Production of high-calorie synthetic natural gas using copper-impregnated iron catalysts

Article abstract of DOI:10.1016/j.molcata.2016.10.013

Fe-Cu catalysts were applied in the production of high-calorie synthetic natural gas (HC-SNG), wherein a proper level of C₂–C₄ hydrocarbon selectivity must be secured. The Fe-Cu catalysts were activated by reduction under diluted CO gas before the reaction, and their catalytically active Fe phases changed according to the reduction temperature: Fe₃O₄ formed when reduced at 300 °C, carbon-deficient FeC_x at 400 °C and Fe₃C at 500 °C. Iron carbide catalysts achieved stronger CO adsorption and higher BET surface area than Fe₃O₄ catalysts, which resulted in higher CO conversion. The carbon-deficient FeC_x was metal-like in its electron structure due to the low number of bonded carbons, and it was attributed to the highest CO and H₂ conversion of FC15-400R by providing H₂ activation ability. The C_n (n ≥ 2) selectivity or carbon chain growth of the hydrocarbons increased as the carburization degree of the active Fe phase increased, which was associated with an increase in the CO chemisorption strength. The impregnated Cu exerted little influence on the product selectivity, but it promoted hydrogen adsorption, thereby improving the paraffin-to-olefin ratio of the produced hydrocarbons.

Full text of DOI:10.1016/j.molcata.2016.10.013

Journal of Molecular Catalysis A: Chemical 425 (2016) 190–198
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Research Paper
Production of high-calorie synthetic natural gas using
copper-impregnated iron catalysts
Yong Hee Leea, Dae-Won Leeb,∗∗, Kwan-Young Leea,c,∗
a Department of Chemical and Biological Engineering, Korea University, Anam-dong, Seongbuk-gu, Seoul 02841, Republic of Korea
b Department of Chemical Engineering, Kangwon National University, 1 Kangwondaehak-gil, Chuncheon-si, Gangwon-do 24341, Republic of Korea
c KU-KIST School of Converging Science and Technology, Korea University, Anam-dong, Seongbuk-gu, Seoul 02841, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 26 July 2016
Received in revised form 6 October 2016
Accepted 9 October 2016
Available online 12 October 2016
Fe-Cu catalysts were applied in the production of high-calorie synthetic natural gas (HC-SNG), wherein
a proper level of C2–C4 hydrocarbon selectivity must be secured. The Fe-Cu catalysts were activated
by reduction under diluted CO gas before the reaction, and their catalytically active Fe phases changed
according to the reduction temperature: Fe3O4 formed when reduced at 300 ◦C, carbon-deficient FeCx at
400 ◦C and Fe3C at 500 ◦C. Iron carbide catalysts achieved stronger CO adsorption and higher BET surface
area than Fe3O4 catalysts, which resulted in higher CO conversion. The carbon-deficient FeCx was metal-
like in its electron structure due to the low number of bonded carbons, and it was attributed to the highest
CO and H2 conversion of FC15-400R by providing H2 activation ability. The Cn (n ≥ 2) selectivity or carbon
chain growth of the hydrocarbons increased as the carburization degree of the active Fe phase increased,
which was associated with an increase in the CO chemisorption strength. The impregnated Cu exerted
little influence on the product selectivity, but it promoted hydrogen adsorption, thereby improving the
paraffin-to-olefin ratio of the produced hydrocarbons.
Keywords:
High-calorie synthetic natural gas
Fe-Cu catalyst
Heating value
© 2016 Elsevier B.V. All rights reserved.
1. Introduction
combustion of other fossil fuels (88.0–97.8 g CO2/kJ for coal and
67.3 g CO2/kJ for gasoline) [3]. As a result, natural gas is proposed
Natural gas is still regarded as the most promising alternative
fossil fuel resource for petroleum. It is estimated that the global pro-
duction of natural gas will increase from 632 billion m3 in 2013 to
1667 billion m3 in 2040 [1]. The reserves-to-production ratio (R/P),
which indicates the remaining time (or amount) to depletion for a
fossil fuel resource, is estimated to be 54.1 years for conventional
natural gas [2]. However, as unconventional resources such as shale
gas, tight gas, coalbed methane, and methane hydrate are discov-
ered, the R/P value of natural gas is extended, making it one of the
most important issues in energy industry. In addition, the use of
natural gas is recommended because it has environmental advan-
tages over the use of other fossil fuel resources. Because methane
has the highest H/C atomic ratio (= 4) among hydrocarbons, the
combustion of natural gas releases the least amount of carbon diox-
ide (CO2) per production of unit energy (50.1 g CO2/kJ) than the
as the only realistic solution for alleviating the carbon dioxide emis-
sions of fossil-fuel-based power stations [4]. Other environmental
benefits of natural gas combustion are the relatively low emissions
of carbon monoxide, particulate matter, sulfur oxides and nitrogen
oxides. Now it is widely recognized that “a Golden Age of the Gas”
will begin in near future [5].
Natural gas is transported by ships or pipeline. To be transported
by ships, natural gas should be liquefied under high-pressure com-
pression, which is called liquefied natural gas (LNG). The natural
gas which is transported in a gas state through pipeline is called
pipeline natural gas (PNG). In general, LNG has a much higher
heating value (approximately 41–42 MJ/Nm3 [6]) than that of PNG
(33.8–38.8 MJ/Nm3) [7–9] because LNG contains more C2+ hydro-
carbons than PNG does. South Korea and Japan import LNG by ships
because they are surrounded by sea, separated from the main con-
tinent. Thus, the standard heating value of natural gas distributed
in these countries is established at high values over 42.7 MJ/Nm3.
However, the average heating value of LNG in the worldwide mar-
ket is on the continuous decrease because LNG suppliers tactically
lower the heating value of LNG by separating C2+ hydrocarbons.
Their purpose is first to respond to the increasing number of PNG
customers by standardizing the heating values downward toward
∗
Corresponding author at: Department of Chemical and Biological Engineering,
Korea University, Anam-dong, Seongbuk-gu, Seoul 02841, Republic of Korea.
∗∗
Co-corresponding author.
E-mail addresses: stayheavy@kangwon.ac.kr (D.-W. Lee), kylee@korea.ac.kr
(K.-Y. Lee).
http://dx.doi.org/10.1016/j.molcata.2016.10.013
1381-1169/© 2016 Elsevier B.V. All rights reserved.

Y.H. Lee et al. / Journal of Molecular Catalysis A: Chemical 425 (2016) 190–198
191
the level of PNG and second, to make more profit by selling C2+
hydrocarbon gas [10]. Due to the continuous decrease in the heat-
ing value of LNG, the rate by which municipal gas distribution is
measured in South Korea was recently changed from a volume-
based rate to the heat-based rate. However, this practice results in
an increase in the total volume consumption to meet the demanded
heating values, which causes other problems, such as an increase
in the cost for transport and storage.
tioned above, Fe catalysts usually produce high olefin-to-paraffin
ratio products compared to Co [39]. This is mostly due to the hydro-
genation activity of Fe, which is the lowest among the first period
of group VIII metals [40]. The low hydrogenation activity of Fe is
not a disadvantage in Fischer-Tropsch applications because an oth-
erwise high hydrogenation activity rather leads to an increase in
the methane selectivity, which is unfavorable from the standpoint
of FT synthesis. On the contrary, the low hydrogenation activity
of the catalyst could be a disadvantage in the production of high-
calorie SNG. First, it results in an increase in the olefin selectivity. It
is desirable for the catalysts of the production of high-calorie SNG to
maintain low olefin selectivity because an olefin has lower heating
value than its analogous paraffin. Second, if a catalyst is not properly
active for hydrogenation, the carbon chain growth of hydrocarbons
can extend beyond an acceptable level. In SNG production, the for-
mation of hydrocarbons longer than C5 must be prohibited because
they are condensable and cause many problems during the trans-
portation of HC-SNG. Thus, unlike for Fischer-Tropsch catalysts, a
proper level of hydrogenation activity is a type of prerequisite for
the catalysts for the production of high-calorie SNG. In Fe-based
catalysts, the promotion by Cu can be a way to augment the poor
hydrogenation activity of Fe.
In this work, Fe-Cu catalysts were applied to the production of
high-calorie SNG. The Fe-Cu catalysts were prepared by impreg-
nating the precipitated ␣-Fe2O3 with Cu in a wide range of Fe/Cu
compositions. We anticipated that CO would be activated to grow
into C2–C4 hydrocarbons over the Fe sites, while Cu would pro-
vide the hydrogenation sites required to suppress the carbon chain
growth into undesirably high hydrocarbons (≥ C5) and to pro-
mote the formation of paraffin. The catalysts were activated via
CO reduction at various temperatures, by which different types of
active Fe phases were obtained, and their activities were tested
in the production of high-calorie SNG. The activities were evalu-
ated in terms of CO conversion, C2–C4 selectivity, chain growth
probability (␣), heating value (MJ/Nm3) based on the produced
hydrocarbons, etc. The promoting effect of Cu was examined
especially in terms of the paraffin-to-olefin ratio in hydrocarbon
products. Characterization techniques based on X-ray spectroscopy
and temperature-programmed reduction/desorption were applied
to explain the activity of the tested catalysts.
Synthesizing high-calorie natural gas has been suggested as a
solution for the problem of low-calorie natural gas. To state the
obvious, the production of high-calorie gas should be based on
inexpensive carbon resources, such as coal or biomass. Most exist-
ing technologies for synthetic natural gas (SNG) synthesize pure
methane (i.e., methanation) from coal-based synthesis gas (syn-
gas). On the basis of the commercial SNG (methanation) processes
[11], the production of high-calorie synthetic natural gas (HC-SNG)
may be divided into a three-step process: (1) a coal gasifier, where
coal is converted into syngas, (2) a water-gas shift (WGS) reactor
to increase the H2/CO proportion to 3, which is the stoichiometric
ratio of the methanation reaction, and (3) a series of ‘high-calorie
SNG production’ reactors to convert the syngas into a methane and
C2+ hydrocarbon gas mixture. The heating value of SNG can be
enhanced by increasing the fraction of C2–C4 hydrocarbons because
the heating value of individual hydrocarbons simply increases as
the carbon number increases (the heating values of ethane, propane
and n-butane are 64.3, 92.2 and 121.3 MJ/Nm3, respectively [12]).
Thus, the activity of the catalysts for the production of high-calorie
SNG should be based on methanation, but at the same time, the
catalyst should possess a short-chain (C2–C4) Fisher-Tropsch (FT)
activity.
The four group VIII metals, Fe, Co, Ni and Ru, are well known
to be the most active for hydrogenation of carbon monoxide to
hydrocarbons [13]. Because Ni catalysts are highly selective for
methane [13,14], they are not suitable for the production of high-
calorie SNG. Inui et al. studied the Co-based catalyst to convert
coke oven gas into high-calorie SNGs [15,16]. However, the Co-Mn-
Ru/Al2O3 catalysts are limited in securing high C2–C4 selectivity.
Ru has been reported to provide a strong H2 spillover effect on
the active Co phase, which resulted in increases in the CO conver-
sion and CH4 selectivity but also caused a decrease in the carbon
chain growth [17,18]. The SNG produced using Co-Mn-Ru/Al2O3
contained 12.8 vol.% of C2–C4 hydrocarbons (24.6% in selectivity) to
result in a heating value of 43.8 MJ/Nm3 [18], which narrowly satis-
fied the required standard heating value (>42.7 MJ/Nm3). However,
this heating value is not considered high enough to mix with other
low-calorie gases to increase their heating values.
2. Materials and methods
2.1. Catalyst preparation
Hence, we focused on Fe-based catalysts, which are known to be
especially active for the production of C2–C4 hydrocarbons (olefins)
[19–29] in FT synthesis. It was reported that a commercial Fe-
based, high-temperature Fischer-Tropsch (HTFT) process produces
49 wt.% C2–C4 hydrocarbons with 36 wt.% gasoline and diesel [30].
Therefore, the short-chain FT activity can be regarded as an advan-
tage of Fe when it is used as an active component of the catalyst for
the production of high-calorie SNG.
The FT activity of Fe catalysts is dependent on the type of active
species. Magnetite (Fe3O4) and various types of iron carbides (FeCx)
are known to be catalytically active in FT synthesis. Most recent
studies (published in the years 2005–2015) claimed that FeCx has
higher FT activities than that of Fe3O4 [31–34]. However, the active
phases of Fe catalysts are frequently observed to be a mixture of
Fe3O4 and FeCx (especially ␹-Fe5C2). As a promoter component
for Fe, Cu is a promising candidate because Cu promotes the dis-
sociation of hydrogen (reactant) to improve the reducibility of
the Fe species [35,36]. At the same time, it could act as a struc-
tural promoter to stabilize the magnetite and iron carbide phases
against thermal degradation or agglomeration [37,38]. As men-
Fe-Cu catalysts were prepared by impregnating ␣-Fe2O3 with
Cu. First, ␣-Fe2O3 was synthesized by the precipitation method as
follows: 100 mL of deionized water was poured into a 500 mL three-
neck flask with three necks. The water temperature was adjusted
to 80 ◦C, and the pH was adjusted to 7 using a 1 M aqueous solu-
tion of (NH4)2CO3 (Aldrich). Then, 60 mL of an aqueous Fe(NO3)
3·9H2O (Sigma-Aldrich) solution (Fe metal concentration = 1 M)
was dropped into the flask at a rate of 2 mL/min. Iron hydrox-
ides were precipitated by adding a precipitating agent, (NH4)2CO3
(aq., 1 M). The pH was maintained at 7 0.03 while the precipi-
tation continued. The precipitates were aged at 80 ◦C for 2 h and
then washed sufficiently using 6 L of deionized water. The washed
precipitates were dried at 110 ◦C overnight and then oxidized to
␣-Fe2O3 at 550 ◦C under static air for 1 h. The ␣-Fe2O3 was impreg-
nated with Cu by the incipient wetness method using aqueous
solutions of Cu(NO3)2·2.5H2O (Sigma-Aldrich). The impregnated
samples were dried at 110 ◦C overnight and calcined at 550 ◦C for
5 h. For convenience, Fe–Cu catalysts are abbreviated to FC cat-
alysts, and are denoted as FCx-yR, where x and y represent the
Fe/Cu ratio (x = 40, 15, 6) and reduction temperature (y = 300, 400,

192
Y.H. Lee et al. / Journal of Molecular Catalysis A: Chemical 425 (2016) 190–198
Table 1
Fe/Cu atomic ratios and the Fe-Cu catalyst notation.
Catalyst
Measured atomic ratioa(Fe/Cu)
Reduction temperatureb (◦C)
Notation
–
Fe2O3
not reduced
300
400
500
not reduced
300
400
500
not reduced
300
400
unreduced Fe2O3
Fe2O3-300R
Fe2O3-400R
Fe2O3-500R
unreduced FC40
FC40-300R
FC40-400R
FC40-500R
unreduced FC15
FC15-300R
FC15-400R
FC40
FC15
FC6
39.5
15.0
6.0
500
not reduced
300
FC15-500R
unreduced FC6
FC6-300R
400
FC6-400R
500
FC6-500R
a
The concentration of each component was measured by ICP-AES.
Reduced for 3 h by an 8% CO/He gas mixture at each temperature.
b
500 ◦C), respectively. The prepared catalyst notations are listed in
Table 1.
were used in the heating value calculation because hydrocarbons
longer than C5 can exist in their gaseous form only if their concen-
tration is very thin, i.e., below 0.5 vol.% (The typical concentration
of C5+ hydrocarbons in natural gas is below 0.34 vol.% [42,43].). The
calculation was carried out using heat of combustion (ꢀHC ) data
from NIST chemistry WebBook [12].
2.2. Activity test
The activity of the catalysts was examined in a continuous fixed
bed reactor. Each catalyst was sieved in the range of 53 to 75 ␮m,
and 0.36 g of the catalyst was packed in the middle of a tubular
stainless-steel reactor (I.D. = 9.5 mm). A catalyst was reduced at
different temperatures (300, 400, 500 ◦C) at the ambient pressure
before the reaction test with flowing 8% CO/He gas. The feed gas for
the reaction tests was composed of CO:H2:N2 = 24:72:4 (H2/CO = 3)
and flowed into the reactor at a flow rate of 36 mL/min. The reac-
tion pressure was maintained at 1 MPa using a back-flow regulator
(TESCOM 26–1700 series, Emerson Process Management, Shang-
hai, China) during the entire reaction. The compositions of the feed
and product gases were analyzed by gas chromatography (Younglin
YL6500GC) equipped with a thermal conductivity detector (TCD)
and a flame ionization detector (FID). Carboxen 1000 and HP plot
Q columns were used as separation columns, which were con-
nected to the TCD and FID, respectively. Nitrogen was used as the
internal standard for determining gas compositions and the outlet
flow rate. The sample gas was injected to GC without separation of
water vapor but the composition of water vapor was not quanti-
fied, because it was not necessary. The CO conversion, selectivity for
products, and carbon mole balance were calculated by Eqs. (1)–(4).
ꢀ
ꢅ
ꢆ
4
ꢄ
ꢂ
ꢃ
ꢁHC,i
Vm,i
Heating value MJ/Nm3
=
xi ×
(6)
i=1
xi = volume fraction Ci hydrocarbons i : 1 − 4
(
)
ꢂ
ꢃ
ꢁHC,i = heat of combustion of Ci hydrocarbon
MJ/mol
ꢂ
ꢃ
Vm,i = molar volume of Ci hydrocarbon Nm3/mol
2.3. Characterization
Temperature-programmed reduction (TPR) analysis was carried
out using a BELCAT-M-77 instrument (BEL Japan Inc.). Samples
were purged at 100 ◦C for 1 h under a 30 mL/min flow of He gas prior
ꢁ
molar flow rate of unreacted CO in the product gas
CO conversion carbon mole % = 1 −
× 100(1)
(
)
molar flow rate of CO in the feed gas
n × molar flow rate of Cn hydrocarbonin the product gas
Selectivity for hydrocarbons with carbon number of n carbon mole %
=
× 100(2)
(
)
molar flow rate of total carbon − unreacted CO in the product gas
(
)
molar flow rate of CO2 in the product gas
molar flow rate of total carbon − unreacted CO in the product gas
Selectivity for carbon dioxide carbon mole % =
× 100(3)
(
)
(
)
Total moles of carbon in the product gas per unit time
Total moles of carbon in the feed gas per unit time
Carbon mole balance =
(4)
The chain growth probability (␣) was determined from an
Anderson-Schulz-Flory (ASF) plot using Eq. (5), where Wn is the
mass fraction of hydrocarbons containing n carbon atoms [41].
to analysis. TPR data were recorded with the increase in tempera-
ture from 60 to 810 ◦C at 5 ◦C/min under a 30 mL/min flow of an 8%
CO/He gas mixture (CO-TPR) at the ambient pressure. Temperature-
programmed desorption (TPD) analysis was performed using the
same instrument. Samples were pre-reduced at 300, 400 and 500 ◦C
for 3 h under a 30 mL/min flow of an 8% CO/He gas mixture. For
CO-TPD, CO was adsorbed on a sample at 35 ◦C for 30 min using
an 8% CO/He gas mixture. H2 was adsorbed via same procedure
log(Wn/n) = nlog˛ + log[(1 − ˛)2/˛]
(5)
The heating value of the SNG product gas was calculated without
consideration of non-hydrocarbon species (H2, CO, CO2) in esti-
mating the volume fractions of the hydrocarbon species (Eq. (6))
because SNG should be separated from those species in its recovery
step. Among the hydrocarbon species, only the C1–C4 hydrocarbons

Y.H. Lee et al. / Journal of Molecular Catalysis A: Chemical 425 (2016) 190–198
193
Fig. 1. CO-TPR profiles of calcined Fe-Cu catalysts (unreduced).
Fig. 2. X-ray diffraction results of the FC15 catalysts (a) after calcination, and CO
reduction at (b) 300 ◦C (300R), (c) 400 ◦C (400R) and (d) 500 ◦C (500R) for 3 h: ᭿
␣-Fe2O3; ᮀ Fe3O4; ꢀ carbon-deficient FeCx; ᭹ Fe3C; ꢀ CuO; ꢁ metallic Cu.
using a 10% H2/He gas mixture for H2-TPD. Physisorbed CO or H2
molecules were removed at 120 ◦C for 1 h with pure He gas flowing
at 30 mL/min. TPD data were recorded with the increase in tem-
perature from 120 to 810 ◦C at 5 ◦C/min under a 30 mL/min flow of
pure He gas.
X-ray diffraction (XRD) patterns were obtained using a D/MAX-
2500/PC diffractometer (Rigaku) operated at 40 kV and 100 mA
with Cu-K␣1 irradiation (ꢂ = 0.15406 nm). All the samples were
mixed an internal standard (Si powder) at 10 wt.% of total weight.
The diffraction data were collected by scanning from 20 to 50◦ (2␪)
at a rate of 1.5◦/min. The domain size of the crystallites was calcu-
lated by applying the Scherrer analysis to the largest XRD peak of
the crystallite (Eq. (7)) [44]. For iron carbide peaks, because their
largest peaks were closely overlapping between 2␪ = 43.0 and 45.1◦,
the Scherrer analysis was applied after peak deconvolution.
metallic Cu near 130 ◦C, which was attributed to peak 1, whose
intensity increased with the increasing Cu content in the catalysts.
Peak 2 (∼260 ◦C) was attributed to the reduction of ␣-Fe2O3 to
Fe3O4 (magnetite), which was confirmed by the existence of Fe3O4
in the XRD results of FC15-300R (Fig. 2(b)). The carbon-deficient
FeCx species (x < 1/3) were detected in the XRD results when the
catalyst was reduced by CO at 400 ◦C (FC15-400R, Fig. 2(c)), while
the amount of fully carburized Fe3C noticeably increased when the
reduction temperature was raised to 500 ◦C (FC15-500R, Fig. 2(d)).
Thus, it was deduced that Fe3O4 was consecutively carburized to
Fe3C via the carbon-deficient iron carbides (FeCx) in the 300–500 ◦C
section of the CO-TPR run. Therefore, peak 3 (Fig. 1), which spreads
broadly over 350–550 ◦C, might correspond to the carburization
of Fe3O4 to Fe3C (via FeCx). The negative signal of peak 3 is due to
the differences of the CO2/CO stoichiometry among the constituent
carburization reactions, which was explained in detail in our pre-
vious publication [18]. The 2␪ positions of carbon-deficient FeCx in
Fig. 2(c) (FC15-400R) were close to those of Fe1.86C0.14 (JCPDS #44-
1289) and Fe1.945C0.055 (JCPDS #44-1290), but it was not strictly
correct to conclude that the carbon-deficient FeCx species were
exactly composed of those two species. Thus, they were labeled
collectively as “carbon-deficient FeCx (x < 1/3)” in Fig. 2. Peak 4
was attributed to the reduction of Fe3C to metallic Fe, which was
supposed to take place with the precipitation of carbon [45].
In summary, the active phase of the FC catalysts changed
according to the reduction temperature (300, 400 and 500 ◦C). The
reduction sequence of Fe oxides in the CO-TPR was explained in
further detail in our previous publication [18].
Kꢂ
t =
(7)
Bcosꢃ
t = the mean size of the crystalline domains (nm)
K = dimensionless shape factor (= 0.9)
ꢂ = wavelength of X-ray (Cu K␣1, 0.15406 nm)
B = full width at half maximum (FWHM) in radians
ꢃ = Bragg angle
The mass fractions of Fe and Cu in the Fe/Cu catalysts was
analyzed using an inductively coupled plasma atomic emission
spectrophotometer (ICP-AES) supported by the Seoul center of
the Korea Basic Science Institute (KBSI). X-ray photoelectron spec-
troscopy (XPS) analysis was performed on an Axis Nova instrument
(Kratos) using the monochromatic Al-K␣ line as X-ray source. The
specific surface areas (BET areas) of the catalysts were determined
via BET analysis using a BELSORP-max instrument (BEL Japan). Sam-
ples were degassed overnight at 4 mmHg and 100 ◦C prior to the
analysis. N2 adsorption was carried out at −196 ◦C.
3.1.2. Surface properties of the reduced FC catalysts
The XPS analysis results of the reduced FC15 catalysts are pre-
sented in Fig. 3. The valance states of Fe and Cu were examined using
the Fe 2p and Cu 2p XPS data. FC15-300R (Fe3O4) showed a single
3. Results and discussion
3.1. The analysis of the FC catalyst active phases
Fe 2p
peak positioned at 711.0 eV. When the reduction temper-
3
⁄2
ature was raised (FC15-400R (FeCx) and FC15-500R (Fe3C)), the Fe
3.1.1. The change in the active Fe phases according to the
reduction temperature
2p peak was divided into two, with a lower binding energy peak
3
⁄2
(707.3 eV). The lower binding energy peak was especially intense
Fig. 1 displays the CO-TPR curves of the FC catalysts. To inter-
pret the CO-TPR data, the oxide components of the FC catalysts
should first be identified. The identification was made using XRD
data (Fig. 2). Fig. 2 shows the XRD data for the FC15 catalyst, but the
XRD profiles of the FC6 and FC40 catalysts (not shown here) were
similar to those of the FC15 catalyst. The unreduced FC catalysts
(FC15 (calcined), Fig. 2(a)) were composed of ␣-Fe2O3 (hematite)
and CuO. On the CO-TPR run (Fig. 1), CuO was first reduced to
for the FC15-400R catalyst, which can be interpreted as the carbon-
deficient FeCx are more metal-like in their electron structure (2p
3
⁄2
B.E of Fe0 = 706.6 eV) than Fe3O4 and Fe3C. Thus, an active site (Fe
species, where FT carbon chain-growth might occur) of FC15-400R
could be expected to be more electron-rich than those of FC15-
300R and 500R. On the other hand, the Cu concentration (Cu 2p
peak area) was the largest when the catalyst was reduced at the

194
Y.H. Lee et al. / Journal of Molecular Catalysis A: Chemical 425 (2016) 190–198
Table 2
CO conversion, product selectivities, and carbon balance of FC catalysts after a
10 h reaction test; the reaction conditions: 300 ◦C, 10 bar, w/f = 0.01 g/mL/min,
CO:H2:N2 = 24:72:4.
Catalysts
Reduction
CO conv. (%)
H2 conv. (%)
Carbon balance
(%)a
temperature (◦C)
FC40
300
400
500
300
400
500
300
400
85.7
97.4
90.5
94.2
97.5
96.6
84.4
97.7
95.4
36.4
45.9
40.5
43.1
48.8
43.6
34.1
48.2
99.5
100.4
96.0
99.7
99.6
95.5
99.3
100.4
95.9
FC15
FC6
500
42.4
ꢂ
ꢃ
moles of carbons ininlet flow
moles of carbons inoutlet flow
a
Carbon balance %
(
=
.
)
3.1.3. Catalytic activity of reduced FC catalysts in the production
of high-calorie SNG
Fig. 3. Fe 2p and Cu 2p and C 1 s XPS spectra for FC15-300R, −400R, and −500R.
CO and H2 conversion of FC catalysts after 10 h reaction test were
listed in Table 2. Both CO and H2 conversion were the highest in FCx-
400R. Compared to FCx-300R (Fe3O4), FCx-400R (carbon-deficient
iron carbides) and FCx-500R (Fe3C) had stronger CO adsorption site
(CO-TPD in Fig. 4), and it resulted in higher CO conversion. In addi-
tion, the BET surface of FCx-400R and −500R was much higher than
FCx-300R (Table S1), it also enhanced CO conversion of iron car-
bide catalysts [18]. However, it seemed that CO conversion was not
only influenced by the strength of CO adsorption, because FCx-400R
achieved higher CO conversion than FCx-500R (Fe3C), although FCx-
500R had stronger CO adsorption site than FCx-400R (CO-TPD in
Fig. 4). Because H2 conversion was much higher in FCx-400R than
FCx-500R, this higher CO conversion might be resulted from higher
H2 activation ability of FCx-400R (H2-assisted conversion of CO).
The surface basicity (electron donor) promotes H2 activation [48],
thus, the electron-rich state of carbon-deficient iron carbides pro-
moted the H2 adsorption (discussed in detail below in the chapter
‘3.2. Effect of Cu on the activity of the FC catalysts’). As a result, the
electron-rich state of carbon-deficient iron carbides might cause
higher H2 activation ability of FCx-400R, which resulted in higher
H2 and CO conversion.
Fig. 4. CO-TPD results of FC15-300R, 400R and 500R.
The influence of the type of Fe species on the reaction selec-
tivity at similar CO conversion (97.2–97.5%) was examined using
FC15 catalysts reduced at different temperatures, and presented in
Table 3. The C2+ hydrocarbon selectivity (the sum of C2–C4 and C5+
selectivity) increased from 41.9% to 45.2% and 49.8% as the degree of
the carburization increased (FC15-300R (Fe3O4-rich) < FC15-400R
(carbon-deficient, FeCx-rich) < FC15-500R (Fe3C-rich)) while CH4
(27.8, 26.3, and 22.1%, respectively) and CO2 selectivity (30.3, 28.4,
and 28.1%, respectively) decreased. Also, the ␣ values showed that
FC15-300R produced hydrocarbons that were relatively short in
their carbon number (␣ = 0.15) compared to those produced by
FC15-400R or 500R (␣ = 0.20 for both). Overall, the carbon chain
growth of hydrocarbons was promoted as the carburization degree
of the Fe phase increased. It is known in the literature that CO reacts
middle temperature (400 ◦C) (FC15-300R > 400R > 500R). The effect
of Cu promotion is discussed below in the chapter ‘3.2. Effect of Cu
on the activity of the FC catalysts’.
A CO-TPD analysis was performed for the reduced FC cata-
lysts, and the results are presented in Fig. 4. The total amount of
adsorbed CO was the largest with FC15-300R and decreased in the
order FC15-300R > 500R > 400R. However, the strong CO adsorption
sites, from which CO was desorbed at a temperature higher than
550 ◦C, were the largest with FC15-500R, followed by FC15-400R
and FC15-300R. This result implies that CO adsorption is stronger
on iron carbides than on iron oxides, and the adsorption strength
becomes stronger as the carburization degree of the carbide species
increases. This result agrees well with the literature results: The cal-
culations based on density functional theory (DFT) indicated that
the CO adsorption energy over the (100) surface of Fe3C ranges
from −1.22 to −1.77 eV [46], while those on the Fe3O4 (001) surface
range from −0.19 to −0.81 eV [47], which means that CO adsorbs
on the Fe3C (100) surface more easily than on Fe3O4 (001). Com-
paring carbon-deficient FeCx (FC15-400R) with Fe3C (FC15-400R)
in the CO-TPD results, carbon-deficient FeCx was weaker and had
fewer strong chemisorption sites than Fe3C. To summarize the CO-
TPD results, the CO chemisorption strength increased in the order
Fe3O4 < carbon-deficient FeCx < Fe3C.
with the carbon sites of iron carbides (FeCx, Fe3C) to form C
C O
(ketenylidene) species, which is the key intermediate of the car-
bon chain growth [49,50]. On the contrary, CO adsorbs as a bent or
straight CO2-like molecule on the surface of iron oxide (Fe3O4) [47],
which is unfavorable to carbon chain growth, thus the formation of
methane is promoted. Based on CO-TPD results, it was concluded
that the CO chemisorption strength and the formation of C
C O
species increased as the degree of carburization increased, which
resulted in the increase of the carbon chain growth in the reaction.
The carbon balance of each FC catalyst after 10 h reaction test
is presented in the last column of Table 3. It shows that the bal-
ance was well preserved (99.3–100.4%) for FCx-300R and FCx-400R.

Y.H. Lee et al. / Journal of Molecular Catalysis A: Chemical 425 (2016) 190–198
195
Table 3
Product distributions of FC catalysts at similar CO conversion (∼97%): the reaction conditions: 300 ◦C, 10 bar, w/f = 0.01 g/mL/min, CO:H2:N2 = 24:72:4, reaction time for
FC15-300R: 1 h; FC15-400R: 10 h; FC15-500R: 4 h; FC40-400R and FC6-400R: 10 h.
b
Reduction temp. (◦C)
CO conv. (%)
Selectivity (%)
␣
p/oc
CH4
C2–C4
C5+
CO2
a
C2H6
C2H4
C3H8
C3H6
C4H10
C4H8
FC15
300
400
500
400
97.2
97.5
97.5
97.4
97.5
97.7
27.8
26.4
22.1
25.8
26.4
25.3
34.4
10.6
34.4
9.9
36.6
9.4
34.6
9.6
34.4
9.9
7.5
30.3
28.4
28.1
30.0
28.4
28.9
0.15
0.20
0.20
0.19
0.20
0.19
2.04
2.51
1.63
1.98
2.51
2.04
1.5
0.9
1.5
1.4
0.9
1.3
6.8
8.1
6.6
7.2
8.1
7.3
7.1
6.5
9.0
7.4
6.5
7.3
4.1
5.3
5.1
4.7
5.3
4.7
4.3
3.7
5.0
4.3
3.7
4.2
10.8
13.2
9.6
FC40
FC15
FC6
10.8
11.2
34.6
9.8
a
Only n-C4H10 was detected.
b
c
The ␣ values of the FC catalysts were calculated based on the Anderson-Flory-Shultz distribution in the C1–C6 range.
The molar ratio of paraffin to olefin in C2–C4 fractions.
Fig. 6. H2-TPD results of FCx-400R (x = 40, 15, 6) and pure Fe2O3.
Fig. 5. Fe 2p and Cu 2p and C 1 s XPS spectra for FC40, FC15 and FC6-400R.
were generated by the Cu promotion. The ratio between the peak
for a metal-like Fe and that for normal Fe (711.0 eV) was differed by
the Fe/Cu ratio, and was the largest in FC15-400R, while the Cu con-
centration on the surface simply increased as Cu/Fe ratio increased
(Fe/Cu ratio decreased). Because the carbon concentration on the
surface of FCx-400R catalysts was the largest in FC15-400R (Table
S2), it implies that the electron-rich FeCx was generated by hinder-
ing the carbon incorporation and Fe-C bond formation during the
pre-reduction. It also implies that Cu blocked the carbon incorpo-
ration, but it is not clear why the electron-rich FeCx was the most
abundant in FC15-400R. This order might be associated with how
the thermal agglomeration of copper (Tamman temperature of Cu
≈ 500 ◦C) is balanced by its interaction with the iron species on the
catalyst surface.
Fig. 6 presents the H2-TPD results of the FC catalysts reduced at
400 ◦C. Two peaks appeared in all samples at 450–500 ◦C and at over
500 ◦C. It was observed that Fe2O3-400R also chemisorbed H2, but
the chemisorption increased (i.e., the peak intensity increased) and
became stronger (i.e., the peak temperature increased) as the Cu
loading of the catalyst increased. As mentioned above, the electron-
rich surface promoted the adsorption of H2 [48], thus, the amount of
chemisorbed H2 was the largest in FC15-400R, which had the most
electron-rich Fe site. For FC40-400R and FC6-400R, XPS spectra of
Fe for these two catalysts were almost identical, but the amount of
chemisorbed H2 was larger on FC6-400R than FC40-400R. Because
FCx-500R was slightly deficient in the carbon balance (95.5–96.0%),
which might be due to the occurrence of coke formation on the cat-
alyst surface. It implies that the strong chemisorption of CO might
cause a coking problem. A Fischer-Tropsch study revealed that
the rate of coke deposition increases as the carburization degree
of the Fe phase increases (␪-Fe3C < ␹-Fe5C2 < ␧-Fe2.2C) [51]. This
agrees well with the carbon balance of the reaction results in our
study (Table 3): FCx-300R and −400R were almost 100% in carbon
balance, whereas FCx-500R (Fe3C), which had the highest carbur-
ization degree, was deficient by 4% from having a perfect carbon
balance.
In summary, the CO conversion was the highest with FC15-400R,
while the highest hydrocarbon selectivity and yield were achieved
using FC15-500R. During the 10 h reaction tests, the coke deposi-
tion was almost absent except FCx-500R catalysts.
3.2. Effect of Cu on the activity of the FC catalysts
3.2.1. The influence of Cu loading on surface properties of the
reduced FC catalysts
Fig. 5 presents the XPS analysis results of the FCx-400R (x = 40,
15 and 6) catalysts. First, the peak for a metal-like Fe (707.3 eV)
in Fe2O3-400R (without Cu) was much smaller than those in FCx-
400R, which indicated that the metal-like, electron-rich Fe site

196
Y.H. Lee et al. / Journal of Molecular Catalysis A: Chemical 425 (2016) 190–198
Table 4
Calculated heating values of the product gases.
Catalyst
Heating value of the product gasa
52.3 MJ/Nm3
Product gas composition (vol%; without N2b, H2Oc)
H2
69.7
14.8
15.5
68.7
13.5
17.8
71.4
13.5
15.1
FC15-300R
COx
Hydrocarbons
53.4 MJ/Nm3
56.4 MJ/Nm3
FC15-400R
FC15-500R
H2
COx
Hydrocarbons
H2
COx
Hydrocarbons
a
The heating values were calculated considering C1–C4 fractions only. The values are meaningful because C5+ fractions are minor in amount, and all the non-hydrocarbon
species will be separated in the recovery of SNG.
b
N2 was included as an internal standard for GC analysis.
H2O(v) was assumed to be separated.
c
Table 5
H2 conversion and moles of H2 converted and generated during the reaction using FC15 catalysts; the reaction conditions: 300 ◦C; 1 MPa; w/f = 0.01 g/mL/min; feed composition
CO:H2:N2 = 24:72:4.
p/oa
b
CO
conv.
(%)
H2
conv.
(%)
The amount of H2 in the flow stream (mmol/min)
Reduction
temperature
(◦C)
Time on
Stream
␣
H2 input
(feed)
Converted by
hydrogenation of
CO
Generated
by WGSR
Remaining
in the product
stream
FC15
300
400
500
1 h
10 h
4 h
97.2
97.5
97.5
51.8
48.8
46.9
2.04
2.51
1.63
0.15
0.20
0.20
1.059
0.662
0.614
0.593
0.114
0.097
0.096
0.511
0.542
0.562
a
The molar ratio of paraffin to olefin in C2–C4 fractions.
The ␣ values of the FC catalysts were calculated based on the Anderson-Flory-Shultz distribution in the C1–C6 range.
b
the surface Cu concentration was larger in FC6-400R than FC40-
400R, it implies that H2 also chemisorbed on Cu sites.
H2 is inversely proportional to CO conversion (i.e., H2 conversion
increases as the CO conversion increases): the H2 conversion was
in the same order of CO conversion (Table 2). At similar CO con-
versions (shown in Table 5), the concentration of remaining H2
is determined according to the carbon chain growth (␣) and the
paraffin-to-olefin ratio (p/o) of the hydrocarbon fraction of a prod-
uct gas. First, the increase in the carbon chain growth (␣) lowered
the H2 conversion. As given in Eqs. (8) and (9), the molar ratio of
H2/CO (converted) is between 2 (n = ∞) and 3 (n = 1).
3.2.2. The influence of Cu loading on the activity of the reduced
FC catalysts
The influence of Cu loading on the product selectivity was exam-
ined using FCx-400R catalysts. The CO conversion was adjusted at
ca. 97% by allowing a different reaction time for each catalyst. The
results are presented in Table 3, showing the product selectivity did
not change much according to the Cu loading of the catalysts. How-
ever, the p/o ratio increased in the order FC15 > FC6 > FC40 (from
the tests of FCx-400R, Table 3), which was identical to the order of
the catalysts regarding the H2 adsorption amount (peak areas) in
the H2-TPD tests (Fig. 6). These results imply that the major role
of adsorbed H2 was to saturate olefins rather than to restrict the
carbon chain growth of hydrocarbons.
nCO + (2n + 1)H2 → CnH2n+2 + nH2O
(8)
1
(H2/CO (converted) = 2 + ⁄n for production of paraffins)
nCO + 2nH2 → CnH2n + nH2O(H2
/CO(converted) = 2 for production of olefins)
(9)
In summary, the p/o ratio was the highest with FC15-400R,
which was the most prominent for its H2 adsorption. Therefore,
the impregnation of Cu was deduced to not be as effective as the
change in the Fe active phase at improving C2–C4 selectivity, but
it was effective at increasing the paraffin-to-olefin (p/o) ratio of
hydrocarbon products.
Because the molar ratio of H2/CO (converted) decreases as the
carbon number of produced paraffin (n) increases (Eq. (8)), the con-
centration of the remaining hydrogen increases as the carbon chain
growth (␣) increases only if the CO conversion is fixed at a certain
value. This fixed value corresponds to the data in Table 5, in which
the amount of H2 converted by hydrogenation of CO decreased from
0.662 to 0.614 mmol/min as the ␣ value increased from 0.15 (FC15-
300R in 1 h reaction) to 0.20 (FC15-400R in 10 h reaction). In the
cases of producing olefins, the molar ratio of H2/CO (converted) is
always 2, regardless of the carbon number of the olefin (Eq. (9)).
Thus, an increase in the olefin selectivity would accompany a slight
decline in the H2 conversion. This trend corresponds well to the
aforementioned data of Table 5: the amount of H2 converted by
the hydrogenation of CO decreased from 0.614 to 0.593 mmol/min
as the p/o ratio decreased from 2.51 (FC15-400R with a 1 h reaction)
to 1.63 (FC15-500R with a 4 h reaction).
3.3. Evaluation of heating values for the produced SNG gases
Table 4 presents the heating values of the product gases of the
FC15 catalysts reduced at different temperatures. Excluding the
contributions from non-hydrocarbon gases (H2, COx, and H2O), and
C5+ hydrocarbons, the heating values of the product gases (SNG
fraction, C1–C4) were in the range of 52.3–56.4 MJ/Nm3, which
exceeds the standard heating value (41–44.4 MJ/Nm3) of the nat-
ural gas currently distributed over South Korea. However, the
product gases contained very high concentrations of remaining H2
(68.7–71.4 vol.% for FC15 catalysts, N2-free and dry basis). This is
unavoidable in HC-SNG synthesis because the unreacted hydrogen
always remains if Cn (n > 2) hydrocarbons are formed from feed gas
with a H2/CO ratio of 3. Basically, the concentration of remaining
The concentration of the remaining hydrogen is also influenced
by the degree to which the water-gas shift reaction (WGSR) pro-
gressed over the catalyst. Fe-Cu catalysts are well known for their
WGSR activity [52–56]. Hence, H2 can be generated in situ by the

Y.H. Lee et al. / Journal of Molecular Catalysis A: Chemical 425 (2016) 190–198
197
parallel occurrence of the WGSR (Eq. (10)) during the production
of high-calorie SNG. The generated moles of H2 by WGSR could
be calculated directly from the moles of CO2 in the product gas.
CO2 can also be generated via the Boudouard reaction (CO dispro-
portionation) (Eq. (11)), producing solid carbon as a co-product.
However, the formation of solid coke was considered negligible
because the carbon balances were close to 100% for most of the
catalysts (Table 2). Therefore, the contribution of the Boudouard
reaction was disregarded in the analysis of the WGSR that occurred
over the FCx catalysts.
which consequently resulted in an increase in the carbon chain
growth of hydrocarbons. On the other hand, the loading amount
of Cu exerted little influence on the product selectivity, but the
impregnated Cu promoted H2 adsorption, leading to an increase in
the p/o ratio of the produced hydrocarbons. As a result, the heating
values of the SNGs produced by Fe-Cu catalysts were estimated to
be 52.3–56.4 MJ/Nm3, which satisfied the heating value standard
of commercial LNG distributed over South Korea.
Acknowledgments
CO + H2O → CO2 + H2
2CO → C + CO2
(10)
(11)
This work was supported by the National Research Foundation
of Korea (NRF) grant funded by the Korea government (MSIP) (NRF-
2015R1A2A1A13001856).
Finally, the amount of remaining H2 in the product gas could
be attained by subtracting the amount of converted H2 from the
CO hydrogenation from the amount of H2 in the feed and adding
the amount of generated H2 via WGSR. Table 5 shows that most
of the remaining H2 (mmol/min) was from the unreacted H2, and
the contribution of WGSR-generated H2 was comparatively small.
Therefore, it could be summarized that the carbon chain growth
and the olefin formation can mostly be attributed to the high con-
centration of remaining H2 in the high-calorie product gas.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.molcata.2016.10.
013.
References
The fact that a high concentration of remaining H2 is unavoid-
able means a post-process such as hydrogen recovery or additional
methanation should be considered in designing or installing HC-
SNG production facilities. The use of post-methanation reactor
units is the most promising method because it not only reduces
H2 concentration but also makes additional methane. In general,
the commercial methanation processes consist of 2–6 methanation
reactors, which can obtain nearly complete conversions of hydro-
gen and carbon oxides [11]. The methanation reactors are linked via
an optimized set of recycle or bypass streams. For example, recy-
cling a part of product stream into the feed stream is recommended
for the first methanation reactor because the first methanation
reactor requires a feed with a high H2/CO ratio (=3), and its prod-
uct stream is more H2-rich than those of following methanation
reactors. The product stream of high-calorie SNG is much more H2-
rich (H2/CO = 29.4–61.4) than those of simple methanation. If such
a H2-rich product stream is recycled to the first reactor, it would
be possible to reduce the size of the WGSR unit (or to remove
the unit, in the best-case scenario), which is positioned in front
of the first methanation unit to adjust the H2/CO ratio for metha-
nation. In addition, a part of the feed gas (a synthesis gas from a
coal gasification unit, which is relatively CO-rich, H2/CO = ∼1) to
the WGSR reactor can be by-passed to the feed stream to the post-
methanation reactors, by which the H2/CO ratio required for the
reactors can be adjusted (3, most desirably).
[1] World Energy Outlook 2015, International Energy Agency, 2015.
[2] Statistical Review of World Energy, British Petroleum Company, 2015.
[3] How much carbon dioxide is produced when different fuels are burned?, U.S.
Energy Information Administration (EIA), 2016.
[4] R. Muller, Energy for Future Presidents: The Science Behind the Headlines, 1st
ed., W.W. Norton, New York, 2012.
[5] World Energy Outlook 2011, International Energy Agency, 2011.
[6] Y. Nishiyama, Energy in Japan: On the Introduction of Shale Gas, Credit Suisse,
2012.
[7] Heat Content of Natural Gas Consumed, U.S. Energy Information
Administration (EIA), 2015.
[8] M. Korchemkin, Gazprom Unlikely to Win a Price War, East European Gas
Analysis (EEGA), 2016.
[9] Gazprom, Scale Does Matter: Gazprom Investor Day, 2012.
[10] Description of the city gas calories system documentation, Korea Gas
Corporation (KOGAS), 2012.
[11] J. Kopyscinski, T.J. Schildhauer, S.M.A. Biollaz, Fuel 89 (2010) 1763–1783.
[12] National Institute of Standards and Technology (U.S.), NIST chemistry
webbook, in: NIST Standard Reference Database 69, National Institute of
Standards and Technology, Washington, D.C, 2016.
[13] M.E. Dry, FT catalysts, in: M.E. Dry, A.P. Steynberg (Eds.), Fischer-Tropsch
Technology, 2004, pp. 533–600.
[14] B.C. Enger, A. Holmen, Catal. Rev. 54 (2012) 437–488.
[15] T. Inui, A. Sakamoto, T. Takeguchi, Y. Ishigaki, Ind. Eng. Chem. Res. 28 (1989)
427–431.
[16] Y. Ishigaki, M. Uba, S. Nishida, T. Inui, Appl. Catal. 47 (1989) 197–208.
[17] Y.H. Lee, H. Kim, H.S. Choi, D.W. Lee, K.Y. Lee, Korean J. Chem. Eng. 32 (2015)
2220–2226.
[18] Y.H. Lee, D.-W. Lee, H. Kim, H.S. Choi, K.-Y. Lee, Fuel 159 (2015) 259–268.
[19] H.M.T. Galvis, J.H. Bitter, C.B. Khare, M. Ruitenbeek, A.I. Dugulan, K.P. de Jong,
Science 335 (2012) 835–838.
[20] J.Z. Lu, L.J. Yang, B.L. Xu, Q. Wu, D. Zhang, S.J. Yuan, Y. Zhai, X.Z. Wang, Y.N. Fan,
Z. Hu, ACS Catal. 4 (2014) 613–621.
[21] N. Jiang, G.H. Yang, X.F. Zhang, L. Wang, C.Y. Shi, N. Tsubaki, Catal. Commun.
12 (2011) 951–954.
4. Conclusions
[22] X.Q. Chen, D.H. Deng, X.L. Pan, Y.F. Hu, X.H. Bao, Chem. Commun. 51 (2015)
217–220.
[23] Y. Liu, J.F. Chen, J. Bao, Y. Zhang, ACS Catal. 5 (2015) 3905–3909.
[24] H.M.T. Galvis, J.H. Bitter, T. Davidian, M. Ruitenbeek, A.I. Dugulan, K.P. de Jong,
J. Am. Chem. Soc. 134 (2012) 16207–16215.
[25] S.H. Kang, J.W. Bae, P.S.S. Prasad, K.W. Jun, Catal. Lett. 125 (2008) 264–270.
[26] Z.Q. Yang, X.L. Pan, J.H. Wang, X.H. Bao, Catal. Today 186 (2012) 121–127.
[27] Y. Liu, J.F. Chen, Y. Zhang, RSC Adv. 5 (2015) 29002–29007.
[28] C. Wang, Q.X. Wang, X.D. Sun, L.Y. Xu, Catal. Lett. 105 (2005) 93–101.
[29] J.D. Xu, K.T. Zhu, X.F. Weng, W.Z. Weng, C.J. Huang, H.L. Wan, Catal. Today 215
(2013) 86–94.
[30] M.E. Dry, A.P. Steynberg, Commercial FT process applications, in: M.E. Dry,
A.P. Steynberg (Eds.), Fischer-Tropsch Technology, 2004, pp. 406–481.
[31] A. Sarkar, D. Seth, A.K. Dozier, J.K. Neathery, H.H. Hamdeh, B.H. Davis, Catal.
Lett. 117 (2007) 1–17.
In this work, the activity of Fe-Cu catalysts in the production
of high-calorie SNG was investigated. The major active Fe phase of
the Fe-Cu catalysts changed according to the CO reduction (pre-
treatment) temperature: Fe3O4 at 300 ◦C, carbon-deficient FeCx at
400 ◦C and Fe3C at 500 ◦C. The carburization of iron oxide resulted
in stronger CO adsorption and higher BET surface area, as a result,
iron carbide catalysts (FCx-400R and −500R) achieved higher con-
version than Fe3O4 (FC-300R). Also, the carbon-deficient FeCx was
metal-like in its electron structure due to the low number of bonded
carbons, and it was attributed to the highest CO and H2 conver-
sion of FC15-400R by providing H2 activation ability. For product
selectivity, the production of Cn (n > 2) hydrocarbons increased as
the degree of carburization increased: FC15-500R > 400R > 300R.
As the degree of carburization increased (Fe3C > carbon-deficient
FeCx > Fe3O4), the chemisorption strength of CO also increased,
[32] M.Y. Ding, Y. Yang, B.S. Wu, J. Xu, C.H. Zhang, H.W. Xiang, Y.W. Li, J. Mol. Catal.
A-Chem. 303 (2009) 65–71.
[33] M.S. Luo, H. Hamdeh, B.H. Davis, Catal. Today 140 (2009) 127–134.
[34] M.Y. Ding, Y. Yang, Y.W. Li, T.J. Wang, L.L. Ma, C.Z. Wu, Appl. Energy 112
(2013) 1241–1246.

198
Y.H. Lee et al. / Journal of Molecular Catalysis A: Chemical 425 (2016) 190–198
[35] C.-H. Zhang, Y. Yang, B.-T. Teng, T.-Z. Li, H.-Y. Zheng, H.-W. Xiang, Y.-W. Li, J.
Catal. 237 (2006) 405–415.
[36] K.-D. Jung, A.T. Bell, J. Catal. 193 (2000) 207–223.
[37] S. Li, W. Ding, G.D. Meitzner, E. Iglesia, J. Phys. Chem. B 106 (2002) 85–91.
[38] S. Li, S. Krishnamoorthy, A. Li, G.D. Meitzner, E. Iglesia, J. Catal. 206 (2002)
202–217.
[39] H. Schultz, Fischer-Tropsch synthesis in a modern perspective, in: A. Beller
(Ed.), Catalysis: From Principles to Applications, 2011.
[40] C.N. Satterfield, Heterogeneous Catalysis in Industrial Practice, 2nd ed.,
McGraw-Hill, New York, 1991.
[41] Y. Liu, J. Luo, J. Girleanu, O. Ersen, C. Pham-Huu, C. Meny, J. Catal. 318 (2014)
179–192.
[46] J.G. Wang, H.J. Jian, J. Mol. Catal. A-Chem. 269 (2007) 169–178.
[47] P.Y. Xue, Z.M. Fu, X.L. Chu, Y.X. Zhang, Z.X. Yang, Appl. Surf. Sci. 317 (2014)
752–759.
[48] A. Travert, H. Nakamura, R.A. van Santen, S. Cristol, J.F. Paul, E. Payen, J. Am.
Chem. Soc. 124 (2002) 7084–7095.
[49] J.W. Kolis, E.M. Holt, D.F. Shriver, J. Am. Chem. Soc. 105 (1983) 7307–7313.
[50] W.C. Wu, Z.L. Wu, C.H. Liang, X.W. Chen, P.L. Ying, C. Li, J. Phys. Chem. B 107
(2003) 7088–7094.
[51] A. Machocki, Appl. Catal. 70 (1991) 237–252.
[52] A. Khan, P.G. Smirniotis, J. Mol. Catal. A-Chem. 280 (2008) 43–51.
[53] M. Estrella, L. Barrio, G. Zhou, X.Q. Wang, Q. Wang, W. Wen, J.C. Hanson, A.I.
Frenkel, J.A. Rodriguez, J. Phys. Chem. C 113 (2009) 14411–14417.
[54] L.Z. Zhang, J.M.M. Millet, U.S. Ozkan, Appl. Catal. A-Gen. 357 (2009) 66–72.
[55] D.W. Jeong, V. Subramanian, J.O. Shim, W.J. Jang, Y.C. Seo, H.S. Roh, J.H. Gu, Y.T.
Lim, Catal. Lett. 143 (2013) 438–444.
[42] A. Demirbas, Methane Gas Hydrate, Springer, London; New York, 2010.
[43] Sasol, Natural Gas: Addendum to specification.
[44] B.D. Cullity, Elements of X-ray Diffraction, Addison-Wesley Pub Co., Reading,
Mass, 1956.
[56] Z.H. Bao, W.Z. Ding, Q. Li, Int. J. Hydrogen Energy 37 (2012) 951–955.
[45] G. Routschka, Refractory Materials: Pocket Manual; Design, Properties,
Testing, 3rd ed., Vulkan-Verlag GmbH, 2008, 2016.