Letter pubs.acs.org/Langmuir
Size and Composition Control of Pt−In Nanoparticles Prepared by Seed-Mediated Growth Using Bimetallic Seeds Ferenc Somodi, Sebastian Werner, Zhenmeng Peng, Andrew Bean Getsoian, Anton N. Mlinar, Boon Siang Yeo, and Alexis T. Bell* Department of Chemical and Biomolecular Engineering, 201 Gilman Hall, University of California, Berkeley, California 94720-1462, United States S Supporting Information *
ABSTRACT: A two-step method has been developed for precise size and composition control of bimetallic Pt−In nanoparticles. Very small (1.62 nm) PtIn seed nanoparticles with 1:1 metal ratio were prepared in the absence of capping agents followed by growth of Pt on their surface in the presence of oleyl amine as reducing and stabilizing agent. Nanoparticles with bulk compositions of Pt4In, Pt3In, and Pt2In could be synthesized with average diameter smaller than 3 nm. TEM, EDX, and XPS provided evidence for homogeneous growth without separate nucleation of pure platinum nanoparticles in the reaction solution. Pt3In nanoparticles were deposited onto SiO2 surface by incipient wetness impregnation. Temperature-induced changes in the particle surface were monitored by in situ IR spectroscopy and CO adsorption. It was found that surface alloy composition of the particles could be tuned by using oxidizing or reducing atmospheres.
■
INTRODUCTION Indium is an important promoter for supported Pt catalysts. Pt−In bimetallic catalysts exhibit high activity for the denitrification of water,1 dehydrogenation of isobutane2 and light alkanes.3 The presence of indium in Pt or in bimetallic PtRe catalysts has been found to increase the selectivity toward aromatic hydrocarbons in naphtha reforming,4,5 and Incontaining PtSn/C catalysts have been shown to have high activity for direct oxidation of ethanol in a fuel cell.6 In spite of the catalytic importance of bimetallic and trimetallic alloys of indium, the precise control of size and composition of the catalytically active nanoparticles is not well established. The reason for this situation is the absence of a suitable preparation method for achieving complete size and composition control with the high temperature one-pot synthesis routes, generally used for the preparation of bimetallic nanoparticles. For example, it has been found that the broad compositional distribution of PtFe nanoparticles was sharpened significantly when trioctylphosphine instead of oleic acid was used in the iron-pentacarbonyl preparation method.7 In the case of CdSe quantum dots, elemental analysis of the reaction mixture after separation of nanoparticles revealed that a considerable fraction of the Cd was unreacted in the presence of an excess of oleic acid.8 In a recent study, we have shown that the size and composition of Pt−Sn nanoparticles could not be controlled at the same time by changing the concentration of capping agents (oleyl amine and oleic acid).9 These observations suggest that the nature of metal ion-capping agent interaction and its role in the particle formation is not well understood, leading to difficulties in the simultaneous control of the composition and size of bimetallic particles. We described here a novel, two-step method for preparing Pt−In bimetallic nanoparticles with controlled size and © 2012 American Chemical Society
composition. First, ultrasmall PtIn (Pt/In = 1:1) nanoparticles (average size: 1.62 nm) are prepared at high temperature in the absence of capping agents. Pt is then grown on the surface of these particles by reduction of a Pt precursor in the presence of oleyl amine under mild conditions.
■
EXPERIMENTAL SECTION
PtIn seed nanoparticles were prepared according to the following procedure. The solvent, 20 mL of dioctyl ether, was loaded into a 50 mL three-necked flask, and then 0.574 g of 1,2-hexadecanediol, 0.0973 g of platinum(II) acetylacetonate (with 97% purity) (Pt(acac)2), and 0.0989 g of In(III) acetylacetonate (In(acac)3) were added. The flask was connected to a Schlenk line via a water-cooled condenser. The mixture was then heated to 290 at 5 °C/min under N2 atmosphere and refluxed for 60 min. The synthesis was terminated by removing the heating mantle and cooling the colloidal suspension to room temperature under a nitrogen atmosphere. The sample was stored without further purification. To prepare particles with different metal ratio (Pt2In Pt3In, Pt4In), Pt(acac)2 was introduced into a three-necked flask containing 15 mL of dioctyl ether and 5 mL of the PtIn seed colloid was added to this mixture at room temperature under flowing N2 atmosphere and heating started. After complete dissolution of Pt(acac)2, a 6-fold excess of oleyl amine (OAm) relative to the overall amount of platinum was injected into the flask at 130 °C, and the mixture was stirred at 160 °C for 60 min. The list of prepared samples can be found in Table 1. One Pt3In colloid (Table 1) was selected for deposition onto SiO2 (SiliCycle) by incipient wetness impregnation to prepare samples containing 1 wt % Pt. Then 2.9 mL of the colloid (Pt concentration: 0.009 mol/L) was centrifuged with 3-fold excess amount of ethanol, and the precipitate was dissolved in n-hexane and contacted with 0.5 g Received: December 8, 2011 Revised: January 20, 2012 Published: February 2, 2012 3345
dx.doi.org/10.1021/la204838q | Langmuir 2012, 28, 3345−3349
Langmuir
Letter
Table 1. List of Nanoparticles Prepared Using Bimetallic Seeds PtIn Pt2In Pt3In Pt4In
nPt2+ [mmol]a
nPts [mmol]b
nIns [mmol]c
Pt/Ind
OAm [mmol]e
theor. size [nm]f
size (TEM) [nm]
Pt/In (EDX)g
0.06 0.12 0.18
0.06 0.06 0.06 0.06
0.06 0.06 0.06 0.06
1 2 3 4
∼0.72 ∼1.08 ∼1.44
1.85 2.04 2.20
1.62 1.87 2.15 2.44
1.00 1.95 3.13 3.94
a
Amount of Pt(acac)2 dissolved in 15 mL dioctyl ether. bAmount of Pt in the 5 mL seed solution. cAmount of In in the 5 mL seed solution. dTotal molar ratio of Pt and In: (nPt2+ + nPts)/nIns. eAmount of oleyl amine added during the synthesis. fTheoretical average particle size (2Rd) in the case of complete Pt deposition on the surface of seeds with 1.62 nm average size (2Rs) calculated based in Figure 1. gMolar ratio determined by EDX analysis. of SiO2. The material was dried at room temperature overnight before the IR spectroscopic investigation. A detailed description of sample preparation for TEM, XPS, and FT-IR measurements can be found in the Supporting Information.
calculating the anticipated final size of the particles is shown in Figure 1. The results obtained from both techniques were
■
RESULTS AND DISCUSSION The initial objective of our work was to prepare PtIn (Pt/In = 1:1) particles in the complete absence of capping agents. Surprisingly, a stable colloid formed with an average size of 1.5 nm and a narrow distribution of particle sizes. It was hypothesized that the adsorption of diol or its oxidation product on the nanoparticle surface was responsible for the formation of small particles. To confirm this idea, samples were prepared with decreasing 1,2-hexadecanediol concentration. It was found that the diol had no significant influence on the average particle size and stability of the colloid (see the Supporting Information). Experiments with a higher Pt/In ratio (Pt/In = 3) were also carried out, but in this case the colloid exhibited a broad size distribution and the deposition of a metal film on the flask wall after the synthesis. These results suggest that alloy nanoparticles with Pt/In = 3 molar ratio cannot be prepared by this one-pot synthesis route. It is very likely that Pt2+ ions reduce first forming nuclei, which are then surrounded by unreacted In3+ ions or their complexes forming an electrostatic or steric stabilizing layer. Similar, condensed organic layers have been proposed previously by others.10 The reduction of indium cations at high temperature and the formation of PtIn alloy nanoparticles is highly favored because the enthalpy for formation of this composition is a minimum for Pt−In alloys.11 It is also plausible that the surface energy of this composition is the lowest. These considerations might explain the outstanding stability and sharp particle size distribution of this sample. The PtIn nanoparticles prepared without capping agents and a 1:1 stoichiometric ratio of the metals should be suitable seeds for further deposition of Pt in order to produce PtIn nanoparticles with controlled size and composition. It has been reported that excellent size control could be achieved by seed-mediated growth of monometallic Pd particles.12 The same concept should be applicable for the deposition of Pt on PtIn seed particles, assuming complete reduction and homogeneous growth of Pt without separate nucleation of pure platinum particles in the solution. All samples were prepared using a low concentration of oleyl amine in order to avoid the nucleation of Pt particles, but sufficiently high to enable complete reduction Pt2+ cations present in solution. The samples listed in Table 1 were all prepared with a 6-fold excess of oleyl amine relative to Pt2+, since it was found that this was sufficient to reduce all of the Pt2+ cations on the seed particles as confirmed by TEM measurements of particle size and EDX. The method for
Figure 1. Particle formation model: the volume ratio of the deposited Pt shell (Vd − Vs) and the Pt in the bimetallic seeds (∼0.5Vs) is equal to the molar ratio of the Pt2+ in solution and Pt metal in the seeds. Knowing the seed radius (Rs) and the metal ratio ([Pt2+]/[Pts], the radius of the desired particle (Rd) can be calculated.
found to be in good agreement with what was expected (see Table 1). The TEM pictures and particle size distribution of the samples are presented in Figure 2. The broad size distribution of the PtIn seed sample is a consequence of chain-like association of a portion of the particles. Evidence for such a process can be seen in Figure 1. The association of nanoparticles is reversible since it is not observed in the other samples prepared using the same seeds. The similarity of the standard deviation in particle size observed for the other samples (Pt2In, Pt3In, Pt4In) confirms the absence of platinum nucleation in solution. The observed XPS binding energies for Pt and In are reported in Table 2 for the as-prepared samples. Also shown is the Pt/In ratio determined from the intensities of the XPS signals of the two elements. Consistent with the results of EDX analysis, the Pt/In ratio increases with increasing concentration of Pt in the samples. However, the Pt/In ratio determined from XPS is consistently lower than that determined by EDX. This trend is in accord with that deduced by consideration of the sampling depths for electrons ejected from Pt and In. While the X-ray beam can penetrate the whole particle, the spectrometer only detects the electrons that can escape from the sample. The inelastic mean free path (IMFP) of the photoelectrons (Pt 4f, In 3d) depends on the electron kinetic energy, which can be estimated from calculations for the electronic levels of each metal.13 According to these calculations (see the Supporting Information), the IMFP for the In 3d5/2 electrons is higher than that for the Pt 4f7/2 electrons (2.6 vs 1.68 nm). This means that the In 3d electrons comes from deeper sections of the sample, which explains the lower Pt/In ratio determined from EDX 3346
dx.doi.org/10.1021/la204838q | Langmuir 2012, 28, 3345−3349
Langmuir
Letter
Figure 2. TEM images and corresponding particle size distributions. Scale bars represent 20 nm.
electrons would be to use different X-ray wavelengths for each element, as has been done in the analysis of supported PtSn particles.14 Based on these considerations and on the fact that the average particle size lies in the range of the sampling depth of the method, we conclude that XPS does not give an accurate estimate of the Pt/In molar ratio for either the surface or the bulk metal ratio of the bimetallic nanoparticles prepared in this study. It was found that the Pt and In binding energies were slightly higher than the corresponding values for the bulk metals
Table 2. Binding Energies and Surface Compositions of the Metals Determined by XPS Measurements Pt 4f PtIn Pt2In Pt3In Pt4In
7/2
[eV]
71.25 71.16 71.01 70.79
In 3d
5/2
[eV]
444.14 444.35 444.23 443.95
Pt/In 0.54 1.35 2.39 2.40
than XPS measurements. The only way in which the same IMFP (1 nm) could be achieved for both Pt 4f and In 3d 3347
dx.doi.org/10.1021/la204838q | Langmuir 2012, 28, 3345−3349
Langmuir
Letter
(70.70 eV for Pt 4f7/29 and 443.9 eV for In 3d5/215). However, the binding energy (BE) of both metals decreased with increasing Pt content, that is, with increasing average particle size. This suggests that the effect of alloying compared to the effect of the particle size on the BE is insignificant in this size regime. These findings are in accord with the literature, which shows that the binding energies of metals in nanoparticles are slightly higher than those in the bulk.16 It has also been proposed that in small alloy particles (
