Can Mesoporous TiO2-Al2O3-Supported NiMoS OR

0 downloads 0 Views 537KB Size Report
of the origin of crude oil and the conditions of its cracking [4]. Most of the sulfur compounds can be easily hydrodesulfurized by using traditional catalysts. Alkyl.
‫ھﺬا اﻟﺒﺤﺚ ﻣﻨﺸﻮرﻓﻰ ﻣﺠﻠﺔ ﻋﻠﻤﯿﺔ ﻣﺘﺪاوﻟﺔ ﻣﺘﺨﺼﺼﺔ و ﺗﺼﺪر ﻋﻦ ھﯿﺌﺔ ﻋﻠﻤﯿﺔ و ﺑﮭﺎ ھﯿﺌﺔ‬ ‫ﺗﺤﺮﯾﺮ و ﻟﺠﻨﺔ ﺗﺤﻜﯿﻢ ﻣﺘﺨﺼﺼﺔ و اﻟﺒﺤﺚ ُأﺟﺮى ﻓﻰ ﻧﻄﺎق اﻟﺨﻄﺔ اﻟﺒﺤﺜﯿﺔ ﻟﻠﻘﺴﻢ‬

‫رﺋﯿﺲ اﻟﻘﺴﻢ‬

6 ‫ﺑﺤﺚ رﻗﻢ‬

‫رﺋﯿﺲ اﻟﻤﻌﻤﻞ‬

ISSN 0965-5441, Petroleum Chemistry, 2018, Vol. 58, No. 5, pp. 387–394. © Pleiades Publishing, Ltd., 2018. Published in Russian in Neftekhimiya, 2018, Vol. 58, No. 3.

Can Mesoporous TiO2-Al2O3-Supported NiMoS OR CoMoS Effectively Perform in Ultra-Deep Desulfurization of Gas Oil?1 Nasser H. Shalabya, *, Samia A. Hanafia, Salah A. Hassanb, and Mamdouh S. Elmelawya aEgyptian bDepartment

Petroleum Research Institute, Nasr City, Cairo, 11727 Egypt of Chemistry, Faculty of Science, Ain Shams University, Abbassia, Cairo, Egypt *e-mail: [email protected] Received February 28, 2017

Abstract—Waste aluminum foil was used for preparation of mesoporous TiO2-Al2O3 using starch as a textural modifier. The catalytic species, Mo and Ni or Co were loaded onto the mesoporous support, following incipient wetness sequential impregnation. To gain an insight into the pore dimensions effect, Ni and Mo species with the same mass ratio were loaded onto the TiO2-Al2O3, prepared from analytical grade chemicals without templating. TPR spectra, TEM images and BET analysis showed how the promoter (Ni or Co), TiO2 and the template (starch) affect the ease of reduction of Mo species, the morphology of the active MoS2 phase and the pore dimensions of the catalysts. The catalysts were employed in hydro-desulfurization process of gas oil using a fixed bed down flow microreactor at varying operating conditions, viz., temperature (320–400°C), Liquid hourly space velocity (0.5–4 h–1), H2/oil ratio of 450 v/v, and 6 MPa operating pressure. The results showed that the promotion effect prevails over the textural effect, where Ni promoted catalyst (with lower surface parameters) exhibits higher activity than Co promoted one. The dual layer catalytic bed system achieved the sulfur level less than 10 ppm. Keywords: hydrodesulfurization, titania alumina support, promoter effect, dual catalytic bed reactor DOI: 10.1134/S0965544118050158

INTRODUCTION In recent decades clean fuels have attracted more attention of researchers worldwide. Sulfur is one of the undesirable contaminants of gas oil which causes the corrosion of machines and SOx emissions in the automobile exhaust gas, leading to air pollution and many diseases to human, animals and plants [1]. Ultra-deep hydrodesulfurization (HDS) has become a more important target for environmental catalysis studies [2]. Generally, the efficiency of catalytic desulfurization depends largely on the catalyst design as well as the nature of sulfur compounds. The main sulfur compounds in diesel oils are thiols, sulfides, thiophene and its derivatives, thiophenols, benzothiophene and dibenzothiophene. These compounds may result during the cracking of long alkyl chain thiophenes and/or the addition of H2S to olefins, followed by cyclization [3]. It is known that the proportion of these compounds depends on the nature of the origin of crude oil and the conditions of its cracking [4]. Most of the sulfur compounds can be easily hydrodesulfurized by using traditional catalysts. Alkyl substituents of dibenzothiophene, such as 4-methyl 1 The article is published in the original.

and 4,6-dimethyl, seem to be stubborn compounds in hydrodesulfurization, as the alkyl groups shield the sulfur atom and hinder its catalytic removal. In addition, the crude oil quality continues to decrease with the more involvement of heavier crude oil, due to the higher sulfur, nitrogen and aromatic contents [5]. The challenges for gas oil hydrodesulfurization are to abide the new EPA Tier-II regulations for sulfur and aromatic contents, without affecting the octane number of produced gasoline in a facile manner [6]. Accordingly, the applied techniques and the employed catalysts need to be developed and to become properly applicable for such type of chemical reactions of the feedstock. The employed catalysts can be developed through improving their surface characteristics (surface area and pore dimensions) to affect their catalytic functionality or by designing composite catalytic systems, involving creation of new active sites, e.g., via synergistic-linked metallic species [7]. On the other hand, the applied catalytic technologies can be developed by using modified designs for the catalytic convertors to achieve the optimum catalytic performance. Also, convertors with successive catalytic beds containing a variety of catalysts with different catalytic characteristics can be used.

387

388

NASSER H. SHALABY et al.

In the present work, for getting promising more active catalysts, a modified support was suggested by combining alumina, from a cheap source, with another component, such as titania. Another aim was also attempted, namely, searching for a catalyst material of larger pore dimensions, compatible with the bulky aromatic molecules of the gas oil. Thus, the main objective of this work was focused on the study of the behavior of NiMoS and CoMoS supported on mesoporous TiO2–Al2O3, a modified binary oxide system, in the HDS of straight-run gas oil (SRGO), through using single and dual catalytic beds. EXPERIMENTAL Preparation of Titania-Alumina Support (TiO2(15%)–Al2O3) Preparation of AlCl3 from waste aluminium foils. Waste aluminium foils (used in food preservation at restaurants) were collected and washed throughly to remove the adheared fats and hogwash. The clean foils were cut into smal pices and dissolved in aqua ragia. The resulted solution was filtered and then evaporated under vaccum to obtain AlCl3. The obtained salt was dissolved in demineralized water and evaporated again to obtain pure recrystalized AlCl3 · 6H2O. Preparation of binary oxides support. The binary oxide support was prepared by the co-precipitation method from aqueous solutions of titanium chloride (Sigma-Aldrich) and aluminum chloride (as prepared). The as-prepared AlCl3 · 6H2O and TiCl4 were dissolved in a minimum amount of demineralized water with stirring at 60°C to obtain a clear mixed solution. After cooling to ambient temperature, starch (Sigma-Aldrich) dispersed in demineralized water was added to the mixed solution in a molar ratio of C6H10O5 : (Al + Ti) = 2.5, with agitation [8]. Aqueous NH4OH solution as a precipitating agent was added portion wise with continuous agitation at pH adjusted at ≈8.5. For removal of chlorides and excess ammonia, the obtained product was washed several times by decantation. Finally, the slurry was dried at 80°C, where the mixture swelled and gradually increased in volume forming puffed white product. The final product was crushed and calcined at 550°C for 4 h. The finished composite support was denoted as (TiO2(15%)–Al2O3)w. Applying the described procedure, the binary oxide support with the same TiO2/Al2O3 ratio (15 : 85 wt./wt., respectively), was prepared from analytical grade chemicals without templating with starch. The obtained support was denoted as (TiO2(15%)–Al2O3)a. Catalysts Preparation NiMo and CoMo catalysts were prepared using incipient wetness sequential impregnation. First, the

molybdenum impregnation onto the support, (TiO2(15%)–Al2O3)w, was carried out using an aqueous solution of (NH4)6Mo7O24 · 4H2O (99.8%, Aldrich). The heptamolybdate was dissolved in de-ionized water with stirring and hydrogen peroxide was then added to increase the heptamolybdate solubility. Also, HNO3 was added to adjust pH at 2 for avoiding any precipitation of molybdenum compounds. Thereafter, (TiO2(15%)–Al2O3)w was added to the clear solution with slow stirring at 60°C for 3 h where the Mo precursor was homogenously dispersed and incorporated into the pores of the support. The obtained solid was dried at 120°C [9]. For further incorporating nickel oxide or cobalt oxide, impregnation technique was applied by using an aqueous solution of nickel nitrate (Sigma–Aldrich) or cobalt nitrate (Aldrich). Ni(NO3)2 · 6H2O or Co(NO3)2 · 6H2O was dissolved in de-ionized water, then, the supported Mo on (TiO2(15%)–Al2O3)w was dispersed with stirring for 2 h, dried at 120°C for 6 h and calcined at 550°C for 4 h. The obtained catalysts, NiMo or CoMo with NiO or CoO loading of 3% and MoO3 loading of 10% were or denoted as NiMo/(TiO2(15%)–Al2O3)w CoMo/(TiO2(15%)–Al2O3)w respectively. Following the same procedures, NiO and MoO3 were loaded onto (TiO2(15%)–Al2O3)a surface. Catalysts Characterization Techniques The BET surface area determination and pore size analysis were carried out by the aid of adsorption– desorption isotherms of N2 at –196°C recorded on a Quantachrome Autosorb-1 adsorption analyzer. Pretreatment of the samples involved outgassing at 200°C for 3 h. The X-ray powder diffraction (XRD) patterns were recorded on a Brucker AXS-D8 Advance (Germany), using nickel-filtered copper radiation (λ = 1.5405 Å) at 60 kV and 25 mA with a scanning speed of 8°/20 min over diffraction angle range. The differential scanning calorimetry (DSC) and the thermo-gravimetric analysis (TGA) were performed adopting SDT-Q600 V20.5 Pouild 15 apparatus, with heating rate 5°C min–1 at a nitrogen flow rate of 100 mL min–1. The morphological as well as microstructural features of the prepared samples were imaged by scanning electron microscope (SEM) of JEOL 5300 (Japan) SEM model and by transmission electron microscope (TEM) of JEOL-2000EX model at an accelerating voltage of 100 kV, respectively. Hydrogen temperature-programmed reduction (H2–TPR) experiments for precursor catalysts were carried out from ambient temperature to 1000°C, with a flow of 10% H2/N2 (85 mL min–1) and a heating rate of 10°C min–1 using Chembet-3000 analyzer. PETROLEUM CHEMISTRY

Vol. 58

No. 5

2018

CAN MESOPOROUS TiO2-Al2O3-SUPPORTED

389

Table 1. Physico-chemical properties of gas oil feedstock Item

Value

Specific gravity, 60/60°F Refractive index, 20° Component analysis, %: total saturates total aromatics Total sulfur, content, wt %

0.8340 1.46417 74 26 0.76

Catalytic Performance Assessment The GO feedstock used in this investigation was gifted by Cairo Petroleum Refining Company. Its physicochemical characteristics are given in Table 1. The catalyst performance was evaluated by using two systems: (i) Single stage operation, where each catalyst was tested individually. (ii) Dual catalytic bed system where CoMo/(TiO2(15%)–Al2O3)w and NiMo/(TiO2(15%)– Al2O3)w in a dual layers operation. In this process the catalyst system consists of both CoMo and NiMo catalyst in spatial arrangement with mass ratio of 1 : 1. A series of experiments were conducted to explore the influence of the operating factors on the qualities and quantities of the obtained products. The reactions have been performed using a continuous high pressure micro-reactor. The apparatus consists mainly of 50 cm length stainless steel reactor, with internal and external diameter of 1.9 and 2.7 cm, respectively. It divided into three zones; each of them has its individual heating element and temperature controller. Hydrogen gas was supplied to the unit from a H2 cylinder, while liquid feed way pumped to the top of the reactor by means of a piston pump. After charging the reactor with the predetermined quantity of the catalyst, the unit was first flushed with nitrogen then kept for 4 h under a nitrogen gas pressure to check any leakage. The catalyst bed was pre-sulfided with 2% dimethyl disulfide/cyclohexane mixture which was passed through the catalyst bed at the following conditions: hydrogen flow rate = 0.5 mL min–1, feed flow rate = 0.5 mL min–1 and Liquid hourly space velocity (LHSV) = 4 h–1. Sulfidation was performed at two different temperatures; the first at 260°C for 3 h and the second at 360°C for 3 h. The reaction conditions for all HDS processes were: reaction temperature 325–400°C, LHSV 0.5−4 h–1, operating pressure 6.0 MPa and H2/oil ratio 450 v/v. In all tests, the liquid products were analyzed according to ASTM standard methods as shown in Table 1. PETROLEUM CHEMISTRY

Vol. 58

No. 5

2018

Method ASTM D 289-65 ASTM D 1218-92 Silica-gel column

ASTM D 294-90 (X-ray fluorescence)

RESULTS AND DISCUSSION Catalyst Characterization Textural Analysis. The surface parameters calculated by BET method for the mixed oxide supports, (TiO2(15%)–Al2O3)w with a template or (TiO2(15%)– Al2O3)a without template, as well as for NiMo or CoMo loaded catalysts are presented in Table 2. It is evident from this Table that textural modification of mixed oxide support by using starch template could lead to a marked increase in both surface area (SBET) and pores dimensions, viz., pore volume (Vp) and average pore diameter (Dp). The optimum calcination temperature applied in this study was 550°C, as the lower calcination temperatures could lead to incomplete starch combustion; increasing thus the carbon residual and decreasing the surface area and pore dimensions. Higher calcination temperatures could lead to complete starch combustion and less carbon residual. Yet, this was also associated with a decrease in specific area and pore dimensions apparently due to sintering [10]. The extent of N2 adsorption decreases with loading of NiMo or CoMo on the oxide supports with subsequent reduction in surface areas but with less influence on the pore parameters. This may be referred to the incorporation of NiMo or CoMo nanoparticles in the pore system, which could in turn hinder the accessibility of further amounts of adsorbed N2. X-ray Diffraction (XRD). Figure 1 depicts the XRD patterns for the binary oxide support systems as well as the NiMoS and CoMoS containing catalysts. The XRD pattern of templated (TiO2(15%)–Al2O3)w support indicates the presence of γ-Al2O3 phase (JCPDS Card no. 77-396) [11]. Typical peaks of anatase TiO2 (JCPDS Card no. 04-0477) [11] could not be detected. This could be attributed to the high degree of dispersion of Ti NPs within the Al matrix and in addition, it may be linked with the possible overlap of Ti and Al peaks in the amorphous phases. Regarding the NiMo and CoMo sulfide phases, no characteristic bands for their crystalline phases could be detected, according to the standard (PDF-IDD) files [11], in spite of their moderate loading percentages. This may point to the existence of metallic phases in an extremely small nanoparticle scale and

390

NASSER H. SHALABY et al.

Table 2. Surface parameters of various samples under study SBET, m2 g–1

Vp, mL g–1

Dp, nm

(TiO2(15%)–Al2O3)w

272

0.56

8.42

(TiO2(15%)–Al2O3)a

218

0.35

6.32

NiMo/(TiO2(15%)–Al2O3)w

236

0.42

7.33

CoMo/(TiO2(15%)–Al2O3)w

243

0.39

7.41

NiMo/(TiO2(15%)–Al2O3)a

188

0.26

6.11

moreover it seems that their stronger interaction with Ti may be displayed at the expense of the Al2O3 phase, i.e., with the formation of an amorphous phase containing Ni, Mo, O, and S in both cases [12]. Thermal Analysis. Figure 2 shows the TGA and DSC traces of the as-dried titania-alumina precursor (without template) in the range of 35–1000°C. The observed weight losses apparently took place at temperatures below 550°C, after which no weight loss could be noticed. The first weight loss (16%) up to 200°C, accompanied with an endothermic peak maximized at 100°C, seemed to be due to the removal of physically adsorbed water. The second weight loss (40%) in the range ∼200– 550°C, accompanied also with an endothermic peak maximized at 280°C, indicated beside the loss of chemically bound water, the complete decomposition of hydroxide precursor, most probably contaminated with the precipitated materials [13]. The sample showed, in addition, three DSC exothermic peaks around 430, 500, and 794°C, corresponded to the crystallization of the anatase phase [14], γ-Al2O3 and κ-Al2O3 overlapped with rutile phase, respectively [15]. In general, the observed thermal behavior indicated the justified choice of 550°C degree as a calcination temperature for all the supports adopted throughout this study.

NiMo/(TiO2(15%)–Al2O3)w

On the other hand, the morphology and average particle sizes of the studied catalysts in sulfide form NiMo/(TiO2(15%)–Al2O3)w and CoMo/(TiO2(15%)– Al2O3)w) were represented in the TEM images in Figs. 3b and 3c. The obtained micrographs indicated the typical layered hexagonal crystallographic nature characteristic of MoS2 phase. The edge-bonded MoS2 aggregates, perpendicular to the support surface, seemed to have a weaker metal-support interaction, as compared with the basally bonded MoS2 ones onto the support. For the NiMoS phase (Fig. 3b), mainly the basal-bonded MoS2 single layers (NiMo-S type I) were formed, together with some basal bonded MoS2 multi-layers (NiMo-S type II) [17]. The sample evidently possessed the nanoparticles of MoS2 phase of an average size of ~10 nm (viz., between 9 and 13 nm). Shortly, the sulfided NiMo/(TiO2(15%)–Al2O3)w catalyst appeared as MoS2like structure made up of small ordered sections, inclined and most probably interacted with the support surface. Alumina, on the other hand, was shown as separated fine particles due to the introduction of titania (15 wt %). The produced mesoporous structure 120 TGA

100 Weight, %

Intensity, a.u.

(TiO2(15%)–Al2O3)w

Morphology Investigation. The SEM image of calcined (TiO2(15%)-Al2O3)w support, with a template, given in Fig. 3a revealed the homogeneous microstructure of γ-Al2O3 [16], where the introduction of titania into the alumina matrix seemed to develop the mesoporous structure of the mixed oxide system.

NiMoS CoMoS

80

DSC

16%

γ-Al2O3 κ-Al2O3 anatase rutile phase phase

40%

60 40 20

20

30

40

50 2θ

60

70

Fig. 1. X-ray patterns of the prepared samples.

80

0

200

400 600 800 Temperature, °C

0.6 0.4 0.2 0 –0.2 –0.4 –0.6 –0.8 –1.0 –1.2 1000

Heat flow, cal/gm s

Sample

Fig. 2. TGA and DSC traces of (TiO2(15%)–Al2O3)a. PETROLEUM CHEMISTRY

Vol. 58

No. 5

2018

CAN MESOPOROUS TiO2-Al2O3-SUPPORTED

391 4.1 nm

15.2 nm

2.4 nm 10.3 nm 9.1 nm

1 μm

(а)

(b)

50 nm

260 kx

220 kx

100 nm

(c)

Fig. 3. (a) SEM image of (TiO2(15%)–Al2O3)w support; (b) TEM image of NiMo/(TiO2(15%)–Al2O3)w; (c) CoMo/(TiO2(15%)– Al2O3)w.

seemed to enhance the penetration of reactant species to be accessible for reaction in the inner pore system.

respectively. These Co species seem also to be entrapped in the support pore system.

In the case of sulfided CoMo/(TiO2(15%)–Al2O3)w catalyst (Fig. 3c), several parallel and perpendicular black nanotubes appeared in uneven distributions, some of which had central hollow open ended and multilayer forms [18]. Moreover, separate spherical groups of the CoMoS nanoparticles of ∼2 nm average size were located on this nanotubular support structure (TiO2(15%)–Al2O3)w. The active phase apparently existed in a mono-layered profile.

On the other hand, for Mo supported sample, two main peaks and a shoulder were observed. The lowtemperature peak (around 642°C) is assigned to the partial reduction (Mo6+ → Mo4+) of amorphous highly reactive, multilayered Mo oxides or octahedral Mo species [22]. The high temperature peaks (958– 1020°C) may be related to the deep reduction of Mo species, including highly dispersed tetrahedral Mo species. The shoulder may be due to the intermediate-reducible crystalline phase of orthorhombic MoO3 [20].

As well for cobalt contained catalysts, it is clear that the cobalt species- support (TiO2(15%)–Al2O3)w or (TiO2(15%)–Al2O3)a interactions can lead to a considerable increase in the reduction temperature of the cobalt oxide phases (to 578, 818, and 1014°C). It was reported recently [21], that reduction of Co3O4 to Co occurs in two stages: Co3O4 → CoO → Co (hcp) and Co (fcc) at temperature ranges of 200 to 400°C and 220 to 330°C for unsupported and supported catalysts, PETROLEUM CHEMISTRY

Vol. 58

No. 5

2018

A comparison of the TPR spectra of the Ni or Co supported on Mo/(TiO2(15%)–Al2O3)w, Figs. 4a, 4b, show that Ni promotes the reduction of Mo species (specially the first peak of the reactive octahedral Mo) than cobalt, where they are reduced around 489°C in case of Ni supported catalyst, while it has just reduced around 578°C in case of Co supported catalyst. This may interpret the more activity of Ni than Co in many hydrodesulfurization processes. The change in reduc(a) H2-Consumption, a.u.

Hydrogen Temperature-Programmed Reduction (H2-TPR). To assess the interaction profiles in the assynthesized catalytic systems, the H2-TPR study has been undertaken for MoNi and MoCo oxides loaded onto both the (TiO2(15%)–Al2O3)w and the (TiO2(15%)– Al2O3)a supports, as demonstrated in Fig. 4. For comparison purposes, single metal oxide phases of Mo, Co, and Ni loaded on the same supports were subject to this reduction study. Regarding the single Ni supported sample Ni/(TiO2(15%)–Al2O3)w and taking into consideration the single reduction peak of bulk NiO around 621°C [19], the obtained three reduction processes indicate most likely that Ni2+ species exist in different structural environments [20]. Moreover, the higher reduction temperatures observed, viz., at 723 and 888°C suggest the formation of some different Ni species in a stronger interaction with the modified support, probably entrapped in its pore system.

(b)

(c) NiMo/(TiO2(15%)–Al2O3)w (f) CoMo/(TiO2(15%)–Al2O3)w (b) NiMo/(TiO2(15%)–Al2O3)a (e) Co/(TiO2(15%)–Al2O3)w (a) Ni/(TiO2(15%)–Al2O3)w (d) Mo/(TiO2(15%)–Al2O3)w 972

(c) (b) (a)

667 912

463 888 621

578

400 800 T, °C

818

(f) (e)

723

0

1018

611

489

1014 642 832

958

(d) 0 200 400 600 800 1000 1200

T, °C

Fig. 4. TPR patterns of various prepared samples.

392

NASSER H. SHALABY et al.

Table 3. Catalytic Performance of Prepared Catalysts as a Function in Temperature and LHSV at Fixed Pressure (6 MPa) and H2/Oil Ratio (450 v/v) Temperature, °C 320

360

LHSV, h–1

Catalyst

400

LHSV, h–1

LHSV, h–1

effciency*, effciency*, effciency*, 0.5 1 2 4 0.5 1 2 4 % % % sulfur in product, ppm sulfur in product, ppm sulfur in product, ppm (a) (b) (c) (d)

0.5

1

2

4

55 67 253 –

132 146 415 –

212 240 551 –

341 370 812 –

99.3 99.1 96.7 –

29 33 142 –

60 75 233 –

124 139 380 –

206 228 563 –

99.62 99.57 98.1 –

16 26 64 6

28 38 111 –

41 46 208 –

58 77 293 –

99.8 99.65 99.15 99.92

* Efficiency (%) calculated at LHSV of 0.5. (a)NiMo/(TiO2(15%)–Al2O3)w. (b) NiMo/(TiO2(15%)–Al2O3)a. (c) CoMo/(TiO2(15%)–Al2O3)w. (d) CoMo/(TiO2(15%)–Al2O3)w//NiMo/(TiO2(15%)–Al2O3)w (Dual Catalytic Bed).

ibility of Mo-Ni and Mo-Co may be related to the difference in the type of surface interaction in the two cases. Therefore, it can be suggested that the Ni–Mo–O or Co–Mo–O phases formed by this interaction, are the precursors of the Ni–Mo–S or Co–Mo–S phase known as the HDS active site [23]. Catalytic Performance The catalytic performance tests in HDS of GO using the as-prepared NiMo and CoMo supported catalysts were monitored via the changes in sulfur content. The HDS efficiencies were thus expressed as follows:

Sfeed – Sproduct    × 100,  Sfeed where, Sfeed and Sproduct represent the sulfur concentration in the feed and the products, respectively [24]. The influence of different operating conditions on catalytic performances is discussed below, namely, the reaction temperature (320–400°C), LHSV (0.5–4 h–1), keeping the pressure constant at 6.0 MPa and the H2/oil ratio at 450 v/v. HDS =

Single Stage Operation Effect of the Reaction Temperature. The reaction temperature plays an important role in HDS. To study the influence of the reaction temperature on the HDS of gas oil feedstock, experiments using the prepared catalysts were conducted at 320, 360, and 400°C. As shown in Table 3, it is obvious that the HDS increases rapidly with increasing reaction temperature. The HDS exceeds 99% and is close to 100% when the reaction temperature reaches 400°C at LHSV 0.5 h–1. High temperature favors the formation of vacancies on

MoS2 edge surfaces, which are the coordinately unsaturated molybdenum active sites located at the edge of MoS2 crystallites [25]. The superiority of NiMo/(TiO2(15%)–Al2O3)w catalyst among the other catalyst systems under study seems to be attributable to the findings that Ni facilitates the reduction of Mo species, more than Co at the same composition. It is worth to recall that the unsaturated sulfur vacancies increase in the Ni promoted catalyst than in the Co promoted one [26]. The superiority of this Ni- templated catalyst may also be linked with its textural characteristics, where the pores dimensions are compatible with the large size of GO aromatic molecules. But we must note that the promotion effect is the predominant and then comes the textural effect, Tables 2 and 3. Effect of LHSV. Table 3 shows also the HDS efficiencies as a function of the LHSV value varied between 0.5 and 4 h–1 at different temperatures, from 320 to 400°C, keeping the pressure constant at 6.0 MPa and the H2/oil ratio at 450 v/v. The conversion of sulfur compounds decreases as the LHSV value increases.

By definition, the LHSV is a measure of the time that lapses for the reactant to reach the catalyst active sites. As the LHSV value increases, the reaction time decreases and the conversion is reduced. On the contrary, the residence time increases with decreasing the LHSV value. The most suitable LHSV is found to be 0.5 h–1, where the sulfur removal exceeds 99% at all reaction temperatures (Table 3). At LHSV of 0.5 h–1, the NiMo/(TiO2(15%)–Al2O3)w catalyst exhibits the highest HDS catalytic activity in the single stage operation (the S% is reduced from 55 to 16 ppm, (Table 3) with the increase of temperature from 320 to 400°C. The lowest HDS activity of CoMo/(TiO2(15%)–Al2O3)w PETROLEUM CHEMISTRY

Vol. 58

No. 5

2018

CAN MESOPOROUS TiO2-Al2O3-SUPPORTED

(the S% is reduced from 281 to 75 ppm at the same operating conditions mentioned) can be understood, where Co seems to facilitate the interaction of Mo species with the support, affecting therefore their reduction profile (Fig. 5) and consequently the overall HDS efficiency. As the results of this part show that the sulfur content S% is still higher than 10 ppm, which is the worldwide limit, so the dual stage operation was the other alternative to be applied as given below. Binary-Layer Operation The main task of this part of the study is to reduce the sulfur species in the gas oil, to be less than 10 ppm, the international permissible limit. For this aim, the modern ultra-low-sulfur diesel (ULSD) unit was used; adopting staged and layered catalysts beds [27]. The first stage or layer aims to efficiently reduce the reactive sulfur level by direct desulfurization route and additionally to reduce the refractory sulfur species by hydrogenation route on Co–Mo or Ni–Mo combination as much as possible [28]. Inhibition of the copresent species of the feed gas oil should be moderated by selecting the proper catalyst systems and reaction conditions. The influence of H2S produced by HDS of the reactant was taken into account. Ni–Mo is superior in the practical desulfurization, while suffering much less inhibitions than Co–Mo supported catalyst [29]. In this study, the two-layer catalyst systems were arranged such that the Co–Mo/(TiO2(15%)–Al2O3)w was the upper layer, while as the Ni–Mo/(TiO2(15%)– Al2O3)w was the lower layer, in the operation. The catalysts used in the two-layer systems were in a ratio of 1 : 1. It is evident from Table 3 that, the binary-layer catalyst beds, adopted at the above mentioned conditions, namely at 320–400°C, 0.5–4 h–1, 6.0 MPa and 450 v/v, are shown to achieve deep HDS efficiency of GO, reaching the desirable sulfur level (