Experimental verification of 2-dimensional computational fluid dynamics modeling of supercritical fluids Fischer Tropsch reactor bed

Catalysis Today xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Catalysis Today journal homepage: www.elsevier.com/locate/cattod

Experimental verification of 2-dimensional computational fluid dynamics modeling of supercritical fluids Fischer Tropsch reactor bed Aya E. Abusrafaa, Mohamed S. Challiwalaa,b, Hanif A. Choudhurya, Benjamin A. Wilhiteb, ⁎ Nimir O. Elbashira,c, a

Petroleum Engineering Program & Chemical Engineering Program, Texas A&M University at Qatar, Education City, Doha, 23874, Qatar Texas A&M University, College Station, 77840, TX, USA c TEES Gas & Fuels Research Center, Texas A&M University, College Station, Texas-77843-3122, USA b

ARTICLE INFO

ABSTRACT

Keywords: Reactor scale-up Fischer Tropsch 2-D CFD modeling Supercritical fluids Exothermic reactions

A 2-D Computational Fluid Dynamics (CFD) model of a Packed Bed (PB) Fischer Tropsch (FT) reactor was developed using a non-conventional Supercritical Fluid (SCF) as reaction media. The model was used to study the effect of using SCF-FT reactor bed in alleviating hot spot formation, typically occurring in conventional Gas Phase FT reactors (GP-FT). The potential of scaling-up a typical industrial 1.5-inch diameter reactor bed to a larger tube diameter (up to 4″ ID) was studied as a first step towards process intensification of the FT technology. The high fidelity 2-D model developed in this work was built on experimental data generated at a variety of FT operating conditions both in conventional GP-FT and in SCF-FT reactor bed. Results showed that the maximum Al2 O3 temperature rise in SCF-FT for a 4″ ID bed was just 15 K compared to ˜800 K in GP-FT bed for 15% Co / based catalyst at 500 GHSV and 518.15 K. The enhancement in thermal performance in SCF-FT reactor bed is attributed to the high thermal capacity of SCF media (˜2500 J/kg/K) compared to GP (˜1300 J/kg/K), which resulted in the elimination of hotspot formation. These results provide the first evidence for the application of SCF-FT in larger tube reactor beds while overcoming issues resulting from hotspot formation.

1. Introduction

Although implemented commercially, both reactor configurations suffer from issues related to transport properties, and their handling of catalyst and hydrocarbon products which poses significant bottlenecks for further reactor scale-up [16]. More importantly, compared to the SB (ORYX GTL) reactors in which the whole reaction takes place in liquid phase, the multi-tubular PBR (The Pearl GTL) reactor faces significant challenges due to poor bed thermal conductivity [17–19]. In the literature, various studies have been dedicated towards model development of PB reactors, in which the primary objective was to enable understanding of reactor performance under a set of operational conditions not easily achievable on an experimental scale. Majority of these models were developed under the assumption of Pseudo-homogenous phase to evaluate axial profiles of temperature, product distribution and pressure drop profiles [19–27], with some being more advanced to include diffusional limitations for particle scale assessment of the reactor profile [28–33]. Due to the changes in the reaction phase (from gas phase to liquid/solid phase) and relatively larger size of the catalyst particles, significant temperature gradients in the radial dimension have been observed and reported in the literature [28,29,34]. In contrast to axial temperature gradients, radial gradients lead to

Fischer Tropsch (FT) is an exothermic catalytic reaction which converts synthesis gas (or syngas, a mixture of H2 and CO) to a variety of hydrocarbons- mainly paraffin and olefin products [1–4]. It is an integral part of the Gas to Liquid (GTL) process to produce ultra clean fuels and value-added chemicals [5] including but not limited to paraffins, olefins, and oxygenates. After more than a century of extensive research and advancement both in catalysis and reactor design, FT continues to attract significant attention from the scientific community due to the need for further process intensification and more efficient catalyst to meet global need for cleaner fuels [6–14]. Commercially, FT plants have been implemented using three types of reactor configurations; 1) Fluidized bed, 2) Multi-tubular fixed bed (or Packed bed reactor configuration (PBR)) and 3) Slurry Bubble column reactors (or Slurry Bed (SB)). The State of Qatar hosts two of the world’s largest GTL facilities, with the largest owned jointly by Shell and Qatar Petroleum (the Pearl GTL plant) and the other (ORYX GTL plant) owned jointly by Qatar Petroleum and Sasol with a combined capacity exceeding 180,000 barrels per day of ultra-clean fuels and chemicals [15]. ⁎

Corresponding author at: Petroleum Engineering Program & Chemical Engineering Program, Texas A&M University at Qatar, Education City, Doha, 23874, Qatar. E-mail address: [email protected] (N.O. Elbashir).

https://doi.org/10.1016/j.cattod.2019.05.027 Received 5 February 2019; Received in revised form 22 April 2019; Accepted 13 May 2019 0920-5861/ © 2019 Elsevier B.V. All rights reserved.

Please cite this article as: Aya E. Abusrafa, et al., Catalysis Today, https://doi.org/10.1016/j.cattod.2019.05.027

Catalysis Today xxx (xxxx) xxx–xxx

A.E. Abusrafa, et al.

hotspot formation impacting reaction rates, diffusional resistances and thermal resistances resulting in huge concentration gradients. The secondary impact is realized on the catalyst as it leads to irreversible damages like sintering, coking, phase changes etc that demands more maintenance cycles (downtimes). One of the primary objective of this study is to closely understand the improved heat transfer characteristics achieved by utilization of non-conventional Supercritical Fluids Fischer Tropsch (SCF-FT) reaction systems. It has been suggested that the introduction of the supercritical solvent in the reaction media significantly changes both the transport and thermodynamic behavior of the reactor bed [35–41]. This phase manipulation is achieved by introduction of a solvent while operating the reaction at near critical and supercritical condition to manipulate the physical properties of the reactor bed in such a way that its density and heat capacity behaves like liquid phase, while viscosity and diffusivity behaves like that of gas phase. Thus, the characteristics of SCF-FT reactor bed are generally supposed to be an intermediate between FT SB and FT PBR. The primary advantages of such unique reactor technology is that it provides an opportunity to control both the heat and the mass transfer limitations while facilitating an opportunity to control the hydrocarbon product distribution. In particular, the SCF-FT reactor bed facilitates, 1) in situ extraction of waxy hydrocarbons from catalyst pores due to improved solubility of reaction media [42–44]; 2) elimination of transport limitation promoting selectivity towards heavier hydrocarbons [44]; 3) desorption of primary products prior from undergoing secondary reactions resulting in significant increase in -olefin selectivity [42,45–47]. In a typical commercial scale FT process, there are two modes of operation; High Temperature Fischer Tropsch (HTFT) (utilizes a fused Iron based catalyst in a fluidized-bed reactor to produce lighter hydrocarbons such as olefins, oxygenates and gasoline) and Low Temperature Fischer Tropsch (LTFT) (utilizes cobalt-based catalyst to produce heavy hydrocarbons such as Diesel and Wax) [48]. The original Sasol reactor was of PB LTFT configuration comprising of a large vertical shell of 118″ ID embedding several tubes of ˜ 2″ ID [49]. The capacity of this reactor was 500 barrel per day (BPD) that operated at 27 bar pressure and 503 K temperature [49]. The major limitation of this design was high pressure drop of 2–7 bar due to long tube lengths of ˜ 472″, resulting in high compression costs [49]. In addition to this, due to large amount of heat liberated during the FT reaction (˜160 kJ/ mol), thermal non-uniformity was another concern that effected product selectivity and demanded higher catalyst loading to meet production rate. In order to overcome the aforementioned challenges associated with the operation of the PB reactor, Sasol developed a SB reactor that utilizes wax produced in situ in the process as a media for the reaction rather than complete gas phase operation. This provided much better temperature control due to elevated thermal capacity compared to gas phase [4]. Another benefit of reduction in compression cost was realized as a result of significant reduction in pressure drop due to extremely large hydraulic head of the SB reactor. However, this technology posed several new challenges related to catalyst separation from product wax, and much faster rates of catalyst attrition, as compared to PB technology. Therefore, the main goal of this work is to utilize the SCF-FT mode of operation, which combines the heat transfer benefits of SB, while retaining the operational benefits of PB mentioned above. In our previous study [34], it was demonstrated that the high thermal conductivity of Microfibrous Entrapped Cobalt Catalyst (MFECC) material helps in alleviating the hotspot formation tendency of a conventional catalyst bed. This utility of the MFECC material was then demonstrated to aid in performing radial scale-up to 4″ ID reactor. The objective of the present study is to conduct a similar assessment by utilizing the SCF-FT reactor bed using conventional cobalt based catalyst to first demonstrate its superior capability in mitigating the hotspot formation tendency, and then to perform a scale-up study for process intensification purpose. In contrast to the previous studies on SCF-FT

process, the present study accounts for mass and heat resistances in the radial dimension to investigate the potentiality for scaling up the reactor bed diameter. The approach used in this work was to both develop a 2-D model for SCF-FT and verify it with the experimental data generated from our high pressure reactor bed. In the second stage, the model has been used to scale-up the reactor bed diameter. The 2-Dimensional (2-D) pseudo-homogeneous reactor bed model was developed in COMSOL® Multiphysics and was verified for both modes of operation (GP-FT and SCF-FT). The outcome of this study proves that the introduction of SCF media not only reduces hotspot formation tendency (and consequently methane selectivity), but also provides an excellent scope towards radial scale-up of the reactor bed. 2. Experimental methodology 2.1. Catalyst preparation The cobalt catalyst (15%Co/ 0.5%Ru/ γ-Al2O3) used in this work was prepared by using commercial Alfa Aesar γ-Al2O3 (#43,832) support, cobalt nitrate (CO (NO3 ) 2. 6H2 O) and ruthenium (III) nitrosyl nitrate (Ru (NO )(NO3 )3) precursors. The procedure involved to prepare a final composition of 15% Co/ 0.5%Ru/ γ-Al2O3 catalyst are provided below: 1 Preparation of grounded and sieved (120–170 mesh, 88–125 μm) support followed by calcination in static air at 500 °C for 5 h. This is done to remove hydroxyl groups on the support. 2 A metal nitrate solution was prepared by mixing 3.685 g of CO (NO3 ) 2. 6H2 O and 1.66 mL (Ru (NO)(NO3 )3) solution in 22.73 mL of acetone. 3 Wet impregnation method was used to load the precursor solution on 4 g of the sieved support. This was done in three steps, in the first step; 1/3rd of the precursor solution was added to the support and stirred to dry slowly under ambient conditions. After drying, calcination was done at 300 °C for 3 h in static air. After this, the above mentioned procedure was repeated twice (for the remaining 1/3rd X 2 batches) in which the leftover metal precursor solution was loaded, dried and calcined at 300 °C for 3 h in both the batches. Detailed information about the techniques used to characterize the catalyst are given in the supplementary material. 2.2. High pressure bench scale fixed bed FT reactor Experimental results required for model validation were obtained from a high pressure bench scale Fixed Bed reactor at Texas A&M University at Qatar. A picture of the rig and its components are shown in Fig. 1 below. In this setup, there are three main sections; i) Feed section with a provision of four Mass Flow Controllers (MFC’s), ii) Reaction section comprising of reactor tube with a furnace, iii) Product collection and separation section with hot and cold traps. The reactor tube is made up of SS316 stainless steel tube of 0.688 Inch diameter and 16 Inch total height. The overall capacity of the bench scale unit is about 1 L per day of GTL products. Entire rig is fully automated with Programmable Logic Controllers (PLC) by the use of SCADA interface, and logic design via GE® Proficy hybrid program. The design of the reactor is such that it can run for long time on streams without manual intervention except only during liquid and wax sampling. As the reactor works in an operating condition in the range of 653–773 K, and 1–85 bar condition, the maximum design temperature and design pressure that the reactor can handle is at about 873 K and 150 bar. In addition to this, there are various safety interlocks and alarms designed to handle safety hazards arising due to high pressure leak, flammable gas leak detectors and runaway trip logics which ensure safe operation during exothermic high pressure FT reaction. For SCF-FT reaction, a separate provision for 2

Catalysis Today xxx (xxxx) xxx–xxx

A.E. Abusrafa, et al.

Fig. 1. (a): Geometry of the reactor bed model containing Prepacking, Catalyst bed and Postpacking zone. (b) Bench scale high pressure fixed bed reactor at Texas A& M University at Qatar.

HPLC pump with a liquid vaporizer and static mixture is also provided to ensure that supercritical solvent is vaporized and mixed properly before it enters the reactor at desired pressure condition. Online sampling is conducted using a custom-made Shimadzu® GC, while the offline analysis of liquid and wax samples were done using Agilent® GC. The carbon number range that can be analyzed by the combination of offline and online analysis are in the range of C1-C60 hydrocarbons with excellent resolution and accuracy. More details on the method adopted in analyzing the results, material balance calculations and other aspects related to the experimental scope of this work are covered elsewhere [25,39]. Two independent experimental campaigns were conducted for SCFFT mode of operation in which 15% Co/ 0.5%Ru/ - Al2 O3 catalyst was used. For each SCF-FT campaign, the reaction was carried out at 518 K, 80 bar with 20 bar syngas partial pressure, 2:1 H2:CO molar feed ratio and, 3:1 solvent to syngas molar feed ratio. Steady state data for analyzing CO conversion and CH4 selectivity was measured by the on-line Gas Chromatograph (GC) system. The emerging reaction mixture from reactor was separated in hot trap (at 433 K) and a cold trap (at 277 K) acting as flash drums. The hot trap separates the heavy hydrocarbon from the reaction mixture, and collects at the bottom of the vessel, whereas the cold trap separates light gases and volatile hydrocarbons from the middle distillate products. On-line GC sampling was done at every 1.5 h time interval from the gases leaving top of the hot trap. All the GC lines and the 6-way sampling valve utilized for on-line GC sampling were maintained at high temperature (473 K) to avoid wax condensation. The on-line GC is equipped with one Flame Ionization Detector (FID) and two Thermal Conductivity Detectors (TCD) for analyzing hydrocarbons (C1-C15) and permanent gases (O2, N2, CH4, CO, CO2, C2, and H2S), respectively. The on-line GC system (Make: Shimadzu) is composed of two GCs that have been custom made in such a way that the same sample is simultaneously injected into the three detectors described above. After reaching steady state, liquid samples were collected from both hot-trap and cold trap for off-line GC analysis. For each campaign, the catalyst was diluted with quartz sand (1:10 ratio of catalyst: sand) and loaded in the above mentioned reactor tube (0.688 Inch SS316 reactor tube) to form same reactor volume. The

reactor system was then flushed with n-hexane and then purged with nitrogen followed by in situ reduction under H2 flow (100 mL/min) at a temperature 653 K for 10 h. Subsequently, the system was purged and flushed again and the solvent (n-hexane) was introduced at 2.17 mL/ min rate, while the reactor pressure was varied between 20–80 bar. The temperature was slowly varied between 473–518 K as per the required reaction conditions for each campaign. After the temperature and pressure had been stabilized, syngas was allowed to flow in the range of 50–140 mL/min depending on the desired gas hourly space velocity (GHSV). 3. Mathematical model The fixed bed reactor model developed in this work was assumed to follow pseudo-homogeneous dynamics in 2D cylindrical coordinates system. The model geometry comprised of three zones pertaining to pre-packing, catalytic bed and post-packing respectively as shown in Fig. 2. Fluid flow direction was kept downward to maintain similarity with the experimental reactor rig. 3.1. Momentum transport expressions In order to model porous media flow and the free flow domains, a built-in module in COMSOL® Multiphysics version 5.3a called as “Brinkman Equation” was adopted (Eqs. (1) and (2)). This physics comprises of two terms; the Forchheimer drag term and the convective term which take into account compressibility of the gas phase and incompressibility of the liquid phase and thus makes it an ideal choice for modelling flow fields in GTL systems such as the FT process [50]. The 2D single-phase fluid flow through the PB is described in terms of the velocity (u ) and pressure fields (p) , which are computed via solving the momentum equation and continuity equations along with boundary conditions (Eq. (3)) pertaining to (i) radial symmetry, (ii) no-slip condition at the wall, (iii) inlet mass flow, (iv) outlet pressure.

3

Catalysis Today xxx (xxxx) xxx–xxx

A.E. Abusrafa, et al.

that this study considers a cobalt based catalyst where the rate of watergas shift reaction is assumed to be negligible as CO2 selectivity is very low (< 2%) [52]. Therefore, selectivity calculations of CO2 have not been considered in this modeling study. The local mass balance for species i ( H2, CO , H2 O , (CH2 )n , N2, C6 H14 ) was described by (Eq. 7) using a built in physics module “Transport of concentrated species” which accounts for mass transport through convection and diffusion in the axial and radial directions. The equation provided in COMSOL® Multiphysics for transport mechanism is as follows:

ji +

f (u

Ni = ji + ji =

)wi =

(7)

i ri

(8)

f uwi

f wi

Dik dk

(9)

The diffusion model selected in this case was the Maxwell Stefan diffusion model, where the relative mass flux vector is calculated using (Eq. 9). where Dik represents the multicomponent diffusivity and dk is the diffusional driving force defined as follows:

dk =

Fig. 2. Axisymmetric cut section of the 2D FT reactor bed model cylindrical geometry comprising of i) Pre-packing, ii) Catalyst bed, iii) Post-packing. Fluid flow direction is downward.

1 bed

f

(u .

1

)u

=

pI + µ f

bed

1

( u + ( u)T)

bed

2 1 µ ( 3 f bed

u) I

= 0.1504 +

0.2024

+

p

(

1.0814 dt Dp

+ 0.1226

)

(3)

Dik =

bed

=

150(1 Dp2

bed

bed) 2

2

+

1.75 f u (1

2

(4)

Dp µ f

bed

(5)

Hrxn =

160

kJ mol

(10)

(11)

+ +

1 MWk 1 Vc , k 3 ) 2

× 10

7

(12)

cm3 mol

respectively.

T

(keff T ) = (

Hrxn ) rCO

(13)

where keff , represents the effective thermal conductivity of the bed, T represents the temperature of the reactor bed, Hrxn represents the heat of the reaction and rco represents the rate of consumption of carbon monoxide. The effective thermal conductivity keff of the bed was calculated using a volume based average sum of the bulk fluid mixture kf and the solid domain thermal conductivity ks as follows:

Mass conservation equations for the pseudo homogenous reaction (assuming catalyst effectiveness as unity) is defined for each component of the reaction mixture pertaining to the following simplified FT reaction stoichiometric equation:

+ H2 O ,

1

Energy balance within the 2-D reactor domain was considered to account for the transport of heat through convection and conduction to and through the reactor wall. Balance equations were solved using the simplified “Heat transfer in porous media physics (Eq. 13):

3.2. Mass transport expressions

[(CH2)n]

wi )) MWi

3.3. Heat transfer expressions

f u.

CO + 2H2

1 MWi 1 p(Vc, i 3

12.7,18.9, 7.07, 17.9 and 131.63

bed) 3

(

where, MWi represents the molar mass and Vc, i represents the molar volume for each specie. The molar volume of the representative (CH2 ) n monomer was calculated as a molar weight average sum using the molar volumes of the individual hydrocarbon species based on the Anderson-Schulz-Flory (ASF) product distribution, whereas the molar volumes of H2 O , CO , H2 , N2 and C6 H14 used in this model are:

where p is the sphericity taken as 1 for spherical particles, dt is the tube diameter and Dp is the particle diameter which was 200 µm . The calculated porosity was found to be 0.352. The permeability of the porous medium was calculated using the modified Ergun equation:

1

wk ( MWk

The binary diffusivities Dik for the Maxwell Stefan diffusion model are estimated with an empirical correlation proposed by Fuller [53].

where, µ f is the dynamic viscosity of the fluid, bed is the porosity, f is the density of the fluid, bed is the permeability of the porous medium and m is the mass flowrate at reactor inlet conditions. The bed porosity was calculated using a predictive mean voidage correlation by Benyahia and O’Neill [51]: bed

wk ) p] and xk =

wi = 0 @ r = rbed, wi = wi, o @ z = z 4 r

(2)

u = 0 @ r = rbed, m = mo @ z = z 4 , p= pe @z = z1

1 [(xk p

(Eqs. 7–10) are solved using appropriate boundary conditions corresponding to (i) axial symmetry, (ii) no-flux at the wall, (iii) inlet composition.

(1)

( f u) = 0

xk +

(6)

keff =

bed ks

+ (1

bed ) k f

(14)

The solid phase thermal conductivity ks was found using the volume-average of the thermal conductivities of the inert packing silica and the catalyst. (Eq. 13) was solved along with boundary conditions corresponding to (i) radial symmetry (ii) external cooling (heat transfer between the

where (CH2 )n is the methylene group polymerizing into a different hydrocarbon chain. Stoichiometric coefficient ( i ) of -2, -1, +1, +1 are used for H2, CO , H2 O and (CH2 )n respectively, while C6 H14 was set as the mass constraint component for the SCF-FT case, while in the GPFT N2 was set as the mass constraint component. It is important to note 4

Catalysis Today xxx (xxxx) xxx–xxx

A.E. Abusrafa, et al.

reactor and a constant temperature cooling medium) (iii) inlet temperature (iv) open outflow.

T = Uoverall (Tc r

er

T )@ r = r

bed,

k (T ) = 4 × 10 4 . exp[ 124979(

T = Tinlet @ z = z 4

K1 (T ) = 0.169. exp 6025.98(

The overall heat transfer coefficient Uoverall represents the overall heat transmittance in the vicinity of the wall and is defined using the following correlation suggested by Mamonov et al. 2013 [19]:

Uoverall =

dt + 8 er

1

dw

+

w , int

(15)

w, ext

er

=

kf +

Dp

8.65 1 + 19.4( d ) 2

bed

+

t

(1 0.22

bed

bed) 2

+

2kf

P=

N

kf

3kbed

w , int Dp

kf

= 1.3 + 5

Dp kbed + 0.19Re 3/4Pr 1/3 dt k f

kstainless

(16)

(23)

(28)

(bii + bjj )

(29)

2

T 2 ] Tci

= [1 + mi 1

0.176

i

N

=2

i

(30) (31)

2

N

x i aij and

i

=2

j =1

Z3

(25)

(27)

cij )

Z=

x i xj bij

0.08664RTci Pci

bij = (1

i

N

(26)

xj bij

bm

(32)

j=1

a P b P PVm ;A= m 2;B= m RT (RT ) (RT ) Z2

lnfi = ln

Z (B 2

B + A)

i

ln(

(34)

AB = 0

x i RT + Vm b m Vm RTbm

(33)

i

bm

Vm + bm ) Vm

+

i

i

RTbm2

ln

(Vm + bm ) Vm

bm Vm + bm (35)

For the conventional gas phase reaction, the rate of carbon monoxide consumption is calculated using the kinetic model by Yates and Satterfield [10]:

1

rCO =

2

+ K3 fCO

1 bar 0.5

j=1 I=1

mi = 0.48 + 1.574

2 f2 KfCO H2

+

1 ) 513

(22)

(24)

N

kij ) aii ajj

i (T )

For the SCF-FT case the CO disappearance rate per unit mass of catalyst is calculated using a fugacity based model by Elbashir and Roberts [35,56]:

1+

1 bar 0.5

Pci

+

1 2 K2 fCO

N

aij = (1

(17)

3.4. Kinetic model

1 K1 f H22

1 T

(21)

2 2 i (T )0.42748R Tci

bii =

where, kstainless represents thermal conductivity of stainless steel.

rCO =

1 ) 513

x i xj aij & bm =

aii =

(18)

T

1

1 T

1 bar 0.5

am (T ) Vm (Vm + bm )

bm

j=1 I=1

It should be recognized that industrial reactor tubes are mostly jacketed in which heat is supplied/removed by constant flow of thermic fluid. (Eq. 15) above represents jacketed heat transfer equation that was used for thermal profile comparisons and for scale-up studies. However, for model validation studies, the Fourier’s law of heat conduction was used as the experimental data obtained from the high pressure FT rig utilized a furnace programmed to set the skin temperature at a constant value of specified temperature. Fourier’s law of heat conduction:

q=

RT Vm

am =

where Repa is the Reynolds number, Pr is the Prandtl number, kf is the thermal conductivity of the gaseous mixture and kbed is the heat conductivity of the bed. The first term in (Eq. 16) represents the dynamics contribution to heat transfer at the wall, while the second terms represents the static contribution. The wall heat transfer coefficient w, int is estimated from a correlation of Martin and Nilles [55]:

Nu w =

17981.71(

1 ) 513

(20)

The fugacity of the reaction mixture was estimated by coupling the Modified Soave-Redlich-Kwong EOS (MSRK) along with appropriate mixing rules proposed by Yermakova and Anikeev [57]. The ability of the MSRK EOS to predict the phase behavior of the SCF-FT reaction mixture was well demonstrated in literature [57–59], which was the primary reason for particularly selecting the MSRK EOS for this analysis. Modified Soave Redlich Kwong Equation of State:

where dt is the inner tube diameter, er is the effective radial heat coefficient of the catalyst bed, w, int is the heat transfer coefficient from the bed to the inner wall of the tube, d w is the wall thickness, w is the thermal conductivity of the wall, w, ext is the heat transfer coefficient from the tube wall to the cooling liquid. The effective radial heat coefficient er is one of the main parameters that determine the rate of heat transfer in PB. A correlation that adequately predicts the effective radial heat transfer coefficient in PB was taken from Specchia and Baldi [54] and was used in this modeling study:

Repa Pr

1 mol )] 513 gcat . min . bar

1 T

K3 (T ) = 1.5 × 10 4 . exp 11911(

1

+

w

K2 (T ) = 0.2. exp

1 T

( )P P (1 + b exp ( ) P ) a 0exp

H 2 CO

Eb RT

0

(19)

Ea RT

CO

2

(36)

where PCO , and PH2 are the partial pressures of CO and H2, a0 and b0 are pre-exponential coefficients and Ea and Eb are activation energies. Values for the kinetic parameters used in (Eq. 36) were reported by Sehabiague et al. [60] as given in Eq. 37–40.:

where rCO is the rate of CO consumption; fco and fH2 are the fugacity’s of CO and H2 respectively; K, K1, K2, K3 are kinetic constants. Numerical values of the kinetic rate constant k and the equilibrium constants Ki in (Eq. 19) were estimated by Mogalicherla et al. [40] from experimental data with 15% Co/Al2O3 catalyst. The temperature dependence of the kinetic parameters is given below:

a 0 = 8.037 × 10 5

9

mol kg.s. Pa2

(37)

Catalysis Today xxx (xxxx) xxx–xxx

A.E. Abusrafa, et al.

Ea = 37369

J mol

b0 = 1.243 × 10

Eb =

68478

Table 2 Validation results for (a) SCF-FT bed model in a temperature range of 503–518 K, Solvent: Syngas = 3:1, Syngas ratio = 2:1, Ptot = 80 bar, syngas flow = 138 mL/min (STP) and Solvent flow = 1.14 ml/min (b) GP-FT bed model in a temperature range of 498–528 K, Syngas ratio = 2:1, Ptot = 20 bar, GHSV = 5000 1/hr.

(38) 12

1 Pa

(39)

J mol

(40)

The product distribution was calculated using the ASF product distribution, where the mass fraction of the hydrocarbon cuts produced during the FT reaction are estimated based on the carbon number (n) and the chain growth probability ( ) .

(wn) = n

n 1 (1

SCF-FT

GP-FT

(41)

)2

Based on experimental analysis, the chain growth probability( ) was set as 0.83 [61]. The hydrocarbon product distribution was assumed to be lumped in C1 representing CH4, C2-C4 representing C3H8, C5-C15 representing C10H22 and C22 representing wax and heavy hydrocarbons C22H46. Additionally, other species like H2 O , N2, CO and H2 for gas phase reaction were considered, while for SCF-FT, n-hexane (C6 H14) was used. The physical and chemical properties of each specie in terms of the bulk temperature was obtained from relevant sources in the literature. More detailed information is given in the supplementary material.

4.1. Experimental results 4.1.1. Reproducibility study The objective of the reproducibility study was to confirm the model validation results. For this, we conducted two sets of experiments at constant catalyst weight to GHSV ratio. Table 1 provides a comparison between the performances of the catalyst in terms of % CO conversion of the two SCF-FT runs. For the first run, 1 g of catalyst was used at 1000 GHSV, while in the second run, 0.5 g of catalyst was used at 500 GHSV. It can be observed that similar activity of the catalyst was obtained in terms of % CO conversion in both the runs. These results provides a supporting evidence that results obtained were reproducible. More details on the performance of the reactor bed under different operational conditions are provided elsewhere [62]. 4.2. Modeling SCF FT reactor bed The radial size of the conventional GP-FT reactor tube is limited to a maximum of 1.5-2″ ID as hotspot formation increases drastically at any given temperature with increase in reactor tube diameter [34]. In this section a similar assessment for the SCF-FT reactor bed to compare its thermal performance with larger diameter GP-FT reactor bed (4″ ID) was provided. The results in this section first provide a model validation study for GP-FT and SCF-FT reactor beds, followed by side by side comparison of the 2-D reactor beds at base case condition of 0.688″ ID. Finally, the scale-up results for both the reactor beds are provided to demonstrate the superior ability of the SCF-FT reactor in controlling hotspot formation due to its extremely high thermal capacity (˜2500 J/ kg/K). Table 1 Performance comparison of SCF-FT operated for 0.5 g and 1 g catalyst. Temperature (K)

Pressure (bar)

GHSV (h−1)

% CO Conversion

0.5 1

513-518 513-518

˜80 ˜80

˜500 ˜1000

˜30 ˜30

% CO, experimental

% CO, model

ΔTmax (K) model

503 508 513 518 498 508 518 528

16.18 15.1 23.78 29.1 35.72 53.13 86.82 99.32

17.97 22.15 25.51 28.26 9.89 27.08 91.00 95.84

1.86 2.27 2.6 2.8 1.74 6.57 97.35 102.48

4.2.1. Comparison of model predictions with experimental data The developed model was validated with experimental data obtained from a single tube of 0.688″ ID PB reactor as described in experimental methodology section. The reactor was loaded with 1 g 0.5% Ru promoted 15% Co/ –Alumina supported catalyst, and the bed was diluted with 10 g quartz silica to maintain homogeneous distribution of the catalyst in active bed volume of 3.5 cm height. The experimental data was obtained at different inlet and wall temperatures in the range of 503–518 K at a constant total pressure of 80 bar. Hexane was used as supercritical solvent with a constant solvent/syngas ratio of 3:1, while the syngas ratio (H2/CO) was also kept constant at 2:1. The CO conversions predicted from the model for the SCF-FT case showed good agreement with the experimental data at different inlet temperatures as shown in Table 2. The model validation of the GP-FT was performed over a range of 498–528 K temperature at 20 bar pressure, and the results are also provided in Table 2. For the GP-FT case, 0.59″ ID reactor was used with an active bed length of 10.16 cm at a constant GHSV of 5000 h−1 at STP. The experimental data shown in Table 2 were adopted from Sheng et al. 2013 [61], in which the catalyst comprised of 0.5% Ru/ 15 wt % Cobalt/ -alumina catalyst particles. The modeling results were obtained from the 2-D reactor bed model developed in a previous publication [34] using Yates and Satterfield kinetic model. The simulation results from this work indicate that the developed model is robust and is capable of predicting the trends in SCF-FT mode of operation. On a similar note, the modeling results for % CO conversion of the GP-FT validates well with experimental data at higher temperature conditions (518 and 528 K), while at 498.15 and 508.15 K temperatures, model predictions are lower than experimentally obtained results. This deviation in model predictions from experimental data could be attributed to the sensitivity of the kinetics parameters that are generated for a different catalyst loading reported by Sehabigue et al. [60]. A further analysis of these results also indicated that the hotspot formation tendency swiftly increases (bed ignition) as the reactor temperature is increased. As can be noted from the maximum temperature (ΔTmax) data in Table 2, the temperature rise for the SCF-FT reactor bed is order of magnitude lower than the temperature of the GPFT reactor bed operated at same reactor temperature of 518 K. However, at low temperature conditions (498–508 K) of the GP-FT case, the hotspot formation is not evident, and is of the same order of magnitude to that of the SCF-FT case. A close comparison between the conversion levels at these conditions (498–518 K) for both the beds indicate that the SCF-FT provides much lower conversions compared to the GP-FT case for the same reactor temperature. As recognized earlier in a previous publication [34], the rapid bed ignition in the GP-FT increases conversion to almost 100%, however it leads to significant rise in methane selectivity. As methane is one of the components that produces syngas from reforming reaction [63–65], conversion of syngas to

4. Results and discussion

Catalyst wt. (g)

Temperature (K)

6

Catalysis Today xxx (xxxx) xxx–xxx

A.E. Abusrafa, et al.

methane is highly undesirable in FT reaction. Due to bed ignition which forms a hotspot, a hysterical change in the catalyst activity generally happens as is evident from considerable conversion loss reported in experimental study by Sheng et al. [61]. These challenges limits GP-FT processes to be operated at low syngas conversion levels of 50–70% (a typical of industrial setup [66]). This challenge of high methane selectivity at high conversion levels in conventional GP-FT reactor beds can therefore be mitigated by utilization of the non-conventional SCFFT based reactor that offers significantly high thermal capacity resulting in temperature homogeneity in the reactor bed.

found to be only around 3 K higher than the inlet temperature along the axial dimension, while the GP-FT reactor bed suffers from extreme temperature rise of approximately 500 K as it reaches the reaction zone. While comparing radial temperature gradients, almost 50 K temperature rise was observed in GP-FT case in contrast to only 1 K rise in SCFFT. Superior temperature homogeneity in case of SCF-FT in both axial and radial dimension therefore demonstrates the ability of supercritical media in facilitating effective heat removal compared to GP-FT case. It should be recognized that the CO conversions are functions of temperature, and due to abrupt increase in reactor temperatures, CO conversions as high as 100% could be achieved. However, most of the conversion goes toward methane formation as it is favorable at high temperature conditions [70]. Owing to controlled temperature rise, and uniformity in fluid density (Fig. 4a) throughout the reaction zone, a moderate %CO conversion (˜30%) is achieved in the SCF-FT runs, while suppressing methane selectivity. On the other hand, the fluid density in the GP-FT reactor bed (Fig. 4b) is shown to vary abruptly along the reaction zone due to hot spot formation as shown in Fig. 3b. Similar results were reported in previous studies by Robert and his coworkers [36,47,71]. In their study on SCF-FT using different catalyst bed, it was shown that dense supercritical media facilitated axial thermal uniformity resulting in significant suppression in hot spot formation. Similarly, Yan et al. 1998 [69] conducted a study on the SCF-FT reaction media to investigate the syngas concentration profile across reactor bed to identify the role of diffusional resistances on methane selectivity. They observed that diffusional resistances in the gas phase operation resulted in higher syngas ratio in catalyst pellet and consequently led to increase in methane selectivity. Fig. 5 depicts the centerline temperature profiles and hot spots of the GP-FT and SCF-FT reactor beds simulated at reactor temperatures in the range of 508 K–528 K. It can be observed that the magnitude of the hot spot under all conditions stated above in case of GP-FT is almost 500 K higher than its inlet temperatures, while only a mild temperature rise of ˜3-5 K is observed in SCF-FT. On a similar note, it is observed in Fig. 6 that an increase in the GHSV leads to a decrease in % CO in SCFFT, while an opposite trend is observed in case of GP-FT. It should be noted that the conversion achieved in SCF-FT runs is almost 10–30% to that of GP-FT. This is because of the fact that hotspot formation in GPFT (shown in Fig. 6b) leads to high CO conversion and predominance of methane formation, which is undesirable for FT reaction. Although supercritical media provides an exceptional reaction

4.2.2. Comparison of thermal profiles of reactor beds operated in SCF-FT and GP-FT runs In this section, the validated 2-D reactor bed model was used to compare the thermal profiles of SCF-FT and the GP-FT reactor beds as a function of reactor temperature at constant GHSV of 500 h−1. The two reactor beds are compared in terms of their axial and radial profiles to better understand the heat transfer effects under different sets of operating conditions. It should be noted that the reported data is at reactor conditions of GHSV value, and not at STP condition. This is due to the fact that the calculation of GHSV at STP condition requires all the reactant species to be in gaseous state, however at STP condition, nhexane is at liquid condition, therefore all the reported GHSV values are at reactor condition to avoid any calculation errors due phase changes. More details on the calculations of GHSV are provided in the supplementary material. It has been observed in the previous literature studies [42,43,47,67–69], that SCF-FT mode of operation suppresses both methane and carbon dioxide selectivity relative to the GP-FT due to its ability to maintain homogeneity in both; temperature, as well as syngas concentration across the reactor bed. From our 2-D modeling study, we observe a homogeneous distribution of both, temperature as well as syngas concentration in the SCF-FT reactor bed, which is not observed in the conventional GP-FT reactor bed as shown in Figs. 3 and 4. In both simulations, the bed diameter was kept as 0.688″ ID for GHSV of 500 h−1, while the total pressure of the SCF-FT run is considered 80 bar and the GP-FT run is 20 bar. It can be observed from Fig. 3, there is a smooth temperature transition in the SCF-FT reactor as temperature progresses through the reaction zone, on the other hand an abrupt temperature transition is observed in the GP-FT reactor bed. Additionally, the maximum temperature rise in the case of SCF-FT run was

Fig. 3. (a) Hot spot in SCF-FT and (b) Hotspot in GP-FT for 0.688″ ID (0.0174 m), 500 GHSV calculated at reactor conditions, H2/CO 2:1, Solvent/syngas 3:1, Inlet temperature: 518.15 K. Ptot: 80 bar for SCF-FT, Ptot: 20 bar for GP-FT. 7

Catalysis Today xxx (xxxx) xxx–xxx

A.E. Abusrafa, et al.

Fig. 4. (a) CO mass concentration profile in SCF-FT and (b) CO mass concentration in GP-FT for 0.688″ ID (0.0174 m), 500 GHSV calculated at reactor conditions, H2/CO 2:1, Solvent/syngas 3:1, Inlet temperature: 518.15 K. Ptot: 80 bar for SCF-FT, Ptot: 20 bar for GP-FT.

Fig. 6. (a) %CO conversion in SCF-FT and GP-FT Vs GHSV (b) Maximum Temperature rise in SCF-FT and GP-FT Vs GHSV for 0.688 inch ID (0.0174 m), GHSV range: 100–1000 GHSV calculated at reactor conditions, H2/CO = 2:1, Solvent/syngas 3:1, Inlet temperature 518.15 K. Ptot: 20 bar for GP-FT. Ptot: 80 bar for SCF-FT.

Fig. 5. Centerline temperature in (a) GP-FT, Ptot = 20 bar, and (b) SCF-FT, Solvent/syngas 3:1, Ptot = 80 bar. Both the reactors are 0.688 inch ID (0.0174 m) at constant flow of 500 GHSV calculated at reactor conditions, H2/ CO 2:1.

solvent separation sequence in downstream of the FT reactor utilizing relatively less energy for separation is reported thereby making SCF-FT a competitive option for producing GTL products.

control compared to a conventional GP-FT reaction, extremely low yields are obtained from SCF-FT reaction discouraging its industrial implementation. This is due to the fact that almost 80-90 mass percent of the reaction media in SCF-FT reaction comprises of supercritical solvent. A separate techno-economic study conducted in a previous publication addresses this challenge [72]. In their study, an alternative

4.2.3. Application of SCF-FT for process intensification In the previous section, a comparison between the centerline temperature profiles of the base case (of 0.688″ ID) at variable operational 8

Catalysis Today xxx (xxxx) xxx–xxx

A.E. Abusrafa, et al.

0.688 inch to 4 inch. As the procedure for reactor scale-up is computationally expensive [34], this study was limited only to a 4″ diameter to establish an understanding of the role of SCF media in thermal management of large size reactor beds. For all the aforementioned cases, the effect of variation in GHSV and operational temperatures were recorded in terms of % CO conversion and the hotspot temperatures as shown in Figs. 7 and 8. For all the cases considered, the GHSV was varied in the range of 100–300 h−1, while the operational temperatures were varied from 508 to 518 K. For the SCF-FT case of 4″ ID, it was observed that only a slight increase in % CO conversion is achieved with increase in the reactor temperature (wall and inlet temperature). As 80–90 mass percent of the reaction mixture comprises of supercritical solvent n-hexane and due to strong homogeneity in fluid density (Fig. 5a), the effect of temperature rise on %CO conversion is not very pronounced. Additionally, the variation in GHSV from 100 to 300 h−1 results in decrease in % CO conversion (Fig. 7a) and the maximum temperature rise in the reactor as shown in Fig. 7b. It can be seen that the rate of increase in %CO conversion and maximum temperature rise in the reactor with respect to GHSV for a same reactor diameter also remained constant over all the three temperatures considered in this study. This indicates a linear relationship of residence time with the maximum temperature of the reactor and the %CO conversion. Also, the increase in the tube diameter from 0.688″ to 4″ at constant GHSV does not result in a significant increase in the %CO conversion levels, which could be attributed to excellent concentration homogeneity obtained in SCF-FT reactor bed. In contrast to the SCF-FT case, opposite trends were observed in GPFT case as shown in Fig. 8a and b. As discussed in the previous section, hot spot formation occurs in GP-FT case under all temperature conditions for the base case study of 0.688″ ID (Fig. 5). Similar observation could be made for 4″ ID reactor as shown in Fig. 8 b, in that; for all the wall temperatures the hot spot temperature rise are beyond 500 K. As a consequence, to the hotspot formation, % CO conversion beyond 90% is achieved that mostly leads to higher methane selectivity. Due to this effect, any increase in the reactor temperature does not affect the % CO conversion to a greater extent compared to that of SCF-FT case. With increase in the GHSV (Fig. 8 b), the maximum temperature achieved in the GP-FT case showed an increasing trend, however a decreasing trend is shown in case of SCF-FT. In addition to this, a comparison of hot spot formation between the SCF-FT and the GP-FT case at 4″ID reveals orders of magnitude difference between the hot spot formation tendencies (maximum 15 K temperature rise in SCF-FT Vs. 800 K in GP-FT). Hot spot formations of large magnitude in the GP-FT as demonstrated in this work indicates the inability of the current industrial infrastructure to operate FT reaction in larger diameter tubes despite their numerous benefits. SCF-FT process on the other hand provides an alternative solution in controlling hot spot formation for a larger diameter reactor which reduces temperature impact on hydrocarbon selectivity (reduction in methane selectivity) while at the same time opens a new prospective towards radial reactor scale-up.

Fig. 7. Comparison of (a) %CO conversion (b) Maximum temperature rise in SCF-FT 4″ ID (0.1016 m) with base case of 0.688 inch ID (0.0174 m) at 100–300 GHSV calculated at reactor conditions, H2/CO 2:1, Solvent/syngas 3:1, Ptot = 80 bar.

5. Conclusions In this work, a 2D Pseudo-homogeneous SCF-FT reactor was simulated in COMSOL® Multiphysics v5.3a to compare its thermal performance with equivalent conventional GP-FT reactor bed. The model was validated by experimental data collected under variety of FT reaction conditions, and was further scaled-up to 4″ ID for both SCF-FT and GPFT reactor beds. The impact of reaction media in controlling the hot spot formation for 4″ ID was investigated and correlated with the catalyst activity measured by the % CO conversion over a wide range of GHSVs and wall temperatures. The simulation results showed that the SCF-FT demonstrate exceptional reduction in hotspot formation with a maximum radial bed temperature variation < 15 K for a 4″ ID reactor bed as opposed to 800 K in an equivalent GP-FT reactor bed. Thermal stability in the SCF-FT mode of operation supports previous

Fig. 8. Comparison of (a) %CO conversion (b) Maximum temperature(hot spot) rise in GP-FT 4″ ID (0.1016 m) with base case of 0.688″ ID (0.0174 m) at 100–300 GHSV calculated at reactor conditions, H2/CO 2:1, Ptot = 20 bar.

temperatures in the range of 508–528 K was provided. In the present section, a number of simulation runs for both SCF-FT and GP-FT reactor beds were conducted, where the tube diameter was scaled up from 9

Catalysis Today xxx (xxxx) xxx–xxx

A.E. Abusrafa, et al.

experimental evidence that claimed improved catalyst stability, hydrocarbon selectivity and reactor control under scaled-up conditions. These results provide first confirmation for process intensification in which up to 16-fold reduction in the number of tubes required to achieve a targeted compared to a conventional 1″ ID reactor bed. Moreover, owing to more efficient temperature control, the productivity of the heavy hydrocarbon cuts could be achieved, thus increasing the profitability of the plant. This could lead to significant savings in capital and operating costs associated with existing FT reactor bed technologies. Symbol

Pc , i Critical pressure of species (i), [bar] Pr Prandtl numberr Radial dimension, [m] rbed Bed radius, [m] rCO, Elbashir & Roberts 2004 Rate of carbon monoxide consumption [mol/gcat/ min] rCO, Yates and Satterfield 1991 Rate of carbon monoxide consumption [mol/ kgcat/s] ri Rate of consumption or production of species i [mol/kgcat/s] Repa Reynolds number R Universal gas constant [J/mol/K] T , Tc Local temperature/ Coolant Temperature, [K] Tinlet Inlet temperature, [K] Tc, i Critical temperature of species (i), [bar] u Local velocity vector, [m/s] Uoverall Overall heat transfer coefficient, [W/m2/K] Molar volume of species (i), [cm3/mol] Vc, i Stoichiometry coefficient of species (i) i wi Weight fraction of each species (i) Acentric factor i xi Mole fraction of species (i) z Axial dimension, [m] Hrxn Enthalpy of FT reaction, [kJ/mol] Z Compressibility factor

Chain growth probability Parameter in MSRK Eos Heat transfer coefficient from the bed to the inner wall of the w, int tube, [W/m2/K] Heat transfer coefficient from the tube wall to the cooling w, ext liquid, [W/m2/K] Bed porosity bed Bed permeability, [m2] bed µf Fluid viscosity, [Pa. s] Sphericity p Effective radial heat coefficient [W/m/K] er Thermal conductivity of reactor wall, [W/m/K] w Parameter in MSRK Eos i Density of the fluid mixture [kg/m3] f Nomenclature i

Acknowledgements This paper was made possible by a NPRP award [NPRP 7-843-2312] from the Qatar National Research Fund (a member of the Qatar Foundation). The statements made herein are solely the responsibility of the authors. We extend our gratitude to Prof. Bruce J. Tatarchuk in providing us the catalyst for conducting this research which was a part of this NPRP project between Texas A&M University at Qatar and Auburn University.

Pre-exponential kinetic parameter, [mol/ (kg. s. Pa2)] Binary interaction parameter between species (i) in a mixture Binary interaction parameter between species (i) and (j) in a mixture am Parameter in MSRK Eos bm Parameter in MSRK Eos bii Binary interaction parameter between species (i) in a mixture bij Binary interaction parameter between species (i) and (j) in a mixture b0 Pre-exponential kinetic parameter, [mol/ (kg. s. Pa2)] Cp, f Fluid heat capacity, [J/mol/K] cij Binary interaction parameter between species (i) and (j) in a mixture dk Diffusional driving force of species (k) Dp Average particle diameter, [m] dt Tube diameter, [m]d w Wall thickness, [m] Dik Binary pair Maxwell Stefan diffusivities, [m2/s] Ea Activation energy factor in kinetic expression, [J/mol] Eb Activation energy factor in kinetic expression, [J/mol] fco Fugacity of CO, [bar] fH2 Fugacity of H2, [bar] ji Diffusive flux vector, [kg/ (m2. s)] k Kinetic parameter [mol/(g.min.bar] kij Binary interaction parameter between species (i) and (j) in a mixture K1, K2, K3 Kinetic parameters, [1/bar0.5] keff Effective bed thermal conductivity, [W/m/K] ks Thermal conductivity of solid phase, [W/m/K] kbed Thermal conductivity of the bed, [W/m/K] kf Thermal conductivity of fluid phase, [W/m/K] MWi Molecular weight of species (i), [kg/mol] m Mass flow rate, [kg/s] mi Parameter in MSRK Eos n Carbon number Ni total flux of species i, [kg/ (m2. s)] p Local reactor pressure, [Pa] Pco Partial pressure of CO, [bar] PH2 Partial pressure of H2, [bar]

a0 aii aij

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.cattod.2019.05.027. References [1] M.E. Dry, High quality diesel via the Fischer–Tropsch process–a review, J. Chem. Technol. Biotechnol. 77 (2002) 43–50. [2] H.H. Storch, R.A. Anderson, N. Golumbic, The Fischer-Tropsch and Related Syntheses, Wiley, New York, 1951. [3] C.H. Bartholomew, R.J. Farrauto, Fundamentals of Industrial Catalytic Processes, John Wiley & Sons, 2011. [4] B. Jager, R. Espinoza, Advances in low temperature Fischer-Tropsch synthesis, Catal. Today 23 (1995) 17–28. [5] A.S. Alsuhaibani, S. Afzal, M. Challiwala, N.O. Elbashir, M.M. El-Halwagi, The impact of the development of catalyst and reaction system of the methanol synthesis stage on the overall profitability of the entire plant: a techno-economic study, Catal. Today (2019). [6] M. Dry, T. Shingles, L. Boshoff, Rate of the Fischer-Tropsch reaction over iron catalysts, J. Catal. 25 (1972) 99–104. [7] H.E. Atwood, C.O. Bennett, Kinetics of the Fischer-Tropsch reaction over iron, Ind. Eng. Chem. Process. Des. Dev. 18 (1979) 163–170. [8] J. Barrault, C. Forquy, V. Perrichon, Effects of manganese oxide and sulphate on olefin selectivity of iron supported catalysts in the Fischer-Tropsch reaction, Appl. Catal. 5 (1983) 119–125. [9] B. Sarup, B. Wojciechowski, 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 (1989) 62–74. [10] I.C. Yates, C.N. Satterfield, Intrinsic kinetics of the Fischer-Tropsch synthesis on a cobalt catalyst, Energy Fuels 5 (1991) 168–173. [11] E. Iglesia, S.L. Soled, R.A. Fiato, Fischer-Tropsch synthesis on cobalt and ruthenium. Metal dispersion and support effects on reaction rate and selectivity, J. Catal. 137 (1992) 212–224. [12] R. Zennaro, M. Tagliabue, C.H. Bartholomew, Kinetics of Fischer–Tropsch synthesis on Titania-supported cobalt, Catal. Today 58 (2000) 309–319. [13] G. Jacobs, T.K. Das, Y. Zhang, J. Li, G. Racoillet, B.H. Davis, Fischer–Tropsch synthesis: support, loading, and promoter effects on the reducibility of cobalt catalysts, Appl. Catal. A Gen. 233 (2002) 263–281.

10

Catalysis Today xxx (xxxx) xxx–xxx

A.E. Abusrafa, et al. [14] J. Wu, H. Zhang, W. Ying, D. Fang, Thermal conductivity of cobalt-based catalyst for Fischer–Tropsch synthesis, Int. J. Thermophys. 31 (2010) 556–571. [15] D.A. Wood, C. Nwaoha, B.F. Towler, Gas-to-liquids (GTL): a review of an industry offering several routes for monetizing natural gas, J. Nat. Gas Sci. Eng. 9 (2012) 196–208. [16] B.H. Davis, Fischer–Tropsch synthesis: overview of reactor development and future potentialities, Top. Catal. 32 (2005) 143–168. [17] A. Jess, C. Kern, Influence of particle size and single-tube diameter on thermal behavior of Fischer-Tropsch reactors, Chem. Eng. Technol. 35 (2012) 379–386. [18] W.A. Khan, I. Shamshad, U. Shafiq, A. Mukhtar, Design methodology for the fischertropsch reactor for the production of green diesel from coal syngas, Eur. J. Adv. Eng. Technol. 2 (2015) 20–24. [19] N. Mamonov, L. Kustov, S. Alkhimov, M. Mikhailov, One-dimensional heterogeneous model of a Fischer-Tropsch synthesis reactor with a fixed catalyst bed in the isothermal granules approximation, Catal. Ind. 5 (2013) 223–231. [20] A. Nanduri, P.L. Mills, Comparison of diffusion flux models for Fischer-Tropsch synthesis, Fuel 1 (2019) C2. [21] Y.-N. Wang, Y.-Y. Xu, H.-W. Xiang, Y.-W. Li, B.-J. Zhang, Modeling of catalyst pellets for Fischer-Tropsch synthesis, Ind. Eng. Chem. Res. 40 (2001) 4324–4335. [22] W. Jianmin, S. Qiwen, Z. Zongsen, P. Lifeng, Diffusion and reaction model of catalyst pellets for Fischer-Tropsch Synthesis, China Pet. Process. Petrochem. Technol. 15 (2013) 77–86. [23] G. Chabot, R. Guilet, P. Cognet, C. Gourdon, A mathematical modeling of catalytic milli-fixed bed reactor for Fischer–Tropsch synthesis: influence of tube diameter on Fischer Tropsch selectivity and thermal behavior, Chem. Eng. Sci. 127 (2015) 72–83. [24] B.B. Hallac, K. Keyvanloo, J.D. Hedengren, W.C. Hecker, M.D. Argyle, An optimized simulation model for iron-based Fischer–Tropsch catalyst design: transfer limitations as functions of operating and design conditions, Chem. Eng. J. 263 (2015) 268–279. [25] M.M. Ghouri, S. Afzal, R. Hussain, J. Blank, D.B. Bukur, N.O. Elbashir, Multi-scale modeling of fixed-bed Fischer Tropsch reactor, Comput. Chem. Eng. 91 (2016) 38–48. [26] R. Hussain, J. Blank, B. Todic, N.O. Elbashir, D.B. Bukur, Development of gas-toliquid technologies from micro- to macro-scale, in: M. Weichold, K.R. Hall, E. Masad (Eds.), Excellence and Impact of Research at Texas A&M at Qatar, QScience, Doha, Qatar, 2013, pp. 55–78. [27] N.O. Elbashir, L. Bani Nassr, M.M. Ghouri, Gas-to-liquid technology research at Texas A&M Qatar and its potentials, in: I.M. Mujtaba, R. Srinivasan, N.O. Elbashir (Eds.), The Water-Food-Energy Nexus: Processes, Technologies, and Challenges, CRC Press Taylor & Francis Group, Bota Racon, Florida, USA, 2017, pp. 427–455. [28] M. Mandić, B. Todić, L. Živanić, N. Nikačević, D.B. Bukur, Effects of catalyst activity, particle size and shape, and process conditions on catalyst effectiveness and methane selectivity for Fischer–Tropsch reaction: a modeling study, Ind. Eng. Chem. Res. 56 (2017) 2733–2745. [29] M. Stamenić, V. Dikić, M. Mandić, B. Todić, D.B. Bukur, N.M. Nikačević, Multiscale and multiphase model of fixed bed reactors for Fischer–Tropsch synthesis: intensification possibilities study, Ind. Eng. Chem. Res. 56 (2017) 9964–9979. [30] M. Stamenić, V. Dikić, M. Mandić, B. Todić, D.B. Bukur, N.M. Nikačević, Multiscale and multiphase model of fixed-bed reactors for Fischer–Tropsch synthesis: optimization study, Ind. Eng. Chem. Res. 57 (2018) 3149–3162. [31] B. Todic, M. Mandic, N. Nikacevic, D.B. Bukur, Effects of process and design parameters on heat management in fixed bed Fischer-Tropsch synthesis reactor, Korean J. Chem. Eng. 35 (2018) 875–889. [32] D.B. Bukur, M. Mandić, B. Todić, N. Nikačević, Pore diffusion effects on catalyst effectiveness and selectivity of cobalt based Fischer-Tropsch catalyst, Catal. Today (2018). [33] M. Mandić, V. Dikić, M. Petkovska, B. Todić, D.B. Bukur, N.M. Nikačević, Dynamic analysis of millimetre-scale fixed bed reactors for Fischer-Tropsch synthesis, Chem. Eng. Sci. 192 (2018) 434–447. [34] M.S. Challiwala, B.A. Wilhite, M.M. Ghouri, N.O. Elbashir, Multidimensional modeling of a microfibrous entrapped cobalt catalyst Fischer-Tropsch reactor bed, AIChE J. 64 (2017) 1723–1731. [35] N.O. Elbashir, D.B. Bukur, E. Durham, C.B. Roberts, Advancement of Fischer‐Tropsch synthesis via utilization of supercritical fluid reaction media, AIChE J. 56 (2010) 997–1015. [36] N.O. Elbashir, P. Dutta, A. Manivannan, M.S. Seehra, C.B. Roberts, Impact of cobaltbased catalyst characteristics on the performance of conventional gas-phase and supercritical-phase Fischer-Tropsch synthesis, Appl. Catal. A Gen. 285 (2005) 169–180. [37] N.O. Elbashir, C.B. Roberts, Catalyst in Near-Critical and Supercritical Hexane Media, (2005), pp. 505–521. [38] N.O. Elbashir, C.B. Roberts, Enhanced incorporation of α-olefins in the FischerTropsch synthesis chain-growth process over an alumina-supported cobalt catalyst in near-critical and supercritical hexane media, Ind. Eng. Chem. Res. 44 (2005) 505–521. [39] A. Kasht, R. Hussain, M. Ghouri, J. Blank, N.O. Elbashir, Product analysis of supercritical Fischer-Tropsch synthesis: utilizing a unique on-line and off-line gas chromatographs setup in a bench-scale reactor unit, Am. J. Anal. Chem. 6 (2015) 659. [40] A.K. Mogalicherla, N.O. Elbashir, Development of a kinetic model for supercritical fluids Fischer-Tropsch synthesis, Energy Fuels 25 (2011) 878–889. [41] A.K. Mogalicherla, E.E. Elmalik, N.O. Elbashir, Enhancement in the intraparticle diffusion in the supercritical phase Fischer-Tropsch synthesis, Chem. Eng. Process. Process. Intensif. 62 (2012) 59–68. [42] G. Jacobs, K. Chaudhari, D. Sparks, Y. Zhang, B. Shi, R. Spicer, T.K. Das, J. Li,

[43] [44] [45] [46] [47] [48] [49] [50] [51] [52]

[53] [54] [55] [56]

[57] [58] [59] [60]

[61] [62] [63] [64] [65]

[66] [67]

[68] [69] [70] [71] [72]

11

B.H. Davis, Fischer–Tropsch synthesis: supercritical conversion using a Co/Al2O3 catalyst in a fixed bed reactor, Fuel 82 (2003) 1251–1260. K. Yokota, Y. Hanakata, K. Fujimoto, Supercritical phase Fischer-Tropsch synthesis, Chem. Eng. Sci. 45 (1990) 2743–2749. D.J. Bochniak, B. Subramaniam, Fischer‐Tropsch synthesis in near‐critical n‐hexane: pressure‐tuning effects, AIChE J. 44 (1998) 1889–1896. D.B. Bukur, X. Lang, A. Akgerman, Z. Feng, Effect of process conditions on olefin selectivity during conventional and supercritical Fischer− Tropsch synthesis, Ind. Eng. Chem. Res. 36 (1997) 2580–2587. X. Lang, A. Akgerman, D.B. Bukur, Steady state Fischer-Tropsch synthesis in supercritical propane, Ind. Eng. Chem. Res. 34 (1995) 72–77. X. Huang, C.B. Roberts, Selective Fischer–Tropsch synthesis over an Al2O3 supported cobalt catalyst in supercritical hexane, Fuel Process. Technol. 83 (2003) 81–99. M.E. Dry, The Fischer–Tropsch process: 1950–2000, Catal. Today 71 (2002) 227–241. A. Steynberg, Introduction to Fischer-tropsch Technology, Studies in Surface Science and Catalysis, Elsevier, 2004, pp. 1–63. M.K. Das, P.P. Mukherjee, K. Muralidhar, Modeling Transport Phenomena in Porous Media with Applications, Springer, 2018. F. Benyahia, K.E. O’Neill, Enhanced voidage correlations for packed beds of various particle shapes and sizes, Part. Sci. Technol. 23 (2005) 169–177. N.O. Elbashir, P. Dutta, A. Manivannan, M.S. Seehra, C.B. Roberts, Impact of cobaltbased catalyst characteristics on the performance of conventional gas-phase and supercritical-phase Fischer–Tropsch synthesis, Appl. Catal. A Gen. 285 (2005) 169–180. E.N. Fuller, P.D. Schettler, J.C. Giddings, New method for prediction of binary gasphase diffusion coefficients, Ind. Eng. Chem. 58 (1966) 18–27. V. Specchia, G. Baldi, S. Sicardi, Heat transfer in packed bed reactors with one phase flow, Chem. Eng. Commun. 4 (1980) 361–380. H. Martin, M. Nilles, Radiale Wärmeleitung in durchströmten schüttungsrohren, Chem. Ingenieur. Technol. 65 (1993) 1468–1477. N.O. Elbashir, D.B. Borough, R. Gupta, C.B. Roberts, Reaction Pathway and Kinetic Modeling of Fischer-tropsch Synthesis Over an Alumina Supported Cobalt Catalyst in Supercritical-hexane Media, C1 Chemistry for the Production of Ultra-clean Liquid Transportation Fuels and Hydrogen, (2004), p. 14. A. Yermakova, V.I. Anikeev, Thermodynamic calculations in the modeling of multiphase processes and reactors, Ind. Eng. Chem. Res. 39 (2000) 1453–1472. A. Ermakova, V. Anikeev, G. Froment, Supercritical Fischer-Tropsch synthesis: the effect of nonideality of the reaction mixture on the reaction rate, Theor. Found. Chem. Eng. 34 (2000) 180–188. E. Elmalik, A. Mogalicherla, N. Hamad, S. Nicola, N. Elbashir, S. Fischer, Tropsch synthesis phase behavior in non-conventional reaction media, Preprint PaperAmerican Chemical Society, Division Fuel Chem. 55 (2010) 173–175. L. Sehabiague, R. Lemoine, A. Behkish, Y.J. Heintz, M. Sanoja, R. Oukaci, B.I. Morsi, Modeling and optimization of a large-scale slurry bubble column reactor for producing 10,000 bbl/day of Fischer–Tropsch liquid hydrocarbons, J. Chin. Inst. Chem. Eng. 39 (2008) 169–179. M. Sheng, H. Yang, D.R. Cahela, W.R. Yantz, C.F. Gonzalez, B.J. Tatarchuk, High conductivity catalyst structures for applications in exothermic reactions, Appl. Catal. A Gen. 445 (2012) 143–152. H.A. Choudhury, X. Cheng, S. Afzal, A.V. Prakash, B.J. Tatarchuk, N.O. Elbashir, Understanding the deactivation process of a microfibrous entrapped cobalt catalyst in supercritical fluid Fischer-Tropsch synthesis, Catal. Today (2019). M.S. Challiwala, M.M. Ghouri, P. Linke, N. Elbashir, Kinetic and thermodynamic modelling of methane reforming technologies: comparison of conventional technologies with dry reforming, 2015, AIChE Annual Meet. (2015). M.S. Challiwala, M.M. Ghouri, P. Linke, M.M. El-Halwagi, N.O. Elbashir, A combined thermo-kinetic analysis of various methane reforming technologies: comparison with dry reforming, J. CO2 Util. 17 (2017) 99–111. M.S. Challiwala, M.M. Ghouri, D. Sengupta, M.M. El-Halwagi, N.O. Elbashir, A process integration approach to the optimization of CO2 utilization via tri-reforming of methane, in: a. Espuña, in: M. Graells, L. Puigjaner (Eds.), Computer Aided Chemical Engineering, Elsevier, 2017, pp. 1993–1998. A.M. Saib, A. Borgna, J. van de Loosdrecht, P.J. van Berge, J.W. Niemantsverdriet, XANES study of the susceptibility of nano-sized cobalt crystallites to oxidation during realistic Fischer–Tropsch synthesis, Appl. Catal. A Gen. 312 (2006) 12–19. N.O. Elbashir, C.B. Roberts, Enhanced incorporation of α-Olefins in the Fischer− Tropsch synthesis chain-growth process over an alumina-supported cobalt catalyst in near-critical and supercritical hexane media, Ind. Eng. Chem. Res. 44 (2005) 505–521. H. Tang, H. Liu, X. Yang, Y. Li, Supercritical phase Fischer-Tropsch synthesis reaction over highly active fused Iron catalyst at low temperature, Chin. J. Catal. 29 (2008) 174. S. Yan, L. Fan, Z. Zhang, J. Zhou, K. Fujimoto, Supercritical-phase process for selective synthesis of heavy hydrocarbons from syngas on cobalt catalysts, Appl. Catal. A Gen. 171 (1998) 247–254. G.P. Van Der Laan, A. Beenackers, Kinetics and selectivity of the Fischer–tropsch synthesis: a literature review, Catal. Rev. 41 (1999) 255–318. X. Huang, N.O. Elbashir, C.B. Roberts, Supercritical solvent effects on hydrocarbon product distributions from Fischer− Tropsch synthesis over an alumina-supported cobalt catalyst, Ind. Eng. Chem. Res. 43 (2004) 6369–6381. T. Katbeh, N.O. Elbashir, M. El-Halwagi, An Energy Integrated Approach to Design Supercritical Fischer-Tropsch Synthesis Products Separation and Solvent Recovery System, (2017).