Numerical Study of CO2 Conversion to SAF in a Fixed Bed Catalytic Reactor

Ruiqin Shan1, Shengwei Ma1, Van Bo Nguyen1, Chang Wei Kang1, Teck-Bin Arthur Lim1
1
Institute of High Performance Computing (IHPC), Agency for Science Technology and Research (A*STAR), 1
Fusionopolis Way, #16-16 Connexis, Singapore 138632, Republic of Singapore

ABSTRACT
CO2 hydrogenation to hydrocarbon refers to an indirect pathway of CO2 utilization. Among them, the conversion of CO2
with green H2 to sustainable aviation fuel (SAF) with high energy density has gained much attention. It offers a promising
way to reduce greenhouse gas emissions, address the fossil fuel crises, and transform a climate killer into valuable
products. However, this low-carbon technology is intrinsically complicated. It involves the development of a catalyst, the
design of a reaction system, and its operation and product refining. Hence, it is important to understand the chemical
process of CO2 hydrogenation in the reactor. In this study, numerical simulations of a fixed bed catalytic reactor for CO2-
to-SAF conversion are conducted by coupling CFD with heterogeneous catalytic reactions at the catalytic surface. The
heat and mass transfer between the catalyst surface and surrounding fluid flow are resolved in the simulation. A detailed
understanding of the reacting flow and catalytic processes is obtained from this study. The impact of operating parameters,
i.e., temperature, pressure, mass flow rate, and the ratio between CO2 and H2, is also explored, which provides important
insights into the catalytic reactor design and operation.

1. Introduction + ∇ ∙ [𝐮𝜌] = 0, (1)

Carbon capture and utilization are essential in global ( 𝐮)
efforts to mitigate greenhouse emissions and achieve + ∇ ∙ [𝐮(𝜌𝐮)] + ∇𝑝 + ∇ ∙ 𝐓 = 𝟎, (2)
sustainable development. Carbon capture is a process ( )
+ ∇ ∙ [𝐮(𝜌𝑌 )] + ∇ ∙ −𝐷 , ∇(𝜌𝑌 ) − 𝐷
∇
=
,
designed to capture CO2 emissions from sources before
they are released into the atmosphere in the energy and 𝜔̇ , 𝑖 = 1, … , 𝑁 − 1, (3)
industrial sectors [1, 2]. The captured CO2 can then be 𝜌𝐶 + 𝜌𝐶 𝐮∇𝑇 + ∇ ∙ [𝐮𝑝] + ∇ ∙ (𝐓 ∙ 𝐮) + ∇q̇ = 𝜔̇ .
stored underground in geological formations. Carbon (4)
capture alone is financially demanding for industry. In the above equations, 𝐓 is the viscous stress tensor.
Carbon utilization aims to offset the cost by transforming 𝑌 and 𝐷 , are the mass fraction of the i-th species
CO2 to value-added products, such as building materials, and the mass diffusion coefficient for the i-th species in
valuable chemicals, and high energy fuels, which can also
the mixture, respectively. 𝐷 , is the thermal (Soret)
create revenue streams and economic opportunities.
diffusion coefficient of the i-th species. N represents the
Aviation is a significant contributor to greenhouse
number of gas-phase species in the reaction mechanism.
emissions, and CO2 conversion to sustainable aviation
𝜔̇ is the mass production rate of the i-th species. In the
fuel (SAF) has been a very promising research topic for
reducing the carbon footprint of aviation and dependence energy equation, 𝐶 is the mixture-averaged specific
on fossil fuels [3]. For the conversion of CO2 to SAF, it is heat capacity at constant and q̇ is the heat flux from
critical to understand the physical and chemical processes heat conduction and mass diffusion of species with
to facilitate the chemical system design, catalyst design, different enthalpy. 𝜔̇ is the chemical heat release rate
and performance optimization. However, the related due to the gas-phase reactions. The equation of state
research is very limited due to the intrinsic complexities 𝑝 = 𝜌𝑅𝑇 is used here for the calculation of pressure,
for multi-scale and multi-physics problems. In the present where R is the specific gas constant of the mixture.
paper, numerical simulations are conducted for a fixed- For heterogeneous surface reactions, the governing
bed reactor to resolve the catalytically reacting flow and equation that describes the conservation of the site
interactions between the flow and catalytic pellets. The fraction of adsorbed surface species is shown below:
simulations are expected to provide a detailed 𝜎 =𝜔 , 𝑖 = 1, … , 𝑀 − 1, (5)
understanding of the chemical processes of CO2-to-SAF where 𝜃 is the site fraction of the i-th adsorbed species
conversion and the effects of operating conditions. and 𝜔 is the production rate of the species from the
heterogeneous reactions. 𝜎 represents the site density
2. Numerical Methodologies of the catalyst, and M is the number of adsorbed species
2.1 Governing Equations and Numerical Methods in the heterogeneous reaction mechanism. The
A CFD tool for simulating the CO2-to-SAF process is interaction between the gas-phase bulk flow and the
developed here based on the solver catalyticFOAM [4]. catalyst takes place on the catalytic surface. The mass
The solver allows for CFD simulations of reactive flow flux and heat flux at the catalytic surfaces are calculated
involving both gas-phase reactions and heterogeneous based on the heterogeneous reactions.
reactions at catalytic surfaces. For the gas-phase bulk The governing equations are solved using segregated
flow, the governing equations that describe the algorithms with an operating-splitting method, which
conservation of total mass, momentum, species, and applies to species equation and energy equations
energy are shown below: involving chemical reactions. The reaction operator is
separated from the transport operator allowing each

Corresponding author: Ruiqin Shan

E-mail address: shanr@ihpc.a-star.edu.sg
operator to employ the most suitable numerical method. is 50 cm long and 2.54 cm wide. The catalyst pellets have
Details of the methodologies used in the solver can be a diameter of 5 mm.
found in the reference [4].
2.2 Kinetic Model
SAF, as an alternative jet fuel, has almost identical
physical and chemical properties to conventional jet fuel Fig. 1 Computational configuration for the fixed-bed
based on fossil fuels while featuring low-carbon reactor loaded with catalyst pellets.
emissions. SAF can reduce CO2 emissions by up to 80%
compared with fossil fuels [5]. In terms of compositions, The reactant enters the reactor through the inlet and the
jet fuel consists of hundreds of hydrocarbons with carbon product exits through the outlet. The top and bottom walls
number ranging from 8 to 16, which are mainly paraffins are inert walls with a fixed temperature, while the
[6]. The accurate composition varies based on feedstock surfaces of the catalyst pellets are catalytic walls where
sources and refining processes. The pathway to SAF heterogeneous reactions take place. The interior of the
production from CO2 generally consists of two major catalyst is not resolved here, and the gas mixture flows
steps, i.e., the reverse water gas shift reaction (RWGS) through the void space in the reactor.
and the Fisher-Tropsch (FT) synthesis [6]. In the present 3. Results and Discussion
paper, the reaction model for the hydrogenation of CO2 to In the present study, the reactor is fed with CO2, H2,
hydrocarbons is obtained from the reference [7], as and N2. The operating conditions include the molar ratio
shown below: between CO2 and H2, pressure and temperature of the
𝐶𝑂 + 𝐻 ↔ 𝐶𝑂 + 𝐻 𝑂 (6) reactant, and inlet velocity. A parametric study is
𝐶𝑂 + 2𝐻 ↔ (𝐶𝐻 ) + 𝐻 𝑂 (7) conducted on the operating conditions. The conditions for
where (𝐶𝐻 ) represents long chain hydrocarbons here. the baseline case are: (H2/CO2)mol = 9, pressure = 20 bar,
The kinetic model and parameters shown in Eqs. (8-11) temperature = 900 K, and inlet velocity = 0.1 m/s. The
and Table 1 are obtained through experiments in a fixed- temperature of the top and bottom walls is fixed and set
read reactor loaded with an alumina supported iron to be the same as the feed gas temperature, i.e., 900 K for
catalyst Fe/K@γ-Al2O3. the baseline case. The flow field of the baseline case and
.
.
the sensitivity of the chemical process to the operating
𝑟 = (8) conditions will be discussed in this section.
( / )
3.1 Grid Independence Study
𝑟 = (9) A grid independence study is conducted under the
( )
operating condition for the baseline case. Four mesh grid
𝑘 =𝑘, exp (− ,
( − )) (10) sizes are tested with the total number of cells being 56000,
80000, 188000, and 372000, respectively. Comparisons
log 𝐾 = 3.933 − (11) of the averaged properties, including temperature and key
.
where 𝐾 is the equilibrium constant and T is in unit chemical species, along the reactor length with different
Kelvin. p is the partial pressure of different species. grid sizes are plotted in Fig. 2. The profiles are obtained
Details about the development of the kinetic model and by averaging the flow properties over the cross-sectional
the catalyst characteristics can be found in the reference area of the reactor when the simulation reaches a steady
[7]. state. As shown in Fig.2, there are only negligible
differences in these important properties from coarser
Table 1 Kinetic parameters for the kinetic model in Eqs. mesh to finer mesh. The grid size with 80000 cells is
(8-11) [7]. selected to ensure both computational accuracy and
efficiency.

2.3 Physical Model

The reactor model used in the present paper is a
typical fixed-bed tubular reactor loaded with catalyst
pellets. It is simplified to a 2-dimensional rectangular
channel as shown in Fig. 1 to save computational cost
while still being adequate to study the physical and
Fig. 2 The comparisons of averaged temperature, mass
chemical processes inside the reactor. The reactor model
fraction of CO2, H2 and hydrocarbon (HC) profile along The consumption of CO2 and H2 and the production of
the reactor with different grid sizes of meshing. CO, H2O, and hydrocarbons can be clearly observed from
3.2 Flow Field of the Baseline Case the contours. The more catalytic surface area the feed gas
The reactor is initially filled with a non-reactive mixture passes through from the inlet to the outlet, the more
of H2 and N2 at a pressure of 20 bar and a temperature of heterogeneous reactions take place on the surface. For the
900 K. The feed gas contains 50% N2 by mass. The time temperature, it can be seen in Fig. 5 that it drops up to
evolution of the hydrocarbon mass fraction is shown in 10.2 K in the interior of the reactor due to endothermic
Fig. 3 for the baseline case. As the mixture of CO2, H2, RWGS reaction. Although the FT synthesis reaction is
and N2 enters the reactor from the inlet, the two reactions, exothermic, the heat absorbed by the RWGS reaction
RWGS and FT synthesis, take place on the surface of the exceeds the heat release under the present operating
catalyst pellets. CO2 and H2 reacts to form CO which conditions and catalyst loading. The temperature near the
subsequently reacts with H2 to produce hydrocarbons. As top and bottom walls is higher than in the interior due to
shown in Fig.3, the mass fraction of hydrocarbons the fixed temperature boundary condition set for the walls.
increases from the inlet to the outlet and with time as well
until the flow reaches a steady state.

Fig.5 Flow field of temperature and species mass

Fig.3 Time evolution of flow field of hydrocarbon mass fractions at steady state for the baseline case.
fraction for the baseline case.

Figure 4 depicts the profiles of the area-averaged

mass fraction of chemical species with time at the outlet.
At the initial stage, the amount of H2 remains the same
because the reactive gas has not yet reached the outlet,
and the reactor is initially filled with a non-reactive
mixture of H2 and N2. The mass fraction of H2 starts to
drop, and the other species increase rapidly at t = 2 s when
the reactive feed gas and upstream products arrive at the
outlet. The average properties achieve a steady state at
around t = 3 s.

Fig.6 The profile of area-averaged (a) temperature and (b)

mass fraction of chemical species at steady state along the
reactor length for the baseline case.

The steady-state profiles of the averaged temperature

and chemical species mass fraction are given in Fig. 6. It
can be seen that the temperature starts to drop rapidly at
Fig.4 Time evolution of area-averaged mass fraction of the inlet by up to 6.4 K due to the intensive endothermic
chemical species at the outlet for the baseline case. RWGS reaction near the inlet. The temperature gradually
rises as more CO is produced and reacts with H2 to form
The steady-state flow fields of the mass fraction of hydrocarbons, which release heat into the reactor. For the
chemical species and the temperature are shown in Fig. 5. chemical species, it is seen in Fig. 6b that the amounts of
CO2 and H2 decrease along the reactor, while the water gas shift reaction dominates leading to further
hydrocarbon, CO, and H2O increase due to the CO2 reforming.
hydrogenation process on the catalytic surfaces. It is
4. Concluding Remarks
noted that the decreasing and increasing trend of feed and
The present paper studies the chemical process of CO2
product gases in Fig. 6b continues to the outlet. In
conversion to SAF using CFD simulations. A kinetic
addition, it can be found that there is still a significant
model with two-step heterogeneous reactions for CO2
amount of CO2 and H2 left at the outlet, indicating that
hydrogenation is implemented into the solver for
the conversion of CO2 is not yet complete and will
catalytically reactive flow. A rectangular channel loaded
continue with longer reactors and more catalyst pellets.
with catalytic pellets is employed to mimic a fixed-bed
3.3 Effects of Operating Conditions
reactor. The simulations provide insights into the
A range of different operating conditions is employed
hydrogenation process of CO2 in the reactor facilitating a
to study their effects on the flow field of the reactor and
more detailed understanding of the flow field. A series of
the conversion performance. The results can provide
different operating conditions is applied to the reactor to
insight into the optimization of the chemical process.
understand their effects on the flow field and reactor
Figure 7 shows the profiles of CO2 conversion and
performance, supporting reactor design. It is also worth
hydrocarbon selectivity at the reactor outlet under
noting that the simulation results are not compared with
different conditions, including pressure, temperature,
experimental data in the present paper, which will be
inlet velocity, and molar ratio between H2 and CO2.
conducted in future studies with available detailed reactor
setup and catalyst characteristics, e.g., loading density
and surface area.
5. Acknowledgements
This work is supported by the RIE2025 USS LCER
PHASE 2 PROGRAMME HETFI DIRECTED
HYDROGEN PROGRAMME under Grant ID
U2305D4001, and Carbon Capture and Utilisation
Translational Testbed (CCU-TT) Front-End Loading
Phase 3 (FEL 3) under Grant ID C220415017.

References
[1] E. Martin-Roberts, V. Scott, S. Flude, G. Johnson, R.
S. Haszeldine, and S. Gilfillan, "Carbon capture and
storage at the end of a lost decade," One Earth, vol. 4,
no. 11, pp. 1569-1584, 2021, doi:
10.1016/j.oneear.2021.10.002.
[2] A. Sood and S. Vyas, "Carbon Capture and
Sequestration- A Review," IOP Conference Series:
Earth and Environmental Science, vol. 83, no. 1, p.
012024, 2017/08/01 2017, doi: 10.1088/1755-
1315/83/1/012024.
[3] M. F. Shahriar and A. Khanal, "The current techno-
economic, environmental, policy status and
perspectives of sustainable aviation fuel (SAF)," Fuel,
vol. 325, p. 124905, 2022/10/01/ 2022, doi:
https://doi.org/10.1016/j.fuel.2022.124905.
[4] M. Maestri and A. Cuoci, "Coupling CFD with
detailed microkinetic modeling in heterogeneous
catalysis," Chemical Engineering Science, vol. 96, pp.
Fig. 7 Conversion of CO2 and selectivity of hydrocarbon 106-117, 2013/06/07/ 2013, doi:
at different operating conditions. https://doi.org/10.1016/j.ces.2013.03.048.
[5] IATA, "Annual Review 2020."
In general, the conversion of CO2 and the selectivity [6] K. P. Brooks et al., "Chapter 6 - Low-Carbon Aviation
of hydrocarbon increase monotonically with higher Fuel Through the Alcohol to Jet Pathway," in Biofuels
pressure, higher temperature, lower inlet velocity or for Aviation, C. J. Chuck Ed.: Academic Press, 2016,
longer residence time, and a higher ratio of H2 and CO2 pp. 109-150.
within the present studied range. For pressure effect, the [7] L. Brübach, D. Hodonj, and P. Pfeifer, "Kinetic
conversion of CO2 reaches a maximum at 20 bar within Analysis of CO2 Hydrogenation to Long-Chain
the studied range of 10 to 40 bar. The behavior of CO2 Hydrocarbons on a Supported Iron Catalyst,"
with respect to pressure is attributed to the kinetic model Industrial & Engineering Chemistry Research, vol.
of RWGS in Eq. (8). The partial pressure of CO2 and CO 61, no. 4, pp. 1644-1654, 2022/02/02 2022, doi:
can affect the reaction direction. With more CO produced, 10.1021/acs.iecr.1c04018.