Probing the Role of a Non‐Thermal Plasma (NTP) in the Hybrid NTP Catalytic Oxidation of Methane

Abstract Three recurring hypotheses are often used to explain the effect of non‐thermal plasmas (NTPs) on NTP catalytic hybrid reactions; namely, modification or heating of the catalyst or creation of new reaction pathways by plasma‐produced species. NTP‐assisted methane (CH4) oxidation over Pd/Al2O3 was investigated by direct monitoring of the X‐ray absorption fine structure of the catalyst, coupled with end‐of‐pipe mass spectrometry. This in situ study revealed that the catalyst did not undergo any significant structural changes under NTP conditions. However, the NTP did lead to an increase in the temperature of the Pd nanoparticles; although this temperature rise was insufficient to activate the thermal CH4 oxidation reaction. The contribution of a lower activation barrier alternative reaction pathway involving the formation of CH3(g) from electron impact reactions is proposed.

Methane is am ajor contributor to climate change,w ith ag lobal warming potential at least 21 times higher than that of CO 2 ;consequently,its release into the atmosphere must be stringently controlled. In addition to controlling CH 4 release from landfills,biomass burning, and leakage from natural gas storage and distribution, emission abatement arising from CH 4 slip in automotive vehicles must also be addressed.
One solution is to use catalytic total oxidation to produce CO 2 and water. Palladium is ak nown efficient catalyst for CH 4 oxidation and has been studied extensively.V arying hypotheses of the active phase have been reported, from aP dO-like phase to Pd 0 . [1] Unfortunately,o fa ll the catalysts currently reported, none are sufficiently active under coldstart conditions,w ith most catalysts requiring light-off temperatures of around 400 8 8C. [2] Such high temperatures are required because of the high activation barriers to CH 4 dehydrogenation;particularly formation of surface adsorbed CH 3 *a nd H*, which is thought to be the rate-determining step.F or example,J ørgensen and Grçnbeck predicted that the extraction of the first Hfrom CH 4 had activation barriers of 0.99 and 0.79 eV over Pd(111) and Pd(100), respectively. [3] One known method for inducing catalytic activity in kinetically restricted reactions at low temperatures is by coupling non-thermal plasmas (NTPs) with catalysis.Recent examples are the selective catalytic reduction of NO x ,v olatile organic carbon (VOC) removal, and water gas shift without the need for an external heating source.S imilarly,N TP-assisted CH 4 oxidation has been reported at low temperature,w here no additional heat source is applied, and at elevated temperatures,where the catalyst is also heated to temperatures up to 300 8 8C. [4] Three recurring hypotheses are often proposed to explain the assistance which NTP gives to catalytic reactions:1 )the plasma modifies the catalyst, 2) the plasma heats the catalyst, and 3) the assistance of the plasma permits occurrence of new reaction pathways.N TP has been shown, in some cases,t o alter the catalysts itself by changing the oxidation state [5] or metal surface area [6,7] of the components.S everal attempts have been made to determine the temperature of ac atalyst during NTP reactions,using thermocouples placed near or in the catalyst bed, or by observation with infrared cameras or optical emission spectroscopy. [8] However,n os tudy has yet directly measured the temperature of the metal nanoparticles within the catalyst and compared this with the overall bed temperature during aN TP-assisted catalytic reaction. The interaction of radicals,electrons,orphotons produced by the NTP with the catalyst and the adsorbed molecules may open up new reaction pathways.For instance,the direct reaction of gas-phase radicals with adsorbed species (that is,adirect Eley-Rideal mechanism) could occur. [9] Desorption from the catalyst surface may also be aided by electron impact. [10] To investigate which of these hypotheses are operating under NTP-assisted catalysis,insitu investigations are crucial. Very few in situ plasma catalytic studies have been performed. [11] One recent example involved investigation of the hydrocarbon selective catalytic reduction (HC-SCR) of NO x over Ag/Al 2 O 3 . [12] This investigation used am odified diffuse reflectance infrared Fourier transform (DRIFTS) setup to follow the adsorbates during the thermal and the NTPenhanced reactions;p roviding invaluable information on adsorbed species and the mechanism of the reaction. However,tothe best of our knowledge,noinsitu structural studies have been undertaken to characterize the catalyst during NTP-assisted reactions.Inthe study reported herein, we have probed NTP-assisted CH 4 oxidation, in the absence of any applied external heating, using in situ X-ray absorption fine structure (XAFS) information. Theresults provide significant new insights into the role of plasma-induced heating effects in the NTP-assisted process.
TheNTP-activated oxidation of CH 4 over a2%Pd/Al 2 O 3 (sample 1; Supporting Information, Table S1) was examined at applied voltages of 6and 7kVat22.5 kHz. In the absence of an externally applied heating source,55% (6 kV) and 67 % (7 kV) CH 4 conversions were observed. In the presence of the catalyst, high selectivities were found (CO:CO 2 ,1:10.8 (6 kV) and 1:11.5 (7 kV)) and negligible amounts of H 2 were formed. Additionally,atemperature-programmed oxidation measurement of the catalyst following CH 4 oxidation showed little carbon deposition with no significant oxidation above > 600 8 8C( Supporting Information, Figure S1). Notably,i n the absence of the catalyst, although the conversion was similar (68 %a t7kV), significantly more CO was formed (CO:CO 2 ,1:1.1). These results may also be compared with the reaction over the discrete support in the presence of the plasma, which had ar educed conversion of 59 %a nd aC O:CO 2 ratio of 1:2.1. These observations demonstrate the importance of the catalyst in determining the selectivity of the reaction. Thes pecific energy input (SEI) calculated for the 6a nd 7kVp lasma in the presence of the catalyst was 2.133 kJ L À1 and 2.637 kJ L À1 ,respectively.
Similar conversions/selectivities were also obtained during the X-ray absorption spectroscopy (XAS) investigation monitored by end-of-pipe mass spectrometry (MS) analysis (sample 1, 52 %c onversion at 6kV; Supporting Information, Figure S2). Thes etup,s hown in Figure 1a nd Figure S3 (Supporting Information), allowed XAFS measurements along the length of the packed catalyst bed to be monitored. XAFS measurements were performed at the Pd K-edge on the B18 beamline at the Diamond Light Source, United Kingdom.
TheX -ray absorption near-edge structure (XANES) of the fresh catalyst (sample 1) under thermal CH 4 oxidation reaction (352 8 8C) compared to that under CH 4 oxidation using the plasma are very similar, as shown in Figure 2. On first inspection, there is little impact of the plasma on the catalyst, with the spectra closely resembling that of PdO.Whenoxygen is removed from the system under plasma, leaving just 5000 ppm CH 4 and He balance,t he catalyst is reduced, as shown by ashift in the edge position of 2eV, and the spectra are very similar to that of the Pd foil reference.Analysis of the extended X-ray absorption fine structure (EXAFS) region reveals only very subtle differences in the spectra when comparing the plasma-activated and thermally activated catalysts under CH 4 oxidation conditions.S pectra collected at two positions,2.5 and 7.5 mm from the start of the catalyst bed, are shown in Figure 3a nd the fitting parameters are shown in Table 1. In both cases the oscillations are dampened  when the plasma is on compared to when it is off.N oo ther differences are observed in the spectra;f or example no shift in phase or additional features are observed that would indicate changes in distance to nearest neighbors or changes in the coordination around the absorber atom. Furthermore, on turning off the plasma, no change was found compared with the fresh catalyst (Supporting Information, Figure S4), demonstrating the reversibility of the changes and indicating that no significant permanent changes to the nanoparticle structure had occurred. Notably,X AFS is ab ulk-averaging technique;t herefore,w ec annot exclude that some minor non-reversible changes to the Pd nanoparticles may occur, which are below the detection level of the technique. However,the EXAFS results are consistent with transmission electron micrographs (TEM) of the catalyst before and after the plasma treatment, which showed similar particle sizes 2.1 AE 1.0 and 2.9 AE 1.4 nm and no change in the shape of the nanoparticles observed (Supporting Information, Figure S5).
We propose that the subtle dampening of the oscillations is due to an increase in the temperature of the Pd under the application of the plasma, when an increase in the meansquared thermal disorder parameter (s 2 )c orresponds to ad ecrease in the amplitude of the EXAFS oscillations. Similarly,weak oscillations were also observed for arange of catalyst loadings and particle sizes (samples 1-4;S upporting Information, Table S1) under plasma conditions.Asnoother measurable changes occur to the catalyst, the change in s 2 can be used to estimate the temperature of the Pd nanoparticles on the catalyst. Ad ata set was obtained on cooling of the catalyst from 500 8 8Ct or oom temperature under air,w hich indicated no structural changes on cooling and was used to determine ac alibration curve of the variation of s 2 with temperature.
Thecalibration curve and fitting parameters are shown in the Supporting Information, Figures S6-S9 and Tables S2 and S3. These data were then used to determine the temperature of the catalysts when activated by the plasma. To fit the data upon cooling, the EXAFS fit was performed, allowing the coordination numbers (CN), s 2 values,and distances to refine. Thed etermined CN values were then fixed when fitting the plasma-activated spectrum, allowing only s 2 and distances to refine.T his value was then used to estimate the temperature of the catalyst when activated by the plasma (Supporting Information, Table S4). Fora ll the catalysts studied the estimated temperatures ranged from 138 to 179 8 8Ci nt he middle of the bed. Additionally,measurements were made for sample 1comparing the front and middle of the bed, and the temperatures were determined to be 203 AE 32 and 152 AE 28 8 8C, respectively.T he fitting parameters (Table 1) and the data and fit are shown in Figure S7 (Supporting Information). The higher temperature at the front of the bed is expected as the CH 4 oxidation is exothermic and the rate will be highest towards the start of the catalyst bed. Using Aspen software (Aspen Technology), as imulation of the reaction using the same reactant concentrations and assuming thermodynamic equilibrium (that is,f ull CH 4 conversion) provided ar eactor temperature of approximately 210 8 8C, which is consistent with the exothermicity of the CH 4 oxidation.
This measurement is not surprising as the IR camera predominantly measures the outside wall of the reactor and will significantly underestimate the temperature within the packed bed.
To determine if the observed heating of the catalyst was because of the exothermicity of the CH 4 oxidation reaction or induced by the plasma, the reaction was performed in the absence of O 2 .Under these conditions the reaction of CH 4 is endothermic and results in coupling products. [13] During the endothermic non-oxidative CH 4 coupling,t he EXAFS data ( Figure 2, Table 1) obtained when the plasma is both on and off resembles that of the Pd foil;t herefore,t he PdO catalyst has been, unsurprisingly,reduced on removal of oxygen from the feed gas.U sing the value of s 2 calculated from this data, the estimated temperature during the plasma was 207 AE 32 8 8C (Table 1). From these results we conclude that the plasma is responsible for the observed heating effect.
Thefact that there are no significant nanoparticle size-or catalyst loading-dependent changes on the temperatures calculated from the XAFS may suggest that the surface of the whole catalyst (nanoparticle and oxide) is being heated. TheXAFS data only probes the Pd, which, however, does not preclude the alumina surface from also being heated. In this case,t he support and nanoparticle would be in thermal equilibrium, thereby leading to similar changes in temperature for all the catalysts studied.
Taking account of all the data acquired, the estimated temperature of the nanoparticles during NTP-activated CH 4 oxidation is 162 AE 24 8 8C, which is within the error of the calculated temperature from the exothermicity of the CH 4 oxidation reaction. Therefore,aclear difference in temperature is observed between the EXAFS estimation and that measured by the IR camera. Almost at wo-fold increase in temperature of the nanoparticles is measured compared to the overall temperature of the catalyst bed. However,t his temperature (162 8 8C) is not high enough to activate the thermal reaction, as observed from the light-off curve of the thermal reaction for sample 1( Figure 5).
In summary,t his in situ study has provided evidence for the role of NTP in the hybrid NTP catalytic oxidation of CH 4 . Herein, it is clear that no significant structural changes are found within the catalyst on application of NTP under CH 4 and CH 4 + O 2 conditions.Additionally,the NTP heats the Pd nanoparticles but the temperature of the nanoparticles is lower than that necessary to activate the thermal CH 4 oxidation reaction. Therefore,i ti sl ikely that an alternative CH 4 activation pathway is in operation, with al ower activation barrier than the thermal activation reaction. As noted, the rate-limiting step for the thermal reaction over Pd is the formation of CH 3(a) + H (a) .T his is found above 227 8 8C, whereas carbon oxidation is rate limiting below 227 8 8C. [3] Interestingly,t he major effect of the plasma on CH 4 has recently been reported to be the formation of CH 3 (g) by electron impact reactions. [14] Given the fact the nanoparticle temperatures are approximately at the transition point where CH 4 activation becomes rate limiting, it is likely that the plasma activation of CH 4 in the gas phase then leads to ar educed activation barrier for the surface process and thus the ability of the NTP process to occur at much reduced temperatures.I tc annot be discounted that the Pd nanoparticles may become more defective in the presence of the plasma and more open faces have been reported to offer al ower activation barrier for CH 4 dehydrogenation. [3] However,t his effect is likely to be small compared with the preactivation of CH 4 in the gas phase under plasma conditions. Figure 4. Temperatureprofile along the catalyst bed (x direction) using an IR camera;the units of x are arbitrary and are dependent on the camera focus. Figure 5. Light-off profile 5000 ppm CH 4 ,5%O 2 ,5%Ar, and He balance.