Highly selective production of olefins from syngas with modified ASF distribution model

Article abstract of DOI:10.1016/j.apcata.2018.07.006

The Fischer-Tropsch to olefins (FTO) process is a promising alternative non-petroleum route to produce value-added olefins. In this study, Co₂C nanoprisms formed via the CoMn spinel structure were used as the active phase for the FTO reaction and the product distribution was thoroughly investigated. A break from the traditional ASF model was found with much lower methane selectivity than that predicted by the ideal ASF law. In addition, the as-obtained hydrocarbons were mainly concentrated in the range of C₂ to C₁₂ with the C_2-12 selectivity as high as 91% while the C₁₃₊ selectivity as low as 5%, indicating a much narrower distribution, which was rarely observed in the FT reaction. Moreover, ～86% of the hydrocarbon products were olefins with very high olefin to paraffin ratios, suggesting a promising route for the selective production of olefins directly from syngas. A modified ASF distribution model was proposed with three different chain growth probabilities to rationalize the non-ASF phenomenon.

Full text of DOI:10.1016/j.apcata.2018.07.006

Accepted Manuscript
Title: Highly Selective Production of Olefins from Syngas
with Modified ASF Distribution Model
Authors: Fei Yu, Tiejun Lin, Xinxing Wang, Shenggang Li,
Yongwu Lu, Hui Wang, Liangshu Zhong, Yuhan Sun
PII:
S0926-860X(18)30329-6
DOI:
Reference:
https://doi.org/10.1016/j.apcata.2018.07.006
APCATA 16732
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
24-4-2018
15-6-2018
3-7-2018
Please cite this article as: Yu F, Lin T, Wang X, Li S, Lu Y, Wang
H, Zhong L, Sun Y, Highly Selective Production of Olefins from Syngas
with Modified ASF Distribution Model, Applied Catalysis A, General (2018),
https://doi.org/10.1016/j.apcata.2018.07.006
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.

Highly Selective Production of Olefins from Syngas with Modified
ASF Distribution Model
Fei Yu,1,2 Tiejun Lin,1 Xinxing Wang,1,2 Shenggang Li,1,3 Yongwu Lu,1 Hui Wang,1 Liangshu
Zhong,1,3,* and Yuhan Sun1,3,
*
1 CAS Key Laboratory of Low-Carbon Conversion Science and Engineering, Shanghai
Advanced Research Institute, Chinese Academy of Sciences, Shanghai 201203, PR China
2 University of the Chinese Academy of Sciences, Beijing 100049, PR China
3 School of Physical Science and Technology, ShanghaiTech University, Shanghai 201203,
PR China.
*Email for L. Z.: zhongls@sari.ac.cn.
*Email for Y. S.: sunyh@sari.ac.cn.
Graphical Abstract
Co2C nanoprisms benefited for highly selective production of olefins from syngas with
modified ASF distribution model.
Highlights

1. A negative deviation of CH4 selectivity from the ideal ASF model was found over Co2C
nanoprisms.


2. The products featured a very narrow carbon distribution.
3. About 91wt% of the hydrocarbons were in the range of C2 to C12.



4. The overall olefins selectivity was as high as 86%.
5. A modified ASF distribution model with three chain growth probabilities was
proposed.
ABSTRACT: The Fischer-Tropsch to olefins (FTO) process is a promising alternative non-
petroleum route to produce value-added olefins. In this study, Co2C nanoprisms formed via
the CoMn spinel structure were used as the active phase for the FTO reaction and the product
distribution was thoroughly investigated. A break from the traditional ASF model was found
with much lower methane selectivity than that predicted by the ideal ASF law. In addition, the
as-obtained hydrocarbons were mainly concentrated in the range of C2 to C12 with the C2-12
selectivity as high as 91% while the C13+ selectivity as low as 5%, indicating a much narrower
distribution, which was rarely observed in the FT reaction. Moreover, ~86% of the
hydrocarbon products were olefins with very high olefin to paraffin ratios, suggesting a
promising route for the selective production of olefins directly from syngas. A modified ASF
distribution model was proposed with three different chain growth probabilities to rationalize
the non-ASF phenomenon.
1. Introduction
Keywords:cobalt carbide; CO hydrogenation; Fischer‐Tropsch; Fischer‐Tropsch to olefins;
modified ASF distribution.
Olefins are vital basic chemicals, and are extensively used in the chemical industry.
Lower olefins (C2-4=) are the most important monomers to produce plastics, synthetic rubbers,
and synthetic fibers [1,2]. Long-chain α-olefins (C5+=) with a linear carbon chain and with the
double bond positioned at the terminal position are important raw materials for producing
value-added products, such as lubricants and higher alcohols [3-5]. Traditionally, lower

olefins are produced from thermal or catalytic cracking of a range of hydrocarbon feedstocks,
such as naphtha, gas oil, condensates and light alkanes [6], whereas most of the commercial
α-olefins are produced via the oligomerization of ethylene. With increasing concerns on the
availability of petroleum reserves, alternative routes for olefins production starting with
syngas derived from natural gas, coal or renewable biomass are highly desirable [7,8].
Fischer-Tropsch (FT) synthesis is an important alternative process to produce fuels and
value-added chemicals from syngas [9-12]. With the expected diminution of the petroleum
reserves, the Fischer−Tropsch to olefin (FTO) process is becoming increasingly attractive due
to its simplified operation and low energy consumption. Traditional Fe-based FT catalysts are
usually chosen to produce olefins after some modification [13-16]. The Na- and S-promoted
Fe-based catalysts supported by inert materials discovered by Torres Galvis et al. exhibited a
selectivity to lower olefins up to 60 wt% [17]. They suggested that Na could reduce methane
selectivity by increasing the chain growth probability, while S might reduce the hydrogen
coverage of the catalyst leading to even lower methane selectivity and higher olefins content
[18]. The Zn- and Na-modulated Fe5C2 catalyst reported by Zhai et al. was very active to
produce olefins with the selectivity to olefins up to 79%, attributed to the effect of Na in
suppressing the hydrogenation of double bonds and promoting the desorption of olefins [19].
Gao et al. found that the addition of polar liquid could effectively increase the α-olefin
selectivity on Fe-based catalysts [20]. They suggested that the addition of polar liquid
promoted the separation of wax products from the catalyst particles and thus suppressed α-
olefin secondary reactions [21]. Linghu et al. found that the addition of n-paraffins to the Co-
based FT reaction system could remarkably enhance the selectivity to α-olefins [22-24].

Based on the polymerization mechanism, the FT products generally follow the classical
Anderson-Schulz-Flory (ASF) distribution with a wide carbon number range. Consequently, a
lower value of α will result in higher selectivity to undesired methane. In addition, the olefins
produced can be easily hydrogenated into saturated paraffins during the FT reaction. The
olefin/paraffin (O/P) ratio will decrease due to the higher secondary hydrogenation ability of
long-chain olefins with the increasing carbon number. Therefore, high O/P values are needed
for all hydrocarbons with different carbon numbers to maximize the olefins selectivity.
Meanwhile, the content of methane should be kept as low as possible.
Recently, we found that Co2C nanoprisms with exposed {101} and {020} facets could
effectively convert syngas to lower olefins (61%C) with relatively low methane (5%C) at
mild reaction conditions [25]. The Na and Mn promoters played vital roles in the formation of
the Co2C nanoprisms [26,27]. The addition of Na, which acted as an electronic donor to
cobalt, resulted in stronger CO adsorption and enhanced CO dissociation, which also
benefited the formation and stabilization of the Co2C nanoprisms and the transition of the
catalytic performance from traditional FTS to FTO [26]. In addition, Mn not only acted as a
typical electronic or structural promoter, but also had a strong effect on the morphology, as it
resulted in the formation of Co2C nanoprisms with specifically exposed facets of (101) and
(020) [27]. However, in our previous studies, we mainly focused on the formation of lower
olefins. As the O/P ratios for C2-4 were as high as 30, we speculate that the O/P ratio for C5+
should also be rather high. The detailed product distribution for C5+ has not been thoroughly
investigated. In addition, a negative deviation in methane selectivity has not been previously
reported, which did occur on Co2C nanoprisms due to the facet effect in our previous work

[25-27]. As the obtained products deviated greatly from the classical ASF distribution, a new
mathematical model is needed to describe the special product distribution. In this study, the
CoMn catalyst was selected to perform the FTO reaction, and the total product distribution
was investigated in great detail. The effects of different reaction conditions, such as the
H2/CO ratio, reaction temperature and pressure, on the product distribution were also studied.
The products featured a very narrow carbon distribution, and ~91wt% of the hydrocarbons
were in the range of C2 to C12. Almost 86 wt% of the hydrocarbons were olefins with
relatively high O/P ratios. Thus, the product distribution breaks the traditional ASF model
with the selectivity to methane or C2 hydrocarbons much lower than that predicted from the
ASF law. A modified ASF distribution model with three different chain growth probabilities
was proposed to rationalize this new non-ASF distribution phenomenon.
2. Experimental
Catalyst Preparation
The CoMn catalyst with the spinel structure (CoxMn3-xO4) was prepared by co-
precipitation. Typically, a 2M mixed salt solution (Co/Mn (mol/mol) = 2/1) was prepared
with cobalt nitrate (Co(NO3)2·6H2O, Sinopharm Chemical Reagent Co., Ltd) and manganese
nitrate (50 wt% Mn(NO3)2, aq, Sinopharm Chemical Reagent Co., Ltd). The precipitant was a
2M alkali solution, which was prepared by dissolving sodium carbonate anhydrous (Na2CO3,
Sinopharm Chemical Reagent Co., Ltd) in deionized water. The above-mentioned salt and
alkali solution were simultaneously added into a beaker with 100 ml pre-adding deionized
water dropwise under continuous stirring. After aging for 2 h at 30 °C, the obtained

suspension was centrifuged and washed with deionized water for 7 times, dried at 100 °C for
10 h. Then the samples were further calcined in a muffle furnace at 330 °C for 3 h under static
air.
Catalyst characterization
X-ray powder diffraction test were performed on a Rigaku Ultima IV X-ray powder
diffractometer using Cu Kα radiation with the wavelength of 1.54056 Å at 40 kV and 40 mA.
Phase characterizations were realized from powder patterns using the PDF4-2015 database.
The standard PDF card used in the qualitative analysis were: MnCO3 (PDF#01-086-0173),
CoCO3 (PDF#01-78-0209), MnCo2O4.5 (PDF#00-032-0297), MnO (PDF#01-075-0625), CoO
(PDF#01-071-1178), Co (PDF#01-089-7093) and Co2C (PDF#03-065-1457). The XRD
pattern were recorded ex-situ, and XRD measurements on the catalysts were performed
immediately after they were taken out of the reactor.
Transmission electron microscopy (TEM) measurements were performed on a JEOL JEM
2011 electron microscope with 200 kV accelerating voltage. The sample number for particle
size distribution was over 300. Energy dispersive X-ray spectroscopy (EDX) was performed
to locate elemental distribution of Co, Mn and O with an SDD-type EDX detector on an
ARM-200CF (JEOL, Tokyo, Japan) transmission electron microscope operated at 200 keV
and equipped with double spherical aberration (Cs) correctors. The attainable energy-
resolution of the EDX detector is 130 eV. Spatial drift was corrected with a simultaneous
image collector.
Catalytic evaluation

Syngas conversion was carried out in a fixed-bed reactor with an inner diameter of 9 mm.
Generally, 1.5 g catalyst of 40-60 mesh mixed with 2 ml silica sand was loaded into a
stainless steel tube reactor. The catalyst was reduced in situ in 10% H2/N2 (v/v) at atmosphere
pressure at 300 °C for 5 h. Then the temperature was lowered to selected value (250 °C, 260
°C or 270 °C) in He (99.999%) flow for 30 min to purge the residual reduction gas.
Subsequently, the syngas with different H2/CO (v/v) ratio (H2/CO = 2, 1, 0.5) was introduced
to the reactor as the feed gas. 3 vol% N2 was mixed in the syngas as internal standard.
The outlet gas passed through a hot trap (120 °C) and a cold trap (0 °C) before analyzing
online by gas chromatography (GC). Atlow pressure (1 bar), we could notobtain any products
in the hot or cold trap. In order to obtain the detailed information on heavy hydrocarbons and
oxygenates, the tail gas was passed through three absorption bottles in series using water and
toluene as the absorbents for ca. 24 h, and then was analyzed by online GC. The medium and
high pressure test was performed under the pressure of 3 bar~10 bar. Catalytic activity and
product selectivity at the steady state were analyzed by offline GC of the products up to C45
(oil products and wax products).
The outlet gas was analyzed by an Agilent 7890B GC with TCD and FID. TheTCD with a
carbon molecular sieves column (TDX-1) using He as the carrier gas was used to analyze H2,
N2, CO, CH4, and CO2. The hydrocarbons with the carbon number from C1 to C8 were
analyzed through a KCl-modified alumina capillary column (19095P-K25) with Ar carrier on
FID. N2 was added as the internal standard, as its amount remained constant during the
reaction. Thus, CO conversion was calculated according to the amount of N2. The liquid and
solid products were analyzed by three additional offline GCs. The aqueous products were

analyzed using two Porapak Q columns furnished with a FID (for C1-C5 oxygenates
recognition) and TCD (for H2O and MeOH recognition), respectively. The liquid oil products
were analyzed through a HP-1 column with N2 carrier by FID. The mass balance, carbon
balance and oxygen balance calculated at each test were between 97% and 101%.
CO conversion was calculated on a carbon atom basis with the formula of:
ꢀꢁꢊꢋꢌꢍꢎ ꢏ ꢀꢁꢐꢑꢎꢌꢍꢎ
ꢀꢁ ꢀꢂꢃꢄꢅꢆꢇꢈꢂꢃ ꢉ
ꢒ 100%
ꢀꢁꢊꢋꢌꢍꢎ
where COinlet and COoutlet represent moles of CO at the inlet and outlet, respectively.
CO2 selectivity was calculated according to:
ꢀꢁꢓ ꢐꢑꢎꢌꢍꢎ
ꢀꢁꢓ ꢔꢅꢕꢅꢖꢗꢈꢄꢈꢗꢘ ꢉ
ꢒ 100%
ꢀꢁꢊꢋꢌꢍꢎ ꢏ ꢀꢁꢐꢑꢎꢌꢍꢎ
where CO2 outlet denotes moles of CO2 at the outlet.
The selectivity of the individual product (hydrocarbon or oxygenate) CnHm based on CO2-
free was obtained according to:
ꢃꢀꢋꢙꢚ ꢐꢑꢎꢌꢍꢎ
ꢀꢋꢙꢚ ꢔꢅꢕꢅꢖꢗꢈꢄꢈꢗꢘ ꢉ
ꢒ 100%
ꢀꢁꢊꢋꢌꢍꢎ ꢏ ꢀꢁꢐꢑꢎꢌꢍꢎ ꢏ ꢀꢁꢓ ꢐꢑꢎꢌꢍꢎ
where CnHm outlet represents moles C of the individual product (hydrocarbon or oxygenate) at
the outlet.
3. Modified ASF distribution model
In order to describe the deviation of C1 and C2 components from the predicted line by the
ideal ASF model, a modified distribution model is needed. The ASF distribution curve of the
practical FTO reaction indicates that ln(Wn/n) vs. n (carbon number) for C3+ could well be

fitted to a straight line. Thus, we assume that the chain growth probabilities for C3+ are still
independent of the carbon number, so a single chain growth probability was used to describe
the ASF distribution of C3+. However, for the C1 and C2 component, the chain growth
probabilities are related with the carbon number. This leads to three distinct chain growth
probabilities, α1 for C1 to C2, α2 for C2 to C3 and α3 for the growth of longer chains, for the
description of the modified ASF distribution model.
Scheme 1 illustrates the reaction scheme of the selected chain length n with two parallel
competing reactions, the chain growth and termination. Generally, the chain will be
terminated in the form of paraffin, olefin or oxygenate with carbon chain length of n. To
simplify the deduction, Cn is adopted to represent all the type of chain termination products.
When the reaction of the selected section reaches stability, the steady-state conditions can be
written as:
ꢆꢛ,ꢋꢜꢝ ꢉ ꢆꢛ,ꢋ ꢞ ꢆꢟ,ꢋ
(1)
here ꢆꢛ,ꢋ is the rate of chain growth and rT,n the rate of chain termination. The chain growth
probability α is defined as the ratio of chain growth rate to the total turnover rate:
ꢆꢛ,ꢋ
ꢠꢋ ꢉ
(2)
ꢆꢛ,ꢋ ꢞ ꢆꢟ,ꢋ
Note that the product distribution is involved with the chain termination rate, and the relation
of the consecutive chain termination rate can be calculated as:
ꢆꢟ,ꢋ
1 ꢏ ꢠꢋ
ꢉ
∙ ꢠꢋꢜꢝ
(3)
ꢆꢟ,ꢋꢜꢝ 1 ꢏ ꢠꢋꢜꢝ

The differential form of product Cn with carbon number n can be written as:
ꢢ
ꢣ
ꢡ ꢀꢋ /ꢡꢗ
1 ꢏ ꢠꢋ
(4)
ꢉ
∙ ꢠꢋꢜꢝ
ꢢ
ꢣ
ꢡ ꢀꢋꢜꢝ /ꢡꢗ 1 ꢏ ꢠꢋꢜꢝ
We define d[C0]/dt as the total number of product molecules formed, combined with the
former hypothesis of distinct α1, α2 and α3 values, the following formula can be obtained:
ꢢ
ꢣ
ꢢ
ꢣ
ꢡ ꢀꢝ
ꢡ ꢀꢤ
ꢉ
∙ ꢥ1 ꢏ ꢠꢝꢦ
ꢡꢗ
ꢡꢗ
ꢢ
ꢣ
ꢣ
ꢢ ꢣ
ꢡ ꢀꢓ
ꢡ ꢀꢝ 1 ꢏ ꢠꢓ
ꢉ
∙
∙ ꢠꢝ
∙ ꢠꢧ
ꢡꢗ
ꢡꢗ 1 ꢏ ꢠꢝ
ꢢ
ꢢ
ꢣ
ꢡ ꢀꢧ
ꢡ ꢀꢓ 1 ꢏ ꢠꢧ
(5)
ꢉ
∙
ꢡꢗ
ꢡꢗ 1 ꢏ ꢠꢓ
ꢢ
ꢣ
ꢢ
ꢣ
ꢡ ꢀꢨ
ꢡ ꢀꢧ
ꢉ
∙ ꢠꢧ
ꢡꢗ
ꢡꢗ
⋯
ꢢ
ꢣ
ꢢ
ꢣ
ꢡ ꢀꢋ
ꢡ ꢀꢋꢜꢝ
ꢉ
∙ ꢠꢋꢜꢝ
ꢡꢗ
ꢡꢗ
And the molar based distribution is given by:
ꢢ
ꢣ
ꢡ ꢀꢝ /ꢡꢗ
ꢘꢝ ꢉ
ꢉ ꢥ1 ꢏ ꢠꢝꢦ
ꢢ
ꢣ
ꢡ ꢀꢤ /ꢡꢗ
ꢢ
ꢣ
d Cꢓ /dꢗ
ꢘꢓ ꢉ
ꢉ ꢠꢝ ∙ ꢥ1 ꢏ ꢠꢓꢦ
(6)
ꢢ
ꢣ
d Cꢤ /dꢗ
ꢢ
ꢣ
d Cꢩ /dꢗ
ꢘꢋ ꢉ
ꢉ ꢠꢝ ∙ ꢠꢓ ∙ ꢠꢧꢋꢜꢧ ∙ ꢥ1 ꢏ ꢠꢧꢦ
ꢢ
ꢣ
d Cꢤ /dꢗ

where yn is the molar fraction of product Cn. In order to obtain the distribution based on the
mass fraction, Eq. (6) needs to be transformed by the following equation:
ꢃ ∙ ꢘꢋ
ꢪ ꢉ
(7)
ꢋ
ꢫ
∑
ꢊꢬꢝ ꢃ ∙ ꢘꢋ
The mass fraction based on the separated chain growth probabilities can be calculated from
Eq. (6) and Eq. (7):
ꢥ
ꢦꢥ ꢦ
1 ꢏ ꢠꢝ 1 ꢏ ꢠꢧ
ꢪ ꢉ
ꢝ
1 ꢞ ꢠꢝ ꢞ ꢠꢝꢠꢓ ꢏ ꢠꢧ ꢏ ꢠꢝꢠꢧ
ꢥ
ꢦꢥ ꢦ
ꢪ
ꢠꢝ 1 ꢏ ꢠꢓ 1 ꢏ ꢠꢧ
ꢓ
(8)
ꢉ
2
1 ꢞ ꢠꢝ ꢞ ꢠꢝꢠꢓ ꢏ ꢠꢧ ꢏ ꢠꢝꢠꢧ
ꢋꢜꢧꢥ
ꢦꢓ
1 ꢞ ꢠꢝ ꢞ ꢠꢝꢠꢓ ꢏ ꢠꢧ ꢏ ꢠꢝꢠꢧ
ꢪ
ꢠꢝꢠꢓꢠꢧ 1 ꢏ ꢠꢧ
ꢋ
ꢉ
ꢥꢃ ꢭ 3ꢦ
ꢃ
The chain growth probabilities, α1, α2 and α3, would be obtained by separated steps. First, α3is
calculated from the slope of ln(Wn/n) vs. n fitted from the products with carbon number of>3.
The values of W1 and W2are obtained from the actual reaction data, so α1 and α2 can be
calculated by Eq. (8).
Result and discussion
Figure 1a shows the XRD results of the CoMn catalyst at different stages. The precursor
was found to consist of CoMn composite carbonate. The MnCo2O4.5 spinel structure was
formed after calcination. CoxMn1-xO with lower metal valence was observed after reduction at
300 °C. In addition, a small amount of FCC Co was also detected. CoxMn1-xO, MnO, and
Co2C coexisted in the spent catalyst, suggesting Co2C to form during the reaction process due

to the residual Na [26,27]. In addition, the diffraction peaks of CoxMn1-xO shifted to lower
diffraction angles (Figure 2), which might result from the decrease of the Co to Mn ratio in
the CoxMn1-xO for the spent sample.
The TEM results of various samples at the different stages were displayed in Figure 1b-i.
The CoMn precursor showed severe agglomeration with the average size of the nanoparticles
of ca. 6.7 nm (Figure 3). The corresponding HRTEM image presented a lattice distance of the
CoMn composite carbonate. After calcination, the particle size slightly increased to 6.9 nm
and the phase turned into the MnCo2O4.5 spinel. The particle size grew to 12.1 nm after
reduction, and MnCo2O4.5 was reduced to the low valence composite oxide with a lattice
distance of 2.13 Å (Figure 1h). Co, Mn and O were uniformly distributed according to the
EDX mapping of the reduced catalyst (Figure 4), further confirming that Co and Mn had
close contact and existed in the form of a composite oxide. The TEM images of the spent
catalyst displayed in Figure 1e and 1i suggested the formation of Co2C nanoprisms, which
have been reported as the active phase for the FTO reaction [25]. Figure 5 indicates the
nanoparticles with spherical or irregular shape around the Co2C nanoprisms were manganese
oxides.
The catalytic performance of the CoMn catalyst under different reaction conditions is
shown in Table 1 and Table 2. The catalyst displayed severe WGS activity with the CO2
selectivity up to 45-50 C%, showing it to be more appropriate for syngas with low H2/CO
ratio derived from coal or biomass. CO conversion decreased sharply and the O/P ratio
increased significantly with the decreasing H2/CO ratio in the feed gas. On the other hand,
thechains in the products became longer, and the C2-4= selectivity decreased with the

decreasing H2/CO ratio of syngas. Owing to the influence of the increasing O/P ratio and
chain growth probability, the selectivity to C2-4o decreased with decreasing H2/CO ratio. In
addition, the oxygenates selectivity increased with the decreasing H2/CO ratio.
The effect of pressure and temperature on CO conversion and product selectivity was also
studied. When the pressure increased from 1 bar to 10 bar with the feed gas of H2/CO=0.5 at
250 °C, CO conversion and CH4 selectivity increased to 28.5% from 6.3% and 5.5% from
2.4%, respectively. Owing to the enhanced hydrogenation ability at higher pressure, the O/P
ratio dropped with increasing pressure, while the C3 component maintained the highest O/P
ratio among all the hydrocarbons. In addition, the O/P ratio dropped dramatically with
increasing carbon chain length at 1 bar, and it decreased slightly with the carbon chain length
at higher pressure. The increase in the pressure also resulted in the enhancement of the carbon
chain growth, which further led to decreasing C2-4= selectivity. For compensation, the content
of paraffins and C5+ component increased at higher pressure.
Both CO conversion and CH4 selectivity increased with increasing reaction temperature.
In addition, the products shift to lower hydrocarbons and the C5+ selectivity decreased
gradually as the reaction temperature increased. However, there was no obvious change in the
C2-4= selectivity and O/P ratio at different reaction temperatures.
The hydrocarbon distribution as a function of the carbon number displayed in Figure 6
suggested significant deviation from the theoretical ASF model, especially for the C1 and C2
components. The blue dash line was the ideal ASF line according to the chain growth
probability obtained by the component with the carbon number of >2. C1 and C2 dropped

dramatically from the ideal predicted line, especially for the tests at lower temperature,
pressure and H2/CO ratio of syngas.
Figure 7a shows the hydrocarbon selectivity as a function of the carbon number, which
clearly displays the suppression of CH4 formation, and the distribution of hydrocarbons in a
narrow range with high selectivity to olefins. CH4 selectivity was no more than 5 wt% though
CH4 was the maximum for all the paraffins. CH4 and C13+ only accounted for 9.2 wt%, and
the products mainly consisted of C2 to C12. The C2-12= selectivity was up to 83 wt% and the
overall olefins amounted to 86 wt%. According to the classical ASF model, the chain growth
probability would be 0.78 when CH4 selectivity was 5 wt%, and C13+ selectivity would be
17.8 wt%, much higher than the actual selectivities.
For mostreported Fischer-Tropsch catalysts, the CH4 selectivity always follows the ASF
mode or exhibits a positive deviation [28-34]. However, according to our experimental
observation on the great deviation from the theoretical ASF model, we suggest that the chain
growth probability is not constant with respect to the carbon number, especially for the C1 and
C2 components. Therefore, we propose a modified distribution model based on the
independent chain growth probability of α1 for C1 to C2, α2 for C2 to C3 and α3 for the growth
of longer chains (Scheme 2). It is assumed that the chain growth probability is constant when
the carbon number is >3. The modified distribution based on the mass fraction and the carbon
number is shown in Figure 7b.

The separated chain growth probabilities α1 and α2 were then calculated by the modified
distribution model based on the W1, W2, and α3, obtaining from the experimental data as
displayed in Figure 6 and Table 3. For all the product distributions of the selected reaction
conditions, both the C1 and C2 components deviated from the predicted line from the ideal
ASF model. Based on our previous study [25], the effective barrier for methane formation
from CH2 and H on Co2C (020) surface was the highest of all the selected facet, so the Co2C
nanoprisms effectively suppressed methane formation. As for the C2 component, re-
absorption of olefins would be the main reason for the phenomenon that the value for the C2
component was much lower than the prediction [28,35-39]. In addition, the O/P value of C2
was much lower than those of C3 and C4 components further confirmed the above-mentioned
viewpoint.
The effect of the different reaction conditions on the degree of deviation from the typical
ASF model was also investigated. Due to the deviation for the C1 and C2 components, both α1
and α2 are higher than α3. Under the selected reaction conditions, except for 250 °C, 10 bar
and the feed gas of H2/CO=0.5, α1 is always higher than α2, indicating the degree of deviation
for the C1 component to be more severe than that for the C2 component. When the H2/CO
ratio decreased (Table 3, Line 1-3), all three chain growth probabilities increased. As α1
increased more remarkably, the deviation for the C1 component of the typical ASF model was
the maximum (Figure 6a-c). In addition, the influence of the pressure was also more
remarkable on α1. As shown in Table 3 (Line 3-6), α1 decreased sharply with increasing
pressure, while α2 remained almost unchanged. For the effect of reaction temperature shown

in Table 3, (Line 1, 7 and 8), all three chain growth probabilities are sensitive to reaction
temperature and higher reaction temperature leads to lower chain growth probability.
Conclusion
In summary, Co2C nanoprisms formed via the CoMn spinel structure were studied for the
Fischer-Tropsch to olefins reaction, and the product distribution was investigated in detail. A
non-ASF distribution was found, which contained much lower C1 and C2 content than that
predicted by the ideal ASF law. The product distribution also featured a narrower carbon
distribution, with ca. 91 wt% of the hydrocarbons in the range of C2 to C12 and almost 83 wt%
of them being olefins. Meanwhile, the CH4 selectivity was as low as 4.8 wt% and the C13+
selectivity was less than 5 wt%. A modified ASF distribution model with three different chain
growth probabilities was proposed to rationalize the non-ASF phenomenon. Based on the
modified ASF model, the chain growth probabilities were calculated, and the values of α1 and
α2 were much higher than that of α3, suggesting that the C1 and C2 surface intermediates
prefers participating the chain growth rather than forming methane or C2 species. The
influence of the reaction condition was also investigated. It was found that low reaction
temperature, low pressure, and high concentration of CO in the feed gas were beneficial in
suppressing methane formation with notable deviation from the ideal ASF distribution.
However, an appropriate increase in the reaction pressure could shift the products to longer
carbon chains and produce more linear α-olefins. The suppression to CH4 formation, the
narrower carbon distribution of hydrocarbons and the higher selectivity to olefins further

show that the Co2C nanoprisms are promising FTO catalysts for producing not only lower
olefins but also long-chain α-olefins.
Notes
The authors declare no competing financial interest.
Acknowledgment
This work has been supported by the National Key R&D Program of China
(2017YFB0602202), Natural Science Foundation of China (21573271, 91545112), and the
Chinese Academy of Sciences (XDA21020600, QYZDB-SSW-SLH035).
References
[1] H. M. Torres Galvis, K. P. de Jong, ACS Catal. 3 (2013) 2130-2149.
[2] J. J. H. B. Sattler, J. Ruiz-Martinez, E. Santillan-Jimenez, M. Weckhuysen, Chem.
Rev. 114 (2014) 106.
[3] M. E. Dry, Catal. Today 71 (2002) 227-241.
[4] W. Keim, Angew. Chem. Int. Ed. 52 (2013) 12492-12496.
[5] R. Franke, D. Selent, A. Börner, Chem. Rev. 112 (2012) 5675-5732.
[6] Y. An, T. Lin, F. Yu, Y. Yang, L. Zhong, M. Wu, Y. Sun, Sci. China. Chem. 60
(2017) 887-903.
[7] F. Jiao, J. Li, X. Pan, J. Xiao, H. Li, H. Ma, M. Wei, Y. Pan, Z. Zhou, M. Li, S. Miao,
J. Li, Y. Zhu, D. Xiao, T. He, J. Yang, F. Qi, Q. Fu, X. Bao, Science 351 (2016) 1065-1068.

[8] K. Cheng, B. Gu, X. Liu, J. Kang, Q. Zhang, Y. Wang, Angew. Chem. Int. Ed. 55
(2016) 1-5.
[9] E. de Smit, B. M. Weckhuysen, Chem. Soc. Rev. 37 (2008) 2758-2781
[10] Q. Zhang, J. Kang, Y. Wang, ChemCatChem 2 (2010) 1030-1058
[11] A. Y. Khodakov, W. Chu, P. Fongarland, Chem. Rev. 107 (2007) 1692-1744
[12] Q. Zhang, K. Cheng, J. Kang, W. Deng, Y. Wang, ChemSusChem, 7 (2014) 1251–
1264.
[13] 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.
[14] Y. Liu, J. Chen, J. Bao, Y. Zhang, ACS Catal. 5 (2015) 3905-3909.
[15] X. Chen, D. Deng, X. Pan, Y. Hu, X. Bao, Chem. Commun. 51 (2015) 217-220.
[16] H. M. Torres Galvis, A. C. J. Koeken, J. H. Bitter, T. Davidian, M. Ruitenbeek, A. I.
Dugulan, K. P. de Jong, Catal. Today 215 (2013) 95-102.
[17] H. M. Torres Galvis, J. H. Bitter, C. B. Khare, M. Ruitenbeek, A. I. Dugulan, K. P.
de Jong, Science 335 (2012) 835-838.
[18] H. M. Torres Galvis, A. C. J. Koeken, J. H. Bitter, T. Davidian, M. Ruitenbeek, A. I.
Dugulan, K. P. de Jong, J Catal. 303 (2013) 22-30.
[19] P. Zhai, C. Xu, R. Gao, X. Liu, M. Li, W. Li, X. Fu, C. Jia, J. Xie, M. Zhao, X.
Wang, Y. Li, Q. Zhang, X. Wen, D. Ma, Angew. Chem. Int. Ed. 55 (2016) 9902–9907.

[20] J. Gao, B. Wu, L. Zhou, Y. Yang, X. Hao, J. Xu, Y. Xu, L. Cao, Y. Li, Catal.
Commun. 12 (2011) 1466-1470.
[21] J. Gao, B. Wu, L. Zhou, Y. Yang, X. Hao, Y. Xu, Y. Li, Chin. J. Catal. 32 (2011)
1790-1802.
[22] W. Linghu, X. Li, K. Asami, K. Fujimoto, Fuel Process. Technol. 85 (2004) 1121-
1138.
[23] W. Linghu, X. Liu, X. Li, K. Fujimoto, Catal. Lett. 108 (2006) 11-13.
[24] X. Liu, W. Linghu, X. Li, K. Asami, K. Fujimoto, Appl. Catal. A 303 (2006) 251-
257.
[25] L. Zhong, F. Yu, Y. An, Y. Zhao, Y. Sun, Z. Li, T. Lin, Y. Lin, X. Qi, Y. Dai, L. Gu,
J. Hu, S. Jin, Q. Shen, H. Wang, Nature 538 (2016) 84-87.
[26] 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.
[27] Z. Li, T. Lin, F. Yu, Y. An, Y. Dai, S. Li, L. Zhong, H. Wang, P. Gao, Y. Sun, M.
He, ACS Catal. 7 (2017) 8023-8032.
[28] E. Iglesia, S. C. Reyes, R. J. Madon, J. Catal. 129 (1991) 238-256.
[29] C. G. Visconti, E. Tronconi, L. Lietti, R. Zennaro, P. Forzatti, Chem. Eng. Sci. 62
(2007) 5338-5343.
[30] B. W. Wojciechowski, Catal. Rev. 30 (1998) 629-702.

[31] M. Claeys, E. van Steen, Catal. Today 71 (2002) 419-427.
[32] E. W. Kuipers, I. H. Vinkenburg, H. Oosterbeek, J. Catal. 152 (1995) 137-146.
[33] M. J. Keyser, R. C. Everson, R. L. Espinoza, Appl. Catal. A 171 (1998) 99-107.
[34] D. Förtsch, K. Pabst, E. Groß-Hardt, Chem. Eng. Sci. 138 (2015) 333-346.
[35] G. P. van der Laan, A. A. C. M. Beenackers, Catal. Rev. 41 (1999) 255-318.
[36] H. Schulz, Top. Catal. 26 (2003) 73-85.
[37] T. Komaya, A. T. Bell, J. Catal. 146 (2003) 237-248.
[38] J. Patzlaff, Y. Liu, C. Graffmann, J. Gaube, Appl. Catal. A 186 (1991) 109-119.
[39] J. Patzlaff, Y. Liu, C. Graffmann, J. Gaube Catal. Today 71 (2002) 381-394.
Table:
Table 1. Performance of the CoMn catalyst under different reaction conditions.
Product selec.
(wt%)
Ratio
of
H2/CO
CO
CO2
Hydrocarbon distribution (wt%, CO2-free)
T
P
Conv. Selec.
(°C) (bar)
CH4
5.7
C2-4
C2-4
2.0
C5+
C5-12
—
C5-12
—
C13+
—
HCa
Oxy.b
=
o
=
o
(C%)
31.8
(C%)
47.3
250
1
2
1
60.9
50.5
46.3
31.4
99.2
0.8
c
250
1
11.5
6.3
48.0
48.3
4.3
2.8
1.3
0.9
43.9
50.0
—
—
—
—
—
—
98.5
98.0
1.5
2.0
c
250
1
0.5
c
250
250
250
260
270
3
5
10
1
0.5
0.5
0.5
2
17.4
23.5
28.5
32.6
41.4
43.9
43.3
43.3
48.9
49.6
4.8
5.1
7.2
9.6
14.2
50.9
49.6
44.2
57.3
57.6
5.1
6.0
8.2
2.1
2.5
39.2
39.3
40.4
31.1
25.8
32.1
32.6
27.8
—
2.8
3.7
4.5
—
4.4
3.0
8.1
—
86.7
86.0
87.5
—
13.3
14.0
12.5
—
1
2
—
—
—
—
—
a Hydrocarbon. b Oxygenates. c The data is from reference [25]
Table 2. Ratios of olefin to paraffin under different reaction conditions.
Ratio
of
Olefin/paraffin ratio
T
P

(°C)
250
250
250
250
250
250
260
270
(bar)
1
H2/CO
2
C2
19
27
35
4
C3
41
53
68
16
13
8
C4
29
36
45
13
10
6
C5
27
32
38
13
10
7
C6
24
30
34
12
10
6
C7
23
24
26
11
9
C8
—
—
—
11
8
C9
—
—
—
11
8
C10
—
—
—
10
7
C11
—
—
—
10
7
C12
—
—
—
9
C13
—
—
—
8
C14
—
—
—
7
C15
—
—
—
5
1
1
1
0.5
0.5
0.5
0.5
2
3
5
3
6
6
5
4
10
1
2
6
6
6
6
5
5
4
3
2
18
15
41
39
27
25
26
24
23
20
26
23
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
1
2
Table 3. Value of α1, α2 and α3 calculated by the modified ASF distribution model under different reaction
conditions.
Chain growth probability
T
(°C)
P
(bar)
Ratio of
H2/CO
α1
α2
α3
250
250
250
250
250
250
260
270
1
1
2
1
0.79
0.87
0.90
0.82
0.81
0.74
0.71
0.63
0.72
0.80
0.82
0.79
0.81
0.82
0.64
0.57
0.41
0.43
0.45
0.68
0.64
0.62
0.35
0.34
1
0.5
0.5
0.5
0.5
2
3
5
10
1
1
2
Figure:
Scheme 1.Reaction scheme of selected carbon chain length of n.

Figure 1. Catalysis characterization: (a) XRD patterns of the CoMn catalyst at different stages; (b-e) TEM images
of the precursor and the samples after calcination, reduction, and reaction, respectively; (f-i) The corresponding
high resolution TEM images.
CoxMn1-x
O
Co2C
MnO
Spt.
Red.
39
40
41
42
2θ (degree)
43
44
45
Figure 2.Illustration of the shift of the diffraction peak of CoxMn1-xO to lower angle after the reaction.

Figure 3. Size distribution based on the TEM characterization. a-d) represent the catalyst precursor and the
samples after calcination, reduction, and reaction, respectively.
Figure 4.TEM images of the reduced catalyst. a) Scanning TEM image. b) Corresponding HRTEM image. c)
Overlay EDX mapping of Co, Mn and O. d-f) EDX mapping of Co, Mn and O, respectively.
Figure 5.TEM images of the spent catalyst. a) Scanning TEM image. b-d) EDX mapping of Co, Mn,and Co plus
Mn.

Figure 6. Product distribution of CoMn catalyst under different reaction conditions. (a) 250 °C, 1 bar and
H2/CO=2; (b) 250 °C, 1 bar and H2/CO=1; (c) 250 °C, 1 bar and H2/CO=0.5; (d) 250 °C, 3 bar and H2/CO=0.5; (e)
250 °C, 5 bar and H2/CO=0.5; (f) 250 °C, 10 bar and H2/CO=0.5; (g) 260 °C, 1 bar and H2/CO=2; (h) 270 °C, 1
bar and H2/CO=2.
Figure 7. (a) Distribution of hydrocarbons and ratio of olefin to paraffin with the carbon number, reaction
conditions: 250 °C, 3 bar, H2/CO=0.5 and 2000 ml gcat-1 h-1; (b) Analytical expression of the modified ASF
distribution.

Scheme 2. Fischer-Tropsch to olefin reaction mechanism based on the modified ASF distribution model.