Chapter 4

Copolymerization

4.1 Kinetics of Copolymerization

4.1.1 Involved Chemical Reactions

Initiation I2 + M −→ 2 R•1 r = 2 f kd I2 ≡ RI
kpAA
Propagation A•m,n + A −→ A•m+1,n r = kpAA AA•m,n
kpAB
A•m,n + B −→ B•m,n+1 r = kpAB BA•m,n
kpBA
B•m,n + A −→ A•m+1,n r = kpBA AB•m,n
kpBB
B•m,n + B −→ B•m,n+1 r = kpBB BB•m,n
kfsA
Chain Transfer A•m,n + S −→ R•1 + Pm,n r = kfsA S A•m,n
kfsB
B•m,n + S −→ R•1 + Pm,n r = kfsB S B•m,n
ktAA
Termination A•m,n + A•r,q −→ Pm+r,n+q or Pm,n + Pr,q r = ktAA A•m,n A•r,q
ktAB
A•m,n + B•r,q −→ Pm+r,n+q or Pm,n + Pr,q r = ktAB A•m,n B•r,q
ktBB
B•m,n + B•r,q −→ Pm+r,n+q or Pm,n + Pr,q r = ktBB B•m,n B•r,q

where n, m, r, q = [1...∞] and kt = ktc + ktd

Note that a similar notation as for homopolymerization has been used above with, for example,
A•m,n indicating the concentration of active chains containing m units of monomer A and n units
of monomer B, and terminating with a monomer A unit. The latter unit, where the radical is
located, defines the reactivity of the active chains, and accordingly determines the rate constant
of all the reactions above. This is usually referred to as the terminal kinetic model. So for
example, kpAB represents the rate constant for the propagation of radical chain of type A and
any length with a monomer B.

34
CHAPTER 4. COPOLYMERIZATION

4.1.2 Long-Chain Approximation

The long-chain approximation (LCA) starts from the observation that the following equality
holds along a polymer chain:

(number of transitions A → B) = (number of transitions B → A) ± 1 (4.1)

These geometrical considerations lead to the following condition on reaction rates:

kpAB A• B = kpBA B• A (4.2)

Let us define
R• = A• + B• (4.3)

where A• represents the total concentration of radicals of type A, that is radicals terminating
with a monomer of type A:
∞  ∞
•
A = A•m,n (4.4)
m=1 n=1

Therefore we obtain:
kpAB A• B = kpBA (R• − A• ) A (4.5)

kpBA A
⇒ A• =  R• = pA R• (4.6)
kpAB B + kpBA A
kpAB B
⇒ B• =   R• = pB R• (4.7)
kpAB B + kpBA A
where pA is the probability of having a radical of type A. Now, let us write the mass balance
for A• and B• :
dA•
= RIA + kpBA B• A − kpAB A• B − ktAA A•2 − ktAB A• B• (4.8)
dt

dB•
= RIB + kpAB A• B − kpBA B• A − ktBB B•2 − ktAB A• B• (4.9)
dt
The rates of initiation for the monomers A and B are defined as the product of the overall
rate of initiation and the respective monomer mole fractions in the monomer phase, i.e., RIA =
RI XA and RIA = RI XB , where
A B
XA = and XB = (4.10)
A+B A+B

35
CHAPTER 4. COPOLYMERIZATION

The summation of equation (4.8) and (4.9) gives:

dR• dA• dB•

= + = RIA + RIB − ktAA A•2 − ktBB B•2 − 2ktAB A• B• (4.11)
dt dt dt

Applying the pseudo steady state assumption to R• and using RI = RIA + RIB we obtain:

RI − ktAA p2A R•2 − ktBB p2B R•2 − 2ktAB pA pB R•2 = 0 (4.12)

and therefore 
RI
R• = (4.13)
kt∗

where kt∗ is defined as follows:

kt∗ = ktAA p2A + 2ktAB pA pB + ktBB p2B (4.14)

To conclude, we obtain the same result as in the case of homopolymerization with a pseudo-
termination rate constant, kt∗ defined by equation (4.14) replacing the actual termination rate
constant, kt .

4.1.3 Rate of Monomer Consumption

Since propagation reactions dominate monomer consumption, we have:

 
RpA = kpAA A• A + kpBA B• A = kpAA pA + kpBA pB AR• (4.15)

 
RpB = kpBB B• B + kpBA A• B = kpBB pB + kpAB pA BR• (4.16)

 ∗
RpA = kpA AR• ∗ •
⇒ ∗ • ⇒ Rp,tot = RpA + RpB = kp M R (4.17)
RpB = kpB BR

∗
⇒ Rp,tot = RpA + RpB = kp M R• (4.18)
∗ ∗
Introducing the pseudo-propagation rate constants kpA and kpB we obtain again a similar ex-
pression as in the case of homopolymerization.

36
CHAPTER 4. COPOLYMERIZATION

4.2 Chain Length Distribution of Copolymers

4.2.1 Population Balance Equation
In this section we calculate the CLD in the case of a copolymerization process. Let us therefore
consider the population balance of chain radicals terminating with a monomer unit A and
containing r units of any type (A or B):
dA•r
= kpAA · A · A•r−1 + kpBA · A · B•r−1
dt  
− kpAA · A + kpAB · B A•r
−kfsA · S · A•r − (ktAA · A• + ktAB · B• ) A•r (4.19)

and similarly for the radicals terminating with a B unit:

dB•r
= kpBB · B · B•r−1 + kpAB · B · A•r−1
dt  
− kpBB · B + kpBA · A B•r
−kfsB · S · B•r − (ktBB · B• + ktBA · A• ) B•r (4.20)

Applying the long-chain approximation and using R•r = A•r + B•r we obtain:
dR•r     
= kpAA · pA + kpBA · pB · A + kpBB · pB + kpAB · pA · B · R•r−1
dt     
− kpAA · pA + kpBA · pB · A + kpBB · pB + kpAB · pA · B · R•r
− (kfsA · pA + kfsB · pB ) · S · R•r
 
− ktAA · p2A + 2ktAB · pA · pB + ktBB · p2B · R•r · R• (4.21)

A further simplification can be achieved when using the pseudo-homopolymerization rate

constants:
dR•r ∗ ∗
= kp M R•r−1 − kp M R•r − kfs∗ S R•r − kt∗ R•r R• (4.22)
dt
This expression is identical to the population balance equation (PBE) for a homopolymer
containing r monomer units. We can extrapolate this result to all PBEs involved in computing
the CLD for copolymers. This method is the so-called pseudo-homogeneous approach, it will
be discussed in the next section.

4.2.2 Pseudo-Homogeneous Approach

From the results above we can conclude as follows:

37
CHAPTER 4. COPOLYMERIZATION

When using the long chain approximation, the PBE for the CLD of a copolymeriza-
tion process can be reduced to the corresponding ones of a homopolymerization pro-
cess, where the actual polymerization rate constants are replaced by suitable averages
(pseudo-kinetic parameters) of the actual copolymerization rate constants.

These averages are different depending upon the reaction order with respect to the radical
species:

• zero-order processes
RI = RIA + RIB (4.23)

• first-order processes
∗    
kp M = kpAA pA + kpBA pB · A + kpBB pB + kpAB pA · B (4.24)

kfs∗ = kfsA pA + kfsB pB (4.25)

• second-order processes

kt∗ = ktAA p2A + 2ktAB pA pB + ktBB p2B (4.26)

This implies that in order to compute the CLD of a copolymer we can use the same equations
developed above for a homopolymer, by simply replacing the actual homopolymerization con-
stants with the pseudo-homopolymerization ones (function of the actual copolymerization rate
constants). This procedure is referred to as the pseudo-homopolymerization approach. Note
that all rate constants of the pseudo-homopolymerization are functions of the composition of
the monomer phase which in principle changes during the reaction.

4.2.3 Kinetics of Multi-Monomer Polymerization

The current polymer technology requires, for applications of increasing complexity, the pro-
duction of chains containing two or more monomer. As the long chain assumption cannot be
extended to the case where more than two monomers are involved one has to take a different
approach.
Let us consider the case of three monomer polymerization, and write the molar balances of
the radical chains with the same terminal unit and any chain length. Since the cross propaga-
tion reactions dominate these balances we consider the following reactions:

38
CHAPTER 4. COPOLYMERIZATION

kpAB
A• + B −→ B• r = kpAB A• B
kpAC
A• + C −→ C• r = kpAC A•C
kpBA
B• + A −→ A• r = kpBA B• A
kpBC
B• + C −→ C• r = kpBC B•C
kpCA
C• + A −→ A• r = kpCAC • A
kpCB
C• + B −→ B• r = kpCBC • B
The following balances follow:
dA•  
= kpBA B• A + kpCAC • A − kpAB B + kpACC A• (4.27)
dt

dB•  
= kpAB A• B + kpCBC • B − kpBA A + kpBCC B• (4.28)
dt

R• = A• + B• + C • (4.29)
Note that the last equation has been introduced to replace the molar balance of C • which would
have been a linear combination of the first two. We cannot compute the absolute values of A• ,
B• and C • but only their ratios. Thus, introducing the probabilities (or mole fractions):

pA = A• /R• , pB = B• /R• , pC = C • /R•

we get:  
⎧
⎪
⎪
⎪ k Ap + k Ap − k B + k C p =0
⎪
⎨
pBA B pCA C
 pAB pAC
 A
⎪
⎪
⎪ kpAB BpA + kpCB BpC − kpBA A + kpBCC pB = 0 (4.30)
⎪
⎩ p +p +p
A B C =1
which is a linear system with three unknowns. Note that in the case of binary copolymeriza-
tion, the system above reduces to:

kpBA ApB − kpAB BpA = 0
(4.31)
pA + pB =1
which leads to the same solution
kpBA A
pA = (4.32)
kpBA A + kpAB B
obtained earlier using LCA. Which means that LCA is simply a short-cut form of the pseudo
steady state approximation in the case of binary copolymers. This procedure can be general-
ized to any number of monomer species greater or equal than three.

39
CHAPTER 4. COPOLYMERIZATION

Note that in equation (4.27) and (4.28) we have assumed that the cross propagation reactions
are much faster than all terminations. This is generally the case in polymerization systems,
with the only exception of block polymerization.
The pseudo-homopolymerization rate constants are then computed using the same rules
indicated above:

• zero-order average
RI = RIA + RIB + RIC + ... (4.33)

• first-order average
kfs∗ = kfsA pA + kfsB pB + kfsC pC + ... (4.34)

∗
kpA = kpAA pA + kpBA pB + kpCA pC + ... (4.35)

∗
kpB = kpAB pA + kpBB pB + kpCB pC + ... (4.36)

... (4.37)

• second-order average

kt∗ = ktAA p2A + ktBB p2B + ktCC p2C

+2ktAB pA pB + 2ktAC pA pC + 2ktBC pB pC
+... (4.38)

4.3 Chain Composition Distribution of Copolymers

4.3.1 Instantaneous Chain Composition
The mole fraction of A in the differential amounts of polymer produced at a given instant of
time is given by:
RpA
FA = (4.39)
RpA + RpB
where the rate of polymerization of A and B can be computed using equation (4.17):
 
kpAA pA + kpBA pB · A
FA =     (4.40)
kpAA pA + kpBA pB · A + kpBB pB + kpAB pA · B

40
CHAPTER 4. COPOLYMERIZATION

As seen in equation (4.6) and (4.7) the probability of having a radical of type A or B is given
by:
kpBA A
pA =
kpBA A + kpAB B
kpAB B
pB =
kpBA A + kpAB B
If we now introduce the reactivity ratios, which are defined as the ratio between the direct and
the cross propagation rate constant for a given radical,

kpAA kpBB
rA = and rB = (4.41)
kpAB kpBA

we can rewrite equation (4.40) as follows:

(rA A + B) · A
FA =
(rA A + B) · A + (rB B + A) · B
(rA XA + XB ) · XA
= (4.42)
(rA XA + XB ) · XA + (rB XB + XA ) · XB

Equation (4.42) is the so-called Mayo-Lewis equation which contains only two kinetic pa-
rameters, rA and rB . In order to understand the effect of the reactivity ratios on the Mayo-Lewis
equation we will derive the expression for dFA /dXA . Therefore we rewrite equation (4.42) us-
ing the expression XB = 1 − XA

(rA − 1)XA2 + XA
FA = (4.43)
(rA − 2)XA2 + 2XA + rB (1 − XA )2

and then differentiate with respect to XA

 
dFA 2(rA − 1)XA + 1 den − nom 2(rA − 2)XA + 2 − 2rB (1 − XA )
= (4.44)
dXA den2

where den and nom are the denominator and nominator of equation (4.43), respectively. Now
we will consider the limits of dFA /dXA for XA → 0 and XA → 1:

dFA  1
X →0 = (4.45)
dXA A rB
dFA  1
X →1 = (4.46)
dXA A rA

In the following we want to distinguish three different cases:

41
CHAPTER 4. COPOLYMERIZATION

   
dFA  dFA 
I (rA > 1) ∧ (rB < 1) :  >1 ∧  <1
dXA XA →0 dXA XA →1
both radicals ”prefer” to propagate with A
(A is more reactive)
   
dFA  dFA 
II (rA < 1) ∧ (rB > 1) :  <1 ∧  >1
dXA XA →0 dXA XA →1
both radicals ”prefer” to propagate with B
(B is more reactive)
   
dFA  dFA 
III (rA < 1) ∧ (rB < 1) :  >1 ∧  >1
dXA XA →0 dXA XA →1
both radicals ”prefer” to propagate with
the other monomer
The Mayo-Lewis plot represents the instantaneous composition of a polymer, FA produced
from a given monomer phase composition, XA and it can be represented by the following plot.

FA III

II

0 XA 1

4.3.2 Composition Drift

Let us analyze the behavior of the chain composition distribution in a batch reactor using the
Mayo-Lewis diagram. Suppose that we charge initially in the reactor a monomer mixture
where the molar fraction of A is XA0 . As soon as the reaction starts we produce polymer
chains with mole fraction of A equal to F A . However, since F A > XA0 the monomer phase

42
CHAPTER 4. COPOLYMERIZATION

becomes poorer of A, and XA decreases below the initial value XA0 . With the progress of the
polymerization the XA value decreases further, leading to the production of polymer chains
continuously poorer in A, as indicated by the arrow in the plot.

FA

XoA

0 XoA 1

Note that the composition drift produces substantial variations in the chain compositions. This
can be appreciated by noting that complete monomer conversion, the average composition of
all chains has to be equal to the initial monomer composition.

4.3.3 Chain Composition Control

In order to avoid the composition drift we have to use a semibatch reactor, where we introduce
a flow rate Ṅ of the two monomers in order to keep constant the monomer phase mole fraction,
XA and consequently also the composition of the produced polymer chains.

NA NB

For this we consider the mass balances in a semibatch reactor where ṄA and ṄB are the molar

43
CHAPTER 4. COPOLYMERIZATION

flow rates of the two monomers:

dNA
= ṄA − RpA V (4.47)
dt

dNB
= ṄB − RpB V (4.48)
dt
where Ni represents the moles of the i-th monomer in the reactor, V is the volume of the
∗
reactor content, and RpA and RpB the monomer polymerization rate given by RpA = kpA AR•
∗
and RpB = kpB BR• , respectively, so that:

dNA ∗
= ṄA − kpA R• NA (4.49)
dt

dNB ∗
= ṄB − kpB R• NB (4.50)
dt
where NA = AV and NB = BV. Let us now introduce the dimensionless time τ = t/tf , where tf
is the duration of the process. So that we get:
dNA  
∗
= tf ṄA − kpA R• NA (4.51)
dτ

dNB  
∗
= tf ṄB − kpB R• NB (4.52)
dτ
with the initial conditions NA (0) = NA0 and NB (0) = NB0 . Assuming that the pseudo kinetic
propagation rate constants do not change during the process we can integrate the equations
above to obtain: ⎛ ⎞
⎜⎜⎜ 0 Ṅ ⎟⎟   ṄA
NA = ⎜⎜⎝NA − ∗ • ⎟⎟⎟⎠ exp −kpA ∗
R• tf τ + ∗ •
A
(4.53)
kpA R kpA R
⎛ ⎞
⎜⎜⎜ 0 Ṅ ⎟⎟   ṄB
NB = ⎜⎜⎝NB − ∗ • ⎟⎟⎟⎠ exp −kpB ∗
R• tf τ + ∗ •
B
(4.54)
kpB R kpB R
The question is now how to operate the reactor in order to keep the monomer phase composi-
tion constant, i.e., NA /NB = const. There are two possible operation modes.

Starved Operation

This operation modus, which is frequent in reaction engineering, is based on the observation
that the reactor under consideration is a self-regulated system, in the sense that it admits only

44
CHAPTER 4. COPOLYMERIZATION

one stable pseudo-steady state. Accordingly, as τ → ∞ the system autonomously approaches

such a steady state, defined as follows:

∗
ṄA = kpA R• NApss = RpA V (4.55)

∗
ṄB = kpB R• NBpss = RpB V (4.56)

This means that the system adjusts itself in order to satisfy the above pseudo-steady state
condition, which means that the monomer amounts inside the reactor change from the initial
values NA0 and NB0 , to the values NApss and NBpss defined by the equations above which are such
that the rate of of consumption of each monomer by the polymerization process is equal to
the rate of monomer feed to the reactor which we can of course control. Accordingly, in order
to produce a copolymer with composition F A , in these conditions, we simply need to take
 
ṄA /(ṄA + ṄB ) = F A constant in time so as to have RpA / RpA + RpB = F A and than produce
a copolymer with instantaneous composition equal to the desired value. Note that the values
 
NApss and NBpss are such that the monomer mole fraction XApss = NApss / NApss + NBpss is equal to the
value that according to the Mayo-Lewis plot leads to the desired copolymer composition, F A .
This is done autonomously by the system, we do not need to know this plot for the particular
copolymer under examination.
The problem is that this condition is satisfied only at the pseudo-steady state, i.e., at τ → ∞,
while during the transient period the copolymer composition is not equal to the desired value.
However, by looking at the exponential functions in equations (4.53) and (4.54), we see that
the pseudo-steady state condition is achieved at the time

1 1
τ= ∗ and ∗ (4.57)
kpA R• tf kpB R• tf

for the two monomers. Thus, if we take a very long process time, i.e.,

1 1
tf  ∗ or ∗ (4.58)
kpA R• kpB R•

we have that the pseudo-steady state conditions are achieved after a very short fraction of the
process time, i.e., τ 1, so that the amount of polymer produced in the transient condition
is small, and we can actually make it negligible by increasing tf at will. The starved equation
made is then based on the second condition, that is the feed time should be sufficiently long,

45
CHAPTER 4. COPOLYMERIZATION

in particular such as to satisfy the condition above. A more intuitive explanation of the mech-
anism of this operation mode can be obtained by looking directly at the mass balances (4.51)
and (4.52). By considering the first one in the form
dNA  
∗
= tf kpA R• NApss − NA (4.59)
dτ
 
∗
we see that if we have NA < NApss = ṄA / kpA R• then dNA /dτ > 0 and NA increases, while if
NA > NApss then dNA /dτ < 0 and NA decreases. This means that NA is always attracted towards
the pseudo-steady state value NApss (auto regulated system) with a ”speed” which is higher the
∗
larger is the value of tf kpA R• . A similar argument holds true for B.

qout

A simple example of an auto regulated system, which admits only one stable steady state is
that of a tank with a hole, where feed the volumetric flow rate, Q of water. The mass balance
leads to:
dM
A = Q − qout (4.60)
dt
where M is the level of the liquid, A the cross section and the outlet flow rate is given by

Bernoulli equation as qout = Ah 2gM, where Ah is the cross section of the hole. The mass
balance above can be rewritten as follows:
dM Q Ah 
= − 2gM (4.61)
dt A A
Introducing the steady state value
√ Q
M ss =  (4.62)
Ah 2g
this reduces to: 
dM Ah 2g  √ ss √ 
= M − M (4.63)
dt A

46
CHAPTER 4. COPOLYMERIZATION

which indicates that whatever is the initial level in the tank, this changes in order to reach the
steady state value, which is the one that makes qout = Q. This is in perfect analogy with the
starved reactor, which is however a semibatch and not a continuous system, and therefore does
not admit a real steady state.
   
∗
Note that by taking the feed time tf  1/ kpA R• or 1/ kpB ∗
R• we are imposing to the
system a dynamics much slower than its intrinsic one, which is the dynamics of monomer
   
∗
consumption (whose characteristic times are 1/ kpA R• and 1/ kpB
∗
R• for the two monomers).
Accordingly, we do not allow the monomer accumulation in the reactor, and this is the reason
why this operation modus is referred to as starved. The time evolutions of conversion (solid
line) and amount of monomer introduced in the reactor (dashed line) are shown in the figures
below.
Batch Semibatch

100% 100%

time time

In other words we are forcing the reactor to produce much less than its potential in order to
have a narrow chain composition distribution.

Power Feed Operation

In this operation modus we try to realize two objectives simultaneously:

• narrow chain composition distribution

• maximum productivity of the reactor

We can consider that we have two scalar and two reactor optimization variables, that is the
amounts of the two monomers initially introduced in the reactor, NA0 and NB0 , and the two flow
rates as a function of time, i.e., ṄA (t) and ṄB (t) for t > 0. We need to define them in order to
satisfy the two objectives above. For the first objective function, i.e., chain composition equal
to F A , we can first impose to have initially in the reactor the monomer phase composition,

47
CHAPTER 4. COPOLYMERIZATION

X A which leads to the desired polymer composition, F A . For this we need first to know the
Mayo-Lewis relation to compute X A and then we impose:

NA0
= XA (4.64)
NA0 + NB0

Next we operate on the flow rates ṄA and ṄB in order to keep the monomer phase composition
constant during the entire process, that is:
d (NA /NB )
=0 , for t > 0 (4.65)
dt
or:
dNA dNB
NB = NA , for t > 0 (4.66)
dt dt
which from the mass balances (4.49) and (4.50) leads to:
   
∗
NB ṄA − kpA R• NA = NA ṄB − kpB
∗
R • NB , for t > 0 (4.67)

This constraint allows keeping constant the composition of the produced polymer. A third
constraint is on the total weight of the polymer produced in one batch, WP which is imposed
by the volume of the reactor:
     
NA +
0
ṄA dt MA + NB +
m 0
ṄB dt MBm = WP (4.68)

We can now use the last degree of freedom to reach the second objective, that is maximum
polymerization rate at any time during the process. This is a complex problem of demanding
calculations whose simple solution is that the less reactive monomer (B in our case) should all
be introduced at the beginning of the process, that is

ṄB = 0 , for t > 0 (4.69)

which introduced in equation (4.67)leads to:

 
∗ ∗
ṄA = kpA − kpB R• NA , for t > 0 (4.70)

which indicates that we have to add to the reactor the amount of A needed to equilibrate the
change in composition induced by its higher reactivity with respect to B.
Thus, summarizing, if we need to produce a total amount of polymer WP with uniform chain
w
composition F A , we proceed as follows. Using the corresponding weight fractions F A and

48
CHAPTER 4. COPOLYMERIZATION

w
F B , we compute the total amounts (weights) of the two monomers needed WA and WB . From
Mayo-Lewis relation we compute the monomer phase mole fraction X A which corresponds to
F A . From this we compute the total moles of the two monomers to be initially introduced in
the reactor:
WB
NB0 = (4.71)
MBm

NA0
= XA (4.72)
NA0 + NB0
The remaining amount of A, given by WA /MAm − NA0 , is fed during the process according to the
ṄA value given by equation (4.70).
This operation modus is economically more convenient than the starved operation but it
requires to know the kinetics of the process (cf. Mayo-Lewis) and to be able to implement
the condition (4.65) which requires either to measure on-line or to model the kinetic of the
process. An additional problem which often arises with this procedure is that the very high
productivity leads to a very large heat production which may be difficult to handle with a
given heat exchange area. If the reactor cooling capacity is insufficient and fixed, one should
reconsider the above optimization process and substitute the maximum productivity objective
with the condition that the heat production rate should never exceed the potential heat cooling
of the reactor, so as to be able to maintain isothermal operating conditions. This would be of
the following form:
RpA (−ΔHA ) + RpB (−ΔHB ) ≤ US (T − T c ) (4.73)

where ΔHi is the polymerization enthalpy of the i-th monomer, U and S are the reactor heat
transfer coefficient and surface, respectively and T c is the temperature of the cooling medium.

4.3.4 Reactor Monitoring and Control

The composition control problem can be solved with a closed-loop controller, provided that we
have a reliable on-line sensor for the monomer phase composition. The measured composition
 
could be compared to the set-point X A and then a proper action taken on the manipulated
variables ṄA and ṄB . Such a sensor could be a gas chromatograph for the liquid phase or for
the phase in the head space. The second is probably faster and more robust, but definitely not
sufficiently for its routine use in a production environment. More easy is to measure on-line
the global rate of polymerization in the reactor. For this we can use the reactor heat balance

49
CHAPTER 4. COPOLYMERIZATION

as follows:
 
(−ΔH) Rp V = Ṅc cpc T cout − T cin (4.74)

where Ṅc and cpc are the flow rate and the heat capacity of the cooling medium, and ΔH is
the global reaction enthalpy. From this equation, by measuring on-line the outlet and inlet
temperatures of the cooling medium, we can compute the instantaneous global rate of poly-
merization. Clearly, we have no way to distinguish the individual contribution of the two
monomers from this energy balance. The above equation is approximate and in applications
we should account for heat losses and for the heat accumulations. In addition, it is generally
difficult to use a reactor as a calorimeter. On one hand in fact we need T cout and T cin to be very
close so as to have a uniform cooling of the reactor. On the other hand, we need to have an
appreciable difference of T cout and T cin in order to avoid an excessive propagation of the tem-
perature measurement error in the estimated value of the global rate of polymerization. Other
equivalent sensors are used in applications which estimate the rate of reaction from density
measurements which are obtained by measuring the ultrasound propagation velocity in the re-
acting system. The complementary information needed to compute the rate of polymerization
of the two individual monomers can be obtained using a kinetic model. The rate of monomer
consumption is in fact given by:

dNA ∗
= −RpA = −kpA R• AV (4.75)
dt
In practice, the difficulty in using this equation is the evaluation of R• which is the source of
many irreproducibilities, due to unpredictable impurities, inhibitors or others. On the other
hand, the polymer composition
∗
RpA kpA A
FA = = ∗ ∗ (4.76)
RpA + RpB kpA A + kpB B

50
CHAPTER 4. COPOLYMERIZATION

does not depend on R• and therefore can be evaluated a priori with much higher reliability. The
best way to monitor the reactor behavior is therefore to combine a sensor for measuring on-line
the global polymerization rate with the equation above to split the total monomer consumption
into the contributions of the two monomers.

51