Selective Transfer Hydrogenation of Biomass ... - ACS Publications

May 26, 2017 - KEYWORDS: copper catalyst, furfural, 5-hydroxymethylfurfural, methanol, transfer hydrogenation. □ INTRODUCTION. The depletion of foss...
12 downloads 15 Views 3MB Size
Research Article pubs.acs.org/journal/ascecg

Selective Transfer Hydrogenation of Biomass-Based Furfural and 5‑Hydroxymethylfurfural over Hydrotalcite-Derived Copper Catalysts Using Methanol as a Hydrogen Donor Jun Zhang† and Jinzhu Chen*,†,§ †

Guangzhou Institute of Energy Conversion, Chinese Academy of Sciences. CAS Key Laboratory of Renewable Energy, Guangdong Provincial Key Laboratory of New and Renewable Energy Research and Development. No. 2 Nengyuan Rd, Wushan, Tianhe District, Guangzhou 510640, People’s Republic of China § College of Chemistry and Materials Science, Jinan University. No. 601 Huangpu Avenue West, Tianhe District, Guangzhou 510632, People’s Republic of China S Supporting Information *

ABSTRACT: A series of inexpensive copper-based catalysts, derived from hydrotalcite precursors, were screened for selective transfer hydrogenation of biomass-based furfural (FFR) to furfuryl alcohol (FA) and 2-methyl furan (MF), with methanol as both the solvent and a hydrogen donor. The specific textural characteristics of the prepared catalysts were characterized by various techniques such as X-ray diffraction (XRD), scanning electron microscopy (SEM), transmission electron microscopy (TEM), thermogravimetric analysis/differential thermal analysis (TGA/DTA), nitrogen physisorption, temperature-programmed desorption in an ammonia atmosphere (NH3-TPD), temperature-programmed reduction in a hydrogen atmosphere (H2-TPR), infrared (IR) spectra of pyridine adsorption, and X-ray photoelectron spectroscopy (XPS). The copper catalyst showed excellent transfer hydrogenation selectivity toward FFRto-FA transformation by giving a FA yield of 94.0 mol % at 473 K in methanol. Treating the copper catalyst with H2/N2 at 773 K led to the formation of a metallic Cu species in the activated sample. Interestingly, the resulting activated copper catalyst favored FFR-to-MF transformation by producing MF yield of 94.1 mol % at 513 K. The FFR-to-MF transformation involved a tandem reaction of FFR hydrogenation and successive FA hydrogenolysis; meanwhile, a synergistic effect between metallic Cu species and acidic sites on the catalyst surface played a key role in MF formation. In addition, the activated copper catalyst exhibited outstanding performance toward another biomass-related important conversion of 5-hydroxymethylfurfural (HMF) to 2,5dimethylfuran (DMF) with the desired DMF yield of 96.7 mol %. The recycling experiments revealed that both the copper and the activated copper catalysts maintained good activity and stability after five-time recycling. The research thus highlights a new perspective for a green, efficient, and biomass-related hydrogenation reaction using the readily available and inexpensive copper− methanol catalytic system, while, without any external hydrogen supplies. KEYWORDS: copper catalyst, furfural, 5-hydroxymethylfurfural, methanol, transfer hydrogenation



INTRODUCTION The depletion of fossil carbon resources, huge energy demands, and environmental deterioration have encouraged the development of alternative renewable sources, as well as novel approaches to the sustainable supply of fuels and fine chemicals.1−3 In this regard, the catalytic conversion of lignocellulosic biomass into a wide variety of biofuels and chemicals has recently gained increasing interests. A recent review critically discussed nanoscale metal catalysts for valorization of biomass-related cellulose, chitin, lignin, and lipids.3 Among these biomass-derived platform molecules, furfural (FFR), primarily obtained by acid-catalyzed dehydration of biomass-derived pentoses,4−6 was identified as a promising building block for valuable products. Particularly, selective transformation of FFR provides an important route to prepare a wide range of target derivatives, such as furfuryl © 2017 American Chemical Society

alcohol (FA), tetrahydrofurfuryl alcohol (THFA), 2-methyl furan (MF), and cyclopentanone.7−9 These FFR-derived compounds are broadly employed in the polymer industry, synthetic fibers, rubber resins, the energy chemical industry, and so on.10−12 Conventional process for FFR hydrogenation requires the use of pressurized molecular hydrogen in combination with various metal catalysts, such as Pd/SiO2, Ru/C, Pt/C, Cu/SBA15, Cu-MgO, Mo2C, and Ni−Fe−B.13−20 Generally, the precious-metal catalysts exhibit appreciable activity toward FFR conversion, and FFR can be easily and quantitatively consumed after the reaction. However, several noticeable Received: March 13, 2017 Revised: May 5, 2017 Published: May 26, 2017 5982

DOI: 10.1021/acssuschemeng.7b00778 ACS Sustainable Chem. Eng. 2017, 5, 5982−5993

Research Article

ACS Sustainable Chemistry & Engineering

Inspired by the above findings, herein, we reported the CTH of FFR to FA and MF with methanol as both a solvent and hydrogen donor, using inexpensive copper-based catalysts derived from hydrotalcite precursors. A systematic characterization of the catalysts including structural properties, copper valence state, surface acid amount, and surface acid type were conducted; moreover, various reaction parameters on the CTH reaction were investigated. The copper catalyst showed excellent selectivity toward FFR-to-FA transformation by giving a FA yield of 94.0 mol %. Treatment the copper catalyst with H2/N2 led to the formation of metallic Cu species in the activated sample. Interestingly, the resulting activated copper catalyst favored FFR-to-MF transformation by producing a MF yield of 94.1 mol %. The FFR-to-MF transformation involved a tandem reaction of FFR hydrogenation and successive FA hydrogenolysis; while a synergistic effect between metallic Cu species and acidic sites on the catalyst surface played a key role on MF formation. In addition, the activated copper catalyst, again, efficiently promoted another biomass-related important conversion of HMF to DMF with the target product yield of 96.7 mol %. The recycling experiments revealed that both the copper and the activated copper catalysts maintained good activity and stability after five-time recycling.

drawbacks, such as poor selectivity of the transformations due to undesired side reactions and high cost of the catalysts caused by the limited abundance of precious metals, are prominent barriers for large-scale applications in FFR conversion. In comparison, nonprecious copper-based catalysts have received attractive attention and are reported to be very active for FFR hydrogenation. For instance, Lesiak et al. found that addition of Cu into Pd/SiO2 significantly improved the hydrogenation rate of FFR and reduced the decarbonylation rate.21 Zhu et al. prepared three supported Cu catalysts (Cu/SiO2, Cu/Al2O3, and Cu/ZnO) by a typical precipitation method for the selective hydrogenation of FFR, and Cu/SiO2 exhibited the best catalytic performance affording a MF yield of 89.5% under the synergistic effect of metal and weak Lewis acid sites of the catalyst.22 Yan et al. reported that Cu−Cr bimetallic catalysts, derived from hydrotalcite precursors, revealed a synergistic effect in selective hydrogenation of FFR, producing a 95% yield of FA with 99% conversion of FFR at a temperature of 473 K.23 Notably, the hydrotalcite has a clear advantage to produce a high loading amount of the supported Cu−Cr catalysts through proper calcination and reduction. Though high conversion and excellent product selectivity for FFR hydrogenation were achieved over copper-based catalysts, the use of high-pressure molecular hydrogen still presented several obvious issues, such as H2 storage, safety, transportation, solubility, and so on. On this basis, alternative hydrogen donors of formic acid and alcohols for catalytic transfer hydrogenation (CTH) purposes have received extensive concern, especially in recent years.24−29 Particularly, the utilization of alcohols as a hydrogen donor offers significant benefits when compared with formic acid in some ways related to corrosiveness. In fact, alcohols, such as methanol, ethanol, 1-butanol, 2-propanol, and 2-butanol, are extensively applied in the reductive upgrading of biomass as a hydrogen donor. Generally, secondary alcohols including 2-propanol and 2-butanol require lower reaction temperature and sometimes help to reduce side reactions. For instance, Hermans reported Pd/Fe2O3-promoted transfer hydrogenation/hydrogenolysis for reductive upgrading of FFR and 5-hydroxymethylfurfural (HMF) to MF and 2,5-dimethylfuran (DMF), respectively, using 2-propanol or 2-butanol as hydrogen donors.30 Certainly, methanol is also a preferred hydrogen source, from the perspectives of cost reduction and environmental protection. Moreover, methanol can be produced from renewable biomass resources and the coproducts generated in copper-catalyzed methanol reforming are gaseous compounds, i.e., CO, CO2, and CH4 (see Scheme S1 in the Supporting Information).31 Thus, extensive efforts have been devoted to seeking efficient catalysts for upgrading of biomass using methanol as a clean H-transfer agent. For example, a Cu-doped porous metal oxide (Cu-PMO) from hydrotalcite-like precursor was developed for hydrogenation of HMF via a hydrogen transfer from clean methanol, the transformation afforded a combined yield of 61% to DMF, 2,5dimethyltetrahydrofuran (DMTHF), and 2-hexanol.32 Furthermore, Cu-PMO also served as an effective catalyst for the depolymerization of organosolv lignin in methanol medium by giving good-to-excellent yields of bio-oils with similar molecular weight distributions.33 More recently, Lin et al. examined commercial CuO for methyl levulinate (ML) conversion into γvalerolactone (GVL), using methanol as both a solvent and hydrogen donor, and a GVL selectivity of 87.6% was attained at nearly a quantitative ML conversion.34



EXPERIMENTAL SECTION

Materials. Furfuryl alcohol (FA, 98%), 2-methyl furan (MF, 98%), 5-hydroxymethylfurfural (HMF, 99%), and 2,5-dimethylfuran (DMF, 99%) were purchased from Aladdin Reagent Co. Ltd. (Shanghai, PRC) and used as received. CuO (AR), MeOH (AR), and all other chemicals were provided by Guangzhou Chemical Factory (Guangzhou, PRC) and used without further purification. Furfural (FFR) was obtained from Aladdin Reagent Co. Ltd. (Shanghai, PRC) and distilled under vacuum before use. Catalyst Synthesis. Four samples of copper-based hydrotalcitelike compounds (HTlc) with Cu/Al molar ratios of 1:1, 2:1, 3:1, and 4:1 were synthesized by the coprecipitation method described earlier.35 Two aqueous solutionsone containing appropriate amounts of metal nitrates [Cu(NO3)2·3H2O and Al(NO3)3·9H2O] and the other containing precipitating agents (NaOH and Na2CO3), were dropped slowly at a pH value of ∼9. After precipitation, the resulting suspension was maintained at 348 K for 10 h with stirring. The precipitate was washed with distilled water until the pH value of the filtrate was ∼7 and then dried at 383 K for 12 h under air flow. The precipitate was ground to fine powders and then calcined at 773 K in a muffle furnace for 3 h. The samples obtained in this way were designated as CuxAl, where x was the Cu/Al molar ratio. For example, a calcined copper catalyst, derived from the hydrotalcite precursor, with a Cu/Al molar ratio of 2, was denoted as Cu2Al. The resulting Cu2Al sample further activated in a 40% H2/N2 flow at 773 K was labeled as Cu2Al-A. The activation process was performed in a heating rate of 2 K min−1 under a flowing of 40% H2/N2 with the final temperature maintained for 3 h. Typical Procedure for the CTH of FFR in Methanol. The experiments were performed in a 50 mL cylindrical stainless steel highpressure reactor. In a typical run, a methanol (15 mL) solution of furfural (115 mg, 1.2 mmol) and a certain amount of copper-based catalysts were added into the reactor. The sealed reactor was purged three times with N2 and finally pressurized to 1 MPa. Thereafter, the reactor was brought to the desired temperature by means of a heating jacket. When the reaction was completed at desired times, the mixture was cooled, filtered, and stored for analysis. In addition, the gaseous products of the reaction were collected for GC analysis. Catalyst Characterization. Powder X-ray diffraction (XRD) was performed on a Bruker D8 Avance diffractometer, using Cu Kα radiation generated at 40 kV and 40 mA. X-ray photoelectron spectra (XPS) and Auger electron spectroscopy were made on a Kratos Ultra system employing an Al Kα radiation source. The binding energies for 5983

DOI: 10.1021/acssuschemeng.7b00778 ACS Sustainable Chem. Eng. 2017, 5, 5982−5993

Research Article

ACS Sustainable Chemistry & Engineering

quantitative analysis of in situ-generated H2, CO, and CO2, according to the calibrated curves of standard gases. The corresponding formulas for calculating conversion and yield were defined as below:

each spectrum were calibrated with a C 1s spectrum of 284.6 eV. Notably, the sample was exposed to an air atmosphere at 298 K before XPS measurement. The morphology of the fresh and recycled samples was observed by scanning electron microscopy (SEM, Hitachi S4800). High-resolution transmission electron microscopy (HR-TEM, JEOL, Model JEM-2100HR) was conducted for morphological analysis of copper-based catalysts. As for sample preparation, a small amount of catalysts was first dispersed in ethanol solution, and then a drop of the suspension was placed onto carbon-coated molybdenum grids, followed by evaporating the solvent. Chemical analyses of metal elements were measured by a Thermo Elemental ICP-AES spectrometer. The Brunauer−Emmett−Teller (BET) surface area measurements were performed with N2 adsorption−desorption isotherms at 77 K (Quantachrome, Model SI-MP-10). Before test, the samples were degassed under vacuum at 473 K for 18 h. The measurements of catalyst acidity were carried out on a Micromeritics Autochem II 2920 chemisorption analyzer, following a temperature-programmed desorption in ammonia (NH3-TPD) method. For each run, the sample was heated up to 573 K at a rate of 10 K min−1 and kept for 0.5 h in a He flow to remove adsorbed impurities. Then the sample was cooled down to 373 K for the adsorption of NH3. After flushing with He for 1 h to remove physically adsorbed NH3, the TPD data were collected from 373 K to 1023 K with a ramp of 20 K min−1. The temperatureprogrammed reduction in hydrogen (H2-TPR) analysis of the prepared copper catalysts was performed on Autochem II 2920. In a typical experiment, ∼0.15 g of catalyst was placed in a quartz sample cell; the sample was then flushed with pure argon at a flow rate of 50 mL min−1 and 573 K for 0.5 h and then cooled to room temperature. Subsequently, a 10% H2/Ar (30 mL min−1) was flown through the sample while the temperature was increased to 800 K at 10 K min−1 ramp rate and held at final temperature for 0.5 h. Pyridine adsorption Fourier-transform infrared (Py-IR) analysis was conducted on Bruker FT-IR Tensor 27 equipped with a BX-5 in situ transmission FT-IR device to identify the acid type and concentration of the prepared catalyst. An ∼20 mg sample was pressed into a self-supporting disk 13 mm in diameter, and then the disk was placed in a quartz cell connected to a conventional closed gas-circulation system. The sample disk was pretreated at 473 K for 2 h in a pressure of 1.0 × 10−3 Pa, and then the temperature was cooled to 313 K for collecting a singlechannel background spectrum. For the adsorption experiment, pyridine was introduced to the sample for 20 min. IR spectra were measured at 313 K after adsorption balance, followed by evacuation to remove adsorbed pyridine for collecting IR spectra at 423 K (heating rate, 10 K min−1) and 673 K (heating rate, 10 K min−1). Thermogravimetric/differential thermal analysis (TGA/DTA) processes were performed by using a Model SDT Q 600 instrument over fresh and spent catalysts in order to identify the amounts of adsorbed organic compounds over the catalyst surface. Specifically, a sample of ∼100 mg was typically employed for the measurement at temperatures from room temperature up to 1173 K with a heating rate of 20 K min−1 in air. Products Analysis. The quantity of collected samples was analyzed using a Fuli Model GC 9790II system that was equipped with a flame ionization detector (FID) and a KB-5 capillary column (30.0 m × 0.32 μm × 0.25 μm), using nitrogen as the carrier gas. The GC-MS analysis was recorded on Trace GC-MS 2000 system that was equipped with a 30-m Model HP-5 system (0.25 mm internal diameter). The operating conditions for GC-MS were as follows: injector port temperature, 533 K; column temperature and initial temperature, 323 K (3 min); gradient rate, 10 K min−1; final temperature, 493 K (2 min); flow rate, 75 mL min−1. The assignment of all of the products and intermediates was confirmed by the database for each MS chart. The analysis of gaseous products was performed on an Agilent Model system 7890A equipped with an Agilent Model CP-7429 column, a flame ionization detection (FID) system, and a thermal conductivity detection (TCD) device. The carrier gas was helium, and the column temperature was maintained at 333 K for 15.2 min for the

⎡ moles of reactant converted ⎤ × 100 conversion (%) = ⎢ ⎣ moles of reactant fed ⎥⎦

⎡ moles of product produced ⎤ yield (%) = ⎢ × 100 ⎣ moles of reactant fed ⎥⎦ In addition, the product formation rate was defined as the amount of formed target product with the unit of μmol per gram of copper in the catalyst per minute, in which copper component was determined by ICP-AES analysis and presumably contributed to the reactions.



RESULTS AND DISCUSSION Catalyst Characterization. Figure 1 shows the XRD patterns of the CuxAl-HTlc. The prepared copper catalyst

Figure 1. Powder XRD patterns of CuxAl-HTlc.

precursors revealed a highly ordered crystalline structure with distinct hydrotalcite features consistent with the literature.36 Featured diffraction peaks at 11.6°, 23.3°, 35.6°, 40.1°, 48.0°, 58.3°, and 61.7° could be clearly observed, which were assigned to the (003), (006), (012), (015), (018), (110), and (113) reflections of hydrotalcite (JCPDS File No. 22-0700), respectively.37,38 Furthermore, isolated phases of individual metal hydroxides were undetected in the CuxAl-HTlc samples with an x range from 1 to 3, suggesting the formation of a pure hydrotalcite phase. Evidently, overexcess addition of the Cu component in the CuxAl-HTlc would destroy the layered hydrotalcite structure, because of the change in crystal form, as shown in Figure 1 for Cu4Al-HTlc. In addition, Bragg reflection at 2θ = 38.8° in Cu4Al-HTlc was attributed to the gibbsite phase (JCPDS File No. 07-0324), further indicating the formation of amorphous hydrate phase.39 Copper catalyst CuxAl was obtained by calcination of the corresponding CuxAl-HTlc at 773 K for 3 h in air. However, the activated copper catalyst (CuxAl-A) was prepared by treating CuxAl with a 40% H2/N2 flow at 773 K. Figure 2 gives the XRD patterns of the CuxAl and CuxAl-A samples. Initially, CuO was investigated for comparison purposes (Figure 2). Diffraction peaks at 32.4°, 35.7°, 38.9°, 48.6°, 53.7°, 58.1°, 61.5°, 66.1°, 67.9°, 72.3°, and 75.1° were observed in commercial CuO, which revealed the typical reflections ascribed to monoclinic CuO (JCPDS File No. 65-2309).40 As for CuxAl (x = 1−4) samples, characteristic peaks assigned to 5984

DOI: 10.1021/acssuschemeng.7b00778 ACS Sustainable Chem. Eng. 2017, 5, 5982−5993

Research Article

ACS Sustainable Chemistry & Engineering

monoclinic CuO could also be obviously found. Notably, the peaks corresponding to copper hydroxide and other copper species were unidentified. In the case of Cu3Al-A sample, typical crystallization peaks centered at 43.4°, 50.6°, and 74.2° were ascribed to the (111), (200), and (220) planes of fcc crystal structures of metallic Cu (JCPDS File No. 04-0836).41,42 Interestingly, the XRD patterns of spent Cu2Al show metallic Cu planes, demonstrating that a portion of elemental Cu existed as Cu0 species after the hydrogen transfer reaction. In addition, color change for fresh and recycled Cu2Al evidently suggested the variation of copper valence state (Figure S1 in the Supporting Information). Therefore, XRD analysis revealed the presence of a pure hydrotalcite phase for CuxAl-HTlc samples (x = 1−3), monoclinic CuO for CuxAl samples (x = 1−4), and face-centered cubic (fcc) crystal structures of metallic Cu for the Cu3Al-A sample. Moreover, a portion of the Cu component in Cu2Al was reduced from the oxidation

Figure 2. Powder XRD patterns of CuxAl and CuxAl-A.

Figure 3. XPS spectra of the prepared samples: (A) full spectra, (B) Cu 2p, and (C) Cu LMM. See text for additional details. 5985

DOI: 10.1021/acssuschemeng.7b00778 ACS Sustainable Chem. Eng. 2017, 5, 5982−5993

Research Article

ACS Sustainable Chemistry & Engineering state of Cu(II) to Cu(0) after the Cu2Al-promoted hydrogen transfer reaction. XPS characterizations were undertaken to investigate the surface chemical state of the copper and the activated copper catalysts. As depicted in Figure 3A, full spectra of Cu2Al including C 1s, O 1s, Al 2p, Cu 2p, and Cu LMM were conducted. In the cases of Cu2Al and Cu3Al catalysts, the peaks located at the binding energies of 933.7 and 953.7 eV with strong satellite peaks were assigned as Cu 2p3/2 and Cu 2p1/2, respectively (see Figures 3B1 and 3B2), which indicates the presence of an oxidation state of Cu2+ species.43 The results were in good accordance with XRD analysis (Figure 2). However, for spent Cu3Al, both Cu0 (35.6%) and Cu+ (50.7%) species were additionally observed after the reaction (Figure 3B3), suggesting a partial reduction of surface Cu2+ species during the transfer hydrogenation. In the case of the Cu3Al-A sample, the binding energies of Cu 2p3/2 and Cu 2p1/2 downshifted separately, to 932.4 and 952.2 eV (Figure 3B4), if compared with Cu3Al. Besides, the intensity of the satellite peaks in Cu3Al-A became rather weakened. As is well-known, Cu0 species cannot be distinctly distinguished from Cu+ through Cu 2p XPS analysis;44 thus, Cu LMM Auger spectra were performed to accurately clarify the copper species. According to the Cu LMM spectra, as shown in Figure 3C, the peak with a kinetic energy of ∼918.6 eV was indexed to metallic copper, while the peak at 916.9 eV was assigned to Cu+,45,46 demonstrating the presence of both Cu0 and Cu+ species in Cu3Al-A sample. An in-depth study of the Cu 2p spectra further confirmed the existence of a small amount of Cu+ (6.8%) and Cu2+ (9.9%) species on the surface of the fresh Cu3Al-A (Figure 3B4). However, the H2-TPR analysis indicated that prepared copper catalysts could be completely activated in a H2 atmosphere at ∼600 K (Figure S2 in the Supporting Information). Therefore, it was inferred that the existence of Cu+ and Cu2+ species in the Cu3Al-A sample was presumably attributed to a partial oxidation of surface Cu0 species upon the sample exposure to air before XPS measurement. This assumption was further supported by an evident color change from dark red for the fresh Cu3Al-A to almost black after the air exposure of Cu 3 Al-A (Figure S3 in the Supporting Information). Generally, the XPS results were consistent with XRD analysis, revealing the oxidation state of Cu2+ for both Cu2Al and Cu3Al, as well as the presence of Cu0, Cu+, and Cu2+ species for both Cu3Al-A and spent Cu3Al. The temperature-programmed desorption profiles of desorbed ammonia on CuxAl and CuxAl-A samples were described in Figure 4 (the inset reflects the assigned acid amount). Generally, the higher desorption temperature of NH3-TPD reflects the stronger acid strength of the catalyst. With respect to the CuxAl (x = 1−4) samples, a characteristic peak at a desorption temperature of ∼670 K was clearly observed (Figure 4A). Calculated results further demonstrated that doping of the Al component significantly enhanced the catalyst acidity. As a result, the Cu1Al catalyst, containing the highest levels of Al sites in the investigated samples, presented the maximum acid amount. In the case of CuxAl-A samples (x = 1−4), the intensity of desorption peaks at high-temperature zone became weaker, thus indicating the moderate acid strength (Figure 4B). This observation can be attributed to a reduced Cu valence state for CuxAl-A samples, which gave much metallic Cu species, but at the expense of both acidic CuO species and acidic OH groups.

Figure 4. NH3-TPD profiles of (A) CuxAl and (B) CuxAl-A.

With in-depth study, the IR spectra of adsorbed pyridine (Py-IR) on some typical samples were performed to characterize acid type and the corresponding concentration. As indicated in Figure 5, two bands, at ∼1442 and 1490 cm−1, which were

Figure 5. IR spectra of pyridine adsorbed on Cu2Al, Cu2Al-A, and Cu3Al-A.

attributed to adsorbed pyridine molecules, were apparently displayed, suggesting the presence of large amounts of Lewis acid (L) sites in Cu2Al, Cu2Al-A, and Cu3Al-A. In addition, a weak absorption peak at 1552 cm−1 depicted as Brønsted acid (B) sites was also observed, which indicated a lower concentration of acid sites. Notably, Lewis acid and Brønsted acid sites, obtained by Py-IR (Figure 5), on the sample surface 5986

DOI: 10.1021/acssuschemeng.7b00778 ACS Sustainable Chem. Eng. 2017, 5, 5982−5993

Research Article

ACS Sustainable Chemistry & Engineering revealed much lower acidity in comparison with those described in NH3-TPD analysis (Figure 4), which could be attributed to the presence of a large quantity of acidic OH groups on the sample surface.47 The mesoporosity of copper-based catalysts was confirmed by N2 adsorption−desorption isotherm analysis, as illustrated in Figure S4 in the Supporting Information. All samples displayed type IV isotherms with a type H3 hysteresis loop, which was characteristic of mesoporous structure in a size range of 2−50 nm. The physicochemical properties of the prepared samples were listed in Table 1. Generally, the Brunauer−Emmett− Table 1. Structural Properties of CuxAl and CuxAl-A Samples entry

sample

BET surface area [m2 g−1]

1 2 3 4 5 6 7 8 9 10 11 12

Cu1Al Cu2Al Cu3Al Cu4Al Cu1Al-A Cu2Al-A Cu3Al-A Cu4Al-A Cu2Ala Cu2Alb Cu3Al-Ac Cu3Al-Ad

113.3 66.5 69.6 50.4 96.3 61.7 56.4 66.7 153.7 85.6 46.9 44.0

average pore diameter, PD [nm]

total pore volume, PV [cm3 g−1]

6.5 7.9 38.8 44.0 6.6 9.3 37.5 20.9 3.8 3.7 13.1 18.8

0.31 0.30 0.35 0.32 0.28 0.26 0.18 0.19 0.31 0.27 0.23 0.19

Figure 6. TGA-DTA profiles for the fresh and spent Cu2Al.

probably due to the removal of intercalated water. As the temperature rose, two exothermic peaks at ∼642 K and ∼845 K were observed in DTA profile of the spent catalyst, as compared to the fresh one, indicating the presence of organic compounds adsorbed on the Cu2Al surface after the reaction. Therefore, effectively removing carbon deposits was necessary for catalyst regeneration when concerning recycling experiments. Activity Test. The developed CuxAl and CuxAl-A catalysts were investigated for the transfer hydrogenation of FFR with methanol as both the solvent and the hydrogen source; the results were summarized in Table 2. Initially, commercially

a

Used once. bUsed once and calcined again. cUsed twice. dUsed twice, calcined, and reduced again.

Table 2. Catalytic Activities of CuxAl and CuxAl-A on Transfer Hydrogenation of FFRa

Teller (BET) surface area trended downward as the Cu loading or both CuxAl (Table 1, entries 1−4) and CuxAl-A (Table 1, entries 5−7) samples increased; meanwhile, the average pore diameter increased significantly with Cu loading (Table 1, entries 1−4 and 5−7). Recycled Cu2Al catalysts gave higher BET surface areas, if compared with the fresh one (Table 1, entries 9 and 2), which could be attributable to the in situ formed copper nanoparticles. High-temperature calcination was considered to be an effective way to remove adsorbed organics on the catalyst surface. On the other hand, the calcination treatment led to particle sintering, which is presumably responsible for the decline of BET surface area (Table 1, entries 9 and 10). For regenerated Cu3Al-A under a H2 atmosphere, both BET surface area and average pore diameter of the sample were relatively lower, namely, 44.0 m2 g−1 and 18.8 nm, respectively (Table 1, entry 12). Catalyst regeneration presumably accelerated particle agglomeration during calcination and reduction procedures. The surface structure changes for recycled Cu2Al and Cu3Al-A catalysts were also verified by SEM images (Figure S5 in the Supporting Information) and HR-TEM images (Figure S6 in the Supporting Information). In addition, HR-TEM analysis further indicated that the copper particles existed in an irregular form in the Cu3Al-A sample. In order to investigate the carbonaceous components on the catalyst surface, we analyzed the fresh and spent Cu2Al using a TGA/DTA technique, which was performed by heating the samples from room temperature to 1173 K in air. As shown in Figure 6, the analysis results revealed an initial marginal weight loss (1−4 wt %) before heating to 450 K, which is due to physisorbed water and surface hydroxyl groups. A further weight loss of 2−3 wt % was registered for fresh and spent Cu2Al over the temperature range of 535−560 K, which is

entry

catalyst

FFR conversion [%]

FA yield [mol %]

MF yield [mol %]

1 2 3 4 5 6 7 8 9

CuO Cu1Al Cu2Al Cu3Al Cu4Al Cu1Al-A Cu2Al-A Cu3Al-A Cu4Al-A

86.5 >99 >99 >99 98.7 98.1 >99 97.7 98.0

41.8 47.4 54.7 44.9 38.0 36.9 26.6 4.1 21.0

3.5 1.9 4.7 3.7 0.87 16.5 47.9 88.2 40.7

a

Reaction conditions: catalyst (90 mg), FFR (115 mg, 1.2 mmol), methanol (15 mL), 513 K, N2 (1 MPa), 600 rpm, 1.5 h.

available CuO was investigated as the catalyst for comparison purposes and a FFR conversion of 86.5% with an FA yield of 41.8 mol % was obtained (Table 2, entry 1). In contrast, full conversions of FFR were observed over CuxAl catalysts (Table 2, entries 2−5). However, the selectivity to target product FA was still not improved significantly, which could be attributed to the presence of parallel reactions, such as acetal, hemiacetal, etherification, and aldol (Figure S7 in the Supporting Information) induced by acid sites of the catalyst (see Figures 4 and 5). Among the CuxAl series catalysts examined, Cu2Al revealed preferred activity toward FA production via FFR transfer hydrogenation (Table 2, entry 3). Based on XRD and XPS analyses of CuxAl before and after the reaction (Figures 2 and 3), the Cu component in the catalyst initially existed in the 5987

DOI: 10.1021/acssuschemeng.7b00778 ACS Sustainable Chem. Eng. 2017, 5, 5982−5993

Research Article

ACS Sustainable Chemistry & Engineering oxidation state of Cu2+ and promoted methanol decomposition to produce hydrogen molecules. Meanwhile, Cu2+ species would consume certain amounts of in situ-generated hydrogen to form active Cu0 centers for the subsequent FFR hydrogenation. Therefore, lower copper loading in CuxAl series catalysts (such as Cu1Al) led to insufficient hydrogenation ability toward FFR. On the other hand, much effort should be devoted to the reduction of Cu2+ into Cu0 species at the expense of generated hydrogen, with respect to higher Cu loading samples (such as Cu3Al and Cu4Al), which was evidently not conducive to the subsequent hydrogenation of FFR. Thus, among the CuxAl series catalysts examined, Cu2Al catalyst was the most active one for FA production. Both FA and MF were observed in CuxAl-promoted transfer hydrogenation of FFR. The hydrogenation and saturation of CO double bond in FFR molecule led to the formation of FA; while, the subsequent hydrogenolysis of saturated C−O bond in FA gave MF product. However, as shown in Table 2, only a trace amount of MF was detected with the CuxAl series catalysts (Table 2, entries 2−5), indicating that the catalyst with more active Cu0 species was required for FA hydrogenolysis to give MF. Therefore, CuxAl-A series samples, obtained by treating CuxAl series samples with H2/N2 flow at 773 K, were further investigated for FFR hydrogenation. Surprisingly, the MF yield was remarkably increased under the identical reaction conditions when using CuxAl-A catalysts for FFR transformation (Table 2, entries 6−9). A reasonable explanation is that more active Cu0 species in CuxAl-A catalyst participated in the reaction. These Cu0 species efficiently promoted the FA hydrogenolysis to give MF, as well as methanol decomposition to afford hydrogen (see Figure S8 and Table S1 in the Supporting Information). Also, properly increasing the Cu loading level facilitated the transformation of FA into MF, presumably under the synergistic action of active Cu0 centers and acid sites on the catalyst surface. However, excessive Cu loading in the catalyst (such as Cu4Al-A) did not favor MF formation (Table 2, entry 9), because of the weaker acidity of Cu4Al-A, as confirmed by NH3-TPD analysis. Previous work also verified that acid sites play a crucial role in the deoxygenation process.48,49 By comparison, Cu2Al, possessing a proper copper loading level, was the most active catalyst for FFR-to-FA transformation by producing a FA yield of 54.7 mol % at 513 K (Table 2, entry 3). However, under the identical reaction conditions, Cu3Al-A, rendering more active Cu0 and moderate amounts of acid sites, was the most active catalyst for FFR-to-MF transformation by resulting in a MF yield of 88.2 mol % (Table 2, entry 8). The FFR-to-MF transformation involves a tandem reaction of FFR hydrogenation by saturation of the CO double bond in FFR molecule and subsequent FA hydrogenolysis by the deoxygenation of saturated C−O bond in the FA compound. Based on the above observations, Cu2Al and Cu3Al-A were chosen for the following work to determine the optimum conditions. Herein, the influence of reaction temperature on FFR conversion was investigated in the range of 463−518 K, and the results are shown in Figure 7. In the case of FA formation over Cu2Al catalyst (Figure 7A), increasing the reaction temperature promoted transfer hydrogenation of FFR in the initial stage, which was mainly attributed to a continuous hydrogen supply from rapid methanol decomposition under elevated temperatures. A satisfactory result of 67.5 mol % FA yield was achieved at a temperature of 473 K. Higher temperatures promoted the formation of active Cu0 sites,

Figure 7. Effect of reaction temperature on FFR transfer hydrogenation over (A) Cu2Al and (B) Cu3Al-A catalysts. Reaction conditions: catalyst (90 mg), FFR (115 mg, 1.2 mmol), methanol (15 mL), N2 (1 MPa), 600 rpm, 1.5 h.

resulting in subsequent FA hydrogenolysis to MF. In addition, higher temperatures were associated with a chain of undesired reactions, such as aldol condensation and etherification.27 In terms of MF production over Cu3Al-A (Figure 7B), a low MF yield was observed at relatively lower temperatures with an appreciable amount of intermediate FA. Notably, a quantitative transformation of the intermediate FA was observed with the rise in reaction temperature. Therefore, abundant active Cu0 sites in Cu3Al-A, combined with a sufficient supply of H2, originated from methanol reforming under higher temperatures, accounted for this aspect. In this case, a promising MF yield of 88.2 mol % with a FFR conversion of 97.7% was reported at a temperature of 513 K. As expected, a further increased temperature to 518 K allowed a full FFR conversion; however, both yield and selectivity toward the targeted MF declined, because of deep hydrogenolysis. Taking into consideration of practical cost and efficiency, the optimal temperature for this reaction was suggested to be 513 K. The influence of catalyst dosage on the CTH of FFR was subsequently surveyed to determine the suitable conditions to improve the yields of FA and MF. As depicted in Figure 8, the target compound yields for both FA (Figure 8A) and MF (Figure 8B) were adequately enhanced for higher catalyst dosage during the initial stage of the reaction, because of the fact that much more H2 molecules were available from methanol reforming with the sufficient loading of copper catalysts. As for FA production, a significant rise in FA yield from 39.0 mol % to 63.9 mol % was attained with the increase in the catalyst dosage from 75 mg to 105 mg after 1 h (Figure 5988

DOI: 10.1021/acssuschemeng.7b00778 ACS Sustainable Chem. Eng. 2017, 5, 5982−5993

Research Article

ACS Sustainable Chemistry & Engineering

formation, a higher Cu3Al-A dosage greatly promoted FFR-toMF transformation (Figure 8B), since high dosage rendered abundant active Cu0 sites for both hydrogen supply and subsequent hydrogenation/hydrogenolysis within a relatively short period. Therefore, an excellent MF yield of 94.1 mol % was easily attained at a Cu3Al-A dosage of 105 mg in 1.5 h at 513 K in methanol (Figure 8B). Extending the reaction time was unfavorable to increasing the MF yield, because of the formation of side products, such as 2-methyltetrahydrofuran, 2pentanol, THFA, and so on (see Figure S9 in the Supporting Information). Notably, a prolonged time was necessary to achieve reaction equilibrium when using lower catalyst dosages of 75 and 90 mg. Therefore, the Cu2Al catalyst, derived from hydrotalcite precursor, showed outstanding transfer hydrogenation selectivity toward the FFR-to-FA transformation by giving a FA yield of 94.0 mol % with methanol as the hydrogen donor. Meanwhile, the activated Cu3Al-A catalyst favored the FFR-to-MF transformation by producing a MF yield of 94.1 mol %. Recently, Yan and co-workers reported selective hydrogenation of FFR using hydrotalcite-derived Cu catalyst and H2 by giving 90% yield of FA and 51% yield of MF.50 Evidently, the hydrotalcite-derived Cu2Al and Cu3Al-A catalysts in this research showed comparable or even higher catalytic performance in transfer hydrogenation than in the case of previously reported hydrogenation. Given the outstanding performance of copper-methanol catalytic system in FFR transformation, another important biomass-based platform chemical HMF was also subjected to this system for the synthesis of highly valuable DMF without any external hydrogen supplies. As shown in Table 3, the temperature was found, again, to play a key role in HMF-toDMF transformation over Cu3Al-A, with methanol as both the solvent and the hydrogen source. Low reaction temperature gave low selectivity of DMF (Table 3, entries 1 and 2), many intermediates including 5-hydroxymethyl-2-methyl furan and a series of condensation compounds were detected as confirmed by gas chromatography-mass spectroscopy (GC-MS) analysis (see Figure S10 in the Supporting Information). An exciting result on the desired DMF production was observed at 513 K, which afforded an excellent DMF yield of 96.7 mol % with the highest formation rate of 246.6 μmolDMF gCu−1 min−1 (Table 3, entry 5). Also, a low catalyst dosage of 75 mg gave a poor DMF yield of 60.1 mol % (Table 3, entry 6). Therefore, the copper catalyst developed in this research showed excellent performance toward biomass-related transfer hydrogenation reaction, by giving a FA yield of 94.0 mol % in FFR-to-FA transformation

Figure 8. Effect of catalyst dosage on FFR transfer hydrogenation over (A) Cu2Al and (B) Cu3Al-A catalysts. Reaction conditions: FFR (115 mg, 1.2 mmol), methanol (15 mL), N2 (1 MPa), 600 rpm, (Cu2Al and 473 K) for panel (A), (Cu3Al-A and 513 K) for panel (B).

8A). Under the optimized reaction conditions, a FA yield of 94.0 mol % with a full conversion of FFR was obtained by using a dosage of 105 mg of Cu2Al at a reaction temperature of 473 K after 2.5 h in methanol. Increasing the Cu2Al dosage remarkably promoted FFR-to-FA transformation. However, a longer reaction time of 2.5 h was still required to achieve maximum FA yield, regardless of Cu2Al dosage (Figure 8A), which is attributed to the rate-determining step of Cu2+ species reduction to active Cu0 in the Cu2Al catalyst, rather than FFR hydrogenation. A further increase in reaction time from 2.5 h to 5.0 h led to a steep decrease in FA yield, which was elucidated as subsequent hydrogenolysis of FA to produce ∼16.5 mol % MF and the formation of various etherification products (see Figure S7 in the Supporting Information). With regard to MF

Table 3. Transfer Hydrogenation of Biomass-Derived HMF into DMF

entry

temperature, T [K]

usage [mg]

conversiona [%]

yieldb [mol %]

product formation rate [μmolDMF gCu−1 min−1]

1 2 3 4 5 6 7

473 483 493 503 513 513 513

90 90 90 90 90 75 105

80.1 92.0 93.4 96.8 >99 95.7 >99

13.5 34.7 49.9 78.1 94.9 60.1 96.7

35.0 90.2 129.7 203.0 246.6 187.4 215.4

a

: HMF conversion. b: DMF yield. Reaction conditions: Cu3Al-A (90−105 mg), HMF (151 mg, 1.2 mmol), methanol (15 mL), N2 (1 MPa), 600 rpm, 1.5 h. 5989

DOI: 10.1021/acssuschemeng.7b00778 ACS Sustainable Chem. Eng. 2017, 5, 5982−5993

Research Article

ACS Sustainable Chemistry & Engineering

activity of Cu3Al-A after repeated use was observed, presumably because of a minor decline in BET surface area caused by particle sintering during reactivation procedure. Thus, sufficient contact between FFR and active metal centers could be partially suppressed. Notably, both Cu2Al and Cu3Al-A catalysts exhibited exceptional stability with almost no leaching of Cu (