Full Papers
porosity of the shell with a maximum in pore-width distribu-
tion at approximately 2.5 nm, which is in good agreement
the cobalt core for material C is approximately 17.3 nm (23 nm
for Co O ), which is similar to the results from material B and in
3
4
[
17]
with the literature.
The specific cobalt surface area of
agreement with the rough observation from the electron mi-
croscopic images. The comparable cobalt particle size offers
a unique comparability of the catalytic properties of both ma-
terials.
2
À1
2.8 m g was measured by H chemisorption, which proves
2
the accessibility of the active material by gaseous molecules
through the porous silica shell of material B. Taking the cobalt
content of approximately 7.5 wt% [by energy-dispersive X-ray
spectroscopy (EDX) for calcined material; atomic Si/Co-ratio of
A comparison of the FT activity of material B and C, with re-
spect to selectivity and reaction rate, is show in Table 1. The
16.5 by EDX and 14 by inductively coupled plasma optical
emission spectrometry (ICP-OES)] into account the size of the
cobalt core can be calculated from the active surface area and
amounts to about 18 nm, which in turn corresponds to
[
a]
Table 1. Comparison of material B and C during FT reaction.
[
c]
[7b]
Selectivity
CH CO
Reaction rate
a Co O particle size of about 24 nm [d(Co)=0.75 d(Co O ) ].
[b]
3
4
3
4
T
X
CO
4
2
m
cat
m
Co
S
Co
The difference to the results from the electron microscopic
images can be explained by the uncertainty of the cobalt-con-
tent measurements. Furthermore, TEM images of the as-syn-
thesized material suggest that the core is present as an ag-
glomerate of small Co O nanoparticles of approximately 7 nm,
À1
À1
À2
[
8C]
[mmolkg s] [mmolkg s] [mmolm s]
material B, 0.51 g catalyst
2
10 0.022 0.211 0.029
2.66
3.79
6.97
3.55
5.06
9.29
0.95
1.35
2.49
4.18
225 0.031 0.214 0.051
2
2
40 0.057 0.212 0.075
50 0.096 0.198 0.091 11.70
3
4
15.60
rather than one single particle (Figure S1). Consequently, the
cobalt mass of the porous agglomerate corresponds to a small-
er single particle, which is formed during reduction prior to
the chemisorption experiments.
material C, 0.55 g catalyst
00 0.023 0.188 0.005
225 0.084 0.163 0.004
2
2.55
9.42
2.55
9.42
0.65
2.42
7.59
2
2
40 0.263 0.157 0.011 29.59
50 0.374 0.180 0.022 42.00
29.59
42.00
10.77
During the transformation step of material B into C sulfuric
acid was used to reduce the pH of the synthesis mixture,
À1
[
a] Conditions: 2.1 MPa, H
CO conversion. [c] Reaction rate of CO consumption is related to: mcat
catalyst mass), mCo (cobalt mass), and SCo (cobalt surface area).
2
/CO=2, 0.5 LSTP
h
CO feed flow rate. [b] XCO:
which in turn reduces the rate of dissolution of SiO . This
2
(
allows a permanent separation of the individual Co O cores
3
4
by a SiO or zeolite shell during the whole synthesis, by either
2
partial dissolution of the SiO shell or direct formation of zeolit-
2
ic material around the core particles. However, the prevailing
separation mechanism is a matter of ongoing research.
methane selectivity of material B and C is comparable and
almost independent of temperature, with slightly higher values
for material B. The rather high values (at approximately 20%)
are comparable to literature data, in which 24% is reported for
A characteristic TEM image of material C is shown in
Figure 2. It is clear that the cobalt oxide nanoparticles are in-
corporated inside a zeolitic matrix (zeolite particle size 1–5 mm)
and are, therefore, completely embedded inside a zeolite shell.
The macroscopic observation of the supernatant after centrifu-
gation shows a clear solution, indicating the absence of residu-
al colloidal cobalt and complete immobilization of cobalt
nanoparticles at the external surface or inside the zeolite crys-
[17]
[7b]
material B and 19% for Co/zeolite at 2408C. The CO se-
2
lectivity appears to be unusual for Co catalysts and is higher
for material B than for C. However, values in the same order of
[7b]
magnitude are also reported in literature. The specific reac-
tion rate observed for material C is higher than that for materi-
al B, which holds for catalyst mass, cobalt mass, as well as for
cobalt surface area as a reference quantity. The order of magni-
tude agrees well with the literature, where a value of approxi-
mately 3.1 mmol /kg s can be derived for 10 wt% Co sup-
tal. The presence of ZSM-5 and Co O4 crystallites was con-
3
firmed by XRD measurements. Physisorption experiments with
Ar show a maximum in pore-width distribution of approxi-
mately 6 ꢂ (Figure 4), which agrees well with the maximum in-
CO
cat
[7b]
ported on zeolite at 2408C and 1.5 MPa. The analysis of the
[25]
cluded sphere diameter of MFI framework-type zeolites. Fur-
thermore, the modal pore size of the precursor material B ap-
pears to be negligible. The elemental composition of the cal-
cined material was measured by EDX, resulting in an atomic Si/
Co/Al-ratio of 93:6.3:1 and a cobalt content of 10.0 wt% (ICP-
OES results: 8.1 wt% cobalt, Si/Co/Al-ratio of 119:11.5:1). The
Si/Al-ratio corresponds well with the desired value in the syn-
thesis mixture of 100. The atomic Si/Co-ratio decreases slightly
from 16.5 to 14.8 (by EDX; from 14 to 10.3 by ICP-OES) during
transformation of material B into C, suggesting that Si partially
remains in solution. The cobalt content of material C is higher
than for material B, as confirmed by a specific cobalt surface
temperature dependency of the reaction rate reveals a higher
À1
apparent activation energy for material C (119 kJmol ) than
À1
for material B (77 kJmol ). A possible explanation could be
the different product distribution, which is shifted to lighter
hydrocarbons for material C (see below). Thus, the diffusion
through the liquid-filled pores for material C is improved, lead-
ing to a higher reaction rate and activation energy. This finding
is supported by the slightly higher methane selectivity for ma-
terial B, which indicates more pronounced diffusion limitations.
However, the different nature of the support material might
also cause significant differences in reaction rate and selectivi-
ty. In addition, repeated temperature cycles between 200 and
2608C during approximately 1100 h on stream revealed an ex-
ceptional catalytic stability of the FT component in material C
regarding conversion and selectivity (Figure S5).
2
À1
area of 3.9 m g for material C obtained from chemisorption
measurements. Hence, significant leaching of cobalt during
zeolite synthesis can be ruled out. The corresponding size of
ChemCatChem 0000, 00, 0 – 0
3
ꢁ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ