Effect of Mn promoter on Rh/tungsten carbide on product distributions of alcohols and hydrocarbons by CO hydrogenation

Abstract

Roles of Mn promoter on Rh-impregnated mesoporous tungsten carbides (W_xC) have been investigated for direct synthesis of mixed alcohols by CO hydrogenation of syngas. Mn/Rh ratio on the mesoporous W_xC, prepared using a hard-template of mesoporous SBA-15, significantly altered CO conversion and product distributions. An optimal Mn content of 3 wt% on Rh/W_xC was responsible for a high CO conversion of 8.1% and selectivity for higher alcohols of 54.4% compared with other catalysts. The enhanced activity and selectivity towards alcohols at an optimal Mn content were mainly attributed to close interaction of Rh and Mn nanoparticles with partially reduced Rh nanoparticles on the W_xC, which enhanced CO adsorption as well as suppressed CO hydrogenation to produce light hydrocarbons. This superior activity originated from a higher dispersion of Rh nanoparticles, which intimately interacted with Mn species, maintaining an appropriate surface ratio of metallic Rh to oxidized Rh⁺ surfaces. The structures of the ordered mesoporous W_xC support greatly altered the chemical states of Rh nanoparticles by generating a strong metal–support interaction as well, and Rh–Mn(3)/W_xC showed a high production rate of higher alcohols by synergistic effects of the supported Rh–Mn nanoparticles with an appropriate interaction with the mesoporous W_xC support.

---

1. Introduction

Recent demands to develop economically feasible chemical processes for clean fuel production from alternative feedstock like renewable resources with fewer environmental problems have been much increased by the designing of efficient catalytic systems.¹ Utilization of biomass-derived syngas has enormous potential to produce value-added chemicals such as ethanol, which is an attractive fuel additive and an intermediate in the petrochemical industry.^2,3 Therefore, the selective synthesis of higher alcohols from syngas has continuously attracted attention owing to their various applications such as high-quality fuels, fuel additives, and chemical intermediates.⁴ The synthesis of higher alcohols from syngas has been carried out. Over the past few decades, many researchers have made efforts to develop efficient catalysts with a satisfactory selectivity for production of higher alcohols by CO hydrogenation using various supported metal catalysts – mainly cobalt and copper.^5–7 In addition, precious Rh-based heterogeneous catalysts are also known to be effective for the selective synthesis of C₂₊ oxygenates from syngas.² Active Rh metal is known to have intrinsic characteristics enabling easy formation of oxygenates compared with other transition metals such as Ir, Pd and Pt, which hardly dissociate CO molecules. Even though transition metals such as cobalt and iron species can also easily dissociate CO molecules, the main products are higher hydrocarbons owing to a high hydrogenation activity as well.^8–10 Rh-based catalysts without addition of appropriate promoters are generally known to produce methane, with a lower selectivity for oxygenates as well.¹¹ An appropriate promoter such as K, Ti, V or Mn on the Rh-based catalysts can greatly improve catalytic performances with moderate selectivity for oxygenates.^12–14 Active Rh metal with manganese oxides as promoter on SiO₂ or MCM-41 revealed a higher activity and selectivity for C₂₊ oxygenates, especially for ethanol,¹⁵ and manganese species on the Rh/Al₂O₃ have also been reported as an effective promoter to improve CO dissociation as well as to enhance dispersion of Rh nanoparticles.¹⁶

Because of a relatively large pore mouth and specific surface area of ordered mesoporous materials, they have been widely used as support materials for catalytic reactions¹⁷ since the first synthesis of the ordered mesoporous MCM-41.¹⁸ In addition, the tungsten carbides also have many advantageous properties for support materials, such as a high melting point, superior hardness and good electrical conductivity and so on.^19–22 In addition, active-metal-modified ordered mesoporous carbons produced by incorporating metal nanoparticles into the mesoporous carbon matrix by a hard template method using ordered mesoporous silicas have been widely investigated.²³ In the present investigation, an ordered mesoporous tungsten carbide (W_xC), which mainly contained a carbon matrix, was synthesized by using a hard template method and catalytic properties of the bimetallic Rh–Mn supported mesoporous W_xC were investigated for the synthesis of oxygenates from syngas to elucidate the roles of W_xC and Rh–Mn distribution. Since, as far as we know, the influences of Mn promoter on the Rh/W_xC catalytic system have rarely been reported, Rh–Mn/W_xC was investigated to explain the reasons for its superior activity compared with other previously reported support materials and to verify the effects of Rh–Mn interaction and dispersion on selective oxygenate formation on the partially formed W_xC surfaces.

2. Experimental

2.1. Catalyst preparation and activity measurement

The ordered mesoporous tungsten carbide (W_xC) support was synthesized by impregnating a hard template of ordered mesoporous silica SBA-15 with carbon and tungsten precursors. In more detail, a hard template of SBA-15 was completely filled with an aqueous sulfuric acid solution mixed with sucrose as a source of carbon (C₁₂H₂₂O₁₁, Aldrich) and ammonium metatungstate precursor ((NH₄)₆[H₂W₁₂O₄₀], Aldrich) as a source of tungsten at a fixed carbon to tungsten ratio (C/W) of 10 to prepare partially formed W_xC in a mainly carbon matrix. The precursor-filled SBA-15 was dried in an oven at 100 °C for 6 h and subsequently dried at 160 °C for 6 h, and this pillaring procedure was repeated 2 times consecutively. For preparing the mesoporous W_xC support, typical carburization steps were carried out using a well-controlled temperature profile of 300, 600 and 900 °C with a ramping rate of 1 °C min⁻¹ for 5 h for each step under a flow of 5 vol% H₂/N₂. After the carburization step, the hard template SBA-15 was completely removed using 1.0 M NaOH aqueous solution several times, and the as-prepared W_xC support was dried at 110 °C overnight. For the subsequent preparation of bimetallic Rh–Mn-containing mesoporous W_xC (denoted Rh–Mn/W_xC), the precursors of the metals Rh and Mn were simultaneously deposited on the mesoporous W_xC pores by the incipient wetness impregnation method. In more detail, an aqueous solution of mixed precursors of rhodium (rhodium(III) chloride hydrate, RhCl₃·xH₂O, Aldrich) as well as manganese (manganese(II) nitrate hydrate, Mn(NO₃)₂·xH₂O, Aldrich) was prepared with a desired Rh/Mn ratio. The loading amount of Rh metal was fixed at 5 wt% on all Rh–Mn/W_xC catalysts, and the loading amount of Mn metal was varied from 0 to 10 wt% based on the total weight of W_xC support. The as-prepared Rh–Mn/W_xC with an average particle size of ∼60 μm and without significant mass-transport limitations was further dried at 70 °C in an oven overnight, and the as-prepared catalysts without calcination treatment were denoted Rh–Mn(y)/W_xC, where y represents the wt% Mn.

Catalytic activity on the Rh–Mn/W_xC was measured using a flow fixed-bed reactor having an outer diameter of 12.7 mm with 0.2 g catalyst. Before the reaction, the catalyst was reduced by using 10 vol% H₂ in N₂ at 300 °C for 5 h at atmospheric pressure. After cooling down the reduced sample, it was purged with syngas having a molar ratio of H₂/CO = 2, and the reaction was carried out at 300 °C and 5.0 MPa with a space velocity of 8000 L (kg_cat h)⁻¹ for more than 35 h on stream. The effluent gases from the reactor were analyzed by using on-line gas chromatography (YoungLin 6100GC) connected with a thermal conductivity detector (TCD) with a Carboxen 1000 capillary column for the analysis of nonflammable gases as well as a flame ionization detector (FID) with Plot-Q capillary column for the analysis of alcohols and hydrocarbons. The CO conversion and product distribution based on the carbon mol% (C-mol%) were calculated according to the following equations with an assumption of insignificant coke formation;

where M_i represents mol% of product i, n_i the number of carbons in the product i detected by FID, and M_CO represents mol% of CO in the feed gases detected by TCD. The productivity of C₁–C₄ alcohols (g (kg_cat h)⁻¹) was calculated as: (converted moles of CO) × Σ[(selectivity for each alcohol) × (molecular weight of each alcohol)]/[weight of catalyst (kg_cat) × (hours)].

2.2. Characterization of Rh–Mn/W_xC catalysts

Powder X-ray diffraction (XRD) patterns of the fresh (without calcination and before reaction) and used (after reaction) Rh–Mn/W_xC were obtained by using an X-ray diffractometer operating at 2 kW and 55 mA with Cu-Kα radiation of 0.15406 nm at a scanning rate of 5° min⁻¹ in the scanning range 2θ = 5–90° to elucidate the crystalline structures of the rhodium, manganese and tungsten carbides. The crystallite sizes of tungsten carbide (W₂C) and rhodium oxide (Rh₂O₃) were also calculated using the values of full width at half maximum (FWHM) of the most intense diffraction peaks with the help of Scherrer's equation.

N₂ adsorption–desorption isotherms of the fresh Rh–Mn/W_xC were measured at the liquid nitrogen temperature of −196 °C using a Tristar (Micromeritics) instrument. Surface area and pore volume were also calculated from the adsorption isotherm using the Brunauer–Emmett–Teller (BET) equation, and the Barrett–Joyner–Halenda (BJH) method was further applied to determine pore size distributions of the Rh–Mn/W_xC from the desorption isotherm.

Temperature-programmed reduction (TPR) profiles of the fresh Rh–Mn/W_xC were obtained using 0.05 g sample to evaluate the reducibility of the supported active metals Rh and Mn, and of the tungsten carbides as well. Before the TPR analysis, the sample was pretreated using He gas at 350 °C for 1 h at a gas flow rate of 30 mL min⁻¹. After pretreatment, 5 vol% H₂ in N₂ was passed over the catalyst at a flow rate of 30 mL min⁻¹ with a ramping rate of 10 °C min⁻¹ from 100 °C to 850 °C. The consumption of H₂ was measured by TCD after passing through a molecular sieve trap to remove water formed during the TPR experiment.

The amount of CO adsorbed by the fresh Rh–Mn/W_xC was measured by a CO chemisorption method using ASAP 2020C (Micromeritics). A 0.1 g sample was evacuated at 110 °C for 0.5 h, and subsequently reduced at 350 °C for 0.5 h under a pure H₂ flow. After purging at the same temperature, the reduced Rh–Mn/W_xC was analyzed at 35 °C under static conditions. The amount of adsorbed CO molecules was measured using a first isotherm curve by extrapolating to zero pressure with an assumption of a stoichiometry of 1.0 for CO/metal.

The mesoporous structures and crystallite sizes of the supported metals on the W_xC support were measured by field emission transmission electron microscopy (FE-TEM) using a JEOL JEM-2100F instrument operated at an accelerating voltage of 200 kV. In addition, the distribution of active metals was also measured by an energy dispersive spectroscopy (EDS) mapping method to confirm the dispersion of the supported metal components.

The oxidation states of Rh and Mn species together with their binding energies (BEs) and the surface concentrations on the used Rh–Mn/W_xC after air exposure were characterized using X-ray photoelectron spectroscopy (XPS) using a VG Multilab 2000 instrument with Mg/Al twin anodes as a source of X-ray radiation. The BE of C 1s (284.4 eV) was used as a reference BE to correct the BE shifts of other chemical species. The ratios of surface components were calculated using the integrated area of each characteristic peak, i.e. the ratios of Rh/C and Mn/C are represented by the respective ratios of integrated peak areas of Rh 3d_5/2 to C 1s and Mn 2p_3/2 to C 1s, and the Rh⁺/Rh⁰ ratio was obtained by using the deconvoluted area ratio of the main peak at ∼308 eV (Rh⁰) to a shoulder peak at a higher BE of around 309 eV (Rh⁺), respectively.

3. Results and discussion

3.1. Catalytic activity for direct synthesis of higher alcohols on Rh–Mn/W_xC

CO hydrogenation to higher alcohols on the mesoporous Rh–Mn/W_xC with an average particle size of ∼60 μm which shows an insignificant mass transport limitation^24,25 was carried out using syngas composed of H₂/CO = 2 : 1 at P = 5.0 MPa and T = 300 °C; CO conversion and product distribution are summarized in Table 1. Compared with other support materials such as W₂C, Al₂O₃, mesoporous SBA-16, and CMK-3, with a fixed weight ratio of Rh/Mn = 5/3, the mesoporous W_xC support interestingly gave a higher productivity of C₁–C₄ alcohols, as summarized in ESI Table S1.† All supported Rh–Mn catalysts except for those with the W_xC support, where the W_xC support itself revealed a little activity of hydrogenation to hydrocarbons as well, showed a higher CH₄ selectivity, above 56%, owing to a preferential CO hydrogenation activity through a possible methanation reaction on the active metal sites¹¹ in the present reaction conditions. However, active Rh–Mn supported on Al₂O₃ and SBA-15 supports showed a reasonable production rate of alcohols of around 40–63 g (kg_cat h)⁻¹ with a higher CO conversion of 40.7–46.5% owing to easy CO dissociation due to an appropriate Rh–support interaction,^12–16 and a lower CO conversion of around 6.5% on the ordered carbon-based mesoporous CMK-3 with Rh–Mn species was observed, as shown in ESI Table S1.† Even though Rh–Mn supported on the mesoporous W_xC support showed a lower CO conversion of around 8.1% compared with the other supports, a higher C₁₊ alcohol selectivity was observed, with a maximum production rate of 172 g (kg_cat h)⁻¹, which was strongly affected by the Mn content, as summarized in Table 1. In the case of the W₂C support (Table S1†), the catalyst showed a higher CO conversion of 89% without any formation of oxygenates owing to its lower specific surface area, below 5 m² g⁻¹, with a lower dispersion of Rh–Mn species. On the ordered mesoporous CMK-3 support having a similar structure to the mesoporous W_xC, a higher selectivity for methane and CO₂ was also observed.^15,16 We believe that the appropriate metal dispersion and interaction of the supported active Rh and Mn species with the mesoporous W_xC support are important to obtain a higher selectivity towards alcohols, and the present mesoporous W_xC support with close interactions between the active Rh and Mn nanoparticles and the surface tungsten species could possibly be a major reason for the high productivity of alcohols. As summarized in Table 1, mesoporous W_xC at an optimal weight ratio of Rh/Mn (Rh–Mn(3)/W_xC) increased the selectivity towards alcohols such as methanol and ethanol, mainly due to the well-known contribution of the Mn promoter to increasing the activity of CO adsorption and dissociation.¹¹ In the case of Rh/W_xC, CO conversion was found to be 11.4% with a larger hydrocarbon selectivity, above 45.4%, and a production rate of C₁–C₄ alcohols around 86.1 g (kg_cat h)⁻¹. However, a lower CO conversion of around 0.1% was observed on Mn(3)/W_xC, with a higher selectivity of byproducts such as aldehydes, around 10.7%, which suggests that CO hydrogenation to hydrocarbons, methanol and oxygenates (mainly aldehydes and acids) was mainly due to the Rh contribution instead of the Mn promoter and hydrogenation of the W_xC support itself. On the Rh–Mn/W_xC catalysts, an optimal 3 wt% Mn promoter on the Rh–Mn(3)/W_xC was responsible for higher selectivity towards alcohols, around 54.4 mol% (45.4 mol% methanol), with the highest productivity of C₁–C₄ alcohols being 171.8 g (kg_cat h)⁻¹, compared with other Rh–Mn/W_xC. In addition, all Rh–Mn/W_xC showed stable activity after 20 h on-stream till ∼35 h without significant deactivation, as shown in ESI Fig. S1.† The significant initial deactivation on Rh–Mn/W_xC, especially on the Rh–Mn(3)/W_xC, was attributable to the possible rearrangements of the exposed active Rh nanoparticles by interaction with the Mn and with the W_xC support as well. At a higher or lower Mn content, CO conversion and productivity of C₁–C₄ alcohols were decreased, and the trend of alcohol productivity showed a volcano-curve pattern in terms of Mn content on the Rh/W_xC. For example, the productivity of C₁–C₄ alcohols was found to be 21.8 g (kg_cat h)⁻¹ on Rh–Mn(1)/W_xC and 41.8 g (kg_cat h)⁻¹ on Rh–Mn(10)/W_xC, as summarized in Table 1. In general, the Mn promoter can generate appropriate interactions between Rh and Mn and the mesoporous W_xC surfaces and it significantly changed the productivity of higher alcohols by adjusting CO adsorption strengths as well. It is generally known that the Rh⁺ species is more active for undissociative CO activation than the Rh⁰ species to selectively form C₂₊ alcohols by CO hydrogenation.²⁶ Therefore, to verify the electronic states of the Rh species and its metal–support interaction such as reduction behaviors of Rh and its distribution on the Rh–Mn/W_xC, some related experiments such as CO chemisorption, TPR, XPS and TEM-EDS analysis were carried out and the roles of metal–support interactions on the mesoporous W_xC support were explained in terms of amount of adsorbed CO molecules, ratio of Rh⁺/Rh⁰, and reduction patterns of the Rh–Mn/W_xC. The synergistic effects of a partially reducible Rh species with a close interaction between Rh and Mn species on the mesoporous W_xC were well correlated with CO conversion and product distribution on the Rh–Mn/W_xC.

Table 1 Catalytic activity and product distribution on the Rh–Mn/W_xC catalysts^a

Catalysts	CO conv. (C-mol%)	Product distribution^b (C-mol%)								Productivity^d [g (kgcat h)^−1]
		CO2	C1	C2+	MeOH	EtOH	PrOH	BtOH	Others^c	C1–C4 alcohols
a Reaction was carried out under reaction conditions of T = 300 °C, P = 5.0 MPa, weight hourly space velocity (WHSV) = 8000 L (kgcat h)^−1, H2/CO = 2 with 0.2 g catalysts after reduction at 300 °C for 5 h.b C1, C2+, MeOH, EtOH, PrOH and BtOH stand for methane, light hydrocarbons, methanol, ethanol, propanol and butanol, respectively.c Others' stands for various oxygenated chemicals such as aldehydes and acids.d The productivity of C1–C4 alcohols (g (kgcat h)^−1) was calculated as: (converted moles of CO) × Σ[(selectivity for each alcohol) × (molecular weight of each alcohol)]/[weight of catalyst (kgcat) × (hours)].
Rh/WxC	11.4	29.1	35.4	10.0	15.2	2.5	1.1	0.01	6.69	86.1
Rh–Mn(1)/WxC	1.0	0	29.5	8.8	60.0	0	0	0	1.7	21.8
Rh–Mn(3)/WxC	8.1	15.6	23.6	5.3	45.4	7.2	1.8	0	1.10	171.8
Rh–Mn(6)/WxC	7.4	20.1	26.9	6.3	33.7	10.0	2.3	0.01	0.68	140.1
Rh–Mn(10)/WxC	4.4	14.5	54.5	9.4	4.6	12.5	1.9	0.05	2.55	41.8
Mn(3)/WxC	0.1	0	45.4	14.8	29.1	0	0	0	10.7	1.1

---

3.2. Active sites for the production of higher alcohols on Rh–Mn/W_xC

N₂ adsorption–desorption isotherms and pore size distributions of the fresh (uncalcined) Rh–Mn/W_xC are displayed in Fig. 1(A) and (B), respectively. Typical type IV isotherms were clearly observed, as shown in Fig. 1(A), which represent the mesoporous structures of the Rh–Mn/W_xC. The pore size distribution was measured by the BJH method from the desorption isotherm as shown in Fig. 1(B); the pore size showed a uniform mesopore diameter in the range 3–4 nm with a large pore size around 8 nm, which seems to be attributable to incomplete formation of mesoporous structures during the preparation step of W_xC using a hard template of SBA-15, especially on the Rh–Mn(1)/W_xC. The W_xC support having a C/W ratio of 10 itself showed a typical ordered mesoporous structure with an average pore diameter of 5.1 nm and specific surface area of 267.3 m² g⁻¹ with pore volume of 0.30 cm³ g⁻¹, as shown in ESI Fig. S2† and Table 2. However, with an increase of Mn content, the specific surface area of the Rh–Mn/W_xC gradually decreased from 84.5 m² g⁻¹ on Rh/W_xC to 13.9 m² g⁻¹ on Rh–Mn(10)/W_xC and similar trends in pore volume were observed, in the range 0.04–0.14 cm³ g⁻¹, owing to possible pore mouth blockages through the deposition of active metals. However, the average pore diameter increased inversely, in the range 6.5–12.5 nm, with an increase of Mn content owing to newly formed large pores, as shown in Fig. 1(B). For Mn(3)/W_xC, the surface area was somewhat large, with the value of 92.9 m² g⁻¹, and the pore volume and diameter were found to have intermediate values of 0.09 cm³ g⁻¹ and 6.5 nm, respectively. These variations of physical properties were attributed to mesopore blockages and newly formed inter-particular pores at a higher Mn content of the fresh Rh–Mn/W_xC, which can also alter the dispersion of active metals. Even though CO conversion on Rh–Mn/W_xC seems to be related with the surface area and pore volume, such as higher surface area with higher CO conversion, the active sites of Rh nanoparticles and their electronic states can be well related with CO conversion and product distribution, which were also significantly affected by these surface properties.

Fig. 1 N₂ adsorption–desorption isotherms (A) and pore size distribution (B) of the fresh Rh–Mn/W_xC catalysts.

Table 2 Characteristics of the Rh–Mn/W_xC catalysts before and after reaction

Catalysts	N2 sorption method			Adsorbed amount of CO from chemisorption (cm^3 g^−1)	H2 consumption from TPR (mmol g^−1)		BE (eV) from XPS (used catalysts)^a		Intensity ratios of elements (used catalysts)
	Surface area (m^2 g^−1)	Pore volume (cm^3 g^−1)	Average pore diameter (nm)		Rh^0 (<300 °C, T1)	Rh^+ (300–550 °C, T2)	Rh	Mn	Rh^+/Rh^0	Rh/C	Mn/C	Rh/Mn
a The binding energy (BE) values of Rh 3d5/2 and Mn 2p3/2 were obtained for the used Rh–Mn/WxC catalysts.
WxC	267.3	0.30	5.1	0.00	—	—	—	—	—	—	—	—
Rh/WxC	84.5	0.14	7.2	3.94	0.11	0.73	307.7	—	0.54	0.017	—	—
Rh–Mn(1)/WxC	56.6	0.11	6.5	2.27	0.03	0.43	307.4	640.5	0.40	0.014	0.026	0.53
Rh–Mn(3)/WxC	35.5	0.07	6.5	3.07	0.18	0.56	306.7	640.6	1.80	0.016	0.028	0.57
Rh–Mn(6)/WxC	36.1	0.08	9.1	1.95	0.15	0.45	307.7	641.1	1.54	0.017	0.035	0.49
Rh–Mn(10)/WxC	13.9	0.04	12.5	1.32	0.13	0.46	307.6	640.9	0.85	0.018	0.042	0.43
Mn(3)/WxC	92.9	0.09	6.5	0.00	—	—	—	641.4	—	—	0.017	—

---

XRD patterns of fresh and used Rh–Mn/W_xC are displayed in Fig. 2; all catalysts even after reaction exhibited four different characteristic diffraction peaks at 2θ = 40.4, 58.3, 73.2 and 87.0°, which are in accordance with the W₂C, WC and W phases formed by mixing tungsten carbides with metallic tungsten as well on the W_xC support.^27,28 However, peaks due to active metals such as Rh and Mn species were not clearly observed, owing to the metal nanoparticles being highly dispersed on the Rh–Mn/W_xC and also to their amorphous character.²⁹ Interestingly, the main crystalline tungsten phases on the fresh Rh–Mn/W_xC were found to be W₂C crystallites, with a larger peak intensity compared with WC phases, which seems to be attributable to the low thermal treatment temperature used for preparing tungsten carbide, as shown in Fig. 2(A). These phases on the W_xC support were stably preserved even after reaction, owing to the high temperature required for degradation of the W₂C phase to metallic W species, as confirmed in the following TPR analysis. However, the crystallites of Rh₂O₃ and Mn₂O₃ were observed on all the used Rh–Mn/W_xC catalysts, with diffraction peaks²⁹ at 2θ < 40° as shown in Fig. 2(B), which strongly suggests a significant aggregation of Rh and Mn metals during the reaction. Although the aggregation of Rh nanoparticles was found to be significant, the aggregated size of Rh crystallites was below 2 nm according to the crystallite size calculated with the help of the Scherrer equation using the values of FWHM after reaction; this was further verified by TEM analysis.

Fig. 2 XRD patterns of Rh–Mn/W_xC for (A) fresh and (B) used catalysts.

The comparative reduction patterns of the Rh and Mn metals on other supports such as SBA-15, CMK-3 and mesoporous W_xC at a fixed 5 wt% Rh and 3 wt% Mn are shown in ESI Fig. S3;† the characteristic reduction peaks of Rh metals were observed at around 130 °C for Rh–Mn/SBA-15, which was similar to the previous work.³⁰ However, much higher temperature reduction peaks above 500 °C were observed on Rh–Mn/CMK-3 and Rh–Mn/W_xC, which suggests the possible formation of strongly interacted Rh nanoparticles or the possible hydrogenation of carbon species from the W_xC support.^28,30 These results indicate that tungsten carbides such as W₂C and WC can prohibit a facile reduction of Rh nanoparticles by forming strongly interacted Rh on the W_xC support. TPR patterns for the Rh–Mn/W_xC with different Mn contents are also displayed in Fig. 3, and the reduction peaks at around 250 °C can be assigned to the partially reduced and strongly interacted Rh nanoparticles on the W_xC support. The reduction peaks at above 500 °C were in accordance with the reduction patterns of Rh nanoparticles with a high oxidation state,³¹ and corresponded to the formation of strongly interacted Rh nanoparticles on the surfaces. Compared with Rh/W_xC, Rh–Mn/W_xC showed characteristic reduction peaks of the Rh nanoparticles at around 265 and 370 °C, which shifted to higher temperature with an increase of Mn content as well. However, the reduction peak of Rh nanoparticles was not clearly observed on Rh–Mn(10)/W_xC, possibly owing to interaction of Rh nanoparticles with Mn species. A small reduction peak of Mn was also observed at around 430 °C on Mn/W_xC. The intermediate reduction temperatures of around 265–370 °C could also originate from the partially reduced Rh⁺ nanoparticles by forming a stronger metal–support interaction, which was known to be active for formation of oxygenates on the Rh-based catalysts.^15,16,30 Especially on the Rh–Mn(3)/W_xC, broad reduction peaks at about 405 and 460 °C were observed, and they can be attributed to the partially reduced Rh nanoparticles through the formation of closely interacted Rh–Mn species at an optimal Rh/Mn ratio. The peak at a much higher reduction temperature, above 600 °C, seems to be attributable to the possible hydrogenation of W₂C species on the W_xC support. An increased interaction between Rh and Mn on the W_xC support can lead to selective formation of surface Rh⁺ species rather than Rh⁰ on Rh–Mn/W_xC. However, an increased Mn content above 6 wt% can induce facile reduction of Rh nanoparticles by forming separate large Rh and Mn nanoparticles with abundant Rh⁰ species rather than Rh⁺ on the Rh–Mn/W_xC. As summarized in Table 1, the variation of hydrogen consumption in the two ranges <300 °C (T₁) and 300–550 °C (T₂), which can be respectively assigned to metallic Rh⁰ nanoparticles and strongly interacted Rh⁺ species with higher oxidation states, showed characteristic trends for CO conversion and alcohol selectivity. With an increase of hydrogen consumption assigned to the T₁ peak, CO conversion was increased especially on the Rh–Mn(3)/W_xC, with the value of 0.18 mmol g⁻¹, owing to facile hydrogenation of CO molecules. The hydrogen consumption assigned to the T₂ peak decreased with increasing Mn content on the Rh–Mn/W_xC, which suggests too strongly interacted Rh nanoparticles with a high oxidation state and difficult reducibility, especially on Rh/W_xC, with the value of 0.73 mmol g⁻¹. The relative sizes of the T₁ and T₂ peaks from bulk properties seem to be related with CO conversion and productivity of alcohols; however, the ratio of the surface-exposed metallic Rh to the partially oxidized Rh nanoparticles as confirmed by XPS analysis seems to be directly related with catalytic activity.

Fig. 3 TPR patterns of fresh Rh–Mn/W_xC catalysts.

To further verify the distributions and interactions of the Rh and Mn nanoparticles and their oxidation states on the W_xC support, XPS analyses on used Rh–Mn/W_xC were carried out; XPS spectra are displayed in Fig. 4 and the results are summarized in Table 2 as well. The BEs of Rh 3d_5/2 and Mn 2p_3/2 were found to be 306.7–307.7 eV and 640.6–641.4 eV, respectively. As shown in the ESI Fig. S4,† the BE of Rh 3d_5/2 on the fresh Rh–Mn(3)/W_xC was found to be at around 310.1 eV, which was in accordance with Rh³⁺ species.^31,32 This BE of Rh species was attributed to the use of RhCl₃ precursor on the W_xC without calcination treatment of the fresh Rh–Mn/W_xC. However, on the used Rh–Mn/W_xC, the BEs of Rh 3d_5/2 shifted to much lower values at around 306.7–307.7 eV for Rh⁰ species and 308.5–309.5 eV for Rh⁺ species owing to the presence of partially reduced Rh nanoparticles with different interactions between Rh and Mn species or Rh and W_xC species, without significant variations of BEs of Mn species. Interestingly, a slight shift to a higher BE with an increase of Mn content on the Rh–Mn/W_xC originated from the Mn contribution to enhancing the oxidation states of Rh nanoparticles.^30,33,34 Moreover, the ratio of Rh⁺/Rh⁰ on the used Rh–Mn/W_xC revealed that the surface oxidation states of Rh nanoparticles were higher on the most active Rh–Mn(3)/W_xC, with Rh⁺/Rh⁰ = 1.80, and the ratio was decreased on other Rh–Mn/W_xC with lower or higher Mn content, being in the range Rh⁺/Rh⁰ = 0.40–1.54, as summarized in Table 2. Therefore, Rh–Mn/W_xC containing abundant Rh⁺ species with reduced Rh nanoparticles seems to be effective for increasing oxygenate formation,^15,16 which showed a similar trend to the TPR results for the amount of Rh⁰ and Rh⁺ species even though the absolute values were not exactly same owing to the surface sensitive characterization nature of XPS analysis. Mn species have an important role in maintaining higher oxidation states of Rh nanoparticles as well. As shown in ESI Fig. S4,† the BE of Mn 2p_3/2 on fresh Rh–Mn(3)/W_xC was observed at 641.8 eV, which suggests the presence of Mn³⁺ species mainly in the form of Mn₂O₃.³⁵ After the reaction, the BEs of Mn 2p_3/2 on the used Rh–Mn/W_xC shifted a little bit lower to around 640.6–641.4 eV without significant variation, which corresponds to Mn²⁺ species mainly in the form of MnO.³⁵ The partial reduction of Mn oxides can alter the oxidation states of Rh nanoparticles as well by closely interacting with Rh nanoparticles. In addition, the C 1s peak at around 284.6 eV, which can be assigned to the characteristic graphitic carbon species,³⁵ was not significantly altered for any of the used Rh–Mn/W_xC except for the most active Rh–Mn(3)/W_xC owing to its high dispersion of intimately interacted Rh and Mn nanoparticles as shown in ESI Fig. S5.† Furthermore, the surface concentrations of each element, Rh, Mn, and carbon, were verified by comparing the ratios of Rh/C, which did not vary significantly, Mn/C, which increased with increase of Mn content, and Rh/Mn, as summarized in Table 2. Even though Rh/C ratios (in the range 0.014–0.018) did not alter significantly, which also confirmed a similar amount of Rh deposition on the Rh–Mn/W_xC by the incipient wetness impregnation method, Rh/Mn ratios altered greatly, in the range 0.43–0.57. The supported Rh nanoparticles were greatly exposed on the outer surfaces of the most active Rh–Mn(3)/W_xC, which was confirmed by a high value of Rh/Mn = 0.57. Therefore, the dominantly exposed active Rh surfaces with higher oxidation states due to closely interacting with Mn oxides on the Rh–Mn(3)/W_xC were mainly responsible for higher CO conversion and oxygenate formation. These superior surface properties of Rh–Mn(3)/W_xC were attributed to less aggregation of Rh nanoparticles with higher oxidation state, and could also originate from the interaction of Mn promoter at an optimal content as well as the interaction with W and W₂C species by effectively prohibiting the aggregation of Rh nanoparticles even after reaction, as confirmed by TPR as well as XPS analysis.

Fig. 4 XPS spectra of used Rh–Mn/W_xC for (A) Rh 3d_5/2 and (B) Mn 2p_3/2.

The assumption of close and strong interactions of Rh and Mn species on the W_xC support was further verified by CO chemisorption and TEM analysis. As summarized in Table 2, the amount of adsorbed CO molecules on the Rh/W_xC was found to be higher, with a value of 3.94 cm³ g⁻¹, than for other catalysts; these CO molecules can be adsorbed selectivity on the Rh nanoparticles. However, the larger amount of CO adsorption mainly leads to the formation of methane selectively on the Rh/W_xC owing to the abundant presence of fully reduced Rh nanoparticles as verified by the Rh⁺/Rh⁰ value from XPS analysis. An appropriate amount of adsorbed CO molecules, with the value of 3.07 cm³ g⁻¹, and highly exposed Rh nanoparticles with partially reduced states on the Rh–Mn(3)/W_xC were responsible for improved productivity of higher alcohols.^15,16 The distribution of Rh and Mn species on the fresh and used Rh–Mn(3)/W_xC was further verified by TEM analysis, and the TEM images and EDS mapping results are displayed in Fig. 5. As seen in the TEM images of the fresh and used Rh–Mn(3)/W_xC, oxides of the active metals Rh and Mn were homogeneously dispersed on the fresh Rh–Mn(3)/W_xC (not the reduced one) with an average particle size of 2–4 nm, and the Rh nanoparticles were more homogeneously dispersed after the reaction, as shown in Fig. 5(A) and (B). Interestingly, based on the TEM–EDS mapping analysis of the fresh and used Rh–Mn(3)/W_xC, even after the reaction the active metal crystallites of Rh and Mn oxides were dispersed on the mesoporous W_xC support, with intimate interactions between Rh and Mn species, but with a little bit of aggregation of metallic W species, as shown in Fig. 5(C) and (D). The homogenous distributions of Rh and Mn nanoparticles with partially reduced Rh nanoparticles due to a strong interaction with the W_xC support an Mn species could be a major reason for the high activity and large production rate of C₁–C₄ alcohols on the Rh–Mn(3)/W_xC without significant deactivation as well.

Fig. 5 TEM and EDS mapping images of (A and C) fresh Rh–Mn(3)/W_xC and (B and D) used Rh–Mn(3)/W_xC.

In summary, the observed higher CO conversion and production rate of alcohols on the Rh–Mn(3)/W_xC was attributed to the strong interaction between Rh nanoparticles and W_xC support forming partially reduced and high oxidation state of Rh species. The higher oxidation states of the partially reduced Rh nanoparticles were also affected by intimately distributed Mn oxides on the Rh–Mn(3)/W_xC surfaces. The ratio of Rh⁺/Rh⁰ and reducibility of Rh nanoparticles measured by XPS and TPR analyses together with TEM-EDS analysis on the Rh–Mn(3)/W_xC were responsible for the higher catalytic performance. The synergistic effects of the Rh nanoparticles with Mn species on the W_xC support, which adjusted the oxidation states of the Rh nanoparticles, could be an important variable for obtaining higher productivity of oxygenates at an optimal content of Mn on the Rh/W_xC catalytic systems.

4. Conclusions

The effects of an optimal content of Mn on Rh/W_xC with ordered mesoporous structure were verified to elucidate the different product distribution and CO conversion for direct synthesis of mixed alcohols by CO hydrogenation of syngas. An optimal Mn content of 3 wt% on the Rh/W_xC revealed a superior CO conversion and high productivity of higher alcohols. The enhanced catalytic activity at an optimal Mn content was attributed to a close interaction of Rh and Mn nanoparticles having partially reduced Rh oxides resulting in enhancement of the adsorption capacity of CO molecules. The synergistic effects of the active metal Rh and Mn nanoparticles on the novel W_xC support were the key control factor for obtaining higher productivity of oxygenates by adjusting the Rh and Mn distribution and strong metal–support interaction as well.

Acknowledgements

The authors acknowledge the financial support from the National Research Foundation of Korea (NRF) grants funded by the Korea government (NRF-2014R1A1A2A16055557 and NRF-2016M3D3A1A01913253). This work was also supported by an institutional program grant (2E24834-14-048) from the Korean Institute of Science and Technology. The present work was supported by the R&D Center for Valuable Recycling (Global-Top R&D Program) of the Ministry of Environment of Korea (Project No. E616-00140-0602-2).

References

1. P. Robert, J. Catal., 2015, 329, 87–94.
2. J. J. Spevey and A. Egbebi, Chem. Soc. Rev., 2007, 36, 1514–1528.
3. J. Liu, R. Tao, Z. Guo, J. Regalbuto, C. Marshall, R. Klie, J. Miller and R. Meyer, ChemCatChem, 2013, 5, 3665–3672.
4. Y. Yang, L. Wang, K. Xiao, T. Zhao, H. Wang, L. Zhong and Y. Sun, Catal. Sci. Technol., 2015, 5, 4224–4232.
5. G. Liu, D. Pan, T. Niu, A. Cao, Y. Yu and Y. Liu, RSC Adv., 2015, 5, 31637–31647.
6. X. Q. Qiu, N. Tsubaki, K. Fujimoto and Q. M. Zhu, Fuel Process. Technol., 2004, 85, 1193–1200.
7. T. Matsuzaki, K. Takeuchi, T. Hanaoka, H. Arawaka and Y. Sugi, Appl. Catal., A, 1993, 105, 159–184.
8. M. Bwoker, Catal. Today, 1992, 15, 77–100.
9. M. A. Vannice, J. Catal., 1975, 37, 449–461.
10. M. Mavrikakis, M. Baumer, H. F. Freund and J. K. Norskov, Catal. Lett., 2002, 81, 153–156.
11. V. Subramani and S. K. Gangwal, Energy Fuels, 2008, 22, 814–839.
12. Y. Wang, H. Y. Luo, D. B. Liang and X. H. Bao, J. Catal., 2000, 196, 46–55.
13. J. Hu, Y. Wang, C. Cao, D. C. Elliott, D. J. Stevens and J. F. White, Catal. Today, 2007, 120, 90–95.
14. X. Pan, Z. Fan, W. Chen, Y. Ding, H. Luo and X. H. Bao, Nat. Mater., 2007, 6, 507–511.
15. H. Ma, Z. Yuan, Y. Wang and X. Bao, Surf. Interface Anal., 2001, 32, 224–227.
16. M. Ojeda, M. L. Granados, S. Rojas, P. Terreros, F. J. Garcıa-Garcıa and J. L. G. Fierro, Appl. Catal., A, 2004, 261, 47–55.
17. A. Taguchi and F. Schuth, Microporous Mesoporous Mater., 2005, 77, 1–45.
18. J. S. Beck, C. T. W. Chu, I. D. Johnson, C. T. Kresge, M. E. Leonowicz, W. J. Roth and J. W. Vartuli, WO Patent 91/11390, 1991.
19. J. S. Moon, Y. W. Lee, S. B. Han and K. W. Park, Int. J. Hydrogen Energy, 2014, 39, 7798–7804.
20. V. Nikolova, I. Nikolov, P. Andreev, V. Najdenov and T. Vitanov, J. Appl. Electrochem., 2000, 30, 705–710.
21. L. J. Prakash, Int. J. Refract. Met. Hard Mater., 1995, 13, 257–264.
22. A. R. Ko, J. K. Oh, Y. W. Lee, S. B. Han and K. W. Park, Mater. Lett., 2011, 65, 2220–2223.
23. X. Song, Y. Ding, W. Chen, W. Dong, Y. Pei, J. Zang, L. Yan and Y. Lu, Catal. Commun., 2012, 19, 100–104.
24. R. J. Madon and M. Boudart, Ind. Eng. Chem. Fundam., 1982, 21, 438–447.
25. S. C. Baek, S. H. Kang, J. W. Bae, Y. J. Lee, D. H. Lee and K. Y. Lee, Energy Fuels, 2011, 25, 2438–2443.
26. S. C. Chuang, R. W. Stevens and R. Khatri, Top. Catal., 2005, 32, 225–232.
27. A. Kuma, K. Singh and O. P. Pandey, Phys. E, 2010, 42, 2477–2483.
28. C. B. Rodella, D. H. Barrett, S. F. Moya, S. J. A. Figueroa, M. T. B. Pimenta, A. A. S. Curveol and V. T. Silva, RSC Adv., 2015, 5, 23874–23885.
29. M. Sun, B. Lan, T. Lin, G. Cheng, F. Ye, L. Yu, X. Cheng and X. Zheng, CrystEngComm, 2013, 15, 7010–7018.
30. Y. Wang, Z. Song, D. Ma, H. Luo, D. Liang and X. Bao, J. Mol. Catal. A: Chem., 1999, 149, 51–61.
31. M. J. Ledoux, C. Pham-Huu and J. D. Guille, J. Catal., 1992, 134, 383–398.
32. Y. V. Larichev, O. V. Netskina, O. V. Komova and V. I. Simagina, Int. J. Hydrogen Energy, 2010, 35, 6501–6507.
33. L. Han, D. Mao, J. Yu, Q. Gue and G. Lu, Appl. Catal., A, 2013, 454, 81–87.
34. P. Reyes, I. Concha, G. Pecchi and J. L. G. Fierro, J. Mol. Catal. A: Chem., 1998, 129, 269–278.
35. J. F. Moulder, W. F. Stickle, P. E. Sobol and K. D. Bomben, Handbook of X-ray Photoelectron Spectroscopy: A Reference Book of Standard Spectra for Identification and Interpretation of XPS Data, Physical Electronics Division, Perkin-Elmer Corp., 1995.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra22144k

---