Alexey
Kurlov‡
a,
Dragos
Stoian
b,
Ali
Baghizadeh
c,
Evgenia
Kountoupi
a,
Evgeniya B.
Deeva
a,
Marc
Willinger§
c,
Paula M.
Abdala
*a,
Alexey
Fedorov
*a and
Christoph R.
Müller
*a
aDepartment of Mechanical and Process Engineering, ETH Zürich, Leonhardstrasse 21, CH 8092 Zürich, Switzerland. E-mail: abdalap@ethz.ch; fedoroal@ethz.ch; muelchri@ethz.ch
bSwiss-Norwegian Beamlines at the European Synchrotron Radiation Facility, 71 Avenue des Martyrs, Grenoble, France
cScientific Center for Optical and Electron Microscopy, ETH Zürich, Auguste-Piccard-Hof 1, CH 8093 Zürich, Switzerland
First published on 1st August 2022
The thermal carburization of MoO3 nanobelts (nb) and SiO2-supported MoO3 nanosheets under a 1:
4 mixture of CH4
:
H2 yields Mo2C-nb and Mo2C/SiO2. Following this process by in situ Mo K-edge X-ray absorption spectroscopy (XAS) reveals different carburization pathways for unsupported and supported MoO3. In particular, the carburization of α-MoO3-nb proceeds via MoO2, and that of MoO3/SiO2via the formation of highly dispersed MoOx species. Both Mo2C-nb and Mo2C/SiO2 catalyze the dry reforming of methane (DRM, 800 °C, 8 bar) but their catalytic stability differs. Mo2C-nb shows a stable performance when using a CH4-rich feed (CH4
:
CO2 = 4
:
2), however deactivation due to the formation of MoO2 occurs for higher CO2 concentrations (CH4
:
CO2 = 4
:
3). In contrast, Mo2C/SiO2 is notably more stable than Mo2C-nb under the CH4
:
CO2 = 4
:
3 feed. The influence of the morphology of Mo2C and its dispersion on silica on the structural evolution of the catalysts under DRM is further studied by in situ Mo K-edge XAS. It is found that Mo2C/SiO2 features a higher resistance to oxidation under DRM than the highly crystalline unsupported Mo2C-nb and this correlates with an improved catalytic stability. Lastly, the oxidation of Mo in both Mo2C-nb and Mo2C/SiO2 under DRM conditions in the in situ XAS experiments leads to an increased activity of the competing reverse water gas shift reaction.
CH4 + CO2 ⇄ 2CO + 2H2 ΔH298K = +247 kJ mol−1 | (1) |
CH4 → C + 2H2 ΔH298K = +75 kJ mol−1 | (2) |
CO2 + H2 ⇄ CO + H2O ΔH298K = +41 kJ mol−1 | (3) |
Approaches to yield nanostructured Mo2C dispersed on a support can result in catalysts with an improved performance relative to the unsupported bulk carbides.17–19 Mo2C-based catalysts are typically obtained by the carburization of a molybdenum oxide precursor such as MoO3 (supported or unsupported).7,20–22 The structure and morphology of the pre-catalyst may influence the carburization pathways, the structure of the activated catalyst, its stability under DRM conditions and, thus, its performance.7,11,16–18 The further rational development of Mo2C-based catalysts requires an understanding of their activation and deactivation routes, i.e., understanding of the structural evolution under pre-treatment and operating conditions.11
In this work, using Mo K-edge X-ray absorption spectroscopy (XAS) we compare the carburization pathways of unsupported α-MoO3 nanobelts (α-MoO3-nb) to that of silica-supported delaminated nanosheets of MoO3 to yield, respectively, β-Mo2C-nb and Mo2C/SiO2. We further study the structural evolution of the prepared TMC-based materials under DRM conditions. By following the local structure and the oxidation state of Mo during carburization and DRM conditions by in situ Mo K-edge XAS, assisted with a multivariate curve resolution – alternating least squares (MCR-ALS) method, we reveal a difference in the carburization pathways of unsupported α-MoO3-nb and silica supported MoO3/SiO2. In particular, while α-MoO3-nb carburizes to Mo2C-nb via intermediate (bulk) oxide phases, the carburization of MoO3/SiO2 proceeds via the formation of a MoOx phase highly dispersed onto SiO2. Although both catalysts undergo oxidation under DRM conditions leading to deactivation, we elucidate how the dispersion of Mo2C onto SiO2 increases its stability under DRM conditions.
Delaminated MoO3 nanosheets, d-MoO3, were obtained by the exfoliation of the synthesized α-MoO3 according to a reported method.24 α-MoO3 (1 g) was ground in an agate mortar with acetonitrile (0.2 mL, Sigma-Aldrich, ACS reagent grade, ≥99.5% purity) and the resulting material was dispersed by sonication in 50% aqueous ethanol (15 mL) for 2 h. After sonication, the suspension was centrifuged (8000 rpm, 30 min) and the supernatant containing dispersed d-MoO3 nanosheets was collected and used for impregnation onto a SiO2 support (150–300 μm particle size fraction of Aerosil 300 that had been calcined at 950 °C, 194 m2 g−1 surface area by nitrogen physisorption).
MoO3/SiO2 was obtained via incipient wetness impregnation (IWI) of the supernatant solution of d-MoO3 (ca. 1.5 mg mL−1 determined by thermogravimetric analysis) onto the SiO2 support. Mo2C-nb and Mo2C/SiO2 were obtained by the carburization of α-MoO3-nb and MoO3/SiO2, respectively, using a mixture of H2 and CH4 (H2:
CH4 = 4
:
1, heating at a rate of 5 °C min−1 from room temperature to 300 °C and from 300 °C to 800 °C at a rate of 2 °C min−1, 1 h), as presented in Fig. 1a.
Transmission electron microscopy (TEM) was performed with a double CS corrected JEOL JEM-ARM 300 kV TEM/STEM microscope equipped with two energy-dispersive X-ray spectroscopy (EDX) detectors with a total solid angle of 1.6 sr. The samples for TEM analysis were prepared by embedding nanobelts in the microscopy resin, sandwiched between two dummy silicon wafers, followed by tripod mechanical polishing and gentle Ar ion milling with a GATAN Precision Ion Polishing System (PIPS II) down to electron transparency. Most images were post-processed using HREM-Filters Pro, a commercially available plug-in for Digital Micrograph Package by applying Wiener filter, to reduce noise and background contributions.
In situ XAS experiments were performed at the Swiss-Norwegian Beamlines (SNBL, BM31) at the European Synchrotron Radiation Facility (ESRF, Grenoble, France). XAS spectra were collected at the Mo K-edge using a double-crystal Si (111) monochromator with continuous scanning in transmission mode. The in situ carburization followed by dry reforming of methane (DRM) experiment was performed in a quartz capillary reactor.11,17,25 Energy calibration of the XAS data was based on Mo foil, set at 20000.0 eV. XAS spectra were collected in the range of 19
800.0–20
800.0 eV, with a total acquisition time of 60 s and a step of 1 eV. In the in situ XAS experiments, ca. 2 mg of the sample was placed between two quartz wool plugs in the capillary reactor (outer diameter 1.5 mm, wall thickness 0.1 mm). The carburization step was performed in a mixture of H2
:
CH4 = 4
:
1 in the temperature range from 50 to 750 °C (5 mL min−1, ca. 9 °C min−1 at 50–400 °C and 2 °C min−1 at 400–750 °C) at atmospheric pressure. DRM tests were performed at 8 bar (CH4
:
CO2 He = 4
:
3
:
3) at 730 °C with the total flow rate varying in the range 1.75–3.5 mL min−1. All the flow rates are express at standard conditions of temperature and pressure. Catalyst weight/volume flow rate (W/F, in ms gMo mL−1) are reported in the text below based on the nominal Mo loading. After the DRM reaction, the spent catalysts were exposed to a pure stream of CO2 or CH4 (5 mL min−1, 730 °C). The composition of the outlet gases was monitored online by a mass spectrometer (MS). Ex situ XAS data were collected on pellets of reference materials (MoO2, α-MoO3 and Mo2C) with an optimized amount of sample mixed with cellulose. As-carburized Mo2C was handled in a N2-filled glove box to prepare specimen for XAS analysis in air-tight sealed Al bag. XAS data were processed using the Athena software (Demeter 0.9.25 software package).26 The extracted extended X-ray absorption fine structure (EXAFS) data were k3-weighted and the Fourier transform performed in the k-range 3–10 Å−1 for data collected during thermal treatment and 3–12 Å−1 for that collected at room temperature. The in situ time resolved normalized X-ray absorption near edge structure (XANES) data were analyzed using a MCR-ALS method.27 MCR-ALS analysis was performed with a MATLAB software package using the multivariate curve resolution toolbox.28 Non-negative constraints were applied for both the phase concentration and spectra profiles. The MCR-ALS analysis was complemented with principal component analysis (PCA).
![]() | (4) |
Carburization of α-MoO3-nb in a mixture of CH4 and H2 (H2:
CH4 = 4
:
1) yielded Mo2C-nb. SEM analysis of the carburized (and passivated) material revealed that the carburization process does not affect notably the nanobelt morphology (Fig. 1d). XRD analysis of Mo2C-nb revealed the presence of a single-phase that was assigned to β-Mo2C, consistent with the complete carburization of α-MoO3-nb (Fig. S2†). Similarly, the carburization of MoO3/SiO2 led to Mo2C/SiO2 (vide infra). STEM imaging of Mo2C/SiO2, (after exposure of the specimen to air during sample transfer) revealed the formation of particles of ca. 5–10 nm in diameter (Fig. 1e).
Mo2C/SiO2 shows a stable catalytic performance over 8 h of time on stream (TOS) under a CH4-rich flow (CH4:
CO2
:
N2 = 4
:
3
:
3; W/F = 5.25 ms gMo mL−1), with a methane conversion of ca. 55% (Fig. S5†). However, similarly to Mo2C-nb, the obtained H2
:
CO ratio of ca. 0.7 indicates that the RWGS and DRM reactions compete under these conditions.
![]() | ||
Fig. 2
In situ Mo K-edge XANES (contour plots of the intensity as a function of energy and time) during carburization of a) α-MoO3-nb and b) MoO3/SiO2 under H2![]() ![]() ![]() ![]() |
The XANES spectrum of MoO3/SiO2 showed a similar pre-edge feature (at ca. 20006.5 eV) and edge energy position (at 20
016.4 eV) as the reference spectrum α-MoO3, however it exhibited different white line features, notably with only one well-defined peak at 20
037.5 eV. We hypothesize that the different white line peaks are linked to the dispersion of MoO3 nanosheets onto SiO2,32 leading to a different local structure around Mo (Fig. 2b and S7†). In line with this hypothesis, the EXAFS analysis of MoO3/SiO2 showed a different local structure when compared to that of α-MoO3-nb, showing a significantly lower magnitude of the second coordination sphere pointing to a lower Mo–Mo coordination which can be linked to the high dispersion over SiO2 (Fig. S10 and Table S1†).
Interestingly, two notable changes are observed in the in situ XANES spectra: one at a low temperature between 200 and 250 °C, and one at a higher temperature at ca. 650 °C. The low temperature transition is associated with the loss of the initially more well-defined feature around 20037.5 eV with broadening of the white line and a more intense pre-edge feature that is also shifted to lower energies, ca. 20
006.2 eV (Fig. 2b and S7†). Noteworthy, the intensity of the pre-edge feature increases. The XANES spectrum of this intermediate phase does not correspond to the reference spectrum of MoO2 and neither is it similar to the previously reported spectra of bulk Mo oxides.31,33 However, it resembles the Mo K-edge spectra of isolated Mo-oxo species (Mo oxidation state between +5 and +6) on zeolites supports (i.e. ZSM-5 and H-SSZ-13).34–36 Based on this observation, we propose that the MoO3 phase in MoO3/SiO2 forms small oxo clusters and/or highly dispersed Mo oxo sites on the SiO2 surface when the temperature reaches 200–250 °C. This is in line with the EXAFS data indicating a sudden change in the second coordination sphere at ca. 200 °C (Fig. S11†). The high temperature transition is associated with the disappearance of the well-defined pre-edge feature and a shift of the Mo K-edge energy from 20
015.4 eV to ca. 20
000.0 eV, indicating the reduction and carburization of Mo. The XANES spectrum at 750 °C corresponds to the reference spectrum of β-Mo2C, confirming the complete carburization of MoO3/SiO2 to Mo2C/SiO2. Quantification of the phase evolution will be discussed based on MCR-ALS analysis of the XANES data (vide infra).
The XANES changes during DRM of Mo2C/SiO2 were also evaluated (W/F = 0.8 ms gMo mL−1). In contrast to the rapid oxidation of Mo2C-nb, a slower Mo oxidation was observed for Mo2C/SiO2, identified by a gradual shift of the Mo K-edge to higher energies (Fig. 3b). The slower oxidation rate is consistent with the higher stability of Mo2C/SiO2 under DRM conditions. After ca. 60 min TOS, a pre-edge feature appears in the XANES spectra of this catalyst. Interestingly, consistent with the evolution of the XANES data, the MS data reveals an increase of the H2O signal starting from TOS = 60 min. Therefore, this indicates that the onset of the RWGS reaction coincides with a phase transformation of the catalyst (for the nature and quantitative analysis of the oxidized species vide infra MCR-ALS analysis).11 Additional exposure of the catalyst to a stream of pure CO2 shifts the Mo K-edge to higher energies, indicating a further oxidation of the material (Fig. S13†). Additionally, the intensity of the pre-edge feature increases significantly and the XANES spectrum at the end of the reaction corresponds clearly to the intermediate phase observed during carburization. Interestingly, contrary to the deactivated Mo2C-nb catalyst, an oxidized Mo2C/SiO2 can be partially reduced (and likely, recarburized) under a pure stream of CH4 (Fig. S13†). This observation indicates that a dynamic oxidation-recarburization process may in principle take place under DRM conditions,7 yet it has been reported that addition of transition metals such as Ni is required to increase the rate of recarburization, in order to match the oxidation rate.12,13
![]() | ||
Fig. 4 Phase evolution of a) α-MoO3-nb and b) MoO3/SiO2 during the carburization process and DRM according to MCR-ALS of the XANES data. |
Turning to the Mo2C/SiO2 catalyst, only three distinct components are involved in the carburization process (Fig. S16†). Those three spectra (,
and
) extracted using MCR-ALS analysis are presented in Fig. S17.† Component
is assigned to the initial MoO3 phase dispersed onto SiO2 (vide supra). The intermediate
component shows a unique white line shape which resembles the Mo K-edge spectra of previously reported molybdenum single sites (MOx/SiO2).34–36 The edge energy (20
015.4 eV) indicates Mo in an oxidation state between +5 and +6. Component
shows similar features to that of the β-Mo2C reference and thus is assigned to the carburized Mo (Mo2C) supported on SiO2. Fig. 4b plots the changes of the fraction of the
,
and
components during the carburization process. The intermediate
component starts to appear at ca. 120 °C and reaches its maximum level of ca. 0.8 at 400 °C. The fraction of the component
, corresponding to the initial state of the material, reduces rapidly until it disappears at ca. 425 °C. Interestingly, the MCR analysis indicates that the molybdenum carbide phase (
) appears already at ca. 200 °C and reaches a fraction of ca. 0.2 at 425 °C during the initial stage of carburization. The concentration of
starts to increase slowly replacing
until it reaches a weight fraction of 0.3 at ca. 600 °C. Afterwards, a rapid carburization of
to
occurs yielding a pure
phase at ca. 680 °C (Fig. 4b).
The MCR analysis indicates that the changes in the XANES spectra under DRM conditions can be described with the same components that have been identified in the carburization process (Fig. 4a and b). In particular, for Mo2C-nb the evolution of the in situ XANES data during DRM conditions can be fully described with the components MoO2 (C3) and Mo2C (C4). The MCR analysis indicates that under DRM conditions Mo2C is rapidly oxidized to MoO2 yielding a fraction of MoO2 of >0.6 after 120 min TOS (Fig. 4a).
The evolution of the in situ XANES spectra of Mo2C/SiO2 during DRM can be fully described by the components and
(Mo2C) (Fig. 4b). However, the MCR analysis indicates no changes in the phase composition during 45 min TOS revealing that only
(Mo2C) is present until up to 45 min under DRM conditions. Subsequently, a slow oxidation of Mo2C into
(MOx/SiO2) is revealed, reaching asymptotic fractions of 0.75 for Mo2C and 0.25 for
after ca. 120 min TOS (Fig. 4b). To summarize, our studies reveal that the degree of oxidation of the carbidic phase in Mo2C/SiO2 is lower than for Mo2C-nb under DRM conditions. It should be highlighted that although in the in situ XANES setup requires a different W/F ratio as compared to the laboratory reactor system (i.e. parameters of the synchrotron experimental setup yield a ca. one order of magnitude lower W/F ratio as compared to the laboratory reactor), and this can affect the degree of catalysts oxidation, both laboratory reactor and synchrotron capillary experiments provide qualitatively similar results. The observed MoO2 phase via ex situ XRD analysis of the spent catalyst (Fig. S4†) correlates well with the in situ XANES analysis.
It has been previously reported that the carburization of MoO3/C to Mo2C/C and the evolution of Mo2C/C under DRM conditions involves an oxycarbide intermediate.11 In contrast, we do not observe oxycarbide intermediates in the current study, despite similar experimental conditions and the same starting Mo precursor (d-MoO3). This may be explained by the support effect, that is, the carbon support in MoO3/C and Mo2C/C enables carburization and DRM pathways that involve a Mo2CxOy intermediate that seems to be hindered for Mo2C-nb and Mo2C/SiO2 of this work. That being said, we cannot exclude the formation of surface oxycarbides that can bring about the RWGS reactivity of in situ oxidized Mo2C-nb and Mo2C/SiO2.
Footnotes |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d2cy00729k |
‡ Current address: Laboratory for Bioenergy and Catalysis, Paul Scherrer Institute (PSI), 5232 Villigen PSI, Switzerland. |
§ Current address: Department of Chemistry, Technical University Munich, Lichtenbergstrasse 4, 85748 Garching, Munich, Germany. |
This journal is © The Royal Society of Chemistry 2022 |