Hydroprocessing of heavy gas oils using FeW.pdf

8/14/2019 Hydroprocessing of heavy gas oils using FeW.pdf

1/11

Pleasecite this article in press as:P.E. Boahene, et al., Hydroprocessingof heavy gasoils using FeW/SBA-15 catalysts:Experimentals,optimization

ofmetals loading, and kinetics study, Catal. Today (2012), http://dx.doi.org/10.1016/j.cattod.2012.04.064

ARTICLE IN PRESSGModel

CATTOD-8043; No.of Pages11

Catalysis Today xxx (2012) xxxxxx

Contents lists available at SciVerse ScienceDirect

Catalysis Today

journal homepage: www.elsevier .com/ locate /cattod

Hydroprocessing ofheavy gas oils using FeW/SBA-15 catalysts:

Experimentals, optimization ofmetals loading, and kinetics study

Philip E. Boahenea, Kapil K. Soni a, Ajay K. Dalai a,, John Adjaye b

a Catalysis and Chemical Reaction Engineering Laboratories, Department of Chemical Engineering, University of Saskatchewan, Saskatoon, SK, S7N 5A9, Canadab Syncrude Edmonton Research Centre, Edmonton,AB, T6N1H4, Canada

a r t i c l e i n f o

Article history:

Received 25 January 2012Received in revised form 27 April 2012

Accepted 30 April 2012

Available online xxx

Keywords:

SBA-15

FeW Catalyst

HDS

HDN

Kinetics

Power Law

Multi-parameter models

a b s t r a c t

In the present work, a series ofFeW/SBA-15 catalysts were prepared and screened for their hydrodesul-

furization and hydrodenitrogenation activities using bitumen-derived heavy gas oil from Athabasca. A

systematic process optimization study has been conducted to investigate the optimum process condi-

tions required to evaluate kinetic parameters for these reactions. Catalyst metal loadings were varied

from 0 to 5 and 15 to 45 wt.% for Fe and W, respectively; resulting in an optimum catalyst (Cat-5) with

metal loadings of 3.0 and 30.0 wt.% for Fe and W, respectively. Several techniques were employed to

characterize the prepared catalysts and activity results have been correlated with that obtained from

characterization. Hydrotreating experiments were performed in a continuous flow micro trickle-bed

reactor at the temperatures, pressures, and LHSVs of 633693 K, 7.69.6 MPa, and 0.52 h1, respec-tively, with H2 flow rate and catalyst weight maintained constant at 50 mL/min and 1.5 g, respectively,

in all cases. Three kinetic models were applied to fit experimental data obtained from HDS and HDN

reaction studies evaluated within temperature range of633693 K. The optimum operating conditionsfor maximum sulfur and nitrogen conversions occurred at temperature, pressure, and LHSV of 673 K,8.8MPa, and 1 h1, respectively. Experimental data fitted with the Power Law model yielded reactionorders of 2.0 and 1.5 for HDS and HDN reactions, respectively; and activation energies of 129.6 kJ/mol

and 150.6 kJ/mol, respectively. By fitting a modified power law model (Multi-parameter model) to the

kinetic data yielded hydrodesulfurization (HDS) and hydrotreating (HDN) reactions orders of2.2 and 1.8,with respective activation energies of126.7 kJ/mol and 118.8 kJ/mol.

2012 Elsevier B.V. All rights reserved.

1. Introduction

Mounting worldwide concern to meetthe increasinglystringent

regulations on transportation fuels such as dieseland gasoline have

explored deephydrodesulfurization (HDS) undersevere conditions

[13], and utilization of better catalysts for gas oils hydrotreatment

[4]. Indeed, hydrotreatment targets the removal of heteroatomic

species (S, N, etc.) and also aims to minimize the detrimental poi-

soning effectof catalystsused in downstreamrefineryprocesses. By

virtue of the growing demand on quality light to middle distillateoilfractions, catalytichydroprocessing of heavyoil fractionscontin-

uesto provide benefitto themodern petroleumrefinery. Asa result,

extensive effort has been attributed worldwide toward character-

izing such heavy oil fractions from the standpoints of feedstock

properties and the resultant kinetic properties [57]. In lieuof this,

for instance, the maximum permissible sulfur content in diesel

fuels is now targeted to the ultra low levels (1015ppm) [810].

Correspondingauthor. Tel.: +1 306966 4771; fax: +1 306966 4777.E-mail address: [email protected](A.K. Dalai).

Thus, the development of highly active and selective HDS catalysts,

capable of processing low quality heavier feedstocks, is a perti-

nent challenge encountered by the petroleum industry of recent

times.

In the conventional hydrotreating (HDT) process, compounds

containing heteroatoms such as organic sulfur and nitrogen

undergo surface catalytic reactions with pre-adsorbed hydrogen

to form hydrogen sulfide and hydrocarbon [11,12]. Commercially

used catalysts to perform HDT reactions are typically composed of

sulfides of molybdenum or tungsten (1020wt.%) supported on-Al2O3, and mostly promoted by either cobalt or nickel (35wt.%)

[1316]. The catalytically active phase formed thereof in a pro-

moted Co (Ni) Mo sulfide catalyst system is the so-called CoMoS

or NiMoS phase, in which the promoter atoms ( Ni or Co) deco-

rate the edge of well-dispersed MoS2nanoparticles on the support

[14,17]. Due to tighter environmental regulations regarding sul-

fur reduction in fuels, researchers and refineries need to develop

much higher performance catalysts for HDS [12,16]. In this regard,

numerous studies have been conducted on the development of

new catalyst systems with greater activity than the current indus-

trial catalysts. Strategies employed to achieve this goal include

0920-5861/$ see front matter 2012 Elsevier B.V. All rights reserved.

http://dx.doi.org/10.1016/j.cattod.2012.04.064
http://localhost/var/www/apps/conversion/tmp/scratch_2/dx.doi.org/10.1016/j.cattod.2012.04.064http://localhost/var/www/apps/conversion/tmp/scratch_2/dx.doi.org/10.1016/j.cattod.2012.04.064http://localhost/var/www/apps/conversion/tmp/scratch_2/dx.doi.org/10.1016/j.cattod.2012.04.064http://www.sciencedirect.com/science/journal/09205861http://www.elsevier.com/locate/cattodmailto:[email protected]://localhost/var/www/apps/conversion/tmp/scratch_2/dx.doi.org/10.1016/j.cattod.2012.04.064http://localhost/var/www/apps/conversion/tmp/scratch_2/dx.doi.org/10.1016/j.cattod.2012.04.064mailto:[email protected]://www.elsevier.com/locate/cattodhttp://www.sciencedirect.com/science/journal/09205861http://localhost/var/www/apps/conversion/tmp/scratch_2/dx.doi.org/10.1016/j.cattod.2012.04.064http://localhost/var/www/apps/conversion/tmp/scratch_2/dx.doi.org/10.1016/j.cattod.2012.04.064


2/11

Pleasecite this article in press as:P.E. Boahene, et al., Hydroprocessingof heavy gasoils using FeW/SBA-15 catalysts: Experimentals, optimization




2 P.E. Boahene et al./ Catalysis Todayxxx (2012) xxxxxx

active phase formulation, new preparation methods, and variation

or modification of the catalyst support [16,1822].

Catalytic performance improvement via the option of support

modification has been found to be very crucial [20,23]; thus, con-

tributing immensely to the overall HDT activity. Potential catalyst

supports thathave beeninvestigated overthe yearsincludecarbon-

based materials [24,25], mixed oxides [26,27], zeolites [28] and

ordered mesoporous silica materials like MCM-41 [29,30], HMS

[22,31], KIT-6 [32,33], and SBA-15 [16,34]. The latter catalyst sup-

porthas garnered significant attentionin the fieldof heterogeneous

catalysis and related fields of nano-materials syntheses due to its

attractive textural properties. SBA-15 is characterized by its high

surface area (6001000 m2/g), high hydrothermal stabilityand uni-

formly distributed hexagonal array of cylindrical pore channels

withtunableporesintherangeof530nm [35,36]. Inadditiontoits

high thermal and hydrothermal stability [37,38], it is conceivable

that the large andordered pore diameter of SBA-15would enhance

therelatively easy access of reactant molecules into thepores; thus

increasing therate of HDT reactions. Furthermore,SBA-15 with rel-

atively large pore diameters could be envisaged to minimize the

effects of catalyst coking by pore mouth blocking, which is pro-

foundwith-Al2O3support during the HDT reactions [39]. Moreso,

the high surface area of SBA-15 would enhance great dispersion of

supported metals so as to increase the amount of catalyst metals

converted from the oxide phase to the sulfide phase. These attrac-

tive properties of SBA-15 made it a potential worth exploring in

real feedstock hydrotreatment applications [16].

In the hydrodesulfurization of dibenzothiophene and the hydro-

genation (HYD) of toluene using pure SBA-15 supported NiW-S

catalyst, Vradman et al. reported an increased catalytic activity of

the order of 1.4 and 7.3 times higher, respectively, than that of

the sulfided commercial CoMo/Al2O3. Their findings evidenced the

excellent potential of high loading sulfided NiW/SBA-15 catalysts

for deep hydrotreatment of real petroleum feedstocks [40]. Also,

Dhar et al. evaluated the catalytic performance of purely siliceous

SBA-15 supported Mo, CoMo, and NiMo catalysts for the HDS of

thiophene and HYD of cyclohexene, and correlated the catalytic

activities with the quantity of oxygen chemisorbed on the vacan-cies of the respective sulfided catalysts [34]. The good correlation

found between the catalytic activities and oxygen chemisorption

was attributed to the formation of a patchy monolayer as a result

of oxygen chemisorption at the anionic vacant sites of the MoS2catalysts. An activity comparison with-Al2O3-supported catalysts

clearly indicated the SBA-15-supported catalysts to be superior to

its -Al2O3 counterpart prepared in similar manner. Nonetheless,

the aforementioned limited number of studies in which pure SBA-

15 was applied as a catalyst support for HDT reactions employed

model compounds as the feedstocks [34,40]. However, a study con-

ducted by Sundaramuthy et al. tested the catalytic functionality of

AlSBA-15-supportedNiMo catalyst by screening with a light gas oil

(derived from Athabasca bitumen) petroleum fraction. The catalyst

with 17wt.% Mo and 3.4wt.% of Ni was found to produce the bestHDNand HDSactivities,whichwas comparable tothe conventional

-Al2O3-supported NiMo catalyst at industrial conditions [41].

It is worth mentioning that notable among the aforementioned

studies using SBA-15 catalyst support [34,40,41], is the fact that

the pore diameters of the supports varied in the range of 58nm.

However, in the HDT of heavier petroleum fractions, it is generally

more practicable to usecatalysts with largerporediameters so as to

enhance efficient species diffusion and also to minimize the possi-

bilities of pore-plugging via coke deposition [39]. Even though our

previous investigations on the effectiveness of different pore diam-

eter FeW/SBA-15 catalyst concluded that the catalyst with pore

diameter of at least 10nm wassuperioramongstthe catalystsstud-

ied for the hydrotreatment of bitumen-derived heavy gas oils, due

to sufficientmass transfer of reactant liquids andgases through the

catalysts poreswhile stillmaintaining a highsurface areanecessary

for metal dispersion [16], one should note that the potential indus-

trial application of such catalyst systems would require thorough

kinetic studies.

Several kinetic studies documented in the open literature on

hydrotreating reactions using real feedstocks have mostly used the

PowerLaw model [4246], LangmuirHinshelwoodmodel [4547],

and Multi-parameter model[42,46]to determinesignificant kinetic

parameters. It is well known that in the hydrotreating processes,

adsorption of reactant species on the catalyst active sites is known

to be the rate-determining step in the reaction process [48]. It also

noteworthy that hydrogen sulfide tends to adsorb strongly on the

catalyst active sites; thus, inhibiting the adsorption of nitrogen and

other sulfur molecules during hydrotreating [47,48]. One would

therefore expect that the LangmuirHinshelwood and the Multi-

parameter models which account for such inhibition contributions

during the HDT process would be more representative and thereby

preferredforkineticanalysesin theHDS andHDNof real feedstocks.

The present study is an extension of our previous investiga-

tions on the effectiveness of different pore diameter FeW/SBA-15

catalyst [16]. The principle goal of this study is to investigate the

optimum process conditions required for the HDT of heavy gas oils

using a series of prepared FeW/SBA-15 catalysts, and also to con-

duct kinetic studies using the Power Law, LangmuirHinshelwood,

and Multi-parameter models to ascertain the effects of process vari-

ables on the rates of HDS and HDN reactions in a way to provide

in-depth understanding of HDT reactions as they occur on hetero-

geneous FeW/SBA-15 hydrotreating catalysts.

2. Experimental

2.1. Preparation of supports and catalysts

In the synthesis of the siliceous SBA-15 materials, the proce-

dure described in our previous paper was followed [16,36], using

a triblock copolymer Pluronic P123 (Mav =5800, EO20PO70EO20,

Aldrich) as the structure-directing agent (SDA) and tetraethylorthosilicate (TEOS) as the silica source. The nominal molar

ratio of the chemicals used in the synthesis mixture was

1.0TEOS:0.0168P123:4.02C6H14:0.0295NH4F:4.42HCl:186H2O. In

a typical synthesis procedure, 9.8 g P123 and 0.109g NH4F were

dissolved in 335 mLof 1.3M aqueous HCl solution at room tem-

perature. This solution was transferred to a constant temperature

bath (CTB) maintained at 288 K and a mixture of 20.8g TEOS and34.6g C6H14was slowly added under vigorous mechanical stirring.

After 24h of mechanical agitation of the content in the CTB, the

gel formed was isolated and subjected to hydrothermal treatment

in a teflon-lined autoclave for 3 days. The solid product was fil-

tered, washed with deionized water, and dried for 24h at room

temperature. The organic template was then removed by calcining

the powdered sample at 823K for 5h ata heating rateof 2 K/min.

2.2. Synthesis of FeW/SBA-15 catalysts

The series of SBA-15 supported FeW catalysts were prepared by

an incipient wetness impregnation technique. The calcined SBA-

15 support was impregnated successively using aqueous solutions

of ammonium metatungstate (AMT), (NH4)6H2W12O40 (Fluka) and

iron nitrate, Fe (NO3)3.9H2O (Aldrich) as a W and Fe source, respec-

tively. After each impregnation, the catalysts were dried at 373 Kfor 24 h . To prepare the catalyst with 2 and 15 w t.% Fe and W,

respectively, 2.55g of the pristine SBA-15 support was added to

a homogenous solution made by dissolving 0.514 g AMT in 15mL

de-ionized water. This mixture was dried in an oven at 373 K for

24h and the required amount of Fe was also loaded by a similar
http://localhost/var/www/apps/conversion/tmp/scratch_2/dx.doi.org/10.1016/j.cattod.2012.04.064http://localhost/var/www/apps/conversion/tmp/scratch_2/dx.doi.org/10.1016/j.cattod.2012.04.064


3/11





P.E. Boahene et al. / Catalysis Todayxxx (2012) xxxxxx 3

Table 1

Elemental compositions and physical properties of FeW/SBA-15 catalysts determined from ICP/MS, N2 sorption and XRD analyses.

Sample ID Composition (wt.%) SBET(m2/g)Cat NSBET PV (cm

3/g) PDads(nm) d10 0(nm) a0(nm) (nm)

Fe W

Cat-1 2a (1.91) 15a (14.07) 498 0.94 0.89 10.1 12.4 14.4 4.3

Cat-2 2a (1.89) 20a (14.20) 457 0.95 1.01 10.4 13.0 15.0 4.6

Cat-3 2a (1.86) 30a (29.84) 403 0.96 0.91 10.3 12.8 14.8 4.5

Cat-4 2a (1.93) 45a (43.01) 284 0.87 0.83 10.4 12.7 14.7 4.3

Cat-5 3a (2.82) 30a (29.24) 382 0.92 0.89 10.2 12.9 14.9 4.7

Cat-6 5a (4.76) 30a (29.51) 361 0.90 0.79 10.3 12.8 14.8 4.5

SBET, specific surface area calculated by theBET method.

(SBET)supp=617672m2 /g.

NSBET(Normalized surface area) were calculated by using the equation,NSBET= (SBETof the catalysts)/(1x). SBETof thesupport.PV, pore volume determined by nitrogen adsorption at a relative pressure of 0.98.

PDads, mesopore diameter corresponding to themaximum of thepore size distribution obtained from theadsorption isotherm by theBJH method.

a0, unit-cell parameter determined from thepositionof the(1 0 0) diffraction line asa0=d10 0 2/

3

, pore wall thickness calculated as d10 0=a0 Dads.a Targeted loading.

approach. The powdered sample was then calcined at 823 K for5h at a heating rate of2 K/min. Table 1 gives the designation of allcatalysts prepared and their respective metal contents.

2.3. Metals loading optimization

A sequential metal loading approach was adopted to obtain

the ideal Fe and W loadings required for the preparation of the

optimum FeW/SBA-15 catalyst. In this approach, the optimum W

loading was first obtained by varying it in the range of 1545 wt.%,

while maintaining the Fe loading at an initial constant loading of

2 wt.%. Once theW metal loading wasoptimized, theoptimum pro-

moter(Fe) loading wasthenfoundin conjunction with theobtained

optimum W loading. Specific loading combinations tested at con-

stant Fe loading of 2wt.% and W loadings of 20.0, 30.0, and 45.0,

respectively, resulted in catalysts designated as Cat-2, Cat-3, and

Cat-4, respectively. The obtained optimum W loading of 30wt.%

for these catalyst series was used to load 3.0 and 5.0wt.% Fe on

the optimized support to yield resulting catalysts designated asCat-5 and Cat-6, respectively. Details of the optimization regimen

are summarized in Table 1. The optimum catalyst (Cat-5) found

from the metal loading optimization was used to carry out kinet-

ics and long-term deactivation studies for this catalyst system. A

thorough catalyst characterization by N2-sorption, ICP/MS, SAXS,

powder XRD of similar FeW/SBA-15 catalysts has been discussed

in our previous work [16].

2.4. Catalytic activity studies

Kinetic studies were conducted in a fixed-bed reactor using

bitumen-derived heavy gas oil (HGO) from Athabasca as feed-

stock (supplied by Syncrude Canada Ltd). The feed properties are

given in Table 2. Typical industrial process conditions were main-tainedin allexperimentsusing 5 mLof catalyst in a downward flow

micro-trickle-bed reactor system (10mm ID; 285mm length). This

reactor system is a high pressure and temperature reaction set-up

whose operation mimics industrial hydrotreaters. The set-up con-

sists of liquid and gas feeding sections, a high pressure reactor, a

heater with temperature controller, a scrubber for removing the

ammonium sulfidefrom the reaction products, and a high pressure

gasliquid separator. Details of catalyst loading into the reactor are

described elsewhere [16,41]. Typically, the catalyst bed was packed

using 5 mLof FeW catalyst (1.5 g) and silicon carbide (SiC) parti-cles as diluents. Dilution of the catalyst bed with SiC particles was

necessary to enhance the heat and mass transfer of the system.

SiC particle size selection was based on earlier published work

by Bej et al. [49]. After the catalyst was loaded, the reactor was

mounted vertically inside an electrically heated furnace equipped

with thermocouple to monitor the temperature of the catalyst bed.

Before catalytic activity study, an initial catalyst pre-wetting

protocol was performed by pumping 100mLof a sulfiding solu-tion (composed of 2.9vol.% butanethiol in straight run gas oil

VOLTESSO 35) at a high flow rate of 2.5 m L/min into the reac-

tor. The high oil flow rate was subsequently adjusted to 5 mL/h

and maintained to obtain a liquid hourly space velocity (LHSV) of

1 h1. Prior to feed introduction into the reactor, catalyst sulfida-tion was carried out for a period of 48h at temperatures of 466 Kfor 24 h and 616 K for another 24h using the sulfiding solutionand operating hydrogen gas flow rate at 50mL/min. Subsequent

to sulfidation, precoking was carried out at temperature and pres-

sure of 648 K and 8.8 MPa, respectively, for 5 days to stabilize thecatalytic activity using the real feed (HGO) at a flow rate of 5 mL/h.

Afterprecoking, the effects of temperature on HDN andHDS activity

were conducted at five different temperatures (633, 648, 661, 673,

and 693 K) for 3 days each. The pressure, gas/oil ratio and LHSVwere maintained constant at 8.8 MPa, 600, and 1 h1, respectively.

Table 3 summarizesthe process conditionsemployedto investigatetheeffects of theotherprocess variablesstudied.In cases where the

effect of oneprocess variable was under study, allthe other process

conditions were held constant.

Hydrotreated products were collected after 24h and stripped

with nitrogen gas to remove dissolved ammonia andhydrogen sul-

fide gases. A stabilization period of 24h was allowed throughout

the experiments when there was a change in process conditions.

Liquid products taken within these stabilization periods were not

analyzed. Samples were analyzed for total nitrogen and sulfur con-

tentsusingan Antek 9000NS analyzer.Total nitrogen content ofthe

Table 2

Characteristics of heavy gas oil derived from Athabasca bitumen.

Characteristic Heavy gas oil

Nitrogen (wppm) 3615

Sulfur (wppm) 42,302

Density (g/ml) 0.99

Aromatic content (wt.%) 31.4

Asphaltene content (wt.%) 1.55

Boiling point distribution

IBP (K) 483.8FBP (K) 870.3

Boiling range(K) (wt.%)IBP478 (Gasoline) 0

478533 (Kerosene) 1

533588 (Light gas oil) 5

588698 (Heavy gas oil) 39

698873 (Vacuum gas oil) 55


4/11






Table 3

Process parameters variation for kinetic study (using 5 mLof FeW/SBA-15 catalyst).

Process parameter under study Effect of temperature Effect of pressure Effect of H2/oil ratio Effect of LHSV

Temperature (K) 633693 661 661 661Pressure (MPa) 8.8 7.69.6 8.8 8.8

H2/oil ratio (mL/mL) 600 600 4001000 600

LHSV (h1) 1.0 1.0 1.0 0.52.0

liquid product was measured by the combustion/chemiluminencetechnique following ASTM D4629 method and the sulfur content

was measured using the combustion/fluorescence technique fol-

lowing ASTM 5463 method. The instrumental error in N and S

analysis was 3%.

3. Results and discussion

3.1. Optimization of metals loading for HGO hydrotreating

Catalyst metalsoptimization was importantin order forthe cat-

alyst displaying maximum HDS and HDN activities to be selected

for further kinetic studies. The loading of active metal (tungsten)

was first optimized at a constant promoter loading of 2wt.% Fe,

followed by varying the promoter metal (iron) loading on the opti-mumW loading obtained. A total of six catalysts were investigated

in this study; five prepared using the same pristine SBA-15 sup-

port and the remaining one was the best performed catalyst, Cat-1

(215), selected from our previous study [16] to benchmark cat-

alytic performance as a function of metals(Fe andW) optimization.

Figs. 1 and 2 show results of the HDS and HDN activities for these

catalyst series. It can be concluded from these results that Cat-3,

with 30wt.% W, yielded the optimum W-loading on the SBA-15

catalyst support. On the basis of the optimum W-loading obtained,

the Fe loading was also optimized at three different loadings (2, 3,

and 5 wt.% Fe), maintaining a constant tungsten loading of 30wt.%

on the optimum SBA-15 support. Figs. 3 and 4 compare the steady-

state HDS and HDN activities, respectively, for each of the three

catalysts. Itis evident from these figuresthatincreasing theFe load-ing beyond3.0 wt.% resulted in a decreased HDS andHDN activities

of the SBA-15-supported FeW catalyst. This could be associated

with the corresponding decrease in surface area with increased

metal loadings, which is indicative of the fact that metal loadings

Fig. 1. Effect of HDS activities of SBA-15-supported FeW catalysts as a function

of variable tungsten loading at constant Fe loading of 2wt.%. (Catalyst= 5cm3,

P=8.8MPa,LHSV= 1h1

and H2/oil ratio= 600mL/mL).

Fig. 2. Effect of HDN activities of SBA-15-supported FeW catalysts as a function

of variable tungsten loading at constant Fe loading of 2wt.%. (Catalyst= 5cm3,

P=8.8MPa, LHSV= 1h1 and H2/oil ratio= 600mL/mL).

of 3wt.% Fe and 30wt.% W were the optimum loadings required

to yield the best catalytic performance. For descending reaction

temperatures of 673, 661, and 648 K, this optimum catalyst pro-duced sulfur conversions of 73.4, 64.1, and 52.9%, respectively, and

nitrogen conversions of 38.3, 26.1, and 21.7%, respectively.

Fig. 3. Effect of HDS activities of SBA-15-supported FeW catalysts as a function of

variable iron loading atconstantW loading of 30wt.%. (Catalyst= 5cm3, P= 8.8MPa,

LHSV=1h1

and H2 /oil ratio= 600mL/mL).


5/11






Fig. 4. Effect of HDN activities of SBA-15-supported FeW catalysts as a function of

variableironloadingat constantW loading of30 wt.%.(Catalyst =5 cm3, P=8.8MPa,

LHSV=1h1 and H2/oil ratio= 600mL/mL).

3.2. Influence of process parameters variation with hydrotreating

conversions

As aforementioned, hydrotreating is the key upgrading process

employed in refineries to reduce the sulfur and nitrogen contents

in petroleum fractions so as to produce cleaner fuels using vari-

ous operating conditions. Thus, kinetic experiments were designed

to focus on studying the effects of important process variables on

the extent of total sulfur and nitrogen removal from the heavy

gas oil feed. As can be seen from Table 3, the process parameters

studied arein therangeof temperatures, pressures, LHSVs,and gas-

to-oil ratios of 360420 C, 7.69.6 MPa, 0.52 h1

, and 4001000,respectively.

3.2.1. Effect of temperature

A simple and cost-effective way to enhance hydrotreating con-

versions is by manipulating process temperature. However, an

excessively high operating temperature may lead to activity loss

and shortening of catalyst life [50]. Thus, the effect of temperature

on sulfur and nitrogen conversion was studied. For the purposes

of kinetic studies, the effects of temperature has been studied by

varying it from 633 to 693 K. The other process variables namelypressure, LHSV and hydrogen/heavy gas oil volumetric ratio were

respectively constant at 8.8MPa, 1.0h1 and 600 during theseexperiments. The results shown in Fig. 5 indicate that an increase

in temperature favors the percent sulfur and nitrogen conversionsas expected. However, the rate of increase in HDS and HDN tend to

be slightly slow at higher temperature ranges as compared to that

at lower temperature ranges. Thus, it is quite obvious from Fig. 5

that a 20 K rise in temperature from 673 to 693K only resultedin about 4 wt.% change in sulfur conversion, which is suggestive

of the fact that operating the HDT process at about 673 K, a max-imum sulfur and nitrogen removal of about 73 and 38wt.% could

be achieved usingthe optimumFeW/SBA-15 catalyst.Experimental

runs were repeatedat 633, 648, 661, 673 and 693 K and the resultscompared with those of the previous runs in order to ascertain the

reproducibility of the data obtained. Results of the error analyses

as presented with Fig. 5 evidences the fact that the data are quite

reproducible, especially at higher temperatures with only a nar-

row margin of error. In the case of sulfur, the highest error margin

Fig. 5. Effect of temperature on sulfur and nitrogen conversions (catalyst= 5 cm3,

T= 633693 K, P=8.8MPa, LHSV= 1h1 and H2/oil ratio= 600mL/mL).

was

2.6wt% whereas those of nitrogen gave an average maximum

error of1.2 wt%.

3.2.2. Effect of liquid hourly space velocity

The liquid hourly spacevelocitygives the hourly volumetricflow

rate of liquid to the volume of catalyst in the reactor [51]. Thus,

the effects of liquid hourly space velocity (LHSV) on both sulfur

and nitrogen conversions were studied by varying LHSV from 0.5

to2.0h1 at a constant temperature of 661 K, pressure of 8.8 MPaanda hydrogen/heavy gasoil volumetric ratio of 600. Experimental

results depicted in Fig. 6 indicate that by decreasing the LHSV, the

extent of sulfur and nitrogen conversions increased due to the fact

that the contact time of the liquid with the catalyst increased. It

could be concluded from the LHSV studies that at these conditions

about 77.4 and 43.6 wt.% sulfur and nitrogen, respectively could be

achieved over the optimum FeW/SBA-15 catalyst.

3.2.3. Effect of pressure

One crucial process parameter that directly affects the

hydrotreating conversions is the hydrogen partial pressure. For

LHSV (h-1

)

2.52.01.51.00.50.0

Conversion

(%)

0

20

40

60

80

100

Sulfur

Nitrogen

Fig.6. Effect of liquid hourly space velocityon sulfur andnitrogenconversions (Cat-

alyst=5cm3

, T=661 K, P= 8.8MPa, LHSV =0.52h1

and H2/oil ratio = 600mL/mL).


6/11






0

20

40

60

80

10.09.08.07.0

Con

version(%)

Pressure (MPa)

Sulfur

Nitrogen

Fig. 7. Effect of pressure on sulfur and nitrogen conversions (catalyst= 5cm3,

T=661 K, P= 7.69.6MPa, LHSV= 1h1 and H2/oil ratio= 600mL/mL).

instance, in the hydrotreating of naphtha feedstock, Topsoe et al.

[52] observed that below a certain hydrogen partial pressure,it becomes impossible to reduce the product nitrogen to levels

required to be used as reformer feedstock even when the operating

temperature is increased. As a result, the effect of pressure on the

hydrotreating of heavy gas oil was studied at temperature, LHSV,

andhydrogen-to-gasoil ratio of 661K, 1 h1, and 600, respectively,and is represented in Fig. 7.

As was expected, increasing the hydrogen partial pressure

resulted in an increase in the extent of sulfur and nitrogen con-

versions through an increase in catalytic activity. It is known that

the major role of the catalyst is to provide the required reaction

interface for the reactants, thereby promoting interaction between

thefeedstock constituents(i.e. sulfur, nitrogen,etc.) andthe hydro-

gen [53]. However, it is noteworthy that excessively high hydrogen

pressures may only serve to saturate the catalyst surface and anyfurther increase in hydrogen partial pressure tend to affect the

hydrotreating conversions only by a slight margin [54]. Thus in the

present study, one could observe from Fig. 7 that by increasing the

hydrogen pressure from 9.0 to 9.6 MPa only resulted in less than

5wt.% conversion in both HDS and HDN processes. The nature of

the HDS process is such that the conversion rate tends to increase

with increasing partial pressure of the hydrogen. However, Botch-

wey concluded from his study on the inter-stage hydrogen sulfide

removal in a two-stage hydrotreating of heavy gas oil that exces-

sively increasing hydrogen pressure may result in relatively high

concentrations of ammonia and hydrogen sulfide in the vicinity

of the catalyst, which tends to have detrimental effect on catalyst

activity [55].

3.2.4. Effect of hydrogen-to-heavy gas oil ratio

In any hydrotreating reaction, the choice of proper ratio of

hydrogen/hydrocarbon is very crucial. Generally, an increase in

the hydrogen partial pressure increases the rate of hydrogenation,

which in turn increases the rate of removal of sulfur and nitrogen

compounds. The use of higher hydrogen pressure also enriches the

catalyst and reduces deactivation. Conversely, a very high value

of hydrogen/hydrocarbon ratio may increase the overall operat-

ing cost of the hydrotreating process. Thus, an optimum value of

the hydrogen/hydrocarbon ratio is always desired. In this regard,

the effect of hydrogen/heavy gas oil ratio on the HDS of heavy gas

oil has been determined by varying it from 400 to 100, maintain-

ing temperature, pressure and LHSV at 661 K, 8.8 MPa and 1.0 h1,

respectively. It can be observed fromFig. 8 thattheremovalof sulfur

Hydrogen gas-to-oil ratio (mL/mL)

1000800600400

Conversion

(%)

10

20

30

40

50

60

70

80

Sulfur

Nitrogen

Fig. 8. Effect of hydrogen gas-to-oil ratioon sulfur and nitrogen conversions (Cata-

lyst=5cm3, T=661 K,P=8.8MPa,LHSV=1h1 andH2/oilratio= 4001000mL/mL).

and nitrogen compounds increased significantly as the hydrogen-

to-heavy gas oil volumetric ratio is increased up to 800. However,

the effect began to level off beyond this threshold, rendering fur-

ther increments economically non-beneficial to the process. Bej

et al. attributed a similar observed trend to the pseudo-first order

dependency of the rates of HDS and HDN reactions on hydrogen

partial pressure corresponding to such a maximum value of hydro-

gen/heavy gas oil ratio [44]. Thus, the present study suggests that a

hydrogen/heavy gas oil ratio of about 800 needs to be maintained

if a maximum sulfur and nitrogen removal is to be achieved.

3.3. Kinetic parameters evaluation by different models

Information available in the literature on kinetic parameters

evaluation and modeling for the hydrodesulfurization and hydro-

denitrogenation reactions are normally derived using the power

law and the LangmuirHinshelwood models [2,4349]. In this

study, both the power law and LangmuirHinshelwood models

were employed in the kinetic analyses. In cases whereby inhibition

of other compounds was disregarded, the power law model was

employed for the kinetic parameters evaluation. However, in sce-

narios whereby inhibitive compounds such as hydrogen, hydrogen

sulfide removal, etc., were considered, the LangmuirHinshelwood

model was regarded appropriate in kinetic parameters evaluation.Another kinetic model of industrial importance investigated was

the Multi-parameter model. Kinetic results obtained are discussed

in the subsections that follow and are compared with those found

by other researchers (see Table 4).

3.3.1. The Power Lawmodel

The kinetics of HDS and HDN of heavy gas oil feedstocks is com-

plex, due to the presence of diverse kinds of sulfur and nitrogen

compounds with different reactivities [48]. Due to its simplicity,

the PL model is mostly used by many researchers in kinetic stud-

ies and modeling of the HDS and HDN reactions [4349]. However,

this model does not account for inhibition effects of other compo-

nents present in the feedstock [48,56]. The format of the PL model


7/11






Table 4

Some literature data of reaction orders and activation energies for heavy petroleum fractions for different catalysts.

References Feed boiling range (K) Kinetic model Reaction order Activation energy (kJ/mol)

HDS HDN HDS HDN

Present work 484870 PL 2.0 1.5 129.6 150.6

Present work 484870 LH 1st pseudo 1st pseudo 147.2 165.8

Present work 484870 MP 2.2 1.8 126.7 118.8

[46] 533865 PL 2.0 1.5 101 79

[46] 533865 LH 1st pseudo 1st pseudo 99 69

[46] 533865 MP 2.68 2.02 119 112

[45] 458849 LH 1.0 1.5 87 74

559814 PL 1.5 1.0 151 132

[42] 487832 MP 1.5 1.6 141 94

[47] 483873 LH 1st pseudo 1st pseudo 114.2 93.5

LGO/SRGO PL 1.5 1.5 77.8 64.2

[43] 483938 PL 1.5

[49] 483938 PL 2.0 80

is shown in Eq. (8), yielding three solutions depending on the value

ofn as shown in Eq. (8) (ac):

ri= dCidt = kiCin (8)

where Ci = concentration of species i (S or N) in petroleum frac-

tion, Cf and Cp = concentrations of heteroatomic species i (S or N,wt.%) in feed and hydrotreated products, respectively,ki = apparent

rate constant of species i, n= reaction order, t= residence time, and

LHSV= liquid hourly space velocity (i.e. the inverse of residence

time).

Cf Cp=ki

LHSVfor n = 0 (8a)

ln

CpCf

= ki

LHSVfor n = 1 (8b)

1

Cpn1

1

Cfn1

= (n 1) ki

LHSVfor n /= 0,1 (8c)

The main kinetic parameters that can be determined from thePLmodel arethe apparentrate constantand reaction order. Table5

summarizes the reaction orders determined by the PL model

for the removal of sulfur and nitrogen from the feedstock over

FeW/SBA-15 catalyst investigated in the present study. The values

of reaction orders were determined from the best fit of experi-

mental data. Using the general solution for the nth order kinetic

equation developed for the PL model (Eq. (8c)), a trial and error

approach was adopted by varying the value ofn until the highest

regression coefficient, R2, was obtained. Table 5 also reports values

ofR2 obtained from the fitting of these equations to experimental

data. Generally, R2 is regarded as a statistical measure of fitness;

the closer it is to unity, the better is the fitness. From the different

values of n tested, the best values ofn for the HDS and HDN reac-

tions were selected to be those with the highest R2

values. Thus,HDS follows a 2nd order, whereas HDN follows a 1.5th order. The

fitting of rate data for equations having the aforementioned orders

of reaction is represented in Fig. 9. Hence, the rate constants for

different temperatures were calculated using Eqs. (4) and (5) for

HDN and HDS, respectively, (as given in Table 5).

The activation energy can then be determined from the Svante

Arrhenius equation:

ki(T) = koeEa/RT (9)

where ko = pre-exponential factor; Ea = activation energy (kJ/mol);

R= universal gas constant (kJ/mol K); and T= temperature (K). To

determine the activation energies and pre-exponential factors for

the HDS and HDN reactions, a plot of lnk against 1/Twas com-

puted and represented in a graph as shown in Fig. 10. Parameters

1/LHSV (h)

2.52.01.51.00.50.0

0.0

0.2

0.4

0.6

0.8

1.0

1.2

0.0

0.2

0.4

0.6

0.8

1.0

1.2

HDS 2nd order

HDN 1.5th order

Fig. 9. Fitting of expe rimen tal data to th e PL model having or der s 1. 5 and2.0 f or HDN an d HDS, r esp ectively. (C atalyst= 5 cm3, T=661 K, P= 8.8MPa,LHSV=0.52h1 and H2/oil ratio= 600mL/mL).

1/T*1000 (K-1

)

0.001600.001580.001560.001540.001520.001500.00148

lnkHDS

-2.5

-2.0

-1.5

-1.0

-0.5

0.0

0.5

lnkHDN

-2.5

-2.0

-1.5

-1.0

-0.5

0.0

0.5

HDS

HDN

Fig. 10. Arrhenius plot of HDS and HDN rate constants obtained from the Power

Law model. (Catalyst= 5cm3, T= 648693 K, P=8.8MPa, LHSV=1 h1 and H2/oilratio= 600mL/mL).


8/11






Table 5

Determination of reaction orders by the PL model using their regression coefficients.

Equation # Order of reaction (n) Kinetic equations Value of R2

HDS HDN

1 0.0 k= [CF CP] LHSV 0.9907 0.96442 0.5 k= 2 [CP 0.5 CF 0.5] LHSV 0.9797 0.99373 1.0 k= l n [CF CP] LHSV 0.9905 0.99614 1.5 k=2 [1/CP

0.5 1/CF0.5] LHSV 0.9968 0.99875 2.0 k=[1/CP

1/CF]

LHSV 0.9992 0.9977

6 2.5 k=0.667[1/CP 1.5 1/CF1.5] LHSV 0.9983 0.99717 3.0 k=0.50 [1/CP

2 1/CF 2] LHSV 0.9947 0.9968

obtained from the Arrhenius plot within the range of temperatures

studied (633693K) for HDS and HDN yielded activation energiesof 129.6 and 150.6 kJ/mol, respectively. The pre-exponential fac-

tors computed from the intercept on the lnk axis were found to be

4.18E + 10 and 7.63E + 12 for HDS and HDN, respectively. However,

discrepancies in activation energy may indicate a change in mech-

anism of reaction or interference of a physical phenomenon such

as diffusion (Ferdous et al.), which tends to decrease the activation

energy as a result of inherent mass transfer limitations in packed

beds [45].

3.3.2. The LangmuirHinshelwood model

As a result of the diverse nature of composition of compounds

present in petroleum fractions, species inhibition to the HDS and

HDN reactions becomes a critical consideration for the kinetic

model development. Furthermore, the differentreactivities of these

compounds may result in different rates of adsorption of various

heteroatomic species on the catalyst surface [48]. This contributes

to the margin of error of kinetic parameters obtained from the

power model since it does not account for the competitive adsorp-

tion rates of the various species in the petroleum fraction. The LH

model takes into account thepercentage of catalyst activesites that

are occupied by the adhered reactant species at steady state, as

well as the percentage of sites that are vacant or inhibited by other

adhered compounds from the feed stream. Though, various forms

of the LH models exist in the open literature, a simplified form

mostlycited to represent the HDS and HDN reactions is depicted in

Eq. (10).

ri= dCidt = kiKiKH2PH2Ci

1+ KiCi + KH2PH2+ KH2SPH2S(10)

where ri = reaction rate of species i (S or N); Ki, KH2 , KH2S= adsorption equilibrium constants of species i, H2 and H2S;

ki = apparent rate constant.

The LH model assumes that both HDS and HDN reactions are

irreversible and proceed according to a pseudofirst-order of reac-

tion [5759]. It is noteworthy to mention that the LH type of rate

equation for representing the hydrogenation kinetics of industrial

feedstocks is complicated; and the fact that too many coefficientsare involved to be determined makes it quite a challenging task to

undertake [58,59]. However, using Maple V software, the solution

to Eq. (10) can be obtained as:

Ci(t) =

(1+ KH2PH2+ KH2SPH2S)LambertW

Kiexp

t+

Ci0Kiln(Ci0)+ln(Ci0 )KH2 PH2+ln(Ci0 )KH2SPH2 SKH2

PH2Kiki

KH2

PH2Kiki

1+KH2 PH2+KH2SPH2 S1+KH2PH2+KH2SPH2 S

Ki(10a)

where

Lambert W(x)

=x

x2

+

3

2

x3

8

3

x4

+

125

4

x5

54

5

x6

+(0)7 (10b)

Table 6

Calculated values ofkS , kN , KN , KS, and KH2 Sfrom theLH model.

HDN reaction

T(K) kN KN KH2 KH2 S

633.15 0.15 1.62 1.67 105.01

648.15 0.35 1.48 1.55 98.75

661.15 0.73 1.40 1.48 92.50

673.15 1.05 1.31 1.40 86.25

693.15 1.67 0.97 1.16 78.00

HDS reaction

T(K) Ks Ks KH2 KH2 S

633.15 0.15 4.77 2.77 105.01

648.15 0.42 4.00 2.40 98.50

661.15 0.72 3.32 2.06 92.00

673.15 1.19 2.62 1.75 85.00

693.15 1.80 2.00 1.50 78.00

and

x =

Ki exp

t+

Ci0Kiln(Ci0)+ln(Ci0 )KH2 PH2+ln(Ci0 )KH2SPH2SKH2

PH2Kiki

KH2

PH2Kiki

1+KH2 PH2+KH2 SPH2S

1+ KH2PH2+ KH2SPH2S(10c)

Excel solver was used to solve Eq. (10). The values ofki, Ki,KH2 ,

and KH2S were obtained by rigorous iterative procedure. The calcu-lated parameters for HDS and HDN are given in Table 6. The values

of calculated Ci predicted by the model were in agreement with

those obtained from experimental data used. It is observed from

this table that kS and kN increased with the increase in temper-

ature; i.e., the rates of HDS and HDN increased with the increase

in temperature. Moreover, Fig. 11(a and b) shows that KH2S, KH2 ,KS, and KNdecreased with the increase in temperature. This trend

indicates that increase in temperature decreased the inhibition of

these parameters on HDN and HDS reactions. The activation ener-

gies from this model were calculated from the Arrhenius plot in

Fig. 12. The activation energies for HDN and HDS reactions werefound to be 165.8 and 147.2kJ/mol, respectively.

3.3.3. The multi-parameter model

The MP kinetic model is similar to the generic PL model;

however, to improve the degree of accuracy of kinetic parameters


9/11






1/T*1000 (K-1

)

1.601.561.521.481.44

ln(kS,KS,K

H2,KH2S

)

-4

-2

0

2

4

6a

b

kS

KS

KH2

KH2S

1/T*1000 (K-1

)

1.601.561.521.481.44

ln(kS,KS,KH2,KH2S

)

-4

-2

0

2

4

6

kS

KS

KH2

KH2S

Fig. 11. (a) Plots of constants kS , KS , KH2 , and KH2 S from the LH model for

HDS reaction. (Catalyst= 5cm3, T= 633693K, P=8.8MPa, LHSV=1 h1 and H2/oilratio = 600mL/mL).

(b)Plotsof constantskN ,KN ,KH2 ,andKH2 Sfromthe LHmodelfor HDSreaction.(Cat-

alyst=5cm3, T= 633693K,P=8.8MPa,LHSV=1h1 andH2/oilratio= 600mL/mL).

obtained from the PL model, additional hydrotreating operating

conditions namely hydrogen partial pressures and gas/oil ratio are

taken into consideration in the overall rate expression as well asLHSV and temperature. Obviously, this model would give a better

representation of the kinetics of the hydrotreating reaction due to

the fact that the effect of all process variables can be observed. The

multi-parameter model is shown below in Eq. (11) [42,59]:

ri= dCidt = ki PmH2 Ci

G

O

q(11)

The solutions for Eq. (11) for different values ofn are given

below:

lnCfCp =

ko e(s/T) PmH2 (G/O)q

(LHSV)c ; n = 1 (11a)

1/T*1000 (K-1 )

1.601.561.521.481.44

ln

(kN,kS

)

-2

-1

0

1

kS

kN

9665.0

234.26701.17

2=

+=

R

xy

9587.0

936.26118.18

2=

+=

R

xy

Fig. 12. Arrhenius plot of HDS and HDN rate constants obtained from the

LH model. (Catalyst= 5cm3 , T= 633693K, P=8.8MPa, LHSV=1h1 and H2/oilratio= 600mL/mL).

1

n 1

1

Cn1p 1

Cn1f

=

ko e(s/T) PmH2 (G/O)q

(LHSV)c ; n > 1

(11b)

where s=Ea/R, Ea = activation energy; R= gas constant, m, q, and

c= empirical regression factors; PH2 = reactor pressure; G/O = gas-to-oil ratio; and all other parameters have their usual meaning.

Kinetic parameters predicted for the HDS and HDN reactions by

the multi-parameter model are compiled in Table 7. Experimental

data for the FeW/SBA-15 catalyst was analyzed using the non-

linearregression methodin Polymath 5.1software under extended

experimental conditions of temperatures, pressures, LHSVs, and

gas/oil ratios of 633693 K, 7.69.6 MPa, 0.52h1

, and 4001000,respectively. The activation energies for HDS and HDN reactions

were computed to be 126.7kJ/mol and118.8 kJ/mol, with predicted

reaction orders of 2.2 and 1.8, respectively. R2 for HDS and HDN

were 0.96 and 0.88, respectively. The value ofR2 adjusted tells how

well themodelcouldbe used to predict sulfurand nitrogen product

distributions under conditions not experimented, and were found

to be 0.94 and 0.83, respectively.

3.4. Comparison of results from the various models studied

Compilation of the activation energies and Arrhenius constants

obtained from the PL, LH, and MP models can be found in

Table 8. It can be observed from this table that for HDN and HDS

reactions activation energies from the LH model are higher than

those from the PL model. This could be due to the fact that in the

former model adsorption of sulfur, nitrogen, hydrogen, and H2S

Table 7

Multi-parameter model results for HDS and HDN reactions.

Parameter HDS HDN

n 2.2 1.8

k 722 12.73

s 1.52103 1.43103m 3.39 4.18

q 1.19 1.54

c 0.92 0.67

R2 0.96 0.88

R2adj 0.94 0.83


10/11






Table 8

Comparison of the PL, LH,and MP models.

PL model LH model MP model

n Ea (kJ/mol) ko

* n Ea (kJ/mol) ko

* n Ea (kJ/mol) ko

*

HDS 2.0 129.6 4.18E10 Pseudo 147.2 2.47E11 2.2 126.7 722

HDN 1.5 150.6 7.63E12 Pseudo 165.8 4.99E11 1.8 118.8 12.73

Activation energy*

Pre-exponential factor

were considered in the model development. Also, in this model it

was assumed that H2S inhibits HDN and HDS reactions (Botchwey,

[55]). Higher activation energies for HDN and HDS reactions from

the LH model than that obtained from the power law model indi-

cate that nitrogen and H2S adsorptions have significant inhibition

effects on HDN and HDS. This suggests that the inhibitive species

such as H2S tend to increase the minimum activation energy bar-

rier required forreactants to overcome so as to form products.With

regard to the models investigated, the computed minimum amount

of energy increases in order of MP


11/11

Pleasecite this article in press as:P E Boahene et al Hydroprocessingof heavy gasoils using FeW/SBA-15 catalysts:Experimentals optimization




[22] P.E. Boahene, K. Soni, A.K. Dalai, J. Adjaye, Applied Catalysis B: Environmental101 (2011) 294305.

[23] Y.Okamoto, K.Ochiai,M. Kawano, K. Kobayashi, T. Kubota, Applied CatalysisA226 (2002) 115127.

[24] R. Prins,V.J.H. de Beer, G.A. Somorjai,Catalysis Reviews, Science andEngineer-ing31 (1989) 141.

[25] S. Sigurdson, A.K. Dalai, J. Adjaye, CanadianJournal of Chemical Engineering89(2011) 562575.

[26] S. Damyanova, L. Petrov, M.A. Centeno, P. Grange, Applied Catalysis A: General224 (2002) 271284.

[27] M.S. Rana, E.M.R. Capitaine, C. Leyva, J. Ancheyta, Energy and Fuels 86 (2007)

12541262.[28] D. Li, A. Nishijima, D.E. Morris, Journal of Catalysis 182 (1999) 339348.[29] T. Klimova, M. Calderon, J. Ramirez, Applied Catalysis A: General 240 (2003)

2940.[30 ] K. S on i, P. E. Bo ahe ne, A.K. Dalai, Pro duction an d pur ificatio n of ultr -

aclean transportation fuels, ACS Symposium Series (2011) 1529,http://dx.doi.org/10.1021/bk-2011-1088.ch002.

[31] T. Chiranjeevi, P. Kumar, M.S. Rana, G.M. Dhar, T.S.R. Prasada Rao, Journal ofMolecular Catalysis A: Chemical 181 (2002) 109117.

[32] K. Soni, B.S. Rana, A.K. Sinha, A. Bhaumik, M. Nandi, M. Kumar, G.M. Dhar,Applied Catalysis B: Environmental 90 (2009) 5563.

[33] K. Soni, K.C. Mouli, A.K. Dalai, J. Adjaye, CatalysisLetters 136 (2010)116125.[34] G.M. Dhar,G.M. Kumarana, M. Kumara,K.S. Rawat,L.D. Sharma,B.D. Raju, K.S.R.

Rao, Catalysis Today 99 (2005) 309314.[35] D. Zhao, J. Feng, Q. Huo, N. Melosh, G.H. Fredrickson,B.F. Chmelka, G.D. Stucky,

Science 279 (1998) 548552.[36] M. Kruk, L. Cao, Langmuir 23 (2007) 72477254.[37] N. Brodie-Linder, G. Dosseh, C. Alba-Simonesco, F. Audonnet, M. Imperor-Clerc,

Materials Chemistry and Physics 108 (2008) 7381.

[38] J. S un, H. Zh ang , D. Ma, Y. Chen, X. Bao, A. Klein-Hoffmann, N. Pfanderb, D.Sheng, Chemical Communications (2005) 53435345.

[39] J.H. Gary, G.E. Handwerk, M.J. Kaiser, Petroleum Refining, 5th ed., 2004, pp.195197.

[40] L.Vradman,M.V.Landau, M. Herskowitz, V. Ezersky,M. Talianker, S. Nikitenko,Y. Koltypin, A. Gedanken, Journal of Catalysis 213 (2003) 163175.

[41] V. Sundaramurthy, I. Eswaramoorthi, A.K. Dalai, J. Adjaye, Microporous Meso-porous Materials 111 (2008) 560568.

[42] A. Duan, X. Chun-ming, L. Shi-xing, H.C. Keng, Chinese Journal of ChemicalEngineering 11 (2005) 743746.

[43] S.K. Bej, A.K. Dalai, J. Adjaye, Energyand Fuels 15 (2001) 377383.[44] S.K. Bej, A.K. Dalai, J. Adjaye, Petroleum Science and Technology 20 (2002)

867877.[45] D. Ferdous,A.K. Dalai, J. Adjaye,Industrial andEngineering ChemistryResearch

45 (2006) 544552.[46] M. L. Mapiou r, MSc. Th esis, Unive rs ity o f S askatch ewan , S askatoo n,

2009.

[47] C. Botchwey, A.K. Dalai, J. Adjaye, Industrial and Engineering ChemistryResearch 43 (2004) 58545861.

[48] M.J.Girgis,B.C. Gates, Industrial andEngineeringChemistry Research30 (1991)20212058.

[49] S.K. Bej, A.K. Dalai, S.K. Maity, Catalysis Today 64 (2001)333345.[50] A. Gruia, D.S.J. Jones, P.R. Pujado, Handbookof Petroleum Processing, Springer,

Netherlands, 2006, Chapter 8.[51] H.S. Fogler, Elements of Chemical Reaction Engineering, 4th ed., 2006, pp.

6869.[52] H. Topsoe, B.S. Clausen, F.E. Massoth, Hydrotreating Catalysis Science and

Technology, Springer-Verlag, New York, 1996.[53] H.S. Fogler, Elements of Chemical Reaction Engineering, 4th ed., 2006, pp.

647648.[54] J.G. Speight, The Desulfurization of Heavy Oils and Residua, Marcel Dekker Inc.,

NewYork,2000.[55] C. Botchwey, MSc. Thesis, University of Saskatchewan, Saskatoon, 2003.[56] C. Botchwey, A.K. Dalai, J. Adjaye, International Journal of Chemical Reactor

Engineering 4 (2006) A20.[57] C.P. Chu, K.M. Ng, AIChE Journal 35 (1989) 148158.

[58] M. Egorova, R. Prins, Journal of Catalysis 224 (2004) 278287.[59] D.D. Whitehurst, I. Takaaki, I. Mochida, Advances in Catalysis 42 (1998)

344368.[60 ] M. Mapiour, V. Sundaramurthy, A.K.Dalai,J. Adjaye,Energyand Fuels 24(2010)

772784.
http://localhost/var/www/apps/conversion/tmp/scratch_2/dx.doi.org/10.1021/bk-2011-1088.ch002http://localhost/var/www/apps/conversion/tmp/scratch_2/dx.doi.org/10.1021/bk-2011-1088.ch002

Documents

Hydroprocessing of heavy gas oils using FeW.pdf