chemical engineering research and design 9 2 ( 2 0 1 4 ) 25982608

Contents lists available at ScienceDirect

Chemical Engineering Research and Design

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

Methanol synthesis from CO2 and H2 in

multi-tubular xed-bed reactor and multi-tubular
reactor lled with monoliths
Soane Arab , Jean-Marc Commenge, Jean-Francois Portha, Laurent Falk
Laboratoire Ractions et Gnie des Procds, 1 rue Grandville, BP 20451, 54000 Nancy, France

a b s t r a c t
This work investigates the impact of catalyst structuring into particles or monoliths on methanol production from
only CO2 and H2 at a large scale. Methanol synthesis in multi-tubular reactors is evaluated using packed-bed and
monolithic reactors by modeling heat and mass transfer in each reactor. The obtained simulation results show that,
at low gas hourly space velocity (GHSV = 10,000 h1 ), the performances of both reactor technologies are similar. In this
case, the packed-bed reactor technology is the most appropriate technology due to its simplicity of installation and
operation. At high GHSV (25,000 h1 ), the packed-bed reactor technology is limited by a considerable pressure drop
that causes an important loss in productivity due to thermodynamic equilibrium, whereas the monolithic reactors
exhibit negligible pressure drop and achieve far better performances.
2014 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

Keywords: Methanol synthesis; Monolith; Packed bed; Reactor modeling; Carbon dioxide; Hydrogen

1.

Introduction

The production of electricity from renewable sources such as

wind or sun rises the issue of production irregularity and the
gap between production and consumption. Conversion of the
excess of produced electricity to valuable chemicals may be an
interesting alternative to its storage in batteries which is generally not viable from an economic standpoint (Beaudin et al.,
2010). Production of hydrogen by electrolysis of water can be
the rst step for the conversion of the excess of electricity.
The storage of large amounts of hydrogen under high pressure
is expensive and requires high safety precautions (Sakintuna
et al., 2007). Therefore, the second step of the electricity conversion process can be the use of the synthesized hydrogen to
hydrogenate CO2 , stemming from industrial sources, to valuable and safer chemicals such as methanol (Mignard et al.,
2003).

Methanol is a key chemical intermediate that can be used

to produce alternative fuels such as dimethyl ether (DME)
(Olah, 2005). Currently, methanol synthesis is achieved at
industrial scale in multitubular packed-bed reactors that operate at high pressure (5080 bar) and intermediate temperature
(200300 C).
Methanol synthesis from carbon oxides and hydrogen
is usually achieved over copper and zinc oxides catalysts (CuOZnOAl2 O3 ) according to the following main
reactions:

CO + 2H2 CH3 OH

(A)

CO2 + H2 CO + H2 O

(B)

CO2 + 3H2 CH3 OH + H2 O

(C)

Abreviations: CPA, cubic-plus-association; DME, dimethyl ether; GHSV, gas hourly space velocity, h1 ; PC-SAFT, perturbed-chain
statistical associating uid theory; SRK, SoaveRedlichKwong; WHSV, weight hourly space velocity, h1 .

Corresponding author. Tel.: +33 383175104; fax: +33 383175086.

E-mail addresses: soane.arab@univ-lorraine.fr, soane.arab@live.fr (S. Arab).
Received 14 June 2013; Received in revised form 17 December 2013; Accepted 6 March 2014
Available online 15 March 2014
http://dx.doi.org/10.1016/j.cherd.2014.03.009
0263-8762/ 2014 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

chemical engineering research and design 9 2 ( 2 0 1 4 ) 25982608

List of symbols

heat of reaction, J mol1

interfacial area, m2 m3
concentration, mol m3
molar heat capacity of gas species, J mol1 K1
half thickness of the monolith wall, m
hydraulic diameter of the channel, m
coefcient of axial dispersion, m2 s1
tube diameter, m
molecular diffusivity of species i in the gas mixture, m2 s1
diffusivity, m2 s1
coating thickness, m
tube thickness, m
fugacity, bar
friction factor
heat-transfer coefcient between the uid and
the internal wall of the tube, W m2 K1
reaction rate constant
coefcient of interphase mass transfer, m s1
adsorption equilibrium constant, bar1
equilibrium constant based on partial
pressures
tube length, m
hydraulic diameter of the channel, m
mass, kg
number of tubes
total pressure, Pa
power per unit of length generated by the reactions, W m1
ideal gas constant, J mol1 K1
1
rate of reaction j, mol kg1
cata s

Hr
asf
C
Cp
d
dh
Dax
dt
Dm,i
D
e
et
f
F
h
k
kd
K
Kp
Lt
l
M
Nt
P
Q
R
rj

udh


Redh =
S

Sc = D
Sh =
T
u
u
U
v
V
Vp
x
z
Z

Kdp
D

Reynolds number
surface, m2
Schmidt number

Sherwood number
temperature, K
uid velocity, m s1
uid supercial velocity, m s1
global heat-transfer coefcient, W m2 K1
molar volume, m3 mol1
reactor volume, m3
volume of a single catalyst particle, m3
depth in the solid phase, m
axial position, m
compressibilty factor

Greek letters

global efciency of the catalyst

volume ratio

porosity of the catalytic bed


denominator of the kinetics rates

thermal conductivity, W m1 K1

viscosity, Pa s

stoichiometric coefcient

open cross section offered to the uid in the

reactor, m2

density, kg m3

2599

Subscripts
ax
axial
A, B, C reaction
cata
catalyst
eff
effective
i
chemical species number
j
reaction number
m
mixture
p
particle
t
tube
w
internal wall of the tube
Superscripts
ce
center of the solid
f
uid
int
internal
s
solid
su
surface of the solid

The methanol synthesis process is globally exothermic.

Skrzypek et al. (1995) carried out thermodynamic and kinetic
studies of gas mixtures containing CO, CO2 and H2 . They show
that conversion of carbon oxides to methanol is promoted by
high pressures and low temperatures (150200 C) while the
reaction kinetics are promoted by high temperatures.
The use of monolithic reactor technology observed an
important growth during the last two decades, and they will
have increasing applications in mass production of chemicals.
2006)
Through a state of the art study, (Tomasic and Jovic,
pointed out that the monolithic reactors are widely used for
NOx abatement, CO oxidation, and ne chemicals production.
It seems worth to know to which extent this reactor technology may be advantageous for methanol synthesis from only
CO2 and H2 .
The present work investigates the impact of catalyst structuring on the performances and the size of methanol synthesis
reactors. The conventional packed-bed reactors are compared
to monolithic reactors for methanol production at a large
scale. A parametric study is carried out in order to get more
insight on the optimal operating conditions and geometry of
both reactors. More than a traditional performance comparison of these technologies, these types of reactors both exhibit
advantages and drawbacks with respect to operational and
maintenance criteria that must be considered for proper evaluation of the potentials of each technology.

2.

Modeling of reactors

To perform a reliable comparison of both technologies, in conjunction with a parametric study, a detailed model of the
system rst must be developed.

2.1.

Description of technologies

Cooled multi-tubular reactors are considered in the present

work, this reactor technology is well-established and widely
used in chemical industry. The reactors are cooled by a
counter-current stream of liquid water on the shell side.
The tubes of the packed-bed reactor are lled with spherical catalyst particles while those of the monolithic reactor
contain several pieces of cylindrical monoliths with square

2600

chemical engineering research and design 9 2 ( 2 0 1 4 ) 25982608


rCH3 OH,C =

kC KCO2  1


3/2
fCO2 fH
2

fCH3 OH fH2 O
3/2

(fH Kp,C )

(4)

1/2

 = (1 + KCO fCO + KCO2 fCO2 )[fH

Fig. 1 Monolithic reactor technology.

channels as depicted in Fig. 1. Monoliths are either made of
inert material that is coated with the catalytic phase, or fully
extruded from the catalytic material.

2.2.

Thermodynamics

Based on the data presented in Table 1, thermodynamic computations have been carried out in order to understand the
evolution of chemical equilibria of methanol synthesis with
pressure ranging between 30 and 90 bar, and temperature
between 120 and 320 C. The considered gas mixture is composed of CO2 (24 mol%), H2 (72 mol%), and N2 (4 mol%).
Methanol synthesis from CO2 and H2 is favored at high
pressure and low temperature while the rates of reactions
become larger as the operating temperature increases. Thus,
there should be an optimal temperature for which the production of methanol is maximal.
The SoaveRedlichKwong equation of state (SRK) is used
to estimate the compressibility factor Z for the gas mixture at
high pressure and temperature (Skrzypek et al., 1995), and it
is found close to unity (Eq. (1)). Thus, the use of the ideal gas
law is a good approximation to represent the thermodynamic
behavior of the uid.
Pv = ZRT  RT

2.3.

(1)

Kinetics

To describe the kinetics of methanol synthesis, the kinetic

model developed by Graaf et al. (1990) is used, the results
obtained by this model are similar to those obtained by
the models proposed by Bussche and Froment (1996), and
Skrzypek et al. (1995) for methanol synthesis since the outlet
composition of the reactor is close to thermodynamic equilibrium. The rates of reactions (A), (B) and (C) are respectively
given by Eqs. (2), (3) and (4), in mol s1 kg1
.
catalyst

rCH3 OH,A = kA KCO  1 fCO fH

3/2
2

fCH3 OH

(2)

1/2

(fH Kp,A )
2


rH2 O,B = kB KCO2  1 fCO2 fH2

fH2 O fCO
Kp,B


(3)

Table 1 Equilibria constants and heats of reactions

(Graaf et al., 1986; Gallucci et al., 2004).
Reaction
(A)
(B)
(C)

Log10 (Kp )
5139
T 12.621
2073
+ 2.029
T
3066
T 10.592

(bar2 )
()
(bar2 )

1/2

+ (KH2 O /KH )fH2 O ]

2

where r denotes the rates of reactions and fi are the fugacities

of the chemical species involved in the reactions.
Table 2 gathers the parameters of the kinetic model proposed by Graaf et al. (1990) for temperatures ranging between
210 and 275 C.
Even if the ideal gas law is used to predict the uid behavior
in the reactor, fugacities are used to evaluate the rates of reactions given by Eqs. (2)(4). In fact, kinetics proposed by Graaf
et al. rely on fugacities and the use of partial pressures instead
of fugacities leads to an error on species proles at the outlet
of the reactor (around 9% for methanol, 7% for water, and less
than 5% for other species).
The gas mixture in the reactors contains polar molecules
such as CH3 OH and H2 O. These species lead to association
phenomena that change the thermodynamic behavior of the
uid. Association thermodynamic models such as CPA or PCSAFT should be used to get a more accurate prediction of
the uid behavior (Lundstrm et al., 2006). The use of such
thermodynamic models to compute fugacities of species in
rates of reactions may lead to considerable deviations since
the parameters of the kinetic model developed by Graaf et al.
(1990) were tted by using the SRK model.

2.4.

Description of the reactor model

A two-dimensional reactor model is developed to evaluate

both considered technologies under a set of geometric parameters and operating conditions. Fig. 2 shows that the 2D
heterogeneous model considers three regions in the reactor:
a owing gas region, a solid region, and the interface between
these two regions. The solid phase can either represent a catalyst particle in the packed-bed reactor or the wash-coat in one
channel of the monolithic reactor.
For the packed-bed reactor, a depth x = 0 corresponds to
the center of the spherical particle, while for the monolithic
reactor, it corresponds to the interface between the catalytic
coating layer and the monolithic support. In the case of spherical particles, the sphericity is taken into account and the size
of control volumes depends on the depth x.
In the 2D model, heat and mass transfer phenomena are
described by a set of partial differential equations that are
solved by the nite volume method developed by Patankar
(1980). The reactor model assumes that thermal and concentration gradients in the radial direction of the tubes are
negligible. As a result, ow maldistribution through the channels of the monolithic reactor, as well as well-related effects
on uid velocity in the packed bed are not taken into account.

2.5.

Mass-transfer modeling

H r298K (kJ mol1 )

90.70
+41.19
49.51

The mass balance in the gas phase leads to Eq. (5) that considers mass convection with a mean velocity u, axial dispersion
of species i with a coefcient Dax,i , mass transfer between the
owing gas and the catalyst as a function of the interphase

2601

chemical engineering research and design 9 2 ( 2 0 1 4 ) 25982608

Table 2 Parameters of the kinetic model used in simulations (Graaf et al., 1990).
Parameter

Expression

kA
kB
kC
KCO
KCO2

KH2 O /

(7.05 1.39) 10

KH2

exp +

mass-transfer coefcient kd,i , the interfacial area asf , and the

catalyst external porosity .

(uCi )

Unit



(4.89 0.29) 10 exp 113,000300
RT


(9.64 7.3) 1011 exp 152,9001800
RT


(1.09 0.07) 105 exp 87,500300
 RT

(2.16 0.44) 105 exp + 46,800800
RT


(6.37 2.88) 107 exp + 61,700800
RT


84,0001400
9

mol s1 bar1 kg1

cata

Dax,i

2 Ci

z2

kd,i asf

(Ci Csu
)
i

(5)

At the reactor inlet (z = 0), molar ows of species are considered constant and equal to those upstream the inlet. That
leads to the boundary condition described by Eq. (6).
f

uCi |z=0 = u0 Ci,0

(6)

Eq. (7) represents the equality of species molar ows at the

reactor outlet (z = Lt ) and downstream the reactor.
f

(uCi )
z

|z=Lt = 0

(7)

The mass transfer at the interface between the solid and

the owing gas is modeled by a stagnant lm in which the
mass transfer is described by Eq. (8) resulting from equality of
species molar ows at the gassolid interface.
f

kd,i (Ci Csu

) = Deff,i
i

Csu
i
x

|x=R

(8)

Mass transfer inside the catalyst is governed by pure diffusion. The diffusion and reaction equation at a depth x in the
solid is given by Eq. (9).

(Deff,i Sx i )dx +
i,j rj s dV s = 0
x
x
s

(9)

where Deff,i denotes the effective diffusion coefcient in the

catalyst, Sx the solid surface at depth x, i,j the stoichiometric
coefcient of reagent i in reaction j, rj the rate of reaction j

mol s1 bar1/2 kg1

cata
mol s1 bar1 kg1
cata
bar1
bar1
bar1/2

RT

per mass unit of solid, s the density of the solid, and dVs the
elementary solid volume between depths x and x + dx.
At a depth x = 0 in the catalyst, the mass ow by molecular
diffusion is set equal to zero. This assumption leads to the
boundary condition represented by Eq. (10).

Deff,i

Csi
x

|x=0 = 0

(10)

The reactor is assumed to be a closed-closed vessel with

respect to axial dispersion, and the axial dispersion coefcient
of each component is estimated using the correlation of Gunn
(1987): the coefcient of radial dispersion in the packed-bed
reactor is one order of magnitude lower than that of axial
dispersion. The correlation of Taylor and Aris is used to estimate the axial dispersion coefcient in the monolithic reactor
channels.
In the packed-bed reactor, the interphase mass-transfer
coefcient kd,i is estimated by the correlation of Ranz (1952).

kd,i dp
= 2 + 1.8
Dm,i

p
f ud


1/2 


f Dm,i

1/3
(11)

In this correlation, dp represents the particle diameter, Dm,i

the molecular diffusivity in the gas mixture, f the uid density, u the uid supercial velocity, and  is the uid viscosity.
Uberoi and Pereira (1996) proposed a correlation to estimate
mass transfer in long monoliths which is used to evaluate the
interphase mass-transfer coefcient in the monolithic reactor.
In Eq. (12), dh and Lt represent the hydraulic diameter of the
channel and the tube length, respectively.

kd,i dh
ud2h
= 2.696 1 + 0.139
Dm,i
Dm,i Lt

Fig. 2 Considered phases in the 2D model.

0.81
(12)

2602

2.6.

chemical engineering research and design 9 2 ( 2 0 1 4 ) 25982608

Heat-transfer modeling

The tubes are all considered similar and possess the same
global heat-transfer coefcient U. The reactors are cooled by
pressurized boiling water at a constant temperature on the
shell side. Therefore, tubes external wall temperature Tw is
considered constant along the axial direction z. Heat balance
over the Nt cooled tubes leads to Eq. (13).

Nt dint
t U(T Tw ) =
z

f
uCi Cp,i T

+Q

(13)

Q=

1
Vp

Vp

( Hr,j )rj s dV s

(18)

dint
t

where et and t are respectively the tube thickness and its

thermal conductivity.

2.7.

Pressure drop modeling

The pressure drop in the packed-bed reactor is estimated by

Erguns equation (Ergun, 1952).
2

1 f u 2
(1 ) u
dP
+ 1.75 3
= 150
3
2
dz
dp

dp

(19)

The pressure drop over the monolithic reactor is calculated

according to the formula developed by Bird et al. (2007):
dP
f u2
=F
dz
2dh

 e

(15)

The Blasius correlation is used to estimate the Fanning friction factor F for turbulent ow (2.1 103 < Redh < 104 ) in the
monolithic reactor:

( Hr,j )rj s (l + 2x) dx

F=
where Vp is the volume of a single catalyst particle, Hr,j is
the heat of reaction j, and dint
is the internal diameter of the
t
tubes.
The temperature at the reactor inlet (z = 0) is assumed constant and that at the outlet (z = L) is assumed equal to the
temperature just after the reactor. These assumptions are
included in the integration of the heat balance equations
through the following boundary conditions:

(20)

0.0791

(21)

Re0.25
d
h

The 2D model can be made equivalent to a onedimensional pseudo-homogeneous model that does not
consider mass transfer limitations. The one-dimensional
model results will be used as a reference to evaluate masstransfer limitations in the reactors.

3.
Geometric comparison of both reactor
technologies

(16)

and
T(z)
|z=Lt = 0
z

dint
t + 2et

(14)

T(z)|z=0 = T0

2
Nt (dint
t )

(l + 2e + 2d)

dint
1
1
= + t ln
U
h
2t

The left side of Eq. (13) represents the amount of heat transferred per unit of reactor length from the tubes to the shell,
while the right side of this equation describes heat transport
by convection through the reactor open cross-section , and
the power per unit of length Q generated by the reactions. The
expressions of Q for the packed-bed reactor and the monolithic reactor are given by Eqs. (14) and (15), respectively:

Q=

On the shell side, the temperature of the tubes is assumed

constant and equal to that of the boiling point of water at
48 bar.

(17)

The developed model includes the three main reactions

(A)(C). Properties of a commercial catalyst (Haldor Topse
MK-101) have been integrated in the model to describe the
catalytic phase (Dybkjr, 1981).
Heat-transfer coefcient h between the catalyst and the
tubes is estimated by the correlation of Specchia et al. (1980) in
the packed-bed reactor, while in the gap between the monoliths and the internal wall of tubes it is assumed constant and
equal to 500 W m2 K1 , in line with the results reported by
Groppi et al. (2012). In fact, two different approaches could be
used to estimate the global heat-transfer coefcient between
the monoliths and the tubes, the rst one consists in combining the thermal resistances from the channel to the external
wall of the tubes, while the second approach considers only a
mean heat-transfer coefcient in the gap and the conductive
resistance of the tubes. However, the second approach was
used in order to keep the same formalism of the global heattransfer coefcient U in both reactor technologies given by Eq.
(18).

For the same mass of catalyst, a comparison of the two considered reactor technologies is performed in order to identify the
impact of catalyst structuring on the volume of each reactor
and its interfacial area. This comparison is general and independent of the considered operating conditions and reactions
since it only considers geometrical aspects. The catalyst volume in the packed-bed reactor and in the monolithic reactor
are given by Eqs. (22) and (23), respectively.
Vpacked bed =

Vmonolith =

Mcata 1
cata 1

Mcata (l + 2e + 2d)
cata
4e(l + e)

(22)
2

(23)

The catalyst density cata is the same for both reactors since
the catalyst internal structure is supposed not to change with
catalyst structuring into particles or coating layers.
For ratios dp /dt lower than 0.15, the external porosity of the
packed-bed reactor ranges between 0.4 and 0.417. Thus, the
external porosity of the packed-bed can be assumed constant
and equal to 0.41. In other words, for a given mass of catalyst,
the gas volume in the packed-bed reactor can be considered
constant for ratios dp /dt lower than 0.15 and the impact of
the porosity change near the tube walls can be neglected. The

2603

Coating thickness (mm)

1
0.8
0.6
0.4

Radial position (cm)

chemical engineering research and design 9 2 ( 2 0 1 4 ) 25982608

0.9
1
1.1
1.2
1.3
1.5
1.8
2.2

0.2
1
2
3
4
Channel hydraulic diameter l (mm)

0 270

260
265

267

1
0

263

4
Axial position (m)

261

Fig. 5 Radial temperature gradients in tubes estimated

with Grossman and Chambr criterion for packed-bed
reactor. Tw = 260 C, dt = 2 cm.

Fig. 3 Volume ratio of the monolithic reactor to that of the

packed-bed reactor. d = 0.25 mm, = 0.41.
same internal porosity of the catalyst particles or layer is used
for both reactor technologies, and is set equal to 0.43.
A parameter is dened as the ratio of the volume of the
monolithic reactor to the volume of the packed-bed reactor.
This ratio enables to evaluate the degree of intensication of
the packed-bed reactor due to catalyst structuring.

4.

Vmonolith
(l + 2e + 2d)
=
=
(1 )
Vpackedbed
4e(l + e)

4l
(l + 2e + 2d)

(25)

Fig. 4 shows that the variation of the coating thickness

yields to an opposite variation of interfacial area and volume
of the monolithic reactor. That means that an optimal value
of the coating thickness should be set in order to minimize the
reactor volume and to maximize its interfacial area.
In the case where the monoliths are extruded from the catalytic phase (d = 0 mm), the volume of the monolithic reactor

Coating thickness (mm)

400

0.8
500
0.6
0.4
0.2

600
800
1000
1200
1
2
3
4
Channel hydraulic diameter l (mm)

Heat and mass-transfer study

(24)

The evolution of the ratio is reported in Fig. 3. When the

ratio is lower than unity, the monolithic reactor is smaller
than the packed-bed reactor and reciprocally when it is larger
than unity. For most geometric congurations, the volume of
the monolithic reactor is larger than that of the packed-bed
reactor. A value of may be chosen so that the monolithic
reactor volume becomes smaller than that of the packed-bed
reactor. However, miniaturization is not the only criterion that
should be considered. In fact, the interfacial area per unit
volume of tubes is also a key parameter that ensures good
external mass transfer. The interfacial area in monoliths with
square channels is given by Eq. (25).
asf =

can be reduced by at least 35% and the interfacial area can be

multiplied by a factor 2. Zamaniyan et al. (2010) discussed the
possibility and advantages of structuring the catalytic phase
into monoliths. For comparison with the packed-bed reactor, the interfacial area increases from 590 up to 3540 m2 m3
when the particle diameter decreases from 6 to 1 mm.

Fig. 4 Evolution of monolithic reactor interfacial area in

m2 m3 as a function of channels hydraulic diameter l and
coating thickness e. d = 0.25 mm.

This section aims at dening windows of operating conditions

and geometric parameters for which mass-transfer limitations are negligible and the temperature difference between
the center and the wall of tubes is less than 10 K for the two
considered reactor technologies.

4.1.

Heat transfer

High temperature gradients in the tubes are not desirable

since they accelerate catalyst deactivation by sintering, make
temperature control more difcult, and lead sometimes to
formation of undesirable byproducts. The 2D model is based
on a main hypothesis that assumes the absence of temperature gradient between the center and the walls of the tubes.
This hypothesis may be acceptable only for tubes with a small
diameter. One may wonder what is the largest tube diameter
for which temperature gradient remains negligible.
Chambr and Grossman (1955) developed an analytical
solution for the radial heat transfer in a tube lled with
catalytic phase (particles or monoliths) that provides an estimation of the radial temperature proles in cooled or heated
tubes. This method overestimates the radial temperature difference since it only considers the radial heat transfer and
not axial convection. More details of the solution are given by
Mears (1971). In the Grossman criterion calculations, only activation energy and heat of reaction of the reaction (C) have been
considered. In fact, the overall heat generated by reactions (A)
and (B) is in the same range than that generated by reaction
(C). The estimated gradients using Grossman and Chambr
criterion are shown in Fig. 5.
Fig. 5 shows that a hot spot is expected at the inlet of the
packed-bed reactor and the maximum temperature difference
in the gas phase is around 10 K. Identical temperature distributions are obtained in the tubes lled with monoliths. The
assumption of the absence of temperature gradient can be
considered valid for tube diameters smaller than 2 cm.
To complete the heat-transfer study in both reactor technologies, an estimation of temperature gradients inside and
outside the catalytic phase is carried out. The maximal difference between the center temperature Tce of a catalyst particle

2604

External

Internal

6
Particle diameter (mm)

Internal + external

(K)

max

chemical engineering research and design 9 2 ( 2 0 1 4 ) 25982608

Packedbed reactor

Tmax (K)

0
10
Internal + external
5

External

Monolithic reactor

Internal
0
240

250

260

270

280

290

77

80

85
88

93
95

97

300

Surface temperature (C)

Fig. 6 Maximum temperature difference in the catalyst.

Particle: dp = 6 mm, monolith: e = 0.6 mm, tubes: dt = 2 cm.
and its surface temperature Tsu can be estimated according to
Prater criterion depicted by Eq. (26).

Deff,i ( Hr )(Csu,i Cce,i )

eff

(26)

Eq. (27) is used to evaluate the temperature difference in

the stagnant gaseous lm.
(Tsu Tex )max

kd,i Cex,i ( Hr )
=
hf

12

Outlet mass ow rate of CH3 OH using the 2D model

Outlet mass ow rate of CH3 OH at chemical regime

(28)

(27)

where eff is the effective thermal conductivity in the catalyst,

hf the heat-transfer coefcient between the uid and the particle, Cce , Csu and Cex the concentrations of the species at the
center of the particle, at its surface, and in the uid phase,
respectively.
The overall temperature gradient is dened as the sum of
temperature differences in the stagnant lm and inside the
catalyst. The temperature gradients dened above are estimated using concentrations obtained by the 2D model close
to the inlet of the reactor (z = 0.05Lt ) for different values of
Tsu . Simulations results are represented in Fig. 6. The maximal temperature difference between the gas phase and the
center of the particles is less than 6 K.
At the axial position corresponding to half the length of
the packed-bed reactor, the temperature difference between
the gas phase and the center of the particles is at most 3.5 K
for tube diameter equal to 2 cm. The temperature difference
between the tube wall and the center of particles far from the
hot spot at the inlet of the reactor is estimated between 7 and
12 K.
In the case of the monolithic reactor, the temperature difference is close to that observed in the packed-bed reactor. At
the inlet of the monolithic reactor, which corresponds to the
hot spot, a maximum difference of 7 K is obtained between the
wash-coat and the external gas phase.

4.2.

10

limitations occur for large catalyst thickness (particle radius or

coating thickness). A parameter is dened to describe both
kinds of mass-transfer limitations as a function of catalyst
thickness and WHSV.
=

6
8
1
WHSV (h )

Fig. 7 Global efciency of the catalyst in the packed-bed

reactor as a function of WHSV and particle diameter.
Lt = 30 cm, dt = 2 cm.

Mass-transfer study

An efcient mass transfer is a key feature that governs the

performances of the reactors. In order to estimate the internal
and external mass-transfer limitations, simulations have been
performed using the 2D model, and methanol production has
been compared to that obtained with the kinetic regime for
the same operating conditions and geometric parameters.
External mass-transfer limitations depend on the weight
hourly space velocity (WHSV) and interfacial area (external surface area/tubes volume), while internal mass-transfer

Mass-transfer limitations become negligible as the parameter gets close to unity. It should be noted that the geometric
parameters used for mass-transfer study lead to a pressure
drop lower than 3 bar in the packed-bed reactor and a pressure
drop of 0.02 bar in the monolithic reactor. The tube diameter
is set equal to 2 cm in order to satisfy the 2D model hypothesis related to the absence of radial thermal gradients in the
tubes. Results of the mass-transfer study in the packed-bed
reactor and the monolithic reactor are reported in Figs. 7 and 8,
respectively.
For the packed-bed reactor, internal mass-transfer limitations become signicant for particle diameter greater than
2 mm. Regarding the monolithic reactor, the internal masstransfer limitations appear for thicker coating layers.
The catalyst efciency is quite similar in both reactors
for almost all congurations. Nevertheless, the differences
that appear in catalyst efciency are related to the difference between the interfacial area in the reactors. In fact, for a
channel hydraulic diameter equal to 2 mm and a coating thickness of 1.5 mm, the interfacial area of the monolithic reactor
is around 270 m2 /m3 while that of the packed-bed reactor

2 65
Coating thickness (mm)

(Tce Tsu )max =

90

75
1.5

80

85
89
92

95
97

0.5

99

6
8
WHSV (h1)

10

12

Fig. 8 Global efciency of the catalyst in the monolithic

reactor as a function of the coating thickness and the
WHSV. Lt = 30 cm, dt = 2 cm, d = 0.25 mm, l = 2 mm.

2605

chemical engineering research and design 9 2 ( 2 0 1 4 ) 25982608

Tube
Particle

Monolith

Range

dt
Lt
dp

[1; 2] cm
[0.5; 8] m
[1; 3] mm

l
e
d

[1; 2] mm
[0.1; 0.6] mm
[0; 0.5] mm

Constraints
Max(Tw T) < 10 K
Min P
Global efciency of the
catalyst > 90%
Min P
Small reactor

with particles of 3 mm exceeds 1100 m2 /m3 . This explains the

decrease of the global efciency observed in the monolithic
reactor for a WHSV equal to 4 h1 .
It should be highlighted that the contact time in the monolithic reactor, for a constant WHSV, depends on the coating
thickness. In fact, the number of channels increases as the
washcoat layer becomes thin. In other words, the overall open
frontal area varies with the coating thickness and yields to the
variation of the velocity in the channels to keep the WHSV of
the owing gas constant. The contact time in the packed-bed
reactor remains roughly constant when particle diameter is
changed. Computations carried out using Dixons correlation
(Dixon, 1988) show that for a tube diameter equal to 2 cm, the
bed porosity lies between 0.4 and 0.45 when the diameter of
particles changes from 0.5 to 6 mm. The choice of the geometric parameters for the catalyst structure should be a trade-off
between the mass transfer, the pressure drop, and the volume
of the reactor.
In this section, results rising from heat and mass-transfer
study, and geometric comparison of the two reactor technologies, have been used to dene operating conditions
and windows of geometric parameters for which heat and
mass-transfer limitations are negligible in both reactors.
Tables 3 and 4 summarize the retained windows of parameters
and operating conditions for which heat and mass-transfer
limitations become negligible for each technology.

5.

Methanol production

The production of methanol using the two studied reactor

technologies is then evaluated for the same mass of catalyst
(1000 kg). This evaluation aims at nding the geometric congurations of the reactor that minimize the number of tubes
and the operating costs on one hand and maximize the productivity of methanol on the other hand.
Simulations are performed for equal mass of catalyst in
both reactors and the same operating conditions gathered in
Table 5. The tube length and the catalyst thickness (dp or e) are
varied while all other parameters are kept constant. The number of tubes is introduced as an additional criterion for the
choice of the reactor technology. Indeed, this criterion takes
into account the technical feasibility during the fabrication
and the maintenance ability of reactors. Packed-bed reactors

Table 4 Operating conditions and constraints.

Condition
T0
Tw
P0
[CO2 ; H2 ; N2 ]
WHSV

Range

[240; 260] C
[250; 270] C
80 bar
[24; 72; 4] mol%
[2; 10] h1

Parameter

Value

Unit

dt
Lt
dp
l
e
d
T0
Tw
P0
[CO2 ; H2 ; N2 ]
WHSV

2
[0.5; 8]
[1.3; 3]
1.5
[0.1; 0.6]
0.25
250
260
80
[24; 72; 4]
4 and 10

cm
m
mm
mm
mm
mm

C
bar
mol%
h1

used for methanol synthesis at industrial scale are operated

at GHSV close to 10,000 h1 . The production of methanol is
evaluated at WHSV = 4 h1 , and WHSV = 10 h1 successively.
These values of WHSV correspond to GHSV of 10,000 h1 and
25,000 h1 , respectively. Thereby, four case studies are compared for both reactor technologies.

5.1.
4 h1

Case A1: packed-bed reactor with WHSV equal to

Fig. 9 shows the results of simulations for the packed-bed reactor. Methanol production is represented as a function of tube
length ranging from 0.5 to 8 m for different particle diameter.
The productivity is large for short tubes and large diameters of
particles and it decreases slightly as the tubes become longer
and as the particles become smaller. The obtained results for
the packed-bed reactor can be explained by the pressure drop
that is reported in Fig. 10. In fact, the pressure drop moves
the thermodynamic equilibria towards a lower production of
methanol.

5.2.
4 h1

Case A2: monolithic reactor with WHSV equal to

The same simulations have been carried out in the monolithic

reactor and the results are depicted in Fig. 11. The amount
of methanol produced in the monolithic reactor is similar to
that obtained in the packed-bed reactor, this can be explained
by the fact that the outlet composition of both reactors is
close to the same thermodynamic equilibrium. Moreover,
pressure drop in the monolithic reactor remains negligible
and the thermodynamic equilibria are not impacted, this

Parameter

Table 5 Parameters used for comparison of reactor

technologies.

Methanol production (kg.h )

Table 3 Design parameters and constraints.

540

Case A1
Particle diameter (mm)

530
520

3
2.5
2

510

1.5
1.3

500

Tube length (m)

Fig. 9 Production of methanol in the packed-bed reactor as
a function of the tube length for various particle diameters
ranging from 1.3 up to 3 mm. dt = 2 cm, WHSV = 4 h1 .

2606

chemical engineering research and design 9 2 ( 2 0 1 4 ) 25982608

Case A1

Methanol production (kg.h )

Particle diameter (mm)

0.5

2.5

1
2

3
4

1.5

3
4
5
6
Tube length (m)

Methanol production (kg.h )

Coating thickness (mm)

0.1

538

0.2

536

0.3

534

0.4

1000

3
2.5
2

800

1.5
1.3

600

4
6
Tube length (m)

Fig. 13 Production of methanol in the packed-bed reactor

as a function of the tube length for various particle
diameters ranging from 1.3 up to 3 mm. dt = 2 cm,
WHSV = 10 h1 .

0.5

532

0.6

530

3
4
5
6
Tube length (m)

Fig. 11 Production of methanol in the monolithic reactor

as a function of the tube length for various coating
thicknesses ranging from 0.1 up to 0.6 mm. l = 1.5 mm,
d = 0.25 mm, dt = 2 cm, WHSV = 4 h1 .
explains why methanol production is almost independent of
the tube length. It should be noticed that the number of tubes
required in the monolithic reactor is generally larger since the
monolithic support of the catalyst contributes to increase the
volume of the reactor.
Fig. 12 emphasizes the evolution of tubes number in both
reactor technologies as a function of catalyst structuring and
tube length. In the packed-bed reactor, the tubes number
is almost independent of particles size since the bed external porosity remains roughly constant when the diameter
of particles ranges from 1.3 to 3 mm. For short tubes and

3000
2000
1000
15000

Particle diameter (mm)

3
1.3

Coating thickness (mm)

0.1

5000
0.6

4
5
6
Tube length (m)

Simulation of methanol production results at high WHSV in

the packed-bed reactor are gathered in Fig. 13. In this case, the
production of methanol decreases sharply as the diameter of
particles becomes smaller and the tubes become longer.
This loss of productivity is due to the high pressure drop
in the reactor, as represented in Fig. 14. The pressure drop is
very large for long tubes and makes the use of the packed-bed
reactor not viable from a process standpoint.

5.4.
Case B2: monolithic reactor with WHSV equal to
10 h1
Simulations have been carried out for the monolithic reactor and show that an enhanced production of methanol can
be achieved. Fig. 15 describes the evolution of methanol production with the geometric parameters of monoliths. The
production of methanol in the monolithic reactor remains

Packedbed reactor

Monolithic reactor

10000

5.3.
Case B1: packed-bed reactor with WHSV equal to
10 h1

Particle diameter (mm)

Tubes number

Particle diameter (mm)

thin coating layers, tubes number in the monolithic reactor

increases sharply and makes its fabrication difcult. Tubes
number of both reactors remains in the same range for long
tubes and thick coating layers.
The packed-bed reactor technology seems the most appropriate for methanol production at industrial space velocity
(WHSV = 4 h1 ) since it is less expensive, easier to fabricate,
to operate and to maintain.

Case A2

540

Case B1

Fig. 10 Pressure drop (bar) at the outlet of the packed-bed

reactor as a function of the tube length and particle
diameter. dt = 2 cm, WHSV = 4 h1 .

542

1200

Fig. 12 Tubes number in both reactors for different

geometric congurations. l = 1.5 mm, d = 0.25 mm, dt = 2 cm.

Case B1

2.5

3
5
8

12
15

20
30
40

1.5
1

3
4
5
6
Tube length (m)

Fig. 14 Pressure drop (bar) at the outlet of the packed-bed

reactor as a function of the tube length and particle
diameter. dt = 2 cm, WHSV = 10 h1 .

Methanol production (kg.h1)

chemical engineering research and design 9 2 ( 2 0 1 4 ) 25982608

1110

Case B2
Coating thickness (mm)

0.5

1100
1090

0.4
0.3

0.6

0.2

1080
1070

0.1

3
4
5
6
Tube length (m)

Fig. 15 Production of methanol in the monolithic reactor

as a function of the tube length for various coating
thickness ranging from 0.1 up to 0.6 mm. l = 1.5 mm,
d = 0.25 mm, dt = 2 cm, WHSV = 10 h1 .
roughly constant when the length of tubes is changed. A slight
change in methanol production is observed when the coating
thickness is changed from 0.1 to 0.6 mm, this change is due to
a variation of the residence time with the coating thickness in
the monolithic reactor. The pressure drop in the monolithic
reactor is less than 0.2 bar for all geometries and operating
conditions considered in this section. The monolithic reactors can operate at high WHSV and their performance is not
impacted by the pressure drop.

6.

Additional design considerations

Washcoat thickness (mm)

1102
1099
1095

0.4

0.3

0.6

0.2

0.5

0.1

6000
4000

1091
1088
1084

8000

2000

more investments and operating costs through energy consumption and compressors.
The monolithic reactor may raise the issue of the cost of
the catalyst manufacturing and regeneration. If the rate of the
catalyst deactivation is high, and the regeneration requires
the extraction and treatment of monoliths out of the reactor,
the operating cost will rise and make the monolithic reactor
less appropriate from an economic point of view. The use of
extruded monoliths made of the catalytic phase will cost less
than using a supported catalytic phase (Boger et al., 2004). The
choice between the considered reactor technologies should be
completed by an accurate economic study in order to evaluate
the viability of the chosen reactor technology. More aspects
such as the reactor fabrication cost, the catalyst structuring
and regeneration feasibility should be evaluated more precisely.

7.

Conclusion

Two reactor technologies for methanol synthesis from CO2

and H2 at an industrial scale have been compared. Heat and
mass transfer studies have been carried out in order to dene
geometries and operating conditions for which heat and mass
transfer limitations are negligible. The production of methanol
using the two reactor technologies has been compared under
the same operating conditions and for the same mass of catalyst. The obtained results show that the pressure drop is the
main drawback of the packed-bed reactor technology.
For WHSV used currently for methanol synthesis at industrial scale, the use of the packed-bed reactor technology is the
most appropriate. This reactor technology is easy to manufacture, to operate and it is widely used. The use of monolithic
reactors at low WHSV is not suitable since they are more
expensive to manufacture and lead to similar performances
as the packed-bed reactors.
At high WHSV, packed-bed reactors are penalized by a high
pressure drop while the monolithic reactors operate at high
space velocities with high performances and negligible pressure drop. This technology is suitable to be used for methanol
synthesis in this case. More technical and economic survey of
the monolithic reactors should be carried out in order to dene
more precisely the viability of this technology. The extrusion of
the catalyst into monoliths could enable to maintain the size
of the monolithic reactor close to that of the packed-bed reactor in one hand, and it can help to reduce the manufacturing
cost of the monolithic reactor in the other hand.

Acknowledgments

Tube number

Methanol production (kg.h )

A maximum tube length should be considered in order to minimize their number in the shell. For example, if a tube length
of 8 m is considered, the number of tubes is around 500 for the
packed-bed reactor and around 700 for the monolithic reactor (for e = 3 mm). The high number of tubes in the monolithic
reactor makes their fabrication more expensive than that of
the packed-bed reactor. Fig. 16 shows that this issue of fabrication does not appear for the monoliths extruded from the
catalytic phase and the number of tubes is slightly lower than
that for the packed-bed reactor.
For tubes longer than 6 m, the mass of methanol produced
in the monolithic reactor is at least 10% higher than that
obtained using the packed-bed reactor at a WHSV of 10 h1 .
Moreover, high pressure drop in packed-bed reactor is not suitable from the standpoint of process viability since it induces

2607

0.1

This work has been supported by French Research National

Agency (ANR) through Efcacit nergtique et rduction des
missions de CO2 dans les systmes industriels program
(project VITESSE2 no ANR-10-EESI-06). The authors want to
thank the ANR for their nancial support.

0.6

References

Tube length (m)

Fig. 16 Production of methanol and number of tubes in
monoliths extruded from the catalytic phase as a function
of the tube length for various half wall thickness ranging
from 0.1 up to 0.6 mm. l = 1.5 mm, d = 0 mm, dt = 2 cm,
WHSV = 10 h1 .

Beaudin, M., Zareipour, H., Schellenberglabe, A., Rosehart, W.,

2010. Energy storage for mitigating the variability of
renewable electricity sources: an updated review. Energy
Sustain. Dev. 14 (December (4)), 302314.
Bird, R., Stewart, W., Lightfoot, E., 2007. Transport Phenomena,
Wiley International ed. J. Wiley, New York, USA, pp. 189.

2608

chemical engineering research and design 9 2 ( 2 0 1 4 ) 25982608

Boger, T., Heibel, A., Sorensen, C., 2004. Monolithic catalysts for
the chemical industry. Ind. Eng. Chem. Res. 43 (16),
46024611.
Bussche, K., Froment, G., 1996. A steady-state kinetic model for
methanol synthesis and the water gas shift reaction on a
commercial Cu/ZnO/Al2 O3 catalyst. J. Catal. 161 (1), 110.
Chambr, P.L., Grossman, L.M., 1955. On limiting temperatures in
chemical reactors. Appl. Sci. Res. 5 (4), 245254.
Dixon, A.G., 1988. Correlations for wall and particle shape effects
on xed bed bulk voidage. Can. J. Chem. Eng. 66 (5), 705708.
Dybkjr, I., 1981. Topsoe methanol technology. Chem. Econ. Eng.
Rev. 13, 1725.
Ergun, S., 1952. Fluid ow through packed columns. Chem. Eng.
Prog. 48 (2), 8994.
Gallucci, F., Paturzo, L., Basile, A., 2004. An experimental study of
CO2 hydrogenation into methanol involving a zeolite
membrane reactor. Chem. Eng. Process. Process Intensif. 43
(August (8)), 10291036.
Graaf, G., Scholtens, H., Stamhuis, E., Beenackers, A., 1990.
Intra-particle diffusion limitations in low-pressure methanol
synthesis. Chem. Eng. Sci. 45 (4), 773783.
Graaf, G., Sijtsema, P., Stamhuis, E., Joosten, G., 1986. Chemical
equilibria in methanol synthesis. Chem. Eng. Sci. 41 (11),
28832890.
Groppi, G., Tronconi, E., Cortelli, C., Leanza, R., 2012. Conductive
monolithic catalysts: development and industrial pilot tests
for the oxidation of o-xylene to phthalic anhydride. Ind. Eng.
Chem. Res. 51 (22), 75907596, 10.1021/ie2021653.
Gunn, D., 1987. Axial and radial dispersion in xed beds. Chem.
Eng. Sci. 42 (2), 363373.
Lundstrm, C., Michelsen, M.L., Kontogeorgis, G.M., Pedersen,
K.S., Sorensen, H., 2006. Comparison of the SRK and CPA

equations of state for physical properties of water and

methanol. Fluid Phase Equilib. 247 (1), 149157.
Mears, D.E., 1971. Diagnostic criteria for heat transport
limitations in xed bed reactors. J. Catal. 20 (2), 127131.
Mignard, D., Sahibzada, M., Duthie, J., Whittington, H., 2003.
Methanol synthesis from ue-gas CO2 and renewable
electricity: a feasibility study. Int. J. Hydrogen Energy 28 (4),
455464.
Olah, G.A., 2005. Beyond oil and gas: The methanol economy.
Angew. Chem. Int. Ed. Engl. 44 (April (18)), 26362639.
Patankar, S., 1980. Numerical heat transfer and uid ow. In:
Series in Computational Methods in Mechanics and Thermal
Sciences. Taylor & Francis, New York, USA.
Ranz, W., 1952. Friction and transfer coefcients for single
particles and packed beds. Chem. Eng. Prog. 48, 247253.
Sakintuna, B., Lamari-Darkrim, F., Hirscher, M., 2007. Metal
hydride materials for solid hydrogen storage: a review. Int. J.
Hydrogen Energy 32 (9), 11211140.
Skrzypek, J., Lachowska, M., Grzesik, M., Sloczynski, J., Nowak, P.,
1995. Thermodynamics and kinetics of low pressure
methanol synthesis. Chem. Eng. J. 58 (2), 101108.
Specchia, V., Baldi, G., Sicardi, S., 1980. Heat transfer in packed
bed reactors with one phase ow. Chem. Eng. Commun. 4,
361380.
V., Jovic,
F., 2006. State-of-the-art in the monolithic
Tomasic,
catalysts/reactors. Appl. Catal. A: Gen. 311, 112121.
Uberoi, M., Pereira, C.J., 1996. External mass transfer coefcients
for monolith catalysts. Ind. Eng. Chem. Res. 35 (1), 113116.
Zamaniyan, A., Mortazavi, Y., Khodadadi, A., Mana, H., 2010.
Tube tted bulk monolithic catalyst as novel structured
reactor for gassolid reactions. Appl. Catal. A: Gen. 385 (1/2),
214223.