Polish Journal of Chemical Technology, 14, 3, 59 —Pol.

65, J.
10.2478/v10026-012-0085-2
Chem. Tech., Vol. 14, No. 3, 2012 59

Simulation and sensitivity analysis for biodiesel production in a reactive

distillation column
Madhu Agarwal1, Kailash Singh2, S.P. Chaurasia3
Malaviya National Institute of Technology, Department of Chemical Engineering, Jaipur-302017, India
Corresponding author’s: e-mail: 1 [email protected], 2 [email protected], 3 [email protected]

The conventional process for biodiesel production by transesterification is still expensive due to a need
of high excess of alcohol required and its recovery by distillation. The use of a reactive distillation pro-
cess can reduce the amount of alcohol in the feed stream as it works on a simultaneous reaction and
separation. In the present study, a mathematical model has been developed for biodiesel production
from triglycerides in a reactive distillation column, which has been validated with the reported data and
CHEMCAD results. The effects of process parameters such as methanol to oil feed ratio, feed tempera-
ture, and reaction time have been investigated. The sensitivity analysis shows that yield of ester increases
with methanol to oil ratio and number of stages, however, it decreases with flow rate. The MATLAB
simulated results show that methanol to oil molar ratio of 5:1 produces 90% (by wt.) of methyl ester in
a residence time of 4.7 minutes.

Keywords: Biodiesel, reactive distillation, modeling, simulation, transesterification, sensitivity analysis.

INTRODUCTION zed reactive distillation process for the production of

isopropyl palmitate by palmitic acid esterification with
Need for renewable energy has gained much impor-
isopropanol based on the experimental and simulation
tance nowadays due to increased demand of energy to
results. The RD technique can reduce the use of excess
sustain the human development. Biodiesel is one of such
alcohol in the feeding stream, which reduces the cost
renewable energy that can substitute diesel oil produced
in downstream alcohol recovery processes; maintain a
from petroleum and is produced from vegetable oils and
high alcohol-to-oil molar ratio inside the RD reactor,
animal fats. Generally, it is formed by transesterification
which ensures the completion of the transesterification
reaction of triglycerides in the vegetable oil or animal fat
of seed oil to biodiesel11. Catalytic reactive distillation
with short chain aliphatic alcohols. It is biodegradable,
has several advantages such as shifting the chemical
nontoxic, and essentially free of sulfur and aromatics.
equilibrium by a continuous removal of the products,
Methanol and ethanol are the preferred alcohols to be
easy separation of the catalyst from the reaction mixture,
used for the transesterification reaction1. The commercia-
utilization of heat of reaction for mass transfer in case
lization production of biodiesel from vegetable oils/fats
of exothermic reaction, and absence of undesired side
still have drawbacks due to a high cost of the vegetable oil
reactions12. The synthesis of fatty acid esters by heavy
and the purification of biodiesel product. Both batch and
alcohols such as propanol, butanol, etc. using catalytic
continuous processes utilize almost 100% excess alcohol
reactive distillation has been studied by many researchers.
than the stoichiometric molar requirement (3:1) in order
The synthesis of 2-ethylhexyl dodecanoate in the reactive
to drive the transesterification reaction to a completion
distillation column (RDC) was reported by Omota et
and to produce the maximum amount of biodiesel per
al.8. Theoretically, the possibility of producing 2-ethyl
unit consumption of oil2. This excess alcohol has to be
hexanol and methanol ester with oleic acid and lauric
recovered by a separate distillation column. The use of
acid has been investigated by Matallana et al.13 using
separate distillation column for alcohol recovery increases
Aspen Plus. Thotla and Mahajani14 used RDC with side
capital as well as operating cost. Therefore, new alterna-
draw for esterification of lactic acid, aldol condensation
tives are needed to be explored for the reaction and the
of acetone, and esterification of fatty acid by methanol.
product separation to minimize the cost of the process
They reported better yield with side draw than without
without affecting the yield and quality of biodiesel for
it. Singh et al.11 developed and used a laboratory scale
which reactive distillation (RD) or catalytic distillation
sieve-tray RDC to study the transesterification of seed oil.
could be a prominent alternative technique.
They showed that methanol to oil molar ratio of 4:1 and
Application of the reactive distillation process for
a column temperature of 65°C produces biodiesel that
biodiesel production can save the cost by eliminating
met the ASTM standards. Steinigeweg and Gmehling15
the need of the reactor and the separation unit separa-
used reactive distillation process for the production of
tely. It is a hybrid process that combines the reaction
decanoic acid methyl esters by esterification of the fatty
and separation in a single equipment. The product
acid decanoic acid with methanol. They varied opera-
is removed at the same time when it is formed. This
ting conditions to determine the influence of reactant
characteristic makes it possible to overcome the equ-
ratio, reflux ratio, pressure, and distillate to feed ratio,
ilibrium thermodynamics of the reaction, reaching high
size of the reactive section, and role of a prereactor on
conversion and selectivity3–5. Reactive distillation tech-
yield of ester. He16 designed and tested a bench-scale
nique is more effective for reversible reactions such as
RD reactor for biodiesel preparation. He showed that
esterification of fatty acids6–8 and transesterification to
the bench-scale RD process can produce biodiesel at a
produce biodiesel9. Bhatia et al.10 proposed an optimi-
production rate of about 75 ml/min continuously with
Unauthentifiziert | Heruntergeladen 18.03.20 18:11 UTC
60 Pol. J. Chem. Tech., Vol. 14, No. 3, 2012

an alcohol to oil ratio of 4:1. Kiss et al.2 demonstrated VI. Vapor phase hold up is assumed to be negligible
through simulations that combining metal oxide catalysts as compared to the liquid phase hold up on each stage.
with reactive distillation technology is a feasible and
advantageous solution for biodiesel production. San- Mass balance
tander et al.17 used the response surface methodology Mass balance on each stage of column is given by
and Aspen Plus software for simulating the castor oil
biodiesel production by reactive distillation and found
ethanol to oil molar ratio 31.6% lower than that in the (1)
conventional process.
In the present study, a steady state model has been where, v i,j = y i,jV j and ,
developed for the production of biodiesel in a RD co-
lumn from triglyceride (triolein) and methanol in the j= 1…N and i=1…nc
presence of NaOH. A MATLAB program was written to
solve the resulting equations. This model was validated For the top stage (j=1), W1 = 0. For the total condenser,
by comparing the simulated results with the reported V1 = 0. The distillate rate was assumed to be zero for
data from literature16. This model was also used to in- the total reflux condition. As there is no side stream wi-
vestigate the effect of the reactant ratio, the number of thdrawal from stage 2 onwards, s2=s3=…=sN=0.
equilibrium stages, the feed tray location, and the feed Equilibrium relationships
flow rate on the yield of biodiesel.
The most commonly used vapour-liquid equilibrium
relationship is the modified Raoult’s law, which is valid
MATHEMATICAL MODEL for low to moderate pressure:
To develop a mathematical model of reactive distillation
(2)
for biodiesel production, mass and energy balance as well
as the thermodynamic equilibrium equations on each tray
are written. The stages are counted from top to bottom For partial condenser, j = 1…N and i= 1… nc. For
as shown in the block diagram of the RDC in Fig. 1.
The model is valid under the following assumptions: total condenser, j = 2…N and i= 1…nc and .
I. All the plates are equilibrium stages. Enthalpy balance
II. Vapor phase is an ideal gas mixture.
III. Reaction takes place only in the liquid phase.
IV. Constant operating pressure and total condenser (3)
with no sub-cooling.
V. Heat generated due to a chemical reaction is taken
where j=1…N. If bottoms rate is specified in place of
into consideration.
reboiler heat duty then
(4)

Summation equations

(5)

(6)

Reaction kinetics
Triglyceride (oils/fats) reacts with alcohol in the pre-
sence of the catalyst to give fatty acid alkyl esters and
glycerol. The reaction proceeds in three steps as shown
below:
Triglyceride + ROH Diglyceride + RCOOR1

Diglyceride + ROH Monoglyceride + RCOOR2

Monoglyceride + ROH Glycerol + RCOOR3

where R1, R2, and R3 are the fatty acid chains associated
with glycerol in triglyceride. K1, K2, K3 are equilibrium
constants, K1=k1/k2, K2=k3/k4, and K3=k5/k6 where k1,
k2,….,k6 are forward and backward rate constants. The
reaction rates of each reaction are as follows:
Figure 1. Block Diagram of Reactive Distillation Column
Unauthentifiziert | Heruntergeladen 18.03.20 18:11 UTC
 Pol. J. Chem. Tech., Vol. 14, No. 3, 2012 61

r1 =k1CTGCROH – k2CDGCAE (7) where, A, B, C, D, and E are the Antoine constants;

r2 =k3CDGCROH – k4CMGCAE (8) T and P are absolute temperature and vapor pressure,
respectively. The algorithm flow chart for solving the
r3 =k5CMGCROH – k6CGCAE (9)
steady state equations is shown in Fig. 2.
where CTG, CDG, CMG, CG, CAE and CROH represent the
concentration of triglyceride, diglyceride, monoglyceride,
glycerol, alkyl ester, and alcohol respectively. Forward
and backward the rate constants are given by
(10)

The kinetic parameters are as shown in Table 1. The

heat of the reaction is J/mol for each
reaction, calculated from the heat of the formation data.

Table 1. Frequency factor and activation energy [17]

Degree of Freedom Analysis

The number of equations in Mass balance = Nnc
The number of equations in Equilibrium relationship
= Nnc
The number of equations in Heat balance = N
The total number of equations = N(2nc + 1)
The number of variables
Figure 2. Algorithm flow chart for solving model equations
vi,j , i=1…nc, j=1…N : N+nc variables
Tj , j=1…N : N variables
li,j , i=1…nc, j=1:N : N+nc variables RESULTS AND DISCUSSION
The total number of variables = N(2nc + 1) The model equations were solved in MATLAB and
Therefore the Number of equations = the Number the results were also compared with those obtained in
of variables. CHEMCAD software.
In the present study, the variables which have been
specified are pressure, distillate rate, reflux ratio, Fj, Tower profile
and zi,j. A reactive distillation column with 10 numbers of trays,
the total condenser, and the reboiler was considered for
Simulation Methodology
simulation. The feed containing oil, methanol, and the
The column was simulated assuming 11 stages including catalyst was fed on the top tray. The reactive zone thus
the reboiler. The feed was introduced at the upper stage lied from the top tray to the reboiler. The column speci-
and the reaction took place on all the stages. Each reactive fications are shown in Table 2. The steady state column
stage has 300 ml holdup The kinetic model proposed composition and temperature profile were obtained by
by Bambase et al.18 was employed. The simulation was running the MATLAB simulation as shown in Fig. 3(a)
carried out by solving the system algebraic equations and (b). For the calculation of weight fraction in the
simultaneously using Newton’s method19 in MATLAB®. column, glycerol and methanol have been excluded as
Composition, temperature, and alkyl ester purity were these will be separated out in a successive unit.
obtained on all stages assuming infinite reflux ratio. The From Fig. 3a it is clear that the concentration (wt.
temperature and vapor phase composition were calcula- fraction) of methyl ester, i.e. product concentration
ted by the bubble-point calculation using the “Extended increases from the top to the bottom tray and is maxi-
Antoine” equation: mum in the reboiler. The intermediates DG and MG
remain in the low amount throughout the column due
(11)
to fast consumption in the chemical reactions. The TG
reactant shows a continuous decreasing trend confirming
Unauthentifiziert | Heruntergeladen 18.03.20 18:11 UTC
62 Pol. J. Chem. Tech., Vol. 14, No. 3, 2012

Figure 3. (a) Composition (b) Temperature profile in the column

Table 2. Column specification for biodiesel production biodiesel is shown in Fig. 5a. It can be seen from the
figure that the yield of biodiesel increases with increase
in methanol amount. Similar results have been shown
by Singh et al.11. The yield increases from 72.42 wt% to
84 wt % on increasing the molar ratio of methanol to
oil from 3 to 4, however, it increases only up to 90.3 wt
% when the ratio is increased from 4 to 5. Beyond the
ratio of 5:1, there is only a slight change in the yield.
Therefore, the ratio of 5:1 is sufficient in a reactive
distillation beyond this ratio. Simasatitkul et al.23 also
reported the molar ratio of 4.5:1 to be optimum for the
production of biodiesel in three reactive staged reactive
distillation column. Moreover, a high amount of alco-
its consumption in reaction. The maximum yield of 84 hol will need a large column to handle the unreacted
wt% of methyl ester was obtained in the reboiler. A methanol. The TG reactant and the intermediates DG
conversion of 97% was achieved in the column. and MG continuously decrease with an increase in the
The temperature profile of the column is shown in methanol amount due to enhancement in conversion.
Fig. 3b. It can be seen from the figure that the tempe- The effect of the feed flow rate on the yield of bio-
rature increases steadily from top to bottom and shoots diesel obtained is as shown in Fig. 5b. It can be seen
up in the reboiler due to the accumulation of the high from this figure that the yield of biodiesel decreases with
boiling components such as methyl ester, tri-, di-, mono- the increase in the feed flow rate due to a decrease in
-glycerides, and glycerol in the bottoms. the residence time. The decrease in the yield is more
pronounced after 0.012 mol/s of feed flow rate. The
Model validation intermediates formed remain in a low amount.
The column profile was obtained with respect to resi- The effect of the feed tray location on the yield of
dence time in the column in order to validate the model biodiesel from RDC is shown in Fig. 5c. It can be ob-
with the experimental data available in the literature. served from this figure that the yield of biodiesel slightly
The experimental data of He16 and Singh et al.11 were decreases with the change in the feed tray location from
considered for model validation as shown in Fig. 4a–d. the top towards the bottom. The amount of the interme-
A good match of the simulation and experimental diates formed remains almost constant. The maximum
data is evident from these figures. It can be seen that yield is obtained with the top tray being the feed tray as
most of the reaction takes place in the column with a the number of reactive stages is maximum in this case.
reaction time of less than 2 min. The reboiler having To find the effect of the total number of trays in RDC
3 minutes of residence time shows little reaction due on the yield of biodiesel, simulation results were obtained
to the vaporization of methanol from the liquid phase. by changing the total number of trays in the column
The concentration profile obtained for RDC for TG, as shown in Fig. 5d. It can be seen from these figures
DG, MG, and ME is similar to the batch reactor as that with increasing the number of trays in the column
reported by Noureddini and Zhu20, Freedman et al.21, from 8 to 18, the yield of biodiesel increases from 83 to
and Darnoko and Cheryan22. 85.5% only. This may be due to the fact that most of
the reaction is carried out in the upper trays because of
Sensitivity analysis of the column a large fraction of methanol lying on these trays.
The effects of various parameters on the yield of
biodiesel were investigated as shown in Fig. 5a–d. The
effect of molar ratio of methanol to oil on the yield of
Unauthentifiziert | Heruntergeladen 18.03.20 18:11 UTC
 Pol. J. Chem. Tech., Vol. 14, No. 3, 2012 63

Figure 4. Experimental and simulated profile in the reactive distillation column (a) triglyceride (b) methyl ester (c) mono-glyceride
(d) di-glyceride

Figure 5. Effect of (a) methanol to oil molar ratio (b) feed flow rate (c) feed stage location (d) total number of stages on yield
Unauthentifiziert | Heruntergeladen 18.03.20 18:11 UTC
64 Pol. J. Chem. Tech., Vol. 14, No. 3, 2012

CONCLUSIONS Abbreviations
TG Triglyceride
A model developed to predict the composition and the
DG Diglyceride
yield of biodiesel production from pure TG (triolein) in
MG Monoglyceride
a RDC using methanol and NaOH catalyst was simula-
ME Methyl ester
ted using MATLAB to obtain the column profile. The
G Glycerol
tower profile obtained in MATLAB was also compared
with the CHEMCAD result. It was observed that both
results are similar. This model was validated with the LITERATURE CITED
reported data. The RDC can be effectively used for the 1. Fangrui, M. & Milford, A.H. (1999). Biodiesel production:
transesterification of vegetable oils as 97% conversion a review. Bioresource Technology. 70 (1), 1–15. doi:10.1016/
of triglyceride was predicted in a residence time of 4.7 S0960-8524(99)00025-5.
minutes. Most of the reaction takes place in the column 2. Kiss, A.A., Dimian, A.C. & Rothenberg, G. (2008). Bio-
and small conversion in reboiler. Sensitivity analysis of diesel by catalytic reactive distillation powered by metal oxides.
the column revealed that the yield of ester increases Energy Fuels. 22 (1), 598–604.
3. Giessler, S., Danilov, R.Y., Pisarenko, R.Y., Serafimov,
with methanol to oil ratio, the number of trays and
L.A., Hasebe, S. & Hashimoto, I. (2001). Systematic struc-
decreases with the flow rate. Methanol to oil ratio of ture generation for reactive distillation processes. Computers
5:1 gives 90.3% yield, however beyond this ratio, yield is & Chemical Engineering. 25 (1), 49–60. doi:10.1016/S0098-
not much improved. High methanol to oil ratio (>5:1) 1354(00)00632-3.
can lead to methanol flooding in the column as it is the 4. He, B.B., Singh, A.P. & Thompson, J.C. (2006). A novel
most volatile component in the system. continuous-flow reactor using reactive distillation for biodiesel
production. Transactions of the ASABE. 49 (1), 107–112.
Nomenclature 5. Pai, R.A., Doherty, M.F. & Malone, M.F. (2002). Design
Ci concentration of ith component (mol/litre) of reactive extraction systems for bioproduct recovery. AIChE
D distillate flow rate (mol/s) Journal. 48 (3), 514–526. DOI: 10.1002/aic.690480310.
6. Chin, S.Y., Mohamed, A.R., Ahmad, A.L. & Bhatia, S.
Gj Volume of jth stage (liter)
(2006). Esterification of palmitic acid with iso-propanol in a
heat of reaction (J/mol) catalytic distillation column: Modeling and simulation studies.
HLj liquid enthalpy on jth stage International Journal of Chemical Reactor Engineering. 4 (4), 32.
HVj vapour enthalpy on jth stage 7. Kiss, A.A., Omota, F., Dimian, A.C. & Rothenberg, G.
HFj Feed enthalpy on jth stage (2006). The heterogeneous advantage: biodiesel by catalytic
iref,k reference component number in kth reaction reactive distillation. Topics in Catalysis. 40 (1), 141–150. DOI:
Ki,j Equilibrium constant= γi,j Pi,jsat/P 10.1007/s11244-006-0116-4.
8. Omota, F., Dimian, A.C. & Bliek, A. (2003). Fatty acid
li,j liquid flow rate of ith component from jth stage
esterification by reactive distillation. Chemical Engineering
Lj Total liquid flow rate from jth plate
Science. 58, 3159–3174. doi:10.1016/S0009-2509(03)00165-9.
(mol/s)=∑li,j, i=1: nc
9. He, B.B., Singh, A.P. & Thompson, J.C. (2005). Experi-
nc no. of components = 6 mental optimization of a continuous-flow reactive distillation
nr no. of reactions involved reactor for biodiesel production. Transactions of the ASAE.
N no. of stages including reboiler 48 (6), 2237–2243.
P pressure (N/m2) 10. Bhatia, S., Ahmad, A.L., Mohamed, A.R. & Chin, S.Y.
Pi,jsat vapor pressure of ith component on jth stage (2006). Production of isopropyl palmitate in a catalytic distil-
Qj Heat removed from stage j lation column: Experimental studies. Chemical Engineering
R i,k,j rate of reaction of ith component on jth plate for Science. 61 (22), 7436–7447. doi:10.1016/j.ces.2006.08.039.
kth reaction 11. Singh, A.P., Thompson, J.C. & He, B.B. In A continuous-
sj Uj/Lj -flow reactive distillation reactor for biodiesel preparation from
Sj Wj/Vj seed oils, ASAE/CSAE Annual International Meeting, Ottawa,
Ontario, Canada, 1–4 August, 2004; Ottawa, Ontario, Canada,
T temperature (K)
2004. Paper number 046071.
Uj Flow rate of liquid side stream withdrawn from 12. Taylor, R. & Krishna, R. (2000). Modelling reactive
jth stage
distillation. Chemical Engineering Science. 55 (22), 5183–5229.
vi,j Vapor flow rate of ith component from jth stage doi:10.1016/S0009-2509(00)00120-2.
Vj Total vapor flow rate leaving jth tray (mol/s) 13. Mattalana, L.G., Gutierrez, L.F. & Cardona C.A. In
Fj Feed flow rate on jth tray (mol/s) Biodiesel Production by reactive distillation, R Distillation –
Wj vapour flow rate withdrawn from jth tray (mol/s) enpromer2005.eq.ufrj.br., 2005; 2005; pp 1–9.
zi,j mole fraction of ith component in liquid phase 14. Thotla, S. & Mahajani, S. (2009). Reactive distillation
on jth plate with side draw. Chemical Engineering and Processing: Process
γi,j activity coefficient of ith component on jth tray Intensification. 48 (4), 927–937. DOI: 10.1016/j.cep.2008.12.007.
15. Steinigeweg, S. & Gmehling, J. (2003). Esterification of
a fatty acid by reactive distillation. Ind. Eng. Chem. Res. 42
Subscripts
(15), 3612–3619.
f forward
16. He, B. (2006.). A novel continuous-flow reactor using
i component number (1 to nc) reactive distillation for biodiesel production. A report, Natio-
j stage number (0 to N+1) nal Institute of Advanced Transport Technology, University
r reverse of Idaho.
17. Santandera, C.M.G., Sandra Marcela Gómez Rueda,
Nívea De, Lima da Silvaa, Aline Carvalho da Costa, Rubens

Unauthentifiziert | Heruntergeladen 18.03.20 18:11 UTC

 Pol. J. Chem. Tech., Vol. 14, No. 3, 2012 65

Maciel Filhoa, Regina, M. & Maciela., W. (2010). Simulation

of the reactive distillation process for biodiesel production.
20th European Symposium on Computer Aided Process Engi-
neering – ESCAPE20.
18. Bambase, M.E., Nakamura, N., Tanaka, J. & Matsumura,
M. (2007). Kinetics of hydroxide-catalyzed methanolysis of
crude sunflower oil for the production of fuel-grade methyl
esters. Journal of Chemical Technology & Biotechnology. 82 (3),
273–280. DOI: 10.1002/jctb.1666.
19. Gupta, S.K., Numerical methods for engineers. New Age
International: New Delhi, 2006; p 422.
20. Noureddini, H. & Zhu, D. (1997). Kinetics of transeste-
rification of soybean oil. Journal of the American Oil Chemists’
Society. 74 (11), 1457–1463. DOI: 10.1007/s11746-997-0254-2.
21. Freedman, B., Butterfield, R.O. & Pyrde, E. (1986).
Transesterificationkinetics of soybean oil. J. Am. Oil Chem.
Soc. 63, 1375–1380. DOI: 10.1007/BF02679606.
22. Darnoko, D. & Cheryan, M. (2000). Kinetics of palm
oil transesterification in a batch reactor. Journal of the Ame-
rican Oil Chemists’ Society. 77 (12), 1263–1267. DOI: 10.1007/
s11746-000-0198-y.
23. Simasatitkul, L., Siricharnsakunchai, P., Patcharavora-
chot, Y., Assabumrungrat, S. & Arpornwichanop, A. (2011).
Reactive distillation for biodiesel production from soybean
oil. Korean Journal of Chemical Engineering. 28 (3), 649–655.
DOI: 10.1007/s11814-010-0440-z.

Unauthentifiziert | Heruntergeladen 18.03.20 18:11 UTC