Low-Temperature Activity and PdO-PdO x Transition in Methane Combustion by a PdO-PdO x / γ-Al 2 O 3 Catalyst

The search to discover a suitable catalyst for complete combustion of methane at low temperature continues to be an active area of research. We prepared a 5 wt % PdO-PdOx/γ-Al2O3 catalyst by a modified Vortex-assisted Incipient Wetness Method. X-ray Photoelectron Spectroscopy showed that the original catalyst contained PdO (38%) and PdOx (62%) on the surface and indicated that PdOx originated from the interaction of PdO with the support. Scanning Transmission Electron Microscopy confirmed the catalyst had an average particle size of 10 nm and was well-dispersed in the support. The catalyst exhibited exceptional low-temperature activities with 90–94% methane conversion at 300–320 ◦C. The catalyst was active and stable after several catalytic runs with no signs of deactivation by steam in this narrow temperature range. However, the conversion decreased in the temperature range 325–400 ◦C. The surface composition changed to some extent after the reaction at 325 ◦C. A tentative mechanism proposes PdOx (Pd native oxide) as the active phase and migration of oxide ions from the support to PdO and then to PdOx during the catalytic oxidation. The high methane conversion at low temperature is attributed to the vortex method providing better dispersion, and to catalyst–support interaction producing the active phase of PdOx.


Introduction
Methane is a greenhouse gas with a global warming potential 25 times higher than carbon dioxide.One remediation technique involves the use of catalysts to convert methane into a less harmful or shorter-lived species.Of these catalysts, palladium shows promise for conversion of methane at low temperatures, lending hope to automakers and industries using turbines powered by natural gas.While the idea of low-temperature combustion is appealing, finding a catalyst that can perform in "lean-burn" conditions has been a challenge.To date, an ideal catalyst that performs optimally under 400 • C remains undiscovered, but palladium, supported on various materials, comes close [1].Gélin and Primet [2] conducted a comprehensive review of Pt-and Pd-supported catalysts, and reported that many researchers used alumina supports.Simplicio et al. [3] prepared a Pd catalyst on a y-alumina support using the incipient wetness method and reported 50% and 90% conversion of methane at 320 • C and 400 • C, respectively.Yoshida et al. [4] reported 60% methane conversion at 380 • C by a Pd/Al 2 O 3 catalyst.Gholami and Smith [5] prepared Pd/SiO 2 catalysts by the incipient wetness method and reported 50% conversion at 550 • C and 100% at 600 • C. Goodman et al. [6] developed a Pd/γ-Al 2 O 3 catalyst that converted 90% methane at 400 • C, but only 30% at 325 • C. Several researchers have also used other precursors and supports for the combustion of methane.Stefanov et al. [7] reported a Pd-Co/γ-Al 2 O 3 catalyst with 70% methane conversion at 400 • C; Ercolino et al. [8] developed a Pd/Co 3 O 4 catalyst on ZrO 2 open cell foam, and the catalyst converted about 90% methane at 325 • C and 100% at 400 • C. Cargnello et al. [9] used a supramolecular approach to develop a novel Pd@CeO 2 /hydrophobic-Al 2 O 3 core-shell catalyst that converted 50% methane at 325 • C and 100% at 400 • C. A Pd/Ceria nanocatalyst supported on alumina and prepared by the solution combustion method showed 50% activity at 300 • C and 90% at 400 • C [10].A nanozeolite silicalite-1 coating on cordierite catalyst and a nitrogen-modified perovskite-type composite catalyst converted 90% methane at 350 • C [11,12].Zhang et al. [13] reported a Pd catalyst on an alumina-ceria support that converted 80% methane at 320 • C and 100% at 400 • C. In a recent review, Venezia et al. [14] discussed effects of oxides on supported Pd catalysts, support-metal contributions, and distribution of the active sites.
Deactivation by steam has been reported as a major issue with supported Pd catalysts [6,[15][16][17][18] particularly at temperatures above 325 • C.However, Pt/Pd bimetallic catalysts were active for methane combustion in the presence of steam [6].
The nature and identity of the active phase in Pd catalysts during methane combustion is currently under dispute [2,6,7,9,[18][19][20][21]. Several researchers recognize PdO as the active phase [2,6,7,9]; some advocate for Pd [18,19], and others for the Pd to PdO ratio [20,21].Another consideration is the possibility of the support lending its constituents to the catalytic active site during the course of the reaction [20].For example, under some situations, the oxygen from the aluminum oxide could migrate to the palladium, forming PdO or some form of PdO x , and/or migrate to the gas-phase molecule providing oxygen for the combustion [22].This mechanism is underexplored and is one that this research seeks to advance evidence to support or reject.
A review of the literature on the catalytic combustion of methane indicates the development of some promising catalysts.Even though Pd/Al 2 O 3 catalysts have been well-studied for methane combustion, this group of catalysts exhibit poor activities at temperatures below 325 • C. One of the objectives of our research has been to address this issue.We developed a catalyst containing PdO-PdO x on a gamma-alumina support that showed low-temperature activity at 275-325 • C. Another objective was to identify the surface compounds and their role in catalytic activity and the reaction mechanism.We hypothesize that the active phase of the catalyst is Pd native oxide (PdO x ).We also propose a new mechanism for the catalytic reaction at lower temperatures below 325 • C involving the migration of oxide ions from PdO x to adsorbed CH 4 and from the support (γ-Al 2 O 3 ) to PdO to form PdO x .

Catalyst Preparation
The Incipient Wetness Method (IWM) has been used extensively for the preparation of supported Pd catalysts [3,8,20,22].As a part of the preliminary investigations, we prepared a 5% Pd/γ-Al 2 O 3 catalyst by IWM and determined the Pd content (by Energy Dispersive X-ray Spectroscopy (EDS)) and activity for methane combustion.This catalyst had a total Pd content of 3.2%, indicating that the mixing of the precursor (Palladium nitrate hydrate) with the support γ-Al 2 O 3 was not uniform during the small-scale preparation.The catalyst also gave a very low conversion of methane below 400 • C.These preliminary results prompted us to look for a better mixing method while using IWM.The concept of vortex mixing is to improve the dispersion of particles by applying powerful shear flows [23,24].Vortex mixing has been used in catalytic preparation for hydrogen production [23] and for the dispersion of carbon nanotubes [24].We developed a modified Vortex-assisted Incipient Wetness Method (VIWM) for preparation of the catalyst and details of this method are given in the "Materials and Methods" section.The catalyst prepared by VIWM and calcined at 500 • C showed a higher Pd content (4.3%) and much better activities at lower temperatures.The use of vortexing in conjunction with IWM helped in better mixing and dispersion of the PdO/PdO x particles in the support.The catalyst was calcined at three different temperatures (500 • C, 750 • C, and 900 • C) to study the effect of calcination temperature.XPS showed that the surface contained both PdO and PdO x .However, the percentage of PdO x was more at 500 • C and it originated from the interaction of PdO with the γ-Al 2 O 3 support.The change in surface composition was more likely a calcination-surface interaction effect.The "Discussion Section" provides further details.

Scanning Transmission Electron Microscopy (STEM) and EDS Elementary Mapping
Figure 1 shows the Energy Dispersive X-ray Spectra (EDS) of the catalyst.The signals of Ti and Cu in Figure 1 are from the TEM grids and TEM sample holder.The average of the scans resulted in a 0.93 atomic percent of Pd, which corresponded to a 4.3 wt % Pd.The catalyst was prepared with a 5.0 wt % Pd loading.This shows the effectiveness of the modified Vortex-assisted Incipient Wetness method in retaining and dispersing the metal in the supported catalyst.High-angle annular dark-field Scanning transmission electron microscopy (HAADF-STEM) has been used by researchers to determine particle sizes and the particle size distribution of metal/metal oxides and support materials in catalysts [3,6,9,10] Goodman et al. [6] used HAADF-STEM images to determine the crystalline structure and metal dispersion in a Pd/γ-Al 2 O 3 nanocatalyst.Cargnello et al. [9] demonstrated dispersion of Pd@CeO 2 on alumina using HAADF-STEM images and EDS mapping.Using STEM measurements, Khader et al. [10] identified Pd/CeO 2 nanocrystals of 1-50 nm in size.The STEM images and EDS maps of the catalyst are shown in Figures 2-6.HAADF-STEM images (Figure 2a, Figure 3A, Figure 4, and Figure 5a) show PdO/PdO x as bright spots on the surface of γ-Al 2 O 3 crystallites .The PdO/PdO x crystals have an average size of 5-10 nm (Figure 4) and are well-dispersed in the support (Figure 6). Figure 5 gives the contrast HAADF and Bright-field STEM images of PdO/PdO x and γ-Al 2 O 3 ; these images demonstrate that the support is crystalline and contains nanosize particles.Our STEM data on average particle sizes are close to the particle size of 10 nm for a PdO/γ-Al 2 O 3 catalyst [3,22] and 5 nm for a Pd-Co/γ-Al 2 O 3 catalyst [7].
Catalysts 2017, 7, x FOR PEER REVIEW 3 of 18 of calcination temperature.XPS showed that the surface contained both PdO and PdOx.However, the percentage of PdOx was more at 500 °C and it originated from the interaction of PdO with the γ-Al2O3 support.The change in surface composition was more likely a calcination-surface interaction effect.The "Discussion Section" provides further details.

Scanning Transmission Electron Microscopy (STEM) and EDS Elementary Mapping
Figure 1 shows the Energy Dispersive X-ray Spectra (EDS) of the catalyst.The signals of Ti and Cu in Figure 1 are from the TEM grids and TEM sample holder.The average of the scans resulted in a 0.93 atomic percent of Pd, which corresponded to a 4.3 wt % Pd.The catalyst was prepared with a 5.0 wt % Pd loading.This shows the effectiveness of the modified Vortex-assisted Incipient Wetness method in retaining and dispersing the metal in the supported catalyst.High-angle annular darkfield Scanning transmission electron microscopy (HAADF-STEM) has been used by researchers to determine particle sizes and the particle size distribution of metal/metal oxides and support materials in catalysts [3,6,9,10] Goodman et al. [6] used HAADF-STEM images to determine the crystalline structure and metal dispersion in a Pd/γ-Al2O3 nanocatalyst.Cargnello et al. [9] demonstrated dispersion of Pd@CeO2 on alumina using HAADF-STEM images and EDS mapping.Using STEM measurements, Khader et al. [10] identified Pd/CeO2 nanocrystals of 1-50 nm in size.The STEM images and EDS maps of the catalyst are shown in Figures 2-6.HAADF-STEM images (Figures 2a, 3A, 4, and 5a) show PdO/PdOx as bright spots on the surface of γ-Al2O3 crystallites.The PdO/PdOx crystals have an average size of 5-10 nm (Figure 4) and are well-dispersed in the support (Figure 6). Figure 5 gives the contrast HAADF and Bright-field STEM images of PdO/PdOx and γ-Al2O3; these images demonstrate that the support is crystalline and contains nanosize particles.Our STEM data on average particle sizes are close to the particle size of 10 nm for a PdO/γ-Al2O3 catalyst [3,22] and 5 nm for a Pd-Co/γ-Al2O3 catalyst [7].

X-ray Photoelectron Spectroscopy (XPS)
XPS is widely used in catalytic research for identifying surface compounds and their composition [3,7,9,10].An important objective of this research is to find out the change in surface composition of the catalyst before and after methane combustion and to identify the active phases in the catalytic reaction.We used the Pd 3d5/2 peak deconvolution technique to estimate the surface composition of the catalyst.The quality of the fit is good for XPS peaks, particularly with the very small signals of Pd observed and the "roughness" of nanoparticle specimens, which block a significant fraction of the outgoing photoelectron signal (smoother surfaces, such as Si, yield much higher S/N spectra).All of the XPS peak fits reported have chi-squared values <3.0.Further, there was no attempt to "guide" the peak fit to a preconceived surface chemistry.The peak fit was set to float freely in terms of binding energy, peak width, and combination of Gaussian/Lorentzian peak envelopes to model XPS peak shapes.
XPS of the catalyst gave a surface elemental composition of 0.37 atomic % Pd and 1.64 wt % of Pd.The catalyst contained 4.3 wt % Pd based on EDS data.This means that 35.8% of total Pd was on the surface and the rest of the Pd was in the bulk.
The binding energies (BE) for Pd 3d are 335.78eV for Pd native oxide (PdOx); 336.96 eV for PdO; and 338 eV for metallic Pd [25].The vertical lines in the XPS Figure 7 identify the BE locations for standard, well-characterized Pd surface compounds [25].PdOx (Palladium native oxide) is designated as "PdNtv" in this research article.A high-resolution XPS spectrum (Figure 7) of the deconvoluted Pd3d5/2 peak of the catalyst produced two distinct peaks with BE of 337 ev for PdO and 335.8 ev for PdNtv.The XPS data also confirmed the surface composition to be 38% PdO and 62% Pd native oxide.XPS did not show any peak for elemental Pd, thereby confirming the absence of metallic palladium in the catalyst.This result is in contrast to previous research reports on the presence of metallic Pd in Pd/γ-Al2O3 catalysts [18][19][20]26].In addition to metallic Pd and stoichiometric PdO, palladium surfaces exhibit non-stoichiometric native oxides.Most native oxides are not stoichiometric and commonly designated as PdOx.The bulk of a catalyst could be micro-structured with PdO while the surface composition exists as non-stoichiometric (native) palladium oxide [25].The BE for Pd 3d5/2 peaks for a 3% PdO/γ-Al2O3 catalyst were in the range 335.2-336.8ev and assigned to Pd and PdO; a peak with BE of 337.8 ev was due to PdO2 [3] To the best our knowledge, this is the first research report for the presence of both PdO and Pd native oxide (PdOx) on the surface of the

X-ray Photoelectron Spectroscopy (XPS)
XPS is widely used in catalytic research for identifying surface compounds and their composition [3,7,9,10].An important objective of this research is to find out the change in surface composition of the catalyst before and after methane combustion and to identify the active phases in the catalytic reaction.We used the Pd 3d5/2 peak deconvolution technique to estimate the surface composition of the catalyst.The quality of the fit is good for XPS peaks, particularly with the very small signals of Pd observed and the "roughness" of nanoparticle specimens, which block a significant fraction of the outgoing photoelectron signal (smoother surfaces, such as Si, yield much higher S/N spectra).All of the XPS peak fits reported have chi-squared values <3.0.Further, there was no attempt to "guide" the peak fit to a preconceived surface chemistry.The peak fit was set to float freely in terms of binding energy, peak width, and combination of Gaussian/Lorentzian peak envelopes to model XPS peak shapes.
XPS of the catalyst gave a surface elemental composition of 0.37 atomic % Pd and 1.64 wt % of Pd.The catalyst contained 4.3 wt % Pd based on EDS data.This means that 35.8% of total Pd was on the surface and the rest of the Pd was in the bulk.
The binding energies (BE) for Pd 3d are 335.78eV for Pd native oxide (PdO x ); 336.96 eV for PdO; and 338 eV for metallic Pd [25].The vertical lines in the XPS Figure 7 identify the BE locations for standard, well-characterized Pd surface compounds [25].PdO x (Palladium native oxide) is designated as "PdNtv" in this research article.A high-resolution XPS spectrum (Figure 7) of the deconvoluted Pd3d5 /2 peak of the catalyst produced two distinct peaks with BE of 337 ev for PdO and 335.8 ev for PdNtv.The XPS data also confirmed the surface composition to be 38% PdO and 62% Pd native oxide.XPS did not show any peak for elemental Pd, thereby confirming the absence of metallic palladium in the catalyst.This result is in contrast to previous research reports on the presence of metallic Pd in Pd/γ-Al 2 O 3 catalysts [18][19][20]26].In addition to metallic Pd and stoichiometric PdO, palladium surfaces exhibit non-stoichiometric native oxides.Most native oxides are not stoichiometric and commonly designated as PdO x .The bulk of a catalyst could be micro-structured with PdO while the surface composition exists as non-stoichiometric (native) palladium oxide [25].The BE for Pd 3d5/2 peaks for a 3% PdO/γ-Al 2 O 3 catalyst were in the range 335.2-336.8ev and assigned to Pd and PdO; a peak with BE of 337.8 ev was due to PdO 2 [3] To the best our knowledge, this is the first research report for the presence of both PdO and Pd native oxide (PdO x ) on the surface of the supported catalyst before reaction.In-situ XPS of a palladium catalyst on La-modified aluminum oxide confirmed PdO as the main surface compound in 'lean (oxygen rich)' conditions [27].
Catalysts 2017, 7, x FOR PEER REVIEW 7 of 18 supported catalyst before reaction.In-situ XPS of a palladium catalyst on La-modified aluminum oxide confirmed PdO as the main surface compound in 'lean (oxygen rich)' conditions [27].

X-ray Powder Diffraction
Figure 8 gives the XRD patterns of the catalyst, pure PdO, and γ-Al2O3.The important peaks for γ-Al2O3 are the peaks at 46.0° and 67.4° and the identifying peaks for PdO are a strong peak at 34.1° and smaller peaks at 42.2°, 54.9°, 61.1°, and 71.7°.The catalyst gave two weak XRD peaks at 34.1° (strongest peak for PdO) and 46.0° (strongest peak for γ-Al2O3), confirming the presence of PdO and γ-Al2O3.The catalyst as well as the standard PdO also showed an unexpected peak at 26.0°.This is due to graphite contamination originating from the carbon tape.The catalyst was in nanopowder form, and did not stick to the XRD sample plate.Hence, a carbon tape was used as an adhesive, and part of it remained in the sample during XRD.This contamination was much less during sample preparation of PdO; the gamma alumina support was sticking to the plate and hence no carbon tape was needed.Our XRD data were similar to reported values [3,4].The catalyst as well as the standard PdO also showed an unexpected peak at 26.0 • .This is due to graphite contamination originating from the carbon tape.The catalyst was in nanopowder form, and did not stick to the XRD sample plate.Hence, a carbon tape was used as an adhesive, and part of it remained in the sample during XRD.This contamination was much less during sample preparation of PdO; the gamma alumina support was sticking to the plate and hence no carbon tape was needed.Our XRD data were similar to reported values [3,4].

Physisorption and Chemisorption
We used the Brunauer-Emmett-Teller (BET) method of nitrogen physisorption to calculate the total surface area; and Barrett-Joyner-Halenda (BJH) method to estimate pore volume and pore size of the catalyst.CO pulse chemisorption experiments provided data to calculate the dispersion and particle size (Crystallite size) of the catalyst.The calculations were done based on 5% Pd loading and

Physisorption and Chemisorption
We used the Brunauer-Emmett-Teller (BET) method of nitrogen physisorption to calculate the total surface area; and Barrett-Joyner-Halenda (BJH) method to estimate pore volume and pore size of the catalyst.CO pulse chemisorption experiments provided data to calculate the dispersion and particle size (Crystallite size) of the catalyst.The calculations were done based on 5% Pd loading and chemisorption of one molecule of CO on each Pd atom on the surface.Table 1 gives the calculated physical and chemical properties of the catalysts.Our values of dispersion and particle size of the catalyst are similar to the data reported in the literature for Pd/γ-Al 2 O 3 catalysts [3,4].Figure 9 shows the temperature-programmed oxidation (TPO) profile of the Pd/γ-Al 2 O 3 catalyst.The plot is noisy, displaying no definitive oxygen uptake.Miller et al. [20] reported a TPO peak around 286 • C due to the Pd-PdO transition.The more likely explanation of the plot's noise is that the surface already had a high content of PdO prior to the oxidation procedure, and hence oxygen update would be limited.Figure 10 displays the temperature-programmed reduction (TPR) profile for the catalyst.The TPR experiment was conducted without any prior reduction of the catalyst by hydrogen.As stated in the methods section, after pretreatment with helium, hydrogen gas was flown through the catalyst at 30 • C, and then the temperature was raised up to 500 • C. The TCD signals were recorded from 30 • C Hydrogen consumption to reduce PdO/PdO x on the catalyst surface, which should happen at 0-30 • C as reported by Miller et al. [20].However, this peak is missing in the TPD graph since measurements were started only from 30 • C. The TPD showed a peak around 60 • C and this is attributed to the decomposition of palladium hydride [20].

Activity Studies
The activity of the catalyst was tested with two separate reacting gas mixtures: (a) 1% methane, 4% oxygen, and balance helium; (b) 2% methane, 4% oxygen, and balance helium.The first gas mixture resembles a "lean-burn" condition, and the second one is the stoichiometric ratio for the catalytic combustion reaction.Each activity run was performed in a fixed reaction temperature for a reaction period of 20 min and a space velocity of 3.0 × 10 5 cc•g −1 •h −1 .Figure 11 gives the conversions of methane at different temperatures.

Activity Studies
The activity of the catalyst was tested with two separate reacting gas mixtures: (a) 1% methane, 4% oxygen, and balance helium; (b) 2% methane, 4% oxygen, and balance helium.The first gas mixture resembles a "lean-burn" condition, and the second one is the stoichiometric ratio for the catalytic combustion reaction.Each activity run was performed in a fixed reaction temperature for a reaction period of 20 min and a space velocity of 3.0 × 10 5 cc•g −1 •h −1 .Figure 11 gives the conversions of methane at different temperatures.The catalyst initiated combustion at 250 °C and attained 60% conversion at 275 °C and 90% at 300 °C with both gas mixtures.The highest conversion was 95% at 325 °C.All values of conversions were between 90-94% in a narrow temperature range of 300-320 °C and we did not get 100% conversion under the experimental conditions.The reaction became very exothermic and water condensation was visible in the reactor in the narrow temperature range of 325 °C and above.The conversion of methane decreased from 95% to 70% for the 2% methane gas mixture as the temperature increased from 325 °C to 400 °C; the 1% gas mixture had a more pronounced dip in conversion in this temperature range.The catalytic reaction did not produce any carbon monoxide The catalyst initiated combustion at 250 • C and attained 60% conversion at 275 • C and 90% at 300 • C with both gas mixtures.The highest conversion was 95% at 325 • C. All values of conversions were between 90-94% in a narrow temperature range of 300-320 • C and we did not get 100% conversion under the experimental conditions.The reaction became very exothermic and water condensation was visible in the reactor in the narrow temperature range of 325 • C and above.The conversion of methane decreased from 95% to 70% for the 2% methane gas mixture as the temperature increased from 325 • C to 400 • C; the 1% gas mixture had a more pronounced dip in conversion in this temperature range.The catalytic reaction did not produce any carbon monoxide with the 1% and 2% methane gas mixtures.
An activity run with pure PdO was conducted in the catalytic reactor using a gas mixture of 1% methane, 4% oxygen, and balance helium.Methane conversion was 60% at 230 • C and 100% at 260 • C. In comparison, the 5% Pd/γ-Al 2 O 3 catalyst could convert only 30% methane at 260 • C.
Table 2 gives a brief review of some promising catalysts developed in recent years.These catalysts showed high activities for methane conversion of 90% or more at 400 • C [3,[7][8][9]11,28,29].These catalysts have been developed using different preparation methods, such as Pd loading, and tested under different reaction conditions.Hence, it is not feasible to compare the relative effectiveness of these catalysts.However, among these catalysts, only two exhibited excellent low-temperature activity with a methane conversion of 90% and above at 325 • C. A Pd/Co 3 O 4 on a ZrO 2 foam catalyst converted 90% of methane [8].Our PdO-PdO x catalyst converted 95% methane at 325 • C. The high methane conversion at low temperature is partly attributed to the vortex method providing better dispersion.
The efficiency and stability of the catalyst were tested on a preliminary basis by carrying out the catalytic reaction for six catalytic cycles under the best conversion conditions (at 320 • C for 20 min with 2% methane, 4% oxygen gas mixture, and a space velocity of 3.0 × 10 5 cc•g −1 •h −1 ).The data in Table 3 show the catalyst was efficient in maintaining high conversion (90-94%) and was stable without much change in the surface composition after a reaction period of 2 h.The thermal stability of the catalyst was also tested by heating the catalyst up to 550 • C for 2 h in a flow of helium.XPS of the catalyst after the thermal treatment in a helium atmosphere did not show any change in the surface composition.XPS (Figure 12) of the catalyst after five reaction cycles indicated no change in the surface Pd content.However, there was a small change in the surface composition of the catalyst (PdO, 49%; PdNtv oxide, 51%) compared to the original composition of the catalyst before reaction (PdO, 38%; PdNtv oxide, 62%).
The thermal stability of the catalyst was also tested by heating the catalyst up to 550 °C for 2 h in a flow of helium.XPS of the catalyst after the thermal treatment in a helium atmosphere did not show any change in the surface composition.XPS (Figure 12) of the catalyst after five reaction cycles indicated no change in the surface Pd content.However, there was a small change in the surface composition of the catalyst (PdO, 49%; PdNtv oxide, 51%) compared to the original composition of the catalyst before reaction (PdO, 38%; PdNtv oxide, 62%).Many researchers reported deactivation of catalysts by steam at higher temperatures (above 400 °C) resulting in sintering and or poisoning [6,15,16].At temperatures above 325 °C, the conversion decreased, but the surface Pd content did not change (Table 4).The catalyst after reaction with 2% Many researchers reported deactivation of catalysts by steam at higher temperatures (above 400 • C) resulting in sintering and or poisoning [6,15,16].At temperatures above 325 • C, the conversion decreased, but the surface Pd content did not change (Table 4).The catalyst after reaction with 2% methane at 320 • C did not show any loss of surface Pd.This indicates that the catalyst was stable at lower temperatures up to 320 • C and was not deactivated by steam.It appears that deactivation by steam may not be the only reason for the decrease in catalytic activity at 325-450 • C. Catalyst before reaction (Figure 7) 1.64 38 62 Catalyst after reaction at 325 • C (Figure 12) 1.62 49 51 Catalyst after reaction at 450 • C (Figure 14) 1.62 68 32 Pd nitrate calcined at 500 • C (Figure 13) -100 0 Pure PdO before and after reaction -100 0

Active Phase of the Catalyst and Proposed Mechanism
Disputes exist on the active phases in methane catalytic combustion.Some researchers reported PdO as the active Pd species [2,6,7,9,27].However, others attribute the highest catalytic activity to the Pd metallic state in methane-rich conditions [18,19,27] or to the ratio of Pd to PdO [20,21].Another consideration is the possibility of the support itself lending its constituents to the catalytic active site during the course of the reaction [20].For example, under some situations, oxygen from the aluminum oxide support could migrate to the palladium, forming PdO or some form of PdO x , and/or migrate to the gas-phase molecule providing oxygen for the combustion [5,22].
In order to explore the support effect, we did XPS of standard PdO (Sigma Aldrich, St. Louis, MO, USA) before and after reaction with 2% methane and 4% oxygen gas mixture at 325 • C. The best-fit data indicated the presence of only 100% PdO with a binding energy of 336.96 ev in both XPS spectra.This indicates that PdO remained as the active phase in the catalytic reaction of methane with pure PdO without the support.Pd (II) nitrate was the precursor for the catalyst.Since the catalyst was prepared by calcination in air at 500 • C, we treated the precursor under the same conditions by calcining at 500 • C and then did the XPS of the calcined Pd (II) nitrate (Figure 13).
The best-fit data indicated the presence of 100% Pd in the calcined Pd (II) nitrate with a binding energy of 336.96 ev.XPS (Figure 7) of the supported catalyst confirmed the presence of PdO and Pd native oxides (PdO x ) and that the native oxide formed only in the presence of the support.The XPS of calcined Pd nitrate showed only PdO and no native oxide.This confirms the supposition that the γ-Al 2 O 3 support facilitated oxygen migration from the support to the stoichiometric PdO to form Pd native oxide (PdO x ).Other researchers reported the presence of Pd and/or PdO in Pd/γ-Al 2 O 3 catalysts [2,6,7,9,18,19].Miller et al. [20] and Schwartz et al. [22] suggested that the support could facilitate oxygen mobility and affect composition of the surface catalytic sites at temperatures <400 • C.  In order to identify the active phases in the catalytic combustion process, we did XPS of the catalyst after reaction with 2% CH4/4% O2 at 450 °C for 20 min.XPS of the deconvoluted 3d5/2 peak gave a surface composition of 68% PdO and 32% Pd native oxide (Figure 14).Table 4 summarizes the surface composition at various reaction conditions based on XPS data.The evidences from XPS suggest that the surface composition of the catalyst changed after the reaction and the surface contained more PdO than Pd native oxide (PdOx).In order to identify the active phases in the catalytic combustion process, we did XPS of the catalyst after reaction with 2% CH 4 /4% O 2 at 450 • C for 20 min.XPS of the deconvoluted 3d 5/2 peak gave a surface composition of 68% PdO and 32% Pd native oxide (Figure 14).Table 4 summarizes the surface composition at various reaction conditions based on XPS data.The evidences from XPS suggest that the surface composition of the catalyst changed after the reaction and the surface contained more PdO than Pd native oxide (PdO x ).In order to identify the active phases in the catalytic combustion process, we did XPS of the catalyst after reaction with 2% CH4/4% O2 at 450 °C for 20 min.XPS of the deconvoluted 3d5/2 peak gave a surface composition of 68% PdO and 32% Pd native oxide (Figure 14).Table 4 summarizes the surface composition at various reaction conditions based on XPS data.The evidences from XPS suggest that the surface composition of the catalyst changed after the reaction and the surface contained more PdO than Pd native oxide (PdOx).(ii) During the reaction at lower temperatures below 325 • C, oxide ions from PdO x migrate to react with methane; PdO x forms PdO; and oxide ions from the support (γ-Al 2 O 3 ) migrate to PdO to form PdO x .This process maintains the surface composition of both PdO and PdO x ; the catalytic activity remains more or less steady since the active phase of PdO x does not deplete.(iii) At a higher temperature when catalytic activity decreases in the temperature range 325-450 • C, oxide ions from PdOx migrate to react with methane; and PdO x forms PdO.However, due to unknown reasons (may be due to deactivation by steam) the support does not provide oxide ions to PdO and thereby the number of active sites of PdO x decreases and hence activity decreases.This hypothesis supports the correlation between decrease in catalytic activity and depletion of PdO at temperatures in the range 325-400 • C as shown by XPS.
Based on isotope exchange evidences, Schwartz and Pfefferle [22] suggested that at temperatures below 350 • C, any oxygen exchange in reaction is due to oxygen from the support.Therefore, our results and suggested mechanism are in agreement with these researchers.

Catalyst Synthesis
We developed a modified Vortex-assisted Incipient Wetness Method to improve mixing and dispersion of the catalyst precursor in the support.The vortex unit is an Analog Vortex Mixer with Single Tube Holder (Fisher Scientific; Catalog number 02-215-430, Asheville, NC, USA).The vortex unit is given in Figure 15.The vortex mixer (bottom part) is attached to a tube holder that holds and secures a vortex tube.The solid support is taken in the vortex tube and the catalyst precursor solution is added to the support drop-by-drop.An analog controller controls the speed (0-3000 rpm), which allows for high-speed mixing for vigorous vortexing of samples.
Catalysts 2017, 7, x FOR PEER REVIEW 14 of 18 decreases.This hypothesis supports the correlation between decrease in catalytic activity and depletion of PdOx at temperatures in the range 325-400 °C as shown by XPS.
Based on isotope exchange evidences, Schwartz and Pfefferle [22] suggested that at temperatures below 350 °C, any oxygen exchange in reaction is due to oxygen from the support.Therefore, our results and suggested mechanism are in agreement with these researchers.

Catalyst Synthesis
We developed a modified Vortex-assisted Incipient Wetness Method to improve mixing and dispersion of the catalyst precursor in the support.The vortex unit is an Analog Vortex Mixer with Single Tube Holder (Fisher Scientific; Catalog number 02-215-430, Asheville, NC, USA).The vortex unit is given in Figure 15.The vortex mixer (bottom part) is attached to a tube holder that holds and secures a vortex tube.The solid support is taken in the vortex tube and the catalyst precursor solution is added to the support drop-by-drop.An analog controller controls the speed (0-3000 rpm), which allows for high-speed mixing for vigorous vortexing of samples.A modified Vortex-assisted Incipient Wetness Method was developed to prepare a 5 wt % Pd/γ-Al2O3 catalyst.A solution of 3.0 mL deionized (DI) water and 0.328 g of palladium(II) nitrate hydrate (Sigma Aldrich, MW: 230.43 g/mol; anhydrous basis) was added in 20 μL increments to 2.702 g of gamma phase aluminum oxide nanopowder (<50 nm particle size, Sigma Aldrich, St. Louis, MO, USA) under vortex mixing set to 900 rpm for 2 h.The resulting "ball" of dough-like mixture was dried in an oven at 103 °C overnight, crushed with a mortar and pestle, and then calcined at 500 °C for 6 h in a Thermolyne Benchtop Muffle Furnace.This catalyst was then passed through a sieve to A modified Vortex-assisted Incipient Wetness Method was developed to prepare a 5 wt % Pd/γ-Al 2 O 3 catalyst.A solution of 3.0 mL deionized (DI) water and 0.328 g of palladium(II) nitrate hydrate (Sigma Aldrich, MW: 230.43 g/mol; anhydrous basis) was added in 20 µL increments to 2.702 g of gamma phase aluminum oxide nanopowder (<50 nm particle size, Sigma Aldrich, St. Louis, MO, USA) under vortex mixing set to 900 rpm for 2 h.The resulting "ball" of dough-like was dried in an oven at 103 • C overnight, crushed with a mortar and pestle, and then calcined at 500 • C for 6 h in a Thermolyne Benchtop Muffle Furnace.This catalyst was then passed through a sieve to achieve a particle size of <150 µm.
Palladium oxide (99.97% trace metals basis, Sigma Aldrich, St. Louis, MO, USA) was used as a standard for the XPS, XRD, and catalytic activity tests.
XPS measurements were conducted in a load-locked Kratos XSAM 800 (Kratos Analytic Ltd., Manchester, U.K.) surface analysis system equipped with a hemispherical energy analyzer and a MgK α (1253.6 eV) radiation source and following a procedure reported earlier [30].The C1s binding energy of 285.0 eV was used as the reference.The XPS data reported were run under two instrumental setups.For survey spectra (broad-spectrum cases), we scanned 20 times at an energy step of 0.5 eV.For the higher-resolution spectra where peak deconvolution was performed, we scanned only over the peak of interest with a varying number of scans (determined by the peak size) with a higher energy resolution step of 0.1 eV.All other instrumental parameters (X-ray flux, tilt angle, etc.) were the same.
X-ray analysis of the catalyst, standard PdO, and standard γ-Al 2 O 3 was conducted in a Bruker AXS D2 Phaser using a Cu radiation source at 30 kV and 10 mA.Samples were irradiated at a 2 range from 15 to 90 • at a speed of 2.4 • /min.Samples were either applied to a zero-plate or to double-sided tape to ensure even application to the XRD slide.
Chemisorption measurements were performed in a Micromeritics AutoChem II 2920 (Micromeritics Corporation, Norcross, GA, USA) and details of the procedure have been reported [20,31].For the pulse chemisorption experiment, the catalyst was first reduced in hydrogen at 450 • C and then pulses of CO were injected.Dispersion and particle size (Crystallite size) of the catalyst were calculated from the CO uptake, wt % of Pd loading, and assuming a 1:1 ratio between the CO-adsorbed and Pd-active site.The temperature-programmed reduction (TPR) experiment involved pretreatment by helium at 110 • C followed by a flow of 10% H 2 /Ar at 30 • C and then continuing heating up to 500 • C in a flow of 10% H 2 /Ar.For the TPO experiment, the TPR procedure was followed except using 10% O 2 /He instead of H 2 /Ar.The physisorption experiment was performed in a Micromeritics TriStar II 3020 3.02 (Micromeritics Corporation, Norcross, GA, USA) surface and porosity analyzer using nitrogen gas as adsorbate at minus 198.8 • C. Surface area was calculated using the BET method; the pore volume from the BJH adsorption cumulative volume of pores between 17,000 Å and 3,000,000 Å width; and the pore size from the BJH adsorption average pore width (4V/A).
HAADF-STEM images and EDS elementary mapping were obtained in a Hitachi HD2700 (Hitachi Ltd., Tokyo, Japan) aberration-corrected scanning transmission electron microscope.The electron beam convergent angle α was ~27 mrad and the HAADF detector collection angle β = 70-370 mrad.

Activity Measurements
The catalytic activity measurements were performed in a temperature-controlled tubular fixed-bed quartz reactor.A catalyst bed of 0.200 g was secured in the middle of the reactor by quartz wool/beads.Helium gas (1.0 L/min) was passed through the catalytic reactor and the reactor temperature was raised to a desired value by a temperature-controlled furnace.The temperature of the reaction bed

Figure 1 .
Figure 1.EDS spectra of the catalyst.Figure 1. EDS spectra of the catalyst.

Figure 1 .
Figure 1.EDS spectra of the catalyst.Figure 1. EDS spectra of the catalyst.

Figure 3 .
Figure 3. (A) HAADF STEM image from the same area of EDS mapping; (B) SEM image from the same area of EDS mapping.

Figure 3 .
Figure 3. (A) HAADF STEM image from the same area of EDS mapping; (B) SEM image from the same area of EDS mapping.

Figure 3 .
Figure 3. (A) HAADF STEM image from the same area of EDS mapping; (B) SEM image from the same area of EDS mapping.

Figure 5 .
Figure 5. (a) z-contrast HAADF STEM image of PdOx and γ-Al2O3; (b) phase-contrast Bright-field STEM image of PdOx and γ-Al2O3.Both images are from the same area; the γ-Al2O3 particles are crystalline, not in amorphous phase.

Figure 5 .
Figure 5. (a) z-contrast HAADF STEM image of PdOx and γ-Al2O3; (b) phase-contrast Bright-field STEM image of PdOx and γ-Al2O3.Both images are from the same area; the γ-Al2O3 particles are crystalline, not in amorphous phase.

Figure 5 .
Figure 5. (a) z-contrast HAADF STEM image of PdO x and γ-Al 2 O 3 ; (b) phase-contrast Bright-field STEM image of PdO x and γ-Al 2 O 3 .Both images are from the same area; the γ-Al 2 O 3 particles are crystalline, not in amorphous phase.

Figure 8
Figure 8 gives the XRD patterns of the catalyst, pure PdO, and γ-Al 2 O 3 .The important peaks for γ-Al 2 O 3 are the peaks at 46.0 • and 67.4 • and the identifying peaks for PdO are a strong peak at 34.1 • and smaller peaks at 42.2 • , 54.9 • , 61.1 • , and 71.7 • .The catalyst gave two weak XRD peaks at 34.1 • (strongest peak for PdO) and 46.0 • (strongest peak for γ-Al 2 O 3 ), confirming the presence of PdO and γ-Al 2 O 3 .The catalyst as well as the standard PdO also showed an unexpected peak at 26.0 • .This is due to graphite contamination originating from the carbon tape.The catalyst was in nanopowder form, and did not stick to the XRD sample plate.Hence, a carbon tape was used as an adhesive, and part of it remained in the sample during XRD.This contamination was much less during sample preparation of PdO; the gamma alumina support was sticking to the plate and hence no carbon tape was needed.Our XRD data were similar to reported values[3,4].

18 Figure 8 .
Figure 8. XRD of the catalyst, PdO, and the support.

Figure 8 .
Figure 8. XRD of the catalyst, PdO, and the support.

Figure 11 .
Figure 11.Activity of the catalyst at different temperatures.

Figure 14 .
Figure 14.XPS (deconvoluted Pd3d5/2 peak) of catalyst after reaction at 450 °C.We are proposing the following hypotheses and tentative mechanism based on the experimental data: (i) The active phase of the catalyst is Pd native oxide (PdOx); during the catalyst preparation and calcination process, the oxide ions from the support, γ-Al2O3, migrate to PdO to form PdOx.This mechanism supports the presence of both PdO and PdOx in the catalyst before reaction.(ii) During the reaction at lower temperatures below 325 °C, oxide ions from PdOx migrate to react with methane; PdOx forms PdO; and oxide ions from the support (γ-Al2O3) migrate to PdO to form PdOx.This process maintains the surface composition of both PdO and PdOx; the catalytic activity remains more or less steady since the active phase of PdOx does not deplete.(iii) At a higher temperature when catalytic activity decreases in the temperature range 325-450 °C, oxide ions from PdOx migrate to react with methane; and PdOx forms PdO.However, due to unknown reasons (may be due to deactivation by steam) the support does not provide oxide ions to PdO and thereby the number of active sites of PdOx decreases and hence activity

Figure 14 .
Figure 14.XPS (deconvoluted Pd3d5/2 peak) of catalyst after reaction at 450 °C.We are proposing the following hypotheses and tentative mechanism based on the experimental data: (i) The active phase of the catalyst is Pd native oxide (PdOx); during the catalyst preparation and calcination process, the oxide ions from the support, γ-Al2O3, migrate to PdO to form PdOx.This mechanism supports the presence of both PdO and PdOx in the catalyst before reaction.(ii) During the reaction at lower temperatures below 325 °C, oxide ions from PdOx migrate to react with methane; PdOx forms PdO; and oxide ions from the support (γ-Al2O3) migrate to PdO to form PdOx.This process maintains the surface composition of both PdO and PdOx; the catalytic activity remains more or less steady since the active phase of PdOx does not deplete.(iii) At a higher temperature when catalytic activity decreases in the temperature range 325-450 °C, oxide ions from PdOx migrate to react with methane; and PdOx forms PdO.However, due to unknown reasons (may be due to deactivation by steam) the support does not provide oxide

Table 1 .
Properties of the catalyst.

Table 2 .
Review of recent literature on catalytic activity for methane combustion.
[a] This research report.

Table 3 .
Methane conversion in six catalytic cycles.

Table 4 .
Change in surface composition with reaction conditions