Influence of Re–M interactions in Re–M/C bimetallic catalysts prepared by a microwave-assisted thermolytic method on aqueous-phase hydrogenation of succinic acid

Xin Di a, Chuang Li a, Gwendoline Lafaye *b, Catherine Especel b, Florence Epron b and Changhai Liang *a
aLaboratory of Advanced Materials and Catalytic Engineering, Dalian University of Technology, Dalian 116024, P.R. China. E-mail: changhai@dlut.edu.cn
bInstitut de Chimie, des Milieux & Matériaux (IC2MP), Université de Poitiers, Poitiers 86073, France. E-mail: gwendoline.lafaye@univ-poitiers.fr

Received 23rd May 2017 , Accepted 15th September 2017

First published on 18th September 2017


Carbon-supported Re–M (M = Pt and Rh) bimetallic catalysts with controlled size and composition were synthesized by using a microwave-assisted thermolytic method and evaluated in the aqueous phase hydrogenation of succinic acid. The Re–M interaction contributes to the inhibition of aggregation of particles and to the improvement in the catalytic activity for succinic acid hydrogenation through decreasing the activation energy. The Re–M interaction favors the ring opening of γ-butyrolactone, an intermediate product, to 1,4-butanediol instead of the hydrogenation and dehydration to tetrahydrofuran observed over a Re/C catalyst. The kinetic study proves that the Re–M interaction can increase the relative formation rate of 1,4-butanediol more than that of tetrahydrofuran, while the strength of the Re–M interaction has a limited influence on the product selectivity. It was shown that the Re–Rh interaction can reduce the direct hydrogenolysis of succinic acid, but it cannot avoid the hydrogenolysis of 1,4-butanediol, thus limiting the selectivity to this product. According to the kinetic mechanism, ring opening of γ-butyrolactone is favored at low temperature while direct hydrogenation to tetrahydrofuran is favored at high temperature.


1. Introduction

Platform chemicals derived from biomass are promising alternatives to unsustainable fossil resources for the production of high value added chemicals. Succinic acid (SA) belongs to a new class of bio-derived building-block chemicals, which can replace the current maleic anhydride C4 platform according to the data released by the U.S. Department of Energy. The market potential for products based on succinic acid is estimated to be around 250[thin space (1/6-em)]000 t per year.1 Furthermore, the functionalization step is not required for biomass platform molecules which are usually involved in a productive process when using petroleum feedstock as raw materials.2 The biomass platform molecules can offer numerous synthetic pathways for the production of a large variety of useful chemicals and thus show high flexibility. The derivatives of succinic acid such as γ-butyrolactone (GBL), tetrahydrofuran (THF), 1,4-butanediol (BDO) and others, are widely used as intermediates and in the synthesis of polymers.3–5 Furthermore, research on the direct transformation of succinic acid in aqueous solution will contribute to decreasing the high cost of purifying succinic acid from organic reaction media and increasing its competitiveness with the conventional process using butane as the raw material.6

In addition to the application potential cited above, SA being a dicarboxylic acid is an ideal model molecule for studying the hydrogenation of carboxylic acids. Research on the hydrogenation of SA experienced a shift from industrial development to theoretical study less than twenty years ago.7–9 In recent years it has witnessed great breakthroughs in developing novel catalytic systems and materials for the hydrogenation of SA.10–23 While both mono and bimetallic catalysts are active for SA hydrogenation, the bimetallic ones are of particular interest because they frequently exhibit superior activity and their selectivity can be changed as a result of the synergistic effect between the two metals.11,16,17,19–21,23 The intermediate product GBL can be transformed into BDO or THF via different catalytic processes, which play a decisive role in the product selectivity for SA hydrogenation. However, research on how metallic interactions will affect the choice of the catalytic process and the catalytic activity is still limited, although some works on the influence of the chemical state and location of the Re–Pd bimetallic catalyst were reported.19,21 Considering this, kinetic studies contribute to the comprehension of the reaction mechanism and are a great way to explain the influence of metallic interactions. Besides, industrial applications can also benefit from such studies. Previous investigations on the hydrogenation of carboxylic acids explored the catalytic activity of Re based catalysts and important promotion effects by addition of a noble metal were observed for the Re component. However, research on the hydrogenation of SA over supported Re based catalysts is not sufficient and systematic, because most researchers have focused their attention on the study of novel reaction system and modification of original catalysts. As a result, there is still a lot of space for further exploring traditional catalytic components such as Re, Pt and Rh. In addition to the strong influence of the choice of the active components, the preparation methods are reported to greatly affect the performance of the catalysts for the transformation of SA. Recently, we have reported the preparation of Re based catalysts by a microwave-assisted thermolytic (MAT) method, which is generally more convenient than some traditional methods.22,23 The main advantages of this method are the reduction of the preparation time and making use of the favourable heat effect. Meanwhile, this method contributes to the control of the textural properties of catalysts and to the achievement of highly dispersed catalysts.24,25

In this work, we aimed to provide a MAT method to synthesize Re–M/C (M = Pt or Rh) bimetallic catalysts in order to study the influence of Re–M interactions on the hydrogenation of SA. The Re–M/C catalysts were characterized and a kinetic study was performed to provide insight into the effect of Re–M interactions and reaction mechanism. Finally, the relationship between the textural properties of the catalysts and catalytic activity was discussed in depth by using the experimental data obtained.

2. Experimental

2.1 Catalyst preparation

Re–M/C bimetallic catalysts were prepared by a MAT method using dirhenium decacarbonyl [Re2(CO)10] (98.0%, Alfa Aesar), platinum(II) acetylacetonate [Pt(acac)2] (98.0%, Acros Organics) and rhodium(I) dicarbonylacetylacetonate [Rh(acac)(CO)2] (99.0%, Acros Organics) as precursors. The coconut shell based activated carbon (surface area: 1400 m2 g−1, pore volume: 0.7 cm3 g−1, pore diameter: 1.9 nm) was used as a support in these experiments. The activated carbon was pre-treated with HNO3 solution at 90 °C to remove impurities before use and to create oxygen-based anchoring groups for the metal precursors. Therefore, functional groups such as carboxylic, lactonic and phenolic groups can be introduced to the surface of activated carbon by this type of treatment.26 The precursors were impregnated on the carbon support and dried at 50 °C for 8 h. The homogeneous mixtures were then transferred into a fluidized quartz tube reactor and fluidized with Ar to make sure that the reaction mixture was maintained under inert atmosphere. Thereafter the samples were irradiated using a microwave (800 W, 2.45 GHz) for 5 min in a fluidized state and then cooled down naturally. The scheme of the MAT method is given in the ESI. The monometallic catalysts with 2 wt% metal loading were identified as Re/C, Rh/C and Pt/C. The Re–M/C bimetallic catalysts were prepared with a 1[thin space (1/6-em)]:[thin space (1/6-em)]1 atomic ratio and the same total metal loading (2 wt%) as the monometallic ones. They were identified as Re–Pt/C and Re–Rh/C.

2.2 Characterization method

In our experiments, the metal loading of catalysts was analysed using inductively coupled plasma atomic emission spectroscopy (ICP-AES). CO chemisorption was used to measure the amount of active sites and dispersion of the catalysts. The morphology of the samples was analysed by powder X-ray diffraction (XRD) and transmission electron microscopy (TEM). HAADF-STEM and the corresponding elemental mapping were performed to investigate the surface composition and fine structures. The metallic interaction and chemical state of the catalysts were studied by H2-TPR and XPS. H2-TPD was carried out to study the ability for hydrogen adsorption and activation of the Re–M/C bimetallic catalysts. More specific information is listed in the ESI.

2.3 Catalyst evaluation

Aqueous-phase hydrogenation of SA was carried out in a 50 mL autoclave (Hastelloy alloy) with a magnetic stirrer and a temperature controller unit. The catalysts were previously treated with H2 for 120 min at 300 °C for Rh/C and Re–Rh/C, at 400 °C for Re/C and Re–Pt/C and at 350 °C for Pt/C. After reduction, the catalysts were transferred into the autoclave under the protection of an Ar atmosphere as soon as possible and then a water solution of SA was added. The batch autoclave was purged with H2 to remove traces of air and then set to the desired pressure. After that, the reactor was rapidly heated to the reaction temperature and the reaction was conducted with constant stirring. When the reaction was finished, the reactor was cooled down to room temperature by water circulation and the spent catalysts were collected. For kinetic studies, the reaction was performed under constant pressure by using a pressure control system.

The reactant and products during conversion of succinic acid were analysed using a liquid chromatograph (Waters 1525, USA) equipped with an ultraviolet detector and a C18 column (Welch, Ultimate LP-C18, China). Aqueous phosphoric acid solution (5 mM) was used as the mobile phase and the wavelength of the detector was set at 204 nm. The reaction products were analysed using a gas chromatograph (Tianmei GC7890, China) equipped with a FFAP capillary column. Isopropanol was used as the internal standard. Products were also identified by GC-MS (Agilent 7000B, USA) and LC-MS (Thermo Scientific TSQ Quantum Ultra, USA).

The conversion, selectivity, relative concentration and TOFs were calculated as follows:

image file: c7cy01039g-t1.tif

image file: c7cy01039g-t2.tif

image file: c7cy01039g-t3.tif

image file: c7cy01039g-t4.tif
where n0 represents the initial molar quantity of SA and nt represents the molar quantity of SA at time t. mi represents the molar quantity of product i. ci represents the molar concentration of component i and ct represents the total molar concentration of all components. V represents the amount of active sites calculated from CO chemisorption. The TOFs were calculated based on the SA conversion at 220 °C, which is controlled below 20%. In the experiments, the carbon balance was around 90% for all catalysts.

3. Results and discussion

3.1 Phase structure and morphology

From the XRD patterns in Fig. 1, no diffraction peaks corresponding to the active metal were observed for monometallic Re/C, Pt/C and all Re–M/C bimetallic catalysts. This indicates that the metallic particles are not crystallized or too small to be resolved (less than 4 nm) by XRD, i.e.: highly dispersed on the support. Besides that, the low metal loading (2 wt%) tends to disadvantage the observation of diffraction peaks. Very small diffraction peaks were observed only on the Rh/C catalyst, which may exhibit aggregation of metallic particles on this catalyst. The TEM images in Fig. 2 confirm that the metal particles are well dispersed on the surface of the carbon support. One can note that the graphs of the particle size distribution exhibit non-symmetrical shapes. The average particle size of the Re–M bimetallic catalysts was significantly smaller than those of the monometallic Pt/C and Rh/C catalysts, whereas Re/C presents the smallest particle size. Furthermore, the size distribution histograms indicated that the Re–M/C bimetallic catalysts have a narrower distribution than the monometallic Pt/C and Rh/C catalysts. It can be inferred that there is an interaction between Re and M (M = Pt and Rh), favouring metal dispersion at the support surface. The higher dispersion of the metal in the Re-based catalysts could be explained by the use of the carbonyl precursor (Re2(CO)10), which could be decomposed into the active metal rapidly without noteworthy aggregation under microwave heating. Meanwhile, the acetylacetonate precursors (Pt(acac)2, Rh(acac)(CO)2) tend to partly aggregate when decomposed via microwave heating. The aggregation of acetylacetonate precursors during the preparation process can be inhibited by the mixed carbonyl precursors when the Re precursor is present, which contributes to the decrease of the particle size as shown in Fig. 2. The high dispersion of Re on Re/C may also be explained by the oxophilicity of this metal, favouring its strong interaction with the oxygenated functional groups at the carbon surface created during the nitric acid pretreatment.
image file: c7cy01039g-f1.tif
Fig. 1 The XRD patterns of Re/C, M/C monometallic and Re–M/C bimetallic catalysts (M = Pt or Rh).

image file: c7cy01039g-f2.tif
Fig. 2 The TEM images of Re/C, M/C and Re–M/C catalysts (M = Pt or Rh).

The Re–Pt/C and Re–Rh/C samples were characterized with the HAADF-STEM dark field images and the corresponding elemental maps shown in Fig. 3 and 4. The presence of oxygen is due to the oxygen-containing groups on activated carbon obtained by nitric acid pretreatment. According to elemental mapping, Re and M (M = Pt and Rh) did not exist as isolated species thus indicating the presence of bimetallic entities, where both metals are interacting. The linear scan in Fig. 3(B) showed that the atomic ratio of Re to Pt was about 1[thin space (1/6-em)]:[thin space (1/6-em)]1 for the atomic cluster, which is similar to the Re/Pt atomic ratio of the surface phase and bulk phase of the Re–Pt/C catalyst from the results in Table 1. A low degree of elemental segregation was detected on the Re–Rh/C catalyst since the Re/Rh atomic ratio was around 1[thin space (1/6-em)]:[thin space (1/6-em)]1 for both the atomic cluster and bulk phase as seen in Fig. 4(B) and Table 1. The results of linear scanning also demonstrate that the Re–M interaction exists in Re–M/C bimetallic catalysts. Meanwhile, the HAADF-STEM images and elemental maps also prove the fairly small particle size and good dispersion of the active metal.


image file: c7cy01039g-f3.tif
Fig. 3 The TEM image (A), linear scan (B), HAADF-STEM image and elemental mapping (C) of Re–Pt/C.

image file: c7cy01039g-f4.tif
Fig. 4 The TEM image (A), linear scan (B), HAADF-STEM image and elemental mapping (C) of Re–Rh/C.
Table 1 Characteristics of the Re–M/C bimetallic catalysts and their TOFs for the hydrogenation of SA
Catalysts Re[thin space (1/6-em)]:[thin space (1/6-em)]M molar ratio of bulk phasea Re[thin space (1/6-em)]:[thin space (1/6-em)]M molar ratio of surface phasea Average particle diameter (nm) Amount of CO uptake (μmol per g-catalyst) Dispersionb (%) TOFsb (h−1)
a The molar ratio of the bulk phase and surface phase was calculated by ICP and XPS, respectively. b Calculated by assuming a stoichiometry factor of CO[thin space (1/6-em)]:[thin space (1/6-em)]metal atom = 1. The TOFs are calculated when the conversion of SA is 20% at 220 °C, and with an initial pressure of 8.0 MPa.
Pt/C 0[thin space (1/6-em)]:[thin space (1/6-em)]1 0[thin space (1/6-em)]:[thin space (1/6-em)]1 3.3 9.5 9.3 646
Re–Pt/C 1[thin space (1/6-em)]:[thin space (1/6-em)]1 1[thin space (1/6-em)]:[thin space (1/6-em)]0.9 1.9 12.8 12.2 1419
Re/C 1[thin space (1/6-em)]:[thin space (1/6-em)]0 1[thin space (1/6-em)]:[thin space (1/6-em)]0 1.8 22.0 20.5 420
Re–Rh/C 1[thin space (1/6-em)]:[thin space (1/6-em)]1 2.7 27.6 20.0 1162
Rh/C 1[thin space (1/6-em)]:[thin space (1/6-em)]0 1[thin space (1/6-em)]:[thin space (1/6-em)]0 3.6 26.6 13.7 304


The catalytic activity for cyclohexane dehydrogenation was used as a probe to investigate the composition and accessibility of the surface metal as this model reaction performed in the gas phase is considered as structure insensitive. The TOFs for cyclohexane dehydrogenation are listed in Table S1 (the reaction is detailed in the ESI). The results showed that the activity of Pt/C (154 min−1) and Rh/C (22 min−1) was significantly higher than that of Re/C (9.5 min−1). The TOF values of the bimetallic catalysts can be calculated by linear combination of the TOF values of the corresponding monometallic catalysts modulated by the proportion of Re and M (M = Pt or Rh) for the bimetallic catalysts.27 The experimental TOFs for Re–Pt/C (87 min−1) and Re–Rh/C (20 min−1) are between those of the monometallic catalysts and similar to the calculated ones for Re–Pt/C (82 min−1) and Re–Rh/C (16 min−1), thus indicating that the surface composition of these catalysts are in accordance with the results obtained by XPS and linear scanning, with an equidistribution of Re and Rh/Pt at the surface of the bimetallic particles, i.e.: accessible to the reactants.

H2-TPR experiments were carried out to further investigate the Re–M interaction of bimetallic catalysts and the results are shown in Fig. 5. The peak at high temperature (500–800 °C) was attributed to methane formation coming from the reaction of the carbon support with hydrogen. Meanwhile, the temperature for the formation of methane in the carbon support is higher than those in monometallic and bimetallic catalysts, indicating that the reduced metal can also contribute to the formation of methane under H2/Ar atmosphere.28 The main peaks detected at 240, 385 and 345 °C for monometallic Rh/C, Re/C and Pt/C catalysts, respectively, were assigned to the reduction of monometallic oxides caused by re-oxidation of the metal after exposure to air.23,29,30 The higher reduction temperature for the Re/C catalyst than those of the Rh/C and Pt/C catalysts manifests that Re is easy to oxidize and difficult to reduce. The main reduction temperature for Pt/C is higher than those of some monometallic Pt catalysts whose reduction temperature is around 250–300 °C.31,32 The oxygenated groups on the surface of carbon could be bonded with Pt species, thereby strengthening the interaction between the metal species and support, leading to a higher reduction temperature. Some studies also confirmed the influence of the Pt–support interaction on reduction temperature.33–35 The first reduction temperature (283 °C) for bimetallic Re–Rh/C catalysts was between those of the two corresponding monometallic catalysts. The reduction of metallic oxides at this first reduction temperature can be attributed to the co-reduction of Re and Rh for bimetallic catalysts. This exhibits that Re is in close interaction with Rh metal, which can also be identified by linear scanning and elemental mapping. The promotion of the Re reduction can be attributed to the hydrogen spillover from beforehand reduced Rh to Re species.28 It is noteworthy that for the bimetallic Re–Pt/C catalyst the reduction starts at a low temperature, and continuously increases up to a first maximum at 420 °C which is higher than those of the corresponding monometallic catalysts. This could be due to the smaller size of bimetallic Re–Pt particles (1.9 nm), resulting from strong Re–Pt interactions, compared to that of the Pt particles (3.3 nm) in the monometallic ones, inducing a stronger metal–support interaction via the oxygen groups.


image file: c7cy01039g-f5.tif
Fig. 5 H2-TPR results of Re/C, M/C monometallic and Re–M/C bimetallic catalysts (M = Pt or Rh) (Y-axis on different scales).

XPS was used to further characterize the Re–M interaction in the catalysts. The distribution of the chemical states in the catalysts was estimated using the curve fit of Re 4f, Rh 3d and Pt 4f, as presented in Fig. 6. Though the XPS measurements were not carried out under in situ conditions but after air exposure, the characterization results can give useful information about the structure properties of the catalysts. Before the measurement, the samples were stored under the same conditions and were exposed to air for the same period of time. These samples were analysed by XPS simultaneously, in order to characterize them under the same conditions. The Rh 3d profile fitting (Fig. 6(C)) showed that Rh0 was the major rhodium species36 and the weak signal of Rh was detected on the Re–Rh/C catalyst which may result from the less amount of rhodium. The deconvoluted peaks at 41.2, 42.5, and 45.2 eV in Fig. 6(A) are assigned to Re 4f7/2 lines and indicate the existence of Re0, Re4+ and Re6+ species, respectively.23,37,38 The spectrum in Fig. 6(B) can be fitted in terms of three different chemically active species with Pt 4f binding energies at 71.2, 72.5 and 74.0 eV, assigned to Pt0, Pt2+ and Pt4+, respectively.39,40 According to Fig. 6(A), a position shift (0.2–0.5 eV) in binding energy can be observed for the Re–Pt/C and Re–Rh/C bimetallic catalysts with respect to the monometallic Re/C one indicating the existence of intermetallic interactions. The oxidation state ratios are quantified and summarized in Table S2. Re is easy to oxidize when exposed to air compared with Pt and Rh according to the results. However, the Re–Pt/C catalyst exhibits a much higher proportion of zero-valence Re than the Re/C catalyst, which demonstrates that Re–Pt interactions can contribute to the stabilization of metallic Re thereby avoiding the oxidation of Re. Similarly, the Re–Rh interaction was also beneficial to maintaining Re in its metallic state, with a higher proportion of Re0 than in the Re/C catalyst. Furthermore, the proportion of Re0 for Re–Pt/C is higher than that of Re–Rh/C, which shows that the interaction between Re and Pt for Re–Pt/C is stronger than the interaction between Re and Rh for Re–Rh/C. In a previous work dedicated to the study of Re-based catalysts supported on TiO2,19 XPS analysis performed under the same conditions, i.e.: without in situ reduction of the samples, has revealed the presence of Re species only in an oxidized state. A Re0 peak was observed only after reduction, with a proportion of 15% for the monometallic Re/TiO2 catalyst. Consequently, in the present study, it seems that the carbon support favours maintaining Re in its metallic state, this being also more reducible than Re deposited on TiO2.


image file: c7cy01039g-f6.tif
Fig. 6 XPS results of the Re/C, M/C monometallic and Re–M/C bimetallic catalysts (M = Pt or Rh).

The adsorption and activation of hydrogen on the active sites of the Re/C, Pt/C and Rh/C monometallic and Re–M/C bimetallic catalysts were studied by H2-TPD and the thermal desorption curves of hydrogen from the various samples, the carbon support included, are presented in Fig. 7. A broad peak at high temperature (650–800 °C) was observed on the carbon support which can correspond to the reduction of the functionalized carbon on the surface caused by hydrogen spillovers according to the results of H2-TPR. In our experiments, the thermal desorption curves can be divided into four peaks (peak I, II, III and IV) by simulation. Peak IV corresponds to the methanation of the support according to the dissociation of adsorbed hydrogen on a carbon support. Peak III with the maximum desorption temperature of 600 °C can be attributed to desorption of hydrogen spilt over the support. The hydrogen adsorbed on the metal surface can migrate in the form of atomic hydrogen onto the carbon support via surface diffusion, and the migrated atomic hydrogen is stabilized by the oxygen functional groups on the surface of the support. At high temperature, the spillover hydrogen is removed from the surface of supports.41–43 The main desorption temperatures are 270 and 450 °C for peaks I and II, respectively, which reflect the relatively different adsorption degrees of chemisorbed hydrogen according to the desorption temperature. Thereafter, peak I is identified as weak adsorption and peak II as medium adsorption. The peak area of peak II on the Re/C catalyst was used as the standard for quantitative analysis. To eliminate the influence of the molar quantity of the active metal on the calculation, the amount of relative hydrogen desorption for peaks I and II was calculated according to the following equation:

image file: c7cy01039g-t5.tif


image file: c7cy01039g-f7.tif
Fig. 7 H2-TPD results of the carbon support, Re/C and M/C monometallic and Re–M/C bimetallic catalysts (M = Pt or Rh).

The amounts of relative hydrogen desorption are listed in Table S3. The Re–Pt/C and Re–Rh/C bimetallic catalysts exhibited a higher total amount of desorbed hydrogen than the monometallic catalysts with high predominance of the medium adsorption (peak II) of hydrogen resulting from the influence of the Re–M interaction. This phenomenon is more significant over Re–Pt/C compared with Re–Rh/C because of the stronger intermetallic interaction for Re–Pt/C. This effect is even more important considering that the relative incremental amount of medium hydrogen adsorption for Re–Pt/C (from Pt/C [1.6] to Re–Pt/C [2.5]) is larger than that for Re–Rh/C (from Rh/C [1.0] to Re–Rh/C [1.9]). Thus, it can be seen that the Re–M interaction not only strengthens the hydrogen activation and adsorption, but also contributes to the generation of medium hydrogen adsorption on the metallic active sites.18,23

3.2 Kinetic study on hydrogenation of SA and GBL over Re–Pt/C and Re–Rh/C catalysts

Analysis of the kinetics is essential to correlate structure properties and catalytic activity. We have performed a kinetics study of the main reaction pathways as shown in Table 2. Details for each experiment are provided in the following subsections and the ESI. In these experiments, the influence of external and internal transport was eliminated as much as possible by controlling the reaction conditions (ESI). The specific fitting method is fully described in the ESI.
Table 2 The kinetic data for the hydrogenation of SA and GBL over Re–Pt/C and Re–Rh/C

image file: c7cy01039g-u1.tif

image file: c7cy01039g-t6.tif

image file: c7cy01039g-t7.tif

image file: c7cy01039g-t8.tif

Re–Pt/C α 1 = 0.6 α 2 = 0.4 α 3 = 1.0
β 1 = 1.4 β 2 = 0.9 β 3 = 0.9
E 1a = 45 kJ mol−1 E 2a = 20 kJ mol−1 E 3a = 91 kJ mol−1
Re–Rh/C α 1 = 0.3 α 2 = 0.2 α 3 = 0.8
β 1 = 2.7 β 2 = 1.2 β 3 = 1.1
E 1a = 51 kJ mol−1 E 2a = 38 kJ mol−1 E 3a = 117 kJ mol−1


3.2.1 Kinetic study on the hydrogenation of SA. The hydrogenation of SA to GBL (step 1, Table 2) was firstly studied, and the conversion of SA was controlled to be under 30% to make sure that the major SA hydrogenated product was GBL (>90% selectivity). Fig. S1 shows the effect of the initial concentration of SA on the reaction rate concentration over Re–Pt/C and Re–Rh/C. The reaction order with respect to the initial SA concentration over Re–Pt/C and Re–Rh/C was estimated to be 0.6 and 0.3 respectively. This observation is consistent with previous studies indicating that the apparent reaction order between zero and one are generally observed for the hydrogenation of C[double bond, length as m-dash]O over noble metals.38,44–46 As a result, it is strongly suggested that SA is more strongly adsorbed on Re–Rh/C than on Re–Pt/C, and the reaction rate is more influenced by the SA concentration for Re–Pt/C. However, the difference is not so significant and the reaction order remains close for the two catalysts. Moreover, the reaction order can reflect the saturation of available metal surface sites by hydrogenated intermediates.45 The nearer a catalyst approximates to zero reaction order, the higher its surface saturation is. Compared with the monometallic Re/C catalyst for which the reaction order is one,23 Re–M/C (M = Pt and Rh) bimetallic catalysts are more benefit to the surface saturation by hydrogenated intermediates, meaning that the Re–M interaction contributes to the activation of the intermediates of SA hydrogenation.

Fig. S2 illustrates the dependence of the reaction rate constant (k1) on hydrogen pressure over Re–Pt/C and Re–Rh/C, and the details are listed in Fig. S3. An apparent reaction order of 1.4 and 2.7 with respect to hydrogen pressure was observed for Re–Pt/C and Re–Rh/C, respectively. As far as we know, a high reaction order of hydrogen pressure is often observed for the hydrogenation of carboxyl.21,23,38 The positive reaction order indicates that the reaction rate was strongly influenced by hydrogen pressure, which means that dissociation of H2 or the reaction of the dissociated hydrogen species is involved in the rate-determining step. Considering the fast dissociation of H2 over Re based catalysts, the reaction of the dissociated hydrogen species will be the rate-determining step.21,38 According to the results, the Re–Pt interaction is much more favorable to the reaction of dissociated hydrogen species because of the significantly smaller reaction order (β1 = 1.4) compared with Re–Rh (β1 = 2.7). This phenomenon implies that the stronger metallic interaction of Re–Pt/C is more beneficial to the activation of hydrogen and reaction of hydrogen species. Moreover, the hydrogenation of SA is greatly influenced by hydrogen pressure compared with the effect of the SA concentration according to the higher reaction order of hydrogen pressure than the SA order.

Fig. S4 shows the Arrhenius plots for Re–Pt/C and Re–Rh/C, and the detailed fitting data are provided in Fig. S5. According to the data, the apparent activation energy is 45 kJ mol−1 for the Re–Pt/C catalyst while it is 51 kJ mol−1 for the Re–Rh/C catalyst. The hydrogenation of SA is much easier over Re–M/C than over the Re/C catalyst (65 kJ mol−1 (ref. 23)) as a result of the lower activation energy. The Re–M interaction can change the reaction route to increase the reaction rate. The activation energy of Re–Pt/C is lower than that of Re–Rh/C because of the stronger metallic interaction between Re and Pt. In addition, the apparent activation energy also reveals the influence of temperature on the rate of reaction. It can be seen that the temperature has a more significant influence on Re–Rh/C than on Re–Pt/C.

3.2.2 Kinetic study on the hydrogenation of GBL. On the basis of the results above, the kinetics study of hydrogenation of GBL to BDO and THF (steps 2 and 3 in Table 2) was also performed. The study on the kinetics of GBL hydrogenation is useful to help explain the reaction mechanism because GBL is the intermediate product of SA hydrogenation. The major GBL hydrogenation products were THF and BDO (total selectivity >90%) by controlling the reaction conditions. The details of the reaction conditions are listed in the ESI.

Fig. S6 and S7 show the influence of GBL concentration and hydrogen pressure on the initial reaction rate, and the fitting results are summarized in Table 2. The apparent reaction order of the GBL concentration was 1.0 for the generation of THF over Re–Pt/C, which was higher than that (0.4) for the generation of BDO. The same tendency is observed with the Re–Rh/C catalyst. This proves that the transformation of GBL into THF by hydrogenation and dehydration is more influenced by the GBL concentration than the ring opening of GBL to BDO. A high concentration of GBL, the intermediate product of SA hydrogenation, is favourable to the generation of THF. Being different than the effect of GBL concentration, hydrogen pressure has limited influence on the selectivity since the reaction orders of hydrogen pressure for the generation of THF and BDO over Re–Pt/C and Re–Rh/C were similar, around one.

Table 3 summarizes the reaction rate constant k and initial reaction rate r for the hydrogenation of GBL. The Re–M interaction obviously promotes the transformation of GBL into BDO since the values of kBDO/kTHF and rBDO/rTHF of the Re–M/C catalysts are double that of the Re/C catalyst under different temperatures. The kinetic data prove that the Re–M interaction changes the activation of the intermediate components on the surface of Re–M/C, thereby contributing to the ring opening of GBL by hydrogenation to BDO instead of dehydration–hydrogenation to THF. Re–M/C increases the generation rate of BDO more than Re/C does, as shown in Scheme 1, which can explain the difference in selectivity over the various catalysts. In addition, kBDO/kTHF and rBDO/rTHF of Re–Pt/C are nearly the same as those of Re–Rh/C. This result exhibits that the selectivity to BDO and THF is not influenced by the strength of the metallic interaction.

Table 3 The reaction rate constants (k) and reaction rate (r) for the hydrogenation of GBL over Re/C and Re–M/C catalysts under 8.0 MPa
Temperature Re–Pt/C Re–Rh/C Re/Cb
k BDO/kTHF r BDO/rTHFa k BDO/kTHF r BDO/rTHFa k BDO/kTHF r BDO/rTHFa
a The initial reaction rate was calculated at a 10 wt% concentration of GBL. b The data of Re/C come from ref. 23 and were obtained under the same conditions as those in the present study.
160 °C 8.5 7.8 9.2 8.4 4.8 4.1
170 °C 4.8 4.3 6.0 5.4 2.4 2.1
180 °C 3.0 2.8 3.3 3.0 1.4 1.2
190 °C 2.3 2.1 2.2 2.0
200 °C 1.6 1.4 1.5 1.3



image file: c7cy01039g-s1.tif
Scheme 1 Proposed reaction mechanism of SA hydrogenation over Re/C, Re–Pt/C and Re–Rh/C catalysts, which is focusing on the bimetallic effect.

The values of kBDO/kTHF and rBDO/rTHF for all catalysts gradually decrease when the temperature increases. This proves that low temperature is more beneficial to the ring opening of GBL to BDO than to the direct hydrogenation of GBL to THF. Arrhenius plots for the Re–Pt/C and Re–Rh/C catalysts are shown in Fig. S8, and the apparent activation energies are displayed in Table 2. For all catalysts, the value of the apparent activation energy for the generation of THF is much higher than that of the generation of BDO. This shows that the activity of GBL direct hydrogenation to THF is more sensitive to the reaction temperature than that of GBL ring opening. The generation rate of THF will change more than that of BDO when decreasing the reaction temperature and thus it is possible to increase the selectivity to BDO. It is worth noting that the apparent activation energy for Re–Pt/C is lower than for Re–Rh/C, which is similar to that observed for the hydrogenation of SA. It can be concluded that compared with Re–Rh/C the stronger Re–M interaction of Re–Pt/C contributes to the decrease in the activation energy of SA and GBL conversion, increasing the catalytic activity.

3.3 Comparison for aqueous-phase hydrogenation of SA over Re/C, Rh/C, Pt/C, Re–Pt/C and Re–Rh/C catalysts

The product distributions and TOF values determined during succinic acid conversion at 220 °C and an initial pressure of 8.0 MPa in the presence of the various catalysts are summarized in Fig. 8 and Table 1. The products observed in addition to GBL, THF and BDO are propanoic acid (PA), butyric acid (BA), n-propanol (NPA) and n-butanol (NBA), produced according to Scheme S1. The TOF values for the Re–M/C bimetallic catalysts are significantly higher than those of monometallic catalysts according to the results in Table 1. This is in accordance with the results of the kinetic study. This result is probably not only related to the presence of a second metal but also to the Re–M interaction. The Re–M interaction is beneficial to the stabilization of Re0 by strengthening the interaction between the metal and support,19 especially for Re–Pt/C which shows a stronger metallic interaction than Re–Rh/C according to characterization results.
image file: c7cy01039g-f8.tif
Fig. 8 Product distributions of Pt/C (A), Re–Pt/C (B), Re/C (C), Re–Rh/C (D), Rh/C (E). Reaction conditions: 220 °C, initial pressure 8.0 MPa, 20 g solutions (5 wt% SA), 0.1 g of catalyst for Re–Rh/C and 0.2 g of catalyst for other catalysts.

The results in Fig. 8 confirm that hydrogenation of SA is a consecutive reaction and the final product is THF coming from the further hydrogenation of the GBL intermediate over Re/C and Pt/C, in accordance with the scheme presented in Table 2. Over Rh/C, the reaction mechanism is different; the main products are GBL and PA. The Rh/C catalyst is then beneficial to the hydrogenolysis of SA to PA, compared with Re/C and Pt/C. The hydrogenolytic properties of Rh, i.e.: its capacity to break C–C bonds, are well-known.47 The introduction of Re to Rh improves the hydrogenation performance, since the major final products with the Re–Rh/C bimetallic catalyst are THF and NPA, the latter being produced from BDO. It also reduces the direct hydrogenolysis of SA, PA being barely detected on Re–Rh/C. The results show that the interaction between Re and Rh is beneficial to the hydrogenation of SA to GBL that is then hydrogenated at the expense of hydrogenolysis. As NPA is produced from BDO, in the following we will consider 3 processes: the route involving direct hydrogenolysis, leading to BA and PA; the route to BDO, leading to NPA and NBA; and the route to THF.

The generation of BDO and THF was observed for Re–M/C bimetallic catalysts, which is different from that observed with monometallic Re/C, Rh/C and Pt/C catalysts. As a consequence, the Re–M intermetallic interaction seems to favour the hydrogenation of GBL, especially the route to BDO, in accordance with our kinetic results. The Re–M/C can increase the relative generation rate of BDO more than that of THF according to the kinetic study. Furthermore, the medium adsorbed hydrogen also played a key role in the generation of BDO considering that the selectivity to BDO is higher for the Re–M/C samples that possess a significantly larger amount of medium adsorbed hydrogen than monometallic catalysts based on the results of H2-TPD. The Re–M interaction can contribute to the increase of medium adsorbed hydrogen and the hydrogen-binding sites are beneficial to the ring opening of the GBL intermediate. Our previous work also proves that the medium hydrogen adsorption can increase the catalytic activity to generate diols.23 What is of interest is that the selectivity to BDO of Re–Pt/C is higher than that of Re–Rh/C, whereas the kinetic study exhibited that the selectivity to BDO and THF is not significantly influenced by the strength of the Re–M interaction. Besides, H2-TPD analysis also confirmed that the ratio of medium hydrogen adsorption to the total hydrogen adsorption is almost the same for Re–Pt/C and Re–Rh/C. This phenomenon can be attributed to the remarkable performance of Re–Rh/C for the hydrogenolysis of BDO yielding a large amount of NPA (Fig. 8(D)), thus suppressing the selectivity to BDO. In spite of this, Re–Rh/C can reduce the direct hydrogenolysis of SA to PA, but it could not avoid the hydrogenolysis of BDO to NPA. The study on the transformation of the intermediate products reported in Table S4 also further proved this conclusion. While the Re–Pt interaction is stronger than the Re–Rh one, it was found that the selectivity to BDO for these two catalysts was slightly different when using GBL as a reactant. This result exhibits that the selectivity to BDO is not significantly influenced by the strength of the Re–M interaction, which is consistent with the value of rBDO/rTHF (Table 3). However, the selectivity to NPA of Re–Rh/C is significantly higher than that of Re–Pt/C due to the high hydrogenolysis activity of Rh. Furthermore, the main product was THF and a few NPA molecules were detected on the Re–Pt/C catalyst when using BDO as the reactant, compared to a high amount of NPA obtained on the Re–Rh/C catalyst. This gives evidence that although Re–Rh/C can reduce the direct hydrogenolysis of SA, it favours the hydrogenolysis of BDO, decreasing the selectivity to BDO. Besides, Re–Pt/C and Re–Rh/C can transform BA and PA into NBA and NPA, respectively, by hydrogenation. PA will be quickly hydrogenated to NPA because of the Re–M interaction, which may also result in the increase in the amount of NPA to a certain extent. Research on the transformations of intermediates confirms the above conclusions and the corresponding main reaction pathways shown in Scheme S1.

3.4 Evaluation of reaction conditions

A higher apparent activation energy was observed for the generation of THF compared with that of BDO (Table 2). As a result, the improvement of selectivity to BDO can benefit from an appropriate decrease in temperature. In contrast, increasing the temperature contributes to the formation of THF. Furthermore, increasing the temperature can promote the cyclodehydration of BDO to THF. For these reasons, the effect of the temperature on the performances of the bimetallic catalysts was studied between 160 and 240 °C. The results are reported in Table 4. Taking Re–Pt/C as an example, the selectivity to BDO decreased from 80.0% to 12.1% while the selectivity to THF increased from 9.3% to 60.6% when the temperature increased from 160 °C to 240 °C. However, the hydrogenolysis of BDO was also strengthened when increasing the temperature, with the selectivity to NPA rising from 1.9 to 9.6%. So, low temperature is not only beneficial to the generation of BDO, but also contributes to the reduction of the by-products. Compared with Re–Rh/C, Re–Pt/C is more favourable to obtain BDO or THF with high selectivity considering that the Re–Rh interaction on Re–Rh/C cannot fully suppress the hydrogenolysis of BDO. Indeed, the selectivity to NPA increases from 18.5% to 39.4% as the temperature increases over Re–Rh/C (Table 4). The stronger metallic interaction between Re and Pt in Re–Pt/C contributes to controlling the side reactions and an optimum of 80% selectivity to BDO was obtained on Re–Pt/C. But low temperature will obviously decrease the catalytic activity. Because of this, it is important to balance the catalytic activity and selectivity.
Table 4 The influence of temperature on the hydrogenation of SA over Re–M/C
Catalysts Temp (°C) Con. (%) Sel. (%)
GBL BDO THF NBA NPA
a Reaction conditions: initial pressure 8.0 MPa, 20 g solutions (5 wt%), 0.2 g catalyst; *20 g solutions (2.5 wt%). b Reaction conditions: initial pressure 8.0 MPa, 20 g solution (5 wt%), 0.1 g catalyst; *20 g solution (2.5 wt%), 0.2 g catalyst.
Re–Pt/Ca 160* >99 0.1 80.0 9.3 8.7 1.9
180* >99 5.1 67.6 16.3 9.0 2.0
220 >99 5.0 30.1 45.9 9.7 9.3
240 >99 7.2 12.1 60.6 10.5 9.6
Re–Rh/Cb 160* >99 7.3 61.4 5.1 7.7 18.5
180 >99 2.7 48.7 14.0 7.0 27.6
220 >99 3.0 8.7 44.7 4.1 39.4


3.5 Stability of bimetallic catalysts

The reusability of Re–Pt/C and Re–Rh/C was evaluated with four recycle times. After reaction, most of the solution was centrifuged from the reaction system and a small part of the solution was left to protect the catalyst from air. Then the reaction was performed again with a fresh solution and a small quantity of catalyst was added into the system to make up for the loss of catalyst. It can be found that, only a slight drop in catalytic activity was observed for Re–Pt/C and Re–Rh/C catalysts in our experiments as shown in Fig. S9. By analysis of the reactant solution with ICP-AES, no significant leaching of the active metal was detected between each cycle. The results indicate that the Re–M/C catalysts prepared with the MAT method are stable for aqueous-phase hydrogenation of SA.

Succinic acid can be obtained from the fermentation of bio-wastes and some impurities such as acetic acid and pyruvic acid are involved in the fermentation broth.48 For this purpose, it is meaningful to investigate the influence of these impurities on the catalytic activity. In these experiments, the hydrogenation reaction of SA over Re–Pt/C was performed with the presence of 1 wt% appropriate acetic acid or pyruvic acid in the reactive system. Because of the competitive effect of acetic acid or pyruvic acid with SA, the catalytic activity decreased in different degrees. The active sites are not only occupied by SA but also by acetic acid or pyruvic acid and the two additives will be further transformed. According to our results, acetic acid is hydrogenated to ethanol while the main products are 1,2-propanediol and n-propanol after the hydrogenation of pyruvic acid. SA continues to be transformed into the desired products after complete transformation of the impurities and the selectivity is finally not significantly influenced.

4. Conclusions

Re–M/C (M = Pt or Rh) catalysts prepared by a MAT method present very well dispersed metal particles compared to M/C catalysts, with a high degree of Re–M interaction, as evidenced by TEM, EDX, XPS, TPR and TPD analysis. The Re–M interaction contributes to the improvement in the catalytic activity for SA hydrogenation by decreasing the apparent activation energy, especially for Re–Pt/C which exhibits a stronger Re–M interaction than Re–Rh/C. The Re–M interaction favours the activation of intermediates compared with the Re/C monometallic catalyst and the stronger Re–Pt interaction is more favourable to the reaction of hydrogen species compared with the Re–Rh interaction. The kinetic study proves that the Re–M interaction can increase the relative generation rate of BDO more than that of THF, and the strength of the Re–M interaction has a limited influence on the selectivity, the rBDO/rTHF ratio being nearly the same for Re–Pt/C and Re–Rh/C. In spite of this, the Re–Rh interaction can reduce the direct hydrogenolysis of SA to PA, but it cannot avoid the hydrogenolysis of BDO to NPA, thus limiting the selectivity to BDO. Being different with the concentration of GBL, hydrogen pressure does not affect the selectivity as a result of the reaction order of H2 for the generation of THF and BDO being consistent. Besides, low temperature is advantageous to the ring opening of GBL to BDO and high temperature is more advantageous to the direct hydrogenation of GBL to THF, due to the fact that the value of the apparent activation energy for the generation of THF is much higher than that of the generation of BDO for both Re–Pt/C and Re–Rh/C.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

We gratefully acknowledge the financial support provided by the National Natural Science Foundation of China (21573031), the Program for Excellent Talents in Dalian City (2016RD09) and the Project Partenariats Hubert Curien (PHC) Cai Yuanpei supported by the French Embassy in China and the China Scholarship Council.

References

  1. J. J. Bozell and G. R. Petersen, Green Chem., 2010, 12, 539–554 RSC.
  2. J. C. Serrano-Ruiz, R. Luque and A. Sepulveda-Escribano, Chem. Soc. Rev., 2011, 40, 5266–5281 RSC.
  3. A. Cukalovic and C. V. Stevens, Biofuels, Bioprod. Biorefin., 2008, 2, 505–529 CrossRef CAS.
  4. M. Patel, M. Crank, V. Dornburg, B. Hermann, L. Roes, B. Hüsing, L. Overbeek, F. Terragni and E. Recchia, The BREW project, 2006 Search PubMed.
  5. U. Herrmann and G. Emig, Ind. Eng. Chem. Res., 1998, 37, 759–769 CrossRef CAS.
  6. C. Delhomme, D. Weuster-Botz and F. E. Kühn, Green Chem., 2009, 11, 13–26 RSC.
  7. J. R. Budge, T. G. Attig and A. Dubbert, US Pat., 5969164, 1999 Search PubMed.
  8. M. Konishi, K. Yokota and E. Ueno, US Pat., 6706932, 2004 Search PubMed.
  9. T. Werpy, J. G. Frye, Y. Wang and H. Zacher, US Pat., 6670300, 2003 Search PubMed.
  10. U. G. Hong, J. K. Kim, J. Lee, J. K. Lee, J. H. Song, J. Yi and I. K. Song, J. Ind. Eng. Chem., 2014, 20, 3834–3840 CrossRef CAS.
  11. L. Corbel-Demailly, B. K. Ly, D. P. Minh, B. Tapin, C. Especel, F. Epron, A. Cabiac, E. Guillon, M. Besson and C. Pinel, ChemSusChem, 2013, 6, 2388–2395 CrossRef CAS PubMed.
  12. U. G. Hong, H. W. Park, J. Lee, S. Hwang, J. Yi and I. K. Song, Appl. Catal., A, 2012, 415–416, 141–148 CrossRef CAS.
  13. U. G. Hong, J. K. Kim, J. Lee, J. K. Lee, J. H. Song, J. Yi and I. K. Song, Appl. Catal., A, 2014, 469, 466–471 CrossRef CAS.
  14. B. Tapin, F. Epron, C. Especel, B. K. Ly, C. Pinel and M. Besson, ACS Catal., 2013, 3, 2327–2335 CrossRef CAS.
  15. S. H. Chung, Y. M. Park, M. S. Kim and K. Y. Lee, Catal. Today, 2012, 185, 205–210 CrossRef CAS.
  16. D. P. Minh, M. Besson, C. Pinel, P. Fuertes and C. Petitjean, Top. Catal., 2010, 53, 1270–1273 CrossRef CAS.
  17. B. K. Ly, D. P. Minh, C. Pinel, M. Besson, B. Tapin, F. Epron and C. Especel, Top. Catal., 2012, 55, 466–473 CrossRef CAS.
  18. K. H. Kang, U. G. Hong, Y. Bang, J. H. Choi, J. K. Kim, J. K. Lee, S. J. Han and I. K. Song, Appl. Catal., A, 2015, 490, 153–162 CrossRef CAS.
  19. B. K. Ly, B. Tapin, M. Aouine, P. Delichere, F. Epron, C. Pinel, C. Especel and M. Besson, ChemCatChem, 2015, 7, 2161–2178 CrossRef CAS.
  20. Z. Shao, C. Li, X. Di, Z. Xiao and C. Liang, Ind. Eng. Chem. Res., 2014, 53, 9638–9645 CrossRef CAS.
  21. Y. Takeda, M. Tamura, Y. Nakagawa, K. Okumura and K. Tomishige, Catal. Sci. Technol., 2016, 6, 5668–5683 CAS.
  22. X. Di, Z. Shao, C. Li, W. Li and C. Liang, Catal. Sci. Technol., 2015, 5, 2441–2448 CAS.
  23. X. Di, C. Li, B. Zhang, J. Qi, W. Li, D. Su and C. Liang, Ind. Eng. Chem. Res., 2017, 56, 4672–4683 CrossRef.
  24. H. J. Kitchen, S. R. Vallance, J. L. Kennedy, N. Tapia-Ruiz, L. Carassiti, A. Harrison, A. G. Whittaker, T. D. Drysdale, S. W. Kingman and D. H. Gregory, Chem. Rev., 2014, 114, 1170–1206 CrossRef CAS PubMed.
  25. M. B. Gawande, S. N. Shelke, R. Zboril and R. S. Varma, Acc. Chem. Res., 2014, 47, 1338–1348 CrossRef CAS PubMed.
  26. J. S. Noh and J. A. Schwarz, Carbon, 1990, 28, 675–682 CrossRef CAS.
  27. N. Hérault, L. Olivet, L. Pirault-Roy, C. Especel, M. A. Vicerich, C. L. Pieck and F. Epron, Appl. Catal., A, 2016, 517, 81–90 CrossRef.
  28. S. G. Wettstein, J. Q. Bond, D. M. Alonso, H. N. Pham, A. K. Datye and J. A. Dumesic, Appl. Catal., B, 2012, 117–118, 321–329 CrossRef CAS.
  29. S. R. Bare, S. D. Kelly, F. D. Vila, E. Boldingh, E. Karapetrova, J. Kas, G. E. Mickelson, F. S. Modica, N. Yang and J. J. Rehr, J. Phys. Chem. C, 2011, 115, 5740–5755 CAS.
  30. K. Chary, C. Srikanth and V. Venkatrao, Catal. Commun., 2009, 10, 459–463 CrossRef CAS.
  31. M. J. Taylor, L. J. Durndell, M. A. Isaacs, C. M. A. Parlett, K. Wilson, A. F. Lee and G. Kyriakou, Appl. Catal., B., 2016, 180, 580–585 CrossRef CAS.
  32. L. Wang, H. Wan, S. Jin, X. Chen, C. Li and C. Liang, Catal. Sci. Technol., 2015, 5, 465–474 CAS.
  33. L. W. Ho, C. P. Hwang, J. F. Lee, I. Wang and C. T. Yeh, J. Mol. Catal. A: Chem., 1998, 136, 293–299 CrossRef CAS.
  34. B. M. Nagaraja, C. H. Shin and K. D. Jung, Appl. Catal., A, 2013, 467, 211–223 CrossRef CAS.
  35. G. Neri, C. Milone, S. Galvagno, A. P. J. Pijpers and J. Schwank, Appl. Catal., A, 2002, 227, 105–115 CrossRef CAS.
  36. S. Parres-Esclapez, I. Such-Basañez, M. J. Illán-Gómez, C. Salinas-Martínez de Lecea and A. Bueno-López, J. Catal., 2010, 276, 390–401 CrossRef CAS.
  37. J. Okal, W. Tylus and L. Kepinski, J. Catal., 2004, 225, 498–509 CrossRef CAS.
  38. Y. Takeda, M. Tamura, Y. Nakagawa, K. Okumura and K. Tomishige, ACS Catal., 2015, 5, 7034–7047 CrossRef CAS.
  39. A. Arevalo-Bastante, M. A. Álvarez-Montero, J. Bedia, L. M. Gómez-Sainero and J. J. Rodriguez, Appl. Catal., B, 2015, 179, 551–557 CrossRef CAS.
  40. A. S. Aricò, A. K. Shukla, H. Kim, S. Park, M. Min and V. Antonucci, Appl. Surf. Sci., 2001, 172, 33–40 CrossRef.
  41. W. Rachmady and M. A. Vannice, J. Catal., 2000, 192, 322–334 CrossRef CAS.
  42. D. S. Park, D. Yun, T. Y. Kim, J. Baek, Y. S. Yun and J. Yi, ChemSusChem, 2013, 6, 2281–2289 CrossRef CAS PubMed.
  43. R. Prins, Chem. Rev., 2012, 112, 2714–2738 CrossRef CAS PubMed.
  44. W. Rachmady and M. A. Vannice, J. Catal., 2000, 192, 322–334 CrossRef CAS.
  45. O. A. Abdelrahman, A. Heyden and J. Q. Bond, ACS Catal., 2014, 4, 1171–1181 CrossRef CAS.
  46. L. Fabre, P. Gallezot and A. Perrard, J. Catal., 2002, 208, 247–254 CrossRef CAS.
  47. Y. Shinmi, S. Koso, T. Kubota, Y. Nakagawa and K. Tomishige, Appl. Catal., B, 2010, 94, 318–326 CrossRef CAS.
  48. J. B. McKinlay, C. Vieille and J. G. Zeikus, Appl. Microbiol. Biotechnol., 2007, 76, 727–740 CrossRef CAS PubMed.

Footnote

Electronic supplementary information (ESI) available. See DOI: 10.1039/c7cy01039g

This journal is © The Royal Society of Chemistry 2017