Size and promoter effects on iron nanoparticles confined in carbon nanotubes and their catalytic performance in light olefin synthesis from syngas

Article abstract of DOI:10.1016/j.cattod.2019.05.054

Light olefins are important products of high temperature Fischer-Tropsch synthesis over iron catalysts. In this paper, we found that the catalytic performance of iron catalysts was strongly affected by iron nanoparticle size, their encapsulation inside of carbon nanotubes and promotion with bismuth or lead. The presence of promoters and iron nanoparticle confinement leads to a major increase in the reaction rate. A gradual increase in the TOF numbers with the increase in the iron nanoparticle sizes from 2.5 to 12 nm was observed at both 1 and 10 bar over the carbon nanotubes containing encapsulated monometallic or Bi- or Pb-promoted iron nanoparticles. The size of monometallic iron nanoparticles encapsulated in carbon nanotubes does not show any noticeable effect on the light olefin selectivity, while in the Bi- and Pb-promoted catalysts, the light olefin selectivity was higher over smaller encapsulated iron nanoparticles and decreased with the increase in the nanoparticle size. The stability of iron nanoparticles versus sintering was also improved by the encapsulation.

Full text of DOI:10.1016/j.cattod.2019.05.054

Accepted Manuscript
Title: Size and Promoter Effects on Iron Nanoparticles
Confined in Carbon Nanotubes and their Catalytic
Performance in Light Olefin Synthesis from Syngas
Authors: Bang Gu, Cheng Zhou, Shun He, Simona Moldovan,
Petr A. Chernavskii, Vitaly V. Ordomsky, Andrei Y. Khodakov
PII:
S0920-5861(18)31394-4
DOI:
Reference:
https://doi.org/10.1016/j.cattod.2019.05.054
CATTOD 12230
To appear in:
Catalysis Today
Received date:
Revised date:
Accepted date:
5 November 2018
16 March 2019
21 May 2019
Please cite this article as: Gu B, Zhou C, He S, Moldovan S, Chernavskii PA, Ordomsky
VV, Khodakov AY, Size and Promoter Effects on Iron Nanoparticles Confined in Carbon
Nanotubes and their Catalytic Performance in Light Olefin Synthesis from Syngas,
Catalysis Today (2019), https://doi.org/10.1016/j.cattod.2019.05.054
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Revision, March 16th 2019
Size and Promoter Effects on Iron Nanoparticles Confined in Carbon
Nanotubes and their Catalytic Performance in Light Olefin Synthesis
from Syngas
Bang Gua, Cheng Zhoub, Shun Heb, Simona Moldovanc, Petr A. Chernavskiid,
Vitaly V. Ordomsky a,* and Andrei Y. Khodakova,*
aUniv. Lille, CNRS, Centrale Lille, ENSCL, Univ. Artois, UMR 8181 – UCCS – Unité
de Catalyse et Chimie du Solide, F-59000 Lille, France.
bCollaborative Innovation Center of Chemistry for Energy Materials, State Key
Laboratory of Physical Chemistry of Solid Surfaces, Xiamen University, 361000
Xiamen, China.
c Groupe de Physique des Matériaux, CNRS, Université Normandie & INSA Rouen Avenue
de l'Université - BP12, 76801 St Etienne du Rouvray.
d Department of Chemistry, Moscow State University, 119992 Moscow, Russia.
*Corresponding authors:
vitaly.ordomsky@univ-lille.fr and andrei.khodakov@univ-lille.fr
1

Graphical_abstract
Iron nanoparticle size/ nm
Highlights
 -Higher TOF numbers over larger iron nanoparticles encapsulated in CNT
 -No effect of iron nanoparticle size on olefin selectivity in unpromoted
catalysts
 -Higher olefin selectivity over smaller iron nanoparticles in the promoted
catalysts
 -Enhancement of the stability of iron nanoparticles encapsulated in CNT
2

Abstract
Light olefins are important products of high temperature Fischer-Tropsch synthesis
over iron catalysts. In this paper, we found that the catalytic performance of iron
catalysts was strongly affected by iron nanoparticle size, their encapsulation inside of
carbon nanotubes and promotion with bismuth or lead. The presence of promoters and
iron nanoparticle confinement leads to a major increase in the reaction rate. A gradual
increase in the TOF numbers with the increase in the iron nanoparticle sizes from 2.5
to 12 nm was observed at both 1 and 10 bar over the carbon nanotubes containing
encapsulated monometallic or Bi- or Pb-promoted iron nanoparticles. The size of
monometallic iron nanoparticles encapsulated in carbon nanotubes does not show any
noticeable effect on the light olefin selectivity, while in the Bi- and Pb-promoted
catalysts, the light olefin selectivity was higher over smaller encapsulated iron
nanoparticles and decreased with the increase in the nanoparticle size. The stability of
iron nanoparticles versus sintering was also improved by the encapsulation.
Keywords: Fischer-Tropsch; Particles size; Confinement; Lower olefins; Iron;
Promoter
3

1. Introduction
Light olefins (ethylene, propylene and butene) are key building-block chemicals, which
are widely used in the chemical industry[1-3]. Currently, light olefins are primarily
produced via stream cracking of naphtha or methanol-to-olefins (MTO) process, while
ethylene is also manufactured from ethane cracking [4]. The growing demand for light
olefins and depletion of oil reserves have stimulated development of processes for light
olefin synthesis from syngas, which might be produced from alternative feedstocks
such as methane, biomass, coal or organic and plastic wastes. Fischer-Tropsch (FT)
synthesis produces hydrocarbons from syngas. High temperature FT synthesis (300-
350 °C) is mainly used for production of gasoline and light olefins [5, 6]. FT synthesis
is one of the most efficient ways to convert syngas directly into light olefins [7-9]. In
the olefin synthesis from syngas, Fe-based catalysts can be the catalysts of choice,
because of their low cost, high selectivity towards olefins and flexibility relative to the
use of syngas with different H2/CO ratios [4, 10-12].
FT synthesis is often considered as a structure sensitive reaction and particle size may
have noticeable effects both on the activity and selectivity [13-16]. The particle size
effects over cobalt [17-20] and ruthenium-based [21-23] catalysts in low temperature
FT synthesis have been addressed in numerous publications. The optimum cobalt and
ruthenium size was found to be about 6-8 nm. The increase in the size of metal
nanoparticles above 8 nm leads to lower overall activity, while the metal particles with
4

the sizes smaller than 6 nm demonstrate both lower intrinsic activity (Turnover
Frequency, TOF) and lower selectivity toward the C5+ hydrocarbons. The particle size
effects over iron-based FT catalysts can be rather complex, because of simultaneous
presence of different iron phases and catalyst deactivation.
Previous reports suggest that the selectivities to methane and lighter hydrocarbons
increase, when iron nanoparticles in supported catalysts are getting smaller than 6–9
nm [24]. TOF increases with the increase in the particle sizes from 2.4 to 6.2 nm and it
remains almost constant up to the particle size of 11.5 nm. The stability of iron particles
in silica-supported catalysts was previously studied [25, 26] in the nanoscale range (9–
20 nm). On the one hand, smaller iron-carbide nanoparticles were found more difficult
to obtain by carbidisation of iron oxides and on the other hand, they were less stable
than larger particles.
The iron particle size effects on the catalytic performance in FT synthesis might be
influenced by catalyst promotion and metal-support interactions. In the oxide-supported
catalysts, the active metal can react with a support and form barely reducible aluminates
or silicates and thus, hinder formation and reduce the stability of active iron carbide
phase. The metal support interactions are usually weaker in the carbon-supported
catalysts. De Jong [27, 28] et al examined the influence of iron carbide particle size in
the promoted and unpromoted carbon nanofiber supported iron catalysts on the syngas
conversion to light olefins. The TOF numbers of unpromoted catalysts increased 6-8
times with particle size decreasing from 7 to 2 nm, while the selectivities to methane or
light olefins were found independent on the particle size [27]. The iron catalysts
5

promoted with sodium and sulphur showed similar variation of TOF as a function of
particle size. The increase in TOF with decrease in iron particle size in the promoted
catalysts was principally due to higher yield of methane over the smallest particles.
Recently, we have discovered strong promoting effects of Bi and Pb on the catalytic
performance of Fe catalysts in FT synthesis [29, 30]. The effects have been attributed
to the enhanced C-O dissociation on the surface of iron carbide nanoparticles in the
presence of Bi and Pb, which significantly increases catalyst activity and selectivity to
light olefins.
A number of previous works suggest important influence of iron nanoparticles
confinement in carbon nanotubes (CNT) on the electronic structure, catalytic
performance and stability of iron FT catalysts [31-33]. Bao [34] et al found that iron
nanoparticles confined in CNT exhibited better reducibility, carbidisation and much
higher activity in FT synthesis compared to the unconfined counterparts, while the
selectivity to light olefins was not much affected by the confinement [35, 36]. Higher
reaction rate was principally attributed to higher extent of iron carbidisation inside
CNTs. The encapsulation of nanoparticles also protects them from sintering [37].
Analysis of the electron distribution in CNT using DFT [38] showed that the curvature
of the CNT induced a significant electron disparity. More electrons were distributed on
the exterior surface of the CNT. This led to stronger bonding of iron species with the
outside CNT surface than those located the inside CNT walls. These electronic effects
might result in different catalytic activity of iron inside and outside CNT.
6

The present work addresses the size effects of iron nanoparticles encapsulated inside of
CNT on syngas conversion to light olefins. To the best of our knowledge, there have
been no studies about the effect of the size of iron nanoparticles confined inside CNT
on their properties and catalytic performance. Both monometallic iron catalysts and
catalysts promoted with bismuth and lead have been studied. The syngas conversion to
lower olefins has been performed at both high and atmospheric pressures at the
temperatures typical of high temperature FT synthesis.
2. Experimental
2.1 CNT pre-treatment
Prior to the deposition of iron and promoters, the CNTs were pre-treated according
to previous literature [34]. CNTs (Iolitec nanomaterial, 95%, inter diameter 5-12 nm,
outer diameter 10-20 nm) were exposed to boiling concentrated nitric acid in order to
open the tube channels, to remove contaminations and to make CNT hydrophilic.
Typically, 3.0 g of CNT were pre-treated in concentrated HNO3 (68%, 210 mL) during
14 h under reflux at 140 oC. Then the samples were filtered, washed with distilled water
until pH = 7 and dried at 100 ºC overnight. CNT with closed tube channels were pre-
treated under milder condition (34 wt. % HNO3 for 6h at 110oC).
2.2 Catalyst preparation
The encapsulated iron catalysts were prepared by incipient wetness impregnation of
CNT support with aqueous solutions of iron nitrate (Fe(NO3)3·9H2O, Sigma-Aldrich).
The amount of iron in the impregnating solutions was calculated in order to obtain 10
7

wt. % iron in the final catalysts. Lead nitrate (Pb(NO3)2, Sigma-Aldrich) and bismuth
nitrate (Bi(NO3)3·5H2O, Sigma-Aldrich) were used for preparation of the Bi- and Pb-
promoted iron catalysts by co-impregnation. The aqueous solutions were drawn into
the CNT channels by capillary forces assisted by ultrasonic treatment and stirring. The
molar ratios of Fe/Pb and Fe/Bi were 100/2. After the impregnation, the samples were
dried in an oven at 80 oC for 12 h followed by thermal treatment for 4 h under a flow
of nitrogen (50 mL/min). The temperature of thermal treatment was varied from 300 to
600 °C in order to obtain different iron particle size. The calcined catalysts are labelled
as FeM/CNT-in-T, where “M” represents the promoters (Bi or Pb), “in” represents iron
inside the CNT tubes and “T” stands for thermal treatment temperature. For
comparison, the Fe/AC and Fe/CNT-out catalysts, where “out” indicates iron particles
located outside the tubes, were prepared by the same procedure. The temperature of
thermal treatment in the N2 flow for those 2 samples is 400 oC; the duration is 4h. The
iron content in these two catalysts is 10 wt.%.
2.3. Catalyst characterization
Low temperature N2 physisorption measurements were carried out using a
Micromeritics Tristar 3020 Surface Area and Porosimetry analyzer. The samples were
degassed at 200 oC for 2 h to remove water before N2 adsorption. The N2 adsorption-
desorption isotherms were measured at -196°C.
The X-ray diffraction patterns (XRD) were recorded using a Siemens D5000
diffractometer. The measurement conditions were: Cu(Kα) (λ = 0.15418 nm), tube
8

pressure 40 kv, tube current 30 mA, and scanning rate 0.02° s-1. The phase composition
analysis of the samples was performed by comparing the XRD Standards Database of
the Joint Committee on Powder Diffraction Standards (JCPDS). The average crystallite
size of iron oxides (Fe2O3) was evaluated by the Scherrer equation [39] using the
diffraction peak of Fe2O3 at 2θ of 30.2o.
Quantitative elemental analyses were performed by inductively coupled plasma-
optic emission spectroscopy 720-ES ICP-OES (Agilent) with axially viewing and
simultaneous CCD detection. The quantitative determination of metal content in the
catalysts was based on the analysis of standard solutions. The ICP ExpertTM software
(version 2.0.4) provided metal concentration in the samples. The minimum detection
limitation is 0.1 ppm and the accuracy is better than 5%.
The H2 temperature-programmed reduction (H2-TPR) experiments were carried out
by the AutoChem II 2920 apparatus (Micromeritics) using 0.10 g of the sample in a
flow of H2/Ar (5 vol. % H2) stream (50 ml/min). The temperature was increased from
room temperature to 900 °C at the rate of 10 °C/min.
The TEM (Transmission Electron Microscopy) analyses were carried out on a
Phillips FEI G2 30 high resolution transmission electron microscope with an
accelerating voltage of 300 KV. The sample was prepared by dispersing in an ethanol
solution. After 1 h of ultrasonication, the sample slurry was dropped onto a 400-mesh
carbon film copper mesh, and dried under an infrared lamp to test. High angle annular
dark field (HAADF)-scanning transmission electron microscopy (STEM) imaging, and
energy dispersive X-ray spectroscopy (EDX) of the calcined analysis were performed
9

on a Jeol-ARM200 transmission electron microscope, operated at 200 kV and scanning
speed 20 μs/px for imaging and 0.05 μs/px for EDX. More than 200 iron particles were
counted to estimate the iron particles size distribution and average iron particle size.
In-situ magnetic characterization has been performed using a Föner vibrating-sample
magnetometer. 1 mg of pure metallic Fe was used for the calibration before each
experiment. 10 mg of the fresh catalysts was heated to 200 °C with the ramping rate
6.6 °C/min in the flow of CO or syngas (15 ml/min) and then sequentially heated to
350 °C (4.7 °C/min) for 2 h in CO or syngas. After activation, the sample was cooled
to RT in the atmosphere of CO or syngas with the saturation magnetization curve
recorded by the magnetometer.
2.4. Catalytic tests
The catalytic tests in FT synthesis were performed in a fixed-bed reactor (8 mm
inner diameter). Typically, 0.1 g of fresh catalyst with the particle size of 100-150 μm
was loaded in the reactor and then reduced in the CO gas flow (50 ml/min, atmospheric
o
pressure) at 350 oC with a heating rate of 2 C/min for 10 h before the reaction. After
the reactor was cooled down to 180 oC, a syngas with the H2/CO ratio of 1/1 and under
pressure of 10 or 1 bar was introduced into the reactor. Nitrogen with a fixed flow rate
of 1 ml/min in the syngas was used as an internal standard for the calculation of CO
conversion. After the pressure and gas flow rate stable, the temperature was raised with
the ramping rate of 1 oC/min to the desired reaction temperature (350 oC). The absence
of mass transport resistances was checked by Weisz-Prater Criterion (NW-P) for internal
diffusion [40-42].
10

2
푅푟
푝
푁푊−푃
=
≤ 0.3
퐷푒푓푓퐶푠
where R = observed reaction rate (molCO g-1 s-1), 푟 = catalyst particle radius (m), 퐷푒푓푓
푝
= effective diffusivity (cm2 s-1), Cs = gas concentration of CO at the external surface of
the catalyst (mol cm-3). We considered a total gas flow rate of 57 ml min-1 with ~ 60 %
CO conversion and a catalyst particle size of 0.10 - 0.15 mm. The pressure was 10 bar
and the reaction rate was calculated from the measured FTY of 6.1*10-4 molCO gFe-1s-1.
Cs and 퐷푒푓푓 are calculated according to the literature [43] with the values of 4.9*10-5
and 0.0229 cm2 s-1, respectively. The Nw-p value is calculated to be 2 ×10-3, which is
much lower than 0.3.
The reaction feed gas and gaseous products were analysed online using a gas
chromatograph (Bruker GC-450), which was equipped with a thermal conductivity
(TCD) and a flame ionization (FID) detectors. The packed CTR-1 column is connected
to the TCD detector for separation of N2, CO, CO2, and CH4. The Rt-Q PLOT capillary
column is connected to the FID detector for analysis of hydrocarbons and oxygenates.
The liquid products were collected in a cold trap and analysed by gas chromatography.
The selectivities were calculated on carbon basis. The CO2 free hydrocarbon
selectivities were calculated considering only hydrocarbon production in FT synthesis.
The carbon balance was better than 90%.
The iron time yields (FTY) were expressed as moles of CO converted per gram of
total iron per second. Apparent turnover frequency (TOF) was calculated according
with previous literature [16, 27] using the bulk density of Fe5C2 (ρ = 7.57 g mL−1) and
assuming the surface density of 14 Fe atoms nm−2.
11

3. Results and discussion
3.1 Catalyst structure
Tables 1 and 2 show textural properties and elemental composition of the CNTs
supported monometallic and Bi- or Pb-promoted iron nanoparticles exposed to different
temperatures. The impregnation of CNTs with iron decreases the surface area, total pore
volume and mesopore diameter. Interestingly, the pore volume is decreased after
introduction of iron from 0.83 cm3/g to about 0.50 cm3/g. This suggests that some iron
particles are located inside of the CNT tubes. Localization of iron nanoparticles inside
CNT leads to the decrease in the pore volume and partial blocking of the tubes. The
presence of the Bi and Pb promoters, in addition to iron, results in only very small effect
on the texture of CNT catalysts. Indeed, the BET surface area and pore volume are
almost not affected by the presence of the promoters.
The elemental analysis (Tables 1 and 2) shows that all monometallic iron catalysts
supported by carbon nanotube treated in nitrogen at different temperatures contain
around 10 wt.% iron. Iron content in the promoted catalysts was also close to10 wt. %,
while the contents of bismuth and lead were respectively ~0.85 wt.% and ~0.90 wt.%.
Figure 1a displays XRD profiles of the CNT containing confined monometallic and
promoted iron nanoparticles treated in nitrogen at different temperatures. The peaks at
26.3° and 43.8° are attributed to the (002) and (101) reflections of the CNT supports
[44]. The diffraction lines at 2θ of 30.2o and 35.6° can be assigned to the hematite phase
(Fe2O3, JCPDS13-0534), while the diffraction lines at 2θ of 35.8o,43.5° and 53.9° can
12

be ascribed to the magnetite phase (Fe3O4, JCPDS 75-0449). The iron oxide peaks
become sharper with higher pre-treatment temperature probably due to increase in the
iron oxide particle sizes. The iron crystallite sizes have been evaluated from the
Scherrer equation (Table 1). The iron crystallite size increases from 2.6-2.8 nm for the
iron catalysts treated at 300 °C to 11-12 nm for the catalysts treated in nitrogen at 600 °C.
Note that the promotion with Bi and Pb does not noticeably affect the iron oxide particle
size (Table 2). Our previous studies [29, 30] also showed insignificant effect of the
bismuth and lead promotions on iron dispersion and iron oxide nanoparticle sizes in the
SiO2 and CNT supported iron catalysts. The confined iron catalysts contain smaller
particles compared with the Fe/AC iron catalysts supported on active carbon (Figure
S1 and Table S1, SI). The difference is more pronounced, when the Fe/AC catalysts
were calcined at higher temperatures. It is interesting to note that maximum size of
encapsulated metal oxide nanoparticles corresponds to the internal diameter of CNT
(14.4 nm). This is consistent with localization of iron oxide nanoparticles inside the
CNT and suggests that the growth of iron oxide crystallites inside the CNT could be
limited by the CNT diameter.
The TEM images and particle size distribution of the encapsulated monometallic
iron catalysts with different calcination temperatures are shown in Figures 2 and 3. It
is clear that most of the particles are located inside of the nanotubes. This clearly
confirms confinement of iron nanoparticles inside the CNT in the prepared catalysts.
The iron oxides particles size increases from 2.8 nm to 11.4 nm with increase in the
pre-treatment temperature. The iron oxide nanoparticle size measured by TEM is
13

consistent with XRD results. The TEM images and particle size distribution for the Bi-
or Pb-promoted iron catalysts at different calcination temperature are shown in Figures
4, S2 and S3, Supporting Information (SI). Importantly, in the catalysts, iron oxide
nanoparticles are selectively confined inside the CNT. Similar effects of the pre-
treatment temperature on the iron oxide particle sizes were observed with and without
promoters (Table 1 and 2). Thus, the characterisation suggests that the promoter does
not affect the size of metal nanoparticles.
STEM-HAADF and STEM-EDS analysis have been performed in order to further
confirm localization of iron and promoters in CNT (Figure 5). Bi is a heavier metal
compared to iron. In the STEM-HAADF images, the Bi species correspond to brighter
spots compared to iron. STEM-EDX indicates that Bi is uniformly distributed on the
surface of Fe nanoparticles. These results are similar to our previous data obtained for
iron-supported catalysts promoted by Bi and Pb over the SiO2 and CNT supports.
Because of low melting points, metallic bismuth and lead migrate to the iron
nanoparticles and form core-shell structures during catalyst activation and reaction.
These core-shell structures were identified by the EDS analysis [29, 30].
The reducibility of iron oxide nanoparticles encapsulated inside CNT was
investigated by H2-TPR. Figure 6a shows the TPR profiles of the monometallic
Fe/CNT-in catalysts with encapsulated iron oxide nanoparticles of different sizes. The
o
TPR profiles exhibit three main broad peaks located at around 390, 450 and 650 C.
According to the literature [33, 45, 46], the first peak can be assigned to the reduction
of Fe2O3 to Fe3O4, the second peak to the reduction of Fe3O4 to FeO and the third peak
14

to the reduction of FeO to metallic Fe. Note that some hydrogen consumption can due
to the reductive removal of carboxyl and other oxygen-containing groups from the
surface of CNT during the treatment in hydrogen. The small and broad peaks located
at 700-800 oC can be caused by gasification of the CNT support [13]. Because of low
amounts of bismuth and lead, no TPR peaks could be assigned to the reduction of the
promoters. The Fe/CNT-out catalyst containing iron nanoparticles outside the CNT
channels has been used as a reference (Figure 6a). Interestingly, the encapsulated iron
catalysts can be reduced at lower temperatures compared with the Fe/CNT-out catalyst.
This suggests that the confinement increases the iron reducibility in agreement with the
works of Bao’ group [34]. Moreover, as the iron particle size inside the CNT channels
increases, the intensity of reduction peaks decreases and they slightly shift to lower
temperatures. Similar effect also has been discovered by Park [24] over alumina
supported iron catalysts with different crystal sizes. This effect might be explained by
easier reducibility of larger iron oxide nanoparticles. Easier reducibility of larger metal
oxide nanoparticles compared to smaller counterparts has been previously observed in
numerous publications [25, 47, 48].
Introduction of the Bi and Pb promoters to the iron catalysts does not change
significantly the relative position of the TPR peaks but shifts their maximums by
approximately 10 ºC to lower temperatures (Figure 6b, Table 2). This indicates that
promotion of the CNT containing confined iron nanoparticles enhances iron
reducibility. Similar effects were also observed recently for the silica and CNT
supported FeBi and FePb catalysts [29, 30]. Note that previously prepared promoted
15

iron catalysts supported over CNT contained iron nanoparticles only outside the CNT
[30]. Thus, Bi and Pb also promote iron reducibility by increasing total hydrogen
consumption and decreasing the reduction temperature in both confined and non-
confined systems.
Iron carbide has been considered active phase for iron-based Fischer-Tropsch
catalysts. Iron carbide can be formed during activation in the presence of CO prior to
the reaction. The XRD for the used catalysts shows a broad peak, which corresponds to
the mixture of different iron carbide phase (χ-Fe5C2, ε-Fe2.2C, Fe3C, or Fe7C3, Figure
S4, SI). Similar results were obtained for the Fe/CNT and FeBi/CNT catalysts [30]. In
order to identify iron carbide phase, the in-situ magnetic measurements [49, 50] were
conducted in the flow of CO or syngas at different temperatures (Figure S5, SI). The
two examined Fe/CNT-in and FeBi/CNT-in catalysts present a similar shape of
thermomagnetic curves with the Curie temperature of ~ 250oC. According to our
previous report [30], the main iron carbide phase has been identified as χ-Fe5C2 in the
catalysts exposed to the flow of CO or syngas. Interestingly, the iron carbide
concentration was higher in the Bi promoted iron catalyst.
3.2 Catalytic results
The catalytic data are shown in Tables 3 and 4 and in Figures 7-11, S6 and S7, SI.
High temperature FT synthesis yields methane, C2-C4 hydrocarbons, C5+ hydrocarbons,
CO2 and water. The CNT with encapsulated monometallic and Bi-or Pb-promoted iron
nanoparticles with different sizes have been tested in a fix-bed reactor both at high (10
16

o
bar) and atmospheric pressure at T = 350 C and H2/CO = 1/1. For comparison,
Fe/CNT-out and Fe/AC have been also tested under the same conditions (Table 3). The
activity of Fe/CNT-out is 1.5 times higher than for the Fe/AC catalyst. Note however
that the confined catalysts present 2-4 times higher reaction rate compared with the
Fe/AC catalyst.
The selectivity to light olefins and C5+ hydrocarbons follows the sequence: Fe/AC<
Fe/CNT-out< Fe/CNT-in, whereas the methane selectivity shows the opposite trend.
These results are consistent with earlier data of Bao’s group [34]. The effect of
confinement on the chain growth probability is relatively small, while the ratio of
olefins to paraffins increases in the confined catalysts. Stronger interaction of CO with
the CNT internal surface [51, 52] can lead to lower H2/CO ratio inside the CNT and
thus decreases the secondary olefin hydrogenation. Figure 7 displays CO conversion
and product selectivities for the CNT containing encapsulated iron monometallic
nanoparticles as functions of iron particle sizes. At high pressure and at GHSV = 17
L/g.h, carbon monoxide conversion over monometallic iron catalysts increases from 18
to 34 % with the increase in the encapsulated iron particle sizes from 2.8 to 6 nm.
Interestingly, further increase in the iron nanoparticle sizes leads to the decrease in the
CO conversion. Note that the FT reaction rates were much lower at atmospheric
pressure compared to the tests conducted at 10 bar. The selectivities to methane and
light olefins over the iron unpromoted nanoparticles at the 10 bar test were about 28 %
and 37 %, respectively. Interestingly, these two values do not change much with
increasing iron particles size. In order to compare the reaction selectivities, the carbon
17

monoxide conversion was adjusted to the 40-50% level by varying GHSV. Relevant
selectivity data are presented in Figure 8. The increase in the size of iron nanoparticles
encapsulated in CNT does not significantly affect the olefin selectivity and olefin to
paraffin ratio over the unpromoted iron catalysts. The selectivity to the C5+
hydrocarbons increases from 16.1 to 19.9% with the iron particles size increase from
2.8 nm to 6.0 nm and then stabilises. Further particle size increase from 6.0 nm to 11.4
nm does not affect the reaction selectivity [24]. For comparison, iron catalysts
supported over active carbon with different particles size (Fe/AC) has been tested in FT
synthesis (Table S1, SI). Similar trend has been found as for the Fe/CNT-in-T catalysts.
Thus, we found that over the unpromoted, confined or unconfined iron catalysts,
particle size does not noticeable affect the light olefin selectivity. These results are
consistent with the data of De Jong [27] who also observed similar selectivity to
methane and lower olefins is over the iron particles of different sizes in FT reaction
over unpromoted iron catalysts.
Different to the unpromoted iron catalysts, the size of Bi-promoted and encapsulated
iron particles has strong effects both on the activity and selectivity (Figure 8). The
presence of the promoters significantly increases the FT reaction rate. The FeBi/CNT-
in catalyst shows more than 2 times higher FTY in comparison with the Fe/CNT-in
catalyst at the pressure of 10 bar (Figures 9 and 10, S6, SI, Table 4). The reaction
selectivity shifts to the C2-C4 light hydrocarbons compared to the monometallic
Fe/CNT-in catalyst. The methane selectivity decreases from ~29 % to ~ 25 % and C5 +
decreases from ~ 20 % to ~12%. In addition, the selectivity of lower olefins increases
18

from ~ 35 % to ~ 45%. The selectivity to CO2 also increases over the FeBi/CNT-in
catalysts compared with the monometallic Fe/CNT-in counterparts. The effect can be
due to easing CO dissociation on iron carbide via oxygen removal from iron carbide by
scavenging with the promoters [29, 30]. The catalytic activity in the presence of the
promoters increases with increase in iron nanoparticle size until 6 nm and then
decreases at the larger iron nanoparticle sizes. More importantly, this promotion effect
is also significant even under atmospheric pressure with the 2-3 times increase in the
FTY in comparison with the non-promoted catalysts. Compared with the Bi promoted
iron catalysts, the FePb/CNT-in catalysts present even much more significant increase
in FTY (~2.5 times higher than Fe/CNT-in). Different to CNT containing monometallic
iron nanoparticles, the C2-C4 olefin selectivity and olefin to paraffin ratio decrease with
increase in the iron particle size both at 10 bar and under atmospheric pressure (Table
4, Figure 8 and 9). Higher, but slightly different, light olefin selectivities over the
promoted iron catalysts with different particle size might be ascribed to different
coverages of the iron carbide surface with the promoters. De Jong [27] also observed
this effect over Na and S promoted iron catalysts with different iron carbide particle
size.
More detailed information about the intrinsic activity of CNT containing
encapsulated monometallic iron nanoparticles and those promoted with Bi and Pb was
obtained from the TOF values. The TOF values assuming the presence of the Fe5C2
under the reaction conditions are presented in Tables 3 and 4. TOFs are plotted as a
function of iron particle size for CNT containing encapsulated monometallic and
19

promoted iron nanoparticles measured at the 10 bar and atmospheric pressure (Figure
11). As expected, higher TOF were obtained at higher reaction pressure. The TOF
numbers for unpromoted catalysts at both high and low pressure reaction conditions
increase rapidly with increase in the particle sizes up to 6 nm and stabilise for larger
metal nanoparticles. These results are consistent with those for cobalt and ruthenium
catalysts for which TOF increases with the particle size up to 6-8 and stabilize at larger
sizes [20, 23]. Park [24] also found the similar results over iron catalysts with varying
particle size from 2 nm to 12 nm in FTS. It is noteworthy that the Bi and Pb promoted
iron catalysts present ~3 times (10 bar) and ~ 5 times (1 bar) higher TOF values
compared with monometallic iron catalysts.
Apart from the enhancement of FT reaction rate, confinement of iron nanoparticles
in CNT also improves the catalyst stability. Figure S8, SI shows micrographs of two
spent Bi-promoted iron catalysts and calculated respective particle size distribution.
The average particle size of spent FeBi/CNT-300 is 3.9 nm, while for the FeBi/CNT-
in-400 the average size after reaction is 6.5 nm. Both the FeBi/CNT-in-300 and
FeBi/CNT-in-400 catalysts show very slight sintering compared with the fresh catalysts.
We can conclude that confinement is also an efficient way to improve the stability of
small iron carbide nanoparticles.
4. Conclusion
Encapsulation inside carbon nanotubes affects the properties of iron nanoparticles,
their catalytic performance and stability in high temperature FT synthesis. The
20

temperature of catalyst thermal treatment has a strong effect on the size of iron
nanoparticles encapsulated inside CNT. Larger iron nanoparticles were obtained after
treatment at higher temperature and they exhibit better reducibility. The promoters (Bi
and Pb) are located in the surface of iron with close contact and the iron sintering degree
is restricted by confinement in CNT.
The presence of promoters strongly influences on the catalytic performance of iron
carbide nanoparticles. The specific reaction rates (TOF) were 3-5 times higher over the
Bi- and Pb –promoted catalysts compared to the unpormoted counterparts. In the
unpromoted iron catalysts with confined iron nanoparticles, the TOF increases with the
increase in the iron nanoparticle size from 2.5 nm to 6-8 nm and then remains stable
with further increase in the iron particle size, while the product selectivity is not affected
by the nanoparticle size variation.
In the Bi and Pb promoted iron catalysts, the size of encapsulated iron nanoparticles
affects both the activity and selectivity. The activity shows similar trend with the
unpromoted iron catalysts, while the selectivity to light olefins decreases with the
increase in the iron particle size. These findings shed further light on the fundamental
effects of confinement on catalysis, and provide more insights into the particle size
effects in iron-based FT catalysts.
21

Acknowledgements
The authors thank Olivier Gardoll, Laurence Burylo and Joelle Thuriot for help with
TPR, XRD measurements and ICP-OES analysis. B. G. thanks the China Scholarship
Council for providing him a stipend for PhD studies in France. The authors
acknowledge financial support from European Union (Interreg V project PSYCHE).
The authors acknowledge financial support of the French National Research Agency
(Project NANO4-FUT, ANR-16-CE06-0013).
22

References
[1] M. Janardanarao, Industrial & Engineering Chemistry Research 29 (1990) 1735-
1753.
[2] K. Cheng, J. Kang, D.L. King, V. Subramanian, C. Zhou, Q. Zhang, Y. Wang, in:
C. Song (Ed.), Advances in Catalysis, Academic Press, 2017, pp. 125-208.
[3] B. Gu, S. He, W. Zhou, J. Kang, K. Cheng, Q. Zhang, Y. Wang, Journal of
Energy Chemistry 26 (2017) 608-615.
[4] H.M. Torres Galvis, K.P. de Jong, ACS Catalysis 3 (2013) 2130-2149.
[5] A. Steynberg, Studies in surface science and catalysis, Elsevier, 2004, pp. 1-63.
[6] S. Abelló, D. Montané, ChemSusChem 4 (2011) 1538-1556.
[7] M. Casavola, J. Hermannsdörfer, N. de Jonge, A.I. Dugulan, K.P. de Jong,
Advanced Functional Materials 25 (2015) 5309-5319.
[8] H.M.T. Galvis, J.H. Bitter, C.B. Khare, M. Ruitenbeek, A.I. Dugulan, K.P. de
Jong, Science 335 (2012) 835-838.
[9] A.C. Koeken, T. Galvis, M. Hirsa, T. Davidian, M. Ruitenbeek, K.P. De Jong,
Angewandte Chemie International Edition 51 (2012) 7190-7193.
[10]C. Wang, L. Xu, Q. Wang, Journal of Natural Gas Chemistry 12 (2003) 10-16.
[11]Q. Chang, C. Zhang, C. Liu, Y. Wei, A.V. Cheruvathur, A.I. Dugulan, J.W.
Niemantsverdriet, X. Liu, Y. He, M. Qing, L. Zheng, Y. Yun, Y. Yang, Y. Li, ACS
Catalysis (2018) 3304-3316.
[12]M.E. Dry, T. Shingles, L.J. Boshoff, Journal of Catalysis 25 (1972) 99-104.
23

[13]J. Lu, L. Yang, B. Xu, Q. Wu, D. Zhang, S. Yuan, Y. Zhai, X. Wang, Y. Fan, Z.
Hu, ACS Catalysis 4 (2014) 613-621.
[14]A. Nakhaei Pour, M.R. Housaindokht, M. Irani, S.M. Kamali Shahri, Fuel 116
(2014) 787-793.
[15]Z.-J. Wang, S. Skiles, F. Yang, Z. Yan, D.W. Goodman, Catalysis Today 181
(2012) 75-81.
[16]V.K. Jones, L.R. Neubauer, C.H. Bartholomew, Journal of Physical Chemistry 90
(1986) 4832-4839.
[17]J. Den Breejen, P. Radstake, G. Bezemer, J. Bitter, V. Frøseth, A. Holmen, K.d.
Jong, Journal of the American Chemical Society 131 (2009) 7197-7203.
[18]J.-Y. Park, Y.-J. Lee, P.R. Karandikar, K.-W. Jun, K.-S. Ha, H.-G. Park, Applied
Catalysis A: General 411-412 (2012) 15-23.
[19]G. Prieto, A. Martínez, P. Concepción, R. Moreno-Tost, Journal of Catalysis 266
(2009) 129-144.
[20]G.L. Bezemer, J.H. Bitter, H.P. Kuipers, H. Oosterbeek, J.E. Holewijn, X. Xu, F.
Kapteijn, A.J. van Dillen, K.P. de Jong, Journal of the American Chemical Society
128 (2006) 3956-3964.
[21]J.M.G. Carballo, J. Yang, A. Holmen, S. García-Rodríguez, S. Rojas, M. Ojeda,
J.L.G. Fierro, Journal of Catalysis 284 (2011) 102-108.
[22]J.L. Eslava, X. Sun, J. Gascon, F. Kapteijn, I. Rodríguez-Ramos, Catalysis
Science & Technology 7 (2017) 1235-1244.
24

[23]J. Kang, S. Zhang, Q. Zhang, Y. Wang, Angewandte Chemie International
Edition 48 (2009) 2565-2568.
[24]J.-Y. Park, Y.-J. Lee, P.K. Khanna, K.-W. Jun, J.W. Bae, Y.H. Kim, Journal of
Molecular Catalysis A: Chemical 323 (2010) 84-90.
[25]K. Cheng, M. Virginie, V.V. Ordomsky, C. Cordier, P.A. Chernavskii, M.I.
Ivantsov, S. Paul, Y. Wang, A.Y. Khodakov, Journal of Catalysis 328 (2015) 139-
150.
[26]P.A. Chernavskii, V.I. Zaikovskii, G.V. Pankina, A.Y. Khodakov,
ChemCatChem 5 (2013) 1758-1761.
[27]H.M. Torres Galvis, J.H. Bitter, T. Davidian, M. Ruitenbeek, A.I. Dugulan, K.P.
de Jong, Journal of the American Chemical Society 134 (2012) 16207-16215.
[28]J. Xie, H.M. Torres Galvis, A.C. Koeken, A. Kirilin, A.I. Dugulan, M.
Ruitenbeek, K.P. de Jong, ACS catalysis 6 (2016) 4017-4024.
[29]V.V. Ordomsky, Y. Luo, B. Gu, A. Carvalho, P.A. Chernavskii, K. Cheng, A.Y.
Khodakov, ACS Catalysis 7 (2017) 6445-6452.
[30]B. Gu, V.V. Ordomsky, M. Bahri, O. Ersen, P.A. Chernavskii, D. Filimonov,
A.Y. Khodakov, Applied Catalysis B: Environmental 234 (2018) 153-166.
[31]W. Chen, Z. Fan, L. Gu, X. Bao, C. Wang, Chemical communications 46 (2010)
3905-3907.
[32]P. Serp, E. Castillejos, ChemCatChem 2 (2010) 41-47.
[33]G. Yu, B. Sun, Y. Pei, S. Xie, S. Yan, M. Qiao, K. Fan, X. Zhang, B. Zong,
Journal of the American Chemical Society 132 (2009) 935-937.
25

[34]W. Chen, Z. Fan, X. Pan, X. Bao, Journal of the American Chemical Society 130
(2008) 9414-9419.
[35]C. Wang, X. Pan, X. Bao, Chinese Science Bulletin 55 (2010) 1117-1119.
[36]Z. Yang, X. Pan, J. Wang, X. Bao, Catalysis Today 186 (2012) 121-127.
[37]H. Tian, X. Li, L. Zeng, J. Gong, ACS Catalysis 5 (2015) 4959-4977.
[38]L. Yu, W.-X. Li, X. Pan, X. Bao, The Journal of Physical Chemistry C 116
(2012) 16461-16466.
[39]B.D. Cullity, Addision-Wesley Publishing Company, London (1978).
[40]D.E. Mears, Industrial & Engineering Chemistry Process Design and
Development 10 (1971) 541-547.
[41]P. Weisz, C. Prater, Advances in Catalysis, Elsevier, 1954, pp. 143-196.
[42]A. Primo, J. He, B. Jurca, B. Cojocaru, C. Bucur, V.I. Parvulescu, H. Garcia,
Applied Catalysis B: Environmental 245 (2019) 351-359.
[43]Y.-N. Wang, Y.-Y. Xu, H.-W. Xiang, Y.-W. Li, B.-J. Zhang, Industrial &
Engineering Chemistry Research 40 (2001) 4324-4335.
[44]R.M.M. Abbaslou, J. Soltan, A.K. Dalai, Applied Catalysis A: General 379
(2010) 129-134.
[45]K. Mai, T. Elder, L.H. Groom, J.J. Spivey, Catalysis Communications 65 (2015)
76-80.
[46]V. Subramanian, V.V. Ordomsky, B. Legras, K. Cheng, C. Cordier, P.A.
Chernavskii, A.Y. Khodakov, Catalysis Science & Technology 6 (2016) 4953-4961.
26

[47]R.M.M. Abbaslou, A. Tavassoli, J. Soltan, A.K. Dalai, Applied Catalysis A:
General 367 (2009) 47-52.
[48]Z. Yu, D. Chen, B. Tøtdal, A. Holmen, Catalysis today 100 (2005) 261-267.
[49]P.A. Chernavskii, A.Y. Khodakov, G.V. Pankina, J.-S. Girardon, E. Quinet,
Applied Catalysis A: General 306 (2006) 108-119.
[50]P. Chernavskii, J.-A. Dalmon, N. Perov, A. Khodakov, Oil & Gas Science and
Technology-Revue de l'IFP 64 (2009) 25-48.
[51]X. Pan, X. Bao, Accounts of Chemical Research 44 (2011) 553-562.
[52]J. Guan, X. Pan, X. Liu, X. Bao, Journal of Physical Chemistry C 113 (2009)
21687-21692.
27

Table 1. Characterization of the supports and CNT confined monometallic Fe
catalysts.
a
b
c
d
e
Sample
SBET
(m2/g) (cm3/ (nm)
g)
Vtot
Dmeso Dmetal Dmetal Total H2
Fe
conte
ntg
(nm)
(nm)
consum
ptionf
(mmol/g) (wt%)
CNT-raw
138.2
230.3
205.3
198.7
192.9
188.3
193.6
199.9
0.43
0.83
0.53
0.51
0.52
0.50
0.51
0.53
14.0
14.4
10.9
10.6
10.8
11.1
10.8
11.2
CNT-HNO3
-
-
-
-
Fe/CNT-in-300
Fe/CNT-in-350
Fe/CNT-in-400
Fe/CNT-in-450
Fe/CNT-in-500
Fe/CNT-in-600
a BET surface area.
-
2.8
3.9
6.0
7.7
9.9
11.4
2.40
-
10.9
10.7
10.3
11.3
11.1
10.7
3.8
6.0
7.8
11.2
12.7
1.80
-
1.60
1.52
b Single point desorption total pore volume of pores, P/P0=0.975.
c The pore diameter in the mesoporous region evaluated by the BJH method.
d Average particle size of iron oxide by XRD.
e Average particle size of iron oxide by TEM.
f The total H2 consumption and iron reducibility degree from TPR analysis.
g The Fe content from ICP-OES.
28

Table 2. Characterization of CNT with confined iron nanoparticles promoted with Bi
or Pb.
a
b
c
d
e
Sample
SBET
Vtot
Dmeso Dmetal Dmetal Total H2
Fe
conte
ntg
Bi or
Pb
(m2/g) (cm3 (nm)
/g)
(nm)
(nm)
consum
Ptionf
conte
ntg
(mmol/g) (wt%)
(wt%)
CNT
230.3 0.83
14.4
11.0
11.2
11.3
10.4
10.8
11.2
10.7
10.2
10.5
-
-
-
-
FeBi/CNT-in-300 200.3 0.52
FeBi/CNT-in-350 193.7 0.52
FeBi/CNT-in-400 183.3 0.52
FeBi/CNT-in-450 173.3 0.50
FeBi/CNT-in-500 176.6 0.48
FeBi/CNT-in-600 179.9 0.45
FePb/CNT-in-300 208.2 0.51
FePb/CNT-in-400 187.8 0.50
FePb/CNT-in-600 180.1 0.48
aBET surface area.
-
2.6
3.0
10.2
10.8
10.5
11.2
11.3
10.9
10.9
10.6
11.0
0.88
0.84
0.92
0.86
0.83
0.87
0.90
0.88
0.87
3.9
6.2
8.3
10.8
11.9
-
4.1
2.83
2.53
2.51
2.40
2.12
2.70
2.18
1.85
5.8
7.8
10.3
11.7
2.7
5.3
11.5
5.9
11.8
b Single point desorption total pore volume of pores, P/P0=0.975.
c The pore diameter in the mesoporous region evaluated by the BJH method.
d Average particle size of iron oxide by XRD.
e Average particle size of iron oxide by TEM.
29