- Email: [email protected]
Vapor-phase hydrodeoxygenation of guaiacol on Al-MCM-41 supported Ni and Co catalysts
Accepted Manuscript Title: Vapor-phase Hydrodeoxygenation of Guaiacol on Al−MCM−41 Supported Ni and Co Catalysts Author: Nga T.T. Tran Yoshimitsu Uemura Sujan Chowdhury Anita Ramli PII: DOI: Reference:
S0926-860X(15)30294-5 http://dx.doi.org/doi:10.1016/j.apcata.2015.12.021 APCATA 15694
To appear in:
Applied Catalysis A: General
Received date: Revised date: Accepted date:
3-9-2015 8-12-2015 18-12-2015
Please cite this article as: Nga T.T.Tran, Yoshimitsu Uemura, Sujan Chowdhury, Anita Ramli, Vapor-phase Hydrodeoxygenation of Guaiacol on AlminusMCMminus41 Supported Ni and Co Catalysts, Applied Catalysis A, General http://dx.doi.org/10.1016/j.apcata.2015.12.021 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Vapor-phase Hydrodeoxygenation of Guaiacol on AlMCM41 Supported Ni and Co Catalysts Nga T.T. Tran1, 4*, Yoshimitsu Uemura1*, Sujan Chowdhury2, Anita Ramli3 1
Centre for Biofuel and Biochemical Research, Universiti Teknologi PETRONAS, 32610 Bandar Seri Iskandar, Perak Darul Ridzuan, Malaysia
2
Department of Chemical Engineering, Universiti Teknologi PETRONAS, 32610 Bandar Seri Iskandar, Perak Darul Ridzuan, Malaysia
3
Department of Fundamental & Applied Science, Universiti Teknologi PETRONAS, 32610 Bandar Seri Iskandar, Perak Darul Ridzuan, Malaysia
4
Faculty of Chemical Engineering, HCMC University of Technology, Ho Chi Minh, Vietnam
*
Yoshimitsu Uemura
Universiti Teknologi PETRONAS, 32610 Bandar Seri Iskandar, Perak Darul Ridzuan, Malaysia, Tel: +605-368-7644. E-mail: [email protected] *
Nga T.T. Tran
Tel: +60 11 18593681. E-mail: [email protected]
1
Graphical abstract:
2
Highlights:
HDO of guaiacol on AlMCM41 supported Ni and Co catalysts at 400 oC and 1 atm.
Co is active in HDO via CO hydrogenolysis, while Ni favors multiple CC hydrogenolysis.
Acid sites catalyze demethylation and transalkylation (methyl transfer).
Coke deposition has stronger deactivate effect in HDO than transalkylation.
3
Abstract Vapor-phase hydrodeoxygenation (HDO) of guaiacol, a typical lignin-derived phenolic compound, was studied on AlMCM41 supported Ni and Co catalysts at 400 oC and atmospheric pressure. Ni was found as an active metal for ring opening activity while Co favored the deoxygenation activity. Besides benzene and phenol were observed as the major deoxygenation products, the existence of toluene, cresol and other methylated C712 products was the result of methyl transfer reaction over the acidic function of support. Co/AlMCM41 catalyzed not only HDO to remove oxygen but also transalkylation to prevent the carbon loss via methanization. In addition, coke deposit mainly gathered on the surrounding area of cobalt particles, leading to the faster deactivation in HDO activity than transalykation. Keywords: Hydrodeoxygenation, Cobalt, Nickel, AlMCM41, Guaiacol.
4
1 INTRODUCTION Fast pyrolysis is an attractive thermal conversion process to generate the alternative liquid fuels (bio-oil) [1,2]. Bio-oil obtained from fast pyrolysis contains a large fraction of ligninderived phenolic compounds [1,3–5]. These compounds can be upgraded to renewable fuel source such as aromatic or naphthenic hydrocarbon by hydrodeoxygenation (HDO) [6]. This is a prominent process because it not only removes oxygen, but also preserves the carbon number in the product [7,8]. The final products can be utilized as petrochemical feedstock or high octane gasoline base materials. For example, C68 aromatics are valuable petrochemicals for use as solvents or intermediate in polyester fiber and film manufacture [9], and C9-12 aromatics are the major blending component (1220 % by volume) in gasoline engine fuels to obtain high octane number because there has been a limitation of low vapor pressure aromatics content (e. g., benzene and toluene) [10]. The HDO of pyrolysis oil and model phenolic compounds were mostly conducted at high hydrogen pressure in a batch-wise autoclave (40200 bar) or in a continuous-flow fixed-bed reactor (1040 bar) [3,11–15]. The high hydrogen pressure was required to give high deoxygenation degree but it led to saturate the aromatic ring before eliminating the oxygen groups [13,16–21]. Besides resulting a large consumption of costly hydrogen, this condition also produced ring saturation products with lower octane number and less valuable for cofeeding in existing oil refineries [22,23]. To date, a few studies have been investigated the HDO at low hydrogen consumption (atmospheric pressure) and found that this process had high selectivity to the deoxygenated aromatics [7,8,12,23]. A study by Zhu et al. [23] pointed out the role of metal function and acidic function in HDO of anisole over Pt/SiO2 and Pt/HBeta to form BTX products. In order to have a high selectivity to toluene and high stability of catalyst, Pt/Al2O3 and Pt/SiO2 showed a better catalytic performance than Pt/HB for HDO of m-cresol [12]. However,
5
noble metal catalysts still showed their strong activity in hydrogenation for ring-saturation and ring-opening instead of deoxygenation activity, even though the reaction was conducted at ambient pressure [6,7,20,24]. In addition, the high cost of noble metal is also a disadvantage in their application. Therefore, the non-noble metal may become a solution to these issues. For example, the addition of Fe to Pd or Ni catalyst directed the selection route to oxygen-free aromatic instead of ring-saturated products [7,8]. Moreover, HDO of m-cresol over Pd/Fe2O3 catalyst showed high BTX selectivity (>80%) and less ring saturation product (<1%) that is resulted by the synergic between Pd and Fe [25]. Nevertheless, the iron-based catalysts had lower HDO activity than the precious metal, and mainly formed phenol as a major product [7,26]. In addition, Ni was significantly less active in deoxygenation activity comparing with Pd and Pt catalyst due to its activities in hydrogenation or hydrogenolysis of CC bond [27]. The HDO of guaiacol over transition metal phosphides changed the major product selectivity from phenol to benzene when increasing the contact time (W/F) about 60 times [20,28]. In view of catalyst deactivation during HDO process, coke formation through the deposition of polyaromatics was found to be the main cause instead of sintering [24]. Moreover, the acidic support was deactivated faster than the inert support due to its strong interaction with the coke precursors [12,23,26]. Besides, a zeolite with mesoporous structure showed a much higher stability than the microporous one in the HDO of phenolic compounds [22,23,29]. In this study, we investigated the activity of AlMCM–41 supported non-noble metal (Co and Ni) catalysts for HDO of guaiacol at 400 oC and ambient pressure. The mesoporous AlMCM–41 was selected as the catalyst support since it has good internal mass-transport, high stability and transalkylation activity [23,29]. Guaiacol was chosen as a model compound because it contains both major functional groups of lignin-derived phenolic such as hydroxyl
6
(OH) and methoxy (OCH3). The role of metal and support on catalytic HDO of guaiacol were clarified. Catalyst deactivation and regeneration behavior were also reported. 1. EXPERIMENTAL 1.1. Catalyst preparation Aluminosilicate Al–MCM–41 support (3–4% Al2O3) supplied by ACS Material was activated in air at 500 oC for 14 h. The catalysts containing 10 wt% metal were prepared by the wetness impregnation method. Aqueous solutions containing metal precursor (i.e. nickel nitrate and cobalt nitrate) were first impregnated on the support overnight and then dried at 105 oC for 8 h, followed by calcination at 500 oC for 14 h in air. 1.2. Catalyst characterization The TriStar II 3020 was used to measure the nitrogen adsorption/desorption isotherms at 77K. Transmission electron microscopy (TEM) image was obtained on a Carl Zeiss LIBRA 200FE (operation voltage 200 kV). The catalyst powder was dispersed ultrasonically in 2-propanol for 1 h and then deposited on a carbon coated copper grid for TEM measurement. X-ray diffraction (XRD) pattern was measured on a LabX XRD-6000 X-ray diffractometer (Shimadzu) using Cu K radiation. XRD pattern was recorded by scanning with 0.02o increments in the 270o diffraction angle range. Temperature programmed reduction (TPR) in hydrogen was carried out using a Thermo Scientific TPDRO 1100 instrument. Samples were pretreated in nitrogen at 120 oC during 60 min. Then, a TPR program was performed with a heating rate of 10 oC/min to 900 oC under 5% H2/N2 flow. The amount of hydrogen consumption during catalyst reduction process was monitored by a thermal conductivity detector (TCD).
7
Thermogravimetric analysis (TGA) under the flow of air was conducted in a TA Instrument model QA50. During the analysis, temperature was increased from room temperature to 900 oC at heating rate of 10 oC/min. 1.3. Catalytic performance Catalytic reaction was conducted in a fixed-bed tubular reactor (internal diameter 13 mm) made of stainless steel under ambient pressure. This reactor has two different operating zones where their temperatures can be controlled independently. The top bed functions as a vaporization zone of the feedstock at 350 oC before reaching to the catalyst bed at the bottom zone. The catalyst sample (3560 mesh) was packed in the bottom bed between two layers of glass wool. The Ni catalyst was reduced at 500 oC for 1 h and the Co and NiCo catalysts were reduced at 450 oC for 1 h in H2 flow. The catalyst bed was then cooled to the desired reaction temperature with N2. Guaiacol was fed at a flow rate of 1.08 mL/h using a syringe pump and vaporized in the top glass wool bed. The H2/guaiacol molar ratio was 25 for all runs. The W/F (gcat/greactanth) was adjusted in the range of 01.67 h through varying the amount of catalyst (02.0 g). The reactor outlet line was heated at 195 oC to avoid the condensation of liquid products before they reach the cold trap. All catalytic runs lasted 210 min, and the gas and liquid samples were collected to analyze at every 30 min interval. The liquid products were quantified by gas chromatography (Shimadzu GC-2014), using a SGE BPX5 capillary column and a flame ionized detector (FID) while non-condensed gas was analyzed with GCTCD (Shimadzu GC8A). The product components were identified with GCMS (Agilent Technologies 6890A, BPX5 capillary column), and verified with organic standards. In all experiments, the carbon balance was between 91 and 97 %, depending on the coke deposit on catalyst and the condensation of heavy components inside reactor and piping. Carbon-based guaiacol conversion (XGua), product distribution (Di) and product yield (Yi) were calculated in MolCarbon% by the following equations [7,23].
8
X
Gua
(%)
D i (%)
Mol
gua in Mol
Mol Mol
i
reacted
Mol
gua in
i Gua
gua out
7
100
100
(1)
(2)
where i is the carbon number in product i. Y i (%)
X
Gua
Di
(3)
100
Catalyst regeneration was carried out after 3 h of guaiacol HDO reaction. An air flow was applied to treat the used catalyst at 450 oC for 2 h. Afterwards the catalyst was re-activated in H2 flow for 1 h, and catalyzed a new HDO reaction cycle. 2. RESULTS AND DISCUSSION 2.1. Catalyst characterization Table 1 demonstrates textural properties of metal-modified catalysts and support. After calcination, the color of nickel catalyst sample became gray while cobalt and bimetallic catalysts had a dark brown due to the formation of nickel oxides and cobalt oxides [30,31]. The metal-modified catalysts showed a slight decrease in surface area and pore volume as expected. Moreover, the pore diameters of three Al–MCM–41-supported catalysts also reduced comparing with support only (Error! Reference source not found.). It may be caused by the partial blockage of pores after metal impregnation. However, the mesoporosity of catalysts had not changed significantly. Fig. 1 shows the powder XRD patterns of Al–MCM–41 supported Ni, Co and NiCo catalysts. Both support and metal-modified catalysts have a broad peak at around 23o which is assigned to the amorphous silica in Al–MCM–41 support [32]. Since only the wide-angle XRD was applied, the (100), (110), (200) and (210) peaks assigned to hexagonal structure of Al–MCM–41 were not obviously detected in the range of 17o [30]. The Ni-modified catalyst have three peaks at 37o, 43o and 64o referring to NiO [32,33]. Meanwhile, the Co-modified
9
catalyst has five diffraction peaks at 31.3o, 36.8o, 44.9o, 59.3o, and 65.2o corresponding to different Co3O4 crystallites [31,34,35]. The five major diffraction peaks of NiCo bimetallic catalyst could be NiO, Co3O4 or their composites since these oxides have similar and overlapping XRD reflections [36–39]. The TEM images of three metal-modified catalysts are displayed in Fig. 2, which include the distribution of metal oxide particle size. The 10Ni/Al–MCM–41 had a narrow distribution from 2 to 9 nm particle size as in Fig. 2A. However, a small amount of the larger bulk NiO particles, of which the size was from 50 to 110 nm, was formed on the support. This phenomenon was also observed in nickel catalysts by Wu et al. [32] and Sankaranarayanan et al. [33]. In contrast, cobalt catalyst shows a broad distribution of metal oxide particle size (Fig. 2B). Although most of the cobalt oxide particles were in a size range between 314 nm, the larger particles from 1525 nm were also observed. In contrast to monometallic catalyst, the NiCo bimetallic has a better distribution of particle size which was from 2 to 14 nm and no observation of larger particles. Meanwhile, the estimated particle sizes via XRD in Table 1 were within the range of size distribution via TEM observation. Among three catalysts, bimetallic NiCo had the smallest particle size, and this result was also compatible with TEM measurement. Since the average pore size of Al–MCM–41 was around 3.2 nm, a large amount of metal oxide particles would be located on the external surface of the support. Hence, the porous structure of catalyst did not extensively change after supporting metal. The TPR-H2 profiles of the three metal-supported catalysts are shown in Fig. 3. The reduction of 10Ni/Al–MCM–41 catalyst has two broad peaks observed at 440 oC and 610 oC which is consistent with previous studies [30,40]. The first peak was the reduction of NiO species to the metallic form of nickel [40]. While the second one could be corresponded to the reduction of a cationic form of nickel (NiSiO3) [6,40] and the small NiO particles inside the pore of support [32]. The cobalt catalyst has two reduction zones that could be defined as:
10
one at low-temperature (370 oC) and one at high-temperature (710 oC) which are associated with the reduction of cobalt oxide species with different dispersions and interactions with the support. The doublet peak at lower temperature may be attributed to the reduction of Co3O4 to CoO, and CoO to cobalt metal [6,31,34]. The broad peak at higher temperature could be ascribed to the strong interaction of small cobalt oxide particles with the surface [35]. The reducibility of cobalt species was higher than that of Ni species in their monometallic catalyst. Moreover, the reduction profile of NiCo bimetallic also has two distinct zones. The triplet peak at low temperature could be the simultaneous reduction of cobalt oxides and nickel oxides. The broad peak at higher temperature, which related to the reduction of strong interaction metal oxides with support, was shifted to higher temperature than that of monometallic. In summary, the presence of cobalt improved the reduction behavior of nickel species since it decreased not only the temperature of first reduction peak but also the amount of reduction metal oxides in the second peak. This trend is consistent with the ones already reported elsewhere [41,42]. 2.2. Hydrodeoxygenation of guaiacol The supported metal catalysts were evaluated for HDO of guaiacol at 400 oC under ambient pressure. The hydrotreatments of guaiacol with blank and support-packed reactors were used as references to verify the effect of catalysts and summarized in Error! Reference source not found.. The reactions observed in the blank test were supposed to be homogeneous and catalytic reactions on the internal wall of stainless steel reactor. The conversion of guaiacol over pure AlMCM–41 support predominantly led to the demethylation of guaiacol and formed final products, i.e., catechol and methane. Moreover, methyl-guaiacol was observed as a secondary liquid product formed via transalkylation catalyzed by the acidic support [23].
11
Fig. 4 compares the conversion of guaiacol and product yields over the supported Ni, Co, and NiCo catalysts at the same conditions. Among the three catalysts, 10Co/Al–MCM–41 exhibited the highest HDO activity due to its highest deoxygenation products yield and lowest gas phase yield (mainly C1 products including CH4, CO and CO2). Meanwhile, Nimodified catalyst showed lower activity for HDO of guaiacol since it favored the methanization reaction to form methane as the major product. Even though both cobalt and nickel are 3d transition metals, nickel was found as an active metal for multiple hydrogenolysis of CC bonds (resulted in ring opening), leading to the formation of methane, especially at high temperature [27,43]. The bimetallic 5Ni5Co/Al–MCM–41 catalyst showed substantially higher HDO yield than the average HDO yield of monometallic 10Ni and 10Co catalysts. It proved that cobalt promoted nickel in bimetallic catalyst to enhance the HDO activity. In terms of deoxygenation yield, 10Co/Al–MCM–41 showed to be more attractive catalyst so that this catalyst was chosen for further investigations. The presence of C712 aromatic and phenolic compounds such as toluene, xylene, pentamethylbenzene, hexamethylbenzene, cresol, xylenol, methyl-guaiacol, etc. was resulted of the acid-catalyzed transalkylation of methoxy group to another phenolic ring [23,44–46]. In HDO of guaiacol, especially at high temperature, the CArOCH3 cleavage and the multiple hydrogenolysis of CC bonds led to formation of methane in gas phase. Therefore, the methyl transfer reaction occurred on acid sites of support helped to minimize the carbon loss via methanization. In addition, the observation of some C8+ aromatic and phenolic indicates that the surface methyl groups produced by demethylation were transferred not only within their own molecules but also to other aromatic rings. Fig. 5 illustrates the evolution of guaiacol conversion and product yields over a freshly prepared 10Co/AlMCM41 as a function of W/F. The guaiacol conversion, oxygenated product yields, and non-oxygenated product yields are shown in Fig. 5A, B and C,
12
respectively. Complete conversion of guaiacol was achieved at W/F of 1.67 h. At lower W/F, the major products were phenol, cresol and methane, indicating that hydrogenolysis of CArOCH3 bond and transalkylation were the primary reaction. Since the bond dissociation energy (BDE) of CArOCH3 bond is 45 kJ/mol lower than that of CArOH bond [47], the removal of hydroxyl group considerably occurred at higher W/F, resulting in higher yield of oxygen-free aromatic products. The yield of C8-12 aromatics resulted by multi methyl transfer reactions increased more slowly than those of benzene and toluene when the W/F increased. However, high W/F promoted not only the hydrogenolysis of CO bond in guaiacol molecule but also the multiple hydrogenolysis of CC bonds, leading to an increase in gasification activity. When the W/F ratio was increased to 1.67 h, CO and CO2 started to be obviously detected in the gas phase, and the molar ratio of methane to aromatic ring in liquid phase became higher than the stoichiometry of demethylation reaction. It proved that the gasification occurred significantly at W/F = 1.67 h, yielding high amount of CH4, CO and CO2 in the gas phase. During catalytic hydrotreatment processes, multiple reactions may occur, including hydrogenation, hydrogenolysis, hydrodeoxygenation, hydrocracking, and polymerization [1,3]. Hydrogenation for ring saturation and hydrocracking for gasification make hydrogen consumption exceed the deoxygenated stoichiometric ratio [3]. In order to reduce the hydrogen consumption, the direct deoxygenation without ring saturation was expected in catalytic HDO. The HDO conducted in high pressure continuous reactor showed not only deoxygenation but also ring saturation or ring coupling activities [13,17,33,48,49]. Even at ambient pressure, the low temperature (<300 oC) was conducive to ring saturation [6–8]. In contrast, our experimental runs were carried out at high temperature (400 oC) and low pressure (1 atm) to prevent the hydrogenation/dehydrogenation for ring saturation [7,12,23].
13
Hence, no aromatic saturation product was observed in our study, and it led to lower hydrogen consumption in HDO of guaiacol. Fig. 6 shows the proposed reaction pathways for guaiacol HDO over AlMCM41 supported Co or Ni catalysts based on the changes of products as the functions of W/F and catalyst type. The overall HDO reaction of guaiacol is complex due to the synergic effect of acidic and metallic sites in catalyst, resulting in various intermediates and products as well as many elementary reactions. Considering that the BDE of CArOH, CArOCH3, CArOCH3 are ranked in descending order from 464 kJ/mol to 263 kJ/mol [47], the primary product was changed from catechol in case of pure support to phenol and cresol in case of metal modified catalyst, and further to oxygen-free aromatics at higher W/F ratio. That is, the acidic sites of support only catalyzed the demethylation in C ArOCH3 bond and methyl transfer reaction, while the deoxygenation reaction was catalyzed by metallic sites which facilitated the hydrogenolysis of CArOR bond. Besides some similar activities with Co-modified catalyst such as hydrodeoxygenation, methylation, and transalykation, Ni-modified catalyst principally favored the multiple CC hydrogenolysis to form methane as the final product. In summary, the bifunctional catalyst can convert guaiacol not only to benzene as completely deoxygenated product but also to other methylated aromatics as transalkylated products. 2.3. Coke deposition and catalyst regeneration As shown in Fig. 7, the conversion of guaiacol and the oxygen-free product yields decreased with time on stream (TOS) due to deactivation of catalyst. Considerable deactivation occurred since the conversion of guaiacol decreased from 99.5% to 85.8% after reaction time of 210 min. In earlier state, the yield of oxygen-free aromatic products was dropped rapidly while the monooxygenated phenolics yield increased quickly. After reaching the maximum, selectivity to phenol and methylated phenol products decreased slowly, leading to an increment in the concentration of deoxygenated products at longer time on
14
stream. Hence, the reduction of active metal sites gave more strongly effect to the breaking of CArOH bond compared to that of CArOCH3 bond. Fig. 2D is a representative TEM image of the spent catalyst after two HDO reaction cycles. Coke deposit could be observed as the indeterminate-shaped objects having a larger size and lighter color than the spherical cobalt particles. This TEM observation shows that more carbon deposits were gathered on the surrounding area of cobalt particles than on support, resulting in blockage of reactants approach to metallic surface sites. It means that the coke formations during HDO reaction were not only affected by the acid sites of support but also considerably by the metal sites. Similarly, Zhu et al. [23] concluded that the coke deposition was dependent on the presence of Pt metal, the acidity and pore structure of support. When using inert silica as the support, Olcese et al. [26] found that carbon deposits were mainly observed in the vicinity of iron metal instead of bare silica support due to its very low acidity. The acid sites of support promoted the carbon deposition around metal sites hence the amount of coke formation over Pt/zeolite catalyst was significantly higher than over Pt/SiO2 catalyst [12,23]. In the present work, the smaller accumulation of coke on AlMCM41 support than on the surrounding area of cobalt led to a lower deactivation of transalkylation than hydrodeoxygenation, as evidenced by the stability of cresol and other methylated phenol products yield and the decrease of deoxygenation product yields with time on stream in Fig. 7. The catalysts were deactivated during HDO reactions due to carbon deposition; hence, regeneration ability becomes an important issue in practical application. Fig. 8 illustrates the details of HDO activities of the regenerated 10Co/AlMCM41 at different W/F. Comparing with the fresh 10Co/AlMCM41 catalyst (Fig. 5), the regenerated one showed insignificant difference in guaiacol conversion but the product distribution considerably changed. That is, yield of all aromatics showed a noticeable decrease, especially at high W/F. Moreover, the
15
higher W/F was used, the more decrease in HDO activity of regenerated catalyst was observed. This result indicated that the thickness of catalytic bed might affect to the regeneration efficiency. Fig. 9 shows a TEM image of the regenerated 10Co/AlMCM41 catalyst. Particle size distribution was extracted from measurement of more than 100 particles to avoid the population variance effect. According to the results, particle size of the regenerated catalyst ranged from 2 to 17 nm, which was narrower than that of the fresh catalyst. In addition, the multi-spot EDX scanning which carried out in SEM analysis (data not shown) shown no significant difference in cobalt content between fresh and regenerated catalysts. Hence, the thermal degradation or sintering was not the primary reason of catalyst deactivation. Moreover, the TEM results of regenerated 10Co/AlMCM41 also shown a small amount of coke deposit still remained in catalyst. According to Zanuttini’s research, the coke formation after HDO of m-cresol over Pt/Al2O3 has two oxidation zones at moderate and high temperatures (250 and 550 oC) corresponding to carbon deposited on the metal and on the support, respectively [12]. Meanwhile, the derivative thermogravimetric (DTG) of used 10Co/AlMCM41 catalyst showed a broad peak from 200 to 650 oC which associated with combustion of coke (Fig. 10). It means that the regeneration at 450 oC could not remove all the coke deposited on used catalyst and some active sites were not recovered. These reasons can explain the decrease in HDO activity of the regenerated catalyst. 3. CONCLUSION Hydrodeoxygenation of guaiacol over Al–MCM–41 supported Ni and Co catalysts has been studied at 400 oC and under atmospheric pressure. The Ni catalyst favored the multiple hydrogenolysis of CC bond, leading to carbon loss via methanization. Cobalt not only facilitated the reduction and dispersion of nickel metal oxides but also enhanced the HDO activity of bimetallic NiCo catalyst. Results showed that Co/Al–MCM–41 catalyst was
16
conducive to HDO of guaiacol reaction by a combination of acidic and metallic functions. The acid sites existing on Al–MCM–41 catalyzed the demethylation and transalkylation reactions while metallic cobalt catalyzed the hydrogenolysis CArOR bond to produce oxygen-free products. During HDO reaction, CArOCH3 cleavage occurred preferentially while more active sites were required to catalyze the CArOH cleavage. Deactivation of catalyst affected more strongly the CArOH hydrogenolysis reaction than CArOCH3 hydrogenolysis and transalkylation reaction, resulting in the decrease of oxygen-free aromatic yield.
Acknowledgment The authors gratefully acknowledge the financial support from Mitsubishi Corporation Educational Trust Fund. We wish to thank Universiti Teknologi PETRONAS for providing a congenial work environment and state-of-the-art research facilities. We also thank Prof. Dr. Taufiq Yap Yun Hin for his support in XRD measurement. The authors are also grateful to the anonymous reviewers for their valuable comments on this paper. References [1] J. Wildschut, I. Melián-Cabrera, H.J. Heeres, Applied Catalysis B: Environmental 99 (2010) 298–306. [2] S.R. Naqvi, Y. Uemura, S.B. Yusup, Journal of Analytical and Applied Pyrolysis 106 (2014) 57–62. [3] P.M. Mortensen, J.-D. Grunwaldt, P.A. Jensen, K.G. Knudsen, A.D. Jensen, Applied Catalysis A: General 407 (2011) 1–19. [4] H. Zhang, R. Xiao, H. Huang, G. Xiao, Bioresource Technology 100 (2009) 1428–1434. [5] C.S. Lira, F.M. Berruti, P. Palmisano, F. Berruti, C. Briens, A.A.B. Pécora, Journal of Analytical and Applied Pyrolysis 99 (2013) 23–31. 17
[6] P.T.M. Do, A.J. Foster, J. Chen, R.F. Lobo, Green Chem. 14 (2012) 1388–1397. [7] J. Sun, A.M. Karim, H. Zhang, L. Kovarik, X.S. Li, A.J. Hensley, J.-S. McEwen, Y. Wang, Journal of Catalysis 306 (2013) 47–57. [8] L. Nie, P.M. de Souza, F.B. Noronha, W. An, T. Sooknoi, D.E. Resasco, Journal of Molecular Catalysis A: Chemical 388–389 (2014) 47–55. [9] O.A. Al-Rashed, M.A. Fahim, M. Shaaban, Fluid Phase Equilibria 363 (2014) 248–262. [10] H. Fukuya, N. Okabe, S. Sasaki, Fuel Composition for Use in Gasoline Engines, US8876920 B2, 2014. [11] N. Tran, Y. Uemura, S. Chowdhury, A. Ramli, Applied Mechanics and Materials 625 (2014) 255–258. [12] M.S. Zanuttini, B.O. Dalla Costa, C.A. Querini, M.A. Peralta, Applied Catalysis A: General 482 (2014) 352–361. [13] A.R. Ardiyanti, S.A. Khromova, R.H. Venderbosch, V.A. Yakovlev, H.J. Heeres, Applied Catalysis B: Environmental 117–118 (2012) 105–117. [14] J. Payormhorm, K. Kangvansaichol, P. Reubroycharoen, P. Kuchonthara, N. Hinchiranan, Bioresource Technology 139 (2013) 128–135. [15] K. Leiva, N. Martinez, C. Sepulveda, R. García, C.A. Jiménez, D. Laurenti, M. Vrinat, C. Geantet, J.L.G. Fierro, I.T. Ghampson, N. Escalona, Applied Catalysis A: General 490 (2015) 71–79. [16] B. Güvenatam, O. Kurşun, E.H.J. Heeres, E.A. Pidko, E.J.M. Hensen, Catalysis Today 233 (2014) 83–91. [17] Y.-K. Hong, D.-W. Lee, H.-J. Eom, K.-Y. Lee, Applied Catalysis B: Environmental 150–151 (2014) 438–445. [18] B.M.Q. Phan, Q.L.M. Ha, N.P. Le, P.T. Ngo, T.H. Nguyen, T.T. Dang, L.H. Nguyen, D.A. Nguyen, L.C. Luu, Catal Lett 145 (2014) 662–667.
18
[19] C. Newman, X. Zhou, B. Goundie, I.T. Ghampson, R.A. Pollock, Z. Ross, M.C. Wheeler, R.W. Meulenberg, R.N. Austin, B.G. Frederick, Applied Catalysis A: General 477 (2014) 64–74. [20] J.-S. Moon, E.-G. Kim, Y.-K. Lee, Journal of Catalysis 311 (2014) 144–152. [21] M. Hellinger, H.W.P. Carvalho, S. Baier, D. Wang, W. Kleist, J.-D. Grunwaldt, Applied Catalysis A: General 490 (2015) 181–192. [22] M.S. Zanuttini, C.D. Lago, C.A. Querini, M.A. Peralta, Catalysis Today 213 (2013) 9– 17. [23] X. Zhu, L.L. Lobban, R.G. Mallinson, D.E. Resasco, Journal of Catalysis 281 (2011) 21–29. [24] D. Gao, C. Schweitzer, H.T. Hwang, A. Varma, Ind. Eng. Chem. Res. 53 (2014) 18658– 18667. [25] Y. Hong, H. Zhang, J. Sun, K.M. Ayman, A.J.R. Hensley, M. Gu, M.H. Engelhard, J.-S. McEwen, Y. Wang, ACS Catal. 4 (2014) 3335–3345. [26] R. Olcese, M.M. Bettahar, B. Malaman, J. Ghanbaja, L. Tibavizco, D. Petitjean, A. Dufour, Applied Catalysis B: Environmental 129 (2013) 528–538. [27] C. Chen, G. Chen, F. Yang, H. Wang, J. Han, Q. Ge, X. Zhu, Chemical Engineering Science 135 (2015) 145–154. [28] H.Y. Zhao, D. Li, P. Bui, S.T. Oyama, Applied Catalysis A: General 391 (2011) 305– 310. [29] Y. Wang, J. Wu, S. Wang, RSC Adv. 3 (2013) 12635–12640. [30] W.-W. Liu, C.-W. Hu, Y. Yang, D.-M. Tong, L.-F. Zhu, R.-N. Zhang, B.-H. Zhao, Applied Catalysis B: Environmental 129 (2013) 202–213. [31] N. Cuello, V. Elías, M. Crivello, M. Oliva, G. Eimer, Journal of Solid State Chemistry 205 (2013) 91–96.
19
[32] C. Wu, L. Wang, P.T. Williams, J. Shi, J. Huang, Applied Catalysis B: Environmental 108–109 (2011) 6–13. [33] T.M. Sankaranarayanan, A. Berenguer, C. Ochoa-Hernández, I. Moreno, P. Jana, J.M. Coronado, D.P. Serrano, P. Pizarro, Catalysis Today 243 (2015) 163–172. [34] O. González, H. Pérez, P. Navarro, L.C. Almeida, J.G. Pacheco, M. Montes, Catalysis Today 148 (2009) 140–147. [35] S. Wang, Q. Yin, J. Guo, B. Ru, L. Zhu, Fuel 108 (2013) 597–603. [36] M. Zhao, T.L. Church, A.T. Harris, Applied Catalysis B: Environmental 101 (2011) 522–530. [37] L. He, H. Berntsen, E. Ochoa-Fernández, J.C. Walmsley, E.A. Blekkan, D. Chen, Top Catal 52 (2009) 206–217. [38] J. Xiang, X. Wen, F. Zhang, Ind. Eng. Chem. Res. 53 (2014) 15600–15610. [39] X. Chen, S.A.C. Carabineiro, S.S.T. Bastos, P.B. Tavares, J.J.M. Órfão, M.F.R. Pereira, J.L. Figueiredo, Journal of Environmental Chemical Engineering 1 (2013) 795–804. [40] R. Wojcieszak, S. Monteverdi, M. Mercy, I. Nowak, M. Ziolek, M.M. Bettahar, Applied Catalysis A: General 268 (2004) 241–253. [41] S. Sengupta, G. Deo, Journal of CO2 Utilization 10 (2015) 67–77. [42] L. Wang, D. Li, M. Koike, H. Watanabe, Y. Xu, Y. Nakagawa, K. Tomishige, Fuel 112 (2013) 654–661. [43] S. Sitthisa, D.E. Resasco, Catal Lett 141 (2011) 784–791. [44] R.C. Runnebaum, T. Nimmanwudipong, R.R. Limbo, D.E. Block, B.C. Gates, Catal Lett 142 (2012) 7–15. [45] S. Boonyasuwat, T. Omotoso, D.E. Resasco, S.P. Crossley, Catal Lett 143 (2013) 783– 791.
20
[46] T. Nimmanwudipong, C. Aydin, J. Lu, R.C. Runnebaum, K.C. Brodwater, N.D. Browning, D.E. Block, B.C. Gates, Catal Lett 142 (2012) 1190–1196. [47] Y.-R. Luo, Comprehensive Handbook of Chemical Bond Energies, CRC Press, 2007. [48] S. Echeandia, B. Pawelec, V.L. Barrio, P.L. Arias, J.F. Cambra, C.V. Loricera, J.L.G. Fierro, Fuel 117 (2014) 1061–1073. [49] Y. Wang, Y. Fang, T. He, H. Hu, J. Wu, Catalysis Communications 12 (2011) 1201– 1205.
21
Fig. 1. XRD patterns of support and metal modified catalysts.
23
Fig. 2. TEM images of (A) fresh 10Ni/Al–MCM–41, (B) fresh 10Co/Al–MCM–41, (C) fresh 5Ni5Co/Al–MCM–41, and (D) used 10Co/Al–MCM–41.
24
Fig. 3. Temperature programmed reduction (TPR-H2) profile of different metal modified AlMCM41 catalysts.
25
Fig. 4. Conversion and product yields of guaiacol HDO on Ni and Co supported Al–MCM– 41 catalysts. Reaction condition: T = 400 oC, P = 1 atm, H2/guaiacol molar ratio = 25, W/F = 1.67 h, TOS = 30 min.
26
Fig. 5. Conversion and product yields of guaiacol HDO on fresh 10Co/Al–MCM–41. (A) Guaiacol conversion, (B) Oxygen products, (C) Oxygen-free products. Reaction condition: T = 400 oC, P = 1 atm, H2/guaiacol molar ratio = 25, TOS = 30 min.
27
Fig. 6. Proposed reaction network for HDO of guaiacol on Ni and Co supported Al–MCM–41 catalysts. HDO reactions catalyzed by metal function are presented by red solid arrows; transalkylation reactions catalyzed by acidic function are presented by blue dash arrows.
28
Fig. 7. Conversion and product yields of guaiacol HDO on fresh 10Co/Al–MCM–41 catalyst. (A) Guaiacol conversion, (B) Oxygen products, (C) Oxygen-free products.
Reaction
condition: T = 400 oC, P = 1 atm, H2/guaiacol molar ratio = 25, W/F = 1.67 h.
29
Fig. 8. Conversion and product yields of guaiacol HDO on regenerated 10Co/Al–MCM– 41. (A) Conversion, (B) Oxygen products, (C) Oxygen-free products. Reaction condition: T = 400 oC, P = 1 atm, H2/guaiacol molar ratio = 25, TOS = 30 min.
Fig. 9. TEM image of regenerated 10Co/Al–MCM–41 catalyst.
30
Fig. 10. DTG curve of fresh 10Co/Al–MCM–41 and used catalyst after two HDO cycles under air atmosphere.
31
Table 1. Textural properties of support and metal modified catalysts. SBET
Pore sizea
Pore volumea
Particle sizeb
(m2/g)
(nm)
(cm3/g)
(nm)
AlMCM41
742
3.21
0.56
10Ni/AlMCM41
605
3.38
0.45
15.70
10Co/AlMCM41
583
3.33
0.43
14.09
5Ni5Co/AlMCM41
599
3.32
0.46
12.54
Samples
a
Calculated by BJH adsorption theory.
b
Calculated from the XRD data using the Scherrer equation at 2 = 36.8o.
32
S0926-860X(15)30294-5 http://dx.doi.org/doi:10.1016/j.apcata.2015.12.021 APCATA 15694
To appear in:
Applied Catalysis A: General
Received date: Revised date: Accepted date:
3-9-2015 8-12-2015 18-12-2015
Please cite this article as: Nga T.T.Tran, Yoshimitsu Uemura, Sujan Chowdhury, Anita Ramli, Vapor-phase Hydrodeoxygenation of Guaiacol on AlminusMCMminus41 Supported Ni and Co Catalysts, Applied Catalysis A, General http://dx.doi.org/10.1016/j.apcata.2015.12.021 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Vapor-phase Hydrodeoxygenation of Guaiacol on AlMCM41 Supported Ni and Co Catalysts Nga T.T. Tran1, 4*, Yoshimitsu Uemura1*, Sujan Chowdhury2, Anita Ramli3 1
Centre for Biofuel and Biochemical Research, Universiti Teknologi PETRONAS, 32610 Bandar Seri Iskandar, Perak Darul Ridzuan, Malaysia
2
Department of Chemical Engineering, Universiti Teknologi PETRONAS, 32610 Bandar Seri Iskandar, Perak Darul Ridzuan, Malaysia
3
Department of Fundamental & Applied Science, Universiti Teknologi PETRONAS, 32610 Bandar Seri Iskandar, Perak Darul Ridzuan, Malaysia
4
Faculty of Chemical Engineering, HCMC University of Technology, Ho Chi Minh, Vietnam
*
Yoshimitsu Uemura
Universiti Teknologi PETRONAS, 32610 Bandar Seri Iskandar, Perak Darul Ridzuan, Malaysia, Tel: +605-368-7644. E-mail: [email protected] *
Nga T.T. Tran
Tel: +60 11 18593681. E-mail: [email protected]
1
Graphical abstract:
2
Highlights:
HDO of guaiacol on AlMCM41 supported Ni and Co catalysts at 400 oC and 1 atm.
Co is active in HDO via CO hydrogenolysis, while Ni favors multiple CC hydrogenolysis.
Acid sites catalyze demethylation and transalkylation (methyl transfer).
Coke deposition has stronger deactivate effect in HDO than transalkylation.
3
Abstract Vapor-phase hydrodeoxygenation (HDO) of guaiacol, a typical lignin-derived phenolic compound, was studied on AlMCM41 supported Ni and Co catalysts at 400 oC and atmospheric pressure. Ni was found as an active metal for ring opening activity while Co favored the deoxygenation activity. Besides benzene and phenol were observed as the major deoxygenation products, the existence of toluene, cresol and other methylated C712 products was the result of methyl transfer reaction over the acidic function of support. Co/AlMCM41 catalyzed not only HDO to remove oxygen but also transalkylation to prevent the carbon loss via methanization. In addition, coke deposit mainly gathered on the surrounding area of cobalt particles, leading to the faster deactivation in HDO activity than transalykation. Keywords: Hydrodeoxygenation, Cobalt, Nickel, AlMCM41, Guaiacol.
4
1 INTRODUCTION Fast pyrolysis is an attractive thermal conversion process to generate the alternative liquid fuels (bio-oil) [1,2]. Bio-oil obtained from fast pyrolysis contains a large fraction of ligninderived phenolic compounds [1,3–5]. These compounds can be upgraded to renewable fuel source such as aromatic or naphthenic hydrocarbon by hydrodeoxygenation (HDO) [6]. This is a prominent process because it not only removes oxygen, but also preserves the carbon number in the product [7,8]. The final products can be utilized as petrochemical feedstock or high octane gasoline base materials. For example, C68 aromatics are valuable petrochemicals for use as solvents or intermediate in polyester fiber and film manufacture [9], and C9-12 aromatics are the major blending component (1220 % by volume) in gasoline engine fuels to obtain high octane number because there has been a limitation of low vapor pressure aromatics content (e. g., benzene and toluene) [10]. The HDO of pyrolysis oil and model phenolic compounds were mostly conducted at high hydrogen pressure in a batch-wise autoclave (40200 bar) or in a continuous-flow fixed-bed reactor (1040 bar) [3,11–15]. The high hydrogen pressure was required to give high deoxygenation degree but it led to saturate the aromatic ring before eliminating the oxygen groups [13,16–21]. Besides resulting a large consumption of costly hydrogen, this condition also produced ring saturation products with lower octane number and less valuable for cofeeding in existing oil refineries [22,23]. To date, a few studies have been investigated the HDO at low hydrogen consumption (atmospheric pressure) and found that this process had high selectivity to the deoxygenated aromatics [7,8,12,23]. A study by Zhu et al. [23] pointed out the role of metal function and acidic function in HDO of anisole over Pt/SiO2 and Pt/HBeta to form BTX products. In order to have a high selectivity to toluene and high stability of catalyst, Pt/Al2O3 and Pt/SiO2 showed a better catalytic performance than Pt/HB for HDO of m-cresol [12]. However,
5
noble metal catalysts still showed their strong activity in hydrogenation for ring-saturation and ring-opening instead of deoxygenation activity, even though the reaction was conducted at ambient pressure [6,7,20,24]. In addition, the high cost of noble metal is also a disadvantage in their application. Therefore, the non-noble metal may become a solution to these issues. For example, the addition of Fe to Pd or Ni catalyst directed the selection route to oxygen-free aromatic instead of ring-saturated products [7,8]. Moreover, HDO of m-cresol over Pd/Fe2O3 catalyst showed high BTX selectivity (>80%) and less ring saturation product (<1%) that is resulted by the synergic between Pd and Fe [25]. Nevertheless, the iron-based catalysts had lower HDO activity than the precious metal, and mainly formed phenol as a major product [7,26]. In addition, Ni was significantly less active in deoxygenation activity comparing with Pd and Pt catalyst due to its activities in hydrogenation or hydrogenolysis of CC bond [27]. The HDO of guaiacol over transition metal phosphides changed the major product selectivity from phenol to benzene when increasing the contact time (W/F) about 60 times [20,28]. In view of catalyst deactivation during HDO process, coke formation through the deposition of polyaromatics was found to be the main cause instead of sintering [24]. Moreover, the acidic support was deactivated faster than the inert support due to its strong interaction with the coke precursors [12,23,26]. Besides, a zeolite with mesoporous structure showed a much higher stability than the microporous one in the HDO of phenolic compounds [22,23,29]. In this study, we investigated the activity of AlMCM–41 supported non-noble metal (Co and Ni) catalysts for HDO of guaiacol at 400 oC and ambient pressure. The mesoporous AlMCM–41 was selected as the catalyst support since it has good internal mass-transport, high stability and transalkylation activity [23,29]. Guaiacol was chosen as a model compound because it contains both major functional groups of lignin-derived phenolic such as hydroxyl
6
(OH) and methoxy (OCH3). The role of metal and support on catalytic HDO of guaiacol were clarified. Catalyst deactivation and regeneration behavior were also reported. 1. EXPERIMENTAL 1.1. Catalyst preparation Aluminosilicate Al–MCM–41 support (3–4% Al2O3) supplied by ACS Material was activated in air at 500 oC for 14 h. The catalysts containing 10 wt% metal were prepared by the wetness impregnation method. Aqueous solutions containing metal precursor (i.e. nickel nitrate and cobalt nitrate) were first impregnated on the support overnight and then dried at 105 oC for 8 h, followed by calcination at 500 oC for 14 h in air. 1.2. Catalyst characterization The TriStar II 3020 was used to measure the nitrogen adsorption/desorption isotherms at 77K. Transmission electron microscopy (TEM) image was obtained on a Carl Zeiss LIBRA 200FE (operation voltage 200 kV). The catalyst powder was dispersed ultrasonically in 2-propanol for 1 h and then deposited on a carbon coated copper grid for TEM measurement. X-ray diffraction (XRD) pattern was measured on a LabX XRD-6000 X-ray diffractometer (Shimadzu) using Cu K radiation. XRD pattern was recorded by scanning with 0.02o increments in the 270o diffraction angle range. Temperature programmed reduction (TPR) in hydrogen was carried out using a Thermo Scientific TPDRO 1100 instrument. Samples were pretreated in nitrogen at 120 oC during 60 min. Then, a TPR program was performed with a heating rate of 10 oC/min to 900 oC under 5% H2/N2 flow. The amount of hydrogen consumption during catalyst reduction process was monitored by a thermal conductivity detector (TCD).
7
Thermogravimetric analysis (TGA) under the flow of air was conducted in a TA Instrument model QA50. During the analysis, temperature was increased from room temperature to 900 oC at heating rate of 10 oC/min. 1.3. Catalytic performance Catalytic reaction was conducted in a fixed-bed tubular reactor (internal diameter 13 mm) made of stainless steel under ambient pressure. This reactor has two different operating zones where their temperatures can be controlled independently. The top bed functions as a vaporization zone of the feedstock at 350 oC before reaching to the catalyst bed at the bottom zone. The catalyst sample (3560 mesh) was packed in the bottom bed between two layers of glass wool. The Ni catalyst was reduced at 500 oC for 1 h and the Co and NiCo catalysts were reduced at 450 oC for 1 h in H2 flow. The catalyst bed was then cooled to the desired reaction temperature with N2. Guaiacol was fed at a flow rate of 1.08 mL/h using a syringe pump and vaporized in the top glass wool bed. The H2/guaiacol molar ratio was 25 for all runs. The W/F (gcat/greactanth) was adjusted in the range of 01.67 h through varying the amount of catalyst (02.0 g). The reactor outlet line was heated at 195 oC to avoid the condensation of liquid products before they reach the cold trap. All catalytic runs lasted 210 min, and the gas and liquid samples were collected to analyze at every 30 min interval. The liquid products were quantified by gas chromatography (Shimadzu GC-2014), using a SGE BPX5 capillary column and a flame ionized detector (FID) while non-condensed gas was analyzed with GCTCD (Shimadzu GC8A). The product components were identified with GCMS (Agilent Technologies 6890A, BPX5 capillary column), and verified with organic standards. In all experiments, the carbon balance was between 91 and 97 %, depending on the coke deposit on catalyst and the condensation of heavy components inside reactor and piping. Carbon-based guaiacol conversion (XGua), product distribution (Di) and product yield (Yi) were calculated in MolCarbon% by the following equations [7,23].
8
X
Gua
(%)
D i (%)
Mol
gua in Mol
Mol Mol
i
reacted
Mol
gua in
i Gua
gua out
7
100
100
(1)
(2)
where i is the carbon number in product i. Y i (%)
X
Gua
Di
(3)
100
Catalyst regeneration was carried out after 3 h of guaiacol HDO reaction. An air flow was applied to treat the used catalyst at 450 oC for 2 h. Afterwards the catalyst was re-activated in H2 flow for 1 h, and catalyzed a new HDO reaction cycle. 2. RESULTS AND DISCUSSION 2.1. Catalyst characterization Table 1 demonstrates textural properties of metal-modified catalysts and support. After calcination, the color of nickel catalyst sample became gray while cobalt and bimetallic catalysts had a dark brown due to the formation of nickel oxides and cobalt oxides [30,31]. The metal-modified catalysts showed a slight decrease in surface area and pore volume as expected. Moreover, the pore diameters of three Al–MCM–41-supported catalysts also reduced comparing with support only (Error! Reference source not found.). It may be caused by the partial blockage of pores after metal impregnation. However, the mesoporosity of catalysts had not changed significantly. Fig. 1 shows the powder XRD patterns of Al–MCM–41 supported Ni, Co and NiCo catalysts. Both support and metal-modified catalysts have a broad peak at around 23o which is assigned to the amorphous silica in Al–MCM–41 support [32]. Since only the wide-angle XRD was applied, the (100), (110), (200) and (210) peaks assigned to hexagonal structure of Al–MCM–41 were not obviously detected in the range of 17o [30]. The Ni-modified catalyst have three peaks at 37o, 43o and 64o referring to NiO [32,33]. Meanwhile, the Co-modified
9
catalyst has five diffraction peaks at 31.3o, 36.8o, 44.9o, 59.3o, and 65.2o corresponding to different Co3O4 crystallites [31,34,35]. The five major diffraction peaks of NiCo bimetallic catalyst could be NiO, Co3O4 or their composites since these oxides have similar and overlapping XRD reflections [36–39]. The TEM images of three metal-modified catalysts are displayed in Fig. 2, which include the distribution of metal oxide particle size. The 10Ni/Al–MCM–41 had a narrow distribution from 2 to 9 nm particle size as in Fig. 2A. However, a small amount of the larger bulk NiO particles, of which the size was from 50 to 110 nm, was formed on the support. This phenomenon was also observed in nickel catalysts by Wu et al. [32] and Sankaranarayanan et al. [33]. In contrast, cobalt catalyst shows a broad distribution of metal oxide particle size (Fig. 2B). Although most of the cobalt oxide particles were in a size range between 314 nm, the larger particles from 1525 nm were also observed. In contrast to monometallic catalyst, the NiCo bimetallic has a better distribution of particle size which was from 2 to 14 nm and no observation of larger particles. Meanwhile, the estimated particle sizes via XRD in Table 1 were within the range of size distribution via TEM observation. Among three catalysts, bimetallic NiCo had the smallest particle size, and this result was also compatible with TEM measurement. Since the average pore size of Al–MCM–41 was around 3.2 nm, a large amount of metal oxide particles would be located on the external surface of the support. Hence, the porous structure of catalyst did not extensively change after supporting metal. The TPR-H2 profiles of the three metal-supported catalysts are shown in Fig. 3. The reduction of 10Ni/Al–MCM–41 catalyst has two broad peaks observed at 440 oC and 610 oC which is consistent with previous studies [30,40]. The first peak was the reduction of NiO species to the metallic form of nickel [40]. While the second one could be corresponded to the reduction of a cationic form of nickel (NiSiO3) [6,40] and the small NiO particles inside the pore of support [32]. The cobalt catalyst has two reduction zones that could be defined as:
10
one at low-temperature (370 oC) and one at high-temperature (710 oC) which are associated with the reduction of cobalt oxide species with different dispersions and interactions with the support. The doublet peak at lower temperature may be attributed to the reduction of Co3O4 to CoO, and CoO to cobalt metal [6,31,34]. The broad peak at higher temperature could be ascribed to the strong interaction of small cobalt oxide particles with the surface [35]. The reducibility of cobalt species was higher than that of Ni species in their monometallic catalyst. Moreover, the reduction profile of NiCo bimetallic also has two distinct zones. The triplet peak at low temperature could be the simultaneous reduction of cobalt oxides and nickel oxides. The broad peak at higher temperature, which related to the reduction of strong interaction metal oxides with support, was shifted to higher temperature than that of monometallic. In summary, the presence of cobalt improved the reduction behavior of nickel species since it decreased not only the temperature of first reduction peak but also the amount of reduction metal oxides in the second peak. This trend is consistent with the ones already reported elsewhere [41,42]. 2.2. Hydrodeoxygenation of guaiacol The supported metal catalysts were evaluated for HDO of guaiacol at 400 oC under ambient pressure. The hydrotreatments of guaiacol with blank and support-packed reactors were used as references to verify the effect of catalysts and summarized in Error! Reference source not found.. The reactions observed in the blank test were supposed to be homogeneous and catalytic reactions on the internal wall of stainless steel reactor. The conversion of guaiacol over pure AlMCM–41 support predominantly led to the demethylation of guaiacol and formed final products, i.e., catechol and methane. Moreover, methyl-guaiacol was observed as a secondary liquid product formed via transalkylation catalyzed by the acidic support [23].
11
Fig. 4 compares the conversion of guaiacol and product yields over the supported Ni, Co, and NiCo catalysts at the same conditions. Among the three catalysts, 10Co/Al–MCM–41 exhibited the highest HDO activity due to its highest deoxygenation products yield and lowest gas phase yield (mainly C1 products including CH4, CO and CO2). Meanwhile, Nimodified catalyst showed lower activity for HDO of guaiacol since it favored the methanization reaction to form methane as the major product. Even though both cobalt and nickel are 3d transition metals, nickel was found as an active metal for multiple hydrogenolysis of CC bonds (resulted in ring opening), leading to the formation of methane, especially at high temperature [27,43]. The bimetallic 5Ni5Co/Al–MCM–41 catalyst showed substantially higher HDO yield than the average HDO yield of monometallic 10Ni and 10Co catalysts. It proved that cobalt promoted nickel in bimetallic catalyst to enhance the HDO activity. In terms of deoxygenation yield, 10Co/Al–MCM–41 showed to be more attractive catalyst so that this catalyst was chosen for further investigations. The presence of C712 aromatic and phenolic compounds such as toluene, xylene, pentamethylbenzene, hexamethylbenzene, cresol, xylenol, methyl-guaiacol, etc. was resulted of the acid-catalyzed transalkylation of methoxy group to another phenolic ring [23,44–46]. In HDO of guaiacol, especially at high temperature, the CArOCH3 cleavage and the multiple hydrogenolysis of CC bonds led to formation of methane in gas phase. Therefore, the methyl transfer reaction occurred on acid sites of support helped to minimize the carbon loss via methanization. In addition, the observation of some C8+ aromatic and phenolic indicates that the surface methyl groups produced by demethylation were transferred not only within their own molecules but also to other aromatic rings. Fig. 5 illustrates the evolution of guaiacol conversion and product yields over a freshly prepared 10Co/AlMCM41 as a function of W/F. The guaiacol conversion, oxygenated product yields, and non-oxygenated product yields are shown in Fig. 5A, B and C,
12
respectively. Complete conversion of guaiacol was achieved at W/F of 1.67 h. At lower W/F, the major products were phenol, cresol and methane, indicating that hydrogenolysis of CArOCH3 bond and transalkylation were the primary reaction. Since the bond dissociation energy (BDE) of CArOCH3 bond is 45 kJ/mol lower than that of CArOH bond [47], the removal of hydroxyl group considerably occurred at higher W/F, resulting in higher yield of oxygen-free aromatic products. The yield of C8-12 aromatics resulted by multi methyl transfer reactions increased more slowly than those of benzene and toluene when the W/F increased. However, high W/F promoted not only the hydrogenolysis of CO bond in guaiacol molecule but also the multiple hydrogenolysis of CC bonds, leading to an increase in gasification activity. When the W/F ratio was increased to 1.67 h, CO and CO2 started to be obviously detected in the gas phase, and the molar ratio of methane to aromatic ring in liquid phase became higher than the stoichiometry of demethylation reaction. It proved that the gasification occurred significantly at W/F = 1.67 h, yielding high amount of CH4, CO and CO2 in the gas phase. During catalytic hydrotreatment processes, multiple reactions may occur, including hydrogenation, hydrogenolysis, hydrodeoxygenation, hydrocracking, and polymerization [1,3]. Hydrogenation for ring saturation and hydrocracking for gasification make hydrogen consumption exceed the deoxygenated stoichiometric ratio [3]. In order to reduce the hydrogen consumption, the direct deoxygenation without ring saturation was expected in catalytic HDO. The HDO conducted in high pressure continuous reactor showed not only deoxygenation but also ring saturation or ring coupling activities [13,17,33,48,49]. Even at ambient pressure, the low temperature (<300 oC) was conducive to ring saturation [6–8]. In contrast, our experimental runs were carried out at high temperature (400 oC) and low pressure (1 atm) to prevent the hydrogenation/dehydrogenation for ring saturation [7,12,23].
13
Hence, no aromatic saturation product was observed in our study, and it led to lower hydrogen consumption in HDO of guaiacol. Fig. 6 shows the proposed reaction pathways for guaiacol HDO over AlMCM41 supported Co or Ni catalysts based on the changes of products as the functions of W/F and catalyst type. The overall HDO reaction of guaiacol is complex due to the synergic effect of acidic and metallic sites in catalyst, resulting in various intermediates and products as well as many elementary reactions. Considering that the BDE of CArOH, CArOCH3, CArOCH3 are ranked in descending order from 464 kJ/mol to 263 kJ/mol [47], the primary product was changed from catechol in case of pure support to phenol and cresol in case of metal modified catalyst, and further to oxygen-free aromatics at higher W/F ratio. That is, the acidic sites of support only catalyzed the demethylation in C ArOCH3 bond and methyl transfer reaction, while the deoxygenation reaction was catalyzed by metallic sites which facilitated the hydrogenolysis of CArOR bond. Besides some similar activities with Co-modified catalyst such as hydrodeoxygenation, methylation, and transalykation, Ni-modified catalyst principally favored the multiple CC hydrogenolysis to form methane as the final product. In summary, the bifunctional catalyst can convert guaiacol not only to benzene as completely deoxygenated product but also to other methylated aromatics as transalkylated products. 2.3. Coke deposition and catalyst regeneration As shown in Fig. 7, the conversion of guaiacol and the oxygen-free product yields decreased with time on stream (TOS) due to deactivation of catalyst. Considerable deactivation occurred since the conversion of guaiacol decreased from 99.5% to 85.8% after reaction time of 210 min. In earlier state, the yield of oxygen-free aromatic products was dropped rapidly while the monooxygenated phenolics yield increased quickly. After reaching the maximum, selectivity to phenol and methylated phenol products decreased slowly, leading to an increment in the concentration of deoxygenated products at longer time on
14
stream. Hence, the reduction of active metal sites gave more strongly effect to the breaking of CArOH bond compared to that of CArOCH3 bond. Fig. 2D is a representative TEM image of the spent catalyst after two HDO reaction cycles. Coke deposit could be observed as the indeterminate-shaped objects having a larger size and lighter color than the spherical cobalt particles. This TEM observation shows that more carbon deposits were gathered on the surrounding area of cobalt particles than on support, resulting in blockage of reactants approach to metallic surface sites. It means that the coke formations during HDO reaction were not only affected by the acid sites of support but also considerably by the metal sites. Similarly, Zhu et al. [23] concluded that the coke deposition was dependent on the presence of Pt metal, the acidity and pore structure of support. When using inert silica as the support, Olcese et al. [26] found that carbon deposits were mainly observed in the vicinity of iron metal instead of bare silica support due to its very low acidity. The acid sites of support promoted the carbon deposition around metal sites hence the amount of coke formation over Pt/zeolite catalyst was significantly higher than over Pt/SiO2 catalyst [12,23]. In the present work, the smaller accumulation of coke on AlMCM41 support than on the surrounding area of cobalt led to a lower deactivation of transalkylation than hydrodeoxygenation, as evidenced by the stability of cresol and other methylated phenol products yield and the decrease of deoxygenation product yields with time on stream in Fig. 7. The catalysts were deactivated during HDO reactions due to carbon deposition; hence, regeneration ability becomes an important issue in practical application. Fig. 8 illustrates the details of HDO activities of the regenerated 10Co/AlMCM41 at different W/F. Comparing with the fresh 10Co/AlMCM41 catalyst (Fig. 5), the regenerated one showed insignificant difference in guaiacol conversion but the product distribution considerably changed. That is, yield of all aromatics showed a noticeable decrease, especially at high W/F. Moreover, the
15
higher W/F was used, the more decrease in HDO activity of regenerated catalyst was observed. This result indicated that the thickness of catalytic bed might affect to the regeneration efficiency. Fig. 9 shows a TEM image of the regenerated 10Co/AlMCM41 catalyst. Particle size distribution was extracted from measurement of more than 100 particles to avoid the population variance effect. According to the results, particle size of the regenerated catalyst ranged from 2 to 17 nm, which was narrower than that of the fresh catalyst. In addition, the multi-spot EDX scanning which carried out in SEM analysis (data not shown) shown no significant difference in cobalt content between fresh and regenerated catalysts. Hence, the thermal degradation or sintering was not the primary reason of catalyst deactivation. Moreover, the TEM results of regenerated 10Co/AlMCM41 also shown a small amount of coke deposit still remained in catalyst. According to Zanuttini’s research, the coke formation after HDO of m-cresol over Pt/Al2O3 has two oxidation zones at moderate and high temperatures (250 and 550 oC) corresponding to carbon deposited on the metal and on the support, respectively [12]. Meanwhile, the derivative thermogravimetric (DTG) of used 10Co/AlMCM41 catalyst showed a broad peak from 200 to 650 oC which associated with combustion of coke (Fig. 10). It means that the regeneration at 450 oC could not remove all the coke deposited on used catalyst and some active sites were not recovered. These reasons can explain the decrease in HDO activity of the regenerated catalyst. 3. CONCLUSION Hydrodeoxygenation of guaiacol over Al–MCM–41 supported Ni and Co catalysts has been studied at 400 oC and under atmospheric pressure. The Ni catalyst favored the multiple hydrogenolysis of CC bond, leading to carbon loss via methanization. Cobalt not only facilitated the reduction and dispersion of nickel metal oxides but also enhanced the HDO activity of bimetallic NiCo catalyst. Results showed that Co/Al–MCM–41 catalyst was
16
conducive to HDO of guaiacol reaction by a combination of acidic and metallic functions. The acid sites existing on Al–MCM–41 catalyzed the demethylation and transalkylation reactions while metallic cobalt catalyzed the hydrogenolysis CArOR bond to produce oxygen-free products. During HDO reaction, CArOCH3 cleavage occurred preferentially while more active sites were required to catalyze the CArOH cleavage. Deactivation of catalyst affected more strongly the CArOH hydrogenolysis reaction than CArOCH3 hydrogenolysis and transalkylation reaction, resulting in the decrease of oxygen-free aromatic yield.
Acknowledgment The authors gratefully acknowledge the financial support from Mitsubishi Corporation Educational Trust Fund. We wish to thank Universiti Teknologi PETRONAS for providing a congenial work environment and state-of-the-art research facilities. We also thank Prof. Dr. Taufiq Yap Yun Hin for his support in XRD measurement. The authors are also grateful to the anonymous reviewers for their valuable comments on this paper. References [1] J. Wildschut, I. Melián-Cabrera, H.J. Heeres, Applied Catalysis B: Environmental 99 (2010) 298–306. [2] S.R. Naqvi, Y. Uemura, S.B. Yusup, Journal of Analytical and Applied Pyrolysis 106 (2014) 57–62. [3] P.M. Mortensen, J.-D. Grunwaldt, P.A. Jensen, K.G. Knudsen, A.D. Jensen, Applied Catalysis A: General 407 (2011) 1–19. [4] H. Zhang, R. Xiao, H. Huang, G. Xiao, Bioresource Technology 100 (2009) 1428–1434. [5] C.S. Lira, F.M. Berruti, P. Palmisano, F. Berruti, C. Briens, A.A.B. Pécora, Journal of Analytical and Applied Pyrolysis 99 (2013) 23–31. 17
[6] P.T.M. Do, A.J. Foster, J. Chen, R.F. Lobo, Green Chem. 14 (2012) 1388–1397. [7] J. Sun, A.M. Karim, H. Zhang, L. Kovarik, X.S. Li, A.J. Hensley, J.-S. McEwen, Y. Wang, Journal of Catalysis 306 (2013) 47–57. [8] L. Nie, P.M. de Souza, F.B. Noronha, W. An, T. Sooknoi, D.E. Resasco, Journal of Molecular Catalysis A: Chemical 388–389 (2014) 47–55. [9] O.A. Al-Rashed, M.A. Fahim, M. Shaaban, Fluid Phase Equilibria 363 (2014) 248–262. [10] H. Fukuya, N. Okabe, S. Sasaki, Fuel Composition for Use in Gasoline Engines, US8876920 B2, 2014. [11] N. Tran, Y. Uemura, S. Chowdhury, A. Ramli, Applied Mechanics and Materials 625 (2014) 255–258. [12] M.S. Zanuttini, B.O. Dalla Costa, C.A. Querini, M.A. Peralta, Applied Catalysis A: General 482 (2014) 352–361. [13] A.R. Ardiyanti, S.A. Khromova, R.H. Venderbosch, V.A. Yakovlev, H.J. Heeres, Applied Catalysis B: Environmental 117–118 (2012) 105–117. [14] J. Payormhorm, K. Kangvansaichol, P. Reubroycharoen, P. Kuchonthara, N. Hinchiranan, Bioresource Technology 139 (2013) 128–135. [15] K. Leiva, N. Martinez, C. Sepulveda, R. García, C.A. Jiménez, D. Laurenti, M. Vrinat, C. Geantet, J.L.G. Fierro, I.T. Ghampson, N. Escalona, Applied Catalysis A: General 490 (2015) 71–79. [16] B. Güvenatam, O. Kurşun, E.H.J. Heeres, E.A. Pidko, E.J.M. Hensen, Catalysis Today 233 (2014) 83–91. [17] Y.-K. Hong, D.-W. Lee, H.-J. Eom, K.-Y. Lee, Applied Catalysis B: Environmental 150–151 (2014) 438–445. [18] B.M.Q. Phan, Q.L.M. Ha, N.P. Le, P.T. Ngo, T.H. Nguyen, T.T. Dang, L.H. Nguyen, D.A. Nguyen, L.C. Luu, Catal Lett 145 (2014) 662–667.
18
[19] C. Newman, X. Zhou, B. Goundie, I.T. Ghampson, R.A. Pollock, Z. Ross, M.C. Wheeler, R.W. Meulenberg, R.N. Austin, B.G. Frederick, Applied Catalysis A: General 477 (2014) 64–74. [20] J.-S. Moon, E.-G. Kim, Y.-K. Lee, Journal of Catalysis 311 (2014) 144–152. [21] M. Hellinger, H.W.P. Carvalho, S. Baier, D. Wang, W. Kleist, J.-D. Grunwaldt, Applied Catalysis A: General 490 (2015) 181–192. [22] M.S. Zanuttini, C.D. Lago, C.A. Querini, M.A. Peralta, Catalysis Today 213 (2013) 9– 17. [23] X. Zhu, L.L. Lobban, R.G. Mallinson, D.E. Resasco, Journal of Catalysis 281 (2011) 21–29. [24] D. Gao, C. Schweitzer, H.T. Hwang, A. Varma, Ind. Eng. Chem. Res. 53 (2014) 18658– 18667. [25] Y. Hong, H. Zhang, J. Sun, K.M. Ayman, A.J.R. Hensley, M. Gu, M.H. Engelhard, J.-S. McEwen, Y. Wang, ACS Catal. 4 (2014) 3335–3345. [26] R. Olcese, M.M. Bettahar, B. Malaman, J. Ghanbaja, L. Tibavizco, D. Petitjean, A. Dufour, Applied Catalysis B: Environmental 129 (2013) 528–538. [27] C. Chen, G. Chen, F. Yang, H. Wang, J. Han, Q. Ge, X. Zhu, Chemical Engineering Science 135 (2015) 145–154. [28] H.Y. Zhao, D. Li, P. Bui, S.T. Oyama, Applied Catalysis A: General 391 (2011) 305– 310. [29] Y. Wang, J. Wu, S. Wang, RSC Adv. 3 (2013) 12635–12640. [30] W.-W. Liu, C.-W. Hu, Y. Yang, D.-M. Tong, L.-F. Zhu, R.-N. Zhang, B.-H. Zhao, Applied Catalysis B: Environmental 129 (2013) 202–213. [31] N. Cuello, V. Elías, M. Crivello, M. Oliva, G. Eimer, Journal of Solid State Chemistry 205 (2013) 91–96.
19
[32] C. Wu, L. Wang, P.T. Williams, J. Shi, J. Huang, Applied Catalysis B: Environmental 108–109 (2011) 6–13. [33] T.M. Sankaranarayanan, A. Berenguer, C. Ochoa-Hernández, I. Moreno, P. Jana, J.M. Coronado, D.P. Serrano, P. Pizarro, Catalysis Today 243 (2015) 163–172. [34] O. González, H. Pérez, P. Navarro, L.C. Almeida, J.G. Pacheco, M. Montes, Catalysis Today 148 (2009) 140–147. [35] S. Wang, Q. Yin, J. Guo, B. Ru, L. Zhu, Fuel 108 (2013) 597–603. [36] M. Zhao, T.L. Church, A.T. Harris, Applied Catalysis B: Environmental 101 (2011) 522–530. [37] L. He, H. Berntsen, E. Ochoa-Fernández, J.C. Walmsley, E.A. Blekkan, D. Chen, Top Catal 52 (2009) 206–217. [38] J. Xiang, X. Wen, F. Zhang, Ind. Eng. Chem. Res. 53 (2014) 15600–15610. [39] X. Chen, S.A.C. Carabineiro, S.S.T. Bastos, P.B. Tavares, J.J.M. Órfão, M.F.R. Pereira, J.L. Figueiredo, Journal of Environmental Chemical Engineering 1 (2013) 795–804. [40] R. Wojcieszak, S. Monteverdi, M. Mercy, I. Nowak, M. Ziolek, M.M. Bettahar, Applied Catalysis A: General 268 (2004) 241–253. [41] S. Sengupta, G. Deo, Journal of CO2 Utilization 10 (2015) 67–77. [42] L. Wang, D. Li, M. Koike, H. Watanabe, Y. Xu, Y. Nakagawa, K. Tomishige, Fuel 112 (2013) 654–661. [43] S. Sitthisa, D.E. Resasco, Catal Lett 141 (2011) 784–791. [44] R.C. Runnebaum, T. Nimmanwudipong, R.R. Limbo, D.E. Block, B.C. Gates, Catal Lett 142 (2012) 7–15. [45] S. Boonyasuwat, T. Omotoso, D.E. Resasco, S.P. Crossley, Catal Lett 143 (2013) 783– 791.
20
[46] T. Nimmanwudipong, C. Aydin, J. Lu, R.C. Runnebaum, K.C. Brodwater, N.D. Browning, D.E. Block, B.C. Gates, Catal Lett 142 (2012) 1190–1196. [47] Y.-R. Luo, Comprehensive Handbook of Chemical Bond Energies, CRC Press, 2007. [48] S. Echeandia, B. Pawelec, V.L. Barrio, P.L. Arias, J.F. Cambra, C.V. Loricera, J.L.G. Fierro, Fuel 117 (2014) 1061–1073. [49] Y. Wang, Y. Fang, T. He, H. Hu, J. Wu, Catalysis Communications 12 (2011) 1201– 1205.
21
Fig. 1. XRD patterns of support and metal modified catalysts.
23
Fig. 2. TEM images of (A) fresh 10Ni/Al–MCM–41, (B) fresh 10Co/Al–MCM–41, (C) fresh 5Ni5Co/Al–MCM–41, and (D) used 10Co/Al–MCM–41.
24
Fig. 3. Temperature programmed reduction (TPR-H2) profile of different metal modified AlMCM41 catalysts.
25
Fig. 4. Conversion and product yields of guaiacol HDO on Ni and Co supported Al–MCM– 41 catalysts. Reaction condition: T = 400 oC, P = 1 atm, H2/guaiacol molar ratio = 25, W/F = 1.67 h, TOS = 30 min.
26
Fig. 5. Conversion and product yields of guaiacol HDO on fresh 10Co/Al–MCM–41. (A) Guaiacol conversion, (B) Oxygen products, (C) Oxygen-free products. Reaction condition: T = 400 oC, P = 1 atm, H2/guaiacol molar ratio = 25, TOS = 30 min.
27
Fig. 6. Proposed reaction network for HDO of guaiacol on Ni and Co supported Al–MCM–41 catalysts. HDO reactions catalyzed by metal function are presented by red solid arrows; transalkylation reactions catalyzed by acidic function are presented by blue dash arrows.
28
Fig. 7. Conversion and product yields of guaiacol HDO on fresh 10Co/Al–MCM–41 catalyst. (A) Guaiacol conversion, (B) Oxygen products, (C) Oxygen-free products.
Reaction
condition: T = 400 oC, P = 1 atm, H2/guaiacol molar ratio = 25, W/F = 1.67 h.
29
Fig. 8. Conversion and product yields of guaiacol HDO on regenerated 10Co/Al–MCM– 41. (A) Conversion, (B) Oxygen products, (C) Oxygen-free products. Reaction condition: T = 400 oC, P = 1 atm, H2/guaiacol molar ratio = 25, TOS = 30 min.
Fig. 9. TEM image of regenerated 10Co/Al–MCM–41 catalyst.
30
Fig. 10. DTG curve of fresh 10Co/Al–MCM–41 and used catalyst after two HDO cycles under air atmosphere.
31
Table 1. Textural properties of support and metal modified catalysts. SBET
Pore sizea
Pore volumea
Particle sizeb
(m2/g)
(nm)
(cm3/g)
(nm)
AlMCM41
742
3.21
0.56
10Ni/AlMCM41
605
3.38
0.45
15.70
10Co/AlMCM41
583
3.33
0.43
14.09
5Ni5Co/AlMCM41
599
3.32
0.46
12.54
Samples
a
Calculated by BJH adsorption theory.
b
Calculated from the XRD data using the Scherrer equation at 2 = 36.8o.
32