Tunable Fe3O4 nanoparticles assembled porous microspheres as catalysts for Fischer-Tropsch synthesis to lower olefins

Article abstract of DOI:10.1016/j.cattod.2020.04.017

Fe catalyzed Fischer-Tropsch synthesis to lower olefins (FTO) has been recognized as a structure-sensitive reaction. Nanostructure Fe-based catalysts can expose more active surface. But bulk nanostructure Fe catalysts without support interaction show easily sintered and agglomerated, which makes hard to investigate size effect. Therefore, the effect of particle size on bulk Fe catalysts have rarely been reported. Herein, nanoparticles assembled Fe₃O₄ bulk catalysts were synthesized via a PAA-mediated solvothermal method to solely investigate the size effect of Fe phase. By tuning preparation parameters, a series of porous-Fe₃O₄ microspheres assembled by different nanoparticle size (8.5–16.5 nm) was obtained, maintaining constant microspheres size. When Fe₃O₄ nanoparticles are smaller than 10 nm, catalysts show similar catalytic activity (FTY). Beyond 10 nm, the FTY obviously decreases with increasing of particle size. Particularly, Fe₃O₄ with nanoparticle size of 9.9 nm performs the highest activity as well as C₂-C₄⁼ selectivity and O/P ratio. The smaller particles attribute to C₅₊ formation, while inhibit CH₄ selectivity. By CO-TPD, TPH-MS and XRD analysis, we discuss size effect on CO adsorption, surface carbon species and carburization degree, building relationship between particle size and catalytic activity. It is found that the carburization degree of iron phase correlates positively with catalytic activity. These results deepen understanding of the structure-performance relationship for bulk iron catalysts in FTO.

Full text of DOI:10.1016/j.cattod.2020.04.017

CatalysisTodayxxx(xxxx)xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Tunable Fe3O4 nanoparticles assembled porous microspheres as catalysts for
Fischer-Tropsch synthesis to lower olefins
Qiao Zhao, Haoting Liang, Shouying Huang*, Xiaoxue Han, Hongyu Wang, Jian Wang,
Yue Wang, Xinbin Ma*
Key Laboratory for Green Chemical Technology of Ministry of Education Collaborative Innovation Center of Chemical Science and Engineering, School of Chemical
Engineering and Technology, Tianjin University, Tianjin 300072, China
A R T I C L E I N F O
A B S T R A C T
Keywords:
Fischer-Tropsch
Lower olefins
Fe3O4 microspheres
Nanoparticles assembly
Size effect
Fe catalyzed Fischer-Tropsch synthesis to lower olefins (FTO) has been recognized as a structure-sensitive re-
action. Nanostructure Fe-based catalysts can expose more active surface. But bulk nanostructure Fe catalysts
without support interaction show easily sintered and agglomerated, which makes hard to investigate size effect.
Therefore, the effect of particle size on bulk Fe catalysts have rarely been reported. Herein, nanoparticles as-
sembled Fe3O4 bulk catalysts were synthesized via a PAA-mediated solvothermal method to solely investigate
the size effect of Fe phase. By tuning preparation parameters, a series of porous-Fe3O4 microspheres assembled
by different nanoparticle size (8.5–16.5 nm) was obtained, maintaining constant microspheres size. When Fe3O4
nanoparticles are smaller than 10 nm, catalysts show similar catalytic activity (FTY). Beyond 10 nm, the FTY
obviously decreases with increasing of particle size. Particularly, Fe3O4 with nanoparticle size of 9.9 nm per-
=
forms the highest activity as well as C2-C4 selectivity and O/P ratio. The smaller particles attribute to C5+
formation, while inhibit CH4 selectivity. By CO-TPD, TPH-MS and XRD analysis, we discuss size effect on CO
adsorption, surface carbon species and carburization degree, building relationship between particle size and
catalytic activity. It is found that the carburization degree of iron phase correlates positively with catalytic
activity. These results deepen understanding of the structure-performance relationship for bulk iron catalysts in
FTO.
1. Introduction
reaction [6–8], which means Fe particle size has obvious influence on
catalytic activity and product distribution. Park et al. [9] studied the
Lower olefins (C2=-C4=) as crucial intermediates in chemical in-
dustry are widely used to produce plastics, polymers, solvents, textiles,
etc [1–3]. Traditionally, these components are mainly produced from
cracking of naphtha or vacuum gas oil. However, due to the excessive
energy-consuming by cracking, growing depletion of crucial oil and
high CO2 emission [4], it is necessary to search for alternative routes to
produce lower olefins. Fischer-Tropsch synthesis to lower olefins (FTO)
is considered as a non-oil based route for the direct production of C2-C4
olefins from syngas. Since syngas can be derived from diverse sources
such as natural gas, coal or biomass, FTO offers an alternative to alle-
viate our dependency on crude oil, which is of great significance
worldwide, especially to oil-scarce countries and regions.
particle size effect of nano-sized γ-Fe2O3 supported on δ-Al2O3. They
found the TOF increased when particle size increased from 2.4 to 6.2
nm, then remained almost constant with further increasing particle size
to 11.5 nm. Torres Galvis et al. [3] selected carbon nanofibers as sup-
port and concluded the apparent TOF decreased with increasing par-
ticle size from 2 to 7 nm, but the selectivity of methane and lower
olefins were not affected. However, in the same group Marianna Ca-
savola et al. [10] reported a totally different result that methane se-
lectivity increased slightly with the increase of colloidal iron oxide size
from 3 to 14 nm over carbon nanotube-supported catalysts. Yuan et al.
[11] firstly prepared Fe3O4 nanoparticles with different particle sizes
from 8.3 to 17.3 nm by thermal decomposition which then were loaded
on α-Al2O3 by impregnation. Among all the catalysts, 12.0 nm nano-
Fe3O4 on α-Al2O3 performed the best catalytic activity whether with or
without promoters (K and S). According to the various conclusions from
different groups, it is noticed that all the performance tendencies with
Iron-based catalysts have been widely studied due to their low price,
broad availability, appreciable selectivity to lower olefins, high water-
gas-shift activity and high resistance to contaminants [1,5]. FTO on
iron-based catalysts has been reported to be a structure-sensitive
⁎ Corresponding authors.
E-mail addresses: huangsy@tju.edu.cn (S. Huang), xbma@tju.edu.cn (X. Ma).
https://doi.org/10.1016/j.cattod.2020.04.017
Received 19 January 2020; Received in revised form 21 March 2020; Accepted 5 April 2020
0920-5861/©2020ElsevierB.V.Allrightsreserved.
Pleasecitethisarticleas:QiaoZhao,etal.,CatalysisToday,https://doi.org/10.1016/j.cattod.2020.04.017

Q. Zhao, et al.
CatalysisTodayxxx(xxxx)xxx–xxx
Fe particle sizes are totally different, which might be ascribed to the
effect of catalyst structure and support. This has also been observed for
Co- and Ru-based catalysts that the structure and support significantly
affect the size effect [12,13]. Although numerous efforts have been
devoted to developing supported catalysts, bulk iron catalysts are still
mainly used in industry until now. So, it is of great significance to ex-
plore size effect of bulk Fe catalysts in FTO reaction. Actually, oversized
particles of traditional bulk Fe catalysts would shield size effect.
Whereas, nano-Fe catalysts tend to agglomeration without anchoring
effect of supports [14]. Therefore, with the help of fabricating nano-
assembled catalysts with stable structure and large active specific area
[15–17], we can solely investigate size effect in Fe-catalyzed FTO re-
action.
room temperature. Finally, the porous-Fe3O4 microspheres were wa-
shed and dried in vacuum.
To investigate the nanoparticle size effect, a series of catalysts with
different nanoparticle size but same microsphere size was synthesized.
During catalyst preparation, the amount of sodium acrylate plays an
important role in controlling the Fe3O4 nanoparticle size, while the
volume ratio of EG/DEG can tune the microsphere size. By increasing
the sodium acrylate amount, more polyacrylate would bind to the
Fe3O4 nanoparticles which inhibited them to aggregate during nitrogen
calcination. Although the Fe3O4 nanoparticle size was decreased, the
microsphere diameter would be enlarged [16]. Therefore, in this case, it
is necessary to adjust both sodium acrylate amount and EG/DEG vo-
lume ratio. This series of catalysts is named as x nm-Fe3O4, in which x
refers to the average nanoparticle size based on XRD calculation.
In this study, a series of porous-Fe3O4 microsphere bulk catalysts
assembled by Fe3O4 nanoparticles with different size were synthesized
through a solvothermal method. By tuning Na acrylate amount and
solvent ratio, the Fe3O4 nanoparticle size could be precisely controlled
in the range of 8.5 nm to 16.5 nm, keeping the assembled microspheres
2.3. Catalyst characterization
Crystal structures of the fresh and used catalysts were determined
by X-ray diffraction (XRD) patterns that was achieved on a Rigaku D/
max-2500 X-ray diffractometer using Cu Kα radiation (λ = 1.5406 Å) at
40 kV and 200 mA. The scanning speed is 5°·min-1 in a 2θ angle range of
10-90°. Scanning electron microscope (SEM) micrographs were per-
formed on a S-4800 field-emission microscope (Hitachi, Japan) at an
accelerating voltage of 3 kV. Transmission electron microscope (TEM)
micrographs were carried out on a JEM-2100 F field-emission micro-
scope at 200 kV. Before SEM and TEM testing, the samples were dis-
persed in ethanol uniformly and then dropped onto a small piece of
copper sheet and ultra-thin carbon membranes. The Brunauer-Emmett-
Teller (BET) surface area and pore size were determined by N2-physi-
sorption at 77 K using a Micromeritics ASAP 2020 surface analyzer.
Prior to physisorption, the fresh catalysts (0.1-0.15 g) were degassed at
573 K and 100 μm Hg for 24 h, and then in-situ degassed at 573 K under
vacuum at < 3 μm Hg for 4 h. Hydrogen temperature programmed
reduction (H2-TPR) was carried out on a Micromeritics AutoChem 2910
equipped with a thermal conductivity detector. Prior to reduction, the
samples were pretreated in Ar at 473 K for 60 min, and then cooled
down to room temperature. The reducing gas (a mixture of 10 % vol. H2
and 90 % vol. Ar) was fed into the samples and the temperature was
increased from 333 K to 1103 K with a ramping rate of 10 K·min-1. X-ray
photoelectron spectroscopy (XPS) was performed on a PHI-1600 ESCA
XPS equipment which was excited by monochromated MgKα X-ray ra-
diation. The binding energy was calibrated with C 1s photoelectron
peak at 284.6 eV as the reference. Carbon monoxide temperature pro-
grammed desorption (CO-TPD) was conducted on a Micromeritics
Autochem II 2920 chemisorption system. He was used as the carrier
gas. The samples were reduced in situ under a mixture of CO and H2 (50
% and 50 % vol.) at 623 K for 2 h. After cooling down to room tem-
perature, CO/Ar (10 % vol.) was injected into the sample tube for 1 h.
Then He was purged until the baseline signal was stable. The CO TPD
signal was recorded by heating the samples from room temperature to
1073 K with a heating rate of 10 K·min-1. Temperature programmed
hydrogen with mass spectrometry (TPH-MS) was conducted on a
Micromeritics Autochem II 2920 chemisorption system connected to a
mass spectrometer. The samples were firstly in-situ reduced by syngas
at 623 K for 2 h. After pretreatment, the samples were cooled down to
room temperature. Then H2/Ar (10 % vol.) was injected into the sample
tube while the temperature increased from room temperature to 1073 K
at 10 K·min-1. CH4 mass signal was monitored using m/z of 16 in the
mass spectrometer.
with
a constant diameter of ∼300 nm. Several characterization
methods, including XRD, SEM, TEM, N2 physisorption, H2-TPR and XPS
were employed to illustrate the textural properties, reducibility and
surface compositions of the fresh catalysts. Together with CO-TPD,
TPH-MS and XRD semi-quantitative analysis, this paper discussed the
particle size effect on catalytic performance over bulk phase Fe-based
catalysts.
2. Experimental section
2.1. Materials
Ferric chloride hexahydrate (FeCl3·6H2O) (> 99 %) was obtained
from Tianjin Guangfu Fine Chemical Research Institute. Sodium acry-
late (NaAcry) (> 98 %) was purchased from Nanjing Shengbicheng
Chemical Technology Co., Ltd. Sodium acetate anhydrous (NaAc),
ethylene glycol (EG), diethylene glycol (DEG), ethanol and tetraethyl
orthosilicate (TEOS) were of analytical grade and purchased from
Tianjin Kermal Chemical Reagent Co., Ltd. All the chemicals and re-
agents were directly used without further purification. Deionized water
was made by a SCD-Ⅱ high purity water system.
2.2. Synthesis of hierarchically porous Fe3O4 microspheres
The hierarchically porous Fe3O4 microspheres were fabricated by a
facile solvothermal method [17]. Typically, 0.54 g FeCl3·6H2O was
dissolved in a mixture solution of 15 mL EG and 5 mL DEG followed by
adding 1.5 g sodium acrylate and 1.5 g sodium acetate under magnetic
stirring. The mixture was vigorously stirred for 1 h at room tempera-
ture. The obtained homogeneous solution was then transferred into a
50 mL Teflon-lined stainless steel autoclave which would be sealed and
placed in an oven at 473 K. After reaction for 12 h, the autoclave was
cooled down to room temperature. The product poly(acrylic acid)-en-
tangled Fe3O4 microspheres (Fe3O4/PAA) was obtained, separated by a
magnet and washed several times with deionized water and ethanol.
Then the Fe3O4/PAA microspheres were coated by a thin silica layer in
order to prevent microspheres aggregation during the following calci-
nation process. In particular, the obtained Fe3O4/PAA microspheres
were dispersed by 9 mL water and 90 mL ethanol under sonic assis-
tance, following by adding 4 mL ammonia solution. Then, the mixture
of 0.4 mL TEOS and 20 mL ethanol was dropped into the solution. After
reaction for 100 min, the product Fe3O4/PAA @SiO2 was obtained and
collected by magnet, then washed repeatedly and dried in vacuum oven
at 333 K overnight. The Fe3O4/PAA @SiO2 microspheres were calcined
at 773 K for 6 h in N2 with rate of 2 K·min-1 in tube furnace to de-
compose the PAA. After cooling to room temperature, the samples were
passivated by a mixture of 1 % vol. O2 and 99 % vol. Ar. The silica layer
of microspheres was dissolved by 0.5 M NaOH aqueous solution at
2.4. Reaction measurements
The catalytic performance over the porous-Fe3O4 microspheres for
FTO reaction was tested in a tubular fixed-bed reactor (i.d. 8.0 mm).
Experiments were performed using 0.1 g of Fe3O4 catalyst, which was
sieved to 40–60 mesh and mixed with 1.9 g of quartz sand. The Fe3O4
2

Q. Zhao, et al.
CatalysisTodayxxx(xxxx)xxx–xxx
catalyst was in-situ reduced by syngas (H2/CO mole ratio = 1) at 623 K
for 10 h with a total flow rate of 15 mL·min-1. Then the reactor tem-
perature was cooled down to 553 K with the flow of inert gas (Ar). The
catalytic reaction was finally performed at 613 K, 1 MPa, with H2/CO
mole ratio of 1 and 10 vol.% Ar as internal standard gas, with a specific
GHSV of 18000 mL·g-1 h-1. The products passed through a hot trap
which was set at 408 K. Then the uncondensed products were analyzed
with an online gas chromatograph Agilent 7890B equipped with a
flame ionization detector (FID) and two thermal conductivity detectors
(TCD). The permanent gases, CH4 and CO2 were separated on an
UltiMetal column and quantified by the internal standard gas Ar with a
TCD detector. The hydrocarbon gases were separated on an HP-AL/S
column and detected with an FID detector.
sharp and strong without other impurity diffraction peak, indicating the
porous-Fe3O4 microspheres have high crystallinity and the amorphous
SiO2 layer has been removed completely. According to the Scherrer
equation, the crystal sizes of the obtained catalysts increase from 8.5
nm to 16.5 nm which are calculated based on (311) facet (2θ = 35°).
The morphology and size distribution of the as-prepared samples
were investigated by TEM and SEM, as shown in Fig. 2 and Fig. S1. It is
clear that all the samples are nanoparticles assembled porous-Fe3O4
microspheres with diameter of ∼300 nm. By tuning the ratio of NaAc/
NaAcry and EG/DEG, the average nanoparticle size of the obtained
catalysts was successfully manipulated in the range of 8-16 nm (cal-
culated by more than 300 nanoparticles, TEM) without changing the
microsphere diameter, as shown in Table 1. Note that the average na-
noparticle size obtained from TEM measurement is slightly larger than
that from XRD calculation, because the TEM measurements mainly
count up nanoparticles on the outside layer of microspheres while the
XRD results represent the average nanoparticle size according to a
certain lattice plane. In the following discussion, we name the catalysts
according to the XRD results.
The pore size distribution and specific surface area of the fresh
catalysts were characterized by N2 physisoprtion. Fig. 3a shows that
isotherms of the four samples demonstrate a combination of type I and
type Ⅳ sorption isotherm, indicating both microporous and meso-
porous characteristics. With increasing of the nanoparticle size, the
nitrogen sorption isotherm more tends to type Ⅳ isotherm. The BET
surface areas of 8.5 nm-Fe3O4 and 9.9 nm-Fe3O4 microspheres are 81
and 89 m2·g-1, respectively (Table 1). These values are relatively high
compared with other reported porous-structure Fe3O4 microspheres
whose BET surface areas were less than 40 m2·g-1 [20,21]. Further in-
creasing the nanoparticle size from 9.9 nm to 16.5 nm, the corre-
sponding specific area will decrease from 89 to 33 m2·g-1. Since the
thermal decomposition of polyacrylate was reported to be beneficial for
improving surface area [17], it is reasonable that the 11.3 nm-Fe3O4
and 16.5 nm-Fe3O4 prepared with less or no NaAcry have much lower
surface areas. Fig. 3b shows the pore size distribution of the catalysts
based on NLDFT adsorption model [22]. With increasing particle size,
the microporous volume decreases while the mesoporous volume is
improved. For 16.5 nm Fe3O4, both micro- and meso- porosity are
limited. These observations demonstrate that the amount of Na acrylate
mainly influences the surface area and mesopore size distribution.
The H2-TPR reveals that all the porous-Fe3O4 microspheres show
three overlapped reduction peaks (Fig. 4a), which are consistent with
the typical reduction characteristics of Fe-based catalysts. The first
small reduction peak at around 623 K is assigned to the reduction to
Fe3O4 from Fe2O3 [23], perhaps because scarce Fe3O4 was oxidized to
Fe2O3 when passivated by the mixture of oxygen and argon. As it is
hard to distinguish the characteristic diffraction peaks between γ-Fe2O3
phase and Fe3O4 phase from XRD, XPS spectra was employed to vali-
date the nature of Fe species. In the Fe 2p spectra (Fig. 4b), the two
peaks at the binding energies of ∼711 eV and ∼ 724 eV could be as-
signed to Fe3O4 phase [24]. That testifies that Fe3O4 phase exists as the
dominated species in the fresh catalysts. The second and third reduction
peaks appearing at 773-873 K and above 973 K represent the reduction
of Fe3O4 to FeO and the formation of metallic iron phase. It is obvious
that the two samples with small nanoparticle size (8.5 and 9.9 nm) have
similar reducibility, while further increasing nanoparticle size makes
the reduction of samples more difficult.
The FTY was calculated based on moles of CO which was converted
to hydrocarbons per gram of Fe per second. The selectivity of hydro-
carbon was determined by carbon products except CO2. The carbon
balance of all reactions were about 95 %. The apparent TOF was esti-
mated according to:
FTY× 7.14·10-20×Na×ρ×ϕ
R×Na
TOF=
=
S/ 7.14·10-20
6×wFe5C2
(1)
where R is reaction rate and S is the total surface area of Fe per gram of
catalyst. The value of '7.14·10-20' stands for surface area per Fe atom in
Fe5C2 [18]. The wFe5C2 is the mass fraction of Fe5C2 in the used catalysts
based on XRD patterns. Then the total surface area (S) was calculated
by 6/ρϕ, in which ϕ and ρ are the diameter and density of Fe5C2 particle
(ρ= 7.57 g·mL-1), respectively. We obtained the diameter of Fe3O4
particles from TEM analysis. However, considering the Fe3O4 phase was
transformed to mixture of oxide-carbide-amorphous composite during
reaction, the change in density was inevitable [19]. Therefore, the
diameter of Fe5C2 particle was calculated by the average Fe3O4 particle
size, as shown below:
1
3
3×ρFe3O4 × ϕ3Fe3O4 × MFe5C2
⎛
⎜
⎞
⎟
ϕFe5C2
=
5×ρFe5C2 × MFe3O4
(2)
⎝
⎠
3. Results and discussion
3.1. Characterization of Fe3O4 microspheres with different nanoparticle size
Fig. 1 shows the XRD patterns of the obtained Fe3O4 microspheres.
All of the diffraction peaks can be indexed to the face-centered-cubic
phase of Fe3O4 (JCPDS no. 19-0629) [17]. The diffraction peaks are
3.2. Catalytic performance
The FTO catalytic properties of the porous-Fe3O4 microspheres with
different nanoparticle size are presented in Table 2 and Fig. 5. The CO
conversion differs from 55.3 % to 6.7 %. The 8.5 nm-Fe3O4 and 9.9 nm-
Fe3O4 exhibit similar CO conversion and FTY value. Further increasing
nanoparticle size, the catalytic activity tends to decrease. This demon-
strates the structure sensitivity of Fe-based catalysts in F-T reaction and
Fig. 1. XRD patterns of the porous-Fe3O4 microspheres obtained with different
nanoparticle size.
3

Q. Zhao, et al.
CatalysisTodayxxx(xxxx)xxx–xxx
Fig. 2. TEM images and particle size distribution of the porous-Fe3O4 microspheres (a) 8.5 nm-Fe3O4, (b) 9.9 nm-Fe3O4, (c) 11.3 nm-Fe3O4, (d) 16.5 nm-Fe3O4.
Table 1
The precursor compositions, average nanoparticle (microsphere) sizes and pore structural parameters.
d
samples
NaAc/NaAcry (mass ratio) EG/DEG (volume ratio) dFe3O4 (nm)a,b Dmicrosphere (nm)b,c dused
(nm)b SBET (m2 g-1
)
Dmean (nm)e Vtotal (cm3 g-1
)
Fe3O4
8.5 nm-Fe3O4
9.9 nm-Fe3O4
11.3 nm-Fe3O4 2/1
16.5 nm-Fe3O4 1.5/0
2/1.2
1.5/1.5
17/3
15/5
20/0
20/0
8.5 / 7.2
9.9 / 8.6
11.3 /10.4
16.5 / 15.6
305 / 337
300 / 340
304 / 345
300 / 330
12.8
14.7
18.0
23.0
81
89
65
33
0.5
0.5
0.5
0.6
0.09
0.10
0.09
0.10
a
Calculated from XRD
Determined by TEM.
Dertermined by SEM.
Calculated according to BET method.
b
c
d
e
Mean micropore size besed on NLDFT method.
the size around 10 nm seems a critical point in this case. In reported
work, CO conversion usually depends on the CO adsorption and asso-
ciation as well as carburization of Fe species, which would be inter-
preted in the following part.
selectivity to CH4 increased accompanied by a decrease of selectivity to
C5+ with increasing the particle size. Besides, the O/P ratio increases
firstly, and then starts to decrease with increasing the particle size,
which is in line with the change trend of catalytic activity. Borg and
Rane reported that C5+ selectivity would increase with the enlarged
particle size (< 8 nm), while further increasing particle size C5+ se-
lectivity turned decreased [26,27]. They ascribed the declined C5+
selectivity of enlarged particles (> 8 nm) to lower surface inter-
mediates reactivity represented by smaller O/P ratio. The CO2
=
For products distribution, 9.9 nm-Fe3O4 shows the highest C2-C4
=
selectivity of 32.4 %, while in fact C2-C4 selectivity of these catalysts
exhibits unobvious change only with a difference of 2.3 %. For all
samples, changes in product distribution as a function of particle size
are consistent with the findings in literatures [10,25], in which the
Fig. 3. N2 physisorption of the porous-Fe3O4 microspheres: (a) nitrogen adsorption-desorption isotherms, (b) pore size distribution.
4

Q. Zhao, et al.
CatalysisTodayxxx(xxxx)xxx–xxx
Fig. 4. (a) H2-TPR curves, (b) XPS spectra of the porous-Fe3O4 microspheres with different nanoparticle size.
selectivity changes in parallel to CO conversion, probably because the
samples with smaller particles exhibit strong CO adsorption and dis-
sociation, which would promote Boudouard reaction (2CO → C +
CO2).
To illustrate the stability of these catalysts, Fig. 5 shows the CO
conversion changes with time on stream. The 8.5 nm-Fe3O4 and 9.9 nm-
Fe3O4 show an obvious induction period and then the CO conversion
tends to be stable, while the CO conversion on 11.3 nm-Fe3O4 shows a
gradual and moderate increase. In contrast, the 16.5 nm-Fe3O4 shows
the lowest activity and fastest deactivation rate. The morphology of
porous 9.9 nm-Fe3O4 microspheres has not be damaged after the 52 h
reaction (Fig. S2a), while the used 16.5 nm-Fe3O4 shows some structure
collapse (Fig. S2b). The collapse of porous microsphere structure would
decrease surface area and increase diffusion resistance, resulting in
declined activity and fast deactivation.
Fig. 5. CO conversion of porous-Fe3O4 microspheres on TOS. (P = 1 MPa, H2/
CO = 1, GHSV = 18000 mL·g-1 h-1, T =613 K).
3.3. CO adsorption, surface carbon species and carburization degree
To analyze carburization behavior of the catalysts with different
particle size, CO adsorption capacity, surface carbon species and phase
compositions were characterized by CO-TPD, TPH-MS and XRD ana-
lysis. Fig. 6a shows the CO-TPD profiles of the porous-Fe3O4 micro-
spheres. An enhanced adsorption strength of CO is observed with the
decrease of nanoparticle size, as evidenced by an improvement of
desorption temperature. Accordingly, the integral area of CO adsorp-
tion also displays the same trends (Fig. 6b). This demonstrates that the
particle size affects not only strength but also capacity of CO adsorp-
tion. The observation can be interpreted by the fact that the smaller Fe
particles expose more surface areas with unsaturated coordination, re-
sulting in more adsorption sites and enhanced electron back-donation
from metal species to CO [25]. It has been proposed that weaker CO
adsorption would benefit higher FT synthesis activity, because the
weaker Fe-C bond facilitated removal and hydrogenation of C and O
atoms from Fe surface [28]. However, too weak CO adsorption may
make it hard to CO activation and dissociation. It has also been pointed
out that the enhanced CO adsorption could facilitate CO dissociation
and accelerate CO conversion [29,30]. So we speculate that appropriate
CO adsorption strength is beneficial to CO conversion and iron carbides
formation.
The chemical phase and structural characterization of the used
catalysts was carried out after 52 h on stream. XRD patterns of all the
used catalysts are shown in Fig. 6c. The diffraction peaks are indexed to
the phases of Fe3O4 and Fe5C2. Hägg carbide (χ-Fe5C2) has been widely
considered to be the active phase for FTO reaction [31–33]. The relative
quantity of Fe5C2 and Fe3O4 was calculated by XRD semi-quantitative
analysis. The most intense peaks of Fe5C2 and Fe3O4 were integrated
respectively from fitting of the XRD patterns [34,35]. Then the mass
fraction of Fe5C2 phase (wx) could be calculated by integral of the most
intense diffraction peak, according to the following Eq. (3). As shown in
Fig. 6d, the calculated mass fraction of Fe5C2 phase in used 8.5 nm-
Fe3O4 and 9.9 nm-Fe3O4 catalysts are similar, and increasing the
Table 2
FTO performance of porous-Fe3O4 microspheres with different nanoparticle size.
Catalysts
CO conv. (%)
FTY (μmolCO gFe-1s-1
)
Apparent TOF ×10(s-1
)
CO2 selectivity (%)
Hydrocarbon selectivity (C-mol%)
O/P
=
CH4
C2-C4
C2-C4°
C5+
Oxy.
8.5 nm-Fe3O4
9.9 nm- Fe3O4
11.3 nm- Fe3O4
16.5 nm- Fe3O4
53.1
55.3
26.2
6.7
44
45
25
8
0.21
0.22
0.19
0.11
46.1
47.2
38.8
22.0
22.4
24.1
25.5
36.8
30.1
32.4
32.1
30.2
13.2
11.9
13.4
14.4
31.7
28.8
26.2
16.4
2.6
2.8
2.8
2.2
2.3
2.7
2.4
2.1
Reaction conditions: 0.1 g cat., 613 K, 1 MPa, H2/CO = 1, GHSV = 18000 mL·g-1 h-1, TOS = 52 h.
5

Q. Zhao, et al.
CatalysisTodayxxx(xxxx)xxx–xxx
Fig. 6. (a) CO-TPD profiles and (b) adsorption capacities of the syngas pretreated porous-Fe3O4 microspheres, (c) XRD patterns and (d) mass fraction of Fe5C2 phase
for the used catalysts.
nanoparticle size gives rise to an obvious decline in Fe5C2 content.
Combining with CO-TPD analysis, this indicates higher capacity and
appropriate strength of CO adsorption are beneficial for the carbur-
ization of Fe phase, increasing Fe5C2 phase content in used catalysts
assembled with smaller particles. Moreover, the extent of transforma-
tion into Fe5C2 phase positively correlates with catalytic activity, which
is consistent with other findings [30,36,37].
Ix
wx =
N
I
i
RIR
RIRx∑
i=1
(3)
i
The HRTEM of used catalyst (Fig. S3) displays that the nanoparticles
are surrounded by a layer of amorphous carbonaceous species produced
during reaction, instead of polycrystalline carbon shell [38]. In addi-
tion, visible lattice fringes with d spacing of 1.48 Å and 2.05 Å were
observed, corresponding to the (440) lattice planes of Fe3O4 and (510)
lattice planes of Fe5C2 respectively [24,39]. In Table 1, the average
particle sizes of the used catalysts are presented. Because of phase
transformation during reaction, the particle sizes are larger than those
of the fresh catalysts. However, the order of the particle size is main-
tained, indicating the size effect is reliable.
The TPH-MS was performed to analyze the surface carbon species of
syngas-treated samples (Fig. 7). The CH4 signal was recorded by using
m/z = 16. We fitted the CH4 curves into five peaks (α, β, γ1, γ2 and δ),
which are associated with different carbon species including adsorbed
atomic carbon, polymeric amorphous aggregate carbon, ε’ and χ iron
carbides, and graphitic carbon [40]. The compositions of carbon species
are collected in Table 3. The compositions of α, β and δ carbon species
are relatively low. The most intense peaks (i.e. γ1, γ2) are assigned to
Fig. 7. TPH spectra of the syngas pretreated porous-Fe3O4 microspheres with
different nanoparticle size.
iron carbides that are generally accepted as the active phases for FTO
reaction. Generally, upon increasing particle size, the amounts of γ1 and
γ2 carbon decrease accordingly, consistent with the XRD results. The
8.5 nm-Fe3O4 shows similar amount of γ1 and γ2 with 9.9 nm-Fe3O4,
but higher β carbon content results in activity difference.
4. Conclusion
The PAA-mediated solvothermal method was employed to synthe-
size the large-surface-area porous-Fe3O4 microspheres in FTO reaction.
6

Q. Zhao, et al.
CatalysisTodayxxx(xxxx)xxx–xxx
Table 3
TPH peak temperatures and percentage compositions of carbon species for porous-Fe3O4 microspheres
samples
carbidic (α)
amorphous (β)
carbide (γ1/γ2)
graphitic (δ)
Temp. (K)
Percent. (%)
Temp. (K)
Percent. (%)
Temp. (K)
Percent. (%)
Temp. (K)
Percent. (%)
8.5 nm-Fe3O4
9.9 nm-Fe3O4
11.3 nm-Fe3O4
16.5 nm-Fe3O4
603
581
584
585
0.5
0.5
0.3
0.3
678
674
679
680
2.3
0.9
0.6
0.6
783/911
779/918
795/889
796/890
18.7/76.2
21.3/76.9
16.8/81.9
13.3/85.2
993
1023
994
2.3
0.4
0.4
0.6
1014
By tuning the amount of sodium acrylate and the volume ratio of EG/
DEG, fabrication of bulk porous-Fe3O4 microspheres (diameter ∼300
nm) assembled by different nanoparticle size (8.5-16.5 nm) was suc-
cessfully achieved. Thanks to abundant porous structure and nano-as-
sembled structure, the catalytic activity and stability perform excellent.
More importantly, this provides possibility to explore the size effect by
eliminating support influence. The catalytic activity (FTY) shows this
trend with varying the nanoparticle size: samples with nanoparticle size
(< 10 nm) exhibits similar FTY, then the larger nanoparticle size, the
lower FTY. The higher CO uptake and adsorption strength have been
observed with smaller nanoparticle size. And the increased CO uptake
and appropriate adsorption strength could promote Fe phase carbur-
ization, which is evidenced by TPH-MS and XRD results. Among all
catalysts, 9.9 nm-Fe3O4 performs the optimal FTY, TOF and highest
lower olefins selectivity as well as O/P ratio, which has greatest Fe
carbide content. Furthermore, the increase of nanoparticle size would
lead to increasing of methane selectivity and decreasing of longer-chain
hydrocarbons selectivity. These conclusions provide a further under-
standing of bulk iron catalyst particle size effect in Fischer-Tropsch to
lower olefins.
[9] J.Y. Park, Y.J. Lee, P.K. Khanna, K.W. Jun, J.W. Bae, Y.H. Kim, J. Mol. Catal. A:
Chem. 323 (2010) 84–90.
[10] M. Casavola, J. Hermannsdörfer, N. de Jonge, A.I. Dugulan, K.P. de Jong, Adv.
Funct. Mater. 25 (2015) 5309–5319.
[11] Y. Yuan, S. Huang, H. Wang, Y. Wang, J. Wang, J. Lv, Z. Li, X. Ma, ChemCatChem 9
(2017) 3144–4152.
[12] J.P. den Breejen, P.B. Radstake, G.L. Bezemer, J.H. Bitter, V. Feøseth, A. Holmen,
K.P. de Jong, J. Am. Chem. Soc. 131 (2009) 7197–7203.
[13] J. Kang, S. Zhang, Q. Zhang, Y. Wang, Angew. Chem. 48 (2009) 2565–2568.
[14] C. Wang, Q. Wang, X. Sun, L. Xu, Catal. Lett. 105 (2005) 93–101.
[15] H. Deng, X. Li, Q. Peng, X. Wang, J. Chen, Y. Li, Angew. Chem. 44 (2005)
2782–2785.
[16] S. Xuan, Y.X.J. Wang, J.C. Yu, K. Cham Fai Leung, Chem. Mater. 21 (2009)
5079–5087.
[17] S. Xuan, F. Wang, J.M. Lai, K.W. Sham, Y.X. Wang, S.F. Lee, J.C. Yu, C.H. Cheng,
K.C. Leung, ACS Appl. Mater. Interfaces 3 (2011) 237–244.
[18] V. Iablokov, Y. Xiang, A. Meffre, P.F. Fazzini, B. Chaudret, N. Kruse, ACS Catal. 6
(2016) 2496–2500.
[19] J. Wang, S. Huang, S. Howard, B.W. Muir, H. Wang, D.F. Kennedy, X. Ma, ACS
Catal. 9 (2019) 7976–7983.
[20] Y. Zhang, L. Ma, T. Wang, X. Li, New J. Chem. 39 (2015) 8928–8932.
[21] Y. Zhang, L. Ma, J. Tu, T. Wang, X. Li, Appl. Catal. A: Gen. 499 (2015) 139–145.
[22] J. Jagiello, M. Jaroniec, J. Colloid Interface Sci. 532 (2018) 588–597.
[23] Y. Wei, C. Zhang, X. Liu, Y. Wang, Q. Chang, M. Qing, X. Wen, Y. Yang, Y. Li, Catal.
Sci. Technol. 8 (2018) 1113–1125.
[24] C. Yang, H. Zhao, Y. Hou, D. Ma, J. Am. Chem. Soc. 134 (2012) 15814–15821.
[25] Q. Cheng, Y. Tian, S. Lyu, N. Zhao, K. Ma, T. Ding, Z. Jiang, L. Wang, J. Zhang,
L. Zheng, F. Gao, L. Dong, N. Tsubaki, X. Li, Nat. Commun. 9 (2018) 3250.
[26] O. Borg, P. Dietzel, A. Spjelkavik, E. Tveten, J. Walmsley, S. Diplas, S. Eri,
A. Holmen, E. Rytter, J. Catal. 259 (2008) 161–164.
Acknowledgements
[27] S. Rane, Ø. Borg, E. Rytter, A. Holmen, Appl. Catal. A: Gen. 437–438 (2012) 10–17.
[28] J. Cheng, P. Hu, P. Ellis, S. French, G. Kelly, C.M. Lok, J. Phys. Chem. C 114 (2010)
1085–1093.
This work was supported by National Natural Science Foundation of
China (U1462204) and Natural Science Foundation of Tianjin
(18JCQNJC05900, 19JCYBJC20700).
[29] Y. Cheng, J. Lin, K. Xu, H. Wang, X. Yao, Y. Pei, S. Yan, M. Qiao, B. Zong, ACS Catal.
6 (2015) 389–399.
[30] J. Lu, L. Yang, B. Xu, Q. Wu, D. Zhang, S. Yuan, Y. Zhai, X. Wang, Y. Fan, Z. Hu, ACS
Catal. 4 (2014) 613–621.
Appendix A. Supplementary data
[31] H.J. Schulte, B. Graf, W. Xia, M. Muhler, ChemCatChem 4 (2012) 350–355.
[32] T.H. Pham, X. Duan, G. Qian, X. Zhou, D. Chen, J. Phys. Chem. C 118 (2014)
10170–10176.
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.cattod.2020.04.017.
[33] B. Chen, D. Wang, X. Duan, W. Liu, Y. Li, G. Qian, W. Yuan, A. Holmen, X. Zhou,
D. Chen, ACS Catal. 8 (2018) 2709–2714.
References
[34] Z. Li, L. Zhong, F. Yu, Y. An, Y. Dai, Y. Yang, T. Lin, S. Li, H. Wang, P. Gao, Y. Sun,
M. He, ACS Catal. 7 (2017) 3622–3631.
[35] M.D. Shroff, D.S. Kalakkad, K.E. Coulter, S.D. Kӧhler, M.S. Harrington,
N.B. Jackson, A.G. Sault, A.K. Datye, J. Catal. 156 (1995) 185–207.
[36] Y. Cheng, J. Lin, T. Wu, H. Wang, S. Xie, Y. Pei, S. Yan, M. Qiao, B. Zong, Appl.
Catal. B: Environ. 204 (2017) 475–485.
[37] X. Duan, D. Wang, G. Qian, J.C. Walmsley, A. Holmen, D. Chen, X. Zhou, J. Energy
Chem. 25 (2016) 311–317.
[38] N.G. Hamilton, R. Warringham, I.P. Silverwood, J. Kapitán, L. Hecht, P.B. Webb,
R.P. Tooze, W. Zhou, C.D. Frost, S.F. Parker, D. Lennon, J. Catal. 312 (2014)
221–231.
[39] S. Zhao, X. Liu, C. Huo, X. Wen, W. Guo, D. Cao, Y. Yang, Y. Li, J. Wang, H. Jiao,
Catal. Today 261 (2016) 93–100.
[1] H.M. Torres Galvis, J.H. Bitter, C.B. Khare, M. Ruitenbeek, A.I. Dugulan,
K.P.d. Jong, Science 335 (2012) 835–838.
[2] H.M. Torres Galvis, K.P. de Jong, ACS Catal. 3 (2013) 2130–2149.
[3] H.M. Torres Galvis, J.H. Bitter, T. Davidian, M. Ruitenbeek, A.I. Dugulan, K.P. de
Jong, J. Am. Chem. Soc. 134 (2012) 16207–16215.
[4] R. Warringham, A.L. Davidson, P.B. Webb, R.P. Tooze, S.F. Parker, D. Lennon,
Catal. Today 339 (2020) 32–39.
[5] S. Abello, D. Montane, ChemSusChem 4 (2011) 1538–1556.
[6] J. Xie, J. Yang, A.I. Dugulan, A. Holmen, D. Chen, K.P. de Jong, M.J. Louwerse, ACS
Catal. 6 (2016) 3147–3157.
[7] J.X. Liu, P. Wang, W. Xu, E.J.M. Hensen, Engineering 3 (2017) 467–476.
[8] L.A. Cano, M.V. Cagnoli, N.A. Fellenz, J.F. Bengoa, N.G. Gallegos, A.M. Alvarez,
S.G. Marchetti, Appl. Catal. A: Gen. 379 (2010) 105–110.
[40] J. Xu, C.H. Bartholomew, J. Phys. Chem. B 109 (2005) 2392–2403.
7