On the morphology of MoS₂ slabs on MoS₂/Al₂O₃ catalysts: the influence of Mo loading

Abstract

Two-dimensional MoS₂ is an important material with diverse catalytic applications. The edge sites exposed by MoS₂ are of great importance to its catalytic performance since the catalytic reactions generally occur on the edge sites rather than on the basal planes. In this work, low temperature (100 K) CO adsorption followed by IR spectroscopy (IR/CO) was used to in situ probe the edge sites of the MoS₂ phase on MoS₂/Al₂O₃ catalysts. It is found that the morphology of the MoS₂ phase on an Al₂O₃ support is a truncated triangle exposing mainly Mo-terminated edges (M-edge). However, the proportion of sulfur-terminated edges (S-edge) increases with increasing the Mo loading, resulting in a more heavily truncated triangle morphology of MoS₂. The change of MoS₂ morphology with Mo loading can be explained by the modification of MoS₂–Al₂O₃ interactions, indicating the importance of phase-support interactions in supported MoS₂ catalysts.

---

1 Introduction

The two-dimensional (2D) molybdenum sulfide (MoS₂) nanostructure (slab) is one of the most important materials with diverse applications. While supported MoS₂ is used in the traditional hydrodesulfurization (HDS) process to remove sulfur from crude oil feedstock,¹ an exciting application of MoS₂ was recently found in the hydrogen evolution reaction (HER).² Since the catalytic reactions usually proceed on the edge sites of MoS₂ slabs and the basal planes are generally catalytically inert,³ the type and number of edges exposed under different synthesis conditions are of crucial importance to MoS₂ catalytic performance.

In its perfect crystallographic (100) plane, MoS₂ exposes principally two types of edges: the Mo terminated edge (M-edge) and the sulfur terminated edge (S-edge).⁴ Recently, it was found that the relative ratio of M-edge and S-edge, i.e. the morphology of MoS₂, is strongly influenced by MoS₂-support interactions. With STM (scanning tunneling microscopy), Walton and co-workers observed that MoS₂ adopts a triangle shape exposing only one edge on Au support, whereas it presents a hexagon shape with both M-edge and S-edge on graphite and TiO₂.⁵ On the support of Al₂O₃, we recently found that the MoS₂ morphology can be stepwise tuned from a slightly truncated triangle to a heavily truncated triangle by addition of chelating agent during catalyst preparation.⁶ In addition, DFT simulations show that the ratio of anchored and free edge sites (including the relative proportion of M-edge and S-edge sites) differs with MoS₂-support interactions.⁷

As an important parameter in supported MoS₂ catalysts, the effect of Mo loading on the catalyst structure and activity was intensively investigated.^8–17 While several studies suggest that the MoS₂–Al₂O₃ interactions may differ with Mo loading,^13,18,19 the correlation of Mo loading and MoS₂ morphology is never considered. Therefore, in this work the effect of Mo loading on the morphology of MoS₂ phase on MoS₂/Al₂O₃ catalysts is investigated. For this purpose, a series of oxidic Mo/Al₂O₃ precursors with different Mo loading was prepared by conventional impregnation method. The MoS₂/Al₂O₃ catalysts were obtained by sulfiding the Mo/Al₂O₃ precursors with 10% H₂S/H₂ at atmospheric pressure or 4.0 MPa. The MoS₂ morphologies on these catalysts were determined with IR spectroscopy using CO as probe molecule that adsorbs on MoS₂ edge sites at low temperature (100 K).

2 Experimental

2.1. Catalyst preparation

A set of oxidic Mo/Al₂O₃ catalyst precursors with variable Mo loading (6%, 9%, and 12%, wt Mo) was prepared by the one-step pore volume impregnation method. Firstly, variable amount of ammonium heptamolybdate tetrahydrate ((NH₄)₆Mo₇O₂₄·4H₂O, Acros Organics) for the designed Mo loading was dissolved in 3.30 mL deionized water. Secondly, 3.00 gram of pretreated γ-Al₂O₃ support (Sasol, specific surface area of 252 m² g⁻¹ and pore volume of 0.84 mL g⁻¹, pre-calcined in air at 723 K for 2 hours) was added into the solutions and strongly shaken for 2 hours. Finally, the matured slurry was dried at 383 K for 3 hours and then calcined at 773 K for 3 hours under air.

2.2. Infrared (IR) spectroscopy characterization

IR characterization was performed on a newly designed setup called CellEx. The CellEx consists of three parts: (i) a stainless steel reactor for catalyst sulfidation and other treatment under different temperatures, pressures and gas phases; (ii) an IR cell equipped with a spectrometer for spectroscopic characterization; and (iii) a transfer connection for transferring the sample from reactor into IR cell under inert gas. With the CellEx, the catalysts can be treated under different conditions and sequentially in situ characterized by IR spectroscopy without any air pollution. More details about CellEx were described in ref. 20.

2.2.1. Catalyst sulfidation. Catalyst sulfidation were performed in the stainless steel reactor of CellEx. Catalyst sample was firstly grounded and pressed into self-supporting pellet. The pellet was introduced into the reactor which was then evacuated to 1.33 Pa to remove the air. After that the pellet was sulfided at 0.1 MPa or 4.0 MPa with the following procedure. (i) 0.1 MPa sulfidation: firstly, a gas mixture of 0.1 MPa 30 mL min⁻¹ 10% H₂S/H₂ (Air Liquide France, H₂S: 9.97% ± 0.3%, v/v) was introduced into the reaction. Then the reactor was heated with a rate of 3 K min⁻¹ to 623 K and maintained for 2 hours. Sequentially, the reactor was flushed with Ar at 623 K and then cooled down to room temperature under Ar. Finally, the sulfided pellet was transferred under Ar to the IR cell for IR characterization. (ii) 4.0 MPa sulfidation: the pressure of reactor was firstly increased to 4.0 MPa with 10% H₂S/H₂ mixture. Then the reactor was heated with a rate of 3 K min⁻¹ to 623 K and maintained for 2 hours. During this period, the flow rate of H₂S/H₂ was fixed at 30 mL min⁻¹. After that, the reactor pressure was decreased to atmospheric pressure. Sequentially, the reactor was flushed with Ar at 623 K and then cooled down to room temperature under Ar. Finally, the sulfided pellet was transferred under Ar to the IR cell for IR characterization.

2.2.2. Low-temperature CO adsorption followed by IR spectroscopy characterization (IR/CO). In the IR cell, the sulfided pellet was firstly heated at 6 K min⁻¹ up to 623 K and kept for 1 hour under evacuation. The final pressure in the IR cell after evacuation should reach 10⁻³ Pa. After that, the pellet was cooled down to 100 K for CO adsorption. CO adsorption was performed by introducing small calibrated doses of CO (Air Liquide France, 99.9973%, purified by trapping in liquid nitrogen before use) at different pressures (0.03–1.20 μmol of CO) and finally with 133 Pa CO at equilibrium in the IR cell. IR spectra of adsorbed CO were recorded using a Thermo Scientific Nicolet FT-spectrometer equipped with a MCT detector with 256 scans. Note that the graphical resolution is 0.5 cm⁻¹. For comparison, all the spectra presented were normalized to a sulfided catalyst pellet of 5 mg cm⁻².

The obtained IR spectra were further decomposed with Peakfit V4.12 using “Autofit peak II—Second derivative Methods”. Then the concentration of each type of MoS₂ edge sites was determined using the molar extinction coefficient of CO adsorbed on M-edge (ε_M-edge) and S-edge (ε_S-edge) previous measured (ε_M-edge and ε_S-edge are 20 ± 3 μmol⁻¹ cm and 35 ± 9 μmol⁻¹ cm, respectively⁶). Details for spectral decomposition, ε_M-edge and ε_S-edge measurement, and edge site concentration calculation were described in ref. 6.

2.3. TEM characterization

Transmission electron microscopy (TEM) was carried out on a JEOL 2010 FEG operated at 200 kV. The catalyst precursors were firstly sulfided with the same procedure as that used in IR/CO characterization. To limit detrimental exposure to air, the sulfided catalysts were unloaded from the sulfidation reactor under argon flow into ethanol. A few drops of a suspension of catalyst were put on a copper grid. Slab length and stacking degree distributions of sulfide slabs were determined manually by measuring at least 300 slabs per sample from the TEM images. All the TEM images were recorded at the same magnification and digitized using a 2k × 2k CCD camera. The image treatment was performed using the commercial software from GATAN (DIGITALMICROGRAPH).

3 Results

3.1. Catalyst characterization

Both the oxidic Mo/Al₂O₃ precursor and the sulfide MoS₂/Al₂O₃ catalyst were widely characterized in the literature.^21–25 For the oxidic precursors, it is generally agreed that the isolated and tetrahedral coordinated molybdenum oxide species present at low Mo loadings while the hepta- and octamolybdates species increase with increasing Mo loading.^23,24 The as-prepared oxidic Mo/Al₂O₃ samples in this work were characterized by Raman spectroscopy, and such conclusion is confirmed (ESI, Fig. S1†). Due to the good dispersion of MoS₂ phase on Al₂O₃ support, the X-ray diffraction (XRD) patterns of MoS₂/Al₂O₃ samples shows mainly the characteristics of Al₂O₃ without or with only weak signals according to crystalline MoS₂ phases up to the MoO₃ loading around 15 wt%.^26–28

The obtained 9 wt% MoS₂/Al₂O₃ catalysts were also characterized by the high resolution TEM (Transmission Electron Microscopy). The MoS₂ phase is imaged as tiny slabs (Fig. 1) with an average length around 2 nm and average stacking around 1.5 layers (Fig. 2). On calcined catalyst, the sulfidation pressure has slight effect on the average slab length and the average stacking number (Fig. 2 and ref. 29). Because of the contrast disturbance from alumina support, the HRTEM provides neither the morphology of MoS₂ clusters nor the type of edges that MoS₂ exposes (Fig. 1 and ref. 25).

Fig. 1 TEM images of the MoS₂ slabs on MoS₂/Al₂O₃ catalyst sulfided at 623 K with 0.1 MPa (A) or 4.0 MPa (B) 10% H₂S/H₂. The Mo loading is 9 wt% on this catalyst.

Fig. 2 TEM analysis of the average slab length and stacking of MoS₂ on MoS₂/Al₂O₃ catalyst sulfided at 623 K with 0.1 MPa or 4.0 MPa 10% H₂S/H₂. The Mo loading is 9 wt% on this catalyst. The average slab length is 1.9 nm for the 0.1 MPa sulfided catalyst, and 2.2 nm for the 4.0 MPa sulfided one. The average slab stackings are respectively 1.56 (0.1 MPa) and 1.67 (4.0 MPa).

3.2. IR spectra of low-temperature CO adsorption on MoS₂/Al₂O₃ catalyst

In order to investigate the MoS₂ morphology on MoS₂/Al₂O₃ catalyst, low temperature (100 K) CO adsorption followed by IR spectroscopy (IR/CO) was employed to probe the MoS₂ edge sites. For comparison, the IR spectrum of CO adsorption on pure Al₂O₃ support sulfided by the same procedure is also provided. As shown in Fig. 3, low temperature CO adsorption on pure Al₂O₃ support leads to two ν(CO) bands located at 2189 and 2156 cm⁻¹, which are assigned respectively to CO adsorption on Al³⁺ sites and OH groups on Al₂O₃ support.³⁰ When MoS₂ phase is introduced onto Al₂O₃ surface, distinct IR/CO bands in the region of 2125–1950 cm⁻¹ were observed. The ν(CO) band located at 2111 cm⁻¹, as well as the shoulder at 2064 cm⁻¹, correspond to CO adsorption on the edge sites of MoS₂ phase. Combining experimental IR/CO data with theoretical DFT calculations, Travert and co-workers^31,32 assigned the former band (∼2111 cm⁻¹) to CO adsorption on the Mo-terminated edge (M-edge) of MoS₂ slabs, and the later one (∼2064 cm⁻¹) to CO adsorption on the sulfur-terminated edge (S-edge) of MoS₂ phase. With increasing Mo loading, the intensity of ν(CO) bands on Al³⁺ sites and OH groups stepwise decrease, indicating that more Al₂O₃ surface is covered by MoS₂ phase, which is in good agreement with the augmentation of Mo loading. Meanwhile, the intensity of ν(CO) bands on both M-edge and S-edge gradually increases, indicating more MoS₂ edge sites are formed with the increasing Mo loading. Nevertheless, it should be underlined that these two bands do not increase proportionately. The ν(CO) shoulder band on S-edge becomes more evident with Mo loading.

Fig. 3 IR spectra of CO adsorption (133 Pa CO at equilibrium, 100 K) on MoS₂/Al₂O₃ prepared with different Mo loading. The MoS₂/Al₂O₃ catalysts were obtained by sulfiding the oxidic Mo/Al₂O₃ precursors at 623 K and 0.1 MPa with 10% H₂S/H₂.

In order to go further in the parameters that affects the relative change of M-edge and S-edge with Mo loading, the set of oxidic Mo/Al₂O₃ precursors were further sulfided with 10% H₂S/H₂ at high pressure (4.0 MPa) and then in situ characterized by IR/CO. As shown in Fig. 4, no new IR band appears on the high pressure sulfided MoS₂/Al₂O₃ catalysts. Nevertheless, the ν(CO) band on S-edge (∼2064 cm⁻¹) is more clearly observed after high pressure sulfidation than after sulfidation at atmospheric pressure. With increasing Mo loading, this band is an ill-defined shoulder on 6 wt% MoS₂/Al₂O₃ catalyst and becomes a well-defined band on 12 wt% MoS₂/Al₂O₃ samples, indicating an increasing of S-edge proportion on the MoS₂ phase.

Fig. 4 IR spectra of CO adsorption (133 Pa CO at equilibrium, 100 K) on MoS₂/Al₂O₃ prepared with different Mo loading. The MoS₂/Al₂O₃ catalysts were obtained by sulfiding the oxidic Mo/Al₂O₃ precursors at 623 K and 4.0 MPa with 10% H₂S/H₂.

Fig. 5 compares the IR/CO spectra on 9 wt% MoS₂/Al₂O₃ catalysts sulfided at atmospheric (0.1 MPa) and high pressure (4.0 MPa). As can be seen, the CO adsorption intensity on MoS₂ phase is strongly modified by sulfidation pressure. With elevating the sulfidation pressure, the CO adsorption intensity on M-edge is decreased, whereas the CO band on S-edge becomes more pronounced. Besides, the CO adsorption intensity on OH groups is also significantly enhanced after high-pressure sulfidation.

Fig. 5 Comparison of IR spectra of CO adsorption (133 Pa at equilibrium, 100 K) on MoS₂/Al₂O₃ sulfided at 623 K with 0.1 MPa (thin line) or 4.0 MPa (bold line) 10% H₂S/H₂. The Mo loading is 9 wt% on this catalyst.

To depict the MoS₂ morphologies on MoS₂/Al₂O₃ catalysts, the IR/CO spectra in Fig. 3 and 4 were further decomposed with the software Peakfit V4.12 and the concentration of M-edge and S-edge sites detected by low-temperature CO adsorption was calculated. The details for spectral decomposition and sites calculation were reported previously.⁶ As shown in Fig. 6, the ratio of S-edge/M-edge on the series of MoS₂/Al₂O₃ catalysts are steadily increased with Mo loading. Assuming that the ratio of M-edge and S-edge on a single MoS₂ slab is in line with the overall ratio of each edge detected by CO adsorption, the data in Fig. 6 suggests that with increasing Mo loading, the morphology of MoS₂ slabs changes from a slightly truncated triangle exposing mainly M-edge to a more heavily truncated triangle exposing relatively more S-edge. For the same Mo loading, the S-edge/M-edge ratios on the 4.0 MPa sulfided MoS₂/Al₂O₃ catalysts are always higher than the ones on the 0.1 MPa sulfided samples, indicating that MoS₂ morphology is probably also influenced by sulfidation pressure. Nevertheless, the S-edge/M-edge ratios, even on the high pressure sulfided samples, are much smaller than 1, showing that the MoS₂ phase with typical Mo loadings in HDS catalysts mainly exposes M-edge, which is consistent with the DFT results that the M-edge is energetically more stable than S-edge under HDS conditions.³³

Fig. 6 The ratios of S-edge/M-edge on MoS₂/Al₂O₃ catalysts as a function of Mo loadings.

4 Discussion

4.1. On the assignment of CO adsorption bands on MoS₂ phase

Although the IR/CO data in this work clearly show that the relative intensity of the two CO adsorption bands on MoS₂ phase varies with Mo loading, the conclusion that MoS₂ morphology is changed relies firstly on the assignment of the two ν(CO) bands. Besides the M-edge/S-edge assignment used above, another competing assignment in the literature is the edge/corner assignment. By comparison with CO adsorption on metal particles, Muller and co-workers³⁴ attributed the two ν(CO) bands at 2111 and 2064 cm⁻¹ to respectively edge and corner sites of MoS₂ slabs. This assignment was used in several studies without further validation.^35,36 However, this assignment is contradicted with the experimental results present in this paper. Indeed, whatever the morphology of MoS₂ slabs is (triangle or hexagon), the corner/edge site ratio should decreases with increasing MoS₂ slab size. Since the MoS₂ slab size increases with Mo loading,^8,19,37,38 the 2064/2111 ν(CO) band ratio should decreases with increasing Mo loading. However, as shown in Fig. 6, the 2064/2111 ν(CO) band ratio steadily increases with increasing Mo loading. Therefore, the assignment of these two bands to CO adsorption on edge and corner sites of MoS₂ slabs can be discarded.

Using the density functional method, Zeng and co-workers³⁹ calculated the CO adsorption on triangular MoS_x clusters. Accordingly, they assigned the ν(CO) band at ∼2111 cm⁻¹ to symmetric coupled stretching of double-CO adsorption on corner site (frequency calculated is 2102 cm⁻¹), and the band at ∼2064 cm⁻¹ to either mono-adsorption stretching or asymmetric coupled stretching on the corner site. Firstly, this assignment of double CO adsorption is contradicted with the co-adsorption results of ¹²CO/¹³CO mixtures on MoS₂/Al₂O₃ catalyst.^40,41 Secondly, the assignment of these two ν(CO) bands to CO adsorption on corner sites makes it impossible the opposite intensity change of this two bands. However, our recent experimental data on citric acid effect⁶ show a decrease of the ν(CO) band at ∼2111 cm⁻¹ with an increase of the ν(CO) band at ∼2064 cm⁻¹. Therefore, this assignment can also be discarded.

The assignment of CO adsorption on other species, such as Mo carbide, can also be excluded, since the catalysts used in the present work were prepared without any carbon additives and calcined at 773 K under air before sulfidation. Another assignment to be considered for these two CO bands is the effect of MoS₂ stacking. However, this assignment is unlikely since the relative intensity of the two bands varies significantly while the average stacking of MoS₂ slabs stays comparable (Fig. 2 and ref. 6). In the same way, the ν(CO) band at ∼2064 cm⁻¹ cannot be attributed to the nanometer-sized entities as observed by Kooyman and co-workers,²⁵ since these nanoparticles, which predominantly occur under mild sulfidation, cannot be the major species on well sulfided MoS₂/Al₂O₃ catalyst.

In summary, it is reasonable to assign the two CO adsorption bands at ∼2111 and ∼2064 cm⁻¹ to respectively M-edge and S-edge of MoS₂ slabs. The recent work of Labruyere with parallel studies of IR/CO and DFT calculations on the effect of sulfidation temperature confirms again this assignment.⁴² Consequently, with this assignment, the IR/CO data presented in this work clearly show that the morphology of MoS₂ slab is changed with Mo loading on supported MoS₂/Al₂O₃ catalysts.

4.2. The influence of Mo loading on MoS₂ morphology

The question of the origin of the MoS₂ morphology change with Mo loading arises. Many factors allowed us to interpret this effect by a change in MoS₂–Al₂O₃ interactions. The MoS₂ phase interacts with Al₂O₃ support by the so-called Mo–O–Al linkages, which are located at the edges of MoS₂ slabs rather than at the bulk sites.^7,43 Therefore, the concentration of Mo–O–Al linkages (normalized by the total number of Mo in MoS₂) will decreases with increasing MoS₂ size. Since it is well established in the literature that the MoS₂ size increases with Mo loading,^8,19,37,38 the concentration of Mo–O–Al linkages decreases with increasing Mo loading. Moreover, The Mo–O–Al linkages are formed by the reaction of Mo atoms with hydroxyl groups on Al₂O₃ surface.⁴⁴ Since the hydroxyl groups on Al₂O₃ support have different acidity, the Mo atoms preferentially react with the most basic hydroxyl groups, and the excess Mo atoms will react with less basic or neutral OH groups to form Mo–O–Al linkages.⁴⁴ Therefore, the strength of Mo–O–Al linkages also decreases with Mo loading. All these results suggest that the MoS₂–Al₂O₃ interactions decrease with increasing Mo loading.

According to the Wulff construction,⁴⁵ the morphology of MoS₂ is determined by the relative free energies of M-edge and S-edge. These edge energies are related to the chemical potential of sulfur in sulfidation,³³ but also sensitive to other parameters such as MoS₂-support interactions.⁵ The role of MoS₂-support interactions on MoS₂ morphology can be understood as an extra energy obtained by the edges when interacting with support.⁵ We surmise that such additional energy is not the same on M-edge and S-edge. Therefore, the growth of one edge will be promoted, and thus the MoS₂ morphology is changed when changing MoS₂-support interactions.

The above explanations of MoS₂ morphology change with Mo loadings are consistent with the previous report. Our previous study on citric acid⁶ demonstrated that the reduction of MoS₂–Al₂O₃ interactions by chelating agent addition promotes the growth of S-edge, thus leading to the morphology change of MoS₂ slabs from a slightly truncated triangle exposing mainly M-edge to a hexagon with both M-edge and S-edge. In this paper, the reduction of MoS₂–Al₂O₃ interactions can also be expected with increasing Mo loadings. Accordingly, it is observed that the MoS₂ slabs on MoS₂/Al₂O₃ catalysts expose relatively more S-edge with increasing Mo loading.

4.3. The effect of high pressure sulfidation

The role of the sulfidation pressure on MoS₂ morphology is even more crucial since the effect of Mo loading on morphology is more prominent after 4.0 MPa sulfidation. The results in this work point out that MoS₂ morphology is also influenced by the sulfidation pressure since the S-edge/M-edge ratio is always greater on the high pressure sulfided samples for the same Mo loading. According to Le Chatelier's principle, the effect of sulfidation pressure can be tentatively interpreted as that the increase of the total gas phase pressure (P_H2S + P_H2) will favor the thiolysis of Mo–O–Al linkages toward the formation of Mo-SH and Al-OH,⁴⁶ and thus lowers the MoS₂–Al₂O₃ interactions. Besides, it was also observed that the CO uptake on the high pressure sulfided MoS₂ is lower than that on the atmospheric sulfided one (Fig. 5), implying that sulfidation pressure probably also changes the sulfur coverage on MoS₂ edge sites as suggested by DFT calculations.⁴⁷ The effect of sulfidation pressure on MoS₂ morphology and edge structure will be further investigated in another work.

5 Conclusion

In this work, low temperature CO adsorption followed by IR spectroscopy (IR/CO) is used to demonstrate the MoS₂ morphology change with Mo loading on real type MoS₂/Al₂O₃ catalysts. It is found that the morphology of MoS₂ on Al₂O₃ support is a truncated triangle exposing both M-edge and S-edge, and that the truncation degree, i.e., ratio of S-edge/M-edge, increases with Mo loading. The effect of Mo loading on MoS₂ morphology can be explained by the modification of MoS₂–Al₂O₃ interactions. Nevertheless, because of the strong interactions between MoS₂ and Al₂O₃ support, MoS₂ presents mainly M-edge rather than S-edge even at high Mo loading and high pressure sulfidation, indicating that M-edge is energetically much more stable than S-edge on Al₂O₃ support.

Acknowledgements

The French Ministry of Research is acknowledged for the Ph.D. grant of J. Chen. Yoann Levaque, Valérie Ruaux, and Philippe Bazin are greatly acknowledged for the technical support on IR experiment. Guillaume CLET is greatly acknowledged for the Raman characterization.

References

1. H. Topsoe, B. S. Clausen and F. E. Massoth, Hydrotreating catalysis, Springer, Berlin-Heidelberg, 1996.
2. I. Song, C. Park and H. C. Choi, RSC Adv., 2015, 5, 7495–7514.
3. T. F. Jaramillo, K. P. Jorgensen, J. Bonde, J. H. Nielsen, S. Horch and I. Chorkendorff, Science, 2007, 317, 100–102.
4. J. V. Lauritsen, J. Kibsgaard, S. Helveg, H. Topsoe, B. S. Clausen, E. Laegsgaard and F. Besenbacher, Nat. Nanotechnol., 2007, 2, 53–58.
5. A. S. Walton, J. V. Lauritsen, H. Topsoe and F. Besenbacher, J. Catal., 2013, 308, 306–318.
6. J. Chen, F. Maugé, J. El Fallah and L. Oliviero, J. Catal., 2014, 320, 170–179.
7. D. Costa, C. Arrouvel, M. Breysse, H. Toulhoat and P. Raybaud, J. Catal., 2007, 246, 325–343.
8. L. Jalowiecki-Duhamel, J. Grimblot and J. P. Bonnelle, J. Catal., 1991, 129, 511–518.
9. E. Etienne, E. Ponthieu, E. Payen and J. Grimblot, J. Non-Cryst. Solids, 1992, 147–148, 764–768.
10. Y. Yokoyama, N. Ishikawa, K. Nakanishi, K. Satoh, A. Nishijima, H. Shimada, N. Matsubayashi and M. Nomura, Catal. Today, 1996, 29, 261–266.
11. R. Iwamoto and J. Grimblot, Stud. Surf. Sci. Catal., 1997, 106, 195–210.
12. R. Iwamoto and J. Grimblot, Stud. Surf. Sci. Catal., 1999, 121, 289–294.
13. E. J. M. Hensen, V. H. J. de Beer, J. A. R. van Veen and R. A. van Santen, Catal. Lett., 2002, 84, 59–67.
14. A. Ishihara, F. Dumeignil, D. H. Wang, X. G. Li, H. Arakawa, E. W. Qian, S. Inoue, A. Muto and T. Kabe, Appl. Catal., A, 2005, 292, 50–60.
15. P. Blanchard, C. Lamonier, A. Griboval and E. Payen, Appl. Catal., A, 2007, 322, 33–45.
16. A. Ishihara, J. Jpn. Pet. Inst., 2008, 51, 73–82.
17. J. Mazurelle, C. Lamonier, C. Lancelot, E. Payen, C. Pichon and D. Guillaume, Catal. Today, 2008, 130, 41–49.
18. P. Arnoldy, J. A. M. Vandenheijkant, G. D. Debok and J. A. Moulijn, J. Catal., 1985, 92, 35–55.
19. H. Shimada, N. Matsubayashi, T. Sato, Y. Yoshimura, M. Imamura, T. Kameoka and A. Nishijima, Catal. Lett., 1993, 20, 81–86.
20. L. Oliviero, L. Mariey, M.-A. Lelias, S. Aiello, J. van Gestel and F. Mauge, Catal. Lett., 2010, 135, 62–67.
21. G. M. Dhar, B. N. Srinivas, M. S. Rana, M. Kumar and S. K. Maity, Catal. Today, 2003, 86, 45–60.
22. G. Mestl and T. K. K. Srinivasan, Catal. Rev.: Sci. Eng., 1998, 40, 451–570.
23. H. C. Hu, I. E. Wachs and S. R. Bare, J. Phys. Chem., 1995, 99, 10897–10910.
24. K. V. R. Chary, K. R. Reddy, G. Kishan, J. W. Niemantsverdriet and G. Mestl, J. Catal., 2004, 226, 283–291.
25. P. J. Kooyman, E. J. M. Hensen, A. M. de Jong, J. W. Niemantsverdriet and J. A. R. van Veen, Catal. Lett., 2001, 74, 49–53.
26. B. Liu, Y. M. Chai, Y. P. Li, A. J. Wang, Y. Q. Liu and C. G. Liu, Appl. Catal., A, 2014, 471, 70–79.
27. W. B. Chen, F. Mauge, J. van Gestel, H. Nie, D. D. Li and X. Y. Long, J. Catal., 2013, 304, 47–62.
28. D. Liu, Z. Li, Q. Sun, X. Kong, A. Zhao and Z. Wang, Fuel, 2012, 92, 77–83.
29. A. I. Dugulan, E. J. M. Hensen and J. A. R. van Veen, Catal. Today, 2008, 130, 126–134.
30. K. I. Hadjiivanov and G. N. Vayssilov, Adv. Catal., 2002, 47, 307–511.
31. A. Travert, C. Dujardin, F. Mauge, S. Cristol, J. F. Paul, E. Payen and D. Bougeard, Catal. Today, 2001, 70, 255–269.
32. A. Travert, C. Dujardin, F. Mauge, E. Veilly, S. Cristol, J. F. Paul and E. Payen, J. Phys. Chem. B, 2006, 110, 1261–1270.
33. H. Schweiger, P. Raybaud, G. Kresse and H. Toulhoat, J. Catal., 2002, 207, 76–87.
34. B. Muller, A. D. Vanlangeveld, J. A. Moulijn and H. Knozinger, J. Phys. Chem., 1993, 97, 9028–9033.
35. L. P. A. F. Elst, S. Eijsbouts, A. D. van Langeveld and J. A. Moulijn, J. Catal., 2000, 196, 95–103.
36. F. Cesano, S. Bertarione, A. Piovano, G. Agostini, M. M. Rahman, E. Groppo, F. Bonino, D. Scarano, C. Lamberti, S. Bordiga, L. Montanari, L. Bonoldi, R. Millini and A. Zecchina, Catal. Sci. Technol., 2011, 1, 123–136.
37. Y. P. Li, A. T. Li, F. F. Li, D. P. Liu, Y. M. Chai and C. G. Liu, J. Catal., 2014, 317, 240–252.
38. E. Payen, S. Kasztelan, S. Houssenbay, R. Szymanski and J. Grimblot, J. Phys. Chem., 1989, 93, 6501–6506.
39. T. Zeng, X. D. Wen, Y. W. Li and H. J. Jiao, J. Phys. Chem. B, 2005, 109, 13704–13710.
40. M. A. Lelias, A. Travert, J. van Gestel and F. Mauge, J. Phys. Chem. B, 2006, 110, 14001–14003.
41. X.-D. Wen, T. Zeng and H. Jiao, J. Phys. Chem. B, 2006, 110, 14004–14005.
42. V. Labruyere, Ph.D thesis, Universite of Caen, 2014.
43. C. Arrouvel, M. Breysse, H. Toulhoat and P. Raybaud, J. Catal., 2005, 232, 161–178.
44. N. Y. Topsoe and H. Topsoe, J. Catal., 1993, 139, 631–640.
45. G. Wulff, Z. Kristallogr. Mineral., 1901, 34, 449–530.
46. Y. V. Joshi, P. Ghosh, M. Daage and W. N. Delgass, J. Catal., 2008, 257, 71–80.
47. P. Y. Prodhomme, P. Raybaud and H. Toulhoat, J. Catal., 2011, 280, 178–195.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra14768a

---