Auto-Reduction Behavior of Cobalt on Graphitic Carbon Nitride Coated Alumina Supports for Fischer–Tropsch Synthesis

Article abstract of DOI:10.1002/cctc.201700613

Graphitic carbon nitride (g-C₃N₄)-coated alumina (CN-Al) is investigated as a support material for a cobalt-based Fischer–Tropsch synthesis (FTS) catalyst. The g-C₃N₄ blocks the interaction between cobalt and alumina to hinder the formation of irreducible cobalt oxide species, and improves the dispersion of metallic cobalt species on the CN-Al support. Furthermore, the coated g-C₃N₄ is decomposed during the pre-annealing process even under inert conditions to evolve reductive carbon monoxide gas that directly reduces the cobalt oxide species to metallic cobalt. Hence, the catalyst can be activated without typical hydrogen pretreatment. Moreover, reduction promotion by expensive noble metals such as Pt is not needed either. Because of the improved dispersion and reducibility of the cobalt species on the CN-Al support, the catalytic FTS activity of Co/CN-Al is higher compared with the conventional Co/Al₂O₃ catalyst.

Full text of DOI:10.1002/cctc.201700613

Accepted Article
Title: Auto-reduction Behaviour of Cobalt on Graphitic Carbon Nitride
Coated Alumina Supports for Fischer-Tropsch Synthesis
Authors: Hunmin Park, Kwang Young Kim, Duck Hyun Youn, Yo Han
Choi, Won Young Kim, and Jae Sung Lee
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemCatChem 10.1002/cctc.201700613
Link to VoR: http://dx.doi.org/10.1002/cctc.201700613
A Journal of
www.chemcatchem.org

ChemCatChem
10.1002/cctc.201700613
Auto-Reduction Behaviour of Cobalt on Graphitic Carbon Nitride
Coated Alumina Supports for Fischer-Tropsch Synthesis
†
[b]
†[a]
[a]
[c]
[b]
Hunmin Park,
Kwang Young Kim,
Duck Hyun Youn, Yo Han Choi, Won Young Kim, and Jae
*
[a]
Sung Lee
Abstract: The graphitic carbon nitride (g-C
Al) is investigated as a support material for a cobalt-based Fischer-
Tropsch synthesis (FTS) catalyst. The g-C blocks the interaction
3
N
4
)-coated alumina (CN-
production of liquid hydrocarbons without adding anthropogenic
CO .
2
Fischer-Tropsch synthesis (FTS) is the key and common
reaction in the XTL processes, which produces hydrocarbon
[3a]
3
N
4
between cobalt and alumina to hinder the formation of irreducible
cobalt oxide species, and improves the dispersion of metallic cobalt
chains from H
2
and CO mixtures (synthesis gas or syngas). The
[5]
3 4
species on the CN-Al support. Furthermore, the coated g-C N is
reaction occurs on the catalysts typically composed of Ru,
[6]
[7]
[8]
decomposed during the pre-annealing process even under the inert
condition to evolve reductive carbon monoxide gas that directly
reduces the cobalt oxide species to metallic cobalt. Hence, the
catalyst can be activated without typical hydrogen pretreatment.
Moreover, the reduction promotion by expensive noble metals like Pt
is not needed either. Because of the improved dispersion and
reducibility of the cobalt species on CN-Al support, the catalytic FTS
Co, Fe, or Ni, . The cobalt-based catalysts are commonly
used as an industrial FTS catalyst because of its high selectivity
to long chain hydrocarbons,
a long catalyst life, and a
[9]
reasonable material cost. Metallic cobalt has been regarded as
the active phase of cobalt-based catalysts in FTS reaction, so
the reduction of cobalt without sintering is important to enhance
[10]
the activity of cobalt-based FTS catalysts.
strong interaction with common oxide supports like SiO
and TiO , the cobalt species forms a spinel structure or other
stable oxides, which are hardly reducible by the reduction
Because of the
activity of Co/CN-Al is higher compared to the conventional Co/Al
catalyst.
O
2 3
2
, Al
2 3
O ,
2
[11]
pretreatment even at high temperatures (Scheme 1). Typically,
promotion by noble metals such as Pt, Ru, and Re can lower the
reduction temperature of cobalt species, which greatly elevates
Introduction
[12]
the catalytic activity of cobalt-based catalysts. In the case of
00,000 bpd GTL process, 2.5 tons of platinum is required for
Alternative fuel production from non-petroleum resources is an
urgent issue today, because the petroleum has problems of
long-term supply and emission of environmentally harmful
1
[13]
catalyst per batch.
noble metal
However, high price and scarcity of the
[14]
raise the operation cost of the whole process,
[1]
exhausts to the air. Technologies have been developed to
substitute petroleum with electricity derived especially from
renewable energies as the transportation energy like Li-ion
which consequentially lower cost competitiveness of FTS
[15]
products. Xiong et al. reported the cobalt-based catalysts on
the nitrogen-doped carbon spheres, which facilitate the
reduction of cobalt resulting in improved FTS activity.
[2]
batteries and fuel cells. Another approach is XTL (X-to-liquid)
process to produce liquid fuels from non-petroleum sources
such as natural gas (GTL, gas-to-liquid), and coal (CTL, coal-to-
[3]
liquid), which have more reserves than petroleum. The liquid
fuels from XTL process have advantages that current petroleum
refinery facility could be used to produce derivatives of
petrochemical products. Moreover, diesel or gasoline produced
from syncrude could be environmentally less harmful and of
[4]
better quality compared to traditional oil from petroleum. The
resources of XTL process could also be replaced by biomass
(
BTL), and organic wastes (WTL), which enable the sustainable
†
[a]
K. Y. Kim, Dr. D. H. Youn, Prof. Dr. J. S. Lee
School of Energy and Chemical Engineering
Ulsan National Institute of Science and Technology (UNIST)
Ulsan, 689-798 (Korea)
Scheme 1. Effect of the CN-Al support on the reducibility of cobalt species.
E-mail: jlee1234@unist.ac.kr
†
[
b]
c]
H. Park, Dr. W. Y. Kim
Department of Chemical Engineering
Pohang University of Science and Technology (POSTECH)
Pohang, 790-784 (Korea)
3 4
The graphitic carbon nitride (g-C N ) is composed of carbon and
nitrogen bonds with a graphene-like 2-dimensional (2-D) plane
structure. It contains abundant nitrogen sites with lone pair
[
Y. H. Choi
Division of Advanced Nuclear Engineering
Pohang University of Science and Technology (POSTECH)
Pohang, 790-784 (Korea)
electrons, which would function as basic sites for various
[16]
reactions.
3 4
Also g-C N has many defects at the edge of its
plane. It is widely applied for various chemical and electro-
†
[
[17]
]
These authors are contributed equally to this work.
catalytic reactions. The synthesis of g-C
3 4
N is relatively simple,
but its specific surface area is small because of the stacking and
Supporting information for this article is given via a link at the end of
the document.
[18]
aggregation of the 2-D planes.
To increase the specific
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700613
3 4
surface area of g-C N , the colloidal silica particles can be used
as a nano-sized pore template, which is removed later by HF or
[19]
NaOH treatment.
Previously, we reported iron-based FTS catalysts over g-C
which g-C promoted the reduction of iron oxide and facilitated
the formation Hägg iron carbide (-Fe ,), the active catalytic
However, the structure of g-C
3
4
N , in
3
N
4
5
C
2
[20]
phase of the FTS reaction.
3 4
N
was completely destroyed during the pretreatment and FTS
reaction, and the iron particles were severely aggregated
leading to the catalyst deactivation. In this work, we prepare g-
C
3
N
4
coated alumina support (CN-Al) for cobalt-based FTS
catalyst, where g-C layers block the direct interaction
3
N
4
between cobalt and alumina and facilitate the reduction of cobalt
species (Scheme 1). Moreover, by using alumina as the stable
support material, the catalyst maintains the high specific surface
area and the small cobalt particle size during the reaction.
Figure 1. SEM images of bare Al
2
O
3
(a) and 17CN-Al (d). HRTEM images of
Co/Al O (b, c) and Co/17CN-Al (e, f). The marked regions in (b) and (e) are
2
3
magnified in (c) and (f), respectively.
Results and Discussion
Table 1. Properties of synthesized catalysts.
surface area
N content /
wt.%
C+N content
/wt.%
Cobalt
[b]
Catalyst preparation and characterization
2
-1
[a]
[a]
/m g
size/nm
The g-C
cyanamide precursor, which can be confirmed by XRD patterns
in Figure S1. The prepared g-C coated alumina composite is
denoted as xCN-Al (x = 6, 12, and 17), where x is weight % of g-
in CN-Al as determined by elemental analysis in Table 1.
Surface morphology and crystalline structure of cobalt-loaded
CN-Al (Co/CN-Al) catalysts and reference Co/Al was
investigated by SEM and HRTEM analysis in Figure 1. The
rough surface of CN-Al (Figure 1d) is attributed to the g-C N
4
3 4
N was directly synthesized on alumina by using
[c]
Co/Al
2
O
3
117.61
128.81
116.09
123.26
-
-
8.85
3
N
4
[
c]
Co/6CN-Al
Co/12CN-Al
Co/17CN-Al
4.26
7.97
6.34
12.02
17.04
5.03
C
3
N
4
[d]
7.00
[
e]
11.07
13.40
2
O
3
[
a]
[b]
[c]
0
For bare CN-Al supports. Measured by XRD. Crystallite size of Co calculated
3
0
[21] [d]
from the obtained value of Co
Crystallite size of Co .
3
O
4
by Co = 3/4 * Co
3
O .
4
Crystallite size of CoO.
debris on alumina while alumina itself has a smooth surface
morphology (Figure 1a). The specific surface area of CN-Al
composites decreases considerably with increasing amounts of
[e]
0
0
g-C
Figure S2 and Table S1). Therefore, the surface of Al
covered by g-C coating in CN-Al, which would modify the
3
N
4
on Al
2
O
3
, and pores are also blocked by g-C
3
N
4
coating
CoO with Co peaks are observed as impurity phases. The
Co/12CN-Al catalyst contains partially reduced cobalt (CoO) as
the major phase. Finally, fully reduced cobalt becomes the major
phase in Co/17CN-Al catalyst in agreement with the lattice
distance exhibited in HRTEM images. The XRD patterns
indicates that cobalt species in catalysts is in a more reduced
(
O is well
2 3
3
N
4
physicochemical properties of alumina. When 20 wt. % of cobalt
is loaded on the alumina support by impregnation, cobalt
particles are severely aggregated giving 70-100 nm sizes
(
1
Figure 1b). The dispersion of cobalt is much improved on
7CN-Al with particle size decreasing to 15-20 nm (Figure 1e).
The higher dispersion of cobalt species with addition of g-C
on alumina could be attributable to a strong interaction between
cobalt species and g-C on alumina surface, which retards the
agglomeration of cobalt particles. After annealing under N flow
is
(Figure
c). In contrast, the observed lattice distance in Co/CN-Al
form as the amount of g-C
well represented in Figure 2b by decreasing peak intensity ratio
3 4
N in the support increases, which is
0
3
N
4
3 4 3 4
of Co O /CoO and increasing Co /CoO with increasing g-C N
content in CN-Al composites. The annealing condition in our
o
3
N
4
experiments (450 C, 3 h, N
reduced cobalt species. Yet it is remarkable that the extent of
reduction of cobalt species can be controlled by the g-C
content in the CN-Al support. The calculated crystallite size of
2
flow) is not supposed to form a
2
o
at 450 C, the lattice distance of cobalt particle in Co/Al
0
1
treated under the same conditions is 0.206 nm, which
corresponds to (111) facet of metallic cobalt (Figure 1f). This
unexpected result suggests that the cobalt species on CN-Al
support can be reduced without hydrogen treatment.
2
O
3
3 4
N
3 4
.247 nm, which corresponds to the (113) facet of Co O
[22]
cobalt by Scherrer’s equation
Co/Al to 13.4 nm for Co/17CN-Al (Table 1). The trend is not
simple because of different oxidation states of cobalt species
and specific surface areas (Table S1). Al has a larger surface
area than CN-Al supports (Table S1), but Co/Al contains
increases from 8.85 nm for
2 3
O
2
O
3
2
O
3
The crystal structure of cobalt species in Co/Al
investigated by XRD in Figure 2. The species is Co
generally observed in calcined cobalt-based catalysts. The
Co/6CN-Al catalyst also shows Co as the major phase, but
2
O
3
was
larger Co particles than those in Co/6CN-Al and Co/12CN-Al
catalysts (Table 1). The result indicates that the strong
interaction of Co and g-C N is more important than the surface
3 4
area of the support materials in the formation of small Co
particles. More importantly, the secondary particle aggregation is
3
O
4
as
3
O
4
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700613
o
Figure 3. (a) H
2
-TPR spectra and (b) fractions of peak area below 400 C
(
black column) and relative total areas (■). Relative total area was
calculated from total area of H -TPR in (a) and normalized with the area of
Co/Al at 100 %.
2
2 3
O
Figure 2. (a) XRD patterns for alumina-supported cobalt catalysts with
o
different amounts of carbon nitride. (b) Comparisons of XRD peak intensity
signals below 300 C, which are attributed to the decomposition
of nitrate or various weakly adsorbed molecules on the CN-Al
0
ratio of Co
3
O
4
/CoO (■) and Co /CoO (●). The peak intensity for Co
3
O
4
/CoO
0
o
o
o
and Co /CoO ratios was obtained for peaks at 31 , 42 , and 44 , for Co
3
O ,
4
0
o
CoO, and Co , respectively.
supports. The TCD signals in the range of 300-500 C exhibit
3 4
increasing peak intensity with increasing g-C N contents and
effectively retarded by coated g-C
analysis in Figure 1.
3
N
4
as revealed in HRTEM
thus they should originate from the decomposition of g-C N
3 4
during He-TPR. This is convinced by elemental analyses of the
samples upon interruption of He-TPR at different temperatures
(Figure 5). The overall nitrogen and carbon content (C+N, wt.%)
decreases as the He-TPR analysis progresses.
In Figure 3a, the reduction behaviour of cobalt species in the
catalysts is investigated by H -TPR analysis. The Co/Al
catalyst gives two dominant peaks; at ~300 C attributed to the
2
2 3
O
o
0
o
o
reduction of Co
strongly bound Co
of the latter peak is dominant for Co/Al
rapidly decreases with increasing amounts of g-C
Al supports. Typical reduction treatment in FTS process is below
3
O
4
to Co and at 600 C to the reduction of
The two He-TPR peaks appearing at 350 and 460 C suggests
[23]
x
O species on Al
2
O
3
surfaces. The intensity
and Co/6CN-Al, but
in the CN-
that two reactions occur at the temperatures. To investigate the
o
2
O
3
processes occurring at 350 and 460 C, He-TPR of Co/17CN-Al
o
3
N
4
was interrupted at 280, 400, and 550 C. The XRD pattern in
o
Figure 4b of 280 C sample exhibits Co
3
O
4
as the only cobalt
o
o
4
00 C, therefore a substantial fraction of the cobalt species is
species. The Co/17CN-Al annealed by He-TPR up to 400 C
o
not reduced to metallic cobalt in Co-based FTS catalysts. Since
metallic cobalt is the active phase for cobalt-based FTS
catalysts, those unreduced cobalt oxide species such as Co O
3 4
and CoO cannot contribute to the FTS reaction.
high content of cobalt species reduced below 400 C is an
exhibits pure CoO. Finally in the 550 C sample, the metallic
cobalt species is observed. The XRD patterns of samples
annealed at different temperatures indicate that the He-TPR
[10]
o
Hence, the
peaks at 350 and 460 C originate from the successive reduction
o
3 4
of cobalt species, Co O  CoO  metallic Co. Thus, the
important feature of the active Co-based FTS catalysts.
analyses clearly demonstrate that the decomposition of g-C N
3 4
o
2
The fractions of peak areas below 400 C of H -TPR spectra are
accompany the reduction of cobalt species on the CN-Al support.
As shown in Figure 5, the content of nitrogen and carbon (N+C)
continuously decreases with temperature increase during He-
presented in Figure 3b, which tend to increase with increasing g-
C
H
3
4
N content in Co/CN-Al catalysts. In contrast, the total area of
o
2
-TPR peaks over whole temperature range (100–900 C)
content. This indicates that some of
3 4
TPR, which is ascribed to the decomposition of g-C N . The
decreases with g-C
3
N
4
nitrogen to carbon ratio (N/C), however, remains constant until
o
cobalt species in Co/CN-Al catalysts has been already reduced
during annealing procedure, which is in accordance with the
400 C, then suddenly decreases after the second peak of He-
o
TPR (550 C). This implies that nitrogen is less stable than
XRD analysis. Thus, H
support reduces the amount of irreducible cobalt species and
facilitates the formation of metallic cobalt. Raman spectrum of
2
-TPR indicates that g-C
3
N
4
in the CN-Al
3 4
carbon under heat treatment leading to nitrogen deficient g-C N
o
at 450-500 C (Scheme 2). Then, the neighbouring carbon
atoms adjacent to nitrogen deficient site become unstable and
easily reacts with oxygen atom in CoO resulting in completely
reduced cobalt. This is supported by gas chromatography
analysis, which indicates that the gaseous product evolving at
(2)
2 3 g
Co/Al O in Figure S3 exhibits Co-O stretching peaks of E , F2g ,
(
3)
-1
F
2g , and A1g at 472, 512, 610, and 682 cm , respectively, for
[24]
cobalt oxide species. On the other hand, Co/CN-Al catalysts
-1
o
[25]
show slight red-shift by ~10 cm of these Raman peaks. The red
shift in Co-O stretching indicates weakened Co-O bonding in
450 C is CO (Figure S5). Yang et al.
reported the similar
effect on the mesoporous carbon support, which was referred as
‘auto-reduction behaviour’. The auto-reduction of cobalt oxide is
caused by CO evolved from the decomposition of carbon
supports. Thus mass spectroscopic CO evolution pattern
Co
cobalt becomes easier. Therefore, it is concluded that g-C
CN-Al support promotes the reduction of cobalt species as
revealed by HRTEM, XRD, H -TPR, and Raman analyses.
x
O species on CN-Al supports, thus its reduction to metallic
3
N
4
on
[26]
2
reported by Cheng et al.
is similar to our He-TPR pattern of
Co/CN-Al catalysts. The auto-reduction of cobalt on CN-Al
o
Auto-reduction of cobalt on the graphitic carbon nitride
To investigate unusual reduction behaviour of cobalt under inert
atmosphere on CN-Al supports, unannealed Co/CN-Al catalysts
were analysed by the temperature-programmed reduction under
helium flow (He-TPR) in Figure 4. There are complicated TCD
supports (at 450 C in Figure 4a) is much lower than that on
o
nitrogen-doped carbon (500 C). The enhanced reducibility is
attributed to the reactive and unstable nature of nitrogen atoms
N , which is easily decomposed to gaseous molecules. In
3 4
auto-reduction of cobalt species on the carbon supports,
in g-C
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700613
nitrogen doping is reported to be important to enhance the auto-
[
15,
reducibility of cobalt during the calcination in inert atmosphere.
2
5a]
3 4
The chemical structure of g-C N is analogous to abundantly
nitrogen-doped carbon materials, thus dominant nitrogen doping
effect can be expected on the auto-reducibility.
Scheme 2. The reduction mechanism of cobalt species under the
inert gas flow for Co/CN-Al catalysts.
Degree of reduction of cobalt on graphitic carbon nitride
The degree of reduction (DOR) is quantified by O
2
pulse titration
of the catalysts reduced at 400 C by 10 vol. % H /Ar flow. It
should be noted that g-C is fully combustible with O below
00 C (Figure S4), which is the experimental condition for O
o
2
3
N
4
2
o
4
2
o
pulse titration. The shoulder peaks at 200-350 C in temperature
programmed oxidation (TPO) in Figure S4 increase as the
proportion of g-C
g-C consumes oxygen below 400 C. Therefore, the amount
of O consumption at the first titration must contain a portion for
g-C combustion. Therefore, we denoted the first titration of
Co/17CN-Al as ‘0th (zeroth)’ to exclude the portion of g-C
combustion, and the ‘1st and 2nd’ are counted sequentially. The
3 4
N in CN-Al support increases, indicating that
o
3
N
4
2
3
N
4
3
N
4
amount of consumed oxygen in the ‘1st’ titration (25.4 μmol O
is quite lower than that of the ‘0th’ titration (29.5 μmol O
2
)
)
2
because of the g-C
such difference disappears after the ‘2nd’ titration (25.2 μmol O
Based on O
2.93 % for Co/17CN-Al, which is 1.54 times higher than that of
Co/Al (14.88 %).
3 4
N combustion as mentioned above, but
Figure 4. (a) He-TPR analysis for unannealed samples (b) XRD
patterns of Co/17CN-Al catalysts during the interupted He-TPR.
2
).
2
pulse titration results, DOR is calculated as
2
2
O
3
1.8
1.5
1.2
0.9
0.6
0.3
0.0
20
The improved DOR of cobalt species would increase the active
sites in the catalysts for FTS reaction. Typical oxide support
1
1
8
4
0
6
materials such as SiO
2
,
Al
O
2 3
,
and TiO
2
have a strong
interaction with cobalt species in the catalysts, which retards the
reduction of cobalt during the hydrogen reduction step and the
catalytic reaction. Therefore, even after the reduction treatment,
there is irreducible cobalt species in the catalyst which is
inactive for FTS reaction. On the surface of CN-Al support,
2
3 4 2 3
however, g-C N blocks the direct contact of Al O and cobalt
species to hinder the formation of irreducible cobalt species and,
in addition, directly reduces the cobalt species along with its
decomposition. Hence the reducibility of cobalt is much
improved by using CN-Al support.
200
300
400
500
600
o
Temperature / C
Figure 5. He-TPR (left scale) of Co/17CN-Al. N/C ratio (■) and total
content of N and C (N+C/wt. %, ●) of Co/17CN-Al during He-TPR
o
interrupted at 280, 400, and 550 C.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700613
2
Table 2. Degree of reduction (DOR) of catalysts by O pulse titration.
3
2
1
0
b
Catalysts
Consumed O
2
titration / μmol O
2
DOR /%
Co/CN-Al
a
0th
1st
2nd
Co/CN-Al_Ar
Co/Al
2
O
3
-
14.9
25.4
o
16.5
25.2
14.88
22.93
-1
Co/17CN-Al
29.5
Co/Al O
2
3
-1
o
Pretreatment: ramping rate of 5 C min , 400 C, 3 h, 50 ml min 10 vol. %
o
-1
H
2
/Ar flow. Titration condition: 400 C, 50 ml min He flow, 5 vol. % O
2
/He
Co/Al O _Ar
pulse injection. The pretreatment and titration are repeated until the 2nd
2
3
a
titration is completed. The O
2
consumption of the 0th titration contains
b
0
10
20
30
40
50
contribution of the g-C combustion by O (see the text). DOR is
3
N
4
2
Time on Stream / h
calculated based on the average value of the 1st and 2nd O
2
titration for
each catalyst.
-
1
-1
Figure 6. Cobalt time yield (CTY, molCO
s gco ) for the FTS reaction
o
2
with time on stream. Reaction conditions: 220 C, 20 bar, H /CO=2.
Catalytic Fischer-Tropsch reaction
Catalytic FTS reaction was performed in a 1/2 inch stainless
o
a
steel fixed bed reactor at 220 C, 20 bar, and H
3 4
Co/17CN-Al and Co/Al to demonstrate the effect of g-C N
2
/CO = 2 over
Table 3. Catalytic performance in the FTS reaction.
2
O
3
Catalysts
CTY
/
-1
CO
sel./%
2
CO
2
-free HC sel./%
C -C
O/P
coating on the alumina surface. The Co/17CN-Al catalyst
demonstrates a stable activity of CO hydrogenation for 50 h of
continuous operation as in Figure 6. The cobalt time yield (CTY)
μmols
-1
Co
g
CH4
C5+
2
4
-
1
of Co/17CN-Al at 50 h of time on stream (TOS) is 21.84 μmol s
Co/Al
Co/Al
2
O
3
_H
2
10.75
3.92
2.358
24.10
10.90
26.15
8.46
80.64
34.06
1.539
0.886
-1
-1
g
Co , which is twice higher than that of Co/Al
2
O
3
(10.75 μmol s
without Al is
-1
O
2 3
_Ar
39.79
13.12
gCo ). Notably, the catalytic activity of Co/g-C
3
N
4
2 3
O
lower than Co/Al
2 3
O because of the severe aggregation of cobalt
Co/17CN-
Al_H
2
1.84
2.900
7.744
9.08
9.92
77.80
75.79
2.106
1.812
as g-C is destroyed during the pretreatment (Figure S7). The
3
N
4
2
enhanced activity of Co/17CN-Al is mainly due to the improved
reducibility of cobalt species as discussed above. Since the
metallic cobalt is the active phase for the cobalt-based FTS
catalyst, the catalysts should be treated under reductive
condition to generate the metallic cobalt prior to the FTS
reaction. However, cobalt species in Co/17CN-Al is already
reduced during the annealing process, and thus it does not need
to be reduced before the FTS reaction. Thus, the Co/17CN-Al
catalyst pretreated under Ar flow exhibited almost the same CTY
Co/17CN-
Al_Ar
21.52
14.29
a
The results were obtained at steady state at least 50 h of time on stream.
b
The O/P ratio was calculated for C
2
-C
4
hydrocarbon products. Reaction
/CO = 2, CO conversion = 14±2 %.
o
conditions: 220 C, 20 bar, H
2
surface, which was verified by XRD, BET, and SEM analysis.
The g-C coating on the alumina surface retards the
3
N
4
2 2 3
as that of H treated Co/CN-Al. In contrast, when the Co/Al O is
aggregation of cobalt particles and improves the cobalt
dispersion. Direct contact of cobalt and alumina is blocked by g-
treated with Ar, the activity drops drastically due to the
unreduced cobalt species in the catalyst. Therefore, the
enhanced reducibility and carbothermal reduction induced by g-
3 4
C N coating on CN-Al support, which reduces formation of
C
3
N
4
result in improved catalytic activity on the FTS reaction for
irreducible cobalt oxide species. Moreover, g-C N4 is
3
decomposed under inert condition during the initial annealing
step to produce carbon monoxide that directly reduces the
cobalt species to metallic cobalt even without hydrogen
treatment. This auto-reduction of cobalt oxide species is
facilitated by unstabilized carbon species adjacent to defective
Co/17CN-Al catalysts and the catalyst does not need reduction
treatment prior to the FTS reaction which would lower the
operating cost for FTS process. The XRD pattern of used
2 3
Co/Al O show the metallic cobalt species generated during the
reaction (Figure S6). However the catalytic activity is still lower
than Co/17CN-Al, which suggests that the improved dispersion
of cobalt species on CN-Al support is also the reason for the
enhanced activity.
3 4
nitrogen sites in g-C N . As much more reduced cobalt sites are
formed on CN-Al support, FTS activity of Co/CN-Al catalysts is
improved and remains stable for 50 h of continuous operation.
Because of the readily formed metallic cobalt on CN-Al support
under inert condition, the high FTS activity is obtained without
pre-reduction by hydrogen. The reduction promotion by
expensive noble metals like Pt is not needed either. By using the
CN-Al support, pretreatment process for cobalt activation
Conclusions
The effects of g-C
have been investigated for cobalt-based FTS catalysts. The g-
was successfully synthesized in situ on the alumina
3 4
N coating on the typical alumina support
3 4
C N
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700613
=
1
63.8/31.9/4.3 vol. %) were introduced and the catalysts were aged for
becomes much simpler and cheaper, and the activity is also
much enhanced.
o
o
2 h. Finally, the temperature was slowly increased to 220 C (0.5
C
-1
min ), and CO conversion was controlled by adjusting the gas flow rate.
The catalytic performance was measured at a fixed CO conversion of
14±2 % in order to exclude the effect of conversion on selectivity. The
Experimental Section
CO conversion was controlled by adjusting the gas flow rate. The liquid
o
and wax products were collected by hot trap (150 C, 20 bar) and cold
o
trap (1 C, 1 bar), respectively. The gaseous products were analysed by
Preparation of carbon nitride coated alumina (CN-Al). Carbon nitride
on-line connected gas chromatography (GC, 7890A, Agilent
technologies) equipped with 60/80 Carboxen 1000 column.
coated aluminium oxide support was prepared by direct synthesis of
carbon nitride from cyanamide (CA, Aldrich) on γ-Al
dissolved in deionized water and mixed with γ-Al
stirring, the mixture was heated to evaporate water. Then, the powder
O
2 3
(Aldrich). CA was
2
O . With vigorous
3
o
o
-1
-1
was heated to 550 C (4.58 C min ) for 4 h under N
Three kinds of the carbon nitride coated alumina were prepared with
different CA/Al weight ratios, denoted as xCN-Al (x= 6, 12, and 17 for
CA/Al of 0.8/1.6, 0.6/1.7, and 0.4/1.8, respectively. The value x was
the weight percent of g-C in CN-Al.).
2
flow (100 ml min ).
Acknowledgements
2 3
O
This research was supported by the Climate Change Response
project (2015M1A2A2074663, 2015M1A2A2056824), the Basic
Science Grant (NRF-2015R1A2A1A10054346), Korea Center
for Artificial Photosynthesis (KCAP, No. 2009-0093880) funded
by MSIP, Project No. 10050509 and KIAT N0001754 funded by
MOTIE of Republic of Korea.
2 3
O
3
N
4
Impregnation of support materials with cobalt. 20 wt. % of cobalt was
loaded by a typical incipient wetness impregnation method using cobalt
nitrate hexahydrate (Aldrich) as a cobalt precursor. The catalysts were
o
-1
2
heated at 450 C for 3 h under N flow (100 ml min ).
General characterization. HRTEM images were obtained using the
instrument in UNIST Central Research Facility (JEM2100F, JEOL). SEM
images were obtained using a Quanta 200FEG, FEI instrument. X-ray
diffraction patterns (XRD) were measured on PANalytical PW 3040/60
X’pert. The chemical composition of materials was measured by element
analyzer (EA, Truspec Micro, Leco), and inductively couples plasma (ICP,
Keywords: Fischer-Tropsch synthesis • auto-reduction • cobalt •
carbon nitride • reducibility
[
1]
K. S. Deffeyes, Hubbert's Peak: The Impending World Oil
Shortage, 1st ed., Prinston University Press, New Jersey, 2001.
M. S. Dresselhaus, I. L. Thomas, Nature 2001, 414, 332-337.
a) A. d. K. Peter M. Maitlis, Greener Fischer-Tropsch Processes
for Fuels and Feedstocks, Wiley-VCH Verlag GmbH & Co. KGaA,
Weinheim, 2013; b) S. Shafiee, E. Topal, Energy Policy 2009, 37,
181-189.
7
00-ES, Varian). The oxidation state of surface elements was measured
by X-ray photoelectron spectroscopy (XPS, K-alpha, ThermoFisher). The
specific BET surface area was measured by N physisorption isotherm at
196 C (Nanoporosity-XQ, Mirae SI). Raman spectra were measured at
[2]
3]
[
2
o
-
room temperature with 532 nm radiation laser, Nd-doped yttrium
aluminium garnet, 1 mW potency, CCD type detector, 800 nm spot
diameter with 9 distinguished spots for each sample (alpha300R, WITec).
[4]
5]
[6]
a) B. I. Kamara, J. Coetzee, Energy Fuels 2009, 23, 2242-2247; b)
O. Armas, K. Yehliu, A. L. Boehman, Fuel 2010, 89, 438-456.
C. Xiao, Z. Cai, T. Wang, Y. Kou, N. Yan, Angew. Chem. Int. Ed.
[
2008, 47, 746-749.
Temperature-programmed reductions and oxidation. Evolution of
E. Iglesia, Appl. Catal. A: Gen. 1997, 161, 59-78.
H. M. Torres Galvis, J. H. Bitter, C. B. Khare, M. Ruitenbeek, A. I.
Dugulan, K. P. de Jong, Science 2012, 335, 835.
L. Bruce, J. F. Mathews, Appl. Catal. 1982, 4, 353-369.
A. Y. Khodakov, W. Chu, P. Fongarland, Chem. Rev. 2007, 107,
1692-1744.
[7]
products during the H
He-TPR, temperature-programmed oxidation (TPO), and
titration were monitored by
AutoChem 2920, Micrometrics). The samples were pretreated at 200
2
-temperature-programmed reduction (H
2
-TPR),
pulse
O
2
[8]
a
thermal conductivity detector (TCD,
[9]
o
C
-1
[10]
A. Y. Khodakov, Catal. Today 2009, 144, 251-257.
G. Jacobs, T. K. Das, Y. Zhang, J. Li, G. Racoillet, B. H. Davis,
Appl. Catal. A: Gen. 2002, 233, 263-281.
for 2 h under Ar flow (50 ml min ) and then cooled down to ambient
temperature. The TCD signal was recorded with temperature ramp to
[11]
o
o
-1
9
00 C at 10 C/min under 10 vol. % H
2
/Ar (50 ml min ) flow. He-TPR
[12]
a) S. Vada, A. Hoff, E. Ådnane, S. D. Schanke, A. Holmen, Top.
Catal. 1995, 2, 155-162; b) W. Chu, P. A. Chernavskii, L.
Gengembre, G. A. Pankina, P. Fongarland, A. Y. Khodakov, J.
Catal. 2007, 252, 215-230; c) A. Kogelbauer, J. J. G. Goodwin, R.
Oukaci, J. Catal. 1996, 160, 125-133.
o
was performed by using uncalcined samples with a ramping rate of 5
C
-
1
-1
min and He flow of 50 ml min . TPO was measured under 5 vol. %
-1
2
O /He flow (50 ml min ). Degree of reduction of cobalt in catalysts was
measured by O
ml min ). It was assumed that all reduced cobalt species formed during
the hydrogen treatment (400 C, 3 h, 50 ml min ) were sufficiently
oxidized by 150 pulses of O
o
2
pulse titration at 400 C under 5 vol. % O
2
/He flow (50
[13]
A. Brumby, M. Verhelst, D. Cheret, Catal. Today 2005, 106, 166-
-1
169.
o
-1
[14]
P. C. K. Vesborg, T. F. Jaramillo, RSC Adv. 2012, 2, 7933-7947.
H. Xiong, M. Moyo, M. K. Rayner, L. L. Jewell, D. G. Billing, N. J.
Coville, ChemCatChem 2010, 2, 514-518.
X. Jin, V. V. Balasubramanian, S. T. Selvan, D. P. Sawant, M. A.
Chari, G. Q. Lu, A. Vinu, Angew.Chem. Int. Ed. 2009, 48, 7884-
[
15]
16]
2
.
[
Catalytic performance test of FTS reaction
7887.
[
[
[
17]
18]
19]
Y. Zheng, J. Liu, J. Liang, M. Jaroniec, S. Z. Qiao, Energy Env.Sci.
2012, 5, 6717-6731.
P. Niu, L. Zhang, G. Liu, H.-M. Cheng, Adv. Funct.Mater. 2012, 22,
Catalytic performance was tested in 1/2 inch stainless steel fixed bed
o
reactor under simulated industrial condition (20 bar, 220 C and H
2
/CO =
4763-4770.
2). The catalyst was mixed with 3.0 g of SiC (Aldrich) as diluent to
a) A. Vinu, K. Ariga, T. Mori, T. Nakanishi, S. Hishita, D. Golberg,
Y. Bando, Adv. Mater. 2005, 17, 1648-1652; b) F. Goettmann, A.
Fischer, M. Antonietti, A. Thomas, Angew. Chem. Int.Ed. 2006, 45,
prevent the formation of hot spot in the catalytic bed. Catalysts were pre-
-1
treated under hydrogen flow (50 ml min ) at ambient pressure and 400
o
o
-1
o
4467-4471.
C (5 C min ) for 3 h, then the reactor was cooled down to 200 C (2.22
o
-1
[20]
21]
H. Park, D. H. Youn, J. Y. Kim, W. Y. Kim, Y. H. Choi, Y. H. Lee, S.
H. Choi, J. S. Lee, ChemCatChem 2015, 7, 3488-3494.
D. Schanke, J. Catal.1995, 156, 85-95.
C min ) and argon was filled until the pressure reaches 20 bar. After
2 2
waiting 1 h for stabilization of the temperature, reactant gases (H /CO/N
[
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700613
[
22]
A. L. Patterson, Phys. Rev. 1939, 56, 978-982.
S. M. Olusola O. James, J. Pet. Technol. Altern. Fuels 2016, 7, 1-
2.
B. Jongsomjit, C. Sakdamnuson, J. G. Goodwin, P. Praserthdam,
Catal. Lett. 2004, 94, 209-215.
[
[
[
23]
1
24]
25]
a) Y. Yang, L. Jia, B. Hou, D. Li, J. Wang, Y. Sun, ChemCatChem
2014, 6, 319-327; b) Y. Yang, L. Jia, B. Hou, D. Li, J. Wang, Y.
Sun, Catal. Sci. Tech. 2014, 4, 717-728.
[26]
K. Cheng, V. Subramanian, A. Carvalho, V. V. Ordomsky, Y. Wang,
A. Y. Khodakov, J. Catal. 2016, 337, 260-271.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700613
Entry for the Table of Contents (Please choose one layout)
Layout 1:
FULL PAPER
Text for Table of Contents
Hunmin Park, Kwang Young Kim, Duck
Hyun Youn, Yo Han Choi, Won Young
Kim, and Jae Sung Lee*
g-C
3
N
4
coated on Al
2
O
3
blocks the
and
interaction between Co and Al
2
O
3
is decomposed to evolve reductive
CO gas to generate the metallic cobalt
even under the inert conditions. The
improved reducibility of the cobalt
species on g-C N /Al O raises the
3 4 2 3
the catalytic activity of Co in the
Fischer-Tropsch synthesis.
Page No. – Page No.
Title
Auto-reduction Behaviour of Cobalt
on Graphitic Carbon Nitride Coated
Alumina Supports for Fischer-
Tropsch Synthesis
This article is protected by copyright. All rights reserved.