Yongde Ma, Rengan Liang, Wenquan Wu, Jiayin Zhang, Yanning Cao, Kuan Huang, Lilong Jiang

National Engineering Research Center of Chemical Fertilizer Catalyst (NERC-CFC), Fuzhou University, Fuzhou 350002, China

Keywords:Petroleum Hydrogenation Catalyst Suspended bed Cracking Mixed oxide

A B S T R A C T Developing catalysts with not only hydrogenation activity but also cracking activity is very important for the advancement of suspended-bed hydrocracking technology. Within this respect, MoS2/SiO2-Al2O3 bifunctional catalyst is a kind of typical catalysts with both hydrogenation and cracking activity.Herein, a series of Zr-doped SiO2-Al2O3 mixed oxides were synthesized by a sol-gel coupled with hydrothermal method. The synthesized mixed oxides were characterized for chemical structures and acidic properties. It is found that doping SiO2-Al2O3 with Zr atoms significantly increases the numbers of acidic sites. The Zr-doped SiO2-Al2O3mixed oxides were then combined with dispersed MoS2, which was in-situ produced from oil-soluble Mo precursors, to fabricate a novel kind of bifunctional catalysts for suspended-bed hydrocracking of heavy oils. Owing to the significantly increased numbers of acidic sites in Zr-doped SiO2-Al2O3 mixed oxides, corresponding bifunctional catalysts demonstrate much enhanced activity for suspended-bed hydrocracking of heavy oils in relative to MoS2/SiO2-Al2O3 bifunctional catalysts.

1. Introduction

Given the urgent energy and environmental issue confronted by human beings, the demand for light and clean fuel oils across the world has been constantly increasing. However, conventional oils account for less than 30%of the total oil reserves,and most of them are heavy oils with poor quality[1-4].The conversion of heavy oils into high-quality light oils through hydrocracking technologies is one of the most important direction of petroleum refining industry.Generally,the hydrocracking technologies for heavy oil processing can be divided into four categories: fixed-bed hydrocracking [5],ebullated-bed hydrocracking[6,7]and suspended-bed hydrocracking[8,9].Among these technologies,suspended-bed hydrocracking is quite promising because it overcomes the high pressure drop and large diffusion barrier associated with other hydrocracking technologies, and is also flexible for various heavy oils, simple in process sheets, and cheap of facility investments.

Suitable catalysts are required for suspended-bed hydrocracking to suppress the formation of cokes during the conversion process. Within this respect, nanosized transition metal sulfides are effective catalysts for suspended-bed hydrocracking[10-12].They can be producedex-situ[1,13] orin-situ[14-16] from metal precursors and sulfurization reagents,and then dispersed in heavy oils to impose negligible diffusion barrier for large hydrocarbon molecules to access the active sites. However, metal sulfides are only active for hydrogenation reactions,but inert for cracking reactions.Therefore, suspended-bed hydrocracking is normally operated at extremely high temperatures (＞400 °C) [8], at which thermal cracking reactions can take place. Such high temperatures are unfavorable for the suppression of coke formation. To this end,the development of catalysts with not only hydrogenation activity but also cracking activity is very important.

Some researchers combined metal sulfides with metal oxides to fabricate ‘‘bifunctional catalysts” for suspended-bed hydrocracking. In these bifunctional catalysts, metal oxides with acidic sites can promote the isomerization of aromatic rings, thus showing cracking activity.Usually,there exist synergic effect between metal sulfides and metal oxides.For example,Sanchezet al.[17]designed a class of bifunctional catalysts consisting of dispersed MoS2and amorphous SiO2-Al2O3. The dispersed MoS2wasin-situproduced from oil-soluble Mo precursors. It was found that the addition of SiO2-Al2O3improved the structural and morphological properties of MoS2. In addition, Bdwiet al. [18] and AI-Attaset al. [16,19]investigated the synergic effect between dispersed metal sulfides and commercial W-Ni/Al2O3-SiO2by comparing the product distributions for the hydrocracking of vacuum gas oil. It was found that both the conversions of feed oil and selectivities of valuable products were increased owing to the introducing of acidic sites.

In various bifunctional catalysts developed for suspended-bed hydrocracking,SiO2-Al2O3is the most frequently used mixed metal oxide.Br?nsted acidic sites are formed at the Si-O-Al bridge bonds of SiO2-Al2O3,which are crucial for the promotion of hydrocracking reactions [17,18]. Recently, researchers further found that doping SiO2-Al2O3with P or F could improve the acidity - especially Br?nsted acidity of SiO2-Al2O3[20].ZrO2is also a well-known acidic sites carrier whose hydrothermal stability is high[21-23].Previous studies have demonstrated that the coordinatively unsaturated Zr atoms and neighbouring lattice O atoms in ZrO2can activate C-H bonds[24].Moreover,sulfated ZrO2is a class of superacidic catalyst with extensive application in the industry [25,26]. Given these facts,we envisioned that doping SiO2-Al2O3with Zr atoms can also improve the acidity of SiO2-Al2O3. To the best of our knowledge,there are few reports on the Zr doping of SiO2-Al2O3to improve the acidity of SiO2-Al2O3. The application of Zr-doped SiO2-Al2O3in hydrocracking of heavy oils is also very scarce.

In the present work,a series of Zr-doped SiO2-Al2O3mixed oxides were synthesized, and characterized for chemical structures and acidic properties.The mixed oxides were then combined with dispersed MoS2,which wasin-situproduced from oil-soluble Mo precursors, to fabricate a novel kind of bifunctional catalysts for suspended-bed hydrocracking of heavy oils. The performance of bifunctional catalysts was systematically evaluated in a stirredbatch reactor using a mixture of phenanthrene and decalin as the model system. Phenanthrene undergoes hydrogenation reactions in the presence of metal sulfide catalysts,and decalin undergoes isomerization and cracking reactions in the presence of acid catalysts.The transformations of phenanthrene and decalin can reflect the hydrogenation and cracking processes respectively [17,20]. The bifunctional catalysts collected after reaction were also characterized to examine the state of MoS2and Zr-doped SiO2-Al2O3mixed oxides.

2. Experimental

2.1. Chemicals

Phenanthrene (95%, by mass), tetraethyl orthosilicate (TEOS,98%, by mass) andn-tetradecane (99%, by mass) were supplied by Aladdin. Decalin (99%, by mass) was supplied by Macklin.ZrOCl2·8H2O (99%, by mass), AlCl3·6H2O (99%, by mass), ammonia water(25%-28%,by mass)and anhydrous ethanol(99.5%,by mass)were supplied by Sinopharm.Molybdenum(Ⅳ)2-ethyl hexanoate(15%of Mo,by mass)was supplied by Strem.Sublimed sulfur(99%,by mass) was supplied by Xilong.

2.2. Synthesis

Zr-doped SiO2-Al2O3mixed oxides were synthesized by a solgel coupled with hydrothermal method. In a typical run, TEOS(0.045 mol) was dissolved in a mixture of ethanol/water (1/1 (v/v), 500 ml); the solution was vigorously stirred to make the TEOS be hydrolyzed; ZrOCl2·8H2O and AlCl3·6H2O were added to the solution to make the molar ratio of Si/M(M=Al+Zr)be 5;ammonia water was added to the solution to make the pH be ～4.0; the solution was loaded in an autoclave and treated at 100 °C for 20 h; the solid product was separated from the reaction system by centrifugation, and washed thoroughly by ethanol-water mixture; the product was dried at 120 °C, and finally calcination at 500 °C for 4 h in air. The samples were named as SiO2-ZrO2/Al2O3-x,in whichxrepresents the molar ratio of Zr/Al used for synthesis. Whenx= 0 (no Zr doping), the sample is SiO2-Al2O3. Whenx= ∞(fully Zr doping), the sample is SiO2-ZrO2.

Bifunctional catalysts consisting of dispersed MoS2and mixed oxides werein-situproduced when evaluating the catalysts (see Section 2.4).

2.3. Characterizations

A Panalytical X’Pert powder diffractometer was used to measure the X-ray diffraction (XRD) patterns with CoKα radiation at 40 kV and 40 mA. A Micromeritics ASAP 2020 analyzer was used to measure the N2adsorption-desorption isotherms at -196 °C.The surface areas were calculated according to the Barrett-Emmet-Teller (BET) equation using the adsorption data at 0.05-0.35 relative pressure. The pore volumes and pore size distributions were calculated according to the Barrett-Joyner-Halenda(BJH) model using the desorption data. A JEOL JEM-2020 electron microscope was used to take the field emission transmission electron microscope(TEM)images at an acceleration voltage of 200 kV.A PerkinElmer OPTIMA 8000 spectrometer was used to measure the contents of metal atoms through inductively coupled plasma(ICP) analysis. A Nicolet 6700 spectrometer was used to measure the Fourier transform infrared (FTIR) spectra with KBr tablet. A Bruker Avance III 500 M spectrometer was used to measure the solid-state29Si MAS NMR spectra with TMS as the external standard.A Thermo Fisher EscaLab 250Xi spectrometer with AlKα radiation was used to measure the X-ray photoelectron spectroscopy(XPS) spectra, and the C 1s peak at 284.8 eV was used to calibrate the binding energies.

A Micromeritics AutoChem 2920 analyzer was used to perform the temperature-programmed desorption of ammonia (NH3-TPD)and isopropylamine (IPA-TPD) experiments. In a typical run, a specific amount of sample was fixed in a quartz tube with U shape;the sample was pretreated at 500°C for 60 min under flowing He;after cooling down to 100°C,the He flow was switched to a mixed gas flow with 3%(by volumn)of NH3or IPA balanced in He,and the mixed gas flow was kept for 60 min;the mixed flow was switched to a He flow without changing the temperature, and the He flow was kept for 90 min; the sample was finally heated to 700 °C at 10 °C·min-1under flowing He, and the contents of NH3(m/e= 16) and propylene (m/e= 41) in outlet gas were measured by a Hiden HPR-20 mass spectrometer.

A Setram Setsys Evolution thermal analyzer was used to perform the temperature programmed oxidation (TPO) analysis. Each sample was first treated by flowing N2at 150 °C for 30 min to remove any adsorbed water. The sample was then cooled down to 100°C,and treated by flowing dry air(30 ml·min-1).Simultaneously, the sample was heated from 100 to 850 °C at 10 °C·min-1,and the signals of H2O (m/e= 18) and CO2(m/e= 44) in tail gas were detected online a Hiden HPR-20 mass spectrometer.

2.4. Catalytic hydrocracking

Phenanthrene (4 g), decalin (36 g), molybdenum (IV) 2-ethyl hexanoate (0.16 g), sulfur (0.1 g) and mixed oxide (1.6 g) were loaded in a batch-type high-pressure reactor. The reactor was purged with H2to replace the air,sealed tightly in case of leaking,and fed with H2to a specific pressure. The mixture was stirred at 600 rpm, and heated at 3 °C·min-1. During heating, the mixture was kept at 230 for 30 min, and at 320 °C for another 30 min, to make sure that the Mo precursor was fully sulfurized.The mixture was finally treated at 380 °C and 10 MPa for 4 h. After cooling down,a Shimadzu 2010-Plus GCMS was used to analyze the products withn-tetradecane as the internal standard. The conversions of phenanthrene (XP) and decalin (XD) were calculated by the following equations:

whereAP,ADandASare the areas of phenanthrene, decalin andntetradecane peaks respectively;the subscripts 0 andtrefer to reaction at beginning andthours respectively.

According to the literature [17,27], the hydrocracking products of phenanthrene and decalin can be classified into seven categories: C10-C12ARO, C10ISO, C10RO, C7- C9, C14HYD, C14RO,C14+ (see Fig. S1). It should be pointed out that the most valuable products for the conversion of phenanthrene are C14HYD and C14RO, which are important for the evaluation of the hydrogenation activity of catalysts.The most valuable products for the conversion of decalin are C10ISO,C10RO and C7-C9,which are important for the evaluation of the cracking activity of catalysts. The selectivity(Si) and yield (Yi) for each category of products were calculated by the following equations:

3. Results and Discussion

3.1. Characterization results

Four mixed oxides with different Zr doping amounts were synthesized in this work. The metal contents of synthesized mixed oxides were characterized by ICP analysis. Table 1 lists the molar ratios of Si/(Al + Zr) and Zr/Al in synthesized mixed oxides, which agree well with the theoretical values according to the amounts of chemicals used for synthesis. The crystalline structures of synthesized mixed oxides were characterized by XRD (see Fig. S2). It can be seen that the XRD pattern of SiO2-Al2O3shows only one broad peak at ～26°, suggesting the predominantly amorphous structure of SiO2-Al2O3. Doping SiO2-Al2O3with Zr causes the peak shifting slightly to higher angles,and also becoming slightly weaker.Therefore,doping with Zr partially destroys the local crystalline phase of SiO2-Al2O3,which also suggests that Zr atoms are well dispersed in SiO2-Al2O3. The morphology of synthesized mixed oxides was characterized by TEM, as shown in Fig. 1. As can be seen, lattice fringes cannot be found in the TEM images,and the predominantly amorphous structures of synthesized mixed oxides were also confirmed by the electron diffraction patterns. In addition, there are abundant wormlike nanopores with the size of microscale and mesoscale in synthesized mixed oxides.

Table 1 Porosity parameters of synthesized mixed oxides

The detailed porous structures of synthesized mixed oxides were characterized by low-temperature N2adsorption, as show in Fig. 2A.It can be seen that the N2isotherm of SiO2-Al2O3shows type-III profile with H3-type hysteresis loop in the relative pressure range of 0.7-1.0. This indicated the predominantly meso-macroporous structure of SiO2-Al2O3, which is mainly formed by the random aggregation of plate-like nanoparticles.Doping SiO2-Al2O3with minor Zr atoms (i.e., SiO2-ZrO2/Al2O3-3/7)does not change the type of N2isotherm, but slightly increases the N2uptakes. The increased N2uptakes can be attributed to the substitution of Al atoms by Zr atoms,which changes the bonding structures of mixed oxides and thus creates more nanoscale apertures. However, doping SiO2-Al2O3with more Zr atoms(i.e., SiO2-ZrO2/Al2O3-7/3 and SiO2-ZrO2) changes the N2isotherm from type-III to type-IV, and the hysteresis loop from H3-type to H1-type. This indicates that the macroporosity of SiO2-Al2O3is destroyed while the mesoporosity is well preserved, further suggesting the tighter aggregation of plate-like nanoparticles in mixed oxides after Zr doping. Correspondingly, the N2uptakes in low to medium relative pressure range are increased, while those in high relative pressure range are decreased.

The evolution of meso-macroporosity for synthesized mixed oxides can also be evident from the pore size distributions, as shown in Fig. 2B. The covering range of pore size distributions becomes narrower with the increase of Zr doping amounts. For example, the pore size distribution of SiO2-Al2O3covers a wide range of 0-120 nm, with the largest dV/dlgDpore volume appearing at ～45 nm. However, the pore size distribution of SiO2-ZrO2cover a narrow range of 0-6 nm,with the largest dV/dlgDpore volume appearing at ～4 nm.Table 1 also lists the porosity parameters of synthesized mixed oxides.As can be seen,the BET surface areas increase with the increase of Zr doping amounts,the total pore volumes first slightly increase but then decrease with the increase of Zr doping amounts, and the average pore size decreases with the increase of Zr doping amounts.

The bonding structures of synthesized mixed oxides were characterized by FTIR and solid-state29Si MAS NMR.Fig.3A shows the FTIR spectra of synthesized mixed oxides, which are quite similar in profiles. There is a broad and strong peak at 1450-850 cm-1,which can be ascribed to the overlapping of peaks associated with the asymmetric stretching vibrations Si-O-Si, Si-OH and Si-O-M (M = Al and Zr) bonds. The other two weak peaks at～810 and ～470 cm-1can be ascribed to the symmetric stretching and rocking vibrations of Si-O-Si bonds.These peaks shift slightly to lower wavenumbers with the increase of Zr doping amounts,due to the change of electron environment. The FTIR peak at 1450-850 cm-1can be deconvoluted using the Gauss function,and the dispersion of M atoms can be calculated using the following equation [28,29]:

whereDSi-O-Mis the dispersion of M atoms;SSi-O-MandSSi-O-Siare the areas of deconvoluted peaks at ～940 and ～1210 cm-1respectively;nSi/nMis the molar ratio of Si/M. The deconvolution results are shown in Fig. 3B, and the calculated dispersions are presented in Fig. 3C. It is found that the deconvoluted peak associated with Si-O-M bonds at ～940 cm-1is very weak for SiO2-Al2O3. Correspondingly, the dispersion of M atoms in SiO2-Al2O3is also very low.However, doping SiO2-Al2O3with Zr atoms causes the increase of the intensity for this deconvoluted peak.As a result,the dispersions of M atoms increase with the increase of Zr doping amounts. This result indicates that Zr atoms prefer to form Si-O-M bonds in the mixed oxides in relative to Al atoms.

Fig. 1. TEM images of SiO2-Al2O3 (A-D) and SiO2-ZrO2/Al2O3-7/3 (E-H).

Fig. 2. N2 adsorption-desorption isotherms at -196 °C (A) and pore size distributions (B) of synthesized mixed oxides.

Fig. 4 shows the solid-state29Si MAS NMR spectra of synthesized mixed oxides, which can also be deconvoluted using the Gaussian function [30-33]. The deconvoluted peaks at around-90, -100 and -110 can be ascribed to Si atoms in Q2, Q3and Q4species respectively. Herein, Q2refers to (Si-O)2-Si-(O-H)2or (Si-O)2-Si-(O-M)2, Q3refers to (Si-O)3-Si-O-H or (Si-O)3-Si-O-M,and Q4refers to(Si-O)4-Si.Table 2 summarizes the area ratios of deconvoluted peaks in NMR spectra for synthesized mixed oxides. It is found that the value of Q2/Q4is much lower than that of Q3/Q4for SiO2-Al2O3,suggesting that Al atoms mainly exist in Q3form. Doping SiO2-Al2O3with Zr atoms increases the values of Q2/Q4and Q3/Q4, and the increase of Q2/Q4is much more significant than that of Q3/Q4. Therefore, the doped Zr atoms mainly exist in Q2form. The values of (Q2+ Q3)/Q4increase with the increase of Zr doping amounts,agreeing with the effect of Zr doping amounts on the dispersions of M atoms shown in Fig.3C.This result further validates that Zr atoms prefer to form Si-O-M bonds in the mixed oxides in relative to Al atoms.

The acidic properties of synthesized mixed oxides were characterized by NH3-TPD and IPA-TPD experiments. Fig. 5 depicts the NH3-TPD curves of synthesized mixed oxides,which can be deconvoluted into three peaks based on the desorption temperature.The deconvoluted peaks at 100-250°C,250-450°C and 450-700°C can be ascribed to weak, medium and strong acidic sites respectively.By integrating the deconvoluted peaks,the numbers of acidic sites can thus be calculated, and results are summarized in Table 3. As can be seen, doping SiO2-Al2O3with Zr atoms significantly increases the numbers of weak and medium acidic sites, but decreases the number of strong acidic sties.Overall,the total numbers of acidic sites increase significantly with the increase of Zr doping amounts, which is related to the favorable formation of Si-O-M bonds for Zr atoms.

Fig. 6 shows the IPA-TPD curves of synthesized mixed oxides,which can be used for the differentiation of Br?nsted acidic sites from Lewis acidic sites, since only Br?nsted acidic sites can catalyze the conversion of IPA to propylene and NH3via the Hofmann elimination [34-40]. It can be seen that the MS peaks of NH3appear at slightly higher temperatures than those of propylene,as a result of the re-adsorption of NH3. Doping SiO2-Al2O3with Zr atoms causes the MS peaks of propylene and NH3slightly shifting to lower temperatures, suggesting the slightly increased strength of acidic sites. This is consistent with the increased numbers of the total numbers of acidic sties after Zr doping, as illustrated by NH3-TPD experiments. By integrating the MS peaks of propylene, the numbers of Br?nsted acidic sites can thus be calculated. The numbers of Lewis acidic sites can then be calculated by subtracting the numbers of Br?nsted acidic sites from the numbers of total acidic sites. The calculated results are summarized in Table 4.It can be seen that doping SiO2-Al2O3with Zr atoms causes the simultaneous increase of the numbers of Br?nsted and Lewis acidic sites. However, the B/L ratios remain almost unchanged before and after Zr doping.

Table 4 Numbers of Bronsted and Lewis acidic sites in synthesized mixed oxides

Fig. 6. IPA-TPD curves of synthesized mixed oxides with MS signals for propylene (A) (m/e = 41) and NH3 (B) (m/e = 16).

Table 3 Numbers of acidic sites in synthesized mixed oxides

Fig. 3. FTIR spectra of synthesized mixed oxides (A); deconvolutions of FTIR spectra in wavenumber range of 1450-850 cm-1 (B); dispersions of M atoms in synthesized mixed oxide (C).

Fig. 4. Solid-state 29Si MAS NMR spectra of synthesized mixed oxides.

Fig. 5. NH3-TPD curves of synthesized mixed oxides.

Table 2 Area ratios of deconvoluted peaks in NMR spectra for synthesized mixed oxides

3.2. Hydrocracking activity

The activity of MoS2/mixed oxides bifunctional catalysts was investigated by determining the conversions of model reactants,selectivities and yields of hydrocracking products.For comparison,the activity of bare MoS2which wasin-situproduced from molybdenum (IV) 2-ethyl hexanoate without the addition of mixed oxides was also determined. Fig. 7A shows the conversions of phenanthrene and decalin for bare MoS2and MoS2/mixed oxides bifunctional catalysts. As can be seen, the conversion of decalin for bare MoS2is negligible, because it only has hydrogenation activity.The conversions of decalin for MoS2/mixed oxides bifunctional catalysts are much higher than that for bare MoS2, owing to the addition of mixed oxides with acidic sites.In addition,the conversions of decalin for MoS2/mixed oxides bifunctional catalysts increase significantly with the increase of Zr doping amounts.The highest value is achieved for MoS2/SiO2-ZrO2, in which the Al atoms are fully substituted by Zr atoms. This is undoubtfully a result of the increased numbers of acidic sites - especially Br?nsted acidic sites for Zr-doped SiO2-Al2O3mixed oxides. It is also found that the conversions of phenanthrene for MoS2/mixed oxides bifunctional catalysts increase slightly with the increase of Zr doping amounts. Therefore, doping SiO2-Al2O3with Zr atoms can enhance not only the cracking activity but also the hydrogenation activity of bifunctional catalysts.

Fig.7B and C show the selectivities and yields of hydrocracking products for bare MoS2and MoS2/mixed oxides bifunctional catalysts. As can be seen, the hydrocracking products for bare MoS2are mostly C14HYD with negligible amounts of other products.However,all the other products are observed for MoS2/mixed oxides bifunctional catalysts owing to the addition of mixed oxides with cracking activity.Correspondingly,the selectivities and yields of C14HYD for MoS2/mixed oxides bifunctional catalysts are significantly reduced in relative to those for bare MoS2.The major products for MoS2/mixed oxides bifunctional catalysts are C10ISO and C10RO. It is noted that the selectivities of all the products for MoS2/mixed oxides bifunctional catalysts do not change too much with the increase of Zr doping amounts.The yields of all the products for MoS2/mixed oxides bifunctional catalysts gradually increase with the increase of Zr doping amounts, as a result of the increase of the conversions of phenanthrene and decalin.Therefore, doping SiO2-Al2O3with Zr atoms does not change the product distributions of hydrocracking reactions for bifunctional catalysts,but only improve the conversions of reactants and yields of products.

The conceptual mechanism of MoS2/mixed oxides bifunctional catalysts for the hydrocracking of heavy oils has been well established in the literature [18]. Basically, molecules with large size first undergo thermal cracking reactions or acid-catalyzed cracking reactions on the surface of mixed oxides. The cracked molecules with small size subsequently undergo hydrogenation reactions catalyzed by dispersed MoS2.Some cracked molecules with small size are not completely hydrogenated, and undergo combination reactions to produce molecules with large size again, or even form cokes on the surface of dispersed MoS2and mixed oxides.This will cover the active sites of catalysts, and induce the deactivation of catalysts. The formation of cokes was confirmed by TPO analysis(see Fig. S3), which shows that H2O and CO2were detected in the tail gas of oxidation.In addition,it is found that there are more cokes formed on MoS2/SiO2-ZrO2/Al2O3-7/3 than on MoS2/SiO2-Al2O3, which is probably a result of the enhanced acidity of ZrO2/Al2O3-7/3 in relative to that of MoS2/SiO2-Al2O3due to the doping of Zr atoms.

Fig.7. Conversions of model reactants(A),selectivities(B)and yields(C)of hydrocracking products for bare MoS2 and MoS2/mixed oxides bifunctional catalysts:1,MoS2;2,MoS2/SiO2-Al2O3; 3, MoS2/SiO2-ZrO2/Al2O3-3/7; 4, MoS2/SiO2-ZrO2/Al2O3-7/3; 5, MoS2/SiO2-ZrO2.

To confirmthatMoS2phase wasin-situformed in the bifunctional catalysts,the catalysts were collected from the reaction system by washing the solid residuals thoroughly with toluene.The Mo 3d and S 2p spectra of bifunctional catalysts were then measured,as shown in Fig.8.The Mo 3d spectra can be deconvoluted into three doublets with the Mo 3d5/2peaks appearing at 228.9, 229.9 and 232.3 eV respectively.The first Mo 3d5/2peak at 228.9 eV is associated with the Mo4+of MoS2,while the other two Mo 3d5/2peaks at 229.9 and 232.5 eV are associated with the Mo5+of MoOxSyand Mo6+of MoO3respectively.The S 2p spectra can be deconvoluted into two doublets with the S 2p3/2peaks appearing at 162.0 and 163.2 eV respectively. The first S 2p3/2peak at 162.0 eV is associated with the S2-of MoS2,while the other one S 2p3/2peak at 163.2 eV is associated with the S2-of MoOxSy.The percentages of the Mo4+of MoS2account for 75.8%-85.2% (atom fraction) of the total Mo element,and the percentages of the S2-of MoS2account for 90.2%-95.0%(atom fraction)of the total S element in bifunctional catalysts(see Table S1).Therefore,most of the Mo element exists as MoS2phase in bifunctional catalysts.The Mo 3d and S 2p spectra of bifunctional catalysts are similar to those of bare MoS2(see Fig.S4),which wasinsituproduced from molybdenum(IV)2-ethyl hexanoate without the addition of mixed oxides.

Fig. 8. Mo 3d (A) and S 2p (B) spectra of MoS2/mixed oxides bifunctional catalysts.

The porosity structures and acidic properties of MoS2/mixed oxides bifunctional catalysts were also characterized by low-temperature N2adsorption, NH3-TPD and IPA-TPD experiments respectively (see Figs. S5-S7 and Tables S2-S4). It is found that the porosity parameters and acidic sites numbers of bifunctional catalysts are lower than those of mixed oxides.This is understandable because some mixed oxides are covered by thein-situformed MoS2phase. However, the variation trends of the porosity parameters and acidic sites numbers with Zr doping amounts for bifunctional catalysts are the same as those for mixed oxides.The catalytic activity of mechanically mixed MoS2/SiO2-Al2O3and MoS2/ZrO2(denoted as MoS2/SiO2-Al2O3+ ZrO2) was also investigated, and compared with that of MoS2/SiO2-ZrO2/Al2O3-7/3 (see Fig. S8). It is found that the catalytic activity of MoS2/SiO2-Al2O3+ ZrO2is much lower than that of MoS2/SiO2-ZrO2/Al2O3-7/3. Therefore, the enhanced activity of bifunctional catalysts origins from the doping of Zr atoms, while not the simple mixing of ZrO2.

The recycling performance of MoS2/mixed oxides bifunctional catalysts is very important for their long-term use in the industry.Therefore, the used MoS2/SiO2-ZrO2was washed thoroughly with toluene, and then reused for the slurry-phase hydrocracking of model reactants. The recycling experiments were performed for two times (see Fig. S9). It is found the catalytic activity of MoS2/SiO2-ZrO2partially decreases after recycling. This may be a result of the deposition of heavy products (e.g., cokes) on the catalyst,which cover the active sites of the catalyst. On the other hand,the transportation of used catalyst may lead to the loss of some active components. Actually, the partial decrease of activity is unavoidable for catalysts used in the industry. In practical case,some used catalysts are replaced with fresh ones, and then fed back to reactor, so that the activity of catalysts is kept.

4. Conclusions

In summary,a series of Zr-doped SiO2-Al2O3mixed oxides were synthesized by a sol-gel coupled with hydrothermal method. The Zr-doped SiO2-Al2O3mixed oxides were combined with dispersed MoS2to fabricate a novel kind of bifunctional catalysts for suspended-bed hydrocracking of heavy oils. It is found that Zr atoms are well dispersed in SiO2-Al2O3,and Zr atoms prefer to form Si-O-M bonds in the mixed oxides in relative to Al atoms. As a result, doping SiO2-Al2O3with Zr atoms significantly increases the numbers of acidic sites. Owing to this feature, the catalytic activity of bifunctional catalysts produced from Zr-doped SiO2-Al2O3mixed oxides are much enhanced in relative to that produced from SiO2-Al2O3.The enhanced activity of bifunctional catalysts origins from the doping of Zr atoms, while not the simple mixing of ZrO2.This work provides a facile method for the synthesis of catalysts with promising application in suspended-bed hydrocracking of heavy oils.

Acknowledgements

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This work was financially supported by the National Key Research & Development Program of China (2018YFA0209403),the National Natural Science Foundation of China (U1662108),and the Science and Technology Project of Fujian Province (FG-2016002).

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2021.03.015.