Product distribution of CO2 hydrogenation by K- and Mn-promoted Fe catalysts supported on N-functionalized carbon nanotubes

Article abstract of DOI:10.1016/j.cattod.2016.02.045

An iron based catalyst supported on an N-functionalized carbon nanotube (NCNT) was promoted with potassium and manganese as follows: Fe/NCNT, K/Fe/NCNT, Mn/Fe/NCNT, and K/Mn/Fe/NCNT for CO₂ hydrogenation. Time-resolved reduction X-ray absorption near edge spectroscopy (XANES) showed mixed phases of Fe, FeO, Fe₃O₄, and Fe₂O₃ resulting from K/Fe/NCNT, and of FeO and Fe₃O₄ resulting from Mn/Fe/NCNT. The product distributions and growth probability of n-alkanes during CO₂ hydrogenation indicated that the potassium-promoted iron catalysts performed Fischer-Tropsch (FT) synthesis under steady state at 60?h. 1-Alkenes desorbed from the FT sites with the potassium-promoted catalysts, (K/Fe/NCNT and K/Mn/Fe/NCNT), with low methane formation. Small amounts of 1-alkene, along with high methanation, were produced from the potassium-unpromoted catalysts, (Fe/NCNT and Mn/Fe/NCNT), indicating high local H₂:CO ratios on the catalyst surfaces. K/Fe/NCNT and K/Mn/Fe/NCNT catalysts also produced ethanol. Thus, potassium is a key promoter providing active species of the catalysts for alkene and ethanol formation. Reduced surrounding of the NCNT support, potassium as an electronic promoter together with manganese as a structural promoter made the iron-active phase well suitable for CO₂ hydrogenation producing mainly alkenes and ethanol.

Full text of DOI:10.1016/j.cattod.2016.02.045

G Model
CATTOD-10078; No. of Pages7
ARTICLE IN PRESS
Catalysis Today xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Product distribution of CO2 hydrogenation by K- and Mn-promoted Fe
catalysts supported on N-functionalized carbon nanotubes
Praewpilin Kangvansuraa, Ly May Chewb, Worasarit Saengsuic, Phatchada Santawajad,
Yingyot Poo-arporne, Martin Muhlerb, Hans Schulzf, Attera Worayingyongc,∗
a Scientific Equipment Center, Faculty of Science, Kasetsart University, Bangkok 10900, Thailand
b Laboratory of Industrial Chemistry, Ruhr-University Bochum, 44801 Bochum, Germany
c Department of Materials Science, Faculty of Science, Kasetsart University, Bangkok 10900, Thailand
d Department of Chemistry, Faculty of Science, Kasetsart University, Bangkok 10900, Thailand
e Synchrotron Light Research Institute (Public Organization), Nakhon Ratchasima 30000, Thailand
f
Engler-Bunte Institute, Karlsruhe Institute of Technology, 76131 Karlsruhe, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
An iron based catalyst supported on an N-functionalized carbon nanotube (NCNT) was promoted with
potassium and manganese as follows: Fe/NCNT, K/Fe/NCNT, Mn/Fe/NCNT, and K/Mn/Fe/NCNT for CO2
hydrogenation. Time-resolved reduction X-ray absorption near edge spectroscopy (XANES) showed
mixed phases of Fe, FeO, Fe3O4, and Fe2O3 resulting from K/Fe/NCNT, and of FeO and Fe3O4 resulting
from Mn/Fe/NCNT. The product distributions and growth probability of n-alkanes during CO2 hydro-
genation indicated that the potassium-promoted iron catalysts performed Fischer-Tropsch (FT) synthesis
under steady state at 60 h. 1-Alkenes desorbed from the FT sites with the potassium-promoted cat-
alysts, (K/Fe/NCNT and K/Mn/Fe/NCNT), with low methane formation. Small amounts of 1-alkene,
along with high methanation, were produced from the potassium-unpromoted catalysts, (Fe/NCNT and
Mn/Fe/NCNT), indicating high local H2:CO ratios on the catalyst surfaces. K/Fe/NCNT and K/Mn/Fe/NCNT
catalysts also produced ethanol. Thus, potassium is a key promoter providing active species of the cat-
alysts for alkene and ethanol formation. Reduced surrounding of the NCNT support, potassium as an
electronic promoter together with manganese as a structural promoter made the iron-active phase well
suitable for CO2 hydrogenation producing mainly alkenes and ethanol.
Received 1 September 2015
Received in revised form 20 January 2016
Accepted 29 February 2016
Available online xxx
Keywords:
CO2 hydrogenation
Fischer-Tropsch synthesis
Fischer-Tropsch iron catalyst
In-situ XANES reduction
Product distributions of CO2 hydrogenation
© 2016 Elsevier B.V. All rights reserved.
1. Introduction
Iron is mainly used, due to its activity for both RWGS and FTS.
It forms active species via CO decomposition, producing surface
CO2 hydrogenation plays a significant role in environmental
remediation. It composes of two steps: 1) a reverse water gas shift
(RWGS) reaction over a transition metal such as iron, cobalt, nickel,
etc. and 2) Fischer-Tropsch synthesis (FTS). Those steps are shown
in Eqs. (1) and (2) [1]:
carbon which may react in several ways forming iron carbides.
At the steady state of FTS, the iron catalyst could be constructed
according to the degree of catalyst alkalization and composition
of the syngas [2]. Riedel et al. [2] reported that the episodes (the
period of reaction time in minutes) of Fischer-Tropsch synthesis
(over Fe-Al-Cu-K catalyst) corresponded to the real catalyst struc-
ture and composition. For the reaction of H2/CO2 syngas, the steady
state of the Fischer-Tropsch synthesis is reached in 2500–5000 min,
which is much more than that resulting from H2/CO syngas (about
1000 min), because of less CO being available. The reverse water
gas shift (RWGS) reaction with H2/CO2 syngas occurred during
100–1000 min reaction time producing CO from CO2 (Eq. (1)). It has
been concluded that the consumption of CO by FTS causes overall
low CO yield. The catalyst construction process is the same, either
CO or CO2 being the reactants in the hydrogenation process over
alkaline promoted catalysts.
RWGS : CO2 + H2 ꢀ CO + H2OꢀRH573K = +38 kJ/mol
(1)
FTS : nCO + (2n + 1)H2 → CnH2n+2 + nH2OꢀRH573K = −166 kJ/mol
(2)
FTS is the reaction of syngas (CO + H2) converting into long chain
hydrocarbons, i.e. alkanes, alkenes and/or oxygenated compounds
in the presence of a transition metal, either iron or cobalt, etc.
∗
Corresponding author.
E-mail addresses: fsciarw@ku.ac.th, aworayingyong@gmail.com
(A. Worayingyong).
http://dx.doi.org/10.1016/j.cattod.2016.02.045
0920-5861/© 2016 Elsevier B.V. All rights reserved.
Please cite this article in press as: P. Kangvansura, et al., Product distribution of CO2 hydrogenation by K- and Mn-promoted Fe catalysts
supported on N-functionalized carbon nanotubes, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.02.045

G Model
CATTOD-10078; No. of Pages7
ARTICLE IN PRESS
2
P. Kangvansura et al. / Catalysis Today xxx (2016) xxx–xxx
A catalyst support also plays an important role for active species
dispersion and stabilizing the catalyst in its active state during the
reaction time. Carbon nanotubes (CNTs) are chosen as the support
for many applications, such as magnetic materials, sensors and cat-
alysts [3]. Recently, the conversion of syngas to hydrocarbons and
to alcohols using metal catalysts supported on carbon materials
have been developed [4]. In the catalysis of FTS over an iron cata-
lyst, the reduction of Fe2O3 and Fe3O4 forming iron carbide as the
active catalyst is an important step, which could be more facile for
a carbon-supported iron catalyst, with a weak interaction between
iron and carbon [4,5]. Using CNTs as the support, the carbon-iron
(␲-d) interaction of curve surface of CNTs facilitates the reduction
process [6], and nitrogen of N-functionalized CNT (NCNT) causes
some distortion on such interaction [4]. It is also important in the
catalysis of FTS to have iron in the reduced form during the reaction
time, thus used iron catalysts within the confined environment as
CNTs should be reduced much easily [3,4]. It has been tested that the
method of preparation of Fe/CNT did not affect the conversion of the
syngas [7]. CNT modification with 10 M nitric acid produced high
degree of iron dispersion inside NCNTs, and the catalysts showed
high reducibility, which enhanced the FTS performance [8]. The
catalyst particles do not form strong interactions with the carbon
support, thus their activities are higher than that on oxide supports,
such as SiO2 [1,4]. Chew et al. used H2-temperature programmed
reduction (H2-TPR) and in situ XANES to show that iron oxides
nanoparticles on NCNT were easier to reduce than those on OCNT
indicating low interaction between iron oxides and NCNT. Scan-
ning transmission electron microscopic (STEM) images showed
iron oxide nanoparticles located within the NCNT [1]. CO2 adsorp-
tion, on catalyst surface, also plays an important role in the FTS
reaction. For example, 1,4-bis(hydroxydimethylsilyl) benzene was
used to modified zeolite surface and altered the surface property
of the catalysts for CO2 hydrogenation [9].
Potassium is known as an electronic promoter of the iron cat-
alyst for FTS over promoted Fe/CNTs, producing high molecular
weight alkanes and alkenes [9]. Potassium also decreases methane
production and increases WGS. A potassium promoter increases
the strength of CO chemisorption and suppresses that of hydro-
genation, which causes a high concentration ratio of CO to H2
on catalyst surfaces, therefore causing low hydrogenation activ-
ity and high alkene selectivity [10]. Jiang et al. [11] investigated
the effects of potassium on iron-manganese catalysts by in situ dif-
fuse reflectance Fourier transform infrared spectroscopy (DRIFTS)
using NO, CO and CO + H2 as probes. The potassium species coordi-
nates with iron species and promotes iron oxide reduction forming
fine metallic iron clusters. X-ray photoemission spectroscopy (XPS)
measurements of the K 2p levels showed that potassium forms
oxides on the iron catalyst surface. It is not covered by carbon,
but appears on top of the carbon layer [12]. Xiong et al. [13] found
that sodium- and potassium- promoted Fe/CNT for FTS decreased
the selectivity of methane, increased the selectivity of alkene and
shifted the selectivity of products to higher alkanes and alkenes,
compared to the results of the unpromoted Fe/CNT catalyst. The
sodium-promoted catalyst showed better activity. Yang et al. [10]
showed that the catalysts are further reduced during FTS reaction.
As illustrated by the X-ray diffraction (XRD) analysis, the addition
of potassium promoted the formation of iron carbides during the
FT reaction. When the potassium loading increased, carbon deposi-
tion was significantly enhanced. The influence of potassium content
on the selectivity of alkenes under different temperatures, is con-
sistent with its effect on the strength of carbon monoxide and
hydrogen chemisorption.
agglomeration during the reduction reaction with the syngas. The
effect of manganese promotion of the iron catalyst is in a lim-
ited concentration range [15]. Manganese-to-iron ratios higher
than 0.05 decreases the production of C2-C5, C6+ and oxygenated
products due to high CH4 formation. It was assumed that high
manganese loadings produce high amount of manganese oxide
on the surface during the reaction, which could block CO or CO2
to the active iron phase. The blockage of such phases by the
amorphous MnOx phases should be avoided [15]. Manganese and
potassium promoters coated on carbon nanotubes (MnK-CNTs) as
the catalyst support, then impregnated by iron (Fe/MnK-CNTs with
7.9 wt% Fe, 15.7 wt% Mn and 1.9 wt% K), were used for the direct
conversion of syngas to lower alkenes. The as-synthesized Fe/MnK-
CNTs catalyst compared with the FeMnK/CNTs catalyst prepared
by co-impregnation showed better catalytic performance, leading
to high selectivity for C2-C4 alkenes, which were due to uniform
nanoparticles, well distributed promoters, the weak iron-CNT sup-
port interaction and the defects on CNTs of the unique structural
transformation of MnK-CNTs [16]. However, metal-support inter-
actions do not affect the promoter properties [4].
X-ray absorption near edge spectroscopy (XANES) provides
structural and chemical information on the element of interest,
namely, iron or cobalt of the catalyst [17]. The XANES technique
is suitable for characterization of nanocrystals when limited using
XRD [18]. However, iron catalyst phases can be sensitive to envi-
ronmental conditions, especially oxidation after removing from the
reactor. Thus, iron phases should be monitored using an in situ cell
under the high temperature reduction conditions needed for CO2
hydrogenation.
CO2 hydrogenation, the reverse water gas shift (RWGS) reaction
followed by the FTS under suitable conditions, produces valuable
chemicals such as alkenes or oxygenated hydrocarbons. FTS over
an iron catalyst involves many proposed mechanisms which are
developed from ideas resulting from the observed products and
the conditions, such as high or low pressure and temperature
operations [19]. For low temperature FTS, iron carbide forms dur-
ing the reaction, which involves in the reaction mechanism [19].
The promoted catalyst undergoes restructuring under the reaction
conditions which has an influence on chain growth and hydro-
carbon distribution. Particular product distributions, of alkanes,
alkenes, etc. could be achieved by controlling catalyst composi-
tion which affects the selectivity of secondary reactions [20]. “Time
resolved product selectivity analysis” [21] provides information
on product desorption (methane, alkanes, alkenes and oxygenated
compounds), which is useful for evaluating a promoted catalyst,
especially at the steady state.
Product selectivity and distribution during CO2 hydrogenation
by K- and Mn-promoted Fe catalysts supported on NCNT was
the focus of this work. Fe/NCNT, K/Fe/NCNT, Mn/Fe/NCNT and
K/Mn/Fe/NCNT were used for the hydrogenation reaction. Alkenes
and alcohols, important chemicals for further petrochemical pro-
cesses, recovered from CO2 emission gas, were of interest. Product
distributions of alkanes and alkenes during the CO2 hydrogenation
of all catalysts were used to investigate Fischer-Tropsch (FT) cata-
lyst performances. Time-resolved XANES reduction was also used
for catalyst characterization of iron phases for the reaction.
2. Material and methods
2.1. Catalyst preparation [22]
Manganese, in the form of MnO, has been used as a chem-
ical promoter to develop the chemisorption of the syngas and
Carbon nanotubes (Pyrograf III, diameter 70–200 nm, length
50–100 ␮m from Applied Sciences) were pretreated with HNO3
vapor at 473 K for 24 h to obtain the oxygen-functionalized CNT
(OCNT) [23] and dried at 333 K for 24 h. The OCNT was then treated
as
a structural promoter to stabilize the iron-active phase
during the FTS [14]. It stabilizes the active species against
Please cite this article in press as: P. Kangvansura, et al., Product distribution of CO2 hydrogenation by K- and Mn-promoted Fe catalysts
supported on N-functionalized carbon nanotubes, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.02.045

G Model
CATTOD-10078; No. of Pages7
ARTICLE IN PRESS
P. Kangvansura et al. / Catalysis Today xxx (2016) xxx–xxx
Table 1
3
in flowing NH3 (10 vol% in He) at 673 K for 6 h with a flow rate of
25.50 cm3 min−1, to give the nitrogen-functionalized CNT (NCNT)
with about 5.9% of nitrogen concentration [1]. Iron catalysts were
prepared by an impregnation method using ammonium ferric cit-
rate (C6H8O7·xFe·xNH3) as the iron precursor [24]. The NCNT was
immersed in an aqueous solution of ammonium ferric citrate to
achieve iron loading (40.0 wt% Fe). To prepare a Mn-promoted
iron catalyst, an aqueous solution of 10.0 wt% Mn in manganese(II)
nitrate hydrate, Mn(NO3)2·xH2O, was mixed with an aqueous solu-
tion of ammonium ferric citrate. The as-prepared catalyst was
calcined at 573 K with a heating rate of 2 K min−1 in flowing syn-
thetic air (100.0 cm3 min−1) for 90 min. To obtain the K-promoted
iron catalyst (K/Mn/Fe/NCNT), the calcined sample was suspended
in an aqueous solution of 1.5 wt% K of potassium carbonate (K2CO3).
Subsequently, the impregnated catalyst was dried at 333 K.
Percentage mass of the elemental compositions in the catalysts.
Sample
Mass (%)
Fe
K
Mn
Fe/NCNT
34.2
31.2
32.8
22.5
–
1.3
–
–
–
7.1
5.4
K/Fe/NCNT
Mn/Fe/NCNT
K/Mn/Fe/NCNT
1.0
2.2. Characterization
The catalyst compositions were determined quantitatively
using an atomic absorption spectrometer (AAS) (Varian AA300).
X-ray diffraction (XRD) was performed using a Panalytical MPD
diffractometer to determine the crystalline phases of the catalysts,
using CuK␣ (␭ = 1.54 Å), an electron current of 40 mA and an accel-
erating voltage of 45 kV. The spectra were scanned with step size of
0.02◦ in the 2␪ range of 20–80◦. The crystalline phase was identified
by comparison data from the inorganic crystal structure database,
ICSD.
Fig. 1. XRD patterns of a) Fe/NCNT, b) K/Fe/NCNT, c) Mn/Fe/NCNT and d)
K/Mn/Fe/NCNT after calcination at 573 K.
X-ray absorption spectroscopy (XAS) was performed at beam
line 2.2 of the Synchrotron Light Research Institute (SLRI), Thailand.
Samples were prepared as 4.0 mm diameter pellets. Fe K-edge
XANES measurements were carried out in transmission mode. Iron
metal foil was used for energy calibration at 7112 eV. Data process-
ing was performed using the Athena software. Iron metal foil, FeO,
Fe3O4 and Fe2O3 were used as references.
2.3. Catalytic testing
The experimental method for CO2 hydrogenation set-up and
testing is described in detail elsewhere [25]. A catalyst 40 mg
was mixed with 160 mg of SiC and loaded between quartz plugs
in a reactor. Before the catalytic test, the catalyst mixture was
reduced in flowing H2 (50 cm3 min−1) at 653 K for 5 h. The reac-
tor was then cooled to 633 K in flowing Ar (30 cm3 min−1) to
remove residual H2. The reactant gases, 7.5 cm3 min−1 CO2 mixed
with 22.5 cm3 min−1 H2 and 3.3 cm3 min−1 Ar, flowed continuously
to the reactor through the catalyst bed with a space velocity of
50 dm3 g−1 h−1 at 633 K and 25 bar. The products were directly ana-
lyzed using an online gas chromatograph (Shimadzu, GC-2010 plus)
equipped with a thermal conductivity detector (TCD) and two flame
ionization detectors (FID), with argon as an internal standard.
Fig. 2. In situ XANES spectra of K/Fe/NCNT catalyst reduction from 323 K to 653 K
and 300 min holding time. Inset shows the spectra of iron standards: Fe metal foil,
FeO, Fe3O4 and Fe2O3.
peaks at 35.6◦ and 62.5◦ became stronger. However, Fe3O4 was the
major iron phase of all catalysts.
In situ XANES spectra of K/Fe/NCNT reduction from 323 K to
653 K with 300 min holding time are shown in Fig. 2. Table 2
gives the phase compositions of the iron catalysts, K/Fe/NCNT
and Mn/Fe/NCNT, as deduced from linear combination analyses of
XANES spectra before and after reduction. In the case of K/Fe/NCNT,
Fe2O3 and Fe3O4 phases were detected during reduction for 60 min,
and Fe, FeO, Fe2O3, and Fe3O4 phases were detected during 300 min
holding time. Mn/Fe/NCNT exhibited only the Fe3O4 phase during
60 min reduction time, and a mixture of FeO and Fe3O4 during the
holding time for 300 min.
Figs. 3a and b shows the phase compositions of K/Fe/NCNT and
Mn/Fe/NCNT catalysts deduced from the linear combination analy-
sis of the XANES spectra during reduction and holding time. When
Fe/NCNT was promoted with potassium (K/Fe/NCNT), Fe2O3 was
reduced well to Fe3O4, and then further to FeO until it reached
2.2% at 623 K. The reducibility of FeO increased to 14.5% Fe, 1.2%
FeO, with some remaining 15.7% Fe3O4 and 68.6% Fe2O3 after
300 min holding time. In the case of Mn/Fe/NCNT, 100.0% Fe3O4
was observed until 653 K. Fe3O4 reduction started after holding
at 653 K, reaching 14.8% FeO after holding at 653 K for 300 min.
3. Results and discussion
3.1. Catalyst characterization
The percentage masses of the elemental compositions in the cat-
alysts were analyzed by AAS as shown in Table 1. The XRD patterns
of the unpromoted and promoted Fe/NCNT are illustrated in Fig. 1
[22]. Peaks at 26.3◦ and 44.8◦ can be assigned to the (002) and (111)
planes, respectively, of the hexagonal graphite structures of CNT.
For Fe/NCNT, the peaks at 30.2◦, 35.6◦ and 62.5◦ corresponded to the
magnetite (Fe3O4) phase. When manganese was added to Fe/NCNT,
the intensity of the peaks at 30.2◦, 35.6◦ and 62.5◦ became weaker,
which can be attributed to the loss of crystallinity of Fe3O4. A broad
peak observed between 32◦–34◦ can be attributed to Fe2O3 (33.2◦).
When potassium was added to Mn/Fe/NCNT, the intensity of the
Please cite this article in press as: P. Kangvansura, et al., Product distribution of CO2 hydrogenation by K- and Mn-promoted Fe catalysts
supported on N-functionalized carbon nanotubes, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.02.045

G Model
CATTOD-10078; No. of Pages7
ARTICLE IN PRESS
4
P. Kangvansura et al. / Catalysis Today xxx (2016) xxx–xxx
Table 2
Percentage phase compositions of all catalyst samples deduced from linear combination analyses of XANES spectra during reduction and at 653 K 300 min holding time.
Catalysts
Reduction Temperature (K)
Content (%)
Fe2O3
Reduction holding time (min)
Content (%)
Fe2O3
Fe3O4
FeO
Fe3O4
FeO
Fe
K/Fe/NCNT
323
473
543
623
633
653
80.3
83.0
82.0
76.0
75.9
78.7
19.7
17.0
18.0
21.8
20.2
15.3
–
–
–
2.2
3.9
6.0
10
60
81.0
68.5
69.2
65.1
65.9
66.7
67.1
68.6
5.0
14.0
16.1
16.7
4.4
3.3
2.3
–
–
–
12.0
12.6
13.4
13.7
14.5
15.4
14.1
18.5
18.2
17.6
17.1
15.7
120
180
210
240
270
300
2.1
1.2
Mn/Fe/NCNT
323
473
573
613
653
–
–
–
–
–
100.0
100.0
100.0
100.0
100.0
–
–
–
–
–
30
90
–
–
–
–
–
–
–
–
–
100.0
94.6
90.7
88.2
86.8
86.0
84.6
84.4
85.2
–
5.4
9.3
11.8
13.2
14.0
15.4
15.6
14.8
–
–
–
–
–
–
–
–
–
140
180
210
240
270
290
300
3.2. CO2 hydrogenation
For the hydrogenation reaction, the catalysts a) Fe/NCNT, b)
K/Fe/NCNT, c) Mn/Fe/NCNT and d) K/Mn/Fe/NCNT were used.
Growth probability and product selectivities are discussed.
3.2.1. CO2 conversion and product selectivities at the steady state
CO2 conversion over all catalysts was ca. 30–35%. To study prod-
uct selectivity of CO2 hydrogenation over the catalysts; a) Fe/NCNT,
b) K/Fe/NCNT, c) Mn/Fe/NCNT and d) K/Mn/Fe/NCNT, those of n-
alkanes, alkenes and alcohols at 60 h reaction time are illustrated
in Fig. 4(a–d). Methane produced from Fe/NCNT and Mn/Fe/NCNT
was ca. 70 mol%. Methanation occurs from the metal active sites for
hydrogenation but not for FTS [21]. Besides, a Mn/Fe catalyst sup-
ported on NCNT could cause methane formation by hydrogenolytic
cleavage of alkenes [20], due to a high H2: CO ratio on the metal
surface. The results from K/Fe/NCNT and K/Mn/Fe/NCNT were ca.
25 mol%, due to low hydrogen dissociation corresponding to the
potassium-promoted catalysts suppressing methane formation [5].
Alkane selectivity from the Fe/NCNT and Mn/Fe/NCNT catalysts was
maximum at C3 (ca. 5–10%) but decreased for higher carbon num-
bers indicating that propane products were produced from both
the FT and the readsorption sites [21]. For the reaction over the
K/Fe/NCNT and K/Mn/Fe/NCNT catalysts, n-alkane selectivity was
low at C2 with small amounts of long chain hydrocarbons up to C6.
Among alkene products of different carbon numbers, alkenes
selectivity from the Fe/NCNT and Mn/Fe/NCNT catalysts was high
at C2 (ca. 20%) but decreased for the high carbon numbers, together
with high CH4 selectivity of 70%, indicating that a high local H2:CO
ratio causes CH4 formation by hydrogenolysis of alkenes [20]. With
potassium-promoted catalysts, ethene selectivity was low due to
ethanol formation but the selectivities of higher alkenes were high
and decreased with high carbon numbers.
Fig. 3. Percentage phase compositions of a) K/Fe/NCNT and b) Mn/Fe/NCNT deduced
from linear combination analyses of time-resolved XANES reduction spectra as a
function of temperature (K) and holding time (min).
Alcohol selectivity was obtained in terms of ethanol for 10%
from the potassium-promoted catalysts only. Mn/Fe/NCNT pro-
duced methanol only in small amounts.
From the reduction phases of iron via XANES, it indicated that
with the manganese promoter, the percentage of the phase com-
positions from the reduction decreased. Manganese influences the
iron-containing species by forming MnO, causing high stability of
FeO, which can be reduced at higher temperature [14]. By compar-
ison, potassium-promoted Fe/NCNT was reduced to 14.5% Fe after
holding at 653 K for 300 min (the FTS condition), indicating that the
potassium-promoted Fe/NCNT could provide active iron for carbide
formation available for FTS.
3.2.2. Product distributions and growth probability of n-alkane
Product distributions in terms of log (mole of n-alkane) as a
function of carbon number for Fe/NCNT, K/Fe/NCNT, Mn/Fe/NCNT
and K/Mn/Fe/NCNT at 60 h are shown in Fig. 5a. Fe/NCNT and
Mn/Fe/NCNT catalysts show high values at Nc = 1, low values at
Nc = 2 and straight lines with no further changes of slopes at high
carbon numbers, referred to high CH4 selectivity and at Nc = 2,
explained by C2 dissociation to CH4 [20] or ethene secondary
Please cite this article in press as: P. Kangvansura, et al., Product distribution of CO2 hydrogenation by K- and Mn-promoted Fe catalysts
supported on N-functionalized carbon nanotubes, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.02.045

G Model
CATTOD-10078; No. of Pages7
ARTICLE IN PRESS
P. Kangvansura et al. / Catalysis Today xxx (2016) xxx–xxx
5
Fig. 5. a) Log (mole of n-alkane) from CO2 hydrogenation and b) Growth probabil-
ity, pg over Fe/NCNT, K/Fe/NCNT, Mn/Fe/NCNT and K/Mn/Fe/NCNT as a function of
carbon number (Nc) at 60 h.
Scheme 1. Ethane formation from ethene re-adsorption.
catalysts [2]. The pg-value in the range of 0.6 was independent of
carbon number and higher than those resulting from Fe/NCNT and
Mn/Fe/NCNT. The pg-values at Nc = 2 resulting from the potassium-
promoted catalyst were ca. 0.1 indicating low selectivity for ethane
due to ethanol formation. The products of the potassium-promoted
catalysts corresponded to those reported by Riedel et al. [2], such
that at the steady state with H2/CO2 and H2/CO, pg at Nc = 1 is even
higher than pg at Nc = 2. It indicated that methane forms a little,
meaning that methane desorption (CH3 + H = CH4) is slow even if it
has a high tendency for methane formation according to thermo-
dynamics. Riedel et al. [2] reported that FTS with a H2/CO2 syngas
is feasible only on the alkaline-promoted iron catalysts.
Fig. 4. Product selectivities of n-alkanes, alkenes and alcohols over a) Fe/NCNT, b)
K/Fe/NCNT, c) Mn/Fe/NCNT and d) K/Mn/Fe/NCNT as a function of carbon number
(Nc) at 60 h.
reaction [2]. The n-alkane distribution of each catalyst at 60 h indi-
cated the production of high n-alkanes.
FTS polymerization can be characterized by the growth proba-
bility, pg, [21]. It is shown as a horizontal line in the diagram of pg
over carbon numbers in the case of ideal polymerization behaviour
[20,21,26]. Fig. 5b shows the growth probabilities of n-alkanes from
FTS over Fe/NCNT, K/Fe/NCNT, Mn/Fe/NCNT and K/Mn/Fe/NCNT at
60 h [2,27]. The growth probabilities of the products from four cat-
alysts can be discussed in two groups, Fe/NCNT without and with
potassium promoted catalysts. A low pg-value at Nc = 1 for Fe/NCNT
and Mn/Fe/NCNT referred to high CH4-selectivity and a high pg-
value at Nc = 2, explained by ethene re-adsorption and proceeded
secondary reactions producing ethane in Scheme 1 [2].
The pg-values at Nc = 3 resulting from both Fe/NCNT and
Mn/Fe/NCNT catalysts were ca. 0.2 and increased to 0.4 at Nc = 5,
due to 1-alkene re-adsorption forming n-alkanes. The growth
probability of polymerization on potassium-promoted catalysts
(K/Fe/NCNT and K/Mn/Fe/NCNT) was almost constant at 60 h
indicating the steady state of FTS over the potassium-promoted
3.2.3. Product distributions of alkene content and the growth site
of FT synthesis
Product distributions in terms of alkene content in mol% as a
function of carbon number over all catalysts at 60 h are shown
in Fig. 6. With the potassium-promoted catalysts, the alkene con-
tent was high at Nc = 3 and decreased to 10–2 mol% (for Nc = 4 − 7),
which corresponded to the primary selectivity [2]. With the
non-potassium promoted catalysts, deviation was consistent for
secondary reactions of 1-alkenes, i.e. double bond hydrogenation
and double bond shift (Scheme 2) [2]. With increasing carbon num-
ber, the hydrogenation of double bonds increases, mainly due to the
carbon number increasing the reactor residence time. Alkene sec-
ondary reactions are mostly observed when FT catalyst formation
is slow especially with the H2/CO2 syngas.
The growth site of FT synthesis can be indicated by a 1-alkene
[28] such as 1-butene desorbs from an FT intermediate on an
FT growth site [17]. The mol% of 1-butene produced from each
Please cite this article in press as: P. Kangvansura, et al., Product distribution of CO2 hydrogenation by K- and Mn-promoted Fe catalysts
supported on N-functionalized carbon nanotubes, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.02.045

G Model
CATTOD-10078; No. of Pages7
ARTICLE IN PRESS
6
P. Kangvansura et al. / Catalysis Today xxx (2016) xxx–xxx
Fig. 8. Mol% of alcohols resulted from CO2 hydrogenation over Fe/NCNT, K/Fe/NCNT,
Mn/Fe/NCNT and K/Mn/Fe/NCNT as a function of carbon number (Nc) at 60 h.
Fig. 6. Mol% of 1-alkene over Fe/NCNT, K/Fe/NCNT, Mn/Fe/NCNT and K/Mn/Fe/NCNT
as a function of carbon number (Nc) at 60 h.
Scheme 2. Alkene secondary reaction.
Scheme 3. Alcohol formation from alkylidene.
3.3. Catalyst performance
CO2 conversions over Fe/NCNT, K/Fe/NCNT, Mn/Fe/NCNT and
K/Mn/Fe/NCNT at 60 h were around 30–35%, with K/Mn/Fe/NCNT
giving the best results in terms of high conversion and constant per-
formance during 60 h reaction time. The first step is a reverse water
gas shift (Eq. (1)) and the second step is the FTS (Eq. (2)), which
produced mainly FT products, described in alkanes and alkenes
selectivities distributed over carbon numbers. From the XANES
results, potassium-promoted Fe/NCNT was reduced to 14.5% Fe
after holding at 653 K for 300 min (the FTS condition), so one can
assume that the potassium-promoted Fe/NCNT can provide active
iron for carbide formation available for FTS.
Fig. 7. Mol% of 1-butene as a function of time resulted from CO2 hydrogenation over
Fe/NCNT, K/Fe/NCNT, Mn/Fe/NCNT and K/Mn/Fe/NCNT.
Catalyst performances is divided into two groups; products
resulting from Group
I (iron catalyst without potassium-
promotion: Fe/NCNT and Mn/Fe/NCNT) and those resulting from
Group II (iron catalysts with potassium-promotion: K/Fe/NCNT and
K/Mn/Fe/NCNT). Group I catalysts produced mainly CH4, with small
amounts of alkenes at high carbon number. In a confined environ-
ment with high local H2:CO, hydrogenolysis of alkenes could occur,
due to the hydrogenation site. One assumed that FeO on the catalyst
acts as an alkene adsorption site [20]. With the XANES reduction
of Mn/Fe/NCNT, FeO was the main iron species, along the period of
holding time, thus it can adsorb alkenes producing small alkanes
as final products.
catalyst as a function of time is shown in Fig. 7. 1-Butene result-
ing from K/Fe/NCNT and K/Mn/Fe/NCNT was constant at 10 mol%
during 40–60 h, meaning that the FT growth site was constant at
the steady state. Fe/NCNT and Mn/Fe/NCNT produced negligible
amount of 1-butene. The catalysts with high percentages of iron
and manganese promoted iron produced mainly CH4 from the early
stage and retained the sites until it reached the steady state. In this
case the FTS steady state occurs in a longer reaction time.
The product distributions for the Group II catalysts included high
alkenes from C3-C6. From the growth site of FT synthesis, indicated
by 1-butene desorption from an FT intermediate on an FT growth
site, 1-butene resulting from K/Fe/NCNT and K/Mn/Fe/NCNT was
constant during 40–60 h (Fig. 4), meaning that the FT growth site
was constant at the steady state. The catalyst with the potassium
promoter (K/Fe/NCNT) was reduced to Fe and FeO as the active
sites for RWGS and FT reactions as confirmed by XANES. Potassium
weakens the C O bond in the promoted Fe catalyst [5], producing
CHx species for chain formation. Besides, H2O, the product from the
reverse water gas shift in the first step, can easily adsorb onto the
potassium-promoted catalysts and add to an alkene product such
as ethene to produce ethanol as by product. However, manganese
also plays an important role as a structural promoter inhibiting
3.2.4. Alcohol formation
Mol% of alcohols as a function of carbon number, Nc, resulted
from CO2 hydrogenation over Fe/NCNT, K/Fe/NCNT, Mn/Fe/NCNT
and K/Mn/Fe/NCNT at 60 h are shown in Fig. 8. Ethanol was mainly
produced with the potassium-promoted catalysts. However, the
other two catalysts also produced traces of methanol and propanol.
The formation of oxygenated hydrocarbons (i.e. alcohols)
has been explained as follows: CO dissociates on iron, follow-
ing by adsorption, COad = Cad + Oad, forming adsorbed oxygen,
which reacts with adsorbed hydrogen, Oad + Had = OHad, forming
chemisorbed OH, available on the surface for the reaction with
hydrocarbon species such as alkylidenes (R-CH2-CH=) as shown in
Scheme 3 [2];
Please cite this article in press as: P. Kangvansura, et al., Product distribution of CO2 hydrogenation by K- and Mn-promoted Fe catalysts
supported on N-functionalized carbon nanotubes, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.02.045

G Model
CATTOD-10078; No. of Pages7
ARTICLE IN PRESS
P. Kangvansura et al. / Catalysis Today xxx (2016) xxx–xxx
7
the iron active site on NCNT from agglomeration during the high
temperature reduction as examined by XANES.
References
[1] L.M. Chew, P. Kangvansura, H. Ruland, H.J. Schulte, C. Somsen, W. Xia, G.
Eggeler, A. Worayingyong, M. Muhler, Appl. Catal. A: Gen. 482 (2014)
163–170.
4. Conclusions
[2] T. Riedel, H. Schulz, G. Schaub, K.W. Jun, J.S. Hwang, K.W. Lee, Top. Catal. 26
(2003) 41–54.
[3] W. Chen, Z. Fan, X. Pan, X. Bao, J. Am. Chem. Soc. 130 (2008) 9414–9419.
[4] H. Xiong, L.L. Jewell, N.J. Coville, ACS Catal. 5 (2015) 2640–2658.
[5] H. Xiong, M. Moyo, M.A.M. Motchelaho, L.L. Jewell, N.J. Coville, Appl. Catal. A:
Gen. 388 (2010) 168–178.
[6] M. Menon, A.N. Andriotis, G.E. Froudakis, Chem. phys. Lett. 320 (2000)
425–434.
[7] M.C. Bahome, L.L. Jewell, D. Hildebrandt, D. Glasser, N.J. Coville, Appl. Catal. A:
Gen. 287 (2005) 60–67.
[8] Z. Lia, R. Liua, Y. Xua, X. Ma, Appl. Surf. Sci. 347 (2015) 643–650.
[9] M. Fujiwara, T. Satake, K. Shiokawa, H. Sakurai, Appl. Catal. B 179 (2015)
37–43.
[10] Y. Yang, H. Xiang, Y. Xu, L. Bai, Y. Li, Appl. Catal. A: Gen. 266 (2004) 181–194.
[11] M. Jiang, N. Koizumi, M. Yamada, Appl. Catal. A: Gen. 204 (2000) 49–58.
[12] R.A. Dictor, A.T. Bell, J. Catal. 97 (1986) 121–136.
[13] H. Xiong, M.A. Motchelaho, M. Moyo, L.L. Jewell, N.J. Coville, Fuel 150 (2015)
687–696.
[14] M. Dad, H.O.A. Fredriksson, J. van de Loosdrecht, P.C. Thuene, J.W.
Niemantsverdriet, Catal. Struct. React. 1 (2015) 101–109.
[15] M. Al-Dossary, A.A. Ismail, J.L.G. Fierro, H. Bouzid, S.A. Al-Sayar, Appl. Catal. B.
165 (2015) 651–660.
Product distributions and growth probability of n-alkanes dur-
ing CO2 hydrogenation indicated that the alkaline-promoted iron
catalysts (K/Fe/NCNT, K/Mn/Fe/NCNT) performed Fischer-Tropsch
(FT) steady state around 60 h. The time-resolved XANES reduction
showed mixed phases of Fe, FeO, Fe3O4, and Fe2O3 resulting from
K/Fe/NCNT and FeO and Fe3O4 resulting from Mn/Fe/NCNT. The
potassium-promoted Fe/NCNT can provide active iron for carbide
formation available for FTS. The catalyst performance is divided
into two groups: Group I (the potassium unpromoted catalysts:
Fe/NCNT and Mn/Fe/NCNT) and Group II (the potassium promoted
catalysts: K/Fe/NCNT and K/Mn/Fe/NCNT). Group I catalysts pro-
duced mainly CH4, small amount of alkenes with high carbon
number. The product distributions of Group II included high alkenes
from C3-C6. Potassium promotion on the Fe/NCNT catalyst for CO2
hydrogenation could be assumed to provide active sites for alkene
and alcohol formation. The reduced surrounding of NCNT support,
the potassium electronic promoter together with the manganese
structural promoter maintained the iron phases as the active cata-
lyst well suitable for CO2 hydrogenation producing mainly alkenes
and ethanol.
[16] D. Wang, X. Zhou, J. Ji, X. Duan, G. Qian, X. Zhou, D. Chen, W. Yuan, J. Mater.
Chem. A 3 (2015) 4560–4567.
[17] P. Kangvansura, H. Schulz, A. Suramitr, Y. Poo-arporn, P. Viravathana, A.
Worayingyong, Mater. Sci. Eng. B: Adv. 190 (2014) 82–89.
[18] A. Niltharach, S. Kityakarn, A. Worayingyong, J. T-Thienprasert, W. Klysubun,
P. Songsiriritthigul, S. Limpijumnong, Phys. B: Condens. Matter. 407 (2012)
2915–2918.
[19] B.H. Davis, Catal. Today 141 (2009) 25–33.
[20] H. Schulz, M. Claeys, Appl. Catal. A: Gen. 186 (1999) 91–107.
[21] H. Schulz, Catal. Today 214 (2013) 140–151.
Acknowledgments
[22] P. Kangvansura, Ph.D. Thesis, Kasetsart University, Bangkok, 2014.
[23] W. Xia, C. Jin, S. Kundu, M. Muhler, Carbon 47 (2009) 919–922.
[24] L.A. Boot, A.J. van Dillen, J.W. Geus, F.R. van Buren, J. Catal. 163 (1996)
186–194.
This work was supported by the Synchrotron Light Research
Institute (Public Organization), Thailand (grant no. GS-54-D01)
and was performed under the project “Sustainable Chemical
[25] L.M. Chew, H. Ruland, H.J. Schulte, W. Xia, M. Muhler, J. Chem. Sci. 126 (2014)
481–486.
[26] H. Schulz, T. Riedel, G. Schaub, Top. Catal. 32 (2005) 117–124.
[27] H. Schulz, Z. Nie, F. Ousmanov, Catal. Today 71 (2002) 351–360.
[28] H. Schulz, M. Claeys, Appl. Catal. A: Gen. 186 (1999) 71–90.
Synthesis (SusChemSys), which is co-financed by the European
¨
Regional Development Fund (ERDF) and the state of North Rhine-
Westphalia, Germany, under the Operational Programme ‘Regional
Competitiveness and Employment’ 2007–2013.
Please cite this article in press as: P. Kangvansura, et al., Product distribution of CO2 hydrogenation by K- and Mn-promoted Fe catalysts
supported on N-functionalized carbon nanotubes, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.02.045