Studies on MCM-48 supported cobalt catalyst for Fischer-Tropsch synthesis

Article abstract of DOI:10.1016/j.molcata.2005.08.050

MCM-48 molecule sieve was used as a support of cobalt catalyst for Fischer-Tropsch synthesis (FTS). Co/MCM-48 catalysts were prepared using incipient witness impregnation method with cobalt loadings of 5, 10 and 15 wt.%, respectively. The catalysts were characterized by N₂ physisorption, X-ray diffraction (XRD), temperature programmed reduction (TPR), hydrogen temperature programmed desorption (H₂-TPD) followed by pulse oxygen titration and transmission electron microscopy (TEM). The catalytic properties for FTS were tested in a fixed bed reactor. TPR and oxygen titration indicated different extent of overall reduction of Co cobalt species in different cobalt content catalysts. With increasing Co loading, CO conversion and C5+ selectivity first increased remarkably and then remained nearly invariable. Lower reducibility of smaller cobalt particles on 5Co/MCM-48 is likely to be one of the reasons responsible for the lower CO conversion and highest methane selectivity. Co loading exceeding 10 wt.% had no significant effect on FTS properties of the catalysts.

Full text of DOI:10.1016/j.molcata.2005.08.050

Journal of Molecular Catalysis A: Chemical 244 (2006) 33–40
Studies on MCM-48 supported cobalt catalyst for
Fischer–Tropsch synthesis
Hualan Lia, Shuguo Wanga, Fengxiang Lingb, Jinlin Lia,∗
a Hubei Key Laboratory for Catalysis and Material Science, College of Chemistry and Material Science,
South-Central University for Nationalities, Wuhan 430074, China
b Fushun Research Institute of Petroleum and Petrochemicals, Fushun 113001, China
Received 4 July 2005; received in revised form 30 August 2005; accepted 30 August 2005
Available online 4 October 2005
Abstract
MCM-48 molecule sieve was used as a support of cobalt catalyst for Fischer–Tropsch synthesis (FTS). Co/MCM-48 catalysts were prepared
using incipient witness impregnation method with cobalt loadings of 5, 10 and 15 wt.%, respectively. The catalysts were characterized by N2
physisorption, X-ray diffraction (XRD), temperature programmed reduction (TPR), hydrogen temperature programmed desorption (H2-TPD)
followed by pulse oxygen titration and transmission electron microscopy (TEM). The catalytic properties for FTS were tested in a fixed bed reactor.
TPR and oxygen titration indicated different extent of overall reduction of Co cobalt species in different cobalt content catalysts. With increasing
Co loading, CO conversion and C5+ selectivity first increased remarkably and then remained nearly invariable. Lower reducibility of smaller cobalt
particles on 5Co/MCM-48 is likely to be one of the reasons responsible for the lower CO conversion and highest methane selectivity. Co loading
exceeding 10 wt.% had no significant effect on FTS properties of the catalysts.
© 2005 Elsevier B.V. All rights reserved.
Keywords: Fischer–Tropsch synthesis; MCM-48; Cobalt; Catalyst; Mesoporous silica
1. Introduction
metal catalysts, which were found to have well-defined periodic
mesopores, consequently provide very narrow pore size distri-
butions, and possess large pore volumes of 1–2 cm3/g and high
surface areas reaching 1000 m2/g. The pore diameters can be
controlled in the range of 2–30 nm by using various surfactants,
additives, and different synthetic conditions [16]. The narrow
pore size distributions are suitable for evaluation of pore size
and texture effects on cobalt dispersion and also it may make
it possible to design new catalysts with higher productivity for
long chain hydrocarbons. Thus, the use of periodic mesoporous
silicas as supports for preparing Co-based FTS catalysts has
been recently explored [17–22].
Catalytic behavior of silica supported cobalt catalysts in FT
synthesis was found to depend on the nature of cobalt species,
cobalt particle size and catalyst mesoporous structure. The prop-
erties of the cobalt particles were greatly affected by the pore
size of the mesoporous support. Previous reports showed that
reducibility of cobalt particles supported by silica [20,23,24],
titania [25] and alumina [26] depended mostly on their sizes.
Smaller cobalt particles are usually more difficult to reduce
than larger ones. For example, with alumina supported cobalt
catalysts, even though the true metal cluster size is smaller at
Fischer–Tropsch (FT) synthesis is a process for hydrocarbon
production from carbon monoxide and hydrogen. The synthetic
fuels produced via FT synthesis have high cetane numbers, low
contents of sulfur and aromatics. This makes them suitable for
diesel engines and friendly for environment [1]. Supported Co-
based catalysts have been widely used to achieve high yields of
paraffinic hydrocarbons in FT synthesis [2–6]. Cobalt is usually
supported on a high surface area support such as SiO2 [7–9],
Al2O3 [7], zeolites [10] and TiO2 [7] with microporous and
mesoporous structures in order to obtain a high metal disper-
sion. These conventional mesoporous supports are irregularly
spaced and their pore sizes are broadly distributed. Thus, it is
rather hard to study the effect of support and its porosity on FT
reaction rate and hydrocarbon selectivity [11]. Highly ordered
mesoporous silica, such as MCM-41 [12], FSM-16 [13], SBA-
15 [14] and HMS [15], have been recently used as a support for
∗
Corresponding author. Tel.: +86 27 67843016; fax: +86 27 67842752.
E-mail address: lij@scuec.edu.cn (J. Li).
1381-1169/$ – see front matter © 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2005.08.050

34
H. Li et al. / Journal of Molecular Catalysis A: Chemical 244 (2006) 33–40
lower loadings, the poor extent of reduction leads to a low
cobalt surface metal active site density under FT conditions.
Heavier loadings are often used to increase the cluster size and
thereby break the cobalt oxide–support interaction, resulting in
increased extent of reduction and leading to improved cobalt
surface active site densities. It is often considered [7,15] that the
presence of larger cobalt particles leads to higher selectivity to
heavier hydrocarbons.
Consideringthediscussionabove, largercobaltparticlessizes
formed in wider pore mesoporous supports, such as SBA-15, and
were more reducible leading to catalysts with higher catalytic
activity and lower methane selectivity than smaller particles in
narrower pore materials, such as MCM-41. Lower FT activ-
ity and higher methane selectivity observed on the narrow pore
cobalt catalysts are principally attributed to the lower reducibil-
ity of small cobalt particles [18,20].
2. Experimental
2.1. Synthesis of MCM-48 and catalyst
MCM-48 was synthesized by the conventional hydrothermal
pathway similar to the procedure described by Wang et al. [32].
n-Hexadecyltrimethylammonium bromide (C16H33(CH3)3–
NBr, template) was dissolved in deionized water, sodium
hydroxide and tetraethoxysilane (TEOS) were added. The
molar composition of TEOS:NaOH:C16H33(CH3)3–NBr:H2O
was 1:0.48:0.4:55. The solution was stirred for about 30 min,
charged into a polypropylene bottle and then heated at 383 K for
3 days. The sample was filtered, washed with water, and calcined
at 823 K for 6 h to remove template. The calcined MCM-48 was
used as a support for Co catalyst.
Co was introduced to the supports by incipient wetness
impregnation using aqueous solutions of cobalt nitrate. After
impregnation the catalysts were dried at 373 K and then cal-
cined in a flow of air at 773 K for 5 h. The three catalysts were
prepared with 5, 10 and 15 wt.% cobalt loading and denoted as
5Co/M, 10Co/M and 15Co/M, respectively.
It is usually suggested that the size of supported metal and
oxide particles and thus their catalytic behavior are principally
affected by overall metal content in supported catalysts. It is
generally assumed that an increase in metal loading would
almost automatically result in lowering of metal dispersion
[27]. Co/SBA-15 catalysts cobalt loadings of 10–40 wt.% were
prepared and a maximum CO conversion was obtained for
the catalyst loaded with ca. 30 wt.% Co presenting the high-
est density of surface Co0 sites. This finding highlights the
importance of further considering the impact of the cluster size
on the reducibility of cobalt clusters. That is, at low cobalt
loadings, the cobalt oxide–support interaction is strengthened,
making it difficult to achieve reduction of the cobalt clusters
and therefore, generate surface active sites. Product selectiv-
ities were also influenced by Co loading. The product distri-
bution shifted toward the formation of lighter hydrocarbons
(methane, C2–C4) for the less reducible low-loaded sample
[28].
The synthesis of the silica-based M41S was first reported in
1992 [29]. The specific properties of these materials are large
surface areas and a narrow pore-size distribution. MCM-48, as
a member of M41S, has a cubic structure indexed in the space
group Ia3d [30], recently modelled as a gyroid minimal surface
with its high specific surface area up to 1600 cm2 g−1 specific
pore volume up to 1.2 cm3 g−1 and high thermal stability.
This is because of its interwoven and branched pore structure,
which provides more favourable mass transfer kinetics in
catalytic and separation applications than MCM-41 with its
unidirectional pore system [31]. Various ordered mesoporous
silica for cobalt catalysts have been used, but an application of
MCM-48 as metal support to cobalt-based FT catalysts has not
previously been reported. In this work, MCM-48 was utilized
to investigate how MCM-48, its pore size and Co loading
affect surface reaction parameters during CO hydrogenation.
Co/MCM-48 catalysts with different Co content were prepared,
characterized by N2 physisorption, X-ray diffraction (XRD),
transmission electron microscopy (TEM), temperature pro-
grammed reduction (TPR), hydrogen temperature programmed
desorption (H2-TPD) with pulse oxygen titration, and tested
in a fixed-bed reactor for the Fischer–Tropsch synthesis
reaction.
2.2. Characterization techniques
2.2.1. BET measurements
Pore size distribution, BET surface area and pore volume
were measured by Micromeritics ASAP2405 automatic system
using nitrogen physisorption at 77 K. The specific surface area
was estimated by the BET method. The pore size distribution
and pore volume was determined by the BJH method [33].
2.2.2. XRD
X-ray diffraction patterns were obtained at room temperature
in a Phillips X’pert diffractometer using monochromatized Cu
K␣ radiation. The scan range was 1–80◦ with 0.002◦ steps. The
averageparticlesizesofCo3O4 wereestimatedfromtheScherrer
equation [34] using the most intense reflexion at 2θ = 36.9◦
kλ 180◦
d =
B cos θ
π
where d is the mean crystallite diameter, k (0.89) the Scherrer
constant, λ the X-ray wave length (1.54056 A) and B is the full
˚
width half maximum (FWHM) of Co3O4 diffraction peak.
The Co3O4 particle sizes in the calcined samples were
then converted to the corresponding cobalt metal diameters in
reduced catalysts by considering the relative molar volumes of
Co0 and Co3O4 using the equation: d(Co0) = 0.75 × d(Co3O4).
Then, the Co0 metal dispersions can be calculated from the
mean Co0 particle sizes assuming a spherical geometry of the
metal particles with uniform site density of 14.6 atoms/nm2 as
described in ref. [35] using D = 96/d(Co0), where D is the per-
centage dispersion and d(Co0) is the mean particle size of Co0
in nm.
2.2.3. Temperature programmed reduction
The reduction behavior and the interaction between active
phase and support of each catalyst were examined by using

H. Li et al. / Journal of Molecular Catalysis A: Chemical 244 (2006) 33–40
35
temperature programmed reduction technique. The TPR experi-
ments were carried out with a Zeton Altamira AMI-200 unit. The
catalyst (ca. 0.12 g) was placed in a quartz tubular reactor, fitted
with a thermocouple for continuous temperature measurement.
The reactor was heated with a furnace designed and built to stabi-
lizethetemperaturegradientandminimizethetemperatureerror.
Prior to the hydrogen temperature programmed reduction mea-
surement, the calcined catalysts were flushed with high purity
argon at 423 K for 1 h, and cooled down to 323 K. Then 10%
H2/Ar was switched on and the temperature was raised at a rate
of 10 K min−1 from 323 to 1023 K (hold 30 min). The gas flow
rate through the reactor was controlled by three Brooks mass
flow controllers and was 30 cm3 min−1. The H2 consumption
(TCD signal) was recorded automatically by a PC.
then, increased to 723 K in 2 h and held for 12 h. Subsequently,
the reactor was cooled down to 453 K. The syngas (H2:CO = 2)
was introduced to the reactor and the pressure was increased to
10 bar. The reactor temperature was increased to 503 K in 4 h
and the reaction was carried out at 503 K. The products were
collected in a hot trap and a cold trap in sequence. The effluent
product gas was passed through an Agilent 3000GC for online
analysis. The liquid product analysis was performed with an
Agilent 6890GC equipped with a FID detector. The solid wax
was analyzed with an Agilent 4890GC. The carbon monoxide
conversion (XCO%) was measured at the steady state and XCO%,
hydrocarbon selectivity have been averaged over the period of
constant operation. The ratio of olefin/paraffin was calculated
from the respective chromatogram peak areas.
2.2.4. Hydrogen temperature programmed desorption and
O2 titration
3. Results and discussion
Hydrogen temperature programmed desorption was also car-
ried out in a U-tube quartz reactor with the Zeton Altamira
AMI-200 unit. The sample weight was about 0.2 g. The catalyst
was reduced at 723 K for 12 h using a flow of high purity hydro-
gen and then cooled to 373 K under hydrogen flow. The sample
was held at 373 K for 1 h under flowing argon to remove weakly
bound physisorbed species prior to increasing the temperature
slowly to 723 K. Then, the catalyst was held under flowing argon
to desorb the remaining chemisorbed hydrogen and the TCD
began to record the signal till the signal returned to the baseline.
The TPD spectrum was integrated and the amount of desorbed
hydrogen was determined by comparing to the mean areas of
calibrated hydrogen pulses.
3.1. N2 physisorption
The N2 physisorption isotherms for the 10Co/M sample are
shown in Fig. 1. BET surface areas, pore volumes and average
pore diameters of the catalysts are given in Table 1. The isotherm
of the catalyst presents a sharp inflection at a relative pressure
of about 0.25, indicating a narrow distribution of pores in the
mesoporerangecharacteristicforthismaterial. N2 physisorption
isotherms of the supported Co samples were similar to that of
the original MCM-48, suggesting that the mesoporous structure
of MCM-48 was mostly retained upon cobalt impregnation.
O2 titration was also performed with the Zeton Altamira
AMI-200 unit. After reduction under the conditions (as
described above for H2-TPD), the catalyst was kept in flowing
Ar at 723 K and the sample was reoxidized by injecting pulses
of high purity oxygen in argon. The extent of reduction was
calculated by assuming metal Co was converted to Co3O4. The
results were shown in Table 2.
2.2.5. Transmission electron microscopy
Transmission electron microscopy characterization of the
samples was carried out by using a FEI Tecnai G2 instrument.
The samples were crushed in an agate mortar, dispersed in
ethanol and dropped on copper grids. The Co3O4 average par-
ticle diameter and particle-size distribution were obtained by
TEM. Between 100 and 300 particles coming from at least
four different blind-coded images were measured. Unfocused
or superimposed particles were discarded. From these data, the
statistical average was calculated. Thus, only a rough indication
on the size of the cobalt clusters can be given, due to the limited
number of TEM images.
Fig. 1. Nitrogen physisorption isotherms of 10Co/M.
Table 1
N2 physisorption results
Sample
BET surface
area (m2/g)
Pore volume
(cm3/g)
Average pore
diameter (nm)
2.2.6. FT activity and selectivity
Fischer–Tropsch synthesis was performed in a fixed bed reac-
tor (i.d. 2 cm) at 10 bar. The catalyst (ca. 6.0 g) was mixed
with ca. 36.0 g carborundum and reduced in situ at atmosphere
pressure. The reactor temperature was increased from ambient
MCM-48
5Co/M
10Co/M
15Co/M
1127
776
734
626
1.1
2.6
2.4
2.3
2.4
0.72
0.68
0.58
temperature to 373 K (hold 60 min)ina H2 flow of 6 S L g−1 h−1
,

36
H. Li et al. / Journal of Molecular Catalysis A: Chemical 244 (2006) 33–40
Table 2
Characterization results of Co catalyst
Catalyst
Co3O4 crystallite
diameter (nm)
Co0 dispersion (%)
Extent of overall
reduction (%)
XRD TEM
XRD TPD and O2
titration
O2 titration
5Co/M
10Co/M
15Co/M
5
11
14
3
4.5
6.5
25.6
11.6
9.1
59.2
21.4
14.2
1.0
20.4
21.3
structures of the calcined MCM-48 are still retained to some
extent even by adding 5–15 wt.% Co. But it was found that after
loading Co, the intensity of the (2 1 1) MCM-48 reflection peak
gradually decreased and the peak became broader. It suggested
that the order of the MCM-48 may have partially collapsed dur-
ing calcination and Co loading.
The diffraction peaks at 2θ of 31.3◦, 36.9◦, 45.1◦, 59.4◦ and
65.4◦ (Fig. 3) indicate that after calcinations at 500 ◦C cobalt was
primarily in the form of Co3O4 spinel on all the catalysts and Co
species are highly dispersed. The higher the cobalt content is, the
stronger the XRD intensity of Co3O4 is and the narrower the full
width half maximum of Co3O4 diffraction peak is. It suggests
that crystallite size increases and many larger particles formed
at higher Co loading. The average sizes of Co3O4 crystallites
detectedwerelargerthanthecorrespondingporediameters, indi-
cating that some of Co3O4 clusters are held outside the pores.
It is probable that the Co particles present outside the pores
can readily agglomerate to be crystallized due to weak interac-
tions with surface SiOH groups of the support compared with
those inside the pores [36]. Then when cobalt loading increased
the cobalt particles outside the pore became larger because of
agglomeration.
Fig. 2. XRD profiles for MCM-48 and catalysts.
The BET surface area and pore volumes of the catalysts
decreased with increasing Co loading, the decrease however
being more significant in the case of higher at larger Co load-
ing catalysts. This may be caused by a partial blockage of the
MCM-48 pores by cobalt oxide clusters and/or a partial collapse
of the mesoporous structure.
3.2. X-ray diffraction
The XRD patterns at lower and higher diffraction angles of
the calcined samples are displayed in Figs. 2 and 3, where those
forMCM-48supportisalsoillustratedforcomparison. Theaver-
age crystallite sizes of Co3O4 and cobalt metal (Co0) dispersion
were calculated and listed in Table 2.
The MCM-48 provided distinct two diffraction lines due to
SiO2 (2 1 1), (2 2 0) planes (Fig. 2), showing cubic structures,
as reported early [12,32]. When Co loading in the Co/MCM-48
was increased, the XRD profiles were almost unchanged and
exhibited the highly diffraction peaks at low 2θ angles reflect-
ing the ordered structure of MCM-48. It is evident that the cubic
3.3. Temperature programmed reduction
The TPR profiles for the Co/MCM-48 catalysts are shown in
Fig. 4. There are four H2 consumption peaks in every profile,
Fig. 3. XRD profiles for the catalysts.
Fig. 4. TPR profiles for Co/MCM-48 catalysts.

H. Li et al. / Journal of Molecular Catalysis A: Chemical 244 (2006) 33–40
37
with temperature ranges of about 220–360, 360–450, 500–600
and above 640 ◦C, respectively. The α peak could be assigned to
the reduction of Co3O4 to CoO, and the β one to the subsequent
reduction of CoO to Co0 [37,38]. The relative intensity of the
second reduction peak increased with Co loading, suggesting a
higher reduction degree of CoO to metallic Co with increasing
the mean diameter of Co3O4 particles. The reducibility of TPR
220–450 ◦CisdirectlyrelatedtotheamountofactiveofCo0 sites
available for catalyzing FTS after standard reduction. The γ and
δ peaks suggest the presence of surface Co species with different
degrees of interaction with the support [28]. The γ peak indicates
so a low strength of interaction between Co species and support
which could not form the cobalt silicate species. The δ peak with
the temperature range of 640–800 ◦C might be assigned to the
reduction of cobalt silicate species formed during the catalysts
preparing process.
TheextentofCoreductionandCo0 dispersionestimatedfrom
TPD and O2 titration is given in Table 2. The extent of reduction
of catalysts first increased rapidly and then slightly increased
with increasing Co loading. It attributed to the size of Co3O4
on the catalysts, which larger cobalt oxide particles being much
easily reduced.
Fig. 6. TEM micrographs of 10Co/MCM-48.
dark areas, marked with black arrows, give a smeared contrast
over the channels, which correspond to cobalt oxide particles on
the external surface. Figs. 5–7 show that cobalt oxide particles
in Co/MCM-48 are present in clusters. It can be more easily
found that cobalt oxide clusters concentrated more towards the
external surface of the catalyst granules especially for higher Co
loadings.
The cobalt crystallite size distributions obtained for the cat-
alysts are shown in Fig. 8. The average crystallite size obtained
from the above particle size distributions calculated are given in
Table 2. Many of cobalt oxide particles (Fig. 8) have a diameter
below 2.6 nm, which do not exceed the average pore diame-
3.4. Transmission electron microscopy
Micrographs of 5Co/M, 10Co/M and 15Co/M are shown in
Figs. 5–7, respectively. TEM investigation provides the direct
observation of the morphology and distribution of Co particles
in the MCM-48. Images (Figs. 4–6) show the highly ordered
arrangement of the channels. The MCM-48 typical structure
was maintained after impregnation and calcination, though the
support structure is partially collapsed or destroyed after intro-
ducing the cobalt.
Fig. 5 shows 5Co/M’s structures and areas with a high density
of cobalt crystallites and large areas with no cobalt present. The
Fig. 5. TEM micrographs of 5Co/M.
Fig. 7. TEM micrographs of 15Co/MCM-48.

38
H. Li et al. / Journal of Molecular Catalysis A: Chemical 244 (2006) 33–40
Table 3
Catalytic results for the FTS on Co/MCM-48 catalysts
a
b
Catalyst
XCO
%
SCO
%
Hydrocarbon selectivity (C mol%)
2
C1
C2
C3
C4
C5+
5Co/M
10Co/M
15Co/M
1.5
27.1
25.8
3.96
1.36
1.44
23.84
18.82
17.79
3.68
1.14
1.13
4.72
2.54
2.79
3.92
1.9
2.07
59.88
74.23
74.78
The catalysts were reduced in a flow of H2 at 723 K for 12 h before FTS. Reaction
conditions: H2/CO = 2, space velocity of syngas 2 S L g−1 h−1 (273 K, 0.1 MPa),
temperature 503 K, pressure 10 bar.
a
CO conversion.
CO2 selectivity.
b
was only 1.56% on the 5Co/M catalyst, which was almost inac-
tive. On the other hand, taking into account experimental error,
the 10Co/M and 15Co/M were quite active and no significant
difference in CO conversion and C5+ selectivity.
Fig. 8. Cobalt crystallites size distributions as determined by TEM.
Such a low activity on 5Co/M can be related to the low
reducibility of cobalt species in the catalyst, probably in the form
of cobalt silicates as discussed before. In principle, the activity
of reduced Co catalysts should be proportional to the concen-
tration of surface metal Co sites. When Co loading was low,
increasing Co loading can increase the surface Co0 sites den-
sity rapidly. At higher loading, the number of clusters increases.
Also, since initially increasing the loading assists in breaking the
cobalt oxide–support interaction, this allows the clusters to be
more easily reduced. The true metal dispersion decreased with
increasing Co loading. When Co loading reaches to 10 wt.%,
increasing Co loading does not increase the surface Co0 sites
density further. As evidenced by O2 titration, the reducibility
of 10Co/M and 15Co/M are very similar (i.e., 20.4 and 21.3%,
respectively). So, 10Co/M and 15Co/M showed the similar CO
conversion in FTS.
ter of the support. It suggests that small Co crystallites were
formed and stabilized in the mesoporous silica channels. The
crystallites which exceed 2.6 nm were situated on the external
silica grain surface near the mesopores entries resulting from the
uncontrolled metal particle growth.
A pronounced maximum for particles of 3 nm diameter, rep-
resenting about 40% of the total cobalt oxide particles, is found
on 5Co/M (Fig. 8a). Significant differences in crystallite-size
distribution are evident for three catalysts with higher cobalt
content having larger Co3O4 distribution. The catalysts with
higher cobalt loadings have bigger particles. The Co3O4 crystal-
lites size of 5Co/M, 10Co/M and 15Co/M below 5 nm is about
92, 83 and 79%, respectively. Small Co metal crystallites below
5 nm appear to reoxidize and deactivate rapidly in presence of
water and other reaction products at typical FTS conditions
[3]. Co/MCM-48 (Fig. 4) shows a broad reduction feature in
the range of 500–800 ◦C that could be attributed to the reduc-
tion of the smaller cobalt particles (probably those with less
than 5 nm in diameter) detected by TEM. The very small cobalt
oxide particles confined inside the mesoporous channels would
favor the CoO–support interaction leading to the formation of
hardly reducible cobalt species. Comparing the extent of reduc-
tion listed in Table 2, the three distribution values represent the
crystallites which were difficult to reduce.
As seen in Table 2, XRD crystallite diameters are about
0.7–1.2 times larger than the corresponding TEM values. Usu-
ally, larger values have been reported for average particle sizes
estimated from XRD as compared to other techniques and
these discrepancies have been explained by considering the
limitations and approximations of using XRD line broaden-
ing, which generally tend to overstate the actual particle sizes
[39,40].
3.5.2. Product selectivity
As observed in Table 3, the mesoporous 5Co/M catalyst gives
the highest methane selectivity and the lowest C5+ selectiv-
ity. Increasing Co loading leads to higher C5+ selectivity. For
10Co/M and 15Co/M catalysts, selectivity to lower hydrocar-
bons (C1–C4) or long hydrocarbons (C5+) was similar, irre-
spective of the content of cobalt.
The high methane selectivity was usually reported for cat-
alysts having high metal dispersion and low Co reducibility
[7,28]. This resulted in the presence of unreduced cobalt oxides
which can catalyze the WGS reaction; resulting in an increas-
ing in H2/CO ratio on the catalyst surface. It is known however,
that an increase in effective H2/CO ratio usually leads to both
an increase in methane selectivity and to higher chain termina-
tion probability and to lower coefficient of chain propagation
[23,41].
As seen in Table 3, both 10Co/M and 15Co/M present higher
C5+ selectivity than 5Co/M. It is also considered that the pres-
ence of larger cobalt particles leads to higher selectivity to
heavier hydrocarbons [7,15]. Also from Fig. 9, a decrease of the
ratio of olefin to paraffin is showed with increasing chain length,
mainly owed to the decrease of olefin content. The decrease of
3.5. Fischer–Tropsch synthesis
3.5.1. Activity of Co/MCM-48 catalysts
CO conversion and the selectivity (a carbon basis) to the dif-
ferent FTS products are presented in Table 3. The CO conversion

H. Li et al. / Journal of Molecular Catalysis A: Chemical 244 (2006) 33–40
39
Scientist Foundation of Hubei (2003ABB013), Excellent Young
Teachers Program of Ministry of Education of China, the State
Ethnic Affairs Commission, PR China, and Returnee Startup
Scientific Research Foundation of Ministry of Education of
China.
References
[1] C. Knottenbelt, Catal. Today 71 (2002) 437.
[2] H. Schulz, Appl. Catal. A 186 (1999) 3.
[3] E. Iglesia, Appl. Catal. A 161 (1997) 59.
[4] P.J. van Berge, R.C. Everson, Stud. Surf. Sci. Catal. 107 (1997) 207.
[5] G.P. van der Laan, A.A.C.M. Beenackers, Catal. Rev. Sci. Eng. 41 (1999)
255.
[6] R. Zennaro, M. Tagliabue, C.H. Bartholomew, Catal. Today 58 (2000)
309.
[7] R.C. Reuel, C.H. Bartholomew, J. Catal. 85 (1984) 63.
[8] S. Sun, N. Tsubaki, K. Fujimoto, Appl. Catal. A 202 (2000) 121.
[9] L.B. Backman, A. Rautiainen, A.O.I. Krause, M. Lindblad, Catal. Today
43 (1998) 11.
Fig. 9. Influence of Co loading on the olefin to paraffin ratio in product of
catalysts. (There had no products gathered for low CO conversion of 5Co/M.)
[10] S. Bessell, Appl. Catal. A 126 (1995) 235.
[11] S. Bessell, Appl. Catal. A 96 (1993) 253.
[12] C.T. Kresge, M.E. Leonowicz, W.J. Roth, J.C. Vartuli, J.S. Beck, Nature
359 (1992) 710.
[13] S. Inagaki, Y. Fukushima, K. Kuroda, J. Chem. Soc., Chem. Commun.
(1993) 680.
[14] E. Zhao, J. Feng, Q. Huo, N. Melosh, G.H. Fredrickson, B.F. Chmelka,
G.D. Stucky, Science 279 (1998) 548.
[15] H.H. Nijs, P.A. Jacobs, J. Catal. 65 (1980) 328.
[16] A. Corma, Chem. Rev. 97 (1997) 2373.
olefin content with chain length could be caused by the decrease
in the diffusivities of longer chain hydrocarbons. The increase in
their residence time in the catalyst pores [3], may be caused by
␣-olefin readsorption [42], or the higher solubility of the higher
␣-olefin in the liquid phase [5], leading to their increased con-
version to paraffin. So, in this study, the higher C5+ selectivity
observed over 10Co/M and 15Co/M can be attributed to the fol-
lowing two reasons: (i) the presence of larger cobalt particles
and (ii) the readsorption of the ␣-olefins.
[17] J. Panpranot, J.G. Goodwin Jr., A. Sayari, Catal. Today 77 (2002) 269.
[18] D. Yin, W. Li, W. Yang, H. Xiang, Y. Sun, B. Zhong, S. Peng, Micro-
porous Mesoporous Mater. 47 (2001) 15.
[19] Y. Wang, M. Noguchi, Y. Takahashi, Y. Ohtsuka, Catal. Today 68 (2001)
3.
[20] A.Y. Khodakov, A. Griboval-Constant, R. Bechara, V.L. Zholobenko, J.
Catal. 206 (2002) 230.
[21] A. Griboval-Constant, A.Y. Khodakov, R. Bechara, V.L. Zholobenko,
Stud. Surf. Sci. Catal. 144 (2002) 609.
[22] A.Y. Khodakov, R. Bechara, A. Griboval-Constant, Stud. Surf. Sci. Catal.
142 (2002) 1133.
The chain growth probability during Fischer–Tropsch syn-
thesis on supported cobalt catalysts generally increases with the
size of the reduced Co particles [7,23] up to a certain parti-
cle size above which it remains constant [42]. Khodakov et al.
[27] also thought that for a given catalytic support hydrocarbon
selectivities are almost constant in the wide range of cobalt sur-
face densities. In this study, the C5+ selectivities followed these
results.
[23] A.Y. Khodakov, A. Griboval-Constant, R. Bechara, F. Villain, J. Phys.
Chem. B 105 (2001) 9805.
[24] A.Y. Khodakov, J. Lynch, D. Bazin, B. Rebours, N. Zanier, B. Moisson,
P. Chaumette, J. Catal. 168 (1997) 16.
[25] J. Li, G. Jacobs, Y. Zhang, T.K. Das, B.H. Davis, Appl. Catal. A 223
(2002) 195.
4. Conclusions
The reducibility and FT catalytic behavior of cobalt species
supported by MCM-48 can be affected by Co loading. Lower FT
activity and higher methane selectivity observed on low cobalt
loading catalyst are principally attributed to the lower reducibil-
ity of small cobalt particles. With increasing Co loading the
cobalt crystallite size increases and leads to higher reducibility
and C5+ selectivies. But when Co loading exceed 10 wt.%, the
CO conversion and hydrocarbons selectivies seemed to affected
slightly by Co loading. The extent of overall reduction may thus
be one of the key factors for determining not only CO conversion
but chain growth probability.
[26] G. Jacobs, T.K. Das, Y. Zhang, J. Li, G. Racoillet, B.H. Davis, Appl.
Catal. A: Gen. 233 (2002) 263.
[27] A.Y. Khodakov, R. Bechara, A. Griboval-Constant, Appl. Catal. A: Gen.
254 (2003) 273.
[28] A. Mart´ınez, C. Lo´pez, F. Ma´rquez, I. D´ıaz, J. Catal. 220 (2003)
486.
[29] S. Beck, J.C. Vartuli, W.J. Roth, M.E. Leonowicz, C.T. Kresge, K.D.
Schmitt, C.T.W. Chu, D.H. Olson, E.W. Sheppard, S.B. McCullen, J.B.
Higgins, J.L. Schlenker, J. Am. Chem. Soc. 114 (1992) 10834.
[30] M.W. Anderson, Zeolites 19 (1997) 220.
[31] K. Schumacher, M. Gruun, K.K. Unger, Microporous Mesoporous Mater.
27 (1999) 201.
[32] S. Wang, D. Wu, Y. Sun, B. Zhong, Acta Phys.-Chim. Sin. 17 (2001)
659.
[33] E.P. Barrett, L.G. Joyner, P.P. Halenda, J. Am. Chem. Soc. 73 (1951)
373.
Acknowledgments
[34] B.D. Cullity, Elements of X-Ray Diffraction, Addision–Wesley, London,
1978.
[35] R.D. Jones, C.H. Bartholomew, Appl. Catal. 39 (1988) 77.
The work was supported by the National Natural Science
Foundation of China (20473114, 20590360), Talented Young

40
H. Li et al. / Journal of Molecular Catalysis A: Chemical 244 (2006) 33–40
[36] Y. Ohtsuka, Y. Takahashi, Catal. Today 89 (2004) 419.
[37] P. Arnoldy, J.A. Moulijn, J. Catal. 93 (1985) 38.
[38] B. Viswanathan, R. Gopalakrishnan, J. Catal. 99 (1986) 342.
[39] D.G. Castner, P.R. Watson, I.Y. Chan, J. Phys. Chem. 93 (1989)
3188.
[40] P. Ganesan, H.K. Kuo, A. Saavedra, R.J. De Angelis, J. Catal. 52 (1978)
319.
[41] A. Barbier, A. Tuel, I. Arcon, A. Kodre, G.A. Martin, J. Catal. 200
(2001) 106.
[42] E. Iglesia, S.L. Soled, R.A. Fiato, J. Catal. 137 (1992) 212.