Fischer–Tropsch synthesis in a bench-scale two-stage multitubular fixed-bed reactor: Simulation and enhancement in conversion and diesel selectivity

Chemical Engineering Science 105 (2014) 1–11

Contents lists available at ScienceDirect

Chemical Engineering Science journal homepage: www.elsevier.com/locate/ces

Fischer–Tropsch synthesis in a bench-scale two-stage multitubular fixed-bed reactor: Simulation and enhancement in conversion and diesel selectivity Xiao Ping Dai, Pei Zhi Liu, Yong Shi, Jian Xu, Wei Sheng Wei n State Key Laboratory of Heavy Oil Processing, China University of Petroleum, Box 277, 18 Fuxue road, Changping, Beijing 102249, People's Republic of China

H I G H L I G H T S








A process based on a two-stage multitubular fixed-bed reactor is presented for Fischer–Tropsch synthesis (FTS). A two-dimensional pseudohomogeneous reactor model is proposed for bench-scale FTS. The two-stage multitubular fixed-bed process increases CO conversion, diesel range selectivity and stability. The simulations demonstrate the understandable enhanced CO conversion and products distribution shift toward diesel-range hydrocarbon. We suggest a novel method of improving products distribution and increasing carbon efficiency by multi-stage short reactor rather than one long reactor for FTS.

art ic l e i nf o

a b s t r a c t

Article history: Received 11 July 2013 Accepted 23 September 2013 Available online 14 October 2013

A process based on a two-stage multitubular fixed-bed reactor is presented to enhance in conversion and diesel selectivity for the Fischer–Tropsch synthesis (FTS) process, which is characterized by the condensation and separation of liquid products and water from the outlet of the first-stage reactor. A two-dimensional pseudohomogeneous reactor model is proposed to simulate the temperature profile and CO conversion of the two-stage fixed-bed process. Model calculations indicate that the condensation and separation of liquid products and water between the two stages of the multitubular fixed-bed reactor plays an important role by changing temperature profile and synthesis gas partial pressure of the second-stage reactor. The model validation has been verified on the basis of the bench-scale test data in a single-stage and two-stage fixed-bed reactor, respectively, which demonstrates the understandable products distribution shift toward diesel-range hydrocarbon. The combined action of higher temperature, syngas/H2 partial pressure and readsorption of olefins at the two-stage fixed-bed reactor make CO conversion and C5 þ selectivity increase, while CH4 and CO2 selectivity keep decreasing trends. Compared with the single-stage fixed-bed reactor process, the total CO conversion increases from 79% to 89% in the two-stage fixed-bed reactor process. The high valuable diesel range (C12–C22) increases significantly from 67% to 78%, while the gasoline range (C5–C11) decreases from 27% to 18%. The catalysts exhibit more stability and less deactivation rate over the two-stage fixed-bed process. The results are helpful to obtain more valuable diesel products and effective utilization of the syngas by the multistage fixed-bed process without tail gas recycling, which provides a practical case for multi-stage short reactor rather than one long reactor for FTS to improve products distribution. & 2013 Elsevier Ltd. All rights reserved.

Keywords: Fischer–Tropsch synthesis Two-stage fixed-bed reactor Bench-scale Pseudohomogeneous reactor model Simulation Conversion and selectivity control

1. Introduction The low temperature Fischer–Tropsch synthesis (FTS) is considered nowadays the most promising route for environmentally sound production of transportation fuels and chemical feedstocks from natural gas, coal and biomasses (Bae et al., 2011; dela Osa

n

Corresponding author. Tel.: þ 86 10 89734981; Fax: þ86 10 89734979. E-mail address: [email protected] (W.S. Wei).

0009-2509/$ - see front matter & 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ces.2013.09.057

et al., 2012; Zhang et al., 2010). In recent years, the availability of cheap natural gas and raw materials like coal and biomass has given momentum to FTS. The capacities will increase in the near future with natural gas favoured as feedstock, and around 2015 the global annual production rate of fuels via FTS will be about 30 million tons, mostly produced in countries like South Africa, Malaysia, and Qatar (Jess and Kern, 2012). It is believed that the major impediment of the GTL popularization via FTS is the economical feasibility, which is related to the productivity, lifespan and CH4 selectivity of catalysts besides the

2

X.P. Dai et al. / Chemical Engineering Science 105 (2014) 1–11

price of natural gas. In this respect, the manufacture of syngas is by far the most capital-intensive part of a gas conversion plant. Therefore, the FTS unit should focus on utilizing syngas as efficiently as possible, and selectivity considerations are then extremely important in the design of the FTS section, which is of industrial interest (Geerlings et al., 1999; Lira, et al., 2008; dela Osa et al., 2012; Teiseh et al., 2012). The present emphasis has shifted towards maximizing the yield of high-cetane diesel products from the FTS process, which is virtually free of sulfur and aromatic compounds, low particulate and NOx emissions (Bermúdez et al., 2011; Szeto et al., 2012). Due to its high activity, high C5 þ hydrocarbons selectivity, lower activity for the water-gas shift (WGS) reaction and long life, cobalt-based catalyst is currently the catalyst of choice for the conversion of syngas to liquid fuels (Sadeqzadeh et al., 2012; Wang et al., 2013; Zhang et al., 2005). The highly exothermic nature of the FT reaction combined with the high activity of the Co catalyst makes the removal of heat from the reactor of critical importance, because the local reaction temperature is critical in controlling the process selectivity as the unwanted methanation reaction becomes dominant at high temperatures, while moderate temperatures are crucial to extend the catalyst life time and prevent thermal runaways. Therefore, the effectiveness of heat transfer determines the temperature profile of the reactor and catalyst particle. Various types of reactors, including fixed bed, fluidized bed and slurry phase, are used to efficiently remove the reaction heat by FT reaction, but fixed-bed reactor process remains an attractive approach because it has the highest volumetric catalyst loading volume (catalyst loading/ reactor volume), the highest potential of productivity (Davis, 2005; Mazidi et al., 2013). Moreover, it doesnot require separating the catalyst from the products and is easy to scale up from a single tube test (Yang et al., 2010), which has been persuasively exemplified by the large-scale commercial operations of Sasol and Shell (Dry, 1996; Sie, 1998). In multitubular fixed-bed reactor, a large amount of reaction heat is generated at the entrance of the catalyst bed, where the reactant partial pressures and temperature rapidly declined due to increased water partial pressure (Raq et al., 2011; Sharma et al., 2011; Kölbel, 1959; Wang et al., 2003). To remove the reaction heat and increase the syngas conversion, the multitubular fixed-bed reactor contains many tubes filled with catalyst immersed in boiling water for heat removal, or a considerable fraction of the liquid reaction products has to be recycled to the fixed reactor, which increases pressure drops and makes the reactor more tricky to be operated and less flexible to be scaled up (Jess and Kern, 2009; Sie and Krishna, 1999). Another disadvantage in fixed-bed multitubular reactors is low catalyst utilization (Wu et al., 2010). It is vitally necessary to maximize the products value for diesel distillate and carbon utilization efficiency with rising energy and raw material prices. In FTS, water is the major byproduct. Many researchers have investigated its effects on the catalyst performances in various reaction systems. The water may affect the syngas conversion, hydrocarbon selectivity, product distribution and catalyst longevity by changing syngas adsorption, chain initiation, chain growth, methanation, hydrogenation to paraffins and dehydrogenation to olefins (Dalai and Davis, 2008). The effect of water on the activity of cobalt catalysts depends on catalyst composition, nature of the support, catalyst preparation method and pretreatment. The water effects are reported to be either negligible (Huber et al., 2001), negative (Rahimpour et al., 2011a, 2011b; Bezemer et al., 2010), or positive (Lualdi et al., 2011; Bertole et al., 2002). The negative is linked to the formation of inactive oxides of cobalt or the formation of irreducible cobalt support compounds. Furthermore, water and heavy hydrocarbons are main sources of catalyst deactivation (Bezemer et al., 2010; Zhang et al., 2006; Van Berge et al., 2000). It is well known that the catalyst deactivation is not significant at

low partial pressures of water, but a much more rapid decrease in conversion can be observed in the presence of more significant amounts of water, because it increases the CoO surface coverage and leads to a more severe sintering. Sadeqzadeh et al. (2012) confirmed the absolute water pressure increase in the reactor and the high P H2 O =P H2 ratio will enhance cobalt sintering. Therefore, keeping the P H2 O =pH2 below its critical value by removing the water may enhance the catalyst lifetime significantly. Herein, to maximize reactor volume productivity, a process based on a two-stage multitubular fixed-bed reactor is presented to enhance CO conversion and diesel production from FTS, which is characterized by the condensation and separation of liquid products and water from the outlet of the first-stage reactor. A two-dimensional pseudohomogeneous reactor model is proposed to simulate the temperature profiles and CO conversion profiles of the two-stage multitubular fixed-bed reactor process, which demonstrates the understandable products distribution shift toward diesel-range hydrocarbon. The process is investigated in a bench-scale single-stage and two-stage multitubular fixed-bed reactor using cobalt-based catalyst. The results are helpful to obtain more valuable diesel products and effective utilization of syngas by simple multi-stage multitubular fixed-bed process, which provide a practical case for multi-stage short reactor rather than one long reactor for FTS to improve products distribution, as well as increase the carbon efficiency.

2. Experimental section 2.1. Material preparation 2.1.1. Preparation of Al2O3-modified SiO2 support The SiO2 support (Qingdao Haiyang Chemical Co., average particle size Ø 3.2 mm) was soaked with 0.05 mol/L HNO3 for 3 h by slowly stirring, and washed three times with distilled water. Then the resulting SiO2 support was dried at 373 K for 10 h, followed by calcining at 873 K for 5 h. The pretreated SiO2 support was modified with calculated amounts of Al(NO3)3 9H2O by incipient wetness impregnation, and dried at 323 K for 24 h and at 373 K for 12 h, then calcined at 873 K for 8 h. 2.1.2. Preparation of ZrO2 promoted Co3O4/Al2O3–SiO2 catalyst The resulting Al2O3-modified SiO2 was used as support, and prepared by the incipient wetness co-impregnation method with an aqueous solution of Co(NO3)2 6H2O and ZrO(NO3)2 6H2O (Industrial products, Shandong Zibo Zhaoyi Chemical Co.) for 24 h, followed by drying at 313 K for 10 h, and at 383 K for 2 h, then calcined at 723 K for 10 h. 2.2. Material characterization 2.2.1. X-ray diffraction characterization The crystal structures were determined by a powder X-ray diffractometer (Shimadzu XRD 6000), using Cu Kα radiation (λ ¼1.54184 Å) combined with a nickel filter operating at 40 kV and 10 mA. The diffractometer data were recorded for 2θ values between 101 and 901 with a scanning rate of 4deg/min. 2.2.2. Surface area and pore volume Nitrogen adsorption/desorption was measured by using a Micromeritics, ASAP 2405N analyzer at  196 1C in liquid nitrogen. Prior to N2 sorption measurements, the samples were degassed at 250 1C, under vacuum, for at least 16 h. The Brunauer–Emmett– Teller (BET) surface area was calculated using experimental points at a relative pressure of P/P0 ¼0.05–0.30. The pore size distribution

X.P. Dai et al. / Chemical Engineering Science 105 (2014) 1–11

was derived from the adsorption branch using the Barret–Joyner– Halenda (BJH) method. 2.2.3. Temperature programmed reduction (TPR/MS) TPR/MS was performed on a fixed-bed quartz microreactor (4.5 mm i. d.) packed with 300 mg samples. The sample was dehydrated for 1 h at 150 1C in pure Ar with 30 ml/min, then the mixture of 10 vol% H2/He was used at a gas flow rate of 30 mL/min, and linear temperature was increased at a rate of 15 1C/min. The analysis of the reactants was carried out using an on-line mass spectrometer (AMTEK dycor System 1000). The quadrupole mass spectrometer can detect eight mass channels simultaneously with a minimum dwell time of 3 ms. 2.3. Bench-scale Fischer–Tropsch synthesis process and reactor The Fischer–Tropsch block of GTL process consists of two-stage multitubular fixed-bed reactor, circulating water systems and product collection systems (as shown in Fig. 1 and Supporting Information Fig. 1s). The each stage of two-stage multitubular fixed-bed reactor is kept with equal sizes and catalyst loadings. Each

3

stage comprises five tubes of 6 m in height, 32 mm in external diameter and 3 mm in thickness, which are arrayed in a stainless steel tube with inner diameter of 120 mm (see Supporting Information Figs. 2–4s). The temperature can be measured by thermocouples, which are set along the axial center and axial coordinate. The output signals from temperature controller are converted into digital signal and transferred to the computer by the Ethernet (Supporting Information Fig. 5s). The heat-transfer medium is continuously circulated around the stationary tube bundle, in which the heat of reaction is supplied or removed according to the reactor temperature. The rising saturated steam fills the jacket, and a circulation flow is established due to the difference in density in the downpipe and in the reactor jacket. The saturated steam temperature is regulated by the saturated vapor valve pressure. Gas (including unreacted synthesis gas, methane, CO2 and light hydrocarbon), C5 þ hydrocarbon and water from the first stage outlet are separated by three-phase separator. The liquid hydrocarbon and wax are collected in tank 4 at the outlet of the first stage. Then, the gas from the first-stage reactor enters the second-stage reactor from the bottom. The two-stage multitubular fixed-bed reactor is singlepass reactor without recycling. As a contrast, the single-stage multitubular fixed-bed reactor with five tubes of 12 m in height, 32 mm in

Fig. 1. Process scheme of the bench-scale two-stage fixed-bed reactor for Fischer–Tropsch synthesis (1) compressor, (2) heating device, (3) reactor, (4, 6) liquid product collector, (5) three-phase separator, (7) water collector, (8) condenser, (9) heater, (10) water storage tank, (11) saturated vapor valve, (12) liquid level gauge, and (13) vent valve.

Table 1 Operating condition and reactor parameters for bench-scale Fischer–Tropsch synthesis. Parameter

Value

Two-stage multitubular reactor dimension (mm) Single-stage multitubular reactor dimension (mm) Multitubular number for each Internal diameter of single tube, dt (mm) Total catalyst loading (L) Diameter of spherical cobalt-catalyst particle, dp (mm) Bulk density of catalyst bed, ρ (kg/m3) Bed void fraction, ε Gas viscosity, μ (Pa s) Heat capacity of gas mixture (feed), cp (kJ/(kg K)) Heat capacity of water, cp′ (kJ/kg K) Effective radial thermal conductivity, λe (J(m s K)) Reynolds number (Re) Reaction pressure (MPa) Cooling temperature (K) Inlet temperature (K) Synthesis gas flow (kg/m2/s) Inlet gas composition (%)

Ø 32  3  6000 for each stage Ø 32  3  12000 5 26 30 3.20 630 0.42 4.54  10  5 2.70 4.22 3.6 24.9 1.50 486 493 0.205 63.11H2/30.76CO/2.37CH4/0.62CO2/3.14N2

4

X.P. Dai et al. / Chemical Engineering Science 105 (2014) 1–11

external diameter and 3 mm in thickness is also investigated. The amount of catalyst loading in single-stage multitubular fixed-bed reactor is same as the total amount of two-stage multitubular fixedbed reactor, and the operating conditions are shown in Table 1. The 15%Co–0.2Zr/3%Al2O3–SiO2 catalyst was packed in the tubes about 15 L for each stage. Prior to the reaction, the catalyst was reduced in H2 at 450 1C for 50 h, which was heated by hot-air in the shell using electric furnace. Then the catalyst was cooled to room temperature, and the gas was switched to the feed gas. In our previous work, the ZrO2–Co3O4/Al2O3–SiO2 catalyst exhibits excellent performance with CO conversion of 80.5% and C5 þ selectivity of 74.9% in micro-reactor stability test, and the kinetics of CO consumption has been found. Syngas is prepared by the decomposition of methanol, purified by cooling, desulfurization and deoxidation, and the composition is shown in Table 1. 2.4. Analysis of the products The reactant and gas products are analyzed by online gas chromatographs HP-5890 equipped with Carbon-sieve column and Porapak-Q column. CO, N2, CH4, and CO2 are analyzed by carbon-sieve column with TCD detector. Gaseous hydrocarbons (C1–C5) are analyzed by Porapak Q column with FID detector. Liquid product and wax in the traps is collected and analyzed by a FID GC equipped with an HP-101 capillary column. The carbon monoxide conversion and product selectivity are calculated by these formulas as follows: CO conversion (%) ¼[(moles of inlet COmoles of outlet CO)/moles of inlet CO]  100%) ¼ [1  (CO/N2)out/(CO/N2)in]  100 CO2 selectivity (%)¼[moles of CO2 produced/(moles of inlet CO moles of outlet CO)]  100 CH4 selectivity (%) ¼[moles of CH4 produced/(moles of inlet CO moles of outlet CO)]  100 C2–C4 selectivity (%)¼ [moles of C2–C4 produced/(moles of inlet CO moles of outlet CO)]  100 C5 þ selectivity ¼ð1  SCH4  SC2 –C4 SCO2 Þ  100

3. Fischer–Tropsch multitubular fixed-bed reactor model 3.1. Mass balance and heat balance In order to predict the temperature and concentration profiles of the multitubular fixed-bed reactor, a steady state two-dimensional (2D) pseudo-homogeneous fixed-bed reactor model is presented, which is formulated using cylindrical coordinates. In this steady-state model, a radial dispersive plug flow is described by mass and heat balance equations. In view of the operation characteristics of the fixed-bed reactor, the following assumptions are introduced: (1) the plug flow regime is assumed with no channeling along the bed; (2) the axial dispersion of mass and heat as well as radial concentration gradients are ignored; (3) the radial temperature profile is considered, and the corresponding radial heat transfer within the catalyst bed is expressed in terms of constant effective radial heat conductivity (λe); (4) the temperature of the cooling water is assumed to be constant; (5) the heavy wax is limited for Co-based catalysts in our experiments where the build-up of a wax layer on the catalyst surface is negligible for mass and heat transfer. For a two-dimensional axisymmetric plug flow inside the fixed bed reactor, when mass axial dispersion is negligible, the steady

state mass balance equation is  2  ∂x ∂ x 1 ∂x ρr c u  De  þ ¼0 ∂z r ∂r c0 ∂r 2

ð1Þ

where ðρr c =c0 Þ denotes the consumption of syngas when FTS reaction occurs. The radial effective diffusion coefficient (De) can be taken from a correlation in reference (Wilhelm, 1962).   ∂x De ∂2 x 1 ∂x ρ   þ rc ¼ 0 ð2Þ ∂z r ∂r u ∂r 2 uc0 The two-dimensional heat balance over fixed-bed reactor for FTS is  2  ∂T ∂ T 1 ∂T ∂ ∂T  λe  λe þ ρr c ΔΗ ¼ 0 þ ð3Þ GC p ∂z r ∂r ∂r ∂r ∂r 2 The radial effective thermal conductivity (λe) can be estimated from the following well-known correlation for low thermal conductivity catalyst particles:

λe ¼ 0:199 þ 0:015

dt 0:331 þ Re dp 1 þ 8:54ð dt Þ2 dp

ð4Þ

where, Re ¼

dp ρu0

μ

1 dp G ¼ 1ε μð1  εÞ

ð4Þ

If effective thermal conductivity (λe ) is constant for radial position, it means ð∂λe =∂rÞ=0. The above equation is given as  2  ∂T ∂ T 1 ∂T  λe þ ρr c ΔΗ ¼ 0 þ ð5Þ GC p ∂z r ∂r ∂r 2 where ρr c ΔΗ is the reaction heat released by unit volume of reactants. ΔH can be calculated by PROII software (see Supporting Information, Fig. 6s). The initial conditions and the boundary conditions for Eqs. (2) and (5) are given as for z ¼ 0 and 0 rr rr 0 ; xA ¼ xA0 ; T ¼ T in for r ¼ 0 and 0 r z rL; ∂x∂rA ¼ 0; ∂T ∂r ¼ 0 2π r 0 λe ∂T for r ¼ r 0 and 0 r z r L; ∂x∂rA ¼ 0, ∂T ∂z ¼ FC0p ∂r 3.2. Numerical methods The Crank–Nicolson finite difference method combining with Matlab was applied to solve equations of mass balance and heat balance to obtain the temperature and concentration distribution. Here, the initial r and z are calculated from the axial center and bed inlet. A finite change for catalyst bed in axial direction and radial direction will be referred as Δr and Δz respectively. The subscripts “i and j” indicate the increment number in axial direction and radial direction. The primary and secondary difference of radial temperature and conversion at collocation point (r, z) are carried out, respectively, while only first difference of axial temperature and conversion is performed at collocation point (r, z). The result is shown as follows:   Δz λ e 1 T i;j þ 1 ¼ T i;j þ ðT T  2T þ T þ  T Þ i þ 1;j i;j i  1;j i;j i i þ 1;j ðΔrÞ2 GC p 

ρB ðr c ÞðΔHÞΔz

ð6Þ

GC p

The temperature and conversion are axial symmetry at i¼ 0, which results that ð1=rÞð∂T=∂rÞis indefinite. According to L’ Hôpitol's rule, the limð1=rÞð∂T=∂rÞ ¼ ð∂2 T=∂r 2 Þ can be calculated, and the r-0 temperature at axial center is as follows: T 0;j þ 1 ¼ T 0;j þ

2Δzλe

ðΔrÞ2 GC p

ð2T 1;j  2T 0;j Þ 

ρB ðrc ÞðΔHÞΔz GC p

ð7Þ

X.P. Dai et al. / Chemical Engineering Science 105 (2014) 1–11

Describing the F–T reaction kinetics is quite difficult due to its complicated reaction mechanisms and a large number of chemical species involved. Besides those problems, chemical kinetics studies are affected by many factors, including the catalyst's preparation method, metal loading and support, etc. F–T kinetics studies can be categorized by three different approaches: Mechanistic proposals consisting of a sequence of elementary reactions among surface adsorbents and/or intermediates; Empirical expressions of general power-law kinetics; Semiempirical kinetics expression based on F-T mechanisms. Traditionally, due to low activity of water gas shift (WGS) reaction, the rate expressions for Co-based FT catalysts have only contained the partial pressures of H2 and CO, in which, generally, the reaction order in P H2 is positive while that in PCO is negative (Visconti et al., 2011; Zhen et al., 2009; Zennaro et al., 2000; Yates and Satterfield, 1991; Sarup and Wojciechowski, 1989; Outi et al., 1981). Therefore, the CO consumption kinetics equation incorporated into the resulting model is estimated according to the following irreversible power-law kinetics as Eq. (8): n r c ¼ Ae  Ea=RT P m H2 P CO

ð8Þ

The kinetic parameters A, Ea, m and n are estimated for the adopted catalyst through a non-linear regression on a comprehensive set of FTS experiments carried out. The parameters are obtained as follows: A¼2.45  106 mol/(kg-cat/(MPa0.3 s)), Ea¼ 73 kJ/mol, m¼1.3, n¼  1. The CO consumption kinetics equation is exactly similar to those of many researchers (Visconti et al., 2011; Zhen et al., 2009; Zennaro et al., 2000; Yates and Satterfield, 1991; Sarup and Wojciechowski, 1989; Outi et al., 1981). The negative reaction order with respect to CO and the positive reaction order with respect to H2 are well aligned with the literature data over supported cobalt catalysts (Visconti et al., 2011; Zhen et al., 2009; Zennaro et al., 2000). The activation energy is consistent with the reported value in the literature for cobalt catalysts (Wu et al., 2010).

4. Results and discussions 4.1. Materials physicochemical characterization The isotherms of nitrogen adsorption/desorption and the corresponding pore distribution calculated by the BJH method

5

are shown in Fig. 2. The BET surface area, average pore diameters and mesopore volume are presented in Table 2. All samples exhibit an obvious Langmuir type IV with H1 hysteresis loop in the isotherm curves, indicating that the supports and catalyst are composed of mesopore with 1D cylindrical channel. The steepness of the capillary condensation step of SiO2 indicates uniformity of mesopores with average pore diameter of 2.6 nm. After it is washed with 0.05 mol/L HNO3, and modified with 3% Al2O3, the average pore size is widen to 12.4 m, but it gives wide pore distribution from 4 to 20 nm. It is clearly seen for ZrO2–Co3O4/ Al2O3–SiO2 catalyst that the hysteresis loop shifts to the lower relative pressure, which reflects the decrease of average pore diameter. The ZrO2–Co3O4/Al2O3–SiO2 catalyst has a narrow Gaussian-like unimodal pore distribution. The BET surface area in mesoporous SiO2 is higher than 849 m2/g, whereas the BET surface areas of Al2O3–SiO2 and ZrO2–Co3O4/Al2O3–SiO2 are much lower (253.7 and 179.6 m2/g, respectively). Cobalt and zirconium introduction results only in a slight decrease in pore volume and average pore size relative to the initial Al2O3–SiO2 support. The reducibility of the calcined catalyst is also investigated by temperature programmed reduction, and the reduce profiles are presented in Fig. 3. Two main reduction peaks, centered approximately at 596 and 651 K, are observed in the temperature range of 500–650 K. They are correspond to the two-step reduction process in which the Co3O4 phase is first reduced to CoO and then the CoO phase is reduced to Co0. The catalyst exhibits one unique Co oxide species: a reduction feature with a shoulder at 753 K. Li et al. (2006) and González et al. (2009) attributed the signals to easily reducible surface-Co species at the reduction conditions, or that can be considered as the reduction of the smaller cobalt oxide particles, which can increase the number of active sites available for FTS. The broad reduction shoulder at 880 K could be attributed to the reduction of cobalt silicates formed by the strong interaction between cobalt and the siliceous support, which are stable and difficult to reduce and to detect by XRD characterization. The XRD patterns of ZrO2–Co3O4/Al2O3–SiO2 catalyst are shown in Fig. 4. The fresh catalyst shows five sharp signals of Co3O4 crystalline structure (JCPDS 43-1003) and on broad signal with low intensity located at 2θ ¼231 typical of siliceous materials. Cobalt silicate (Co2SiO4) or cobalt aluminate (CoAl2O4) is not detected over fresh catalyst due to the small crystallite size (Jabłoński et al., 1998).

Fig. 2. Nitrogen adsorption–desorption isotherms and pore size distribution curves of SiO2 (1), Al2O3–modified SiO2 (2) and ZrO2–Co3O4/Al2O3–SiO2 and (3) catalyst for the bench-scale Fischer–Tropsch synthesis.

6

X.P. Dai et al. / Chemical Engineering Science 105 (2014) 1–11

Sample

SiO2 Al2O3–SiO2 ZrO2–Co3O4/ Al2O3–SiO2

Surface area (m2 g  1)

BJH pore diameter (nm)

Mesopore volume (cm3 g  1)

849.4 253.7 179.6

2.6 12.4 10.6

0.56 0.72 0.60

651 K

Temperature profile at axial center (K)

520

Table 2 Textural property of supports and catalyst.

: Experimental values : Calculated values

515 510 505 0.7 m Hot pot

8.2 m Hot pot

500 (b)

495

493 K

490

(a)

The inlet of the second-stage reactor

485

7.2 m

H2 consumption/ a.u.

480

0

2

596 K

4

6

8

10

12

Axial coordinate (m) Fig. 5. Comparison between experimental values and calculated values of temperature of axial center and axial coordinate over single-stage (a) and two-stage (b) fixed-bed reactor for bench-scale Fischer–Tropsch synthesis.

753 K

100

880 K

(b) 80

300

400

500

600

700

800

900 1000 1100 1200

Temperature / K Fig. 3. H2-TPR profiles for ZrO2–Co3O4/Al2O3–SiO2 catalyst for the bench-scale Fischer–Tropsch synthesis.

60

40 Calculated values Experimental values

20

Co

Co3O4

CO conversion (%)

(a)

0

Intensity (a.u.)

0

(c)

(b)

(a)

10

20

30

40

50

60

70

80

90

2θ /° Fig. 4. XRD patterns of (a) fresh catalyst, (b) post-reacted catalyst at the bottom of the catalyst bed over two-stage fixed-bed reactor, and (c) post-reacted catalyst at the bottom of the catalyst bed over single-stage fixed-bed reactor for bench-scale Fischer–Tropsch synthesis.

No characteristic peaks of any zirconium and alumina compounds are observed, indicating that zirconium and alumina are well dispersed on the SiO2 support. 4.2. Model validation The Crank–Nicolson difference method combining with Matlab was applied to solve equations of mass balance and heat balance to obtain the temperature and concentration distribution on the basis of the bench-scale test data. The governing equations of the twodimensional model formed a set of stationary differential an algebraic equations coupled with the non-linear algebraic equations of the

2

4

6

8

10

12

Axial coordinate (m) Fig. 6. Comparison between experimental values and calculated values on CO conversion on single-stage (a) and two-stage (b) fixed-bed reactor for bench-scale Fischer–Tropsch synthesis.

two-dimensional model for mass balance and heat balance. For twostage fixed bed reactor, the syngas partial pressure in second-stage reactor is decided by the separation of liquid and water at room temperature. The concentration and temperature profiles of experimental and calculated values in the axial direction are obtained in Figs. 5 and 6. It can be seen that the simulated values of CO conversions and temperature are in good agreement with the experimental values. It is evident that the simulation results are close to the experimental values. So the two-dimensional pseudohomogeneous mathematical model on single-stage and two-stage multitubular fixed-bed reactor for FTS is reasonable and reliable. 4.3. Temperature profiles in single-stage and two-stage multitubular fixed-bed reactor The axial and radial temperature profiles in single- and twostage multitubular fixed-bed reactor are shown in Fig. 7. The simulated results clearly indicate that the temperature at axial center along the length of the reactor increases (i.e. creating a hot spot) to 503.1 K at the reactor inlet about 0.7 m, but the temperature soon continuously drops to the given wall temperature over single-stage fixed-bed reactor. The hot spot temperature at axial center is 10.1 K higher than that of inlet temperature. Due to decreasing syngas partial pressure and reaction rate along the

X.P. Dai et al. / Chemical Engineering Science 105 (2014) 1–11

7

100

Length=6 m for each Length=12 m

504

80

498 496 12.5 7.5 2.5

(m m)

494 492 490

-2.5

488 0.0 0.4 0.6 1.2

3.0 6.0 7.2 Reactor length (m)

-7.5 9.0

10.2 11.4

-12.5

CO conversion (%)

500

Ra dia lc oo rdi na te

Temperature (K)

502

Length=6 m 60

40

20

0 A

504

Fig. 8. CO conversion on single-stage (A,B) and two-stage (C) fixed-bed reactor for bench-scale Fischer–Tropsch synthesis.

502 500 498

100

496

Single-stage fixed-bed reactor Two-stage fixed-bed reactor

490 488

12.5 7.5 2.5

(m m)

492

-2.5 0.0 0.4 0.6 1.2 -7.5 3.0 6.0 6.6 7.2 8.2 9.0 Reactor length 10.2 11.4 -12.5 (m)

Fig. 7. Temperature profiles over single-stage (a) and two-stage (b) fixed-bed reactor for bench-scale Fischer–Tropsch synthesis.

reactor, the reaction heat is quickly removed by cooling water, which results that the temperature at axial center is lower than syngas inlet temperature after 7.2 m of single-stage fixed-bed reactor. Radial temperature gradually decreases with radius increase, and also decreases continuously along reactor length increase. Furthermore, the axial temperature difference is lower gradually with reactor length increased. Compared with singlestage fixed-bed reactor, the axial temperature profile will be changed on two-stage fixed-bed reactor due to condensation of the products and water from the first reactor, which makes the syngas concentration increased significantly over the second stage of the reactor. The increase of syngas partial pressure corresponds to the increase of syngas concentration, which will lead to reappearance of the hot spot. Compared with single fixed-bed reactor (0.7 m), its position will move to the rear part of the second stage of the reactor (2.2 m). The hot spot temperature at axial center is 9.5 K higher than that of inlet temperature. It is lower slightly than that of the first hot spot at single-stage fixedbed reactor. The axial temperature profile also makes radial temperature profile changed simultaneously. The higher axial and radial temperature makes the bulk temperature increased in the second-stage reactor. Obviously, the temperature over the second stage of the reactor is higher than that of single-stage fixed-bed reactor at the same position. This means that the syngas conversion can be enhanced over two-stage fixed-bed reactor, which can release more reaction heat. The improvement of temperature profiles will contribute favorably to utilize syngas as efficiently as possible. 4.4. Improving diesel-range production in two-stage multitubular fixed-bed reactor The experimental results of 300 h time-on-stream in singlestage and two-stage multitubular fixed-bed reactor are shown in

Conversion and selectivity (%)

494

Ra dia lc oo rdi na te

Temperature (K)

C

B

80

60

40

20

0

X-CO

S-C5+

S-C2-4

S-CH4

S-CO2

Fig. 9. Comparison of CO conversion and selectivity on single-stage and two-stage fixed-bed reactor for bench-scale Fischer–Tropsch synthesis.

Figs. 8 and 9 with syngas GHSV 150 h  1, 1.5 MPa, inlet temperature 493 K, cooling water temperature 486 K and flow 780 kg h  1. The CO conversion is about 64% over the first stage of the reactor, while only 15% CO conversion are obtained over the second stage from Fig. 8. Due to condensation and removal of liquid hydrocarbon and water from the outlet of the first-stage of the reactor, the partial pressure of syngas significantly increases compared with single fixed-bed reactor from 6 m to 12 m in the reactor (see Supporting Information, Fig. 7s). Bartholomew and Farrauto (2005) also confirmed that P H2 O 40.6 MPa can be easily exceeded at medium conversion levels (XCO 460%). Water can dominate the gas phase at higher conversion levels and thus decrease the partial pressure of reactants and residence time by dilution. Although most reported water effects on the FT rate of SiO2-supported Co catalysts are positive (Krishnamoorthy et al., 2002; Das et al., 2005; Storsæter et al., 2005), It should be mentioned that the syngas partial pressure is kept constant by means of an Ar pocket. van Steen and Schulz (1999) proposed that the kinetic effect of water on FT rate for Co catalysts is positive at low partial water pressures, while turning to negative at higher partial pressures. Small amounts of water are suggested to clean the catalyst surface carbon inhibiting the FT reaction, while large amounts of water reduce the active surface carbon concentration. Borg et al. (2006) found water amounts equal to P H2 O =P H2 ¼ 0:7 at the reactor inlet suppressed the reaction rates and led to permanent deactivation. Storsæter et al. (2005) concluded that, above a certain water

8

X.P. Dai et al. / Chemical Engineering Science 105 (2014) 1–11


partial pressure pH2 O =pH2 4 0:6 , the effects of water on CO conversion are generally negative and irreversible for all catalysts. In this case, the CO conversion is about 64% at the first stage of the reactor, which produce about 0.6 MPa for the partial pressure of water (nearly 50% molar fraction, see Supporting Information,

100

CO conversion (%)

90 80 70 Single-stage fixed-bed reactor Two-stage fixed-bed reactor

60 50 40 100

150

200 250 Time on stream (h)

300

Fig. 10. CO conversion and selectivity vis time on stream on single-stage (A) and two-stage (B) fixed-bed reactor for bench-scale Fischer–Tropsch synthesis.

Fig. 7s). The above results suggest that the feed gas in the second stage reactor is similar to that of the large amount of addition of water if we use one long reactor for FTS. The large amount of water will exert negative influence on CO conversion (van Steen and Schulz, 1999). On the other hand, the removal of water and liquid hydrocarbon will decrease mass-transport restrictions, which has been explained by the fact that an intra-pellet hydrocarbon-rich phase would result in a higher diffusivity of the reactants through the wax-filled pores. Therefore, the removal of liquid hydrocarbon and water between the two-stage reactors not only increase the syngas partial pressure (including H2), but also increase the partial pressure of low hydrocarbon (such as olefin) and diffusivity of the reactants in the second-stage reactor. The increase of syngas partial pressure (especially for H2) and improvement of temperature profile significantly increase CO conversion to 89%. The significant effect on the enhancement of CO conversion by syngas partial pressure and reaction temperature has been observed by many researchers (Rohde et al., 2008; Rahimpour et al., 2011a, 2011b; Wang et al., 2003). Although the Co-based catalyst exhibit low WGS reaction in FTS, the WGS reaction is an equilibrium reaction. Under FT conditions, WGS reaction of CO toward CO2 is thermodynamically favorable. It is possible to decrease WGS reaction by water removal from the outlet of the first reactor, which results in CO2 selectivity slightly decreased (see Fig. 9). Mazzone and Fernandes (2006) observed that the rate of WGS reaction is coupled to the H2O partial pressure. Botes (2009) indicated CO2 selectivity solely increases with increasing water

Fig. 11. The distribution of liquid hydrocarbon over single-stage (A) and two-stage (B) fixed-bed reactor for bench-scale Fischer–Tropsch synthesis.

X.P. Dai et al. / Chemical Engineering Science 105 (2014) 1–11

partial pressure. Rahimpour et al. (2011a) also suggested CO2 selectivity decrease with the removal of water by a novel water perm-selective membrane dual-type reactor. Therefore, the decrease of water partial pressure is favorable to decrease CO2 selectivity. In general, past studies have shown that methane selectivity also increases with reaction temperature. Chernobaev et al. (1997) concluded that the formation of unwanted CH4 could be explained on the basis that the surface species desorbed rather than propagated to higher molecular weight compounds at increased temperature. However, the decrease of CH4 selectivity and increase of C5 þ selectivity is obtained over two-stage multitubular fixed-bed reactor. Botes (2009) observed that increasing syngas partial pressure contributes favorably to the formation of heavy hydrocarbons and improvement of reaction activity over alumina-supported cobalt catalyst. According to the reaction network in the FTS, a primary α-olefin can react in two ways: it can either be secondarily hydrogenated to a paraffin or readsorb with subsequent chain initiation. It is generally assumed that readsorption of olefins is an important part of the Fischer–Tropsch reaction network, and addition of olefins to the syngas feed has some positive effect on the long-chain paraffin yield (Rytter et al., 2007). In this case, the olefin/paraffin (o/p) ratio of C2 hydrocarbon just slightly decrease from 0.15 to 0.13 (about 13%), but the total selectivity of C2 hydrocarbon dominantly decrease from 2.7% to 1.6%, compared with the one long reactor in our experiments, which suggest it is dominantly linked to the C5 þ selectivity under the reaction conditions investigated here. Rahimpour et al. (2011a, 2011b) also observed a little effect on the production rates of light hydrocarbons (such as ethene) by the in-situ water removal over fixed-bed water perm-selective membrane reactor. Thus, readsorption of olefins with subsequent chain initiation may play an important role for the FT reaction in our experiments. The stability of the catalyst vs. time on stream and the analysis of liquid hydrocarbon collected after 300 h from single-stage and two-stage fixed-bed reactor are shown in Figs. 10 and 11a, b and c, respectively. The CO conversion decreased from a maximum value of 80–73% ( 8.8%) over single-stage fixed-bed reactor, while it slightly decreased from a maximum value of 89–85% (  4.4%) during 100–300 h on stream over two-stage fixed-bed reactor. The deactivation rate over single-stage fixed-bed reactor is more than 2-fold compared to two-stage fixed-bed reactor. The main difference between the above processes is that water and liquid hydrocarbon has been removed at the two-stage fixed-bed reactor process, which indicates that the water and liquid hydrocarbon exert negative effects on the stability. Though Bezemer et al. (2010) provide direct evidence of water-assisted sintering of cobalt on carbon nanofiber catalysts, or the possible re-oxidation of cobalt active sites at high partial pressure of water (Tsakoumis et al., 2012) to explain the deactivation of cobalt-based catalyst, we donot observe the above behaviors from XRD characterization (Fig. 4b and c). The dominant phase was still metallic Co0 (41.61, 44.31 and 75.61, JCPDS Card no. 89-4308) on post-reacted catalysts from the bottom of catalyst bed over single-stage and two-stage fixed-bed reactor for bench-scale FTS. The characteristic peaks of CoO are not also detected over post-reacted catalysts. The spectrum for liquid hydrocarbon displays the usual distribution on Co-based catalysts where a series of linear n-paraffins are the major products at each carbon number. The condensed products are composed mainly of motor fuel range hydrocarbons (C5–C22) as shown in Fig. 10C. Compared with the results from single-stage fixed-bed reactor, the fraction of condensed products in the diesel range (C12–C22) increased significantly from 67% to 78%, while the gasoline range (C5–C11) decreased from 27% to 18% over two-stage fixed-bed reactor. Such hydrocarbon distribution makes chain growth probability (α) increased from 0.81 to 0.88. Despite higher water partial pressure having positive effects on CO conversion

9

and C5 þ selectivity in some studies (Storsæter et al., 2005; Zhen et al., 2009; Lödberg et al., 2011), the results were usually obtained at constant syngas partial pressure and by external water vapor addition under simulating reaction condition, which might lead to distortion of the reason for changing activation. In our experiments, the total pressure keep constant and water is removed by cooling between two stages of the fixed-bed reactor, which suggests that the syngas partial pressure increased significantly from 0.6 MPa to 1.2 MPa (see Supporting Information Fig. 5s). The removal of water and liquid hydrocarbon, higher syngas partial pressure as well as optimized temperature distribution not only give rise to an improvement in the CO conversion, but also make a important shift to higher molecular weight hydrocarbons products distribution. The above results indicate that, in order to utilize syngas as efficiently as possible, simple multiple-stage and short multitubular reactor should be suggested rather than single-stage and long one for FTS.

5. Conclusion The process based on a bench-scale two-stage multitubular fixed-bed Fischer-Tropsch reactor, which characterized by the condensation and separation of liquid products and water from the outlet of the first-stage reactor, is investigated for enhance in conversion and diesel selectivity, and a two-dimensional pseudohomogeneous model is developed. Simulations of a wall-cooled single-stage and two-stage multitubular FT reactor with cobalt as catalyst demonstrate that the reaction performance is distinctly improved by increasing the temperature and syngas partial pressure in the second stage of reactor. The combined effects significantly increase the CO conversion, C5 þ selectivity and diesel production (C12–C22) in two-stage multitubular fixed-bed reactor. The catalysts exhibit more stability and less deactivation rate over two-stage fixed-bed reactor process as the condensation and separation of liquid products and water. The results are helpful to design multitubular fixed-bed reactor for FTS with simple multiple-stage and short multitubular reactor rather than singlestage and long one.

Nomenclature A C Co Cp Cp' De Ea F G ΔrHm L m, n P R Re r0 rc T Tin Ti,j u x z

pre-exponential factor concentration of syngas, mol/m3 inlet concentration of syngas, mol/m3 heat capacity of mixed gas, kJ/(kg K) heat capacity of water, kJ/(kg K) radial effective diffusion coefficient, m2/s activity energy, kJ/mol circulating water flow, m3/s syngas mass flow, kg/m2/s syngas reaction enthalpy, kJ/mol reactor length, m correlation parameters partial pressure, Pa molar gas constant, kJ/(K mol) Reynolds number tube radius, mm consumption rate of CO, mol/kg/s/ reaction temperature, K inlet temperature, K temperature of collocation point (i, j), K gas velocity, m/s syngas conversion reactor length, m

10

μ λe ρ ε

X.P. Dai et al. / Chemical Engineering Science 105 (2014) 1–11

gas viscosity, Pa s radial effective heat transfer coefficient, kJ(m s K) stacking density of catalyst bed, kg/m3 bed voidage

Acknowledgments Financial support by the China National Natural Science Foundation Programs (Grant no. 21076226), the Science Foundation of China University of Petroleum-Beijing (No. KYJJ2012-03-02) and the China National Petroleum Corp. are gratefully acknowledged. The authors also gratefully acknowledge Dr. Changchun Yu and Prof. Xiaojun Bao, China University of Petroleum (Beijing, China), for the helpful discussion and preparation of some samples.

Appendix A. Supporting Information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.ces.2013.09.057. References Bae, J.W., Park, S.J., Woo, M.H., Cheon, J.Y., Ha, K.S., Jun, K.W., Lee, D.H., Jung, H.M., 2011. Enhanced catalytic performance by zirconium phosphate-modified SiO2supported RuCo catalyst for Fischer–Tropsch synthesis. ChemCatChem 3, 1342–1347. Bartholomew, C.H., Farrauto, R.J., 2005. Fundamentals of Industrial Catalytic Processes. John Wiley & Sons, New York. Bermúdez, V., Lujan, J.M., Pla, B., Linares, W.G., 2011. Comparative study of regulated and unregulated gaseous emissions during NEDC in a light-duty diesel engine fuelled with Fischer–Tropsch and biodiesel fuels. Biomass Bioenergy 35, 789–798. Bertole, C.J., Mims, C.A., Kiss, G., 2002. The effect of water on the cobalt-catalyzed Fischer–Tropsch synthesis. J. Catal. 210, 84–96. Bezemer, G.L., Remans, T.J., van Bavel, A.P., Dugulan, A.I., 2010. Direct evidence of water-assisted sintering of cobalt on carbon nanofiber catalysts during simulated Fischer–Tropsch conditions revealed with in situ mössbauer spectroscopy. J. Am. Chem. Soc. 132, 8540–8541. Borg, O., Storsæter, S., Eri, S., Wigum, H., Rytter, E., Holmen, A., 2006. The effect of water on the activity and selectivity for γ-alumina supported cobalt Fischer– Tropsch catalysts with different pore sizes. Catal. Lett. 107, 95–102. Botes, F.G., 2009. Influences of water and syngas partial pressure on the kinetics of a commercial alumina-supported cobalt Fischer–Tropsch catalyst. Ind. Eng. Chem. Res. 48, 1859–1865. Chernobaev, I.I., Yakubovich, M.N., Tripolskii, A.I., Pavlenko, N.V., Struzhko, V.L., 1997. Investigation of the mechanism of methane formation in the Fischer– Tropsch synthesis on a Co/SiO2-ZrIV catalyst. Theor. Exp. Chem. 33, 38–40. Dalai, A.K., Davis, B.H., 2008. Fischer–Tropsch synthesis: a review of water effects on the performances of unsupported and supported Co catalysts. Catal. Today 348, 1–15. Das, T.K., Conner, W.A., Li, J., Jacobs, G., Dry, M.E., Davis, B.H., 2005. Fischer–Tropsch synthesis: kinetics and effect of water for a Co/SiO2 catalyst. Energy Fuels 19, 1430–1439. Davis, B.H., 2005. Fischer–Tropsch synthesis: overview of reactor development and future potentialities. Top. Catal. 32, 143–168. dela Osa, A.R., De Lucas, A., Sánchez-Silva, L., Díaz-Maroto, J., Valverde, J.L., Sánchez, P., 2012. Performing the best composition of supported Co/SiC catalyst for selective FTS diesel production. Fuel 95, 587–598. Dry, M.E., 1996. Practical and theoretical aspects of the catalytic Fischer–Tropsch process. Appl. Catal. A 138, 319–344. Geerlings, J.J.C., Wilson, J.H., Kramer, G.J., Kuipers, H.P.C.E., Hoek, A., Huisman, H.M., 1999. Fischer–Tropsch technology-from active site to commercial process. Appl. Catal. A 186, 27–40. González, O., Pérez, H., Navarro, P., Almeida, L.C., Pacheco, J.G., Montes, M., 2009. Use of different mesostructured materials based on silica as cobalt supports for the Fischer–Tropsch synthesis. Catal. Today 148 148, 140–147. Huber, G.W., Guymon, C.G., Conrad, T.L., Stephenson, B.C., Bartholomew, C.H., 2001. Hydrothermal stability of Co/SiO2 Fischer–Tropsch synthesis catalysts. Stud. Surf. Sci. Catal. 139, 423–430. Jabłoński, J.M., Wołcyrz, M., Ludwina, K., 1998. On cobalt silicate formation during high-temperature calcination of impregnated cobalt/silica catalysts. J. Catal. 173, 530–534. Jess, A., Kern, C., 2009. Modeling of multi-tubular reactors for Fischer–Tropsch synthesis. Chem. Eng. Technol. 32, 1164–1175.

Jess, A., Kern, C., 2012. Influence of particle size and single-tube diameter on thermal behavior of Fischer–Tropsch reactors. Chem. Eng. Technol. 35, 369–378. Kölbel, H., 1959. Die Fischer–Tropsch synthesis. Winnacker-külchler chemische technologievol. 3. Organische Technologie I, Karl Hauser Verlag, Munich p. 439. Krishnamoorthy, S., Tu, M., Ojeda, M.P., Pinna, D., Iglesia, E., 2002. An investigation of the effects of water on rate and selectivity for the Fischer–Tropsch synthesis on cobalt-based catalysts. J. Catal. 211, 422–433. Li, H.L., Li, J.L., Ni, H.K., Song, D.C., 2006. Studies on cobalt catalyst supported on silica with different pore size for Fischer–Tropsch synthes. Catal. Lett. 110, 71–76. Lira, E., López, C.M., Oropeza, F., Bartolini, M., Alvarez, J., Goldwasser, M., Linares, F.L., Lamonier, J.F., Zurita, M.J.P., 2008. HMS mesoporous silica as cobalt support for the Fischer–Tropsch synthesis: pretreatment, cobalt loading and particle size effects. J. Mol. Catal. A-Chem. 281, 146–153. Lödberg, S., Boutonnet, M., Walmsley, J.C., Järå, S., Holmen, A., Blekkan, E.A., 2011. Effect of water on the space-time yield of different supported cobalt catalysts during Fischer–Tropsch synthesis. Appl. Catal. A 393, 109–121. Lualdi, M., Lögdberg, S., Regali, F., Boutonnet, M., Järås, S., 2011. Investigation of mixtures of a co-based catalyst and a cu-based catalyst for the Fischer–Tropsch synthesis with bio-syngas: the importance of indigenous water. Top. Catal. 54, 977–985. Mazidi, S.K., Sadeghi, M.T., Marvast, M.A., 2013. Optimization of Fischer–Tropsch process in a fixed-bed reactor using non-uniform catalysts. Chem. Eng. Technol. 36, 62–72. Mazzone, L.C.A., Fernandes, F.A.N., 2006. Modeling of Fischer–Tropsch synthesis in a tubular reactor. Lat. Am. Appl. Res. 36, 141–148. Outi, A., Rautavuoma, I., van der Baan, H.S., 1981. Kinetics and mechanism of the Fischer–Tropsch hydrocarbon synthesis on a cobalt on alumina catalyst. Appl. Catal 1, 247–272. Rahimpour, M.R., Mirvakili, A., Paymooni, K., 2011a. A novel water perm-selective membrane dual-type reactor concept for Fischer–Tropsch synthesis of GTL (gas to liquid) technology. Energy 36, 1223–1235. Rahimpour, M.R., Mirvakili, A., Paymooni, K., Moghtaderi, B., 2011b. A comparative study between a fluidized-bed and a fixed-bed water perm-selective membrane reactor with in situ H2O removal for Fischere–Tropsch synthesis of GTL technology. J. Nat. Gas Sci. Eng 3, 484–495. Raq, M.H., Jakobsen, H.A., Schmid, R., Hustad, J.E., 2011. Experimental studies and modeling of a fixed bed reactor for Fischer–Tropsch synthesis using biosyngas. Fuel Sci. Technol. 92, 893–907. Rohde, M.P., Schaub, G., Khajavi, S., Jansen, J.C., Kapteijn, F., 2008. Fischer–Tropsch synthesis with in situ H2O removal-directions of membrane development. Microp. Mesop. Mat. 115, 123–136. Rytter, E., Eri, S., Skagseth, T.H., Schanke, D., Bergene, E., Myrstad, R., Lindvåg, A., 2007. Catalyst Particle Size of Cobalt/Rhenium on Porous Alumina and the Effect on Fischer–Tropsch Catalytic Performance. Ind. Eng. Chem. Res. 46, 9032–9036. Sadeqzadeh, M., Hong, J.P., Fongarland, P., Curulla-Ferré, D., Luck, F., Bousquet, J., Schweich, D., Khodakov, A.Y., 2012. Mechanistic modeling of cobalt based catalyst sintering in a fixed bed reactor under different conditions of Fischer– Tropsch synthesis. Ind. Eng. Chem. Res. 51, 11955–11964. Sarup, B., Wojciechowski, B.W., 1989. Studies of the Fischer–Tropsch synthesis on a cobalt catalyst II. Kinetics of carbon monoxide conversion to methane and to higher hydrocarbons. Can. J. Chem. Eng. 67, 62–74. Sharma, A., Philippe, R., Luck, F., Schweich, D., 2011. A simple and realistic fixed bed model for investigating Fischer–Tropsch catalyst activity at lab-scale and extrapolating to industrial conditions. Chem. Eng. Sci. 66, 6358–6366. Sie, S.T., 1998. Process development and scaleup: IV. Case history of the development of Fischer–Tropsch synthesis process. Rev. Chem. Eng. 14, 109–137. Sie, S.T., Krishna, R., 1999. Fundamentals and selection of advanced Fischer–Tropsch reactors. Appl. Catal. A 186, 55–70. Storsæter, S., Borg, O., Blekkan, E.A., Holmen, A., 2005. Study of the effect of water on Fischer–Tropsch synthesis over supported cobalt catalysts. J. Catal. 231, 405–419. Szeto, K.C., Hardou, L., Merle, N., Basset, J.-M., Thivolle-Cazat, J., Papaioannou, C., Taoufik, M., 2012. Selective conversion of butane into liquid hydrocarbon fuels on alkane metathesis catalysts. Catal. Sci. Technol. 2, 1336–1339. Teiseh, E.A., Capareda, S., Rezenom, Y.H., 2012. Cobalt based hybrid Fischer–Tropsch synthesis catalyst for improved selectivity of hydrocarbons in the JP-8 carbon number range from a synthesis gas obtained from the pyrolysis of the MixAlco process derived sludge. Appl. Catal. A 437–438, 63–71. Tsakoumis, N.E., Voronov, A., Ronning, M., van Beek, W., Borg, O., Rytter, E., Holmen, A., 2012. Fischer–Tropsch synthesis: an XAS/XRPD combined in situ study from catalyst activation to deactivation. J. Catal. 291, 138–148. van Berge, P.J., van de Loosdrecht, J., Barradas, S., van der Kraan, A.M., 2000. Oxidation of cobalt based Fischer–Tropsch catalysts as a deactivation mechanism. Catal. Today 58, 321–334. van Steen, E., Schulz, H., 1999. Polymerisation kinetics of the Fischer–Tropsch CO hydrogenation using iron and cobalt based catalysts. Appl. Catal. A 186, 309–320. Visconti, C.G., Tronconi, E., Groppi, G., Lietti, L., Iovane, M., Rossini, S., Zennaro, R., 2011. Monolithic catalysts with high thermal conductivity for the Fischer– Tropsch synthesis in tubular reactors. Chem. Eng. J. 171, 1294–1307. Wang, H., Zhou, W., Liu, J.X., Si, R., Sun, G., Zhong, M.Q., Su, H.Y., Zhao, H.B., Rodriguez, J.A., Pennycook, S.J., Idrobo, J.C., Li, W.X., Kou, Y., Ma, D., 2013. Platinum-modulated cobalt nanocatalysts for low-temperature aqueous-phase Fischer–Tropsch Synthesis. J. Am. Chem. Soc. 135, 4149–4158.

X.P. Dai et al. / Chemical Engineering Science 105 (2014) 1–11

Wang, Y.N., Xu, Y.Y., Li, Y.W., Zhao, Y.L., Zhang, B.J., 2003. Heterogeneous modeling for fixed-bed Fischer–Tropsch synthesis: reactor model and its applications. Chem. Eng. Sci. 58, 867–875. Wilhelm, R.H., 1962. Progress towards the a priori design of chemical reactors. Pure Appl. Chem. 5, 403–422. Wu, J.M., Zhang, H.T., Ying, W.Y., Fang, D.Y., 2010. Simulation and analysis of a tubular fixed-bed Fischer–Tropsch synthesis reactor with Co-based catalyst. Chem. Eng. Technol. 33, 1083–1092. Yang, J.H., Kim, H.K., Chun, D.H., Lee, H.T., Hong, J.C., Jung, H., Yang, J., 2010. Mass transfer limitations on fixed-bed reactor for Fischer–Tropsch synthesis. Fuel Process. Technol. 91, 285–289.

11

Yates, I.C., Satterfield, C.N., 1991. Intrinsic kinetics of the Fischer–Tropsch synthesis on a cobalt catalyst. Energy Fuels 5, 168–173. Zennaro, R., Tagliabue, M., Bartholomew, C.H., 2000. Kinetics of Fischer–Tropsch synthesis on titania-supported cobalt. Catal. Today 58, 309–319. Zhang, J., Wright, H.A., Jiang, Y., Allison, J.D., York, K.M., Mohedas, S.R., Melquist, V. H., 2006. US Patent 7,001,927. Zhang, Q., Kang, J., Wang, Y., 2010. Development of novel catalysts for Fischer– Tropsch synthesis: tuning the product selectivity. ChemCatChem 2, 1030–1058. Zhang, Y., Xiong, H.F., Liew, K.Y., Li, J.L., 2005. Effect of magnesia on alumina-supported cobalt Fischer–Tropsch synthesis catalysts. J. Mol. Catal. A-Chem. 237, 172–181. Zhen, Y., Zhou, J.W., Bukur, D.B., Goodman, D.W., 2009. Fischer–Tropsch synthesis on a model Co/Si catalyst. J. Catal. 268, 196–200.