Hydrodeoxygenation of Anisole over Silica-Supported Ni2P, MoP, and

Jan 31, 2011 - Recent Advances in Hydrotreating of Pyrolysis Bio-Oil and Its Oxygen-Containing Model Compounds ... Renewable and Sustainable Energy Re...
0 downloads 14 Views 1MB Size
ARTICLE pubs.acs.org/EF

Hydrodeoxygenation of Anisole over Silica-Supported Ni2P, MoP, and NiMoP Catalysts Kelun Li, Rijie Wang, and Jixiang Chen* Tianjin Key Laboratory of Applied Catalysis Science and Technology, Department of Catalysis Science and Engineering, School of Chemical Engineering and Technology, Tianjin University, Tianjin 300072, People’s Republic of China ABSTRACT: Ni2P/SiO2, MoP/SiO2, and NiMoP/SiO2 with different Ni/Mo molar ratios were prepared by temperatureprogrammed reduction (TPR). Their structural properties were characterized by N2 sorption, X-ray diffraction (XRD), CO chemisorption, X-ray photoelectron spectroscopy (XPS), H2 temperature-programmed desorption (H2-TPD), and NH3 temperatureprogrammed desorption (NH3-TPD). Their performances for the hydrodeoxygenation (HDO) of anisole were tested in a fixed-bed reactor. It was found that there were mainly three reactions that occurred during the HDO, i.e., the demethylation of anisole, the hydrogenolysis of phenol, and the hydrogenation of benzene. The HDO activities decreased in the sequence of Ni2P/SiO2 > NiMoP/SiO2 > MoP/SiO2. The NiMoP/SiO2 catalysts with larger Ni/Mo ratios had higher activities. In the phosphides, the Niδþ and Moδþ sites bearing small positive charges acted not only as Lewis acid sites for the demethylation but also as metal sites for the hydrogenolysis and hydrogenation. The Niδþ site was more active than the Moδþ site, and there was no synergy between the Niδþ and Moδþ sites. The superior activity of Ni2P to that of MoP is attributed to the higher d electron density in Ni2P. PO-H groups, which acted as Brønsted sites and provided active hydrogen species, had less activity for the three reactions compared to the metal sites. In comparison to a conventional NiMo/γ-Al2O3 catalyst, the Ni phosphide-containing catalysts had much higher activities. The catalyst deactivation due to water was preliminarily discussed. The oxidation of phosphide by water might lead to the formation of metal oxide and/or phosphate, leading to the catalyst deactivation. The high stability of Ni2P/SiO2 may be related to the ligand effect of P that lowers the electron density of Ni and inhibits the Ni-O combination.

1. INTRODUCTION Nowadays, the fast development of the global economy is accelerating the consumption of energy resources, while the reserve of fossil fuels is dwindling rapidly. Also, with the increase of political and environmental concerns about fossil fuels, it is imperative to develop alternative energy resources. As a kind of renewable energy resource, biomass has drawn great attention.1 Through thermochemical methods, such as pyrolysis, gasification, and liquefaction, biomass can be converted into bio-oils.2 However, the crude bio-oils, especially produced by the fast pyrolysis, contain high oxygen contents (10-45 wt %).3,4 The high oxygen contents result in poor operational characteristics, including low pH, low heating value, high viscosity, thermal and chemical instability, and immiscibility with hydrocarbon fuels.1,5,6 The bio-oils need to be upgraded before used as fuels. Hydrodeoxygenation (HDO) has been widely used for upgrading the bio-oils. Conventional sulfide Ni(Co)-Mo(W)/ Al2O3 catalysts have been investigated for this process.7-9 In these catalysts, the sulfur anion vacancies or coordinatively unsaturated sites, located at the edges of MoS2 slabs, are often believed to be active sites. However, the S loss because of replacement with O resulted in the deactivation of the sulfided catalysts.10,11 To avoid the deactivation, S-containing agents (e.g., H2S and CS2) were usually added to the reactants. However, the agents not only lower the catalyst activity but also incorporate the reactants, forming undesirable S-containing products.8,12,13 The noble metal catalysts have also been investigated.13,14 In comparison to the sulfide CoMo/Al2O3 catalysts, the Rh catalysts had better performance in the HDO r 2011 American Chemical Society

of guaiacol; however, the products had a lower H/C molar ratio than gasoline and diesel.13 Thus, a novel catalyst with high HDO activity and hydrogenation ability needs to be developed. As new hydroprocessing catalysts, transition-metal phosphides have attracted great attention in recent decades. In principle, the special catalytic properties of the metal phosphides are ascribed to the ensemble and/or ligand effects of P.15-17 It has been reported that the phosphide catalysts have excellent activities for hydrodenitrogenation, hydrodesulfurization, and hydrodechlorination.18-22 Recently, the HDO over nickel phosphides has also been investigated under sulfur-involving conditions. Ni2P/ MCM-41 and Ni2P/SiO2 showed high activities for the HDO of phenol and benzofuran, respectively.23-25 However, the existing S species may contribute to the HDO through the formation of phosphosulfide.25 Therefore, the HDO performance of the metal phosphides by themselves is worth investigating. In this work, Ni2P/SiO2, MoP/SiO2, and NiMoP/SiO2 catalysts with different Ni/Mo molar ratios were prepared and characterized. It is known that, in bio-oils, the most refractory compounds under hydrotreatment conditions are phenolic compounds. Anisole, having a representative Cmethyl-O-Caromatic structure of the oxygen-containing compounds was used as a model compound to test the HDO performance of the phosphide catalysts. The active sites of phosphide catalysts were investigated, and the catalyst deactivation was also preliminarily analyzed. Received: May 5, 2010 Revised: December 28, 2010 Published: January 31, 2011 854

dx.doi.org/10.1021/ef101258j | Energy Fuels 2011, 25, 854–863

Energy & Fuels

ARTICLE

Table 1. Nominal Compositions and Properties of Support and Catalysts Ni content

Mo content

(wt %)

(wt %)

PxOy/SiO2a

0

0

MoP/SiO2

0

15

NiMo(0.5)P/SiO2

3.5

11.5

NiMo(1.0)P/SiO2

5.7

9.3

NiMo(2.0)P/SiO2

8.2

6.8

catalyst

BET surface area

average pore

pore volume

CO uptake

(m2/g)

diameter (nm)

(cm3/g)

( μmol/gcatal)

548

5.5

0.8

0

0

288

5

0.49

76

0.5

291

4.9

0.47

83

1

317

4.7

0.49

88

2

351

4.6

0.52

77

296 374

4.9 4.6

0.47 0.56

67 15

459

4.5

0.65

44

Ni/Mo molar ratio

SiO2

Ni2P/SiO2 Ni-POx/SiO2b

15 15

Ni/SiO2

15

0

3

16

NiMo/γ-Al2O3 a

0

0 0 0.3

With the same P content as the Ni2P/SiO2 precursor. b Prepared from the reduction of the Ni2P/SiO2 precursor at 450 °C.

2. EXPERIMENTAL SECTION

X-ray diffraction (XRD) patterns were obtained by a Rigaku D/max 2500v/PC powder diffractometer operated at 45 kV and 200 mA, using Cu KR radiation (λ = 0.154 18 nm). CO chemisorption uptakes were measured using the same apparatus as with H2-TPR. A 100 mg portion of the sample was loaded in the reactor and reduced in a H2 flow (60 mL min-1) at 450 °C for 1 h. Afterward, the sample was flushed with a He flow (80 mL min-1) at 450 °C for 1 h. Subsequently, the sample was cooled to 30 °C. When the TCD signal was stable, pulses of CO (50 μL) were passed through the samples until the effluent areas of consecutive pulses were constant. The total dynamic CO uptake was then calculated. X-ray photoelectron spectroscopy (XPS) was carried out using a PHI-1600 ESCA instrument with Mg KR radiation (1253.6 eV). Binding energies were determined with C 1s (284.6 eV) as the reference. For XPS measurement, the catalysts without passivation were stored in ethanol. Before measurement, the catalysts were transferred into the instrument in air, during which the catalysts could be oxidized. H2-TPD was carried out using the same apparatus as with H2-TPR. A 70 mg portion of sample was reduced with a H2 flow (60 mL min-1) at 450 °C for 1 h and then decreased to 30 °C. After an adsorption of H2 for 30 min, the sample was swept with a N2 flow (20 mL min-1) until the TCD signal was stable. H2-TPD was conducted at a heating rate of 15 °C min-1. The desorbed H2 was detected by a TCD. Before detection, the gas was passed through a trap containing solid NaOH to remove water. NH3-TPD was carried out using the same apparatus as with H2TPR. A 70 mg portion of sample was reduced with a H2 flow (40 mL min-1) at 450 °C for 1 h and then cooled to 100 °C. After NH3 adsorption in a NH3/He flow for 30 min, the sample was swept with a He flow to remove the physically adsorbed NH3. Afterward, NH3-TPD was performed in a He flow (60 mL min-1) at a heating rate of 15 °C min-1. The desorbed NH3 was detected by a TCD. Before detection, the gas was passed through a trap containing solid NaOH to remove water. 2.3. Activity Test of Catalysts. The catalyst activity was tested on a continuous-flow stainless-steel reactor (inner diameter of 12 mm and length of 380 mm). For the experiments, 1.0 or 0.1 g of catalyst sample (0.15-0.3 mm in diameter) was loaded in the reactor and 2.0 g of quartz sand (0.3-0.45 mm in diameter) was placed on the catalyst bed to preheat the reactants. In a H2 flow (>99.9%, 100 mL min-1), the passivated phosphide catalysts were rereduced at 450 °C for 1 h, while the Ni/SiO2, Ni-POx/SiO2, PxOy/SiO2, and NiMo/γ-Al2O3 catalysts were in-situ-prepared. After the temperature and H2 total pressure were adjusted to 300 °C and 1.5 MPa, respectively, a n-octane solution containing 4 wt % anisole was fed into the reactor at a rate of 10 g h-1

2.1. Preparation of Catalysts. The Ni2P/SiO2, MoP/SiO2, and NiMoP/SiO2 catalysts with different Ni/Mo molar ratios were prepared by the temperature-programmed reduction (TPR) method from the supported phosphate precursors. The NiMoP/SiO2 catalysts with different Ni/Mo molar ratios are denoted as NiMo(x)P/SiO2 (x = Ni/Mo molar ratio). First, the silica support was incipiently impregnated with an aqueous solution containing NH4H2PO4, (NH4)6Mo7O24, and/ or Ni(NO3)2. After drying at 120 °C for 12 h and calcination at 500 °C for 4 h, the precursor was obtained. In the precursors, the nominal molar ratio of P/metal (Ni and Mo) was 1.0. Second, the precursor was reduced from 20 to 650 °C at a rate of 1 °C min-1 and then maintained at 650 °C for 3 h. The H2 (>99.9%) flow was set as 320 mL min-1 per gram of precursor. The prepared catalysts were cooled to room temperature in the H2 flow and then passivated in a 0.5 vol % O2/N2 flow (100 mL min-1) for 5 h. For comparison, the Ni2P/SiO2 precursor was also reduced at 450 °C for 2 h. The resulting sample is labeled as Ni-POx/SiO2. For comparison, NiO/SiO2, MoO/SiO2, and PxOy/SiO2 samples were also prepared. Silica was incipiently impregnated with the aqueous solutions of Ni(NO3)2, (NH4)6Mo7O24, and NH4H2PO4, respectively, followed by drying at 120 °C for 12 h and calcination at 500 °C for 4 h. After a reduction of NiO/SiO2 at 450 °C for 2 h, the Ni/SiO2 catalyst was prepared. NiMo/γ-Al2O3, a conventional HDO catalyst, was prepared by the sequential impregnation method. γ-Al2O3 was incipiently impregnated with an aqueous solution of Ni(NO3)2 followed by drying at 120 °C for 12 h. The resulting sample was then incipiently impregnated with an aqueous solution of (NH4)6Mo7O24, followed by drying at 120 °C for 12 h and calcination at 500 °C for 4 h. After a reduction at 450 °C for 2 h, the NiMo/γ-Al2O3 catalyst was prepared. The nominal compositions of the catalysts are listed in Table 1. 2.2. Characterization of Catalysts. The reducibility of precursors was characterized by the H2 temperature-programmed reduction (H2-TPR) using a quartz U-tube reactor (inner diameter of 4 mm), in which 50 mg of catalyst was loaded in the thermostatic zone. Reduction was conducted at a heating rate of 10 °C min-1 in a 10 vol % H2/N2 flow (60 mL min-1). The hydrogen consumption was determined by a thermal conductivity detector (TCD). N2 adsorption-desorption isotherms of the catalysts were measured by a Micromeritics TriStar 3000 apparatus at -196 °C. The BrunauerEmmett-Teller (BET) equation was used to calculate the specific surface area (SBET). The average pore diameter was calculated using the desorption branch of the isotherm according to the BarrettJoyner-Halenda (BJH) method. 855

dx.doi.org/10.1021/ef101258j |Energy Fuels 2011, 25, 854–863

Energy & Fuels

ARTICLE

[normal temperature and pressure (NTP)]. Thus, corresponding to 1.0 and 0.1 g of catalyst loadings, the weight hourly space velocities (WHSVs) of the feed solution were 10 and 100 h-1, respectively. The molar ratio of H2/anisole was maintained at 24. The reaction was found to operate with negligible mass- and heat-transfer limitations. The liquid products were sampled at 1 h intervals and were then analyzed with a SP-3420 gas chromatograph equipped with a flame ionization detector (FID) and a commercial SE-30 capillary column (100% methyl polysiloxane, 50.0 m  0.32 mm  3.0 μm). The n-heptane was used as an internal standard. The conversion of anisole (Xanisole) and product selectivity (Si) were defined as follows: ! nanisole  100% ð1Þ Xanisole ¼ 1nanisole,0 ni Si ¼ nanisole,0 -nanisole

!  100%

ð2Þ

where nanisole,0 and nanisole denote the molars of aisole in the feed and product, respectively, and ni is the molar of product i (for example, benzene and cyclohexane). HDO conversion (XHDO) was used to identify the hydrodeoxygenation ability of the catalyst. Because benzene and cyclohexane were the main products via HDO, XHDO was calculated as follows: XHDO ¼

nbenzene þncyclohexane  100% nanisole,0 -nanisole

Figure 1. H2-TPR profiles of (a) PxOy/SiO2, (b) NiO/SiO2, (c) Ni2P/ SiO2 precursor, (d) NiMo(2.0)P/SiO2 precursor, (e) NiMo(1.0)P/ SiO2 precursor, (f) NiMo(0.5)P/SiO2 precursor, (g) MoP/SiO2 precursor, and (h) MoO/SiO2.

ð3Þ

where nbenzene and ncyclohexane denote the molars of benzene and cyclohexane in the product, respectively. On the basis of the amount of active sites measured by the CO chemisorption, the turnover frequencies for the conversion of anisole and HDO were calculated, which are expressed as TOFanisole and TOFHDO, respectively.

The NiMo(x)P/SiO2 (x = 0.5, 1.0, and 2.0) precursors had two peaks (curves d-f of Figure 1). The first peak mainly belongs to the reduction of Mo6þ to Mo4þ, while the latter peak is related to the co-reduction of Mo4þ, Ni2þ, and phosphates. This is consistent with the study on the reduction mechanism of the MoNiP/SiO2 precursor, during which MoO3, MoO2, Mo, and MoNiP are formed sequentially.17 In comparison to the MoP/ SiO2 precursor, the NiMo(x)P/SiO2 precursors were reduced more easily, especially the reduction of Mo6þ to Mo4þ. With the increase of the Ni/Mo molar ratio, the reduction of Mo6þ to Mo4þ in NiMo(x)P/SiO2 precursors became facile. This phenomenon may be explained as follow: (1) Ni and P species promoted the dispersion of Mo species,30,31 and (2) the facile reducible Ni-Mo species formed.32 3.1.2. XRD. The XRD patterns of the catalysts are shown in Figure 2. A broad diffraction peak between 15° and 35° is caused by amorphous silica. Ni2P and MoP were detected in the Ni2P/ SiO2 and MoP/SiO2 catalysts (curves c and g of Figure 2), respectively. For the NiMo(0.5)P/SiO2 and NiMo(1.0)P/SiO2 catalysts, no obvious diffraction peaks were ascribed to phosphides. However, this cannot rule out the possibility of the existence of Ni2P, MoP, and NiMoP solid solution.32,33 As reported by Ma et al., when Ni was added to MoP, the intensities of diffraction peaks ascribed to MoP were reduced, while Ni2P and NiMoP solid solutions were formed.33 Thus, the formed phosphides might be highly dispersed on the NiMo(0.5)P/ SiO2 and NiMo(1.0)P/SiO2 catalysts. The NiMo(2.0)P/SiO2 catalyst had a similar pattern to Ni2P/SiO2, while its peaks shifted to lower angles. This may be due to the formation of NiMoP solid solutions through a homogeneous substitution of Ni with Mo. As a result, the lattice parameters of NiMoP

3. RESULTS AND DISCUSSION 3.1. Characterization of Catalysts. 3.1.1. H2-TPR. The H2TPR profiles of the catalyst precursors are shown in Figure 1. In the H2-TPR profile of PxOy/SiO2 (curve a of Figure 1), the peak centered at about 878 °C is attributed to the reduction of the high thermally stable P-O bond.26 The NiO/SiO2 sample gave two peaks at 360 and 490 °C (curve b of Figure 1), which are ascribed to the reductions of bulk NiO and nickel silicate, respectively.27 Two peaks were visible in the trace of MoO/SiO2 (curve h of Figure 1). They are ascribed to the reductions of MoO3 to MoO2 and MoO2 to Mo, respectively.28 The Ni2P/SiO2 precursor (curve c of Figure 1) had two peaks centered at about 700 and 800 °C, ascribed to the reductions of nickel species and the P-O bond, respectively. Clearly, the P-O bond in the Ni2P/SiO2 precursor was reduced more easily than that in PxOy/SiO2. This is attributed to the active hydrogen species derived from the dissociation of H2 on metallic Ni and/or nickel phosphides formed earlier during the reduction.25,29 There were two peaks in the H2-TPR profile of the MoP/SiO2 precursor (curve g of Figure 1). Associated with the H2-TPR profiles of PxOy/SiO2 and MoO/SiO2 (curves a and h of Figure 1), the peak centered around 500 °C is ascribed to the reduction of Mo6þ to Mo4þ, while the one centered at 760 °C is related to the co-reduction of Mo4þ species and P-O bond.26 856

dx.doi.org/10.1021/ef101258j |Energy Fuels 2011, 25, 854–863

Energy & Fuels

ARTICLE

adsorb to bridged and 3-fold-coordinated Ni on the metallic Ni.34 However, it is mainly linearly bound to the Ni site on Ni2P.35 This is ascribed to the ensemble effect of P species, which enlarges the Ni-Ni distance in comparison to that of metallic Ni.36 The enlarged Ni-Ni distance might inhibit the adsorption of CO on the bridged and 3-fold-coordinated Ni sites. The linear adsorption favors more CO bound to Ni sites. 3.1.5. XPS. XPS was used to characterize the electronic properties of the catalysts. The XPS spectra in the Ni 2p, Mo 3d, and P 2p regions are shown in Figure 3. In Figure 3a, two Ni 2p peaks at around 853.1 and 856.8 eV are ascribed to reduced Niδþ (0 < δ < 2) and Ni2þ, respectively.37 In nickel phosphides, there is a small electron transfer from Ni to P. Niδþ (0 < δ < 2) covalently bonds with Pδ- (0 < δ < 1), whose binding energy was about 129.3 eV (Figure 3c). Ni-POx/SiO2 had an unremarkable peak at 853.1 eV, which is ascribed to its incomplete reduction. The Ni2þ species on Ni2P/SiO2 was formed when the sample was exposed to air.20 Interestingly, the Ni2þ species were not detected on NiMo(1.0)P/SiO2. Figure 3b shows the XPS spectra in the Mo 3d region for MoP/SiO2 and NiMo(1.0)P/SiO2. Two peaks of Mo 3d5/2 are observed. The peak with lower binding energy (228.3-228.9 eV) corresponds to Moδþ species (0 < δ e 4) bound to Pδ- (0 < δ < 1) in phosphides, while the peak with higher binding energy (231.1-232.0 eV) is assigned to the oxidized molybdenum.38,39 The Mo 3d5/2 binding energies of NiMo(1.0)P/SiO2 were lower than those of MoP/SiO2, indicating the interaction between Ni and Mo, with an electron transfer from Ni to Mo.17,32 Figure 3c shows the XPS spectra in the P 2p3/2 region. Two peaks were observed for the phosphide catalysts. The binding energies of P species at around 129.3 eV indicate the formation of phosphides, in which Ni and/or Mo covalently bond with reduced P. The binding energies around 134.0 eV are assigned to P5þ, which was derived from the oxidation when the samples contacted air. P5þ on the Ni-POx/SiO2 catalyst was mainly attributed to unreduced phosphate. 3.1.6. H2-TPD. The H2-TPD results of different samples are shown in Figure 4 and Table 2. Usually, the hydrogen species desorbed below 400 °C are ascribed to those adsorbed on the metal sites, while those desorbed above 400 °C are attributed to the spilt-over hydrogen species.40 In the H2-TPD profiles of the phosphide catalysts (curves a-e of Figure 4), there were two peaks below 400 °C, indicating two different states of adsorbed H species on the phosphide particles. With the increase of the Ni/Mo molar ratio, the first peak became large and the shoulder peak shifted toward a low temperature. As shown in Table 2, below 400 °C, Ni2P/SiO2 had the largest amount of desorbed H2, while the other phosphide catalysts had similar amounts. Thus, the H2 desorption below 400 °C was promoted by Ni for the phosphide catalysts. Above 400 °C, there were also two peaks centered at around 580 and 800 °C in the H2-TPD profiles of the phosphide catalysts (curves a-e of Figure 4). With the increase of the Ni/Mo molar ratio, the former peak shifted toward a low temperature and the peak area decreased. In Table 2, the relative H2 desorption amount above 400 °C decreased with the increase of the Ni/Mo molar ratio, indicating a decrease of the amount of spilt-over hydrogen species. In short, the higher the Ni/Mo ratio, the less the amount of spilt-over hydrogen and the more easily the spilt-over hydrogen species reversed back to the surface of phosphide particles.

Figure 2. XRD patterns of (a) Ni/SiO2, (b) Ni-POx/SiO2, (c) Ni2P/ SiO2, (d) NiMo(2.0)P/SiO2, (e) NiMo(1.0)P/SiO2, (f) NiMo(0.5)P/ SiO2, and (g) MoP/SiO2.

increased, while the Fe2P-type structure adopted by Ni2P was still maintained.33 The XRD patterns of Ni/SiO2 and Ni-POx/SiO2 are also shown in curves a and b of Figure 2, respectively. Metallic Ni was detected in the Ni/SiO2 catalyst, while no diffraction peaks corresponding to the metallic Ni and nickel phosphides are presented in the pattern of Ni-POx/SiO2. This is ascribed to the incomplete reduction and the high dispersion of reduced nickel in Ni-POx/SiO2. 3.1.3. N2 Adsorption-Desorption. N2 sorption was used to measure the surface areas, pore diameters, and pore volumes of the catalysts. The results are listed in Table 1. In comparison to the Ni2P/SiO2 and MoP/SiO2 catalysts, the NiMo(x)P/SiO2 catalysts possessed larger specific surface areas and smaller average pore diameters. This is related to the high dispersion of phosphides on the NiMo(x)P/SiO2 catalysts as indicated by XRD and CO uptakes (as seen in Table 1). Because of the high preparation temperature, Ni2P/SiO2 had a lower surface area and a larger pore diameter than Ni/SiO2. Ni-POx/SiO2 has a lower surface area than Ni/SiO2 because of the pore blockage of unreduced phosphate. 3.1.4. CO Uptake. CO chemisorption was used to measure the exposed surface metal sites.25 Table 1 presents the CO uptakes of different catalysts. The NiMo(x)P/SiO2 catalysts had larger CO uptakes than the MoP/SiO2 and Ni2P/SiO2 catalysts. This is consistent with the XRD result that NiMo(x)P/SiO2 had a higher dispersion of phosphides. Ni-POx/SiO2 had lower CO uptake than Ni/SiO2 and Ni2P/ SiO2, which might be caused by the coverage of reduced Ni sites with the unreduced phosphate. In comparison to Ni/SiO2, the larger CO uptake of Ni2P/SiO2 might be related to the CO adsorption model. Apart from linear adsorption, CO can also 857

dx.doi.org/10.1021/ef101258j |Energy Fuels 2011, 25, 854–863

Energy & Fuels

ARTICLE

Figure 4. H2-TPD profiles of (a) MoP/SiO2, (b) NiMo(0.5)P/SiO2, (c) NiMo(1.0)P/SiO2, (d) NiMo(2.0)P/SiO2, (e) Ni2P/SiO2, (f) NiPOx/SiO2, (g) Ni/SiO2, and (h) PxOy/SiO2.

Table 2. H2- and NH3-TPD Data of Catalysts relative H2 desorption amounta

catalyst

below

above

400 °C

400 °C

total

desorption amountb

relative NH3

MoP/SiO2

1.00

2.12

3.12

1.00

NiMo(0.5)P/SiO2

1.01

2.05

3.06

0.89

NiMo(1.0)P/SiO2

0.96

2.02

2.98

1.03

NiMo(2.0)P/SiO2

0.99

1.85

2.84

1.02

Ni2P/SiO2

1.14

1.60

2.74

0.94

Ni-POx/SiO2

1.01

Ni/SiO2

0.31

PxOy/SiO2

1.00

The H2 desorption amount of MoP/SiO2 below 420 °C was designated as 1.00. b The NH3 desorption amount of MoP/SiO2 was designated as 1.00. a

The Ni/SiO2 catalyst had a similar H2-TPD profile to Ni2P/SiO2, while it had less desorbed H2 in all temperature ranges. The more H2 desorbed from Ni2P/SiO2 may be ascribed to the PO-H groups. In the H2-TPD profile of Ni-POx/SiO2, the peak centered at 180 °C is ascribed to H2 desorbed from reduced Ni sites, while the peaks above 400 °C may mainly be related to H2 derived from unreduced phosphate. Thus, the H2 desorption from the phosphide catalysts are ascribed to the H species both adsorbed on metal sites and provided by PO-H groups. The increase of the Ni/Mo ratio lowers the H2 desorption temperature. 3.1.7. NH3-TPD. The NH3-TPD results of the catalysts are shown in Figure 5 and Table 2. For each phosphide catalyst

Figure 3. XPS spectra of (a) Ni 2p, (b) Mo 3d, and (c) P 2p regions for MoP/SiO2, NiMo(1.0)P/SiO2, Ni2P/SiO2, and Ni-POx/SiO2.

For comparison, the H2-TPD profiles of Ni-POx/SiO2, Ni/ SiO2, and PxOy/SiO2 are shown in curves f-h of Figure 4. Interestingly, the traces of the PxOy/SiO2 and Ni-POx/SiO2 samples were under the baselines above 600 °C. H2 desorbed from PxOy/SiO2 may derive from the PO-H groups in phosphates. 858

dx.doi.org/10.1021/ef101258j |Energy Fuels 2011, 25, 854–863

Energy & Fuels

ARTICLE

Scheme 1. HDO Network of Anisolea

a

(a) Demethylation, (b and f) hydrogenolysis, (c) hydrogenation, (d) hygrogenation-hydrogenolysis, and (e and g) methyl transfer.

Figure 5. NH3-TPD profiles of (a) MoP/SiO2, (b) NiMo(0.5)P/SiO2, (c) NiMo(1.0)P/SiO2, (d) NiMo(2.0)P/SiO2, (e) Ni2P/SiO2, (f) NiPOx/SiO2, (g) Ni/SiO2, and (h) PxOy/SiO2.

(curves a-e of Figure 5), there was a peak at around 210 °C with a shoulder at a higher temperature. In comparison to the Mo phosphide-containing catalysts, the Ni2P/SiO2 catalyst had a more remarkable shoulder peak at 300 °C (Figure 5e). As shown in Table 2, the MoP/SiO2 catalyst had more acid sites than the Ni2P/SiO2 catalyst. For comparison, the NH3-TPD profiles of Ni-POx/SiO2, Ni/ SiO2, and PxOy/SiO2 are presented in curves f-h of Figure 5. PxOy/SiO2 had a peak centered at 210 °C, which is mainly ascribed to the PO-H groups with Brønsted acidity.35 A peak centered at 320 °C in the profile of Ni/SiO2 might be attributed to Lewis acid sites because of the trace of the unreduced Ni species in the form of nickel silicate.41,42 Because the Ni/SiO2 catalyst had higher NH3 desorption temperature than PxOy/ SiO2, the Lewis acid sites on the Ni/SiO2 catalyst might be stronger than the Brønsted acid sites on PxOy/SiO2. In combination with the acid properties of PxOy/SiO2 and Ni/SiO2, both Brønsted and Lewis acid sites might exist on Ni-POx/SiO2 and are related to the PO-H groups and unreduced Ni species, respectively. Because of the P loss during the reduction, the Ni2P/SiO2 catalyst has less Brønsted acid sites than Ni-POx/ SiO2. Oyama et al.35 found that the Lewis acidity is dominant over the Brønsted acidity (because of PO-H groups) on Ni2P/ SiO2. The Lewis acidity on Ni2P may be related to the Ni species bearing a small positive charge. Because of the electron transfer from metal to P, the other phosphides have a similar acid property to Ni2P. Thus, on the phosphide catalysts, the PO-H group is associated with Brønsted acidity, while the metal site (i.e., Niδþ and Moδþ) is related to Lewis acidity. 3.2. Test of Catalysts. As a model reaction, the HDO of anisole has been investigated, and the reaction network has been

Figure 6. Anisole conversion over (b) MoP/SiO2, (2) NiMo(0.5)P/ SiO2, (1) NiMo(1.0)P/SiO2, ([) NiMo(2.0)P/SiO2, (tilted, leftpointing 2) Ni2P/SiO2, (0) Ni-POx/SiO2, (]) Ni/SiO2, (O) PxOy/SiO2, (g) NiMo/γ-Al2O3, and () γ-Al2O3. Reaction conditions: temperature, 300 °C; pressure, 1.5 MPa; H2/anisole ratio, 24; and WHSV, 10 h-1.

proposed as shown in Scheme 1.43 During the HDO of anisole, four reaction routes, i.e., demethylation, hydrogenation, hydrogenolysis, and methyl transfer, can occur. Because the rupture of the Cmethyl-O bond is easier than that of the Caromatic-O bond, demethylation via the rupture of the Cmethyl-O bond occurs first, giving phenol as a primary intermediate product. Further, phenol is converted to benzene via the direct hydrogenolysis of the Caromatic-O bond (DDO) and to alicyclic hydrocarbons via the hydrogenation of the aromatic ring followed by Calicylic-O hydrogenolysis (HYD).8 In the present test, phenol, benzene, and cyclohexane were detected as main products, while no cyclohexanol and cyclohexanone were detected. Thus, we propose that the HDO of anisole mainly involve two steps, that is, the conversion of anisole to 859

dx.doi.org/10.1021/ef101258j |Energy Fuels 2011, 25, 854–863

Energy & Fuels

ARTICLE

Table 3. Activity Data of Phosphide Catalystsa Xanisole,2b

XHDO,2b

TOFanisole,2b

TOFHDO,2b

TOFanisole,12c

TOFHDO,12c

Danisoled

DHDOd

(%)

(%)

(10-3/s)

(10-3/s)

(10-3/s)

(10-3/s)

(%)

(%)

MoP/SiO2

17.8

6.0

24.1

8.1

16.5

3.4

31.3

58.3

NiMo(0.5)P/SiO2

19.7

9.2

24.4

11.4

15.3

4.2

37.3

63.2

NiMo(1.0)P/SiO2

20.7

12.6

24.2

14.7

19.6

9.5

19.2

35.2

NiMo(2.0)P/SiO2

28.6

24.5

38.2

32.7

33.0

24.7

13.6

24.5

Ni2P/SiO2

76.8

73.6

91.2

80.3

22.6

28.9

catalyst

118

113

Reaction conditions: temperature, 300 °C; pressure, 1.5 MPa; H2/anisole ratio, 24; and WHSV, 100 h-1. b Data at the 2nd hour. c Data at the 12th hour. d Deactivation degree of the catalyst calculated on the basis of TOF. Danisole = ((TOFanisole,2 - TOFanisole,12)/(TOFanisole,2))  100%. DHDO = ((TOFHDO,2 - TOFHDO,12)/(TOFHDO,2))  100%. a

associated with the PO-H group, while the Niδþ site had superior activity to the Moδþ site. 3.2.2. Conversion of Phenol via Hydrogenolysis and Hydrogenation: Selectivities of Products. After the demethylation of anisole, the formed phenol was converted to benzene and cyclohexane through hydrogenolysis and hydrogenation. Panels a-c of Figure 7 show the selectivities of phenol, benzene, and cyclohexane as a function of time-on-stream at the WHSV of 10 h-1. Over the phosphide catalysts, with the increase of the Ni/Mo ratio, the selectivities of phenol and benzene decreased (panels a and b of Figure 7), while the selectivity of cyclohexane increased (Figure 7c). As a result, the total selectivity of cyclohexane and benzene (SCþB) and the molar ratio of cyclohexane/benzene (SC/B) increased (panels d and e of Figure 7) with the Ni/Mo ratio. Figure 8 shows the HDO conversions over the phosphide catalysts at the WHSV of 10 h-1. The increased Ni/Mo ratio favors the XHDO. This trend also existed when the reaction was tested at the WHSV of 100 h-1 (see Table 3). Thus, the Ni site in phosphide was more active than the Mo site for the hydrogenolysis and hydrogenation. In comparison to the Ni-phosphide-containing catalysts, NiMo/γ-Al2O3 had higher phenol selectivity and lower SCþB and SC/B, indicative of its lower activity for the hydrogenolysis and hydrogenation. Different from NiMo/Al2O3, nearly no benzene and cyclohexane were produced over γ-Al2O3; that is, the pure γ-Al2O3 support had not activity for the hydrogenolysis and hydrogenation. The Ni/SiO2 catalyst had lower SCþB than Ni2P/SiO2, while it gave higher SC/B than Ni2P/SiO2 during the reaction from 3 to 7 h. After the reaction for 7 h, the SC/B of Ni/ SiO2 was lower than that of Ni2P/SiO2, indicating a more remarkable deactivation of Ni/SiO2 for hydrogenation. This difference may be related to the participation of PO-H in the reactions, as indicated as follows. The Ni-POx/SiO2 catalyst had lower SCþB and SC/B than the Ni/SiO2 and Ni2P/SiO2 catalysts. This may be related to less reduced Ni sites on Ni-POx/SiO2. PxOy/SiO2 had no metal sites, while it gave benzene and cyclohexane, even though its activity was much lower than that of the metal site. This indicates that the PO-H group, acting as a Brønsted acid site that can provide hydrogen, also catalyzed the hydrogenolysis and hydrogenation. As reported by Cairon et al.,47 the Brønsted acid site plays an important role in the hydrogenation of benzene. In the phosphides, the Ni and Mo sites possess small positive charges. The DFT calculations have shown that the replacement of Ni atoms in Ni2P with Mo atoms decreases the d electron density of metal near the Fermi level because of an electron transfer from Ni to Mo.17,32 On the basis of this, the d electron

phenol via demethylation and the conversion of phenol to hydrocarbons through the hydrogenolysis and hydrogenation. Here, we mainly investigated the activities of the catalysts and the nature of active sites for the demethylation, hydrogenolysis, and hydrogenation. 3.2.1. Conversion of Anisole via Demethylation. Figure 6 shows the anisole conversions as a function of time-on-stream at the WHSV of 10 h-1. Because the anisole conversion in the blank reactor was very low (about 2%), the effect of the reactor was ignored. Among the phosphide catalysts, MoP/SiO2 had the lowest initial anisole conversion of 65%. The Ni-phosphidecontaining catalysts gave initial anisole conversions above 95%. Clearly, as the Ni/Mo molar ratio increased, the anisole conversion increased. This trend also existed as the reaction was performed at the WHSV of 100 h-1 (see Table 3). Table 3 also shows the TOFanisole for the phosphide catalysts. The increased Ni/Mo molar ratio favors the TOFanisole. The Ni2P/SiO2 catalyst had the highest TOFanisole. As references, the Ni-POx/SiO2, Ni/SiO2, and PxOy/SiO2 catalysts were also tested. The results are shown in Figure 6. In comparison to Ni-POx/SiO2 and PxOy/SiO2, Ni/SiO2 had higher anisole conversion that is similar to that of Ni2P/SiO2 before 7 h. However, Ni/SiO2 deactivated after the 7th hour, and the anisole conversion decreased to about 84% at the 10th hour. The superior stability of Ni2P/SiO2 to that of Ni/SiO2 is discussed in section 3.3. The NiMo/γ-Al2O3 catalyst, a conventional catalyst for HDO, was also tested. As shown in Figure 6, the anisole conversion over the NiMo/γ-Al2O3 catalyst was about 95%, which is slightly lower than that of Ni2P/SiO2. Because the γ-Al2O3 support gave a similar anisole conversion to NiMo/γ-Al2O3, the demethylation of anisole over NiMo/γ-Al2O3 may mainly be ascribed to that of the γ-Al2O3 support.43,44 The high activity of the γ-Al2O3 support indicates that the Lewis acid site is mainly the active site. Also, the low anisole conversion over PxOy/SiO2 indicates that the Brønsted acid site is less active than the Lewis acid site because the Brønsted acid site was dominant on PxOy/SiO2. During the demethylation of anisole, the Lewis acid site can accept the free p electron pair of the oxygen atom, forming a weak coordinative bond, followed by the heterolytic scission of the Cmethyl-O bond.45,46 In the phosphide catalysts, the reduced metal sites acting as Lewis acid sites were dominant and account for their high activities for the demethylation. Because the anisole conversion increased with the increase of the Ni/Mo molar ratio, the Niδþ site in phosphide was more active than the Moδþ site. Thus, for the demethylation, the Lewis acid site related to the metal site was much more active than the Brønsted acid site 860

dx.doi.org/10.1021/ef101258j |Energy Fuels 2011, 25, 854–863

Energy & Fuels

ARTICLE

Figure 7. Selectivity of (a) phenol, (b) benzene, (c) cyclohexane, (d) cyclohexane and benzene, and (e) cyclohexane/benzene ratio over (b) MoP/ SiO2, (2) NiMo(0.5)P/SiO2, (1) NiMo(1.0)P/SiO2, ([) NiMo(2.0)P/SiO2, (tilted, left-pointing 2) Ni2P/SiO2, (0) Ni-POx/SiO2, (]) Ni/SiO2, (O) PxOy/SiO2, (g) NiMo/γ-Al2O3, and () γ-Al2O3. Reaction conditions: temperature, 300 °C; pressure, 1.5 MPa; H2/anisole ratio, 24; and WHSV, 10 h-1.

while the Niδþ site was more active than the Moδþ site and PO-H group. 3.3. Preliminary Analysis for Catalyst Deactivation. As shown in Figures 6-8 and Table 3, the decreases of the anisole conversion, HDO conversion, and SC/B indicate that there was a deactivation of the phosphide catalysts with time on stream. The deactivation of the sulfide catalysts has been investigated, and is attributed to coke deposition of S loss and inhibition of water.43,48 Here, only the effect of water is preliminarily analyzed. The other reasons are being investigated. In phosphide catalysts, the metal sites covalently bound with P were active for the demethylation, hydrogenolysis, and hydrogenation. Similar to sulfide, phosphide might be oxidized by the byproduct water.48

density of metal near the Fermi level may increase on the order of MoP < NiMo(0.5)P < NiMo(1.0)P < NiMo(2.0)P < Ni2P. As proposed by Ueckert et al., the metal site with a high electron density favors the formation of the π back bond between the aromatic ring and metal sites, which promotes the hydrogenation of the aromatic ring.34 This may be one reason for the higher activity of the phosphide catalyst with a higher Ni/Mo ratio. Additionally, the NiMo(x)P/SiO2 catalysts had larger CO uptakes than Ni2P/SiO2, while they had less activities, indicating that there was no synergism between the Ni and Mo sites for the hydrogenolysis and hydrogenation. In short, both metal sites (Niδþ and Moδþ) and PO-H groups were active for demethylation, hydrogenation, and hydrogenolysis, 861

dx.doi.org/10.1021/ef101258j |Energy Fuels 2011, 25, 854–863

Energy & Fuels

ARTICLE

During the HDO of anisole, the metal and P in phosphide may be transformed into metal oxide and/or phosphate by water, which might cover the metal sites and transform the Lewis acid sites to Brønsted acid sites, leading to the deactivation of phosphide catalysts. In comparison to MoP/SiO2, Ni2P/SiO2 had better resistance to oxidation.

’ AUTHOR INFORMATION Corresponding Author

*Telephone: þ86-22-27890865. Fax: þ86-22-87894301. E-mail: [email protected].

Figure 8. HDO conversion over (b) MoP/SiO2, (2) NiMo(0.5)P/ SiO2, (1) NiMo(1.0)P/SiO2, ([) NiMo(2.0)P/SiO2, and (tilted, leftpointing 2) Ni2P/SiO2. Reaction conditions: temperature, 300 °C; pressure, 1.5 MPa; H2/anisole ratio, 24; and WHSV, 10 h-1.

’ ACKNOWLEDGMENT The authors acknowledge the support from the Natural Science Foundation of Tianjin (08JCYBJC01600) and the Program of Introducing Talents to the University Disciplines (B06006). ’ REFERENCES (1) Huber, G.; Iborra, S.; Corma, A. Synthesis of transportation fuels from biomass: Chemistry, catalysts, and engineering. Chem. Rev. 2006, 106 (9), 4044–4098. (2) Demirbas, A. Biomass resource facilities and biomass conversion processing for fuels and chemicals. Energy Convers. Manage. 2001, 42 (11), 1357–1378. (3) Furimsky, E. Catalytic hydrodeoxygenation. Appl. Catal., A 2000, 199 (2), 147–190. (4) Bridgwater, A. V.; Meier, D.; Radlein, D. An overview of fast pyrolysis of biomass. Org. Geochem. 1999, 30 (12), 1479–1493. (5) Czernik, S.; Bridgwater, A. V. Overview of applications of biomass fast pyrolysis oil. Energy Fuels 2004, 18 (2), 590–598. (6) Mohan, D.; Pittman, C. U.; Steele, P. H. Pyrolysis of wood/ biomass for bio-oil: A critical review. Energy Fuels 2006, 20 (3), 848–889. (7) Bunch, A. Y.; Wang, X. Q.; Ozkan, U. S. Hydrodeoxygenation of benzofuran over sulfided and reduced Ni-Mo/γ-Al2O3 catalysts: Effect of H2S. J. Mol. Catal. : A: Chem. 2007, 270 (1-2), 264–272. (8) S-enol, O. I.; Ryymin, E. M.; Viljava, T. R.; Krause, A. O. I. Effect of hydrogen sulphide on the hydrodeoxygenation of aromatic and aliphatic oxygenates on sulphided catalysts. J. Mol. Catal. A: Chem. 2007, 277 (1-2), 107–112. (9) Laurent, E.; Delmon, B. Influence of oxygen-, nitrogen-, and sulfur-containing compounds on the hydrodeoxygenation of phenols over sulfided CoMo/γ-A12O3 and NiMo/γ-A12O3 catalysts. Ind. Eng. Chem. Res. 1993, 32 (11), 2516–2524. (10) Yoshimura, Y.; Sato, T.; Shimada, H.; Matsubayashi, N.; Nishijima, A. Influences of oxygen-containing substances on deactivation of sulfided molybdate catalysts. Appl. Catal. 1991, 73 (1), 55–63. (11) Furimsky, E. Deactivation of molybdate catalyst during hydrodeoxygenation of tetrahydrofuran. Ind. Eng. Chem. Prod. Res. Dev. 1983, 22 (1), 34–38. (12) Elliott, D. C. Historical developments in hydroprocessing biooils. Energy Fuels 2007, 21 (3), 1792–1815. (13) Gutierrez, A.; Kaila, R. K.; Honkela, M. L.; Slioor, R.; Krause, A. O. I. Hydrodeoxygenation of guaiacol on noble metal catalysts. Catal. Today 2009, 147 (3-4), 239–246. (14) Fisk, C. A.; Morgan, T.; Ji, Y.; Crocker, M.; Crofcheck, C.; Lewis, S. A. Bio-oil upgrading over platinum catalysts using in situ generated hydrogen. Appl. Catal., A 2009, 358 (2), 150–156. (15) Liu, P.; Rodriguez, J. A.; Asakura, T.; Gomes, J.; Nakamura, K. Desulfurization reactions on Ni2P(001) and R-Mo2C(001) surfaces: Complex role of P and C sites. J. Phys. Chem. B 2005, 109, 4575–4583. (16) Liu, P.; Rodriguez, J. A.; Takahashi, Y.; Nakamura, K. Water-gas-shift reaction on a Ni2P(001) catalyst: Formation of oxyphosphides and highly active reaction sites. J. Catal. 2009, 262 (2), 294–303.

On the one hand, the metal site might interact with oxygen in H2O, inhibiting the reaction. A recent study has shown that the MoOx catalyst gives lower activity than the MoP catalyst in the HDO of 4-methylphenol, which is caused by the lower electron density of MoOx.39 As reported by Liu et al.,16 O atoms deposited on Ni2P(001) by the dissociation of water can interact with both P and Ni atoms but mainly with the P atoms. As a result, oxy-phosphide species are generated, in which an electron transfer from P to O occurs.16 Because of the ligand effect of P, P species may inhibit the combination of Ni with O, enhancing the oxidation resistance of Ni2P. This may account for the superior stability of Ni2P/SiO2 to that of Ni/SiO2. As show in Table 3, the deactivation degree of phosphide catalysts became less as the Ni/ Mo ratio increased. Thus, the Niδþ site may be more resistant to oxidation than the Moδþ site, which can be supported by the higher affinity of Mo with oxygen than that of Ni.10 On the other hand, similar to the formation of sulfate, P in phosphide might be oxidized by water to form phosphate.48 Phosphate, with much less activity, might cover the active sites on the phosphide catalysts.

4. CONCLUSIONS From the phosphate precursors with a nominal P/M (M = Ni and/or Mo) molar ratio of 1, the Ni2P/SiO2, MoP/SiO2, and NiMoP/SiO2 catalysts were prepared. In the NiMoP/SiO2 catalysts, the coexistence of Ni and Mo promoted the dispersion of phosphides and the NiMoP solid solution was formed. With the increase of the Ni/Mo molar ratio, H2 desorbed more easily from the surface of the phosphide catalysts. The metal sites in phosphides present both Lewis acidity and metallic property. Both metal sites (Niδþ and Moδþ) and PO-H groups on phosphides were active for the demethylation, hydrogenolysis, and hydrogenation. The metal sites, especially the Ni sites, were much more active than the PO-H groups. The activity and stability of the phosphide catalyst increased with the Ni/Mo molar ratio. However, no synergism existed between Ni and Mo. In comparison to the conventional NiMo/γ-Al2O3 catalyst, the Ni phosphide-containing catalysts, especially Ni2P/ SiO2, had higher activities and showed a promising application for the HDO. 862

dx.doi.org/10.1021/ef101258j |Energy Fuels 2011, 25, 854–863

Energy & Fuels

ARTICLE

(39) Whiffen, V. M. L.; Smith, K. J. Hydrodeoxygenation of 4-methylphenol over unsupported MoP, MoS2, and MoOx catalysts. Energy Fuels 2010, 24 (9), 4728–4737. (40) Chen, J. X.; Sun, L. M.; Wang, R. J.; Zhang, J. Y. Hydrodechlorination of chlorobenzene over Ni2P/SiO2 catalysts: Influence of Ni2P loading. Catal. Lett. 2009, 133 (3-4), 346–353. (41) Hadjiivanov, K.; Mihaylov, M.; Klissurski, D.; Stefanov, P.; Abadjieva, N.; Vassileva, E.; Mintchevz, L. Characterization of Ni/SiO2 catalysts prepared by successive deposition and reduction of Ni2þ ions. J. Catal. 1999, 185 (2), 314–323. (42) Fang, K. G.; Ren, J.; Sun, Y. H. Effect of nickel precursors on the performance of Ni/AlMCM-41 catalysts for n-dodecane hydroconversion. J. Mol. Catal. A: Chem. 2005, 229 (1-2), 51–58. (43) Viljava, T. R.; Komulainen, R. S.; Krause, A. O. I. Effect of H2S on the stability of CoMo/Al2O3 catalysts during hydrodeoxygenation. Catal. Today 2000, 60 (1-2), 83–92. (44) Laurent, E.; Delmon, B. Study of the hydrodeoxygenation of carbonyl, carboxylic and guaiacyl groups over sulfided CoMo/γ-A12O3 and NiMo/γ-A12O3 catalyst: II. Influence of water, ammonia and hydrogen sulfide. Appl. Catal., A 1994, 109 (1), 97–115. (45) Bredenberg, J. B.; Huuska, M.; R€aty, J.; Korpio, M. Hydrogenolysis and hydrocracking of the carbon-oxygen bond: I. Hydrocracking of some simple aromatic O-compounds. J. Catal. 1982, 77 (1), 242–247. (46) Laurent, E.; Delmon, B. Study of the hydrodeoxygenation of carbonyl, carboxylic and guaiacyl groups over sulfided CoMo/γ-A12O3 and NiMo/γ-A12O3 catalyst: II. Influence of water, ammonia and hydrogen sulfide. Appl. Catal., A 1994, 109 (1), 97–115. (47) Cairon, O.; Thomas, K.; Chambellan, A.; Chevreau, T. Acidcatalysed benzene hydroconversion using various zeolites: Br€onsted acidity, hydrogenation and side-reactions. Appl. Catal., A 2003, 238 (2), 167–183. (48) Laurent, E.; Delmon, B. Influence of water in the deactivation of a sulfided NiMo/γ-Al2O3 catalyst during hydrodeoxygenation. J. Catal. 1994, 146 (1), 281–291.

(17) Rodriguez, J. A.; Kim, J. Y.; Hanson, J. C.; Sawhill, S. J.; Bussell, M. E. Physical and chemical properties of MoP, Ni2P, and MoNiP hydrodesulfurization catalysts: Time-resolved X-ray diffraction, density functional, and hydrodesulfurization activity studies. J. Phys. Chem. B 2003, 107 (26), 6276–6285. (18) Oyama, S. T. Novel catalysts for advanced hydroprocessing: Transition metal phosphides. J. Catal. 2003, 216 (1-2), 343–352. (19) Stinner, C.; Prins, R.; Weber, Th. Binary and ternary transitionmetal phosphides as HDN catalysts. J. Catal. 2001, 202 (1), 187–194. (20) Sawhill, S. J.; Phillips, D. C.; Bussell, M. E. Thiophene hydrodesulfurization over supported nickel phosphide catalysts. J. Catal. 2003, 215 (2), 208–219. (21) Teng, Y.; Wang, A.; Li, X.; Xie, J.; Wang, Y.; Hu, Y. Preparation of high-performance MoP hydrodesulfurization catalysts via a sulfidationreduction procedure. J. Catal. 2009, 266 (2), 369–379. (22) Liu, X. G.; Chen, J. X.; Zhang, J. Y. Hydrodechlorination of chlorobenzene over silica-supported nickel phosphide catalysts. Ind. Eng. Chem. Res. 2008, 47 (15), 5362–5368. (23) Duan, X. P.; Teng, Y.; Wang, A. J.; Kogan, V. M.; Li, X.; Wang, Y. Role of sulfur in hydrotreating catalysis over nickel phosphide. J. Catal. 2009, 261 (2), 232–240. (24) Oyama, S. T.; Wang, X.; Lee, Y. K.; Chun, W. J. Active phase of Ni2P/SiO2 in hydroprocessing reactions. J. Catal. 2004, 221 (2), 263–273. (25) Oyama, S. T.; Wang, X.; Lee, Y. K.; Bando, K.; Requejo, F. G. Effect of phosphorus content in nickel phosphide catalysts studied by XAFS and other techniques. J. Catal. 2002, 210 (1), 207–217. (26) Zuzaniuk, V.; Prins, R. Synthesis and characterization of silicasupported transition-metal phosphides as HDN catalysts. J. Catal. 2003, 219 (1), 85–96. (27) Clause, O.; Bonneviot, L.; Che, M. Effect of the preparation method on the thermal stability of silica-supported nickel oxide as studied by EXAFS and TPR techniques. J. Catal. 1992, 138 (1), 195–205. (28) Arnoldy, P.; De Jonge, J. C. M.; Moulijn, J. A. Temperatureprogrammed reduction of MoO3 and MoO2. J. Phys. Chem. 1985, 89 (21), 4517–4526. (29) Rodriguez, J. A.; Hanson, J. C.; Frenkel, A. I.; Kim, J. Y.; Perez, M. Experimental and theoretical studies on the reaction of H2 with NiO: Role of O vacancies and mechanism for oxide reduction. J. Am. Chem. Soc. 2002, 124 (2), 346–354. (30) Klimova, T.; Calderon, M.; Ramírez, J. Ni and Mo interaction with Al-containing MCM-41 support and its effect on the catalytic behavior in DBT hydrodesulfurization. Appl. Catal., A 2003, 240 (1-2), 29–40. (31) Lizama, L.; Klimova, T. Highly active deep HDS catalysts prepared using Mo and W heteropolyacids supported on SBA-15. Appl. Catal., B 2008, 82 (3-4), 139–150. (32) Wang, R.; Smith, K. J. Hydrodesulfurization of 4,6-dimethyldibenzothiophene over high surface area metal phosphides. Appl. Catal., A 2009, 361 (1-2), 18–25. (33) Ma, D.; Xiao, T. C.; Xie, S. H.; Zhou, W. Z.; Gonzalez-Cortes, S. L.; Green, M. L. H. Synthesis and structure of bimetallic nickel molybdenum phosphide solid solutions. Chem. Mater. 2004, 16 (14), 2697–2699. (34) Ueckert, T.; Lamber, R.; Jaeger, N. I.; Schubert, U. Strong metal support interactions in a Ni/SiO2 catalyst prepared via sol-gel synthesis. Appl. Catal., A 1997, 155 (1), 75–85. (35) Lee, Y. K.; Oyama, S. T. Bifunctional nature of a SiO2supported Ni2P catalyst for hydrotreating: EXAFS and FTIR studies. J. Catal. 2006, 239 (2), 376–389. (36) Lee, Y. K.; Shu, Y. Y.; Oyama, S. T. Active phase of a nickel phosphide (Ni2P) catalyst supported on KUSY zeolite for the hydrodesulfurization of 4,6-DMDBT. Appl. Catal., A 2007, 322, 191–204. (37) Sawhill, S. J.; Layman, K. A.; Van Wyk, D. R.; Engelhard, M. H.; Wang, C. M.; Bussell, M. E. Thiophene hydrodesulfurization over nickel phosphide catalysts: Effect of the precursor composition and support. J. Catal. 2005, 231 (2), 300–313. (38) Phillips, D. C.; Sawhill, S. J.; Self, R.; Bussell, M. E. Synthesis, characterization, and hydrodesulfurization properties of silica-supported molybdenum phosphide catalysts. J. Catal. 2002, 207 (2), 266–273. 863

dx.doi.org/10.1021/ef101258j |Energy Fuels 2011, 25, 854–863