Kinetics of dimethyl carbonate synthesis from methanol and carbon dioxide over ZrO2-MgO catalyst in the presence of butylene oxide as additive

Article abstract of DOI:10.1016/j.apcata.2011.07.004

A kinetic investigation of dimethyl carbonate (DMC) synthesis from methanol and CO₂ over ZrO₂-MgO was performed by using butylene oxide as a chemical trap for the water formed during the reaction. The effect of the catalyst amount, the stirring speed, the temperature, as well as the amount of butylene oxide on the reaction rate and the selectivity to DMC was studied. The analysis of the reaction pathway suggests that DMC and butylene glycol are formed via the reaction of adsorbed mono-methoxycarbonate intermediate and methoxybutanol or methanol. A kinetic model was developed based on the reaction mechanism and it was in agreement with the experimental data. The apparent activation energy for the formation of DMC was 62 kJ/mol.

Full text of DOI:10.1016/j.apcata.2011.07.004

Applied Catalysis A: General 404 (2011) 39–46
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Kinetics of dimethyl carbonate synthesis from methanol and carbon dioxide over
ZrO2–MgO catalyst in the presence of butylene oxide as additive
Valerie Etaa, Päivi Mäki-Arvelaa, Johan Wärnåa, Tapio Salmia, Jyri-Pekka Mikkolaa,b
,
Dmitry Yu. Murzina,∗
a Laboratory of Industrial Chemistry and Reaction Engineering, Process Chemistry Centre, Åbo Akademi University, Biskopsgatan 8, 20500 Turku/Åbo, Finland
b Technical Chemistry, Department of Chemistry, Chemical–Biological Center, Umeå University, SE-90187 Umeå, Sweden
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 6 May 2011
Received in revised form 20 June 2011
Accepted 6 July 2011
Available online 18 July 2011
A kinetic investigation of dimethyl carbonate (DMC) synthesis from methanol and CO2 over ZrO2–MgO
was performed by using butylene oxide as a chemical trap for the water formed during the reaction.
The effect of the catalyst amount, the stirring speed, the temperature, as well as the amount of butylene
oxide on the reaction rate and the selectivity to DMC was studied. The analysis of the reaction pathway
suggests that DMC and butylene glycol are formed via the reaction of adsorbed mono-methoxycarbonate
intermediate and methoxybutanol or methanol. A kinetic model was developed based on the reaction
mechanism and it was in agreement with the experimental data. The apparent activation energy for the
formation of DMC was 62 kJ/mol.
Keywords:
Dimethyl carbonate
Methanol
Carbon dioxide
Butylene oxide
Kinetics
© 2011 Elsevier B.V. All rights reserved.
ZrO2–MgO
1. Introduction
methylating reagent replacing toxic phosgene, methyl iodide and
dimethyl sulphate traditionally used [7,8]. DMC is also an interme-
Efficient transformation of carbon dioxide (CO2) derived from
anthropogenic sources into useful chemicals is an important mile-
stone in research and development, which may contribute to the
reduction of greenhouse gas emissions via carbon recycling [1]. The
utilisation of CO2 as a carbon resource for the benign synthesis of
chemicals is of particular interest, since CO2 is abundant, afford-
able, non-toxic and non-flammable. This synthetic methodology
favours the implementation of more direct reaction pathways with
a high atom efficiency and at the same time, avoiding the use of
toxic chemicals, such as phosgene. In spite of these advantages,
industrial utilisation of CO2 in chemical syntheses remains a sig-
nificant challenge, because it lays in a deep potential energy well
(−400 kJ/mol), requiring huge amounts of energy for activation [2].
To overcome this challenge, high temperatures, extremely reactive
reagents, metal salts, photochemical and electrochemical energy
have been proposed [3–6] as tools to mimic natural phenomena
such as photosynthesis. CO2 can therefore be used in reactions,
such as the synthesis of organic carbonates in which the –COO–
moiety is maintained.
diate for the synthesis of higher carbonates and carbamates and
as an electrolyte in the production of lithium batteries [9]. Sev-
eral routes have been developed for the synthesis of DMC, such
as methanol phosgenation, oxidative carbonylation of methanol,
methanolysis of urea, transesterification of cyclic carbonates, as
well as the direct synthesis from CO2 and methanol. Although
methodologies such as phosgenation are credited for their high
reactivity and high productivity of DMC, the toxicity of phosgene
both to health and the environment limits its applications for
cleaner syntheses.
Among the DMC synthesis technologies, the direct reaction of
CO2 and methanol is the most attractive route from the viewpoint
of green chemistry and sustainable development. The direct route
is among the most efficient in terms of the atom economy, produc-
ing only water as a by-product. This process undoubtedly results
in environmental and economic benefits with waste minimisation.
However, the thermodynamic limitations embedded in the pro-
cess [10], and the inertness of CO2 restrain the synthesis of DMC
in high yields. In order to overcome these obstacles, the removal
of water from the catalytic reaction system using acetonitrile or
trimethoxymethane as promoters [11,12] have been proposed. The
use of butylene oxide as a chemical water scavenger [13] has also
been demonstrated in the presence of ZrO2–MgO catalyst resulting
in 7.2% yield and 53% selectivity to DMC. To understand the chem-
ical processes occurring at the molecular level, a detailed kinetic
Dimethyl carbonate (DMC) has attracted much attention as
an environmentally benign and important carbonylating and
∗
Corresponding author.
E-mail address: dmurzin@abo.fi (D.Yu. Murzin).
0926-860X/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2011.07.004

40
V. Eta et al. / Applied Catalysis A: General 404 (2011) 39–46
Nomenclature
*
adsorption site
c
d
D
i
concentration (mol/l)
diameter (mm)
diffusion coefficient (m2/s)
component index
Scheme 1. Reaction scheme for the synthesis of DMC from methanol and CO2 in the
presence of a chemical water scavenger.
K
k
equilibrium constant
rate constant
M
N
P
Q
r
t
ˇ
ε
Á
molar mass (g/mol)
stoichiometric number
pressure (bar)
ing the pH at 9 by a constant addition of appropriate amounts
of NH4OH. The precipitate was washed several times after age-
ing with distilled water to remove the unreacted NH4OH. The
resulting cake was dried and calcined at 700 ◦C for 3 h to obtain a
white powder (MgO–ZrO2). The specific surface area of the catalyst
was determined by measuring the nitrogen adsorption–desorption
isotherms at 77 K (Sorptomatic 1900) in liquid nitrogen. Prior to
the measurements, the samples were outgassed for 3 h at 473 K to
a residual pressure below 0.01 Pa. The total surface area calculated
according to the Brunauer–Emmett–Teller (BET) [14] equation was
55 m2/g.
objective function (mol/l)2
generation rate (decomposition rate) (mol/hg)
time (h)
mass transfer coefficient (m/s)
dissipation energy (J)
viscosity (Pa s)
coverage
density (g/l)
Â
ꢀ
ꢀB
BC
BG
BO
DMC
MC
MeOH methanol
˛
y
catalyst bulk density (g/l)
butylene carbonate
butylene carbonate
butylene oxide
dimethyl carbonate
mono-methyl carbonate
2.3. Experimental setup and catalytic tests
All the reactions were performed in a high pressure autoclave
(Parr Inc., volume 300 ml) equipped with a stirrer, baffles and a
sampling line controlled by a system of valves. The temperature
in the reactor was regulated to an accuracy of 1 ◦C by a temper-
ature controller. The stirrer operated with a system of four blade
impellers rotating at 700 rpm. In a typical reaction sequence, the
catalyst (1 g), methanol (695 mmol, 28.1 ml) and butylene oxide
(21.3 mmol, 1.8 ml) were introduced into the reactor and purged
with argon and then CO2 to expel any trapped gases. The reactor
was pressurised to ca 4.5 MPa with CO2 and then heated to the
desired temperature (120, 150 or 180 ◦C). At specific time intervals,
samples were withdrawn and cooled in the sampling loop prior
to the analysis. The samples were analysed by gas chromatogra-
phy (Agilent Technologies, 6890 N) with a 60 m × 320 ␮m × 0.5 ␮m
capillary column (J&W 123-506 E DB5) equipped with a flame ion-
isation detector and using 2-propanol as an internal standard. The
products were identified by GC–MS (Agilent Technologies, 6890 N).
K1MKB5/K3
[ ]
MeOH
[
]
ꢀ
ꢁ−1
ꢁ
ꢀ
˛
y+1 DMC
(
)[
]
ω
ꢁ
+ 1
2
ꢀ
(˛ +y) MeOH
[
]
ꢀ
ꢁꢀ
ꢀ
ꢁ
ꢀ
DMC
]
y
[
[
2˛
1 −
ꢀ
ꢀ
MeOH
]
˛ +y
˛ +y
description is required. Knowledge and transformation of such data
into mathematical models can foster the design and reaction inten-
sification aspects of chemical processes.
In this paper, kinetic investigations to understand the detailed
reaction mechanism leading to these improved yields under dif-
ferent conditions using ZrO2–MgO catalysts are reported. The
influence of the reaction parameters on the synthesis of DMC was
examined and the interactions of carbon dioxide and methanol
with the catalyst surface were discussed. In addition, a kinetic
model was developed based on the experimental data. The activa-
tion energy of the reaction was estimated from the kinetic model
at different temperatures.
3. Results and discussion
3.1. Qualitative kinetics
3.1.1. Butylene oxide as water trap
The substrates consisted of a mixture of methanol, CO2 and buty-
lene oxide, and the catalyst composed of ZrO2 and MgO. Mixed
oxides of Zr and Mg have previously demonstrated positive cat-
alytic effects for the synthesis of DMC from methanol and CO2 in the
presence of butylene oxide as water abstracting additive [13]. Since
the concomitant production of water during this reaction essen-
tially limits the synthesis of DMC in high yields, the in situ removal
of the water from the reaction medium shifts the reaction equi-
librium towards higher DMC yields (Scheme 1). Butylene glycol is
formed from the reaction of butylene oxide with water, thereby
limiting the hydrolysis of the DMC formed.
2. Experimental
2.1. Materials
Zirconyl nitrate, aqueous ammonium hydroxide (28%),
methanol (99.8%), butylene oxide (99%), DMC (99%), and magne-
sium nitrate, were purchased from Sigma–Aldrich. CO2 (99.99%)
was purchased from AGA. All chemicals were used as received.
The abstraction of water from the reaction medium leads to
the relaxation of the thermodynamic constraints, consequently
resulting in an increase in the yield of DMC. The water abstract-
ing additive, however, reacts with methanol and CO2, forming
methoxybutanol (1-methoxybutanol and 2-methoxybutanol) and
butylene carbonate, respectively, as reaction intermediates. In
addition, trace amounts of dimethyl ether were identified as a
result of the dehydration of methanol. The maximum conversion
2.2. Catalyst synthesis and characterisation
The catalyst was synthesised by the co-precipitation of 0.5 M
solution of zirconyl nitrate and 0.3 M solution of magnesium nitrate
hydrate in 500 ml followed by the drop-wise addition of ammo-
nium hydroxide under stirring until the pH was adjusted to 9.
The precipitate obtained was aged for 3 days, while maintain-

V. Eta et al. / Applied Catalysis A: General 404 (2011) 39–46
41
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
500 rpm
700 rpm
900 rpm
0
5
10
15
20
25
30
time (h)
Fig. 2. Effect of stirring rate on the synthesis of DMC from methanol (695 mmol)
and CO2 using butylene oxide (21 mmol) as a chemical trap for water at 9.5 MPa and
1 g catalyst.
Fig. 1. Concentration profiles of the reactants and products as
a function of
reaction time. Reaction conditions: 1 g catalyst, 695 mmol methanol, 21 mmol buty-
lene oxide, 150 ◦C, 700 rpm, 9 MPa (MeOH, methanol; BO, butylene oxide; MB,
methoxybutanol; BG, butylene carbonate).
3.1.2. Effect of stirring on DMC formation
Kinetic results correlated to different reaction parameters can
be meaningful only in the absence of mass transfer limitations. The
reaction studied at 500, 700 and 900 rpm did not show any appre-
ciable difference in the formation of DMC. The results indicate that
the experiments were conducted in the absence of mass transfer
limitations when catalyst particles smaller than 100 ␮m were used
(Fig. 2). The mass transfer resulting from the film diffusion of CO2
into methanol and to the solid catalyst particles was minimised
by the fast stirring speed and small catalyst particles used during
the reaction. The presence of baffles ensured efficient mixing of the
catalyst and substrates, and the rate was independent of the stir-
ring speed used indicating that the reaction was performed in the
kinetic regime.
of methanol and butylene oxide obtained after 26 h at 150 ◦C was
11.7% and 100%, respectively. The concentration variations of the
species in the reaction mixture as a function of reaction time at
150 ◦C and 9 MPa are shown in Fig. 1.
The amount of DMC increased gradually as the reaction pro-
ceeded and levelled off after 9 h, indicating the attainment of
equilibrium. The conversion of butylene oxide after 7 h was 90%,
suggesting an increase in the accumulation of water as more DMC
was produced. The decrease in the concentration of butylene oxide
resulted in a corresponding increase in the amount of methoxybu-
tanol or butylene carbonate, which eventually transform into
butylene glycol by reacting with methanol. Methoxybutanol and
butylene carbonate present as reaction intermediates increased
steadily in the beginning of the reaction until 5 h, after which the
concentrations slowly decreased as DMC and butylene glycol were
formed. The low concentration of butylene carbonate in compari-
son to methoxybutanol may be attributed to the slow reactivity of
butylene oxide and CO2, and the decomposition of resulting buty-
lene carbonate to methoxybutanol (Scheme 2). The yield of DMC
obtained after 26 h was 6.2% with 11.7% conversion of methanol.
Comparing the amount of DMC and butylene glycol, the lower
concentration of butylene glycol may be due to the formation of
dibutylene glycol or water uptake by magnesium species in the
catalyst. The results suggest that the reaction does not proceed pri-
marily via the transesterification route although cycloaddition and
transesterification are highly feasible in such catalytic systems.
3.1.3. Effect of catalyst particle size and catalyst loading on DMC
formation
The effect of internal mass transfer limitation was also investi-
gated by comparing the catalyst particles below 63 ␮m (average
particle size 1.5 nm) and 100 ␮m (average particle size 2.1 nm).
The catalyst effectiveness was demonstrated by using minimal
catalyst loading to obtain high yields of DMC. The effect of the cat-
alyst loading and the catalyst particle size on the concentration
of DMC yield over time is shown in Fig. 3. The concentration of
DMC increased steadily as the reaction progressed, but remained
constant after 21 h. This shows that the thermodynamic limitations
initially suppressed by the use of butylene oxide resurface, and thus,
the concentration of DMC remains constant.
The initial rate of 1.2 mmol/h was obtained using 0.5 g catalyst
during 5 h. When the catalyst loading increased to 1 g, the initial
rate also doubled to 2.4 mol/h. The proportional increase of the ini-
tial reaction rate with the catalyst loading suggests that the initial
stage of the reaction is not controlled by gas–liquid mass transfer
but by the intrinsic kinetics. However, the reaction rate for DMC
formation decreased after 7 h for both 1 g and 0.5 g of catalyst and
remained constant after 21 h. Since butylene oxide is completely
used up, the formation of water may lead to further thermody-
namic constraints restricting an increase in the DMC concentration.
During 5 h of the reaction, the concentration of methoxybutanol
showed a maximum at ca 45% conversion of butylene oxide (Fig. 4),
and increasing the concentrations of DMC and butylene glycol.
The results also show that the formation rate DMC was inde-
pendent of the catalyst particle size when efficient stirring speed of
700 rpm was used. The effect of the diffusion resistance on the cata-
lyst with particle sizes <63 ␮m and <100 ␮m was minimal, since the
O
O
2MeOH
DMC
BG
CO2
+
O
CH3
O
H C
CO2
3
CH3OH
CH3OH
O
H3C
CH3
HO
Scheme 2. The cycloaddition of CO2 to butylene oxide to yield butylene carbonate
and the decomposition and/or transesterification with methanol.

42
V. Eta et al. / Applied Catalysis A: General 404 (2011) 39–46
0.8
0.75
0.6
0.45
0.3
0.15
0
0.6
0.5 g < 100 µm
1 g < 100 µm
1 g < 63 µm
0.4
0.2
0
Me:Bu=20
Me:Bu=30
Me:Bu=45
0
5
10
15
20
25
30
time (h)
0
5
10
15
20
25
30
time (h)
Fig. 5. DMC concentration and selectivity obtained at 150 ◦C and 9.5 MPa using
methanol to butylene oxide molar ratios of 20, 30 and 45, and 1 g catalyst.
Fig. 3. The dependence of the DMC concentration on time using different cata-
lyst loadings and catalyst particle size. Reaction conditions: methanol 695 mmol,
butylene oxide 21 mmol, 9.5 MPa, 700 rpm.
limited amount of butylene oxide favours the nucleophilic attack
of methanol on butylene oxide in comparison to the formation of
butylene carbonate, resulting in the formation of methoxybutanol,
which is eventually transformed to DMC and butylene glycol.
An increase in the amount of butylene oxide in the substrate
promoted reactions with both methanol and CO2, thus decreasing
the selectivity of DMC. The concentration of DMC did not increase
linearly as the concentration of butylene oxide in the substrate
increased, showing that the reaction intermediates persist in the
reaction medium decreasing the selectivity to DMC. The amount
of methoxybutanol increased as the amount of butylene oxide in
the substrate increased, owing to the ease of reactions in the liquid
phase. The optimal molar ratio of 30 exhibiting a reasonable yield
and selectivity towards DMC with maximum DMC conversion was
used in the study.
concentrations of DMC obtained were similar. This shows that for
very slow reactions, such as the synthesis of DMC from methanol
and CO2 using catalyst particles less than 100 ␮m, the effect of the
gas–liquid mass transfer may not influence the product formation.
3.1.4. Kinetic analysis of the effect of methanol/butylene oxide
initial molar ratio
The dependence of DMC on butylene oxide as a chemical dehy-
drating additive was investigated with different concentrations
of methanol and butylene oxide. The influence of methanol-to-
epoxide ratio on the yield and selectivity to DMC was studied at
150 ◦C and 9.5 MPa. The consumption of butylene oxide during the
reaction resulted in the formation of methoxybutanol, butylene
carbonate, butylene glycol and DMC in different molar ratios. A
decrease in the molar ratio of methanol-to-butylene oxide in the
substrate from 45 to 20 promoted the yield of DMC via the forma-
tion of methoxybutanol albeit a decrease in the selectivity (Fig. 5).
When comparing the selectivity at the same conversion level of
methanol e.g. 4%, it can be observed that the selectivity to DMC
using Me:Bu = 45, Me:Bu = 30 and Me:Bu = 20 was 61%, 41% and
33% respectively (Fig. 6). The results show that the presence of a
3.1.5. Kinetic analysis of the effect of temperature on DMC
synthesis
The influence of temperature on the reaction kinetics of DMC
synthesis was investigated at 120, 150 and 180 ◦C at 9.5 MPa. The
initial rates for the conversion of methanol to DMC or to all other
0.15
0.12
0.09
0.06
0.03
0
80
70
60
50
40
30
Me:Bu=45
Me:Bu=30
Me:Bu=20
20
10
0
0
20
40
60
80
100
120
0
2
4
6
8
10
12
14
Conversion of BO (%)
Conversion of methanol (%)
Fig. 4. The dependence of the concentration of methoxybutanol (MB) on the con-
version of butylene oxide (BO).
Fig. 6. The effect of the amount of butylene oxide on DMC selectivity as a function
of methanol conversion.

V. Eta et al. / Applied Catalysis A: General 404 (2011) 39–46
43
14
12
10
8
60
50
120 °C
150 °C
180 °C
120 °C
150 °C
180 °C
40
30
120 °C
150 °C
180 °C
6
20
4
10
0
2
0
0
2
4
6
8
10
12
14
0
5
10
15
20
25
30
time (h)
Methanol conversion (%)
Fig. 7. The conversion of methanol to all reaction products (solid lines) and the
Fig. 9. The effect of temperature on the selectivity to DMC as a function of the
conversion of methanol.
calculated methanol conversion only to DMC (broken lines) at 120, 150 and 180 ◦
as a function of time.
C
temperatures and for prolonged reaction times in accordance with
earlier observations [13,15].
products were analysed during 5 h at different temperatures (Eq.
(1))
The formation rates of methoxybutanol and butylene carbonate
also increased as the temperature increased due to the reactions of
methanol and CO2 with butylene oxide. For example, at 5 h, the con-
version of methanol increased from 3.2 mol% at 120 ◦C to 10.3 mol%
at 180 ◦C. Similarly, the butylene oxide conversion increased from
56 mol% to 92 mol% as the temperature increased from 120 to
180 ◦C. The selectivity to DMC therefore decreases with increas-
ing temperature (Fig. 9). Owing to the activation of methanol and
CO2 at high temperature, higher DMC yields with a lower selectivity
were obtained. In addition, dehydration of methanol leading to the
formation of dimethoxymethane is favourable at higher tempera-
tures. The optimum temperature of 150 ◦C was therefore chosen for
the reaction since a maximum yield and selectivity towards DMC
were obtained, and desorption of CO2 and H2O from the catalyst
surface can be achieved at this temperature [16].
The selectivity of DMC increased at the same conversion level
of methanol with decreasing temperature, e.g. at the same con-
version of 4%, a selectivity of 25% was obtained at 180 ◦C while
the selectivity at 120 ◦C increased to 46%. This also shows that a
lower temperature is favourable for the DMC formation although
at a decreasing conversion of methanol. This can also be seen
by comparing the conversion of methanol with the calculated
consumption of methanol based on the formation of DMC. The
amount of butylene glycol formed also increased with an increase
in the temperature (Fig. 10) due to the dehydration of methanol to
dimethyl ether and water, which further reacts with methoxybu-
tanol or butylene oxide to yield butylene glycol and decreasing the
selectivity to DMC.
rDMC
ꢂ
R =
,
(1)
rc
where R is the ratio of the rates, rc is the rate of the various compo-
nents during 5 h at a particular temperature.
The ratios of the conversion of methanol-to-DMC to the conver-
sion of methanol to all products were 0.56, 0.53 and 0.45 for 120,
150 and 180 ◦C, respectively (Fig. 7). This shows that DMC selectiv-
ity decreased with an increase in temperature from 120 to 180 ◦C.
At low temperatures, the catalytic transformation of methanol over
ZrO2–MgO results in the formation of DMC. At the same time,
reactions at low temperatures limit the dehydration of methanol
to dimethoxymethane and retard the liquid phase reaction of
methanol and butylene oxide, thereby increasing the selectivity.
Therefore, the conversion of methanol to DMC, methoxybutanol
and other reaction products, such as dimethoxymethane depend
on the reaction temperature.
The rate of formation of DMC increased steadily with an increase
of the reaction temperature and levelled off after ca 9 h. The ini-
tial rates of DMC formation obtained during 5 h at 120, 150 and
180 ◦C were 0.9 mmol/h, 2.4 mmol/h and 2.7 mmol/h, respectively
(Fig. 8). At 180 ◦C, the concentration of DMC remained constant at
ca 0.7 mol/l after 9 h whereas a slow increase was observed at 120
and 150 ◦C. This result indicates that equilibrium was established
for the formation and consumption or hydrolysis of DMC, at high
0.8
0.7
0.6
0.5
0.4
0.3
0.2
4. Analysis of the reaction mechanism
The weak basic and acidic sites as well as the tetragonal phase of
ZrO2 have been found to be important for the synthesis of DMC [17]
from methanol and CO2 in the presence of butylene oxide as a water
trap. DMC synthesis is initiated by the adsorption of methanol
as mono- or di-methoxide species on the surface of tetragonal
phase of ZrO2, stabilised by MgO. The proton from methanol, at the
same time, is transferred to the oxygen atom of ZrO2. The molec-
ularly adsorbed methoxide species forms loose associations with
the active sites of the catalyst, permitting further reactions with
the carbonyl function of CO2. The dissolution and interaction of
CO2 with the adsorbed methoxide groups result in the formation of
mono-methyl carbonate species. This has been confirmed by spec-
troscopic studies for the elementary steps in DMC formation using
120 °C
150 °C
180 °C
0.1
0.0
0
5
10
15
time (h)
20
25
30
Fig. 8. The effect of temperature on the formation of DMC obtained from methanol
(695 mmol) and CO2 using butylene oxide (21 mmol) as a chemical trap for water at
9.5 MPa and 1 g catalyst.

44
V. Eta et al. / Applied Catalysis A: General 404 (2011) 39–46
O
OH
H2C
O
O
C
H3C
OH
H3C O-
Zr2+
C
H3C
O-
O
CH
Zr2+
O
O
+
O
C
O
H2C
OH
CH3
CH3
H2C
CH
H2C
CH3
CH3
Scheme 3. Proposed reaction pathway for the formation of DMC and butylene glycol (BG) from methanol and CO2 using butylene oxide (BO) as a chemical trap for water.
Zr4+ or Cu+ as catalysts [18], and the in situ infrared and Raman stud-
ies using ZrO2 [19,20]. The reaction step involving the dissolution
and reaction of CO2 is assumed to be slow, since CO2 remained non-
activated and the reaction with the adsorbed methoxide species is
slow. The nucleophilic attack of mono-methyl carbonate species on
methoxybutanol, formed from the reaction of methanol and buty-
lene oxide or cyclic carbonate, results in the formation of DMC and
butylene glycol (Scheme 3). The protonation of methanol can also
lead to the formation of methoxonium ions [CH3OH2]+. The nucle-
ophilic attack of the methoxide group on the methoxonium ion
results in the formation of dimethyl ether (DME) as a by-product.
a factor of 30. Quasi-equilibria were applied to all steps except the
step involving the reaction of CO2, and the liquid-phase reaction
of butylene oxide and methanol. DMC was formed via two parallel
routes. Based on these hypotheses, the reaction steps leading to the
formation of the products can be written as follows:
N1 N2 N3
1
2
3
4
5
MeOH + ∗ ≡ MeOH∗
MeOH ∗ +CO2 ⇔ MC∗
1
1
1
0
0
2
1
0
0
1
0
0
0
1
0
MC ∗ +MB ≡ DMC + BG + ∗
BO + MeOH ⇒ MB
MC ∗ +MeOH∗ ≡ DMC + H2O + 2∗
5. Kinetic modelling
N1 : MeOH + CO2 + MB → DMC + BG
N2 : 2MeOH + CO2 → DMC + H2O
N3 : BO + MeOH → MB,
(2)
5.1. Derivation of kinetic model
The main reactions leading to the formation of DMC, such as,
adsorption of methanol and other species on the catalyst active
sites, the reaction of the activated methanol with CO2 to form
monomethyl carbonate species and the subsequent formation of
DMC are described in Scheme 4.
where * is the free adsorption sites, A* is the adsorption of species
A on the catalyst surface, N1, N2, N3 are reaction routes and the
numbers before the adsorptions/reactions are the reaction steps.
After these steps stoihiometric numbers along the reaction routes
are given.
Three main reaction products observed were DMC, butylene gly-
col and methoxybutanol with trace amounts of butylene carbonate.
The Eley-Rideal type of reaction mechanism was adopted for the
kinetic modelling, according to which the liquid reactants adsorb on
the catalyst active sites and subsequently react with the dissolved
CO2 from the gas phase. The pressure of CO2 in the reactor ensures
that equilibrium is established between the CO2 dissolved in the
liquid phase and the CO2 reacting with the activated methanol.
Hence, it is convenient to express the reaction rate as a function of
the CO2 pressure in the gas phase rather than the concentration of
CO2 in the liquid phase. The mechanism also assumes that methanol
adsorbs on the catalyst active sites preferentially to butylene oxide.
This assumption is reasonable since the molar concentration of
methanol is higher than the concentration of butylene oxide by
The reaction step 2 is the rate-determining catalytic step, since
the reaction of the activated methanol with the diffused CO2 is pos-
tulated to be slow in comparison to the other reaction steps. Step
4 represents the non-catalytic liquid-phase reaction for the forma-
tion of methoxybutanol. The kinetic equation can be derived from
coverage of the reacting species as follows:
ÂMeOH
K1
K3
=
=
⇒ ÂMeOH = K1 MeOH Â
(3)
(4)
[
]
v
MeOH Â
[
[
]
v
DMC BG Â
DMC BG Â
] [
] [
]
[
]
v
v
⇒ ÂMC
=
.
K3 MB
MB Â
[
]
[
]
MC
The balance equation for the sites is
 + ÂMeOH + ÂMC = 1.
(5)
v
8
7
6
5
4
MeOH
H2O
MC
CO2
DMC
120 °C
3
150 °C
MeOH
2
BG
180 °C
1
0
BO
MB
0
0.5
1
1.5
2
2.5
3
3.5
4
Butylene glycol yield (%)
Scheme 4. Reaction network for the formation of DMC from methanol and carbon
dioxide using butylene oxide as a chemical trap for water (MeOH, methanol; BO,
butylene oxide; MC, mono-methyl carbonate; BG, butylene carbonate).
Fig. 10. The dependence of the yield of DMC on butylene glycol yield.

V. Eta et al. / Applied Catalysis A: General 404 (2011) 39–46
45
Table 1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
The estimated value of the rate constant ki [mol/h.gcat.bar] from the kinetic mod-
elling at 150 ◦C and 45 bar of initial CO2 pressure (degree of explanation = 99.98%).
Parameter
Estimated
STD error
Relative STD
error
(%) Parameter
STD error
k2
k4
1.60 × 10−5
1.12 × 10−2
1.09 × 10−6
7.55 × 10−4
6.8
6.7
14.7
14.8
From the reaction rates and the stoichiometric scheme, the rate
for the formation of DMC, methoxybutanol, butylene carbonate,
and the rate of decomposition of butylene oxide and methanol in
the batch reactor can be written as follows:
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
d DMC
[
]
= r2ꢀB
(16)
(17)
(18)
(19)
(20)
dt
[BG]
d MB
[
]
Fig. 11. The test plot of MB DMC
] ([
−
BG / MeOH against [BG] according to Eq.
])
[
]
[
[
]
= (r4 − ꢁr2ꢀB)ω
= r4 − (r4 − ꢁr2ꢀB)ω
= −r4
dt
(15) using the experimental data giving a linear fit. The calculated value of ˛ was
1.3 × 10−3
.
d BG
[
]
dt
Combination of Eqs. (3–5) gives the coverage of the vacant
adsorption sites
d BO
[
]
dt
1
Â
v
=
.
(6)
d MeOH
[
]
1 + K1 MeOH + DMC BG /K MB
] [
The reaction rate for step 2 can be written as follows:
[
]
[
]
[
]
3
= −2r2ꢀB − (r4 − ꢁr2ꢀB)ω,
dt
where ꢀB is the bulk density of the catalyst, and ꢁ and ω are
merged quantities arising from a rigorous treatment of mass bal-
ances (available from authors, see notation).
r2 = k2ÂMeOHPCO − k−2ÂMC
.
(7)
2
Substituting the respective coverages into Eq. (7) gives
ꢃ
ꢄ
k2 [MeOH]PCO − (1/K)([DMC][BG]/[MB])
2
5.2. Parameter estimation
r2
=
,
(8)
1 + ˛[MeOH] + ˇ([DMC][BG]/[MB])
The numerical parameter estimation was based on the kinetic
experiments in which the temperature and pressure were main-
tained constant, and the changes in the concentration of the species
were recorded by gas chromatography. The model predictions of
the kinetic parameters were obtained by solving the ordinary dif-
ferential equations (ODEs; Eqs. (16–19)), which describe the mass
balances. A rapid and stable prediction of the components was
obtained by applying the backward differential method using the
software ModEst [21], employing a stiff ODE solver [22] and the
residual sum of squares (Q) was minimised by a combined simplex
Levenberg–Marquardt algorithm according to Eq. (21):
ꢅ
where K = DMC BG / MB MeOH P 2 ; ˛ = K1, ˇ = 1/K3.
[
] [
]
[
] [
]
CO
The value of K can be obtained from the experimental data.
The rate for step 4 can be expressed as
r4 = k4 BO MeOH .
(9)
[
] [
]
The equilibrium constant for step 5 can be written as
DMC H O Â2 DMC H O Â2
[
] [ ] [
]
[
]
2
v
2
v
K5
=
⇒ ÂMC
=
.
(10)
ÂMeOHÂMC
K5ÂMeOH
Since the coverage of the intermediate adsorbed species ÂMC is
the same in Eqs. (4) and (10), it can be represented as follows:
Q =
(ci(t) − cˆi(t))2,
(21)
DMC BG Â
DMC H O Â2
]
2
v
[
] [
] [
]
]
[
v
=
,
(11)
K3 MB
[
K5ÂMeOH
which together with Eq. (3) gives
K1K5 BG MeOH
where the estimated concentrations (cˆi) were obtained from the
models, and the concentrations (ci) were obtained from the labo-
ratory experiments.
[
] [
]
H O =
.
(12)
[
]
2
K3 MB
[
]
Fig. 12 shows that the simulated curves obtained were in good
agreement with the experimental points. The initial parameter esti-
mation showed that the inclusion of the equilibrium constants,
K1 and K3 in the parameter estimations, resulted in an over-
parameterised system and did not improve the fit of the model
to the experimental data. The constants K1 and K3 correlated with
the rate constant k2, therefore, solving for the parameters resulted
in high relative errors (>1000%). As a result, ˛ and ˇ, a function of
K1 and K3 respectively, were set to zero. The estimated values at
150 ◦C for the kinetic parameters k2 and k4 and the errors are sum-
marised in Table 1. The residual sum of squares obtained was 0.046
and the degree of explanation was 99.98%. The activation energy
was obtained by performing separate regression analysis at 120,
150 and 180 ◦C. The results provided different values for the rate
constant k2, from which the slope (Ea/R) was obtained from the
plot of the natural logarithms versus the inverse of the respective
temperatures. The activation energy estimated was 62 6.6 kJ/mol.
DMC is formed via two parallel routes, i.e. reactions (3) and (5),
thus the following mass balance can be written as
H O + BG = DMC .
(13)
[
]
[
]
[
]
2
Substituting Eq. (12) into Eq. (13) enables the elimination of the
concentration of water. The following relation is obtained:
ꢀ
ꢁ
K1K5 MeOH
[
]
BG 1 +
= DMC ,
(14)
[
]
[
]
K3
MB
]
[
which can be rewritten as
DMC − BG
K1K5
K3
[
[
]
[
[
]
=
MeOH / MB
]
BG = ˛ BG .
(15)
[
]
[
]
]
A test plot (Fig. 11) is linear, thus confirming that Eq. (15) is able
to describe the experimental data. The water concentration can be
obtained from Eq. (13).

46
V. Eta et al. / Applied Catalysis A: General 404 (2011) 39–46
Fig. 12. The fit of the kinetic model compared with the experimental data at 150 ◦C, initial pressure of CO2 at 45 bar. The concentrations of different species as a function of
time for the experimental (points) and estimated values (continuous lines).
6. Conclusions
References
[1] M. Aresta, A. Dibenedetto, Dalton Trans. (2007) 2975–2992.
[2] M. Aresta, Quimica Nova 22 (1999) 269–272.
[3] R. Srivastava, T.H. Bennur, D. Srinivas, J. Mol. Catal. A: Chem. 226 (2005)
199–205.
[4] R. Srivastava, D. Srinivas, P. Ratnasamy, J. Catal. 233 (2005) 1–15.
[5] E. Fujita, B.S. Brunschwig, T. Ogata, S. Yanagida, Coord. Chem. Rev. 132 (1994)
195–200.
[6] M.A. Casadei, A. Inesi, L. Rossi, Tetrahedron Lett. 38 (1997) 3565–3568.
[7] Y. Ono, Catal. Today 35 (1997) 15–25.
[8] P. Tundo, Pure Appl. Chem. 73 (2001) 1117–1124.
[9] D. Aurbacha, Y. Talyosefa, B. Markovskya, E. Markevicha, E. Zin-
igrada, L. Asrafa, J.S. Gnanaraja, H.-J. Kimb, Electrochim. Acta 50 (2004)
247–254.
[10] V. Eta, P. Mäki-Arvela, D. Murzin, Y. Salmi, T.J.-P. Mikkola, in: D.L. Marmaduke
(Ed.), Progress in Heterogeneous Catalysis, Nova Science Publishers, Haup-
pauge, NY, 2008, pp. 135–155.
[11] Z.-F. Zhang, Z.-W. Liu, J. Lu, Z.-T. Liu, Ind. Eng. Chem. Res. 50 (2011)
1981–1988.
[12] M. Honda, A. Suzuki, B. Noorjahan, K. Fujimoto, K. Suzuki, K. Tomishige, Chem.
Commun. (2009) 4596–4598.
The kinetics of the synthesis of DMC from methanol and CO2
using butylene oxide as a chemical trap for water over ZrO2–MgO
was studied with the temperature range 120–180 ◦C. Mass trans-
fer resistances were eluded by using catalysts with particle sizes
<100 ␮m and 700 rpm of stirring speed. The methanol-to-butylene
oxide ratio of 30 showed a reasonable conversion of methanol
and DMC selectivity at 150 ◦C. DMC and butylene glycol were
formed via the reaction of adsorbed mono-methoxycarbonate
intermediate and methoxybutanol. DMC was also formed via
the mono-methoxycarbonate reaction with methanol, producing
water as by-product in the absence of butylene oxide. Regression
analysis performed for the advanced mechanism provided reason-
able description of the concentration dependencies of the species
over time for reactions performed at different temperatures. The
apparent activation energy was determined as 62 kJ/mol.
[13] V. Eta, P. Mäki-Arvela, A.-R. Leino, K. Kordás, T. Salmi, D. Murzin, J.-P. Mikkola,
Ind. Eng. Chem. Res. 49 (2010) 9609–9617.
[14] S. Brunauer, P.H. Emmett, E. Teller, J. Am. Chem. Soc. 60 (1938) 309–319.
[15] Y. Kishimoto, I. Ogawa, Ind. Eng. Chem. Res. 43 (2004) 8155–8162.
[16] H. Hattori, Chem. Rev. 95 (1995) 537–558.
[17] K.T. Jung, A.T. Bell, Top. Catal. 20 (2002) 97–105.
[18] Y. Zhang, A.T. Bell, J. Catal. 255 (2008) 153–216.
[19] K.T. Jung, A.T. Bell, J. Catal. 204 (2001) 339–347.
[20] S. Xie, A.T. Bell, Catal. Lett. 70 (2000) 137–143.
Acknowledgements
This work is part of the activities of Åbo Akademi Process Chem-
istry Centre within the Finnish Centre of Excellence (2006–2011),
and the KETJU Research Programme (2006–2010) by the Academy
of Finland. The authors are grateful to the Academy of Finland for
the financial support under the Grants 120853, 124357 and 128626.
This work is also associated with the activities of Umeå Univer-
sity Chemical-Biological Centre whereupon financial support from
Bio4Energy Programme, Kempe Foundations and Knut and Alice
Wallenberg Foundation are acknowledged.
[21] H. Haario, MODEST User’s Manual
2001.
, Profmath Oy, Helsinki, Finland,
[22] A. Hindmarsh, in: R. Steppleman (Ed.), C. OPEPARK – A Systematized Collec-
tion of ODE Solvers in Scientific Computing, IMACS/North Holland Publishing
Company, Amsterdam, The Netherlands, 1983, pp. 55–64.