Anda di halaman 1dari 10

2 nd Mercosur Congress on Chemical Engineering

th
4 Mercosur Congress on Process Systems Engineering

ALTERNATIVE DESIGN FOR EMULSION POLYMERIZATION


TUBULAR REACTOR OF STYRENE WITH INTERNAL ANGLE
BAFFLES: REACTOR DIAMETER EFFECT
Mendoza Marín, F.L.1*, Lona, L.M.F.2, Maciel Filho, R. 1
1
Laboratorio de Otimização, Projetos e Controle Avançado – LOPCA, UNICAM P
2
Laboratorio de Análise, Simulação e Síntese de Processos Químicos – LASSPQ. UNICAMP

Abstract. Steady state modeling and simulation of emulsion homopolymerization process of styrene with
baffles into tubular reactor as static mixer are presented in this work. It is assumed cylindrical one-dimensional
model, fully developed laminar plug flow, and the model solved by finite volume method. Also it is considered
the Smith-Ewart model to estimate the monomer conversion and Arrhenius chemical kinetics as laminar finite-
rate model to compute chemical source. The objective is to model, simulate and analyze the emulsion
polymerization reactor performance with baffles for different reactor diameter. The simulation data were
validated with experimental results with a good concordance. The performance of baffled reactor leads to
obtain polymer with better properties.

Keywords: Emulsion polymerization, tubular reactor, baffles.

1. Introduction

The fundamentals of emulsion polymerization are sufficiently well understood so that new products can be
made, and old ones re-formulated. An example is the economically and technically importance to produce a
submicron suspension of colloidally stable polymer particles or polymer solution, which is called a latex with 40-
50% by free radical mechanism of the total polymer. It is a heterogeneous reaction process in which unsaturated
monomers or monomer solutions are dispersed in a continuous phase with the aid of an emulsifier system and
polymerized with free-radical initiators. It leads to a high molecular weight polymer and high reaction rates in
many cases. It has low viscosity which is a specific advantage compared to other polymerization techniques;
viscosity reduces pumping and agitation equipment requirements and improves heat transfer to the reactor walls
(Paquet and Ray, 1994 a, b; Chern, 1995; Gilbert, 1995; Scholtens, 2002). Emulsion polymerization is a complex
heterogeneous process involving transport of monomer and other species and free radicals between aqueous and
organic phases. Compared to other heterogeneous polymerization, like suspension or precipitation, it is likely the
most complicated system; all this factor make modeling of this system very difficult (Bockhorn, 1992; Gao and
Penlidis, 2002).
The chemical processes in the chemical industry have growing operational difficulties caused by the
diversification and specification of products and investigations for alternative reactor projects, and analysis of
their behavior under static and dynamic conditions are wellcome. In conventional tubular reactor (TR) most
reaction happens towards the reactor entrance, and the variable reaction for exothermic reactions as well as the

*
To whom all correspondence should be addressed.
Address: Departamento de Processos Químicos, CP.6066, CEP 13081-970, TL. (19) 37883910, Campinas, SP, Brazil

1
2 nd Mercosur Congress on Chemical Engineering
th
4 Mercosur Congress on Process Systems Engineering

limitations to heat transfer near the wall makes the behavior to be very complex (Maciel Filho e Domingues, 1992;
Kiparissides, 1996, Toledo, 1999).
The emulsion homopolymerization of styrene was considered and a steady-state one -dimensional model was
developed to simulate the system. The model is written in cylindrical coordinate with the hypothesis of fully
developed laminar plug flow and solved by finite volume method (FVM). In fact, conventional tubular reactor may
have serious limitations to heat transfer leading to the obtation of product with undesired conditions. Bearing this
in mind in this work it is proposed an alternative TR based on the placement of baffles inside the reactor for
emulsion polymerization of styrene (EPS). The objective is to model, simulate and analyze the emulsion
polymerization reactor performance with baffles for different reactor diameters. The inclined baffle increase the
heat transfer and the mixture of reactants and emulsifier agents. The products were characterized with particles
number (homogeneous and heterogeneous nucleation), molecular weight distribution, polymer particles size and
polymer viscosity distribution. These results were validated with literature results under same or approximate
condition. The performance of baffled reactor leads to obtain polymer with better properties.

2. Reactional System

Conditions of test: In order to evaluate the performance of the proposed design the EPS is considered,
comparison with conventional TR is carried out. Figures 1 and 2 show the schematic of the baffled reactor. To
represent the system, a simplified one dimensional deterministic model is developed with the following
assumptions: flow along the axial direction (negligible diffusion); fully developed axial velocity of fluid flow;
polymer phase is the main locus of polymerization; particle size is monodisperse; and the monomer conversion is
estimated by Smith-Ewart model.

Reactants Products

(a) (b)

Fig. 1 (a) Schematic representation of polymerization TR with baffles and (b) Profile of polymerization TR with internal-
inclined angular baffles.

α
α θ
θ α θ θ
θ α

(a) (b) (c) (d) (e) (f)

Fig. 2 Variation of t ransversal area of fluid flow of EPS inside TR beneath or over baffles.

E-mail: marin@lopca.feq.unicamp.br
2
2 nd Mercosur Congress on Chemical Engineering
th
4 Mercosur Congress on Process Systems Engineering

The work of Bataille et al. (1982) was considered as case study at 60°C with the recipe: potassium persulfate,
0.026 (mol/L); Sodium dodecyl sulfate, 0.070 (mol/L); styrene, 8.39 (mol/L) and water, 161.52 (mol/L). The
bibliographical references of the properties (Table 1) can be found in Mendoza Marín (2004).

Table 1: Physics, chemicals, mechanics, transport, rheologys, thermodynamics, geometrical properties (grid, particles, reactor,
baffles) properties of Styrene as database for simulation of EPS.
Symbol Value, Unit and Description
Na 6.02x1023(molecule/mol); Abogadro’s number; π = 3.14159
Rg 1.987 (cal/mol K); gas constant
ρp, ρm 1.25-0.0004202T (Kg/L), polymer and 0.949-0.00128(T -273.15); (Kg/L), monomer density
anrp 0.5; average number of radicals per particle
CMC 0.008 (mol/L); critical micelle concentration
CMw 0.005 (mol/L); monomer concentration in water phase (=Mw)
fi 0.5; initiator efficiency
Kcm 4πDwr micNA (1/min); rate const of aqueous phase radical capture by micelles
Kcp 4πDprpNA (1/min); rate const of radicals capture by polymer particles
Kd 1.524x1018 exp(-33320/ RgT) (1/min); rate constant of initiator decomposition
Kp 4.703x1011 exp(-9805/RgT) (L/mol min); rate const. of propag. of poly. partis
Kt 1.04619x1010 exp(-2950,45/RgT) (L/mol min); global rate const for termination
[M]P (1-φp)ρm /MWs (mol/L); monomer concentration in polymer particle
MWe 288.38 (g/mol); molecular weight of surfactant
MWi 271.3 (g/mol); molecular weight of initiator
MWs 104 (g/mol); molecular weight of styrene
ncr 5; critical chain length at which water phase radical can be absorbed
nem 60 (No.Emul/mic); number of emuls ifier molecules in a micelle
Sa 3x10-17(dm2); area covered by one molecule of emulsifier
φp, φm 0.4, volume fraction of polymer; 0.6, volume fraction of monomer
Re 5000 (laminar) and 13600 (turbulent) Reynolds number
Vin 0.27027 (laminar) and 0.7351 (turbulent) (m/min) inlet velocity in TR
Dp 1.76x10-12 (dm2/min ); (diffusivity of monomer radicals in polymer phase
Dw 1.76x10-9 (dm2/min ); (diffusivity of monomer radicals in water phase
µ 0.001 (Kg/m s); viscosity of polymer
Tin 333.15 and 363.15 (K); inlet temperatures to the reactor
∆H -16682.2 (cal/mol); polymerization reaction heat of styrene
N 51; number of nodal points
rmic , rp 27.5 (Å); radius of micelle and 275 (Å); radius of polymer
Dr, Lr 1 (m) diameter and 20 (m) length of TR
Lbr 1 (m) length of baffle separation and Nb = 6, 18 number of baffles

3. Mathematical Modeling

In order to represent the proposed reactor shown in Figures 1 and 2, with the case study of emulsion
homopolymerization of styrene the following model equation may be written.

3.1. Chemical reaction


The mechanism of EPS may schematic and brief shown as:

Initiation:  2I −
I 2 →
Kd
I − + M 
→ R1
Kp
(01)

Radical Absorption by Micelles (Micellar Nucleation (MN))

3
2 nd Mercosur Congress on Chemical Engineering
th
4 Mercosur Congress on Process Systems Engineering

Diffusion in the water phase Absorption in the micelle surface


Kcmw Kcm: Kcmd
Rrw → RrwMIC Rr + MIC ←
→ Rr MIC (02)

Radical Absorption by Particles (Homogeneous Nucleation (HN))


Diffusion in the water phase Absorption in the particle surface
Kcpw Kcp:Kcpd
Rrw → RrwPP Rr + PP←
→ Rr PP (03)
Kp
Propagation: Rr + M → Rr +1 (04)

Termination: by combination (Ktc) and disproportionation (Ktd)

Rn + Rm Ktc
 → Pn + m or Rn + Rm Ktd

→ Pn + Pm (05)

3.2. Conservative models


Principle of Mass Conservation in general form for a chemical species j reacting in a flowing fluid with varying

density, temperature, and composition is


∂C j
∂t
( )
r
+ ∇. C j u + ∇.J j = R j (06)

where Cj is the molar concentration of species j; ∂C j / ∂t is the nonsteady-state term expressing accumulation or
r r r
depletion; ∇ is the gradient operator; ∇.u is the divergence of a vector function u ; u is the three-dimensional
r
( )
mass-average velocity vector; ∇. C j u is the transport of mass by convective flow; Jj is the molar flux vector for

species j with respect to the mass-average velocity; ∇. J j is molecular diffusion only; Rj is the total rate of change

of the amount of j because of reaction. Species j occurs in liquid phase. This equation can take in single-phase or
“homogeneous” or “pseudohomogeneous” reactors (Bird, 1960; Froment, 1990).
The generalized Laminar Finite-rate Model was applied to compute the chemical source terms (R j ). The model is
exact for laminar flames, but is generally inaccurate for turbulent flames due to highly non-linear Arrhenius
chemical kinetics. The net source of chemical species j due to reaction Rj is computed as the sum of the Arrhenius
reaction sources over the Ni reactions that the species participate in:

 
∏ [C j,i ]n ∏ [C ]
Ni Ni Ni Ni
Rj = ∑ R j, i = ∑ K
 f ,i
fj ,i
− K b, i j ,i
n bj, i 
 (07)
i =1 i =1  i =1 i =1 

where Rj,i is the Arrhenius molar rate of creation/destruction of species j in reaction i; Kf,i is the forward rate
constant for reaction i, Kb,i is the backward rate constant for reaction i, Ni is the number of chemical specie s in

reaction i, CJ,i is the molar concentration of each reactant and product species j in reaction i, n fj,i is the forward rate
exponent for each reactant and product species j in reaction i, nbj,i is the backward rate exponent for each reactant
and product species j in reaction i. Only non-reversible reactions was considered and the mass balance of
equations gives the different chemical source term as free radical (RRw), initiator (RI ), monomer (RM), surfactant
(RE =0, by to be inert), and polymer (RP ):

4
2 nd Mercosur Congress on Chemical Engineering
th
4 Mercosur Congress on Process Systems Engineering

ncr−1
 
− Kcp[N P ][. R]w − Kcm[MIC ][. R]w − Ktw[R]w
KpM w 2
RRw = RI − R I  
 KpM w + Kcp[N P ] + Ktw[R]w 
(08)

Kp[M ]P N P n Kp[M ]P N P n
R I = − fiKd [I ]w ; RM = − − Kpw[M ]w [R]w ; RP = + Kpw[M ]w [R ]w (09)
N AVP N AVP

Newton’s Second Law of Momentum was applied to a small volume element moving with the fluid that is
accelerated because the forces acting over it. The motion equation in terms of τ is
r
r r
= −∇P + (∇.τ ) + ρ .g
Du
ρ (10)
Dt
r r
here g represent the body force per unit area; ρ is the density; P is the pressure; τ is the extra stress tensor, and

D/Dt is the material or substantial derivative (Bird, 1960; Tucker, 1989; Versteeg, 1998).
In the Micellar Nucleation (MN), RMN is the rate of particle formation by MN, [R]w is the formation rate of total
oligomeric radicals in the aqueous phase, It was obtained by the mechanism of EPS, steady state hypothesis and
geometric progression approximations and αm is the probability for oligomeric radical for micelle propagation
(Gilbert, 1995; Gao and Penlidis, 2002; Mendoza Marín, 2004).

d [Np ]m  1 − α mncr −1 
RM N = = Kcm[MIC][. R]w ; [R]w = RI  
KpMw + Kcm[MIC] + Ktw[R]w  1 − αm  (11)
dt  
KpMw
αm =
KpMw + Kcm [MIC ] + Ktw[R ]w + Kcp [Np ]
(12)

In the Homogeneous Nucleation (HN), RHN is the rate of particle formation by HN, [Rncr-1]w is the formation rate
of oligomeric radicals with critical chain length (ncr) in the aqueous phase, the same hypothesis adopted for [R] w
were considered and αh is the probability for an oligomeric radicals for homogeneous propagation.
d [Np]h
= KpM w [Rncr−1 ]w ; [Rncr−1 ]w =
RI
RH N = α hncr −1 (13)
dt KpMw

KpMw
αh =
KpMw + Ktw[R]w + Kcp[Np]
(14)

3.3. Characterization of polymer particle


The polymer particle (by MN and HN) is determined with Eq. (06) written like Eq. (15) and source term as Eq.
(16):

∂C N P  ∂C N P 1 ∂C N P ∂C N P    ∂C N P  1 ∂ CN P ∂ CN P
2 2 
+  v r + vθ + vz  = D NB  1 ∂  r + + + R (15)
∂t ∂r r ∂θ ∂z   r ∂r  ∂r  r 2 ∂θ 2 ∂z 2  NP
     

Kcm[MIC ]R I  1 − α mncr−1 
  + RI α hncr−1
RNp = RM N + RH N =
KpMw+ Kcm[MIC ] + Ktw[R]w  1− α m  (16)
 
The average swollen (R S)and unswollen(R) particle polymer size radius are

5
2 nd Mercosur Congress on Chemical Engineering
th
4 Mercosur Congress on Process Systems Engineering

1/3 −1 / 3
 3 MW p CPs   ρm 
Rs =   R = Rs  
 ρm + [M ]P MWm 
 4πρ Pφ P N A C N   (17)
 P 

The viscosity of polymer (µ) was estimated (Harkness, 1982) as

2013   1387  3 
Ln( µ ) = −13,04 + + MWP0,18 3,915 X j − 5, 437 X 2j +  0,623 + X j
T  
(18)
T  

The Molecular Weight Distribution is determined with mechanism of EPS, discrete transfomation method and
generating function. The Cumulative Dead Polymer were number-average molecular weight (M nc ) and weight-

average molecular weight ( M cw ).

2aMWs Xj  a + 2b 3a 
M nc = M cw = MWs  − X j  (19)
a−b  b − aX j   b − a 2 (b − a ) 
Ln 
 b 

where: a = Kp[M ] ; b = a + Ktc[R]w (20)

3.4. Reactor and baffles


The internal transversal areas over or beneath baffles were calculated where α is the increment angle from the TR
center point until total diameter, θ is the increment angle to calculate the fluid flow area of EPS beneath or over the
baffles, Avb is fluid flow variable area inside TR beneath or over baffles, Ar is the area of TR without baffles and
Afb is the fixed area inside TR over or beneath baffles (Figure 2 (a-f)).
 ID (1) 
αr (1) = arcSen 2  [rad] (21)
 Dr 

π  αr(1) 
θr (1) = θg (1) = π 1 − 2  [rad] (22)
180  π 

Dr 2
Avb(1) = (θr (1) − Senθ r(1)) (23)
8

π 2
Afb (1) = Ar − Avb(1) = D − Avb(1) (24)
4 r
Calculus of Avb from I=2 to Number of areas beneath or over baffles (Nab)
Dvb(I ) = Dvb( I − 1) + ∆Dvb (25)

 ID (I )   2
αr ( I ) = arcSen  2  = arcSen  (Dvb( I ) − 0.5 * Cr ) (26)
 Dr   Dr 

π  2 
θr (I ) = θg (I ) = π 1 − αr (I ) (27)
180  π 

Dr 2
Afb (I ) = (θ r( I ) − Sen θr (I )) (28)
8

Avb(I ) = Ar − Afb( I ) (29)

6
2 nd Mercosur Congress on Chemical Engineering
th
4 Mercosur Congress on Process Systems Engineering

The axial velocity such as at TR inlet, over or beneath baffles were estimated from Momentum Newton’s
Second Law, thus vin is the axial velocity at TR inlet and vb is the axial velocity over or beneath baffles , both in
isothermal condition.
µ .Re µ. Ar Re
vin = vb = (30)
ρ.Dr ρ. Avb Dr

4. Numerical Method

Finite Volume Method was used as numerical method according to Patankar (1980), Maliska (1995), and
Versteeg (1998). The Discrete approximation was applied to conservative balance of convection and source only,
e.g. the equation of particle number, Eq. (15) and source term or the overall rate of formation of particles by micelle
and homogeneous nucleation, Eq. (16).
The Linear approximation was applied to source term of Eq. (16). The Taylor’s Series method of linearization was
used to linearize Sφ:

The overall equation of source-term linearization, (Sφ = S N P ) in Eq. (16) leads to

S Np = R Np = S Npm + S Nph = SUT − SPT [Np ] P (31)

The source-term linearization of Micellar Nucleation ( S Npm ) in Eq. (11) allows to write

S Npm = SUM − SPM [Np ]P (32)

The source-term linearization of Homogeneous Nucleation ( S Nph ) in Eq. (13) is

S Nph = SUH − SPH [Np ] P (33)

Now the integral form of general transport equations in one-dimensional control volume is given by:
∂(v z Np )

cv
∂z
Adz = ∫S
cv
Np Adz = ∫ (SUT − SPT .Np
cv
P )Adz (34)

( AvNp )e − ( AvNp ) w (
= S Np Az )e − (S Np Az )w = (SUT − SPT .Np P ) AP ∆z (35)

The continuity Equation is:


( Avρ ) e = ( Avρ )w ; F = ρv ; ( AF )e = (AF ) w (36)

In the interpolation was applied the upwind difference scheme to calculate the particle number as:

Fwφw = φW Fw ,0 − φ P − Fw ,0 Feφe = φP Fe ,0 − φE − Fe ,0 (37)

Mass Balance or continuity equation


A w − Fw ,0 = A w F w ,0 − Aw F w ; Ae Fe ,0 = Ae − Fe ,0 + Ae Fe (38)

After integration and interpolation were obtained the linear algebraic equations. In the minimum control volume
the discretised equation for minimum control volume is
a P NpP = a E NpE + S U (39)

7
2 nd Mercosur Congress on Chemical Engineering
th
4 Mercosur Congress on Process Systems Engineering

where: a W = 0 ; a in = Aw Fw ,0 ; a E = Ae − Fe ,0 ; a P = a E + S P (40)

S P = SPT. AP ∆z + a in + ( Ae Fe − Aw Fw ) ; SU = SUT .AP ∆z + a in Npin (41)

In the internal control volume from I=3 to I=N-2, the discretised equation for internal control volume is
a P Np P = a W Np W + a E Np E + S U (42)

where: aW = Aw Fw ,0 ; a E = Ae − Fe ,0 ; a P = aW + a E + S P (43)

S P = SPT. AP ∆z + ( Ae Fe − Aw Fw ) ; SU = SUT. AP ∆z (44)

In the maximum control volume I=N-1, the discretised equation for maximum control volume:
a P Np P = a E Np E + S U (45)

where: aW = Aw Fw ,0 ; a E = 0 ; a P = aW + S P (46)

S P = SPT. AP ∆z + ( Ae Fe − Aw Fw ) ; SU = SUT. AP ∆z (47)

The solution of algebraic equations as a system of linear algebraic equations were obtained by Thomas
Algorithm or the Tri-diagonal Matrix Algorithm (TDMA) as direct method for one-dimensional situation problems.
In isothermal condition the discretization equation, for example Eq.(39), Eq.(42) and Eq.(45) are a linear algebraic
equations, and the set of such equations is solved by the methods of linear algebraic equations with the
following iterative procedures to find the value of the number of particle with or without baffle: 1) Guess a number
of particle, aN p; 2) Calculate the initiator and radicals distribution; 3) Calculate the number of particle, Np; 4)

Compare the Np with the actual number of particle, aN p; 5) If absolute value of N p − aN p is greater than a factor

of convergence, then to set aNp=Np (return to the step 1). We must make another difference is lower.

5. Results and Discussion

The simulation results of conversion (Xj) versus length of the reactor (Z) for styrene without baffles (C0) and
with baffles (C18) in isothermal condition (C) at 60 °C, with reactor diameter (Dr) at Dr=0.5 and Dr=1 m are
displayed in Figures 3 (a) and (b). The conversions at Dr=1 m are higher than the conversions at Dr=0.5 m without
and with baffles.
The comparative results in isothermal condition at 60°C of computational (XcC), experimental (Xe) and
simulation conversion (Xs) (literature results) versus residence time (t) inside TR are shown in Figure 3 (c). When
compared to the model used by Bataile et al. (1982), it can be observed that the three curves have same behavior
and validate the FVM through the comparative of experimental and simulation results.

1 1 1

0,8 0,8 0,8

0,6 0,6 XcC


0,6 C18-d.5
C0-d.5
Xj
Xj
Xj

Xe
0,4 C0-d1 0,4 C18-d1 0,4
Xs
0,2 0,2 0,2

0 0 0
0 5 10 15 20 0 5 10 15 20 0 20 40 60 80
Z(m) Z(m) t(min)
8
2 nd Mercosur Congress on Chemical Engineering
th
4 Mercosur Congress on Process Systems Engineering

Fig. 3 (a) Conversion of monomer without (Nb=0) and (b) with (Nb=18) baffles, and (c) experimental validation in isothermal
(CNb=60°C) conditions.

The prediction results of the polystyrene particles number (Np) versus length of the reactor (Z) with baffles (C18)
in isothermal condition (C) at 60°C with reactor diameter at Dr=0.5 and Dr=1 m are exhibited in Figure 4 (a). The
particle
(a) numbers with Dr=1 m are higher(b)
than the particle numbers with Dr=0.5(c)m, but the simulation results with
Dr=0.5 m are approximately close to the experimental results. The mathematical model of the Eqs. (15) and (16)
were used to simulate the particles numbers with MN and HN.
The simulation of the polystyrene particles size of unswollen particle radius (R) versus length of the reactor (Z)
with baffles (C18) in isothermal condition at 60°C with reactor diameter of Dr=0.5 and Dr=1 m are shown in Figure
4 (b). The particle sizes with reactor diameter Dr=0.5 m are smaller than the particle size with reactor diameter
Dr=1m. The mathematical model of the Eq. (17) was used to estimate the particle sizes for both diameters as shown
in Figure 4(b).
The results of polymer viscosity distribution (Ln(µ)) inside TR versus conversion of monomer (Xj) with baffles
(C18) in isothermal condition (C) at 60°C with reactor diameter Dr=0.5 and Dr=1 m are shown in Figure 4 (c). The
viscosity with reactor diameter Dr=0.5 m is lower than the viscosity with reactor diameter Dr=1 m like expected.
The mathema tical model of the Eq. (18) was used to estimate the viscosity for both diameters as shown in Figure 4
(c).

8,00E+19 5 50
7,00E+19
6,00E+19 4 40 C18-d.5
5,00E+19 C18-d1
R(cnm)
Np(1/l)

3 30
4,00E+19
Ln(MI)

3,00E+19 2 20
2,00E+19 C18-d.5 C18-d.5
1 C18-d1 10
1,00E+19 C18-d1
0,00E+00 0
0
0 5 10 15 20 0 5 10 15 20
0 0,2 0,4 0,6 0,8 1
Z(m) Z(m) -10
(a) (b) (c) Xj

Fig. 4 (a) Number of particles with Nb=18 baffles; (b) Average particle size distribution with Nb=18 baffles and (c) Viscosity
distribution with Nb=18 baffles.

The simulation results of the cumulative average molecular weight distribution (Eq. (19)) versus conversion of
monomer (Xj), with baffles (MnC18, MwC18) in isothermal at 60 °C with reactor diameter Dr=0.5 and Dr=1 m are
shown in Figure 5. The performances of the average molecular weight are different as shown in Figures 5 (a)-(c).
The number (Mn) and weight (Mw) molecular have a better distribution when the reactor diameter effect decrease
in proposed reactor as shown in Figure 5 (c).

4000000 4000000 4000000


3500000 3500000 3500000
Molecular weight
Molecular weight
Molecular weight

3000000 3000000 3000000


2500000 2500000 2500000
2000000 2000000 2000000
1500000 1500000 MnC18-d.5
1500000
1000000 1000000 1000000
MwC18-d.5 9
MnC18-d.5 MnC18-d1 MnC18-d1
500000 500000 MwC18-d1 500000
MwC18-d.5 MwC18-d1
0 0 0
0 0,2 0,4 0,6 0,8 1 0 0,2 0,4 0,6 0,8 1 0 0,2 0,4 0,6 0,8 1
Xj Xj Xj
2 nd Mercosur Congress on Chemical Engineering
th
4 Mercosur Congress on Process Systems Engineering

Fig. 5 Molecular weight distribution with reactor diameter effect of (a) Dr=0.5 m, Nb=18 baffles, (b) Dr=1 m, Nb=18 baffles
and (c) Comparative results of Dr=0.5 and Dr=1 m.

6. Conclusions
The performances of emulsion polymerization proposed reactor is higher in relation to conventional tubular
reactor when desired properties are analyzed. The proposed reactor design consists on the placement of baffles
inside the tube. With the proposed design a better average molecular weight distribution is achieved. The effect
of the diameter is such as that the design parameter can be used to obtain the products with the desired
characteristic. In particular the conversion and particle number decrease when the reactor diameter was
diminished. In general the performance of baffled reactor leads to obtain with better properties.

References
BOCKHORN, H. Mathematical Modeling, ULLMANN’S Encyclopedia of Industrial Chemistry, Vol.B1, 2, 1992.
CHERN, C.S., Diffusion-Controlled semibatch Emulsion Polymerization of Styrene; Journal of Applied Polymer science, Vol.56, 221-230, 1995.
FARINAS, M.I., GARON, A. e LOUIS, K.S., Study of heat transfer in a horizontal cylinder with fins, Ver. Gen. Therm, 36, 398-410, 1997.
FROMENT, G.F. AND BISCHOFF, K.B.; Chemical Reactor Analysis and Design; Second Edition; John Willey-Sons; 1990.
GAO,J. AND PENLIDIS,A.; Mathematical Modeling and Computer Simulator/database for Emulsion Polymerizations; Prog. Polym. Sci. ;27,
403-535, 2002.
GILBERT, R.G.; Emulsion Polymerization, A Mechanistic Approach; Academic Press; London; 1995.
HARKNESS, M.R. A computer simulation of the bulk polymerization of polystyrene in a tubular reactor. M.S. Thesis, Rensselaer Polytechnic
Institute, Troy, New York, 1982.
KIPARISSIDES, C. Polymerization Reactor Modeling: A Review of Recent Developments and Future Directions. Chemical Eng. Science, Vol.51,
No.10, 1996.
MACIEL FILHO, R., and DOMINGUES, A., Multitubular Reactor for Obtención of Acetaldehyde by Oxidation of Ethyl Alcohol, ISCRE 12
(Twelfth International Symposium on Chemical Reaction Engineering), Turin, Itália, 1992.
MENDOZA MARIN, F.L.; Modelagem, Simulação e Analise de Desempenho de Reatores Tubulares de Polimerização com Deflectores Angulares
Internos; Tese de Doutorado, FEQ/UNICAMP, Campinas, 2004.
PAQUET, D.A. AND RAY, W.H.; Tubular Reactors for Emulsion Polymerization: I and II, Experimental Investigation; Reactor, kinetics and
catalysis and Model Comparison with Experiments; AIChE journal, vol.40 No.1, January 1994.
PATANKAR, S.V.; Numerical Heat Transfer and Fluid Flow; Hemisphere Publishing Corporation; New York, 1980.
SCHOLTENS, C. A.; Process Development for Continuous Emulsion Copolymerization; CIP -DATA Library Technische Universiteit Eindhoven,
Ter Verkrijging van de graad van Doctor, 2002.
SIIROLA, J.J. Industrial Applications of Chemical Process Synthesis, AICHE, Vol.23, 1-62, 1996.
TOLEDO, E.C.V., Modelagem, Simulação e Controle de Reatores Catalíticos de Leito Fixo, Tese de Doutorado, FEQ/UNICAMP, Campinas,
1999.
TUCKER III, C.L.; Fundamentals of Computer Modeling for Polymer Processing; Hanser Publishers, Munich Vienna New York, 1989.
VERSTEEG, H.K. AND MALALASEKERA, W.; An Introduction to Computational Fluid Dynamics, The Finite Volume Method; 2nd Ed.;
Addison Wesley Longman Limited; 1998.

Acknowledgement
The author is grateful to the PEC-PG/CAPES and LOPCA-DPQ/FEQ UNICAMP for their financial support.
10