47. Modeling And Simulation Of Citronelly Laurate Estirification In A Batch Reactor

RSCE-SOMCHE 2008 Edited by Daud et al.

955

MODELING AND SIMULATION OF CITRONELLYL LAURATE ESTERIFICATION IN A BATCH REACTOR

S.A. Zulkeflee and N. Aziz* School of Chemical Engineering, Engineering Campus, Universiti Sains Malaysia, 14300 Nibong Tebal, Seberang Perai Selatan, Pulau Pinang, Malaysia *Corresponding author. Phone: +604 5996457, Fax. +604 5941013, Email: [email protected] Keywords: batch reactor; Citronellyl laurate; esterification; modeling; simulation. ABSTRACT Operating batch esterification process at the required operating conditions is very important to ensure the maximum conversion obtained. However, to control this reactor is quite difficult due to its non-linear nature, time varying system and unsteady state operating condition. Model based control strategies such as internal model control (IMC) and model predictive control (MPC) have been revealed as a better control system compared to the conventional method because of its ability to satisfy strict performance required. To implement such model based control, model representing the process is a must. In this work, a detailed dynamic model is developed for Citronellyl Laurate esterification. The models include all major aspects of the description of mass and energy balance, chemical reactions and thermodynamics non-idealities. The kinetic data for the model is obtained from experimental data. The effects of catalyst loading and temperature are also be evaluated by simulating the model using MATLAB®. It is found that the model predictions are in satisfactory agreement with the experimental data with R2=0.982. INTRODUCTION Batch reactors are one of the most major equipment in the biochemical industries. It is quite flexible, it can adopt to small volume production of various products and provides the natural way to scale-up processes from laboratory to industrial manufacturing. One of difficulties in analyzing the dynamic response on biochemical processes is the fact that they are nonlinear (Hua et. al., 2004). Due to these difficulties, linear controllers or control strategies based on a local linearized model e.g. standard PID controller lead to a poor control performance

956 (Liu & Macchietto, 1995). The common way of controlling process systems with strong nonlinear character is to apply model-based controllers such as model predictive controllers (MPC) or internal model controllers (IMC) where a detailed dynamic process model is used (Seborg et. al., 2004; Dowd et. al., 2001; Toivonen et. al., 2003). In this paper, the process of Citronellyl laurate esterification is chosen as a case study (Yadav and Lathi, 2004). The set-ups of the batch reactor are given together with the experimental procedures. The analytical procedures performed to the samples are also explained. Finally, the simulated models were then verified with the experimental data. EXPERIMENTAL SETUPS A schematic diagram of the experimental apparatus is provided in Fig. 1. The reactor was a 1.5L pyrex glass cylindrical vessel. The temperature of the reactor, inlet, and outlet water in jacket were measured by thermocouples. The reactor mixture was agitated by a propeller of 8cm diameter. A variable speed driver permitted the impeller speed to be varied from 100 to 500rpm. The flow rate of the water in jacket was measured by a flowmeter. A computer with A/D and D/A converters was employed to control the temperature of the reactor. Digital Computer

A/D

T

T

T

A/D Triac Module Cooling

Heater

FIGURE 1: Schematic diagram of the esterification reactor system. In the experiments, 0.03M of Citronellol, 0.02M lauric acid are used as reactants, where 0.02M NaOH is used to stop the reaction and phenolphthalein is used for titration indicator. The temperature and stirring rate are adjusted to 37oC (optimal temperature) and 200rpm respectively. The system is operated to reach

957 the steady state condition while samples are taken every five minutes. Then, the flow rate of cooling water is increased and decreased from 40ml/min to 44ml/min and 36ml/min respectively by giving ±10% step change, and the inlet temperature of the cooling water is kept constant. For another test, the inlet temperature of the cooling water is increased and decreased from 22oC to 37oC and 17oC respectively by giving the ±10% step change, and the flow rate of cooling water is kept constant. RESULTS AND DISCUSSION First Principle Model of the Esterification Batch Reactor Obtaining a dynamical model for the considered process addresses objective consists in simulating the process behavior. In practice, dynamical models for chemical reactors are obtained by considering continuity equations. In the case of stirred tank reactors, it is commonly assumed that the reactor is perfectly mixed. The kinetic reaction sequence is (Garcia et. al., 2000); dCA dt

Al r αK A

KA Ac

1

KA K

Ac βK

αK A Ac

Al 1

       1

The energy balance around the reactor is written as; Reactor dynamics:

dT dt

V CA CpA

Jacket dynamics:

dT dt

where

Q

∆H rA V Q CA CpA CE CpE T F Cp ρ T V Cp ρ UA  T

(2) CW CpW

Q

(3)

T                                              4

Model of the reactor is solved using 4th/5th order Runge Kutta method (MATLAB ODE45). Table 1 shows the operating and calculated data. Model Validation: Fig. 2 depicts measurements and model predictions of percent of ester conversion, reaction and jacket temperature profile that yields a leastsquares model fit. The figure shows that the predicted conversion of acid to ester is initially higher than the measured conversion with R2=0.982. This result indicates that the models developed are adequate to describe the esterification of Citronellyl laurate in batch reactor.

958 TABLE 1: Operating Conditions and Calculated Parameters Parameters Units Values Parameters Units Values 18.20871 Cpw J/mol K AAc L mol/s 75.40 24.04675 V L L mol/s AAl 1.5 0.319947 Vj L mol/s 0.8 Ai L 3 -105.405 J mol/K 11.648 J/m EAc Q -66.093 J mol/K 19.84 EAl kJ ΔHrxn -249.944 J mol/K 1 Ei α 294 K 1 Tji β 420.53 J/mol K 2.857 J/s m2 K CpAc U 235.27 J/mol K 0.077 m2 CpAl A 617.79 J/mol K 8.314 J/mol K CpEs\ R 100

conversion (%)

80 60 40

♦

20

Experimental

___ Simulation 0 0

10

20 30 time (min)

40

50

FIGURE 2: Measurements and model predictions profile that yielded a minimum least-squares model fit (R2=0.982) for 3% catalyst loading and temperature 37oC Effect of Reaction Conditions a) Catalyst loading The effect of enzyme concentration was studied in the range of 1-3% (w/v), at 37oC (optimal temperature). The results are illustrated in Fig. 3. It can be seen that conversion of ester increased with increasing catalyst concentration. At the concentration of 3% enzyme the highest initial rate was observed. All further experiments were carried out with 3% (w/v) of lipase from Candida rugosa in order to reach maximum initial rate and utilization of enzyme activity.

959 100 conversion (%)

80 60 1% 2% 3%

40 20 0 0

10

20

30 time (min)

40

50

FIGURE 3: The effect of catalyst loading on the synthesis of Citronellyl laurate b) Effect of temperature The model predictions were also tested for responses to a temperature change. The impact of temperature in the ester yield and velocity is difficult to predict because it may affect reaction efficiency in opposite way. First, a temperature raise would have a positive effect on the kinetic constant as defined by the transition state theory. Conversely, the treatment at high temperatures may disrupt enzyme tertiary structure, losing its catalytic activity. From previous findings, (Aziah Serri et. al., 2006) heating was required for faster reaction and the reaction time may vary from a few minutes to several hours for a temperature range 30oC to 40oC for esterification process. The ester conversion was found to increase with increasing temperature, as shown in Figure 4.

conversion (%)

100 80 60 40

30 35 40

20 0 0

10

20

30

40

50

time (min)

FIGURE 4: The effect of reaction temperature on the synthesis of Citronellyl laurate

960 CONCLUSION A mathematical model for esterification process has been proposed by developing the differential mass and energy balance equation for batch reactor. The models have been solved and validated using 4th/5th order Runge Kutta method (MATLAB ODE45) and the simulation showed good agreement with the experimental results with R2=0.982. The amount catalyst loading and temperature were found to have profound influence on the conversion of Citronellyl laurate. ACKNOWLEDGMENT The authors wish to acknowledge the financial support by Ministry of Science, Technology and Innovation (MOSTI), Malaysia through the NSF scholarship for the first author and sciencefund project no 03-01-05-SF0090. REFERENCES Aziah Serri N., Kamaruddin A.H., and Long W.S. (2006). Studies of reaction parameters on synthesis of Citronellyl laurate ester via immobilized Candida rugosa lipase in organic media. Bioprocess and Biosystems Engineering 29, 253-260 Dowd, J. E., Kwok, K. E., and Piret, J. M. (2001). Predictive modeling and looseloop control for perfusion bioreactors), Biochemical Engineering Journal, 9, 19 Garcia, T., Coteron, A., Martinez, M. and Aracil, J. (2000) “Kinetic model for the esterification of oleic acid and cetyl alcohol using immobilized lipase as catalyst” Chemical Engineering Science, 55, 1411-1423 Hua, X., Rohani S., Jutan A. (2004) Cascade closed-loop optimization and control of batch reactor. Chemical Engineering Science 59, 5695-5708 Seborg, D. E., Edgar, T. F. and Mellichamp, D. A. (2004), On-line Controller Tuning, Process Dynamics and Control, pp 317 – 321. Toivonen H.T., Sandstrom K.V., and Nystrom R.H. (2003). Internal model control of nonlinear systems described by velocity based linearizations. Journal of Process Control 13, 215-224. Yadav G.D., Lathi P.S. (2004) Synthesis of citronellol laurate in organic media catalyzed by immobilizied lipases: kinetic studies, Journal of Molecular Catalysis B:Enzymatic, 27, 113-119.