2023 Volume 64 Issue 10 Pages 2445-2449
Pd-loaded perovskite composite manganese oxide (Pd/AMnO3, A = Ca, Sr, La) catalysts were prepared by coprecipitation method, in order to reveal metal-support interaction (MSI). Particulate PdO with sizes of a few ten nanometers were randomly formed on CaMnO3. On the other hand, fibrous PdO with diameter approximately 20 nm was formed on LaMnO3. Both shape of PdO were formed on SrMnO3. HAXPES measurement showed a down shift of valence band of deposited PdO depending on the composite manganese oxide. Our characterization indicates that the MSI at Pd–AMnO3 interface affects not only the shape but the electronic structure of deposited PdO on AMnO3. The CO oxidation activity was in order of Pd/LaMnO3 > Pd/SrMnO3 > Pd/CaMnO3, which corresponds to the order of the PdO valence band shift trend. We proposed that the observed correlation between the valence band shift and the CO oxidation activity for Pd/AMnO3 can be understood in terms of the CO adsorption strength.
Metal-support interaction at Pd-composite manganese oxide AMnO3 (A = Ca, Sr, La) interface is effective to promote CO oxidation.
Palladium nanoparticles (Pd NPs) are widely used as catalysts in many chemical processes, including Suzuki–Miyaura cross-coupling,1) hydrogenation,2) and automobile exhaust gas purification.3) Especially, automobile exhaust catalyst uses 60% of the world’s Pd output.4) The lowering the amount of Pd usage in exhaust catalyst is highly desired. Moreover, the recent tightening of exhaust gas regulations calls for drastic improvements of catalytic activity. Supported Pd catalyst shows excellent performance for exhaust gas purification,5) however, the aggregation of Pd NPs during the preparation process and time on stream represents a serious problem because it significantly reduces their catalytic activity. The Pd-support interface has been reported to act as catalytically active sites and suppress sintering, e.g., the Pd–O–Ce bonds in CeO2-supported Pd catalysts act as active sites for various reactions6–11) and suppress sintering of Pd NPs under harsh thermal aging conditions.12,13) The effectiveness of Pd–O–Al bonds formed on Al2O3 supports for CO oxidation has also been demonstrate.14,15) To enhance the activity of supported Pd catalysts, a deeper understanding of the role of the Pd-support interface is required.
In our previous study, a Pd catalyst supported on Ca2AlMnO5+δ (CAMO), which is known to have a high oxygen storage capacity16–18) has been reported to show high catalytic activity for purifying automotive exhaust gases.19) Interestingly, the H2 reduction treatment of Pd/CAMO results in the formation of a MnO–CaO solid solution shell around the Pd metal core, and that the solid solution layer contributes to the catalytic activity. This demonstrates the possibility that Pd-composite manganese oxide support interface will be effective to promote purifying exhaust gases. Then in this study, we investigated composite manganese oxide support as one of the promising supports for Pd catalyst.
We prepared Pd NPs deposited composite manganese oxide with perovskite type structure (AMnO3, A = Ca, Sr, La) and characterized the deposited Pd properties using XRD, TEM, STEM, HAXPES and tested CO oxidation activity. This reveals that shape and electronic structure of deposited Pd is significantly affected by the AMnO3 support due to metal-support interaction (MSI).
Composite manganese oxides (AMnO3, A = Ca, Sr, La) were synthesized using a coprecipitation method. Mn(NO3)2·6H2O (10 mmol) and each metal nitrates (20 mmol): Ca(NO3)2·4H2O, Sr(NO3)2 and La(NO3)3·6H2O were added to an distilled water (100 mL), and the aqueous solution was added to an alkali solution (100 mL) contains sodium carbonate (0.1 mol/L) and sodium hydroxide (0.1 mol/L) while being stirred at room temperature for 1 h. The formed precipitates were centrifuged, washed with distilled water and acetone, dried at room temperature overnight, and then calcined at 1000°C for 2 h in air. Pd loading was conducted using an impregnation method. Pd(CH3COO)2 (0.0212 g) was dissolved in acetone (10 mL), and the solution was supplemented with the composite manganese oxides (AMnO3) (0.99 g). The slurry was dried at 50°C for 30 min, and the obtained powder was calcined at 800°C in air for 2 h to fabricate the Pd loaded-composited manganese oxide catalysts. The Pd loading amount was 1 mass% based on the metal.
2.2 Characterization of prepared samplesPhase identification was performed by means of X-ray diffraction (XRD) using Cu Kα (λ = 1.543 Å) operation at 40 kV and 15 mA (Rigaku MiniFlex600-C). Transmission electron microscopy (TEM, STEM) analysis was carried out with a JEOL 2100F microscope with an operating voltage of 200 kV. The composition of the prepared samples was analyzed by an EDS spectrometer, which was attached onto the TEM. Hard X-ray photoemission spectroscopy (HAXPES) was performed using an X-rays with photon energy of 7.940 keV at the beamline BL09XU equipped with a high-resolution electron spectrometer (VG Scienta R4000) of SPring-8, Japan. Sample powder was mixed with carbon nanotube and molded into pellets. Total energy resolution was set to 208 meV. The binding energy of photoelectrons was referenced to the Fermi energy of an Au film that was electrically contacted to the sample. The measurements were performed at room temperature. The valence band centroids were evaluated by applying the following formula.
| \begin{align*} &\text{Valence band centroid}\\ &\quad = \frac{\displaystyle\int_{10\textit{eV}}^{0\textit{eV}} \textit{binding energy}(E) \times \textit{intensity}(E)dE}{\displaystyle\int_{10\textit{eV}}^{0\textit{eV}}\textit{intensity}(E)dE} \end{align*} |
The catalytic test was conducted under stoichiometric conditions of 1000 ppm CO, 500 ppm O2 and He balance. That is, the stoichiometric condition indicated the following chemical reaction:
| \begin{equation*} \text{CO} + \text{1/2O$_{2}$} \rightarrow \text{CO$_{2}$} \end{equation*} |
The catalyst (100 mg; 25–50 mesh) was pretreated in a flow of He gas at 500°C for 30 min and exposed to the reaction gas at 100°C. Then, the reaction temperature was increased from 100 to 500°C in increments of 50°C, held for 20 min, and the outlet gas was analyzed using a gas chromatograph (Agilent, high-speed and compact gas analyzer Micro GC).
Figure 1(a) shows powder X-ray diffraction patterns (XRD) for the prepared composite manganese oxides (AMnO3, A = Ca, Sr, La) and the assigned model structures illustrated by VESTA20) are also shown in Fig. 1(b). ABO3-type perovskite CaMnO3, SrMnO3 and LaMnO3 single phase were obtained, respectively. The CaMnO3 and SrMnO3 are cubic perovskite structure, and the LaMnO3 is rhombohedral distorted perovskite. No other impurity phase was observed like Mn oxide. It is confirmed that all the synthesized composite manganese oxides were retained after the Pd deposition with the impregnation method.
(a) Powder X-ray diffraction patterns (XRD) for the synthesized composite manganese oxides (CaMnO3, SrMnO3, LaMnO3). (b) Model structures assigned with the XRD illustrated by VESTA.20)
Figure 2 shows TEM and HRTEM images for Pd-loaded AMnO3. Nano-sized particulate Pd species were formed on CaMnO3 and SrMnO3. Larger Pd particles were observed on Pd/CaMnO3 than Pd/SrMnO3, where the average Pd particle size was 36.3 and 20.2 nm, respectively as indicated in the histogram in the inset of the TEM image. In part, fibrous Pd was also observed on Pd/SrMnO3. While, Pd on LaMnO3 formed only fiber shape as traced in the TEM image (Fig. 2(c)), where isolated particulate Pd species was rarely observed. The average fiber diameter was 20 nm. The inset in the HRTEM images show the corresponding fast Fourier transform pattern (FFT) obtained from the Pd species area. The FFTs demonstrated that Pd was presence as PdO for all of the AMnO3 surface. The TEM observation confirmed that the shape and size of loaded Pd were much different according to the AMnO3.
TEM and HRTEM images for the Pd loaded-synthesized composite manganese oxide: (a) Pd/CaMnO3, (b) Pd/SrMnO3, (c) Pd/LaMnO3. The inset in the HRTEM shows the corresponding fast Fourier transform pattern obtained from the Pd particle area.
Figure 3 shows STEM-DF and EDS elemental mapping for Pd/AMnO3 particle. We confirmed at the beginning that distribution of Mn (red) mapping overlapped considerably with A (green) = Ca, Sr, La mapping over all the AMnO3 particles as shown in the mix images, showing uniform composition of AMnO3 particles fabricated by the coprecipitation method and also retained through Pd deposition. In addition, Mn and A signal were not appeared to Pd (blue) region on the surface, indicating that Pd–Mn or Pd–A alloying hardly occurred thought Pd deposition. The STEM-EDS revealed the random deposition of particulate PdO on the CaMnO3 (Fig. 3(a)). The PdO particles were isolated each other. On the other hand for LaMnO3, Pd region connected continuously on the LaMnO3 surface (Fig. 3(c)). For on the SrMnO3 (Fig. 3(b)), both isolated particles and connected fiber shape of PdO can be observed. The appearance of the fibrous PdO would be attributed to metal-oxide interaction (MSI) for Pd–AMnO3 interface. Fibrous Pd deposition instead of particle deposition on LaMnO3 and SrMnO3 were considered as evidence of strong interaction at Pd–AMnO3 interface.
STEM-DF and EDS elemental mapping of Pd (blue), Mn (red), A = Ca, Sr, La (green), composite (Pd+Mn+A) images for Pd/AMnO3: (a) Pd/CaMnO3, (b) Pd/SrMnO3, (c) Pd/LaMnO3.
Figure 4(a) shows HAXPES spectrum in the Pd 3d core-level region for the Pd/AMnO3 together with those for reference PdO and metal Pd. The Pd 3d peaks for the Pd/AMnO3 are assigned to PdO state, showing the Pd oxide deposition on the AMnO3 particles. Figure 4(b) shows the valence band spectra for the Pd/AMnO3. The valence band of the reference PdO mainly situated in the biding energy within 3 eV. Thus, we can identify that the deposited PdO valence band region for Pd/AMO3 exists within several binding energy exclusively. Figure 4(c) shows the enlarged valence region within 4 eV for Pd/AMO3. The PdO valence region for Pd/AMO3 was greatly different with respect to AMO3, where it shifted to higher biding energy (downshift) compared to reference PdO in order of Pd/LaMnO3, Pd/SrMnO3 and Pd/CaMnO3. The centroid in valence band spectra was calculated and drawn a line of centroid in Fig. 4(b). The location of centroid was PdO (3.7 eV), Pd/CaMnO3 (4.8 eV), Pd/SrMnO3 (5.4 eV) and Pd/LaMnO3 (5.3 eV), respectively. We estimated contributions from substrate oxide (AMnO3) in valance spectra by subtracting bare AMnO3 from the original Pd loaded AMnO3 spectra and drew a line of centroid as shown in Fig. 4(d). The valence spectra after AMnO3 subtraction are mainly derived from deposited PdO. The centroid in valence band after AMnO3 subtraction were Pd/CaMnO3 (3.6 eV) and Pd/LaMnO3 (4.5 eV). The deposited PdO on CaMnO3 have close centroid to reference PdO. On the other hand, PdO on LaMnO3 showed a rather shift of centroid. We can confirm the downshift of valence band for the deposited PdO from the aspect of centroid shift in valence band spectra.
HAXPES spectra for Pd/AMnO3 (A = Ca, Sr, La) and reference Pd species: (a) Pd 3d core level, (b) Valence region, (c) Enlarged view for Fig. 4(b), (d) Valence spectra after AMnO3 subtraction.
This valance band shift indicates that the MSI at Pd–AMnO3 interface affects not only the shape but the electronic structure of deposited PdO. The MSI at Pd–AMnO3 stabilized fibrous Pd on AMnO3 surface, and accompanying Pd valence band shift contributed to enhancement of catalytic activity as described below.
3.2 CO oxidation activity for Pd/AMnO3Figure 5 shows the catalytic activity of CO oxidation with O2 over Pd/AMnO3. All the samples were pretreated in flow of He at 500°C and exposed to the reaction gas at 100°C. Then, the reaction temperature was increased from 100 to 500°C. The catalytic activity was in the order of Pd/LaMnO3 > Pd/SrMnO3 > Pd/CaMnO3. It should be noted that the CO oxidation activity of Pd/AMnO3 corresponds to the order of the Pd valence band shift trend as shown in Fig. 4(c). There were several reports for CO oxidation activity on PdO surface.21–23) The previous reports showed that PdO (101) surface leads to effective CO oxidation via the Langmuir–Hinshelwood (LH) mechanism. The reaction on the PdO surface has a low activation energy, which is comparable to the lowest activation energies observed on metallic Pd surfaces. At the same time, theoretical study also demonstrated that excess strong CO adsorption on PdO surface is the less favorable for CO oxidation due to surface blocking by CO adlayer, which enables competition between the LH surface reaction and Eley–Rideal (ER) mechanism without O2 adsorption.21) Therefore, an optimal CO adsorption strength on the PdO surface is favorable for CO oxidation. For Pt-alloy catalyst study, it has been reported that CO adsorption strength on the Pt-alloy surface monotonically decreases with shifting of the Pt d-band center because of diminished hybridization of the surface d-band and the lowest-unoccupied molecular orbital (LUMO) of CO.24) It is considered that similarly decreasing of CO adsorption strength will be occurred on the deposited PdO on AMnO3 according to the down shifting of PdO valence band. The lowering of CO adsorption strength will lead to inhibit surface blocking by CO adlayer on PdO surface, which enable co-adsorption of CO and O2 on the surface. From these findings, the observed correlation between the PdO valence band shifting and the CO oxidation activity for Pd/AMnO3 can be understood in terms of the CO adsorption strength. The favorable Pd/LaMnO3 surface for CO oxidation is anticipated to an optimal CO adsorption strength provided by PdO valence band shifting. It is probably that other factor of high activity for Pd/LaMnO3 is structural effect. LaMnO3 is rhombohedral crystal structure different from the other two. In previous report for Pd/LaMnO3 catalyst, the PdO clusters transformed partially to highly dispersed Pd oxides or nonstoichiometric PdOx phases during the catalytic reaction, which are the catalytically active phases for the total oxidation of methane.25) There is a possibility that such catalytically active phases for highly dispersed and/or nonstoichiometric PdOx phases are formed during CO oxidation reaction.
Light-off curve of CO conversion to CO2 over Pd/AMnO3 (A = Ca, Sr, La).
In conclusion, we have investigated the metal-support interaction (MSI) at Pd-composite manganese oxide interface for Pd/AMnO3 (A = Ca, Sr, La) prepared by coprecipitation method. The MSI at Pd–AMnO3 interface affects the shape of deposited PdO and its valence band profile. CO oxidation activity of Pd/AMnO3 corresponds to the order of the PdO valence band shift trend. The Pd–AMnO3 interface is effective to promote CO oxidation. We have proposed the composite manganese oxide support as one of the promising supports for Pd catalyst.
The HAXPES measurements were performed under the approval of the SPring-8 Synchrotron X-ray Station (Proposal 2023A1540). We thank Dr. A. Yasui and Dr. Y. Takagi for them help for HAXPES measurement at the SPirng-8 BL09. This work was supported by JSPS KAKENHI Grant Number JP23K04501, JP22K14756.
