The Influence of the Reaction Conditions on the Photocatalytic Gas‐Phase Conversion of Methanol with Water Vapor over Pt/SrTiO3 in a Continuously Operated Flow Reactor

Continuous methanol photooxidation in the gas phase is a promising method to produce valuable chemicals like formaldehyde or methyl formate in addition to hydrogen under mild conditions. The influence of the reaction conditions on the selectivity of methanol oxidation to formaldehyde is studied using a heated flat‐plate flow photoreactor illuminated by an LED array (λmax = 368 nm) and Pt‐modified SrTiO3. A combination of online analytical methods allowed to quantify all gaseous products during extended time‐on‐stream (> 48 h TOS). The selectivity to formaldehyde is found to be primarily determined by the residence time and the process temperature. At a low methanol to water ratio, methanol conversion and evolution of CO2 are favored, whereas the light intensity primarily influenced the apparent quantum yield from 5.1 to 1.8% at 9.36 to 52.93 mW cm−2, respectively, and the methanol conversion thus determining the economic efficiency of the process. Operation temperatures higher than 110 °C resulted in a strong deactivation of the catalyst while simultaneously the formation of CO at the expense of formaldehyde selectivity is favored. This study demonstrates the importance of understanding the influence of relevant reaction conditions and the potential of selective photocatalytic gas‐phase oxidation of small molecules.


Introduction
In 2015, the UN member states adopted the 2030 agenda for sustainable development including the 17 sustainable development goals (SDGs) which need to be addressed by all countries in a DOI: 10.1002/adsu.202300329global partnership. [1]The achievement of many SDGs directly or indirectly relies on chemical processes that strongly depend on heterogeneously catalyzed reactions.Although semiconductor-based photocatalytic reactions are still not implemented in large-scale chemical processes, they have the potential to improve or replace conventional routes by more sustainable and energy-efficient ways to harvest and convert solar energy into valuable chemical products.
Photooxidation of biomass-derived methanol in the gas phase represents a promising and cost-effective way to convert methanol under milder reaction conditions compared with classical thermally catalyzed routes.Besides a few studies focusing on gas-phase reaction conditions, most studies revolve around liquid-phase batch or semibatch reactors which usually do not match industrial requirements. [2,3][6] Carrying out the reaction in the gas phase, similar to most large-scale industrial processes nowadays under operation, allows for improved mass transfer properties and enables the continuous operation of the reactor, the recycling of reactants, easy catalyst recovery and better separation of (by-)products.Reusable gas-phase microreactors have been successfully developed to study various photocatalytic reactions and to perform efficient catalyst screening. [6,7]Furthermore, gas-phase reactors enable operation at elevated temperatures which has been shown to result in synergistic improvements of the photocatalytic/thermal performance while achieving high conversions and a good selectivity to the target product.These thermally assisted photocatalytic reactions are thermodynamically enabled by UV light irradiation while surface reaction kinetics are improved. [8,9]0] CH 3 OH + H 2 O → CO 2 + 3H 2 (1)  Frequently, CO is reported as the only byproduct originating from the thermal (dark) decomposition of methanol (Equation (2)) or as a product of the stepwise oxidation of methanol to formic acid via formaldehyde (Equations ( 3) and ( 4)) and its subsequent dehydration (Equation ( 5)). HCOOH Additionally, the evolution of byproducts like formaldehyde, methyl formate (Equation ( 6)), formic acid, acetaldehyde (Equation (7)), or dimethyl ether (DME) (Equation ( 8)) has been reported. [11,12]H 3 OH → HCO 2 CH 3 + 2H 2 (6) For example, Chiarello et al. [5,11,12] observed formaldehyde and formic acid accumulation in the liquid phase using a recirculating gas-phase set-up.Nevertheless, CO 2 , CO, and H 2 were still shown to be abundant products.Mills et al. [2,3,10] studied different aspects of the reaction over TiO 2 -based photocatalysts in the gas phase detecting CO 2 and CO as oxidation products.Independent of the reaction conditions, a closed carbon balance was obtained.Consequently, formaldehyde and formic acid are usually only mentioned as intermediates which are further oxidized and do not desorb into the gas phase. [8,9]Although in most studies reaction conditions like the methanol to water ratio or the reaction temperature were varied, the general product spectrum remained similar.Nevertheless, it is important to realize that different reactor geometries, total gas flow rates, or photocatalyst amounts may result in a variation of the product distributions even for similar photocatalysts.
Though formaldehyde is not frequently reported in the context of photocatalytic methanol oxidation, [13] it is widely used in the chemical industry and required for the production of polymers, resins, and solvents or as C1 building block in organic synthesis. [14]The annual formaldehyde production (2022) is 46.6 Mio t, corresponding to a market of 27.57Billion US$ with an expected growth to 40.68 Billion US$ in 2030. [15]Currently, the main production routes are the silver contact process and the formox process that require reaction temperatures of 590 -720 °C and 250 -400 °C, respectively.Here, methanol is either dehydrogenated or oxidized selectively with O 2 achieving a methanol conversion X MeOH of 95-99% with formaldehyde yields of 86 -92%. [16]Thus, the anaerobic photocatalytic conversion of methanol to formaldehyde (Equation (3)) may be a promising and cost-effective alternative allowing for milder reaction conditions compared with classical thermally catalyzed routes.Considering that previous reports, if at all, observed formaldehyde in minor quantities, it is essential to disclose reaction conditions that en-able selective photon-driven formaldehyde formation in conjunction with H 2 production.
Strontium titanate (SrTiO 3 ) is a versatile photoabsorber, which is frequently applied in various photocatalytic applications such as overall water splitting [17] or selective alcohol oxidation [18] in the liquid phase.Recently, SrTiO 3 has been successfully used in gasphase HCl oxidation, [19] but gas-phase oxidation of methanol using SrTiO 3 -based photocatalysts has not yet been reported.[22] As for most photocatalysts, a cocatalyst is required to achieve reasonable H 2 evolution rates (hydrogen evolution reaction, HER). [3]Although the past years have shown great potential in the development of non-precious HER catalysts in various fields, [23] precious metals like Pt, Au, or Ag are considered as most efficient HER cocatalysts in this regard.Among these, Pt usually results in the highest H 2 evolution rates reported from photo-and electrocatalysis [24] due to its high reducibility and suitable hydrogen adsorption/desorption characteristics. [25]Moreover, Pt nanoparticles can be easily immobilized by photodeposition allowing for the formation of a favorable solid-solid interface with SrTiO 3 [20,26] with a good control of size, oxidation state and position of the cocatalyst. [27]uilding on our expertise in the gas-phase photocatalytic conversion of HCl, in this study, we demonstrate the influence of the reaction conditions on the photocatalytic methanol conversion in the presence of water vapor over Pt/SrTiO 3 photocatalysts and disclose process conditions allowing for high formaldehyde selectivity.A self-built gas-phase set-up with appropriate online analytics enabled us to monitor the reaction continuously and to determine the conversion of methanol, the distribution of carbon-containing products and the hydrogen evolution rates with high time resolution.In an initial photocatalyst screening, we showed that 0.1 wt.% Pt/SrTiO 3 produced by reductive photodeposition of H 2 PtCl 6 on commercial SrTiO 3 resulted in the highest methanol conversion.Using the optimized photocatalyst we explored the influence of reaction temperature, residence time, methanol to water ratio, and UV light intensity on methanol conversion and product selectivity in detail.We show that formaldehyde formation can be achieved with high selectivity at temperatures below 120 °C and that the residence time of methanol is essential for controlling product selectivity.Thus, this study highlights the importance of so far unexplored process conditions in photocatalytic methanol oxidation and reveals the potential of thermally assisted photocatalytic methanol conversion for selective oxidation processes.

Results and Discussion
The Pt loading was initially optimized in the range of 0.025 wt.% to 2.0 wt.% (Figure 1).Independent of the deposited Pt amount, the pre-treatment performed prior to the actual methanol oxidation experiments resulted in a stepwise CO 2 evolution as shown exemplarily in Figure S1 (Supporting Information).Particularly, evolution of CO 2 was observed upon exposure to humid N 2 at 100 °C, which was further promoted by photocatalyst illumination.The photocatalyst was considered clean when y CO2 was less than 5 ppm.According to ex situ XPS analysis (Figure S2, Supporting Information, right), the predominant Pt species was found to be Pt 0 .However, in contact with the highly oxidizing Table 1.Process conditions used for photocatalyst optimization.Here, a reactor temperature of 100 °C and high light intensity of 52.93 mW cm −2 were used to achieve a reasonable X MeOH and to avoid water condensation.The total volumetric gas flow of 50 mL min −1 was chosen as a compromise between high residence time and high gas flow needed for a fast response of the analytics.The adjusted methanol mole fraction y MeOH of 1.5% enabled the easy detection of methanol and the products while excess water (y H2O = 4.5%) was expected to mitigate catalyst deactivation by formation of organic residues.atmosphere during the cleaning procedure, the formation of PtO x cannot be fully excluded.
The highest X MeOH amounting to 13% in the gas-phase methanol oxidation under the reaction conditions summarized in Table 1 was obtained with a Pt loading of 0.1 wt.% (denoted hereafter 0.1PtSTO).Correspondingly, the parameter variation was carried out with this sample.The full characterization by means of ICP-OES, TEM/EDX, N 2 physisorption, XRD, and UVvis spectroscopy is shown in the supporting information (Table S1 and Figures S3-S7, Supporting Information).Despite the highest X MeOH obtained with 0.1PtSTO, it is already evident that for all tested photocatalysts a high selectivity toward the selective oxidation product formaldehyde is obtained.Thus, surface polymerization of formaldehyde is likely prevented during thermally assisted photocatalytic methanol oxidation.Carbon dioxide and carbon monoxide selectivities remain ≈10%.
Using the same process conditions (i. e. Table 1), a long-term experiment was performed with 0.1PtSTO (Figure S8, Supporting Information).Without illumination, small amounts of H 2 , CO 2 , and CO were formed by the thermal (dark) reaction with CO 2 showing the highest selectivity.Initiating the photocatalytic reaction by UV-LED illumination after 30 min led to a sharp increase in all product evolution rates and a clear decrease of the effluent molar flow of methanol.Similarly, switching off the UV-LEDs resulted in an increase in the effluent molar flow of methanol to its original level substantiating that UV illumination is essential.
During illumination, a complex X MeOH profile was obtained.After an initial peak in X MeOH , a steady increase during the first 9.5 h was observed.Then, X MeOH decreased, approaching a steady-state value of 9.3% after approx.42 h TOS.Interestingly, the product distribution changed only slightly during the reaction and formaldehyde, methyl formate, CO 2, and CO were obtained as most abundant products with a selectivity of 65%, 22%, 7%, and 5% after 48 h, respectively.Other carbon-containing products like DME, dimethoxy methane (DMM), or methane were only observed in trace amounts and a H 2 evolution rate of 195 μmol h -1 was obtained under these conditions.Importantly, the carbonbalance was close to 1 during illumination showing that all relevant products were quantified, and the deposition of carboncontaining residues on the catalyst was presumably very small.
The variation of reaction conditions was consequently performed after > 48 h TOS.Thus, the catalyst was considered to be at steady state.Note that only one parameter at a time was changed while all other parameters were kept constant at the corresponding starting value summarized in Table 1.
The influence of the residence time is often not addressed in the literature related to photocatalytic reactions, but it is known that the residence time is of great importance in traditional thermally driven selective oxidation reactions.Therefore, understanding the influence of the residence time is crucial.Figure 2a shows X MeOH and the product distribution obtained at total gas flows varying from 25 to 100 mL min −1 , which correspond to mean residence times from 36.0 to 9.0 s assuming an ideal plugflow behavior.A lower total gas flow rate resulted in a higher X MeOH accompanied by higher selectivity to CO 2 , methyl formate, and DMM and lower selectivity to formaldehyde.The selectivity to CO was not influenced by the residence time variation.Thus, a high residence time favors the conversion of formaldehyde to follow-up products being in agreement with thermal catalytic formaldehyde production where short contact times are considered beneficial. [28] residence time variation can also help to find mass transfer limitations in a photocatalyzed reaction.Mostly, this topic is addressed for microreactors as these are prone to severe diffusion and/or advection limitations but should also be considered while using larger reactors. [6,29]Mass transfer limitations arise when the rate of diffusion or the rate of advection is smaller than the reaction rate.In this case, only an apparent reaction rate can be derived from the measured data.By plotting the reaction rate (consumption of methanol) against the total volumetric gas flow (Figure 2b), a moderate linear increase was observed pointing to mass transport limitations.This is not surprising due to the high light intensity (52.93 mW cm −2 ) and the elevated temperature (100 °C) used here leading to a high reaction rate while the mass transport independent of the gas flow rate in the reactor primarily relies on diffusion from the gas phase to the photocatalyst surface.A detailed analysis of mass transport limitations and  their dependence on light intensity, temperature, and flow dynamics (laminar/turbulent flow) has been recently published. [29]lthough this is beyond the scope of this work, further improvements in X MeOH despite short contact times can be envisioned after further optimization of the reactor geometry.
The influence of temperature on X MeOH , product distribution, and catalyst deactivation was investigated by performing the reaction at 100, 125, 150, and 175°C using four individual freshly prepared catalyst plates which were not in steady-state conditions.The temporal course of X MeOH and the selectivities toward the different products are shown in Figure 3.The thermal evolution of CO 2 , CO, and H 2 before and after each photocatalytic reaction are additionally summarized in Table 2.
It is evident that the thermal gas evolution rates are much smaller than the photocatalytic gas evolution rates and close to the detection limit of the analytics.The initial X MeOH under illumination strongly increased by increasing the temperature from 100 to 175 °C.Nevertheless, at temperatures > 100 °C severe catalyst deactivation was observed and already after 7 h TOS  Thermocatalytic and photocatalytic reactions are marked in red and green, respectively.The depicted reaction pathways are strongly simplified and still under debate. [2,31]creased, whereas the selectivity toward the valuable products formaldehyde and methyl formate decreased.In comparison to earlier reports on the synergistic effect of temperature and light in thermally assisted photocatalytic reactions using Ru/TiO 2 , [8] Cu/TiO 2 , [9] or Pt/TiO 2 [2] photocatalysts, the influence of light irradiation is significant for the conditions used here.For example, a high thermal (dark) X MeOH was reported even for 100 °C using Pt/TiO 2 and the improvements of light irradiation at a reaction temperature of 160 °C were small. [2]The pronounced differences were probably caused by a higher catalyst amount (800 vs 300 mg), a lower light intensity (19 vs 52.93 mW cm −2 ) and a lower volumetric gas flow (10 vs 50 mL min -1 ) used by Hardacre and co-workers. [2]Furthermore, a synergistic effect between thermocatalytic and photocatalytic processes occurring simultaneously is assumed in several studies. [2,9,30] Although the sole photocatalytic activity for instance at room temperature is not accessible in this work, the combined photothermocatalytic reaction rate was much higher than the thermocatalytic reaction rate alone, which was below or close to detection limit.Furthermore, the initial X MeOH strongly increased at higher temperatures.Thus, a beneficial effect of temperature on the observable overall photocatalytic reaction rate constant k photo is likely.However, this effect is superimposed by the strong deactivation at higher temperatures which prevented us from comparing steady-state reaction rates at different temperatures.Thus, the exact origin of the increased initial rate must be addressed in future studies.Improved adsorption and desorption kinetics [32] or faster diffusion of protons on the catalyst surface to the reduction sites [12] may represent possible explanations as the activation energies of photocatalytic reactions are often considered small. [33]he temperature-dependent deactivation of the photocatalyst was also studied on one photocatalyst plate using a step experiment performed in the temperature range from 80 to 150 °C (Figure S9, Supporting Information).At 80 °C, steady-state conditions were obtained.Subsequently, the reactor temperature was increased by 10 °C every 90 min while observing the transient changes in the product distribution and X MeOH .With every temperature step, the (initial) X MeOH increased, and for the temperature range between 80 and 110 °C a slight activation, i.e., an increase in methanol conversion with time, was observed.For temperatures > 110 °C, deactivation during TOS was noticed with X MeOH decreasing at a similar rate for all temperature steps per-formed above 120 °C.Considering that the overall trend obtained at one individual photocatalyst plate is similar to the results presented in Figure 3, it is clear that catalyst deactivation should be a major concern.The deactivation of the photocatalyst also resulted in a lower X MeOH obtained at a reactor temperature of 80 °C and even after an additional 6 h TOS at lower reaction temperature, a full recovery of the catalysts was not observed.Thus, deactivation appears to be partially irreversible, and catalyst stability depends strongly on the chosen reaction temperature.Most likely, deactivation is caused by the formation of organic residues like formates, CO, or other carbon deposits which block the catalyst surface. [2,34]It seems that their formation became faster than their (oxidative) removal at higher temperatures leading to the rapid deactivation of the catalyst which, if at all, was only slowly reversed at temperatures < 110 °C.
Even though X MeOH was strongly dependent on the reaction temperature and the rate of deactivation, a rather constant product distribution was finally obtained at each individual temperature used.Moreover, the temperature ramp measurements revealed that the formaldehyde selectivity increased in the temperature range between 80 and 110 °C.Likely, minor temperature increases enhanced formaldehyde desorption, thereby reducing its further oxidation toward CO 2 in agreement with a decrease in the CO 2 selectivity in the respective temperature range.It is also evident that the water-gas shift reaction (WGSR, Equation ( 9)) is not of relevance for the conditions used here.Platero et al. [9] studied the effect of temperature on the evolution of CO 2 , CO, and H 2 between 20 and 200 °C over Cu/TiO 2 and observed a decrease in the CO selectivity with increasing temperature due to the WGSR becoming relevant at elevated temperatures.As we observed an increasing CO selectivity with increasing temperature, we assume that the WGSR is not catalyzed by 0.1PtSTO in the studied temperature range.
The influence of different mole fractions of methanol (y MeOH ) and water (y H2O ) was studied by changing the mole fraction of one component while keeping the other one constant.Both the variation of y MeOH and y H2O are illustrated in Figure 5 and showed similar trends.With an increasing y MeOH /y H2O ratio X MeOH decreased.Furthermore, the formation of formaldehyde and methyl formate was favored over CO 2 formation and only at high y MeOH /y H2O ratios significant amounts of DMM were detected.Despite these changes, the CO selectivity appeared to be independent of the precise y MeOH /y H2O ratio and was constant throughout the experiment.
To further evaluate the role of water, a measurement without water was performed (Figure 5b).X MeOH dropped to almost zero, and the product distribution was significantly changed.CO 2 formation was inhibited and the selectivity to formaldehyde dropped while a slight increase in CO and methyl formate selectivities was observed.The results generally resemble the findings of Chiarello et al. [11,12] who studied the influence of the y MeOH /y H2O ratio for gas-phase methanol oxidation in a recirculating gas-phase reactor at 55 °C using noble metal-modified TiO 2 .They observed a similar product spectrum with H 2 , formaldehyde, formic acid, CO 2 , and CO as most abundant products.Also, a high CO 2 selectivity was observed at low y MeOH /y H2O ratios, whereas an optimum in the evolution rates of formaldehyde, formic acid, and H 2 was found at a medium y MeOH /y H2O ratio.The group proposed that CO is formed via the dehydration of formic acid according to Equation (5).As the formation of formic acid requires the uptake of oxygen from H 2 O, the CO evolution rate dropped at a high y MeOH /y H2O ratio.In our case, no formic acid was observed likely due to the presence of Pt-containing catalysts causing thermal decomposition at the higher temperatures used in this study (100 °C) Also, we observed an increase in the CO selectivity in the absence of water indicating that CO was also formed by another reaction pathway which may be the decomposition of methanol (Equation ( 2)).
Importantly, the oxidation of methanol to formaldehyde is strongly dependent on the presence of water in the feed gas.Although the direct oxidation of methanol by photogenerated holes is generally feasible, almost no reaction was observed in the absence of H 2 O indicating a more complex role of H 2 O. On the one hand, H 2 O or adsorbed OH groups can be oxidized by photogenerated holes.The resulting hydroxyl radicals are considered as the most reactive oxygen species and indirectly oxidize methanol by weakening the C-H bond. [35,36]This mechanism likely dominates at low y MeOH /y H2O ratios as indicated by the increased selectivity to CO 2 in this regime.On the other hand, transport of protons (H + ) is essential and water adsorption facilitates the diffusive transport to the reduction sites thus enabling H 2 formation.In this way, water acts as an efficient diffusion medium for H + through neighboring hydroxyl groups and drastically improves the overall reaction rate. [12]inally, a variation of the light intensity was performed and the corresponding apparent quantum yield (AQY, Equation ( 10)) was calculated.AQYs were derived from the product evolution rates ṅi and the number of incident photons ṅph which depends on the power P, the catalyst area A cat and the photon energy multiplied with the Avogadro number E ph • N A .The number of photons  i needed for the formation of a particular product i was defined based on the current reaction mechanisms. [37]For the oxidation of methanol to CO 2 , CO, formaldehyde, methyl formate, and DMM, the transfer of 6, 4, 2, 4, and 2 photogenerated holes was considered, respectively.Furthermore, scattering effects were neglected.
Chorkendorff and co-workers [38] studied the difference between front side illumination (FSI) and backside illumination (BSI) using methane photooxidation over thin TiO 2 layers (7 nm -1.8 μm).The direct comparison showed a wavelengthindependent asymptotic behavior of CH 4 conversion as function of the layer thickness for the FSI configuration, whereas the BSI configuration revealed an optimum layer thickness.Here, the exact determination of the 0.1PtSTO layer thickness was not possible due to the high roughness of the sand-blasted quartz substrate.Nevertheless, with a density of  SrTiO3 = 4.81 g cm −3 and mean catalyst loading of 5 mg cm −2 , the theoretical thickness was calculated to be ≈10 μm.Thus, we work in the asymptotic regime described for the FSI configuration, and small layer thickness variations should not influence the activity significantly.Figure 6a illustrates the correlation between AQY, X MeOH, and UV light intensity.A higher UV light intensity led to a higher X MeOH and lower AQY following a logarithmic dependence as observed before. [19,39]This behavior can be explained by the increased concentration of photogenerated charge carriers in SrTiO 3 which led to a high X MeOH , while simultaneously charge carrier recombination was accelerated.Additionally, the observed curves may be flattened due to kinetic limitations at high light intensity. [40]he product distribution (Figure 6b) was only slightly changed toward a higher CO 2 selectivity which is in contrast to a previous study by El-Roz et al. [41] who observed significantly more CO 2 at higher light intensities compared with low intensities in the  photooxidation of methanol with O 2 .This was attributed to the high amount of reactive O 2 − which was produced from the reduction of O 2 by photogenerated electrons.Without O 2 , the most reactive oxidant is the hydroxyl radical produced by the oxidation of H 2 O from photogenerated holes (see Equation (12)) resulting in the formation of CO 2 . [36]Generally, the production of hydroxyl radicals is not the rate-limiting step for CO 2 formation.
Based on the presented process parameter variation, process parameters sets were chosen to either achieve high formaldehyde selectivity or to optimize the H 2 yield by maximizing the CO 2 selectivity.The conditions of the respective experiments are summarized in Table 3, and the selectivities toward the various products are presented in Figure 7 as a function of time.
Process optimization toward selective formaldehyde formation resulted in a minor contribution of the thermal reaction and an overall low photocatalytic X MeOH of ≈5.9 % (12 h TOS).Already after a short induction period, both X MeOH and the product distribution remained stable throughout the measurement and a selectivity toward formaldehyde of 76% was obtained, while methyl formate was formed with a selectivity of 18%.Importantly, CO 2 and CO were only minor byproducts under the conditions employed here.The formation of carbon-based products was accompanied by a H 2 evolution rate of 75 μmol h −1 in agreement with the low X MeOH and the general reaction stoichiometry for methanol dehydrogenation to formaldehyde (Equation ( 3)).Overall, the results appear to be similar to the values obtained during the residence time variation shown in Figure 2 for a flow rate of 100 mL min −1 .The similarities are caused by the minor influence of higher light intensities on the product distribution and the compensating effect of the reaction temperature.A slightly higher reaction temperature may be beneficial for formaldehyde formation, yet it would also lead to an increase in CO selectivity.A lower reaction temperature would further decrease X MeOH and favor methyl formate formation.Thus, it seems that the selectivity to formaldehyde is generally limited to values < 100 % in the gas-phase reactor used here.
Contrasting the optimized reaction conditions used for high formaldehyde selectivity (Figure 7a,b), optimization of the conditions toward hydrogen evolution (Figure 7c,d) already resulted in a high thermal X MeOH of ≈9% with 80% selectivity to CO 2 .Besides that, full conversion was achieved for the first 7 h of illumination which is not surprising considering that for similar reaction conditions almost full conversion of methanol had been observed before.The initial selectivities to CO 2 , CO, and formaldehyde were 53%, 42%, and 5%, respectively.Interestingly, after 7 h of continuous illumination a steady decrease in X MeOH was noticed, which was accompanied by a decrease in CO 2 and CO selectivity, while the selectivity of formaldehyde and methyl formate constantly increased.Again, the origin of the strong initial changes is presumably related to the formation of organic residues on the photocatalyst surface, which are formed by secondary reactions of formaldehyde.
The presented results clearly demonstrate the strong influence of the reaction conditions on the observed photocatalytic performance.It is particularly important to note that short contact times, moderate temperature and a medium y MeOH /y H2O ratio favor selective formaldehyde evolution over subsequent oxidation to CO and CO 2 .Longer contact times as usually used in the available literature as well as low methanol to water ratios favor the evolution of CO 2 resulting in optimized H 2 yields.In the light of our investigation, the high selectivity to formaldehyde reported by Chiarello et al. [5,11,12] compared with other studies may thus be attributed to the lower reaction temperature (55 °C), short contact times due to the low catalyst mass (14 mg), and a relatively large y MeOH /y H2O ratio.Because of the mode of operation and the recirculating gas-phase reactor design, formaldehyde was trapped in the liquid phase without catalyst, and followup reactions to CO 2 and methyl formate were prevented in their work.3), X MeOH a) and product distribution with H 2 evolution b).Photocatalytic gas-phase methanol conversion with optimized conditions for high CO 2 selectivity (Table 3), X MeOH c) and product distribution with H 2 evolution d).Note, the difference in the MeOH molar amounts is caused by the differing reaction conditions required to enable either high formaldehyde selectivity a,b) or high X MeOH c,d).
Overall, thermally assisted photocatalytic methanol conversion appears to be a highly flexible process that enables the production of formaldehyde even in a continuous operation mode.Followup studies should focus on the general reaction mechanism on SrTiO 3 photocatalysts, the cause of photocatalyst deactivation, which has been observed particularly for high reaction rates, and a detailed evaluation of the process economics and energy demand.In this study, the energy demand is primarily defined by the illumination intensity, whereas variations in the temperature contribute only to a minor extent.
Although the reaction mechanism for methanol oxidation (Equations ( 1) -( 7)) is considered well understood for TiO 2 , the abundance of different surface intermediates and spectator species may differ for SrTiO 3 and even result in photocatalyst deactivation by a high surface coverage of, among others, formates and CO poisoning of Pt nanoparticles. [34,42]Photocatalyst deactivation may also stem from transient changes in the Pt oxidation state, [43] or changes in the defect concentration in the SrTiO 3 lattice. [44]Improvements in thermally assisted photocatalytic methanol conversion can be obtained by modification of the surface and bulk properties of SrTiO 3 .[22] In future studies, we will therefore focus on the mechanism of the thermally assisted photocatalytic methanol conversion over SrTiO 3based catalysts and connect it to the results presented in this study.Also, changes in the defect structure or the Pt oxidation state and their influence on the catalytic reaction will be evaluated.

Conclusion
Gas-phase photooxidation of methanol using water vapor was investigated over 0.1 wt.% Pt/StTiO 3 covering a broad process parameter space, which allowed us to identify the significant effect of residence time on methanol conversion and, more importantly, on product distribution.Formation of formaldehyde was favored especially at short contact times allowing for a lightdriven production of a valuable bulk chemicals from methanol.We also demonstrate the importance of achieving steady-state conditions typically requiring extended time on stream.Especially during the initial phase, the conversion of methanol was changing strongly due to catalyst activation and deactivation which was even more pronounced at temperatures above 110 °C.Generally, our results highlight that the correlation between residence time, light intensity, and catalytic performance is often not linear, and understanding the effect of each parameter is therefore required for process optimization.

Experimental Section
Photocatalyst Synthesis: Photodeposition of Pt on commercial SrTiO 3 (Sigma Aldrich, 99%) was carried out in a double-walled liquid-phase reactor equipped with a Hg lamp (500 W, 59 mW cm −2 , UV Consulting Peschl) described in detail elsewhere. [46]Platinum photodeposition was performed using a suspension of 2 g SrTiO 3 in 500 mL H 2 O (HPLC grade) and 50 mL methanol (VWR Chemicals, 100%) containing H 2 PtCl 6 (Sigma Aldrich 99.995%) in the respective amount to achieve Pt loadings ranging from 0.025 to 2.0 wt.%.Illumination was performed for 1 h under continuous stirring.During illumination, the temperature of the reactor walls was maintained at 30 °C while the immersion Hg lamp was cooled to 10 °C.The modified photocatalyst was obtained after filtration, washing (HPLC grade water, > 500 mL) and freeze-drying (−50 °C for 24 h).The samples were labeled according to their Pt content in wt.%.For example, 0.1PtSTO refers to 0.1 wt.% Pt deposited on SrTiO 3 .Extensive characterization of the different photocatalyst, as described in the supporting information, was performed by N 2 physisorption measurements (BET method), X-ray diffraction (XRD), diffuse reflectance UV-vis spectroscopy, X-ray photoelectron spectroscopy (XPS), high-resolution and high angle annular dark-field scanning transmission electron microscopy (HR-TEM, HAADF-STEM), and energy-dispersive X-ray spectroscopy (EDX).
Photocatalytic Gas-Phase Measurements: photocatalytic methanol reforming in the gas-phase was performed using a flow reactor described in detail in the supporting information (see Figure S10, Supporting Information for a flow chart).The amounts of H 2 O and methanol in the N 2 stream were controlled via two separate temperature-controlled saturators (Figures S11 and S12, Supporting Information).The photocatalyst (≈300 mg suspended in 2-propanol) was spray coated onto a quartz glass plate and placed into the aluminum flat-plate reactor (Figure S13, Supporting Information), allowing for top illumination using an LED array with 30 individual LEDs (Figures S14 and S15, Supporting Information).Quantification of gaseous compounds in the effluent gas stream with high time resolution was achieved by a combination of gas-phase FTIR spectroscopy (Table S2, Supporting Information) and a multichannel gas analyzer (X-STREAM, Emerson).Photocatalytic measurements were performed after a pretreatment to remove organic residues. [47]Before the pretreatment, the reactor was flushed with N 2 (50 mL min −1 ) to ensure an O 2 -free atmosphere and heated to 100 °C to prevent H 2 O condensation and enable easier product desorption.After 1.5 h, the dry gas feed was changed to humid N 2 (4.5% H 2 O), and the sample was illuminated by UV light for 18 h.After that, the reactor was flushed with N 2 .Photocatalytic methanol oxidation in the gas phase was subsequently performed in the presence of water vapor.The molar amounts of methanol and H 2 O were defined using preheated saturators.To allow for a reliable determination of the photocatalytic activity, the thermal activity (light off) was recorded for at least 30 min before and 60 min after illumination.To achieve reliable trends while studying the influence of the various reaction conditions on the photocatalytic performance, each measurement series was usually performed with one freshly prepared catalyst layer.If not stated otherwise, individual measurement points resemble mean values (30 min) obtained after at least 3 h of reaction time.
Methanol conversion X MeOH as well as the selectivities to the main products S P were derived from the molar feed flow of methanol ṅMeOH,0 , the effluent methanol molar flow ṅMeOH and the products ṅP according to Equations ( 13) and ( 14), respectively.The stoichiometry factor  depends on the specific compound as shown in Equations (1-8).

Figure 2 .
Figure 2. Influence of the total gas flow rate in the range of 25 to 100 mL min −1 on X MeOH and product distribution a) and correlation between reaction rate and total gas flow rate b); T = 100 °C, y MeOH = 1.5%, y H2O = 4.5%, I = 52.93mW cm −2 .

Figure 4 .
Figure 4. Schematic illustration of the thermocatalytic (left) and photothermocatalytic (right) processes.Thermocatalytic and photocatalytic reactions are marked in red and green, respectively.The depicted reaction pathways are strongly simplified and still under debate.[2,31]

Figure 4
schematically illustrates the processes in a strongly simplified way where k thermal and k photo describe the observable overall reaction rate constants of the thermocatalytic and photocatalytic reactions.

Figure 7 .
Figure 7. Photocatalytic gas-phase methanol conversion with optimized conditions for high formaldehyde selectivity (Table3), X MeOH a) and product distribution with H 2 evolution b).Photocatalytic gas-phase methanol conversion with optimized conditions for high CO 2 selectivity (Table3), X MeOH c) and product distribution with H 2 evolution d).Note, the difference in the MeOH molar amounts is caused by the differing reaction conditions required to enable either high formaldehyde selectivity a,b) or high X MeOH c,d).

Table 2 .
CO 2 , CO, and H 2 evolution rates of the thermal reaction before and after UV illumination of the catalyst at 100, 125, 150, and 175 °C.
X MeOH appeared to be independent of the reaction temperature.Despite the similarities in X MeOH , the product selectivities toward carbon-based products were strongly temperature-dependent.With increasing temperature, CO and CO 2 selectivities

Table 3 .
Reaction conditions favoring formaldehyde selectivity or maximizing CO 2 selectivity and H 2 yield.