Catalytic Hydrogenation of Levulinic Acid in Water into-Valerolactone over Bulk Structure of Inexpensive Intermetallic Ni-Sn Alloy Catalysts

A bulk structure of inexpensive intermetallic nickel-tin (Ni-Sn) alloys catalysts demonstrated highly selective in the hydrogenation of levulinic acid in water into -valerolactone. The intermetallic Ni-Sn catalysts were synthesized via a very simple thermochemical method from non-organometallic precursor at low temperature followed by hydrogen treatment at 673 K for 90 min. The molar ratio of nickel salt and tin salt was varied to obtain the corresponding Ni/Sn ratio of 4.0, 3.0, 2.0, 1.5, and 0.75. The formation of Ni-Sn alloy species was mainly depended on the composition and temperature of H2 treatment. Intermetallics Ni-Sn that contain Ni3Sn, Ni3Sn2, and Ni3Sn4 alloy phases are known to be effective heterogeneous catalysts for levulinic acid hydrogenation giving very excellence -valerolactone yield of >99% at 433 K, initial H2 pressure of 4.0 MPa within 6 h. The effective hydrogenation was obtained in H2O without the formation of by-product. Intermetallic Ni-Sn(1.5) that contains Ni3Sn2 alloy species demonstrated very stable and reusable catalyst without any significant loss of its selectivity. © 2015 BCREC UNDIP. All rights reserved.


Introduction
-valerolactone (GVL) has been identified as one of the most promising renewable molecules that can be converted into a variety of intermediate chemicals, from which a diverse range of biofuels as well as commodities and fine chemicals [1][2][3][4][5][6][7].GVL has been accepted to use a solvent for lacquers, insecticides and adhesives and some use in cutting oil, brake fluid and as a coupling agent in dye bath [2][3][4].
Previously, GVL is typically obtained from LA by catalytic hydrogenations in liquid phase in a batch reactor system.High temperature, high H2 pressures, and noble metal-based catalysts were required to obtain high yield of GVL [4][5][6][7][8][9].Christian et al. have reported LA hydro-genation at 493 K, 48 bar H2 pressure over Raney Ni catalyst to give 94% yield of GVL, while over copper-chromite catalyst at 523 K and 202 bar H2 pressure, resulted in a complex mixture of products composed of 11% GVL, 44% 1,4-PeD (1,4-pentanediol) and 22% water containing small amount of methyl tetrahydrofuran (MTHF) [8].Schulte and Thomas reported the hydrogenation of LA by using platinum oxide catalyst in different organic solvents for 44 h, at 3 atm of H2 to give 87% GVL yield [9]. Recent reported results showed that even though reaction conditions has been mild conditions however the employing noble metal catalysts are required to achieve high yield of GVL [10][11][12].Bourne et al. used supercritical CO2 for the hydrogenation of LA over 5% Ru/SiO2 at 473 K and 100 bar hydrogen pressure with high yield of GVL [10].The liquid phase hydrogenation of LA to GVL has been reported over 5% Ru/C in a batch reactor and obtained 99% selectivity to GVL at 92% conversion of LA at 403 K and 12 bar H2 pressure in methanol solvent [11].Most recently, Upare et al. reported the vapor phase hydrogenation of LA to VGL in continuous down flow over 5% Ru/C with almost 100% GVL selectivity [12].Although several works have been reported as described above, the noble metal catalysts such as Rh, Ru, Pd, and Pt were mainly employed making it high cost and less favorable in point of view of industrial application.Therefore the search of a new facile, cost effective without the employing of noble metal catalyst has really been attracted so far.
We have reported a facile and efficient synthesis method of bimetallic Ni-Sn alloys both bulk and supported without organometallic tin precursor as starting materials via hydrothermal route from two types nickel metal precursor and its catalytic performances on [13,14] and selective hydrogenation of biomass-derived furfural into furfuryl alcohol with excellent activity and selectivity [15].Firstly, from nickel chloride hexahydrate (NiCl2.6H2O)and tin chloride dihydrate (SnCl2•2H2O) to produce bulk and supported Ni-Sn alloy catalysts [13] and secondly, from Raney nickel supported on aluminium hydroxide (R-Ni/AlOH) and tin chloride dihydrate (SnCl2•2H2O) to obtain Ni-Sn alloy supported on aluminium hydroxide (Ni-Sn/AlOH) [14,15] and applied for the chemoselective hydrogenation of α,βunsaturated carbonyl compounds into unsaturated alcohol and selective hydrogenation of biomass-derived furfural into furfuryl alcohol with excellent activity and selectivity.
In this present report, we continue to investigate the catalytic behaviour of bulk structure of intermetallic Ni-Sn alloy in selective hydrogenation of biomass-derived levulinic acid (LA) in water.Bulk structure intermetallic Ni-Sn catalysts were synthesised according to our previous procedure that has been reported elsewhere [13].Five types of intermetallic Ni-Sn with different Ni/Sn molar ratio were synthesised and the effect of reaction parameters such as temperature, initial H2 pressure, time profile, solvent used, and reusability test are investigated.A commercially available of 5% Pd/C was also used as a catalyst for hydrogenation of LA to GVL as a comparison.

Synthesis Ni-Sn Alloys
A general procedure of the synthesis of Ni-Sn alloy with Sn/Ni ratio of 0.67 is described as follows according to our procedure that had been reported elsewhere [13].A typically NiCl2•6H2O (7.2 mmol) was dissolved in deionised water (denoted as solution A), and SnCl2•2H2O (4.8 mmol) was dissolved in ethanol/2-methoxy ethanol (2:1) (denoted as solution B) at room temperature.Solutions A and B were mixed at room temperature; the temperature was subsequently raised to 323 K and the mixture was stirred for 12 h.The pH of the mixture was adjusted to 12 through the dropwise addition of an aqueous solution of NaOH (3.1 M).The mixture was then placed into a sealed-Teflon autoclave for the hydrothermal reaction at 423 K for 24 h.The resulting black precipitate was filtered, washed with distilled water, and then dried under vacuum overnight.Prior to the catalytic reaction, the obtained black powder was treated under hydrogen at 673K for 90 min.

Catalyst Characterization
Powder X-ray diffraction (XRD) measurements were recorded on a Mac Science M18XHF instrument using monochromatic CuKα radiation (λ = 0.15418 nm).The XRD was operated at 40 kV and 200 mA with a step width of 0.02 o and a scan speed of 4 o min -1 (α1 = 0.154057 nm, α2 = 0.154433 nm).ICP measurements were performed on an SPS 1800H plasma spectrometer of Seiko Instruments Inc.

Bulletin of Chemical Reaction Engineering & Catalysis, 10 (2), 2015, 193
Copyright © 2015, BCREC, ISSN 1978-2993 (Ni: 221.7162 nm and Sn: 189.898 nm).The BET surface area (SBET) and pore volume (Vp) were measured using N2 physisorption at 77 K on a Belsorp Max (BEL Japan).The samples were degassed at 473 K for 2 h to remove physisorbed gases prior to the measurement.The amount of nitrogen adsorbed onto the samples was used to calculate the BET surface area via the BET equation.The pore volume was estimated to be the liquid volume of nitrogen at a relative pressure of approximately 0.995 according to the Barrett-Joyner-Halenda (BJH) approach based on desorption data [16] SEM images of the synthesised catalysts were taken on a JEOL JSM-610SEM after the samples were coated using a JEOL JTC-1600 autofine coater.
The H2 uptake was determined through irreversible H2 chemisorption.After the catalyst was heated at 393 K under vacuum for 30 min, it was treated at 673 K under H2 for 30 min.The catalysts were subsequently cooled to room temperature under vacuum for 30 min.The H2 measurement was conducted at 273 K, and H2 uptake was calculated according to the method described in the literature [17,18].

General Procedure for the Transformation of LA
Catalyst (0.05 g), LA (2.0 mmol), H2O (3 ml) as a solvent, and 1,8-octanediol (0.2 mmol) as an internal standard were placed into a glass reaction tube, fitted inside a stainless steel reactor.After H2 was introduced into the reactor with an initial H2 pressure of 4.0 MPa at room temperature, the temperature of the reactor was raised to 433 K.After 6 h, the conversion of LA and the yield of GVL were determined via GC analysis.For reusability test, the used Ni-Sn(1.5)catalyst was easily separated using either simple centrifugation or filtration in air, and then it was utilized repeatedly without any additional treatments.
Analysis of GVL for the hydrogenation of LA in H2O was performed on a Shimadzu GC-8A equipped with a flame ionization detector and with Silicone OV-101 packing.A Shimadzu 14A with a flame ionization detector equipped with an InertCap® capillary column was used for product analyses for the hydrogenation of LA in alcohol or EtOH/H2O solvents.Gas chromatography-mass spectrometry (GC-MS) was performed on a Shimadzu GC-17B equipped with a thermal conductivity detector and with an RT-βDEXsm capillary column. 1 H and 13 C NMR spectra were obtained on a JNM-AL400 spectrometer at 400 MHz and 101 MHz, respectively; samples were dissolved in chloroform-d1 or D2O with TMS as an internal standard.Products were confirmed by the comparison of their GC retention time, mass, 1 H and 13 C NMR spectra with those of authentic samples

BET Surface Area and Ni Active Surface Area
Bulk composition, identified Ni-Sn alloy phases, specific surface area SBET, and crystallite sizes of the synthesized Ni-Sn alloy catalysts are summarized in Table 1.The bulk composition was determined by ICP-AES and the results are closely to the nominal value of the precursors.Five types of intermetallic Ni-Sn systems were successfully synthesized with Ni/Sn ratio of 4.0, 3.0, 2.0, 1.5, and 0.75.H2 maximum uptake for Ni/Sn ratio of 3.0, 1.5, and 0.75 was 12 μmol.g - , 9 μmol.g - , and 5 μmol.g -1 , respectively.The increase of tin amount in Ni-Sn system enhanced SBET significantly, but reduced the H2 uptake reflecting the decrease of the Ni surface activity due to the presence of Sn or the formation of Ni-Sn alloy.We found that nickel active surface area (SNi) was 2.9 m 2 .g - cat, 2.0 m 2 .g - cat, and 0.3 m 2 .g - cat, respectively.It should be noted that our results are comparable to the previous study reported by Komatsu et al. [19] and Takenaka [21].The average Ni-Sn crystallite sizes was derived from Scherrer`s equation to the selected diffraction peaks of Ni-Sn alloy face.In the case of Ni/ Sn ratio of 4.0 and 3.0 Ni3Sn(201) crystallite sizes was 12 nm and 14 nm, respectively.For Ni/Sn ratio of 2.0 and 1.5, Ni3Sn2(101) crystallite sizes was 28 nm and 27 nm, respectively, and Ni3Sn4(112) of Ni/Sn ratio of 0.75 was 23 nm.
TG-DTA measurements were carried out for the bulk Ni-Sn(1.5)sample that obtained after hydrothermal at 423 K for 24 h and the results are shown in Figure 1.TG curves showed two successive weight loss of 8.77% at 333-420 K and 10.52% at 473-675 K which can be attributed to the evaporation of H2O and solvent and transformation of remained NiCl2.6H2Oand SnCl2.2H2Ointo NiO and SnO, respectively [20,21].DTA curves confirmed that two exothermic peaks were observed at 548.7 K and 648.8K which occurred simultaneously with the weight loss as shown in TG curves.Furthermore, the less of weight loss during preheated at 473-673 K indicated that Ni-Sn alloy formed during a hydrothermal treatment at 423 K for 24 h (Figure 1).
Low-magnification scanning electron micrographs of Ni-Sn(1.5) after H2 treatment at 673 The XRD patterns of the synthesized Ni-Sn alloys with Ni/Sn ratio of 4.0, 3.0, 2.0, 1.5, and 0.75 after H2 treatment at 673 K for 1 h are shown in Figure 3. Single phase of intermetallic Ni-Sn was formed for each Ni/Sn ratio.No diffraction peaks of metallic nickel or tin were observed in Figure 3a-e.In the case of Ni/Sn ratio of 2.0 and 1.5, a single phase of Ni3Sn2 alloy was formed (Figure 3a-b), while for Ni/Sn ratio of 4.0.and 3.0 gave a single phase Ni3Sn alloy (Figure 3c-d) and single phase of Ni3Sn4 alloy was obtained for Ni/Sn ratio of 0.75 (Figure 3e).

Bulletin of Chemical Reaction Engineering & Catalysis, 10 (2), 2015, 195
Copyright © 2015, BCREC, ISSN 1978-2993 a Determined by ICP-AES.b Based on the crystallographic data ICCD-JCPDS.c Determined by N2 adsorption at 77 K. d Total H2 uptake at 273 K (noted after corrected for physical and chemical adsorption); the value in the parenthesis is nickel active surface area (SNi); nd = not determined.e Ni-Sn alloy crystallite size derived from the Scherrer`s equation.f Ni3Sn(201).g Ni3Sn2(101).h Ni3Sn4(112).It should be noted that our method was able to synthesize a single phase Ni-Sn alloy at 673 K which was much lower than the arc-melting or CVD methods [19,22].

Selective Hydrogenation of LA to GVL
The catalytic activity of the synthesised Ni-Sn alloys was evaluated in the selective hydrogenation of LA to GVL in water according to the reaction of Scheme 1.The results of the selective hydrogenation of LA to GVL over various Ni-Sn alloy catalysts are summarized in Table 2.
In order to understand the insight into the specific catalytic reaction of the bulk structure of intermetallic Ni-Sn alloy system, the aqueous phase hydrogenation of LA to GVL at different reaction conditions was also investigated.The effect of reaction temperature, initial H2 pressure, time profile, and effect of solvent were evaluated for Ni-Sn alloy with Ni/Sn ratio of 1.5.

Effect of Reaction Temperature
The effect of reaction temperature on the product composition in the selective hydrogenation of LA to GVL by means of Ni-Sn(1.5)alloy catalyst is shown in Figure 4. Reaction temperature was varied in the range of 373 K to 435 K at the initial H2 pressure of 4.0 MPa for 6 h.As increase of reaction temperature from 373 K to 403 K, the conversion of LA remarkably increased from 27% to 99%.LA was converted to GVL completely at 413 K after 6 h.We noted here that no by-products were formed even reaction temperature was raised up to 453 K.

Time Profile
We also examined the time profile in the selective hydrogenation of LA to GVL at 453 K, initial H2 pressure of 4 MPa for 1-10 h and the composition of the reaction product are shown in Figure 5.The LA conversion within 1 h was 15% and remarkably increased to 95% LA after 2 h at the same reaction condition.LA was converted completely after 6 h with yield and selectivity of >99%.No by-product was observed within the reaction time of 1 h to 10 h.

Effect of Initial H2 Pressure
The effect initial H2 pressure on the product composition in the selective hydrogenation of LA to GVL by means of Ni-Sn(1.5)alloy catalyst is shown in Table 3.Initial H2 pressure was varied in the range of 1.0-5.0MPa.At the initial H2 pressure of 1 MPa, only 8% of LA was converted (entry 1).The LA conversion remarkably increased almost 10 times from 7.8% to 79% when initial H2 pressure increased from 1 MPa to 2.0 MPa (entry 2).At the initial H2 pressure of 4.0 MPa, an excellent yield of GVL (>99%) at the completed reaction was achieved (entry 4. Further increase of the initial H2 pressure up to 5.0 MPa, the LA conversion slightly decreased to 88% at the same reaction condition (entry 5).

Solvent Effect
The effect of solvent used in the selective hydrogenation of LA to GVL by means of Ni-Sn (1.5) alloy catalyst was carried out and the product composition are summarized in Table 4.In ethanol system, 97% of LA was converted and gave 86% of GVL and 14% of by-products (entry 1).LA conversion and GVL selectivity increased significantly to 99% and 97%, respectively when the mixture of ethanol/H2O with volume ratio of 1.5/2.0 was used as a solvent (entry 2).It should be noted that LA conversion and GVL selectivity were >99% when H2O was used as a solvent (entry 3).

Reusability Test
A reusability test was performed on the Ni-Sn(1.5)catalyst, and the results are summarised in Table 5.The used Ni-Sn(1.5)catalyst was easily separated by either simple centrifugation or filtration after the reaction.The activity of the catalyst decreased while the high selectivity was maintained for at least four con-secutive runs.Treatment of the used Ni-Sn(1.5)catalyst (after five runs) with H2 at 673 K for 1 h restored the catalyst's original activity and selectivity.
XRD analysis of the recovered Ni-Sn(1.5)catalyst after the third run of the reaction confirmed that the Ni-Sn alloy structure was maintained.No change Ni-Sn alloy species was observed as shown in Figure 6.The amount of Ni and Sn that leached into the reaction solution was 1.58% and 5.3% after four runs, respectively.

Conclusions
Five types bulk structure of intermetallic Ni-Sn with different Ni/Sn molar ratio were synthesised via a thermochemical method.Intermetallics Ni-Sn that contain Ni3Sn, Ni3Sn2, and Ni3Sn4 alloy phases are known to be effective heterogeneous catalysts for levulinic acid (LA) Engineering & Catalysis, 10 (2), 2015, 194 Copyright © 2015, BCREC, ISSN 1978-2993 K for 1.5 h showed typical granular characteristics of Ni-Sn alloy phases (Figure 2(a).Highmagnification micrographs of the Ni-Sn(1.5)fresh catalyst (Figure 2a-inset) show a predominant porous region and a less-extensive solid phase distributed on top of the recovered Ni-Sn(1.5)catalyst appears to have a predominant porous region and a diffuse phase covering parts of the porous region (Figure 2b-inset).

Figure 1 .Figure 2 .
Figure 1.TG-DTA profiles for the obtained Ni-Sn(1.5) from hydrothermal treatment at 423 K for 24 h Figure 2. SEM images of Ni-Sn(1.5)(a) fresh and (b) recovered after the reaction.

Figure 4 .
Figure 4. Effect of reaction temperature on the product composition over Ni-Sn (1.5) alloy catalyst.Initial H2 pressure of 4.0 MPa and reaction time of 6 h.
Reaction conditions: catalyst, 0.042 g; LA, 1.2 mmol; H2O, 3.5 ml; initial H2 pressure, 4.0 MPa; temp.433 K; reaction time, 6 h. a Determined by ICP-AES.b Determined by GC using an internal standard technique.c R-Ni/AlOH and SnCl2•2H2O were used as starting materials, the value in the parenthesis is Ni/Sn molar ratio [14].d R-Ni/AlOH was prepared by alkali leaching of Raney Ni-Al alloy using a dilute aqueous solution of NaOH according to Petro et al. [23].e Commercially available of Pd/C (5 wt%) and used as received SnCl2

Figure 5 .Figure 6 .
Figure 5.Time profile of the selective hydrogenation of LA to GVL over Ni-Sn(1.5)alloy catalyst at reaction temperature of 433 K and initial H2 pressure of 4.0 MPa

Table 1 .
Bulk composition, identified Ni-Sn alloy phases, SBET, and crystallite sizes of the synthesised intermetallic Ni-Sn catalysts

Table 2 .
Results of the selective hydrogenation of LA to GVL by means of various intermetallic Ni-Sn catalysts

Table 3 .
Effect of initial H2 pressure on the product composition over Ni-Sn(1.5)alloy catalyst.Reaction temperature of 433 K and reaction time of 6 h Reaction conditions: catalyst, 0.042 g; LA, 1.2 mmol; solvent, 3.5 ml; initial H2 pressure, 4.0 MPa; temp.433 K; reaction time, 6 h. a Determined by GC using an internal standard technique.b Determined by GC using an internal standard technique.

Table 4 .
Results of the effect of solvent used in the selective hydrogenation of LA over Ni-Sn(1.5)alloy catalyst.
a Determined by GC using an internal standard technique.b Determined by GC using an internal standard technique.c GC area ratio, 1,4-pentanediol (1,4-PeD) and 2methyltetrahydrofuran (2-MTHF) were detected by GC-MS

Table 5 .
Results of reusability test of Ni-Sn(1.5)alloy catalyst in in hydrogenation of LA into GVL The used catalyst was treated by H2 at 673 K for 1 h before reaction.b Determined by GC using an internal standard technique hydrogenation giving -valerolactone (GVL) yield of >99% at 433 K, initial H2 pressure of 4.0 MPa within 6 h.The effective hydrogenation was obtained in H2O without the formation of by-product.Intermetallic Ni-Sn(1.5) that contains Ni3Sn2 alloy species demonstrated high stability and good reusability without any significant loss of its selectivity. a