Download as pdf or txt
Download as pdf or txt
You are on page 1of 15

Development of Kinetic

Models for Acid-Catalyzed


Methyl Acetate Formation
Reaction: Effect of Catalyst
Concentration and Water
Inhibition
B. GANESH, K. YAMUNA RANI, B. SATYAVATHI, CH. VENKATESWARLU
Chemical Engineering Sciences, Indian Institute of Chemical Technology, Hyderabad 500 007, India
Received 24 September 2010; revised 19 January 2011; accepted 22 January 2011
DOI 10.1002/kin.20555
Published online 31 March 2011 in Wiley Online Library (wileyonlinelibrary.com).

ABSTRACT: The reaction kinetics of reversible liquid-phase esterification of acetic acid with
methanol is investigated in the temperature range 26–50◦ C using sulfuric acid catalyst. The main
goal of this work is to study the effect of catalyst concentration and sensitivity to the presence of
water on the rate expression of this industrially important reaction. Experiments are conducted
in an isothermal batch reactor and a second-order kinetic model is used to correlate the
experimental data, which are found to fit well with the assumed kinetic model in terms of the
concentrations of reactants and products. Furthermore, an activity-based kinetic model is also
developed employing the UNIQUAC (universal quasi-chemical equation) model to compute
the activities. It is observed that the rate constant is influenced by the concentration of catalyst,
and the reaction rate increased with an increase in the catalyst concentration. It is also observed
that the catalyst activity is slightly inhibited by the water present in the reaction mixture. The
performance of the proposed models is compared with that of other models reported in the
literature, and it is found that the proposed models outperformed all the other models reported
in the literature.  C 2011 Wiley Periodicals, Inc. Int J Chem Kinet 43: 263–277, 2011

INTRODUCTION ticizers, and medicinal and surface-active agents. It is


produced via esterification of methanol with acetic acid
Methyl acetate is widely used in various industrial according to the following scheme:
products including fragrances, flavors, solvents, plas-
k1
CH3 COOH + CH3 OH ↔ CH3 COOCH3 + H2 O
k−1
Correspondence to: K. Yamuna Rani; e-mail: kyrani@iict.res.in,
k y rani@yahoo.com.
Contract grant sponsor: DST, New Delhi, India. The reaction is a reversible homogeneous liquid-phase
c 2011 Wiley Periodicals, Inc. process, where the limiting conversion of the reactants
264 GANESH ET AL.

is determined by equilibrium. The reaction rate is very [12] on other esterification and trans-esterification re-
slow; hence addition of a catalyst is necessary to in- actions reveal that the presence of water deactivates the
crease the rate of reaction. catalyst activity, which decreases the reaction rate. Liu
The kinetics of the esterification reaction between et al. [9] investigated the impact of water on the liquid-
acetic acid and methanol was investigated as early as phase sulfuric acid-catalyzed esterification of acetic
1934 by Rolfe and Hinsshelwood [1], who based their acid with methanol, by adding water initially to ap-
theory on molecular statistics of the reaction in al- proximate different levels of conversion. The catalytic
coholic and nonhydroxylic media by considering hy- activity of sulfuric acid has been found to be strongly
drochloric acid as the catalyst. Hilton and Smith [2] inhibited by water; the catalyst lost its activity up to
summarized that the kinetics of the catalyzed ester- 90% as the amount of water increased. Liu et al. [9]
ification reaction of normal aliphatic acid in methyl reported that the diminished catalytic activity observed
alcohol is influenced by the length of carbon chain as the concentration of water increases is likely to be
while considering different buffer solutions such as a consequence of acid strength decline due to strong
anisol and ammonium acetate. The kinetic model for solvation of protons by water molecules.
this esterification reaction using a homogeneous hy- Although a large number of researchers have stud-
drogen iodide catalyst has been reported by Ronnback ied the esterification reaction, some uncertainties still
et al. [3], where it is also stated that the rate-initiating appear with regard to the kinetic models reported. The
step in the reaction mechanism is the protonation of present investigation aims to develop a kinetic model
carboxylic acid. Popken and Gmehling [4] investi- of the esterification reaction in a homogeneous liquid
gated the reaction kinetics and chemical equilibrium of phase using sulfuric acid as the catalyst by incorporat-
homogeneously and heterogeneously catalyzed acetic ing the effect of the catalyst concentration and the effect
acid esterification with methanol and methyl acetate of water formation during the reaction on the reaction
hydrolysis and developed a rate expression for non- kinetics by carrying out experimental studies at differ-
catalytic homogeneous reaction that suggests catalysis ent temperatures and catalyst concentrations without
by molecular acetic acid instead of solvated protons adding water initially. Experimental studies are con-
and also indicates that the noncatalytic reaction is ex- ducted in an isothermal batch reactor at atmospheric
tremely slow and takes days to reach equilibrium. pressure. Kinetic models, based the concentration and
Sulfuric acid has been reported as a preferred cat- activity, are developed, and the model predictions are
alyst for this esterification reaction due to its greater compared with the experimental data. A comparison
density of acid sites per gram [5] and also because it is of different kinetic models from the literature with the
found to be more active than some heterogeneous cata- newly developed models is also presented.
lysts. Process intensification by reactive distillation has
been adopted for many industrially important chemi-
cals including methyl acetate. Investigations have been EXPERIMENTAL
carried out by Agreda et al. [6], who proposed a rate
expression for this reaction using sulfuric acid (homo- Materials
geneous) catalyst while carrying out the reaction in a Methyl acetate (purity >99.436%), methanol (purity
reactive distillation unit. The proposed rate expression >99.811%), acetic acid (purity >99.746), and sulfu-
incorporates nonlinear dependence on the catalyst con- ric acid (purity >98%) were purchased from SD Fine
centration; however, kinetic parameters have not been Chemicals Ltd. (Mumbai, India) and used as such.
reported. Engell and Fernholz [7] and Kreul et al. [8]
have modified the model proposed by Agreda et al. [6]
Apparatus
for application using a heterogeneous sulfonic acid ion
exchange resin as the catalyst. More recently, Liu et al. The experiments to study the effect of the temperature
[9] developed a rate expression for this reaction us- and catalyst concentration on the acid-catalyzed ester-
ing homogeneous sulfuric acid catalyst by considering ification of acetic acid with methanol are performed
linear dependence of the catalyst concentration on the in an isothermal batch reactor. A pictorial representa-
reaction kinetics, and Elgue et al. [10] have proposed tion of the experimental setup is shown in Fig. 1. The
a similar dependence and applied it for intensification reactor is a 1-L glass-jacketed vessel equipped with
of ester production in a continuous reactor. a digital overhead stirrer (a direct controlled model)
Another important issue that needs to be addressed with a motor rating of 83/75 W, maximum torque of
while describing the kinetics for this esterification reac- 85 N-cm, and a speed range of 200–3000 rpm. The im-
tion is the effect of water produced during the reaction. peller used is a propeller made of Teflon. The dimen-
Studies by Aafaqi et al. [11] and Kusdiana and Saka sions of the impeller are 5-cm diameter, blade length

International Journal of Chemical Kinetics DOI 10.1002/kin


KINETIC MODELS FOR ACID-CATALYZED METHYL ACETATE FORMATION REACTION 265

Figure 1 Schematic representation of the reactor assembly (R, jacketed reactor; TI, thermo well and temperature indicator;
SW, sample withdrawal provision; C, condenser; S, stirrer; TS, thermostat for coolant circulation; TIC, temperature indicator
and controller).

1.25 cm, blade width 1 cm, and 0.1-cm shaft diameter. varying catalyst concentrations of 1%, 2%, and 5% to
The agitator is located at 3.17 cm from the bottom of find the effect of the catalyst concentration on the ki-
the vessel (one third of the reactor diameter). The reac- netics. The final set of experiments was conducted at
tor assembly is provided with a four-necked glass lid different temperatures with different catalyst concen-
with provision for the liquid-phase temperature mea- trations (50◦ C: 1%, 40◦ C: 5% and 35◦ C: 4%) and is
surement in the reactor and periodic withdrawal of liq- used to validate the kinetic models developed and to
uid samples during the course of the reaction. To one compare their performance with the literature-reported
of the outlets, a condenser is connected wherein cold models. During the last experiment (35◦ C: 4%), the ini-
water is circulated to minimize losses. Temperature is tial reaction mixture consists of water also in addition
measured using a Chromel Alumel type thermocou- to methanol and acetic acid so that it can serve to com-
ple. A Polyscience refrigerating and heating circulator pare with the model incorporating inhibition effect of
thermostat model 9702 with a digital temperature con- water through initial addition of water.
troller is used to maintain the temperature (±0.01◦ C
accuracy) in the jacketed reactor. Analysis
The acetic acid concentration was determined by titra-
tion with standard solution of NaOH using phenolph-
Procedure
thalein as the indicator.
In a typical experiment, equimolar quantities of
methanol (224 g) and acetic acid (420 g) are mixed
and taken in the reactor. The desired amount of sulfu- RESULTS AND DISCUSSION
ric acid (as a percentage with respect to the weight of
acetic acid) is added to initiate the reaction. Samples Prior to the development of a new model, the mod-
are drawn after the temperature is stabilized at close els reported in the literature were explored to check
intervals of time and are analyzed for the acetic acid whether the experimental data generated in the present
concentration. The reaction is continued for 4 h. Af- study match with any of the model predictions. How-
ter the reaction is completed, the reactor contents are ever, it is found that none of the reported models is
cooled, weighed, and analyzed. found to represent the kinetics of this reaction with suf-
The first set of experiments was performed at dif- ficient accuracy. Therefore, we attempted to develop a
ferent temperatures 26, 40, and 50◦ C, with 2% catalyst new model based on the experimental data generated.
concentration to determine the kinetic parameters. The A comparison of the performance of other models is
next set of experiments was conducted at 40◦ C with reported at the end of this section.

International Journal of Chemical Kinetics DOI 10.1002/kin


266 GANESH ET AL.

Effect of Different Factors on Reaction catalyst influences the reaction rate because more H+
Kinetics ions are available when the amount of catalyst in the
reaction mixture increases.
To develop a kinetic model of the present reacting sys-
tem, a second-order kinetic approach is adopted and
the kinetic model is obtained for a fixed catalyst con- Water Sensitivity
centration in a manner similar to that employed for a
noncatalytic elementary reaction. Later, the approach It is observed from the results that the formation of
is extended to develop the rate equation for the cat- water as a by-product inhibits the catalytic activity.
alyzed reaction by considering factors including ef- Similar results have been observed by Liu et al. [9] for
fects of temperature, catalyst concentration, and water the acid-catalyzed esterification reaction with varying
sensitivity. Each of these effects is discussed in detail amounts of water added initially to the reaction mix-
below. ture.

Development of the Concentration-Based


Reaction Temperature Kinetic Model
The experimental results obtained at different tempera- From thermochemistry, the equilibrium conversion ob-
tures with the fixed catalyst concentration are depicted tained is 69% at 25◦ C for the conventional methyl ac-
in Fig. 2. As is evident from the plot, conversion of etate process as per the following reaction:
acetic acid increases almost linearly at all tempera-
tures with an increase in the temperature initially and k1
is asymptotic later. A + B↔C + D
k−1

The reaction is carried out at 1:1 molar ratio of acetic


Catalyst Concentration acid to methanol; hence a second-order elementary ki-
Experiments with varying concentrations of catalyst netic expression is assumed. The reaction rate expres-
show that the increase in the catalyst concentration en- sion (−rA ) of the homogeneous reaction is
hances the reaction rate. The experimental results ob-
tained with different catalyst concentrations at a fixed −rA = (k1 CA CB − k−1 CC CD ) (1)
temperature are shown in Fig. 3. As is evident from
the plot that an increase in the catalyst concentration where CA is the acetic acid concentration, CB is the
increased the conversion of acetic acid and the equi- methanol concentration, CC is the methyl acetate con-
librium conversion is attained earlier. The amount of centration, and CD is the water concentration. k1 is the

Figure 2 Effect of temperature on reaction kinetics at 2% catalyst concentration.

International Journal of Chemical Kinetics DOI 10.1002/kin


KINETIC MODELS FOR ACID-CATALYZED METHYL ACETATE FORMATION REACTION 267

Figure 3 Effect of catalyst concentration on reaction kinetics at 40◦ C.

rate constant of the forward reaction, and k−1 the rate k1 and k−1 to the Arrhenius equation. The activation
constant of the backward reaction. It is also assumed energies for both the forward and backward reactions
that water and ester are not present at the beginning of are obtained from Figs. 4 and 5 as 63,062 and 63,012
the reaction. The etherification of alcohol is negligible, J/mol, respectively. The activation energy for forward
and the reaction volume during the reaction is consid- and backward reactions are nearly equal, and therefore
ered to be constant. The temperature dependency of the equilibrium constant (Ke ) for this reaction is nearly
both the rate constants is expressed by the Arrhenius temperature independent (similar to the approach con-
equation. sidered earlier by Agreda [6]) and is approximately
equal to 5 for equimolar quantities of initial reactants.
k = k 0 exp(−Ea /RT ) (2) To quantify the effect of the catalyst concentration
on the reaction kinetics, the preexponential factor at
where k 0 is the preexponential factor, Ea is the ac- each of the three catalyst concentrations considered
tivation energy, R is the gas constant, and T is the in the second set of experiments for a fixed tempera-
temperature. ture is determined after eliminating the temperature-
The standard procedure for second-order kinetic dependent term by using the activation energy deter-
model development [13] for elementary reactions is mined earlier. The preexponential factor (k10 ) is plotted
followed to determine the specific reaction rates k1 as a function of the catalyst concentration as shown
and k−1 at each temperature and catalyst concentra- in Fig. 6. It is found that this dependency can be ap-
tion based on the first two sets of experiments. The proximated as a quadratic function passing through the
procedure followed to determine the specific reac- origin, and the equation is expressed as
tion rate k1 is as follows: (a) plot ln((XAe − (2XAe −
1)XA )/(XAe − XA )) versus time, where XAe is the k10 = −9.405338 × 106 (wc )2 + 1.344269 × 108 (wc )
equilibrium conversion; (b) find the slope of the straight
(3)
line obtained; and (c) calculate k1 according to
    where wc is the weight% of catalyst with respect to the
1 weight of acetic acid in the initial reaction mixture.
k1 = Slope 2 − 1 CAo .
XAe The model incorporating the temperature and cat-
alyst concentration dependence is employed to pre-
Calculation of k−1 is straightforward using equi- dict the conversion of acetic acid for both the first
librium constant and k1 . The influence of temperature and second sets of experiments mentioned above, and
on the specific reaction rate is determined by fitting the predicted conversion profiles are plotted along

International Journal of Chemical Kinetics DOI 10.1002/kin


268 GANESH ET AL.

Figure 4 Logarithm of the forward reaction rate constant (k1 ) for esterification of acetic acid with methanol as a function of
reciprocal of temperature.

Figure 5 Logarithm of the backward reaction rate constant (k−1 ) for esterification of acetic acid with methanol as a function
of reciprocal of temperature.

with the experimental data in Figs. 7 and 8, respec- mismatch is found to decrease. As observed earlier,
tively. The figures clearly indicate that there is a with an increase in the catalyst concentration, the reac-
considerable deviation between the predicted and ex- tion rate increases, indicating formation of more water
perimental values during the initial region, whereas in (being one of the products) at an early stage. There-
the later part, the match is reasonably good. The mis- fore, a decrease in mismatch with an increase in the
match can be attributed to the inhibition effect of water catalyst concentration implies that although the pres-
formed during the reaction until equilibrium is reached. ence of water inhibits the reaction, as the water con-
Furthermore, it can be observed from Fig. 7 that as the centration increases its impact on the reaction rate de-
temperature is increasing, the mismatch is decreas- creases, further implying an inverse dependence on the
ing, indicating an inverse dependence. Similarly in water concentration. Keeping in view these two ef-
Fig. 8, as the catalyst concentration is increasing, the fects, a correction term is included in the reaction rate

International Journal of Chemical Kinetics DOI 10.1002/kin


KINETIC MODELS FOR ACID-CATALYZED METHYL ACETATE FORMATION REACTION 269

Figure 6 Dependence of preexponential factor on the catalyst concentration.

Figure 7 Comparison of experimental profiles of acetic acid conversion at different temperatures with predictions of the
concentration-based kinetic model without water effect.

expression including temperature dependence in the sidered. The conversion profiles are predicted using the
form of specific reaction rate and water concentration proposed rate expression for both sets of experiments
dependence explicitly. After exploring different forms and are plotted along with the experimental data in
of dependence, the following expression is found to Fig. 9. This figure clearly indicates that the proposed
match the experimental results accurately: rate expression is able to represent the reaction kinetics
accurately.
  
k1 CC CD
−rA = √ CA CB −
(12.5k1 + CD ) Ke Development of the Activity-Based
(4) Kinetic Model
Equation (4) together with Eqs. (2) and (3) constitute When modeling the reaction for a liquid-phase sys-
the reaction kinetics for the esterification reaction con- tem where the mixture is nonideal, a correction must

International Journal of Chemical Kinetics DOI 10.1002/kin


270 GANESH ET AL.

Figure 8 Comparison of experimental profiles of acetic acid conversion at 40◦ C at different catalyst concentrations with
predictions of the concentration-based kinetic model without water effect.

be made to the concentration to indicate the departure the difference between Gibbs energy of mixing for the
from ideal case. Nonideality exists from the difference real-liquid mixture by subtracting the Gibbs energy of
in the interaction between the molecules as well as mixing of an ideal mixture at the same temperature,
size and shape difference in the molecules participat- pressure, and mole fraction.
ing in the liquid mixture. Usually phase models such as The nonideality of the mixture then needs to be
the UNIFAC (universal functional activity coefficient), incorporated in the rate equation. Generally, the rate
UNIQUAC (universal quasi-chemical equation), and equation is then written in terms of the activity of
NRTL (nonrandom two liquid) phase equilibrium mod- each component (ai ) present in the reaction mixture.
els are used to predict this nonideality factor. The ac- The activity coefficients of the components of the
tivity coefficient (γi ) is determined from the excess main reaction (CH3 COOH, CH3 OH, CH3 COOCH3 ,
Gibbs energy (GE ); this excess energy originates from and H2 O) are estimated using the UNIQUAC activity

Figure 9 Comparison of experimental profiles of acetic acid conversion at different temperatures at varying catalyst concen-
trations with predictions of the concentration-based kinetic model with water effect.

International Journal of Chemical Kinetics DOI 10.1002/kin


KINETIC MODELS FOR ACID-CATALYZED METHYL ACETATE FORMATION REACTION 271

Table I UNIQUAC r i and q i Values in terms of the activities is expressed as follows:


Component ri qi    act 
−rA,act = k1,act aA aB − aC aD Keq (6)
Acetic acid 2.2024 2.0720
Methanol 1.4311 1.4320 where ai = xi γi ; Keq
act
is the activity-based reaction
Methyl acetate 2.8042 2.5760 equilibrium constant defined as
Water 0.9200 1.4000 eq eq eq eq
xC γC xD γD
act
Keq = eq eq eq eq (7)
xA γA xB γB
Based on an equilibrium conversion of 69%, the mole
model:
fractions and activities at equilibrium point are com-
φi z θi puted and substituted in Eq. (7) to obtain the value
ln γi = ln + qi ln + li of Keqact
. The model development procedure adopted
xi 2 φi
⎡ ⎛ ⎞ for the concentration-based model is extended to the
φi   case of the activity-based model also, and the preex-
nc nc
− xj lj + qi ⎣1 − ln ⎝ θj τij ⎠ ponential factor for the forward reaction is computed
xi j =1 j =1 according to
⎤  eq eq 

nc
θj τij CA CB
− nc ⎦ (5)
0
k1,act = k10 eq eq (8)
 γA γB
j =1 k=1 θk τkj

where k10 is defined according to Eq. (3), and “eq”


where refers to the equilibrium concentrations and activity
coefficients at the corresponding compositions, which
ri xi qi xi
φi = nc , θi = nc , are determined from the equilibrium conversion. The
k=1 rk xk k=1 qk xk final rate expression incorporating inhibition effect of
q  xi water activity is given as
θi = nc i  ,
k=1 qk xk  
  k1,act aA aB − aC aD /Keq
act

τij = exp aij +


bij
+ cij T , −rA,act = (9)
T 5(0.9 + aD )
z The predicted conversion profiles of the reaction based
li = (ri − qi ) − (ri − 1) , z = 10 and nc = 4 on the activity-based kinetic model at 26, 40, and 50◦ C
2
with 2% catalyst concentration and at 40◦ C with 1%
The parameters ri and qi values of pure components and 5% catalyst concentration along with experimen-
are reported in Table I, and the interaction coefficients tal values without a correction term for water activity
are reported in Table II [4]. The basic rate expression are shown in Figs. 10 and 11, respectively, whereas the

Table II UNIQUAC Interaction Parameters


i j aij (K) bij cij (K−1 )
Acetic acid Methanol 390.26 0.97039 −3.0613 × 10−3
Methanol Acetic acid 65.245 −2.0346 3.1570 × 10−3
Acetic acid Methyl acetate −62.186 −0.43637 2.7235 × 10−4
Methyl acetate Acetic acid 81.848 1.1162 −1.3309 × 10−3
Acetic acid Water 422.38 −0.051007 −2.4019 × 10−4
Water Acetic acid −98.120 −0.29355 −7.6741 × 10−5
Methanol Methyl acetate 62.972 −0.71011 1.1670 × 10−3
Methyl acetate Methanol 326.60 0.72476 −2.354 × 10−3
Methanol Water −575.68 3.1453 −6.0713 × 10−3
Water Methanol 219.04 −2.0585 7.0149 × 10−3
Methyl acetate Water 593.70 0.010143 −2.1609 × 10−3
Water Methyl acetate −265.83 0.96295 2.0113 × 10−4

International Journal of Chemical Kinetics DOI 10.1002/kin


272 GANESH ET AL.

Figure 10 Comparison of experimental profiles of acetic acid conversion at different temperatures with predictions of the
activity-based kinetic model without water effect.

predicted conversion profiles incorporating the correc- 2% catalyst concentration are employed to validate the
tion term for both the sets are shown in Fig. 12. The concentration and activity-based models and compare
results clearly indicate that the activity-based model them with the models cited in the literature, as shown
with the correction term also represents the experi- in Figs. 13–16. The models reported in the literature
mental data very well. considered for comparison are listed in Table III along
with the parameters reported. In some cases, although
the models have been reported, the parameters have
Comparative Validation of Models not been reported and in such cases, we attempted in
The data generated during the third set of experiments the present study to refit the parameters and employ
at different temperatures with varying catalyst concen- the models for comparison. The parameters computed
trations, that is, 50◦ C: 1%, 40◦ C: 5%, and 35◦ C: 4% in the present study for such models are also reported
in addition to the data for the experiment at 40◦ C with in Table III.

Figure 11 Comparison of experimental profiles of acetic acid conversion at 40◦ C at different catalyst concentrations with
predictions of the activity-based kinetic model without water effect.

International Journal of Chemical Kinetics DOI 10.1002/kin


Table III Kinetic Models Considered in the Present Study (Units as Per Notation)
Literature Model Model Structure Reported Parametersa Estimated Parameters
Agreda et al. −rA = k0 exp(−Ea /RT ).(CA CB − CC CD /Ke ), where Ke = 5.2 k3 = 2.0 × 1012
k0 = k1 (Xcat )2 + k2 (Xcat ) for Xcat ≤ 0.004 and k4 = 1.99543 × 1012

International Journal of Chemical Kinetics


k0 = k3 − k4 exp(k5 /(Xcat − 0.004)) for Xcat > 0.004 k5 = 2.457182 × 10−7
  Ea = 63, 062.4
Liu et al. −rA = Ccat 0.38
CD
(CA CB − CC CD /Ke ) Ke = 6.22 at 60◦ C –
  1  2 
−E −E
Elgue et al. −rA = Ccat k01 exp RTa CA CB − k02 exp RTa CC CD k01 = 4.21
k02 = 0.322

DOI 10.1002/kin
Ea1 = 53,800
  1  2  Ea2 = 52,580 –
−E −E
Bonnaillie et al. −rA = Ccat k01 exp RTa CA CB − k02 exp RTa CC CD k01 = 0.0055 –
k02 = 0.011
Ea1 = 41800
   Ea2 = 41800
√ k1 CC CD
Present work −rA = (12.5k1 +CD )
CA CB − Ke Ea = 63062.4
Concentration-based model   Ke = 5
−Ea
k1 = k10 exp RT , k10 = −9.405338 × 106 (wc )2 + 1.344269 × 108 (wc )
 
act
k1,act aA aB −aC aD /Keq
Activity-based model −rA,act = 5(0.9+a , Ea = 63062.4
   eqD ) eq  eq eq eq eq
0 −Ea 0 C CB act xC γC xD γD
k1,act = k1,act exp , k1,act , Keq = eq eq eq eq
RT = k10 γ Aeq γ eq
A B xA γA xB γB
a
Units for k01 and k02 are lit mol−1 min−1 mL−1 L H2 SO4 .
KINETIC MODELS FOR ACID-CATALYZED METHYL ACETATE FORMATION REACTION
273
274 GANESH ET AL.

Figure 12 Comparison of experimental profiles of acetic acid conversion at different temperatures at varying catalyst concen-
trations with predictions of the activity-based kinetic model with water effect.

Figures 13–16 clearly illustrate that the proposed where the deviation is drastic throughout the reaction.
new models based on the concentration as well as The reason for the drastic deviation in this case is that
activity are able to predict the conversion achieved the catalyst concentration falls in another region of the
comparable with experimental data with reasonable preexponential factor dependency on the catalyst con-
accuracy throughout the reaction. The model reported centration for which parameters are not available. The
by Agreda et al. [6] with parameters as determined in model reported by Elgue et al. [10] is found to deviate
the present study is found to predict the conversion rea- considerably both with respect to the equilibrium con-
sonably well during the initial 10–15 min after which version as well as the rate at which it is approached. In
it exhibits mismatch until equilibrium is attained in all the cases except for 35◦ C: 4% catalyst concentra-
all cases except for 50◦ C: 1% catalyst concentration, tion and initial water presence, the response is found to

Figure 13 Comparison of different kinetic models at 40◦ C at 2% catalyst concentration.

International Journal of Chemical Kinetics DOI 10.1002/kin


KINETIC MODELS FOR ACID-CATALYZED METHYL ACETATE FORMATION REACTION 275

Figure 14 Comparison of different kinetic models at 40◦ C at 5% catalyst concentration.

Figure 15 Comparison of different kinetic models at 50◦ C at 1% catalyst concentration.

be very slow compared to the experimental data. Only the experimental data. However, the equilibrium con-
in the latter case is the predicted conversion profile in version in this case is found to match reasonably well
the initial region found to match the experimental data with the experimental data. The only case where a com-
to some extent. However, the latter part of the profile parison could be made with the model reported by Liu
is found to deviate considerably in this case also. The et al. [9] is the reaction at 35◦ C: 4% catalyst concentra-
model reported by Bonnaillie et al. [14] (as cited by tion with some initial quantity of water because in all
Elgue et al. [10]) is found to be slightly faster than other cases, the initial water concentration is zero and
that reported by Elgue et al. [10] but is still unable therefore the rate expression as reported by them be-
to match the speed of response of the conversion of comes undefined at time = 0, and hence the conversion

International Journal of Chemical Kinetics DOI 10.1002/kin


276 GANESH ET AL.

Figure 16 Comparison of different kinetic models at 35◦ C at 4% catalyst concentration.

cannot be determined. Figure 16 clearly illustrates that all% absolute deviation, and correlation coefficient of
the model reported by Liu et al. [9] is unable to pre- fractional conversion prediction in comparison with
dict the conversions to match the experimental data. the experimental values for six models (including two
This could be because of the linear catalyst concen- models proposed in the present study) are reported in
tration dependence and water inhibition dependence to Table IV. This table, together with Figs. 13–16, clearly
the power of −0.83 in contrast to the model proposed illustrates the accuracy of the kinetic models devel-
in the present study, which has a quadratic catalytic oped in the present study. Although the performance
concentration dependence and water inhibition depen- of both the proposed models is equally good, the use of
dence approximately to the power of −0.5. concentration-based model is recommended because
Furthermore, it can be observed that at high cata- of its simplicity.
lyst concentration (in the absence of water initially),
the reaction rate is very fast and all the models ex-
hibit a closer match with the experimental data (as CONCLUSIONS
in Fig. 14), whereas both the models proposed in this
study have exhibited predicted conversion profiles con- On the basis of comprehensive experimental investiga-
sistently matching with the experimental data in all tion of the esterification of acetic acid with methanol
the cases studied. Based on the data from seven ex- using sulfuric acid as the catalyst, a second-order ap-
periments, the overall absolute mean deviation, over- proach in terms of the concentration and activities is

Table IV Comparison of Prediction Accuracy of Different Models for Fractional Conversion


Overall Absolute Overall % Absolute Correlation
Model Average Deviation Deviation Coefficient (R 2 )
Agreda et al. 0.086 20.9 0.7671
Liu et al. 0.124 32.2 0.5015
Elgue et al. 0.104 23.9 0.8114
Bonnaillie et al. 0.076 18.3 0.8313
Present work
Concentration-based model 0.025 5.8 0.9646
Activity-based model 0.027 6.7 0.9582

International Journal of Chemical Kinetics DOI 10.1002/kin


KINETIC MODELS FOR ACID-CATALYZED METHYL ACETATE FORMATION REACTION 277

used to develop a model describing the reaction ki- Greek Symbols


netics. Concentration-based and activity-based kinetic
γi activity coefficient of component i
models are developed incorporating the effects of tem-  
θi , i , θi , i Uniquac model parameters of
perature, catalyst concentration, and water inhibition
component I
and are validated and compared with the literature-
cited models. Both the newly developed models predict
the experimental data very well when compared with Subscripts
the literature-cited models. The study shows that it is
A acetic acid
necessary to take into account the quadratic influence
B methanol
of the catalyst on the rate constants along with the in-
C methyl acetate
corporation of a correction term considering inhibitory
D water
effect of water formed.
cat catalyst

NOTATION BIBLIOGRAPHY
ai activity of ith component
1. Rolfe, C.; Hinsshelwood, C. N. Trans Faraday Soc 1934,
Ci concentration of ith component
30, 935.
Ea activation energy (J mol−1 )
2. Hilton, S.; Smith, H. A. J Am Chem Soc 1939, 61, 254.
k10 , k−1
0
preexponential factors (lit mol−1 3. Ronnback, R.; Sami, T.; Vuori, A.; Haario, H.; Lehtonen,
min−1 ) J.; Sundqvist, A.; Tirronen, E. Chem Eng Sci 1997, 52,
k1 esterification reaction rate constant 3369.
(lit mol−1 min−1 ) 4. Popken, T.; Gmehling, G. J. Ind Eng Chem Res 2000,
k−1 hydrolysis reaction rate constant(lit 39, 2601–2611.
mol−1 min−1 ) 5. Liu, Y.; Lotero, E.; Goodwin Jr. G. J. J Catal 2006, 242,
Ke equilibrium constant 278–286.
Keq act
equilibrium constant in 6. Agreda, V. H.; Partin, L. R.; Heise, W. H. Chem Eng
activity-based model Prog 1990, 86 (2), 40–46.
7. Engell, S.; Fernholz, G. Chem Eng Process 2003, 42,
R gas constant (J mol−1 K−1 )
201–210.
rA rate of reaction based on 8. Kreul, L. U.; Górak, A.; Dittrich, C.; Barton, P. I. Com-
concentration (mol lit−1 min−1 ) put Chem Eng 1998, 22, S371-S378.
rA,act rate of reaction based on activity 9. Liu, Y.; Lotero, E.; Goodwin Jr. G. J. J Mol Catal A:
(mol lit−1 min−1 ) Chem. 2006, 245, 132–140.

ri, qi , qi , li UNIQUAC model parameters 10. Elgue, S.; Devatine, A.; Prat, L.; Cognet, P.; Cabassud,
aij , bij , cij , τij Binary interaction parameters M.; Gourdon, C.; Chopard, F. Int J Chem React Eng
T temperature (K) 2009, 7, Article A24.
xi liquid-phase mole fraction of 11. Aafaqi, R.; Mohamed, A. R.; Bhatia, S. J Chem Technol
component i Biotechnol 2004, 79, 1127.
XA ,XAe conversion and equilibrium 12. Kusdiana, D.; Saka, S. Bioresource Technol 2004, 91,
289–295.
conversion
13. Levenspiel, O. Chemical Reaction Engineering, 3rd ed;
yi vapor-phase mole fraction of Wiley: New York, 1999.
component i 14. Bonnaillie, L.; Meyer, X. M.; Wilhelm, A. M. Proceed-
wc weight% of catalyst with respect to ings of IMSR2, Nuremberg: Germany, 2001 (as cited by
weight of acetic acid Elgue et al., 2009).

International Journal of Chemical Kinetics DOI 10.1002/kin

You might also like