Article pubs.acs.org/ac
Behavior of Supported Palladium Oxide Nanoparticles under Reaction Conditions, Studied with near Ambient Pressure XPS Astrid Jürgensen,*,† Niels Heutz,‡ Hannes Raschke,† Klaus Merz,‡ and Roland Hergenröder† †
Leibniz Institut für Analytische Wissenschaften − ISAS − e.V., Bunsen-Kirchhoff-Straße 11, 44139 Dortmund, Germany Lehrstuhl für Anorganische Chemie I, Fakultät für Chemie und Biochemie, Ruhr-Universität Bochum, Universitätsstraße 150, D-44801 Bochum, Germany
‡
S Supporting Information *
ABSTRACT: Near ambient pressure X-ray photoelectron spectroscopy (NAP-XPS) is a promising method to close the “pressure gap”, and thus, study the surface composition during heterogeneous reactions in situ. The specialized spectrometers necessary for this analytical technique have recently been adapted to operate with a conventional X-ray source, making it available for routine quantitative analysis in the laboratory. This is shown in the present in situ study of the partial oxidation of 2-propanol catalyzed with PdO nanoparticles supported on TiO2, which was investigated under reaction conditions as a function of gas composition (alcohol-to-oxygen ratio) and temperature. Exposure of the nanoparticles to 2-propanol at 30 °C leads to immediate partial reduction of the PdO, followed by a continuous reduction of the remaining PdO during heating. However, gaseous oxygen inhibits the reduction of PdO below 90 °C, and the oxidation of 2-propanol to carboxylates only occurs in the presence of oxygen above 90 °C. These results support the theory that metallic palladium is the active catalyst material, and they show that environmental conditions affect the nanoparticles and the reaction process significantly. The study also revealed challenges and limitations of this analytical method. Specifically, the intensity and fixed photon energy of a conventional X-ray source limit the spectral resolution and surface sensitivity of lab-based NAP-XPS, which affect precision and accuracy of the quantitative analysis.
A
In this study, such a custom-built instrument (SPECS)6 was used to investigate PdO nanoparticles supported on TiO2 under reaction conditions over a temperature range from 30 to 120 °C, to study the role of PdO in the catalyzed oxidation of 2-propanol. Changes in the oxidation state of palladium as a function of temperature were followed by NAP-XPS (Near Ambient Pressure XPS) measurements at 0.5 mbar in gas-phase environments with selected reactant ratios. By correlating the oxidation state of palladium, obtained from the Pd 3d spectra, with changes in the C 1s spectra, which provide information on the adsorbed species, it is possible to gain insight on the active phase of the catalyst. The study also revealed challenges and limitations of the experimental method, and their impact on the quantitative analysis is discussed. Although irrelevant from an industrial viewpoint, 2-propanol, a secondary alcohol, was used. Primary alcohols decompose via decarbonylation on Pd and PdO surfaces,12,13 releasing CO, a good reducing agent for PdO14,15 Secondary alcohols have a lower tendency for decarbonylation, decreasing the effect of Pd reduction by CO, and 2-propanol, the smallest secondary
challenge in the study of heterogeneous reactions is the analysis of the surface composition during the reaction process. Analysis methods suitable for in situ studies may not have sufficient surface sensitivity, whereas surface science techniques usually require ultrahigh vacuum (UHV) for operation.1 However, the surface structure in UHV is not necessarily the same as under ambient pressure conditions. For example, many adsorbed species and reaction intermediates only exist on surfaces in equilibrium with gases.1,2 Thus, the study of surfaces under more realistic conditions is necessary to gain a better understanding of heterogeneous reactions. To close this “pressure gap”, X-ray photoelectron spectroscopy (XPS) has been adapted to operate under near ambient pressure (NAP) conditions.1 XPS provides quantitative elemental and chemical information about the near-surface region of solids including the makeup of adsorbed species; and the method is suitable for the investigation of the surface composition at increased pressure,2,3 using a synchrotron source and spectrometers designed to operate at pressures in the mbar region.1 Recently, a growing number of papers report the adaptation of these spectrometers for operation with a laboratory X-ray source.4−8 The in situ study of heterogeneous catalytic reactions9−11 is one application of these lab-based instruments. © 2015 American Chemical Society
Received: April 23, 2015 Accepted: July 5, 2015 Published: July 5, 2015 7848
DOI: 10.1021/acs.analchem.5b01531 Anal. Chem. 2015, 87, 7848−7856
Article
Analytical Chemistry
s/pt) regions. The total measurement time for a set of spectra was ∼36 min. The first step, a measurement in UHV, established the starting surface composition. Then 2-propanol was let into the analysis chamber, and a set of spectra was measured (second step), followed by the addition of oxygen and the measurement of another set (third step). This was followed by stepwise heating (∼1.5 °C/min) with measurements taken at 45, 60, 70, 80, 90, 100, 110, and 120 °C (steps 4 to 11). Vacuum was restored, and the sample was allowed to cool down. After ∼15 h, another measurement was collected in high vacuum (∼1 × 10−7 mbar) at ∼30 °C to determine the final surface composition (12th step). Four gas mixtures of 2propanol (C3H8O) and oxygen (O2) were investigated with P(C3H8O):P(O2) having the following values: (1) 0.24 mbar:0.25 mbar, (2) 0.31 mbar:0.22 mbar, (3) 0.46 mbar:0.00 mbar, and (4) 0.16 mbar:0.34 mbar, and to assist in the data analysis, the O 1s, Ti 2p, Pd 3d, and C 1s spectra of several standards, TiO2, Na2CO3, and PdO powders and Pd metal foil, were measured in UHV at 24 °C, and the O 1s, and C 1s spectra of the reactant gases (P = 0.25 mbar) were measured separately without the presence of a solid sample. All spectra were measured in Fixed Analyzer Transmission mode, and because of the significant signal intensity loss under NAP conditions an analyzer pass energy of 80 eV was used. The silver 3d peaks have a fwhm of 1.69 eV at this pass energy. Data Analysis. The spectra were analyzed using CasaXPS version 2.3.15,33 using a Shirley-type background and 70% Gaussian/30% Lorentzian (product form) curves, unless stated otherwise. The spectra of the solid standards (TiO2, Pd, PdO, and Na2CO3) were charge corrected using the C 1s peak of adventitious carbon as internal standard: it was set to 284.8 eV. The binding energies obtained for the solids (Supporting Information) agree well with the values in the literature,34 whereas those of the gases (Supporting Information) are shifted by ∼4.1 eV relative to the literature values35 due to the spectrometer work function of 4.1 eV.6 For the spectra of the catalyst samples, the Ti 2p3/2 peak was used as internal standard, because TiO2 was a spectator in the reaction. Charge correction was achieved by setting the Ti 2p3/2 peak position to 459.0 eV, and to compare spectra measured under different pressure conditions, the intensities were normalized by using the Ti 2p3/2 peak intensity as internal standard. The Pd 3d spectra of the catalyst samples were fitted with the corresponding spectra of the standard compounds. This has the advantage that satellite peaks are modeled properly, resulting in a more accurate fit of the experimental data. Similarly the Pd 3p3/2/O 1s spectra of the Pd and PdO standards were used for fitting the O 1s spectra of the catalyst samples. To model the broadening of the O 1s peak of TiO2 properly, the peaks used to fit the TiO2 standard spectrum, rather than the spectrum itself, were used to fit the spectra of the nanoparticles. Gaussian−Lorentzian peaks were used for the gas phase and the adsorbed carbon species. Further details on the peak fitting procedure are provided in the Supporting Information.
alcohol, has a high vapor pressure, another crucial aspect in the study of gas phase oxidation. Catalyzed oxidation processes are studied as an alternative green method in organic synthesis,16 and transition metal nanoparticles, including palladium, are among the materials investigated.16,17 Specifically, the investigation of the chemical state of palladium under reaction conditions is of interest, as there is still some debate, whether Pd or PdO is the active species.2,18−30 Investigations on the behavior of catalyst materials under reaction conditions have been performed using a synchrotron source: methods like EXAFS22−27 and XANES25−30 are well established for this purpose. In situ XPS studies have also been performed, using Pd(111) as a model catalyst surface.2,10
■
EXPERIMENTAL SETUP Sample Preparation. Colloidal deposition was used for the production of the palladium-based carrier catalysts on a P-25 TiO2 support,31,32 using Na2PdCl4 (Chempur, 99.9%) as a source for palladium, poly(vinyl alcohol) (Sigma-Aldrich, MW: 10 000, 80% hydrolyzed) as a stabilizer, and NaBH4 (Merck, 99%) as a reducing agent. The Pd/TiO2 was then calcinated at 500 °C under a steady-state supply (10 mL/min) of synthetic air (Air Liquide, N2/O2: 80%/20%) to prepare PdO/TiO2. For the in situ NAP-XPS analysis, the PdO/TiO2 was deposited onto a silicon substrate (∼1.0 cm × ∼2.0 cm Si(100)) from suspension using the drop-casting method. A detailed description of the sample preparation process is provided in the Supporting Information. NAP-XPS Instrument. The spectrometer is equipped with a modified SPECS Phoibos 150 electron analyzer mounted behind a wide angular acceptance “prelens” section. The nozzle aperture to the analyzer system in front of the prelens has a diameter of 0.6 mm, and the distance between nozzle and sample surface is also ∼0.6 mm. A high vacuum (≤3 × 10−7 mbar) is maintained at the detector with the aid of differential pumping. Monochromatized Al Kα X-rays (1486.74 eV) focused onto a 0.6 mm Ø spot was provided by a SPECS μFocus 600 photon source operating at 12 kW and 120 W. The windowed (0.1 μm Al coated Si3N4) exit slit is very close (
Behavior of Supported Palladium Oxide Nanoparticles under Reaction Conditions, Studied with near Ambient Pressure XPS Astrid Jürgensen,*,† Niels Heutz,‡ Hannes Raschke,† Klaus Merz,‡ and Roland Hergenröder† †
Leibniz Institut für Analytische Wissenschaften − ISAS − e.V., Bunsen-Kirchhoff-Straße 11, 44139 Dortmund, Germany Lehrstuhl für Anorganische Chemie I, Fakultät für Chemie und Biochemie, Ruhr-Universität Bochum, Universitätsstraße 150, D-44801 Bochum, Germany
‡
S Supporting Information *
ABSTRACT: Near ambient pressure X-ray photoelectron spectroscopy (NAP-XPS) is a promising method to close the “pressure gap”, and thus, study the surface composition during heterogeneous reactions in situ. The specialized spectrometers necessary for this analytical technique have recently been adapted to operate with a conventional X-ray source, making it available for routine quantitative analysis in the laboratory. This is shown in the present in situ study of the partial oxidation of 2-propanol catalyzed with PdO nanoparticles supported on TiO2, which was investigated under reaction conditions as a function of gas composition (alcohol-to-oxygen ratio) and temperature. Exposure of the nanoparticles to 2-propanol at 30 °C leads to immediate partial reduction of the PdO, followed by a continuous reduction of the remaining PdO during heating. However, gaseous oxygen inhibits the reduction of PdO below 90 °C, and the oxidation of 2-propanol to carboxylates only occurs in the presence of oxygen above 90 °C. These results support the theory that metallic palladium is the active catalyst material, and they show that environmental conditions affect the nanoparticles and the reaction process significantly. The study also revealed challenges and limitations of this analytical method. Specifically, the intensity and fixed photon energy of a conventional X-ray source limit the spectral resolution and surface sensitivity of lab-based NAP-XPS, which affect precision and accuracy of the quantitative analysis.
A
In this study, such a custom-built instrument (SPECS)6 was used to investigate PdO nanoparticles supported on TiO2 under reaction conditions over a temperature range from 30 to 120 °C, to study the role of PdO in the catalyzed oxidation of 2-propanol. Changes in the oxidation state of palladium as a function of temperature were followed by NAP-XPS (Near Ambient Pressure XPS) measurements at 0.5 mbar in gas-phase environments with selected reactant ratios. By correlating the oxidation state of palladium, obtained from the Pd 3d spectra, with changes in the C 1s spectra, which provide information on the adsorbed species, it is possible to gain insight on the active phase of the catalyst. The study also revealed challenges and limitations of the experimental method, and their impact on the quantitative analysis is discussed. Although irrelevant from an industrial viewpoint, 2-propanol, a secondary alcohol, was used. Primary alcohols decompose via decarbonylation on Pd and PdO surfaces,12,13 releasing CO, a good reducing agent for PdO14,15 Secondary alcohols have a lower tendency for decarbonylation, decreasing the effect of Pd reduction by CO, and 2-propanol, the smallest secondary
challenge in the study of heterogeneous reactions is the analysis of the surface composition during the reaction process. Analysis methods suitable for in situ studies may not have sufficient surface sensitivity, whereas surface science techniques usually require ultrahigh vacuum (UHV) for operation.1 However, the surface structure in UHV is not necessarily the same as under ambient pressure conditions. For example, many adsorbed species and reaction intermediates only exist on surfaces in equilibrium with gases.1,2 Thus, the study of surfaces under more realistic conditions is necessary to gain a better understanding of heterogeneous reactions. To close this “pressure gap”, X-ray photoelectron spectroscopy (XPS) has been adapted to operate under near ambient pressure (NAP) conditions.1 XPS provides quantitative elemental and chemical information about the near-surface region of solids including the makeup of adsorbed species; and the method is suitable for the investigation of the surface composition at increased pressure,2,3 using a synchrotron source and spectrometers designed to operate at pressures in the mbar region.1 Recently, a growing number of papers report the adaptation of these spectrometers for operation with a laboratory X-ray source.4−8 The in situ study of heterogeneous catalytic reactions9−11 is one application of these lab-based instruments. © 2015 American Chemical Society
Received: April 23, 2015 Accepted: July 5, 2015 Published: July 5, 2015 7848
DOI: 10.1021/acs.analchem.5b01531 Anal. Chem. 2015, 87, 7848−7856
Article
Analytical Chemistry
s/pt) regions. The total measurement time for a set of spectra was ∼36 min. The first step, a measurement in UHV, established the starting surface composition. Then 2-propanol was let into the analysis chamber, and a set of spectra was measured (second step), followed by the addition of oxygen and the measurement of another set (third step). This was followed by stepwise heating (∼1.5 °C/min) with measurements taken at 45, 60, 70, 80, 90, 100, 110, and 120 °C (steps 4 to 11). Vacuum was restored, and the sample was allowed to cool down. After ∼15 h, another measurement was collected in high vacuum (∼1 × 10−7 mbar) at ∼30 °C to determine the final surface composition (12th step). Four gas mixtures of 2propanol (C3H8O) and oxygen (O2) were investigated with P(C3H8O):P(O2) having the following values: (1) 0.24 mbar:0.25 mbar, (2) 0.31 mbar:0.22 mbar, (3) 0.46 mbar:0.00 mbar, and (4) 0.16 mbar:0.34 mbar, and to assist in the data analysis, the O 1s, Ti 2p, Pd 3d, and C 1s spectra of several standards, TiO2, Na2CO3, and PdO powders and Pd metal foil, were measured in UHV at 24 °C, and the O 1s, and C 1s spectra of the reactant gases (P = 0.25 mbar) were measured separately without the presence of a solid sample. All spectra were measured in Fixed Analyzer Transmission mode, and because of the significant signal intensity loss under NAP conditions an analyzer pass energy of 80 eV was used. The silver 3d peaks have a fwhm of 1.69 eV at this pass energy. Data Analysis. The spectra were analyzed using CasaXPS version 2.3.15,33 using a Shirley-type background and 70% Gaussian/30% Lorentzian (product form) curves, unless stated otherwise. The spectra of the solid standards (TiO2, Pd, PdO, and Na2CO3) were charge corrected using the C 1s peak of adventitious carbon as internal standard: it was set to 284.8 eV. The binding energies obtained for the solids (Supporting Information) agree well with the values in the literature,34 whereas those of the gases (Supporting Information) are shifted by ∼4.1 eV relative to the literature values35 due to the spectrometer work function of 4.1 eV.6 For the spectra of the catalyst samples, the Ti 2p3/2 peak was used as internal standard, because TiO2 was a spectator in the reaction. Charge correction was achieved by setting the Ti 2p3/2 peak position to 459.0 eV, and to compare spectra measured under different pressure conditions, the intensities were normalized by using the Ti 2p3/2 peak intensity as internal standard. The Pd 3d spectra of the catalyst samples were fitted with the corresponding spectra of the standard compounds. This has the advantage that satellite peaks are modeled properly, resulting in a more accurate fit of the experimental data. Similarly the Pd 3p3/2/O 1s spectra of the Pd and PdO standards were used for fitting the O 1s spectra of the catalyst samples. To model the broadening of the O 1s peak of TiO2 properly, the peaks used to fit the TiO2 standard spectrum, rather than the spectrum itself, were used to fit the spectra of the nanoparticles. Gaussian−Lorentzian peaks were used for the gas phase and the adsorbed carbon species. Further details on the peak fitting procedure are provided in the Supporting Information.
alcohol, has a high vapor pressure, another crucial aspect in the study of gas phase oxidation. Catalyzed oxidation processes are studied as an alternative green method in organic synthesis,16 and transition metal nanoparticles, including palladium, are among the materials investigated.16,17 Specifically, the investigation of the chemical state of palladium under reaction conditions is of interest, as there is still some debate, whether Pd or PdO is the active species.2,18−30 Investigations on the behavior of catalyst materials under reaction conditions have been performed using a synchrotron source: methods like EXAFS22−27 and XANES25−30 are well established for this purpose. In situ XPS studies have also been performed, using Pd(111) as a model catalyst surface.2,10
■
EXPERIMENTAL SETUP Sample Preparation. Colloidal deposition was used for the production of the palladium-based carrier catalysts on a P-25 TiO2 support,31,32 using Na2PdCl4 (Chempur, 99.9%) as a source for palladium, poly(vinyl alcohol) (Sigma-Aldrich, MW: 10 000, 80% hydrolyzed) as a stabilizer, and NaBH4 (Merck, 99%) as a reducing agent. The Pd/TiO2 was then calcinated at 500 °C under a steady-state supply (10 mL/min) of synthetic air (Air Liquide, N2/O2: 80%/20%) to prepare PdO/TiO2. For the in situ NAP-XPS analysis, the PdO/TiO2 was deposited onto a silicon substrate (∼1.0 cm × ∼2.0 cm Si(100)) from suspension using the drop-casting method. A detailed description of the sample preparation process is provided in the Supporting Information. NAP-XPS Instrument. The spectrometer is equipped with a modified SPECS Phoibos 150 electron analyzer mounted behind a wide angular acceptance “prelens” section. The nozzle aperture to the analyzer system in front of the prelens has a diameter of 0.6 mm, and the distance between nozzle and sample surface is also ∼0.6 mm. A high vacuum (≤3 × 10−7 mbar) is maintained at the detector with the aid of differential pumping. Monochromatized Al Kα X-rays (1486.74 eV) focused onto a 0.6 mm Ø spot was provided by a SPECS μFocus 600 photon source operating at 12 kW and 120 W. The windowed (0.1 μm Al coated Si3N4) exit slit is very close (
