Plasma-Induced Catalyst Support Defects for the Photothermal Methanation of Carbon Dioxide

The presence of defects in a catalyst support is known to benefit catalytic activity. In this work, a He-plasma treatment-based strategy for introducing and stabilising defects on a Ni/TiO2 catalyst for photothermal CO2 hydrogenation was established. The impact of pretreatment step sequence—which comprised He-plasma treatment and reduction/passivation—on defect generation and stabilisation within the support was evaluated. Characterisation of the Ni/TiO2 catalysts indicated that defects created in the TiO2 support during the initial plasma treatment stage were then stabilised by the reduction/passivation process, (P-R)Ni/TiO2. Conversely, performing reduction/passivation first, (R-P)Ni/TiO2, invoked a resistance to subsequent defect formation upon plasma treatment and consequently, poorer photothermal catalytic activity. The plasma treatment altered the metal-support interaction and ease of catalyst reduction. Under photothermal conditions, (P-R)Ni/TiO2 reached the highest methane production in 75 min, while (R-P)Ni/TiO2 required 165 min. Decoupling the impacts of light and heat indicated thermal dominance of the reaction with CO2 conversion observed from 200 °C onwards. Methane was the primary product with carbon monoxide detected at 350 °C (~2%) and 400 °C (~5%). Overall, the findings demonstrate the importance of pretreatment step sequence when utilising plasma treatment to generate active defect sites in a catalyst support.


Introduction
The transformation of carbon dioxide (CO 2 ) to valuable hydrocarbon products is a promising approach to mitigate CO 2 emissions to the atmosphere [1,2]. Amongst the fuels and chemicals which can be produced, methane is a desirable product as it can be used directly within our existing natural gas infrastructure [3,4]. One approach to driving the methanation reaction is through exploiting photothermal conditions, which relies on a temperature increase from light absorption by the catalyst. The photothermal conditions allow the reaction to occur without external heating and promote a sustainable approach for catalytic activity.
To date, studies surrounding photothermal CO 2 methanation have focused on understanding the role of light and catalyst design, encompassing different active metals and catalysts [5][6][7]. Amongst the active metals studied for thermal CO 2 hydrogenation, nickel (Ni) has been probed expansively due to its performance in CO 2 hydrogenation and relative abundance [3,[8][9][10][11]. With regard to catalyst supports, titania (TiO 2 ) has been studied comprehensively for thermal CO 2 methanation [9,12,13]. TiO 2 , a semiconductor support, can harness light primarily in the ultraviolet (UV) region [14]. While catalysts have been prepared with varied metals and supports, the methods of catalyst modification, such as to induce defects, have not been reported extensively. The photothermal reaction is influenced by factors such as light harnessing, ease of catalyst reduction, and interaction with CO 2 and H 2 [15][16][17]. Therefore, pretreatments to alter catalyst structure can enhance the reaction.
One approach to modifying catalytic properties is by pretreating the catalyst [18][19][20][21]. Plasma treatment is capable of rapidly form defects without requiring additional chemicals or complex reactor set-ups. Non-thermal plasma pretreatment induce defects by bombarding the catalyst surface with excited species, typically easily ionised gases such as argon (Ar) or helium (He) [22,23]. This surface bombardment can result in a range of physical and chemical changes to the catalyst surface including etching, reducing the surface and altering metal deposit size and support interaction [22,23]. Plasma pretreatment has shown significant impacts on both metal oxide supports alone, as well as metal-supported catalysts. For example, Horlyck et al. examined the use of He plasma pretreatment on TiO 2 /SiO 2 composites and found a significant increase in defect formation as a result of prolonged plasma treatment [24]. In the case of metal-supported catalysts, the effect of cold Ar plasma treatment on Pt/CeO 2 for the water-gas shift reaction has been studied. Compared to calcined Pt/CeO 2 , the plasma treatment step altered the Pt species and resulted in an increase in electron density, stronger Pt-Ce interaction, and higher CO adsorption [25].
Plasma treatment, as a means of defecting catalysts and altering Ni properties, has not been investigated for pretreating catalysts used in the photothermal methanation of CO 2 . Overall, applying a plasma at different stages of catalyst preparation has been reported to impact material characteristics such as structural properties and metal-support interaction; subsequently influencing the catalytic performance.
In this work, plasma treatment was utilised to induce defects on Ni/TiO 2 catalysts for the photothermal conversion of CO 2 to methane. A TiO 2 support was prepared by flame spray pyrolysis (FSP) and loaded with Ni. Plasma treatment was applied to either: (i) as-prepared NiO/TiO 2 prior to reduction/passivation; or (ii) after reduction/passivation of the as-prepared NiO/TiO 2 . The plasma pretreatment can induce defects in the TiO 2 support as well as influence the Ni catalyst properties. Its application at different points within the catalyst preparation process enables us to examine its influence on the Ni/TiO 2 catalyst characteristics and their subsequent catalyst activity. The Ni/TiO 2 catalysts were assessed for CO 2 reduction under photothermal conditions in a batch-circulated reactor system, where light was the sole driving force for the reaction. To decouple the influence of light and heat on the catalytic process, the reaction was undertaken in a continuous flow reactor system. Within the continuous flow reactor, catalyst activity/selectivity under only either thermal stimulus or combined thermal-light stimuli, can be determined. The materials were characterised to gain an appreciation of the impact plasma treatment had on inducing defects in the TiO 2 support and altering the Ni catalyst properties, and the ensuing catalytic activity for photothermal CO 2 methanation.

Catalyst Synthesis
TiO 2 was prepared by flame spray pyrolysis (FSP), with the synthesis conditions described elsewhere [26]. Briefly, titanium (IV) isopropoxide (97%, Sigma-Aldrich, St. Louis, MO, USA) was mixed with absolute ethanol (Chem-Supply, Adelaide, Australia) to form a 1.26 M precursor solution. The mixture was then fed into the flame via a syringe pump at a rate of 5 mL/min. A 5 L/min O 2 sheath (oxygen, >99.9%, Coregas, Yennora, Australia) aided upward movement of the synthesised particles into the vacuum hood. The flame was supported with a mix of 1.5 L/min CH 4 (>99.95%, Coregas) and 3.2 L/min O 2 . The precursor was dispersed using a 5 L/min gas flow of O 2 , with the pressure drop across the nozzle maintained at 150 kPa. A vacuum pump connected to a Whatmann filter paper located above the flame was used to collect the product.
NiO/TiO 2 catalyst was prepared via impregnation and calcination. During the synthesis, nickel (II) nitrate hexahydrate (99.999%, Sigma-Aldrich) was dissolved in Milli-Q water (18.2 MΩ·cm, Merck Millipore, Billerica, MA, USA) to form a 0.3 M solution. The FSP-prepared TiO 2 was added to the Ni solution and stirred at 120 • C until a paste formed. The paste was further dried at 110 • C in an oven for 16 h and then ground with a mortar and pestle. The powder was loaded into a glass reactor with a 10 mm internal diameter and calcined at 400 • C for 3 h under dry air (Coregas, 50 mL/min) at ramp rate of 5 • C/min. The NiO/TiO 2 catalysts were reduced and passivated in a Micromeritics Autochem 2910 (Micromeritics, Norcross, GA, USA). For the reduction step,~300 mg of NiO catalyst was reduced in 10.18% H 2 /Ar (20 mL/min) ramped to 500 • C at 5 • C/min, held for 1 h, and then cooled under Ar. The reduced Ni catalyst was then passivated under 10 mL/min 0.97% O 2 /He (Coregas) for 12 h at room temperature, after which it was ground with a mortar and pestle.
Plasma treatment was undertaken using a Dielectric Barrier Discharge (DBD) plasma system (CTP-2000K Plasma Generator, Corona Lab, Nanjing, China), with a discharge gap of~8 mm, as reported elsewhere [24]. Typically,~250 mg of catalyst was loaded into the plasma chamber and subjected to plasma under a 30 mL/min He (>99.996%, Coregas) flow for 20 min. Plasma treatment was conducted on either: (i) the as-prepared NiO/TiO 2 prior to reduction and passivation; or (ii) samples that had been initially reduced and passivated. The catalysts are labelled based on the pretreatment order where reduced/passivated and then plasma treated is (R-P)Ni/TiO 2 , plasma treated and then reduced/passivated is (P-R)Ni/TiO 2 . A catalyst that has been only reduced/passivated (i.e., no plasma treatment) is referred to as (R)Ni/TiO 2 . A catalyst that has been only plasma-treated (i.e., no reduction/passivation) is referred to as (P)NiO/TiO 2 .

Catalyst Characterisation
To determine the actual Ni loading on the catalyst, the as-prepared NiO/TiO 2 was digested in aqua regia using microwave assistance and subject to inductively coupled plasma optical emission spectrometry (ICP-OES) using a PerkinElmer OPTIMA 7300 ICP-OES instrument (PerkinElmer, Waltham, MA, USA).
Specific surface area (SSA) was measured using N 2 adsorption/desorption isotherms (Micromeritics Tristar 3030) at −196 • C. The samples were pretreated under vacuum for 3 h at 150 • C prior to analysis. The SSA was determined by the Brunauer-Emmet-Teller (BET) method. Material phases and crystallinity were analysed using X-ray diffraction (XRD) with a PANalytical Xpert Multipurpose X-ray Diffraction System (MPD) (Malvern, UK). The XRD instrument was set at 45 kV and 40 mA with a Cu Kα source. The pattern was collected over the range of 2θ = 20 • -100 • with a scan rate of 0.01 • /min and a step size of 0.026 • . The Scherrer equation was used to calculate the crystal size using the 100% intensity peak and a shape factor of 0.9.
As light absorbance by the Ni catalysts is critical for driving the photothermal methanation reaction, the materials were characterised by ultraviolet-visible-near infrared (UVvis-NIR) spectrometry using a Shimadzu UV-Vis 3600 (Shimadzu, Kyoto, Japan). The collected reflectance spectra were converted to absorbance spectra via the Kubelka-Munk equation, referenced to barium sulphate. Subsequently, the converted spectra of nickelloaded samples were normalised against absorbance of TiO 2 .
Catalyst reducibility was characterised using a Micromeritics Autochem 2910 (Micromeritics, Norcross, GA, USA). For the hydrogen-temperature programmed reduction (H 2 -TPR) procedure, approximately 50 mg of catalyst was loaded into a quartz U-tube, supported by a plug of quartz wool. The catalyst was pretreated under an argon (>99.997%, Coregas) flow (20 mL/min) where it was heated to 150 • C at a rate of 10 • C/min and held for 0.5 h. The sample was then cooled to 50 • C prior to introducing 10.18% H 2 /Ar (Coregas) at 20 mL/min. Heating from 50 • C to 700 • C at a ramp rate of 5 • C/min was applied and the H 2 consumed was measured by a TCD. CO 2 -temperature programmed desorption (CO 2 -TPD) experiments were also conducted on the Micromeritics Autochem 2910. Here, approximately 50 mg of sample was placed into the quartz U-tube. The catalyst was initially reduced in 10.18% H 2 /Ar (20 mL/min) ramped at 5 • C/min up to 500 • C, held for 1 h, and cooled to 50 • C in 20 mL/min He. The sample was next exposed to 20 mL/min CO 2 (>99.5%, Coregas) for 1 h at 50 • C after the gas flow was swapped to He for 1 h. The CO 2 -TPD was conducted in 20 mL/min He from 50 • C to 700 • C at a ramp rate 10 • C/min. Surface species present on the pretreated Ni/TiO 2 catalysts were identified by Xray photoelectron spectroscopy (XPS). Shifts in binding in energy were referenced to the Carbon 1s peak (284.8 eV) and the spectra were collected using an ESCALAB 250Xi (Thermo Scientific, Waltham, MA, USA) equipped with an Al Kα X-ray source. Transmission electron microscopy (TEM) images and energy-dispersive spectroscopy (EDS) mapping were taken by a JEOL JEM-ARM200F microscope (JEOL Ltd., Tokyo, Japan), operating at 200 kV. Prior to the imaging, the Ni/TiO 2 catalysts were ultrasonically dispersed in ethanol and dropcast onto lacy carbon-coated copper grids. Defects present in the TiO 2 were evaluated using electron paramagnetic resonance (EPR) and Raman spectroscopies. EPR was conducted at −153 • C (cooled by liquid nitrogen) using 20 mg of sample loaded in a 4 mm (internal diameter) quartz tube and analysed by a Bruker EMX X-Band ESR spectrometer (Bruker, Billerica, MA, USA). Raman spectroscopy was performed using a Renishaw inVia 2 Raman Microscope (532 nm) (Renishaw, Wotton-under-Edge, UK).

Activity Tests
Two styles of activity testing were conducted to gain an understanding of the impacts of plasma treatment on catalyst performance for the photothermal methanation of CO 2 : (i) the first utilised a batch-circulated reactor system; (ii) the second employed a continuous flow reactor system where the effects of light and heat on catalyst performance could be decoupled. Details on the two reactor configurations and associated catalyst activity/selectivity evaluation are available in an earlier study [16].
In the case of the batch-circulated reactor system, illumination was provided from the top of the reactor using a 300 W Xenon lamp (Peccell CX-04E with an Eagle R300-3J lamp housing) (Peccell Technologies, Yokohama, Japan). No other heat source was employed. The catalyst was supported on a glass fibre filter (Merck Millipore, Billerica, MA, USA) within the reactor. To load the catalyst, 100 mg of sample was ultrasonically dispersed in Milli-Q water and drop-cast on to a 7 cm 2 region of the filter, dried and then placed in the reactor. A reactant mixture comprising~15 kPa CO 2 and~60 kPa hydrogen (H 2 , >99.99%, Coregas) was injected into, and continuously circulated around, the reactor. The reactor was illuminated using the Xenon lamp (20 A current output) to initiate the photothermal methanation reaction. Reactants and products were analysed using a gas chromatograph (Shimadzu GC-2010 Plus with a Supelco Carboxen 1010 column) (Shimadzu, Kyoto, Japan). The gas chromatograph was equipped with a methaniser, flame ionisation detector (FID) and TCD. The change in temperature of the catalyst was monitored using a thermocouple in direct contact with the catalyst-loaded filter.
In the case of the continuous flow reactor system, heat was supplied from beneath the reactor using a tube furnace. The catalyst was illuminated by a Xenon lamp (300 W, 20 A current output) through a quartz window in the top of the reactor. The catalyst (100 mg) was loaded on an (1.5 cm × 1.5 cm) aluminium plate and placed in the reactor facing upwards. Under thermal-only reaction conditions, the temperature increased from 50-400 • C, at 50 • C intervals, under a 4 mL/min CO 2 and 16 mL/min H 2 reactant flow. The temperature was ramped at 2.5 • C/min and held for 1 h at each temperature interval. The impact of light was studied by replicating the thermal reaction conditions with the addition of catalyst illumination from above. Reactants and products from the reactor effluent were identified using a Shimadzu GC-2010 gas chromatograph (Shimadzu, Kyoto, Japan) equipped with an Agilent J&W HP-PLOT Q capillary column, methaniser and FID. A schematic of the apparatus can be found in Tan et al. [27].

Catalyst Properties
The structural properties of Ni/TiO 2 are displayed in Table 1. As determined from ICP-OES, the actual Ni loading was 9.2 wt.%, close to the nominal 10 wt.% loading. The as-prepared FSP TiO 2 had a SSA of 104 m 2 /g. The type III N 2 adsorption/desorption isotherms of TiO 2 ( Figure S2) indicates a non-porous structure. The pore size distribution ( Figure S2) shows a maximum at approximately 3.46 nm for TiO 2 . After loading the TiO 2 with Ni and subsequent calcination (NiO/TiO 2 ), the SSA decreased to 66 m 2 /g. Plasmatreating the NiO/TiO 2 ((P)NiO/TiO 2 ) saw the SSA increase slightly to 75 m 2 /g. Both NiO/TiO 2 and (P)NiO/TiO 2 displayed type V isotherms. Two peaks, centred at 4.2-5.1 nm and 8.7-9.1 nm, were observed in the pore size distribution of the nickel catalysts. As the catalysts are non-porous, the porosity observed from N 2 adsorption/desorption isotherm is likely to stem from spaces between the particles. The catalyst structure and metal distribution are discussed at the end of this section. a "Reduced" refers to "reduction/passivation". b BET specific surface area. c Calculated from 100% intensity peak of XRD pattern ( Figure S2).
The crystal phase and size of the TiO 2 support and Ni deposits were determined from the XRD patterns ( Figure S2). Both anatase and rutile TiO 2 were detected; however, the anatase phase was dominant, which is consistent with TiO 2 synthesised by FSP [26]. The TiO 2 crystal phase was not affected by plasma treatment or Ni addition and calcination. Plasma treatment had a minimal impact on the NiO crystallite size (NiO/TiO 2 : 7.7 nm, (P)NiO/TiO 2 : 9.3 nm) and the TiO 2 crystal size (21 nm) remained unchanged. Upon reduction/passivation, the Ni crystal sizes were similar ((P-R)Ni/TiO 2 : 13.7 nm, (R)Ni/TiO 2 : 13.1 nm, (R-P)Ni/TiO 2 : 12.7 nm), indicating that the plasma treatment had little impact on this characteristic. The BET and XRD results indicate the plasma pretreatment does not impart any undesirable morphological changes.
As plasma treatment has the potential to induce defects on the TiO 2 support, the asprepared, plasma-treated and reduced/passivated TiO 2 (without Ni loading) were studied by EPR and Raman spectroscopy ( Figure 1). The EPR spectrum exhibits a peak at g = 2.0020 for the as-prepared TiO 2 , associated with the presence of oxygen vacancies [28,29]. The peak representing Ti 3+ (g ≈ 1.94) was not observed [28]. The EPR spectrum remains unchanged after plasma treating the TiO 2 ((P)TiO 2 ), with no shift in the g = 2.0020 peak position and no new peaks appearing. Similarly, the EPR spectrum for the passivated/reduced sample ( Figure S3a) remains unchanged.
Raman spectroscopy was used to further understand the influence of plasma treatment on the TiO 2 (Figure 1b,c). Peaks corresponding to the anatase phase of TiO 2 were evident for both samples (143/144 cm −1 , 396/397 cm −1 , 516 cm −1 , and 638/639 cm −1 ) [30,31]. This is consistent with the presence of anatase from XRD ( Figure S2). The peak at 143/144 cm −1 is the linear combination of asymmetric bending of O-Ti-O bonds, while 197/198 cm −1 can be attributed to the presence of the brookite phase [30]. Consistent with the XRD results, no significant peak shifts or broadening was evident as a consequence of the plasma treatment, indicating no major structural changes. A peak at 1050 cm −1 appeared for (P)TiO 2 , which is not commonly observed in TiO 2 . Surmacki et al. [32] observed a peak at 1048 cm −1 in an electron beam irradiated N-doped TiO 2 sample which was not observed in the nonirradiated sample. They attributed it to a new species from the interaction of the TiO 2 and N-dopant under irradiation, although it was not assigned to a specific feature of TiO 2 .
Plasma treatment of FSP SiO 2 -TiO 2 saw the emergence of a new Raman peak at 440 cm −1 for that sample, which was attributed to structural distortion [24]. Although the peak at 1050 cm −1 (Figure 1c) has not been defined in the literature, the formation the additional peak for (P)TiO 2 indicates that a structural change occurred due to the plasma treatment. A similar change in the Raman spectra of (R)TiO 2 , (P-R)TiO 2 , and (R-P)TiO 2 was not apparent, suggesting that the reduction/passivation may constrain the effect of plasma on the TiO 2 . Raman spectroscopy was used to further understand the influence of plasma treatment on the TiO2 (Figure 1b,c). Peaks corresponding to the anatase phase of TiO2 were evident for both samples (143/144 cm −1 , 396/397 cm −1 , 516 cm −1 , and 638/639 cm −1 ) [30,31]. This is consistent with the presence of anatase from XRD ( Figure S2). The peak at 143/144 cm −1 is the linear combination of asymmetric bending of O-Ti-O bonds, while 197/198 cm −1 can be attributed to the presence of the brookite phase [30]. Consistent with the XRD results, no significant peak shifts or broadening was evident as a consequence of the plasma treatment, indicating no major structural changes. A peak at 1050 cm −1 appeared for (P)TiO2, which is not commonly observed in TiO2. Surmacki et al. [32] observed a peak at 1048 cm −1 in an electron beam irradiated N-doped TiO2 sample which was not observed in the non-irradiated sample. They attributed it to a new species from the interaction of the TiO2 and N-dopant under irradiation, although it was not assigned to a specific feature of TiO2. Plasma treatment of FSP SiO2-TiO2 saw the emergence of a new Raman peak at 440 cm −1 for that sample, which was attributed to structural distortion [24]. Although the peak at 1050 cm −1 (Figure 1c) has not been defined in the literature, the formation the additional peak for (P)TiO2 indicates that a structural change occurred due to the plasma treatment. A similar change in the Raman spectra of (R)TiO2, (P-R)TiO2, and (R-P)TiO2 was not apparent, suggesting that the reduction/passivation may constrain the effect of plasma on the TiO2.
As the photothermal CO2 hydrogenation reaction utilises light to heat conversion, the light absorbance of Ni/TiO2 was studied by UV-vis-NIR spectroscopy (Figure 2a). The asprepared TiO2 absorbed light in the UV region and exhibited a bandgap of 3.4 eV. The calculated bandgap was consistent with the literature, particularly for anatase TiO2 at ~3.2 eV. [33,34], which was the dominant phase present in the materials. Upon Ni addition and pretreatment, the light absorbance extends from the UV region into the visible and near infrared. The absorbance in the Vis and NIR regions can be attributed to the presence of metallic Ni deposits, consistent with group VIII metals [35]. Specifically, the broadband absorbance is driven by intraband transition within overlaid partially filled 3d sub-orbitals [36]. All Ni/TiO2 catalysts exhibited full spectrum absorbance, coinciding with the spectrum profile of the Xenon lamp, which allows utilisation of the full solar spectrum for the photothermal CO2 reaction. As the photothermal CO 2 hydrogenation reaction utilises light to heat conversion, the light absorbance of Ni/TiO 2 was studied by UV-vis-NIR spectroscopy (Figure 2a). The as-prepared TiO 2 absorbed light in the UV region and exhibited a bandgap of 3.4 eV. The calculated bandgap was consistent with the literature, particularly for anatase TiO 2 at~3.2 eV. [33,34], which was the dominant phase present in the materials. Upon Ni addition and pretreatment, the light absorbance extends from the UV region into the visible and near infrared. The absorbance in the Vis and NIR regions can be attributed to the presence of metallic Ni deposits, consistent with group VIII metals [35]. Specifically, the broadband absorbance is driven by intraband transition within overlaid partially filled 3d sub-orbitals [36]. All Ni/TiO 2 catalysts exhibited full spectrum absorbance, coinciding with the spectrum profile of the Xenon lamp, which allows utilisation of the full solar spectrum for the photothermal CO 2 reaction.
To understand the impact of plasma treatment on the support, CO 2 -TPD analyses were conducted on the neat TiO 2 (Figure 2b). A CO 2 desorption peak was centred at 190 • C and 194 • C for TiO 2 and (P)TiO 2 , respectively, which indicates medium-strength basic sites [37]. The desorption temperature is favourable, as it is within the temperature range of photothermal CO 2 methanation. Relative to TiO 2 , the area under the peak for (P)TiO 2 increased by 11%, indicating a mild increase in CO 2 adsorption and the effectiveness of plasma treatment in inducing defects in TiO 2 . Defects present on TiO 2 can behave as sites for CO 2 adsorption [38]. Subsequently, the increase in CO 2 desorption peak area can be a contributing factor to the improved activity.
The reducibility of NiO/TiO 2 and pretreated Ni/TiO 2 catalysts was characterised by H 2 -TPR (Figure 2c,d). As shown in Figure 2c, the as-prepared TiO 2 has a low intensity peak between 200-250 • C, which may arise from the partial reduction in TiO 2 [39] or the removal of residual surface species resulting from the flame synthesis. For NiO/TiO 2 , peaks are observed at~205 • C, 335 • C, and 425 • C. The ratio of integrated H 2 -TPR peak area was 0.97:1 for (P)NiO/TiO 2 :NiO/TiO 2 , which indicated that slightly less H 2 was required to reduce (P)NiO/TiO 2 . The lower temperature reduction peaks for NiO/TiO 2 can be attributed to the reduction in NiO that weakly interacts with TiO 2 . The reduction in nickel oxide occurs at~400 • C [40]. The broad peak extending up to~500 • C can be attributed to NiO with a strong NiO-TiO 2 interaction [41]. After plasma treatment, (P)NiO/TiO 2 ( Figure 2c) displayed a slight decrease in reduction temperature of all peaks to~195 • C, 310 • C, and 400 • C, implying an ease in reducibility and a slight variation to the metalsupport interaction. An increase in the ratio of the lowest temperature peak (~200 • C) to the higher reduction peak (~250 • C to~500 • C) was observed for (P)NiO/TiO 2 . The peak at 205 • C for NiO/TiO 2 accounted for 27% of its reduction profile, while the peak at~195 • C for (P)NiO/TiO 2 accounted for 32% of its reduction profile. The 5% increase in the low temperature peak in (P)NiO/TiO 2 relative to NiO/TiO 2 depicts an increase in NiO with weak metal-support interaction due to plasma treatment. To understand the impact of plasma treatment on the support, CO2-TPD analyses were conducted on the neat TiO2 (Figure 2b). A CO2 desorption peak was centred at 190 °C and 194 °C for TiO2 and (P)TiO2, respectively, which indicates medium-strength basic sites [37]. The desorption temperature is favourable, as it is within the temperature range of photothermal CO2 methanation. Relative to TiO2, the area under the peak for (P)TiO2 increased by 11%, indicating a mild increase in CO2 adsorption and the effectiveness of plasma treatment in inducing defects in TiO2. Defects present on TiO2 can behave as sites for CO2 adsorption [38]. Subsequently, the increase in CO2 desorption peak area can be a contributing factor to the improved activity.
The reducibility of NiO/TiO2 and pretreated Ni/TiO2 catalysts was characterised by H2-TPR (Figure 2c,d). As shown in Figure 2c, the as-prepared TiO2 has a low intensity peak between 200-250 °C, which may arise from the partial reduction in TiO2 [39] or the removal of residual surface species resulting from the flame synthesis. For NiO/TiO2, peaks are observed at ~205 °C, 335 °C, and 425 °C. The ratio of integrated H2-TPR peak area was 0.97:1 for (P)NiO/TiO2:NiO/TiO2, which indicated that slightly less H2 was required to reduce (P)NiO/TiO2. The lower temperature reduction peaks for NiO/TiO2 can be attributed to the reduction in NiO that weakly interacts with TiO2. The reduction in nickel oxide occurs at ~400 °C [40]. The broad peak extending up to ~500 °C can be attributed to NiO with a strong NiO-TiO2 interaction [41]. After plasma treatment, (P)NiO/TiO2 (Figure 2c) displayed a slight decrease in reduction temperature of all peaks Following reduction/passivation, the Ni/TiO 2 was reducible at temperatures below 250 • C (Figure 2d). The oxygen passivation was conducted to prevent uncontrolled oxidation after catalyst reduction. The decrease in the reduction temperature post-passivation, relative to as-prepared NiO/TiO 2 , is important, as the onset temperature reached in the photothermal system is approximately 200 • C [16]. The post-passivation reduction peaks for (R)Ni/TiO 2 and (P-R)Ni/TiO 2 can be deconvoluted into two peaks ((R)Ni/TiO 2 :~190 • C and 205 • C, (P-R)Ni/TiO 2 :~210 • C and 220 • C. The post-passivation reduction profile for (R-P)Ni/TiO 2 was different, with only one reduction peak present at~215 • C and a shoulder at 200 • C. These peaks are indicative of the re-reduction in the surface oxidised Ni deposits after passivation. The peak ratio, integrated from~180 • C to~270 • C and normalised to (R)Ni/TiO 2 , was (R-P)Ni/TiO 2 (1.01) > (R)Ni/TiO 2 (1) > (P-R)Ni/TiO 2 (0.83). The extent of re-oxidation from the oxygen passivation was calculated: (R-P)Ni/TiO 2 (7.5%) ≈ (R)Ni/TiO 2 (7.4%) > (P-R)Ni/TiO 2 (6.3%). Whilst the percent oxidation was comparable between the catalysts, a clear difference in profile features is evident for the (R-P)Ni/TiO 2 compared to (R)Ni/TiO 2 and (P-R)Ni/TiO 2 , where two reduction peaks are present, whilst in the case of the (R-P)Ni/TiO 2 sample, a peak with a lower temperature shoulder is evident. This may be an impact of the re-reduction in different Ni sizes (vide infra). The change in reduction profile for (R-P)Ni/TiO 2 compared to (R)Ni/TiO 2 and (P-R)Ni/TiO 2 indicates that the plasma treatment had the most significant impact on the Ni deposits when implemented after reduction and passivation. However, (P-R)Ni/TiO 2 required the least H 2 for re-reduction and subsequently ease of catalyst activation.
XPS was conducted on the Ni/TiO 2 to examine their surface oxidation state (Figure 3). Peaks corresponding to Ni 2+ (nickel oxide) and Ni 0 (metallic nickel) were identified as seen in Figure 3a, while the unlabelled peaks (at 861.4 eV and 856.6 eV) are satellite peaks [42]. The ratio of Ni 2+ to Ni 0 was calculated to be (R-P)Ni/TiO 2 (6.0) > (R)Ni/TiO 2 (4.6) > (P-R)Ni/TiO 2 (4.3). The ratio was consistent with the H 2 -TPR results for Ni/TiO 2 (Figure 2d) for which (P-R)Ni/TiO 2 had the lowest quantity of NiO. As is evident in Figure 3b, only a peak corresponding to Ti 4+ was present in all samples. A shift in the binding energy of Ti 4+ was observed as a consequence of the impact of plasma, indicating the presence of oxygen vacancy (Ti 3+ ) formation [43]. The shift in the Ti 4+ peak for (P-R)Ni/TiO 2 (from 458.6 eV to 458.4 eV) and (R-P)Ni/TiO 2 (from 458.6 eV to 458.4 eV), compared to (R)Ni/TiO 2 indicates a change in Ti oxidation state arising from plasma treatment. The O 1s spectrum of (R)Ni/TiO 2 can be deconvoluted to lattice oxygen (529.7 eV), -OH bond (531.2 eV), and adsorbed water (532.7 eV) [44,45]. Corresponding shifts of the oxygen lattice from plasma treatment were observed in (P-R)Ni/TiO 2 and (R-P)Ni/TiO 2 , from 529.7 eV in (R)Ni/TiO 2 to 529.6 eV. A slight decrease in oxygen lattice peak position due to plasma treatment has been previously observed [46].  To further understand the interaction between Ni and the TiO2 support, TEM images were taken and EDS mapping was performed (Figure 4). The Ni deposits display little difference in size between the (R)Ni/TiO2 (17 ± 5 nm) and (P-R)Ni/TiO2 (18 ± 5 nm) samples, consistent with the XRD crystallite sizes ( Table 1). The (R-P)Ni/TiO2 sample also exhibited Ni deposits of a similar size (16 ± 5 nm). However, also present were additional finely dispersed Ni deposits. The sizes of the fine Ni deposits were not able to be accurately measured and, as such, were not included in the associated histogram ( Figure 4). To further understand the interaction between Ni and the TiO 2 support, TEM images were taken and EDS mapping was performed (Figure 4). The Ni deposits display little difference in size between the (R)Ni/TiO 2 (17 ± 5 nm) and (P-R)Ni/TiO 2 (18 ± 5 nm) samples, consistent with the XRD crystallite sizes ( Table 1). The (R-P)Ni/TiO 2 sample also exhibited Ni deposits of a similar size (16 ± 5 nm). However, also present were additional finely dispersed Ni deposits. The sizes of the fine Ni deposits were not able to be accurately measured and, as such, were not included in the associated histogram ( Figure 4). . TEM images and EDS mapping of (a) (R)Ni/TiO2, (b) (P-R)Ni/TiO2 and (c) (R-P)Ni/TiO2 including: (i,v) bright field images; (ii,vi) dark field images; (iii,vii) EDS mapping; and (iv,viii) particle size distribution histograms with range of particle counts = 50-100. R = reduced/passivated, P = plasma treated.
Characterisation of the catalyst properties indicates that plasma treatment modifies the TiO2. The modification is evident in changes to the surface texture of the TiO2 ( Figure  4) and the shift in the Ti 2p peak shift in the XPS spectra (Figure 3) for the plasma-treated samples. Plasma treatment also leads to an increase in CO2 adsorption by the neat TiO2 . TEM images and EDS mapping of (a) (R)Ni/TiO 2 , (b) (P-R)Ni/TiO 2 and (c) (R-P)Ni/TiO 2 including: (i,v) bright field images; (ii,vi) dark field images; (iii,vii) EDS mapping; and (iv,viii) particle size distribution histograms with range of particle counts = 50-100. R = reduced/passivated, P = plasma treated.
Characterisation of the catalyst properties indicates that plasma treatment modifies the TiO 2 . The modification is evident in changes to the surface texture of the TiO 2 ( Figure 4) and the shift in the Ti 2p peak shift in the XPS spectra (Figure 3) for the plasma-treated samples.
Plasma treatment also leads to an increase in CO 2 adsorption by the neat TiO 2 (Figure 2b) and structural modification as illustrated by the emergence of a new peak for (P)TiO 2 in the Raman spectra ( Figure 1). Plasma treatment affected the Ni deposit properties, impacting the reducibility of (R-P)Ni/TiO 2 (Figure 1c,d). To understand the impact of the altered characteristics of the Ni/TiO 2 catalyst arising from the plasma treatment, the activity/selectivity of the catalysts for CO 2 methanation under (i) photothermal (batchcirculated reactor system) and (ii) decoupled photo and thermal (continuous flow reactor system) conditions was evaluated.

CO 2 Methanation Results
The effects of the pretreatment conditions on catalyst activity for the photothermal methanation reaction within the batch-circulated reactor system are shown in Figure 5a,b.  Figure 2b) and structural modification as illustrated by the emergence of a new peak for (P)TiO2 in the Raman spectra ( Figure 1). Plasma treatment affected the Ni deposit properties, impacting the reducibility of (R-P)Ni/TiO2 (Figure 1c,d). To understand the impact of the altered characteristics of the Ni/TiO2 catalyst arising from the plasma treatment, the activity/selectivity of the catalysts for CO2 methanation under (i) photothermal (batchcirculated reactor system) and (ii) decoupled photo and thermal (continuous flow reactor system) conditions was evaluated.

CO2 Methanation Results
The effects of the pretreatment conditions on catalyst activity for the photothermal methanation reaction within the batch-circulated reactor system are shown in Figure 5a,b. Within the batch-circulated reactor system, the methane formation trend followed the order: (P-R)Ni/TiO2 > (R)Ni/TiO2 > (R-P)Ni/TiO2 (Figure 5a). (P-R)Ni/TiO2 and (R)Ni/TiO2 required 75 min and 90 min, respectively, to achieve maximum CH4 formation with ≈ 100% selectivity (with trace CO). A longer time period was required for methane formation to stabilise in the case of (R-P)Ni/TiO2. The temperature profile for each of the Ni/TiO2 samples is included in Figure 5b. The temperature for all catalysts reached 175- Within the batch-circulated reactor system, the methane formation trend followed the order: (P-R)Ni/TiO 2 > (R)Ni/TiO 2 > (R-P)Ni/TiO 2 (Figure 5a). (P-R)Ni/TiO 2 and (R)Ni/TiO 2 required 75 min and 90 min, respectively, to achieve maximum CH 4 formation with ≈ 100% selectivity (with trace CO). A longer time period was required for methane formation to stabilise in the case of (R-P)Ni/TiO 2 . The temperature profile for each of the Ni/TiO 2 samples is included in Figure 5b. The temperature for all catalysts reached 175-200 • C within the first 15 min due to the initial photothermal heating of the catalyst combined with the later exothermic heat from methane formation. Given the Ni in the pretreated Ni/TiO 2 catalysts was reducible below 250 • C (Figure 2d) following reduction and passivation of the catalysts, it is anticipated that catalysts are converted to their reduced form during the batch reaction. Absorbance by the Ni/TiO 2 catalysts in the infrared region contributed to the localised heating [47]. Consequently, the capacity of the catalysts for full spectrum absorbance was beneficial for the reaction. With increasing methane production, the temperature further increased and reached 250-280 • C after 3 h. The further temperature increase arising from the exothermic nature of methanation can promote better catalyst reduction and contribute to greater methane production [16]. Both (P-R)Ni/TiO 2 and (R)Ni/TiO 2 exhibit a similar temperature profile while the temperature profile for (R-P)Ni/TiO 2 was generally lower. It is apparent that the temperature profile can be correlated to methane formation with (P-R)Ni/TiO 2 achieving the highest temperature, followed by (R)Ni/TiO 2 and (R-P)Ni/TiO 2 .
The influences of light and heat were decoupled in the continuous flow reactor system, where a furnace (separate from the light source) was used to heat the catalyst bed. CO 2 conversion and methane selectivity for thermal-only input are shown in Figure 5c,d. CO 2 conversion under thermal-only conditions was observable from 200 • C onwards, which is consistent with the findings from the batch photothermal catalytic reactor. A significant increase in CO 2 conversion was observed with increasing temperature for all catalysts, particularly from 300 • C to 350 • C, with methane as the sole product up to 350 • C. A significant difference in CO 2 conversion is not apparent over the temperature range 250-400 • C for any of the catalysts. Despite the different pretreatments inducing variations to the catalyst properties, such as reducibility and defect density, the CO 2 conversion and methane selectivity were not altered. The indifference indicates that the catalyst modification was only effective under photothermal conditions.
Adding illumination to the thermal stimulus can provide better understanding on the influence of light on the reaction, particularly across the UV-visible region. The addition of light, Figure S5, to the thermal conditions did not lead to any significant increase in CO 2 conversion or change to product selectivity. Subsequently, photothermal CO 2 conversion by the Ni/TiO 2 catalysts appears to be thermally dominant. Under both conditions-thermal-only and light-plus-thermal conditions-methane selectivity decreased from 350 • C onwards in favour of CO formation due to the reverse-water gas shift reaction [11].
The thermal dominance during photothermal CO 2 conversion indicates that the absorbance of light from the NIR portion of the spectrum is crucial for reaction feasibility. While catalyst modifications induced by pretreatment, such as greater CO 2 adsorption, TiO 2 defects, and a change in metal-support interaction, did not promote greater CO 2 conversion under the decoupled photo-and-thermal condition (continuous flow reactor), they do significantly influence the photothermal catalytic activity (cyclic batch reactor). Importantly, the order in which the plasma and reduction/passivation treatments were applied influenced the catalytic activity of the Ni/TiO 2 .
Others have examined defects generated in TiO 2 as a consequence of plasma treatment. Bharti et al. reported that plasma treatment in an air environment could induce Ti 3+ and oxygen vacancies in a TiO 2 thin film [46]. The higher activity of (P-R)Ni/TiO 2 compared to (R-P)Ni/TiO 2 in our work can be related to additional defects formed and their ensuing stability arising from the different pretreatment sequences. Chen et al. found that plasma treating Pt/CeO 2 followed by 20% H 2 /N 2 thermal treatment led to higher toluene catalytic oxidation, compared to thermal then plasma treatment. The plasma treatment affected the surface oxygen defects and Pt-Ce interaction. When thermally treating the catalyst after plasma treatment, more defects could be generated via greater hydrogen dissociation on the plasma-generated defect sites [48]. In our case, while the impact of plasma treatment was most prominent for (R-P)Ni/TiO 2 , the changes were not necessarily beneficial for enhancing photothermal CO 2 hydrogenation relative to (R)Ni/TiO 2 . The thermal treatment can stabilise the catalyst structure and subsequently minimise the impact of any ensuing plasma treatment [24]. Consequently, the results suggest that the most effectual approach for the Ni/TiO 2 catalyst is to thermally treat the defected catalyst to stabilise the defects generated from the plasma treatment.

Conclusions
The influence of Ni/TiO 2 catalyst pretreatment strategy on the generation and stabilisation of surface defects active for the photothermal CO 2 methanation reaction was examined. The sequence of two pretreatment steps, involving a He-plasma stage and a reduction/passivation stage, were switched, with the ensuing impact on defect generation and, in turn, catalyst activity, assessed. The catalyst comprised a TiO 2 support prepared via FSP which was loaded with a Ni catalyst. When the as-prepared NiO/TiO 2 was treated with plasma, the subsequent reduction/passivation step stabilised the plasma-generated defects on the TiO 2 support, as evident from XPS, and Raman analysis. This increase in defects resulted in an increased CO 2 interaction (shown in the CO 2 -TPD of the plasma-treated TiO 2 ). Ultimately, these increased surface TiO 2 defects resulted in higher photothermal catalytic performance. When reduction/passivation was performed prior to plasma treatment, the reduction/passivation imbued the NiO/TiO 2 with a greater resistance to subsequent defecting by the plasma treatment. It was evident, through H 2 -TPR of the Ni/TiO 2 samples (after reduction and passivation), that a change in metal-support interaction occurred as a result of plasma treatment after reduction. Deconvolution of the light and heat effects demonstrated that the photothermal catalytic reaction was thermally dominant with the light as a heat source. The work highlights the importance of the material pretreatment strategy when using plasma as a means to generate active defect sites within a catalyst. If not performed correctly, the potential benefits of plasma treatment can be lost.