Journal of Nanoparticle Research

, Volume 11, Issue 4, pp 965–980

PtPb nanoparticle electrocatalysts: control of activity through synthetic methods

Authors

  • Tanushree Ghosh
    • Department of Chemistry and Chemical Biology, Baker LaboratoryCornell University
  • Futoshi Matsumoto
    • Department of Chemistry and Chemical Biology, Baker LaboratoryCornell University
  • Jennifer McInnis
    • Department of Chemistry and Chemical Biology, Baker LaboratoryCornell University
  • Marilyn Weiss
    • Department of Chemistry and Chemical Biology, Baker LaboratoryCornell University
  • Hector D. Abruña
    • Department of Chemistry and Chemical Biology, Baker LaboratoryCornell University
    • Department of Chemistry and Chemical Biology, Baker LaboratoryCornell University
Research Paper

DOI: 10.1007/s11051-008-9557-y

Cite this article as:
Ghosh, T., Matsumoto, F., McInnis, J. et al. J Nanopart Res (2009) 11: 965. doi:10.1007/s11051-008-9557-y

Abstract

Solution phase synthesis of intermetallic nanoparticles without using surfactants (for catalytic applications) and subsequent control of size distribution remains a challenge: of growing interest, but not widely explored yet. To understand the questions in the syntheses of Pt containing intermetallic nanoparticles (as electrocatalysts for direct fuel cells) by using sodium naphthalide as the reducing agent, the effects of the Pt precursors’ organic ligands were investigated. PtPb syntheses were studied as the model case. In particular, methods that lead to nanoparticles that are independent single crystals are desirable. Platinum acetylacetonate, which is soluble in many organic solvents, has ligands that may interfere less with nanoparticle growth and ordering. Interesting trends, contrary to expectations, were observed when precursors were injected into a reducing agent solution at high temperatures. The presence of acetylacetonate, from the precursor, on the nanoparticles was confirmed by ATR, while SEM imaging showed evidence of morphological changes in the nanoparticles with increasing reaction temperature. A definite relationship between domain size and extent of observed residue (organic material and sodium) present on the particles could be established. By varying post-reaction solvent removal techniques, room temperature crystallization of PtPb nanoparticles was also achieved. Electrochemical activity of the nanoparticles was also much higher than that of nanoparticles synthesized by previous reaction schemes using sodium naphthalide as the reducing agent. Along with the above mentioned techniques, BET, TEM, CBED, SAED, and XRD were used as characterization tools for the prepared nanoparticles.

Keywords

ElectrocatalysisPtPbIntermetallicsNanoparticlesFuel cells

Introduction

The need for more efficient and alternative energy sources to cope with the issues arising from fossil fuel usage is becoming more and more evident each day. One potential technology for higher efficiency is direct fuel cells operating near room temperature. So called direct fuel cells use small molecules as the fuel source, completely oxidizing the fuel to CO2 and H2O (Lamy and Srinivasan 2001). Fuel cells are not heat engines so their efficiency is not bound by Carnot cycle limitations; on the other hand the entire free energy of a redox reaction can (in principle) be converted to electrical energy. Real fuel cells with efficiencies of about 50–60% at their operating point with 10–50 kW outputs are already realizable (Shah and Kandlikar 2003). Cost and durability/longevity factors, however, still keep these fuel cells from being practically viable, which is inhibiting their introduction into the market. Nevertheless, the potential of polymer electrolyte membrane fuel cells (PEMFCs) in the transportation market and as small portable power sources is being explored by many. The basic components of the fuel cell, including the electrocatalyst, present a challenge for most fuel cell technologies. The common industry standard for the anode electrocatalyst is still Pt (for H2 as a fuel) or a 1:1 PtRu alloy (for MeOH as a fuel) usually in the form of 3–5 nm particles supported on carbon to achieve high surface areas and generate the high current densities needed (Liu et al. 2006). The use of nanoparticles also diminishes the amount (and therefore cost) of Pt used.

Recently, the authors and their collaborators have shown that high performance electrocatalysts can be obtained from the family of Pt-based ordered intermetallic compounds (Casado-Rivera et al. 2003; Casado-Rivera et al. 2004; Volpe et al. 2004; Oana et al. 2005; Roychowdhury et al. 2005; Blasini et al. 2006). Two of these intermetallics, PtBi, and PtPb have, indeed, shown not only higher tolerance to S and CO poisoning (surface poisoning of the catalyst by CO is a pertinent issue when carbon-containing fuels are considered) but also superior electrocatalytic activity toward formic acid oxidation compared with Pt, 1:1 PtRu and Pd (Casado-Rivera et al. 2003; Casado-Rivera et al. 2004; Volpe et al. 2004; Roychowdhury et al. 2005; Roychowdhury et al. 2006; Alden et al. 2006a, b). However, the electrocatalytic activity, along with the size and surface properties of the nanoparticles were found to be highly dependent on the details of the synthesis methods. Initially, sodium borohydride was used as a reducing agent for preparing PtPb and PtBi intermetallics, (Roychowdhury et al. 2005; Roychowdhury et al. 2006; Alden et al. 2006b) but in order to explore intermetallics of metals with a wider range of reduction potentials as anode electrocatalysts, stronger reducing agents needed to be considered. Also, it is important to prepare the particles by a number of different chemical routes to investigate the role of synthesis in determining the surface composition, morphology, impurity contamination, and subsequently, electrochemical activity of the nanoparticles.

PtPb nanoparticle synthesis using sodium naphthalide as the reducing agent and dimethyl(1,5-cyclooctadiene)platinum and lead(II)2-ethylhexanoate as the precursors was also undertaken previously in the group (Alden et al. 2006a, b), and interestingly, the domain size of the nanoparticles were found to be dependent on stirring time and reaction temperature. For reductions done at room temperature, crystallization of the nanoparticles giving an identifiable PtPb phase by XRD was observed only when the particles rapidly got warmed by surface oxidation when exposed to air quickly post synthesis. These observations indicated a ‘barrier’ to crystallization of the PtPb nanoparticles that required a thermal treatment. Some observations suggested that the presence of organics from the precursors (or the reducing agent) bound to the nanoparticle surfaces (or even inside the nanoparticles) contributed to this crystallization ‘barrier’. Varying the platinum precursors and studying the resulting nanoparticles is one way to explore this phenomenon. Also, if these organics are the main inhibitors of crystallization, quantification of the amount of organics in or on the nanoparticles, and possible correlation with particle domain sizes would be revealing. Moreover, finding precursors with ligands having smaller binding affinity to platinum would be useful for further syntheses of Pt-based intermetallics. Another important issue with these solution-based syntheses is the solubility of platinum and other metal precursors. Using sodium naphthalide as the reducing agent limits solvent options to tetrahydrofuran (THF) or glymes. The platinum precursor (dimethyl(1,5-cyclooctadiene)platinum), though considerably soluble in THF, has fairly low solubility in diglyme (a THF substitute for higher temperature reactions). Comparable solubility of both metal precursors in the reaction solvent is highly desirable for the purpose of co-reduction of metals with widely different reduction potentials. Hence, higher solubility is a desirable attribute for the platinum precursors to be investigated. Up to now, studies have been done for syntheses of unsupported PtPb nanoparticles. Once a suitable method is established and understood, the study would be extended toward supported intermetallic nanoparticle syntheses. The above research has led us to a plausible mechanistic picture of the processes that occur in the synthesis of intermetallic nanoparticles from solution by chemical co-reduction of the precursors (Scheme 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9557-y/MediaObjects/11051_2008_9557_Sch1_HTML.gif
Scheme 1

Key issues in synthesis of intermetallic nanoparticles by chemical co-reduction of precursors

We started our investigation by synthesizing a number of Pt precursors having 1,5-cyclooctadiene derivatives in place of the 1,5-cyclooctadiene (hoping that alkyl-based derivatives would make the molecule less symmetric and more soluble) and screened them primarily based on their solubility in THF and diglyme. We also started working with platinum acetylacetonate (obtained commercially) following our quest for a platinum precursor with ligands having lower affinity for platinum (on the basis of Pt–L bond enthalpies) (Mortimer 1984). We used the precursors having considerable solubility in THF and diglyme (Table 1) to synthesize PtPb. We also explored different synthetic strategies with platinum acetylacetonate (the most soluble one of all). Interestingly, the resulting nanoparticle domain sizes showed different behaviors with different syntheses or annealing temperatures as well as different electrocatalytic activities depending on the method of synthesis employed. It was also possible to identify some organic groups from the precursor in the products. Quantitative studies further revealed a clear correlation between the amount of organics and crystal domain sizes supporting our hypothesis about organic interference in PtPb nanoparticle crystallization. Sodium was also detected in the product, but could be removed by water wash. However, evidence suggested that sodium may also play a role in the crystallization dynamics. A modification in the method of solvent removal post reduction resulted in lower organic and sodium contamination. Thermogravimetric analyses and IR studies of the particles confirmed the presence of organics. At the same time, organics from the reducing agent/lead precursors have not yet been detected. The SEM and TEM images of the nanoparticles prepared from Pt acetylacetonate offered evidence of the change in nature and composition of the particles crystallizing when the reaction temperature was increased. Although the particles are always agglomerated (expected, since stabilizers/surfactants were not employed), crystal grain boundaries were observable as the smallest features in the images (with domain sizes corresponding to those measured from XRD of the particles). The CBED and SAED studies confirmed the crystalline nature of these particles.
Table 1

Solubility of different platinum precursors in THF and diglyme

Compound

Solubility in mmol/mL

 

In THF

In Diglyme

PtMe2https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9557-y/MediaObjects/11051_2008_9557_Figa_HTML.gif

0.108

0.045

PtCl2https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9557-y/MediaObjects/11051_2008_9557_Figb_HTML.gif

0.006

PtCl2−https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9557-y/MediaObjects/11051_2008_9557_Figc_HTML.gif

0.011

0.010

PtCl2−https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9557-y/MediaObjects/11051_2008_9557_Figd_HTML.gif

0.002

0.002

Pt(acac)2

0.049

0.041

More interestingly, the PtPb nanoparticles prepared from Pt acetylacetonate had the highest electrocatalytic activity (mass activity normalized to the BET surface area) compared to nanoparticles previously synthesized using sodium naphthalide or sodium borohydride as reducing agents. In each case, however, different precursors were used (Roychowdhury et al. 2006; Alden et al. 2006a, b).

Experimental section

Platinum precursors

All materials used were at least reagent grade. All platinum precursors, except Pt acetylacetonate and Pt hexafluoroacetylacetonate (Aldrich), were synthesized in the laboratory starting from Platinum (Englehard, 99.9 %). Platinum (1 g, 5.13 mmol) was dissolved in 20 mL of aqua-regia (three parts of 12 M HCl with one part of 12 M HNO3), and stirred in an open flask at 120 °C for 30 min. The resulting orange solution was cooled down to room temperature and solid KCl (0.848 g, 11.4 mmol) was added. The resulting mixture was cooled in an ice bath for 30 min, and the yellow precipitate was collected by suction filtration, and washed with chilled water, ethanol, and diethyl ether. XRD showed the dried precipitate to be pure K2PtCl6. The K2PtCl6 was reduced to K2PtCl4 with H2SO3 (commercially obtained from Aldrich) following a standard reduction procedure (Keller 1946). The K2PtCl4 was dissolved in glacial acetic acid (Aldrich) and water (purified with a Millipore Milli Q system) in a 12.5:8 by volume solution, and then 1,5-cycloctadiene, or 3-methyl 1,5-cycloctadiene, or 3,4-dimethyl 1,5-cyclooctadiene (all from Aldrich) was added to obtain the corresponding platinum organic dichloride compound. Platinum acetate was synthesized by refluxing Pt tetrachloride (0.406 g, 1.21 mmol) and silver acetate (0.1 g, 0.6 mmol) for 3 h in 24 mL glacial acetic acid and worked up according to a reported procedure (Basato 2003). However, we found that washing the final product with acetone improved its purity. A recrystallization of the product from acetone was also attempted. However, the attempt failed due to the high solubility of platinum acetate in acetone. Platinum acetate was found to decompose to platinum metal (which precipitated out) when its solution in acetone was heated to just 60 °C. Thus, all reactions with this precursor were carried out by injecting the precursors into reducing agent at room temperature, heating up only later, if necessary.

Naphthalide reduction

Materials

Naphthalene was purchased from Fisher. Sodium metal and diglyme were purchased from Aldrich. THF and diglyme were freshly distilled over sodium prior to use. Lead(II)2-ethylhexanoate was purchased from STREM Chemicals Inc. Because of the high viscosity of lead(II)2-ethylhexanoate, 0.05 M stock solutions were prepared by dissolving 2.496 g in 100 mL THF.

Synthesis of PtPb nanoparticles

A solution of sodium naphthalide was prepared by weighing out stoichiometric amounts of naphthalene (0.1923 g, 1.50 mmol) and sodium metal (0.0345 g, 1.50 mmol) in an argon atmosphere glovebox and loading these reactants into a flask containing 50 mL diglyme or THF. The flask was sealed and removed from the glovebox and stirred overnight under argon to produce a dark green sodium naphthalide solution.

Platinum acetylacetonate (0.0988 g, 0.25 mmol) was dissolved in a 5.0 mL aliquot of a 0.05 M stock solution of lead(II)2-ethylhexanoate in diglyme (0.25 mmol lead(II)2-ethylhexanoate). Meanwhile, an argon-purged flask containing the freshly prepared sodium naphthalide was connected to a water-cooled condenser. If the reaction was to be carried out above room temperature, the flask was heated to the desired reaction temperature in an oil bath. The aliquot of metal precursors was injected into the dark green sodium naphthalide solution which immediately became black and opaque. The solution was stirred for 30 min. The reaction mixture was being heated (either before—labeled series 1, or immediately after the reduction step—labeled series 2), was cooled to ambient by removing the oil bath. The solvent and any volatile side products were carefully pumped out of the flask using first static, and then followed by dynamic vacuum, leaving a black residue. After the flask was pumped down to a suitably low pressure (~200 m Torr), 40 mL of degassed ethanol was added to the flask, which was then sonicated in an immersion bath sonicator for 10 min. The contents of the flask were then transferred to a centrifuge tube with a cannula and centrifuged (2 K rpm, 10 min). The supernatant (colored yellow) was removed and 40 mL of hexanes were distilled into the black precipitate. The tube was sonicated for 10 min and centrifuged again. The supernatant (usually colorless) was removed and the black precipitate was dried under vacuum. The tube was then backfilled with argon, and was left overnight with a needle in the septum to allow slow exposure of the nanoparticles to air. Another series (labeled series 3) of reactions were done in which the preparation of reducing agent and injection of precursors were done in the above-mentioned way. After the injection, the flask was heated up to the desired temperature using an oil bath (it took approximately 10–15 min for the bath to reach the desired reaction temperature). After the desired temperature was reached, the flask was stirred for 10 min and the solution was allowed to cool to 40 °C. The reaction mixture was cannula transferred into a centrifuge tube and was centrifuged for 10 min at 2,000 rpm. The diglyme was syringed off leaving a black residue. Forty milliliters of degassed ethanol was added to the flask, which was then sonicated in an immersion bath sonicator for 10 min. The contents of the flask were then transferred to a centrifuge tube with a cannula and centrifuged (2,000 rpm, 10 min). The supernatant was removed with a syringe following which the usual washings and slow exposure to air were carried out.

Table 2 gives a summary of each of the series described above.
Table 2

Different method of synthesis followed for PtPb nanoparticles and associated observations

Series

Method

Domain size variance with reaction temperature

TGA mass loss observed (around 150–200 °C)

Series 1

Precursors injected into a reducing agent solution stirring at the desired temperature. Solvent evaporated off post reaction

Domain size falls on increasing/decreasing reaction temperature from 120 °C

Lowest (120 °C)—10.03%

Highest (155 °C)—19.91%

Series 2

Precursors injected into a reducing agent solution stirring at room temperature, solution then heated up to desired temperature. Solvent evaporated off post reaction

Domain size increases on increasing reaction temperature

Lowest (155 °C)—5.75%

Highest (80 °C)—13.52%

Series 3

Precursors injected into a reducing agent solution stirring at room temperature, solution then heated up to desired temperature. Solution centrifuged post reaction, supernatant solvent syringed off

Domain size remains constant with respect to change in reaction temperature. PtPb domain size measured to be 18 nm even on doing a room temperature reaction

Lowest (155 °C)—8.64%

Highest (r.t.)—9.18%

Characterization

X-ray powder diffraction powder patterns (Scintag XDS 2000) were taken of the black PtPb nanoparticles to confirm the composition and structure of the intermetallic phase. The particle morphology and size were studied by scanning electron microscopy (SEM) using a LEO-1550 Field emission SEM (FSEM). Energy dispersive X-ray analysis (EDX) was also done on the particles on the same device. For EDX analysis, unless otherwise specified, particles in dry powder form were used on GaAs wafers. A JEOL 8900 EPMA Microprobe was used for preliminary EDX on most of the PtPb samples. Scanning TEM images and convergent beam electron diffraction (CBED) data were collected on a VG HB501UX UHV STEM. Dry particles, in powdered form dispersed on ultra thin TEM grid, were used for STEM and SEM images. The TEM images and selected area electron diffraction (SAED) were taken on a JEOL 1200EX TEM on microtomed (60 nm) sections of the nanoparticles dispersed on an ultrathin TEM grid. To measure the surface area of the samples, a Micromeritics ASAP 2020 was used to collect a partial adsorption isotherm at liquid nitrogen temperature (−196 °C) with krypton as the adsorption gas over the pressure ratio (P/P0) range of 0.06–0.5. Prior to measurements, the sample was degassed under vacuum at room temperature for 48 h. The specific surface area was determined according to the Brunauer–Emmett–Teller (BET) method in the relative pressure range of 0.08–0.185.

Electrocatalytic activity

The electrocatalytic activity of the PtPb nanoparticles toward the oxidation of formic acid was examined. Before each experiment, a suspension of the nanoparticle catalyst was prepared as follows: to 4 mg of the dried nanoparticle sample were added 3.98 mL of distilled water and 1 mL of isopropyl alcohol (Aldrich). In addition, 20 μL of a 5% w/w Nafion solution in alcohols (Aldrich, EW: 1100) and water were added to this mixture. The resulting mixture was sonicated in a bath type ultrasonicator, for 15 min. Each nanoparticle suspension described above was coated onto a 3-mm diameter glassy carbon (GC) electrode. The electrode had been previously polished with diamond paste (METADI-Buehler, ø = 1 μm) and ultrasonicated in Millipore water (18 MΩ cm−1, Millipore Milli-Q) for 10 min. The electrode was then rinsed with Millipore water and allowed to dry in air. Coating with 70 μg cm−2 of the nanoparticles (6.1 μL of nanoparticle suspension) was performed onto the clean glassy carbon electrode. The electrode was then spin dried at 600 rpm under a nitrogen gas atmosphere.

Before studying fuel oxidation, electrochemical pretreatment of nanoparticle-coated electrodes is sometimes necessary to obtain optimal activity (Murugan and Yasonath 2005; Prabhuram et al. 2006). A partially optimized procedure involved generating hydrogen by holding the nanoparticle ink-coated electrode at −0.7 V versus Ag/AgCl for 30 min in 0.1 M H2SO4 to activate the catalyst surface presumably by removing surface oxides and other contaminants ostensibly originating from the synthesis.

Formic acid oxidation on the nanoparticle-coated GC electrode was examined in a solution of 0.5 M formic acid (Mallinckrodt, 88% analytical reagent)/0.1 M sulfuric acid at a sweep rate of 10 mV s−1. All solutions were prepared with Millipore water and deaerated with prepurified nitrogen for at least 10 min before each experiment. The measurements were conducted at room temperature. All potentials are referenced to a sodium-chloride saturated Ag/AgCl electrode without regard for the liquid junction.

Results and discussion

Investigating platinum precursors

The platinumdichloride(1,5-cyclooctadiene) derivatives synthesized were compared with the commercially obtained platinum acetylacetonate with respect to solubility in diglyme and THF (Table 1). The solubility of (3,4-dimethyl-1,5-cycloctadiene)platinumdichloride was lower than (3-methyl-1,5-cycloctadiene)platinumdichloride in both solvents, showing the expected lowering of solubility with an increase in symmetry of the ligand. Platinum acetylacetonate showed the highest solubility in both solvents, and was hence chosen as the precursor to be studied further. Platinum acetate was found to be unstable over extended periods of time or at higher temperatures and hence was excluded from consideration after a few initial experiments. After investigating platinum acetylacetonate, platinum hexafluoroacetylacetonate was explored in the hopes that the organic ligand from the precursor would be less likely to adsorb to the nanoparticles, because the precursor itself was a more polar molecule. The new precursor was also expected to be more soluble than platinum acetylacetonate in the solvents compatible with sodium naphthalide, THF, and diglyme. When the solubility of Pt hexafluoroacetylacetonate was tested, it was found to be more soluble than platinum acetylacetonate in these solvents. However, upon using this new precursor to synthesize PtPb nanoparticles, the resulting particles were hygroscopic and showed high amounts of sodium and fluorine (EDS) as well as characteristic fluoroacetylacetonate peaks (FTIR). This apparent sodium and fluorine complex was removed by subjecting the particles to water/double MeOH wash. At this point, the particles were no longer hygroscopic and did not contain the characteristic peaks of fluoroacetylacetonate in the FTIR. However, the particles were highly polar to begin with (probably owing to Na/F containing polar/ionic species sticking to them), making their separation from the polar wash solvents very difficult and leading to extremely low yields. We believe that the lone, now highly acidic hydrogen, left on the hexafluoroacetylacetonate moiety is responsible for the hygroscopic property of the resulting nanoparticles. A per-fluorinated acetylacetonate may behave differently. However, this was not further pursued because the nanoparticles were not found to have any enhanced properties than the ones made from platinum acetylacetonate and our goal is to find a precursor convenient and simple to work with.

Synthesis of PtPb nanoparticles

Previous PtPb nanoparticles synthesized using dimethyl(1,5-cyclooctadiene)platinum and lead(II)2-ethylhexanoate showed higher electrocatalytic activity and larger domain size of PtPb when they were synthesized by injecting the precursors into a sodium naphthalide solution at a higher temperature (135 °C) rather than room temperature injection followed by annealing at 135 °C in the solvent (Alden et al. 2006a). Accordingly, an initial series of reactions was done by injecting the precursors into a reducing agent solution stirring at the desired temperature. A stirring time of 30 min was maintained in each case to ensure good crystallization based on previous trends observed in the group (Alden et al. 2006a), following which, the solvent was evaporated off (series 1). We expected the temperature of crystallization to be lower when Pt(acac)2 was used, because acetylacetonate is anionic and the Pt–O bond is weaker than Pt-ene bonds. Surprisingly, the domain size variation with reaction temperature was not similar to that observed when dimethyl(1,5-cyclooctadiene)platinum was used as the Pt precursor. Nor was lowering of crystallization temperature observed. Reactions carried out at temperatures above and below 120 °C resulted in pXRD patterns with low intensity and broad peaks, approximately matching the more intense peak positions of the PtPb structure. However, given the broad nature of the peaks, the presence/absence of other phases could not be established. Annealing the product, after sealing it in an evacuated quartz tube and heating at 600 °C for 12 h, showed PtPb as the only crystalline phase by pXRD with a mean crystal domain size of 30 nm. The SEM images of the as-prepared samples showed a high degree of agglomeration. EDX analysis of the particles showed the presence of Pt and Pb in 1:1 ratio (±5%) in multiple micron square regions along with the presence of some sodium (Fig. 1). For the reaction carried out at 120 °C, XRD patterns showed peaks which could be clearly indexed only to PtPb. The mean domain size calculated from the peak widths using the Scherrer equation (Warren 1990) was 8.4 nm. A domain size versus reaction temperature curve showed a sharp maximum around 120 °C (Fig. 2). EDX data again showed the presence of Pt and Pb in 1:1 ratio (±5%) in multiple micron square areas along with the presence of some sodium. However, the amount of sodium present was negligible in this case. All EDX studies were done on dry samples on GaAs wafers to allow rough quantification of the amount of carbon present in the samples. Samples prepared at 120 °C showed the lowest amount of carbon. A point to keep in mind here is the much higher molecular weights of the metals compared to carbon. Even a low weight% of carbon would indicate a considerable atomic ratio of carbon to the metals. Although absolute quantification of carbon using EDX is difficult owing to background signals, we used the carbon signal from nearby areas of GaAs as a background correction. The background signal on GaAs was a small fraction of the total carbon signal (less than 0.1 wt%). In order to identify the source of the carbon content, ATR measurements were obtained from the solid powder samples. The acetylacetonate ligand on the PtPb surface could be clearly identified in the resulting spectra from asymmetric and symmetric νcoo– stretching peaks at the 1,600–1,400 cm−1 region in all samples except in the one prepared at 120 °C (Fig. 3). Aliphatic –C–H stretches were also observed (~2,900 cm−1), but no indication of ethylhexanoate (1,735 cm−1 ester peak) from the lead precursor or sodium naphthalide (aromatic –C–H stretching bands) was observed. We assume, therefore, that aromatic groups are present only in insignificant amounts, if at all. The SEM images of the as-prepared samples (dry powder on GaAs wafer), prepared at 120 °C gave clearly observable particles with sizes agreeing well with the calculated domain sizes from XRD. CBED patterns of this region confirm that the particles retain their crystallinity at the nano level (Fig. 4a, b). The particles, as expected, were seen to be highly agglomerated (since surfactants were not used during the syntheses in order to preserve the catalytic surface of the nanoparticles). However, the point to remember is that the particles remained suspended in solution until centrifuged/ solvent evaporated off. So the agglomeration is not high enough to cause the particles to precipitate out of solution. Images of samples prepared at other temperatures also appeared as agglomerated networks, but smaller nanometer shapes were not identifiable under imaging conditions (Fig. 5). As mentioned previously, all samples were found to contain sodium by EDX (evidence of sodium was also found on the walls of the tubes used for annealing). EDX confirms that the sodium could be completely removed by water wash. However, the IR spectra of the washed samples did not seem to be much different (Fig. 6), suggesting the presence of organics on the nanoparticles even after complete removal of the sodium. TGA of the samples done under nitrogen, showed up to 20% mass loss for the samples, with samples starting to lose mass at as low as 150 °C, again suggesting presence of organic species. The mass loss was the smallest for nanoparticles prepared at 120 °C, thus confirming the IR and EDX results. Thus, lower organic presence in the product is correlated with larger domain-sized particles (Fig. 7).
https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9557-y/MediaObjects/11051_2008_9557_Fig1_HTML.gif
Fig. 1

EDX data of PtPb nanoparticles synthesized by injecting the precursors into reducing agent solution at 135 °C, followed by 30 min stirring at 135 °C

https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9557-y/MediaObjects/11051_2008_9557_Fig2_HTML.gif
Fig. 2

Domain size of PtPb nanocrystals synthesized by injecting the precursors into reducing agent solutions at different temperatures followed by stirring for 30 min at those temperatures

https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9557-y/MediaObjects/11051_2008_9557_Fig3_HTML.gif
Fig. 3

FTIR spectra of PtPb nanoparticle powders. Asymmetric and symmetric νcoo– stretching peaks at the 1600–1400 cm−1 region in all samples prepared from platinum acetylacetonate except for the one prepared at 120 °C. PtPb prepared from dimethyl(1,5-cyclooctadiene)platinum does not show the νcoo– stretching peaks

https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9557-y/MediaObjects/11051_2008_9557_Fig4_HTML.jpg
Fig. 4

a SEM image of PtPb nanoparticles prepared from platinum acetylacetonate by injecting the precursors into reducing agent solution at 120 °C. b CBED pattern of a PtPb nanocrystal from the same sample

https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9557-y/MediaObjects/11051_2008_9557_Fig5_HTML.jpg
Fig. 5

SEM image of PtPb nanoparticles synthesized from platinum acetylacetonate by injecting the precursors into the reducing agent solution at 135 °C

https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9557-y/MediaObjects/11051_2008_9557_Fig6_HTML.gif
Fig. 6

FTIR spectra of PtPb nanoparticle powders before and after water wash (particles were washed with water to remove the sodium initially detected on the particles by EDX). Asymmetric and symmetric νcoo– stretching peaks at the 1600–1400 cm−1 region still observed, so water wash removes the sodium completely, but removal of sodium did not result in removal of the organics (at least not completely). The ‘quick exp to air’ refer to the samples exposed to air quickly (instead of slowly as described in the experimental section) post reaction and wash with solvents. It however, seemed to have made little or no difference in terms of the organic material detected by FTIR on the nanoparticles

https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9557-y/MediaObjects/11051_2008_9557_Fig7_HTML.gif
Fig. 7

Thermogravimetric analysis plot of the PtPb nanoparticle samples prepared by injection of precursors into reducing agent solution at different temperatures showing mass loss (in %) of the samples with increasing temperature

Subsequent to the unexpected results observed when the precursors were injected into the reducing solution stirred at temperatures up to 150 °C, another series of reactions was done by injecting the precursors into a reducing agent solution stirred at room temperature, followed by heating up to the desired temperature. This was done to check our hypotheses about the precursor itself undergoing changes when injected into a hot reducing agent solution (temperatures above 120 °C) in presence of the Pb precursor. If so, this might lead to new species being formed which were interfering with nanoparticle crystallization at higher temperatures.

Typically, the solution would take 10–15 min to heat up to the desired temperature, after which it would be stirred for 10 min at that temperature and the reaction solvent would be evaporated off (series 2). The domain size variation with reaction temperature, in this case, looked much more familiar to previous observations (Fig. 8). EDS data showed Pt and Pb to be present in a 1:1 ratio down to the nanoscale (20 nm × 20 nm regions) for all the samples. The amount of carbon and sodium observed on EDX remained almost constant irrespective of the domain size, but were much lower than the previous series (except for the 120 °C reactions, which were comparable). The IR spectra of the samples again showed the presence of the acetylacetonate ligand on the nanoparticles. SEM images showed observable morphological changes suggesting gradual formation of the ‘nanoparticles’ with increasing reaction temperature (Fig. 9a–c). Interestingly, the smallest observable particles corresponded to the respective domain sizes calculated from Scherrer equation for all the samples. CBED patterns confirmed that the crystalline nature of the particles was retained at the nanoscale.
https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9557-y/MediaObjects/11051_2008_9557_Fig8_HTML.gif
Fig. 8

Domain size of PtPb nanocrystals synthesized by injecting the precursors into reducing agent solution at room temperature followed by heating up to different temperatures. Inset shows the similar domain size versus temperature trend observed with PtPb nanoparticles prepared from dimethyl(1,5-cyclooctadiene)platinum(Alden et al. 2006a)

https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9557-y/MediaObjects/11051_2008_9557_Fig9_HTML.jpg
Fig. 9

SEM images of PtPb nanoparticles prepared by injecting precursors into reducing agent solution at room temperature followed by heating up to different temperatures a 120 °C, b 135 °C, and c 155 °C. The gradual formation of nanoparticles on going to higher temperatures can be observed in the images

The long time needed for solvent removal (diglyme took around 4 h to evaporate off at 40 deg, ~150 m Torr) caused us to seek an alternative method of solvent removal. Interestingly, the properties of the resulting nanoparticles were significantly altered by changing this procedure. In this method, as described in the experimental section, post injection of precursors into the reducing agent solution (injection done at room temperature followed by heating up, if necessary) and 15 min of stirring (typically), the mixture would be cannula transferred into a centrifuge tube, centrifuged, and the supernatant would be syringed off. For this method, the mean domain size of the PtPb nanoparticles obtained by a room temperature reaction was 14 nm as determined by the Scherrer equation. The domain size variation with the temperature to which the solution was heated up to post injection, was minimal (Fig. 10). Also, SEM images confirmed the domain sizes calculated from X-ray diffraction. Interestingly, this confirms that the organics present on these nanoparticles are on the surface and not trapped inside. Interesting changes were again observed in the nanoparticle shapes and agglomeration with increases in the reaction temperature (Fig. 11a–d; Fig. 12). A plausible explanation for crystallization at room temperature and the overall ease of crystal formation that was observed with this particular work-up procedure could be the removal of the reaction solvent and organic products. Solvent removal by evaporation concentrates the organic product. This can be understood in terms of a ligand equilibrium between the reaction solvent and the solids in the system (L(solvent)↔L(solid)), which shifts toward more L(solid) when only the solvent is removed (by evaporation). Indeed the IR signal from acetylacetonate is lowered following a THF wash. TEM studies done on microtomed sections (uniformly 60 nm in thickness) of the nanoparticles from the 80 °C reaction, showed darker and lighter regions. SAED confirmed that all regions studied were crystalline (Fig. 13a, b). For both series 2 and 3, TGA showed much lower mass losses (averaging around 9%) than what was observed for series 1. This agrees well with the corresponding EDS studies. Also, for series 1 and 2 the TGA mass losses corresponded to the domain size variation with temperature (highest mass losses were observed for temperatures at which particle domain size was found to be the lowest and vice versa). For series 3, TGA mass loss is almost constant with respect to reaction temperature (as is the domain size variation) (Table 2). These observations fit in well with the model of surface-bound organics interfering with nanoparticle crystallization.
https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9557-y/MediaObjects/11051_2008_9557_Fig10_HTML.gif
Fig. 10

Domain size of PtPb nanocrystals synthesized by injecting the precursors into reducing agent solution at room temperature followed by heating up to different temperatures; solvent syringed off post reaction after precipitating out the nanoparticles. Domain size observed to be largely invariant of temperature, with clearly identifiable PtPb crystals (with mean domain size of 18 nm) was obtained for even the room temperature reaction)

https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9557-y/MediaObjects/11051_2008_9557_Fig11_HTML.jpg
Fig. 11

SEM images of PtPb nanoparticles prepared by injecting precursors into reducing agent solution at room temperature followed by heating up to different temperatures a room temperature, b 80 °C, c 130 °C, and d 155 °C; solvent syringed off post reaction after precipitating out the nanoparticles. The gradual formation of nanoparticles on going towards higher temperature can again be very clearly observed. The observed morphological changes might be arising from compositional changes. If so, the temperature dependent property changes observed in all these materials can be explained

https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9557-y/MediaObjects/11051_2008_9557_Fig12_HTML.jpg
Fig. 12

(enlarged) The needle like artifacts were found to disappear on washing the sample with water suggesting presence of Na+ and other ionic species (organic) which can be removed by water wash along with the PtPb nanoparticles. The exact nature of the interfering species are still unknown to us. The artifacts disappear on going to higher reaction temperatures (Fig. 9c; Fig. 11c and d)

https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9557-y/MediaObjects/11051_2008_9557_Fig13_HTML.jpg
Fig. 13

a SAED pattern of PtPb nanocrystal prepared by injecting precursors into reducing agent solution at room temperature followed by heating up to 130 °C; solvent syringed off post reaction after precipitating out the nanoparticles. b Microtomed section (of uniform 1 μm thickness) of the same sample

Electrocatalytic activity

Electrocatalytic activity of the nanoparticles prepared by injecting precursors into the reducing agent solution at different temperatures and worked up by evaporating off the reaction solvent (series 1) toward the oxidation of formic acid was examined. The onset potential for PtPb nanoparticles is the same as that of all PtPb nanoparticles examined previously in the group (i.e. −0.2 V for PtPb nanoparticles prepared by both sodium borohydride and sodium naphthalide reduction). Hence, to compare the electrocatalytic activity of the particles, the magnitude of current density at a potential positive of onset (+0.2 V) was compared (Table 3). Higher mass activities at a given potential indicate a better catalyst. However, the values compared here are in units of mA/μg and need to be normalized with respect to the BET surface area (which would be different for each sample owing to different particle sizes and degrees of agglomeration) for an absolute comparison. The highest mass activity is observed for the nanoparticles prepared at 120 °C. This mass activity is much higher (>200%) than what was previously obtained from PtPb nanoparticles made by sodium naphthalide reduction (using dimethyl(1,5-cyclooctadiene)platinum precursor) but is lower (<200%) than what was obtained from PtPb nanoparticles made by sodium borohydride reduction in the group.
Table 3

Mass activities of PtPb samples prepared by injection of precursors into reducing agent solution at different temperatures

Sample injection at reaction temperature

Mean domain size (nm)

Mass activity at 0.2 V/mA μg−1

PtPb from Pt(acac)2, 80 °C

4

0.036

PtPb from Pt(acac)2, 135 °C

4

0.092

PtPb from Pt(acac)2, 120 °C

8.4

0.153

PtPb from Pt(acac)2, 155 °C

4

0.030

The electrocatalytic activity of the nanoparticles prepared by heating up the solution to different temperatures post injection of precursors into a room temperature reducing agent solution and worked up by evaporating off the solvent (series 2) toward the oxidation of formic acid was compared by comparing the magnitude of current density at +0.2 V (Table 4). Again, the highest mass activity was observed for the nanoparticles prepared at 120 °C. Also, the mass activity obtained for this sample was much higher than the PtPb nanoparticles made by high temperature injection of precursors into the sodium naphthalide solution and is comparable to current densities obtained from PtPb nanoparticles made by sodium borohydride reduction. Again, for a more meaningful comparison, currents were normalized to the BET surface area of the respective catalysts. The current density obtained for the present PtPb product is the highest obtained to date. In terms of the highest obtainable electrocatalytic activity, room temperature injection into a solution of the reducing agent solution followed by heating works better for platinum acetylacetonate than for platinum dimethyl cyclooctadiene.
Table 4

Mass activities of PtPb samples prepared by injection of precursors into reducing agent solution at room temperature followed by heating up to different temperatures

Sample room temperature injection, followed by heating up

Mean domain size (nm)

Mass activity at 0.2 V/mA μg−1

PtPb from Pt(acac)2 ,80 °C

5

0.051

PtPb from Pt(acac)2, 135 °C

13.7

0.34

PtPb from Pt(acac)2, 120 °C

6

0.52

PtPb from Pt(acac)2, 155 °C

20

0.21

Similarly, the electrocatalytic activity of the nanoparticles prepared by heating the solution to different temperatures post injection, and worked up by centrifuging and syringing off the reaction solvent (series 3) toward the oxidation of formic acid was assessed by comparing the mass activity at a potential positive of onset(−0.2 V). The highest mass activity was observed for the nanoparticles prepared at 135 °C (Table 5). Also, the highest mass activity obtained in this case is much lower than the highest obtained in the case of room temperature injection of precursors following drying of the reaction mixture under vacuum, but is comparable to the highest obtained by high temperature injection of precursor method (Table 6). Figure 14a, b shows the rotating disk voltammograms of the PtPb nanoparticles synthesized by different methods as well as for Pd black and PtRu alloy standards. The variation in current densities of PtPb particles prepared differently may be due to different crystal planes exposed in different samples, different degree of agglomeration or different surface composition of the nanoparticles. To eliminate the possibility of the oxidation current to be coming from residual organics on the nanoparticles, cyclic voltammogram of PtPb-coated GC electrode in 0.1 M H2SO4 was obtained. The difference between voltammograms obtained in formic acid solution and in H2SO4 clearly suggests that the oxidation current is for formic acid oxidation (Fig. 15). In addition, the cyclic voltammograms obtained in 0.1 M H2SO4 did not show any oxidation peak for Pb indicating that Pb atoms do not dissolve electrochemically between −0.2 and +0.2 V (Matsumoto et al. 2008).
Table 5

Mass activities of PtPb samples prepared by injection of precursors into reducing agent solution at room temperature followed by heating up to different temperatures (solvent syringed off post reaction)

Sample room temperature injection, solvent syringed off

Domain size in nm

MA/mA μg−1 at 0.2 V

PtPb from Pt(acac)2, 130 °C

16

0.179

PtPb from Pt(acac)2, 80 °C

14

0.016

PtPb from Pt(acac)2, 155 °C

19

0.064

Table 6

Comparison of electrochemical data of PtPb samples prepared in group till date by various methods

 

Domain size in nm

MA/mA μg1 at 0.2 V

MA/mA cm2 at 0.2 V

PtPb from Pt(acac)2, 120 °C, series 1

9.1

0.153

2.43

PtPb from Pt(acac)2, 120 °C, series 2

6

0.52

5.95

PtPb from Pt(acac)2, 130 °C, series 3

16

0.179

1.73

PtPb (produced with H2PtCl6 and Pb(MOEEAA)3, NaBH4 reduction*

10.6

0.132

5.5

PtPb (produced with H2PtCl6 and Pb(C2H3O2)2), NaBH4 reduction***

12

0.441

2.94

PtPb (produced with PtMe2https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9557-y/MediaObjects/11051_2008_9557_Fige_HTML.gif and Pb(C2H3O2)2, 135 °C, 10 min), NaNp reduction**

24

0.045

1.98

*, **, ***: these results were referred to reference Liu et al. (2006), Matsumoto et al. (2008), Mortimer (1984), respectively

https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9557-y/MediaObjects/11051_2008_9557_Fig14_HTML.gif
Fig. 14

a, b Rotating disk voltammograms for formic acid oxidation on Pd Black(1), PtRu/C (2), 120 °C reaction, series 1 (3), 120 °C reaction, series 2 (4), 135 °C reaction series 3 (5), PtPb prepared from PtPb (produced with PtMe2https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9557-y/MediaObjects/11051_2008_9557_Figf_HTML.gif and Pb(C2H3O2)2, sodium naphthalide reduction (6), PtPb (produced with H2PtCl6 and Pb(MOEEAA)3, NaBH4 reduction (7), PtPb (produced with H2PtCl6 and Pb(C2H3O2)2), NaBH4 reduction (8)

https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9557-y/MediaObjects/11051_2008_9557_Fig15_HTML.gif
Fig. 15

Cyclic voltammograms for the oxidation of FA on PtPb-ordered intermetallic nanoparticles-coated loading of 70 g cm−2 on a rotating GC electrode in 0.5 M FA and 0.1 M H2SO4 aqueous solution at 10 mV s−1 and 2000 rpm. Curves a, b, and c represent the first, second, and third scans, respectively. Voltammogram d was obtained in 0.1 M H2SO4 solution without FA under the same conditions as voltammograms a–c (Matsumoto et al. 2008)

Conclusions

In conclusion, we have found a more convenient precursor (platinum acetylacetonate) for PtPb intermetallic syntheses using sodium naphthalide, or similar reducing agents that require inert atmosphere techniques and etheral solvents. PtPb synthesized with this precursor has a higher specific activity for formic acid oxidation when compared with those prepared previously. Agglomeration, as expected, was observed in all cases in the as-prepared samples owing to exclusion of surfactants from the syntheses. Also, we have shown adsorbed organic fragments from the precursors to be one of the factors interfering with nanoparticle crystallization, and have been able to establish a method to minimize their effects. Interesting trends in particle morphology and electrochemical activity were observed but need further investigation. Research is ongoing in the group for further identification and characterization of the organic fragments that bind to the particle surfaces (MS, TGA, and FTIR studies along with nanoscale EDS). We are also extending our work with this precursor towards syntheses of other platinum intermetallics starting with the Pt–Ti system. What is novel to this study is the identification of the importance of ligands and anions from the precursors in the crystallization of the intermetallic nanoparticles. We show that the previous puzzle of the lack of in situ crystallization of nanoparticles at room temperature can be solved and domain size control can be achieved by appropriate techniques to remove surface interacting organics. This opens up a generalized method for syntheses of intermetallic nanoparticles and increases the ease of commercial scaling up of the same, given the need to do so.

Acknowledgments

This study was supported by the Basic Energy Sciences Division of the Department of Energy through Grants: DE-FG02-03ER46072 and DE-FG02-87ER4529. The authors thank Mick Thomas for help with the UHV-STEM data, John Hunt for the EPMA microprobe and John Grazul for the TEM data for which we made use of the UHV- STEM, TEM, and EPMA microprobe, laboratories of Cornell Center for Materials Research (CCMR). The BET surface area for the PtPb samples was measured in Prof. Ulrich Wiesner’s laboratory in the department of Materials Science and Engineering, at Cornell University. The BET surface area of one of the PtPb compounds (series 1, 120 °C reaction) was measured at Primet Precision Materials Inc. in Ithaca, NY. The authors would also like to thank Thomas McCarrick and Prof. J. Meinwald for helpful discussions on IR spectra.

Copyright information

© Springer Science+Business Media B.V. 2009