(205a) Synthesis of Improved Fuel Cell Catalysts Using Electroless Deposition Methods

Introduction

In an effort to reduce the amount of noble precious metals used in electrochemical catalysts and to increase catalytic activity, a novel method of synthesis that involves electroless deposition (ED) of a Pt salt on a carbon support that has been previously seeded with a group VIIIB metal has been developed.  Our previous work [1,2] has focused on the ED of Pt onto other noble metals such as Rh and Pd.  This methodology has been extended to include deposition of Pt on base metals as well as deposition of base metals on Pt. Others [1] have shown that electronic interactions between Pt and base metals such as Co, Ni, and Cu may occur, which can increase the activity of the Pt towards oxygen reduction.  However, the highly acidic environment of a fuel cell typically corrodes the base metal, leading to reduced performance.  Thus, a core-shell structure of a Pt shell over a base metal core would be advantageous to gain the activity enhancements of the Pt-base metal interaction while mitigating the corrosion of the base metal, now protected by the Pt shell.  This presentation illustrates aspects of ED to give this preferred core-shell type structure as well as how the controlled deposition of Pt gives an overall reduction in metal particle size and a more efficient utilization of the expensive Pt component.  X-ray photoelectron spectroscopy (XPS) is used to provide strong evidence of a core/shell and solid solution for Pt electrolessly deposited on Co and Pd, respectively, while HRTEM confirms the synthesis of smaller Pt-containing particles using ED methods.  The oxygen reduction activity of the synthesized ED catalysts is determined using a rotating disc electrode (RDE) to correlate the promoter effect of possible electronic interactions between the precursor base metal and the electrolessly-deposited Pt on overall fuel cell performance.

Materials and Methods

The preparative method consists of using commercial carbon black which has been seeded with a variety of Group VIIIB metals using conventional, wet impregnation techniques.  Platinum metal deposition from reduction of PtCl₆^2- ions occurred on the seeded metal particles by the Group VIIIB metal-catalyzed activation of a reducing agent, dimethylamine borane (DMAB) in our studies, at the seed sites.  Therefore, the catalytically active metal site serves as both anode and cathode for the reduction of PtCl₆^2- ions.  This methodology should result in a core/shell geometry for the catalyst particles, assuming the deposited Pt metal is itself catalytic (autocatalysis) for further electroless deposition.  Deposition does not occur on the carbon support since there are no sites capable of activating the reducing agent.  The kinetic dependencies of the Pt deposition on metal surface include pH of the ED solution as well as concentrations of reducing agent, Pt⁴⁺ salt source, and Pt⁴⁺ complexing agent.

Prepared catalysts were analyzed using AA, TEM and HRTEM, selective H₂ chemisorption, oxygen reduction with rotating disk electrode (RDE), and XPS.   AA was used to determine final weight loadings of both platinum and the Group VIIIB seed metal.  Particle sizes and metal dispersions were determined for the Pt and the Group VIIIB metal using, TEM, HRTEM, and H₂ chemisorption.  Electrochemically-active surface area and oxygen reduction activity were established by RDE studies.  Lastly, verification of core/shell structures, surface compositions, and evidence of electronic interactions between Pt and the precursor metal were established with XPS.

Results and Discussion

TEM Analysis

TEM characterization has shown that ED methods form Pt particles smaller than those present in a commercially-available 20% Pt/C electrocatalyst to give 30% more Pt surface sites than the commercial 20% Pt/C sample.  Further, the ED-derived samples also have considerably narrower particle size distributions, as seen in Figure 1.  The overall smaller Pt particles as well as the narrower particle size distributions provide increased dispersion of Pt.  Table I tabulates the important data taken from Figure 1 which shows that the number of Pt particles/g_cat for the ED catalysts and the standard sample are quite similar, despite the ED samples containing only 40% of the total Pt in the standard 20% Pt/C sample.  Also, the concentration of Pt surface sites for the best ED catalysts is approximately 60% that of the standard catalyst while containing approximately 60% less Pt.  Thus, the concentration of surface Pt sites and Pt particles, once normalized to Pt loading, are greater for the ED catalysts compared to the standard catalyst. 

Figure 1: Particle size distribution curves for the ED catalysts (8.0% Pt on 5.0%, 2.5%, 1.0%, and 0.5% Pd seeds on carbon) and the standard catalysts (20% Pt on carbon).

Table I: Tabulated data from analysis of TEM images and Figure 1.

Determining Particle Structure with XPS

The structure of the catalyst particles synthesized by ED was examined by XPS.  The Pd 3d orbitals were analyzed by XPS for 0.5%, 1.0%, 2.5%, and 5.0% Pd/C samples and 7% Pt-0.5% Pd/C, 8% Pt-1.0% Pd/C, 8% Pt-2.5% Pd/C, 8% Pt-5.0%Pd/C samples; thus, nearly identical quantities of Pt have been electrolessly deposited on varying quantities of Pd.  As shown previously in Figure 1, the average Pt particle size decreases with increasing Pd loading.  Another factor to consider here is the escape depth of photoelectrons.  Seah and Dench [2] estimated the photoelectron escape depth to be approximately 6 to 7 monolayers (corresponding to 18 to 21 angstroms) through an element at a binding energy of approximately 330 eV.  Thus, the XPS is not a strictly surface-sensitive instrument and is instead a near-surface experiment. 

It was originally surmised that the Pt-Pd particles would have a core/shell structure because Pt must be initially deposited on top of the Pd precursor catalyst during the process of electroless deposition.  Furthermore, the reducing agent DMAB kinetically favors deposition of Pt on Pd rather than Pt on Pt; thus, the Pd should be first covered with Pt before autocatalytic deposition of Pt occurs.  Figure 2 shows an example of some early data which supported this first hypothesis.  The plots show the change in the Pd 3d orbital as analyzed by XPS for 0.5%, 1.0%, 2.5%, and 5.0% Pd/C before and after approximately 8% Pt is electrolessly deposited on the Pd precursor.  In the case of 0.5% and 1.0% Pd/C, a strong (although not complete) suppression of the Pd 3d orbital is observed which is not true for the 2.5% and 5.0% Pd/C samples.  Deconvolution of the Pd 3d peaks after electroless deposition of Pt, indicated that every peak contained both Pd⁰ and Pd²⁺ states.  These results shed doubt on the core/shell structure hypothesis because, if fully encapsulated, the Pd should remain in a metallic (Pd⁰) state since it should be protected from the oxidizing environment (room temperature air) by the Pt overlayer.

Figure 2: Pd XPS spectra for similar weight loadings of Pt deposited on varying amounts of Pd.

            The results of deconvolution are presented below in Table II as the fraction of Pd which is in the metallic state.  This fraction is very low prior to electroless deposition of Pt; however, the fraction only reaches approximately 60% even after the Pt is deposited.  It is also interesting to note that the fraction of Pd in the metallic state does not change appreciably with Pd loading which dictates the relative amount of Pt in each Pt-Pd particle.

Table II: Fraction of metallic Pd at near surface of Pt-Pd nanoparticles

Another series of experiments using in-situ pretreatments was carried out.  For this experiment, all bimetallic compositions were reduced at 200^oC in H₂; however, Pt on 0.5% Pd/C and 1.0% Pd/C were also reduced at a lower temperature (75^oC) before reduction at 200^oC to ensure that any change in composition was the result of reduction only and not intraparticle diffusion.  Further, these two samples were subjected to a heat treatment in vacuum at 400^oC (Pt on 0.5% Pd/C), 540^oC, and cooling from 540^oC to room temperature (Pt on 1.0% Pd/C) after reduction.  The results of these experiments are summarized in Table III.  All samples showed some increase (approximately 10 - 20%) in surface molar concentration of Pd after the initial reduction.  Subsequent treatments (higher reduction temperature, heat treatment, cooling) had only marginal effects on the surface composition.  Finally, the observed Pd surface concentration was always less than the bulk concentration of Pd (aside from Pt on 5.0% Pd/C after reduction) indicating a Pt-rich near surface region.  The composition of this near surface region is clearly proportional to bulk composition (e.g. increasing Pd loading from 0.5% to 1.0% approximately doubles the near-surface concentration of Pd).  However, the apparent equilibrium reached after the first reduction suggests the presence of a Pt-rich solid solution in the near surface area of the nanoparticles which is unaffected by heat treatment. 

Table III: Effect of Reduction and Heat Treatment on Near-Surface Composition

References

1.      T. Toda, H. Igarashi, H. Uchida, M. Watanabe, Journal of the Electrochemical Society, 146 (1999) 3750 - 3756.

2.      M.P.~Seah, and W.A.~Dench, Surface and Interface Analysis, 1 (1979) 2-11.