Selective Transfer Hydrogenation of Furfural into Furfuryl Alcohol on Zr-Containing Catalysts Using Lower Alcohols as Hydrogen Donors.

A series of zirconium-based catalysts were prepared for the selective transfer hydrogenation of biomass-derived furfural (FFR) into furfuryl alcohol with lower alcohols as hydrogen sources. The sample structures were clearly characterized using various methods, such as X-ray powder diffraction, thermogravimetric analysis, scanning electron microscope, NH3-temperature-programmed desorption (TPD), CO2-TPD, and nitrogen physisorption. Excellent furfuryl alcohol yield of 98.9 mol % was achieved over Zr(OH)4 using 2-propanol as a hydrogen donor at 447 K. The poisoning experiments indicated that basic centers displayed pronounced effect for FFR transfer hydrogenation. Moderate monoclinic phase content in ZrO2-x enhanced the conversion rate and furfuryl alcohol selectivity, whereas acid-basic site density ratio had slight influence on FFR conversion. Besides, Zr(OH)4 revealed good performance and stability after being repeated four times. The possible mechanism for this transfer hydrogenation process over Zr(OH)4 catalyst with 2-propanol as the hydrogen source was proposed.

Introduction

The rapid depletion of fossil fuel resources and significant deterioration of environmental problems have motivated the researchers worldwide to explore alternative renewable feedstocks for the sustainable manufacture of liquid fuels and chemicals.[1−5] In recent decades, the catalytic conversion of biomass and related derivatives to chemicals has received wide range of attention. Furfural (FFR), a potential platform for value-added compounds, is mainly obtained by acid-catalyzed hydrolysis and dehydration of C5 carbohydrates.[6−10] Among the conversion routes, catalytic reduction of FFR is of great interest, which gives a variety of value-added compounds, such as 2-methylfuran, cyclopentanone, 2-methyltetrahydrofuran, and furfuryl alcohol (FA).[11−14] Particularly, FA is considered as one of the most important derivatives that serves as precursor for the production of synthetic fibers, plastics, resins, liquid fuels, lubricants, dispersing agents, fine chemicals, and so on.[15,16]

Generally, FA can be produced by the gas-phase or liquid-phase hydrogenation of FFR over supported metallic (i.e., Pd, Cu, Co, Ru, Pt) catalysts in pressurized hydrogen atmosphere.[17−21] For instance, Yan et al. reported that hydrotalcite-derived Cu–Cr bimetallic catalysts could give 95% FA at 473 K.[22] Lesiak et al. synthesized Pd–Cu/Al2O3 for liquid-phase hydrogenation of FFR in water at 363 K and 2 MPa H2.[23] Parikh and co-workers developed Cu–Co/SBA-15 to catalyze liquid-phase hydrogenation of FFR, which offered a FA yield of 96.7% under optimized conditions.[24] Recently, catalytic transfer hydrogenation (CTH) of biomass-derived FFR via Meerwein–Ponndorf–Verley (MPV) reduction has received increasing attention, in which high-pressure gaseous hydrogen is replaced by alcohols to avoid defects, such as storage, safety, solubility, etc. For this reason, much effort has been devoted to the expansion of various catalysts for upgrading FFR via the hydrogen transfer method.[25−31] Marchi et al. found that Cu–Mg–Al catalyst could catalyze FFR hydrogenation using 2-propanol as hydrogen donor.[25] Panagiotopoulou et al. observed that homogeneous DyCl3 afforded nearly 100% FFR conversion and 97% FA selectivity with 2-propanol as a hydrogen donor.[26] Lee et al. synthesized alumina–carbon composite catalysts for the conversion of FFR in 2-propanol, with a desired FA yield of 95.8% at 403 K for 6 h.[27] Recently, low-cost eco-friendly metal hydroxides/oxides are found to allow efficient transformation of biomass-derived chemicals via MPV reduction in various alcohols.[32−35] For example, Dumesic’s group described the CTH of levulinic acid and its esters into γ-valerolactone via MPV reaction over metal oxides using alcohols as the hydrogen sources.[32] Lin et al. reported an effective CTH route to convert ethyl levulinate into γ-valerolactone over ZrO(OH)2·xH2O and ZrO2 in subcritical alcohols.[33,34] Particularly, Zr-containing catalysts are much more efficient for the CTH of biomass derivatives in the presence of alcohols.

Inspired by the above findings, a series of low-cost and environmentally benign Zr-based materials were synthesized for the CTH of FFR. The influences of varied reaction parameters, such as hydrogen source, catalyst type, temperature, and catalyst dosage, were systematically discussed to gain high FFR conversion and FA selectivity. Meanwhile, a comprehensive structure characterization of the fresh and spent catalysts was conducted. The results indicated that Zr(OH)4 exhibited excellent performance toward FFR transfer hydrogenation with 2-propanol as the hydrogen source and was reusable over multiple cycles without significant loss of activity. Furthermore, a plausible reaction mechanism for the CTH of FFR to FA via MPV reduction over Zr-containing materials in 2-propanol was proposed.

Experimental Section

Materials

Furfural (FFR, 99%) and furfuryl alcohol (FA, 98%) were provided from Aladdin Reagent Co. Ltd. (Shanghai, China). ZrOCl2·8H2O (98%), ammonia water (25–28%), butanol (anhydrous, 99%), 2-butanol (anhydrous, 99%), ethanol (anhydrous, 99.5%), 2-propanol (anhydrous, 99.5%), and other commercially available chemicals were purchased from Guangzhou Chemical Reagent Co. Ltd. (Guangzhou, China). Notably, FFR was distilled under vacuum before use.

Catalysts Preparation

First, ZrOCl2·8H2O was dissolved in deionized water to obtain a 0.4 mol L–1 solution, and then NH3·H2O was added dropwise to regulate the pH value to 9–10 under vigorous agitation, followed by aging at room temperature for 24 h. The resulting precipitate was thoroughly washed with deionized water until residual Cl– was completely removed, based on 0.05 mol L–1 AgNO3 detection. The precipitate was dried at 383 K overnight and was further ground to obtain Zr(OH)4 powder. Then, Zr(OH)4 sample was calcined at a heating rate of 2 K min–1 in air flow for 3 h for the synthesis of ZrO2-x (x represents the given calcination temperatures), in which x was varied from 623 to 1023 K with an interval of 100 K.

Catalyst Characterization

X-ray powder diffraction (XRD) was performed on a Panalytical X’pert Pro diffractometer using a Cu Kα radiation source at 40 kV and 40 mA. X-ray photoelectron spectroscopy (XPS) measurements were obtained on a Kratos Ultra system using an Al Kα radiation source. The binding energies for each spectrum were calibrated with a C 1s spectrum of 284.6 eV. The Brunauer–Emmett–Teller (BET) surface area was determined with N2 adsorption–desorption isotherms at 77 K (SI-MP-10, Quantachrome). Before test, the samples were degassed under vacuum at 433 K for 20 h. The sample morphology was realized on field emission scanning electron microscope (FESEM, Hitachi S-4800). Chemical analysis of leached Zr4+ was measured by a thermo elemental inductively coupled plasma optical emission spectrometry (ICP-AES) spectrometer.

The quantitative analysis of acid site density was carried out on a Quantachrome ChemStar chemisorption analyzer through NH3 temperature-programmed desorption (TPD). 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 to remove physically adsorbed NH3, the TPD data were collected from 373 to 1273 K with a ramp of 10 K min–1. CO2-TPD test was performed by using a similar procedure, where NH3 was replaced with CO2 during the adsorption process.

Using a SDT Q 600 instrument, thermogravimetric/differential thermic analyses (TGA/DTA) were measured over fresh and spent Zr(OH)4 to identify the amount of absorbed carbohydrate over the catalyst surface. Specifically, a sample of around 200 mg was typically used for the test at temperatures from room temperature to 1273 K, with a heating rate of 10 K min–1 in an air atmosphere.

Catalytic Test

In a typical run, catalyst, substrate FFR (1.2 mmol), and hydrogen source (15 mL) were put into a 100 mL stainless steel reactor. Subsequently, the air in the reactor was flushed out using inert atmosphere (N2) five times. Then, the reactor was initiated at a designated temperature with a stirring rate of 600 rpm under 1 MPa N2 atmosphere. After predetermined time, the reactor was cooled to room temperature in an ice water, and the filtered samples from the mixtures were further analyzed.

Product Analysis

The quantity analysis of liquid samples was completed on Fuli GC9790 II equipped with a flame ionization detector and a KB-5 capillary column (30.0 m × 0.32 μm × 0.25 μm) using nitrogen as the carry gas. The operating conditions were as follows: injector port temperature, 533 K; column temperature, initial temperature 343 K (0.5 min), gradient rate 5 K min–1, end temperature 428 K (0.5 min), flow rate 75 mL min–1.

The formulas for calculating FFR conversion, FA yield, and selectivity are defined below

Results and Discussion

Catalyst Characterization

Figure shows X-ray diffraction patterns of prepared zirconium-based materials. Specifically, fresh and recycled Zr(OH)4 exhibited similar patterns with two broad and weak peaks in an angle range from 25 to 60°, suggesting that the samples were characteristic of an amorphous structure and no new phase was observed after reuse. Through calcination treatment, typical peaks composed of monoclinic and tetragonal phases were obviously formed in ZrO2-x (x = 623, 723, 823, 923, and 1023). Noticeably, XRD spectrum of ZrO2-623 revealed typical peaks of tetragonal zirconia at 30.2 and 50.3° that agreed well with that of the standard pattern of JCPDS file no. 79-1769.[36,37] Furthermore, characteristic peaks at 2θ of 17.4, 28.2, 31.5, 34.4, 49.3, 54.2, and 55.7° (JCPDS file no. 37-1484), indicative of the presence of monoclinic zirconia, were observed in ZrO2-x (x = 723, 823, 923, and 1023 K). The percentages of the monoclinic and tetragonal phases are also given in Table , as calculated from corresponding peak areas. Apparently, higher temperatures facilitated monoclinic zirconia formation, and an increasing intensity of these peaks was observed along with the increase in calcination temperatures.

Figure 1

XRD patterns of prepared zirconium-based samples.

Table 1

Textural Properties of Prepared Zirconium-Based Catalysts

sample	aBET (cm2 g–1)	bPV (cm3 g–1)	cA (mmol g–1)	dB (mmol g–1)	eA/B	fTP (%)	gMP (%)
Zr(OH)4	242.3	0.21	1.29	1.36	0.95
hZr(OH)4	193.2	0.16
ZrO2-623	78.9	0.078	0.85	0.69	1.23	29.8	70.2
ZrO2-723	38.6	0.085	0.65	0.38	1.71	25.9	74.1
ZrO2-823	20.7	0.083	0.31	0.22	1.41	10.2	89.8
ZrO2-923	11.5	0.075	0.26	0.16	1.63	7.7	92.3
ZrO2-1023	9.0	0.070	0.13	0.11	1.18	7.6	92.4

BET, specific surface area.

PV, total pore volume.

A, acid site density, determined by NH3-TPD.

B, basic site density, determined by CO2-TPD.

Ratio of acid site density to basic site density.

Percentage of tetragonal phase (TP) ZrO2.

Percentage of monoclinic phase (MP) ZrO2.

After the third run.

XPS characterization was undertaken to investigate the surface chemical state of selected Zr-based materials. As depicted in Figure A, the full range XPS spectra of prepared samples were monitored to find out C 1s, O 1s, and Zr 3d species. Clearly, high-resolution scans of the XPS spectra of O 1s and Zr 3d with different intensity scales as ordinate are observed in Figure B and C. High-resolution O 1s spectra (Figure B) verified two surface species with binding energies of 529.8 and 531.3 eV for the fresh Zr(OH)4, characteristic of ZrO2 and Zr-OH, respectively. This result further indicated that prepared Zr(OH)4 might actually existed in the form of ZrO(OH)2·xH2O. However, calcination favored surface dehydroxylation at high temperatures, which produced zirconium oxides in various crystalline forms. Therefore, it is clearly seen that characteristic peak at 531.3 eV assigned to Zr-OH gradually disappeared in the ZrO2-x (x = 623, 723, 823, 923, and 1023) samples. With respect to Zr 3d species, the binding energies of corresponding photoelectron peaks were allocated to 181.8 eV for the Zr 3d5/2 line and 184.2 eV for the Zr 3d3/2 line, indicating the presence of Zr4+ species in ZrO2 and Zr-OH (Figure C).

Figure 2

XPS characterization of prepared Zr-based samples: (A) full spectra; (B) O 1s; (C) Zr 3d.

The textural characterizations of zirconium-based materials are further analyzed via N2 physisorption. As indicated in Figure A, weak interaction between synthesized Zr-based samples and N2 in low-pressure region (P/P0: 0–0.1) was observed, especially for ZrO2-x obtained at higher temperatures above 723 K. Typically, Zr(OH)4 and ZrO2-x (x = 623 and 723) were characteristic of IV isotherm with a H2 hysteresis loop, whereas ZrO2-823, ZrO2-923, and ZrO2-1023 revealed V isotherm with a H3 hysteresis loop. Table summarizes the resulting BET surface area and total pore volume, it is clearly seen that fresh Zr(OH)4 had larger BET surface area and total pore volume of 242.3 m2 g–1 and 0.21 cm3 g–1 respectively, as compared with those of calcined ZrO2-x samples. Actually, higher calcination temperatures (>623 K) cause a significant drop in the BET surface area and total pore volume, presumably attributable to particles crystallization. For the spent Zr(OH)4, an acceptable drop in the BET surface area and total pore volume was mainly due to the presence of carbon residues. Figure B also gives the pore width distributions of prepared Zr-based materials. In comparison, the pore diameters of Zr(OH)4 and ZrO2-x (x = 623 and 723) centered at around 3–5 nm, whereas larger pore diameters in a size range of 15–45 nm appeared for ZrO2-x (x = 823, 923, and 1023) samples. It is inferred that small particles crystallization and sintering under higher temperatures might be responsible for the occurrence of larger pore diameters.

Figure 3

N2 adsorption–desorption isotherms (A) and pore width distributions (B) of prepared zirconium-based materials.

Figure presents the SEM images of zirconium-based materials. It is apparent that fresh and spent Zr(OH)4 consisted of amorphous crystallites in a size of around 20 nm, which was well consistent with that of XRD test. No obvious variation in the primary crystallite size and the surface morphology were detected before and after use (Figure A,B,F). For the calcined samples, the crystallite size increased to 50−100 nm with elevating the temperature to 823 K, and then much crystallites in an irregular shape were identified at 1023 K (Figure E).

Figure 4

SEM images of prepared zirconium-based materials: (A) fresh Zr(OH)4; (B) fresh Zr(OH)4; (C) ZrO2-623; (D) ZrO2-823; (E) ZrO2-1023; and (F) spent Zr(OH)4.

TGA–DTA analysis was conducted to evaluate the carbon deposits on the spent Zr(OH)4 (after the third run). As described in Figure , fresh Zr(OH)4 afforded an initial weight loss of 3.5 wt % before being heated to around 385 K, and the existence of physisorbed water and surface hydroxyl groups largely accounted for this phenomenon. A further weight loss of 8.8 wt % was recorded from 390 to 730 K owing to the removal of intercalated water, implying the structure transformation of amorphous Zr(OH)4 into crystalline zirconia,[38] in good accordance with the results from XRD and SEM analyses. Weight loss of 9.3 wt % in the TGA profile of spent Zr(OH)4 was registered in the temperature range of 410–800 K, suggesting that only small quantity of organic carbon compounds accumulated on the catalyst surface during the recycling experiments. Furthermore, TGA curves showed that the practical weight losses of both of the fresh and recycled samples were much lower than the theoretical value of 22.6 wt %, indicating that the actual state of catalyst might existed as ZrO(OH)2·xH2O, rather than Zr(OH)4.[39,40]

Figure 5

TGA–DTA profiles of fresh and spent Zr(OH)4.

NH3-TPD and CO2-TPD measurements were investigated to characterize the acid–basic properties (Scheme ) and concentration in the synthesized materials, and the acid–basic site density and ratio of acidity to basicity were provided in Table . Figure shows the TPD profiles of desorbed ammonia, medium to strong acid sites at desorption temperature range of 640–800 K could be identified for all samples. For fresh Zr(OH)4, intensive desorption peaks at ∼700 K indicated the presence of higher level of acid site density. However, the calcined samples ZrO2-x (x = 623, 723, 823, 923, and 1023) revealed relatively moderate acid site density. Generally, higher temperatures led to the reconfiguration of surface structure, in which much water molecules were removed from the initial structure. In this regard, the amount of coordination unsaturated Zr4+ species increased significantly, but the Brønsted acid sites decreased due to the diminishing of surface hydroxyl groups and protons. On the other hand, higher temperature treatment resulted in a rapid decline in BET surface area, therefore the amount of surface metal ions available was reduced as well, finally leading to a significant decrease in Lewis acid sites. On the basis of above discussion, the total surface acid amount would be much lowered for calcined ZrO2 catalysts, especially for high-temperature-treated ZrO2-x. In the CO2-TPD profiles (Figure ), a broad CO2 desorption peak assigned to the strong basic sites occurred at ∼800 K. It is obvious that medium basic sites at desorption temperature of 450 K were dominant in Zr(OH)4 sample. Intriguingly, in the ZrO2-x (x = 623, 723, 823, 923, and 1023), the peak of medium temperature region shifted to a higher temperature. As is well known, the Zr4+–O2– acid–basic pair sites existed in the prepared catalysts, so total surface basic content also was gradually lowered for ZrO2-x samples, as compared with Zr(OH)4. Also interesting is that A/B value of ZrO2-x first rose and then dropped, with the increase in calcination temperature. It is concluded that Zr-containing samples obtained via calcination treatment, with relatively larger amount of tetragonal phase, could contribute to higher acid–basic concentration.

Scheme 1

Types of Acid and Basic Sites on the Surface of Zr(OH)4 (A, B) and ZrO2 (C, D)

Figure 6

NH3-TPD profiles of zirconium-based samples.

Figure 7

CO2-TPD profiles of zirconium-based samples.

Catalytic Activity

The catalytic activity of synthesized zirconium series for the transfer hydrogenation of FFR was discussed. First, the reaction was conducted at 423 K with ethanol as both the solvent and hydrogen donor, and the catalyst dosage was set to 75 mg. As shown in Table (runs 1–6), different Zr-based catalysts exhibited significant variations in the catalytic transfer hydrogenation performance. Zr(OH)4 catalyst gave a high FA yield of 39.0 mol % after 2.5 h, with a remarkable FA selectivity of 65.9%. In the case of calcined ZrO2-x (x = 623, 723, 823, 923, and 1023), lower selectivity to FA was observed at similar conversions, particularly for ZrO2-923 and ZrO2-1023. The higher activity of Zr(OH)4 could be due to the appropriate acid–basic site concentration, larger surface area, and higher total pore volume, thus facilitating H-transfer process during FFR conversion.

Table 2

Transfer Hydrogenation of FFR Using Various Zr-Based Catalysts and Hydrogen Sourcesa

run	catalyst	solvent	FFR conv. (mol %)	FA yield (mol %)	FA selec. (%)
1	Zr(OH)4	ethanol	59.2	39.0	65.9
2	ZrO2-623	ethanol	58.0	5.8	10.0
3	ZrO2-723	ethanol	62.7	8.3	13.2
4	ZrO2-823	ethanol	59.2	8.4	14.2
5	ZrO2-923	ethanol	59.9	2.5	4.2
6	ZrO2-1023	ethanol	59.0	0.4	0.7
7	Zr(OH)4	methanol	54.4	13.2	24.3
8	Zr(OH)4	2-propanol	95.7	92.0	96.1
9	Zr(OH)4	butanol	72.5	26.5	36.6
10	Zr(OH)4	2-butanol	69.2	42.6	61.6
11	ZrO2-823	2-propanol	52.4	30.7	58.6
12b	Zr(OH)4	2-propanol	97.9	54.1	55.3
13c	Zr(OH)4	2-propanol	97.8	90.8	92.9

Conditions: FFR 1.2 mmol, solvent 15 mL, catalyst dosage 75 mg, 423 K, 2.5 h, 600 rpm, N2 1 MPa.

75 mg boric acid.

75 mg pyridine.

To get more insight toward understanding the activity discrepancy among above materials, probable correlations between ratios of A/B and M/T (monoclinic phase to tetragonal phase) and catalytic activity are put forward in Figure . As indicated in Figure A, A/B ratio had very little impact on FFR conversion. Zr(OH)4 bearing much more medium to strong acid–basic sites effectively catalyzed FFR hydrogenation into FA with a high selectivity, whereas ZrO2-x possessing insufficient acid–basic sites gave poor FA selectivity. Although relative strong acid–basic sites appeared in the calcined ZrO2-x, the limited amount of Zr4+–O2– acid–basic pair sites and higher A/B ratios (above 1.0) reduced the transfer hydrogen capacity. As for M/T ratio effect, it is observed that moderate monoclinic phase content enhanced the conversion rate and FA selectivity. Usually high temperatures caused the phase change that produced large number of monoclinic ZrO2. In this aspect, diminishing of surface hydroxyl groups and BET surface area also occurred for those monoclinic ZrO2, thus resulting in the decrease of acid–basic site density. It is known that acid–basic sites played a crucial role in the MPV reduction of FFR,[41] therefore catalysts containing high content of monoclinic ZrO2 showed weak activity for FFR transfer hydrogenation. Compared comprehensively, amorphous Zr(OH)4, featuring with high acid–basic site density, exhibited outstanding activity for FA synthesis via the H-transfer method.

Figure 8

Effects of A/B and M/T ratios on FFR transfer hydrogenation.

Subsequently, the influences of hydrogen sources on the conversion of FFR into FA were thoroughly investigated (Table , runs 1 and 7–11). Generally, the hydrogen-donating capacity for FFR transfer hydrogenation greatly depends on the reduction potential of applied reducing lower alcohols, i.e., the lower reduction potential of alcohols, the stronger the reducing capacity. According to the previously reported literatures,[42,43] the reduction potential of reducing alcohols were ranked in the following sequence: methanol > ethanol > 1-butanol > 2-butanol > 2-propanol. Thus methanol, with higher reduction potential, showed the poor CTH capacity for FFR upgrading (Table , run 7). The secondary alcohol, especially 2-propanol, had the lowest reduction potential, revealing the great reducing capacity for H-transfer procedure. As expected, conversion of FFR reached 95.7 mol % in 2-propanol, with a FA selectivity of 96.1% (Table , run 8). In the case of 2-butanol, 42.6 mol % yield toward FA was obtained with a 61.6% selectivity, whereas undesirable results were found in the presence of butanol under identical conditions. On the basis of above discussion, 2-propanol was selected as the optimal hydrogen donor for subsequent experiments.

In the following experiments, several poisoning tests were designed by introducing extra additives to gain more insight toward the role of active sites for the CTH of FFR via MPV reduction. Generally, pyridine is an effective agent in passivating active acid sites, whereas boric acid showed good capacity in passivating basic centers. As observed from Table (runs 12 and 13), the addition of pyridine had slight influence on FFR conversion and FA selectivity, whereas significant decline in the activity was observed when same amount of boric acid was added. The lower selectivity to FA after boric acid poisoning could be ascribed to the imbalance acid–base couple sites, and 2,2′-difurfuryl ether and 4-(2-furyl)-3-buten-2-one were formed through acid sites-catalyzed etherification and aldol condensation reactions, respectively. The above results implied that Zr(OH)4 catalyst is rather irritable to boric acid, thus the active basic site of catalyst was demonstrated to display much more pronounced effect for this reaction. On the other hand, it can be concluded that synergistic effect of acid–basic couple sites in Zr-containing materials also was very important for MPV reduction of FFR.

The influence of temperature on the catalytic transfer hydrogenation of FFR with 2-propanol was investigated using Zr(OH)4 in a range of 403–453 K. As indicated in Figure , within a 2.5 h of CTH reaction at 403 K, the conversion of FFR was up to 78.3 mol %. Subsequently, elevating the reaction temperature to 423 and 433 K gave a higher FFR conversion of 95.7 and 97.5 mol %, respectively. Besides, the selectivity to FA remained high (around 90.0%) for reaction temperature below 443 K. A further rise in the temperature to 453 K also significantly increased the FFR conversion to 99.8 mol %, but decreased the FA selectivity to 87.9%, probably caused by the formations of furan-based polymers via acidic–basic sites-catalyzed condensation reactions.[44,45] Therefore, the optimum temperature for this reaction was set to 443 K.

Figure 9

Effect of reaction temperature on transfer hydrogenation of FFR with 2-propanol. Conditions: FFR 1.2 mmol, 2-propanol 15 mL, Zr(OH)4 75 mg, 2.5 h, 600 rpm, N2 1 MPa.

Figure depicts the effect of catalyst dosage ranging from 65 to 85 mg on FA formation as a function of reaction time, with the aim of finding the optimum conditions. One can clearly observe that the FA yield was greatly promoted at higher dosage that achieved a promising level of 81.5 mol % at a dosage of 85 mg in 0.5 h. One reasonable explanation is that the increased total number of active acid and basic sites accelerated the progress of CTH process, thus facilitating the gradual accumulation of targeted product FA. Generally, properly increasing time facilitated the rapid generation of FA. It is noted that lower catalyst dosage might require longer reaction time to reach the equilibrium state due to insufficient active sites. For 75 mg catalyst dosage, the FA yield gradually increased as the reaction proceeded, which achieved the appreciable value of 98.9 mol % yield of FA at nearly complete conversion in 2.5 h. It is noteworthy that remarkable decrease in the yield and selectivity toward FA was observed with the further prolonging of reaction time, which could be ascribable to the appearance of etherification reaction. After 3.0 h of reaction, the total yield of 2,2′-difurfuryl ether and 2-(isopropoxymethyl)furan reached up to ∼15 mol %. Therefore, the optimal catalyst dosage of 75 mg and reaction time of 2.5 h were adopted for the CTH reaction of FFR.

Figure 10

Effect of catalyst dosage and reaction time on transfer hydrogenation of FFR with 2-propanol. Conditions: FFR 1.2 mmol, 2-propanol 15 mL, Zr(OH)4, 443 K, 600 rpm, N2 1 MPa.

Herein, we further evaluated the reusability of the most active catalyst Zr(OH)4 for the transfer hydrogenation of FFR at 443 K. After each recycling experiment, the spent catalyst was separated from the reaction mixture by filtration, followed by being repeatedly washed with plenty of deionized water. After drying at 323 K overnight, the catalyst was tested again under identical conditions. Specific surface area and TGA–DTA characterizations of spent catalyst indicated the adsorption of certain amounts of carbohydrates on the catalyst surface, which prevented the sufficient contact between FFR and acid–basic sites. Thus, slight decrease in the FFR conversion from 99.8 to 92.1 mol % was observed after three consecutive runs, as revealed in Figure . In the following two runs, the catalyst displayed relatively good performance, in which the FFR conversion remained around 90 mol %. Elemental analysis of Zr4+ in the solution after five successive cycles by ICP-AES revealed that Zr4+ concentration was below the limit of detection. The aforementioned results reflected the good stability and reusability of Zr(OH)4.

Figure 11

Reuse of catalyst. Conditions: FFR 1.2 mmol, 2-propanol 15 mL, Zr(OH)4 75 mg, 443 K, 2.5 h, 600 rpm, N2 1 MPa.

Comparison of catalytic activity among various catalysts is presented in the Table . It can be seen that 2-propanol was a common choice and served as an excellent hydrogen donor for the CTH of FFR. In comparison of catalyst type, bifunctional zirconium N-alkyltriphosphate nanohybrid (ZrPN), characteristic of high basicity/acidity, exhibited better activity at 413 K.[30] Homogeneous Lewis acid catalyst (DyCl3) effectively converted FFR into FA using 2-propanol as the solvent and hydrogen donor,[26] however, issues, including product separation and catalyst reuse, were also put forward. Other metallic catalysts, such as γ-Fe2O3@HAP, Fe-L1/C-800, and Cu/AC–SO3H, could give good FFR conversion (above 90%) at relatively higher temperatures or within longer reaction time.[28,29,31] By contrast, Zr(OH)4 synthesized by the simple precipitation method displayed desirable activity toward FFR transfer hydrogenation, and nearly complete conversion rate and 98.9% FA yield were achieved in a 2.5 h reaction.

Table 3

Comparison of Catalytic Activity in FFR Transfer Hydrogenation over Various Reported Catalysts

catalyst	hydrogen donor	T (K)	t (h)	conv. (%)	yield (%)	ref
DyCl3	2-propanol	453	3	100	97	(26)
γ-Fe2O3@HAP	2-propanol	453	3	96.2	91.7	(28)
Fe-L1/C-800	2-butanol	433	15	91.6	76.0	(29)
ZrPN	2-propanol	413	2	98	98	(30)
ZrPPh	2-propanol	413	2	94	78	(30)
Cu/AC–SO3H	2-propanol	423	5	100	99.9	(31)
Zr(OH)4	2-propanol	443	2.5	100	98.9	this work

On the basis of the above experimental results and previous literatures,[32−35,46−48] a plausible reaction mechanism for the CTH reaction of FFR via MPV reduction over Zr(OH)4 was suggested, as illustrated in Scheme . First, adsorption of 2-propanol into the catalyst occurred, followed by dissociation to the corresponding alkoxide A. Next, the aldehyde group in FFR matched well with alkoxide A to produce a six-membered ring transition state B. Then, hydrogen transfer took place between activated FFR and alkoxide in a concerted manner, with the formation of intermediate C. Meanwhile, the new carbonyl chemical acetone was released from intermediate species C and generated the intermediate D. Finally, another 2-propanol molecule participated in the reaction, giving off the end product FA as well as initial active site alkoxide A. As mentioned earlier,[49] each step in the cycle of FFR transfer hydrogenation was supposed to be reversible.

Scheme 2

Proposed Mechanism for the Transformation of FFR into FA Catalyzed by Zr(OH)4 via CTH Reaction with 2-Propanol as a Hydrogen Donor

Conclusions

In summary, we prepared a series of zirconium-based catalysts for the synthesis of FA via a CTH process using 2-propanol as a hydrogen donor. Among these catalysts, Zr(OH)4 with much more medium to strong acid–basic sites and high acid–basic site density was identified as the most active one, in which a yield of 98.9 mol % FA was achieved at nearly 100 mol % conversion. As comparison, ZrO2-x possessing insufficient acid–basic sites only gave poor FA selectivity. High calcination temperatures resulted in the formation of monoclinic ZrO2 in large numbers, but resulting poorer BET surface area and pore volume were not conducive to the transfer hydrogenation of FFR into FA. Relatively, proper amount of monoclinic ZrO2 in calcined samples improved the conversion rate and FA selectivity, and acid–basic site density ratio had slight influence on FFR conversion. Moreover, the poisoning experiments revealed that active basic sites on the catalyst surface displayed pronounced effect for FFR conversion. Furthermore, the Zr(OH)4 catalyst exhibited good performance and stability that was reusable over multiple cycles without significant loss of catalytic activity. The possible reaction mechanism for this CTH process over Zr(OH)4 with 2-propanol as the hydrogen source was proposed.