ARTICLE https://doi.org/10.1038/s41467-020-18721-0 OPEN Exploiting two-dimensional morphology of molybdenum oxycarbide to enable efficient catalytic dry reforming of methane Alexey Kurlov 1, Evgeniya B. Deeva 1, Paula M. Abdala 1, Dmitry Lebedev 2,4, Athanasia Tsoukalou 1, Aleix Comas-Vives 3 ✉, Alexey Fedorov 1 ✉ & Christoph R. Müller 1 ✉ 1234567890():,; The two-dimensional morphology of molybdenum oxycarbide (2D-Mo2COx) nanosheets dispersed on silica is found vital for imparting high stability and catalytic activity in the dry reforming of methane. Here we report that owing to the maximized metal utilization, the specific activity of 2D-Mo2COx/SiO2 exceeds that of other Mo2C catalysts by ca. 3 orders of magnitude. 2D-Mo2COx is activated by CO2, yielding a surface oxygen coverage that is optimal for its catalytic performance and a Mo oxidation state of ca. +4. According to ab initio calculations, the DRM proceeds on Mo sites of the oxycarbide nanosheet with an oxygen coverage of 0.67 monolayer. Methane activation is the rate-limiting step, while the activation of CO2 and the C–O coupling to form CO are low energy steps. The deactivation of 2D-Mo2COx/SiO2 under DRM conditions can be avoided by tuning the contact time, thereby preventing unfavourable oxygen surface coverages. 1 Department of Mechanical and Process Engineering, ETH Zürich, Leonhardstrasse 21, CH 8092 Zürich, Switzerland. 2 Department of Chemistry and Applied Biosciences, ETH Zürich, Vladimir-Prelog-Weg 1-5, CH 8093 Zürich, Switzerland. 3 Department of Chemistry, Universitat Autònoma de Barcelona, 08193 Cerdanyola del Vallès, Catalonia, Spain. 4Present address: Department of Materials Science and Engineering, Northwestern University, Evanston, IL 60208, USA. ✉email: Aleix.Comas@uab.cat; fedoroal@ethz.ch; muelchri@ethz.ch NATURE COMMUNICATIONS | (2020)11:4920 | https://doi.org/10.1038/s41467-020-18721-0 | www.nature.com/naturecommunications 1 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-18721-0 S ince the early work by Sinfelt1 and Boudart2 demonstrated can be controllably removed (de-functionalized) by reduction in that Mo2C and WC exhibit platinum-like catalytic activity H2, giving 2D-Mo2C/SiO2. The oxidation of 2D-Mo2C/SiO2 by in alkane hydrogenolysis and isomerization, a multitude of CO2 at 800 °C gives a 2D molybdenum oxycarbide 2D-Mo2COx/ reports has appeared aiming to exploit molybdenum carbide for SiO2 with a high oxygen surface coverage and an average Mo industrially-relevant reactions such as methane aromatization3,4, oxidation state of +5.5. Subjecting this material to DRM condi- the water-gas shift5,6, and CO2 hydrogenation reactions7,8, to tions reduces the oxygen coverage and forms a catalyst in situ name just a few9. In contrast to middle-to-late transition metal- with an average Mo oxidation state of +4. This material is highly based catalysts that deactivate by the deposition of coke, sintering active in DRM, in contrast to fully reduced 2D-Mo2C/SiO2 and or poisoning10, carbide catalysts typically feature a low rate of 2D-Mo2COx/SiO2 material with highest oxygen coverage (con- coking11, are resistant to high temperature sintering as well as to firmed by a Mo oxidation state of ca. +5) which are essentially sulfur poisoning12,13. However, the high oxophilicity of carbides inactive for the DRM. Our Mo K-edge X-ray absorption near fosters their evolution in reactions with oxygenate substrates, edge structure (XANES) and X-ray photoelectron spectroscopy leading ultimately to deactivation via the formation of oxide (XPS) results verify that molybdenum oxycarbide is the active phases; deactivated catalysts require regeneration by re- phase in Mo2C-based DRM catalysts31 and density functional carburization with CH4/H211. If such deactivation pathways are theory (DFT) calculations identify that an oxygen coverage of ca. avoided while the metal utilization of carbides is maximized, for 0.7 monolayer provides a high activity of 2D-Mo2COx/SiO2 in instance by developing catalysts with a high surface-to-bulk ratio DRM. Reducing or increasing this optimal oxygen coverage and a high dispersion on supports, the replacement of traditional, decreases the reaction rate. 2D-Mo2COx/SiO2 catalyst with more expensive middle-to-late transition metal catalysts by reduced oxygen coverage is also less selective owing to a com- carbide-based catalysts will become viable. peting reverse water-gas shift reaction. Hence, supported 2D- The dry reforming of methane, (DRM, CH4 + CO2 ↔ 2CO + Mo2COx catalyst with an initially high coverage of surface oxygen 2H2) is an example of a particularly challenging reaction for atoms activates first by reduction, that is the oxygen coverage is Mo2C-based catalysts because it combines a CO2-rich feed with lowered until a maximum in DRM performance is reached at a high operating temperatures (typically, 800 °C and above)11,14–21. Mo oxidation state of ca. +4. Reducing the Mo oxidation state Deactivation of Mo2C in DRM conditions proceeds oxidatively further (i.e., lowering the oxygen coverage below ca. 0.7 mono- according to Mo2C + 5 CO2 → 2 MoO2 + 6 CO11,17,18 and in layers) leads to catalyst deactivation. This observation is in sharp order to mitigate it, dry reforming of methane is often conducted contrast to non-layered Mo2C-based catalysts that readily deac- at elevated pressures (2–10 bar)11,14,16,18,21. From a mechanistic tivate by oxidation in a CO2 atmosphere. It is therefore standpoint, an ill-defined molybdenum oxycarbide phase remarkable that deactivated 2D-Mo2COx/SiO2 catalysts do not (MoOxCy) has been reported to play a key role in DRM22, as well contain any deposited coke according to temperature- as in the water-gas shift23, deoxygenation24,25, and CO2-to- programmed oxidation (TPO) experiments and can be regener- methanol8 conversion processes, suggesting that the partial oxi- ated by oxidation in CO2, fully recovering their highest catalytic dation of Mo2C is favorable for those reactions. In addition, activity. Noteworthy, the deactivation of 2D-Mo2COx/SiO2 under recent reports on CO2-propane dehydrogenation catalyzed by DRM conditions can be avoided by optimizing the contact time. Mo2C highlighted that control over the oxygen coverage of the Density functional theory calculations corroborate experimental surface oxycarbide layer allows tuning the catalytic activity and observations and demonstrate that the DRM pathway on an selectivity26,27. Molybdenum oxycarbide species were also pro- oxycarbidic surface with a submonolayer oxygen coverage is posed as intermediates during the dehydroaromatization of energetically favored compared to the pathway occurring on a methane on molybdenum-loaded zeolite ZSM-528. However, fully carbidic surface. The fully oxygen-covered surface is establishing the local structure and the nature of the active sites in unstable in DRM conditions and reduces in situ its oxygen cov- such molybdenum oxycarbide phases has remained challenging, erage below 1 oxygen monolayer, also consistent with our largely because well-defined oxycarbide catalysts, serving as experiments. references, are not available. In this context, a recently discovered family of two- dimensional atomically-thin early transition metal carbides, Results nitrides, and carbonitrides called MXenes29 could be exploited as Synthesis and characterization. Multi-layered Mo2CTx32 (refer- an entry point to supported well-defined 2D carbides and oxy- red to as m-Mo2CTx) was sonicated in ethanol to yield, after carbides. MXene films are nanocrystalline yet form stable col- centrifugation, a transparent purple colloidal solution of dela- loidal solutions in protic solvents, and feature metal-terminated minated 2D-Mo2CTx flakes (Fig. 1a and Supplementary Fig. 2)33. surfaces (vide infra) with a lateral size in the submicron scale, A dried aliquot of this solution was analyzed by transmission which yields highly uniform surface metal sites30. The atomically electron microscopy (TEM) revealing single and a few-layer thin thin nature of MXenes enables optimal metal utilization provided flakes of Mo2CTx, the latter morphology is identified by the that these films are highly dispersed on a support. The 2D characteristic scrolling of edges of the few-layer thin MXene morphology is also advantageous for mechanistic studies since nanosheets (Supplementary Fig. 3)34. The X-ray powder diffrac- bulk metal sites in catalysts derived from MXenes, are minimized. tion pattern of the dried delaminated material is typical of a Therefore, the characterization methods applied (even when not layered Mo2CTx structure (Supplementary Fig. 4)35. The colloidal being inherently surface sensitive) will provide information solution of the Mo2CTx nanoflakes (ca. 0.15 mg mL−1 con- mostly about the surface sites, which is instructive for structure- centration determined by thermogravimetric analysis) was used activity studies. for several consecutive incipient wetness impregnations of a SiO2 In what follows, we describe an approach to yield a supported support (Aerosil 300, calcined at 950 °C, 194 m2 g−1 surface area two-dimensional (2D) model molybdenum oxycarbide material according to nitrogen physisorption), giving 2D-Mo2CTx/SiO2 containing well-defined surface sites that are highly active in the that, after drying at 100 °C in air, contained 0.48 wt% Mo by dry reforming of methane. We prepare this catalyst from colloi- elemental analysis. TEM imaging of 2D-Mo2CTx/SiO2 shows dal, delaminated thin films of the molybdenum carbide MXene agglomerated amorphous silica grains that are homogeneously phase, Mo2CTx, where T stands for OH, O, F surface termination covered with thin Mo2CTx nanosheets (Fig. 1b and Supplemen- groups. After supporting Mo2CTx on SiO2, these surface groups tary Fig. 5). Consistent with an MXene structure, hexagonally 2 NATURE COMMUNICATIONS | (2020)11:4920 | https://doi.org/10.1038/s41467-020-18721-0 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-18721-0 ARTICLE a m-Mo2CTx 2D-Mo2CTx 2D-Mo2CTx /SiO2 2D-Mo2C/SiO2 Impregnation H2, 800 °C onto SiO2 − [Tx ] Aerosil 300 CO2 Sonication d Mo6+ 800 °C EtOH Mo5+ Mo4+ MoCarbidic E-Mo2C 2D-Mo2COx /SiO2 Mo Mo0 C 0.5 eV T = O, OH, F 2D-Mo2COx /SiO2 Intensity [a.u.] b c 1.2 e 1.2 2D-Mo2C/SiO2 2D-Mo2C/SiO2 2D-Mo2CTx /SiO2 2D-Mo2COx /SiO2 0.9 2D-Mo2C/SiO2 2D-Mo2CTx /SiO2 m-Mo2CTx 0.9 750 E-Mo2C Normalized  Normalized  0.5 600 6 2D-Mo2CTx/SiO2 MoO3 0.6 2D-Mo2COx/SiO2 Oxidation state 10 nm 2D-Mo2CTx /SiO2 0.6 450 T [C°] m- Mo2CTx Normalized 4 MoO2 300 °C 2 0.3 0.3 E -Mo2C 750 RT 150 0 2D-Mo2C/SiO2 0.1 19,990 20,000 20,010 20,020 MoO3 Mo Edge position [eV] Energy [eV] 20,000 20,005 20,010 20,015 0.0 0.0 19,980 20,000 20,020 20,040 20,060 20,080 240 235 230 225 220 19,980 20,000 20,020 20,040 20,060 20,080 1 Pm Energy [eV] Binding energy [eV] Energy [eV] Fig. 1 Synthesis and characterization of 2D-Mo2COx/SiO2. a Synthesis of 2D-Mo2COx/SiO2 from delaminated Mo2CTx films via incipient wetness impregnation of their colloidal solution onto silica, followed by the reduction of the Tx groups under H2 (giving 2D-Mo2C/SiO2) and the subsequent oxidation in CO2 (giving 2D-Mo2COx/SiO2); b TEM images of 2D-Mo2CTx/SiO2; c in situ reduction of 2D-Mo2CTx/SiO2 under H2 as followed by Mo K-edge XANES revealing an evolution of the pre-edge feature with temperature (shown in the inset as a contour plot); d oxidation state of Mo in reference and synthesized materials determined from the deconvolution of ex situ XPS data of the Mo 3d core levels as well as from e the edge position in XANES spectra. Mo−O bonds where O belongs to a hydroxy surface termination Table 1 EXAFS fitting of 2D-Mo2CTx/SiO2 (R factor 0.008). site and C is the interlayer carbidic carbon, consistent with the 2.1 Å Mo−C distance in the parent m-Mo2CTx32. The second Mo Shell CN R (Å) σ2 (Å2) ΔE (eV) coordination sphere is fitted with three Mo–Mo scattering paths with interatomic distances of 2.63(1), 2.92(2), and 3.42(2) Å and Mo−O1 1.0 (2) 1.74 (1) 0.005 (2) 6 (2) Mo−C/O2 1.2 (3) 2.03 (1) with coordination numbers of 0.8(4), 1.4(5), and 1.1(4), Mo−Mo1 0.8 (4) 2.63 (1) 0.006 (2) respectively (Table 1, Supplementary Fig. 12). The presence of Mo−Mo2 1.4 (5) 2.92 (2) three different Mo–Mo paths indicates a strong distortion of Mo Mo−Mo3 1.1 (4) 3.42 (2) sites in silica-supported Mo2CTx, deviating notably from the Mo coordination environment in parent m-Mo2CTx that features only two Mo–Mo paths at 2.9 Å and 3.2 Å with coordination ordered Mo atoms are observed by (high resolution) HR-TEM in numbers of 4 and 2, respectively (as determined by EXAFS)32. 2D-Mo2CTx/SiO2 (Supplementary Fig. 5). The X-ray powder This deviation is likely owing to (i) the higher fraction of Mo=O diffraction (XRD) pattern of 2D-Mo2CTx/SiO2 reveals only an sites in 2D-Mo2CTx/SiO2 relative to the m-Mo2CTx, (ii) a amorphous halo corresponding to SiO2 (Supplementary Fig. 6). disorder of 2D-Mo2CTx/SiO2 caused by the geometrical curvature The absence of crystalline peaks of m-Mo2CTx confirms a high of the delaminated nanosheets supported on the silica grains and dispersion of individual delaminated Mo2CTx nanosheets on the (iii) the interaction with the support. All these factors could support. contribute to the lower total CNs of Mo in 2D-Mo2CTx/SiO2 as Next, we fitted the experimentally acquired Fourier- determined by EXAFS. transformed Mo K-edge extended X-ray absorption fine structure Subsequently, we explored the transformation of these silica- (EXAFS) functions of 2D-Mo2CTx/SiO2 with two coordination supported 2D-Mo2CTx sheets in reductive, oxidative, and DRM spheres, i.e., Mo−C/O/OH and Mo−Mo. The fitting of the first conditions. Note that, unless specified otherwise, all these treated coordination sphere of Mo gives two scattering paths with materials were characterized by avoiding exposure to air, using interatomic distances of 1.74(1) and 2.03(1) Å and coordination in situ experimentation, gloveboxes and air-tight transfers for numbers (CN) of 1.0(2) and 1.2(3), respectively (Table 1, characterization (XPS, fourier-transform infrared (FTIR) spectro- Supplementary Fig. 12). The shorter scattering path is attributed scopy, TEM and XANES). Reductive de-functionalization of 2D- to molybdenum oxo (Mo=O) surface sites since the fitted Mo2CTx/SiO2 from the passivating Tx surface groups was first distance 1.74(1) Å is in good agreement with the reported Mo=O studied in H2 flow by in situ XANES (Mo K-edge, Fig. 1c). bond lengths of ca. 1.7 Å found in a molybdenum oxycarbide in a Interestingly, in contrast to the bulk Mo2CTx32, a well-defined zeolite structure31 as well as in molecular Mo complexes36,37 pre-edge feature, typical for molybdenum in a non- (determined by EXAFS and X-Ray crystallography, respectively). centrosymmetric environment38,39, appears in the room tem- The second scattering path of the first coordination shell with an perature XANES spectrum of 2D-Mo2CTx/SiO2 (Supplementary interatomic distance of 2.03(1) Å is attributed to the Mo−C and Fig. 13). Under H2 flow and with increasing temperature (in situ NATURE COMMUNICATIONS | (2020)11:4920 | https://doi.org/10.1038/s41467-020-18721-0 | www.nature.com/naturecommunications 3 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-18721-0 XANES experiment), the Mo K-edge in 2D-Mo2CTx/SiO2 shifts bridges of silica42. Therefore Mo4+ sites in 2D-Mo2C/SiO2 are to lower energies and this process continues up to ca. 450 °C likely formed due to the grafting43 of the reduced nanosheets (from 20015.8 to 20002.0 eV), indicating a gradual reduction of onto the silica surface. Mo due to the de-functionalization of the Tx groups (Fig. 1c). Interestingly, the fitting of the 3d XPS spectrum of Mo in 2D- Simultaneously, the intensity of the characteristic pre-edge feature Mo2COx/SiO2 requires 4 components, namely Mo6+, Mo5+, of 2D-Mo2CTx/SiO2 reduces until its disappearance at ca. 450 °C. Mo4+ and metallic Mo0 (Supplementary Table 2). Note that Mo0 Increasing the temperature further to 750 °C induces only a is not observed in the XPS spectra of 2D-Mo2C/SiO2 and 2D- minor shift of the Mo K-edge position compared to 450 °C, that is Mo2CTx/SiO2. We rationalize this result by a partial removal of from 20002.0 to 20000.8 eV, the latter value is very close to the the carbidic carbon from the 2D-Mo2C layer by CO2 at 800 °C edge position of β-Mo2C (20000.7 eV). This suggests that H2 (forming 2 equiv of CO) and attribute the presence of the Mo6+ treatment at 750 °C leads to a Mo terminated carbidic surface component in the fit to the oxygen-rich oxycarbide phase of 2D- predominantly free from oxo, hydroxy and fluoro terminations, Mo2COx/SiO2. In contrast, heating multi-layer m-Mo2CTx in a supporting the assignment of the aforementioned pre-edge CO2 flow transforms it into a mixture of β-Mo2C and MoO2 feature to molybdenum sites in 2D-Mo2CTx/SiO2 as opposed to already at ca. 550 °C, according to our in situ XRD experiment MoO2 or MoO3 (conceivable oxidation products of 2D-Mo2CTx). (Supplementary Fig. 16). This result underlines again that Noteworthy, the temperature for the de-functionalization of most dispersion and isolation of MXene nanosheets on a support is a of the Tx groups in 2D-Mo2CTx/SiO2 (according to XANES) is critical step, prior to high temperature treatments in both approximately 450 °C, which is 150 °C lower than that of m- reducing and oxidizing conditions, to retain the nanosheet Mo2CTx32, probably owing to the high dispersion of the 2D- morphology of the (oxy)carbide phase. In a sharp contrast to Mo2CTx sheets on the silica surface that eliminates gas diffusion the stability of 2D-Mo2COx/SiO2 in an atmosphere of pure CO2 limitations in the pores of m-Mo2CTx. at 800 °C, other supported Mo2C-based catalysts yield MoO2 Since the DRM reaction typically requires high temperatures to already in CO2-rich DRM feeds, which is the main deactivation yield high equilibrium conversions (>700 °C)40, 2D-Mo2CTx/SiO2 pathway of these catalysts11,19. was treated in a H2 flow at 800 °C for 1 h to obtain 2D-Mo2C/ Next, ex situ Mo K-edge XANES spectra of 2D-Mo2CTx/SiO2, SiO2. This high temperature, reductive treatment did not affect 2D-Mo2C/SiO2 and 2D-Mo2COx/SiO2 were compared (Fig. 1e). significantly the original morphology and high dispersion of the Note that the edge position is defined as the first inflection point silica-supported nanosheets, as confirmed by TEM imaging of the XANES spectra after the pre-edge feature (where the pre- (Supplementary Fig. 8). That said, in few cases small nanopar- edge corresponds to the forbidden 1s-4d transition)39. In ticles (ca. 2–3 nm) decorating molybdenum carbide sheets could agreement with XPS results, 2D-Mo2CTx/SiO2 features an edge be discerned in TEM images (Supplementary Fig. 8) possibly due position at 20015.8 eV that is notably higher than in m-Mo2CTx to the partial sintering of 2D-Mo2C sheets to Mo2C nanoparticles (20010.9 eV) indicating more oxidized Mo sites in supported at areas where nanosheets overlap. Overall, 2D-Mo2CTx/SiO2 Mo2CTx nanosheets32. The observed Mo edge energy is consistent behaves very differently in reductive de-functionalization condi- with an average oxidation state of 2D-Mo2CTx/SiO2 of ca. +5.5 tions when compared to m-Mo2CTx that sinters under these (Fig. 1e, inset and Supplementary Table 1). The reduction of 2D- conditions forming a porous bulk β-Mo2C phase32. Similarly, a Mo2CTx/SiO2 to 2D-Mo2C/SiO2 in H2 leads to the disappearance multi-layer m-V2CTx was reported to sinter around 600 °C, which of the pre-edge feature, as was also observed in the in situ XANES appears to be a general temperature limitation of multi-layer reduction experiment, and shifts the Mo edge positions MXenes41. These results underline the importance of the significantly to lower energies, i.e., from 20015.8 to 20000.8 eV. dispersion of the nanosheets on a support to avoid sintering. While the Mo K-edge position confirms the carbidic nature of Next, we prepared silica-supported model 2D molybdenum 2D-Mo2C/SiO2, the white line region of the XANES spectrum of oxycarbide 2D-Mo2COx/SiO2 by treating 2D-Mo2C/SiO2 in a 2D-Mo2C/SiO2 differs from the spectrum of bulk β-Mo2C flow of CO2 (800 °C, 1 h, Fig. 1a). According to TEM, the CO2 (Fig. 1e) owing to the different morphologies of these two treatment of 2D-Mo2C/SiO2 does not alter the morphology or materials. This observation corroborates the results of TEM dispersion of the supported nanosheets (Supplementary Fig. 9). analysis and confirms further that sintering to the bulk carbide A comparison of the Mo 3d XPS spectra of the prepared silica- phase did not occur during the reductive treatment of 2D- supported materials and β-Mo2C and MoO3 references is given in Mo2CTx/SiO2. Consistent with XPS analysis, the XANES Fig. 1d. The Mo 3d spectrum of 2D-Mo2CTx/SiO2 was fitted with spectrum of 2D-Mo2COx/SiO2 reveals a significant shift of the Mo4+, Mo5+, and Mo6+ components with the latter two being Mo K-edge towards higher energies compared to 2D-Mo2C/SiO2 the dominating oxidation states (Supplementary Table 2). In (i.e., 20015.7 vs 20000.7 eV), suggesting an average Mo oxidation contrast, the Mo 3d spectrum of m-Mo2CTx can be fitted using state for 2D-Mo2COx/SiO2 between Mo5+ and Mo6+ (Fig. 1e only Mo4+ and Mo5+ components (Supplementary Table 2)32 inset, Supplementary Table 1). The XANES spectrum of 2D- indicating that Mo sites are more oxidized in the silica-supported Mo2COx/SiO2 displays a characteristic pre-edge feature, similarly delaminated Mo2CTx nanosheets than in m-Mo2CTx. In contrast, to 2D-Mo2CTx/SiO2. We conclude that all available characteriza- the Mo 3d XPS spectrum of 2D-Mo2C/SiO2 is described by tion data (XANES, XPS, TEM) are consistent with a predomi- mostly carbidic Mo sites in addition to a smaller fraction of Mo4+ nantly nanosheet morphology and a carbidic and oxycarbidic and Mo6+ sites (Supplementary Table 2). To understand better nature of, respectively, 2D-Mo2C/SiO2 and 2D-Mo2COx/SiO2. the nature of the Mo4+ and Mo6+ sites, observed by XPS of the reduced materials, 2D-Mo2C/SiO2 was analyzed by transmission IR spectroscopy revealing a low intensity band at ca. 2290 cm–1 Correlating catalytic activity to the surface oxygen coverage that is tentatively assigned to surface [≡Si–H] sites (Supplemen- of 2D-Mo2COx. The catalytic activities of 2D-Mo2C/SiO2 and tary Fig. 15). We speculate that the high temperature H2 2D-Mo2COx/SiO2 for the dry reforming of methane (800 °C, 1 treatment of 2D-Mo2CTx/SiO2 may give molybdenum hydrides bar, space velocity (SV) 1200 L gMo−1 h−1, (weight/volume flow (carbidic or oxycarbidic) that further react by opening the rate) W/F = 3 ms gMo mL−1 contact time, 1:1 CO2:CH4 ratio) are siloxane bridges of silica forming [Mo–Os] and [≡Si–H] bonds strikingly different (Fig. 2a). While 2D-Mo2C/SiO2 shows negli- (Supplementary Fig. 15, inset). Such reactivity has been observed gible DRM activity, 2D-Mo2COx/SiO2 is highly active with previously for surface hydrides of tantalum that cleaved siloxane methane conversion rates of 0.42 mol(CH4) mol(Mo)−1 s−1 after 4 NATURE COMMUNICATIONS | (2020)11:4920 | https://doi.org/10.1038/s41467-020-18721-0 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-18721-0 ARTICLE a 100 3.0 b 20,015 0.5 H2 DRM CO2 DRM CO2 CO2 2.5 80 CH4 20,012 0.4 Mo K-edge energy [eV] reff [molCH4 mol–1Mo s–1] 2.0 Conversion [%] 60 20,009 0.3 CO/H2 1.5 40 20,006 ca. +4 0.2 Mo average 1.0 ca. +3 oxidation state ca. +0.2 20 20,003 0.1 0.5 0 0.0 20,000 0.0 0 60 120 180 240 300 360 420 /SiO 2 S10 OS 100 /SiO 2 -T O M o 2C /SiO 2 -T Time [min] 2D- CO x CO x Mo 2 Mo 2 2D- 2D- c 1.2 d 3.0 60 CH4 1.0 2.5 CO2 50 H2 0.8 2.0 40 Conversion [%] Normalized  Fully oxidized state CO/H2 Reduced state 0.6 of Mo ca. +0.2 of Mo ca. +5.5 1.5 30 0.4 1.0 20 DRM conditions 0.2 (W/F = 0.1 ms g Mo mL−1) 0.5 –1 10 or pure CO2 W/F = 2 ms gMo mL W/F = 1 ms gMo mL–1 0.0 0.0 0 19,980 20,000 20,020 20,040 20,060 0 4 8 12 16 20 Energy [eV] Time [h] Fig. 2 Catalytic performance of 2D-Mo2COx/SiO2 in DRM. a Catalytic performance of 2D-Mo2C/SiO2 (90–150 min) and 2D-Mo2COx/SiO2 (240–300 min and from 390 min after the CO2 regeneration step) in DRM at a contact time (W/F) of 3 ms gMo mL−1; b correlation of the Mo oxidation state, determined by Mo K-edge XANES, and the catalytic activity of the material; c Mo K-edge XANES spectra of 2D-Mo2COx/SiO2 when exposed to DRM (0.1 ms gMo mL−1 contact time), reductive and oxidative conditions at 730 °C; d stable performance of 2D-Mo2COx/SiO2 at contact times of 1 or 2 ms gMo mL−1. 10 min on stream at 80% CH4 conversion (i.e., ca. 10% below the higher for 2D-Mo2CTx/SiO2 relative to the β-Mo2C catalyst. thermodynamic equilibrium in those conditions, Fig. 2a)40. Note These results provide strong evidence that the improved catalytic that the 2D nature of our catalysts enables a very high utilization activity is due to the different nature of Mo sites in 2D-nanosheet of Mo such that when normalized by weight of Mo, the initial catalyst. methane consumption rate of 2D-Mo2COx/SiO2 is ~10–200 times When using a CH4:CO2 feed ratio of 1:1 and a contact time of higher than the rates that have been reported for other Mo2C- 3 ms gcat L−1, 2D-Mo2COx/SiO2 continuously deactivates, with the based catalysts at similar conditions (Supplementary Table 3). methane consumption rate dropping within 100 min on stream by The reference catalysts β-Mo2C and m-Mo2CTx exhibit a sig- almost a factor of three to 0.15 mol(CH4) mol(Mo)−1 s−1. At this nificantly lower DRM activity in these conditions (ca. 0.0003 and point, the CH4 conversion has declined from ca. 80 to 30%. 0.002 mol(CH4) mol(Mo)−1 s−1 after ca. 5 min of time in stream, However, the initial activity can be fully regenerated by subjecting respectively), and deactivated entirely after 30 min of reaction the catalyst, after 100 min time on stream (TOS), to a flow of pure (Supplementary Fig. 17). The CO chemisorption capacity was CO2 at 800 °C for 1 h. While carbide catalysts are generally not determined for in situ reduced 2D-Mo2CTx/SiO2 and the refer- known for deactivation by coking10, we evaluated the possible ence β-Mo2C catalysts in order to establish if the superior cata- deposition of carbon during the DRM catalytic test in a lytic performance of 2D-Mo2CTx/SiO2 with respect to β-Mo2C temperature programmed oxidation (TPO) experiment performed (normalized per mass of Mo) is due to the improved utilization of in a thermogravimetric analyzer (TGA) with a simultaneous Mo atoms in the 2D-nanosheet catalyst, or due to a higher detection of the off gas by mass spectrometer (MS, Supplementary activity of the Mo sites in 2D-Mo2CTx/SiO2 owing to their dif- Fig. 19). Noteworthy, no increase in the intensity of the CO2 peak ferent structure and/or morphology, or a combination of these was observed for the used catalysts after 10 and 100 min of factors. Our results show that the estimated areal CO capacity in reaction, 2D-Mo2COx/SiO2−TOS10 and 2D-Mo2COx/SiO2−TOS100 2D-Mo2CTx/SiO2 and the reference β-Mo2C catalysts are nearly (note: these materials remained light gray before and after catalytic the same (ca. 0.2 and 0.25–0.5 μmol m–2, respectively; see Sup- testing). Consistent with this, 2D-Mo2COx/SiO2−TOS10 and 2D- plementary Table 4 and “Methods” for details). However, the Mo2COx/SiO2−TOS100 did not lose any weight during the TPO measured methane conversion rate is ca. 3 orders of magnitude experiment. Thus, TPO data provide strong evidence that no NATURE COMMUNICATIONS | (2020)11:4920 | https://doi.org/10.1038/s41467-020-18721-0 | www.nature.com/naturecommunications 5 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-18721-0 detectable carbon deposition occurred on 2D-Mo2COx/SiO2 under highlight the remarkable structural stability of silica-supported the experimental conditions of the DRM tests. 2D-Mo2COx nanosheets under catalytic DRM environments as While the activation of metallic DRM catalysts by CO2 well as in an oxidative CO2 atmosphere at 800 °C. treatment has been reported before44, in particular through the To further interrogate the relationship between the oxygen removal of deposited carbon by gasification45, the reactivation of coverage of 2D-Mo2COx/SiO2 and the catalytic activity in DRM, carbide-based DRM catalysts by CO2 is, to the best of our an in situ XANES experiment was performed (BM31, ESRF)46 in knowledge, unprecedented and remarkable given that Mo2C- a capillary cell reactor wherein 2D-Mo2CTx/SiO2 was first treated based catalysts for DRM are typically oxidized by CO2 yielding under H2 (750 °C, 0.5 h) to prepare 2D-Mo2C/SiO2 followed by inactive MoO211,15. Note that the reference materials could not be exposing the material to DRM conditions at 730 °C (CH4: CO2 = reactivated in CO2, which is explained by the formation of MoO2 1: 1, SV = 3 × 104 L gMo−1 h−1, contact time of 0.1 ms gMo mL−1). layer covering the active phase (Supplementary Fig. 18). We In those DRM conditions, an immediate oxidation of 2D-Mo2C/ hypothesized that the deactivation of 2D-Mo2COx/SiO2 with time SiO2 to 2D-Mo2COx/SiO2 occurred as indicated by a change of on stream (Fig. 2a) occurred due to a continuous depletion of the the Mo oxidation state from ca. +0.2 to +5 (Mo edge positions of oxygen coverage of the CO2-activated molybdenum oxycarbide 20000.7 and 20014.4 eV, respectively), consistent with a high surface. Therefore, to relate the decrease of the catalytic oxygen coverage of the material (Supplementary Fig. 23). performance of our Mo oxycarbide catalyst to the oxidation state Monitoring the off-gas composition using a compact gas of Mo, 2D-Mo2COx/SiO2 that has been exposed for 10 or 100 min chromatograph (GC)47 shows no catalytic activity at those high to DRM conditions was analyzed by XPS and XANES. The oxygen coverages of 2D-Mo2COx. Hence, no reduction of 2D- deconvolution of the Mo 3d spectrum of 2D-Mo2COx/SiO2−TOS10 Mo2COx with TOS occurs at low contact times of the synchrotron reveals Mo6+, Mo5+, Mo4+, and carbidic Mo states with Mo4+ experiment. The oxygen coverage of 2D-Mo2COx/SiO2 is high being the main component at 41% (Supplementary Table 2). Note under those DRM conditions as a change to a CO2 flow does not that metallic Mo is absent in the fit of 2D-Mo2COx/SiO2−TOS10 in affect the XANES spectra significantly (Fig. 2c and Supplemen- contrast to 2D-Mo2COx/SiO2 discussed above. This might be due tary Fig. 24). Switching to a N2 atmosphere and co-feeding ca. 5% to the facile carburization of Mo0 with time on stream or the of H2 instantly reduces Mo to the carbidic state, yielding XANES removal of volatile Mo carbonyl species. The deconvolution of the spectra similar to that of 2D-Mo2C/SiO2 (Fig. 2c). However, Mo 3d spectrum of 2D-Mo2COx/SiO2−TOS100 reveals the same introducing the DRM feed re-forms immediately 2D-Mo2COx/ oxidation states as those found in 2D-Mo2COx/SiO2−TOS10 SiO2 with a Mo oxidation state of ca. +5, inactive for the DRM. (Mo6+, Mo5+, Mo4+, and carbidic states, Supplementary Fig. 23). The major difference between the laboratory (Fig. 2a) and However, carbidic Mo is the main component in 2D-Mo2COx/ synchrotron experiments lies in the space velocity that is roughly SiO2−TOS100 with a fitted fraction of 45% (Supplementary 25 times higher in the synchrotron experiment. We conclude that Table 2). This comparison of 2D-Mo2COx/SiO2 after 10 or 100 while the ex situ XANES data associate the highest DRM activity min of TOS clearly points at a reduced surface oxygen coverage of 2D-Mo2COx/SiO2 with an average Mo oxidation state of +4 and therefore a higher fraction of carbidic Mo in the less active and subsequent deactivation with TOS with the reduction of Mo, catalyst, namely 2D-Mo2COx/SiO2−TOS100. Noteworthy, similarly in situ Mo K-edge DRM experiment revealed that the oxycarbidic to the as prepared 2D-Mo2CTx/SiO2, the XRD pattern of 2D- surface with a high oxygen coverage and an average oxidation Mo2COx/SiO2-TOS100 features no Bragg peaks, thereby further state of +5 is inactive for DRM. The latter observation is confirming that no sintering to crystalline molybdenum oxides or consistent with the reduction of the average Mo oxidation state carbides had occurred after this time (Supplementary Fig. 6). from +5.5 to +4 in 2D-Mo2COx/SiO2 within 10 min TOS in a XANES data agree well with the XPS results as the catalyst with laboratory DRM catalytic test (W/F = 3 ms gMo mL−1). Note that the highest activity (2D-Mo2COx/SiO2−TOS10) features a Mo K- no characteristic MoO2 features were observed with both ex and edge position at 20011.5 eV, which corresponds to an average in situ XANES spectra for the catalyst after DRM catalytic tests, oxidation state of Mo of +4 (Supplementary Fig. 14 and Fig. 2b). indicating that 2D-Mo2COx/SiO2 does not oxidize to bulk Mo Note that the pre-edge feature in 2D-Mo2COx/SiO2 is well- oxides in the presence of both CO2 and H2O (the latter, owing to defined, but poorly resolved in 2D-Mo2COx/SiO2−TOS10, suggest- the parallel reverse water-gas shift (RWGS, CO2 + H2 ↔ CO + ing a change in the geometry of the Mo sites due to the partial H2O) reaction). reduction during DRM conditions (Supplementary Fig. 14). 2D- Next, we varied the contact time of the laboratory DRM Mo2COx/SiO2−TOS100 displays a Mo K-edge position at 20009.0 experiment and found that for a contact time between 1 and 2 ms eV, corresponding to an average Mo oxidation state of ca. +3. gMo mL−1, the performance of 2D-Mo2COx/SiO2 in DRM is stable This means that 2D-Mo2COx/SiO2 does not produce any over 20 h TOS (Fig. 2d). Note that the measured ratio of CO:H2 of measurable amount of coke when operated under a contact time ca. 2 and 3 for, respectively, 2 and 1 ms gMo mL−1, indicates that of 3 ms gMo mL−1 and deactivates by reduction, lowering the the RWGS reaction takes place as a side reaction under these oxygen coverage with TOS compared to the as-CO2-treated 2D- conditions. Noteworthy, the contact times associated with a stable Mo2COx/SiO2 material. Subjecting 2D-Mo2COx/SiO2−TOS100 to a catalytic performance are intermediate between the contact times CO2 atmosphere (800 °C, 1 h) recovers the fully-oxidized 2D- used in the synchrotron experiment and laboratory experiments in Mo2COx/SiO2 characterized by a Mo oxidation state of ca. +5.5, which catalyst deactivation occurred (ca. 0.1 and 3 ms gMo mL−1, that is the same as in the as-prepared 2D-Mo2COx/SiO2 material respectively). This suggests that at short contact times, dissociation (Supplementary Fig. 14). Likewise, the identical XANES spectra of CO2 proceeds fast and yields a too high oxygen coverage of the of 2D-Mo2COx/SiO2 prior to DRM testing and after a CO2-based Mo2COx surface, while at high contact times CH4 and/or DRM regeneration step are also in line with a fully recovered oxygen products (CO and H2) deplete the surface coverage of Mo2COx coverage after regeneration (Fig. 2a and Supplementary Fig. 14). beyond the optimal value, leading to catalyst deactivation. Note TEM images of the regenerated material reveal that the 2D- that 2D-Mo2COx/SiO2 exhibits a high stability over 20 h on Mo2COx nanosheets have a similar morphology to the initial 2D- stream at W/F = 1–2 ms gMo mL−1 despite the presence of steam Mo2CTx nanosheets as the hexagonal ordering of the Mo atoms produced by the competing RWGS reaction. However, while at and the characteristic scrolling of the nanosheets at the edges are these conditions the selectivity and catalytic activity are stable, still observed (Supplementary Fig. 10). Overall, these results both values are lower than the initial (viz. TOS = 10 min) activity 6 NATURE COMMUNICATIONS | (2020)11:4920 | https://doi.org/10.1038/s41467-020-18721-0 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-18721-0 ARTICLE a d b c Fig. 3 Theoretical modeling of DRM pathways. Top view of the a clean 2D-Mo2C, b 2D-Mo2C-1 O ML and c 2D-Mo2C-0.67 O ML surfaces. d Gibbs free energy profile of the dry reforming of methane catalyzed by the 2D-Mo2C-0.67 O ML surface. Gibbs energies are given with respect to initial reactants, CH4 and CO2, at 800 °C and under 1 bar pressure. For brevity, the reaction profile is presented without H–H coupling steps. The full profile is shown in Supplementary Fig. 25. TS stands for transition state. and selectivity of 2D-Mo2COx/SiO2 at W/F = 3 ms gMo mL−1 Theoretical modeling of DRM pathways on 2D-Mo2COx/SiO2. (Fig. 2a, d). Our search for a stable catalytic performance with DFT calculations provided further insight into the role of oxygen increased activity and selectivity as compared to conditions of W/ coverage for DRM activity and the energetically preferred struc- F = 1–2 ms gMo mL−1 uncovered that performing the catalytic test ture of 2D-Mo2C nanosheets after CO2 treatment and under at a higher pressure (i.e., 8 bar instead of 1 bar) and keeping W/ DRM conditions. 2D-Mo2C was modeled by three atomic layers, F = 3 ms gMo mL−1 yields a CO:H2 ratio close to 1.5 and only a with two Mo layers exposed to reactants and a carbon layer in slightly lower methane conversion rate as compared to that at 1 between these Mo layers (Fig. 3). The stability of the orthor- bar (0.42 vs 0.34 mol(CH4) mol(Mo)−1 s−1) (Supplementary hombic Mo2C and Mo2COx phases has been previously evaluated Fig. 22). Thus, oxygen coverage higher than optimal (i.e., by DFT calculations48,49. We derived 2D-Mo2C surface from the corresponding to an Mo oxidation state of +4) not only decreases crystal structure of Mo2Ga2C, slicing it along the 001 direction, the methane conversion rates, but also compromizes the selectivity since it is the most convenient approach to obtain a 2D system due to the competing RWGS reaction. with the desired stoichiometry (Supplementary Fig. 29). The Finally, aiming to induce major structural changes in 2D- computational model used has Mo sites surrounded by three Mo2COx/SiO2, we subjected the material to three subsequent carbon atoms (Mo–C distance of 2.09 Å), with the inner layer of DRM-catalysis-CO2-regeneration cycles with a total TOS of ca. carbon atoms separating the Mo atoms of the top and bottom 30 h using a contact time of 2 ms gMo mL−1 for the DRM step surfaces by 2.88 Å. The coordination number of the Mo surface (Supplementary Fig. 20). Note, that the resulting material atoms is 6; the Mo–Mo distances at the surface are 3.03 Å. Note collected after the DRM step (labeled 2D-Mo2COx/SiO2−spent) that EXAFS fitting of 2D-Mo2CTx/SiO2, as discussed above, was only partially deactivated, featuring a methane conversion revealed that Mo is coordinated with O and with C/O at a dis- rate of 0.08 mol(CH4) mol(Mo)−1 s−1. Similarly to 2D-Mo2COx/ tance of ca. 1.74–2.03 Å, which is close to the 2.09 Å Mo–C dis- SiO2−TOS10, deconvolution of the Mo 3d spectrum of the spent tance of the theoretical model (Supplementary Table 3). The catalyst reveals Mo6+, Mo5+, Mo4+, and carbidic Mo compo- EXAFS analysis revealed Mo–Mo distances in the range of nents, with Mo4+ being the main component at 62% (Supple- 2.63–3.42 Å (Table 1); this splitting of distances, as compared mentary Fig. 23 and Supplementary Table 2). TEM analysis of with the single Mo–Mo shell of the ideal model, is an indication 2D-Mo2COx/SiO2−spent (exposed to air) reveals the presence of of lattice strain in the Mo2CTx sheets. We note that our idealized 2D nanosheets with a hexagonal ordering (Supplementary Fig. 11) model does not contain lattice strain or curvature of the experi- and overlapping signals of Mo, C, and O in the EDX maps, mental 2D-Mo2COx/SiO2 nanosheets, and the model does not confirming the oxycarbidic nature of the catalyst. Additionally, include the interaction of 2D-Mo2COx with dehydroxylated SiO2 areas containing a bulk crystalline phase of MoO2 with a distorted (i.e., covalent grafting observed by IR spectroscopy). Additionally, rutile structure were revealed by FT HR-TEM imaging (Supple- the average Mo oxidation state at one monolayer (ML) oxygen mentary Fig. 11), in agreement with the high fraction of Mo4+ coverage of our model is lower than the experimentally deter- identified in the XPS fittings of this sample and its reduced mined (by XANES) Mo oxidation state for 2D-Mo2COx/SiO2 (ca. activity relative to the most active 2D-Mo2COx/SiO2−TOS10 +5.5, Fig. 1e and Supplementary Table 5). Thus, it is likely that catalyst. In agreement with TEM results, the XRD analysis of 2D-Mo2COx/SiO2 contains more oxygen atoms than 1 O ML. 2D-Mo2COx/SiO2-spent revealed small but discernable peaks That being said, our idealized model is able to rationalize the corresponding to MoO2. That said, supported nanosheets remain experimental trends and it provides insights on the oxygen cov- the dominant morphology of the spent catalyst, underlining its erage (and the oxidation state of Mo), which has a significant remarkable stability. impact on the DRM catalytic activity. Our evaluation of the NATURE COMMUNICATIONS | (2020)11:4920 | https://doi.org/10.1038/s41467-020-18721-0 | www.nature.com/naturecommunications 7 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-18721-0 thermodynamic stability of the different oxygen coverages of the of methane generates an adsorbed hydrogen that can be released 2D-Mo2C slab at 800 °C under CO2 reveals that a molybdenum as an H2 in a strongly entropically favored process, releasing 59 oxycarbide phase with one oxygen monolayer (1 O ML) adsorbed kcal mol−1 at 800 °C per each desorbed H2 molecule (Fig. 3). on each of the Mo layers, i.e., 2D-Mo2CO2, is preferred ther- Besides the activation of CO2 and CH4, another key step in the modynamically (see the Supplementary computational details for dry reforming of methane is the oxidation of either CH* or C* further details). 2D-Mo2CO2 is formed from 2D-Mo2C and CO2 species on the oxycarbide surface by adsorbed oxygen (O* or that react yielding one oxygen atom chemisorbed on the carbide structural oxygen)57 ultimately producing adsorbed CO. In the surface with a concomitant release of one gas-phase CO molecule. case of 2D-Mo2COx, and in contrast to metallic surfaces50, the To reach this full oxygen surface coverage, 18 CO2 molecules preferred mechanism is the direct oxidation of adsorbed C* by react per unit cell (9 CO2 molecules per each side of the slab) in a adsorbed O* to form CO*. This pathway has a Gibbs energy highly exergonic reaction releasing Gibbs energy of −928 kJ barrier of 175 kJ mol−1, and it is endergonic by only 22 kJ mol−1. mol−1 per unit cell (ca. −52 kJ mol−1 per CO2 molecule). The The alternative pathway involving the coupling of CH* and O* to reaction of an additional CO2 molecule yields an overall Gibbs form adsorbed CHO*, followed by its decomposition to co- free energy of −903 kJ mol−1, therefore, this step is unfavorable adsorbed CO* and H*, is less favorable (Supplementary Fig. 26). and the most stable configuration corresponds to 1 O ML cov- Finally, desorption of CO is exergonic by 20 kJ mol−1 per CO erage (i.e. one oxygen atom per Mo atom on the 2D-Mo2C sur- molecule. Overall, the calculated energy barriers of the proposed face), which is in agreement with the observed stability of 2D- mechanism are feasible under the DRM conditions studied here Mo2COx/SiO2 in a CO2 atmosphere at 800 °C. Under DRM (800 °C, CH4: CO2 = 1:1) and are thermodynamically exergonic conditions, the reaction of 2D-Mo2C 1 O ML with one methane by 32 kJ mol−1. The most energy demanding step is the initial molecule has favorable energetics for the formation of one CO methane activation step, which is associated with a significantly and two H2 gas-phase molecules, reducing thereby the surface higher barrier than the direct CO2 activation or the C–O coupling oxygen coverage. Specifically, decreasing the oxygen coverage of step forming CO. In addition, the relative rates for the formation 2D-Mo2COx from a full coverage of 1 O ML to 0.89, 0.78, and and consumption of intermediate oxygen and carbon surface 0.67 O ML proceeds with reaction energetics of −30, −18 and 0 species, and, therefore their ratio on the 2D-oxycarbide surface, kJ mol−1, that is changing from exergonic to isoergonic ener- has to be similar in order to obtain a good catalytic performance. getics. This data confirm that the full coverage of the surface by 1 Our theoretical calculations suggest that the ratio of C* to O* is O ML is thermodynamically unstable under DRM conditions, as another possible factor (besides the oxidation state of Mo atoms observed experimentally. on the surface) that can contribute to the significantly higher In the following, we selected 2D-Mo2C-0.67 O ML as a starting catalytic activity of the 2D-Mo2C-0.67 O ML system in compar- point to calculate the energy profile of the DRM reaction ison to the clean 2D-Mo2C surface. pathways since the oxygen coverage of this structure is consistent We also evaluated the energy profile of the DRM reaction on a with the optimal oxygen coverage observed experimentally clean 2D-Mo2C surface and found that the activation of CO2 (characterized by an average Mo oxidation state of +4). This occurs on this surface easier than on 2D-Mo2C-0.67 O ML. The partial 0.67 O ML oxygen coverage allows for free reactive Mo produced CO* and O* surface species are significantly more sites on the surface. The energetic penalty to form 2D-Mo2C-0.67 stabilized on 2D-Mo2C compared to 2D-Mo2C-0.67 O ML, O ML from 2D-Mo2CO2 is merely +30 kJ mol−1 higher than featuring a lower Gibbs energy barrier for CO2 activation of forming the most stable oxygen coverage under the DRM 109 kJ mol−1 (compare to 173 kJ mol−1 for the 2D-Mo2C-0.67 O conditions, i.e., 2D-Mo2C-0.88 O ML. The adsorption of CO2 ML surface). The resulting adsorbed CO* and O* species are on 2D-Mo2C-0.67 O ML is endergonic by 143 kJ mol−1, leading located at −99 kJ mol−1 for the clean 2D-Mo2C surface, which is to a bent CO2 on the Mo sites of the oxycarbide surface. significantly lower than for the 2D-Mo2C-0.67 O ML surface Dissociation of this bent CO2 to adsorbed CO* and O* products (−12 kJ mol−1, Supplementary Fig. 27). That said, the free energy requires 173 kJ mol−1 of Gibbs free energy, and is overall barrier for CH4 activation is very similar for both surfaces (238 vs. exergonic by 12 kJ mol−1 (Fig. 3). This Gibbs energy barrier is 243 kJ mol−1). Another difference with implications for the even lower than the one obtained for a Ni(111) surface at 650 °C catalytic activity is the substantial stabilization of CH* and C* (209 kJ mol−1)50. In terms of required electronic energy, CO2 species on the 2D-Mo2C surface with respect to the initial cleavage of the adsorbed, bent CO2 intermediate has a very low reactants. The high adsorption energy of C* and O* species lead energy barrier of +31 kJ mol−1. This bent CO2 adsorbate was to a significantly higher energy barrier for the C–O coupling step evaluated previously on several metal carbides51 and was (+219 kJ mol−1) on 2D-Mo2C, in contrast to the respective suggested to enhance the reactivity of CO2 on interfaces between barrier on the 2D-Mo2C-0.67 O ML surface (+175 kJ mol−1). a metal and a metal carbide52 or a metal oxide53–55. DFT This overstabilization of CH* and C* and O* intermediates on a calculations on the clean, Mo-terminated surface of δ-Mo2C(001) clean 2D-Mo2C surface explains its lower activity for DRM. In showed that CO2 activation on molybdenum carbides is also contrast, the 2D-Mo2C-0.67 O ML surface does not bind inter- associated with low activation barriers56. In turn, the activation of mediates too strongly. Figure 4 summarizes our computational CH4 via C–H bond cleavage requires a higher Gibbs activation studies, which agree well with the experimental results discussed energy of 242 kJ mol−1 and the reaction occurs on Mo sites of the above. oxycarbide surface. The calculated energy barrier is comparable to To conclude, we have shown that supporting and dispersing the ones computed for Ni(111), Pd(111), and Pt(111) surfaces50. delaminated 2D-Mo2CTx nanosheets on silica prevents to a large Conversely, CH4 activation on oxygen sites to form *OCH3 and extent their thermal sintering to bulk Mo2C and MoO2 phases, *OH is endergonic by 105 kJ mol−1, which is by 45 kJ mol−1 less paving the way for the application of 2D Mo-carbides and favorable than CH4 activation on the Mo sites of the same surface. oxycarbides for high temperature heterogeneous catalysis. Subsequent CH3 and CH2 activation steps to form CH2* and XANES and XPS analysis indicate that oxycarbidic 2D-Mo2COx CH* species are significantly less energy demanding than the with an average Mo oxidation state of +4 is the active phase for initial C–H activation of CH4, with respective Gibbs energy DRM and oxidizing or reducing this state reduces the methane barriers of 63 and 38 kJ mol−1. However, the cleavage of the C–H consumption rate. DFT calculations suggest that the rate-limiting bond in CH* to form adsorbed C and H has a high Gibbs energy step in the DRM on 2D-Mo2COx is CH4 cleavage and that the barrier of 117 kJ mol−1. Note that each successive C–H activation oxygen coverage resulting from the initial CO2 activation steps 8 NATURE COMMUNICATIONS | (2020)11:4920 | https://doi.org/10.1038/s41467-020-18721-0 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-18721-0 ARTICLE 2D-Mo2COx 50° of 0°, respectively). The survey and narrow scans used a pass energy of 50 eV and 25 eV, respectively, and C 1s peak of adventitious carbon was set at 284.8 eV to O compensate for charge induced shifts. We used a home-build air-tight cell to Mo Mo Mo transfer reduced or used specimens between the glovebox and the XPS instru- CO2 CH4 ment58. CasaXPS Version 2.3.19PR1.0 software was used to analyze XPS data. The C C background was subtracted according to Shirley59, and the application of the Mo Mo Mo Suggested atomic sensitivity factors (ASF) of Scofield allowed to estimate the atomic com- O RDS position60. HAADF STEM images were recorded using FEI Talos F200X with the – CO – H2 accelerating voltage of 200 kV. All samples were prepared by shaking the holey carbon TEM grid in the vial with the dry specimen powder. 2D-Mo2C/SiO2 and Mo Mo Mo CH3 2D-Mo2COx/SiO2 were deposited on TEM grids in the glove box and transferred to O C C the microscope without ambient exposure by using the TEM vacuum transfer Mo Mo Mo Mo Mo Mo holder (Fischione, 2560). FTIR Spectroscopy was conducted using a 200 mL glass reactor with IR-transparent CaF2 windows. A thin pellet of the sample was pressed in a glovebox and loaded to the reactor using a glass sample holder. A spectrum was recorded in an argon atmosphere in a transmission mode on a Nicolet 6700 FTIR spectrophotometer using 64 scans at a resolution of 2 cm−1. CO chemi- – CO O C – H2 sorption was performed using an AutoChem system (Micromeritics) with a ther- Mo Mo Mo mal conductivity detector (TCD). Approximately 100 mg of the material was loaded in a U-shape reactor and reduced at 800 °C under 5% H2/Ar for 1.5 h (10 °C Fig. 4 A simplified proposed mechanism for the dry reforming of min−1, 50 mL min−1). The sample was then cooled down under a He atmosphere to 450 °C and degassed for 2 h. The first CO adsorption isotherm was carried out at methane on 2D-Mo2COx. The most active catalyst has a partial oxygen 50 °C. The reactor was then purged with He for 30 min to remove weakly adsorbed surface coverage, corresponding to Mo(+4) oxidation state. The C–H bond species prior to performing the second CO isotherm, also at 50 °C. The amount of activation of methane is proposed as RDS, i.e., rate determining step, based strongly adsorbed CO was determined as a difference between the first and the on DFT calculations. second adsorption isotherms, which provided the CO adsorption capacity. The surface area of the 2D-Mo2C nanosheets in 2D-Mo2C/SiO2 was estimated based on the crystal structure of Mo2CTx32. Considering the hexagonal ordering of the Mo reduces the binding energy of C* and O* species and lowers, in atoms with a Mo–Mo distance of 2.8629 Å gave an area estimate of 7.098 Å2 per turn, the energy barriers of the C–O coupling step. Importantly, Mo atom. Based on the Mo content in 2D-Mo2CTx/SiO2, determined by ICP, and the mass of the 2D-Mo2CTx/SiO2 specimen used for CO chemisorption, the when deactivation by reduction of the oxygen coverage has amount of strongly chemisorbed CO was normalized per the surface area of the occurred, the catalyst can be reactivated to its initial activity by re- two-dimensional molybdenum nanosheets (Supplementary Table 4). oxidation in a pure stream of CO2. Noteworthy, the nanosheet Experiments using X-ray absorption spectroscopy (XAS) were performed at the morphology of 2D-Mo2COx is maintained during catalysis and Swiss-Norwegian Beamlines (SNBL, BM31)46 at the European Synchrotron Radiation Facility (ESRF, Grenoble, France). A double-crystal Si (111) regeneration. In addition, we identified hydrodynamic conditions monochromator with continuous scanning in transmission mode (unfocused that ensured a stable performance over 20 h time on stream, beam, 1–2 ∙ 109 ph s−1 mm−1) was used to collect XAS spectra at the Mo K-edge. owing to an in situ regeneration of the optimal oxygen coverage Edge position of the Mo foil set at 20000 eV served for calibration of the XAS data. in those conditions. Overall, the results obtained indicate that A quartz capillary reactor (outer diameter 1.5 mm, wall thickness 0.1 mm) was used to perform the in situ DRM experiment (Supplementary Fig. 1)46. 2D-Mo2CTx/ supported 2D-Mo2COx is a very active and robust catalyst for SiO2 (ca. 2 mg), placed between two quartz wool plugs in a quart capillary reactor, high temperature catalytic applications, exhibiting a remarkable was reduced in 20 vol% H2 in N2 (10 mL min−1 total flow rate, 50–750 °C the stability in oxidative environments. temperature range, ca. 9 °C min−1, 20 min holding time). The oxidation was performed in pure CO2 (10 mL min−1, 730 °C, 2 h). DRM catalytic experiments were performed at 730 °C using a mixture of CH4 and CO2 of varied ratio diluted Methods in N2 (5 mL min−1 total flow rate, SV ca. 30000 L gMo−1 h−1). The composition of Materials. Multi-layered Mo2CTx material (m-Mo2CTx) was synthesized following the outlet gas was followed by a portable GC (Global Analyser Solutions) with TCD a reported method32. 2D-Mo2CTx nanosheets were delaminated from m-Mo2CTx and FID detectors. Ex situ XAS data were collected from pellets with an optimized (50 mg) in ethanol (10 mL) using a pulse sonication (1.5 h, 85 W, 20 kHz) (Fisher amount of specimen mixed with cellulose. All reduced materials or used catalysts Scientific, FB120). Resulting suspension was centrifuged (7000 rpm, 7 min) and the derived from 2D-Mo2C/SiO2 were handled in a glovebox and analyzed in an air- supernatant solution containing colloidal delaminated Mo2CTx nanosheets (d- tight sealed bags. XAS data were processed using the Athena software, and EXAFS Mo2CTx) was further used for the incipient wetness impregnation (IWI) onto SiO2 data were fitted using the Artemis software (Demeter 0.9.25 software package)61. support (150–300 μm particle size fraction of Aerosil 300 that had been calcined at The edge position for Mo and Mo2C was defined as the position of the first 950 °C, 194 m2 g−1 surface area by nitrogen physisorption). The IWI procedure maximum of the first derivative curve. The second maximum of the first derivative using ethanol colloidal solution of d-Mo2CTx was repeated multiple times and after curve was chosen as the edge position for oxide and oxycarbide materials. drying (100 °C, 12 h) yielded 2D-Mo2CTx/SiO2 material with Mo loading of 0.48 wt.% (batch 1) and 0.35 wt.% (batch 2) according to elemental analysis (ICP-AES). 2D-Mo2CTx/SiO2 was subsequently reduced in hydrogen (20 vol % H2/N2, 800 °C, Catalytic testing. The catalytic testing of freshly-prepared materials in conditions 1.5 h, 50 mL min−1, 10 °C min−1) and oxidized in CO2 (100% CO2 10 mL min−1, of the dry reforming of methane was carried out in a fixed-bed quartz reactor at 800 °C, 1.5 h) to give 2D-Mo2C/SiO2 and 2D-Mo2COx/SiO2 materials, respectively. atmospheric pressure. In a typical experiment, 100 mg of freshly prepared 2D- To avoid surface modification of the activated materials, 2D-Mo2C/SiO2, 2D- Mo2CTx/SiO2 material was used. Prior to the activity tests, 2D-Mo2CTx/SiO2 was Mo2COx/SiO2, and all activated catalysts were transferred into a nitrogen-filled in situ reduced in 20 vol % H2/N2 (800 °C, 1.5 h, 50 mL min−1, 10 °C min−1) and glovebox (H2O and O2 levels <0.5 ppm) and handled without exposure to air. then oxidized in pure CO2 (800 °C, 1.5 h, isothermal). The activity test was then performed at 800 °C with a variable total flow rate of the feed (10–50 ml min−1, SV = 1200–6000 L gMo–1 h–1, 45% CH4, 45% CO2, and 10% N2). The composition of Characterization. Ex situ XRD data were examined within the 2θ range of 5–90° the off-gas was analyzed via a GC (PerkinElmer Clarus 580) equipped with a (step size and scan time per step were 0.0167° and 3 s, respectively). XRD data were thermal conductivity TCD detector. The carbon balance during the catalytic tests collected on a PANalytical Empyrean X-ray diffractometer equipped with a was generally close to 100 %, maximal deviations noted were ±10%, observed when Bragg–Brentano HD mirror and operated at 45 kV and 40 mA (Cu Kα radiation, λ switching gases. = 1.5418 nm). TGA was performed on Mettler Toledo TGA/DSC 3 instrument. The We note that catalytic results reported were obtained with freshly-prepared 2D- concentration of 2D-Mo2CTx flakes in the colloidal solutions was determined by Mo2CTx/SiO2. When as prepared 2D-Mo2CTx/SiO2 was left for a prolonged time drying a 750 μL aliquot in a sapphire crucible (900 μL) at 80 °C for 1 h (5 °C min−1). (1 month–1.5 years) in air, a complete or partial fragmentation of 2D-Mo2CTx TPO was performed on Mettler Toledo TGA/DSC 1 instrument. Mo material nanosheets into nanoparticles had occurred according to TEM characterization (20 mg) was placed in an alumina crucible (70 μL) and heated to 800 °C under air (Supplementary Fig. 7). The fully fragmented material is inactive in DRM, which flow (10 °C min−1, 125 mL min−1). The H2O and CO2 content in the outlet gas emphasizes the importance of 2D nanosheets morphology for the high catalytic was followed by a mass spectrometer MKS Cirrus 3. Inductively coupled plasma- activity. atomic emission spectroscopy (ICP-AES) analyses were performed in Remagen, Germany, by the Mikroanalytisches Labor Pascher. XPS data were acquired using a Sigma II instrument of Thermo Fisher Scientific, equipped with an UHV chamber Computational details. The Vienna Ab Initio Simulation Package (VASP) was (non-monochromatic 200 W Al Kα source, a hemispherical analyzer, and a seven utilized to conduct the periodic DFT calculations62–64. The projector-augmented- channel electron multiplier, with the analyzer-to-source and the emission angles of wave (PAW) method was used, where pseudopotentials describe interactions NATURE COMMUNICATIONS | (2020)11:4920 | https://doi.org/10.1038/s41467-020-18721-0 | www.nature.com/naturecommunications 9 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-18721-0 between valence electrons and ion cores while the electronic wave functions are to synthesis gas using stoichiometric feedstocks. Chem. Commun. 1, 39–40 expanded as a discrete plane wave (PW) basis set65. All calculations utilized a PW (1997). energy cutoff of 500 eV. The generalized gradient approximation (GGA), including 15. Brungs, A. J., York, A. P. E., Claridge, J. B., Marquez-Alvarez, C. & Green, M. dispersion corrections by the BEEF-vdW density functional, were used to treat L. H. Dry reforming of methane to synthesis gas over supported molybdenum electron exchange and correlation66. carbide catalysts. Catal. Lett. 70, 117–122 (2000). The sampling of the Brillouin zone was performed using Monkhorst–Pack 16. Xiao, T., Wang, H., Da, J., Coleman, K. S. & Green, M. L. H. Study of the grids67 with K-point values equal to 3 × 3 × 1 for the Mo2C surface. Electronic preparation and catalytic performance of molybdenum carbide catalysts occupancies were determined according to a Gaussian Smearing (with a smearing prepared with C2H2/H2 carburizing mixture. J. Catal. 211, 183–191 (2002). value of 0.05 eV). Self-consistent field (SCF) calculations of the electronic structure 17. Xiao, T.-c, Hanif, A., York, A. P. E., Nishizaka, Y. & Green, M. L. H. Study on were considered converged once the electronic energy change between two the mechanism of partial oxidation of methane to synthesis gas over consecutive steps was below 10−5 eV. All geometries were optimized fully, i.e., until molybdenum carbide catalyst. Phys. Chem. Chem. Phys. 4, 4549–4554 (2002). the forces acting on each atom were converged below 10−4 eV Å−1. Dipole 18. LaMont, D. C., Gilligan, A. J., Darujati, A. R. S., Chellappa, A. S. & Thomson, corrections were applied in the z-direction and placed in the center in the unit cell. W. J. The effect of Mo2C synthesis and pretreatment on catalytic stability in The energy of isolated molecules was determined by a Γ-point calculation such that oxidative reforming environments. Appl. Catal., A 255, 239–253 (2003). each species was placed in a box with dimensions 15 × 15.5 × 16 Å. The climbing 19. LaMont, D. C. & Thomson, W. J. Dry reforming kinetics over a bulk image nudge elastic band (CI-NEB) method allowed locating transition states (TS) molybdenum carbide catalyst. Chem. Eng. Sci. 60, 3553–3559 (2005). using eight intermediate images68. A frequency analysis confirmed further that 20. Liang, P. et al. Simple synthesis of ultrasmall β-Mo2C and α-MoC1−x identified TS structures corresponded to saddle points. Normal vibration modes of nanoparticles and new insights into their catalytic mechanisms for dry adsorbed species were calculated by diagonalization of the Hessian matrix, reforming of methane. Catal. Sci. Technol. 7, 3312–3324 (2017). obtained using a central finite difference approximation with displacements equal 21. Kurlov, A. et al. Molybdenum carbide and oxycarbide from carbon-supported to 0.015 Å in the direction of each Cartesian coordinate. All Mo and C atoms from the carbide were kept fixed during the frequency calculations of the adsorbed MoO3 nanosheets: phase evolution and DRM catalytic activity assessed by species. All the reported energy values in the main text correspond to Gibbs TEM and in situ XANES/XRD methods. Nanoscale 12, 13086–13094 (2020). Energies at 800 °C. The theoretical model of the (001) facet of 2D-Mo2C was 22. Wan, W., Tackett, B. M. & Chen, J. G. Reactions of water and C1 molecules on constructed as presented in Supplementary Fig. 29. The experimental bulk carbide and metal-modified carbide surfaces. Chem. Soc. Rev. 46, 1807–1823 structure of Mo2Ga2C was taken as a starting point. Then, the Ga atoms (in green) (2017). were removed from this structure to generate the bulk Mo2C structure. Next, an 23. Liu, P. & Rodriguez, J. A. Water-gas-shift reaction on molybdenum carbide ideal 2D-facet model for Mo2C was obtained by selecting the inner fragment of the surfaces: essential role of the oxycarbide. J. Phys. Chem. B 110, 19418–19425 cell in the 001 direction (Supplementary Fig. 29). This was followed by the full (2006). geometry optimization of the resulting structure. The Mo2C (001) surface slab were 24. Chen, C.-J. & Bhan, A. Mo2C Modification by CO2, H2O, and O2: effects of taken with dimensions equal to 3 × 3, using two Mo layers on the top and the oxygen content and oxygen source on rates and selectivity of m-cresol bottom of the slab and one middle carbon layer. 15 Å of vacuum were added in the hydrodeoxygenation. ACS Catal. 7, 1113–1122 (2017). direction perpendicular to the surface. 25. Schaidle, J. A. et al. Experimental and computational investigation of acetic acid deoxygenation over oxophilic molybdenum carbide: surface chemistry and active site identity. ACS Catal. 6, 1181–1197 (2016). Data availability 26. Sullivan, M. M. & Bhan, A. Effects of oxygen coverage on rates and selectivity The data supporting the findings of this study are available from the corresponding of propane-CO2 reactions on molybdenum carbide. J. Catal. 357, 195–205 authors upon reasonable request. (2018). 27. Sullivan, M. M., Chen, C.-J. & Bhan, A. Catalytic deoxygenation on transition Received: 17 February 2020; Accepted: 31 August 2020; metal carbide catalysts. Catal. Sci. Technol. 6, 602–616 (2016). 28. Lezcano-Gonzalez, I. et al. Molybdenum speciation and its impact on catalytic activity during methane dehydroaromatization in zeolite ZSM-5 as revealed by operando X-ray methods. Angew. Chem. Int. Ed. 55, 5215–5219 (2016). 29. Anasori, B., Lukatskaya, M. R. & Gogotsi, Y. 2D metal carbides and nitrides (MXenes) for energy storage. Nat. Rev. Mater. 2, 16098 (2017). References 30. Li, Z. & Wu, Y. 2D early transition metal carbides (MXenes) for catalysis. 1. Sinfelt, J. H. & Yates, D. J. Effect of carbiding on the hydrogenolysis activity of Small 15, e1804736 (2019). molybdenum. Nat. Phys. Sci. 229, 27–28 (1971). 31. Zheng, Y. et al. Molybdenum oxide, oxycarbide and carbide: controlling the 2. Levy, R. B. & Boudart, M. Platinum-like behavior of tungsten carbide in dynamic composition, size and catalytic activity of zeolite-supported surface catalysis. Science 181, 547–549 (1973). nanostructures. J. Phys. Chem. C. 123, 22281–22292 (2019). 3. Solymosi, F., Cserényi, J., Szöke, A., Bánsági, T. & Oszkó, A. Aromatization of 32. Deeva, E. B. et al. In situ XANES/XRD study of the structural stability of two- methane over supported and unsupported Mo-based catalysts. J. Catal. 165, dimensional molybdenum carbide Mo2CTx: implications for the catalytic 150–161 (1997). activity in the water–gas shift reaction. Chem. Mater. 31, 4505–4513 (2019). 4. Siaj, M., Oudghiri-Hassani, H., Maltais, C. & McBreen, P. H. Thermally stable 33. Halim, J. et al. Synthesis and characterization of 2D molybdenum carbide alkylidene groups on the surface of β-Mo2C: relevance to methane (MXene). Adv. Funct. Mater. 26, 3118–3127 (2016). aromatization and olefin−metathesis catalysis. J. Phys. Chem. C. 111, 34. Shah, S. A. et al. Template-free 3D titanium carbide (Ti3C2Tx) MXene 1725–1732 (2007). particles crumpled by capillary forces. Chem. Commun. 53, 400–403 (2017). 5. Patt, J., Moon, D. J., Phillips, C. & Thompson, L. Molybdenum carbide 35. Wang, H.-W., Naguib, M., Page, K., Wesolowski, D. J. & Gogotsi, Y. Resolving catalysts for water-gas shift. Catal. Lett. 65, 193–195 (2000). the Structure of Ti3C2Tx MXenes through multilevel structural modeling of 6. Moon, D. J. & Ryu, J. W. Molybdenum carbide water-gas shift catalyst for fuel the atomic pair distribution function. Chem. Mater. 28, 349–359 (2016). cell-powered vehicles applications. Catal. Lett. 92, 17–24 (2004). 36. Graham, P. M., Buschhaus, M. S. A., Baillie, R. A., Semproni, S. P. & Legzdins, 7. Porosoff, M. D., Yang, X., Boscoboinik, J. A. & Chen, J. G. Molybdenum P. Hydroperoxide-Initiated Intramolecular Insertions of NO into metal carbide as alternative catalysts to precious metals for highly selective reduction −carbon bonds. Organometallics 29, 5068–5072 (2010). of CO2 to CO. Angew. Chem., Int. Ed. 53, 6705–6709 (2014). 37. Varjas, C. J., Powell, D. R. & Thomson, R. K. Rapid access to an oxido- 8. Posada-Pérez, S. et al. The conversion of CO2 to methanol on orthorhombic β- alkylidene complex of Mo(VI). Organometallics 34, 4806–4809 (2015). Mo2C and Cu/β-Mo2C catalysts: mechanism for admetal induced change in 38. Takenaka, S., Tanaka, T., Funabiki, T. & Yoshida, S. Structures of the selectivity and activity. Catal. Sci. Technol. 6, 6766–6777 (2016). molybdenum species in silica-supported molybdenum oxide and alkali-ion- 9. Hwu, H. H. & Chen, J. G. Surface chemistry of transition metal carbides. modified silica-supported molybdenum oxide. J. Phys. Chem. B 102, Chem. Rev. 105, 185–212 (2005). 2960–2969 (1998). 10. Chorkendorff, I. & Niemantsverdriet, J. W. Concepts of Modern Catalysis and 39. Beale, A. M. & Sankar, G. In situ study of the formation of crystalline bismuth Kinetics (Wiley, 2003). molybdate materials under hydrothermal conditions. Chem. Mater. 15, 11. Claridge, J. B. et al. New catalysts for the conversion of methane to synthesis 146–153 (2003). gas: molybdenum and tungsten carbide. J. Catal. 180, 85–100 (1998). 40. Pakhare, D. & Spivey, J. A review of dry (CO2) reforming of methane over 12. Schaidle, J. A., Lausche, A. C. & Thompson, L. T. Effects of sulfur on Mo2C noble metal catalysts. Chem. Soc. Rev. 43, 7813–7837 (2014). and Pt/Mo2C catalysts: Water gas shift reaction. J. Catal. 272, 235–245 (2010). 41. Thakur, R. et al. Insights into the thermal and chemical stability of 13. Da Costa, P. et al. Tetralin hydrogenation catalyzed by Mo2C/Al2O3 and WC/ multilayered V2CTx MXene. Nanoscale 11, 10716–10726 (2019). Al2O3 in the presence of H2S. Catal. Today 65, 195–200 (2001). 42. Saggio, G. et al. Synthesis, characterization, and reactivity of the highly 14. York, A. P. E., Claridge, J. B., Brungs, A. J., Tsang, S. C. & Green, M. L. H. unsaturated silica-supported trisiloxy tantalum: (⋮SiO)3Ta(III). Molybdenum and tungsten carbides as catalysts for the conversion of methane Organometallics 21, 5167–5171 (2002). 10 NATURE COMMUNICATIONS | (2020)11:4920 | https://doi.org/10.1038/s41467-020-18721-0 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-18721-0 ARTICLE 43. Coperet, C. et al. Surface organometallic and coordination chemistry toward 66. Wellendorff, J. et al. Density functionals for surface science: exchange-correlation single-site heterogeneous catalysts: strategies, methods, structures, and model development with Bayesian error estimation. Phys. Rev. B 85, 235149 (2012). activities. Chem. Rev. 116, 323–421 (2016). 67. Monkhorst, H. J. & Pack, J. D. Special points for Brillouin-zone integrations. 44. Steib, M., Lou, Y., Jentys, A. & Lercher, J. A. Enhanced activity in methane dry Phys. Rev. B 13, 5188–5192 (1976). reforming by carbon dioxide induced metal-oxide interface restructuring of 68. Henkelman, G., Uberuaga, B. P. & Jónsson, H. A climbing image nudged nickel/zirconia. ChemCatChem 9, 3809–3813 (2017). elastic band method for finding saddle points and minimum energy paths. J. 45. Theofanidis, S. A., Batchu, R., Galvita, V. V., Poelman, H. & Marin, G. B. Chem. Phys. 113, 9901–9904 (2000). Carbon gasification from Fe–Ni catalysts after methane dry reforming. Appl. Catal., B 185, 42–55 (2016). 46. van Beek, W., Safonova, O. V., Wiker, G. & Emerich, H. SNBL, a dedicated Acknowledgements beamline for combined in situ X-ray diffraction, X-ray absorption and Raman The authors thank ScopeM, the Laboratory of Surface Science and Technology, the scattering experiments. Phase Transit 84, 726–732 (2011). Laboratory of Surface and Interfacial Chemistry (all from ETH Zürich) for the use of 47. Tsoukalou, A. et al. The structural evolution and dynamics of an In2O3 their electron microscopy, XPS, and high vacuum and facilities, respectively. The Eur- Catalyst for CO2 hydrogenation to methanol: an operando XAS-XRD and opean Synchrotron Radiation Facility (ESRF), through proposal CH5229, is acknowl- in situ TEM study. J. Am. Chem. Soc. 141, 13497–13505 (2019). edged for provision of beamtime. Dr. Wouter van Beek and Dr. Dragos Stoian are 48. Medford, A. J., Vojvodic, A., Studt, F., Abild-Pedersen, F. & Nørskov, J. K. acknowledged for the support during the XAS experiments. Dr. Agnieszka Kierzkowska, Elementary steps of syngas reactions on Mo2C(001): adsorption Dr. Elena Willinger and Dr. Davood Hosseini (all ETH Zürich) are acknowledged for thermochemistry and bond dissociation. J. Catal. 290, 108–117 (2012). collecting TEM data. Dr. Hui Zhou (ETH Zürich) is acknowledged for performing CO 49. Likith, S. R. J. et al. Thermodynamic stability of molybdenum oxycarbides chemisorption experiments. This project has received funding from the Swiss Office of formed from orthorhombic Mo2C in oxygen-rich environments. J. Phys. Energy (BFE; SI/500881-01), ETH Zurich (ETH-40 17-2), Spanish “Ministerio de Chem. C. 122, 1223–1233 (2018). Ciencia, Innovación y Universidades” (PGC2018-100818-A-I00; “Proyecto I+D Gen- 50. Foppa, L., Silaghi, M.-C., Larmier, K. & Comas-Vives, A. Intrinsic reactivity of eración del Conocimiento”), and Spanish MEC and the European Social Fund (RyC- Ni, Pd and Pt surfaces in dry reforming and competitive reactions: Insights 2016-19930; “Ramon y Cajal” Fellowship to A.C.-V.). from first principles calculations and microkinetic modeling simulations. J. Catal. 343, 196–207 (2016). Author contributions 51. Kunkel, C., Viñes, F. & Illas, F. Transition metal carbides as novel materials D.L. and A.F. conceived the research project. A.K., A.F., and C.R.M. planned the for CO2 capture, storage, and activation. Energy Environ. Sci. 9, 141–144 experimental work. A.K. and E.B.D. performed catalytic tests, synthesized materials, and (2016). analyzed data. D.L., A.K., and A.T. performed TEM, XAS, and XPS experiments, 52. Asara, G. G., Ricart, J. M., Rodriguez, J. A. & Illas, F. Exploring the activity of a respectively. P.M.A. supervised XAS experiments. A.C.-V. designed and performed DFT novel Au/TiC(001) model catalyst towards CO and CO2 hydrogenation. Surf. calculations. A.K. and A.F. wrote the original and final draft of the paper with con- Sci. 640, 141–149 (2015). tributions from all authors. A.F. and C.R.M. supervised the experimental research. 53. Silaghi, M.-C., Comas-Vives, A. & Copéret, C. CO2 Activation on Ni/γ–Al2O3 catalysts by first-principles calculations: from ideal surfaces to supported nanoparticles. ACS Catal. 6, 4501–4505 (2016). Competing interests 54. Foppa, L. et al. Contrasting the role of Ni/Al2O3 interfaces in water-gas The authors declare no competing interests. shift and dry reforming of methane. J. Am. Chem. Soc. 139, 17128–17139 (2017). 55. Lam, E. et al. CO2 Hydrogenation on Cu/Al2O3: role of the metal/support Additional information Supplementary information is available for this paper at https://doi.org/10.1038/s41467- interface in driving activity and selectivity of a bifunctional catalyst. Angew. 020-18721-0. Chem., Int. Ed. 58, 13989–13996 (2019). 56. Posada-Perez, S. et al. The bending machine: CO2 activation and Correspondence and requests for materials should be addressed to A.C.-V., A.F. hydrogenation on δ-MoC(001) and β-Mo2C(001) surfaces. Phys. Chem. Chem. or C.R.M. Phys. 16, 14912–14921 (2014). 57. Damaskinos, C. M., Vasiliades, M. A. & Efstathiou, A. M. The effect of Ti4+ Peer review information Nature Communications thanks the anonymous reviewer(s) for dopant in the 5 wt% Ni/Ce1-xTixO2-δ catalyst on the carbon pathways of dry their contribution to the peer review of this work. reforming of methane studied by various transient and isotopic techniques. Appl. Catal., A 579, 116–129 (2019). Reprints and permission information is available at http://www.nature.com/reprints 58. Cossu, G., Rossi, A., Arcifa, A. & Spencer, N. D. Development and application of a cost-effective transfer cell for X-ray photoelectron spectroscopy. Incontro Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in di Spettroscopia Analitica ISA Cagliari (2018). published maps and institutional affiliations. 59. Shirley, D. A. High-resolution X-ray photoemission spectrum of the valence bands of gold. Phys. Rev. B 5, 4709–4714 (1972). 60. Scofield, J. H. Hartree-slater subshell photoionization cross-sections at 1254 and 1487 eV. J. Electron Spectros. Relat. Phenom. 8, 129–137 (1976). Open Access This article is licensed under a Creative Commons 61. Ravel, B. & Newville, M. ATHENA, ARTEMIS, HEPHAESTUS: data analysis Attribution 4.0 International License, which permits use, sharing, for X-ray absorption spectroscopy using IFEFFIT. J. Synchrotron Rad. 12, adaptation, distribution and reproduction in any medium or format, as long as you give 537–541 (2005). appropriate credit to the original author(s) and the source, provide a link to the Creative 62. Kresse, G. & Hafner, J. Ab initio molecular dynamics for liquid metals. Phys. Commons license, and indicate if changes were made. The images or other third party Rev. B 47, 558–561 (1993). material in this article are included in the article’s Creative Commons license, unless 63. Kresse, G. & Hafner, J. Ab initio molecular-dynamics simulation of the liquid- indicated otherwise in a credit line to the material. If material is not included in the metal-amorphous-semiconductor transition in germanium. Phys. Rev. B 49, article’s Creative Commons license and your intended use is not permitted by statutory 14251–14269 (1994). regulation or exceeds the permitted use, you will need to obtain permission directly from 64. Kresse, G. & Furthmüller, J. Efficiency of ab-initio total energy calculations for the copyright holder. To view a copy of this license, visit http://creativecommons.org/ metals and semiconductors using a plane-wave basis set. Comput. Mater. Sci. licenses/by/4.0/. 6, 15–50 (1996). 65. Blochl, P. E. Projector augmented-wave method. Phys. Rev. B 50, 17953–17979 (1994). © The Author(s) 2020 NATURE COMMUNICATIONS | (2020)11:4920 | https://doi.org/10.1038/s41467-020-18721-0 | www.nature.com/naturecommunications 11