Article pubs.acs.org/JACS
Systematic Identification of Promoters for Methane Oxidation Catalysts Using Size- and Composition-Controlled Pd-Based Bimetallic Nanocrystals Joshua J. Willis,† Emmett D. Goodman,† Liheng Wu,†,‡ Andrew R. Riscoe,† Pedro Martins,† Christopher J. Tassone,‡ and Matteo Cargnello*,† †
Department of Chemical Engineering and SUNCAT Center for Interface Science and Catalysis, Stanford University, Stanford, California 94305, United States ‡ Stanford Synchrotron Radiation Lightsource, SLAC National Accelerator Laboratory, Menlo Park, California 94025, United States S Supporting Information *
ABSTRACT: Promoters enhance the performance of catalytic active phases by increasing rates, stability, and/or selectivity. The process of identifying promoters is in most cases empirical and relies on testing a broad range of catalysts prepared with the random deposition of active and promoter phases, typically with no fine control over their localization. This issue is particularly relevant in supported bimetallic systems, where two metals are codeposited onto high-surface area materials. We here report the use of colloidal bimetallic nanocrystals to produce catalysts where the active and promoter phases are colocalized to a fine extent. This strategy enables a systematic approach to study the promotional effects of several transition metals on palladium catalysts for methane oxidation. In order to achieve these goals, we demonstrate a single synthetic protocol to obtain uniform palladium-based bimetallic nanocrystals (PdM, M = V, Mn, Fe, Co, Ni, Zn, Sn, and potentially extendable to other metal combinations) with a wide variety of compositions and sizes based on hightemperature thermal decomposition of readily available precursors. Once the nanocrystals are supported onto oxide materials, thermal treatments in air cause segregation of the base metal oxide phase in close proximity to the Pd phase. We demonstrate that some metals (Fe, Co, and Sn) inhibit the sintering of the active Pd metal phase, while others (Ni and Zn) increase its intrinsic activity compared to a monometallic Pd catalyst. This procedure can be generalized to systematically investigate the promotional effects of metal and metal oxide phases for a variety of active metal-promoter combinations and catalytic reactions.
1. INTRODUCTION Nanostructured supported catalysts are composed of finely divided particles deposited on high-surface area, thermally stable materials. Very often, other substances called promoters or activators are added, which increase the efficiency of the active phase. Promoters enhance the performance of catalytic materials in different ways: (i) by increasing rates;1−5 (ii) improving thermal and/or chemical stability;6−8 and (iii) by increasing the selectivity of the catalytic process.9−11 There are numerous mechanisms through which promoters enhance catalytic properties. Common mechanisms include changing the electronic structure of the active metal,1−3 stabilizing the active phase against sintering,8 separating reaction steps at different sites,5 and influencing the redox properties of the active phase.4,6,11 A notable example includes alkali promoters (Na, K, and Cs) that have been demonstrated to increase the electron density of active metals Ru and Fe, creating a favorable state to activate N2 in ammonia production.1,12 Interfacial effects have been elegantly demonstrated by Chen et al., who showed that an iron−nickel hydroxide promoter phase deposited onto platinum nanoparticles creates unique metal− OH−Pt sites that enhance CO oxidation activity and stability © 2017 American Chemical Society
by separating the reaction steps of the CO-binding and O2activation onto different proximal sites.5 Similar interfacial effects have been shown to be crucial by Seitz et al. in greatly enhancing the oxygen evolution reaction activity of MnOx catalysts. In this reaction, noble metals, such as Au, Pt, Pd, and Ag, influence the oxidation state of manganese at the interface.4 There are also textural promoters, where promoting compounds increase the thermal stability of active phases while not participating in the catalytic cycle, of which Cr in hightemperature water−gas shift catalysts is an example.13 In most cases, however, the colocalization and interfacial interaction between promoter and active phase are essential. Conventional coprecipitation, sequential, or codeposition methods for the preparation of catalysts usually result in a random distribution of promoter and active phase.14−16 This heterogeneity complicates efforts to link properties to structural or compositional changes. In reactions where two active sites are indispensable for reactivity, such as reduction and oxidation reactions, this problem is further accrued, unless a support is Received: June 16, 2017 Published: August 11, 2017 11989
DOI: 10.1021/jacs.7b06260 J. Am. Chem. Soc. 2017, 139, 11989−11997
Article
Journal of the American Chemical Society used as a promoter (as in the notable example of ceria17). This difficulty is reflected in contrasting results in the literature about promoting phases. For example, in a study of the influence of several metals on the methane combustion activity of Pd-based catalysts synthesized via incipient wetness impregnation, no metal showed any improvement compared to monometallic Pd sample.14 The random distribution of the two metals on the support and the potential coverage of some of the Pd sites may be the reason for the apparent lack of promotion. Other reports instead show the improvement in catalytic activity when Pd is promoted by transition metal NiO deposited on Al2O3 and attributed the enhancement to increased PdO phase stabilization in the presence of NiO.15 These representative conflicting results highlight the need for improved synthetic techniques which impart control over the relative locations of the active and promoter phases. Recent advances in synthesis technologies have shown improved control, but often lack simultaneous control of size and composition of the active phase, and are limited in their broad applicability. As an example, strong electrostatic adsorption processes can selectively deposit multiple metal ions onto supports,18,19 but have limited control over the size of the particles. Atomic layer-deposition (ALD), where thin films are deposited on the surface of support materials through self-limiting chemical reactions, can achieve uniform films but is limited to available precursors that can be easily vaporized and requires the optimization of specific deposition conditions.20−22 The use of preformed bimetallic nanoparticles as catalyst precursors is a promising synthetic strategy to achieve this control, since in these materials the active phase-promoter interaction can be predetermined through precise colloidal chemistry and preserved throughout deposition and catalyst activation steps.5,23−28 The precise colocalization of active phase and promoter results in a much improved catalytic performance, as recently shown for the hydrogenation of nitroaromatics using Pt@SnO2 nanocrystals when compared to monometallic platinum catalyst. This effect is due to a strong weakening of product binding in the presence of SnO2 phase on the Pt nanocrystal surface.23 In a further refinement of the active metal-promoter positioning strategy, compositioncontrolled AuMn bimetallics demonstrated a fine control of Au−MnO heterostructure architectures and determined the optimal Au−MnO architecture for enhanced H2O2 detection while maintaining electrical conductivity.28 These colloidalbased systems demonstrate the clear potential of preformed nanocrystals in providing improved catalyst performance and control over the proximity of promoter and active phase. We herein report a generalized procedure to study the effect of transition metals on the activity of Pd-based catalysts for methane combustion with the goal of discovering true promoters in a systematic fashion. The method relies on a single synthetic protocol based on high-temperature thermal decomposition of organometallic precursors to prepare uniform PdM bimetallic nanocrystals (M = V, Mn, Fe, Co, Ni, Zn, and Sn) with tunable compositions and sizes. These monodisperse nanocrystals are deposited onto high-surface area alumina, and a controlled thermal treatment is used to promote the segregation of the transition metal oxide phase from the PdO phase, while maintaining them in close proximity. Using this method, a library of Pd-transition metal oxide catalysts is realized where the promoting effect can be studied in detail. Catalytic characterization of this library of materials reveals how several metals (Fe, Co, Ni, Zn, and Sn) improve Pd thermal
stability and increase catalytic rates up to two-fold on a permass Pd basis and up to 3-fold on a turnover frequency basis compared to the monometallic palladium catalyst, in contrast to literature reports that provide contrasting results. Detailed structural characterization suggests that some of the promoters (Fe, Co, and Sn) inhibit the sintering of the metal phase, while others (Ni and Zn) change the intrinsic activity of the PdO phase. This method demonstrates the utility of precise bimetallic nanocrystals in the systematic study of catalytic promoters and can prove useful in other reactions where active phase-promoter interactions are crucial for the catalytic performance.
2. EXPERIMENTAL SECTION 2.1. Materials. Palladium acetylacetonate (Pd(acac)2; 35% Pd), vanadyl(IV) acetylacetonate (VO(acac)2; 99%), manganese(III) acetylacetonate (Mn(acac)3;97%), iron(III) acetylacetonate (Fe(acac)3; 99+%), cobalt(II) acetylacetonate (Co(acac)2; 99%), zinc(II) acetylacetonate (Zn(acac)2; 25% Zn) and octadecene (90%, ODE) were purchased from Acros Organics. Nickel II acetylacetonate (Ni(acac)2; 95%), tin(IV) bis(acetylacetonate) dichloride (98%, Sn(acac)2Cl2), oleylamine (70%, OLAM), oleic acid (90%, OLAC) and trioctylphosphine (97%, TOP) were purchased from SigmaAldrich. Tetradecene (94%, TDE) was purchased from Alfa Aesar. Al2O3 Pluralox TH100/150 was obtained from Sasol and calcined at 900 °C for 24 h using heating and cooling ramps of 3 °C min−1. All of the solvents were of reagent grade. All reagents were used as-received. 2.2. Synthesis of Bimetallic PdM (M = V, Mn, Fe, Co, Ni, Zn, and Sn) Nanocrystals. All syntheses were performed using standard Schlenk techniques. In a typical synthesis, Pd(acac)2 (0.20 mmol) and M(acac)x (0.05 mmol) were mixed with ODE (10 mL), OLAM, and in certain cases OLAC in a three-neck flask. The mixture was evacuated at room temperature for 15 min under magnetic stirring. TOP was then added under evacuation, and the mixture was heated to 50 °C. The solution was left under vacuum for 30 min to remove all water and other impurities. At this point, the reaction mixture was a transparent solution. The reaction flask was then flushed with nitrogen and heated quickly (∼40 °C min−1) to the desired temperature. After 15 min of reaction at the appropriate temperature under magnetic stirring, the solution was quickly cooled to RT by blowing compressed air on the outside of the flask and adding a water bath when the temperature was below 170 °C. The particles were purified by three rounds of precipitation with isopropanol and ethanol, and separated by centrifugation (8000 rpm, 3 min), with redissolution in a hexanes/ OLAM solution (20 mL Hexanes: 100 μL OLAM) after each centrifugation step. Finally, the particles were dissolved in hexanes producing a deep black solution and stored at RT. Detailed synthetic conditions can be found in Table S1 in the Supporting Information (SI). 2.3. Preparation of Catalyst Powders. An appropriate amount of metal nanoparticles (to give a weight loading of 0.3−0.5 wt % Pd, confirmed by ICP-OES) was added to a dispersion of alumina in hexanes under vigorous stirring. The mixture was left stirring for 15 min. The solid was recovered by centrifugation (8000 rpm, 3 min). Colorless supernatants, signifying complete adsorption, were observed for most catalysts. The powder was then washed once with hexanes (30 mL), with sonication and centrifugation. Final powders were dried at 80 °C overnight and sieved below 180 μm grain size. Powders were calcined at either 500 or 850 °C for 5 h in a static air furnace using a heating and cooling ramps of 3 °C min−1. Final calcined powders were again sieved below 180 μm grain size and collected. 2.4. Characterization Techniques. CO Chemisorption experiments were carried out on a Micromeritics 3Flex. The samples were placed in a U-shaped quartz reactor and then pretreated and degassed in the following manner: evacuated at 110 °C for 30 min, heated in flowing 5% O2 in Ar at 300 °C for 30 min, evacuated at 300 °C for 30 min, reduced in flowing 5% H2 in Ar at 300 °C for 1 h, and then evacuated at 300 °C for 4 h. All experiments were conducted at 35 °C 11990
DOI: 10.1021/jacs.7b06260 J. Am. Chem. Soc. 2017, 139, 11989−11997
Article
Journal of the American Chemical Society
temperature (typically 200 °C) under Ar flow at 45 mL min−1. The reaction mixture was then introduced. After a stable CO2 production was observed, kinetic rates and light-off curves were measured. For kinetic rates, conversions of the limiting reactant were always kept below 2% conversion to guarantee differential working conditions. Reaction rates were calculated on the basis of accessible metal surface area calculated from CO chemisorption measurements. To record light-off curves, the catalyst was heated from 200 to 850 °C at a ramp rate of 10 °C min−1. Carbon balance closes to 99 ± 1%. Temperature-programmed oxidation (TPO) experiments were conducted on the samples calcined to 850 °C. The catalyst powder (∼150 mg) was placed in a U-shaped quartz reactor and cleaned prior to measurement under a flow of O2 (5%) in Ar at 45 mL min−1 for 30 min at 300 °C. After cooling to the initial measurement temperature (200 °C) under Ar flow, the catalyst was exposed to a mixture of O2 (2%) in Ar at 20 mL min−1. The temperature was then raised to 1000 °C at 10 °C min−1 and cooled down using the same rate. Oxygen release-uptake was evaluated using a quadrapole MS (Hiden Analytical HPR20) equipped with a Secondary Electron Multiplier (SEM) detector tracking the parent molecular ion of O2 (32 amu).
in the pressure range from 100 to 450 Torr, using a double isotherm to remove the contribution from physisorption. Adsorption values were obtained by linear extrapolation to zero pressure. Bright-field transmission electron microscopy (TEM) images and energy-dispersive X-ray (EDX) spectra were collected on a FEI Tecnai transmission electron microscope equipped with an Orius CCD camera and an EDAX SUTW EDX detector. Aberration-corrected TEM images, dark-field scanning TEM (DF-STEM) images, EDX maps, and energy-filtered (EF) maps were recorded using an FEI Titan environmental transmission electron microscope equipped with a spherical aberration corrector in the image forming lens, a Gatan OneView camera, an Oxford Xmax SDD EDX detector, and a Gatan Quantum 966 EEL spectrometer. Both instruments were operated in high vacuum at 200 kV. Nanoparticle samples were dropcast onto TEM grids from their hexane solutions. Supported catalyst TEM grids were prepared by dry deposition by lightly shaking a lacey carbon Cumesh TEM grid with catalyst powder in a plastic tube. XRD patterns were obtained on a PANalytical X’Pert PRO X-ray diffractometer in the 2θ range of 20−80° (Cu Kα radiation, λ = 1.5418 Å), and samples were prepared by depositing a thick film of the particles from hexanes solution on a glass substrate. Small Angle X-ray Scattering (SAXS) measurements were performed at Beamline 1−5 at Stanford Synchrotron Radiation Lightsource (SSRL) of SLAC National Accelerator Laboratory. The scattered patterns were collected using a Rayonix 165 SX CCD area detector and analyzed by fitting to a quantitative model using the IRENA package (available at usaxs.xray.aps.anl.gov/staff/ilavsky/irena.html from the APS) (Ilavsky, J. and Jemian, P. R. Irena: Tool suite for modeling and analysis of small-angle scattering.29 The size and size distribution of disperse nanocrystals were modeled using the Modeling II module of the IRENA package. Quantitative elemental analysis of the multimetallic nanoparticles and supported systems were analyzed by inductively coupled plasma optical emission spectroscopy (ICP-OES) (Thermo Scientific ICAP 6300 Duo View Spectrometer). Unsupported and supported systems were digested with a mixture of nitric acid (710 μL) and hydrochloric acid (660 μL) for 7 h, filtered, and then diluted before measurements. 2.5. Catalytic Characterization. All experiments were conducted at atmospheric pressure. All catalytic measurements were performed in a U-shaped quartz microreactor with an internal diameter of 10 mm. The catalyst (20−25 mg) was physically mixed with alumina in a 1:10 dilution ratio to ensure neither mass nor thermal diffusion limited affected the results. The diluted catalyst mixture was then loaded into the reactor to give a bed length of about 1 cm, between two layers of granular acid-washed quartz, used both for preventing displacement of the catalyst powder and preheating the reaction mixture. The reactor was heated by a Micromeritics Eurotherm 2416 furnace, and the temperature of the catalyst was measured with a K-type thermocouple inserted inside the reactor and touching the catalyst bed. No appreciable conversions were found when only quartz or the bare alumina were placed in the reactor in the range of temperatures used for this study. Reaction mixtures (0.5% CH4, 2% O2, balance Ar) were prepared by mixing 5% CH4 (certified standard, Airgas), 5% O2 in Ar (certified standard, Airgas) and Ar (99.999%, Airgas) using electronic thermal mass flow controllers (Brooks SLA5850). H2O vapor was introduced into the gas stream using a saturator maintained at 30 °C. Gas Hourly Space Velocities (GSHV) were held in the range of 40 000 to 75 000 L gPd−1 h−1. Reactant and product concentrations were monitored online using either a quadrupole Mass Spectrometer (MS) (Hiden Analytical HPR20) equipped with a Faraday detector or a gas chromatograph (GC) (Buck Scientific model 910). In the case of MS analysis, the Faraday detector was used to follow the parent molecular ions for CH4 (15 amu), O2 (32 amu), CO2 (44 amu), and H2O (18 amu). In the case of GC analysis, a thermal conductivity detector (TCD) and flame ionization detector with a methanizer were employed using Ar as the carrier gas. Prior to measuring the catalyst performance, each catalyst was cleaned under a flow of O2 (5%) in Ar at 45 mL min−1 for 30 min at 300 °C. Then, the sample is cooled to initial measurement
3. RESULTS AND DISCUSSION 3.1. Synthesis of PdM Bimetallic NCs. Monodisperse palladium-based bimetallic PdM (M = V, Mn, Fe, Co, Ni, Zn, and Sn) nanocrystals (NCs) were prepared by thermal decomposition of transition metal acetylacetonate precursors in a octadecene (ODE) solution in the presence of 1oleylamine (OLAM), trioctylphosphine (TOP), and, in some samples, oleic acid (OLAC) (see Table S1). The high degree of uniformity of the obtained PdM NCs is demonstrated both qualitatively and quantitatively by representative transmission electron microscopy (TEM) analysis and small-angle X-ray scattering (SAXS) data (Figures 1, 2 and Table S2). Energydispersive X-ray spectroscopy (EDX) and inductively coupled plasma optical emission spectroscopy (ICP-OES) (Figure S1, Table S1) confirm the presence of both (or three) metals in the NCs. Energy-filtered TEM (EF-TEM) imaging reveals that the second metal is uniformly distributed in the bimetallic NCs (Figure 3), further demonstrating the uniformity of these obtained NCs in size and composition. A majority of the PdMsmall (PdM-s) NCs (PdV-s, PdMn-s, PdFe-s, PdFeCo-s, PdCos, PdNi-s, and PdZn-s) exhibit a polycrystalline structure, as demonstrated by the broad peak at ca. 40° in their XRD patterns (Figure S2) and the presence of polycrystalline domains in the NCs in the aberration corrected TEM (ACTEM images) (visible in Figure S3), in agreement with previous studies.30−32 In contrast, the obtained PdSn nanoparticles are highly crystalline Pd2Sn intermetallics (Figures S2, S4, and S5). Control experiments were used to investigate the formation mechanism. EDX spectra and representative TEM images demonstrate little to no incorporation of the second metal in syntheses performed at temperature slower than the ones reported in the synthesis section, while monodisperse Pd nanoparticles still form (Figures 4, S6, and S7). This is in agreement with previous literature, as TOP-based Pd nanoparticle syntheses demonstrate that nucleation of pure Pd NCs occurs at 200−250 °C,32−35 explaining the formation of Pd nanoparticles even at lower temperatures than those used for the final PdM NCs. The literature also helps to explain the requirement for elevated temperatures for the inclusion of the second metal. The synthesis of metal-oxide nanoparticles (Mn, Fe, Co, Ni, Zn, and Sn) with OLAC present require higher temperatures (>280 °C) to decompose the stable metal precursor (usually a metal-oleate) .36−38 For the PdV syntheses, 11991
DOI: 10.1021/jacs.7b06260 J. Am. Chem. Soc. 2017, 139, 11989−11997
Article
Journal of the American Chemical Society
Figure 1. TEM images of (a) PdV-s, (b) PdMn-s, (c) PdFe-s, (d) PdFeCo-s, (e) PdCo-s, (f) PdNi-s, (g) PdZn-s, and (h) PdSn-s. Scale bars are all 20 nm.
where OLAC is not present, VO(acac)2 is a highly stable precursor which requires elevated temperatures as well to decompose.39 On the basis of the evidence provided from the control experiments and literature, we propose the following mechanism: Pd NCs first form at lower temperatures (200− 250 °C) followed by the decomposition of the second metal precursor and alloying with the preformed Pd nanoparticles (Figure 4e). Thus, it is expected that controlling the nucleation and growth of the Pd NC can be used to control the size and composition of the PdM NCs. Using this understanding, we demonstrate size control by increasing the size of PdM NCs with varying the TOP:Pd ratio (Figures 5, S8, and Table S2), in agreement with previous studies.33,34 EDX spectra and ICP-OES analysis of the large PdM (M = Mn, Fe, Co, Ni, Zn) nanoparticles confirm similar ratios of second metal incorporation compared to the small PdM nanoparticles (Figure S9 and Table S1). Structural analysis with XRD demonstrates that the large PdM NCs (PdM-L) also exhibit a similar polycrystalline
Figure 2. Small angle X-ray scattering patterns and particle size distributions for PdV-s, PdMn-s, PdFe-s, PdFeCo-s, PdCo-s, PdNi-s, PdZn-s, and PdSn-s. For TEM measurements, N ≥ 100 counts.
structure as their small PdM nanoparticle counterparts (Figure S10). Since the second metal decomposes and alloys at the elevated temperature, increasing the amount of second metal in the reaction solution results in an increase in second metal fraction in the NCs. This control is demonstrated in the particular example of the Pd/Co system in Figure S11, where the Pd/Co ratio is varied from 4:1 to 0.7:1. In comparison, the PdSn system, which forms stable intermetallics, increases in size with an increase in the amount of Sn(acac)2Cl2 in the reaction solution, but the composition remains similar to what was 11992
DOI: 10.1021/jacs.7b06260 J. Am. Chem. Soc. 2017, 139, 11989−11997
Article
Journal of the American Chemical Society
Figure 3. Representative TEM and EF-TEM images of the following: (a) PdMn-s, (b) PdCo-s, and (c) PdNi-L.
Figure 5. TEM images of (a) PdMn-L, (b) PdFe-L, (c) PdCo-L, (d) PdNi-L, (e) PdZn-L, and (f) PdSn-L NCs. Scale bars are all 20 nm.
combustion activity, two complementary methods were employed. Light-off curves were used to qualitatively compare the effects of the different metal oxides on methane combustion activity. Further kinetic measurements, performed in the differential regime (
