sba-15poredia
TRANSCRIPT
-
8/4/2019 SBA-15Poredia
1/10
Applied Catalysis A: General 402 (2011) 3140
Contents lists available at ScienceDirect
Applied Catalysis A: General
journa l homepage : www.elsevier .com/locate /apcata
Application ofdifferent pore diameter SBA-15 supports for heavy gas oil
hydrotreatment using FeW catalyst
Philip E. Boahene a, Kapil K. Sonia, Ajay K. Dalaia,, John Adjaye b
a Catalysis and Chemical ReactionEngineering Laboratories, Department ofChemical Engineering, University ofSaskatchewan, Saskatoon, SK, S7N5A9, Canadab Syncrude Edmonton Research Centre, Edmonton,AB, T6N1H4, Canada
a r t i c l e i n f o
Article history:
Received 5January 2011Received in revised form 6 May 2011
Accepted 8 May 2011
Available online 13 May 2011
Keywords:
SBA-15
Pore diameter
FeW catalyst
Heavy gas oil
HDN
HDS
Hydrotreating
a b s t r a c t
This work focuses on utilizing mesoporous SBA-15 materials of different pore diameters as poten-
tial hydrotreating catalyst supports for heavy gas oil (HGO). Hexane was used as swelling agent for
the preparation of variable pore diameter SBA-15 materials. Four kinds of SBA-15 supported FeW
catalysts with different pore diameters in the range of 520 nm were prepared and designated as
Cats-A to D. The aqueous co-impregnation technique was employed for preparation of the catalysts.
The supports were characterized by several techniques including X-ray powder diffraction (XRD)
and N2 adsorptiondesorption isotherms. The SBA-15 supported FeW catalysts were characterized by
ICP-MS, BET surface area analysis, powder XRD, transmission electron microscopy (TEM), scanning
electron microscopy (SEM) and CO chemisorption. Results from XRD profiles, TEM images, and N2adsorptiondesorption isotherms confirmed the presence of highly ordered two-dimensional hexag-
onal structure with cylindrical arrays of pores. The structural integrity of the samples was preserved
even after loading of2 wt.% Fe and 15wt.% W. Hydrotreating experiments were conducted using bitu-
men derived heavy gas oil under industrial conditions oftemperature, pressure, LHSV, and gas to oil ratio
of 375400 C, 8.8 MPa, 1 h1, and 600 mL/mL, respectively. The SBA-15 supported catalyst with pore
diameter of10 nm (Cat-B) was the best among the supports studied for FeW catalysts, probably due to
sufficient mass transfer ofreactant liquids and gases through the catalysts pores while still maintaining
a high surface area necessary for metal dispersion. 2011 Elsevier B.V. All rights reserved.
1. Introduction
The hydrotreatment of petroleum feedstocks mainly aims to
produce clean gasoline and diesel with low sulfur, nitrogen and
aromatic contents. More stringent environmental regulations have
required improved efficiency of these processes [1]. Industrially,
the most commonly usedhydrotreatingcatalystis the-Al2O3 sup-
ported molybdenum sulfide, MoS2, (1020 wt.%),which is typically
promoted with cobalt or nickel (35 wt.%). However, the tradi-
tional -Al2O3 support has a fairly broad pore size distribution
with a considerable portion ofthe total surface area in small pores,
specifically, pores less than 5 nm in size [2]. The pore diameter
of the catalyst support plays a crucial role in its overall catalytic
performance by enhancing diffusion ofreactant molecules to cat-
alytic active sites predominantly located inside the pores [1]. As
in the case of-Al2O3 support, active metals in such pores may
not be readily accessible to bulkier molecules (such as the alkyl-
substituted dibenzothiophenic compounds) and may also suffer
Corresponding author. Tel.: +1 306 966 4771; fax: +1 306 966 4777.E-mail address: [email protected](A.K. Dalai).
severe sintering in the pores as a result of exposure of these
dispersed nanocrystalline materials to high hydrotreating temper-
atures.
The issue ofrapid initial catalyst deactivation is of prime con-
cern with commercial -Al2O3 supports due to pore plugging of
smaller pores (
-
8/4/2019 SBA-15Poredia
2/10
32 P.E. Boahene et al. / Applied Catalysis A: General 402 (2011) 3140
have been investigated. These include clays [5], carbon [6],
oxides [7,8], mixed oxides [9,10], zeolites [11] and mesoporous
materials like MCM-41 [12], HMS [13], SBA-15 [14], and KIT-6
[15,16].
Ordered mesoporous solids such as HMS, MCM-41 and MCM-
48 [17] have great potential as hydrotreating catalyst supports.
They have several advantages over conventional catalyst supports,
e.g., high surface (area around 1000m2/g) that can provide bet-
ter distribution ofactive metal components and more adsorption
sites, and a large tunable mesopore size (ranging from 5 to 30 nm).
Their large mesoporous channels can overcome the steric hin-
drance and facilitate the diffusion process; thus allowing the large
aromatic molecules to crack into small ones through the acidic cen-
ter. Corma et al. synthesized NiMo/MCM-41 catalysts for HDT and
the results showed that these catalysts were active in comparison
to other supports (like USY or precipitated silicaalumina) [18].
These materials have already been tested for mild hydrocracking
[18], aromatic hydrogenation [19], and hydrodesulfurization (HDS)
ofdibenzothiophene and petroleumresides [20,21]. It has beenalso
reported that Mo on MCM-41 is a good demetallation catalyst [22].
However, hydrodesulfurization activities obtained with NiMo cat-
alysts supported on purely siliceous MCM-41 were less than their
alumina-supported counterparts [20].
SBA-15 type ofordered mesoporous material, obtained by using
a triblock copolymer as structure-directing agent under strongly
acidic conditions, has attracted considerable attention in the fields
of heterogeneous catalysis and nanoscale materials [23]. SBA-15
possesses a high surface area (6001000 m2/g) and is formed by
a hexagonal array ofuniform tubular channels with tunable pore
diameters in the range of530 nm, obviously larger than those of
MCM-41 and HMS. The high surface area and controlled pore size
ofMCM-41 and SBA-15 make them promising supports for appli-
cation in oil refining. Pore wall thicknesses ofaround 36nm and
2D hexagonal packing ofSBA-15 provide higher thermal stability
than displayedby MCM-41, which consists ofhexagonal packing of
one-dimensional (1D) channels with p6mm symmetry [24]. Com-
paredto MCM-41 type ofmesoporous materials, SBA-15 has proven
efficient for the HDS ofdibenzothiophene (DBT) [25], probably dueto their desirable textural properties which enhance the relatively
easier access ofreactant molecules into its pores; thus increasing
the rate ofhydrotreating reactions.
Many works have dealt with the preparation of supported
and unsupported molybdenum-based catalysts [26,27]. However,
the NiW sulfide bulk systems are less studied [28,29]. W-based
catalysts are known to have a more hydrogenating character
than their Mo counterparts. Thus, it is interesting to analyze the
performance of W-based catalysts, in particular W/SBA-15 cat-
alysts, during the HDS and hydrodenitrogenation (HDN) of real
feedstocks with high contents of refractory sulfur and nitrogen
compounds. These molecules can be efficiently removed mainly
via the hydrogenationdesulfurization route.
An enormous amount of knowledge has been accumulated onthe crucial role of cobalt and nickel promoters in the course of
90 years. However, very little studies have been done using Fe
as a promoter in hydrotreating reactions. In lieu of the deduc-
tions made from the study ofMo/MCM-41 catalyst system, which
was promoted by Fe and Ni, for HDS of DBT and hydrogenation
(HYD) of2-methylnaphthalene reactions, Linares et al. [30] noted
that the incorporation ofFe possibly induced Bronsted acidity that
enhancedthe performance ofsucha catalyst.That notwithstanding,
in the comparison oftypical promoters employed in the prepara-
tion ofHDT catalysts, one can accede to the fact that the cost ofFe
per kilogram is less expensive as compared to Ni and Co counter-
parts [1]. Hence, one has to consider the economical benefits to be
derived when Fe is used as a promoter in the hydrotreating catalyst
system formulation.
The aim of this work is to investigate the effect of applying
different pore diameter SBA-15 supported FeW catalysts for the
hydrotreatment ofheavy gas oil derived from Athabasca bitumen.
To the best of our knowledge, this type ofstudy has not yet been
reported in the literature for real feedstocks. In this investigation,
four mesoporous silicaSBA-15 supported catalystswith porediam-
eters in the 520 nm range were prepared and have been used
for hydrotreating ofheavy gas oil (HGO) under industrial process
conditions.
2. Experimental
2.1. Preparation ofsupports and catalysts
The siliceous SBA-15 materials were synthesized using hex-
anes as a micelle expander under acidic conditions, according
to the procedure described elsewhere [31,32]. The triblock
copolymer Pluronic P123 (Mav = 5 800, EO20PO70EO20, Aldrich)
was used as the structure-directing agent (SDA) and tetraethyl
orthosilicate (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.
SBA-15 materials with different pore diameters were synthesized
by varying the molar ratio ofC6H14 and NH4F.In a typical synthesis
procedure, 9.8 g P123 and 0.109 g NH4F were dissolved in 335 mL
of1.3 M aqueous HCl solution at room temperature. This solution
was transferred to a constant temperature bath (CTB) maintained
at 15 C. After 1 h ofmechanical stirring, a mixture of20.8 g TEOSand 34.6 g C6H14 was added. The reaction was allowed to proceed
under mechanical agitation for 24 h in the CTB, after which the
gel formed was isolated and subjected to hydrothermal treatment
in a teflon-lined autoclave for 3 days. The solid product was
filtered, washed with deionized water, and dried for 24 h at room
temperature. To remove the organic template; the sample was
calcined at 550 C for 5 h using a heating rate of2 C/min.
2.2. Synthesis ofFeW/SBA-15 catalysts
FeW catalysts supported on SBA-15 with different pore diam-
eters were prepared by an aqueous impregnation technique.
The calcined SBA-15 support was impregnated successively
using aqueous solutions of ammonium metatungstate (AMT),
(NH4)6H2W12O40 (Fluka) andiron nitrate, Fe (NO3)39H2O(Aldrich)as aWandFe source, respectively. Aftereach impregnation, the cat-
alysts were dried at 100 C for 24 h. In a typical synthesis, 0.514 gAMT was dissolved in 15 mL de-ionized water until a homogenous
solution was formed. 2.55g siliceous SBA-15 was added to this solu-
tion, to get 15 wt.% W/SBA-15. This mixture was dried in an oven
at 100 C. 2 wt.% Fe was also loaded by using same approach. Theprepared catalysts are designated as Cat-A, Cat-B, Cat-C and Cat-D
with pore diameters 5 nm, 10nm, 15nm and 20nm, respectively.
2.3. Characterization
The calcined SBA-15 samples were characterized by small angle
X-rayscattering (SAXS)to ascertaintheir crystal structures. Diffrac-
tion patterns wererecorded witha Bruker Smart 6000 CCD detector
on a 3-circle D8 goniometer using a Rigaku RU 200 Cu rotating
anode generator fitted with parallel focusing cross-coupled mir-
rors anda0.5 mm pinhole collimator. Datawas obtainedusing a still
data collection (Bruker Software: SMART) with an exposure time of
300 s in the 010.0 range. Broad angle XRD patterns ofall the SBA-15 supported catalysts were recorded on a Rigaku diffractometer
using Cu K radiation.
Nitrogen physisorption isotherms were measured on a
Micromeritics ASAP 2000 analyzer at liquid nitrogen temperature
-
8/4/2019 SBA-15Poredia
3/10
P.E. Boahene et al. / Applied Catalysis A: General 402 (2011) 3140 33
of77 K. Prior to the analysis, the catalyst was out-gassed in vacuum
at 200 C until the static pressure remained less than 6.6104 Pa.The BrunauerEmmettTeller (BET) method was used to calculate
the surface area in the range ofrelative pressures between 0.05
and 0.30. The value of0.1620 nm2 was taken for the cross-section
ofthe physicallyadsorbedN2 molecule.The pore diameterand pore
size distributions were calculated from the adsorption and desorp-
tion branches ofthe isotherms using the BarrettJoynerHalenda
(BJH) method. The mesopore volume was determined from the
N2 adsorbed at a P/P0 =0.4. The total pore volume was calculated
from the amount ofnitrogen adsorbed at P/P0 = 0.95, assuming that
adsorption on the external surface was negligible compared with
adsorption in pores. In all cases, correlation coefficients above0.999
were obtained.
The morphological features ofthe support and catalysts were
studied from Transmission electron micrographs (TEM) obtained
with a Tecnai F20 at 200 kV. The powder samples were grounded
smoothly in an agate mortar and dispersed in heptane in an ultra-
sonic bath for several minutes. A few drops were then deposited
on 200 mesh copper grids covered with a holey carbon film. The
electron micrographs were recorded in electron negative films and
in a digital PC system attached to the electron microscope. Scan-
ning electron microscopy (SEM) micrographs were observed on a
Hitachi-S4700 microscope.
Diffuse reflectance infrared Fourier transform spectroscopy
(DRIFTS) of CO adsorption experiments were performed using
a PerkinElmer Spectrum GX instrument equipped with DTGS
detector and a KBr beam splitter was used for this analysis. Approx-
imately 10 mg of powdered catalyst sample was loaded in to a
sample cup inside a Spectrotech diffuse reflectance in situ cell
equipped with ZnSe windows and a thermocouple mount that
allowed direct measurement of the sample surface temperature.
Spectra for each experiment were averaged over 64 scans in the
region 40001000 cm1 with a nominal 4 cm1 resolution. Priorto the CO adsorption, the catalyst was in situ sulfided in the Spec-
trotech diffuse reflectancecell using 10% (v/v) H2S/H2 (50 cm3/min)
at 400 C for 2 h. At this temperature, the flow was switched to
He at a flow rate of50 cm3/min and the temperature decreased to30 C.The backgroundspectrumwas then recorded.The adsorptionprocess was carried out at 30 C by introducing CO (30 cm3/min)into the system for 30 min. After adsorption, the system was sub-
sequently purged with He at a flow rate of50 cm3/min for 30 min.
Spectra were then recorded under He flow. The background spec-
trum was subtracted from the post adsorption spectra.
The carbon monoxideuptake on sulfided catalystswas also mea-
sured using the Micromeritics ASAP 2000 instrument. Prior to the
CO chemisorption measurement, 0.2g of sample was in situ sul-
fided using 10 mol.% H2S in H2 at 400Cfor 4 h,and thenthe sample
was evacuatedat 120 C until the static pressure remained less than6.6104 Pa. The chemisorption was performed by passing pulsesofCO over the sample to measure the total gas uptake at 35 C.
Inductive coupled plasma (ICP-MS) analysis was utilized toquantify the metal composition in the catalysts. An energy-
dispersive X-ray analysis (EDAX) was also used for compositional
analysis using a system attached to the electron microscope which
was operated at 25 kV. The chemical composition determination
was based on the average analytical data ofindividual particles.
2.4. Catalytic activity
The heavy gas oil (HGO) feedstock used for this study was
derived from Athabasca bitumen and supplied by Syncrude Canada
Ltd. The feed properties are given in Table 1. Hydrotreating exper-
iments were conducted in a downward flow micro-trickle bed
reactor system (10mm ID; 285 mm length) using 5 mL ofcatalyst
under typical industrial process conditions. This reactor system is
Table1
Characteristics ofheavy 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 (C) 210.8FBP (C) 597.3
Boiling range(C)
IBP205 (Gasoline) (wt.%) 0
205260 (Kerosene) (wt.%) 1
260315 (Light gas oil) (wt.%) 5
315425 (Heavy gas oil) (wt.%) 39
425600 (Vacuum gas oil) (wt.%) 55
a high pressure and temperature reaction set-up whose operation
mimics industrial hydrotreaters. The set-up consists ofliquid and
gas feeding sections, a high pressure reactor, a heater with temper-
ature controller, a scrubber for removing the ammonium sulfide
from the reaction products, and a high pressure gasliquid sep-
arator. Details of catalyst loading into the reactor are described
elsewhere [33,34]. Typically, the catalyst bed was packed using5 mL ofcatalyst (1.5 g) and silicon carbide (SiC) particles as dilu-ents. Dilution ofthe 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. [34]
Catalyst sulfidation was carried out for a period of48 h at tem-
peratures of193 C for 24 h and 343 C for another 24 h. 2.9 vol.%butanethiol was used as a sulfiding agent in straight run gas oil
(VOLTESSO 35). An initial catalyst pre-wetting protocol was neces-
sary prior to catalyticactivity study. This procedure was carried out
by pumping 100 mL ofthe sulfiding solution at a high flow rate of2.5 mL/min intothe reactor. The high oil flow rate was subsequently
adjusted to 5 mL/h and maintained to obtain a liquid hourly space
velocity (LHSV) of1 h1. Furthermore, the hydrogen gas flow was
operated at a rate that corresponds to a gas/oil ratio of600 mL/mL.Following sulfidation, the catalyst was precoked at a tempera-
ture and pressure of375 C and 8.8 MPa, respectively, for 5 dayswith the real feed (HGO) at a flow rate of5 mL/h. After precoking,
HDN and HDS activity studies were conducted for 3 days each at
temperatures of 375 C, 388 C, and 400 C. The pressure, gas/oilratio and LHSV were maintained constant at 8.8 MPa, 600 mL/mL
and1 h1, respectively. Hydrotreated products were collected after24 h and stripped with nitrogen gas to remove dissolved ammo-
nia and hydrogen sulfide gases. A transient period of 24 h was
allowed throughout the experiments when there was a change
in process conditions. Liquid products taken within these tran-
sient periods were not analyzed. Samples were analyzed for total
nitrogen and sulfur contents using an Antek 9000 NS analyzer.
Total nitrogen content ofthe liquid product was measured by thecombustion/chemiluminence technique following ASTM D4629
method and the sulfur content was measured using the combus-
tion/fluorescence technique following ASTM 5463 method. The
instrumental error in N andS analysis was 3%.
3. Results and discussion
3.1. Support and catalyst characterizations
3.1.1. Elemental analysis
The elemental compositions ofcalcined FeW/SBA-15 catalysts
with different pore diameters were determined by ICP-MS and
EDX analysis. The results ofICP-MS with targeted compositions are
presented in Table 2. The elemental compositions obtained from
-
8/4/2019 SBA-15Poredia
4/10
34 P.E. Boahene et al. / Applied Catalysis A: General 402 (2011) 3140
Table2
Elemental composition, CO uptake, and HDS/HDN steady-state activities ofSBA-15 supported FeW catalysts with different pore diameter at various reaction temperatures
(Catalyst = 5 cm3, P=8.8 MPa,LHSV=1 h1 and H2 /oil ratio =600 (v/v)).
Sample ID Composition CO uptake (mol/g) Sulfur/nitrogen removal at each reaction temperature (wt.%)
Fe (wt.%) W (wt.%) 375 C 388 C 400 C
S N S N S N
Cat-A 2* (1.95) 15* (14.58) 17 44 9 54 14 64 26
Cat-B 2* (1.91) 15* (14.07) 25 45 17 53 17 66 33
Cat-C 2* (1.91) 15* (14.20) 15 31 11 38 12 62 17Cat-D 2* (1.95) 15* (14.44) 10 30 5 38 7 43 3
* Targeted
ICP-MS and EDX correlate well with each other, as well as with
targeted values.
3.1.2. Small-angle X-ray scattering analysis
The XRD measurements were carried out to study the meso-
porous structure of the supports and catalysts. The crystalline
phases ofthe calcined catalysts can also be obtained by this tech-
nique. The low angle XRD pattern ofFeW catalysts supported on
SBA-15 ofdifferent pore diametersare shownin Fig.1 in the interval
between the 2 values of0.510. The small-angle XRD pattern of
the calcined SBA-15 supportedFeW catalyst materials exhibit threewell-resolved peaks, characteristic ofSBA-15 materials [24]. These
XRD peaks are indexable as (d1 0 0), (d1 1 0) and (d2 0 0) reflections
associated with the p6mm symmetry of the hexagonal ordered
pore structure. Hexagonal order remains more or less intact after
pore diameter increase. The peaks corresponding to d1 1 0 and d2 0 0planes exhibit minor changes; revealing a slight structural change
ofSBA- 15 with pore swelling.
The unit-cell parameter (a0) was estimated from the position of
the (d1 0 0) diffraction line by using the equation a0 =d1 0 0 2/
3
and given in Table 3. Pore wall thickness () was assessed by sub-tracting Dads from the unit-cell parameter (a0), which corresponds
to the distance between centers ofadjacent mesopores.The notable
decrease in pore wall thickness suggests the continuous increase in
pore diameter (Table 3). Moreover, it can be observed that diffrac-tionpeaks shifted to lower 2 values, indicatingincrease in unit-cellsize (a0); corresponding to increase in pore diameter of SBA-15
materials [35].
3.1.3. Wide-angle X-ray diffraction
The wide angle X-ray diffraction patterns ofsiliceous SBA-15
and FeW/SBA-15 catalysts with similar metal loading but different
pore diameters recorded in the 2 interval of1080 are shown in
876543210
2
Intensity(a.u
)
Cat-A
Cat-B
Cat-C
Cat-D
Fig. 1. Low-angle X-ray patterns ofSBA-15 supported FeW catalysts.
Fig. 2. It can be seen from this figure that the presence ofcrystalline
phases were not detected in Cats-A to C; suggesting that either
metals werehomogeneouslydispersedon the support or their com-
positions were below the detection limit ofthe X-ray signals. The
absence ofXRDsignals in wide angleregion indicates that the parti-
cle size ofmetals is below the coherence length ofX-ray scattering
i.e., smaller than 4 nm; suggesting that metals were well dis-persed on the support due to the absence ofany crystalline phase.
However, the single broad hump exhibited by all samples and cen-
tered at 2 values of1540 is characteristic ofsiliceous materials
[36]. The high surface area ofsilica support favors dispersion oftheactive phases. However, in the case ofCat-D some crystalline peaks
can be observed due to the poorly crystalline species ofW and Fe
in the oxidic phase appearing at 2 values of23.5 and 34.2, respec-tively. The species observed at 2 value of23.5 can be assigned to
monoclinic WO3 crystalline phase (JCPDS card 35-609); while the
small and weak broad peak observed at 2 value of34.2 has beenattributed to-Fe, which manifests as small iron oxide crystallites
with a spinelstructure [37]. The poor dispersionofthe oxide phases
in Cat-D mayexplain the notable loss ofhydrotreating activity. This
is due to the fact that amongst the four catalysts prepared, Cat-D
has the largest pore diameter; however, its least surface area may
have contributed to the formation ofaggregates as detected by the
X-ray.
3.1.4. N2 adsorptiondesorption isothermsmeasurement
The textural properties ofmesoporous materials can be exam-
ined byN2 adsorptiondesorption isotherms at 77 K. The nitrogen
adsorptiondesorption isotherm ofall SBA-15 supported catalysts
with different pore diameters are shown in Fig. 3. The textural and
structural characteristics (surface area SBET, total pore volume VP,
pore diameter DP, unit-cell parameter a0 and pore wall thickness)
80706050403020102
Intensity(a.u
)
Cat-A
Cat-B
Cat-C
Cat-D
Fig. 2. Wide-angle X-ray diffraction patterns SBA-15 supported FeW catalysts.
-
8/4/2019 SBA-15Poredia
5/10
P.E. Boahene et al. / Applied Catalysis A: General 402 (2011) 3140 35
Table3
Physical properties ofSBA-15 supported FeW catalysts with different pore diameter determined from N2 sorption and XRD analysis.
Sample ID SBET (m2/g)Sup SBET (m
2/g)Cat NSBET VP (cm3/g) PDads (nm) d1 0 0 (nm) a0 (nm) (nm)
Cat-A 895 651 0.88 0.64 5.7 9.9 11.4 5.7
Cat-B 636 498 0.94 0.89 10.1 12.4 14.4 4.3
Cat-C 491 396 0.97 1.00 15.7 16.9 19.5 3.8
Cat-D 473 388 0.99 1.09 18.5 17.9 20.7 2.2
SBET, specific surface area calculated by the BET method; NSBET, (Normalized surface area) were calculated by using the equation, NSBET = (SBET ofthe catalysts)/(1x) SBETofthe support; VP , pore volume determined by nitrogen adsorption at a relative pressure of0.98; PDads, mesopore diameter corresponding to the maximum ofthe pore size
distribution obtainedfrom the adsorption isotherm by the BJH method;a0, unit-cell parameterdetermined from the positionofthe (1 0 0) diffraction line as a0 =d1 0 0 2/3;, pore wall thickness calculated as d1 0 0 =a0 Dads.
of the SBA-15 supports and corresponding catalysts are given in
Table 3.
All the FeW/SBA-15 catalysts exhibited the type IV isotherm
with H1 hysteresis loop, which is characteristic of a well-formed
SBA-15 material. The shape ofthe loop is unchanged after the Fe
and W metals loading indicating that support and catalysts exhibit
uniform textural porosity, which is also in agreement with XRD
results. The height ofthe hysteresis loop is decreased after metals
loading into SBA-15 due to a decreased pore volume indicating the
introduction ofmetal species within the mesopores ofthe support.
The surface area and pore volume ofthe SBA-15 decreased signif-
icantly after metal loading. Lizama and Klimova [38] investigatedthe incorporation ofactive metals (Mo or W) on the SBA-15 support
from either conventional precursors or heteropolyacid sources and
observed that W-containing samples have lower values ofsurface
area and pore volume than the corresponding Mo-containing ones
because the loading ofW metal is higher than that ofMo for main-
taining active metal-promoter ratios. The pore wall thickness (d)
decreases with increasing unit cell parameter (ao) which is mea-
sured by small-angle XRD. A decrease in the pore wall thickness
as observed was due to increase in the pore diameter ofthe sup-
ports. In the case ofCat-D, the pore wall thickness is considerably
less compared to all other catalysts, making it susceptible to easy
pore collapse under the severe hydrotreating reaction conditions.
The results from N2 adsorptiondesorption analysis correlates well
with that obtained from XRD analysis. The pore size distribu-tions ofFeW/SBA-15 catalysts are given in Fig. 4. The sharpness
of the desorption branches is indicative ofthe narrow mesopore
size distribution. However, as can be seen from Fig. 4, the pore
size distribution for Cat-D is not distinctively narrow. This could
be explained by the fact that majority of the pores are concen-
trated around 18 nm as evident in Cat-D. Nonetheless, it is a sound
argument to note that pores ofsmaller channels may be present,
which contributed to the rather broad pore size distribution
for Cat-D.
0
300
600
900
1200
1500
1.00.80.60.40.20.0
Relative pressure, P/P0
Volumeadsorbed(cm
3/g)
Cat-A
Cat-B
Cat-C
Cat-D
Fig. 3. Nitrogen adsorption isotherms ofSBA-15 supported FeW catalysts.
To confirm the presence ofnanoparticles ofFe and W oxides in
the pores ofthe SBA-15 substrate, the normalized SBET values were
calculated using an equation proposed by Vradman et al. [39]:
NSBET =(SBET ofthe catalysts)
(1 x) SBET ofthe support
where NSBET is normalized SBET and x is the weight fraction of
the phases. The values ofnormalized NSBET are given in Table 3.
Values ofNSBET close to unity give an indication of less pore
blockage. As seen in Table 3, the introduction of the Fe and W
particle caused in a reduction in pore volume, which resulted
in the reduction of normalized surface area. The normalizedsurface area of Cat-A was found to be 0.88, which is
-
8/4/2019 SBA-15Poredia
6/10
36 P.E. Boahene et al. / Applied Catalysis A: General 402 (2011) 3140
Fig. 5. Transmission electron micrograph ofSBA-15 supported FeW catalysts.
physisorption results. The WO3 nanoparticles are in the size range
of 35 nm. The average particle sizes determined by TEM images
are in good agreement with the corresponding crystalline sizes
calculated by the DebyeScherrer equation.
The SEM images of the pure SBA-15 support and supported
FeW catalysts showed that all samples contained fibre-like parti-
cles ofseveral tens ofmicrometers in length (Fig. 6A and B). Similar
morphologyofmesoporous SBA-15 materials has been reported by
other researchers [36].
3.1.6. DRIFT spectroscopy for CO adsorption
The nature ofactive species present on the catalysts in the sul-
fided form was studied by DRIFT spectroscopy of adsorbed CO
on these catalysts. DRIFT spectroscopy of CO adsorption on the
Fig. 6. Scanning electron micrograph of(a) SBA-15 support; and (b) FeW/SBA-15 catalyst.
-
8/4/2019 SBA-15Poredia
7/10
P.E. Boahene et al. / Applied Catalysis A: General 402 (2011) 3140 37
Fig. 7. DRIFT spectroscopy for CO adsorption ofSBA-15 supported FeW catalysts.
sulfided FeW catalysts with different pore diameters was per-
formed at 30 C to characterize the surface active sites. Fig. 7 showsDRIFT spectra ofabove the catalysts around CO stretching region.
CO adsorption on sulfided catalysts yields strong bands at 2102
and 2068 cm1. Although there is no literature on CO adsorptionstudies on FeW/SBA-15 catalysts, the literature available on NiMo-
supported catalysts shows two bands at 2098 cm1 correspondingto CO adsorption on unpromoted Mo sulfide sites, whereas the
band at 2062 cm1 is due to the Ni promoted Mo sulfide sites(NiMoS phase) [40,41]. The adsorption profile for FeW/SBA-15
shows similar results, indicating similar adsorption behaviors of
CO molecules on FeW catalysts supported on SBA-15. In the caseof Cat-C and Cat-D, the intense peak observed at 2102 cm1 cor-responds to unpromoted tungsten sulfides, which is indicative of
a higher number ofWS2 phase. However, Cat-A and Cat-B exhibit
a more intense peak at 2068 cm1, which suggests more adsorp-tion ofCO molecules at FeWS sites of FeW/SBA-15 catalysts. The
presence ofhigh FeWS centers in catalysts (Cat-A and Cat-B) show
a higher number of active centers ofsulfided Fe and W species
on the high surface area silica support. In the case of Cat-A and
Cat-B, there is a disappearance of the band ofcorresponding to
WS2. A similar observation was made in the case of Cat-C and
Cat-D in which the FeWS band is not available due to the over-
lapping with dominated phase. It can be deduced from the DRIFT
studies that catalysts with pore diameter in the range of510 nm
is responsible for higher number of co-ordinatively unsaturatedsites.
Carbon monoxide adsorption is also used to quantify exposed
surface metal atoms on sulfided catalysts. The CO uptake (mol/g
of Cat) measured from CO chemisorption is equivalent to the
number of active metal atoms that are accessible to the reac-
tant molecules. The amount ofadsorbed CO increased from Cat-A
to B; yielding a maxima in the case of Cat-B, then gradually
decreased to Cat-D. This could be explained by the fact that
as the pore diameter increased from Cat-B to D, exposed met-
als increasingly became inaccessible to CO molecules due to the
probable pore collapse as a result of increasing the pore diam-
eter. These results clearly indicate that Cat-B has the highest
number of active sites due to its highly dispersed metal phases
(Table 3).
20
40
60
7654321
Time on Stream (day)
Sulfur
conversion(%)
Cat-A Cat-B Cat-D Cat-CA
0
10
20
30
40
50
7654321
Time on stream (day)
Nitrogenconv
ersion(%)
Cat-A Cat-B Cat-C Cat-DB
Fig. 8. (A) Hydrodesulfurization (HDS) and (B) hydrodenitrogenation (HDN) activ-
ities of SBA-15 supported FeW catalysts during pre-coking with HGO at 375 C
(catalyst= 5 cm3, P=8.8 MPa, LHSV= 1 h1 and H2/oil ratio =600 (v/v)).
3.2. Catalytic activity ofdifferent pore diameters FeW/SBA-15catalysts
To determine the optimum pore diameter for FeW/SBA-15
catalysts, four types of FeW/SBA-15 catalysts with varied pore
diameters (5, 10, 15, and 20nm) were prepared and designated
as Cat-A, Cat-B, Cat-C and Cat-D, respectively. By varying the
molar ratio ofhexane (micelle expander) and P123 (template), the
catalysts with inner diameters ranging from 5 to 20 nm were syn-
thesized. The intended Fe and W loadings for these catalysts was
consistent with most commercial NiMo/-Al2O3 catalysts at 2.5
and 13.0 wt.%, respectively.
3.2.1. Effect ofprecoking on sulfur andnitrogen conversion
Precoking of catalysts is necessary to stabilize its initial highactivity. Fig. 8A and B show the effect of precoking on sulfur and
nitrogen conversion ofFeW/SBA-15 catalysts with different pore
diameters. At the start-of-run (SOR) ofthe hydrotreating reactions,
the sulfided catalysts had very high initial activity which dropped
suddenly withtime-on-stream (TOS).The decreasein HDS andHDN
activities was observed in the following order: Cat-A=Cat-B > Cat-C > Cat-D. It can also be observed that the stability ofCats-A and B
were better than that of Cats-C and D. Cat-C and Cat-D had high
initial HDS and HDN activities at the start-of-run, but could not be
maintained and subsequently dropped sharply as the reaction pro-
gressed (within 48 h period). This observation could be attributed
to their larger pore sizes as compared to that of Cats-A and B.
This range ofpore size is large enough for unhindered diffusion
of a larger portion of low-molecular-size asphaltenes and other
-
8/4/2019 SBA-15Poredia
8/10
38 P.E. Boahene et al. / Applied Catalysis A: General 402 (2011) 3140
Fig. 9. (A) Hydrodesulfurization (HDS) and (B) hydrodenitrogenation (HDN) activ-
ities of SBA-15 supported FeW catalysts with HGO at different temperatures
(catalyst= 5 cm3, P=8.8 MPa,LHSV=1 h1 and H2/oil ratio =600 (v/v)).
heteroatom-containing species that can undergo various reactions
on the catalyst surface at the catalytically active centers within the
pores. Some products formedin the reaction, such as carbonaceous
species, can block the active centers. Consequently, deterioration
in hydrotreating activities in the larger pore supports is observed
within the initial 48 h period. A similar study conducted by Callejas
et al. [42] on the long-term (7400h) hydroprocessing ofpetroleum
residue usingcommercial NiMo/Al2O3 catalyst concluded that cokedeposition occurs rapidly on the catalyst surface during the early
hoursofthe run (100 h),reachingas high as 12.4 wt.% ofthe catalyst.
3.2.2. Effects ofpore size on sulfur andnitrogen conversion
Theability offeedstock molecules to diffuse to the active centers
within the pores is a key factor for determining the effectiveness of
the catalyst [43]. Increasing the pore diameter ofthe catalyst facili-
tates diffusion ofbulkier molecules from the bulk fluid to the active
sites. However, this phenomenon solely does not necessarily trans-
late into higher hydrotreating activity. In the case ofFeW/SBA-15,
the results in Table 2 and Fig. 9 show that HDS and HDN activi-
ties generally pass through a maximum with respect to increasing
the pore diameter in the range of5.718.5 nm. The HDS activity
remained almost the same for Cats-A and B ofpore diameters 5.7
and 10.1 nm, respectively. This highest HDS activity recorded for
these catalysts declined as the pore diameter was further increased
from 10.1 to 15.7 nm and, subsequently, to 18.5 nm. However, HDN
activity increased slightly with pore diameter(5.7 and 10.1nm) and
declined as the pore diameter further increased to 18.5 nm. This
observed trend could be explained by the fact that, past a certain
optimal pore diameter, increasing the pore diameter resulted in a
decrease in surface area.It is worth mentioningthat the surface area
of a catalyst offers the platform for the dispersion ofactive metal
components.However, the effective surface area required for metal
dispersion suffered at the expense ofincreasing pore diameter. The
decrease in surface area was responsible for the decrease in HDS
and HDN conversions as the pore diameter increased. Thus, it can
be concluded that by loading the same amount of metals (2 wt.%
Fe and 15wt.% W) on SBA-15 supports ofdifferent pore diameters
resulted in an inhomogeneous dispersion ofthe metals on highest
pore diameter catalysts (i.e., having the lowest surface area). This
result was also confirmed by broad-angle XRD spectra showing the
crystalline peaks ofmetal species in Cat-D.
Furthermore, this observation could also be explained by the
asphaltene content present in the feedstock, which was found to be
1.55 wt.% (Table 1). The amount ofasphaltene present in the HGO
is quite significant as compared to the amount of catalyst loaded
in the micro reactor (1.5 g); since it only requires a little amount
ofasphaltenes to significantly deactivate the catalyst. Due to their
larger molecular weights [44], asphaltenes will preferentially dif-
fuse andreact in larger pores than they would in smaller pores. This
resulted in the production ofsmaller hydrocarbons which are the
main precursors for carbonaceous species deposition on the cata-
lyst active sites. This conversion ofasphaltene molecules to coke
is a plausible explanation for the decreased HDS and HDN activ-
ities in the larger pore diameter catalysts. A similar explanation
was provided by Song et al., who studied the effect of pore struc-
ture ofNi-Mo/Al2O3 catalysts in hydrocracking ofcoal and oil sand
derived asphaltenes[45]. They concluded that as the pore diameter
increases, heavier fractions (mostly the asphaltenes) are strongly
adsorbed on the catalyst surface, inhibiting the adsorption ofless
heavier molecules.This conclusion was also supportedby Diez et al.[46], who showed that a hydrotreating catalyst with large pores
(>14 nm) was more susceptible for asphaltene decomposition.
On the basis of the catalyst pore sizes (520 nm) andhydrotreating results, it can be suggested that an optimum pore
diameter ofapproximately 10 nm was effective for both HDS and
HDN of the HGO. For the HDS activity, a study conducted by
Inoguchi [47] on the control ofpore size ofsupports concluded that
optimalporediametershiftedfrom 10 nm for HDS ofpetroleumdis-
tillates to 15 nm for HDS ofthe residue. Furthermore, Angevine and
Fischer reported that for residue HDS, a small-pore (with 310 nm
pores) catalyst was selective for non-asphaltene sulfur; while a
large-pore (with 1030 nm pores) catalyst was selective for asphal-
tene sulfur [48]. Thus, it may be noted that the present results
in Fig. 9 for the HDS and HDN of heavy gas oil fractions are con-sistent with that reported in the investigation by Takeuchi, using
-Al2O3-supported NiMo catalyst [49].
3.2.3. Effects oftemperature on sulfur and nitrogen conversion
An easy and cost-efficient way ofincreasing hydrotreating con-
versions is by manipulating process temperature. However, an
excessively high operating temperature may lead to activity loss
and shortening ofcatalyst life [50]. Thus, the effect oftemperature
on sulfur and nitrogen conversion was studied. It is well known
that during hydrotreating reactions, rapid catalyst deactivation (or
ageing) occurs due to sintering and coke formation on the cat-
alyst surface [51]. In order to maintain a stable catalyst activity
and desired product quality, operating temperature is gradually
raised to compensate for catalyst deactivation. Thus, the effect of
-
8/4/2019 SBA-15Poredia
9/10
-
8/4/2019 SBA-15Poredia
10/10
40 P.E. Boahene et al. / Applied Catalysis A: General 402 (2011) 3140
[42] M.A. Callejas, M.T. Martinez, T. Blasco, E. Sastre, Appl. Catal. A 218 (2001)181188.
[43] M.Ternan,J. Can,Chem. Eng. 61 (1983) 689696.[44] T. Takahashi, H. Higashi, T. Kai, Catal. Today104 (2005) 7685.[45] C.Song,T. Nihonmatsu,M. Nomura,Ind. Eng. Chem.Res.30 (1991) 17261734.[46] F. Diez, B. Gates, J. Miller, D. Sajcowski, S. Kukes,Ind. Eng. Chem. Res. 29 (1990)
19992004.
[47] M. Inoguchi, Shokubai 20 (1978) 144154.[48] P.J. Angevine, R.H. Fischer, Appl. Catal. 27 (1986) 203275.[49] C. Takeuchi, Kagaku Kogaku50 (1986) 598603.[50] S.E. Moschopedis, S. Parkash, J.G. Speight, Energy Fuels 57 (1978)
431434.[51] J.G. Speight,The Desulfurization ofHeavy Oilsand Residua,Marcel Dekker, New
York, 2000.