Chemical Engineering Science 160 (2017) 80–84

Contents lists available at ScienceDirect

Chemical Engineering Science

journal homepage: www.elsevier.com/locate/ces

Short Communication

Modeling study of gas-liquid mass transfer enhancement by cylindrical

catalyst particles
crossmark
⁎
Dang Chenga, , Steven Wangb, J.A.M.(Hans) Kuipersa
a
Multiphase Reactors Group, Department of Chemical Engineering and Chemistry, Eindhoven University of Technology, The Netherlands
b
Laboratory of Food Process Engineering, ETH Zurich, Switzerland

A R T I C L E I N F O A BS T RAC T

Keywords: In this work, the influences of non-spherical catalyst particles on the enhancement of the gas-liquid mass
Enhancement factor transfer are numerically studied. The 3D unsteady diffusion-reaction equation both in the liquid phase and the
Numerical simulation solid catalyst particles is solved. A pseudo-homogenous model is employed as well. It is found that the SV
Reaction enhanced mass transfer cylindrical particles increase enhancement factor while SRSV cylindrical particles decrease it in comparison
Three-phase transport phenomena
with the corresponding spherical particles. The horizontal cylinder and the vertical cylinder show small
differences. The enhancement factor increases with catalyst concentration, rate coefficient of reaction and
decreasing particle size. The data of pseudo-homogenous model agree well with the corresponding results of the
fully resolved simulations.

1. Introduction On the other hand, the pseudo-homogenous model is widely used in

Gas-liquid mass transfer plays a crucial role in determining the
performance of slurry reactors for numerous industrial applications. It
is widely known that the rate of gas-liquid mass transfer can be
enhanced by the presence of small solid catalyst particles in the liquid,
which catalyze chemical reactions between the dissolving gas and a
component of the liquid (Holstvoogd et al., 1988; Nagy, 1995; Nagy
and Moser, 1995; Ruthiya et al., 2005).
The previous modeling studies are overwhelmingly based on the
continuum approach, where the one or two dimensional steady state
diffusion-reaction equation was solved (Ramachandran, 2007). The
diffusion-reaction process inside the catalyst particles was usually not
considered, though in some cases the influence of internal diffusion
was taken into account by using the effectiveness factor. However, the
complex gas-liquid-solid interactions and the chemical reactions taking
place inside the solid particles would cause the spatial concentration
gradients in every direction, not only perpendicular to the gas-liquid
interface.
With the rapid development of computer hardware and numerical
method, the effect of the solid particles on the gas-liquid mass transfer near
the gas-liquid interface can now be investigated in more detail by fully
resolved numerical simulation. Wenmakers et al. (2016) studied the
enhancement of gas-liquid mass transfer due to a first order heteroge-
neously catalyzed reaction by means of 3D unsteady numerical simulations.
Only spherical catalyst particles were modeled in their work nevertheless.
industry and be of high relevance to chemical engineers for a fast and
first-level analysis. Wenmakers et al. (2016) also showed that the
pseudo-homogeneous model produced comparable results with those
of detailed numerical simulations. However, how the gas-liquid mass
transfer is affected by non-spherical particles and whether the pseudo-
homogeneous model can be applied to slurry systems with non-
spherical catalyst particles are significant but still unresolved problems.
As non-spherical particles are far more frequently encountered in
industry, it is thus highly desired that further study is carried out to
clarify the effect of non-spherical particles. Therefore, this work aims to
fill in this gap, which is expected to make a relevant additional
contribution.
2. Numerical model
2.1. Fully resolved model
This work targets to model the diffusion of the gas into a liquid film
filled with small catalyst particles (the dashed box in Fig. 1), which can
be seen as a small part of a gas-liquid interface neglecting the
curvature.
The diffusion-reaction of the gas from the gas-liquid interface into
the solid-liquid suspension is described by

⁎
Corresponding author. Current address: LCPP-CNRS, C2P2-UMR 5265, University Claude Bernard Lyon 1, France.
E-mail address: dang.cheng@univ-lyon1.fr (D. Cheng).

http://dx.doi.org/10.1016/j.ces.2016.10.052
Received 22 May 2016; Received in revised form 23 October 2016; Accepted 30 October 2016
Available online 11 November 2016
0009-2509/ © 2016 Elsevier Ltd. All rights reserved.
D. Cheng et al. Chemical Engineering Science 160 (2017) 80–84

defined as:
Rcyl kr
ϕc =
3 Dp (7)

The gas-liquid mass transfer enhancement is characterized by an

enhancement factor (E), which is defined as the ratio of gas flux (J) into
the liquid in case of a reaction and the gas flux into the liquid without a
reaction (Westerterp et al., 1987):
Jreaction |z =0
E=
Fig. 1. Schematic representation of the gas-liquid interface.
Jnoreaction |z =0 (8)

For the details of the pseudo-homogeneous model, please refer to

∂C Wenmakers et al. (2016).
= ∇⋅(D∇C ) − k rC
∂t (1)
where C is concentration, D is Dl for the liquid continuous phase and 3. Methodology
Dp for the solid particles. The first-order chemical reaction is modeled
and only occurs in the solid particles while kr=0 in the liquid phase. The computational domain is a rectangular box, which contains the
continuous liquid and the solid catalyst particles. Two types of particle
2.2. Pseudo-homogeneous model are used, one is sphere (the radius Rsph, the diameter Dsph, the volume
Vsph, the external surface Ssph) and the other is cylinder (the radius Rcyl,
The pseudo-homogeneous model does not distinguish liquid phase the diameter is Dcyl and the height is equal to Dcyl, the volume Vcyl). For
and solid particles, and a homogeneous medium is assumed in where the cylindrical particles, two orientations are considered in reference to
chemical reaction occurs. The material properties of the assumed the z axis, i.e., vertical direction (hereinafter called vertical cylinder)
homogeneous medium are constant throughout space. Its diffusion- and horizontal direction (hereinafter called horizontal cylinder).
reaction equation reads In this work, the spherical particle is used as baseline and the sizes
of the cylindrical particles are defined in two ways:
∂C ∂ 2C 1). The cylindrical particle has the same volume as the spherical
= Deff 2 − kC
∂t ∂z (2) particle; thereby a volume diameter is used (hereinafter called SV
whereDeff is the effective diffusion coefficient of the transferred gas cylinder):
species in the liquid-solid suspension (Wenmakers et al., 2016), k is the Dcyl = 0.8736Dsph (9)
effective rate coefficient of reaction defined as
2). The cylindrical particle has the same ratio of external surface to
k = ηk rεp (3) volume as the spherical particle; thereby a Sauter's diameter is used
where η is the effectiveness factor,εp is the total volume fraction of the (hereinafter called SRSV cylinder):
particles. Dcyl = Dsph (10)
For the first order kinetics, the effectiveness factor of spherical
particle is (Levenspiel, 1999) Therefore, once the diameter of the sphere (50 µm, 100 µm) is
defined, the sizes of the corresponding cylinders are automatically
1⎛ 1 1 ⎞
η= ⎜⎜ − ⎟⎟ defined as described above and thus only the values of Dsph are
ϕs ⎝ tanh(3ϕs) 3ϕs ⎠ (4) specified in all cases.
Both random and hexagonal configurations of the particles are
where ϕs is Thiele modulus for sphere, defined as:
considered (Fig. 2).
Rsph kr The geometry is constructed using the MATLAB in-house code, and
ϕs =
3 Dp (5) is transferred to the COMSOL 4.4 via MATLAB LiveLink. In the case of
randomly configured particles, the dimensions of the rectangular box
The effectiveness factor of cylindrical particle is are 0.3×0.3×0.5 mm3 for small particles and 0.8×0.8×0.5 mm3 for
1 I1(2ϕc ) large particles to ensure for a sufficient number of particles within the
η= domain. In the case of hexagonally configured particles, the dimensions
ϕc I0(2ϕc ) (6)
of the computational domain are determined by the number and size of
where I is Bessel function, and ϕc is Thiele modulus for cylinder, particles, and the particle number ranges from 150–320.

Fig. 2. Schematic representation of different particle configurations.

81
D. Cheng et al. Chemical Engineering Science 160 (2017) 80–84

Table 1
Values of kr and k for spherical particles.

xp [wt%] k=η×kr×εp

kr=0.1 [s−1] kr=1 [s−1] kr=5 [s−1] kr=10 [s−1]

Dsph [μm] 50 100 50 100 50 100 50 100

Random particle configuration

50/100 1 0.0025 0.0020 0.017 0.011 0.051 0.030 0.081 0.044
50/100 5 0.012 0.0097 0.078 0.053 0.238 0.141 0.381 0.209
50/100 10 0.022 0.018 0.147 0.100 0.446 0.264 0.713 0.391
Hexagonal particle configuration
50/100 ≈74 0.088 0.073 0.592 0.403 1.797 1.065 2.876 1.578

Table 2
Values of kr and k for SV cylindrical particles.

xp [wt%] k=η×kr×εp

kr=0.1 [s−1] kr=1 [s−1] kr=5 [s−1] kr=10 [s−1]

Dsph [μm] 50 100 50 100 50 100 50 100

Random particle configuration

50/100 1 0.0025 0.0021 0.019 0.013 0.062 0.033 0.094 0.048
50/100 5 0.012 0.010 0.088 0.063 0.291 0.157 0.442 0.227
50/100 10 0.022 0.019 0.165 0.117 0.543 0.293 0.826 0.424
Hexagonal particle configuration
50/100 ≈74 0.0877 0.0754 0.666 0.473 2.192 1.184 3.332 1.710

Table 3
Values of kr and k for SRSV cylindrical particles.

xp [wt%] k=η×kr×εp

kr=0.1 [s−1] kr=1 [s−1] kr=5 [s−1] kr=10 [s−1]

Dsph [μm] 50 100 50 100 50 100 50 100

Random particle configuration

50/100 1 0.0025 0.0021 0.018 0.013 0.054 0.031 0.084 0.044
50/100 5 0.012 0.010 0.084 0.060 0.256 0.145 0.395 0.209
50/100 10 0.022 0.019 0.157 0.113 0.478 0.272 0.739 0.391
Hexagonal particle configuration
50/100 ≈74 0.088 0.076 0.631 0.456 1.929 1.096 2.981 1.578

The values for the reaction rate coefficient (kr) and the effective
reaction rate coefficient (k) are listed in Tables 1–3. As the values of k is
related to the effectiveness factors and the total volume fractions of the
particles (see Eq. (3)), so their values are tabulated according to the
type of particles.
For the geometrical and computational details as well as other
model parameters, please refer to Wenmakers et al. (2016).
4. Results and discussion
4.1. Qualitative evaluation
Typical concentration profiles at t=1 s are shown in Fig. 3. As seen,
there are strong concentration gradients within solid particles, while
the concentration in the liquid phase is almost homogeneous at x-y
plane.
4.2. Quantitative evaluation
4.2.1. Effect of particle shape
The effect of particle shape is shown in Fig. 4. The symbols
correspond to the data obtained from detailed numerical simulations
and the dashed lines (in the same respective colors) to the correspond-
ing pseudo-homogeneous model (this rule applies to all the following
figures). As seen, E rises monotonically with time for all the cases. The
SV cylindrical particles augment E while the SRSV cylindrical particles
reduce E in comparison with the corresponding spherical particles.
This is possibly because the SV cylinder has a larger area of external
surface (1.14 × Ssph ). However, the volume of the SRSV cylinder is
larger than that of sphere (1.5 × Vsph ), which poses larger resistance/
limitation of internal diffusion. It is observed that the differences
between the horizontal cylinder and the vertical cylinder are small in all
the cases.
The predictions by the pseudo-homogeneous model are found to be
close to the results obtained from the fully resolved numerical
simulations, though a slight over-prediction of E is observed (similar
observations in the following figures). It is noted that the results of
pseudo-homogeneous model are exactly the same for horizontal
cylinders and vertical cylinders, between which slight differences exist
by detailed numerical simulations.
The effect of particle size is shown Fig. 4 as well. The particles of
Dsph=50 µm show higher values of E than the particles of Dsph=100 µm

82
D. Cheng et al. Chemical Engineering Science 160 (2017) 80–84

Fig. 3. Example of a numerically computed concentration profile, (unit: mol/m3), (Dsph=100 µm, kr=0.1 s−1, Dl=10−8 m2/s, Dp=10−10 m2/s).

Fig. 4. E as a function of time for hexagonal and random (xp=10 wt%) packing of particles (kr=10 s−1, Dl=10−8 m2/s, Dp=10−10 m2/s).

Fig. 5. E as a function of time for both the randomly (xp=10 wt%) and hexagonally Fig. 6. E as a function of time for various rate coefficients of reaction (Dsph=50 µm,
oriented SV vertical cylindrical particles (kr=10 s−1, Dl=10−8 m2/s, Dp=10−10 m2/s). Dl=10−8 m2/s, Dp=10−10 m2/s).

for both the randomly and the hexagonally configured particles. This is
because large internal diffusion limitation is present, so the apparent
rate of absorption decreases with the increase in particle size. Since the
differences between horizontal cylinders and vertical cylinders are
small, the vertical cylinders are used hereinafter.
4.2.2. Effect of particle configuration
The effect of particle configuration is shown in Fig. 5. Obviously a
larger E is resulted for the hexagonal packing, which is due to higher
particle concentrations near the gas-liquid interface for hexagonal
packing.
4.2.3. Effect of reaction rate coefficient
The influence of reaction rate coefficient is evaluated for both
hexagonal and random configurations as shown in Fig. 6. As seen,
larger overall reaction rate coefficient leads to a larger E.
4.2.4. Effect of particle concentration
The effect of particle concentration is quantified for random
configuration as shown in Fig. 7. As seen, higher particle concentration
results in a higher E.
The predicted values (after t=1 s) by the pseudo-homogeneous
model and the corresponding results obtained from the fully resolved
simulations are compared in Fig. 8. It is found that the deviation is

83
D. Cheng et al.
Fig. 7. E as a function of time for various solid concentrations (kr=10 s−1, Dl=10−8 m2/s,
Dp=10−10 m2/s).
Fig. 8. Parity plot of the enhancement factor.
mostly within 10%. The good agreement indicates that the pseudo-
homogeneous model is capable of describing the trends as well as the
quantitative values of E in the presence of non-spherical catalyst
particles. Therefore, the pseudo-homogeneous model can be used with
84
Chemical Engineering Science 160 (2017) 80–84
better confidence for slurry systems with non-spherical catalyst parti-
cles.
Some limitations of the models are worth mentioning: Both models
do not consider thermal effect and heat transport. The 3D fully resolved
model is more time consuming, which limits the number of particles
modeled in the system. The pseudo-homogeneous model requires that
the particles smaller than the mass transfer film thickness.
5. Conclusions
Particle shape is found to impact on the gas-liquid mass transfer
process, and the SV cylindrical particles increase E while SRSV
cylindrical particles reduce E in comparison with the corresponding
spherical particles. The horizontal cylinder and the vertical cylinder
show small differences. The enhancement factor increases with catalyst
concentration, rate coefficient of reaction and decreasing particle size.
The data of pseudo-homogenous model agree well with the correspond-
ing results of the fully resolved simulations. The results reported in this
work are of relevance for industrial applications.
Acknowledgements
D.C thanks Dr. Wenmakers at DSM Ahead–Research and
Development for his kind help.
References
Holstvoogd, R., Van Swaaij, W., Van Dierendonck, L., 1988. The absorption of gases in
aqueous activated carbon slurries enhanced by adsorbing or catalytic particles.
Chem. Eng. Sci. 43, 2181–2187.
Levenspiel, O., 1999. Chemical Reaction Engineering. John Wiley & Sons, New York.
Nagy, E., 1995. Three-phase mass transfer: one-dimensional heterogeneous model.
Chem. Eng. Sci. 50, 827–836.
Nagy, E., Moser, A., 1995. Three-phase mass transfer: improved pseudo-homogeneous
model. AIChE J. 41, 23–34..
Ramachandran, P., 2007. Gas absorption in slurries containing fine particles: review of
models and recent advances. Ind. Eng. Chem. Res. 46, 3137–3152.
Ruthiya, K.C., van der Schaaf, J., Kuster, B.F., Schouten, J.C., 2005. Model to describe
mass-transfer enhancement by catalyst particles adhering to a gas-liquid interface.
Ind. Eng. Chem. Res. 44, 6123–6140.
Wenmakers, P.W.A.M., Hoorn, J.A.A., Kuipers, J.A.M., Deen, N.G., 2016. Gas-liquid
mass transfer enhancement by catalyst particles, a modelling study. Chem. Eng. Sci.
145, 233–244.
Westerterp, K.R., Van Swaaij, W.P.M., Beenackers, A.A.C.M., 1987. Chemical Reactor
Design and Operation. John Wiley & Sons, New York.