   0.025-1.00 20 0 deoxygenated solution solution

(15) FIG. 1 shows the cyclic voltammograms (CVs) for electrodes with the comparative examples of palladium black and 70 wt. % Pt on carbon and the example of 40 wt. % PdCoW on carbon in Ar purged 0.1 M HClO.sub.4 at 20 mV s.sup.−1 in the hydrogen adsorption/desorption and hydrogen evolution/oxidation region (i.e. between approximately −0.025 and 0.6 V vs. RHE).

(16) Between approximately 0 and 0.4 V vs. RHE all of the catalysts show features relating to adsorption of hydrogen (reduction currents—negative) and desorption of hydrogen (oxidation currents—positive). Below approximately 0 V vs. RHE on all of the catalysts, the onset of hydrogen evolution can be seen as a sharp onset reduction current, this can be seen more clearly for the Pt and PdCoW catalysts in FIG. 1 a, because FIG. 1a is scaled to the currents obtained. On the Pd black catalyst a larger reduction current is seen below 0 V vs. RHE, this is believed to be due to not only hydrogen evolution, but also absorption of hydrogen into the bulk structure of the palladium. The Pd black catalyst also has a potentially larger surface area of active catalyst than the other catalysts, leading to higher currents. The large oxidation peak seen on the reverse sweep is due to oxidation of not only evolved hydrogen from the surface, but also oxidation of hydrogen absorbed within the bulk of the palladium. On the Pt and PdCoW catalysts an oxidation peak just above 0 V vs. RHE is seen, this is due to oxidation of the hydrogen evolved at the electrode surface. It can be seen that the features of the PdCoW catalyst are very similar to those seen on the Pt catalyst and that peaks relating to the absorption of hydrogen and oxidation of absorbed hydrogen do not appear to be present, suggesting no significant absorption of hydrogen is occurring within the palladium alloy. As mentioned above, absorption, in contrast to adsorption, results in large volume changes which damage the catalyst. Accordingly the absence of absorption is advantageous.

(17) FIG. 2 shows the positive going sweeps of the CVs in the same region as FIG. 1 for each of the catalysts with and without rotation of the electrode in Ar purged 0.1 M HClO.sub.4. For the platinum electrode (comparative example) when the electrode is rotated the currents seen are decreased and there is no longer a peak just above 0 V relating to oxidation of hydrogen that has been evolved at the electrode surface. This suggests that the evolved hydrogen is moved away from the electrode during rotation. The remaining currents seen are due to desorption of adsorbed hydrogen on the electrode surface.

(18) A similar effect is seen on the PdCoW catalyst of the invention, this confirms that the electrode does not significantly absorb hydrogen. In contrast the Pd black electrode (comparative example) has quite different features. When the electrode undergoes rotation, the currents seen are also decreased similar to the other two electrodes; this is presumed to be due to the removal of evolved hydrogen on or near the electrode surface under rotation. However a large oxidation current is still observed, due to the oxidation of hydrogen absorbed in the bulk of the material.

(19) When Pd has been alloyed with Au, Pt, Rh or Pt and Au in the prior art, less hydrogen absorption and faster kinetics of the absorption, desorption process has been seen (Lukaszewski, M., K. Hubkowska, and A. Czerwinski, Physical Chemistry Chemical Physics, 2010. 12(43): p. 14567-14572). This effect has been not only associated with changes in the lattice structure due to alloying, but due to electronic effects (i.e. lower binding energies with hydrogen—all of the alloying elements selected bind hydrogen less strongly than Pd). However both Co and W bind hydrogen at least as strongly as Pd, therefore the surprising results here might be explained by changes in the lattice structure of palladium on alloying, which may not allow the hydride to form.

(20) FIG. 3 shows the positive going sweep of cyclic voltammetry experiments for the three different catalysts (i.e. the comparative examples of palladium black and 70 wt. % Pt on carbon and the example of 40 wt. % PdCoW on carbon) in H.sub.2 saturated 0.1 M HClO.sub.4 at 20 mV s.sup.−1 at 400 rpm. FIG. 3a is an enlarged view of the region where the currents first become positive (i.e. as hydrogen oxidation begins to occur). Above approximately 0.4 V vs. RHE all of the catalysts show a flat current profile. In this region the hydrogen oxidation reaction is limited by transport of hydrogen to the surface of the electrode from solution. The difference between the currents in this region for the different electrodes is considered to be mainly due to slightly different geometric surface areas of the ink deposits. The ink deposits do not cover the whole area of the electrode surface and therefore will have formed deposits with differing shapes and sizes. All of the hydrogen that is getting to the surface is reacting in this potential region, therefore differences in current can be explained by the difference in the size of the ink deposits rather than any effect of the specific catalysts.

(21) The plot of FIG. 3a gives more information about which catalysts are more active. At very low potentials between 0-0.005 V vs. RHE, the Pt (comparative example) and PdCoW (example) catalyst traces show a very similar gradient, however the Pd black (comparative example) catalyst has a slightly less steep onset gradient. This might suggest a slightly faster reaction on the Pt and PdCoW catalyst than on the Pd black, however the differences are only slight and the hydrogen oxidation reaction occurs so fast that these differences are hard to quantify. The most significant difference between the three catalysts is the large oxidation peak seen on the Pd black catalyst, due to the oxidation of absorbed hydrogen from the bulk of the material.

(22) FIG. 4 shows a comparison between the cyclic voltammetry in Ar purged 0.1 M HClO.sub.4 at 100 mV s.sup.−1 before and after hydrogen oxidation experiments have been carried out for the PdCoW catalyst (of the invention) and the Pd black catalyst (of a comparative example). All of the voltammograms show features relating to hydrogen adsorption and desorption (between approximately 0 and 0.4 V vs. RHE) and palladium oxide formation and reduction (between approximately 0.6 and 1 V vs. RHE). The charge under the hydrogen adsorption and desorption features give a rough indication of the palladium surface area of the catalysts. The charge under these features on the PdCoW catalyst have decreased slightly after hydrogen oxidation experiments have been carried out. This suggests a small decrease in the palladium surface area, which may be due to a small loss of palladium from the material or due to restructuring of the material. However these changes are much more significant for the Pd black catalyst, suggesting a significant change in the palladium surface area. This is probably due to a significant loss in material during the experiments due to the expansion and contraction of the structure with absorption and desorption of hydrogen.

(23) A MEA (membrane electrode assembly) was produced with a standard Pt catalyst at the cathode and a 40 wt. % PdCoW alloy supported on carbon at the anode (with an anode loading of 0.31 mg (of the alloy) cm.sup.−2, with an ink containing 25 wt. % nafion).

(24) The MEA was operated with a hydrogen and oxygen feed (λH.sub.2=1.5, λO.sub.2=2.5) at the anode and cathode respectively at 80° C. and at pressures of 7 psi, 15 psi, 25 psi and 30 psi. The active electrode area was 50 cm.sup.2 and the membrane used was a standard IRD polymer electrolyte membrane (PFSA 30 μm). FIG. 5 shows the polarisation curves from an un-optimised MEA and shows a promising open circuit potential of around 1 V at all pressures. This result provides exemplification that the catalyst is indeed suitable as an anode catalyst for a hydrogen fuel cell, as this provides direct evidence of the use of the catalyst at the anode in a complete cell suitable for incorporation into a fuel cell stack.

(25) FIG. 5 also shows, for comparison, the polarisation curve containing standard platinum on carbon catalyst at both anode and cathode (with an anode loading of 0.3 mg Pt cm.sup.−2, with an ink containing 30 wt. % nafion, operated at 9 psi, all other conditions as above). The data shown for this standard catalyst has been iR corrected. In comparison to the data obtained for the Pd-alloy anode catalyst a similar open circuit potential is seen for the Pt anode catalyst, suggesting similar activity for the two anode catalysts. Loss in performance for the Pd-alloy catalyst anode containing MEA at higher current density is likely to be due to internal resistance and mass transport limitations. It is likely that these factors can be improved by optimisation of the catalyst ink.

(26) The results set out above reveal that a catalyst according to the invention has good activity for the hydrogen oxidation reaction. The catalyst has been shown to have an advantage over pure palladium catalysts in that it does not significantly absorb hydrogen and therefore will not undergo large volume changes at potentials required for hydrogen oxidation. This advantage, and the fact that a catalyst similar to that of the example has already shown to be stable over a wide range of potentials seen in a fuel cell (in WO2007042841), suggests that these catalysts could find application as a cheaper replacement for platinum electrodes for anodes in fuel cells (PEM fuel cells and the hydrogen oxidation reaction in particular).

(27) Although the catalyst has previously been shown as an effective oxygen reduction catalyst (in WO2007042841) it had been suggested that to produce good hydrogen evolution (and hence hydrogen oxidation) catalysts, palladium should be alloyed with an element that binds hydrogen less strongly. Both Co and W bind hydrogen at least as strongly as palladium and there is therefore a surprising effect that the catalyst appears active and attenuates hydrogen absorption. It is believed that reduced absorption of hydrogen by the palladium may be due to changes in the lattice structure of palladium by the alloying elements.

(28) From the results shown in the current investigation it might be expected that palladium could be alloyed with other elements that bind hydrogen and CO at least as strongly as Pd to provide promising hydrogen oxidation (and evolution) catalysts. This may include but not be limited to ternary palladium alloys with which may have already been shown or suggested to be stable in a fuel cell environment, e.g. PdCoAu, PdCoMo, PdCoCr, PdFeCr. Accordingly, the invention is not limited to the scope of the example set out above, but should be determined by reference to the following claims.

(29) The catalyst according to the invention is used in the anode of a fuel cell.

(30) The detailed construction of PEM fuel cells is well known to those familiar with such technology. Fuel Cell Handbook (Seventh Edition) by EG&G Technical Services, Inc., available to the public from the National Technical Information Service, U.S. Department of Commerce, 5285 Port Royal Road, Springfield, Va. 22161, U.S.A. and incorporated by reference, provides useful information for those less familiar with the art; Section 3, Polymer Electrolyte Fuel Cells is particularly relevant. Typically the fuel cell comprises an anode, a cathode, a proton exchange membrane between the anode and the cathode, and catalysts for the catalytic oxidation of a hydrogen-containing fuel and for the reduction of oxygen.

(31) A typical hydrogen fuel cell has a hydrogen electrode (fuel electrode or anode) and an air electrode (oxidizer electrode or cathode). In between the electrodes, a proton exchange membrane serves as an electrolyte. The proton exchange membrane, the anode and the cathode are generally integrated into one body, so there is no contact resistance between the electrodes and the proton exchange membrane. Electricity is generated by hydrogen oxidation by introducing hydrogen into a hydrogen fuel chamber open to the anode, while oxygen, preferably as air, is introduced into an air chamber open to the cathode. The hydrogen is oxidised at the anode to produce hydrogen ions (protons) which pass through the PEM, and electrons. An electric current is withdrawn from the fuel cell into an outer circuit by current collectors in contact with the electrodes. Hydrogen ions migrate through the acidic proton exchange membrane and react with oxygen and electrons from the outer circuit at the cathode to form water.

(32) The catalyst can be supported on a commercially available support such as but not limited to carbon; carbon supported catalysts are normally used for fuel cells because surface area of catalysts is greatly increased by using carbon support and carbon has sufficient electronic conductivity and chemical stability under fuel cell operating conditions. Alternative supports include metal oxides, for example. The preparation of dispersed alloy on carbon supports may be achieved in a number of ways. For instance, the alloy catalysts may be formed by reduction of a mixture of compounds of component metals, or heat treatment of carbon supported Pd wherein the other metal salts are precipitated or adsorbed onto the surface. Alternatively, the alloy particles may be formed on the carbon supports by physical deposition, such as sputtering, physical evaporation and chemical vapour deposition.

(33) The fuel cell includes an electrolyte such as a proton exchange membrane (e.g. NAFION® available from Du Pont) alloy catalysts may be deposited directly on the surface of the proton exchange membrane for contact with a current collector. Alternatively the alloy catalysts may be deposited on the surface of an anode support, or within the pores of a porous anode support, such as a carbon structure that is placed in contact with the membrane.