Biomass-based furfural hydrogenation to furfuryl alcohol



Selective Production of Furfuryl Alcohol

via Gas Phase Hydrogenation of Furfural over Au/Al2O3

Maoshuai Li, Yufen Hao, Fernando Cárdenas-Lizana and Mark A. Keane*

Chemical Engineering, School of Engineering and Physical Sciences, Heriot-Watt University, Edinburgh EH14 4AS, Scotland

*corresponding author

Tel: +44(0)131 451 4719, e-mail: M.A.Keane@hw.ac.uk

Abstract

We report exclusive continuous gas phase hydrogenation of furfural to furfuryl alcohol over Au/Al2O3 (∆Ea = 45 kJ mol-1). Characterisation measurements demonstrate the formation of Auδ- (from XPS analysis) nanoparticles (1-8 nm, mean = 4.3 nm) with H2 chemisorption capacity of 318 µmol gAu-1 (at 413 K). Chemical kinetic control was established by experimental variation of contact time, catalyst/furfural feed ratio and temperature. At T ≥ 473 K, hydrogenolysis to 2-methylfuran predominated. Selective TOF (298 h-1 at 433 K) over Au/Al2O3 exceeded values reported for benchmark Cu catalysts at higher temperature (453-498 K).

Keywords: Gas phase selective hydrogenation; furfural; furfuryl alcohol; Au/Al2O3.

1. Introduction

Furfuryl alcohol finds application in the manufacture of resins, rubbers and fibres [1] and can be produced by selective hydrogenation of furfural, a biomass derived heterocyclic aldehyde that can serve as a renewable, non-petroleum based raw material. Decarbonylation (path (I)), furan ring reduction (paths (II), (IV) and (VI)) and hydrogenolysis (path (V)) generate a range of by-products (Fig. 1). Noble metals (Pd and Pt) catalyse decarbonylation to furan, hydrogenation to tetrahydrofurfuryl alcohol and hydrogenolysis to 2-methylfuran [2,3]. Incorporation of Ce to Ni-B polarised/activated the carbonyl function resulting in preferential formation of furfuryl alcohol [4]. Taking an overview of the literature, Cu catalysts are the most selective to furfuryl alcohol [1,5-8]. Batch liquid phase reaction has been predominantly conducted at elevated H2 pressure (10-20 bar) [1,9]. A move to continuous processes at atmospheric pressure presents advantages in terms of product/catalyst separation, reduced downtime and higher throughput [10]. In continuous operation, Liu et al. [5] reported 90% selectivity to furfuryl alcohol over commercial copper chromite with the generation of toxic Cr2O3 waste. Application of Cu/MgO [6,7] and Cu/C [8] can overcome this limitation and these systems are used as the benchmark in this study. Supported Au nano-particles have shown unique selectivity for –C=O reduction in the presence of other functionalities (e.g. C=C) [11]. We could not find any reported application of supported Au in furfural hydrogenation beyond the observation by Hong et al. [12] of negligible activity (1% conversion) for Au/SiO2 in liquid phase reaction at high pressure (10 bar). We should note the work of Ohyama et al. [13] who reported preferential alcohol formation from 2-hydroxylmethyl-5-furfural over Au/Al2O3 in batch operation at 38 bar. In previous work [14] we demonstrated full selectivity in the gas phase reduction of benzaldehyde to benzyl alcohol over Au/Al2O3. Here we assess for the first time continuous hydrogenation of furfural over Au/Al2O3 and compare performance against benchmark supported Cu.

2. Experimental

2.1. Materials and catalyst preparation

Au/Al2O3 was prepared by deposition-precipitation using urea (Riedel-de Haën, 99%) as basification agent. An aqueous solution of urea (100-fold urea excess) and HAuCl4 (Sigma-Aldrich, 99%, 4.4 ( 10-6 mol cm-3, 400 cm3) was added to γ-Al2O3 (Puralox, Condea Vista, 30 g). The suspension was heated (2 K min-1 to 353 K) under constant stirring and the solid separated by filtration, washed with distilled water until chlorine free, dried in He (45 cm3 min-1) at 373 K (2 K min-1) for 5 h and sieved to mean particle diameter = 75 μm. Catalyst activation by temperature programmed (2 K min-1 to 603 K) reduction (TPR in 60 cm3 min-1 H2) ensured metal precursor transformation (Au3+→Au0) [14]. The activated sample was passivated in 1% v/v O2/He for 1 h at ambient temperature for ex situ characterisation.

2.2. Catalyst characterisation

Gold content was measured by atomic absorption spectroscopy (Shimadzu AA-6650 spectrometer) from the diluted extract in aqua regia. X-ray diffractograms (XRD) were recorded on a Bruker/Siemens D500 incident X-ray diffractometer using Cu Kα radiation with a scan rate of 0.02º step-1 over the range 20º ≤ 2θ ≤ 80º. The diffractogram was identified against the JCPDS-ICDD reference standards, i.e. Au (04-0784) and γ-Al2O3 (10-0425). Nitrogen physisorption was performed on the Micromeritics Gemini 2390p system and total specific surface area (SSA) calculated using the standard BET method with cumulative pore volume/pore radius from BJH analysis. Hydrogen chemisorption (pulse 10 μl titration at 298 K and 413 K) following TPR (as above) was conducted on the CHEM-BET 3000 (Quantachrome) unit with data acquisition/manipulation using the TPR WinTM software; there was no measurable H2 uptake on Al2O3. X-ray photoelectron spectroscopic (XPS) analysis was performed on a VG ESCA spectrometer equipped with monochromatised Al Kα radiation (1486 eV) under ultra-high vacuum (99%) solution to the reactor via a glass/teflon air-tight syringe and teflon line using a microprocessor controlled infusion pump (Model 100 kd Scientific) with a co-current flow of excess H2; GHSV = 5 × 103 – 3 × 104 h-1. Catalyst mass to inlet furfural molar feed rate (W/F) spanned the range 100 - 300 g mol-1 h. Passage of furfural in a stream of H2 through the empty reactor or over Al2O3 did not result in any detectable conversion. The reactor effluent was condensed in a liquid nitrogen trap for subsequent analysis by capillary GC (Perkin-Elmer Auto System XL) [14]. All gases (O2, H2, N2 and He) were of high purity (BOC, >99.98%). Furfural fractional conversion (X) is defined by

[pic] (1)

and selectivity (S) to product (j) is given by

[pic] (2)

Catalytic activity is quantified in terms of furfural consumption rate (R) obtained from

[pic] (3)

where initial fractional conversion (X0) was extracted from time on-stream measurements [14]. Turnover frequency (TOF, rate per active site) was calculated using

[pic] (4)

where nAu is the number moles of Au in the catalyst bed and D is the Au dispersion determined by STEM [14]. Repeated reactions with different samples from the same batch of catalyst delivered raw data reproducibility and carbon mass balance within ±5%.

3. Results and discussion

3.1. Catalyst characterisation

Analysis by XRD (diffractogram not shown) revealed diffraction peaks at 2θ = 37.6°, 39.5°, 45.9° and 67.0° due to γ-Al2O3. There was no detectable signal for metallic Au (2θ = 38.1°, 44.3°, 64.6° and 77.5), suggesting Au particle size below XRD ( 0.6 characteristic of mesoporous materials (type IV IUPAC) [15]. This is consistent with the mesoporosity of γ-Al2O3 (from Puralox) with mean pore radii in the 30-41 Å range [16]. A decrease in total surface area, pore volume and radius was observed for Au/Al2O3 (Table 1) relative to the starting support (SSA = 191 m2 g-1, pore volume = 0.45 cm3 g-1 and pore radius = 31 Å), which can be linked to partial pore filling during catalyst preparation [17]. TPR of Au/Al2O3 (not shown) generated a positive H2 consumption signal with a temperature maximum (Tmax) at 451 K where H2 consumption (Table 1) matched that (1.5 mol molAu-1) required for the reduction of the Au3+ precursor to metallic Au. The XPS spectrum over the Au 4f BE region (Fig. 2(II)) exhibited two peaks (BE = 83.4 eV and 87.0 eV) corresponding to 4f7/2 and 4f5/2 levels, respectively. The Au 4f7/2 BE is lower than that characteristic of metallic Au (84.0 eV) [18], suggesting electron donation from the support to generate Auδ- [18]. STEM analysis (Fig. 2(III)) confirmed the presence of nano-sized gold particles (1-8 nm, Fig. 2(IV)) with a mean of 4.3 nm. Hydrogen chemisorption on Au/Al2O3 is an activated process and uptake under reaction conditions (413 K, Table 1) was appreciably higher (9-fold) than that recorded at ambient temperature. The characterisation results demonstrate formation of nano-scale Auδ- particles with significant (318 μmol gAu-1) H2 uptake under reaction conditions.

3.2. Catalytic response

Representative time on-stream conversion and selectivity profiles are shown in Fig. 3(I). An initial decrease in conversion was observed with the attainment of steady state activity after 6 h on-stream. Furfural was solely converted to the target furfuryl alcohol over Au/Al2O3 (at 413 K), i.e. exclusive reduction of the carbonyl group (Fig. 1, path (III)). This is a significant result in the light of the negligible activity over Au/SiO2 in batch mode [12] and the reported formation of by-products over Pd [2], Pt [3] and Ni [19] catalysts. A meaningful evaluation of the catalytic response requires that the reaction is operated under kinetic control, free heat/mass transfer constraints. These possible limitations were assessed using standard diagnostic tests [17] and the results are presented in Fig. 3(II). Furfural consumption rate increased with contact time (at τ < 0.36 s, (IIA)), indicative of interparticle transport limitations that inhibit hydrogenation. Rate was insensitive to contact time at τ ≥ 0.36 s, demonstrating minimal external diffusion constraints; τ was set at 0.36 s in subsequent tests. Variation in catalyst mass to furfural feed rate (W/F) was used to probe the effect of possible reactant concentration gradients on performance. Hydrogenation rate (at X ≤ 0.9) displayed a linear correlation with respect to W/F (IIB) and we can discount contributions due to internal and/or external mass transport on the observed activity. The applicability of pseudo-first order kinetics for hydrogenation of substituted aldehydes over supported Au has been established elsewhere [20]. The Arrhenius plot (IIC) using the extracted rate constants (k) delivered an apparent activation energy (∆Ea = 45 kJ mol-1) close to that (50 kJ mol-1) reported for reaction over Cu/SiO2 [21] and higher than the reported furfural adsorption energy ( ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download