See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/244748606

Simulation of styrene polymerization reactors: Kinetic and thermodynamic

modeling

Article in Brazilian Journal of Chemical Engineering · April 2008

DOI: 10.1590/S0104-66322008000200012

CITATIONS READS

14 1,331

3 authors, including:

Argimiro R. Secchi
Federal University of Rio de Janeiro
333 PUBLICATIONS 1,903 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

NATURAL GAS PLANTS ADVANCED CONTROL AND REAL TIME OPTIMIZATION View project

Chromatography and Crystallization View project

All content following this page was uploaded by Argimiro R. Secchi on 29 October 2015.

The user has requested enhancement of the downloaded file.

 Brazilian Journal
of Chemical ISSN 0104-6632
Engineering Printed in Brazil
www.abeq.org.br/bjche

Vol. 25, No. 02, pp. 337 - 349, April - June, 2008

SIMULATION OF STYRENE
POLYMERIZATION REACTORS: KINETIC
AND THERMODYNAMIC MODELING
A. S. Almeida2, K. Wada1 and A. R. Secchi1*
1
GIMSCOP, Departamento de Engenharia Química, Universidade Federal do Rio Grande do Sul,
Phone: +(55) (51) 3308-3528, Fax: +(55) (51) 3308-3277, Rua Sarmento Leite 288/24,
CEP: 90050-170, Porto Alegre – RS, Brasil.
E-mail: [email protected]
2
Engenharia de Processos, INNOVA S/A, Phone: +(55) (51) 3457-5835,
Fax: +(55) (51) 3457-5829, Rod. Tabaí-Canoas, BR 336, Km 419,
Pólo Petroquímico, CEP: 95853-000, Triunfo - RS, Brasil.
E-mail: [email protected]

(Received: June 23, 2006 ; Accepted: April 25, 2007)

Abstract - A mathematical model for the free radical polymerization of styrene is developed to predict the
steady-state and dynamic behavior of a continuous process. Special emphasis is given for the kinetic and
thermodynamic models, where the most sensitive parameters were estimated using data from an industrial
plant. The thermodynamic model is based on a cubic equation of state and a mixing rule applied to the low-
pressure vapor-liquid equilibrium of polymeric solutions, suitable for modeling the auto-refrigerated
polymerization reactors, which use the vaporization rate to remove the reaction heat from the exothermic
reactions. The simulation results show the high predictive capability of the proposed model when compared
with plant data for conversion, average molecular weights, polydispersity, melt flow index, and thermal
properties for different polymer grades.
Keywords: Polystyrene; Free-radical polymerization; Polymeric solution; Auto-refrigerated reactors.

INTRODUCTION polymerization, propagation, and termination

reactions.
Many researchers have been involved in studying When discussing the kinetics of polymerization
the phenomenon of thermo auto-initiation of styrene. of styrene, it is impossible not to mention the kinetic
Flory (1937) was the first who suggested that this model of the thermal initiation proposed by Hui and
polymerization is initiated by the combination of two Hamielec (1972). In their article, the authors made
molecules of styrene to produce a diradical. an experimental study of the thermal polymerization
However, further statistical analysis indicated that of styrene between the temperatures of 100 and
these diradicals react very fast in order to start the 200oC and a range of conversion from 0 to 100%.
polymerization. Mayo (1961) proposed the They established the kinetic constants for each of the
mechanism for this reaction. The first stage of the reaction’s steps, considering that all the constants are
mechanism is the dimerization reaction of Diels independent of the size of the polymer chain, but
Alder; in the next stage, the obtained diradical reacts vary with the conversion. The constants found by the
with a third molecule of styrene, producing two authors to the gel effect are used until today in the
types of radicals responsible for the initiation of the majority of studies of styrene polymerization.

*To whom correspondence should be addressed

338 A. S. Almeida, K. Wada and A. R. Secchi

From the 80’s, with an increased industrial different process operating conditions. The reactors
interest in the application of multi-functional were modeled as reactive separation systems with
initiators for reaction of free radicals, new studies two equilibrium stages. The reacting system
were published about the kinetic behavior of styrene considered is constituted, mainly, by the monomer
polymerization. The importance of these studies is (styrene), the solvent (ethyl-benzene) and the
based on the nature of the free-radical polymer (polystyrene). The heat of the exothermic
polymerization, where high conversions and high reaction is removed, in the form of heat of
molecular weights cannot be, in general, obtained vaporization, by the vapor phase stream leaving the
simultaneously. The practical advantages of these reactor. For the phase equilibrium calculation, the
multi-functional initiators are, in principle, the Peng-Robinson EOS and the ZM mixture rule with
reduction of the polymerization time and the increase the extension of Louli and Tassios (2000) for the
of the molecular weight control, being unnecessary energy and co-volume parameters for the polymer
major modifications in the reactor. Villalobos et al. were used. The kinetic and binary interaction
(1991) developed models for the kinetics of the parameters were estimated with the use of available
polymerization using three types of bi-functional and limited data in the industry. In the next three
initiators. From the comparison between bi- sections, the kinetic, the thermodynamic, and the
functional initiators and one mono-functional, it was process models are presented, respectively.
concluded that distribution of molecular weights is
controlled by the rate of polymerization with bi-
functional initiators. The different kinetics of KINETIC MODEL
decomposition of the two peroxide groups of bi-
functional initiators was considered by the authors The polymerization of styrene can be initiated
by taking into account the change of the activation thermally and by bi-functional peroxide, following
energy for homolysis of the second peroxide group, the conventional kinetics of free-radical initiation,
which takes place only after the homolysis of the propagation, termination, and chain transfer. The
first group. thermal initiation becomes substantial at
In many industrial processes of polymers, data for polymerization temperatures above 100°C (Chen,
the vapor-liquid equilibrium (VLE) are crucial. On 2000). The thermal initiation mechanism proposed
this basis, many equations of state (EOS) and by Hui and Hamielec (1972) is represented by a
coefficients of activities models have been proposed third-order kinetics with respect to the styrene
in the last two decades. Loulli and Tassius (2000) monomer. The decomposition of the bi-functional
used the cubic equation of state of Peng-Robinson; initiator takes place by the bond scission of one
where the parameters a (energy) and b (co-volume) peroxide group and not by the simultaneous bond
for the polymers were obtained by experiments. The scission of the two peroxide groups (Villalobos et al.,
EOS was applied to the low-pressure vapor-liquid 1991). It is also considered that the chain termination
equilibrium for various polymeric solutions through occurs exclusively by combination. Termination by
the use of an interaction parameter with three mixing disproportion is negligible in the polymerization of
rules: van der Walls, Zhong and Masuoka (ZM), and styrene (Chen, 2000). With the above assumptions,
Huron-Vidal of first order (MHV1). These rules the mechanism of the polymerization reactions is
were used together with the cubic equation of state given in Table 1, where I is the bi-functional
and the model of Flory-Huggins. The parameters a peroxide; Ag is the chain transfer agent; Ag* is the
and b were determined by fitting the model to PVT radical transfer agent; AH is 1-phenyltetralin; Pn and
data of pure polymer and considering that a/MW and Dn are the alive and dead polymers of size n without
b/MV are independent of MW (average molecular a non-dissociated peroxide group; Pn and D  are the
n
weight). They applied the VLE model in a large alive and dead polymers of size n with a non-
variety of solvent-polymer systems, and the best
result was obtained with the ZM mixing rule. dissociated peroxide group; D̂n is the temporarily
In the present work, kinetic and thermodynamic dead (or dormant) polymer of size n with two non-
models of styrene polymerization in solution, using dissociated peroxide groups; R ˜ is the primary
data from an industrial plant, were developed. Based radical without a non-dissociated peroxide group
on these models, a dynamic mathematical model for (radical 1); R ˜ is the primary radical with a non-
a series of two auto-refrigerated reactors was built to dissociated peroxide group (radical 2); and S is the
predict the quality of the polymer product for solvent concentration.

Brazilian Journal of Chemical Engineering

 Simulation of Styrene Polymerization Reactors 339

Table 1: Kinetics of styrene polymerization

Decomposition: Transfer to AH:

2k k
d
I  o R ˜  R ˜ (bi-functional peroxide) Pn  AH trAH
o D n  P1 (n t 1)
 k d1 k trAH
P  AH   P
D 
n o R ˜ P n
(n t 2) n oD n 1
(n t 1)
2k d1
D̂ n  o R ˜  Pn (n t 2)
Thermal Initiation: Transfer to solvent:
k k trS
dm
3M  o 2P1 Pn  S  o D n  P1 (n t 1)
k trS
P  S  
n o D n  P1 (n t 1)
Radical Initiation: Transfer to transfer agent:
k k trX
R ˜  M  ioP
1 Pn  Ag  o D n  Ag (n t 1)
k k trX
R ˜  M  ioP
1 P n  Ag    Ag
oD n
(n t 1)
k ix
Ag  M o P1
Propagation: Termination by combination:
k k
p
Pn  M o Pn 1 (n t 1) tc
Pn  Pm  o Dnm (n,m t 1)
kp k tc
P n  M o P n 1 (n t 1) Pn  P m   
oD nm
(n,m t 1)
k tc ˆ
P n  P m  oD nm
(n,m t 1)
Transfer to monomer: Oligomers: (thermal initiation)
k k CT
Pn  M  trm
o D n  P1 (n t 1) AH  M  o cyclic trimers (CT)
k k
P n  M trm
oD P
n 1
(n t 1) DCB
M  M  o dimers (DCB)

The gel effect must be taken into account for free- sub-products was studied by Pryor and Coco (1970),
radical polymerization reactions with high showing that the degree of polymerization at low
conversions, where the termination reaction becomes auto-initiated conversions is much bigger than at
controlled by diffusion, and the effective constant for high conversions. In order to explain this kind of
termination decreases considerably with the increase anomalous behavior, it was suggested that the
of the monomer conversion. The correlation transfer to monomer controls the chain termination
suggested by Hui and Hamielec (1972), as function in low-conversion processes. However, when the
of the conversion (X) and the reaction temperature concentration of AH increases, the transfer to AH
(T), was used in the gel effect for the polymerization becomes more important for chain termination. On
reaction of styrene, represented in Equation 1. this basis, a large parcel of the chain transfer
constant to styrene is, in reality, the result of the
k tc k tco exp ª-2 A1X  A 2 X 2  A3 X3 º (1) chain transfer to AH:
¬ ¼
§ k trm · § k trm · k trAH AH
where ¨ ¸ ¨ ¸  (5)
© k P ¹apparent k
© P ¹ real kP M

A1 2.57 - 5.05 x 10-3 T (K) (2)

However, because the kinetic constants of the
intermediary reactions of thermal initiation are
A2 9.56 - 1.76 x 10-2 T (K) (3) unknown, the empirical form of Hui and Hamielec
(1972) is an alternative to the use of Equation 5.
A3 - 3.03  7.85 x 10-3 T (K) (4) During an experimental study, the authors
considered that all the constants are independent of
the size of the polymer chain but can vary with the
ktco is the kinetic constant without the gel effect.
conversion and they suggest Equations 6 to 8:
The chain transfer to sub-products of the
thermally initiated reaction must be considered in
order to have an effective constant of transfer to § k trm · § k trm ·
¨ ¸ ¨ ¸  B1X (6)
monomer. The importance of the chain transfer to © k P ¹apparent © k P ¹ real

Brazilian Journal of Chemical Engineering Vol. 25, No. 02, pp. 337 - 349, April - June, 2008
340 A. S. Almeida, K. Wada and A. R. Secchi

§ 473.12  T · 1 dQ1V
B1  1.013x103 log ¨ ¸ T < 473K (7) 2kdmM3  2fkd I  fkd1 P o  2Pˆ o  P1 
© 202.5 ¹ V dt
c2 Qo  Q o  ktrXAg Qo  Q o  kP MQo  (17)
0.01E1 § § 1 1 ··
B1 ; E1 0.9755exp ¨ 12180 ¨  ¸¸
1  2E1 © © T 473 ¹ ¹ c2Q1  ktrXAgQ1  ktcQoQ1  ktcQ1Q o
(8)
for T t 473K 1 dQ2V
2kdmM3  2fkd I  fkd1 P o  2Pˆ o  P 2 
V dt
In this experimental study, Hui and Hamielec
(1972) also considered that the propagation reactions c2 Qo  Q o  ktrXAg Qo  Q o  kP M 2Q1  Qo  (18)
are influenced by the conversion and that Equation 6
could be used to obtain the variation of the kinetic c2Q2  ktrXAgQ2  ktcQ2Qo  ktcQ2Q o
constant (kp) with the conversion. However,
according to Chen (2000), it is not necessary to take
into account the effect of the conversion on the 1 dQ o V
2fk d I  k trX AgQ o  c 2 Q o 
propagation reaction for solution polymerization. V dt
(19)
Undertaking a mass balance of the species,
without considering their inflow and outflow at this k tc Q o2  k tc Q o Q o  2fk d1Pˆ o
stage, and by applying the method of moments while
using the assumption of QSSA (quasi-steady-state 1 dQ 1 V
approximation) for the radicals, the final kinetic 2fk d I  k P MQ o  c 2 Q 1 
V dt
model is obtained: (20)
k trX AgQ 1  k tc Q o Q 1  k tc Q1Q o  2fk d1Pˆ 1
1 d I.V
2fk d I (9)
V dt 1 dQ 2 V
2fk d I  k P M(2Q 1  Q o )  c 2 Q 2 
1 d M.V V dt
4fkd I  fkd1 P o  2Pˆ o  2kdm M3  (21)
V dt k trX AgQ 2  k tc Q 2 Q o  k tc Q 2 Q o  2fk d1Pˆ 2
(10)
c1 Qo  Q o  ktrX Ag Qo  Q o  kCT M  kDCBM2
1 dPo V 1
c 2 Q o  k trX AgQ o  k tc Q o2 (22)
1 d Ag.V V dt 2
 k trX Ag Q o  Q o (11)
V dt
1 dP1 V
c 2 Q1  k trX AgQ1  k tc Q1Q o (23)
1 d S.V V dt
 k trSS Q o  Q o (12)
V dt
1 dP 2 V
1 d AH.V c 2 Q 2  k trX AgQ 2 
 k trAH AH Q o  Q o  k CT M (13) V dt
V dt (24)
k tc Q o Q 2  Q12
1 d DCB.V
k DCB M 2 (14)
V dt
1 dP o V
c 2 Q o  k trX AgQ o 
1 d CT.V V dt
k CT M (15) (25)
V dt k tc Q o Q o  k d1P o

1 dQo V
2k dm M3  2fk d I  fk d1 2P o  2Pˆ o  1 dP 1 V
V dt c 2 Q 1  k trX AgQ 1 
(16) V dt
(26)
k trX MAgQ o  c2 Q o  k tc Qo2  k tc Qo Q o
k tc Q o Q 1  Q o Q1  k d1P 1

Brazilian Journal of Chemical Engineering

 Simulation of Styrene Polymerization Reactors 341

1 dP 2 V Qk and Pk are the kth-order moments of alive and dead

c 2 Q 2  k trX AgQ 2  polymer without a non-dissociated peroxide group;
V dt
(27) Q k and P k are the kth-order moments of alive and
k tc Q o Q 2  2Q1Q 1  Q o Q 2  k d1P 2 dead polymer with a non-dissociated peroxide group;
P̂ k is the kth-order moment of temporarily dead (or
1 dPˆ o V 1 dormant) polymer with two non-dissociated peroxide
k tc Q o2  2k d1Pˆ o (28)
V dt 2 groups; V is the volume of the liquid phase; and f is
the initiator efficiency.
1 dPˆ 1 V Different sets of kinetic parameters were found in
k tc Q 1Q o  2k d1Pˆ 1 (29) the literature, showing a high variability in their
V dt
values as illustrated in Table 2 for the propagation
1 dPˆ 2 V rate constants. In order to select an appropriate set of
k tc Q o Q 2  Q 12  2k d1Pˆ 2 (30) kinetic parameters, a sensitivity analysis was carried
V dt
out among several combinations of parameters. The
more suitable set of published kinetic constants
where
employed to fit the available plant data is given in
c1 k p M  k trm M ; c2 k trm M  k trs S  k trAH AH ; Table 3.

Table 2: Propagation rate constants, where kp = kp0 exp(-Eap/T), with T in Kelvin

kp0 Eap Unit Reference

9.806x104 3557.021 kg m3 s1 Villalobos et al. (1991)
1.009x105 3557.000 kg m3 s1 Hui and Hamielec (1972)
1.009x105 3557.000 kg m3 s1 Cabral et al. (2003)
9.179x103 2600.000 kg m3 s1 Estenoz et al. (1996)
1.009x105 3553.095 kg m3 s1 Kim and Choi (1991)
9.044x104 3537.997 kg m3 s1 Freitas Filho et al. (1994)
1.009x105 3537.997 kg m3 s1 Özkan et al. (1998)
1.018x105 3557.000 kg m3 s1 Russo and Bequete (1998)
1.009x105 3553.095 kg m3 s1 Kim and Choi (1998)
1.008x104 3554.000 kg m3 s1 Cavin et al. (1999)
6.346x103 2600.000 kg m3 s1 Peng (1990)
6.913x104 3524.176 kg m3 s1 Moore (1989)

Table 3: Kinetic constants, where k = k0 exp(-Ea/T), with T in Kelvin

k k0 Ea unit Reference
kd 3.470u1015 17128 s-1 Akzo (supplier)
kd1 1.817u1019 21361 s-1 Villalobos et al. (1991)
kp 6.346u103 2600 kg m-3s-1 Peng (1990)
kdm 2.019u101 13810 kg3m-6s-1 Hui e Hamielec (1972)
ktrm 3.110u103 5631 kg m-3s-1 Kim e Choi (1989)
ktc 1.205u107 844 kg m-3s-1 Hui e Hamielec (1972)
ktrS 1.968u104 6302 kg m-3s-1 Moore (1998)
ktrX 4.263u108 4304 kg m-3s-1 Supplier
kCT 1.266u105 11078 s-1 Chen (2000)
kDCB 7.600u104 14087 kg m-3 s-1 Chen (2000)

Brazilian Journal of Chemical Engineering Vol. 25, No. 02, pp. 337 - 349, April - June, 2008
342 A. S. Almeida, K. Wada and A. R. Secchi

THERMODYNAMIC MODEL reactors are also known as chemical reactors with

evaporative cooling. The vapor boiled from the
The Peng-Robinson equation of state (Equation liquid is condensed in a heat exchanger and,
31), with the values of the parameters a (energy) and generally, the condensate is returned to the reactor.
b (co-volume) determined experimentally by Loulli The condenser operates at lower pressure than the
and Tassius (2000) for the polymer, was used to reactor since there must be a pressure gradient to
predict the low-pressure vapor-liquid thermodynamic drive the vapor from the reactor to the condenser.
equilibrium. Several polymerization reactions are carried out in
industry in auto-refrigerated reactors (Beckman,
RT a 1973; Henderson III and Cornejo, 1989; Toledo et
P  2 (31) al., 2005). In the studied process, the condensate is
V  b V  2Vb  b 2
returned to the feed stream, before a heat exchanger
The parameters a and b were obtained by fitting that controls the feed temperature of the first reactor,
the model to PVT data of pure polymer and taking see Figure 1.
into account that a/MW and b/MW are independent Each reactor was modeled as a reactive
of MW (average molecular weight). The equations distillation system with two equilibrium stages, as
that express the mixing rule ZM (Loulli and Tassios, illustrated in Figure 2. The liquid phase in the bottom
2000) are as follow: stage, where the polymerization reactions occur, was
considered at its bubble point. No liquid holdup was
considered in the top stage, acting as an
DD
am Q (32) instantaneous flash drum of the feed stream. The
1  DD reactor temperature is controlled by manipulating the
vaporization rate in the top stage. The heat of the
Q exothermic reaction is removed, in the form of heat
bm (33)
1  DD of vaporization, by the vapor phase stream leaving
the reactor. The liquid stream from the top to the
with bottom stage provides the cooling of the reaction
mixture, since it is at lower temperature due to the
§ a · flashing. The following assumptions were also
Q ¦¦ x
i j
i xj ¨b 
©
¸
RT ¹ij
(34)
considered in the model:
ƒ The reacting medium is in perfect mixing.
ƒ The heats of initiation and termination reactions
ai are negligible when compared to the heat of the
DD ¦x
i
i
bi RT
(35)
propagation reaction.
ƒ The reactions are irreversible.
ƒ The cooler inlet temperature and the condenser
§ a · 1 ª§ ai · § a j ·º
pressure are specified.
¨b  ¸ «¨ bi  ¸  ¨¨ b j  ¸» 1  kij (36)
© RT ¹ij 2 «¬© RT ¹ © RT ¸¹»¼ A CSTR with evaporation is not a full reactor and
it usually operates under slight vacuum. During
these equations consider the interaction between normal operation, the reactor runs isothermically by
polystyrene, styrene, and ethyl-benzene components controlling the evaporation rate and having the feed
through the binary parameters kij. stream at lower temperature as disturbance. The
reactor’s jacket is the provider of the heating and
cooling during startups and shutdowns using thermal
PROCESS MODEL oil, and it is kept at the reaction temperature during
normal operation, not contributing to heat removal,
The mathematical model proposed for two auto- but acting as thermal isolation. Luyben (1999)
refrigerated reactors in series of an industrial plant is presented a detailed study of temperature control of
illustrated in Figure 1. Auto-refrigerated chemical auto-refrigerated CSTRs.

Brazilian Journal of Chemical Engineering

 Simulation of Styrene Polymerization Reactors 343

WO
VACCUM
CHAIN TRANSFER
ADITIVE
CHAIN TRANSFER
INITIATOR

CONDENSATOR

CONDENSATORS

1º REACTOR 2º REACTOR

STYRENE

MATRIZ
PELLETIZING

Figure 1: Simplified diagram of the styrene polymerization process.

Tv, FvI

Fo, TF Tv Pv q
I
V

FvII F LI

Te Pe q
VII

FLII

Figure 2: Scheme to model the auto-refrigerated CSTR.

Brazilian Journal of Chemical Engineering Vol. 25, No. 02, pp. 337 - 349, April - June, 2008
344 A. S. Almeida, K. Wada and A. R. Secchi

For the auto-refrigerated CSTR, the energy with E a 2300 exp 2.4X .
balance equation for Stage I (adiabatic) in the quasi- The thermal, rheological, and mechanical
stationary state is given by: properties of solid polystyrene may depend upon the
residual styrene (xM), dimer, trimer (xCT), oligomers
Fo h o  FVII H II  FLI h I  FVI H I 0 (37) (methanol soluble), mineral oil (xoil), or polymer
molecular weights in the final product. The melt
In relation to Stage II (non-adiabatic), the flow index (MFI), Vicat temperature, heat distortion
following energy balance was written: temperature (HDT), and tensile strength (TS) of the
polystyrene are estimated from Wooden’s
dT II correlations (Chen, 2000).
UIIL VLII c IIpL FLI (h I  h II ) 
dt
(38) §M ·
ln MFI 23.29  3.27 ln ¨ w ¸ 
FVII (H II h ) II
('H r )rp VLII q © 1000 ¹
(42)
I I II
In these equations, h , H , h , H , are the liquid II § M ·
0.877 ln ¨ n ¸  0.233 100x oil  0.562x M
and vapor enthalpies in Stages I and II, respectively, © 1000 ¹
ho is the feed enthalpy; VLII and TII are the liquid
volume and temperature in Stage II; UIIL and c IIpL are Vicat (195  5.71 100x oil 
(43)
the liquid specific mass and heat in Stage II; q is the 16.3x M  2.8x CT ) /1.8 o
C
heat exchanged with the thermal oil during startups
and shutdowns, and 'Hr is the heat of the
propagation reaction with rate rp. The overall mass HDT (173.2  4.74 100x oil 
balance and component balances for each stage were (44)
o
written in the same way, using the kinetic rates given 15.8x M  1.77x CT ) /1.8 C
in Section 2.
21895000
Polymer Properties TS 1213650  
M n 1000
The number and mass average molecular weights (45)
can be calculated by using the radical and polymer 43210 100x oil , MPa
moments:

P1  P 1  Pˆ 1  Q1  Q 1 RESULTS
Mn MM (39)
P0  P 0  Pˆ 0  Q 0  Q 0
The models were implemented in MATLAB,
version 5.3. In the first step, the binary interaction
P 2  P 2  Pˆ 2  Q 2  Q 2
Mw MM (40) parameters were estimated using plant data of four
P1  P 1  Pˆ 1  Q1  Q 1 different polymer grades (two grades, A and B, of
high molecular weight, about 280,000, and two
where M M is the monomer molecular weight. grades, C and D, of low molecular weight, between
The viscosity of the reaction mixture is estimated 170,000 and 240,000). The ethylbenzenestyrene
by Mendelson’s correlation (Chen, 2000) for interaction parameter was estimated by using the
concentrations between 40 and 100% of polystyrene flash model at the top stage and minimizing the
in ethyl-benzene and temperatures between 60 and difference between the reactor pressure and the
200°C: calculated pressure, given the top stage temperature
and vapor fraction. The styrenepolystyrene and the
P 3.31 u 1015 X10.7 M 3.4 ethylbenzenepolystyrene binary interaction
w
(41) parameters were estimated by using the bubble point
ª§ E ·§ 1 1 ·º model and minimizing the difference between
exp «¨ a ¸ ¨ T  473 ¸ » , 1Pa s reactor pressure and the calculated pressure, given
¬© R ¹© ¹¼ the reaction temperature and composition of the

Brazilian Journal of Chemical Engineering

 Simulation of Styrene Polymerization Reactors 345

bottom stage (Wada et al., 2002). The obtained obtained by Louli and Tassios (2000) for polystyrene
binary interaction parameters are given below. with aromatic solvents (benzene, toluene). The
presence of different grades of polymer in the liquid
styreneethylbenzene: 0.0728 phase is well represented by the dependency of the
energy and co-volume parameters with the molecular
styrenepolystyrene: 0.8959 weight of the polymer molecules, as can be observed
in Figures 3 and 4. These figures present the plant
ethylbenzenepolystyrene: 0.8818 data and calculated results for the top and bottom
pressures of both reactors, as well as the vapor
The binary parameters between polystyrene and pressures of styrene and ethylbenzene at different top
the other components are compatible with the ones operating conditions.

FirstFirst
Reactor
Reactor

800
Plant data of top stage

Plant data of bottom stage

Styrene vapor pressure

700
Ethylbenzene vapor pressure

Calculated top pressure

Calculated bottom pressure

Pressure mmHg
Pressure (mmHg)

600
Grades C,D

500

Grades A,B

400

300
120 122 124 126 128 130 132 134 136 138 140

Temperature
Temperature (°C)(°C)

Figure 3: Reactor and vapor pressures versus reactor temperature for the first reactor.

Second
SecondReactor
Reactor

1200
Plant data of top stage

Plant data of bottom stage

1100
Styrene vapor pressure

Ehtylbenzene vapor pressure

1000
Calculated top pressure
Pressure mmHg

Calculated bottom pressure

900
Pressure (mmHg)

Grades C,D
800

700

600

500
Grades A,B

400
130 132 134 136 138 140 142 144 146 148 150 152 154 156 158 160
Temperature
Temperature (°C)
(°C)
Figure 4: Reactor and vapor pressures versus reactor temperature for the second reactor.

Brazilian Journal of Chemical Engineering Vol. 25, No. 02, pp. 337 - 349, April - June, 2008
346 A. S. Almeida, K. Wada and A. R. Secchi

Due to the narrow region of coexistence of phases average molecular weight and conversion in the
for this system, the fitted thermodynamic model second reactor are compared for three different
presents a significant sensitivity to the interaction products of each polymer grade, giving relative
parameters. Despite this, the model describes errors less than 5% and 7%, respectively. Before the
satisfactorily the system behavior for the studied parameter estimation, using data from Table 3, the
operating conditions, as shown in Table 4 for the relative errors were up to 15% for Mw and 22% for
relative errors between the reactor pressure and conversion.
calculated pressure using the process model. In Table Merging the kinetic and the thermodynamic
4, the calculated pressures using the ideal gas model models in the reactor model, the polymer properties
are also presented for comparison purposes, where it and the reactor operating conditions were predicted
can be noticed a significant increasing in the relative for four different grades of resins. The main results
error as the polymer concentration in the liquid phase of these simulations can be found in Figure 5, where
increases, showing the importance of considering the two products of each grade are presented. Similar
polymer-species interactions. results were also obtained for formation of styrene
In the second step, the propagation rate dimer and trimer, density, viscosity, and tensile
coefficient of the selected set of published kinetic strength.
parameters was re-estimated by using plant data in In Table 6, the relative errors of the simulation
order to minimize the differences between the results are shown. The larger error observed in the
measured and calculated mass average molecular MFI, Figure 5a, for the grade with lowest molecular
weight and conversion in the second reactor (where weight (grade D) is attributed to the experimental error
the plant data is available). The obtained kinetic reported in the industrial laboratory for this grade. This
constant for the propagation rate was: assertion can be strengthened by the small error in the
mass average molecular weight for that grade, Figure
kp = (6994 r 162) exp(-2600/T) , kg m-3s-1 5b. Satisfactory results were also obtained for the
conversion, Figure 5c, Vicat temperature, Figure 5d,
In Table 5 the measured and calculated mass and other polymer properties.

Table 4: Reactor pressure relative errors (%): proposed model / ideal gas model.

First Reactor Second Reactor

Top Stage Bottom Stage Top Stage Bottom Stage
2.24 / -4.45 –1.16 / -6.19 –1.57 / -4.75 –0.59 / -21.58
–1.69 / -3.19 –2.80 / -6.03 –4.52 / -6.71 –2.36 / -24.59
–2.36 / -3.72 –4.39 / -6.81 –7.31 / -5.49 –5.50 / -21.16
–1.51 / -3.62 –5.13 / -6.23 –2.22 / -3.63 –1.12 / -31.33
3.28 / -1.82 0.49 / -5.66 4.35 / -5.60 9.22 / -33.02
3.17 / -3.89 –0.42 / -6.25 –0.04 / -5.59 6.97 / -27.04
0.40 / -4.17 –2.68 / -6.13 –2.47 / -5.25 0.50 / -25.52
–0.44 / -3.24 –3.80 / -5.90 –0.56 / -4.65 1.61 / -34.53
–4.24 / -3.25 –5.46 / -5.79 –1.96 / -5.28 0.69 / -33.53
0.63 / -2.27 –0.80 / -6.03 –2.45 / -6.55 –3.41 / -32.64
–2.89 / -3.14 –4.80 / -7.33 –3.12 / -5.78 –9.57 / -29.38
3.05 / -2.60 –1.20 / -6.96 –3.31 / -6.38 –3.14 / -39.15
–1.43 / -5.50 –3.48 / -7.19 –1.91 / -5.36 3.81 / -46.55
–0.50 / -2.72 –2.86 / -6.87 –0.26 / -5.71 7.02 / -51.87
2.79 / -2.11 0.71 / -6.90 –1.54 / -7.14 –0.09 / -39.35
–0.11 / -3.63 –2.83 / -7.44 –4.54 / -6.40 –7.64 / -32.92
–0.32 / -3.50 –3.31 / -7.11 –3.40 / -6.31 –7.74 / -34.62
1.35 / -3.60 –0.61 / -7.04 –2.52 / -7.07 –7.23 / -32.51
0.04 / -2.89 –2.43 / -7.33 –2.60 / -7.00 –4.25 / -40.42
1.71 / 3.33 2.60 / 6.59 2.67 / 5.82 4.34 / 33.24
abs. mean value

Brazilian Journal of Chemical Engineering

 Simulation of Styrene Polymerization Reactors 347

Table 5: Comparative results for the second reactor using the estimated kPo

Product Mwexp x103 Mw x103 Error Xexp X Error

A 273.76 286.83 4.77% 64.6 65.63 1.59%
A 282.56 293.49 3.87% 66.1 64.04 3.12%
A 279.49 286.79 2.61% 63.7 63.86 0.25%
B 276.83 281.09 1.54% 64.1 63.02 1.68%
B 281.33 279.61 0.61% 65.5 64.23 1.94%
B 281.67 284.82 1.12% 67.3 64.73 3.82%
C 239.15 227.63 4.82% 66.0 61.94 6.15%
C 230.04 233.97 1.71% 65.1 63.04 3.16%
C 227.74 234.02 2.76% 65.9 63.08 4.28%
D 181.31 188.94 4.21% 66.1 63.50 3.93%
D 184.42 188.93 2.45% 67.0 65.74 1.88%
D 179.22 177.90 0.74% 66.6 64.12 3.72%

25.0 350.0

20.0 300.0
Mw (Dalton x E+03)

Plant Predicted
MFI (g/10')

15.0 250.0

10.0 200.0

5.0 150.0
Plant Predicted
0.0 100.0
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9

(a) Melt flow index (b) Mass average molecular weight

110.0
70

65 105.0
Conversion (%)

60
100.0
Vicat (°C)

55
95.0
50
90.0
45
Plant Predicted
40 85.0
Plant Predicted
35
80.0
0 1 2 3 4 5 6 7 8 9
0 1 2 3 4 5 6 7 8 9

(c) Conversion in 2nd reactor (d) Vicat temperature

Figure 5: Model predictions versus industrial data of four grades of resins.

Table 6: Relative errors of the model predictions

point Grade Conv R2 MW R2 MFI Vicat

1 A 3.12 3.87 6.0 -0.9
2 A 0.25 2.61 15.6 -1.4
3 B 1.94 0.61 2.0 0.9
4 B 3.82 1.12 -4.3 1.1
5 C 3.16 1.71 6.0 -2.2
6 C 4.28 2.76 9.6 -2.9
7 D 1.88 2.45 -4.8 -1.3
8 D 3.72 0.74 -10.3 -1.9

Brazilian Journal of Chemical Engineering Vol. 25, No. 02, pp. 337 - 349, April - June, 2008
348 A. S. Almeida, K. Wada and A. R. Secchi

CONCLUSIONS Polymerization Reactors – A general Approach,

Chemical Engineering Science, 49, n. 22, p.
The obtained kinetic model is particularly 3745-3755 (1994).
important in order to take into consideration the Henderson III, L. S. and Cornejo, R. A.,
thermal initiation and the bi-functional initiator, as Temperature Control of Continuous, Bulk
well as the transfer to sub-products like AH and the Styrene Polymerization Reactors and the
solvent. The thermodynamic model presented a Influence of Viscosity: An Analytical Study,
significant sensitivity to the interaction parameters Industrial and Engineering Chemistry Research,
between the polymer and the other components and 28, p. 1644-1653 (1989).
described satisfactorily the system behavior for the Hui, A. W. and Hamielec, A. E., Thermal
studied operating conditions. The mathematical Polymerization of Styrene at High Conversions
model developed for the auto-refrigerated reactors and Temperatures. An Experimental Study,
was capable of predicting the main polymer Journal of Applied Polymer Science, 16, p. 749-
properties and operating conditions with satisfactory 769 (1972).
accuracy. Kim, K. J. and Choi, K. Y., Modeling of Free
Radical Polymerization of Styrene Catalized by
Unsymmetrical Bifunctional Initiators, Chemical
ACKNOWLEDGMENTS Engineering Science, 44, n. 2, p. 297-312 (1989).
Louli, V. and Tassios, D., Vapor-Liquid Equilibrium
The authors would like to acknowledge the in Polymer-Solvent System with a Cubic
assistance given by the industrial board of INNOVA Equation of State, Fluid Phase Equilibria, 168, p.
S.A., especially to Sergio de Oliveira Machado. 165-182 (2000).
Luyben, W. L., Temperature control of
autorefrigerated reactors, Journal of Process
REFERENCES Control, 9, p. 301-312 (1999).
Özkan, G., Hapoglu, H. and Alpbaz, M., Generalized
Beckman, J., Design of Large Polymerization Predictive Control of Optimal Temperature
Reactors. Polymerization Kinetics and Profiles in a Polystyrene Polymerization Reactor,
Technology, Advances in Chemistry Series 128; Chemical Engineering and Processing, 37, p. 125-
ACS Publisher, Washington, DC (1973). 139 (1998).
Cabral, P. A., Melo, P. A., Biscaia Jr., E.C., Lima, Mayo, F. R., Polymer Preparation., American
E.L. and Pinto, J.C., Free-Radical Solution Chemical Society, Div. Polym. Chem., 2, p. 55
Polymerization of Styrene in a Tubular Reactor – (1961).
Effects of Recycling, Polymer Engineering and Moore, E. R., ed., Styrene Polymers, The Dow
Science, 43, n. 6, p. 1163-1179 (2003). Chemical Company, Reprinted from
Cavin, L., Rouge, A., Meyer, Th. and Renken, A., Encyclopedia of Polymer and Engineering, vol.
Kinect Modeling of Free Radical Polymerization 16, p. 1-246, 2nd. Ed., John Wiley & Sons, Inc.
of Styrene Initiated by Bifunctional Initiator 2,5- (1989).
dimethyl-2,5-bis(2-ethil hexanoyl peroxy)hexane, Peng, F. M., Polybutadiene Grafting and Crosslinking
Polymer, 41, p. 3925-3935 (2000). in High-Impact Polystyrene Bulk Thermal
Chen, C., Continuous Production of Solid Process, Journal of Applied Polymer Science, 40,
Polystyrene in Back-Mixed and Linear-Flow p. 1289-1302 (1990).
Reactors, Polymer Engineering and Science, 40, Pryor, A. and Coco, J. H., Computer Simulation of
n. 2, p. 441-464 (2000). the Polymerization of Styrene. The Mechanism of
Estenoz, D. A., Valdez, E., Oliva, H. and Meira, Thermal Initiation and the Importance of Primary
G.R., Bulk Polymerization of Styrene in Presence Radical Termination, Macromolecules, 3, n. 5, p.
of Polybutadiene: Calculation of Molecular 500-508 (1970).
Macrostructure, Journal of Applied Polymer Russo, L. P. and Bequette, B. W., Operability of
Science, 59, p. 861-885 (1996). Chemical Reactors: Multiplicity Behavior of a
Flory, P. J., The Mechanism of Vinyl Polymerizations, Jacketed Styrene Polymerization Reactor,
Journal of the American Chemical Society, 59, n. Chemical Engineering Science, 53, n. 1, p. 27-45
2, p. 241-253 (1937). (1998).
Freitas Filho, I. P., Biscaia Jr., E. C. and Pinto, J. C., Toledo, E. C. V., Martini, R.F., Maciel, M.R.W. and
Steady-State Multiplicity in Continuous Bulk Maciel Filho R., Process Intensification for High

Brazilian Journal of Chemical Engineering

 Simulation of Styrene Polymerization Reactors 349

Operational Performance Target: Autorefrigerated p. 629-641 (1991).

CSTR Polymerization Reactor, Computers and Wada, K., A.R. Secchi and A.S. Almeida, “Liquid-
Chemical Engineering, 29, p.1447–1455 (2005). Vapor Equilibrium in Auto-Refrigerated
Villalobos, M.A., Hamielec, A.E. and Wood, P.E., Polymerization Reactor”, In Proceedings do VI
Kinetic Model for Short-Cycle Bulk Styrene Iberoamerican Conference on Phase Equilibrium
Polymerization through Bifunctional Initiators, for Process Design (EQUIFASE 2002), p. 39, Foz
Journal of Applied Polymer Science, 42, do Iguaçu, Argentina and Brazil (2002).

Brazilian Journal of Chemical Engineering Vol. 25, No. 02, pp. 337 - 349, April - June, 2008

View publication stats