Effect of Topochemical Processes in the Synthesis of FeK/C Catalysts on Their Activity and Selectivity in the Fischer–Tropsch Synthesis

Article abstract of DOI:10.1134/S0023158418060022

Abstract: Iron-containing catalysts promoted with potassium supported on a carbon carrier were prepared by changing the sequence of the introduction of components (iron and potassium) into the carrier (activated carbon) using an impregnation method. After calcination, the catalysts contained hematite and magnetite; in this case, the particle sizes of iron oxides depended on the sequence of the introduction of a potassium promoter. A minimum particle size was observed upon the subsequent introduction of initially potassium from nitrate and then iron. The activation of all of the catalysts in a flow of CO/H₂ resulted in the formation of H?gg carbide (Fe₅C₂). It was shown that the catalyst with the smallest carbide particle size was the most active in the hydrogenation of CO.

Full text of DOI:10.1134/S0023158418060022

ISSN 0023-1584, Kinetics and Catalysis, 2018, Vol. 59, No. 6, pp. 828–836. © Pleiades Publishing, Ltd., 2018.
Original Russian Text © P.A. Chernavskii, O.L. Eliseev, R.V. Kazantsev, G.V. Pankina, N.E. Strokova, A.L. Lapidus, 2018, published in Kinetika i Kataliz, 2018, Vol. 59, No. 6,
pp. 795–803.
Effect of Topochemical Processes in the Synthesis of FeK/C Catalysts
on Their Activity and Selectivity in the Fischer–Tropsch Synthesis
P. A . Cher n avs kii a, *, O. L. Eliseevb, R. V. Kazantsevb, G. V. Pankinaa,
N. E. Strokovaa, and A. L. Lapidusc
aDepartment of Chemistry, Moscow State University, Moscow, 119991 Russia
bZelinsky Institute of Organic Chemistry, Russian Academy of Sciences, Moscow, 119991 Russia
cGubkin State University of Oil and Gas, Moscow, 119991 Russia
*e-mail: chern5@inbox.ru
Received March 5, 2018
Abstract—Iron-containing catalysts promoted with potassium supported on a carbon carrier were prepared
by changing the sequence of the introduction of components (iron and potassium) into the carrier (activated
carbon) using an impregnation method. After calcination, the catalysts contained hematite and magnetite; in
this case, the particle sizes of iron oxides depended on the sequence of the introduction of a potassium pro-
moter. A minimum particle size was observed upon the subsequent introduction of initially potassium from
nitrate and then iron. The activation of all of the catalysts in a flow of CO/H2 resulted in the formation of
Hägg carbide (Fe5C2). It was shown that the catalyst with the smallest carbide particle size was the most active
in the hydrogenation of CO.
Keywords: iron oxide nanoparticles, promotion, catalyst, activated carbon, carbides, Fischer–Tropsch syn-
thesis
DOI: 10.1134/S0023158418060022
The Fischer–Tropsch (FT) synthesis is a key step method for producing an active catalyst. In a number
in the production of liquid hydrocarbons from coal, of works [11–14], the influence of the sequence of
natural gas, or biomass [1, 2]. Moreover, the direct introducing components into a catalyst on its activity
conversion of synthesis gas (CO + H2) into light C2– and selectivity was studied. The method used for
introducing promoters and the type of iron com-
pounds play a decisive role in the development and
production of highly active, selective, and stable iron-
based composite systems. Duan et al. [11] obtained a
catalyst, which was a composite of carbon nanotubes
and iron and potassium oxides, by a redox reaction
between carbon nanotubes and K2FeO4 with the sub-
sequent heat treatment. The study of samples prepared
by impregnation with the subsequent heat treatment
showed that the new composite catalyst had relatively
high activity, greater selectivity for hydrocarbons and
short olefins, and better stability, and it increased the
probability of chain growth.
For the conversion of synthesis gas into ethanol,
Wang et al. [12] modified a rhodium catalyst sup-
ported on silicon dioxide with various transition metal
oxides. They found that iron oxide is an effective pro-
moter of ethanol formation. It turned out that the way
of introducing FeOx significantly affects the conver-
sion and selectivity. The catalyst prepared by the
impregnation of a FeOx–SiO2 composite, which was
C4 olefins without intermediate stages, the so-called
Fischer–Tropsch-to-Olefins (FTO) process, is a
promising alternative method for the production of
high-value chemicals [3, 4]. Iron-containing catalysts
for the FT synthesis are more preferable than cobalt
catalysts, which are more active, because they are less
expensive and exhibit higher selectivity for the forma-
tion of olefins. The results of studies [5, 6] have shown
that supported iron-containing catalysts have undeni-
able advantages over unsupported catalysts. In turn,
the use of carbon carriers has good prospects because
of their weak interaction with a catalytically active
phase. It was assumed [7, 8] that the distribution of the
active component of a catalyst and a promoter affects
the process selectivity of the FT synthesis. It was
reported [9, 10] that potassium facilitates the carbon-
ization of iron and increases the probability of chain
growth.
The simplest way to prepare a promoted catalyst is
the co-impregnation of the support with an aqueous
solution of the salts of an active component and a pro-
moter. However, it is not obvious that this is the best synthesized previously by a sol–gel method, with an
828

EFFECT OF TOPOCHEMICAL PROCESSES
829
aqueous solution of Rh(NO3)3 provided better activity rather than calcined. This series of samples was
intended for studying by differential thermal analysis
(DTA) and magnetometry.
in the synthesis of ethanol than that of the samples
prepared by combined impregnation. Ledford et al.
[13] prepared CoLa/Al2O3 catalysts by the following
two methods: impregnation with a salt of La and then
with a salt of Co (Co/La) and impregnation with a salt
of Co and then with a salt of La (La/Co). The La–Co
interaction was absent from the La/Co samples.
Gnanamani and coworkers [14] studied the effect
of the order of phosphorus and cobalt addition to silica
in the course of the synthesis of Fischer–Tropsch cat-
alysts. They found that the preliminary application of
phosphorus increased the dispersity of cobalt and,
thereby, facilitated an increase in the activity of the sam-
ple in the FT synthesis and its selectivity for heavy hydro-
carbons. On the contrary, the introduction of phospho-
rus after cobalt had a negative effect on the conversion of
CO and on selectivity for C5+ hydrocarbons.
The main purpose of this work was to study the
effect of the sequence of introduction of components
on the activity and selectivity of a FeK/C catalyst. The
catalysts were prepared by the following three meth-
ods: the co-impregnation of an activated carbon car-
rier with the solutions of iron and potassium nitrates,
the sequential impregnation with the solutions of
Fe(NO3)3 and then KNO3, and the sequential impreg-
nation with the reverse order of introduction of the
components.
The elemental composition of the surfaces of the
resulting catalysts was determined by energy-disper-
sive X-ray spectroscopy (EDS) using a JSM-6490LV
scanning electron microscope (JEOL, Japan). The
bulk composition of the catalysts was analyzed by an
inductively coupled plasma (ICP) method using an
ICP-5000 instrument (Focused Photonics, South
Korea).
Differential Thermal Analysis
The test sample was heated from 40 to 600°C at a
rate of 10 K/min in an inert atmosphere (in a flow of
argon at 80 mL/min) in a Netzsch STA 449 C Jupiter
thermal analyzer (Netzsch, Germany). In the course
of heating, effluent gases were supplied to an Aeolos
QMS 403C quadrupole mass spectrometer (Netzsch,
Germany) for scanning ionized fragments over an m/z
range from 10 to 50.
Magnetometry
The kinetic studies of the samples were conducted
on a vibrating-coil magnetometer in the in situ mode.
The test samples with a weight of 10 mg were placed in
the measuring cell of a vibrating-coil magnetometer
[15], which was a flow microreactor with a volume of
0.3 cm3. The test sample was tightly fixed between two
porous quartz membranes. For all of the initial cata-
lysts and catalysts after activation in synthesis gas, the
EXPERIMENTAL
Catalyst Preparation
The catalyst synthesis procedure included the fol- dependences of magnetization on the magnetic field
lowing sequence of operations: the preparation of the strength were obtained and the saturation magnetiza-
nitrate solutions of KNO3 (analytical grade) and tion, residual magnetization, and coercive force were
measured. The gas flow rate was 10 mL/min in all of
the experiments. The gases were preliminarily passed
through a column with γ-Al2O3 heated to 300°C for
the removal of iron carbonyls. In a nonisothermal
experiment, the sample was heated to a specified tem-
perature in a flow of CO (or synthesis gas). The heat-
ing to a specified temperature was performed at a rate
of 10 K/min; in this case, changes in the magnetiza-
tion were continuously measured with a frequency of
1 Hz. After reaching the required temperature, the
sample was cooled at a rate of 10 K/min in a CO/H2
gas flow and the temperature dependence of magneti-
zation was measured. In an isothermal experiment,
the test sample was heated to a specified temperature
in a flow of argon; then, the flow of argon was replaced
by a flow of CO (or synthesis gas) and changes in the
magnetization with time were also measured thereaf-
ter. An IR detector was arranged at the reactor outlet
to measure continuously the rate of CO2 formation. In
Fe(NO3)3 ⋅ 9H2O (analytical grade); the impregnation
of activated carbon (AC) (VEB Laborchemie Apolda;
specific surface area, 820 m2/g) with one of the solu-
tions; drying in a rotary evaporator at 80°C; and calci-
nation at 450°C for 3 h in a flow of nitrogen. Then, the
procedure was repeated with the second component,
iron nitrate or potassium nitrate. Thus, catalysts with
different sequences of introducing the potassium pro-
moter were obtained. The concentration of nitrate
solutions was 50%, and the calculated concentrations
of iron and potassium in the catalysts were 15 and 2 wt %,
respectively. In addition, a catalyst was prepared by
co-impregnating the support with a mixture of iron
and potassium nitrates in the same percentage.
Henceforth, the catalysts are referred to as K/Fe/C,
Fe/K/C, and (FeK)/C, where the application
sequence should be read from right to left. The catalyst
samples containing only iron or only potassium and a
catalyst prepared by co-impregnation were also syn-
thesized; in this case, the catalysts were only dried some cases, when it was necessary to more accurately
KINETICS AND CATALYSIS Vol. 59 No. 6 2018

830
CHERNAVSKII et al.
determine the Curie temperature, the sample after lytic activity was evaluated after 20 h of operation upon
reaching room temperature was repeatedly heated in a reaching a pseudo-stationary regime. The composi-
flow of argon and the change in magnetization was
measured. As a result, we obtained thermomagnetic
functions, which allowed us to determine the Curie
temperature of the resulting phase with a sufficiently
high accuracy; this temperature was found as the inter-
section point of a tangent to an extremum point of the
thermomagnetic curve with the x-axis. Bends in the
thermomagnetic curve indicated the occurrence of
several magnetic phases. In this case, the type of a
magnetic phase according to the Curie temperature
can be determined based on the position of a bend in
the curve. An analysis of thermomagnetic curves in
the presence of two magnetic phases also makes it pos-
sible to determine the quantitative composition of
magnetic phases in the test sample [16].
tion of the gaseous products (C1–C4 hydrocarbons
and CO2) was determined on an LKhM-80 gas chro-
matograph (Russia) with columns (1 × 3 mm) packed
with molecular sieves 5A and Porapak Q; helium was
a carrier gas, and a katharometer was used as a detec-
tor. The C5+ liquid hydrocarbons were collected in a
receiver at ambient temperature and analyzed on a
Biochrom-1 chromatograph (Russia) using a quartz
capillary column (50 m × 0.25 mm), an OV-101 sta-
tionary phase, nitrogen as a carrier gas, and a flame-
ionization detector. Selectivity for the formation of
C5+ products was calculated by difference between the
total balance weight of carbon and the amount of C1–C4
and CO2 gases.
Transmission Electron Microscopy
RESULTS AND DISCUSSION
The microstructure of the samples was studied by
transmission electron microscopy (TEM) using a
LEO912 AB OMEGA transmission electron micro-
scope (Zeiss, Germany) with an image resolution of
0.2–0.34 nm at an accelerating voltage of 60–120 kV.
The test sample as a suspension in ethanol or distilled
water was applied to a carbon film arranged on copper
microgauze. The particle size distribution was
obtained by a visual analysis of the micrographs. The
distributions found were described by the logarithmi-
cally normal distribution
An analysis of the results of DTA (Figs. 1a and 1b)
revealed a significant endothermic effect in a tempera-
ture range from 106 to 120°C for all of the test cata-
lysts, which was accompanied by the release of water
and a noticeable weight loss of the samples. At tem-
peratures of about 210–215°C, a strong exothermic
effect was noted with the concomitant formation of
CO2 and NO (Fig. 1b). At T > 300°C, the rate of
weight loss decreased; the release of CO2 and NO was
strongly slowed down (Fig. 1b), and the removal of
water completely stopped. The total weight loss of the
samples upon calcination at 450°C was ~30%. The
formation of CO2 and NO indicated the oxidation of
the surface of a carbon matrix with the possible forma-
tion of carboxyl groups. The thermal decomposition of
iron nitrate is a complex multistage process [17] with
the concomitant sequential removal of water and nitric
acid molecules. At an elevated temperature, the
hydroxide oxide Fe4O4(OH)4 was formed, which was
then converted into hematite Fe2O3. In the case of
potassium nitrate supported onto activated charcoal,
the release of CO2 and NO was observed at tempera-
tures of 170–200 and 295°C in the course of calcina-
tion in a flow of Ar. Note that the main weight loss and
an exothermic effect were also observed at these tem-
peratures. A noticeable exothermic effect at T =
120°C, accompanied by a weight loss and the forma-
tion of CO2 and NO, was detected only in the presence
of iron nitrate in the system. The sample weight loss,
the exothermic effect, and the two peaks of the release
of CO2 and NO strongly correlated with each other.
This circumstance allowed us to assume that the pro-
cess of the oxidation of an activated carbon carrier
with the formation of CO2 occurred in two tempera-
2




ln d − μ
2σ2
(
)
1
f (d) =
exp −
,

dσ 2π

and the average crystallite size was determined from
the formula
σ2
2


d = exp μ +
.




Catalytic Tests
The productivity of the catalysts in the Fischer–
Tropsch synthesis was tested in a steel fixed-bed
downflow reactor (internal diameter, 16 mm). A cata-
lyst portion of 200 mg was mixed with 1 cm3 of quartz
sand to prevent the formation of hot spots during the
synthesis of hydrocarbons, and this mixture was
loaded into the isothermal zone of the reactor. The
samples were activated at atmospheric pressure in a
flow of synthesis gas (CO : H2 = 1 : 2) at 400°C for 2 h;
thereafter, they were tested in the synthesis of hydro-
carbon at T = 240°C and P = 20 bar. The initial gas
composition was CO : H2 : Ar = 0.32 : 0.64 : 0.04, and
−1
the specific gas flow rate was 16 L
h–1. The cata- ture ranges and coincided with the appearance of NO.
gCat
KINETICS AND CATALYSIS
Vol. 59
No. 6
2018

EFFECT OF TOPOCHEMICAL PROCESSES
831
According to the results shown in Fig. 1, the main
(а)
DSC, mW/mg
part of the processes of thermolysis was complete at
T ≤ 300°C. However, an increase in magnetization due
to the formation of magnetite Fe3O4 could be observed
in the course of thermolysis in a magnetometer cell
[18]. Thus, the partial reduction of hematite to magne-
tite occurred in the course of thermolysis of iron
nitrate on a carbon substrate along with nitrate
decomposition.
2.0
exo
1.5
1.0
0.5
0
1
2
The elemental composition of the surface of the
catalysts differed significantly from the average bulk
composition (Table 1), mainly in terms of the iron
content. The difference between the surface concen-
tration of iron and the average bulk concentration was
13 wt % for all of the catalysts except for Fe/K/C
(11.5 wt %). In our opinion, this difference was due to
the partial hydrolytic decomposition of iron nitrate in
the mouths of the pores of carrier granules. In con-
trast, the potassium concentration in the catalysts
containing the promoter was slightly lower on the sur-
face than in the bulk. From the data given in Table 1,
it follows that the greatest amount of magnetite was
formed on calcination in the catalysts where iron was
first supported onto the carrier (63 and 81% of theo-
retically possible for Fe/C and K/Fe/C, respectively).
In the samples of Fe/K/C and (FeK)/C, the fractions
of magnetite were much lower: 48 and 31% of the cal-
culated values, respectively. The appearance of mag-
netite upon calcination indicates the partial reduction
of Fe3+ to Fe2+. In this case, only the carbon of the car-
rier, which is in direct contact with iron oxides, can be
a reducing agent. The preliminary application of
potassium markedly decreased the amount of magne-
tite. We believe that this was due to the inhibitory
effect of potassium on the process of hematite reduc-
tion to magnetite [19].
4
3
–0.5
0
100
200
300
400
500
600
T, °C
(b)
Ion current, А ×1010
1.6
1.4
1.2
CO
NO
1.0
0.8
0.6
0.4
0.2
O2
CO2
0
100
200 300 400 500 600 700
T, °C
Fig. 1. (a) Results of the differential thermal analysis of the
catalysts: (1) Fe(NO ) ⋅ 9H O/C, (2) KNO /C, (3)
3 3
2
3
(Fe(NO ) ⋅ 9H O + KNO )/C, and (4) (Fe(NO ) ⋅
3 3
2
3
3 3
9H O/K/C. (b) The mass spectrum of gases released in the
2
course of the DTA of the Fe/C catalyst.
An analysis of the results of transmission electron
microscopy showed a regular decrease in the average
particle size of iron oxides in the order Fe/C >
K/Fe/C > (FeK)/C > Fe/K/C, which was accompa-
nied by a decrease in the dispersion of the distribution.
Figure 2 shows the histograms of the particle size dis-
tributions of iron oxide. In our opinion, the decrease
ters for iron oxides and, correspondingly, to a decrease
in their size than that in the case of Fe/C and K/Fe/C
samples.
Figure 3 shows hysteresis loops for the test cata-
in the particle size of iron oxide (mainly Fe3O4) was lysts. The presence of a coercive force should be noted,
which is indicative of the occurrence of nonsuperpara-
magnetic particles or the agglomerates of superpara-
magnetic particles in the catalysts.
due to a decrease in the rate of hematite reduction in
the presence of potassium. Another reason could be
the interaction of the carrier surface with KNO3 sup-
ported during high-temperature processing. It is likely
that the decomposition of the nitrate ion occurred
simultaneously with the surface oxidation of activated
carbon with the formation of oxygen-containing
groups (hydroxyl, carboxyl, etc.) on the surface. The
Activation of Catalysts on Heating in a Flow
of Synthesis Gas
Figure 4 shows changes in the magnetization with
adsorption of Fe3+ ions on these groups was more temperature in the course of catalyst activation in a
intense than on the pure carrier; this led to the appear- flow of synthesis gas under the conditions of linear
ance of a larger number of the crystallite growth cen- heating. Its initial decrease up to 280–290°C was due
KINETICS AND CATALYSIS Vol. 59 No. 6 2018

832
Table 1. Surface composition and magnetite content of the initial catalysts
Surface composition, wt % Bulk composition, wt %
CHERNAVSKII et al.
Calculated
concentration
of Fe3O4**
Volume, wt %
Fe3O4*
Catalyst
Fe
K
Fe
16
K
Fe/C
29
26
20
27
–
0.05
1.06
0.9
14
14
5.6
5.9
22
K/Fe/C
Fe/K/C
(FeK)/С
1.0
0.7
1.7
12.5
8.5
14
17.2
11.7
19
1.85
* Calculated from the measurements of saturation magnetization.
** The amount of magnetite calculated on the condition that iron entirely occurred as magnetite. A dash indicates the absence of the
element.
to the temperature dependence of the magnetization next maximum indicated the presence of magnetite
of magnetite, which was initially present in the cata- and Hägg carbide χ-Fe5C2. Upon cooling from a tem-
lysts. However, in the course of heating, a nonmono-
tonic change in the magnetization was observed due to
competition between the two processes: the decrease
of magnetization with temperature and the occurrence
of chemical reactions accompanied by the formation
of new phases. The following two temperature ranges
in which the magnetization increased can be recog-
nized: the 300–350 and 420–550°C ranges for the
K/Fe/C and Fe/C catalysts and 270–370 and 370–
570°C for the (FeK)/C and Fe/K/C catalysts, respec-
tively. Thus, all of the samples exhibited similar behav-
iors: an increase in the magnetization with the subse-
quent decrease to a minimum value and then a growth
again. A further increase in the temperature led to a
drop in the magnetization to zero, which corre-
sponded to the Curie temperature of one of the mag-
netic phases. The cooling of the catalyst to room tem-
perature in a flow of synthesis gas made it possible to
obtain thermomagnetic curves from which the Curie
temperatures were found for the magnetic phases
formed as a result of activation. The Curie temperature
was determined as the as the intersection point of a
tangent to an extremum point of the thermomagnetic
curve with the x-axis.
perature corresponding to zero magnetization, the for-
mation of the χ-Fe5C2 carbide with a Curie tempera-
ture of 250°C was observed in all of the catalysts. Thus,
the temperature dependence of the magnetization
observed during activation can be described by the fol-
lowing sequence of chemical reactions:
3Fe2O3 + CO = 2Fe3O4 + CO2,
3Fe2O3 + H2 = 2Fe3O4 + H2O,
Fe3O4 + CO = 3FeO + CO2,
Fe3O4 + H2 = 3FeO + H2O,
5FeO + 9CO = Fe5C2 + 7CO2,
FeO + H2 = Fe + H2O,
5Fe + 4CO = Fe5C2 + 2CO2.
The amount of carbide formed can be determined
from the saturation magnetization. Table 2 shows the
carbide content (wt %) in the samples measured and
calculated under the condition of the complete con-
version of iron into carbide after the activation pro-
cess. The calculation was based on the amount of iron
according to ICP analysis data. However, if we used
the results of the analysis of surface composition, we
obtained other ratios between the calculated and
experimentally determined carbide concentrations
As follows from Fig. 5, all of the catalysts exhibited
the formation of Hägg carbide (Fe5C2) with a charac-
teristic Curie temperature of 250 5°C. Additional because magnetic measurements gave average bulk
experiments were carried out to explain the reason for values (Table 3). Table 3 summarizes the average par-
the nonmonotonic change in the magnetization ticle sizes of iron oxides obtained from TEM data and
during the activation of the samples. Heating was the calculated specific surface area of carbide parti-
interrupted at the extremum points of the curves, and cles. This surface area was calculated based on the
the catalyst was cooled to room temperature. The sub- assumption that the size of carbide particles was
sequent analysis of the thermomagnetic curves exactly the same as that of iron oxides, and the con-
obtained upon cooling showed that the first maximum centration of iron in the catalyst corresponded to its
corresponded to a magnetite phase, which was appar- concentration on the surface. The calculated specific
ently formed upon the partial reduction of hematite. surface area of carbide particles noticeably increased
Upon cooling from a minimum point, the magnetiza- markedly in the Fe/K/C and (FeK)/C catalysts, that
tion at room temperature was very small. Presumably, is, in the cases when potassium was supported onto the
the minimum point corresponded to the formation of carbon carrier preliminarily or simultaneously with
wustite, which is an antiferromagnetic material. The iron nitrate.
KINETICS AND CATALYSIS
Vol. 59
No. 6
2018

EFFECT OF TOPOCHEMICAL PROCESSES
833
Fraction of particles, arb. units
0.08
Magnetization, А × m2 × 10–3
25
20
15
10
K/Fe/AC
(а)
0.07
0.06
0.05
0.04
0.03
0.02
0.01
Fe/AC
Fe/K/AC
(FeK)/AC
5
0
–5
–10
–15
–20
–25
0
10
20
30
40
50
60
–7000 –5000 –3000 –1000 1000 3000 5000 7000
H, Oe
0.30
0.25
0.20
0.15
0.10
0.05
(b)
Fig. 3. Dependence of the magnetization on the field
intensity for the initial catalysts.
addition of potassium to the catalyst. Apparently, this
was due to both the intensification of the water gas
shift reaction to form CO2 and the promoting effect of
the alkaline additive on the formation of higher hydro-
carbons. In contrast, selectivity for methane changed
only slightly upon the addition of the promoter.
0
10
20
30
40
The most interesting result was the effect of the
sequence of introducing active components (iron and
potassium) into the catalysts on their specific activity
and selectivity. Thus, among the promoted catalysts,
the activity (expressed as the amount of CO converted
on 1 mol of Fe per second) was maximal for the
Fe/K/C catalyst or minimal for K/Fe/C, whereas the
(FeK)/C sample prepared by joint impregnation had
an intermediate value of this characteristic. All of the
three samples were more active than the Fe/C free of
the promoter (Fig. 6).
A comparison of data on the activity with the par-
ticle size distributions of magnetite in the samples
showed that a sharper distribution with a small average
particle size of magnetite was characteristic of the
most active catalysts. In contrast, samples with a wide
distribution and an increased average particle size
were less active. It is reasonable to assume that the par-
ticle size of the Fe5C2 carbide, which was formed at the
stage of catalyst activation as an active phase in the
synthesis of hydrocarbons, correlates with the size of
the initial magnetite particles. Thus, an increase in the
specific active surface area regularly leads to an
increase in the activity of the catalyst.
0.30
0.25
0.20
0.15
0.10
0.05
(c)
0
5
10 15 20 25 30 35
(d)
0.4
0.3
0.2
0.1
0
5
10
15
20
25
d, nm
Fig. 2. Histograms of the particle size distributions based
on the transmission electron microscopy data: (a) Fe/C,
(b) K/Fe/C, (c) (FeK)/C, and (d) Fe/K/C.
The catalysts obtained by sequential impregnation
exhibited minimal selectivity for C5+, whereas the
jointly impregnated sample showed almost the same
selectivity as that of the nonpromoted Fe/C. The addi-
tion of potassium led to an increase in the fraction of
light olefins, especially, on the catalysts obtained by
Tables 4 and 5 summarize the results of the cata-
lytic tests. The conversion of carbon monoxide and
selectivity for CO2 significantly increased with the
KINETICS AND CATALYSIS Vol. 59 No. 6 2018

834
Magnetization, А × m2 × 10–3
CHERNAVSKII et al.
Magnetization, А × m2 × 10–3
20
18
16
14
12
10
8
6
4
2
35
30
25
20
15
10
5
1
2
1
2
3
4
4
3
0
0
100
200
300
400
500
600
T, °C
25
75
125 175 225 275 325 375
T, °C
Fig. 4. Dependence of the magnetization on temperature
in the course of catalyst activation in synthesis gas:
(1) K/Fe/C, (2) Fe/C, (3) (FeK)/C, and (4) Fe/K/C.
Fig. 5. Dependence of the magnetization of the catalysts
on temperature after their activation in the synthesis gas:
(1) K/Fe/C, (2) Fe/C, (3) Fe/K/C, and (4) (FeK)/C.
sequential impregnation. On the contrary, the olefin
content of heavier products was almost independent of
the addition of potassium. The molecular weight dis-
tribution of synthesized hydrocarbons corresponded
to the Anderson–Schulz–Flory (ASF) formula [20].
A significant increase in the distribution parameter α
was observed upon the addition of potassium. There-
fore, heavier products were formed on the promoted
catalysts (Table 5).
The aqueous phase formed in the synthesis was
analyzed for the content of organic products. Only
C1–C4 alcohols were detected in it, and their amount
varied greatly depending on the catalyst preparation
method. The catalysts promoted with potassium
afforded a smaller amount of alcohols than that
obtained on Fe/C (Table 4). Alcohols almost exclu-
sively contained linear isomers, and only trace
Table 2. ron concentrations in the samples according to ICP data and the amount of carbide in the catalysts after treatment
in CO/H2, as determined experimentally and calculated from the ICP data
Fe5C2, wt %
theoretical* (ICP)
Fe5C2, wt %
experimental**
Degree of Fe
conversion into Fe5C2
Catalyst
Fe, wt % (ICP)
Fe/C
16
17
14
9.2
15
14
0.82
0.89
1
K/Fe/C
Fe/K/C
(FeK)/C
12.5
8.5
14
12.5
9.2
15
1
* Calculated based on the assumption of the complete conversion of Fe into Fe C carbide.
5
2
** The amount of Fe C carbide experimentally determined from the saturation magnetization data.
5
2
Table 3. The iron content of the near-surface layer of catalysts, the amount of carbide in the bulk of the samples, and the
sizes and specific surface areas of the carbide particles
Fe5C2, wt %
surface theoretical**
Fe5C2, wt %
experimental
Fe, wt %
surface*
Ssp, m2/g****
Catalyst
d, nm***
Fe/C
29
26
21
28
31.5
28
23
14
16
6
4
2.3
5.6
6.8
6
K/Fe/C
Fe/K/C
(FeK)/C
12.5
9.2
15
30.4
6
* Surface concentration of Fe based on the scanning electron microscopy data.
** The amount of Fe C carbide on the surface was calculated assuming the complete conversion of surface Fe into the Fe C carbide.
5
2
5 2
*** Average particle size of iron oxides according to the TEM data.
**** Calculated specific surface area of carbide particles on the catalyst surface (**).
KINETICS AND CATALYSIS
Vol. 59
No. 6
2018

EFFECT OF TOPOCHEMICAL PROCESSES
835
Table 4. Characteristics of the catalytic activity of samples in the Fischer–Tropsch synthesis*
Selectivity for carbon, mol %
Catalyst
Conversion of СО, %
СО2
5.1
13.7
18.8
19.0
hydrocarbons
alcohols
Fe/C
32.2
62.0
87.2
86.1
89.2
84.5
78.9
77
5.7
1.8
2.3
4.0
K/Fe/C
Fe/K/C
(FeK)/C
−1
–1
* Reaction conditions: T = 240°C, P = 20 bar, CO : H : Ar = 0.32 : 0.64 : 0.04, and specific gas flow rate of 16 L
h .
gCat
2
Table 5. Composition of the Fischer–Tropsch synthesis products *
ASF distribution
parameter
Distribution of hydrocarbons, mol % in terms of carbon
O/(O + P), % **
Catalyst
CH4
C2
C3
C4
C5+
C2–C4
C5+
αp ***
αa ****
Fe/C
9.3
9.2
9.7
7.9
5.2
5.9
5.8
4.4
3.2
8.7
9.6
5.4
4.6
4.2
4.6
4.7
77.8
72.1
70.3
77.5
36.6
50.0
51.2
37.3
16.2
16.2
13.0
15.4
0.787
0.852
0.830
0.836
–1
0.284
0.273
0.380
0.347
K/Fe/C
Fe/K/C
(FeK)/C
−1
* Conditions: T = 240°C, P = 20 bar, CO : H : Ar = 0.32 : 0.64 : 0.04, and specific gas flow rate of 16 L
h .
gCat
2
** The proportion of olefins in hydrocarbon fractions.
*** The Anderson–Schulz–Flory distribution parameter for C –C n-alkanes.
5
25
**** The Anderson–Schulz–Flory distribution parameter for C –C alcohols.
1
4
amounts of isopropanol were determined. The relative
amounts of alcohols corresponded to the ASF distri-
bution, but their values of the parameter α were much
lower than those of paraffins (Table 5). This may be
explained by different structures of the surface inter-
mediates involved in the growth of the hydrocarbon
chains of paraffins and alcohols.
CONCLUSIONS
We found that the sequence of the supporting of
active components onto potassium-promoted iron
catalysts affected a number of their physicochemical
and catalytic properties. The preliminary application
of potassium to a carbon carrier led to a significant
decrease in the size of the magnetite particles formed
upon the subsequent application of iron nitrate and
calcination. A difference between the iron content of
the surface of carrier granules and the average bulk
concentration of iron was found. We assumed that this
effect was due to the hydrolytic thermolysis of iron
nitrate in the surface layer of activated carbon granules
in the course of the catalyst synthesis. Potassium
inhibited the reduction of hematite to magnetite on
catalyst activation in CO/H2, which led to the forma-
tion of smaller magnetite particles. The formation of
Hägg carbide (Fe5C2) occurred in the course of cata-
lyst activation in an atmosphere of synthesis gas. We
found that the catalyst with a minimum carbide parti-
cle size and a maximum carbide concentration on the
surface of granules had the greatest activity in the
Fischer–Tropsch synthesis.
А × 103, (mol CO) (mol Fe)–1 s–1
35
30
25
20
15
10
5
0
Fe/K/C
(FeK)/C
K/Fe/C
Fe/C
Fig. 6. Dependence of the specific activity on the type of
catalysts.
KINETICS AND CATALYSIS Vol. 59 No. 6 2018

836
CHERNAVSKII et al.
11. Duan, X., Wang, D., Qian, G., Walmsley, J.C., Hol-
ACKNOWLEDGMENTS
This work was supported by the Russian Founda-
tion for Basic Research (grant no. 16-03-00215-a).
men, A., Chen, D., and Zhou, X., J. Energy Chem.,
2016, vol. 25, p. 311.
12. Wang, J., Zhang, Q., and Wang, Y., Catal. Today, 2011,
vol. 171, p. 257.
REFERENCES
13. Ledford, J.S., Houalla, M., Proctor, A., Hercules, D.M.,
and Petrakis, L., J. Phys. Chem., 1989, vol. 93, p. 6770.
1. Huber, G.W., Iborra, S., and Corma, A., Chem. Rev.,
2006, vol. 106, p. 4044.
14. Kumaran, G.M., Jacobs, G., Graham, U.M., Pen-
dyala, V.R.R., Martinelli, M., MacLennan, A., Hu, Y.,
and Davis, B.H., Appl. Catal., A, 2017, vol. 538, p. 190.
2. van Steen, E. and Claeys M., Chem. Eng. Technol.,
2008, vol. 31, p. 655.
3. Torres Galvis, H.M. and de Jong, K.P., ACS Catal.,
15. Chernavskii, P.A., Lunin, B.S., Pankina, G.V.,
Zakharyan, R.A., and Perov, N.S., Instrum. Exp. Tech.,
2014, vol. 57, no. 1, p. 78.
2013, vol. 3, p. 2130.
4. Snel, R., Catal. Rev. Sci. Eng., 1987, vol. 29, p. 361.
5. Torres Galvis, H.M., Bitter, J.H., Khare, C.B., Ruiten-
beek, M., Dugulan, A.I., and de Jong, K.P., Science,
2012, vol. 335, p. 835.
16. Selwood, P.W., Magnetochemistry, Interscience Pub-
lishing Company: N.-Y., 194, p. 287 p.
17. Melnikov, P., Nascimento, V.A., Arkhangelsky, I.V.,
Consolo, Z., and de Oliveira, L.C.S., J. Therm. Anal.
Calorim., 2014, vol. 115, p. 145.
18. Chernavskii, P.A., Kazak, V.O., Pankina, G.V.,
Strokova, N.E., and Perfil’ev, Yu.D., Kinet. Katal.,
2018, vol. 59, no. 2, p. 251.
19. Kazak, V.O., Pankina, G.V., Chernavskii, P.A., and
Lunin, V.V., Russ. J. Phys. Chem. A, 2017, vol. 91, no. 5,
p. 822.
6. Yang, Z.Q., Pan, X.L., Wang, J.H., and Bao, X.H.,
Catal. Today, 2012, vol. 186, p. 121.
7. Bezemer, G.L., Radstake, P.B., Falke, U., Oosterbeek, H.,
Kuipers, H.P.C.E., van Vandillen, A.J., and de Jong, K.P.,
J. Catal., 2006, vol. 237, p. 152.
8. Morales, F., Grandjean, D., de Groot, F.M.F.,
Stephan, O., and Weckhuysen, B.M., Phys. Chem.
Chem. Phys., 2005, vol. 7, p. 568.
9. Ribeiro, M.C., Jacobs, G., Davis, B.H., Cronauer, D.C.,
Kropf, A.J., and Marshall, C.L., J. Phys. Chem. C,
2010, vol. 114, p. 7895.
20. Friedel, R.A. and Anderson, R.B., J. Am. Chem. Soc.,
1950, vol. 72, p. 1212.
10. Luo, M.S., O’Brien, R.J., Bao, S.Q., and Davis, B.H.,
Translated by V. Makhlyarchuk
Appl. Catal., A, 2003, vol. 239, p. 111.
KINETICS AND CATALYSIS
Vol. 59
No. 6
2018