Mechanistic double ASF product distribution study of FischereTropsch

synthesis on precipitated iron catalyst

Ali Nakhaei Pour
*
, Hamideh Khodabandeh, Mohammad Izadyar,
Mohammad Reza Housaindokht
Department of Chemistry, Ferdowsi University of Mashhad, P.O. Box 9177948974, Mashhad, Iran
a r t i c l e i n f o
Article history:
Received 27 June 2013
Received in revised form
5 September 2013
Accepted 22 September 2013
Available online
Keywords:
Iron catalysts
Kinetics
Products data
Double ASF distributions
a b s t r a c t
Product distributions of iron based catalyst represented by double ASF distributions characterized by two
growth probabilities. Based on enol mechanism and double ASF distributions the complete set of
elementary reactions is given. By using mechanistic kinetic studies of FischereTropsch reaction, the
chain growth probabilities (a
1
and a
2
) for double ASF distributions are formulated. FeeCueLa catalyst
was prepared by co-precipitation procedure, and steady-state FTS reaction rates were measured in a
continuous spinning basket reactor. The experimental results for low carbon monoxide conversions have
been satisfactorily t with the double ASF model. But for higher carbon monoxide conversions deviate
substantially because the FTS reaction highly depends on the hydrogen formed by the WGS as the carbon
monoxide conversion increases.
2013 Elsevier B.V. All rights reserved.
1. Introduction
FischereTropsch synthesis (FTS) has attracted increasing inter-
est as an important route for indirect coal liquefaction and gas to
liquid (GTL) process (Anderson, 1984; Davis, 2005). The use of iron-
based catalysts is attractive due to their high FTS activity as well as
their wateregas shift (WGS) reactivity, which helps to make up the
decit of H
2
in the syngas derived fromcoal gasication (Dry, 2002,
2004).
The products of the FTS are primarily linear hydrocarbons
distributed over a wide range of carbon numbers. In the Andersone
SchulzeFlory (ASF) distribution model, the formation of hydro-
carbon chains was assumed as a stepwise polymerization proce-
dure and the chain growth probability was assumed to be
independent of the carbon number (Davis, 2009; Dry, 1999).
However, signicant deviations from the ideal ASF distribution
have been observed in many studies (Nakhaei Pour et al., 2008;
Gaube and Klein, 2008). Many literatures reported that for both
iron and cobalt catalysts, product distributions can be represented
by superposition of double ASF distributions characterized by two
growth probabilities (Gaube and Klein, 2008, 2010; Huff and
Sattereld, 1984; Pour et al., 2008, 2010a).
The mechanistic kinetic studies for iron catalysts are based on
the formation of the monomer species as the rate-determining step
in the consumption of synthesis gas (Teng et al., 2006; Visconti
et al., 2007; Pour et al., 2010b, 2012). In this approach, all the
mechanistic steps inwhich the COand H
2
are consumed and lead to
the nal products are considered jointly (Nakhaei Pour et al., 2010a;
Yang et al., 2003). Based mechanistic kinetic studies of FTS reaction
on iron catalysts the growth probabilities for two ASF distributions
can be formulated.
In this work the chain growth probabilities of the FTS reaction
are developed by LangmuireHinshelwoodeHougeneWatson
(LHHW) modeling in gradientless reactor over Fe/Cu/La/Si catalyst.
The concept two ASF distributions are used for prediction of
products distribution of iron catalyst in FTS reaction using devel-
oped mechanistic chain growth probabilities. Based on enol
mechanism and two ASF distributions the complete set of
elementary reactions is given. By using the experimental data, the
FTS kinetics model are estimated and discriminated separately.
2. Product distribution model
In this work the carbon number distribution of FischereTropsch
products on iron catalyst was studied by use of a modied Ander-
soneSchulzeFlory (ASF) distribution with two chain growth
* Corresponding author. Tel./fax: 98 5118795457.
E-mail addresses: nakhaeipoura@yahoo.com, a.nakhaei@um.ac.ir (A. Nakhaei
Pour).
Contents lists available at ScienceDirect
Journal of Natural Gas Science and Engineering
j ournal homepage: www. el sevi er. com/ l ocat e/ j ngse
1875-5100/$ e see front matter 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jngse.2013.09.005
Journal of Natural Gas Science and Engineering 15 (2013) 53e58
probabilities. The modied ASF distribution with two chain growth
probabilities method was proposed by Donnelly et al. (1988) and is
used to characterize the carbon number distribution of FTS where
independent ASF distributions with different chain growth proba-
bilities are superimposed.
x
i
Aa
i1
1
Ba
i1
2
(1)
The most important double ASF distribution aspect is a
1
and a
2
hydrocarbon chain as following:
R
CnH2n2
R
CnH2n
R
Cn1H2n
R
Cn1H2n2
a
1
n x (2)
R
CnH2n2
R
CnH2n
R
Cn1H2n
R
Cn1H2n2
a
2
n > x (3)
which RC
n
H
2n
is olen rate, RC
n
H
2n2
is linear parafn rate with n
carbon, and a is a factor of hydrocarbon chain growth probability
(0 < a < 1). For deriving the LangmuireHinshelwoodeHougene
Watson (LHHW) for growth probabilities (a
1
and a
2
) presented in
Table 1, enol theory is utilized to initiate the hydrocarbon chain
reaction. In this complete set of elementary reactions, the break-
down of normal ASF is considered. The concentration of the surface
intermediate species such as [Hs], [C
n
H
2n1
s], [CH
3
s], and[s] must
be determined to achieve rate equations by applying quasi-
equilibrium condition for elementary reactions and then accord-
ing to the partial pressure of CO, H
2
, and H
2
O.
CH
3
s
K
1
K
2
K
3
K
1
2
4
K
5
P
3
2
H2
P
CO
P
H2O
s (4)
Hs
_
K
4
P
H2
_
1=2
s (5)
By applying steady state condition according to the partial
pressure of CO, H
2
, and H
2
O, the concentration of the surface in-
termediate [C
n
H
2n1
s] can be determined and then solve above
equations. Therefore, for [C
n
H
2n1
s] when n x, can be dened:
dC
n
H
2n1
s=dt 0 n x (6)
By applying steady state condition:
k
11
HsC
n
H
2n
s k
11
C
n
H
2n1
ss k
12
C
n
H
2n1
sHs
k
13
C
n
H
2n1
ss
0 n x (7)
Then by rearranging:
C
n
H
2n1
s
C
n1
H
2n3
s

k
11
K
10
CH
3
sHs
k
11
s
2
k
12
Hss k
13
s
2
a
1
nx (8)
According to the above equation, ratio of adjacent carbon spe-
cies concentrations may be assumed to be the same as probability
factor of hydrocarbon chain growth. Probability factor of hydro-
carbon chain growth is attained by replacing (4) and (5) equations
in (8) as following:
a
1

k
11
K
1
K
2
K
3
K
4
K
5
K
10
P
2
H
2
PCO
PH
2
O
k
11
k
12
K
1=2
4
P
1=2
H2
k
13
n x (9)
Also, for [C
n
H
2n1
s] when n > x, can be dened:
dC
n
H
2n1
s=dt 0 n > x (10)
By applying steady state condition:
k
15
HsC
n
H
2n
s k
15
C
n
H
2n1
ss k
16
C
n
H
2n1
sHs
k
17
C
n
H
2n1
ss
0 n > x (11)
Then by rearranging:
C
n
H
2n1
s
C
n1
H
2n3
s

k
15
K
14
CH
3
sHs
k
15
s
2
k
16
Hss k
17
s
2
a
2
n > x
(12)
According to the above equation, ratio of adjacent carbon spe-
cies concentrations may be assumed to be the same as probability
factor of hydrocarbon chain growth. Probability factor of hydro-
carbon chain growth is attained by replacing (4) and (5) equations
in (12) as following:
a
2

k
15
K
1
K
2
K
3
K
4
K
5
K
14
P
2
H
2
PCO
PH
2
O
k
15
k
16
K
1=2
4
P
1=2
H2
k
17
n > x (13)
It is important to notice that two growth probabilities a
1
and a
2
not independent parameters in double ASF model. Instead, we note
that at the break point on ASF diagram, the contributions of each
term in Equation (1) are equal.
Aa
i1
1
Ba
i1
2
i z (14)
It is now possible to obtain kinetic rate equations for proposed
models according to the above assumptions. In break point, i
illustrate as z and necessarily is not an integral carbon number. In
this model A and B were not assumed to correspond directly to the
fractions of products produced froma
1
and a
2
, respectively. Instead
the sum of the mole fractions overall carbon numbers is unity:

N
i 1
x
i

N
i 1
_
Aa
i1
1
Ba
i1
2
_
1 (15)
Methane and ethene does not obey the ASF equation and after
removing C
1
and C
2
products to t theoretical distributions to data
leads to:
Table 1
Elementary reactions for proposed models.
Model Reaction steps Elementary reaction
FT 1 CO s 4COs
2 COs H
2
4H
2
COs
3 H
2
COs H
2
4CH
2
s H
2
O
4 H
2
2s 42Hs
5 CH
2
s Hs 4CH
3
s s
6 rate determining step CH
3
s Hs /CH
4
2s
7 CH
3
s CH
2
s 4C
2
H
5
s s
8 rate determining step C
2
H
5
s Hs /C
2
H
6
2s
9 rate determining step C
2
H
5
s s /C
2
H
4
Hs s
10 CH
3
s C
n1
H
2n3
s 4C
n
H
2n
s s n z
11 C
n
H
2n
s Hs 4C
n
H
2n1
s s n z
12 rate determining step C
n
H
2n1
s Hs /C
n
H
2n2
2s n z
13 rate determining step C
n
H
2n1
s s /C
n
H
2n
Hs s n z
14 CH
3
s C
n1
H
2n3
s 4C
n
H
2n
s s n > z
15 C
n
H
2n
s Hs 4C
n
H
2n1
s s n > z
16 rate determining step C
n
H
2n1
s Hs /C
n
H
2n2
2s n > z
17 rate determining step C
n
H
2n1
s s /C
n
H
2n
Hs s n > z
A. Nakhaei Pour et al. / Journal of Natural Gas Science and Engineering 15 (2013) 53e58 54

N
i 3
x
i

N
i 1
_
Aa
i1
1
Ba
i1
2
_
A1 a
1
B1 a
2

1 x
1
x
2
(16)
Therefore, the determination of growth probabilities of the
bimodal ASF distribution is based on hydrocarbons with carbon
numbers i > 2. The bimodal distribution is characterized by two
growth probabilities (a
1
and a
2
) and the point of intersection x of
both ASF distributions. The mathematical procedure given by
Donnelly et al. (1988) arrives at
A 1 x
1
x
2

_
a
1
2
1 a
1

_
a
1
a
2
_
x1
$
_
a
2
2
1 a
2
__
1
(17)
x
i
1x
1
x
2

_
a
i1
1

_
a
1
a
2
_
x1
$a
i1
2
_
$
_
a
2
1
1a
1

_
a
1
a
2
_
x1
$
_
a
2
2
1a
2
__
1
(18)
However, the intersection point x is a formal parameter without
any physicochemical relevance.
The parameters of model were calculated with the Levenberge
Marquardt (LM) algorithm. The mean absolute relative residual
(MARR) is reported as a measure of the goodness of the t:
MARR 100

n
1

_
lnX
exp
lnX
mod
lnX
exp
_

1
n
(19)
where n is the number of data points included and X
exp
and X
model
are molar fraction of experimental and model results, respectively.
3. Experimental
3.1. Catalyst preparation
FeeCu catalyst was prepared by co-precipitation procedure,
described elsewhere (Nakhaei Pour et al., 2010b). The La pro-
moter was added by incipient wetness impregnation and
calcined in 723 K. Final catalyst compositions were designed in
terms of the atomic ratio as: 100Fe/5.64Cu/2La/19Si. The fresh
catalyst was crushed and sieved to particles with diameters of
250e300 mm, which has been proved to be a compromising
particle size, safe for neglecting the intra-particle transfer limi-
tations and promisingly easy operating for the reactor during the
experiment.
3.2. Experimental apparatus and procedure
Steady-state FTS reaction rates were measured in a continuous
spinning basket reactor (stainless steel, H 0.122 m, D
0
0.052 m,
D
i
0.046 m) with temperature controllers (WEST series 3800). A
detailed description of the experimental setup and procedures has
been provided in our previous work (Pour et al., 2012).
Blank experiments showed that the spinning basket reactor
charged with inert silica sand without the catalyst has no conver-
sion of syngas. The weight of the catalyst loaded was 2.5 g and
diluted by 30 cm
3
inert silica sand with the same mesh size range.
The catalyst samples were activated by a 5% (v/v) H
2
/N
2
gas mixture
with space velocity of 15.1 NL/g
cat
/h at 1 bar and 1800 rpm. The
reactor temperature increased to 673 K with a heating rate of 5 K/
min, maintained for 1 h at this temperature, and then reduced to
543 K. The activationwas followed by the synthesis gas streamwith
H
2
/CO 1 and space velocity of 3.07 NL/g
cat
/h for 24 h at 1 bar and
543 K before setting the actual reaction temperature and pressure.
After catalyst reduction, synthesis gas was fed to the reactor at
conditions operated at 563 K, 17 bar (H
2
/CO) 1 and a space ve-
locity of 10.4 NL/g
cat
/h. A stabilization period of 12 h was conducted
under the reaction conditions, and then the kinetic measurements
were carried out. A total of 16 sets of experimental data with
variation of gas space velocity, and syngas quality were carried out
for kinetic study. Experimental conditions were varied in the
following ranges: GHSV 2e20 NL/g
cat
/h, H
2
/CO feed ratio 0.5e
1.5. Each experiment was replicated 3 times in order to verify the
experimental data accuracy and reproducibility. Conversion of
carbon monoxide and hydrogen, and formation of various products
were measured within a time period of 24 h at each run. At regular
intervals, the standard experiment was repeated to determine
possible deactivation effects of the catalysts. Kinetic samples were
collected during a typical period of 10e16 h. For each operation
condition, it took at least 12 h to ensure the steady-state behavior of
Table 2
The reaction conditions and experimental results of syngas conversion and also, carbon monoxide, hydrogen and water partial pressures.
Run H
2
/CO GHSV (NL/gcat/h) PCO (bar) PH
2
O (bar) PH
2
(bar) Conversion%
CO H
2
1 1 2 1.26 1.53 5.06 91.1 63.9
2 1 4 1.67 1.96 4.95 88.2 64.4
3 1 6 4.27 1.38 6.19 64.2 48.1
4 0.5 6 7.68 1.01 2.47 52.2 69.3
5 1.5 6 2.69 1.44 8.46 70.9 39.1
6 1 8 5.70 1.03 6.88 47.4 36.5
7 0.5 8 9.11 0.86 3.40 37.1 53.1
8 1.5 8 4.05 1.06 8.97 52.2 29.4
9 1 10 6.48 0.82 7.27 36.6 28.9
10 0.5 10 9.74 0.71 3.94 28.8 42.3
11 1.5 10 4.81 0.91 9.20 40.4 24
12 1 12 6.96 0.69 7.51 29.3 23.8
13 1 14 7.33 0.62 7.69 23.2 19.4
14 1 16 7.63 0.53 7.82 18.4 16.4
15 1 18 7.80 0.48 7.91 15.2 14.1
16 1 20 7.89 0.48 7.95 13.7 13.1
Reaction condition:
Total pressure 17 bar, GHSV 2e20 NL/g
cat
/h, H
2
/CO 0.5e1.5 (feed mol ratio).
Temperature 543 K.
A. Nakhaei Pour et al. / Journal of Natural Gas Science and Engineering 15 (2013) 53e58 55
the catalyst after a change in the reaction conditions. The water
partial pressure was determined by collecting the water in the trap,
separating it fromthe oil, and determining its amount by weighing.
The weight of water was converted to partial pressure in the reactor
based on the ideal gas law.
The out gas was analyzed by a gas chromatograph (Varian CP-
3800) equipped with TCD and FID detectors. The CO, CO
2
, N
2
, and
O
2
were analyzed through two packed column in series (Molecular
sieve13x CP 81025 with 2 m length, and 3 mm OD, and Hayesep Q
CP1069 with 4 m length, and 3 mm OD) connected to TCD detector.
The C
1
eC
5
hydrocarbons were analyzed via a capillary column (CP
fused silica with 25 m 0.25 mm 0.2 mm lm thickness) con-
nected to FID detector. Hydrogen was analyzed through Shimadzu,
GC PTF 4C, equipped with TCD detector and two column in series
(Propack-Q with 2 m length, and 3 mm OD for CO
2
, C
2
H
4
and C
2
H
6
separation and molecular sieve-5A with 2 m length, and 3 mm OD
for CO, N
2
, CH
4
and O
2
separation), which were connected to each
other via a three way valve.
The collected liquid (Including hydrocarbons and oxygenates)
were analyzed ofine with Varian CP-3800 gas chromatograph
equipped with capillary column (TM DH fused silica capillary col-
umn, PETRO COL 100 m 0.25 mm 0.5 mm lm thickness) con-
nected to FID detector. Both total mass and atomic material
balances were performed with the consideration that a run could
be accepted for further analysis if the carbon material balance
closed between 97 and 103%. This criterion was adopted since
compounds containing carbon and hydrogen may accumulate in
the reactor, in the form of high molecular weight hydrocarbons.
4. Results and discussion
The reaction conditions and experimental results of syngas
conversion and also, carbon monoxide, hydrogen and water partial
pressures are listed in Table 2. As shown in this table, the FTS re-
action evaluated in spread ranges of feed conversions. Based on
experimental results, products distributions and chain growth
probabilities (a
1
and a
2
) for produced hydrocarbons evaluated.
Fig. 1 compared calculated double ASF model based on experi-
mental data and experimental results. In this Fig., the double ASF
model results calculated from experimental for hydrocarbons
production rates results by using Equation (18).The model results is
considered for hydrocarbons with carbon number (n) greater than
3, which is calculated using Equation (18). As shown in this Fig. the
calculated two ASF model are tted with experimental results
carefully. Thus, the two ASF model is a useful model for prediction
of products distribution on lanthanum promoted iron catalyst in
our experimental conditions. Also, two chain growth probabilities
(a
1
and a
2
) are calculated by minimize the MARR (Equation (19))
and listed in Table 3.
Parameter estimation obtained for the model provides a satis-
factory tting between experimental and calculated results. The
calculated chain growth probabilities (a
1
and a
2
) using LHHW
method are listing in Table 3 and compared with experimental
results. For tting the two ASF model derived based on LHHW
method, X
1
and X
2
is calculated from experimental results (Equa-
tion (19)). The parameter values of the kinetic models are listed in
Table 4 and all of them are statistically signicant. The kinetic pa-
rameters for the models listed in this table were calculated fromthe
experimental results by minimizing the following objective func-
tion, MARR. Negative adsorption and kinetic parameters were
excluded because the estimation of these parameters must have
physical relevance. Furthermore, the surface fractions of adsorbed
species should be realistic and the residuals between model and
experimental data should be normally distributed with zero
average and may not observe trends as a function of the indepen-
dent variables.
Fig. 2 shows the comparison of experimental and calculated
product distributions. As shown in this Figure and Table 3, the
experimental results is better t with model for run 16 and become
far to run1. One way to explain the deviation of the model results
from experimental data of some experiments is the inuence of
Fig. 1. Comparison between calculated double ASF model results () and experimental data (-). The model results are considered for hydrocarbons with carbon number (n) n > 3.
Total pressure 17 bar, GHSV 2e20 NL/gcat/h, H
2
/CO 0.5e1.5 (feed mol ratio). Temperature 543 K.
Table 3
Calculated chain growth probabilities (a
1
and a
2
) from experimental results and
LHHW derived results.
Run Experimental Calculated MARR
a
2
a
1
x a
2
a
1
x a
2
a
1
1 0.72 0.52 12 0.014 0.011 12 97.9 98.1
2 0.75 0.54 12 0.013 0.011 12 98.0 98.2
3 0.77 0.55 12 0.086 0.069 12 87.6 88.7
4 0.81 0.52 12 0.020 0.017 12 96.7 97.5
5 0.72 0.58 12 0.116 0.091 12 84.3 83.8
6 0.78 0.57 12 0.203 0.161 12 71.9 74.1
7 0.80 0.55 12 0.063 0.053 12 90.5 92.1
8 0.76 0.60 12 0.275 0.215 12 64.0 63.6
9 0.80 0.60 12 0.333 0.263 12 56.0 58.6
10 0.82 0.58 12 0.121 0.099 12 82.9 85.3
11 0.78 0.61 12 0.406 0.316 12 48.5 48.3
12 0.82 0.61 12 0.460 0.363 12 41.0 44.1
13 0.84 0.64 12 0.576 0.453 12 29.2 31.6
14 0.86 0.66 12 0.729 0.573 12 12.8 15.4
15 0.87 0.67 12 0.858 0.674 12 0.0 1.4
16 0.88 0.69 12 0.880 0.691 12 0.0 0.0
Reaction condition:
Total pressure 17 bar, GHSV 2e20 NL/g
cat
/h, H
2
/CO 0.5e1.5 (feed mol ratio).
Temperature 543 K.
A. Nakhaei Pour et al. / Journal of Natural Gas Science and Engineering 15 (2013) 53e58 56
wateregas-shift (WGS) reaction on FTS reaction. Iron-based is a
suitable catalysts for FTS and WGS reaction. The wateregas-shift
(WGS) reaction can be shown as:
CO H
2
O4CO
2
H
2
(20)
The WGS reaction is a reversible paralleleconsecutive reaction
with respect to CO and assumed that carbon dioxide is essentially
formed by this reaction (Davis, 2009; An et al., 2007; Herranz et al.,
2006; Yang et al., 2006). It is generally accepted that the FTS and
WGS reactions take place on different active sites over a precipi-
tated iron catalyst and the two reactions will only inuence each
other via the gas phase (Lox and Froment, 1993; Zhang et al., 2009;
van der Laan and Beenackers, 2000). As shown in this Fig., the re-
sults for higher carbon monoxide conversions deviate substantially
because the FTS reaction highly depends on the hydrogen formed
by the WGS as the carbon monoxide conversion increases.
The inuence of carbon monoxide and hydrogen conversions on
calculated MARR of model results is showed in Fig. 3. As shown in
this Fig. the calculated MARR is increased linearly with increasing
of hydrogen conversion instead carbon monoxide. These results are
expected by considering the developed models for chain growth
factors (a
1
and a
2
). Equations (9) and (13) show that hydrogen
pressure is inuence on model results with exponent of 3. Thus any
change in hydrogen partial pressure due to interference of WGS
reaction show distractive effect on calculated results.
The overall FTS reaction rate is highly affected by the rate/extent
of the WGS reaction at high FTS conversions. The rate of FTS and
WGS reactions must be written as:
R
WGS
R
CO2
(21)
R
FTS
R
CO
R
WGS
(22)
where R
CO2
is the rate of CO
2
formation (wateregas shift reaction)
and R
CO
is the rate of CO consumption. The rate of FTS and WGS
reactions are listed in Table 5. The R
FTS
/R
WGS
ratio indicates the
degree of deviation of calculated results from experimental and is
show in Fig. 4. As shown in this gure, with increasing the r
FTS
/r
WGS
ratio the deviation of calculated results from experimental
increased rapidly. This Fig. shows that the presence of WGS reaction
in parallel of FTS reaction is the main reason for the deviation from
expected results and the FTS reaction rate is unaffected by the WGS
Table 4
Isothermal the parameters of the FTS kinetic models.
Parameter Value Unit
k
11
1.16 mmol/g
cat
/h
k
11
0.10 mmol/g
cat
/h
k
12
2830.2 mmol/g
cat
/h
k
13
0.2 mmol/g
cat
/h
k
15
97.52 mmol/g
cat
/h
k
15
1.95 mmol/g
cat
/h
k
16
13.11 mmol/g
cat
/h
k
17
8.62 mmol/g
cat
/h
K
1
0.74 bar
1
K
2
0.74 bar
1
K
3
0.74 bar
1
K
4
5.04 bar
1
K
5
0.74 bar
1
K
10
0.86 bar
1
K
14
6.7 10
5
bar
1
Fig. 2. Comparison between calculated double ASF model based on LHHW method (-) and experimental results (). Total pressure 17 bar, GHSV 2e20 NL/gcat/h, H
2
/CO 0.5e
1.5 (feed mol ratio). Temperature 543 K.
Fig. 3. The inuence of carbon monoxide and hydrogen conversions on calculated MARR of model results. Total pressure 17 bar, GHSV 2e20 NL/gcat/h, H
2
/CO 0.5e1.5 (feed
mol ratio). Temperature 543 K.
A. Nakhaei Pour et al. / Journal of Natural Gas Science and Engineering 15 (2013) 53e58 57
reaction only at low conversions (van der Laan and Beenackers,
2000).
It should be noted that the current kinetic model has not
considered the effect of diffusivity and solubility of olens with
different chain sizes on the parafn/olen ratio on the basis of the
understanding of different orientations of kinetic and reactor
models in considering the intrinsic mechanisminformation and the
transportation effects. The kinetic model developed is signicant
only for the cases excluding the diffusivity and solubility factors,
which are believed to be signicant in enhancing the olen read-
sorption and secondary reaction and, therefore, in changing the
olen and parafn selectivities. However, kinetic experiments can
never exclude the diffusivity and solubility (mobility) especially of
heavy olens. It can be seen that the kinetic model without
considering the transportation effect (diffusivities and solubilities)
has not predicted the correct dependence compared with experi-
mental results.
5. Conclusion
In this work the carbon number distribution of FischereTropsch
products on iron catalyst was studied by use of a modied Ander-
soneSchulzeFlory (ASF) distribution with two chain growth
probabilities. Based on enol mechanism and two ASF distributions
the complete set of elementary reactions is given. Using mecha-
nistic kinetic studies of FTS reaction the chain growth probabilities
(a
1
and a
2
) for two ASF distributions formulated. The calculated
two ASF model are tted with experimental results carefully at low
carbon monoxide. Thus, the two ASF model is a useful model for
prediction of products distribution on lanthanum promoted iron
catalyst in our experimental conditions at low carbon monoxide
conversions. The results for higher carbon monoxide conversions
deviate substantially because the FTS reaction highly depends on
the hydrogen formed by the WGS as the carbon monoxide con-
version increases.
References
An, X., Wu, B.-s., Wan, H.-J., Li, T.-Z., Tao, Z.-C., Xiang, H.-W., Li, Y.-W., 2007. Catal.
Commun. 8, 1957e1962.
Anderson, R.B., 1984. Academic Press, Orlando.
Davis, B., 2005. Top. Catal. 32, 143e168.
Davis, B.H., 2009. Catal. Today 141, 25e33.
Donnelly, T.J., Yates, I.C., Sattereld, C.N., 1988. Energy Fuels 2, 734e739.
Dry, M.E., 1999. Appl. Catal. A Gen. 189, 185e190.
Dry, M.E., 2002. Catal. Today 71, 227e241.
Dry, M.E., 2004. Appl. Catal. A Gen. 276, 1e3.
Gaube, J., Klein, H.F., 2008. J. Mol. Catal. A Chem. 283, 60e68.
Gaube, J., Klein, H.F., 2010. Appl. Catal. A Gen. 374, 120e125.
Herranz, T., Rojas, S., Prez-Alonso, F.J., Ojeda, M., Terreros, P., Fierro, J.L.G., 2006.
J. Catal. 243, 199e211.
Huff Jr., G.A., Sattereld, C.N., 1984. J. Catal. 85, 370e379.
Lox, E.S., Froment, G.F., 1993. Ind. Eng. Chem. Res. 32, 11.
Nakhaei Pour, A., Shahri, S.M.K., Bozorgzadeh, H.R., Zamani, Y., Tavasoli, A.,
Marvast, M.A., 2008. Appl. Catal. A Gen. 348, 201e208.
Nakhaei Pour, A., Housaindokht, M.R., Zarkesh, J., Tayyari, S.F., 2010. J. Nat. Gas Sci.
Eng. 2, 79e85.
Nakhaei Pour, A., Housaindokht, M.R., Tayyari, S.F., Zarkesh, J., Alaei, M.R., 2010.
J. Nat. Gas Sci. Eng. 2, 61e68.
Pour, A.N., Zamani, Y., Tavasoli, A., Kamali Shahri, S.M., Taheri, S.A., 2008. Fuel 87,
2004e2012.
Pour, A.N., Zare, M., Zamani, Y., 2010. J. Nat. Gas Chem. 19, 31e34.
Pour, A.N., Housaindokht, M.R., Tayyari, S.F., Zarkesh, J., 2010. J. Nat. Gas Chem. 19,
441e445.
Pour, A.N., Housaindokht, M.R., Zarkesh, J., Irani, M., Babakhani, E.G., 2012. J. Ind.
Eng. Chem. 18, 597e603.
Teng, B.-T., Chang, J., Zhang, C.-H., Cao, D.-B., Yang, J., Liu, Y., Guo, X.-H., Xiang, H.-W.,
Li, Y.-W., 2006. Appl. Catal. A Gen. 301, 39e50.
van der Laan, G.P., Beenackers, A.A.C.M., 2000. Appl. Catal. A Gen. 193, 39e53.
Visconti, C.G., Tronconi, E., Lietti, L., Zennaro, R., Forzatti, P., 2007. Chem. Eng. Sci. 62,
5338e5343.
Yang, J., Liu, Y., Chang, J., Wang, Y.-N., Bai, L., Xu, Y.-Y., Xiang, H.-W., Li, Y.-W.,
Zhong, B., 2003. Ind. Eng. Chem. Res. 42, 5066e5090.
Yang, J., Sun, Y., Tang, Y., Liu, Y., Wang, H., Tian, L., Wang, H., Zhang, Z., Xiang, H.,
Li, Y., 2006. J. Mol. Catal. A Chem. 245, 26e36.
Zhang, R., Chang, J., Xu, Y., Cao, L., Li, Y., Zhou, J., 2009. Energy Fuels 23, 4740e4747.
Table 5
The experimental results for R
FTS
and R
WGS
reactions.
Run R
FTS
(mol/g
cat
/h) R
WGS
(mol/g
cat
/h)
1 0.0223 0.0183
2 0.0448 0.0338
3 0.0499 0.0361
4 0.0513 0.0420
5 0.0456 0.0304
6 0.0500 0.0347
7 0.0503 0.0380
8 0.0455 0.0291
9 0.0490 0.0327
10 0.0497 0.0360
11 0.0455 0.0267
12 0.0479 0.0306
13 0.0458 0.0269
14 0.0423 0.0233
15 0.0404 0.0208
16 0.0416 0.0196
Reaction condition:
Total pressure 17 bar, GHSV 2e20 NL/g
cat
/h, H
2
/CO 0.5e1.5 (feed mol ratio).
Temperature 543 K.
Fig. 4. The inuence of rate of wateregas shift reaction (R
WGS
) and FischereTropsch
synthesis (R
FTS
) on calculated MARR of model results.
A. Nakhaei Pour et al. / Journal of Natural Gas Science and Engineering 15 (2013) 53e58 58