Kinetic modeling of homogeneous methanol synthesis catalyzed by base-promoted nickel complexes

Article abstract of DOI:10.1139/v01-032

Nickel complexes promoted by alkoxide bases can affect facile conversion of synthesis gas into methanol under thermodynamically favorable temperatures. At 100°C, syngas conversion of 99% with turnover frequency of 66 h^-1 is reported. The reaction is truly catalytic in Ni as well as in base. A kinetic study of the Ni(CO)₄-KOMe catalyst system in 1,2-bis(2-methoxy ethoxy)ethane (triglyme)-MeOH solvent mixture is presented. The kinetic expression includes terms of zero-order in H₂ and first-order in CO with syngas of H₂-CO stoichiometry of ～2:1 and less than first-order in Ni (0.03-0.10 M), ～third-order in base (2.0-4.75 M), and an exponential dependence on methanol concentration (13.0-23.7 M at 2.0 M base; 15.8-22.6 M at 3.0 M base). The activation energy of 42.2 kJ mol^-1 was estimated from the Arrhenius plot of the data between 374 and 393 K. Gas phase IR spectrum at the end of each run showed an intense signature peak at 2060 cm^-1 for Ni(CO)₄. Process uncertainties to commercialization of this versatile homogenous catalyst system for CO hydrogenation to methanol are discussed.

Full text of DOI:10.1139/v01-032

848
Kinetic modeling of homogeneous methanol
synthesis catalyzed by base-promoted nickel
complexes
Devinder Mahajan, Viwat Krisdhasima, and Robert D. Sproull
Abstract: Nickel complexes promoted by alkoxide bases can affect facile conversion of synthesis gas into methanol un-
der thermodynamically favorable temperatures. At 100°C, syngas conversion of 99% with turnover frequency of 66 h–1
is reported. The reaction is truly catalytic in Ni as well as in base. A kinetic study of the Ni(CO)4–KOMe catalyst sys-
tem in 1,2-bis(2-methoxy ethoxy)ethane (triglyme)–MeOH solvent mixture is presented. The kinetic expression includes
terms of zero-order in H2 and first-order in CO with syngas of H2–CO stoichiometry of -2:1 and less than first-order
in Ni (0.03–0.10 M), -third-order in base (2.0–4.75 M), and an exponential dependence on methanol concentration
(13.0–23.7 M at 2.0 M base; 15.8–22.6 M at 3.0 M base). The activation energy of 42.2 kJ mol–1 was estimated from
the Arrhenius plot of the data between 374 and 393 K. Gas phase IR spectrum at the end of each run showed an in-
tense signature peak at 2060 cm–1 for Ni(CO)4. Process uncertainties to commercialization of this versatile homogenous
catalyst system for CO hydrogenation to methanol are discussed.
Key words: homogeneous catalysis, syngas conversion, methanol synthesis, alkoxides, carbonyls of nickel.
Résumé : Des complexes de nickel activés par des bases de type alcoolate peuvent effectuer facilement la conversion
du gaz de synthèse en méthanol, dans des conditions de température thermodynamiquement favorables. À 100°C, on
observe une conversion à 99% du gaz de synthèse avec une fréquence de renouvellement de 66 h–1. La réaction est
vraiment catalytique tant en Ni qu’en base. On présente les résultats d’une étude cinétique du système Ni(CO)4–KOMe
dans un mélange de solvant 1,2-bis(2-méthoxyéthoxy)éthane (triglyme)–MeOH. L’expression cinétique comprend des
termes d’ordre zéro en H2 et du premier ordre en CO dans le gaz de synthèse de stoechiométrie H2–CO d’environ 2:1,
d’ordre inférieur à l’unité en Ni (0,03 à 0,10 M), du troisième ordre en base (2,0 à 4,75 M) et une dépendance expo-
nentielle sur la concentration de méthanol (13,0 à 23,7 M à 2,0 M en base; 15,8 à 22,6 M à 3,0 M en base). On a
évalué une énergie d’activation de 42,2 kJ mol–1 à l’aide d’une courbe d’Arrhenius des données entre 374 et 393 K.
Les spectres infrarouges en phase gazeuse des produits à la fin de chaque réaction montrent tous une bande signature
intense à 2060 cm–1 pour le Ni(CO)4. On discute des incertitudes qui entourent les possibilités de commercialisation de
ce système versatile de catalyseur homogène pour l’hydrogénation du CO en méthanol.
Mots clés : catalyse homogène, conversion du gaz de synthèse, synthèse du méthanol, alcoolates, carbonyls de nickel.
[Traduit par la Rédaction] Mahajan et al. 853
Introduction
lytic synthesis gas conversions (also known as C1
chemistry).
A vast amount of natural gas is stranded in remote loca-
tions and economical transport of this clean fuel is highly
desirable. Methanol and Fischer–Tropsch (F–T) hydrocar-
bons are considered shippable liquids that can be produced
at remote locations from natural gas-derived synthesis gas
(eqs. [1]–[3]) and as such, there is a renewed interest in cata-
[1]
[2]
[3]
CH4 + O2 (H2O) → CO + nH2 (n = 2, 3)
CO + 2H2 → CH3OH
xCO + yH2 → (-CH2-)n + pCO2 + qH2O
Methanol has particularly attractive attributes as a fuel, an
additive in conventional fuels, and a versatile building block
in the synthesis of chemicals (1, 2). Methanol is presently
manufactured from synthesis gas using supported Cu–ZnO
heterogeneous catalyst at 250–300°C and between 5–
10 MPa operating pressure (3). Reaction [2] is highly exo-
thermic (∆H298 = –128.6 kJ mol–1) that poses thermody-
namic limitations and together with poor process heat
control allows <20% per pass gas conversion during the re-
action (4). A homogeneous catalyst that affords high reac-
tion rates at low temperatures would offer superior reaction
heat management, higher yields per pass, and may afford at-
tractive economics. Efforts have been made to synthesize
Received August 8, 2000. Published on the NRC Research
Press Web site at http://canjchem.nrc.ca on June 30, 2001.
Dedicated to Professor Brian R. James on the occasion of his
65th birthday.
D. Mahajan,1 V. Krisdhasima, and R.D. Sproull. Building
815, Advanced Fuels Group, Energy Sciences and
Technology Department, Brookhaven National Laboratory,
Upton, NY 11973–5000, U.S.A.
1Corresponding author (telephone: (631) 344-4985; fax: (631)
344-7905; e-mail: dmahajan@bnl.gov).
Can. J. Chem. 79: 848–853 (2001)
DOI: 10.1139/cjc-79-5/6-848
© 2001 NRC Canada

Mahajan et al.
849
oxygenates from synthesis gas using homogeneous systems.
The use of transition-metal carbonyls especially those of Ru
(5) and Rh (6) for synthesis of methanol, ethanol, other
higher alcohols, and ethylene glycol is well documented for
synthesis gas conversion but with these systems the overall
reaction parameters are not economically attractive.
We previously reported (7) our approach to methanol syn-
thesis from synthesis gas that utilized base-activated nickel
tetracarbonyl as a catalyst for operation in homogeneous
mode. In a follow-up paper (8), we reported on the kinetics
of methanol synthesis catalyzed by the Ni(CO)4–KOMe sys-
tem in 1,4-dioxane–MeOH solvent mixture. Our further
work on this catalyst suggested that 1,4-dioxane–MeOH was
not particularly a good solvent mixture for this reaction be-
cause the solution homogeneity could not be maintained un-
der a wide range of Ni and base concentrations. In this
paper, we report on our work with other solvents for the re-
action including a kinetic study in the MeOH–1,2-bis(2-
methoxy ethoxy)ethane (triglyme) solvent mixture.
gassed with N2 and the weighed KOCH3 in the vial was
mixed with the degassed solvent mixture to obtain a light
yellow homogeneous solution. The solution was poured into
the pressure vessel and the vessel was quickly sealed and
purged several times with N2. A known amount of nickel
tetracarbonyl was then injected into the base solution
through the catalyst injection port and the vessel was imme-
diately pressurized with syngas to about 750 psig. After
10 min, a gas sample was taken to establish the initial gas
composition in the vessel. The automatic temperature con-
troller and the stirrer were set at 120°C and 1600 rpm re-
spectively. The stirring speed of 1600 rpm ensured that the
reaction was operating out of the mass transfer regime. The
vessel contents were then stirred for 15 min at room temper-
ature to monitor initial gas absorption by the solvent and the
base. At this time, methanol synthesis was initiated by heat-
ing the solution. The temperature and pressure were re-
corded as a function of time until the reaction became very
slow. The reaction was quenched by replacing the heater
with an ice water bath. Once the vessel cooled to room tem-
perature, the final pressure and temperature were recorded.
Gas and liquid samples were taken for analysis and the final
liquid volume was measured. For this study, methanol syn-
thesis was carried out under conditions with respect to
nickel, base, and methanol concentrations that allowed col-
lection of several data points during the runs. The data
points were collected by analyzing the vapor-phase compo-
sition every 2 to 3 min during the run.
Gas samples were analyzed for H2 and CO using Gow-
Mac 550 gas chromatographs in the TCD mode: a molecular
sieve column (8′ × 1/8′′) with He carrier gas for CO and a
molecular sieve column (6′ × 1/8′′) with N2 carrier gas for H2
with data reproducibility within ±2%. Liquid phase samples
(2µl) were analyzed on chromosorb-102 (8′ × 1/8′′) column
fitted in a PerkinElmer Model 8500 GC in the TCD mode
with the data reproducibility within ±5%. All gas
chromatographs were connected to HP Model 3390A inte-
grators for peak quantification. A qualitative detection of
Ni(CO)4 in the vapor phase was achieved as follows. A 1-cm
gas cell with KBr windows was evacuated and filled with the
gas sample from the pressure vessel. The IR spectrum of the
sample was recorded between 4000 and 400 cm–1 on a
PerkinElmer Model 1330 IR spectrophotometer. The broad
peaks at 2170, 2120 cm–1 (free CO), 2060 cm–1 (Ni(CO)4),
and 1760 cm–1 (CH3O2CH) were monitored.
Experimental
Materials
Potassium methoxide (>95% pure) and nickel
tetracarbonyl (>99.9% pure) were purchased from Alfa and
used as received. Methanol, triglyme, and other solvents
were obtained from Aldrich. For the kinetic study, methanol
was distilled in 400mL batches with magnesium turnings
and iodine under dry argon and triglyme was stored over
molecular sieves prior to use. Pure CO, H2, and syngas (H2
and CO) mixtures were purchased in aluminum cylinders (to
avoid any iron pentacarbonyl impurity) from Scott Specialty
Gases.
Apparatus
A stainless steel MagneDrive II Packless Zipperclave
0.5 L reactor manufactured by Autoclave Engineers, Inc.
(AE) was modified for this study. The reactor was equipped
with a Dispersamax six-blade impeller mounted on a shaft
driven by an AE variable speed motor. Ports were provided
for loading nickel tetracarbonyl, reactant gases, and sam-
pling vapor and liquid phases. The temperature of the reac-
tor was maintained with a Parr 4831 temperature controller,
which controlled an external heating mantle and an internal
air-cooling system. The pressure was monitored with a Setra
204 pressure transducer (accuracy to ±1 psig). The trans-
ducer and temperature outputs were recorded using an
Omega dual pen continuous chart recorder. The unit ratings
were as follows: stirrer = 0–2000 rpm, T = 25–350°C, P =
0.1–14 MPa.
For the kinetic study, the Redlich–Kwong equation of
state was used to calculate the number of moles of H2 and
CO in the gas phase from the experimental data of syngas
consumption vs. time. These data were further processed to
yield rate constant values.
Run procedure and analysis
Results and discussion
All runs containing nickel tetracarbonyl were performed
under a well-ventilated fume hood (Caution: Ni(CO)4 is vol-
atile and extremely toxic!).2 In a typical kinetic run, a speci-
fied amount of KOCH3 was weighed inside a Labconco
glove box under N2 and was brought out in a closed vial.
The methanol–triglyme solvent mixture (100mL) was de-
Catalyst performance
It has been shown in an earlier study (8) that the
Ni(CO)4–KOMe system is a versatile catalyst to affect meth-
anol synthesis from syngas. Table 1 lists representative runs
2 Pure Ni is commercially manufactured via the Mond process (9) that involves the intermediacy of Ni(CO)4. Ni(CO)4 is volatile (bp =
43.2°C at 25°C; vapor pressure @ 20°C = 315 mmHg (1 mmHg = 133.322 Pa)) and extremely toxic. The OSHA limit for Ni(CO)4 is set at
1 ppb in air for an 8-h daily exposure.
© 2001 NRC Canada

850
Can. J. Chem. Vol. 79, 2001
Table 1. Methanol synthesis catalyzed by the Ni(CO)4–alkoxide base systems.a
Run
Solventb (mL)
[Ni] (mmol)
Alkoxide (KOR)R (mmol)
Gas conversion (%)c
TOF (h–1)d
1
2
3
4
5
6
7
THF
THF
THF
THF
5
10
1
5
5
2-Methyl-2-butyl (40)
2-Methyl-2-butyl (100)
2-Methyl-2-butyl (40)
2-Methyl-2-propyl (40)
Methyl (40)
93
99
92
81
97
96
94
24
66
6
36
41
27
14
THF
1,4-Dioxanee
50% 1,4-Dioxane
50% MeOH
MeOH
5
5
2-Methyl-2-butyl (40)
Methyl (400)
8
9
5
5
Methyl (600)
Methyl (200)
92
95
1.4
2.2
Triglymee
aAll runs were carried out in 100mL solvent at 100°C under 2.1 MPa syngas (66% H2:33%CO) per charge in a pressure vessel
stirred at 1600 rpm. Typically, the reactor was charged twice in a given run.
bTHF: tetrahydrofuran; triglyme: 1,2-bis(2-methoxyethoxy)ethane.
cBased on the final pressure at the end of the first charge.
dTurnover frequency is defined as g-mol MeOH g-mol catalyst (Ni + base)–1 h–1.
that demonstrate the performance of the alkoxide base-
activated Ni catalyst for CO hydrogenation. A typical run
proceeded as follows. For example in Run 1, on adding
tetracarbonyl nickel to a light yellow solution of potassium
2-methyl-2-butoxide in tetrahydrofuran (THF) under argon,
a deep red solution resulted. The red solution yielded an ac-
tive catalyst when heated at 100°C under 2.1 MPa syngas
(66% H2:34% CO). During heating, syngas consumption
commenced and within minutes the reactor pressure dropped
below 0.2 MPa corresponding to an average gas consump-
tion rate of 1 mmol s–1. Gas chromatographic analysis of the
final liquid and gas phases yielded >98% selectivity to
methanol with the rest comprising of methyl formate and
dimethyl ether as by-products. The total gas consumption
during the run was 93%. Such analysis resulted in the data
presented in Table 1 for all the runs. Gas phase IR spectrum
taken at room temperature at the end of each run showed a
sharp and intense signature peak at 2060 cm–1 for Ni(CO)4.
These data suggest the presence of free Ni(CO)4 during
these runs. The IR spectrum of the final solutions from these
runs was somewhat complex that are being analyzed. It is
known (10, 11) that basic solutions of Ni(CO)4 yield
multinuclear Ni species that are in facile equilibrium with
Ni(CO)4:
the slowest rate measured in neat methanol as solvent even
at 6.0 M KOMe concentration. Thus in the present system, a
suitable balance between methanol and base concentrations
is needed to attain high reaction rates while maintaining so-
lution homogeneity during the methanol synthesis reaction.
The screening runs in Table 1 also show that as a general
trend, the reaction rate was dependent on Ni (0.01–0.10 M)
and base (0.4–6.0 M) concentrations. Furthermore, the rate
was affected by the nature of the alkyl group associated with
potassium alkoxide. For example, in THF, at fixed Ni and
alkoxide concentrations, the rate dependence followed the
order: KOMe > KO-2-methyl-2-propyl > KO-2-methyl-2-
butyl (Runs 5, 4, 1). KOMe was the preferred alkoxide be-
cause irrespective of the starting alkoxide, once MeOH is
formed, methoxide is invariably produced via rxn. [6] and
exists in equilibrium with the added alkoxide (13):
[6]
KOR + MeOH
KOMe + ROH
It is thus clear that the use of KOMe as the alkoxide base
yields the highest rates and avoids the complication of equi-
librium [6].
The catalyst was evaluated in several different solvents
(Runs 5–9). The reaction rate was solvent dependent and fol-
lowed the order: 1,4-dioxane - THF > triglyme > MeOH.
As stated earlier, methanol is necessary as a solvent to
solubilize the base via rxns. [5] and [6] but the presence of a
cosolvent appears to decrease methoxide association (rxn.
[5]) and yield higher rates. Initially, THF was the solvent of
choice based on the rate data in Table 1 but its similar boil-
ing point (65°C) limited its use in the downstream separa-
tion from product methanol (bp: 65°C). 1, 4-dioxane (bp:
102°C) was then selected and a kinetic study in this solvent
was reported (8). But the utility of the MeOH–1, 4-dioxane
solvent mixture was limited due to its base solubility con-
straints. Triglyme boils at a higher temperature (216°C) and
the MeOH–triglyme solvent mixture appeared to satisfy sev-
eral reaction and process criteria. This led us to undertake a
kinetic study on the Ni(CO)4–KOMe catalyst system in
MeOH–triglyme solvent mixture.
Base
[4]
Ni(CO)4
[Nix(CO)y ]n−
CO
Metal alkoxides as precursors for electronic and ceramic
materials have been extensively studied (12). Alkali metal
alkoxides are known to form hydrogen bonded
homoconjugate anion. Specific hydrogen bonding is thought
to contribute greatly to the solvation of alkoxides in protic
solvents (13):
[5]
RO–M(s) + HOR
[ROLHLOR] + M+(s)
Methoxides are strongly associated due to reduced steric
effect and increased localization of charge compared to
higher alkoxides. In the present catalyst system containing
KOMe–MeOH, methanol is necessary to dissolve KOMe but
the presence of methanol is likely to have a pronounced ef-
fect on the reaction rate. The data of Runs 6, 7, and 8 (Ta-
ble 1) confirm the negative effect of methanol on rate with
Kinetic study
Methanol synthesis requires syngas with H2:CO ratio of
2:1 (eq. [2]). Therefore, all runs in this study were con-
© 2001 NRC Canada

Mahajan et al.
851
Fig. 3. Effect of nickel concentration on the observed rate con-
stant. Solvent: triglyme–MeOH = 100 mL. [KOMe] = 4.0 M,
[MeOH] = 21.7 M, T = 110°C.
Fig. 1. Typical log plots of CO consumption as a function of
time during a typical kinetic run for the Ni(CO)4–KOMe catalyst
system. Solvent: 10% triglyme:90% MeOH = 100 mL. [Ni] =
0.05 M, [KOMe] = 3.0 M, T = 110°C. Stirring speed =
1600 rpm. Syngas: 66% H2:34% CO = 5.0 MPa at 25°C.
-2
-2.2
-2.4
-2.6
-2.8
-3
6
4
2
0
-4
-3.6
-3.2
-2.8
-2.4
-2
ln [Ni]
0
20
40
60
80
under the following reaction conditions: [Ni] = 0.05 M,
[KOMe] = 3.0 M, T = 110°C; solvent: 100 mL triglyme–
MeOH (10%:90%). The log of CO partial pressure vs. time
plot yielded a straight line suggesting a first-order disappear-
ance of CO:
Time (min)
Fig. 2. Normalized CO concentration as a function of time dur-
ing a typical kinetic run with the Ni(CO)4–KOMe catalyst sys-
tem. Solvent: 50% triglyme:50% MeOH = 100 mL. [Ni] =
0.05 M, [KOMe] = 3.0 M, T = 110°C. Stirring speed =
1600 rpm.
[7]
–dCco/dt = kCco
where Cco is the liquid-phase CO concentration and k is a
first-order rate constant. Such analysis was carried out for all
the runs. The data points could be used to construct a nor-
malized CO concentration curves as a function of time dur-
ing a kinetic run. Such a curve depicting CO consumption
vs. time data is shown in Fig. 2 for a run that was carried out
at: [Ni] = 0.05 M, [KOMe] = 3.0 M, T = 110°C; solvent:
100 mL triglyme–MeOH (50%:50%).
1.0
0.8
0.6
0.4
0.2
0.0
-0.38t
C(t) / C(0) = e
Effect of Ni concentration
The Ni concentration was varied between 0.03 and 0.10 M
at fixed values of 4.0 M KOMe, 21.7 M MeOH, and a con-
stant temperature of 110°C. In this Ni concentration range,
the expression simplifies to:
0
2
4
6
8
10
Time (min)
[8]
k = kNiCNαi
ducted with
a constant syngas composition of 66%
H2:34% CO. The data in Table 1 show that under certain
conditions, syngas consumption can be fast. For the kinetic
runs, the catalyst composition was adjusted to allow enough
reaction time for collection of several data points during
each run. As shown below, four kinetic parameters were var-
ied to collect the kinetic data on the Ni(CO)4–KOMe system
in MeOH–triglyme solvent mixture.
where α is the exponent of Ni concentration CNi and kNi is a
constant. A plot of the experimental data in Fig. 3 shows
that log k varied linearly with log CNi. The slope of this line
yielded α of 0.534.
Effect of base concentration
The concentration of KOMe was varied between 2.0 and
4.75 M at fixed values of 0.05 M Ni(CO)4 at a constant tem-
perature of 110°C. Note that in this set, the methanol con-
centration was adjusted at 23.7, 24.0, 22.6, 21.7, 21.1 M,
respectively, at KOMe concentrations of 2.0, 2.5, 3.0, 4.0,
4.75 M. Since the solution volume changes on dissolution of
KOMe in the solvent mixture, this correction was necessary
to keep the total solution volume constant. In this base con-
centration range, the rate constant can be modeled as a
power function:
Rate of CO disappearance
The rate constants were calculated from the CO consump-
tion data as follows. First, the Redlich–Kwong equation was
used to calculate the individual H2 and CO values from the
experimentally measured values of the total syngas con-
sumption. The time dependent CO values were corrected for
solvent vapor pressure and liquid phase CO concentration
factors. The variations of CO mol fraction as a function of
time data were then used to determine the rate of disappear-
ance of CO. Figure 1 shows one such plot for a run that was
charged with 5.0 MPa syngas pressure at room temperature
[9]
k = kbC βb
© 2001 NRC Canada

852
Can. J. Chem. Vol. 79, 2001
Fig. 4. Effect of base concentration on the observed rate con-
stant. Solvent: triglyme–MeOH = 100 mL. [Ni] = 0.05 M,
[MeOH] = 21.7 M, T = 110°C.
Fig. 6. An Arrhenius plot of the observed rate constant data at
101, 110, and 120°C. Solvent: triglyme–MeOH = 100 mL. [Ni]
= 0.05 M, [KOMe] = 4.0 M, [MeOH] = 21.7 M.
-1
-2
-3
-4
-5
-6
-2
-2.2
-2.4
-2.6
-2.8
-3
2.5
2.54
2.58
2.62
2.66
2.7
1000/T (K-1
)
0.5
0.8
1.1
1.4
ln [Base]
1.7
2
Table 2. Parameters in the kinetic model for metha-
nol synthesis.
Fig. 5. Effect of methanol concentration on the observed rate
constant. Solvent: triglyme–MeOH = 100 mL. [Ni] = 0.05 M,
[KOMe] = 2.0, 3.0 M, T = 110°C.
Parameter
Experimental
Calculateda
ko
ki
α
β
Eact
—
0.248
0.534
–2.89
42.23
1.267 × 106
0.256
0.548
–2.81
42.22
-1
-2
-3
aFor these values, regression constant (R2) was 0.98.
-4
2M Cb;0.05M CT
Predicted
quency factor (A) and the energy of activation (Eact) of the
reaction.
3M Cb; 0.05M Cc
Predicted
-5
[11] k = Aexp(Eact/RT)
12
14
16
18
20
22
24
Eact was calculated by linear regression to be 42.2 kJ mol–1.
This value appears low compared to what is expected for a
typical chemical reaction (83–126 kJ mol–1). This value is
also smaller compared to our previous data (97.6 kJ mol–1)
for the Ni––OMe catalyst system in 1,4-dioxane–MeOH sol-
vent mixture (8). Interestingly, the measured Eact value is in
the range of that measured for carbonylation of methanol to
methyl formate (67.7 kJ mol–1 between 323 and 343 K) cata-
lyzed by sodium methoxide (13). We are further analyzing
the data to explain this discrepancy in the Eact values.
[MeOH] (M)
where β is the exponent of base concentration Cb and kb is a
constant. A log–log plot of the experimental data in Fig. 4
shows that log k varied linearly with log Cb. The slope of
this line yielded β of 2.89.
Effect of methanol concentration
We discussed earlier that a complex dependence of metha-
nol and methoxide concentrations was noted during the
methanol synthesis reaction. We, therefore, collected two
sets of data, at 2.0 M and 3.0 M KOMe, in which MeOH
concentration was varied. The plots in Fig. 5 show that there
is a linear relation between log k and the liquid-phase meth-
anol concentration (Cm). The slopes of the lines for both the
both 2.0 M and 3.0 M data sets were the same (–0.248).
From these data, the relationship between methanol concen-
tration and the rate constant can be expressed as:
Kinetic model
The experimental data in Figs. 3–6 were used to estimate
the respective parameters, α, β, ki, and Eact. Next, all of the
parameters were determined simultaneously by utilizing all
the data points at one time. The proposed model is based on
a combination of eqs. [8]–[11] that yields the following ex-
pression:
[12] k = koexp(–Eact/RT)CNαiC bβexp (–kiCm)
[10] k = kmexp(–kiCm)
A better estimate of the parameters in this model can be
calculated by using multiple linear regression of eq. [11]
(14). The parameter values from experimental data and the
calculated values are given in Table 2. The final expression
is then modified to:
where km and ki are methanol concentration dependent con-
stants.
Effect of temperature
Rate constants were measured at 374, 383, and 393 K at
constant 0.05 M Ni, 4.0 M base, 21.7 M MeOH in triglyme.
An Arrhenius plot of the data yielded a straight line as
shown in Fig. 6. This plot was used to calculate the fre-
[13] k = (1.27 × 106)exp(–10 084/RT)CN0.i548C 2b.81
exp(–0.256Cm)
© 2001 NRC Canada

Mahajan et al.
853
Fig. 7. A comparison of the observed and the predicted rate con-
stants.
that result in methoxycarbonyl and hydroxycarbonyl adduct
formation respectively via rxn. [16]:
0.20
0.16
0.12
0.08
0.04
0.00
[16] Mx(CO)y + RO– → Mx(CO)y – 1(CO2R)–
(R = H or Me)
Rate constants have been reported for the reaction of
–
Ru3(CO)12, Ru(CO)5, Fe3(CO)12, and Fe(CO)5 with OMe
and –OH ions at 25°C (16). The Ru carbonyls do yield meth-
anol even in the absence of methoxide though at higher tem-
peratures and pressures (5). A preliminary mechanistic study
has appeared in literature for methanol synthesis catalyzed
by the Ni(CO)4–NaOMe catalyst system but the rate deter-
mining step (RDS) was not identified (17). The rate expres-
sion in eq. [13] is empirical and, as such, only limited
information can be derived from the presented kinetic data.
We are presently establishing a detailed mechanism of the
KOMe–Ni catalyst system that will allow us to correlate the
observed kinetic dependence and identify the rate-
controlling step in methanol synthesis. A detailed study of
our catalyst system will be the subject of a forthcoming pub-
lication3 in which we also seek to establish the role of
multinuclear Ni species (10, 11) in methanol synthesis.
0.00
0.04
0.08
0.12
0.16
0.20
kpred
In Fig. 7, the observed rate constants (kobs) are compared
to the line based on the predicted rate constants (kpred) from
eq. [13]. The data are in excellent agreement.
The kinetic expression in eq. [13] includes terms for tem-
perature, nickel, alkoxide base, and methanol concentrations.
A less than first-order dependence on nickel may indicate
the presence of more than one Ni species that are in equilib-
rium in solution. An approximate third-order dependence of
rate on alkoxide base and a negative exponential methanol
term likely result from ion-pairing due to hydrogen bonding
in methanol, a protic solvent. Thus, the RO––ROH interac-
tion results in slow reaction kinetics that is consistent with
the presented data. Further work with different alkoxide
bases in combination with various alkali metals in methanol
will likely establish the effect of these parameters on ther-
modynamics and kinetics of the system.
Acknowledgments
The authors wish to thank the Electric Power Research In-
stitute and the U.S. Department of Energy under Contract
no. DE-AC02–98CH10886 for partial support of this work.
References
1. W. Keim (Editor). Catalysis in C1 chemistry. D. Reidel Pub-
lishing Company, Dordrecht, The Netherlands. 1983; W. Keim.
J. Organomet. Chem. 372, 15 (1989).
2. I. Wender. Catal. Rev. Sci. Eng. 26, 3 (1985).
3. S. Lee. Methanol synthesis technology. CRC Press, Inc., Boca
Raton, Florida. 1990.
4. S Strelzoff. Chem. Eng. Prog. Symp. Ser. 66, 98 (1970).
5. J.S. Bradley. J. Am. Chem. Soc. 101, 7419 (1979).
6. R.L. Pruett and W.E. Walker. U.S. Patent 3 957 857 to Union
Carbide, 1976.
7. D. Mahajan, R.S. Sapienza, W.A.R. Slegier, and T.E. O’Hare.
U.S. Patent 4 935 395, 1990.
8. J.E. Wegrzyn, D. Mahajan, and M. Gurevich. Catal. Today, 50,
97 (1999).
9. L. Mond. U.S. Patent 455230, 1891.
10. G. Longoni, P. Chini, and A. Cavalieri. Inorg. Chem. 15, 3025
(1976), and refs. therein.
11. J.K. Beattie, T.W. Hambley, A.F. Masters, and J.T. Meyer.
Inorg. Chim. Acta, 213, 49 (1993).
Future direction
The potential use of methanol as a clean fuel including its
use as a peak-shaving fuel in the integrated gasification com-
bined cycle (IGCC) power plants is justifying further re-
search in developing a simple and cost-effective methanol
synthesis process. A catalyst that yields high reaction rates,
high CO conversion, and high methanol selectivity would be
applicable to such systems. In this paper, we have presented
data to demonstrate that it is possible to develop such a cata-
lyst system that operates in the homogeneous mode. The
base activated Ni(CO)4 system, in its present formulation, is
not yet commercially attractive for two reasons. These are:
(i) the presence of toxic Ni(CO)4; and (ii) the sensitivity of
the catalyst to CO2 and H2O. The latter is due to interaction
of these molecules with the alkoxide base that yield alkyl
carbonate and hydrolysis products respectively:
12. D.C. Bradley. Chem. Rev. 89, 1317 (1989).
13. S.P. Tonner, D.L. Trimm, M.S. Wainwright, and N.W. Cant. J.
Mol. Catal. 18, 215 (1983), and refs. therein.
14. J. Neter, W. Wasserman, and M. Kutner. Applied linear regres-
sion models. Richard D. Irwin, Inc., U.S.A. 1983.
15. C. Faurholt. Z. Phys. Chem. Stoechiom. Verwandtschaftsl.
126, 72 (1927).
16. D.C. Gross and P.C. Ford. Inorg. Chem. 21, 1704 (1982).
17. M. Marchionna, L. Basini, A. Arogno, M. Lami, and F.
Ancilotti. J. Mol. Catal. 75, 147 (1992).
[14] –OMe + CO2 → –OCO2Me
[15] –OMe + H2O → –OH + MeOH
Reaction [14] lies far to the right and the equilibrium con-
stant (Keq) for the NaOMe–CO2 reaction is known to be -2
× 10–8 (15). Both these reactions compete with presumably
the initial methanol synthesis step that requires the availabil-
ity of alkoxide base. This reaction has been studied for other
metal carbonyls both with methoxide and hydroxide bases
3 D. Mahajan. Manuscript in preparation.
© 2001 NRC Canada