Hydrogenation of Levulinic Acid Using Formic Acid as

0 downloads 0 Views 611KB Size Report
The Ni/SiO2 catalyst prepared by the citric acid assisted method and calcined in inert gas flow was the most efficient for the hydrogenation of levulinic acid ...
Research Article

Mohan Varkolu1,2,* David Raju Burri1 Seetha Rama Rao Kamaraju1 Sreekantha B. Jonnalagadda2 Werner E. van Zyl2,*

719

Hydrogenation of Levulinic Acid Using Formic Acid as a Hydrogen Source over Ni/SiO2 Catalysts Several Ni/SiO2 catalysts were developed for the hydrogenation of levulinic acid using formic acid as the hydrogen source. The catalysts were prepared by a variety of methods including impregnation, co-precipitation, deposition-precipitation, and citric acid assisted impregnation combustion. The morphological properties were investigated by XRD, N2 sorption, HRTEM, and H2 pulse chemisorption measurements. XRD patterns of the calcined material revealed the presence of NiO particles, while calcination in an inert atmosphere produced Ni particles through in situ reduction of NiO. The reaction proceeded without external H2 flow using formic acid as hydrogen source. The Ni/SiO2 catalyst prepared by the citric acid assisted method and calcined in inert gas flow was the most efficient for the hydrogenation of levulinic acid without external H2 flow. The high catalytic performance was attributed to the high dispersion of cheap and earth-abundant Ni nanoparticles and optimal porosity. Keywords: Biomass, Heterogeneous catalysis, Hydrogenation, Levulinic acid, Nickel Received: August 08, 2016; revised: November 09, 2016; accepted: February 06, 2017 DOI: 10.1002/ceat.201600429

1

Introduction

The 20th century was marked by excessive use of nonrenewable resources originating from the industrial revolution to maintain and support a number of emerging technologies such as energy and transport. The 21st century needs to refocus and shift toward a more efficient utilization of renewable resources to aid in energy sustainability and, among other things, lower global warming effects by reducing greenhouse gases. The scientific challenge is to look for alternative, sustainable, and carbon-neutral sources of energy. In the pursuit of virtually inexhaustible renewable energy sources, biomass has emerged as an abundant, inexpensive, and efficient resource due to its facile and ongoing photosynthetic production. The utilization of biomass, derived from a molecular platform as a carbonneutral source to synthesize useful chemicals and fuels, is a topic of current importance. Among the biomass-derived platform molecules, levulinic acid (LA) is one of 12 primary chemicals under consideration due to its enormous downstream utilization [1, 2]. Hydrogenation of levulinic acid to g-valerolactone (GVL) was previously reported over various homogeneous and heterogeneous catalysts in both batch and continuous processes [3], and notably hydrogenation of levulinic acid can proceed by using either molecular H2 or formic acid as hydrogen source. However, the utilization of formic acid as a hydrogen source vastly enhances the entire process, since formic acid is co-produced with levulinic acid by the acid hydrolysis of cellulose and furfuryl alcohol [4, 5].

Chem. Eng. Technol. 2017, 40, No. 4, 719–726

The pioneering work of levulinic acid hydrogenation by Horva´th et al. using homogeneous Ru complexes and formic acid as a hydrogen source at 140 C achieved only a 25 % yield of GVL, but set a key preliminary benchmark [6]. This was followed by a number of other catalyst sources, different operating temperatures, and a range of yields to obtain GVL [7–14]. We note that the majority of reports to date have used expensive Ru-based catalysts in order to scale up the process; our work focuses on a cheaper earth-abundant variant in the development of a base-metal (Ni) catalyst. Another significant observation is that the majority of the results reported to date operated at high pressures, are batch processes, and mostly utilize H2 as an external hydrogen source. A number of catalysts used for the conversion of levulinic acid to GVL with formic acid as hydrogen source were homogeneous, leading to problems associated with the limited recyclability. Recently, we

– 1

Dr. Mohan Varkolu, Dr. David Raju Burri, Dr. Seetha Rama Rao Kamaraju [email protected] Inorganic & Physical Chemistry Division, Indian Institute of Chemical Technology, Uppal Road, Tarnaka, Hyderabad 500607, India.

2

Dr. Mohan Varkolu, Prof. Sreekantha B. Jonnalagadda, Prof. Werner E. van Zyl [email protected] School of Chemistry & Physics, University of KwaZulu-Natal, Westville Campus, Chiltern Hills, King George V Ave, Durban 4000, South Africa.

ª 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.cet-journal.com

Research Article

720

reported a continuous process for the hydrogenation of levulinic acid to GVL using supported Ni catalysts [15–19] including the role of impurities such as water and formic acid [16, 19]. These results seemed to bolster the idea that supported Ni-based catalysts will work for the continuous hydrogenation of levulinic acid without an external hydrogen source. Shell Oil Company previously patented the hydrogenation of levulinic acid and its esters using formic acid as a sole hydrogen source [20], but additional studies reported in the open literature are limited and include the hydrogen-independent hydrogenation of levulinic acid [21, 22]. A few studies have suggested the hydrogenation of levulinic acid using formic acid as hydrogen source [23, 24] due to the fact that formic acid has clear benefits in terms of renewability, cost, and safety compared with gaseous hydrogen derived from petroleum derivatives. However, formic acid decomposes to H2 and CO2 but, on the other hand, in situ liberated H2 is more reactive than ex situ H2 sources. Moreover, it is well known that supported Ni-based catalysts can aid in the decomposition of formic acid [24, 25]. Against this background, we have extended our two aims, namely, to achieve good to excellent yields of GVL without using an external hydrogen source. The reaction is shown in Scheme 1. The silica-supported non-noble-metal (Ni) catalysts were found to be effective for the hydrogenation of levulinic acid compared with the other metal oxide supported Ni catalysts, which depend significantly on the metal dispersion, metal surface area, and particle size [16]. On the other hand, metal dispersion is intimately dependent on the preparation method, which in turn influences the catalytic performance, as this study shows. Thus, we investigated the influence of different preparation methods of Ni/SiO2 catalysts and their subsequent use as hydrogenation catalysts for levulinic acid without external hydrogen (H2 gas) flow, but using formic acid instead.

2

Experimental Section

2.2 Deposition-Precipitation Method In this method, the admixture of Ni(NO3)26H2O (7.4 g) and SiO2 (3.5 g) was stirred in deionized water. To this mixture, Na2CO3 solution was added dropwise until a pH of 8.5 was reached. The resulting gel was filtered, washed several times with deionized water to remove all impurities, and then dried at 100 C for 12 h. The oven-dried resultant solid was then calcined in a muffle furnace at 450 C for 5 h. The final solid is designated as DP.

2.3 Impregnation Method In this method, an admixture of Ni(NO3)26H2O (7.4 g) and SiO2 (3.5 g) was stirred in deionized water. The resultant gel was kept on a hot plate at 80 C until complete dryness, and then dried further in an oven at 100 C for 12 h. Finally, the oven-dried solid was calcined in a muffle furnace at 450 C for 5 h. The final solid is designated as IM.

2.4 Surface Impregnation Combustion Method In this method, citric acid was used as both a reductant and capping agent for the SiO2 supported Ni nanoparticles during synthesis. An aliquot mixture of precursors [Ni(NO3)26H2O (7.4 g), SiO2 (3.5 g); Ni/citric acid molar ratio: 1:1.5] was added to deionized water [26]. The resultant suspension was kept on a hot plate at 80 C until complete dryness. The resultant solid was dried in an oven at 120 C for 12 h, and the dried solid was then divided into two portions. One portion was calcined in static air in a muffle furnace and the other portion was calcined in an inert atmosphere in a downflow fixed bed reactor (30 mL min–1 of N2 flow) at 450 C for 5 h. The final solids are designated as SG O2 and SG N2, respectively.

2.1 Co-precipitation Method 2.5 Catalyst Characterization In this method, an aqueous solution of Ni(NO3)26H2O (7.3 g) was added dropwise to an aqueous solution of sodium silicate (7.1 g) with vigorous stirring. The resultant precipitation was filtered and washed with deionized water. The resultant solid was then dried in an oven at 100 C for 12 h. The oven-dried sample was calcined in a muffle furnace with a temperature ramping of 10 K min–1 to 450 C held for 5 h. The final solid is designated as CP.

XRD patterns of the catalysts were recorded on a Rigaku Ultima-IV (Rigaku Corporation, Japan) X-ray diffractometer using Ni-filtered Cu-Ka radiation (l = 1.5406 Å) with a scan speed of 4 min–1 and a scan range of 2–80 at 40 kV and 20 mA. The BET surface area, pore size, and pore volume of all the catalysts were determined by N2 physisorption at liquid N2 temperature (–196 C) with an ASAP 2020 Adsorption unit (Micromeritics, USA). Prior to physisorption, the catalyst samples were degassed under vacuum at 250 C for 1 h to remove the physisorbed moisture. Temperature-programmed reduction (TPR) of the fresh and spent catalysts was performed on a homemade reactor setup. About 50 mg of catalyst was placed in a quartz reactor and pretreated in an Ar flow at 100 C for 2 h. A flow of a 5 % H2/Ar gas mixture Scheme 1. Hydrogenation of levulinic acid with formic acid (as hydrogen source), forming (60 cm3 min–1) with a temperature GVL in high yield.

Chem. Eng. Technol. 2017, 40, No. 4, 719–726

ª 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.cet-journal.com

Research Article

721

ramping of 10 K min–1 was maintained. The hydrogen consumption was monitored using a thermal conductivity detector (TCD). The morphological features of the catalysts were obtained using a JEOL JEM 2000EXII transmission electron microscope, operating between 160 and 180 kV. The specimens were prepared by dispersing the samples in acetone using an ultrasonic bath and evaporating a drop of the resultant suspension onto the carbon support grid. A pulse (100 mL) titration procedure was carried out for H2 chemisorption at 40 C on an Autosorb-iQ automated gas sorption analyzer (Quantachrome Instruments, USA) to determine the dispersion and metal particle size and metal surface area of the catalyst. Prior to the experiment, the catalyst was reduced at 500 C for 2 h followed by evacuation for 2 h.

2.6 Reaction Procedure Catalytic tests were performed at atmospheric pressure in a glass downflow fixed bed reactor (14 mm inner diameter, 200 mm length) loaded with 1 g of catalyst and mixed with 1 g of quartz beads. Before starting, the catalysts were reduced at 500 C online for 4 h in H2 flow (30 mL min–1). The feed solution consisting of the required molar ratio of levulinic acid and formic acid was fed at a flow rate of 1 mL h–1 through the microperfusor feed pump in a stream of N2 flow (30 mL min–1). The product samples were collected at the outlet of the reactor and analyzed with a gas chromatograph equipped with a flame-ionization detector, and in particular, formic acid decomposition to H2 and CO2 was analyzed by a GC equipped with a TCD. Furthermore, the conversion of formic acid is not considered because formic acid is used as the hydrogen source.

3

Results and Discussion

Figure 2. XRD patterns of reduced Ni/SiO2 catalysts prepared by various methods.

3.1 Structural and Textural Properties The XRD patterns (Fig. 1) of as-synthesized samples from the co-precipitation and deposition-precipitation methods did not

Figure 1. XRD patterns of calcined Ni/SiO2 catalysts prepared by various methods.

Chem. Eng. Technol. 2017, 40, No. 4, 719–726

show any diffraction peaks associated with Ni/NiO species due to their being either too highly dispersed or below the detection limit of the XRD. However, for samples prepared by the impregnation and sol-gel methods, diffraction peaks attributed to Ni/NiO species were clearly observed. Interestingly, the presence of diffraction peaks at 19.7, 26.9, 33.9, 53.2, and 60.4 indexed as (100), (103), (110), (200), and (300) planes, in the XRD patterns of the co-precipitation and deposition-precipitation samples after calcination were assigned to nickel silicates [27]. No reflections were noted for Ni in the samples from the co-precipitation and deposition-precipitation methods due to their being either too highly dispersed or below the detection limit of the XRD. To confirm the presence of Ni in these catalysts, we reduced the samples and analyzed the XRD patterns of the resulting catalysts. The results indicated the presence of Ni in these catalysts (Fig. 2). The reflections observed at 36.7, 42.7, 62.2, 74.83, and 79.23 were indexed as (111), (200), (220), (311), and (222) (ICDD No. 88-2326) and attributed to NiO in the case of the impregnation and citric acid assisted method calcined in static air.

All the diffraction patterns of the reduced catalysts are presented in Fig. 2. All the patterns that showed peaks at 2q = 44.5, 53.0, and 78.3 were indexed as the (111), (200), and (220) planes (ICDD No. 88-2326) and correspond to Ni. However, there existed a small amount of NiSiO3 in the case of the co-precipitation and deposition-precipitation methods after reduction [27]. The XRD patterns of Ni particles formed by means of CP, DP, and sol-gel (calcined in N2 flow) methods were rather broad and weak compared to IM and sol-gel (calcined in O2 flow) methods. While the peak pertaining to Ni in the case of the IM method was the strongest, it implies aggregation of Ni particles (i.e., larger Ni particles) over the SiO2 support. The specific surface areas, pore sizes, and pore volumes of the Ni/SiO2 catalysts prepared by the various methods are shown in Tab. 1. The specific surface areas of the Ni/SiO2 catalysts prepared by the IM, CP, DP, and sol-gel (SG O2 and SG N2 samples) methods were calculated as 138, 326, 213, 187, and 301 m2 g–1, respectively. In particular, the high specific surface areas of the SG O2 and SG N2 samples were attributed to

ª 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.cet-journal.com

Research Article

722

the porosity generated through the citric acid [28]. The N2 adsorption-desorption isotherms (Fig. 3) of the SG O2 and N2 catalysts showed characteristic type-IV isotherms, which was ascribed to the capillary condensation of pores formed during the synthesis of catalysts with H1-type hysteresis according to IUPAC classification [29], and the pore size also fell in the mesoporous range (Fig. 4). These results are in agreement with previous reports [28, 30]. In addition, the co-precipitation samples showed characteristic type-IV isotherms assigned to slitlike pores, which presumably arise through the use of tetraethyl orthosilicate. In the samples obtained through the impregnation and deposition-precipitation methods, no such hysteresis was observed. Table 1. Physicochemical properties of Ni/SiO2 catalysts prepared by various methods. Catalyst

Composition of Ni by ICP [wt %]

SBET [m2 g–1]

Vp [cc g–1]

Dp [nm]

CP

28.8

325.7

0.86

10.9

DP

29.1

213.3

0.24

17.8

IM

28.9

138.1

0.26

3.2

SG N2

28.5

301.2

0.42

4.6

SG O2

28.6

187.4

0.25

4.5

The pore diameters obtained through the IM, CP, and DP methods were in the microporous range, while those obtained through the SG N2 and SG O2 methods could be in the mesoporous range, which was attributed to the citric acid.

3.2 Reduction Behavior of Catalysts The H2 TPR profiles of all the catalysts prepared by the various methods are shown in Fig. 5. The NiO was directly reduced to Ni without any intermediate steps and therefore the hydrogen consumption at various temperature ranges could be attributed to the interaction with the support. Two peaks centered at 422 and 630 C were observed in the case of the deposition-precipitation method. The peak at the lower temperature is attributed to a weak interaction while the peak at the higher temperature is assigned to a strong interaction with the SiO2 support [31]. The same two peaks were observed for the impregnation method, with the major peak at around 430 C and a smaller peak at 550 C [32]. Three peaks were detected in the case of

Figure 3. N2 adsorption-desorption isotherms of Ni/SiO2 catalysts prepared by various methods.

Ni/SiO2 prepared via the citric acid assisted method and calcined in air at ca. 500, 680, and 800 C ascribed to weak and strong interaction with SiO2, while one small peak at 500 C was observed in the case of Ni/SiO2 prepared via the citric acid assisted method and calcined under inert atmosphere. Additionally, one negative peak was observed, which was assigned to the methanation of carbon [33] through incomplete combustion of citric acid and presumably formed during the calcination step. Calcination is an important step in obtaining highly dispersed supported metal oxide catalysts, and a number of alternative processes have been reported. The influence of inert gas flow on the dispersion of metal oxide catalysts such as Ni, Co, and CuO has been reported and better enables dispersion during calcination [34–38], while O2 flow can also promote the sintering during calcination. The results of our study are in good agreement with these reports and it is known that the cal-

Figure 4. Pore size distribution of Ni/SiO2 catalysts prepared by various methods.

Chem. Eng. Technol. 2017, 40, No. 4, 719–726

ª 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.cet-journal.com

Research Article

723

Figure 5. TPR profiles of calcined Ni/SiO2 catalysts prepared by various methods.

cination of catalysts in an inert gas atmosphere in the presence of citric acid generates carbon species (around 3 wt % as evidenced from the CHN analysis). The TEM and HRTEM analysis of the sample prepared by SG N2 is shown in Fig. 6. As seen in Fig. 6 a, dark spots attributed to Ni particles were dispersed throughout the silica matrix; analogous results have been reported [28, 39]. In order to get a clear indication of particle distribution, we performed HRTEM studies as seen in Figs. 6 b and 6 c. The Ni particles are spherical and homogeneously distributed. The average particle size of Ni nanoparticles is 5.7 ± 3 nm (Fig. 6 d) and the results are in agreement with H2 pulse chemisorption studies at 9.85 nm. For comparison, we performed the TEM analysis of the IM catalyst (Fig. 6 e) and found aggregated Ni particles. The results obtained for H2 chemisorption including hydrogen uptake (Nm), active metal surface area (AMSA), dispersion (D), and average particle size are presented in Tab. 2. The Ni dispersion (D, %), Ni metal surface area (AMSA), and particle size were calculated based on the assumption that every Ni atom chemisorbs one hydrogen atom. We have previously reported the calculation procedure [15]. The dispersion of catalysts prepared by various methods follows the order SG N2 > SG O2 > IM > CP > DP. The AMSA values are also in agreement with the dispersion.

erably from 53.5 to 93 % as the molar ratio (LA:FA) increased from 1:1 to 1:5. With further increase of the molar ratio, the conversion is almost complete but we noticed a lower yield of GVL, presumably due to the formation of side products (i.e., valeric acid) attributed to the Brønsted acid nature of formic acid. We have previously observed valeric acid formation by the utilization of the bifunctional Ni/HZM-5 catalyst [15]. Furthermore, as we previously reported, the reaction proceeds through two different pathways, namely, (i) cyclization (angelica lactone intermediate) followed by hydrogenation [15], and (ii) hydrogenation (4-hydroxyvaleric acid intermediate) followed by cyclization [16]; the former dominated over acidbased catalytic sites and the latter dominated over metallic sites. In the present study, we noticed the acid sites dominated the mechanism because the one reactant acts as either catalyst or as a hydrogen source. The findings in our approach are in agreement with reported results [40]. The results indicated that a molar ratio of 1:5 is optimum for both levulinic acid conversion and the GVL yield. In contrast, a

a)

b)

d)

c)

e)

3.3 Activity Measurements The results for the hydrogenation of levulinic acid without an external hydrogen source are summarized in Fig. 7. The yield of GVL increased consid-

Chem. Eng. Technol. 2017, 40, No. 4, 719–726

Figure 6. Representative TEM and HRTEM images of Ni/SiO2 catalysts prepared by the surface impregnation combustion method and calcined in an inert gas flow, i.e., SG N2. (a) TEM, (b,c) HRTEM and (d) particle size distribution, (e) IM catalyst.

ª 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.cet-journal.com

Research Article

724

Table 2. H2 uptake (Nm), dispersion (D), metal surface area (AMSA), and average particle size of Ni/SiO2 catalysts prepared by various methods. Catalyst

Nm [mmol g–1]

AMSA [m2 gcat–1]

AMSA [m2 gNi–1]

Average particle size [nm]

CP

183.50

14.35

47.85

14.09

7.18

DP

177.31

13.87

46.24

14.58

6.94

IM

195.01

15.25

50.85

13.26

7.63

SG N2

262.48

20.53

68.44

9.85

10.27

SG O2

210.66

16.48

54.93

12.27

8.24

Conversion (%)

Conversion (%)

Yield (%)

D [%]

Yield (%)

100 100 90

80

80

60

70 60

40

50 40

20

30

0

20

01:01

10

01:03

01:05 Molar rao (LA: FA)

01:07

CP

Figure 7. Catalytic performance of the SG N2 catalyst with various molar ratios. Reaction conditions: catalyst = 1 g, temperature = 250 C, pressure = 1 atm.

lower molar ratio of 1:1 resulted in a lower GVL yield due to insufficient H2 generation, which is required for the hydrogenation of levulinic acid. This allowed us to identify the intermediates as byproducts (i.e., angelica lactone). A further increase in the molar ratio led to a decrease in intermediates. It is well known that levulinic acid hydrogenation proceeds through an intermediate cyclization step and results in the formation of angelica lactone. Through hydrogenation of the intermediate, the target product is formed in the presence of a Brønsted acid. Conversely, levulinic acid hydrogenation can proceed through hydrogenation of the keto group present in levulinic acid (this results in the formation of 4-hydroxyvaleric acid), which leads to GVL upon dehydration [16]. In the present study, we observed a cyclization followed by hydrogenation of the intermediate (as evidenced by GC-MS) to form the GVL product. The acidity is also another factor to obtaining GVL. Among the tested catalysts, the SG N2 catalyst led to the highest selectivity for GVL, which was attributed to its high acidity (data not shown) compared with the other catalysts used in this study. Inspired by the aforementioned results, we then tested the hydrogenation of levulinic acid without external hydrogen (i.e., using formic acid as hydrogen source) over Ni/SiO2 catalysts, as shown in Fig. 8. The catalyst prepared by the citric acid

Chem. Eng. Technol. 2017, 40, No. 4, 719–726

0 DP

SG N2 SG O2

IM

Figure 8. Catalytic performance of Ni/SiO2 catalysts prepared by various methods. Reaction conditions: catalyst = 1 g, temperature = 250 C, levulinic acid:formic acid = 1:5, pressure = 1 atm.

assisted method calcined under inert atmosphere appeared to be superior in affording the highest yield of GVL among all catalysts tested. The high catalytic performance was most likely due to the high dispersion of Ni nanoparticles (HRTEM) and lower particle size (TEM and H2 chemisorption) achieved, and also the optimum porosity achieved by the citric acid that was used as both a reducing agent and capping agent (N2 adsorption-desorption isotherms) [28]. In addition to the maximum catalytic performance, stability of the catalyst system is crucial for industrial applications. In this context, a time-on-stream study was performed to determine the stability of the catalyst for the hydrogenation of levulinic acid without an external hydrogen source (Fig. 9). A considerable decrease in conversion was noticed during the time on stream, which could be attributed to the coking of the catalysts as a consequence of acidity (caused by the formic acid), which in turn promotes the condensation of reaction intermediates; this type of deactivation is possible through the presence of Brønsted acids [15]. Furthermore, some of us have reported that this can be avoided by decreasing the acidity through the carbon coating [18]. There is still room for the development of a more efficient catalyst to enhance the stability of the system.

ª 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.cet-journal.com

Research Article

725

cess at atmospheric pressure. The catalytic results revealed that the prepa100 ration method plays a decisive role in the catalytic performance. Among the 80 tested catalysts, the Ni/SiO2 catalyst prepared by the citric acid assisted 60 method and calcined in inert gas flow was the most efficient for the hydro40 genation of levulinic acid without 20 external hydrogen flow. The high catalytic performance is presumably due 0 to the high dispersion of cheap and 1 2 earth-abundant Ni nanoparticles and 3 4 5 optimal porosity that could be 6 7 8 achieved. Although a better catalytic 9 10 Time (h) performance was seen by the system parameters as described, it is not Figure 9. Time-on-stream study over SG N2 catalyst. Reaction conditions: catalyst = 1 g, temwithout further challenges, since the perature = 250 C, levulinic acid:formic acid = 1:5, pressure = 1 atm. time on stream showed a decline in We probed the reasons for the deactivation of the catalyst by catalytic performance. Further work is in progress to remove using a number of different techniques, e.g., CHNS analysis, or decrease this deactivation. regeneration of catalyst (calcination followed by reduction), ICP-MS, and pulse chemisorption. CHNS analysis revealed Acknowledgment 10 % coking of the catalyst. The result for the regenerated catalyst closely resembled the fresh catalyst, which we ascribed to M.V. is grateful to the Council of Scientific and Industrial the coking of the catalyst through the condensation of reaction Research, New Delhi, India, and the School of Chemistry & intermediates. These results collectively suggest that the coking Physics, University of KwaZulu-Natal, South Africa, for the might be responsible for the deactivation of catalytic activity facilities and financial support. during the time on stream. Furthermore, to get additional insight into the deactivation process, we performed pulse The authors have declared no conflict of interest. chemisorption of the spent catalyst and observed that there was no change in the dispersion of the catalyst (10.02 %). This also suggests that the coking is only responsible for the deactiAbbreviations vation of catalytic performance during the time on stream. However, one cannot ignore possible leaching of the catalyst, to AMSA active metal surface area which end we performed ICP-MS analyses of both fresh and D dispersion spent catalysts and did not find any leaching of Ni species. GVL g-valerolactone It is noted that the majority of reports are batch processes, LA levulinic acid and a few have utilized formic acid as hydrogen source but TCD thermal conductivity detector with expensive Ru catalysts in comparison to our cheap earthTPR temperature-programmed reduction abundant Ni. Moreover, a few studies have reported that Ni promotion made the catalysts efficient for the hydrogenation of levulinic acid without external hydrogen [14, 24, 41], but only References the present study has to date achieved maximum yield in a continuous process. Patented work over Ni/SiO2 catalysts [1] J. J. Bozell, L. Moens, D. C. Elliott, Y. Wang, G. G. Neuenscreported low yields of GVL, which were improved through wander, S. W. Fitzpatrick, R. J. Bilski, J. L. Jarnefeld, Resour., platinum promotion [20]; in our case, we obtained a maximum Conserv. Recycl. 2000, 28, 227–239. yield, avoiding additive or expensive noble metals. This catalyt[2] P. Gullon, A. Romani, C. Vila, G. Garrote, J. C. Parajo, ic system could be advantageous in view of process economics Biofuels, Bioprod. Biorefin. 2012, 6, 219–232. due to the inexpensive Ni and continuous operation. Conversion (%)

Yield (%)

[3]

4

Conclusions

[4] [5]

Various preparation methods, including impregnation, coprecipitation, deposition-precipitation, and citric acid assisted impregnation combustion were employed for the preparation of 30 wt % Ni/SiO2 catalysts for levulinic acid hydrogenation with formic acid as a hydrogen source in a continuous pro-

Chem. Eng. Technol. 2017, 40, No. 4, 719–726

[6] [7]

X. Tang, X. Zeng, Z. Li, L. Hu, Y. Sun, S. Liu, T. Lei, L. Lin, Renewable Sustainable Energy Rev. 2014, 40, 608–620. D. J. Hayes, Catal. Today 2009, 145, 138–151. Z. Zhang, K. Dong, Z. Zhao, ChemSusChem 2011, 4, 112– 118. H. Mehdi, V. Fa´bos, R. Tuba, A. Bodor, L. T. Mika, I. T. Horva´th, Top. Catal. 2008, 48, 49–54. J. C. Serrano-Ruiz, D. J. Braden, R. M. West, J. A. Dumesic, Appl. Catal., B 2010, 100, 184–189.

ª 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.cet-journal.com

Research Article

[8] [9] [10] [11] [12] [13] [14] [15] [16] [17]

[18] [19] [20] [21] [22] [23] [24]

[25] [26]

726

L. Deng, Y. Zhao, J. Li, Y. Fu, B. Liao, Q.-X. Guo, ChemSusChem 2010, 3, 1172–1175. D. Kopetzki, M. Antonietti, Green Chem. 2010, 12, 656–660. M. Varkolu, Catalytic Applications over Supported Ni Catalysts, LAP Lambert Academic Publishing, Saarbru¨cken 2017. C. Ortiz-Cervantes, J. J. Garcı´a, Inorg. Chim. Acta 2013, 397, 124–128. X.-L. Du, Q.-Y. Bi, Y.-M. Liu, Y. Cao, K.-N. Fan, ChemSusChem 2011, 4, 1838–1843. J. Yuan, S. S. Li, L. Yu, Y. M. Liu, Y. Cao, H. Y. He, K. N. Fan, Energy Environ. Sci. 2013, 6, 3308–3313. A. M. Hengne, A. V. Malawadkar, N. S. Biradar, C. V. Rode, RSC Adv. 2014, 4, 9730–9736. V. Mohan, C. Raghavendra, C. V. Pramod, B. David Raju, K. S. Rama Rao, RSC Adv. 2014, 4, 9660–9668. V. Mohan, V. Venkateshwarlu, C. V. Pramod, B. David Raju, K. S. Rama Rao, Catal. Sci. Technol. 2014, 4, 1253–1259. V. Mohan, B. David Raju, K. S. Rama Rao, in Proc. of Int. Conf. on New Dimensions in Chemistry & Chemical Technologies – Applications in Pharma Industry (NDCT-2014), Spectrum Publishers, Karachi 2014, 94–96. V. Mohan, B. David Raju, K. S. Rama Rao, J. Catal. Catal. 2015, 2, 33–38. M. Varkolu, V. Velpula, S. Ganji, D. R. Burri, S. S. R. Kamaraju, RSC Adv. 2015, 5, 57201–57210. R. J. Haan, J. P. Lange, L. Petrus, C. J. M. P. Hoogenbosch (Shell Oil Company), US Patent 0208183A1, 2007. L. Deng, J. Li, D. M. Lai, Y. Fu, Q. X. Guo, Angew. Chem. Int. Ed. 2009, 48, 6529–6532. X. L. Du, L. He, S. Zhao, Y. M. Liu, Y. Cao, H. Y. He, K. N. Fan, Angew. Chem. Int. Ed. 2011, 50, 7815–7819. M. Varkolu, Ph.D. Thesis, Osmania University, India 2014. P. P. Upare, M. G. Jeong, Y. K. Hwang, D. H. Kim, Y. D. Kim, D. W. Hwang, U. H. Lee, J. S. Chang, Appl. Catal., A 2015, 491, 127–135. E. Iglesia, M. Boudart, J. Catal. 1983, 84, 204–238. L. Shi, K. Tao, T. Kawabata, T. Shimamura, X. J. Zhang, N. Tsubaki, ACS Catal. 2011, 1, 1225–1233.

Chem. Eng. Technol. 2017, 40, No. 4, 719–726

[27] [28] [29] [30] [31] [32] [33]

[34]

[35] [36]

[37]

[38]

[39] [40]

[41]

Z. Liu, M. Li, X. Yang, M. Yin, J. Ren, X. Qu, Biomaterials 2011, 32, 4683–4690. R. Takahashi, S. Sato, T. Sodesawa, M. Suzuki, N. Ichikuni, Microporous Mesoporous Mater. 2003, 66, 197–208. F. Rouquerol, J. Rouquerol, K. S. W. Sing, Adsorption by Powders and Porous Solids, Academic Press, London 1999. P. H. Blanco, C. Wu, J. A. Onwudili, V. Dupont, P. T. Williams, Waste Biomass Valorization 2014, 5, 625–636. A. M. Diskin, R. H. Cunningham, R. M. Ormerod, Catal. Today 1998, 46, 147–155. F. Pompeo, N. N. Nichio, M. G. Gonzalez, M. Montes, Catal. Today 2005, 107–108, 856–862. V. Mohan, C. V. Pramod, M. Suresh, K. Hari Prasad Reddy, B. David Raju, K. S. Rama Rao, Catal. Commun. 2012, 18, 89–92. J. R. A. Sietsma, J. P. Meeldijk, J. P. den Breejen, M. V. Helder, A. J. Dillen, P. E. de Jongh, K. P. de Jong, Angew. Chem. Int. Ed. 2007, 46, 4547–4549. J. R. A. Sietsma, P. E. de Jongh, A. J. van Dillen, K. P. de Jong, Stud. Surf. Sci. Catal. 2006, 162, 95–102. J. R. A. Sietsma, J. D. Meeldijk, M. V. Helder, A. Broersma, A. J. Dillen, P. E. de Jongh, K. P. de Jong, Chem. Mater. 2008, 20, 2921–2931. P. Munnik, M. Wolters, A. Gabrielsson, S. D. Pollington, G. Headdock, J. H. Bitter, P. E. de Jongh, K. P. de Jong, J. Phys. Chem. C 2011, 115, 14698–14706. J. R. A. Sietsma, H. Friedrich, A. Broersma, M. V. Helder, A. J. Dillen, P. E. de Jongh, K. P. de Jong, J. Catal. 2008, 260, 227–235. M. A. Ermakova, D. Y. Ermakov, Appl. Catal., A 2003, 245 (2), 277–288. J. M. Bermudez, J. A. Mene´ndez, A. A. Romero, E. Serrano, J. Garcia-Martinez, R. Luque, Green Chem. 2013, 15, 2786–2792. M. Varkolu, V. Velpula, D. R. Burri, S. R. R. Kamaraju, New J. Chem. 2016, 40, 3261–3267.

ª 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.cet-journal.com