Reaction Mechanism and Role of Chlorine in Ethylene Epoxidation Revealed by in situ XPS

Reaction Mechanism and Role of Chlorine in Ethylene Epoxidation Revealed by in situ XPS | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Reaction Mechanism and Role of Chlorine in Ethylene Epoxidation Revealed by in situ XPS Luca Artiglia, Man Guo, Nanchen Dongfang, Frank Krumeich, Marcella Iannuzzi, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7141633/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Ethylene epoxidation is a cornerstone process in the chemical industry and the primary route for ethylene oxide (EO) production. Despite its significance, the reaction mechanism and the structure of the active surface remain under debate. Here, we employ in situ photoelectron spectroscopy supported by DFT to probe surface intermediates on Ag/Al2O3 under ethylene epoxidation conditions, with and without vinyl chloride promotion. Under these conditions, combustion is triggered by ethylene dehydrogenation at surface defects, assisted by lattice oxygen and yielding CO2. Vinyl chloride suppresses this pathway by reducing defect density, hindering lattice oxygen formation, and inhibiting ethylene adsorption. Chlorine binds preferentially at defect sites, passivating undercoordinated silver atoms to disfavor C–H activation. Ethylene reacts with one of the oxygen atoms of adsorbed dioxygen on reconstructed silver surfaces through a direct oxidation pathway to produce EO. These findings provide molecular-level insights and offer a rational basis for optimizing ethylene epoxidation catalysts. Physical sciences/Chemistry/Catalysis/Heterogeneous catalysis Physical sciences/Chemistry/Surface chemistry/Surface spectroscopy Physical sciences/Materials science/Theory and computation/Electronic structure Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Ethylene epoxidation (EPO) is a fundamental reaction in the chemical industry, producing ethylene oxide (EO)—a vital building block for polymers, surfactants, and antifreeze. 1 As a central pillar bridging hydrocarbon feedstocks to high-value downstream markets, it sustains global sectors like textiles, packaging, and automotive manufacturing. 2, 3 A key technical challenge in this process lies in enhancing the selectivity of silver-based catalysts to simultaneously boost process economics and mitigate CO 2 emissions. 4 While unpromoted silver catalysts exhibit limited selectivity, industrially relevant promoted catalysts can achieve EO selectivity approaching 90%. Among various promoters, chlorine-containing molecules, such as vinyl chloride (VC) and dichloroethane (DCE), are particularly effective. Kinetic analyses showed that, when cofed at ppm levels, such promoters enhance the selectivity by approximately 25% through poisoning over-oxidation sites while preserving active epoxidation centers. 5, 6 Other promoters, such as cesium and rhenium, are added in trace amounts during catalyst synthesis and also have a positive impact on the selectivity toward EO. 7, 8 , 9 Despite decades of research, the underlying mechanism governing the selectivity of ethylene oxidation remains a topic of ongoing debate. A long-standing hypothesis suggests that the reaction proceeds through the formation of an oxametallacycle (OMC) intermediate, which serves as a common precursor for both EO and acetaldehyde (AA), with acetaldehyde further oxidizing to CO 2 . 10, 11, 12 However, several key issues challenge this mechanism. First, direct in situ experimental evidence for OMC formation and its role under realistic conditions is lacking. 13, 14, 15 Second, the asymmetric structure of OMC resembles AA more than EO, suggesting a higher activation barrier for EO formation. This has led researchers to explore specific oxygen species—such as sulfate impurities or the O 5 surface oxide phase—that might lower EO’s activation barrier. 16, 17, 18, 19 Since lattice oxygen is widely known for its unselective behavior, it is necessary to identify how oxygen activation on silver leads to the selective oxidation of ethylene. 20 Previous kinetic studies show that EO and CO 2 have similar activation energies (~90–100 kJ/mol) but differ in reaction orders: –0.4 to 1 for C 2 H 4 and 0.5–1 (EO) vs. 0.2–1 (CO 2 ) for O 2 . 1, 21 Both Langmuir–Hinshelwood and Eley–Rideal mechanisms have been proposed. 4, 22 The addition of chlorine raises activation barriers and reduces oxygen species coverage, as indicated by an increased O 2 reaction order (0.7 to 1) with rising ethyl chloride (C 2 H 5 Cl) concentration. 5, 23, 24, 25 Consequently, three fundamental questions remain: (i) what are the molecular-level active sites for selective versus unselective pathways on silver surfaces? (ii) how do different oxygen species formed under reaction conditions cooperate to control selectivity? and (iii) what is the mechanistic role of chlorine promoters? In this work, in situ X-ray photoelectron spectroscopy (XPS) was used to investigate surface intermediates under reaction conditions on a powder Ag/Al 2 O 3 catalyst. 26 27, 28 Experiments were carried out with X-rays generated by a synchrotron light source in order to improve the surface sensitivity, the signal intensity and the overall resolution. In the case of EPO, most of the studies primarily focused on the collection and interpretation of O 1s spectra exposing model surfaces (silver single crystals and/or silver powder) to reaction mixtures. 19, 29, 30 In our previous in situ XPS works, we utilized silver foil as a model catalyst to investigate the nature of surface oxygen and carbon species. 31, 32 Adsorbed diatomic oxygen was found to exhibit an electrophilic behavior and to react with adsorbed ethylene to yield EO. In parallel, adsorbed ethylene underwent dehydrogenation to C 2 H x (x=1-3) which was correlated with the production of CO 2 . However, the nature of surface intermediates and active sites on supported Ag/Al 2 O 3 catalysts during in situ EPO —particularly in the presence of ppm-level chlorine containing promoters—remains largely unexplored and presents significant experimental challenges. Capitalizing on the knowledge gathered on model silver surfaces, we employed ambient pressure (AP)-XPS to directly probe catalytically active sites and reaction intermediates on Ag/Al 2 O 3 catalysts, both in the presence and absence of chlorine promotion. Combined with theoretical calculations, our findings reveal that ethylene spontaneously adsorbs and undergoes dehydrogenation at coordinatively unsaturated sites on the silver surface, leading to CO 2 formation. The introduction of chlorine promoters effectively blocks these surface defect sites and consequently enhances the selectivity towards EO. Furthermore, our results suggest that EO forms by the direct reaction between ethylene and dioxygen species activated on the reconstructed silver surface oxide. Results Deep dehydrogenation of ethylene on silver nanoparticles The Ag/Al 2 O 3 catalyst used in this work was synthesized using a low-surface area α-Al 2 O 3 support (5.2 m²/g, see Supplementary Fig. 1 and Supplementary Table 1), following industrially-relevant protocols. 33, 34 After silver deposition, the Ag/Al 2 O 3 catalyst exhibited a surface area of 4.6 m²/g. Along with the reflections of the α-Al 2 O 3 support, X-ray diffraction (XRD) displayed predominantly metallic silver (Supplementary Fig. 2), consistent with literature reports. 35 The catalytic performance of Ag/Al 2 O 3 with and without VC was evaluated in a flow reactor setup. The Ag/Al 2 O 3 catalyst was first exposed to a reaction mixture (C 2 H 4 + O 2 ) until steady state EO production rate and selectivity were established. Upon co-feeding 0.5 ppm VC, the EO production rate slightly decreased, while CO 2 formation was significantly suppressed, resulting in an overall increase in EO selectivity from 38% to 69% (Fig. 1a). Then, the gas hourly space velocity (GHSV) was adjusted at a constant temperature of 230°C to maintain identical ethylene conversion in both the absence and presence of VC. As illustrated in Fig. 1b, EO selectivity increased by approximately 20% upon VC addition at the same ethylene conversion level. These results align with typical studies reporting that chlorine-containing promoters enhance EO selectivity primarily by suppressing CO 2 formation. 8, 11, 43 To investigate the electronic structure of all elements and the reaction intermediates under in situ EPO conditions, we conducted AP-XPS experiments. Two freshly synthesized Ag/Al 2 O 3 samples were sequentially introduced in the XPS flow measurement cell. High vacuum (HV) photoemission spectra were collected and used as references before dosing the reaction feed. Each sample was then exposed to a C 2 H 4 :O 2 (2:1) gas mixture (1 mbar total pressure), either in the presence or in the absence of 2 ppm VC, respectively, while the temperature was gradually increased from 150°C to 400°C. Steady-state measurement conditions were ensured at each temperature before starting spectra acquisition by monitoring the main core-level peaks as a function of time. The energy axis of all spectra was calibrated using the Al 2p peak at 74.6 eV as a reference (see Supplementary Fig. 3). 36 To probe the evolution of surface carbon intermediates, C 1s spectra were collected in situ at kinetic energies (KE) of 100 eV and 300 eV, corresponding to mean escape depths (MED) of 3.64 Å and 5.68 Å, respectively. Under HV conditions, all spectra exhibited a main peak at 285.3 eV, assigned to adventitious carbon, and a low intensity feature at 289.6 eV, which can be ascribed to carbonates (Supplementary Fig. 4 and Supplementary Table 2). 37, 38 Given that XPS is highly surface sensitive, the detection of a carbon contamination on a freshly synthesized powder sample was expected. In previous experiments performed on silver foil, carbon contamination was efficiently removed heating the sample in vacuum. 31, 32 Experimental results showed a relevant decrease of the carbon signal between 100 and 300°C. In this case, we did not pretreat Ag/Al 2 O 3 samples at high temperature to preserve the morphology of silver NPs. However, because in situ photoemission spectra were acquired at industrially relevant temperatures (starting from 150°C) in the presence of oxygen, we can expect a substantial decrease of adventitious carbon. Upon introducing the C 2 H 4 :O 2 reactant mixture, the line shape of C 1s spectra acquired at a KE of 100 eV (Fig. 2a) significantly changed compared to those obtained in vacuum (Supplementary Figure 4), revealing a broad feature between 283.0 eV and 286.0 eV. When VC was co-fed, the C 1s spectra acquired in the 150-250 °C temperature range showed a significantly suppressed low-binding-energy (BE) shoulder (BE<284.0 eV). Additionally, a low intensity peak at 289.6 eV was detected. Through deconvolution of the C 1s spectra (Supplementary Fig. 5, fitting parameters reported in Supplementary Table 3), a sharp peak at ~286.0 eV was discerned, corresponding to gas-phase C 2 H 4 , 31 along with six distinct peaks (Fig. 2b–g). The chemical nature of surface species assigned to each C 1s peak component was analyzed based on DFT calculations and on a previous study carried out on a silver foil. 31 Peaks at BE < 284.5 eV (C EPO-0 and C EPO-1 ) were attributed to dehydrogenation and C-C cleavage products formed on the surface of silver nanoparticles (NP) upon ethylene adsorption. As shown in Fig. 2h–k, DFT calculations confirmed that consecutive dehydrogenation and cleavage of C-C bond led to a gradual negative BE shift of the corresponding adsorbates. This trend was consistently observed on Ag(111) (Fig. 2h-k), on the reconstructed p(4×4) surface oxide reconstruction, and on defective silver surfaces (Supplementary Fig. 6-7). Combining experimental results with BE simulations, the peak at 284.0 eV (C EPO-1 ) was assigned to C 2 H x (x=1-3) fragments, while the peak at 282.9 eV (C EPO-0 ) matched well with CH y (y=1-2) fragments adsorbed on the silver surface. The peak at 285.0 eV (C EPO-2 ) was assigned to adventitious carbon or long-chain carbonaceous species, with minor contributions from weakly adsorbed ethylene. 31 Carbonate species (C EPO-5 ) were always detected at 289.6 eV under reaction conditions. Weak features at 287.3 eV and 288.1 eV (C EPO-3 and C EPO-4 , respectively) were attributed to surface-adsorbed carbon monoxide and EO, respectively, as previously observed. 31 The integrated peak areas of all carbon species were normalized by photon flux, C 1s cross-section, and MED, to obtain values proportional to surface atomic percentages. As shown in Figures 2b–c, in the absence of VC, C 2 H x and CH y species were abundant at 150°C. With increasing temperature, the CH y coverage diminished, while C 2 H x accumulated on the surface. Notably, the introduction of 2 ppm VC drastically suppressed the amount of C 2 H x and CH y species, particularly CH y , which remained at extremely low levels across all temperatures. This suggests that VC strongly inhibited both C–H and C–C bond cleavage of ethylene on the surface of silver. This result is further supported by the consistent trends observed in the C 1s spectra collected at KE of 300 eV (Supplementary Fig. 8–10 and Supplementary Table 4). To further investigate the surface modifications induced by VC, we analyzed the Ag 3d spectra collected at a KE of 300 eV across different temperatures (Supplementary Fig. 11–13 and Supplementary Table 5). Approximately 70% of silver remained metallic (Ag 0 ) under reaction conditions both in the presence and absence of VC, with the fraction of Ag 0 increasing with temperature. Supplementary Figure 11 shows that the full width at half maximum (FWHM) of the Ag 3d spectra gradually decreased with temperature both in the presence and in the absence of VC. Taking the spectra acquired at 400 °C as a reference for metallic silver NP, two main peak components could be deconvoluted in the 150-300 °C temperature range, which were labelled as Ag α+β and Ag γ . In agreement with previous results, the Ag α+β component at 367.4 eV was attributed to surface silver oxides. 31 Ag γ at 368.9 eV was newly identified and assigned to silver coordinated to CH y fragments, based on the similar temperature-dependent behavior as that of C EPO-0 (see Fig. 2b). In the absence of VC, Ag γ accounted for approximately 10% of the total silver species below 200 °C, whereas in the presence of VC, the Ag γ signal was nearly undetectable. The Cl 2p spectra (Supplementary Fig. 14) acquired while feeding VC exhibited a faint signal due to the low VC concentration (2 ppm). Even after 210 repeated scans (over 2 hours), the signal-to-noise ratio remained low, and the doublet (2p 3/2 centered at approximately 198.0 eV, suggesting the presence of chloride) could hardly be resolved. The reference O 1s spectrum acquired in HV (Supplementary Figure 15) displayed a main peak at 531.4 eV (dashed line), corresponding to lattice oxygen in Al 2 O 3 . 39 Under reaction conditions, the spectra broadened, due to the presence of a shoulder appearing above 532.0 eV, with the VC-treated sample showing a less prominent change. Difference spectra, obtained by using the O 1s in HV as a reference, revealed two broad features at 532.2 eV and 533.9 eV (Supplementary Fig. 16). Hydroxyls groups and oxygenated carbon species (e.g. adsorbed C-O) typically appear almost superimposed in the 531.5-532.5 eV BE range. 40 Since the introduction of VC did not significantly affect the fraction of oxygenated carbon species observed in the C 1s spectra (Figs. 2e and 2f), the peak at 532.2 eV was attributed to surface hydroxyl groups while the peak at 533.9 eV was assigned to adsorbed water. 41, 42 These findings suggest that VC inhibited the formation of both hydroxyl groups and adsorbed water on the catalyst surface under reaction conditions, in particular the formation of water below 250 °C. Overall, introducing VC into the reaction feed simultaneously suppressed the cleavage of C-H and C-C bonds and inhibited the accumulation of adsorbed hydroxyls, most likely contributing to the reduction in CO 2 formation. Surface defect sites as active centers for dehydrogenation Employing DFT calculations, we investigated the sites that activated ethylene through the dehydrogenation pathway on the surface of silver —negatively influencing the selectivity towards EO. The Ag 3d spectra revealed that approximately 70% of silver remained metallic under in situ EPO conditions (supplementary Fig. 12). Given that these spectra were collected at a KE of 300 eV, the estimated photoelectrons MED was 5.7 Å, corresponding to approximately two silver atomic layers. 44 Thus, around 30% of the top two silver layers contains oxide species, indicating the coexistence of silver oxide and metallic silver under reaction conditions. Previous STM studies reported that exposing the thermodynamically most stable facet of silver (Ag(111)) to oxygen yields a dispersed p(4×4) reconstruction, forming isolated surface oxide islands. 45, 46, 47, 48 Based on these observations, we constructed a series of atomistic slab models that expose a range of potential reactive sites and evaluated the energetics of ethylene activation on three representative sites: pristine Ag(111), reconstructed p(4x4) surface oxide and the interface between Ag(111) and p(4x4). The model systems included pristine Ag(111), Ag(111) with edge sites, Ag(111) with adsorbed molecular oxygen, reconstructed p(4×4) and c(4x8) surfaces, and configurations featuring coexisting p(4x4) and metallic silver on Ag(111) substrates, as shown in Supplementary Fig. 17. To evaluate the relative thermodynamic stability of these models, we calculated the surface energies as a function of temperature (see thermodynamic diagram in Supplementary Fig. 18). The p(4×4)-edge-3 structure—which incorporates coordinatively unsaturated oxygen atoms on the reconstructed surface—was the most stable substrate at EPO reaction temperature. Therefore, this configuration was selected as a representative p(4x4)-edge site for subsequent activation energy calculations. The adsorption of ethylene was relatively weak on most of the considered sites (see Supplementary Fig. 19), including pristine and stepped Ag(111), and oxygen-reconstructed flat surfaces. Adsorption energies ranged from ΔE = –0.30 to –0.55 eV. At these sites, the C=C bond typically resided atop a silver atom or bridged between two surface oxygen atoms. The strongest adsorption energy, –1.05 eV, occurred on the p(4×4)-edge model, where the ethylene C=C bond was atop a subsurface silver atom and surrounded by edge-located oxygen atoms. We then evaluated the dehydrogenation process, by calculating the minimum energy pathway (MEP) using nudged elastic band (NEB) simulations. The resulting energy values simulated for the transition state and the optimized products state provided the energy barrier and overall reaction energy. As previously reported by Huš et al., the activation barrier for direct dehydrogenation on pristine Ag(111) is relatively high (1.89 eV), while the presence of oxygen is expected to lower it, due to its interaction with split off hydrogen. 49 We observed that the physisorbed oxygen molecule on pristine Ag(111) reduced the dehydrogenation energy barrier to 1.30 eV (Supplementary Fig. 20). On the flat reconstructed p(4x4) surface, the barrier decreased further to 1.17 eV (Supplementary Fig. 21), while the lowest barrier of 0.71 eV was observed on the p(4×4)-edge surface (Fig. 3a). These results point to the critical role of surface defects in combination with coordinated oxygen. On all substrates, dehydrogenated ethylene bound to one silver atom with the unsaturated carbon atom and tilted into an almost vertical geometry. The dissociated hydrogen bound to a nearby oxygen atom, when present, or adsorbed at an fcc site. The adsorption energies of the dehydrogenated intermediates, relative to gas-phase ethylene molecule, are reported in Fig. 3b, while the optimized configurations of these intermediates are shown in Fig. 3c-g. On both the pristine and stepped Ag(111), the adsorption energies are positive (1.08 eV and 0.89 eV, respectively), indicating unstable states. As expected, the presence of oxygen significantly stabilized the dehydrogenated intermediates, primarily due to the formation of O–H bonds (see the trend in Fig. 3b). The most favorable dehydrogenation site turned out to be at the p(4×4)-edge (adsorption energy of -1.05 eV), where the undercoordination of the oxygen atoms further strengthened the O-H bond. Since p(4x4) defective sites are identified as the most active, we simulated the MEP for three possible full combustion processes, as shown in the Supplementary Fig. 22. The direct C-C bond cleavage appears unlikely due to the high energy of the initial intermediate (CH + CH 2 + OH). In contrast, the interaction of C 2 H 3 with lattice oxygen significantly stabilized the intermediates, ultimately leading to C-C bond cleavage and the formation of CH fragments and carbon dioxide. Chlorine-mediated promotion of ethylene epoxidation To highlight the role of chlorine in promoting the selectivity during ethylene epoxidation, we investigated the interaction of chlorine with the Ag/Al 2 O 3 catalyst. As shown in Supplementary Fig. 23, EDS mapping of a Ag/Al 2 O 3 catalyst after reacting under EPO conditions in the presence of 2 ppm VC for 48 hours revealed a chlorine signal, which spatially overlapped with that of silver. This suggests that chlorine preferentially reacted with the silver NP rather than with the alumina support. We utilized model surfaces (silver foil samples) and exposed them to a reaction gas mixture of C 2 H 4 :O 2 = 2:1, in the presence/absence of 10 ppm DCE, to enhance the photoemission signal of chlorine and investigate the interaction of chlorine with silver and its influence on the rection mechanism. The use of DCE was based on similar kinetic behavior observed for VC and DCE reported in the literature. 50 Upon dosing the EPO reaction mixture, the temperature was raised stepwise from room temperature to 400 °C. At each temperature, photoemission spectra were collected under steady state reaction conditions. A survey scan under HV confirmed the cleanliness of the surface, as evidenced by low intensity C 1s and O 1s signals on both samples (Supplementary Fig. 24). Upon introducing the reaction gas, Cl 2p spectra were collected at each temperature until signal stabilization was achieved. As shown in Supplementary Fig. 25-26, the Cl 2p signal remained weak but stable over time below 200 °C. When the temperature exceeded 300 °C, the Cl 2p signal significantly increased in intensity, yet still showed no signs of accumulation during data acquisition, indicating that the surface coverage of chlorine was stable throughout the experiments. Ag 3d spectra acquired at various temperatures (Supplementary Fig. 27) confirmed that the silver foil remained predominantly metallic. A shoulder centered at 367.4 eV, attributed to surface AgO x species, 32, 51, 52 was clearly detected in the absence of DCE, and almost disappeared upon DCE introduction. To investigate reaction intermediates, we collected C 1s spectra (Supplementary Fig. 28–30). Variations in ethylene gas-phase peak intensity are attributed to small differences in sample-to-analyzer distances between the two experiments. C 2 H x species (C EPO-1 ) were detected from room temperature, further confirming that spontaneous dehydrogenation of ethylene took place on the surface of silver. In contrast to the Ag/Al 2 O 3 catalyst, the intensity of C EPO-0 (CH x species) was negligible on silver foil; the foil contained less defects than NP and did not facilitate C–C bond cleavage. Moreover, co-feeding DCE significantly suppressed the surface C 2 H x signal, further proving that chlorine promoters inhibit ethylene dehydrogenation and in good agreement with the results obtained on the Ag/Al 2 O 3 catalyst. O 1s spectra acquired at various temperatures (Supplementary Fig. 31–33) revealed that the peaks at 532.1 eV and 533.7 eV—assigned to surface hydroxyl species and adsorbed water, respectively—were significantly suppressed in the presence of DCE. 53, 54 As observed on Ag/Al 2 O 3 (Supplementary Fig. 15-16), co-feeding DCE suppressed ethylene dehydrogenation, thus the resulting formation of adsorbed hydroxyls and water. Supplementary Fig. 34 a-b show Cl 2p spectra acquired at KE of 200 and 480 eV, respectively, at increasing temperatures (from room temperature to 400 °C). At both KEs, the chlorine signal markedly intensified above 250 °C, indicating that DCE rapidly decomposed on silver. Normalized spectra acquired at a KE of 480 eV (Supplementary Fig. 34c) show that the Cl 2p 3/2 peak was consistently centered at 198.5 eV, matching well the BE of chloride. 55, 56, 57 Under EPO reaction conditions, DCE decomposed to partially decorate the silver foil surface with chloride. Supplementary Figure 33 a shows that the O 1s peak component at 529.3 eV, attributed to lattice oxygen (O L ), increased less sharply at T>250 °C in the presence of DCE. This, combined with the rapid increase of the Cl 2p signal detected above 250 °C (Supplementary Fig. 34), suggests that chloride partially substituted lattice oxygen species. Based on these results, we explored the impact of chloride species using DFT simulations on the previously discussed p(4x4)-edge slab model. Notably, substituting one lattice oxygen with chlorine lowers the total energy of the system by more than 1 eV (Supplementary Fig. 35), indicating a thermodynamically favorable process. As the next step, we calculated the adsorption energies of VC at different sites (Supplementary Fig. 36). Similar to ethylene, VC preferentially adsorbed at the edge between Ag(111) and p(4×4) terraces (Supplementary Fig. 36f). Moreover, owing to the high electronegativity of chlorine, VC consistently exhibited stronger adsorption than ethylene. This implies that chlorine promoters suppress ethylene adsorption and, consequently, its activation. Using transmission electron microscopy (TEM), van Hoof et al. revealed that voids/pores formed within silver NP during EPO in the absence of a chlorine promoter. Introducing VC significantly suppressed void formation. 11 The authors proposed that oxygen adsorption at silver NP grain boundaries induced defects formation during the reaction, whereas VC mitigated this effect. To validate this phenomenon on the Ag/Al 2 O 3 catalyst used in this work, we conducted TEM measurements on the fresh catalysts and on samples treated under 3 different EPO reaction conditions for 3 hours. As shown in Supplementary Fig. 37a–d, the fresh Ag/Al 2 O 3 catalyst exhibited relatively large silver particle sizes without observable internal voids. After 3 hours EPO reaction at 200 °C without co-feeding the VC promoter (Supplementary Fig. 37e–h), small void-like structures (~10 nm) started to appear inside silver particles. At 230 °C, the extent of void formation with diameters up to 50 nm increased. Notably, the formation of voids within silver particles introduced additional surface defect sites. In contrast, co-feeding 0.5 ppm VC at 230 °C markedly suppressed void formation, in agreement with previous literature. 11 The addition of a chlorine promoter increased the selectivity toward EO through a twofold effect: i) it suppressed ethylene activation by limiting the formation of lattice oxygen and inhibiting the adsorption of ethylene on the surface of silver, ii) it minimized the number of surface defect sites, where ethylene activation (dehydrogenation and C-C bond cleavage) is favored. Direct oxidation pathway to EO The presence of dioxygen species on reconstructed silver oxide surface under EPO reaction conditions was revealed by means of AP-XPS and Raman spectroscopy by Pu et al. and Chen et al. 4, 21, 30 Chen et al. reported that the activation energy for oxygen dissociation on silver surfaces typically exceeds 1.1 eV. 58 Similarly, experimental studies on silver foil showed that weakly adsorbed dioxygen species embedded within reconstructed surfaces only formed at temperatures above 150 °C. 32 The formation of such oxygen species strongly correlated with the EPO process. To verify this hypothesis, we simulated the ethylene oxidation MEP on both p(4×4) and c(4×8) reconstructions, where weakly bound dioxygen inserts into the surface oxide lattice (orange in Fig. 4). As expected, ethylene initially physisorbed on both substrates, with weak adsorption energies of –0.35 eV and –0.41 eV, respectively. In the presence of additional oxygen species, the oxidation reaction leading to adsorbed EO occurred over a relatively low activation barrier, i.e., 0.41 eV and 0.24 eV on p(4×4) and c(4×8) surfaces, respectively. These findings suggest that once the selective dioxygen species forms, ethylene rapidly undergoes epoxidation, meaning that the rate-determining step in the EO formation process is the generation of dioxygen species from gaseous oxygen activation. 32, 58 This mechanistic insight aligns with prior work. 59 Based on these findings, we propose that two distinct silver-catalyzed mechanisms are responsible for EPO and ethylene combustion (Fig. 5). These mechanisms deviate from the reaction pathway involving a common OMC intermediate, as further demonstrated by the temperature-programmed surface reaction experiments shown in Fig. 38 and described in the SI. 11, 12 Under in situ EPO conditions, two main types of oxygen species are formed on the surface of silver. Ethylene interacts with electrophilic dioxygen species to form EO through a direct oxidative pathway, in which the formation of surface dioxygen is the rate-determining step. In contrast, at surface defect sites, ethylene readily adsorbs and undergoes deep dehydrogenation assisted by lattice oxygen and C-C cleavage, leading to the formation of carbon dioxide through an unselective pathway. Chlorine suppresses the combustion pathway by preferentially binding to defect sites as chloride, thereby passivating undercoordinated silver atoms, reducing lattice oxygen formation, and inhibiting C–H activation. Furthermore, the addition of VC hinders ethylene adsorption and limits the formation of porous structures during the reaction, thereby reducing the number of surface defect sites. Conclusion In this study, we combined in situ XPS and DFT calculations to elucidate the reaction mechanism of EPO on Ag/Al 2 O 3 catalysts. Under in situ EPO conditions, C 2 H x and CH y fragments were identified alongside adsorbed hydroxyls and water, which were confirmed as key precursors to carbon dioxide formation based on reactivity studies. These results demonstrate that ethylene preferentially undergoes activation on the surface of silver via oxygen-assisted dehydrogenation pathways to form carbon dioxide. In the presence of chlorine, the surface concentrations of dehydrogenation fragments of ethylene, adsorbed hydroxyls and water were significantly suppressed. This indicates that chlorine-containing promoters enhance EO selectivity primarily by inhibiting the dehydrogenation of ethylene. DFT calculations revealed that ethylene strongly adsorbs on defect sites, particularly at the interface between metallic and oxygen reconstructed surface domains. These sites facilitated the reaction of ethylene with lattice oxygen, confirming that defects trigger carbon dioxide formation. We further demonstrated that chlorine-containing promoters suppressed ethylene activation by reducing the fraction of surface lattice oxygen, inhibiting ethylene adsorption, and decreasing the density of defect sites. These combined effects significantly enhanced the selectivity toward EO. Moreover, DFT simulations indicated that EO formation may proceed through a direct oxidation pathway involving adsorbed dioxygen species formed on oxygen reconstructed silver surfaces. These findings provide new insights into the mechanistic pathways of EPO on silver-based catalysts, proposing parallel reaction mechanisms. This work does not only advance the understanding of the EPO process but also suggests a strategy for enhancing the selectivity toward EO by hindering the formation of defect sites on the silver surface. Methods Synthesis of Ag/Al 2 O 3 catalysts The 15 wt% Ag/Al 2 O 3 catalyst was prepared via incipient wetness impregnation using silver oxalate (Ag 2 C 2 O 4 ) as the silver precursor, following widely adopted protocols in the literature. 34, 60 Ag 2 C 2 O 4 was synthesized by mixing aqueous solutions of oxalic acid (Sigma Aldrich, 99%) and silver nitrate (Merck, 99%) in a 1:2 molar ratio. The precipitate was collected by centrifugation, washed three times with deionized water, and dried overnight in static air at 60 °C. A solution composed of silver oxalate and ethylenediamine (Sigma Aldrich, 99%) in a 1:4 molar ratio, diluted in 2 cm 3 of deionized water, was impregnated onto α-Al 2 O 3 (Thermo Scientific, 99.9%) via incipient wetness. The impregnated support was dried overnight at 60 °C and subsequently calcined in static air at 400 °C for 6 hours. Characterization of Ag/Al 2 O 3 catalysts XRD patterns were recorded using a Bruker D8 ADVANCE diffractometer with Cu K α radiation (λ = 1.54 Å) at a scan rate of 0.02° per step over the 2θ range of 20–80°. Nitrogen physisorption measurements were conducted at 77 K using a Micromeritics 3Flex instrument. Prior to analysis, the samples were degassed at 120 °C overnight. The specific surface area was determined by the Brunauer–Emmett–Teller (BET) method by fitting the isotherms. Transmission electron microscopy (TEM) and scanning transmission electron microscopy (STEM) analyses were performed on a JEOL JEM-F200 microscope operated at 200 kV and equipped with two silicon drift detectors (SDDs) for energy-dispersive X-ray spectroscopy (EDX). Samples were prepared by dispersing the catalyst powder in ethanol followed by drop-cast onto a copper grid. In situ AP-XPS measurements of Ag/ Al 2 O 3 and silver foil AP-XPS measurements of Ag/Al 2 O 3 and silver foil were performed at the X07DB “In Situ Spectroscopy” beamline at the Swiss Light Source (SLS), Paul Scherrer Institute. Ag/Al 2 O 3 powders were pressed onto two separate 10 × 10 mm gold meshes, which were then mounted on a manipulator and introduced into the solid-gas interface endstation. Initially, Al 2p, Ag 3d, C 1s, and O 1s spectra were collected under high vacuum at kinetic energies of 100 eV and 300 eV as references. The sample was then exposed in situ to a gas mixture of ethylene and oxygen in a 2:1 ratio at a total pressure of 1 mbar. For one set of samples, 2 ppm of VC were co-dosed with the reactants feed. Gas mixtures were introduced using mass flow controllers, while the total pressure was regulated downstream by a tunable diaphragm valve connected to a root pump, enabling precise control over partial pressures and gas ratios. The pressure was monitored using Baratron gauges. Samples were heated by an infrared laser targeting the rear of the sample holder, with temperature measured by a Pt100 sensor. Spectra were collected sequentially while exposing the samples to the gas mixture at various temperatures. Linearly polarized synchrotron light was used throughout the experiments. C 1s spectra were acquired using photon energies of 385 and 595 eV (kinetic energies of 100 and 300 eV). O 1s, Al 2p, Ag 3d, and Cl 2p spectra were collected at hν of 840, 385, 675, and 505 eV, respectively, all corresponding to kinetic energies of 300 eV to ensure comparable MEDs. The BE scale was calibrated using the Al 2p peak at 74.6 eV. All spectra were background-subtracted using a Shirley function and fitted with Voigt functions. Fitting parameters are summarized in Supplementary Tables 2–5. For experiments employing silver foil, silver foils (Alfa Aesar, 99.99%) were cut into 10 × 10 mm squares and cleaned sequentially with acetone, isopropanol, and deionized water, followed by plasma cleaning in oxidizing (O 2 /Ar) and reducing (H 2 /He) atmospheres to remove surface contaminants. The cleaned foils were mounted onto the manipulator and introduced into the solid-gas interface endstation. Prior to measurements, foils were preheated to 400 °C under vacuum to desorb residual adsorbates, and surface cleanliness was verified by XPS. The subsequent experimental procedure was identical to that used for Ag/Al 2 O 3 , with the exception that 10 ppm of dichloroethane were co-fed with the reactant gases for one set of experiments. Ag 3d, O 1s, and C 1s spectra were collected at hν of 575, 735, and 490 eV (KE = 200 eV), while Cl 2p spectra were collected at hν of 400 and 680 eV (KE = 200 and 480 eV). Fitting parameters are summarized in Tables S6–S7. Catalytic Performance and Kinetic Analysis The catalytic performance was evaluated in a commercial plug-flow CATLAB reactor. An online gas chromatograph (MicroGC, INFICON) equipped with two TCD detectors was connected to the reactor outlet for product analysis. Oxygen, ethylene, EO, carbon dioxide, and water were monitored at 5-minute intervals. The feed gas consisted of 15 vol% ethylene and 3 vol% oxygen, balanced with nitrogen to a total flow rate of 20 mL/min; optionally, 2.5 vol% of 20 ppm VC in He was added. The reactor was loaded with 5–20 mg of catalyst, heated to 230 °C with a rate of 5 °C/min, and maintained at this temperature for 3–48 hours under continuous flow. Catalysts were evaluated at same ethylene conversions to compare EO selectivity. Ethylene conversion and ethylene oxide selectivity were calculated using equations (1) and (2), respectively. (1) (2) where the subscripts in and out refer to the concentrations measured at the reactor inlet and outlet, respectively. Temperature-programmed surface reaction experiments were performed using the same CATLAB reactor, with outlet gases monitored by an online mass spectrometer. The Ag/Al 2 O 3 catalyst was first exposed to 7.5 vol% oxygen in argon at room temperature. It was then heated to 400 °C and held until the carbon dioxide and water signals were no longer detected. After cooling to room temperature in the O 2 /Ar mixture, the catalyst was exposed to a gas mixture of 5 vol% ethylene and 7.5 vol% oxygen in argon for 10 minutes. The ethylene feed was then cut off, and the catalyst was purged with O 2 /Ar until the ethylene signal fully disappeared. Subsequently, the temperature was ramped at 5 °C/min while recording mass signals for ethylene (m/z = 27), O 2 (m/z = 32), EO (m/z = 43, minor fragment), and carbon dioxide (m/z = 44). DFT simulations All density functional theory (DFT) calculations were performed using the CP2K program package (Quickstep) 56 , employing the Kohn-Sham formalism within the hybrid Gaussian and plane wave (GPW) framework. Goedecker-Teter-Hutter (GTH) pseudopotentials 57 and molecularly optimized (MOLOPT) Gaussian basis sets 58 were used to describe the valence electrons. Specifically, DZVP-MOLOPT-GTH basis sets were applied for Ag, O, C, and H atoms. A plane-wave cutoff energy of 600 Ry was used for the auxiliary basis set. All calculations were performed under periodic boundary conditions, using only the Gamma point, and were spin-polarized. Since an asymmetric slab model was used—featuring the oxidized-reconstructed surface and adsorbates on one side only—a dipole correction was applied along the z-axis to account for the surface dipole 59 . The structural and electronic properties were computed at the generalized gradient approximation (GGA) level of theory, using the Perdew-Burke-Ernzerhof (PBE) functional 60 , with Grimme-D3 dispersion corrections to account for long-range van der Waals interactions 61 . Geometry optimizations were carried out using the Broyden-Fletcher-Goldfarb-Shanno (BFGS) algorithm 62 , with a force convergence criterion of 10⁻³ Hartree/Bohr³. The Climbing-Image Nudged Elastic Band (CI-NEB) method was applied to find minimum-energy pathway and energy barriers of the reaction process, and the transition states 63 . To accurately simulate various oxygen species and intermediates on an Ag/Al₂O₃ substrate, we considered both the clean Ag(111) surface and two thermodynamically stable oxidized-reconstructed Ag(111) surface structures—namely, the p(4×4) and c(4×8) reconstructions—as used in our previous study 28 . Additionally, to mimic defective Ag surfaces with unsaturated Ag sites exposed to the gas phase, we constructed step-edge models of the Ag(111) and p(4×4) surfaces, referred to as Ag(111)-edge and p(4×4)-edge, respectively. The models consisted of a 5-layer Ag(111) slab, where the top layer was modified by introducing oxygen atoms to form the p(4×4) and c(4×8) reconstructions. Various carbon-containing species were then adsorbed onto these surfaces to study their interactions. The lateral dimensions of the simulation cell were set to 23.19 × 23.19 Ų (for Ag(111)), and a vacuum region of 20 Å was included both above and below the slab to eliminate spurious interactions between periodic images. The calculation of the C and O K-edge, used to reproduce the binding energies of 1s electrons, was achieved by applying the Gaussian augmented plane wave method (GAPW). For the C and O elements, all electrons were explicitly considered (no pseudopotentials) and 6-311G** all-electron basis sets were employed to expand the molecular orbitals, as used in previous works 27,28 . The calculations of the energies of the core states were based on the Slater transition potential method with half-core hole approximation, where initial and final state effects were accounted for by electronic energy eigenvalue calculations after removing half an electron from the core state. 64,65 Declarations Acknowledgements M.G. and L.A. acknowledge the Swiss National Science Foundation (project number 196946) for their support. N.D. and M. I. thank for the generous allocation of computing resources from the Swiss National Supercomputing Center (CSCS) under Project ID s1279, and both CSCS and the Alfred Werner-Legat for the computing resources under Project ID uzh35. References Pu TC, Tian HJ, Ford ME, Rangarajan S, Wachs IE. Overview of Selective Oxidation of Ethylene to Ethylene Oxide by Ag Catalysts. ACS Catal 2019, 9 (12) : 10727-10750. Herzberger J, Niederer K, Pohlit H, Seiwert J, Worm M, Wurm FR , et al. Polymerization of Ethylene Oxide, Propylene Oxide, and Other Alkylene Oxides: Synthesis, Novel Polymer Architectures, and Bioconjugation. Chem Rev 2016, 116 (4) : 2170-2243. Stadler BM, Wulf C, Werner T, Tin S, de Vries JG. Catalytic Approaches to Monomers for Polymers Based on Renewables. ACS Catal 2019, 9 (9) : 8012-8067. Pu T, Setiawan A, Foucher AC, Guo M, Jehng JM, Zhu M , et al. Revealing the Nature of Active Oxygen Species and Reaction Mechanism of Ethylene Epoxidation by Supported Ag/alpha-Al 2 O 3 Catalysts. ACS Catal 2024, 14 (1) : 406-417. Iyer KR, Bhan A. Interdependencies Among Ethylene Oxidation and Chlorine Moderation Catalytic Cycles Over Promoted Ag/α-Al 2 O 3 Catalysts. ACS Catal 2021, 11 (24) : 14864-14876. Lockemeyer JR, Lohr TL. Ethylene Oxide Catalysis Under Commercial Conditions – A Guide for Researchers. ChemCatChem 2023, 15 (13). Diao W, DiGiulio CD, Schaal MT, Ma S, Monnier JR. An investigation on the role of Re as a promoter in Ag Cs Re/α-Al 2 O 3 high-selectivity, ethylene epoxidation catalysts. J Catal 2015, 322: 14-23. Jankowiak J, Barteau M. Ethylene epoxidation over silver and copper–silver bimetallic catalysts: II. Cs and Cl promotion. J Catal 2005, 236 (2) : 379-386. Jalil A, Happel EE, Cramer L, Hunt A, Hoffman AS, Waluyo I , et al. Nickel promotes selective ethylene epoxidation on silver. Science 2025, 387: 869–873. Linic S, Barteau MA. Control of Ethylene Epoxidation Selectivity by Surface Oxametallacycles. J Am Chem Soc 2003, 125: 4034-4035. van Hoof AJF, Filot IAW, Friedrich H, Hensen EJM. Reversible Restructuring of Silver Particles during Ethylene Epoxidation. ACS Catal 2018, 8 (12) : 11794-11800. Christopher P, Linic S. Engineering Selectivity in Heterogeneous Catalysis: Ag Nanowires as Selective Ethylene Epoxidation Catalysts. J Am Chem Soc 2008, 130: 11264–11265. Jones GS, Mavrikakis M, Barteau MA, Vohs JM. First Synthesis, Experimental and Theoretical Vibrational Spectra of an Oxametallacycle on a Metal Surface. J Am Chem Soc 1998, 120: 3196-3204. Wu G, Stacchiola D, Kaltchev M, Tysoe WT. The adsorption and reaction of 2-iodoethanol on Ag(111). Surf Sci 2000, 463: 81–92. Linic S, Medlin JW, Barteau MA. Synthesis of Oxametallacycles from 2-Iodoethanol on Ag(111) and the Structure Dependence of Their Reactivity. Langmuir 2002, 18: 5197-5204. Torres D, Lopez N, Illas F, Lambert RM. Why Copper Is Intrinsically More Selective than Silver in Alkene Epoxidation: Ethylene Oxidation on Cu(111) versus Ag(111). J Am Chem Soc 2005, 127: 10774-10775. Derouin J, Farber RG, Turano ME, Iski EV, Killelea DR. Thermally Selective Formation of Subsurface Oxygen in Ag(111) and Consequent Surface Structure. ACS Catal 2016, 6 (7) : 4640-4646. Chen DX, Kang PL, Liu ZP. Active Site of Catalytic Ethene Epoxidation: Machine-Learning Global Pathway Sampling Rules Out the Metal Sites. ACS Catal 2021, 11 (13) : 8317-8326. Jones TE, Wyrwich R, Bocklein S, Carbonio EA, Greiner MT, Klyushin AY , et al. The Selective Species in Ethylene Epoxidation on Silver. ACS Catal 2018, 8 (5) : 3844-3852. Pu T, Setiawan A, Rangarajan S, Wachs IE. Progress and perspective on the fundamental understanding of structure–activity/selectivity relationships for Ag catalyzed ethylene epoxidation. Catalysis Today 2025, 455 . Chen C-T, Sviripa A, Verma S, Paolucci C, Flaherty DW. Reactions of Surface Peroxides Contribute to Rates and Selectivities for C 2 H 4 Epoxidation on Silver. ACS Catal 2025 : 1387-1398. Petrov L, Eliyas A, Shopov D. A Kinetic Model of Steady State Ethylene Epoxidation over a Supported Silver Catalyst. Applied Catalysis 1985, 18: 87-103. Chen CJ, Harris JW, Bhan A. Kinetics of Ethylene Epoxidation on a Promoted Ag/alpha-Al 2 O 3 Catalyst-The Effects of Product and Chloride Co-Feeds on Rates and Selectivity. Chem EurJ 2018, 24 (47) : 12405-12415. Harris JW, Wang L, DeWilde JF. Kinetic Modeling of Ethylene Epoxidation Kinetics as a Function of Chlorine Coverage over a Highly Promoted Silver Catalyst. Ind Eng Chem Res 2019, 58 (41) : 19061-19071. Harris JW, Herron JA, DeWilde JF, Bhan A. Molecular characteristics governing chlorine deposition and removal on promoted Ag catalysts during ethylene epoxidation. J Catal 2019, 377: 378-388. Foster AJ, Lobo RF. Identifying reaction intermediates and catalytic active sites through in situ characterization techniques. Chem Soc Rev 2010, 39 (12) : 4783-4793. Roy K, Artiglia L, van Bokhoven JA. Ambient Pressure Photoelectron Spectroscopy: Opportunities in Catalysis from Solids to Liquids and Introducing Time Resolution. ChemCatChem 2018, 10 (4) : 666-682. Artiglia L, Sushkevich VL, Palagin D, Knorpp AJ, Roy K, van Bokhoven JA. In Situ X-ray Photoelectron Spectroscopy Detects Multiple Active Sites Involved in the Selective Anaerobic Oxidation of Methane in Copper-Exchanged Zeolites. ACS Catal 2019, 9 (8) : 6728-6737. Carbonio EA, Rocha TCR, Klyushin AY, Pis I, Magnano E, Nappini S , et al. Are multiple oxygen species selective in ethylene epoxidation on silver? Chem Sci 2018, 9 (4) : 990-998. Pu T, Setiawan A, Mosevitzky Lis B, Zhu M, Ford ME, Rangarajan S , et al. Nature and Reactivity of Oxygen Species on/in Silver Catalysts during Ethylene Oxidation. ACS Catal 2022, 12 (8) : 4375-4381. Guo M, Dongfang N, Liu Q, Iannuzzi M, van Bokhoven JA, Artiglia L. Reaction Intermediates Involved in the Epoxidation of Ethylene Over Silver Revealed by In Situ Photoelectron Spectroscopy. Small 2025, 21 (5) : e2408432. Guo M, Dongfang N, Iannuzzi M, van Bokhoven JA, Artiglia L. Structure and Reactivity of Active Oxygen Species on Silver Surfaces for Ethylene Epoxidation. ACS Catal 2024, 14: 10234-10244. Alzahrani HA, Bravo-Suárez JJ. In situ Raman spectroscopy study of silver particle size effects on unpromoted Ag/α-Al 2 O 3 during ethylene epoxidation with molecular oxygen. J Catal 2023, 418: 225-236. Keijzer PH, van den Reijen JE, Keijzer CJ, de Jong KP, de Jongh PE. Influence of atmosphere, interparticle distance and support on the stability of silver on α-alumina for ethylene epoxidation. J Catal 2021. van Hoof AJF, Hermans EAR, van Bavel AP, Friedrich H, Hensen EJM. Structure Sensitivity of Silver-Catalyzed Ethylene Epoxidation. ACS Catal 2019, 9 (11) : 9829-9839. Peng J, Sun Q, Zhai Z, Yuan J, Huang X, Jin Z , et al. Low temperature, solution-processed alumina for organic solar cells. Nanotechnology 2013, 24 (48). Chen X, Wang X, Fang D. A review on C1s XPS-spectra for some kinds of carbon materials. Fullerenes, Nanotubes Carbon Nanostruct 2020, 28 (12) : 1048-1058. Gengenbach TR, Major GH, Linford MR, Easton CD. Practical guides for x-ray photoelectron spectroscopy (XPS): Interpreting the carbon 1s spectrum. J Vac Sci Technol 2021, 39 (1) : 013204. Lee D, Park JW, Cho NK, Lee J, Kim YS. Verification of Charge Transfer in Metal-Insulator-Oxide Semiconductor Diodes via Defect Engineering of Insulator. Sci Rep 2019, 9 (1) : 10323. Henderson JD, Payne BP, McIntyre NS, Biesinger MC. Enhancing Oxygen Spectra Interpretation by Calculating Oxygen Linked to Adventitious Carbon. Surf Interfaces Analysis 2025, 57 (3) : 214-220. David DGF, Pinault-Thaury M-A, Ballutaud D, Godet C. Sensitivity of photoelectron energy loss spectroscopy to surface reconstruction of microcrystalline diamond films. Applied Surface Science 2013, 273: 607-612. Casalongue HS, Kaya S, Viswanathan V, Miller DJ, Friebel D, Hansen HA , et al. Direct observation of the oxygenated species during oxygen reduction on a platinum fuel cell cathode. Nature Communications 2013, 4 (1). Rocha TCR, Hävecker M, Knop-Gericke A, Schlögl R. Promoters in heterogeneous catalysis: The role of Cl on ethylene epoxidation over Ag. J Catal 2014, 312: 12-16. Soares EA, Nascimento VB, Carvalho VEd, Castilho CMCd, Carvalho AVd, Toomes R , et al. Structure determination of Ag(111) by low-energy electron diffraction. Surf Sci 1999, 419 89-96. Andryushechkin BV, Shevlyuga VM, Pavlova TV, Zhidomirov GM, Eltsov KN. Structural transformations on an oxidized Ag(111) surface. JETP Letters 2017, 105 (5) : 292-296. Turano ME, Juurlink LBF, Gillum MZ, Jamka EA, Hildebrandt G, Lewis F , et al. Oxygen-induced surface reconstructions on curved Ag(111). J Vac Sci Technol 2021, 39 (5). Derouin J, Farber RG, Heslop SL, Killelea DR. Formation of surface oxides and Ag 2 O thin films with atomic oxygen on Ag(111). Surf Sci 2015, 641: L1-L4. Martin NM, Klacar S, Grönbeck H, Knudsen J, Schnadt J, Blomberg S , et al. High-Coverage Oxygen-Induced Surface Structures on Ag(111). J Phys Chem C 2014, 118 (28) : 15324-15331. Hus M, Grilc M, Terzan J, Gyergyek S, Likozar B, Hellman A. Going Beyond Silver in Ethylene Epoxidation with First-Principles Catalyst Screening. Angew Chem Int Ed 2023, 62 (31) : e202305804. Santos VP, Plauck A, Gold J, Majumdar P, McAdon MH, Calverley T. The Complex Chlorination Effects on High Selectivity Industrial EO Catalysts: Dynamic Interplay between Catalyst Composition and Process Conditions. ACS Catal 2024, 14 (14) : 10839-10852. Hoflund GB, Weaver JF, Epling WS. Ag 2 O XPS Spectra. Surf Sci Spectra 1994, 3 (2) : 157-162. Hoflund GB, Weaver JF, Epling WS. AgO XPS Spectra. Surf Sci Spectra 1994, 3 (2) : 163-168. Idriss H. On the wrong assignment of the XPS O1s signal at 531–532 eV attributed to oxygen vacancies in photo- and electro-catalysts for water splitting and other materials applications. Surf Sci 2021, 712 . Guo LQ, Qin SX, Yang BJ, Liang D, Qiao LJ. Effect of hydrogen on semiconductive properties of passive film on ferrite and austenite phases in a duplex stainless steel. Sci Rep 2017, 7 (1) : 3317. Zhu M, Chen P, Liu M. Sunlight-driven plasmonic photocatalysts based on Ag/AgCl nanostructures synthesized via an oil-in-water medium: enhanced catalytic performance by morphology selection. Journal of Materials Chemistry 2011, 21 (41). Fu S, Chu Z, Huang Z, Dong X, Bie J, Yang Z , et al. Construction of Z-scheme AgCl/BiOCl heterojunction with oxygen vacancies for improved pollutant degradation and bacterial inactivation. RSC Adv 2024, 14 (6) : 3888-3899. Yu X, Zhou C, Huang Z, Xin C, Lin Y, Fu F , et al. Rational design of AgCl@Zr 3+ -ZrO 2 nanostructures for ultra-efficient visible-light photodegradation of emerging pollutants. Appl Catal B 2023, 325 . Chen D, Chen L, Zhao Q-C, Yang Z-X, Shang C, Liu Z-P. Square-pyramidal subsurface oxygen [Ag 4 OAg] drives selective ethene epoxidation on silver. Nat Catal 2024. Esposito J, Bhan A. Isotopic Studies of Reaction Pathways within Propylene Epoxidation over Promoted Silver Catalysts. ACS Catal 2024 : 14339-14349. van Hoof AJF, van der Poll RCJ, Friedrich H, Hensen EJM. Dynamics of silver particles during ethylene epoxidation. Appl Catal B 2020, 272 . T. D. Kuhne, M. Iannuzzi, M. Del Ben, V. V. Rybkin, P. Seewald, F. Stein, T. Laino, R. Z. Khaliullin, O. Schutt, F. Schiffmann, D. Golze, J. Wilhelm, S. Chulkov, M. H. Bani-Hashemian, V. Weber, U. Borstnik, M. Taillefumier, A. S. Jakobovits, A. Lazzaro, H. Pabst, T. Muller, R. Schade, M. Guidon, S. Andermatt, N. Holmberg, G. K. Schenter, A. Hehn, A. Bussy, F. Belleflamme, G. Tabacchi, A. Gloss, M. Lass, I. Bethune, C. J. Mundy, C. Plessl, M. Watkins, J. VandeVondele, M. Krack, J. Hutter, CP2K: An electronic structure and molecular dynamics software package - Quickstep: Efficient and accurate electronic structure calculations J. Chem. Phys. 2020, 152 , 194103. S. Goedecker, M. Teter, J. Hutter, Separable dual-space Gaussian pseudopotentials. Phys. Rev. B 1996, 54 , 1703-1710. J. VandeVondele, J. Hutter, Gaussian basis sets for accurate calculations on molecular systems in gas and condensed phases. J. Chem. Phys. 2007, 127, 114105. L. Bengtsson, Dipole correction for surface supercell calculations. Phys. Rev. B 1999, 59 , 12301-12304. J. P. Perdew, K. Burke, M. Ernzerhof, Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 1996, 77 , 3865-3868. S. Grimme, Semiempirical GGA-type density functional constructed with a long-range dispersion correction. J. Comput. Chem. 2006, 27 , 1787-1799. Richard H. Byrd, P. L., Jorge Nocedal, and Ciyou Zhu. A Limited Memory Algorithm for Bound Constrained Optimization. Phys. Rev. Lett. 1995, 16 (5), 1190-1208. Graeme Henkelman, Blas P. Uberuaga, Hannes Jónsson, A climbing image nudged elastic band method for finding saddle points and minimum energy paths. J. Chem. Phys. 2000; 113 (22): 9901–9904. M. Iannuzzi, J. Hutter, Inner-shell spectroscopy by the Gaussian and augmented plane wave method. Phys. Chem. Chem. Phys. 2007, 9 , 1599-1610. J. C. Slater, K. H. Johnson, Self-Consistent-Field Xn Cluster Method for Polyatomic Molecules and Solids. Phys. Rev. B 1972, 5 , 844-853. Additional Declarations There is NO Competing Interest. Supplementary Files SI.docx Supplementary Information TOC.png Table of Content Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7141633","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":498999898,"identity":"dde0c9f5-e1c6-4928-b220-41a79a21910b","order_by":0,"name":"Luca Artiglia","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA6ElEQVRIiWNgGAWjYBACPhiDjYG54QBDBYgBQngATFaCjYERqOUMKVoYgFoYGNtQBHFokUh++IGh5l4dH//CxgM/592x52NgfvYAv5Y0YwmGY8USbBIPGw72bnuW2MbAZm6AX0sO0FFsCUAtBxsO8G47nMDGwMMmQUAL8w+GfxAtB//OOWxPjBY2CcY2oBb+xobDvA2HgSFASAvPMzOLxL4EyTYJxobDMseAfmFmM8OrhZ89+fGND98S+OX7Dx/++Kbmjr18e/MzvFrAIAFESIDJAwwMzATVw+07ANUyCkbBKBgFowANAAB3E0HIowMNmQAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0003-4683-6447","institution":"Paul Scherrer Institute","correspondingAuthor":true,"prefix":"","firstName":"Luca","middleName":"","lastName":"Artiglia","suffix":""},{"id":498999899,"identity":"415aa658-d74e-4d6d-a8a9-46a97f7fa4b7","order_by":1,"name":"Man Guo","email":"","orcid":"https://orcid.org/0009-0001-7272-3772","institution":"Paul Scherrer Institute","correspondingAuthor":false,"prefix":"","firstName":"Man","middleName":"","lastName":"Guo","suffix":""},{"id":498999900,"identity":"917585a6-d02f-4156-9ad8-1bb0df9347c6","order_by":2,"name":"Nanchen Dongfang","email":"","orcid":"","institution":"University of Zurich","correspondingAuthor":false,"prefix":"","firstName":"Nanchen","middleName":"","lastName":"Dongfang","suffix":""},{"id":498999901,"identity":"054066a2-44bd-4fac-89a7-bd7d6b655188","order_by":3,"name":"Frank Krumeich","email":"","orcid":"https://orcid.org/0000-0001-5625-1536","institution":"Swiss Federal Institute of Technology in Zurich","correspondingAuthor":false,"prefix":"","firstName":"Frank","middleName":"","lastName":"Krumeich","suffix":""},{"id":498999902,"identity":"17bf09f8-8a2e-4b5e-bee7-54828e5a392d","order_by":4,"name":"Marcella Iannuzzi","email":"","orcid":"https://orcid.org/0000-0001-9717-2527","institution":"University of Zurich","correspondingAuthor":false,"prefix":"","firstName":"Marcella","middleName":"","lastName":"Iannuzzi","suffix":""},{"id":498999903,"identity":"837d368f-465c-4d20-b09f-56f14c0b2dad","order_by":5,"name":"Jeroen A. van Bokhoven","email":"","orcid":"https://orcid.org/0000-0002-4166-2284","institution":"ETH","correspondingAuthor":false,"prefix":"","firstName":"Jeroen","middleName":"A. van","lastName":"Bokhoven","suffix":""}],"badges":[],"createdAt":"2025-07-16 15:37:29","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7141633/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7141633/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":88907625,"identity":"a159b4c7-c678-4c42-a9c9-6acf24216b91","added_by":"auto","created_at":"2025-08-12 14:52:37","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":59762,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCatalytic performance of Ag/Al\u003c/strong\u003e\u003csub\u003e\u003cstrong\u003e2\u003c/strong\u003e\u003c/sub\u003e\u003cstrong\u003eO\u003c/strong\u003e\u003csub\u003e\u003cstrong\u003e3\u003c/strong\u003e\u003c/sub\u003e\u003cstrong\u003e with vinyl chloride doping. a\u003c/strong\u003e,\u003cstrong\u003e \u003c/strong\u003eEO and CO\u003csub\u003e2\u003c/sub\u003e production rates, and EO selectivity over Ag/Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e at 230°C, with the introduction of\u0026nbsp; 0.5 ppm VC at time=110 min. (Reaction conditions: 15% C\u003csub\u003e2\u003c/sub\u003eH\u003csub\u003e4\u003c/sub\u003e, 3% O\u003csub\u003e2\u003c/sub\u003e balanced by N\u003csub\u003e2\u003c/sub\u003e; space velocity: 10000 ml·min\u003csup\u003e-1\u003c/sup\u003e·g\u003csup\u003e-1\u003c/sup\u003e) \u003cstrong\u003eb\u003c/strong\u003e, C\u003csub\u003e2\u003c/sub\u003eH\u003csub\u003e4\u003c/sub\u003e conversion (circles) and EO selectivity (diamonds) over Ag/Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e in the absence (grey) and presence (brown) of 0.5 ppm VC at 230°C. EO selectivity were evaluated under steady-state conditions, with space velocities adjusted to ensure same levels of C\u003csub\u003e2\u003c/sub\u003eH\u003csub\u003e4\u003c/sub\u003e conversion. (Reaction conditions: 15% C\u003csub\u003e2\u003c/sub\u003eH\u003csub\u003e4\u003c/sub\u003e, 3% O\u003csub\u003e2\u003c/sub\u003e balanced by N\u003csub\u003e2\u003c/sub\u003e; space velocity without VC:\u0026nbsp; 12000 ml·min\u003csup\u003e-1\u003c/sup\u003e·g\u003csup\u003e-1\u003c/sup\u003e; space velocity with VC:\u0026nbsp; 5000 ml·min\u003csup\u003e-1\u003c/sup\u003e·g\u003csup\u003e-1\u003c/sup\u003e)\u003c/p\u003e","description":"","filename":"1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7141633/v1/6d4b7b6106d7c7b71b9bd70d.jpg"},{"id":88908946,"identity":"097d8905-1df1-4f57-9117-e9eb2d6bf496","added_by":"auto","created_at":"2025-08-12 15:00:37","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":170055,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eReaction intermediates on Ag/Al\u003c/strong\u003e\u003csub\u003e\u003cstrong\u003e2\u003c/strong\u003e\u003c/sub\u003e\u003cstrong\u003eO\u003c/strong\u003e\u003csub\u003e\u003cstrong\u003e3\u003c/strong\u003e\u003c/sub\u003e\u003cstrong\u003e under in situ EPO conditions. a, \u003c/strong\u003eC 1s spectra acquired at a C₂H₄/O₂ ratio of 2:1 (1 mbar total pressure) from 150 °C to 400 °C, with and without 2 ppm VC (kinetic energy = 100 eV). \u003cstrong\u003eb-g,\u003c/strong\u003e normalized peak area of different C 1s peak components detected on Ag/Ag\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e as a function of temperature under in situ EPO conditions. \u003cstrong\u003eh-k,\u003c/strong\u003e optimized structure, and calculated BE of different dehydrogenated species on the Ag (111) facet.\u003c/p\u003e","description":"","filename":"2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7141633/v1/f66316ce58bc0e6eb13ae4ed.jpg"},{"id":88907626,"identity":"fef19903-23c3-4769-acfd-b86834f99779","added_by":"auto","created_at":"2025-08-12 14:52:37","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":93221,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDFT calculations of C\u003c/strong\u003e\u003csub\u003e\u003cstrong\u003e2\u003c/strong\u003e\u003c/sub\u003e\u003cstrong\u003eH\u003c/strong\u003e\u003csub\u003e\u003cstrong\u003e4\u003c/strong\u003e\u003c/sub\u003e\u003cstrong\u003e activation at interfacial defect sites. a, \u003c/strong\u003eCalculated energy profile for ethylene dehydrogenation at the interfacial defect sites between p(4×4) and Ag(111), facilitated by lattice oxygen. Asterisks denote adsorbed species. \u003cstrong\u003eb,\u003c/strong\u003e Adsorption energies of dehydrogenation intermediates from C₂H₄ on representative Ag substrates, and dehydrogenation activation energies for oxygen-assisted processes in orange. \u003cstrong\u003ec-g,\u003c/strong\u003e Optimized configurations of dehydrogenation intermediates derived from C\u003csub\u003e2\u003c/sub\u003eH\u003csub\u003e4\u003c/sub\u003e on various Ag surfaces. Color code includes surface Ag, silver; subsurface Ag, purple; O, red; C, blue; and H, white.\u003c/p\u003e","description":"","filename":"3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7141633/v1/990cbb762f5e783e662beea5.jpg"},{"id":88907627,"identity":"31208bcb-2e1f-4e23-ab55-472e634e11d4","added_by":"auto","created_at":"2025-08-12 14:52:37","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":76565,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDFT calculations of the ethylene epoxidation reaction pathway. a, \u003c/strong\u003eCalculated energy profiles for ethylene epoxidation on p(4×4) and c(4×8) oxygen reconstructed surfaces with dioxygen participation. The adsorption of ethylene (ΔE\u003csub\u003e1\u003c/sub\u003e), activation (E\u003csub\u003ea\u003c/sub\u003e), reaction (E\u003csub\u003er\u003c/sub\u003e) and dissociation of ethylene oxide (E\u003csub\u003ed\u003c/sub\u003e) energies (in eV) on p(4×4) and c(4×8) are reported in the inset table. \u003cstrong\u003eb-c,\u003c/strong\u003e Ball and stick illustration of reactant, transition and final state geometries on p(4×4) and c(4×8) (top and side views). The color code follows that of Fig. 3.\u003c/p\u003e","description":"","filename":"4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7141633/v1/69090e8633ad4a552348fd19.jpg"},{"id":88907629,"identity":"60d468e0-0df7-4f71-98e0-0abe78e191f8","added_by":"auto","created_at":"2025-08-12 14:52:37","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":41173,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSchematic illustration of the reaction pathway. \u003c/strong\u003eReaction mechanism of ethylene reacting with oxygen on the Ag surface to form EO and CO\u003csub\u003e2\u003c/sub\u003e. The schematic highlights the role of chlorine in hindering lattice oxygen formation and inhibiting ethylene adsorption, thereby enhancing selectivity toward EO\u003cstrong\u003e.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7141633/v1/7610c20f5591baf9bf67aa05.jpg"},{"id":91432834,"identity":"1b1150ae-8c9b-448a-9412-a9e3cfe9d51a","added_by":"auto","created_at":"2025-09-16 12:33:52","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1452160,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7141633/v1/2196fca9-c9ae-497f-9950-c318b2cddc48.pdf"},{"id":88907631,"identity":"4e652c5f-be28-40c2-b667-5e5ad65c24d1","added_by":"auto","created_at":"2025-08-12 14:52:37","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":24747401,"visible":true,"origin":"","legend":"Supplementary Information","description":"","filename":"SI.docx","url":"https://assets-eu.researchsquare.com/files/rs-7141633/v1/fddc51a4cc15743df1d7576d.docx"},{"id":88907630,"identity":"860588f9-c9b0-4f1a-8590-6498f3e5fb28","added_by":"auto","created_at":"2025-08-12 14:52:37","extension":"png","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":505804,"visible":true,"origin":"","legend":"Table of Content","description":"","filename":"TOC.png","url":"https://assets-eu.researchsquare.com/files/rs-7141633/v1/c49744ea346492a465d54348.png"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Reaction Mechanism and Role of Chlorine in Ethylene Epoxidation Revealed by in situ XPS","fulltext":[{"header":"Introduction","content":"\u003cp\u003eEthylene epoxidation (EPO) is a fundamental reaction in the chemical industry, producing ethylene oxide (EO)—a vital building block for polymers, surfactants, and antifreeze.\u003csup\u003e1\u003c/sup\u003e As a central pillar bridging hydrocarbon feedstocks to high-value downstream markets, it sustains global sectors like textiles, packaging, and automotive manufacturing.\u003csup\u003e2, 3\u003c/sup\u003e A key technical challenge in this process lies in enhancing the selectivity of silver-based catalysts to simultaneously boost process economics and mitigate CO\u003csub\u003e2\u003c/sub\u003e emissions.\u003csup\u003e4\u003c/sup\u003e While unpromoted silver catalysts exhibit limited selectivity, industrially relevant promoted catalysts can achieve EO selectivity approaching 90%. Among various promoters, chlorine-containing molecules, such as vinyl chloride (VC) and dichloroethane (DCE), are particularly effective. Kinetic analyses showed that, when cofed at ppm levels, such promoters enhance the selectivity by approximately 25% through poisoning over-oxidation sites while preserving active epoxidation centers.\u003csup\u003e5, 6\u003c/sup\u003e Other promoters, such as cesium and rhenium, are added in trace amounts during catalyst synthesis and also have a positive impact on the selectivity toward EO.\u003csup\u003e7, 8\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e \u003csup\u003e9\u003c/sup\u003e\u003c/p\u003e\n\u003cp\u003eDespite decades of research, the underlying mechanism governing the selectivity of ethylene oxidation remains a topic of ongoing debate. A long-standing hypothesis suggests that the reaction proceeds through the formation of an oxametallacycle (OMC) intermediate, which serves as a common precursor for both EO and acetaldehyde (AA), with acetaldehyde further oxidizing to CO\u003csub\u003e2\u003c/sub\u003e.\u003csup\u003e10, 11, 12\u003c/sup\u003e However, several key issues challenge this mechanism. First, direct in situ experimental evidence for OMC formation and its role under realistic conditions is lacking.\u0026nbsp;\u003csup\u003e13, 14, 15\u003c/sup\u003e Second, the asymmetric structure of OMC resembles AA more than EO, suggesting a higher activation barrier for EO formation. This has led researchers to explore specific oxygen species—such as sulfate impurities or the O\u003csub\u003e5\u003c/sub\u003e surface oxide phase—that might lower EO’s activation barrier.\u0026nbsp;\u003csup\u003e16, 17, 18, 19\u003c/sup\u003e Since lattice oxygen is widely known for its unselective behavior, it is necessary to identify how oxygen activation on silver leads to the selective oxidation of ethylene.\u003csup\u003e20\u003c/sup\u003e Previous kinetic studies show that EO and CO\u003csub\u003e2\u003c/sub\u003e have similar activation energies (~90–100 kJ/mol) but differ in reaction orders: –0.4 to 1 for C\u003csub\u003e2\u003c/sub\u003eH\u003csub\u003e4\u003c/sub\u003e and 0.5–1 (EO) vs. 0.2–1 (CO\u003csub\u003e2\u003c/sub\u003e) for O\u003csub\u003e2\u003c/sub\u003e.\u003csup\u003e1, 21\u003c/sup\u003e Both Langmuir–Hinshelwood and Eley–Rideal mechanisms have been proposed.\u003csup\u003e4, 22\u003c/sup\u003e\u0026nbsp; The addition of chlorine raises activation barriers and reduces oxygen species coverage, as indicated by an increased O\u003csub\u003e2\u003c/sub\u003e reaction order (0.7 to 1) with rising ethyl chloride (C\u003csub\u003e2\u003c/sub\u003eH\u003csub\u003e5\u003c/sub\u003eCl) concentration.\u003csup\u003e5, 23, 24, 25\u003c/sup\u003e Consequently, three fundamental questions remain: (i) what are the molecular-level active sites for selective versus unselective pathways on silver surfaces? (ii) how do different oxygen species formed under reaction conditions cooperate to control selectivity? and (iii) what is the mechanistic role of chlorine promoters?\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn this work, in situ X-ray photoelectron spectroscopy (XPS) was used to investigate surface intermediates under reaction conditions on a powder Ag/Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e catalyst.\u0026nbsp;\u003csup\u003e26\u003c/sup\u003e \u003csup\u003e27, 28\u003c/sup\u003e Experiments were carried out with X-rays generated by a synchrotron light source in order to improve the surface sensitivity, the signal intensity and the overall resolution. In the case of EPO, most of the studies primarily focused on the collection and interpretation of O 1s spectra exposing model surfaces (silver single crystals and/or silver powder) to reaction mixtures.\u003csup\u003e19, 29, 30\u003c/sup\u003e In our previous in situ XPS works, we utilized silver foil as a model catalyst to investigate the nature of surface oxygen and carbon species.\u003csup\u003e31, 32\u003c/sup\u003e Adsorbed diatomic oxygen was found to exhibit an electrophilic behavior and to react with adsorbed ethylene to yield EO. In parallel, adsorbed ethylene underwent dehydrogenation to C\u003csub\u003e2\u003c/sub\u003eH\u003csub\u003ex\u003c/sub\u003e (x=1-3) which was correlated with the production of CO\u003csub\u003e2\u003c/sub\u003e. However, the nature of surface intermediates and active sites on supported Ag/Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e catalysts during in situ EPO —particularly in the presence of ppm-level chlorine containing promoters—remains largely unexplored and presents significant experimental challenges.\u0026nbsp;Capitalizing on the knowledge gathered on model silver surfaces, we employed ambient pressure (AP)-XPS to directly probe catalytically active sites and reaction intermediates on Ag/Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e catalysts, both\u0026nbsp;in the presence\u0026nbsp;and\u0026nbsp;absence\u0026nbsp;of chlorine promotion. Combined with theoretical calculations, our findings reveal that ethylene spontaneously adsorbs and undergoes dehydrogenation at coordinatively unsaturated sites on the silver surface, leading to CO\u003csub\u003e2\u003c/sub\u003e formation. The introduction of chlorine promoters effectively blocks these surface defect sites and consequently enhances the selectivity towards EO. Furthermore, our results suggest that EO forms by the direct reaction between ethylene and dioxygen species activated on the reconstructed silver surface oxide.\u0026nbsp;\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eDeep dehydrogenation of ethylene on silver nanoparticles\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe Ag/Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e catalyst used in this work was synthesized using a low-surface area α-Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e support (5.2 m²/g, see Supplementary Fig. 1 and Supplementary Table 1), following industrially-relevant protocols.\u003csup\u003e33, 34\u003c/sup\u003e After silver deposition, the Ag/Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e catalyst exhibited a surface area of 4.6 m²/g. Along with the reflections of the α-Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u0026nbsp;\u003c/sub\u003esupport, X-ray diffraction (XRD) displayed \u0026nbsp;predominantly metallic silver (Supplementary Fig. 2), consistent with literature reports.\u003csup\u003e35\u003c/sup\u003e The catalytic performance of Ag/Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e with and without VC was evaluated in a flow reactor setup. The Ag/Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e catalyst was first exposed to a reaction mixture (C\u003csub\u003e2\u003c/sub\u003eH\u003csub\u003e4\u003c/sub\u003e + O\u003csub\u003e2\u003c/sub\u003e) until steady state EO production rate and selectivity were\u0026nbsp;established.\u0026nbsp;Upon co-feeding\u0026nbsp;0.5 ppm VC, the EO production rate\u0026nbsp;slightly decreased,\u0026nbsp;while CO\u003csub\u003e2\u003c/sub\u003e formation was significantly suppressed, resulting in an overall increase in EO selectivity from 38% to 69% (Fig. 1a). Then, the gas hourly space velocity (GHSV) was adjusted at a constant temperature of 230°C to maintain identical ethylene conversion in both the absence and presence of VC. As illustrated in Fig. 1b, EO selectivity increased by approximately 20% upon VC addition at the same ethylene conversion level. These results align with typical studies reporting that chlorine-containing promoters enhance EO selectivity primarily by suppressing CO\u003csub\u003e2\u003c/sub\u003e formation.\u003csup\u003e8, 11, 43\u003c/sup\u003e\u003c/p\u003e\n\u003cp\u003eTo investigate the electronic structure of all elements and the reaction intermediates under in situ EPO conditions, we conducted AP-XPS experiments. Two freshly synthesized Ag/Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e samples were sequentially introduced in the XPS flow measurement cell. High vacuum (HV) photoemission spectra were collected and used as references before dosing the reaction feed. Each sample was then exposed to a C\u003csub\u003e2\u003c/sub\u003eH\u003csub\u003e4\u003c/sub\u003e:O\u003csub\u003e2\u003c/sub\u003e (2:1) gas mixture (1 mbar total pressure), either in the presence or in the absence of 2 ppm VC, respectively, while the temperature was gradually increased from 150°C to 400°C. Steady-state measurement conditions were ensured at each temperature before starting spectra acquisition by monitoring the main core-level peaks as a function of time. The energy axis of all spectra was calibrated using the Al 2p peak at 74.6 eV as a reference (see Supplementary Fig. 3).\u003csup\u003e36\u003c/sup\u003e To probe the evolution of surface carbon intermediates, C 1s spectra were collected in situ at kinetic energies (KE) of 100 eV and 300 eV, corresponding to mean escape depths (MED) of 3.64 Å and 5.68 Å, respectively. Under HV conditions, all spectra exhibited a main peak at 285.3 eV, assigned to adventitious carbon, and a low intensity feature at 289.6 eV, which can be ascribed to carbonates (Supplementary Fig. 4 and Supplementary Table 2).\u003csup\u003e37, 38\u003c/sup\u003e Given that XPS is highly surface sensitive, the detection of a carbon contamination on a freshly synthesized powder sample was expected. In previous experiments performed on silver foil, carbon contamination was efficiently removed heating the sample in vacuum.\u003csup\u003e31, 32\u003c/sup\u003e Experimental results showed a relevant decrease of the carbon signal between 100 and 300°C. In this case, we did not pretreat Ag/Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e samples at high temperature to preserve the morphology of silver NPs. However, because in situ photoemission spectra were acquired at industrially relevant temperatures (starting from 150°C) in the presence of oxygen, we can expect a substantial decrease of adventitious carbon. Upon introducing the C\u003csub\u003e2\u003c/sub\u003eH\u003csub\u003e4\u003c/sub\u003e:O\u003csub\u003e2\u003c/sub\u003e reactant mixture, the line shape of C 1s spectra acquired at a KE of 100 eV (Fig. 2a) significantly changed compared to those obtained in vacuum (Supplementary Figure 4), revealing a broad feature between 283.0 eV and 286.0 eV. When VC was co-fed, the C 1s spectra acquired in the 150-250 °C temperature range showed a significantly suppressed low-binding-energy (BE) shoulder (BE\u0026lt;284.0 eV). Additionally, a low intensity peak at 289.6 eV was detected. Through deconvolution of the C 1s spectra (Supplementary Fig. 5, fitting parameters reported in Supplementary Table 3), a sharp peak at ~286.0 eV was discerned, corresponding to gas-phase C\u003csub\u003e2\u003c/sub\u003eH\u003csub\u003e4\u003c/sub\u003e,\u003csup\u003e31\u003c/sup\u003e along with six distinct peaks (Fig. 2b–g). The chemical nature of surface species assigned to each C 1s peak component was analyzed based on DFT calculations and on a previous study carried out on a silver foil.\u003csup\u003e31\u003c/sup\u003e Peaks at BE \u0026lt; 284.5 eV (C\u003csub\u003eEPO-0\u003c/sub\u003e and C\u003csub\u003eEPO-1\u003c/sub\u003e)\u0026nbsp;were attributed to dehydrogenation and C-C cleavage\u0026nbsp;products formed on the surface of silver nanoparticles (NP) upon ethylene adsorption. As shown in Fig. 2h–k, DFT calculations confirmed that consecutive dehydrogenation and cleavage of C-C bond led to a gradual negative BE shift of the corresponding adsorbates. This trend was consistently observed on Ag(111) (Fig. 2h-k), on the reconstructed p(4×4) surface oxide reconstruction, and on defective silver surfaces (Supplementary Fig. 6-7). Combining experimental results with BE simulations, the peak at 284.0 eV (C\u003csub\u003eEPO-1\u003c/sub\u003e) was assigned to C\u003csub\u003e2\u003c/sub\u003eH\u003csub\u003ex\u003c/sub\u003e (x=1-3) fragments, while the peak at 282.9 eV (C\u003csub\u003eEPO-0\u003c/sub\u003e) matched well with CH\u003csub\u003ey\u003c/sub\u003e (y=1-2) fragments adsorbed on the silver surface. The peak at 285.0 eV (C\u003csub\u003eEPO-2\u003c/sub\u003e) was assigned to adventitious carbon or long-chain carbonaceous species, with minor contributions from weakly adsorbed ethylene.\u003csup\u003e31\u003c/sup\u003e Carbonate species (C\u003csub\u003eEPO-5\u003c/sub\u003e) were always detected at 289.6 eV under reaction conditions. Weak features at 287.3 eV and 288.1 eV (C\u003csub\u003eEPO-3\u003c/sub\u003e and C\u003csub\u003eEPO-4\u003c/sub\u003e, respectively) were attributed to surface-adsorbed carbon monoxide and\u0026nbsp;EO, respectively, as previously observed.\u003csup\u003e31\u003c/sup\u003e The integrated peak areas of all carbon species were normalized by photon flux, C 1s cross-section, and MED, to obtain values proportional to surface atomic percentages. As shown in Figures 2b–c, in the absence of VC, C\u003csub\u003e2\u003c/sub\u003eH\u003csub\u003ex\u003c/sub\u003e and CH\u003csub\u003ey\u003c/sub\u003e species were abundant at 150°C. With increasing temperature, the CH\u003csub\u003ey\u003c/sub\u003e coverage diminished, while C\u003csub\u003e2\u003c/sub\u003eH\u003csub\u003ex\u003c/sub\u003e accumulated on the surface. Notably, the introduction of 2 ppm VC drastically suppressed the amount of C\u003csub\u003e2\u003c/sub\u003eH\u003csub\u003ex\u003c/sub\u003e and CH\u003csub\u003ey\u003c/sub\u003e species, particularly CH\u003csub\u003ey\u003c/sub\u003e, which remained at extremely low levels across all temperatures. This suggests that VC strongly inhibited both C–H and C–C bond cleavage of ethylene on the surface of silver. This result is further supported by the consistent trends observed in the C 1s spectra collected at KE of 300 eV (Supplementary Fig. 8–10 and Supplementary Table 4).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo further investigate the surface modifications induced by VC, we analyzed the Ag 3d spectra collected at a KE of 300 eV across different temperatures (Supplementary Fig. 11–13 and Supplementary Table 5). Approximately 70% of silver remained metallic (Ag\u003csub\u003e0\u003c/sub\u003e) under reaction conditions both in the presence and absence of VC, with the fraction of Ag\u003csub\u003e0\u003c/sub\u003e increasing with temperature. Supplementary Figure 11 shows that the full width at half maximum (FWHM) of the Ag 3d spectra gradually decreased with temperature both in the presence and in the absence of VC. Taking the spectra acquired at 400 °C as a reference for metallic silver NP, two main peak components could be deconvoluted in the 150-300 °C temperature range, which were labelled as Ag\u003csub\u003eα+β\u003c/sub\u003e and Ag\u003csub\u003eγ\u003c/sub\u003e. In agreement with previous results, the Ag\u003csub\u003eα+β\u003c/sub\u003e component at 367.4 eV was attributed to surface silver oxides.\u003csup\u003e31\u003c/sup\u003e Ag\u003csub\u003eγ\u003c/sub\u003e at 368.9 eV was newly identified and assigned to silver coordinated to CH\u003csub\u003ey\u003c/sub\u003e fragments, based on the similar temperature-dependent behavior as that of C\u003csub\u003eEPO-0\u0026nbsp;\u003c/sub\u003e(see Fig. 2b). \u0026nbsp;In the absence of VC, Ag\u003csub\u003eγ\u003c/sub\u003e accounted for approximately 10% of the total silver species below 200 °C, whereas in the presence of VC, the Ag\u003csub\u003eγ\u003c/sub\u003e signal was nearly undetectable. The Cl 2p spectra (Supplementary Fig. 14) acquired while feeding VC exhibited a faint signal due to the low VC concentration (2 ppm). Even after 210 repeated scans (over 2 hours), the signal-to-noise ratio remained low, and the doublet (2p\u003csub\u003e3/2\u003c/sub\u003e centered at approximately 198.0 eV, suggesting the presence of chloride) could hardly be resolved.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe reference O 1s spectrum acquired in HV (Supplementary Figure 15) displayed a main peak at 531.4 eV (dashed line), corresponding to lattice oxygen in Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e.\u003csup\u003e39\u003c/sup\u003e Under reaction conditions, the spectra broadened, due to the presence of a shoulder appearing above 532.0 eV, with the VC-treated sample showing a less prominent change. Difference spectra, obtained by using the O 1s in HV as a reference, revealed two broad features at 532.2 eV and 533.9 eV (Supplementary Fig. 16). Hydroxyls groups and oxygenated carbon species (e.g. adsorbed C-O) typically appear almost superimposed in the 531.5-532.5 eV BE range.\u003csup\u003e40\u003c/sup\u003e Since the introduction of VC did not significantly affect the fraction of oxygenated carbon species observed in the C 1s spectra (Figs. 2e and 2f), the peak at 532.2 eV was attributed to surface hydroxyl groups while the peak at 533.9 eV was assigned to adsorbed water.\u003csup\u003e41, 42\u003c/sup\u003e These findings suggest that VC inhibited the formation of both hydroxyl groups and adsorbed water on the catalyst surface under reaction conditions, in particular the formation of water below 250 °C. Overall, introducing VC into the reaction feed simultaneously suppressed the cleavage of C-H and C-C bonds and inhibited the accumulation of adsorbed hydroxyls, most likely contributing to the reduction in CO\u003csub\u003e2\u003c/sub\u003e formation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSurface defect sites as active centers for dehydrogenation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eEmploying DFT calculations, we investigated the sites that activated ethylene through the dehydrogenation pathway on the surface of silver —negatively influencing the selectivity towards EO. The Ag 3d spectra revealed that approximately 70% of silver remained metallic under in situ EPO conditions (supplementary Fig. 12). Given that\u0026nbsp;these\u0026nbsp;spectra were collected at a KE of 300 eV, the estimated photoelectrons MED was 5.7 Å, corresponding\u0026nbsp;to approximately two silver atomic layers.\u003csup\u003e44\u003c/sup\u003e Thus, around 30% of the top two silver layers contains oxide species, indicating the coexistence of silver oxide and metallic silver under reaction conditions. Previous STM studies reported that exposing the thermodynamically most stable facet of silver (Ag(111)) to oxygen yields a dispersed p(4×4) reconstruction, forming isolated surface oxide islands.\u003csup\u003e45, 46, 47, 48\u003c/sup\u003e Based on these observations, we constructed a series of atomistic slab models that expose a range of potential reactive sites and evaluated the energetics of ethylene activation on three representative sites: pristine Ag(111), reconstructed p(4x4) surface oxide and the interface between Ag(111) and p(4x4). The model systems included pristine Ag(111), Ag(111) with edge sites, Ag(111) with adsorbed molecular oxygen, reconstructed p(4×4) and c(4x8) surfaces, and configurations featuring coexisting p(4x4) and metallic silver on Ag(111) substrates, as shown in Supplementary Fig. 17. To evaluate the relative thermodynamic stability of these models, we calculated the surface energies as a function of temperature (see thermodynamic diagram in Supplementary Fig. 18). The p(4×4)-edge-3 structure—which incorporates coordinatively unsaturated oxygen atoms on the reconstructed surface—was the most stable substrate at EPO reaction temperature. Therefore, this configuration was selected as a representative p(4x4)-edge site for subsequent activation energy calculations.\u003c/p\u003e\n\u003cp\u003eThe adsorption of ethylene was relatively weak on most of the considered sites (see Supplementary Fig. 19), including pristine and stepped Ag(111), and oxygen-reconstructed flat surfaces. Adsorption energies ranged from ΔE = –0.30 to –0.55 eV. At these sites, the C=C bond typically resided atop a silver atom or bridged between two surface oxygen atoms. The strongest adsorption energy, –1.05 eV, occurred on the p(4×4)-edge model, where the ethylene C=C bond was atop a subsurface silver atom and surrounded by edge-located oxygen atoms.\u0026nbsp;We then evaluated the dehydrogenation process, by calculating the minimum energy pathway (MEP) using nudged elastic band (NEB) simulations. The resulting energy values simulated for the transition state and the optimized products state provided the energy barrier and overall reaction energy. As previously reported by Huš et al., the activation barrier for direct dehydrogenation on pristine Ag(111) is relatively high (1.89 eV), while the presence of oxygen is expected to lower it, due to its interaction with split off hydrogen.\u003csup\u003e49\u003c/sup\u003e We observed that the physisorbed oxygen molecule on pristine Ag(111) reduced the dehydrogenation energy barrier to 1.30 eV (Supplementary Fig. 20). On the flat reconstructed p(4x4) surface, the barrier decreased further to 1.17 eV (Supplementary Fig. 21), while the lowest barrier of 0.71 eV was observed on the p(4×4)-edge surface (Fig. 3a). These results point to the critical role of surface defects in combination with coordinated oxygen. On all substrates, dehydrogenated ethylene bound to one silver atom with the unsaturated carbon atom and tilted into an almost vertical geometry. The dissociated hydrogen bound to a nearby oxygen atom, when present, or adsorbed at an fcc site. The adsorption energies of the dehydrogenated intermediates, relative to gas-phase ethylene molecule, are reported in Fig. 3b, while the optimized configurations of these intermediates are shown in Fig. 3c-g. On both the pristine and stepped Ag(111), the adsorption energies are positive (1.08 eV and 0.89 eV, respectively), indicating unstable states. As expected, the presence of oxygen significantly stabilized the dehydrogenated intermediates, primarily due to the formation of O–H bonds (see the trend in Fig. 3b). The most favorable dehydrogenation site turned out to be at the p(4×4)-edge (adsorption energy of -1.05 eV), where the undercoordination of the oxygen atoms further strengthened the O-H bond.\u003c/p\u003e\n\u003cp\u003eSince p(4x4) defective sites are identified as the most active, we simulated the MEP for three possible full combustion processes, as shown in the Supplementary Fig. 22. The direct C-C bond cleavage appears unlikely due to the high energy of the initial intermediate (CH + CH\u003csub\u003e2\u003c/sub\u003e + OH). In contrast, the interaction of C\u003csub\u003e2\u003c/sub\u003eH\u003csub\u003e3\u003c/sub\u003e with lattice oxygen significantly stabilized the intermediates, ultimately leading to C-C bond cleavage and the formation of CH fragments and carbon dioxide.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eChlorine-mediated promotion of ethylene epoxidation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo highlight the role of chlorine in promoting the selectivity during ethylene epoxidation, we investigated the interaction of chlorine with the Ag/Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e catalyst. As shown in Supplementary Fig. 23, EDS mapping of a Ag/Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e catalyst after reacting under EPO conditions in the presence of 2 ppm VC for 48 hours revealed a chlorine signal, which spatially overlapped with that of silver. This suggests that chlorine preferentially reacted with the silver NP rather than with the alumina support. We utilized model surfaces (silver foil samples) and exposed them to a reaction gas mixture of C\u003csub\u003e2\u003c/sub\u003eH\u003csub\u003e4\u003c/sub\u003e:O\u003csub\u003e2\u003c/sub\u003e = 2:1, in the presence/absence of 10 ppm DCE, to enhance the photoemission signal of chlorine and investigate the interaction of chlorine with silver and its influence on the rection mechanism. The use of DCE was based on similar kinetic behavior observed for VC and DCE reported in the literature.\u003csup\u003e50\u003c/sup\u003e Upon dosing the EPO reaction mixture, the temperature was raised stepwise from room temperature to 400 °C. At each temperature, photoemission spectra were collected under steady state reaction conditions. A survey scan under HV confirmed the cleanliness of the surface, as evidenced by low intensity C 1s and O 1s signals on both samples (Supplementary Fig. 24). Upon introducing the reaction gas, Cl 2p spectra were collected at each temperature until signal stabilization was achieved. As shown in Supplementary Fig. 25-26, the Cl 2p signal remained weak but stable over time below 200 °C. When the temperature exceeded 300 °C, the Cl 2p signal significantly increased in intensity, yet still showed no signs of accumulation during data acquisition, indicating that the surface coverage of chlorine was stable throughout the experiments. Ag 3d spectra acquired at various temperatures (Supplementary Fig. 27) confirmed that the silver foil remained predominantly metallic. A shoulder centered at 367.4 eV, attributed to surface AgO\u003csub\u003ex\u003c/sub\u003e species,\u003csup\u003e32, 51, 52\u003c/sup\u003e was clearly detected in the absence of DCE, and almost disappeared upon DCE introduction. To investigate reaction intermediates, we collected C 1s spectra (Supplementary Fig. 28–30). Variations in ethylene gas-phase peak intensity are attributed to small differences in sample-to-analyzer distances between the two experiments. C\u003csub\u003e2\u003c/sub\u003eH\u003csub\u003ex\u003c/sub\u003e species (C\u003csub\u003eEPO-1\u003c/sub\u003e) were detected from room temperature, further confirming that spontaneous dehydrogenation of ethylene took place on the surface of silver. In contrast to the Ag/Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e catalyst, the intensity of C\u003csub\u003eEPO-0\u003c/sub\u003e (CH\u003csub\u003ex\u003c/sub\u003e species) was negligible on silver foil; the foil contained less defects than NP and did not facilitate C–C bond cleavage. Moreover, co-feeding DCE significantly suppressed the surface C\u003csub\u003e2\u003c/sub\u003eH\u003csub\u003ex\u003c/sub\u003e signal, further proving that chlorine promoters inhibit ethylene dehydrogenation and in good agreement with the results obtained on the Ag/Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e catalyst. O 1s spectra acquired at various temperatures (Supplementary Fig. \u0026nbsp;31–33) revealed that the peaks at 532.1 eV and 533.7 eV—assigned to surface hydroxyl species and adsorbed water, respectively—were significantly suppressed in the presence of DCE.\u003csup\u003e53, 54\u003c/sup\u003e As observed on Ag/Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e (Supplementary Fig. 15-16), co-feeding DCE suppressed ethylene dehydrogenation, thus the resulting formation of adsorbed hydroxyls and water. Supplementary Fig. 34 a-b show Cl 2p spectra acquired at KE of 200 and 480 eV, respectively, at increasing temperatures (from room temperature to 400 °C). At both KEs, the chlorine signal markedly intensified above 250 °C, indicating that DCE rapidly decomposed on silver. Normalized spectra acquired at a KE of 480 eV (Supplementary Fig. 34c) show that the Cl 2p\u003csub\u003e3/2\u003c/sub\u003e peak was consistently centered at 198.5 eV, matching well the BE of chloride.\u003csup\u003e55, 56, 57\u003c/sup\u003e Under EPO reaction conditions, DCE decomposed to partially decorate the silver foil surface with chloride. Supplementary Figure 33 a shows that the O 1s peak component at 529.3 eV, attributed to lattice oxygen (O\u003csub\u003eL\u003c/sub\u003e), increased less sharply at T\u0026gt;250 °C in the presence of DCE. This, combined with the rapid increase of the Cl 2p signal detected above 250 °C (Supplementary Fig. 34), suggests that chloride partially substituted lattice oxygen species. Based on these results, we explored the impact of chloride species using DFT simulations on the previously discussed p(4x4)-edge slab model. Notably, substituting one lattice oxygen with chlorine lowers the total energy of the system by more than 1 eV (Supplementary Fig. 35), indicating a thermodynamically favorable process. As the next step, we calculated the adsorption energies of VC at different sites (Supplementary Fig. 36). Similar to ethylene, VC preferentially adsorbed at the edge between Ag(111) and p(4×4) terraces (Supplementary Fig. 36f). Moreover, owing to the high electronegativity of chlorine, VC consistently exhibited stronger adsorption than ethylene. This implies that chlorine promoters suppress ethylene adsorption and, consequently, its activation.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eUsing transmission electron microscopy (TEM), van Hoof et al. revealed that voids/pores formed within silver NP during EPO in the absence of a chlorine promoter. Introducing VC significantly suppressed void formation.\u003csup\u003e11\u003c/sup\u003e The authors proposed that oxygen adsorption at silver NP grain boundaries induced defects formation during the reaction, whereas VC mitigated this effect. To validate this phenomenon on the Ag/Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e catalyst used in this work, we conducted TEM measurements on the fresh catalysts and on samples treated under 3 different EPO reaction conditions for 3 hours. As shown in Supplementary Fig. 37a–d, the fresh Ag/Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e catalyst exhibited relatively large silver particle sizes without observable internal voids. After 3 hours EPO reaction at 200 °C without co-feeding the VC promoter (Supplementary Fig. 37e–h), small void-like structures (~10 nm) started to appear inside silver particles. At 230 °C, the extent of void formation with diameters up to 50 nm increased. Notably, the formation of voids within silver particles introduced additional surface defect sites. In contrast, co-feeding 0.5 ppm VC at 230 °C markedly suppressed void formation, in agreement with previous literature.\u003csup\u003e11\u003c/sup\u003e The addition of a chlorine promoter increased the selectivity toward EO through a twofold effect: i) it suppressed ethylene activation by limiting the formation of lattice oxygen and inhibiting the adsorption of ethylene on the surface of silver, ii) it minimized the number of surface defect sites, where ethylene activation (dehydrogenation and C-C bond cleavage) is favored.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDirect oxidation pathway to EO\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe presence of dioxygen species on reconstructed silver oxide surface under EPO reaction conditions was revealed by means of AP-XPS and Raman spectroscopy by Pu et al. and Chen et al. \u003csup\u003e4, 21, 30\u003c/sup\u003e\u0026nbsp; Chen et al. reported that the activation energy for oxygen dissociation on silver surfaces typically exceeds 1.1 eV.\u003csup\u003e58\u003c/sup\u003e Similarly, experimental studies on silver foil showed that weakly adsorbed dioxygen species embedded within reconstructed surfaces only formed at temperatures above 150 °C.\u003csup\u003e32\u003c/sup\u003e The formation of such oxygen species strongly correlated with the EPO process. To verify this hypothesis, we simulated the ethylene oxidation MEP on both p(4×4) and c(4×8) reconstructions, where weakly bound dioxygen inserts into the surface oxide lattice (orange in Fig. 4).\u0026nbsp;As expected, ethylene initially physisorbed on both substrates, with weak adsorption energies of –0.35 eV and –0.41 eV, respectively. In the presence of additional oxygen species, the oxidation reaction leading to adsorbed EO occurred over a relatively low activation barrier, i.e., 0.41 eV and 0.24 eV on p(4×4) and c(4×8) surfaces, respectively. These findings suggest that once the selective dioxygen species forms, ethylene rapidly undergoes epoxidation, meaning that the rate-determining step in the EO formation process is the generation of dioxygen species from gaseous oxygen activation.\u003csup\u003e32, 58\u003c/sup\u003e This mechanistic insight aligns with prior work.\u003csup\u003e59\u003c/sup\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eBased on these findings, we propose that two distinct silver-catalyzed mechanisms are responsible for EPO and ethylene combustion (Fig. 5). These mechanisms deviate from the reaction pathway involving a common OMC intermediate, as further demonstrated by the temperature-programmed surface reaction experiments shown in Fig. 38 and described in the SI.\u003csup\u003e11, 12\u003c/sup\u003e Under in situ EPO conditions, two main types of oxygen species are formed on the surface of silver. Ethylene interacts with electrophilic dioxygen species to form EO through a direct oxidative pathway, in which the formation of surface dioxygen is the rate-determining step. In contrast, at surface defect sites, ethylene readily adsorbs and undergoes deep dehydrogenation assisted by lattice oxygen and C-C cleavage, leading to the formation of carbon dioxide through an unselective pathway. Chlorine suppresses the combustion pathway by preferentially binding to defect sites as chloride, thereby passivating undercoordinated silver atoms, reducing lattice oxygen formation, and inhibiting C–H activation. Furthermore, the addition of VC hinders ethylene adsorption and limits the formation of porous structures during the reaction, thereby reducing the number of surface defect sites.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eIn this study, we combined in situ XPS and DFT calculations to elucidate the reaction mechanism of EPO on Ag/Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e catalysts. Under in situ EPO conditions, C\u003csub\u003e2\u003c/sub\u003eH\u003csub\u003ex\u003c/sub\u003e and CH\u003csub\u003ey\u003c/sub\u003e fragments were identified alongside adsorbed hydroxyls and water, which were confirmed as key precursors to carbon dioxide formation based on reactivity studies. These results demonstrate that ethylene preferentially undergoes activation on the surface of silver via oxygen-assisted dehydrogenation pathways to form carbon dioxide. In the presence of chlorine, the surface concentrations of dehydrogenation fragments of ethylene, adsorbed hydroxyls and water were significantly suppressed. This indicates that chlorine-containing promoters enhance EO selectivity primarily by inhibiting the dehydrogenation of ethylene. DFT calculations revealed that ethylene strongly adsorbs on defect sites, particularly at the interface between metallic and oxygen reconstructed surface domains. These sites facilitated the reaction of ethylene with lattice oxygen, confirming that defects trigger carbon dioxide formation. We further demonstrated that chlorine-containing promoters suppressed ethylene activation by reducing the fraction of surface lattice oxygen, inhibiting ethylene adsorption, and decreasing the density of defect sites. These combined effects significantly enhanced the selectivity toward EO. Moreover, DFT simulations indicated that EO formation may proceed through a direct oxidation pathway involving adsorbed dioxygen species formed on oxygen reconstructed silver surfaces. These findings provide new insights into the mechanistic pathways of EPO on silver-based catalysts, proposing parallel reaction mechanisms. This work does not only advance the understanding of the EPO process but also suggests a strategy for enhancing the selectivity toward EO by hindering the formation of defect sites on the silver surface.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cstrong\u003eSynthesis of Ag/Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e catalysts\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe 15 wt% Ag/Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e catalyst was prepared via incipient wetness impregnation using silver oxalate (Ag\u003csub\u003e2\u003c/sub\u003eC\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e4\u003c/sub\u003e) as the silver precursor, following widely adopted protocols in the literature.\u003csup\u003e34, 60\u003c/sup\u003e Ag\u003csub\u003e2\u003c/sub\u003eC\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e4\u003c/sub\u003e was synthesized by mixing aqueous solutions of oxalic acid (Sigma Aldrich, 99%) and silver nitrate (Merck, 99%) in a 1:2 molar ratio. The precipitate was collected by centrifugation, washed three times with deionized water, and dried overnight in static air at 60 \u0026deg;C. A solution composed of silver oxalate and ethylenediamine (Sigma Aldrich, 99%) in a 1:4 molar ratio, diluted in 2 cm\u003csup\u003e3\u003c/sup\u003e of deionized water, was impregnated onto \u0026alpha;-Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e (Thermo Scientific, 99.9%) via incipient wetness. The impregnated support was dried overnight at 60 \u0026deg;C and subsequently calcined in static air at 400 \u0026deg;C for 6 hours.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCharacterization of Ag/Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e catalysts\u003c/strong\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eXRD patterns were recorded using a Bruker D8 ADVANCE diffractometer with Cu K\u003csub\u003e\u0026alpha;\u003c/sub\u003e radiation (\u0026lambda; = 1.54 \u0026Aring;) at a scan rate of 0.02\u0026deg; per step over the 2\u0026theta; range of 20\u0026ndash;80\u0026deg;. Nitrogen physisorption measurements were conducted at 77 K using a Micromeritics 3Flex instrument. Prior to analysis, the samples were degassed at 120 \u0026deg;C overnight. The specific surface area was determined by the Brunauer\u0026ndash;Emmett\u0026ndash;Teller (BET) method\u0026nbsp;by\u0026nbsp;fitting the isotherms. Transmission electron microscopy (TEM) and scanning transmission electron microscopy (STEM) analyses were performed on a JEOL JEM-F200 microscope operated at 200 kV and equipped with two silicon drift detectors (SDDs) for energy-dispersive X-ray spectroscopy (EDX). Samples were prepared by dispersing the catalyst powder in ethanol followed by drop-cast onto a copper grid.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eIn situ AP-XPS measurements of Ag/ Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e and silver foil\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAP-XPS measurements of Ag/Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e and silver foil were performed at the X07DB \u0026ldquo;In Situ Spectroscopy\u0026rdquo; beamline at the Swiss Light Source (SLS), Paul Scherrer Institute. Ag/Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e powders were pressed onto two separate 10 \u0026times; 10 mm gold meshes, which were then mounted on a manipulator and introduced into the solid-gas interface endstation. Initially, Al 2p, Ag 3d, C 1s, and O 1s spectra were collected under high vacuum at kinetic energies of 100 eV and 300 eV as references. The sample was then exposed in situ to a gas mixture of ethylene and oxygen in a 2:1 ratio at a total pressure of 1 mbar. For one set of samples, 2 ppm of VC were co-dosed with the reactants feed. Gas mixtures were introduced using mass flow controllers, while the total pressure was regulated downstream by a tunable diaphragm valve connected to a root pump, enabling precise control over partial pressures and gas ratios. The pressure was monitored using Baratron gauges. Samples were heated by an infrared laser targeting the rear of the sample holder, with temperature measured by a Pt100 sensor. Spectra were collected sequentially while exposing the samples to the gas mixture at various temperatures. Linearly polarized synchrotron light was used throughout the experiments. C 1s spectra were acquired using photon energies of 385 and 595 eV (kinetic energies of 100 and 300 eV). O 1s, Al 2p, Ag 3d, and Cl 2p spectra were collected at h\u0026nu; of 840, 385, 675, and 505 eV, respectively, all corresponding to kinetic energies of 300 eV to ensure comparable MEDs. The BE scale was calibrated using the Al 2p peak at 74.6 eV. All spectra were background-subtracted using a Shirley function and fitted with Voigt functions. Fitting parameters are summarized in Supplementary Tables 2\u0026ndash;5.\u003c/p\u003e\n\u003cp\u003eFor experiments employing silver foil, silver foils (Alfa Aesar, 99.99%) were cut into 10 \u0026times; 10 mm squares and cleaned sequentially with acetone, isopropanol, and deionized water, followed by plasma cleaning in oxidizing (O\u003csub\u003e2\u003c/sub\u003e/Ar) and reducing (H\u003csub\u003e2\u003c/sub\u003e/He) atmospheres to remove surface contaminants. The cleaned foils were mounted onto the manipulator and introduced into the solid-gas interface endstation. Prior to measurements, foils were preheated to 400 \u0026deg;C under vacuum to desorb residual adsorbates, and surface cleanliness was verified by XPS. The subsequent experimental procedure was identical to that used for Ag/Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e, with the exception that 10 ppm of dichloroethane were co-fed with the reactant gases for one set of experiments. Ag 3d, O 1s, and C 1s spectra were collected at h\u0026nu; of 575, 735, and 490 eV (KE = 200 eV), while Cl 2p spectra were collected at h\u0026nu; of 400 and 680 eV (KE = 200 and 480 eV). Fitting parameters are summarized in Tables S6\u0026ndash;S7.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCatalytic Performance and Kinetic Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe catalytic performance was evaluated in a commercial plug-flow CATLAB reactor. An online gas chromatograph (MicroGC, INFICON) equipped with two TCD detectors was connected to the reactor outlet for product analysis. Oxygen, ethylene, EO, carbon dioxide, and water were monitored at 5-minute intervals. The feed gas consisted of 15 vol% ethylene and 3 vol% oxygen, balanced with nitrogen to a total flow rate of 20 mL/min; optionally, 2.5 vol% of 20 ppm VC in He was added. The reactor was loaded with 5\u0026ndash;20 mg of catalyst, heated to 230 \u0026deg;C with a rate of 5 \u0026deg;C/min, and maintained at this temperature for 3\u0026ndash;48 hours under continuous flow. Catalysts were evaluated at same ethylene conversions to compare EO selectivity. Ethylene conversion and ethylene oxide selectivity were calculated using equations (1) and (2), respectively.\u003c/p\u003e\n\u003cp\u003e\u003cimg width=\"236\" height=\"41\" src=\"data:image/png;base64,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\" alt=\"image\"\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; (1)\u003c/p\u003e\n\u003cp\u003e\u003cimg width=\"193\" height=\"35\" src=\"data:image/png;base64,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\" alt=\"image\"\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;(2)\u003c/p\u003e\n\u003cp\u003ewhere the subscripts in and out refer to the concentrations measured at the reactor inlet and outlet, respectively.\u003c/p\u003e\n\u003cp\u003eTemperature-programmed surface reaction experiments were performed using the same CATLAB reactor, with outlet gases monitored by an online mass spectrometer. The Ag/Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e catalyst was first exposed to 7.5 vol% oxygen in argon at room temperature. It was then heated to 400 \u0026deg;C and held until the carbon dioxide and water signals were no longer detected. After cooling to room temperature in the O\u003csub\u003e2\u003c/sub\u003e/Ar mixture, the catalyst was exposed to a gas mixture of 5 vol% ethylene and 7.5 vol% oxygen in argon for 10 minutes. The ethylene feed was then cut off, and the catalyst was purged with O\u003csub\u003e2\u003c/sub\u003e/Ar until the ethylene signal fully disappeared. Subsequently, the temperature was ramped at 5 \u0026deg;C/min while recording mass signals for ethylene (m/z = 27), O\u003csub\u003e2\u003c/sub\u003e (m/z = 32), EO (m/z = 43, minor fragment), and carbon dioxide (m/z = 44).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDFT simulations\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll density functional theory (DFT) calculations were performed using the CP2K program package (Quickstep)\u003csup\u003e\u0026nbsp;56\u003c/sup\u003e, employing the Kohn-Sham formalism within the hybrid Gaussian and plane wave (GPW) framework. Goedecker-Teter-Hutter (GTH) pseudopotentials\u003csup\u003e57\u003c/sup\u003e and molecularly optimized (MOLOPT) Gaussian basis sets\u003csup\u003e58\u003c/sup\u003e were used to describe the valence electrons. Specifically, DZVP-MOLOPT-GTH basis sets were applied for Ag, O, C, and H atoms. A plane-wave cutoff energy of 600 Ry was used for the auxiliary basis set. All calculations were performed under periodic boundary conditions, using only the Gamma point, and were spin-polarized. Since an asymmetric slab model was used\u0026mdash;featuring the oxidized-reconstructed surface and adsorbates on one side only\u0026mdash;a dipole correction was applied along the z-axis to account for the surface dipole\u003csup\u003e59\u003c/sup\u003e. The structural and electronic properties were computed at the generalized gradient approximation (GGA) level of theory, using the Perdew-Burke-Ernzerhof (PBE) functional\u003csup\u003e60\u003c/sup\u003e, with Grimme-D3 dispersion corrections to account for long-range van der Waals interactions\u003csup\u003e61\u003c/sup\u003e. Geometry optimizations were carried out using the Broyden-Fletcher-Goldfarb-Shanno (BFGS) algorithm\u003csup\u003e62\u003c/sup\u003e, with a force convergence criterion of 10⁻\u0026sup3; Hartree/Bohr\u0026sup3;. The Climbing-Image Nudged Elastic Band (CI-NEB) method was applied to find minimum-energy pathway and energy barriers of the reaction process, and the transition states\u003csup\u003e63\u003c/sup\u003e. \u0026nbsp;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo accurately simulate various oxygen species and intermediates on an Ag/Al₂O₃\u0026nbsp;substrate, we considered both the clean Ag(111) surface and two thermodynamically stable oxidized-reconstructed Ag(111) surface structures\u0026mdash;namely, the p(4\u0026times;4) and c(4\u0026times;8) reconstructions\u0026mdash;as used in our previous study\u003csup\u003e28\u003c/sup\u003e. Additionally, to mimic defective Ag surfaces with unsaturated Ag sites exposed to the gas phase, we constructed step-edge models of the Ag(111) and p(4\u0026times;4) surfaces, referred to as Ag(111)-edge and p(4\u0026times;4)-edge, respectively. The models consisted of a 5-layer Ag(111) slab, where the top layer was modified by introducing oxygen atoms to form the p(4\u0026times;4) and c(4\u0026times;8) reconstructions. Various carbon-containing species were then adsorbed onto these surfaces to study their interactions. The lateral dimensions of the simulation cell were set to 23.19 \u0026times; 23.19 \u0026Aring;\u0026sup2; (for Ag(111)), and a vacuum region of 20 \u0026Aring; was included both above and below the slab to eliminate spurious interactions between periodic images.\u003c/p\u003e\n\u003cp\u003eThe calculation of the C and O K-edge, used to reproduce the binding energies of 1s electrons, was achieved by applying the Gaussian augmented plane wave method (GAPW). For the C and O elements, all electrons were explicitly considered (no pseudopotentials) and 6-311G** all-electron basis sets were employed to expand the molecular orbitals, as used in previous works\u003csup\u003e27,28\u003c/sup\u003e. The calculations of the energies of the core states were based on the Slater transition potential method with half-core hole approximation, where initial and final state effects were accounted for by electronic energy eigenvalue calculations after removing half an electron from the core state.\u003csup\u003e64,65\u003c/sup\u003e\u0026nbsp;\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eM.G. and L.A. acknowledge the Swiss National Science Foundation (project number 196946) for their support. N.D. and M. I. thank for the generous allocation of computing resources from the Swiss National Supercomputing Center (CSCS) under Project ID s1279, and both CSCS and the Alfred Werner-Legat for the computing resources under Project ID uzh35.\u0026nbsp;\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003ePu TC, Tian HJ, Ford ME, Rangarajan S, Wachs IE. Overview of Selective Oxidation of Ethylene to Ethylene Oxide by Ag Catalysts. \u003cem\u003eACS Catal\u003c/em\u003e 2019, \u003cstrong\u003e9\u003c/strong\u003e(12)\u003cstrong\u003e:\u003c/strong\u003e 10727-10750.\u003c/li\u003e\n\u003cli\u003eHerzberger J, Niederer K, Pohlit H, Seiwert J, Worm M, Wurm FR\u003cem\u003e, et al.\u003c/em\u003e Polymerization of Ethylene Oxide, Propylene Oxide, and Other Alkylene Oxides: Synthesis, Novel Polymer Architectures, and Bioconjugation. \u003cem\u003eChem Rev\u003c/em\u003e 2016, \u003cstrong\u003e116\u003c/strong\u003e(4)\u003cstrong\u003e:\u003c/strong\u003e 2170-2243.\u003c/li\u003e\n\u003cli\u003eStadler BM, Wulf C, Werner T, Tin S, de Vries JG. Catalytic Approaches to Monomers for Polymers Based on Renewables. \u003cem\u003eACS Catal\u003c/em\u003e 2019, \u003cstrong\u003e9\u003c/strong\u003e(9)\u003cstrong\u003e:\u003c/strong\u003e 8012-8067.\u003c/li\u003e\n\u003cli\u003ePu T, Setiawan A, Foucher AC, Guo M, Jehng JM, Zhu M\u003cem\u003e, et al.\u003c/em\u003e Revealing the Nature of Active Oxygen Species and Reaction Mechanism of Ethylene Epoxidation by Supported Ag/alpha-Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e Catalysts. \u003cem\u003eACS Catal\u003c/em\u003e 2024, \u003cstrong\u003e14\u003c/strong\u003e(1)\u003cstrong\u003e:\u003c/strong\u003e 406-417.\u003c/li\u003e\n\u003cli\u003eIyer KR, Bhan A. Interdependencies Among Ethylene Oxidation and Chlorine Moderation Catalytic Cycles Over Promoted Ag/\u0026alpha;-Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e Catalysts. \u003cem\u003eACS Catal\u003c/em\u003e 2021, \u003cstrong\u003e11\u003c/strong\u003e(24)\u003cstrong\u003e:\u003c/strong\u003e 14864-14876.\u003c/li\u003e\n\u003cli\u003eLockemeyer JR, Lohr TL. Ethylene Oxide Catalysis Under Commercial Conditions \u0026ndash; A Guide for Researchers. \u003cem\u003eChemCatChem\u003c/em\u003e 2023, \u003cstrong\u003e15\u003c/strong\u003e(13).\u003c/li\u003e\n\u003cli\u003eDiao W, DiGiulio CD, Schaal MT, Ma S, Monnier JR. An investigation on the role of Re as a promoter in Ag Cs Re/\u0026alpha;-Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e high-selectivity, ethylene epoxidation catalysts. \u003cem\u003eJ Catal\u003c/em\u003e 2015, \u003cstrong\u003e322:\u003c/strong\u003e 14-23.\u003c/li\u003e\n\u003cli\u003eJankowiak J, Barteau M. Ethylene epoxidation over silver and copper\u0026ndash;silver bimetallic catalysts: II. Cs and Cl promotion. \u003cem\u003eJ Catal\u003c/em\u003e 2005, \u003cstrong\u003e236\u003c/strong\u003e(2)\u003cstrong\u003e:\u003c/strong\u003e 379-386.\u003c/li\u003e\n\u003cli\u003eJalil A, Happel EE, Cramer L, Hunt A, Hoffman AS, Waluyo I\u003cem\u003e, et al.\u003c/em\u003e Nickel promotes selective ethylene epoxidation on silver. \u003cem\u003eScience\u003c/em\u003e 2025, \u003cstrong\u003e387:\u003c/strong\u003e 869\u0026ndash;873.\u003c/li\u003e\n\u003cli\u003eLinic S, Barteau MA. Control of Ethylene Epoxidation Selectivity by Surface Oxametallacycles. \u003cem\u003eJ Am Chem Soc\u003c/em\u003e 2003, \u003cstrong\u003e125:\u003c/strong\u003e 4034-4035.\u003c/li\u003e\n\u003cli\u003evan Hoof AJF, Filot IAW, Friedrich H, Hensen EJM. Reversible Restructuring of Silver Particles during Ethylene Epoxidation. \u003cem\u003eACS Catal\u003c/em\u003e 2018, \u003cstrong\u003e8\u003c/strong\u003e(12)\u003cstrong\u003e:\u003c/strong\u003e 11794-11800.\u003c/li\u003e\n\u003cli\u003eChristopher P, Linic S. Engineering Selectivity in Heterogeneous Catalysis: Ag Nanowires as Selective Ethylene Epoxidation Catalysts. \u003cem\u003eJ Am Chem Soc\u003c/em\u003e 2008, \u003cstrong\u003e130:\u003c/strong\u003e 11264\u0026ndash;11265.\u003c/li\u003e\n\u003cli\u003eJones GS, Mavrikakis M, Barteau MA, Vohs JM. First Synthesis, Experimental and Theoretical Vibrational Spectra of an Oxametallacycle on a Metal Surface. \u003cem\u003eJ Am Chem Soc \u003c/em\u003e1998, \u003cstrong\u003e120:\u003c/strong\u003e 3196-3204.\u003c/li\u003e\n\u003cli\u003eWu G, Stacchiola D, Kaltchev M, Tysoe WT. The adsorption and reaction of 2-iodoethanol on Ag(111). \u003cem\u003eSurf Sci\u003c/em\u003e 2000, \u003cstrong\u003e463:\u003c/strong\u003e 81\u0026ndash;92.\u003c/li\u003e\n\u003cli\u003eLinic S, Medlin JW, Barteau MA. Synthesis of Oxametallacycles from 2-Iodoethanol on Ag(111) and the Structure Dependence of Their Reactivity. \u003cem\u003eLangmuir\u003c/em\u003e 2002, \u003cstrong\u003e18:\u003c/strong\u003e 5197-5204.\u003c/li\u003e\n\u003cli\u003eTorres D, Lopez N, Illas F, Lambert RM. Why Copper Is Intrinsically More Selective than Silver in Alkene Epoxidation: Ethylene Oxidation on Cu(111) versus Ag(111). \u003cem\u003eJ Am Chem Soc\u003c/em\u003e 2005, \u003cstrong\u003e127:\u003c/strong\u003e 10774-10775.\u003c/li\u003e\n\u003cli\u003eDerouin J, Farber RG, Turano ME, Iski EV, Killelea DR. Thermally Selective Formation of Subsurface Oxygen in Ag(111) and Consequent Surface Structure. \u003cem\u003eACS Catal\u003c/em\u003e 2016, \u003cstrong\u003e6\u003c/strong\u003e(7)\u003cstrong\u003e:\u003c/strong\u003e 4640-4646.\u003c/li\u003e\n\u003cli\u003eChen DX, Kang PL, Liu ZP. Active Site of Catalytic Ethene Epoxidation: Machine-Learning Global Pathway Sampling Rules Out the Metal Sites. \u003cem\u003eACS Catal\u003c/em\u003e 2021, \u003cstrong\u003e11\u003c/strong\u003e(13)\u003cstrong\u003e:\u003c/strong\u003e 8317-8326.\u003c/li\u003e\n\u003cli\u003eJones TE, Wyrwich R, Bocklein S, Carbonio EA, Greiner MT, Klyushin AY\u003cem\u003e, et al.\u003c/em\u003e The Selective Species in Ethylene Epoxidation on Silver. \u003cem\u003eACS Catal\u003c/em\u003e 2018, \u003cstrong\u003e8\u003c/strong\u003e(5)\u003cstrong\u003e:\u003c/strong\u003e 3844-3852.\u003c/li\u003e\n\u003cli\u003ePu T, Setiawan A, Rangarajan S, Wachs IE. Progress and perspective on the fundamental understanding of structure\u0026ndash;activity/selectivity relationships for Ag catalyzed ethylene epoxidation. \u003cem\u003eCatalysis Today\u003c/em\u003e 2025, \u003cstrong\u003e455\u003c/strong\u003e.\u003c/li\u003e\n\u003cli\u003eChen C-T, Sviripa A, Verma S, Paolucci C, Flaherty DW. Reactions of Surface Peroxides Contribute to Rates and Selectivities for C\u003csub\u003e2\u003c/sub\u003eH\u003csub\u003e4\u003c/sub\u003e Epoxidation on Silver. \u003cem\u003eACS Catal\u003c/em\u003e 2025\u003cstrong\u003e:\u003c/strong\u003e 1387-1398.\u003c/li\u003e\n\u003cli\u003ePetrov L, Eliyas A, Shopov D. A Kinetic Model of Steady State Ethylene Epoxidation over a Supported Silver Catalyst. \u003cem\u003eApplied Catalysis\u003c/em\u003e 1985, \u003cstrong\u003e18:\u003c/strong\u003e 87-103.\u003c/li\u003e\n\u003cli\u003eChen CJ, Harris JW, Bhan A. Kinetics of Ethylene Epoxidation on a Promoted Ag/alpha-Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e Catalyst-The Effects of Product and Chloride Co-Feeds on Rates and Selectivity. \u003cem\u003eChem EurJ \u003c/em\u003e2018, \u003cstrong\u003e24\u003c/strong\u003e(47)\u003cstrong\u003e:\u003c/strong\u003e 12405-12415.\u003c/li\u003e\n\u003cli\u003eHarris JW, Wang L, DeWilde JF. Kinetic Modeling of Ethylene Epoxidation Kinetics as a Function of Chlorine Coverage over a Highly Promoted Silver Catalyst. \u003cem\u003eInd Eng Chem Res\u003c/em\u003e 2019, \u003cstrong\u003e58\u003c/strong\u003e(41)\u003cstrong\u003e:\u003c/strong\u003e 19061-19071.\u003c/li\u003e\n\u003cli\u003eHarris JW, Herron JA, DeWilde JF, Bhan A. Molecular characteristics governing chlorine deposition and removal on promoted Ag catalysts during ethylene epoxidation. \u003cem\u003eJ Catal\u003c/em\u003e 2019, \u003cstrong\u003e377:\u003c/strong\u003e 378-388.\u003c/li\u003e\n\u003cli\u003eFoster AJ, Lobo RF. Identifying reaction intermediates and catalytic active sites through in situ characterization techniques. \u003cem\u003eChem Soc Rev\u003c/em\u003e 2010, \u003cstrong\u003e39\u003c/strong\u003e(12)\u003cstrong\u003e:\u003c/strong\u003e 4783-4793.\u003c/li\u003e\n\u003cli\u003eRoy K, Artiglia L, van Bokhoven JA. Ambient Pressure Photoelectron Spectroscopy: Opportunities in Catalysis from Solids to Liquids and Introducing Time Resolution. \u003cem\u003eChemCatChem\u003c/em\u003e 2018, \u003cstrong\u003e10\u003c/strong\u003e(4)\u003cstrong\u003e:\u003c/strong\u003e 666-682.\u003c/li\u003e\n\u003cli\u003eArtiglia L, Sushkevich VL, Palagin D, Knorpp AJ, Roy K, van Bokhoven JA. In Situ X-ray Photoelectron Spectroscopy Detects Multiple Active Sites Involved in the Selective Anaerobic Oxidation of Methane in Copper-Exchanged Zeolites. \u003cem\u003eACS Catal\u003c/em\u003e 2019, \u003cstrong\u003e9\u003c/strong\u003e(8)\u003cstrong\u003e:\u003c/strong\u003e 6728-6737.\u003c/li\u003e\n\u003cli\u003eCarbonio EA, Rocha TCR, Klyushin AY, Pis I, Magnano E, Nappini S\u003cem\u003e, et al.\u003c/em\u003e Are multiple oxygen species selective in ethylene epoxidation on silver? \u003cem\u003eChem Sci\u003c/em\u003e 2018, \u003cstrong\u003e9\u003c/strong\u003e(4)\u003cstrong\u003e:\u003c/strong\u003e 990-998.\u003c/li\u003e\n\u003cli\u003ePu T, Setiawan A, Mosevitzky Lis B, Zhu M, Ford ME, Rangarajan S\u003cem\u003e, et al.\u003c/em\u003e Nature and Reactivity of Oxygen Species on/in Silver Catalysts during Ethylene Oxidation. \u003cem\u003eACS Catal\u003c/em\u003e 2022, \u003cstrong\u003e12\u003c/strong\u003e(8)\u003cstrong\u003e:\u003c/strong\u003e 4375-4381.\u003c/li\u003e\n\u003cli\u003eGuo M, Dongfang N, Liu Q, Iannuzzi M, van Bokhoven JA, Artiglia L. Reaction Intermediates Involved in the Epoxidation of Ethylene Over Silver Revealed by In Situ Photoelectron Spectroscopy. \u003cem\u003eSmall\u003c/em\u003e 2025, \u003cstrong\u003e21\u003c/strong\u003e(5)\u003cstrong\u003e:\u003c/strong\u003e e2408432.\u003c/li\u003e\n\u003cli\u003eGuo M, Dongfang N, Iannuzzi M, van Bokhoven JA, Artiglia L. Structure and Reactivity of Active Oxygen Species on Silver Surfaces for Ethylene Epoxidation. \u003cem\u003eACS Catal\u003c/em\u003e 2024, \u003cstrong\u003e14:\u003c/strong\u003e 10234-10244.\u003c/li\u003e\n\u003cli\u003eAlzahrani HA, Bravo-Su\u0026aacute;rez JJ. In situ Raman spectroscopy study of silver particle size effects on unpromoted Ag/\u0026alpha;-Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e during ethylene epoxidation with molecular oxygen. \u003cem\u003eJ Catal\u003c/em\u003e 2023, \u003cstrong\u003e418:\u003c/strong\u003e 225-236.\u003c/li\u003e\n\u003cli\u003eKeijzer PH, van den Reijen JE, Keijzer CJ, de Jong KP, de Jongh PE. Influence of atmosphere, interparticle distance and support on the stability of silver on \u0026alpha;-alumina for ethylene epoxidation. \u003cem\u003eJ Catal\u003c/em\u003e 2021.\u003c/li\u003e\n\u003cli\u003evan Hoof AJF, Hermans EAR, van Bavel AP, Friedrich H, Hensen EJM. Structure Sensitivity of Silver-Catalyzed Ethylene Epoxidation. \u003cem\u003eACS Catal\u003c/em\u003e 2019, \u003cstrong\u003e9\u003c/strong\u003e(11)\u003cstrong\u003e:\u003c/strong\u003e 9829-9839.\u003c/li\u003e\n\u003cli\u003ePeng J, Sun Q, Zhai Z, Yuan J, Huang X, Jin Z\u003cem\u003e, et al.\u003c/em\u003e Low temperature, solution-processed alumina for organic solar cells. \u003cem\u003eNanotechnology\u003c/em\u003e 2013, \u003cstrong\u003e24\u003c/strong\u003e(48).\u003c/li\u003e\n\u003cli\u003eChen X, Wang X, Fang D. A review on C1s XPS-spectra for some kinds of carbon materials. \u003cem\u003eFullerenes, Nanotubes Carbon Nanostruct\u003c/em\u003e 2020, \u003cstrong\u003e28\u003c/strong\u003e(12)\u003cstrong\u003e:\u003c/strong\u003e 1048-1058.\u003c/li\u003e\n\u003cli\u003eGengenbach TR, Major GH, Linford MR, Easton CD. Practical guides for x-ray photoelectron spectroscopy (XPS): Interpreting the carbon 1s spectrum. \u003cem\u003eJ Vac Sci Technol\u003c/em\u003e 2021, \u003cstrong\u003e39\u003c/strong\u003e(1)\u003cstrong\u003e:\u003c/strong\u003e 013204.\u003c/li\u003e\n\u003cli\u003eLee D, Park JW, Cho NK, Lee J, Kim YS. Verification of Charge Transfer in Metal-Insulator-Oxide Semiconductor Diodes via Defect Engineering of Insulator. \u003cem\u003eSci Rep\u003c/em\u003e 2019, \u003cstrong\u003e9\u003c/strong\u003e(1)\u003cstrong\u003e:\u003c/strong\u003e 10323.\u003c/li\u003e\n\u003cli\u003eHenderson JD, Payne BP, McIntyre NS, Biesinger MC. Enhancing Oxygen Spectra Interpretation by Calculating Oxygen Linked to Adventitious Carbon. \u003cem\u003eSurf Interfaces Analysis\u003c/em\u003e 2025, \u003cstrong\u003e57\u003c/strong\u003e(3)\u003cstrong\u003e:\u003c/strong\u003e 214-220.\u003c/li\u003e\n\u003cli\u003eDavid DGF, Pinault-Thaury M-A, Ballutaud D, Godet C. Sensitivity of photoelectron energy loss spectroscopy to surface reconstruction of microcrystalline diamond films. \u003cem\u003eApplied Surface Science\u003c/em\u003e 2013, \u003cstrong\u003e273:\u003c/strong\u003e 607-612.\u003c/li\u003e\n\u003cli\u003eCasalongue HS, Kaya S, Viswanathan V, Miller DJ, Friebel D, Hansen HA\u003cem\u003e, et al.\u003c/em\u003e Direct observation of the oxygenated species during oxygen reduction on a platinum fuel cell cathode. \u003cem\u003eNature Communications\u003c/em\u003e 2013, \u003cstrong\u003e4\u003c/strong\u003e(1).\u003c/li\u003e\n\u003cli\u003eRocha TCR, H\u0026auml;vecker M, Knop-Gericke A, Schl\u0026ouml;gl R. Promoters in heterogeneous catalysis: The role of Cl on ethylene epoxidation over Ag. \u003cem\u003eJ Catal\u003c/em\u003e 2014, \u003cstrong\u003e312:\u003c/strong\u003e 12-16.\u003c/li\u003e\n\u003cli\u003eSoares EA, Nascimento VB, Carvalho VEd, Castilho CMCd, Carvalho AVd, Toomes R\u003cem\u003e, et al.\u003c/em\u003e Structure determination of Ag(111) by low-energy electron diffraction. \u003cem\u003eSurf Sci\u003c/em\u003e 1999, \u003cstrong\u003e419 \u003c/strong\u003e89-96.\u003c/li\u003e\n\u003cli\u003eAndryushechkin BV, Shevlyuga VM, Pavlova TV, Zhidomirov GM, Eltsov KN. Structural transformations on an oxidized Ag(111) surface. \u003cem\u003eJETP Letters\u003c/em\u003e 2017, \u003cstrong\u003e105\u003c/strong\u003e(5)\u003cstrong\u003e:\u003c/strong\u003e 292-296.\u003c/li\u003e\n\u003cli\u003eTurano ME, Juurlink LBF, Gillum MZ, Jamka EA, Hildebrandt G, Lewis F\u003cem\u003e, et al.\u003c/em\u003e Oxygen-induced surface reconstructions on curved Ag(111). \u003cem\u003eJ Vac Sci Technol\u003c/em\u003e 2021, \u003cstrong\u003e39\u003c/strong\u003e(5).\u003c/li\u003e\n\u003cli\u003eDerouin J, Farber RG, Heslop SL, Killelea DR. Formation of surface oxides and Ag\u003csub\u003e2\u003c/sub\u003eO thin films with atomic oxygen on Ag(111). \u003cem\u003eSurf Sci\u003c/em\u003e 2015, \u003cstrong\u003e641:\u003c/strong\u003e L1-L4.\u003c/li\u003e\n\u003cli\u003eMartin NM, Klacar S, Gr\u0026ouml;nbeck H, Knudsen J, Schnadt J, Blomberg S\u003cem\u003e, et al.\u003c/em\u003e High-Coverage Oxygen-Induced Surface Structures on Ag(111). \u003cem\u003eJ Phys Chem C\u003c/em\u003e 2014, \u003cstrong\u003e118\u003c/strong\u003e(28)\u003cstrong\u003e:\u003c/strong\u003e 15324-15331.\u003c/li\u003e\n\u003cli\u003eHus M, Grilc M, Terzan J, Gyergyek S, Likozar B, Hellman A. Going Beyond Silver in Ethylene Epoxidation with First-Principles Catalyst Screening. \u003cem\u003eAngew Chem Int Ed\u003c/em\u003e 2023, \u003cstrong\u003e62\u003c/strong\u003e(31)\u003cstrong\u003e:\u003c/strong\u003e e202305804.\u003c/li\u003e\n\u003cli\u003eSantos VP, Plauck A, Gold J, Majumdar P, McAdon MH, Calverley T. The Complex Chlorination Effects on High Selectivity Industrial EO Catalysts: Dynamic Interplay between Catalyst Composition and Process Conditions. \u003cem\u003eACS Catal\u003c/em\u003e 2024, \u003cstrong\u003e14\u003c/strong\u003e(14)\u003cstrong\u003e:\u003c/strong\u003e 10839-10852.\u003c/li\u003e\n\u003cli\u003eHoflund GB, Weaver JF, Epling WS. Ag\u003csub\u003e2\u003c/sub\u003eO XPS Spectra. \u003cem\u003eSurf Sci Spectra\u003c/em\u003e 1994, \u003cstrong\u003e3\u003c/strong\u003e(2)\u003cstrong\u003e:\u003c/strong\u003e 157-162.\u003c/li\u003e\n\u003cli\u003eHoflund GB, Weaver JF, Epling WS. AgO XPS Spectra. \u003cem\u003eSurf Sci Spectra\u003c/em\u003e 1994, \u003cstrong\u003e3\u003c/strong\u003e(2)\u003cstrong\u003e:\u003c/strong\u003e 163-168.\u003c/li\u003e\n\u003cli\u003eIdriss H. On the wrong assignment of the XPS O1s signal at 531\u0026ndash;532 eV attributed to oxygen vacancies in photo- and electro-catalysts for water splitting and other materials applications. \u003cem\u003eSurf Sci\u003c/em\u003e 2021, \u003cstrong\u003e712\u003c/strong\u003e.\u003c/li\u003e\n\u003cli\u003eGuo LQ, Qin SX, Yang BJ, Liang D, Qiao LJ. Effect of hydrogen on semiconductive properties of passive film on ferrite and austenite phases in a duplex stainless steel. \u003cem\u003eSci Rep\u003c/em\u003e 2017, \u003cstrong\u003e7\u003c/strong\u003e(1)\u003cstrong\u003e:\u003c/strong\u003e 3317.\u003c/li\u003e\n\u003cli\u003eZhu M, Chen P, Liu M. Sunlight-driven plasmonic photocatalysts based on Ag/AgCl nanostructures synthesized via an oil-in-water medium: enhanced catalytic performance by morphology selection. \u003cem\u003eJournal of Materials Chemistry\u003c/em\u003e 2011, \u003cstrong\u003e21\u003c/strong\u003e(41).\u003c/li\u003e\n\u003cli\u003eFu S, Chu Z, Huang Z, Dong X, Bie J, Yang Z\u003cem\u003e, et al.\u003c/em\u003e Construction of Z-scheme AgCl/BiOCl heterojunction with oxygen vacancies for improved pollutant degradation and bacterial inactivation. \u003cem\u003eRSC Adv\u003c/em\u003e 2024, \u003cstrong\u003e14\u003c/strong\u003e(6)\u003cstrong\u003e:\u003c/strong\u003e 3888-3899.\u003c/li\u003e\n\u003cli\u003eYu X, Zhou C, Huang Z, Xin C, Lin Y, Fu F\u003cem\u003e, et al.\u003c/em\u003e Rational design of AgCl@Zr\u003csup\u003e3+\u003c/sup\u003e-ZrO\u003csub\u003e2\u003c/sub\u003e nanostructures for ultra-efficient visible-light photodegradation of emerging pollutants. \u003cem\u003eAppl Catal B\u003c/em\u003e 2023, \u003cstrong\u003e325\u003c/strong\u003e.\u003c/li\u003e\n\u003cli\u003eChen D, Chen L, Zhao Q-C, Yang Z-X, Shang C, Liu Z-P. Square-pyramidal subsurface oxygen [Ag\u003csub\u003e4\u003c/sub\u003eOAg] drives selective ethene epoxidation on silver. \u003cem\u003eNat Catal\u003c/em\u003e 2024.\u003c/li\u003e\n\u003cli\u003eEsposito J, Bhan A. Isotopic Studies of Reaction Pathways within Propylene Epoxidation over Promoted Silver Catalysts. \u003cem\u003eACS Catal\u003c/em\u003e 2024\u003cstrong\u003e:\u003c/strong\u003e 14339-14349.\u003c/li\u003e\n\u003cli\u003evan Hoof AJF, van der Poll RCJ, Friedrich H, Hensen EJM. Dynamics of silver particles during ethylene epoxidation. \u003cem\u003eAppl Catal B\u003c/em\u003e 2020, \u003cstrong\u003e272\u003c/strong\u003e.\u003c/li\u003e\n\u003cli\u003eT. D. Kuhne, M. Iannuzzi, M. Del Ben, V. V. Rybkin, P. Seewald, F. Stein, T. Laino, R. Z. Khaliullin, O. Schutt, F. Schiffmann, D. Golze, J. Wilhelm, S. Chulkov, M. H. Bani-Hashemian, V. Weber, U. Borstnik, M. Taillefumier, A. S. Jakobovits, A. Lazzaro, H. Pabst, T. Muller, R. Schade, M. Guidon, S. Andermatt, N. Holmberg, G. K. Schenter, A. Hehn, A. Bussy, F. Belleflamme, G. Tabacchi, A. Gloss, M. Lass, I. Bethune, C. J. Mundy, C. Plessl, M. Watkins, J. VandeVondele, M. Krack, J. Hutter, CP2K: An electronic structure and molecular dynamics software package - Quickstep: Efficient and accurate electronic structure calculations\u003cem\u003eJ. Chem. Phys. \u003c/em\u003e2020, \u003cstrong\u003e152\u003c/strong\u003e, 194103. \u003c/li\u003e\n\u003cli\u003eS. Goedecker, M. Teter, J. Hutter, Separable dual-space Gaussian pseudopotentials. \u003cem\u003ePhys. Rev. B \u003c/em\u003e1996, \u003cstrong\u003e54\u003c/strong\u003e, 1703-1710.\u003c/li\u003e\n\u003cli\u003eJ. VandeVondele, J. Hutter, Gaussian basis sets for accurate calculations on molecular systems in gas and condensed phases. J. Chem. Phys. 2007, 127, 114105.\u003c/li\u003e\n\u003cli\u003eL. Bengtsson, Dipole correction for surface supercell calculations. \u003cem\u003ePhys. Rev. B \u003c/em\u003e1999, \u003cstrong\u003e59\u003c/strong\u003e, 12301-12304.\u003c/li\u003e\n\u003cli\u003eJ. P. Perdew, K. Burke, M. Ernzerhof, Generalized Gradient Approximation Made Simple. \u003cem\u003ePhys. Rev. Lett. \u003c/em\u003e1996, \u003cstrong\u003e77\u003c/strong\u003e, 3865-3868.\u003c/li\u003e\n\u003cli\u003eS. Grimme, Semiempirical GGA-type density functional constructed with a long-range dispersion correction. \u003cem\u003eJ. Comput. Chem. \u003c/em\u003e2006, \u003cstrong\u003e27\u003c/strong\u003e, 1787-1799.\u003c/li\u003e\n\u003cli\u003eRichard H. Byrd, P. L., Jorge Nocedal, and Ciyou Zhu. A Limited Memory Algorithm for Bound Constrained Optimization. Phys. Rev. Lett. 1995, \u003cstrong\u003e16\u003c/strong\u003e (5), 1190-1208.\u003c/li\u003e\n\u003cli\u003eGraeme Henkelman, Blas P. Uberuaga, Hannes J\u0026oacute;nsson, A climbing image nudged elastic band method for finding saddle points and minimum energy paths. J. Chem. Phys. 2000; \u003cstrong\u003e113\u003c/strong\u003e (22): 9901\u0026ndash;9904.\u003c/li\u003e\n\u003cli\u003eM. Iannuzzi, J. Hutter, Inner-shell spectroscopy by the Gaussian and augmented plane wave method. \u003cem\u003ePhys. Chem. Chem. Phys. \u003c/em\u003e2007, \u003cstrong\u003e9\u003c/strong\u003e, 1599-1610.\u003c/li\u003e\n\u003cli\u003eJ. C. Slater, K. H. Johnson, Self-Consistent-Field Xn Cluster Method for Polyatomic Molecules and Solids. \u003cem\u003ePhys. Rev. B \u003c/em\u003e1972, \u003cstrong\u003e5\u003c/strong\u003e, 844-853.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-7141633/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7141633/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Ethylene epoxidation is a cornerstone process in the chemical industry and the primary route for ethylene oxide (EO) production. Despite its significance, the reaction mechanism and the structure of the active surface remain under debate. Here, we employ in situ photoelectron spectroscopy supported by DFT to probe surface intermediates on Ag/Al2O3 under ethylene epoxidation conditions, with and without vinyl chloride promotion. Under these conditions, combustion is triggered by ethylene dehydrogenation at surface defects, assisted by lattice oxygen and yielding CO2. Vinyl chloride suppresses this pathway by reducing defect density, hindering lattice oxygen formation, and inhibiting ethylene adsorption. Chlorine binds preferentially at defect sites, passivating undercoordinated silver atoms to disfavor C–H activation. Ethylene reacts with one of the oxygen atoms of adsorbed dioxygen on reconstructed silver surfaces through a direct oxidation pathway to produce EO. These findings provide molecular-level insights and offer a rational basis for optimizing ethylene epoxidation catalysts.","manuscriptTitle":"Reaction Mechanism and Role of Chlorine in Ethylene Epoxidation Revealed by in situ XPS","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-08-12 14:52:33","doi":"10.21203/rs.3.rs-7141633/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"3e6bb116-a9d8-4632-8661-83cf55837828","owner":[],"postedDate":"August 12th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":53001642,"name":"Physical sciences/Chemistry/Catalysis/Heterogeneous catalysis"},{"id":53001643,"name":"Physical sciences/Chemistry/Surface chemistry/Surface spectroscopy"},{"id":53001644,"name":"Physical sciences/Materials science/Theory and computation/Electronic structure"}],"tags":[],"updatedAt":"2025-09-16T12:25:45+00:00","versionOfRecord":[],"versionCreatedAt":"2025-08-12 14:52:33","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7141633","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7141633","identity":"rs-7141633","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}