Systematic Evaluation of Reaction Phase Effects on Photocatalytic CO2 Reduction Using Cu‐Doped SrTiO3 Authors: Mohammed A. M. Bajiri, Niqab Khan, Julio Cesar Camilo Albornoz Diaz, Gabriel Natulini Vieira, Yara Jaqueline Kerber Araújo, Lucas D. Paquini, Abdo Hezam, Fabio H. B. Lima, Heberton Wender, Renato V. Gonçalves Published: 2026 ABSTRACT ABSTRACTCarbon dioxide conversion into value‐added products offers a promising strategy to mitigate greenhouse gas emissions while enabling sustainable fuel production. Among the available approaches, photocatalytic CO2 reduction has attracted significant attention, particularly when combined with advanced catalyst design and reaction engineering. In this work, Cu‐doped SrTiO3 is employed as a model photocatalyst to systematically investigate the influence of reaction phase on CO2 reduction performance across three distinct reaction configurations: solid–liquid (SL), solid–gas (SG), and pressurized solid–gas (PSG). The results demonstrate that the reaction phase strongly affects both activity and product selectivity. The SG configuration exhibits the highest performance, reaching accumulated yields of 11.13 µmol g−1 of CH4 and 2.32 µmol g−1 of CO after 4 h of reaction, owing to enhanced reactant accessibility and reduced mass‐transfer limitations. In contrast, the SL configuration shows the lowest gaseous‐product activity due to limited CO2 solubility and accumulation of dissolved intermediates, while favoring the formation of oxygenated liquid products such as methanol and ethanol. The PSG system displays intermediate performance, where increased CO2 availability is partially offset by hindered product desorption. These findings highlight reaction‐phase engineering as a key factor governing photocatalytic CO2 conversion efficiency and selectivity. Introduction Burning fossil fuels releases large amounts of CO2, a major driver of global warming and one of today's most urgent environmental problems [1, 2]. Resolving the ecological issue by capturing CO2 and converting it to environmentally friendly energy sources is a great topic of interest. Photocatalytic CO2 reduction has attracted much attention because it is a green way to directly convert CO2 to other useful chemicals or fuels like CO, HCOOH, CH3OH, HCHO, and CH4 [3]. Among different developed semiconductors for photocatalytic CO2 conversion, SrTiO3 is considered a potential candidate due to its unique properties like low toxicity, low cost, and high thermal and photochemical stabilities [4]. However, pristine SrTiO3 still exhibits relatively low efficiency, mainly due to its high charge recombination rate, poor CO2 activation at the surface, and limited solar spectrum utilization resulting from its wide bandgap (> 3.0 eV) [5]. Thus, further improvements are required to suppress charge recombination and to engineer the bandgap for harvesting a large portion of the solar spectrum. Doping SrTiO3 with aliovalent cations has proven effective in modifying the band structure, as it introduces defect levels that contribute to bandgap narrowing [6]. For instance, Cu‐doping of SrTiO3 preferentially replaces Ti4+ sites with Cu2+ or Cu+ ions, due to geometric reasons such as ionic radius and coordination number, creating oxygen vacancies and intermediate energy states [7]. These defect levels narrow the effective bandgap, enhancing visible‐light absorption, and act as trapping sites to promote charge separation, thereby reducing electron‐hole recombination [8]. Moreover, Cu dopants improve CO2 adsorption and activation at the catalyst surface, providing additional active sites for photocatalytic reduction [9]. Sun et al. demonstrated that Fe, Co, Ni, and Cu‐doped SrTiO3@TiO2 prepared via hydrothermal synthesis broadened the photoresponse into the visible region, with improved CO2 reduction in water to selectively yield C1 products, attributed to Ti3+ species formed by charge compensation [10]. Ding et al. showed that controlling the Cu+/Cu2+ and Cu+/Cu0 ratios in Cu‐doped SrTiO3 nanofibers modulates methanol selectivity, with a linear correlation between the Cu+/Cu2+ ratio and methanol generation rate [9]. Similarly, Li et al. reported that slight Cu2+ doping at the B‐site of SrTiO3 forms stable CuO6 octahedra, enhances charge transfer, and tunes the electronic structure, resulting in more efficient CO2 adsorption and over an order‐of‐magnitude improvement in CO2 electroreduction activity and selectivity [11]. While these studies demonstrate the intrinsic benefits of Cu‐doping to tailor the electronic structure and surface reactivity of SrTiO3‐based photocatalysts, the reaction environment is equally critical for improving performance in CO2 reduction. In fact, many reaction environments like solid–liquid (SL) and solid–gas (SG) phases, including under pressurized CO2, have been adopted to effectively reduce CO2 to hydrocarbon fuels. The SL configuration has shown limited CO2 solubility, reduction competitions of CO2 and water molecules, and preventing efficient interaction between the catalyst and CO2 [12]. However, the SG configuration reactors provide advantages like enhanced reactant mixing and regulation of the CO2/H2O ratio, resulting in superior CO2 adsorption on the catalyst surface [13, 14]. In addition, the reaction phase (SG, SL, or SG under pressure) substantially affects the production rates of the hydrocarbons. This work highlights the influence of the reaction phase on the photocatalytic CO2 reduction capability of Cu‐doped SrTiO3. Our findings demonstrate that the liquid phase exhibited the lowest activity, while the SG with pressure phase showed comparable activity to the SG phase. These differences could be attributed to the formation of dissolved products in the liquid phase and the slightly more difficult desorption caused by increased surface pressure in the SG with pressure phase. These findings demonstrate the significance of suitable reaction conditions to enhance the efficacy of photocatalytic CO2 reduction. Results and Discussion Structural and Morphological Studies In this work, we employ our previously optimized 2%Cu‐doped SrTiO3 (Cu:STO) sample [15] as a benchmark material to systematically investigate the influence of different reactor configurations on photocatalytic performance. By using this well‐characterized and high‐performing material, we ensure that any observed differences arise primarily from the reaction environment rather than from variations in material properties. The crystal structures of the STO and Cu:STO samples were confirmed by X‐ray diffraction (XRD) analysis, revealing differences in their crystal structures after the Cu‐doping. Figure 1a,b exhibits representative XRD patterns of STO and Cu:STO samples, respectively. The XRD patterns confirmed the presence of SrTiO3 as the main crystalline phase along with TiO2 as a secondary phase in both undoped STO and Cu:STO. The refinements of the crystalline structure of pure STO and Cu:STO using the Rietveld method were performed assuming substitutional doping of Cu into the structure of the STO in the position of Ti, and with a secondary phase of TiO2 [15]. The refinement indicated that the TiO2 fraction was approximately 3.2% in STO and increased to 11.1% in Cu:STO. The detailed Rietveld refinement parameters for STO and Cu:STO are summarized in Table S1. The diffraction peaks were indexed using standard data from the ICSD card no. 19796 for STO and no. 121632 for TiO2, which served as references for Rietveld refinement. The presence of TiO2 is expected, likely resulting from a reaction between titanium isopropoxide and atmospheric moisture or water [16, 17]. Consequently, the addition of Cu(CO2CH3)2.H2O, an essential material for copper doping, directly introduces water into the reaction, which increases the fluidity needed for TiO2 formation. Furthermore, the increased TiO2 content in Cu:STO resulting from the Cu‐doping process may indicate phase segregation or the formation of secondary phases [18]. These structural changes are also reflected in the morphology of the material, as SEM images (Figures S1 and S2) reveal a noticeable reduction in particle size after Cu incorporation, suggesting that the dopant influences both phase evolution and particle growth during synthesis. X‐ray diffraction (XRD) patterns with Rietveld refinement. (a) STO and (b) Cu:STO show the ratio between the SrTiO3 and TiO2 phases in each sample, along with their respective space groups. Photocatalytic CO2 Reduction The photocatalytic CO2 reduction performance of Cu:STO was evaluated under three distinct interfacial configurations: SL, SG, and pressurized solid–gas (PSG) phases, as illustrated in Figure 2. In all experiments, identical catalyst loading, irradiation conditions, and analytical procedures were employed to ensure reliable comparison between the different reactor configurations. Significant differences in activity and selectivity were observed across the three configurations. Schematic illustration of the photocatalytic reactors used in this study: (a) solid–liquid (SL) phase using a quartz reactor with catalyst suspension, (b) solid–gas (SG) phase with catalyst exposed to a flowing CO2 gas stream, and (c) pressurized solid–gas (PSG) phase reactor operated under controlled pressure (up to 3 bar). Arrows indicate CO2 inlet and product outlet, while illumination is provided by a Xenon Lamp at 500 mW.cm−2 through a quartz window. Figure 3 summarizes the photocatalytic performance of the Cu:STO sample under three different reaction environments: SL, SG, and PSG. In the SL configuration (Figure 3a), the photocatalytic activity was the lowest among the three systems, yielding 1.3 µmol g−1 of CO and 6.3 µmol g−1 of CH4 after 240 min of reaction. This reduced performance can be attributed to the limited solubility of CO2 in water, which restricts its availability at the catalyst surface [19]. Additionally, competitive adsorption of water molecules and the accumulation of dissolved intermediates hinder efficient charge transfer and product desorption. These factors not only suppress the formation of gaseous products but also favor the generation of liquid‐phase products such as methanol and ethanol. Photocatalytic CO2 reduction of Cu:STO with (a) SL, (b) SG, and (c) PSG configurations. The error bars are the standard deviation of three independent measurements. The SG configuration (Figure 3b) exhibited the highest photocatalytic performance, yielding 11.1 µmol g−1 of CH4 and 2.3 µmol g−1 of CO after 240 min. This enhanced activity can be attributed to the direct gas‐solid interaction, which maximizes CO2 accessibility at the catalyst surface and promotes more efficient utilization of active sites. Unlike SL systems, the absence of a solvent eliminates diffusion limitations associated with CO2 dissolution and transport, enabling faster mass transfer and more effective reactant delivery [20]. Furthermore, the SG configuration environment reduces competitive adsorption from solvent molecules and suppresses side reactions, allowing for more efficient charge transfer and reaction pathways toward hydrocarbon formation [20, 21, 22]. The improved interfacial conditions also facilitate rapid desorption of products, preventing site blocking and sustaining catalytic turnover. As a result, the SG configuration not only enhances overall activity but also improves selectivity toward CH4, highlighting the critical role of the reaction interface in governing photocatalytic CO2 reduction. The PSG configuration (Figure 3c) exhibited intermediate photocatalytic performance compared to the SL and SG systems. Increasing the CO2 pressure to 3 bar enhances the availability of CO2 at the catalyst surface, which can promote adsorption and potentially facilitate its activation [23]. However, this increased surface coverage also alters the adsorption‐desorption equilibrium, leading to significant changes in interfacial reaction dynamics. In particular, the higher CO2 pressure may stabilize surface‐bound intermediates and products, making their desorption more difficult. This can result in partial blockage of active sites, thereby limiting the continuous turnover of the catalytic cycle. Consequently, although reactant availability is improved under pressurized conditions, the restricted product desorption and modified surface equilibria can offset these benefits, preventing further enhancement in overall activity compared to the SG configuration [24]. Thus, excessive CO2 pressure in our system, while beneficial for increasing reactant concentration, may negatively impact catalytic performance by hindering surface regeneration and altering reaction pathways. Overall, the SG configuration demonstrated superior efficiency, producing approximately twice as much as the SL and 1.5 times more than the PSG configuration. This system also showed particularly high selectivity toward CH4 under anaerobic conditions. These findings highlight the importance of reactor design and reaction conditions in determining product distribution and efficiency in photocatalytic CO2 reduction. The CO2 photoreduction stability of the Cu:STO was evaluated over three consecutive cycles under all reaction configurations (Figure 4). In the SL configuration, the evolution of gaseous products (CO and CH4) is shown in Figure 4a, where both products exhibit stable production over three 4‐h cycles, indicating good catalytic stability under these conditions. However, the overall yield of gaseous products remains relatively low, which can be attributed to the limited solubility of CO2 in water, restricting its availability at the catalyst surface and thereby suppressing gas‐phase product formation. Instead, this behavior is primarily driven by the solubility of CO2 and reaction intermediates in the liquid medium, which enhances their residence time at the catalyst surface and enables further transformation into C1 and C2 products. Recyclability of Cu:STO with different phases under each configuration: (a) SL (gas products), (b) SL (liquid products), (c) SG, and (d) PSG configuration. The error bars are the standard deviation of three independent measurements. In parallel, analysis of the liquid product of the SL configuration (Figure 4b) reveals the formation of oxygenated products, specifically methanol (CH3OH) and ethanol (C2H5OH), after prolonged reaction times (up to 16 h). Methanol reaches a maximum concentration of approximately 70.8 ppmmol g−1 (66.33 µmol g−1) within the first 4 h and subsequently decreases, whereas ethanol shows a continuous increase, reaching ∼500 µmol g−1 after 16 h. This opposite trend suggests a sequential reaction pathway, where C1 intermediates are progressively converted into C2 products via C─C coupling reactions on the catalyst surface. The observed behavior indicates that methanol likely acts as an intermediate, undergoing further transformation into ethanol under irradiation [25, 26]. Such pathways are consistent with the accumulation of dissolved intermediates in the liquid phase, which facilitates secondary reactions that are less favorable in gas‐phase systems. Moreover, the presence of water promotes alternative reaction routes involving proton‐coupled electron transfer, favoring the formation of oxygenated products over hydrocarbons. Thus, these results demonstrate that the SL configuration promotes the formation of liquid‐phase products (CH3OH and C2H5OH), while limiting the evolution of gaseous species (CO and CH4). This behavior is primarily driven by the solubility of CO2 and reaction intermediates in the liquid medium, which enhances their residence time at the catalyst surface and enables further transformation into C1 and C2 products. To confirm this selectivity shift, an electron‐equivalent balance was calculated after 4 h of reaction. Gas products (CO and CH4) accounted for only 5.4% (53.0 µmol e− g−1) of the consumed photogenerated electrons, while the reliably quantified liquid products (methanol and ethanol) accounted for 94.6% (928.50 µmol e− g−1) (calculation details in Supporting Information) [27]. This demonstrates that the catalytic activity in the SL phase is not inherently poor, but is overwhelmingly directed toward liquid products. According to existing literature, the reduction of SL CO2 is assumed to occur through direct sequential surface processes. Contemporary mechanistic investigations indicate that methanol and methane function as competitive, parallel products originating from common surface intermediates, rather than through subsequent partial oxidation of methane [28]. Simultaneously, the CO2 reduction produces essential C1 intermediates like formate (*HCOO), formaldehyde (*CH2O), and methoxy groups (*CH3O). Moreover, the synthesis of multi‐carbon compounds or ethers can transpire through the surface coupling of these C1 fragments. As indicated, the interaction of methoxy intermediates (*CH3O) with formaldehyde (*CH2O) offers a feasible route to produce dimethyl ether (CH3OCH3) and H2O [26, 29]. The stability tests performed under SG and PSG configurations (Figure 4c,d) demonstrate sustained photocatalytic activity over multiple reaction cycles, confirming the robustness of the Cu:STO catalyst under gas‐phase conditions. In both systems, CH4 production remains dominant, while CO selectivity is either slightly reduced or remains nearly constant, indicating a preferential reaction pathway toward methane formation. Notably, in the SG configuration (Figure 4c), an increase in CH4 production is observed during the second cycle, which may be attributed to a surface activation process. Such behavior has been previously reported and is often associated with the formation of additional active sites, improved charge separation, or the stabilization of catalytically favorable surface states during initial operation [30, 31]. The PSG system (Figure 4d) also exhibits stable performance, although without a pronounced activation effect, suggesting that elevated pressure conditions may limit dynamic surface restructuring compared to ambient SG conditions. Additionally, this incremental activation is absent in the SL phase (Figure 4a), where fast adsorption competition by liquid water creates an immediate, though kinetically limited, steady‐state surface coverage. Overall, these results demonstrate that Cu:STO maintains high stability and catalytic efficiency across all reaction environments (SL, SG, and PSG), without significant structural or morphological modification, as further supported by post‐reaction characterization (Figure 6). This highlights the suitability of this material for sustained photocatalytic CO2 reduction under varying operational conditions. Multiple independent methods were used to further evaluate the dynamic stability of the Cu:STO catalyst and to confirm the origin of the carbon products. First, the identification ruled out adventitious carbon contamination and evaluated the operational stability of the photocatalytic process and production by time‐resolved online mass spectrometry (MS) was conducted under a continuous CO2 flow (20 mL min−1). The system was subjected to alternating light ON/OFF cycles to monitor the real‐time generation of volatile species. As depicted in Figure 5a, the generation of characteristic molecular fragments, including m/z = 16 (CH4 ion molecular) and m/z = 26 corresponding to C2 species, such as C2H4 (see Figure S3 and Reference [15]), exhibited a reversible dependence on UV–vis irradiation. The online MS detected the volatilized methanol at m/z = 31 (CH2OH+), even though its vapor‐phase concentration was near the detection limit of the instrument compared to ethanol. Concurrently, the CO2 signal (m/z = 44) is presented as a constant, straight baseline, as it acts as the continuous carrier and reactant gas for the reduction process (Figure 5a and Figure S3). Upon illumination, the ion currents for these product species rise sharply with the same pattern, and they decay rapidly when the light source is turned off. (a) Time‐resolved online mass spectrometry (MS) signals of the gaseous products and volatilized methanol during alternating light ON/OFF cycles. (b) 1H‐NMR spectrum of the liquid‐phase products following 4 h of photocatalytic CO2 reduction in the SL configuration. Furthermore, 1H‐NMR spectroscopy was used to clearly confirm the liquid‐phase oxygenates produced in the SL configuration. The obtained spectrum exhibits the distinctive triplet and quartet splitting patterns that verify the presence of ethanol in addition to the typical singlet resonance associated with methanol, as seen in Figure 5b. The quantitative GC‐FID analysis is fully supported by this independent structural confirmation. The Cu:STO catalyst's exceptional photo‐stability and resistance to deactivation are demonstrated by its ability to consistently switch on product generation over repeated illumination cycles under a continuous flowing gas stream. More significantly, this dynamic, light‐responsive behavior unequivocally demonstrates that the hydrocarbons and oxygenates found are the result of ongoing, light‐driven photocatalytic reduction of the supplied CO2. Instead of sustained, cyclical regeneration precisely timed with the photon flux, the product evolution would show a continuous decay profile if they were produced by the breakdown of leftover organic contaminants on the catalyst surface. Post‐reaction SEM, XRD, and XPS analyses were carried out to further clarify the catalyst's structural and electronic stability (Figures 6 and 7). These findings imply that the Cu:STO is extremely stable regardless of the reaction phase, as is the stable morphology and structure seen in all phases (SL, SG, and PSG). Additionally, the Cu:STO's structural integrity is confirmed by the post‐stability XPS analysis, which offers a clear understanding of the elemental composition. FE‐SEM of Cu:STO after the recyclability of different reaction phases (a) SL, (b) SG, and (c) PSG. (d) The XRD after the recyclability test of Cu:STO with different reaction phases. The XPS of Cu:STO and after the recyclability test with different phases: (a) O1s, (b) Sr 3d, and (c) Ti 2p. PSG: XPS after the recyclability test for a SG reaction under 3 bar of CO2 pressure. SG: XPS after the recyclability test for a SG reaction at ambient pressure. SL: XPS after the recyclability test for a liquid‐phase reaction. The O 1s spectra (Figure 7a) show no significant shift in binding energy for the SG and PSG configurations, indicating that the oxygen environment remains largely unchanged under gas‐phase conditions. In contrast, the SL configuration exhibits noticeable variations in the relative intensity of hydroxyl groups and adsorbed oxygen species, reflecting the strong interaction between the catalyst surface and the aqueous environment. The Sr 3d spectra (Figure 7b) remain essentially unchanged across all configurations, confirming the stability of the SrTiO3 framework. Considering the photocatalytic activities of the SL and SG systems, the PSG configuration, as shown in Figure 3c, has an intermediate photocatalytic activity. While increasing the CO2 pressure to 3 bar enhances the initial availability of CO2 at the catalyst surface, it creates a kinetic bottleneck through the excessive surface stabilization of intermediates. As indicated by post‐reaction XPS analysis (Figure 6c, Table S2), the PSG system exhibits the highest Ti3+ content (Ti3+/Ti4+ ∼2.3), followed by the SG configuration (∼1.6) and the SL phase (∼0.81). This demonstrates that gas‐phase conditions, particularly under elevated CO2 pressure, promote the reduction of Ti4+ to Ti3+ through strong CO2 adsorption and surface reconstruction processes, where the formation of surface carbonate‐like species induces charge transfer from the lattice to maintain charge balance [32]. Importantly, the correlation between the Ti3+/Ti4+ ratio and catalytic activity is non‐linear. The PSG system possesses ∼44% more Ti3+ than the SG system, yet it produces significantly less CH4 (∼8.0 µmol g−1 compared to 11.1 µmol g−1). This inverse relationship‐where a higher Ti3+ concentration results in a lower CH4 yield‐clearly indicates that the reaction is limited by product desorption. Specifically, the elevated pressure induces excessive Ti3+ concentrations, which lead to overly strong binding of reaction intermediates. Because these intermediates are bound too tightly, they fail to desorb efficiently, resulting in the blockage of active sites. This desorption bottleneck restricts the continuous turnover of the catalytic cycle, ultimately offsetting the kinetic benefits of increased reactant availability. This interpretation perfectly aligns with the Sabatier principle, which dictates that optimal catalytic performance requires balanced intermediate binding‐neither too weak to cause poor activation nor too strong to hinder product desorption [33]. Similar volcano relationships driven by adsorption strength have been recently demonstrated in gas‐phase photocatalytic systems, such as the Ti3+/Ti4+ containing TiO2/Ag2S nanowires, was reported that optimal defect‐driven dynamics were shown to govern CO2 to CH4 conversion efficiency [34]. More importantly, the Ti 2p spectra (Figure 7c) reveal pronounced differences in the Ti3+/Ti4+ ratio depending on the reaction phase. The PSG system shows the highest Ti3+ content (Ti3+/Ti4+ ≈ 2.3), followed by the SG configuration (≈1.6), while the SL phase exhibits the lowest fraction (≈0.813). This trend indicates that gas‐phase conditions, particularly under elevated CO2 pressure, promote the reduction of Ti4+ to Ti3+, leading to enhanced surface electron density and modified catalytic properties. The increased Ti3+ content under PSG conditions can be attributed to strong CO2 adsorption and surface reconstruction processes. Under high CO2 pressure, the formation of surface carbonate‐like species induces charge transfer from the lattice, reducing Ti4+ to Ti3+ to maintain charge balance [32]. This process alters the electronic structure and surface reactivity, influencing adsorption strength, intermediate stabilization, and reaction pathways. However, while higher Ti3+ concentrations can enhance CO2 activation, they may also lead to stronger binding of intermediates, which is consistent with the previously observed limitation in product desorption under PSG conditions. In contrast, the SL configuration, which favors the formation of methanol and ethanol, exhibits the lowest Ti3+ fraction. This suggests that liquid‐phase reactions proceed via a distinct mechanism, likely dominated by hydration‐mediated pathways and proton‐coupled electron transfer processes. The presence of water competes with CO2 for adsorption sites and suppresses CO2‐induced surface reduction, thereby stabilizing Ti4+ states while enabling alternative C‐C coupling pathways for oxygenate formation. It is worth noting that Cu species were not clearly detected in the XPS spectra, likely due to their low concentration and dispersion below the detection limit [15]. Nevertheless, the observed changes in Ti oxidation states and catalytic behavior strongly support the role of Cu doping in modulating the electronic structure and facilitating charge transfer processes. Overall, the combined catalytic and spectroscopic results demonstrate that Cu:STO maintains structural stability while undergoing phase‐dependent electronic modifications, with the Ti3+/Ti4+ balance playing a key role in governing activity, selectivity, and reaction pathways across different interfacial environments. Conclusion In this work, we systematically demonstrated that the photocatalytic CO2 reduction performance of Cu‐doped SrTiO3 is strongly governed by the reaction phase, highlighting the critical role of the catalyst‐reactant interface in determining activity and selectivity. Among the investigated configurations, the SG phase exhibited the highest performance, which can be attributed to enhanced CO2 accessibility, efficient mass transport, and the absence of competitive adsorption from solvent molecules. These conditions promote favorable reaction kinetics, efficient charge utilization, and selective formation of gaseous products, particularly CH4. In contrast, the SL phase showed lower activity but favored the formation of oxygenated products such as methanol and ethanol. This behavior is associated with limited CO2 solubility, competitive adsorption of water, and the accumulation of dissolved intermediates, which enable secondary reactions such as C─C coupling. The PSG configuration exhibited intermediate performance, revealing that although increased CO2 pressure enhances surface coverage and adsorption, it also alters interfacial equilibria and can hinder product desorption, thereby limiting catalytic turnover. Post‐reaction XPS analysis provides key insights into the phase‐dependent electronic structure of the catalyst. Gas‐phase conditions, particularly under pressure, promote the formation of Ti3+ species through CO2‐induced surface reduction and charge redistribution, while the SL phase preserves Ti4+ states due to the dominant role of water in surface interactions. These findings establish a clear correlation between the Ti3+/Ti4+ ratio, reaction environment, and catalytic behavior, emphasizing the importance of electronic structure modulation in photocatalytic CO2 conversion. Overall, these findings demonstrate that optimizing the reaction environment is as critical as catalyst design for achieving high‐performance photocatalytic systems. Experimental Section All chemicals and reagents used in this work were of analytical grade and were used as received without further purification. Synthesis and Characterization of Photocatalysts The Cu‐doped SrTiO3 photocatalyst was synthesized following our previously reported method, with minor modifications [15]. Briefly, 7.6 mL of titanium isopropoxide (Acros, 98%) was dissolved in 25 mL of isopropanol. Subsequently, 0.02 M copper acetate (Cu(CO2CH3)2·H2O, (Mallinckrodt, 98.7%) and 1 M strontium nitrate (Sigma–Aldrich, 99%) were added and stirred for 30 min to ensure complete dissolution. The resulting solution was then added dropwise into 15 mL of 5 M aqueous KOH under continuous stirring, followed by an additional 1 h of stirring. The pH was adjusted to 7, and the mixture was transferred to a Teflon‐lined autoclave and heated at 200°C (5°C min−1) for 4 h. After synthesis, the product was washed once with acetic acid and multiple times with deionized water to remove residual impurities, and then dried at 80°C overnight. Undoped SrTiO3 (STO) was prepared using the same procedure without the addition of a copper precursor. The Cu‐doped SrTiO3 sample was denoted as Cu:STO, while the undoped material was referred to as STO. The materials were characterized by X‐ray diffraction (XRD), field‐emission scanning electron microscopy (FESEM), and X‐ray photoelectron spectroscopy (XPS). Additional experimental details are provided in the Supporting Information. Photocatalytic CO2 Reduction Under Different Reaction Phases The photocatalytic CO2 reduction performance of Cu:STO was evaluated under three distinct interfacial configurations: SL, SG, and PSG phases, as illustrated in Figure 2. In all experiments, identical catalyst loading, irradiation conditions, and analytical procedures were employed to ensure reliable comparison between the different reactor configurations. SL Phase Photocatalytic reactions in the SL configuration were carried out in a 100 mL quartz reactor. A suspension was prepared by dispersing 50 mg of Cu‐doped SrTiO3 in 30 mL of deionized water, followed by ultrasonication for 5 min to ensure homogeneous dispersion. Prior to irradiation, the suspension was purged with high‐purity argon (99.99%) for 10 min, followed by vacuum‐air cycling to remove residual gases. Subsequently, humidified CO2 was introduced at a flow rate of 4 mL min−1 for 20 min to saturate the system. During photocatalysis, the CO2 flow was reduced to 1 mL min−1 and maintained throughout the reaction. Illumination was provided by a Xe lamp calibrated to 500 mW cm−2 with an AM 1.5 filter. The reaction proceeded for 4 h under continuous irradiation. Gaseous products were periodically analyzed using an online gas chromatograph (Agilent 7890B) equipped with a flame ionization detector (FID) coupled to a nickel‐based methanizer for enhanced detection of CO and CO2. The transfer line connecting the reactor to the GC was maintained at 80°C to prevent condensation of reaction products. Separation of CO and CH4 was achieved using an HP‐PLOT/Q column, coupled in series with an HP‐MOLESIEVE column for permanent gases. The quantitative analyses of liquid products were conducted using a GC‐FID system with a liquid autosampler (Agilent 8860, G2790A). 1H‐nuclear magnetic resonance (1D 1H‐NMR) measurements were performed to identify and qualitatively evaluate the formation of liquid products generated during the photocatalytic CO2 reduction process, particularly methanol (MeOH) and ethanol (EtOH). Spectra were acquired using an Agilent Technologies 500/54 Premium Shield spectrometer operating at 500 MHz for 1H detection. After the photocatalytic experiments, aliquots of the solution were collected and mixed with deuterated water (D2O). All spectra were recorded at room temperature using standard one‐dimensional proton acquisition sequences with water suppression. From that, 1024 scans were accumulated, employing a relaxation delay of 2s and a spectral width sufficient to cover the 2–14 ppm range, PW90: 7.7 µs, and a total procedure time of 1 h 44 min. SG Phase The SG experiments were conducted in a custom‐designed stainless‐steel reactor (4 mL) equipped with a quartz optical window to allow irradiation of the catalyst. The photocatalyst (50 mg) was evenly distributed as a thin layer at the bottom of the reactor using a mesh support to maximize exposure to the gas phase. Before illumination, the reactor was purged with humidified CO2 at a flow rate of 4 mL min−1 for 20 min to remove residual gases and establish a controlled reaction atmosphere. The gas flow was regulated by a mass flow controller (MFC) positioned upstream of the reactor. During the photocatalytic reaction, the CO2 flow was reduced to 1 mL min−1. The system was irradiated under identical conditions to the SL experiments (Xe lamp, 500 mW cm−2, AM 1.5 filter) for 4 h. Product analysis was performed using the same online GC setup and operating conditions described above. PSG Phase Photocatalytic CO2 reduction under PSG conditions was carried out in a stainless‐steel reactor with a total volume of 60 mL, designed to operate under controlled pressure. The catalyst (50 mg) was uniformly distributed at the bottom of the reactor in a similar manner to the SG configuration. The reactor was initially purged with CO2 at 4 mL min−1 for 20 min to remove residual gases and ensure CO2 saturation. Subsequently, the flow rate was reduced to 1 mL min−1, and the system was gradually pressurized up to 3 bar of CO2. In this configuration, the mass flow controller was installed downstream of the reactor, enabling regulation of gas and maintenance of internal pressure. The reaction was performed at room temperature under continuous illumination using the same Xe lamp setup (500 mW cm−2, AM 1.5 filter) for 4 h. In the SG and PSG configurations, the CO2 gas stream was humidified before entering the reactor to provide water as the electron donor. The gaseous products were analyzed using the same gas chromatography setup. Online MS Online MS measurements were performed using an OmniStar GSD320 quadrupole mass spectrometer (Pfeiffer Vacuum) equipped with a stainless‐steel capillary inlet (0.125 mm inner diameter). The mass spectrometer was directly coupled to the photochemical reactor through a capillary interface, enabling continuous monitoring of volatile and/or gaseous products generated during photocatalytic CO2 reduction. A continuous CO2 flow (20 mL min−1) was controlled using a flowmeter (AALBORG ‐GFC17A‐VAL6‐A0‐3NC‐10‐SS) and supplied to the reactor to maintain solution saturation and stable mass transport conditions throughout the experiments. An initial exploratory analysis was conducted by acquiring full mass spectra (Δm/z: 0–50 m.a.u). The measurements were taken over separate 2‐h intervals under the solar simulator (500 mW/cm−2, using a standard AMG filter), allowing qualitative identification of the volatile species generated during the photocatalytic process. Subsequently, time‐resolved MS experiments were carried out using selected ion monitoring (SIM) mode focused on the characteristic m/z molecular ion or fragments of the main CO2 reduction products and possible side products. For this case, we performed a stability test with a mass spectrum being initially recorded under dark conditions for 30 min. The solar simulator was then alternately switched on for 1 h and off for 30 min over two consecutive cycles to assess the photocatalyst activity and stability under intermittent UV–vis irradiation, as well as the relaxation behavior of the photochemical system. In this case, selected mass‐to‐charge (m/z) signals corresponding to molecular ions and characteristic fragments were continuously monitored, including m/z = 2 (H2), 15 (CH3 +), 16 (CH4 +), 26 (C2H2 +), 31 (CH2OH+), 32 (CH3OH+), and 44 (CO2 +). Conflicts of Interest The authors declare no conflicts of interest. Supporting information Supporting File: gch270145‐sup‐0001‐SuppMat.docx.