Macromolecules 2005, 38, 5485-5492 5485

Theoretical Expression of the Average Activation-Deactivation

Equilibrium Constant in Controlled/Living Free-Radical
Copolymerization Operating via Reversible Termination. Application to
a Strongly Improved Control in Nitroxide-Mediated Polymerization of
Methyl Methacrylate

Bernadette Charleux,*,† Julien Nicolas,† and Olivier Guerret‡

Laboratoire de Chimie des Polymères, UMR 7610 associée au CNRS, Université Pierre et Marie Curie,
T44 E1, 4 Place Jussieu, 75252 Paris Cedex 05, France, and ARKEMA, Groupement de Recherches de
Lacq, B.P. No. 34, 64170 LACQ, France
Received January 14, 2005; Revised Manuscript Received April 25, 2005

ABSTRACT: The average activation-deactivation equilibrium constant, 〈K〉, was determined on a

theoretical basis for controlled free-radical copolymerizations operating via a reversible termination
mechanism (i.e., nitroxide-mediated polymerization or atom transfer radical polymerization), using the
terminal model for the activation-deactivation equilibrium and the terminal model or the implicit
penultimate unit effect model for the propagation reaction. From the equation, it was shown that the
addition of a small fraction of an appropriate comonomer to a monomer with a very large activation-
deactivation equilibrium constant, K, might lead to strong reduction of 〈K〉, providing the added comonomer
exhibits a low K. In nitroxide-mediated polymerization, the monomers with a very high K, such as the
methacrylic esters, do not lead to controlled polymerization in the presence of nitroxides like SG1, despite
the absence of disproportionation reaction between the nitroxide and the growing radical, because of the
too fast irreversible self-termination of the propagating radicals present in high concentration. The
polymerization stops at low conversion. Consequently, a reduction of K might lead to an enhanced quality
of control. The method was indeed successfully applied to the SG1-mediated polymerization of methyl
methacrylate at 90 °C. By adding only 4.4 or 8.8 mol % of styrene, the polymerization could be carried
out to large conversions, while exhibiting all the features of a controlled system.

Introduction roradicals do not strongly differ from those measured

in the analogous conventional free-radical copolymeri-
Controlled/living free radical polymerization (CRP)1
zation, meaning that the composition equation is the
via a reversible termination mechanism consists of two
same.5,10,16 However, some deviation can be found,
main techniques, i.e., nitroxide-mediated polymerization
especially when the traditional methods to measure the
(NMP)2 and atom transfer radical polymerization
(ATRP).3,4 Both are based on an activation-deactivation reactivity ratios are applied, i.e., analysis of the copoly-
equilibrium between dominant dormant species and a mer composition at low monomer conversion for short
small fraction of propagating macroradicals. The in- chains. This is explained by the inadequacy of the
volved mechanisms are now well established for both general assumptions made in the Mayo-Lewis ap-
methods, and the kinetic and thermodynamic param- proach, which are only valid for long-chain copolymer-
eters have been determined for the homopolymerization izations. Only the methods using integrated composition
of the most common monomers. NMP and ATRP have equations, such as the Skeist equation,17 are applicable.
also been applied to the synthesis of statistical copoly- The influence of the short-chain radicals reactivity,
mers, revealing a composition gradient when the comono- which differs from that of the long-chain macroradicals,
mers exhibit a noticeable difference in reactivity.2-11 has been first discussed.11,18 In addition, establishment
However, to our knowledge, despite the opportunity of of both the activation-deactivation equilibrium and the
creating new architectures, the kinetics and thermody- cross-propagation equilibrium (steady-state assumption
namics of controlled/living statistical copolymerization for the concentration of all propagating macroradicals)
have not been so largely investigated as those of might require a certain time, i.e., a certain monomer
homopolymerization.5,10,12-15 In a first approach, the conversion. As proposed by Matyjaszewski for ATRP
reactivity ratios have been studied to compare their systems,19 the deviation is better explained by a slow
values with those of a conventional radical copolymer- establishment of the cross-propagation equilibrium. The
ization. Similar values are actually expected since they same conclusion was also illustrated by Klumperman
correspond to the intrinsic reactivity of the involved et al.18 in the simulated atom transfer radical copoly-
macroradicals and monomers. It is indeed generally merization of styrene and n-butyl acrylate, where 15%
shown that the reactivity ratios determined in a con- overall monomer conversion was needed to reach the
trolled/living radical copolymerization proceeding via an steady-state concentrations in propagating radicals.
activation-deactivation mechanism of the growing mac- When the cross-propagation rates are equal, copolymers
from both conventional and living free-radical copoly-
†Université Pierre et Marie Curie.
merizations should exhibit the same instantaneous
‡ARKEMA. composition. In contrast, even though the ratio of the
* Corresponding author: e-mail charleux@[Link]. macroradicals concentration does not differ for both
10.1021/ma050087e CCC: $30.25 © 2005 American Chemical Society
Published on Web 06/01/2005
5486 Charleux et al. Macromolecules, Vol. 38, No. 13, 2005

methods, there is no reason to have the same overall The overall concentration of propagating radicals is then
radicals concentration and hence the same copolymer- given by eq 4, and combination with eq 1 allows to write
ization rate.5 This last point has not been thoroughly the average equilibrium constant as in eq 5.
discussed in the literature and should be addressed.
In this article, we present a theoretical approach to [A-N] [B-N]
describe the copolymerization kinetics in controlled/ [P•] ) KA + KB (4)
[N] [N]
living copolymerization operating via reversible termi-
nation. Both the average activation-deactivation equi- [A-N] [B-N]
librium constant 〈K〉 and the product with the average 〈K〉 ) KA + KB (5)
C0 C0
rate constant of propagation, 〈kp〉〈K〉, are determined on
a fully theoretical basis. The model is developed for
nitroxide-mediated polymerization but might be easily The rate of copolymerization, Rp, depends on the
extended to ATRP. The terminal model (TM) and the overall concentration of propagating macroradicals ac-
implicit penultimate unit effect model (IPUE) have been cording to the following eq 6, where conv represents the
considered for the propagation reaction. In contrast, the overall molar conversion.
activation-deactivation equilibrium constant was sup-
d([A] + [B])
posed to depend on the terminal monomer unit only. Rp ) - ) 〈kp〉[P•]([A] + [B]) (6)
This last point has been discussed by Heuts et al.,13 who dt
concluded that such an assumption was quite reason-
able, even though not experimentally proven. In their d[ln(1/(1 - conv)] C0
) 〈kp〉[P•] ) 〈kp〉〈K〉 (7)
work, these authors13 had already proposed an equation dt [N]
for 〈K〉 in an ATRP system as a function of the individual
equilibrium constants and of the ratio of the propagating The slope of the usual ln[1/(1 - conv)] vs time plot is a
radicals concentration. Here, we go a step further in direct function of 〈kp〉〈K〉. Like in homopolymerization,
presenting 〈K〉 as a simple function of the comonomer it also depends on the concentrations of both the
feed ratio and of known kinetic and thermodynamic alkoxyamine initiator and the nitroxide. The latter
parameters, for both the terminal model and the IPUE continuously increases throughout the polymerization
model. owing to the so-called persistent radical effect, resulting
The variation of 〈K〉 and 〈kp〉〈K〉 with the comonomer from the irreversible self-termination of the propagating
composition is discussed and further applied to the macroradicals.20 However, when a sufficiently high
nitroxide-mediated copolymerization of methyl meth- concentration of free nitroxide is initially introduced,
acrylate and styrene, where it is shown that a small the concentration [N] might remain close to the initial
fraction of styrene can have a remarkable influence. one; i.e., the kinetics depends in that case on the initial
This experimental application will be presented in the C0/[N]0 ratio.21
second part of the article. 1. Terminal Model for Propagation. Only the two
chain ends A• and B• are considered. When it applies,
Theoretical Approach of the Average the cross-propagation equilibrium allows to write
Activation-Deactivation Equilibrium Constant
We consider two monomers A and B. The free nitrox- kBA[B•][A] ) kAB[A•][B] (8)
ide is noted N, and the macromolecular alkoxyamines
are respectively A-N and B-N, were A and B represent where kBA and kAB are the cross-propagation rate
the last monomer unit. constants.
The average activation-deactivation equilibrium con- With the reactivity ratios rA ) kAA/kAB and rB ) kBB/
stant reads kBA (kAA and kBB are the homopropagation rate con-
stants, also noted kp,A and kp,B), this relationship gives
[P•][N] [P•][N]
〈K〉 ) ) (1) [A•] rA kp,B [A] rA kp,B fA
[P-N] C0 ) ) (9)
[B•] rB kp,A [B] rB kp,A fB
where [P•]) [A•]+ [B•] is the overall concentration of
propagating radicals, with [A•] the concentration of where fA and fB are the molar fraction of A and B,
A-terminated macroradicals and [B•] the concentration respectively, in the comonomer mixture.
of B-terminated macroradicals. [P-N] is the overall From the activation-deactivation equilibria (eqs 2
alkoxyamine concentration, which is considered to and 3), the ratio of the propagating macroradicals
match the initial alkoxyamine concentration, C0 ) concentrations also reads as in eq 10, which leads to
[P-N] ) [A-N] + [B-N]. the ratio of the dormant species concentrations (eq 11),
The activation-deactivation equilibrium constant, KA, when combined with eq 9.
in the N-mediated polymerization of A, is related to [A•]
and to the concentration of A-terminated dormant [A•] KA [A-N]
chains [A-N] according to eq 2; the same relationship •
) (10)
[B ] KB [B-N]
also holds for KB (eq 3).
[A-N] KB rA kp,B fA
[A•][N] ) (11)
KA ) (2) [B-N] KA r B kp,A fB
[A-N]
With C0 ) [A-N] + [B-N], the proportion of each
[B•][N]
KB ) (3) individual type of dormant chain can be calculated as
[B-N] in eqs 12 and 13.
Macromolecules, Vol. 38, No. 13, 2005 Polymerization of Methyl Methacrylate 5487

[A-N] KBrAkp,BfA When they apply, the cross-propagation equilibria

) (12) allow to write the concentration ratios given in eqs 18-
C0 KAr Bkp,AfB + KBrAkp,BfA
20.
[B-N] KArBkp,AfB
) (13) [AA•] rAfA
C0 KAr Bkp,AfB + KBrAkp,BfA ) (18)
•
[AA ] + [BA ] • rAfA + fB/sA
The average activation-deactivation equilibrium con-
[BB•] rBfB
stant of eq 5 is then given by the following relationships ) (19)
(eqs 14 and 15). •
[BB ] + [AB ] • rBfB + fA/sB

〈K〉 )
KAKBrAkp,BfA + KBKArBkp,AfB
KAr Bkp,AfB + KBrAkp,BfA
(14)
[BA•]
[AB ] •
) ( )(
fA + rBfB sBkBBB/rB
rAfA + fB sAkAAA/rA ) (20)

rAfA rBfB The activation-deactivation equilibria in eqs 2 and 3

+ give relationship 21, which can also be written as in eq
kp,A kp,B
〈K〉 ) (15) 22.
rAfA rB f B
+
kp,AKA kp,BKB [A•] KA[A-N]
•
) (21)
[B ] KB[B-N]
It appears that 〈K〉 not only depends on the individual
equilibrium constants and the comonomer composition [A•] [AA•] + [BA•]
but also is affected by the cross-propagation rate ) (22)
constants (actually represented by kp/reactivity ratio). [B•] [BB•] + [AB•]
With 〈kp〉 given by the terminal model (eq 16),22-24 it
After combination with the previous eqs 18-20, the
is possible to calculate the 〈kp〉〈K〉 product as follows:
radical concentration ratio can be expressed by the eq
23.
rAfA2 + 2fAfB + rBfB2

( )( )
〈kp〉 ) (16) [A•] rAfA + fB/sA rBfB + fA fA kp,B rA
fA fB
rA + rB ) (23)
kp,A kp,B
•
[B ] rAfA + fB rBfB + fA/sB fB kp,A rB

With the usual notation given below

rAfA2 + 2fAfB + rBfB2
〈kp〉〈K〉 ) (17)
rAfA rB f B rAfA + fB
+ kp,A ) kp,A (24)
kp,AKA kp,BKB rAfA + fB/sA

2. Implicit Penultimate Unit Effect for Propaga- rBfB + fA

kp,B ) kp,B (25)
tion. We consider here the IPUE for the propagation rBfB + fA/sB
and cross-propagation reactions. This leads to the
existence of four different types of propagating macro- one can simplify the eq 23 in the following manner (eq
radicals noted AA•, AB•, BB•, and BA• and of four 26).
reactivity ratios:5,23,24
[A•] rA kp,B fA
kAAA
•
) (26)
two monomer reactivity ratios: rA ) ) [B ] r B k fB
kAAB p,A

kBAA kABB kBBB In combining eq 21 with eq 26 and considering C0 )

and rB ) )
kBAB kABA kBBA [A-N] + [B-N], the ratios [A-N]/C0 and [B-N]/C0 can
be determined.
kBAA
two radical reactivity ratios: sA ) and sB ) KBrAkp,BfA
kAAA [A-N]
) (27)
kABB C0 KAr Bkp,AfB + KBrA‚kp,BfA
kBBB
[B-N] KArBkp,AfB
where kAAA ) kp,A and kBBB ) kp,B are the homopropa- ) (28)
C0 KAr Bkp,AfB + KBrAkp,BfA
gation rate constants, and the other rate constants
correspond to the different propagation reactions.
In contrast, like before, we consider only two activa- The average activation-deactivation equilibrium con-
tion-deactivation equilibrium constants, KA and KB; i.e., stant is then given by the following relationships.
we assume the absence of the penultimate unit effect
for K. Therefore, all the equations from (1) to (7) remain KAKBrAkp,BfA + KBKArBkp,AfB
the same, with [A•] ) [AA•] + [BA•] and [B•] ) [BB•] + 〈K〉 ) (29)
[AB•]. KAkp,Ar BfB + KBkp,BrAfA
5488 Charleux et al. Macromolecules, Vol. 38, No. 13, 2005

rAfA rBfB the nitroxide. It thus prevents the polymerization to

+ reach large conversions and to produce nitroxide-
kp,A kp,B terminated polymers.25-28 Only copolymerizations with
〈K〉 ) (30) large amounts of styrene allowed to form copolymers
rAfA rB f B
+ with controlled molar mass and narrow molar mass
kp,AKA kp,BKB distribution.29-31 The polydispersity index, however,
significantly increased when the proportion of methyl
〈kp〉 from the IPUE model has already been determin- methacrylate was increased.
ed22-24 (eq 31). It is then possible to give the 〈kp〉〈K〉 In contrast to TEMPO, the second generation of
product, which is an experimentally accessible param- alicyclic nitroxides permits the control of a much
eter from the copolymerization kinetics (eq 32). broader range of monomers, such as the acrylates and
substituted acrylamides.32-34 This nitroxide family in-
rAfA2 + 2fAfB + rBfB2 cludes the N-tert-butyl-N-(1-diethylphosphono-2,2-di-
〈kp〉 ) (31)
fA fB methylpropyl) nitroxide, also called SG1 or DEPN,32,33
rA + rB and the 2,2,5-trimethyl-4-phenyl-3-azahexane 3-nitrox-
kp,A kp,B ide (sometimes called TIPNO).34 Nevertheless, they still
do not allow a good control in the polymerization of
rAfA2 + 2fAfB + rBfB2 methacrylic esters. The interesting point is that the
〈kp〉〈K〉 ) (32) disproportionation reaction between the nitroxide and
rAfA rB f B
+ the growing radical is absent.35 The homopolymerization
kp,AKA kp,BKB of methyl methacrylate remains, however, uncontrolled
because of a too large activation-deactivation equilib-
In ATRP, the equations for 〈K〉 and 〈kp〉〈K〉 (eqs 15, 17, rium constant35 that leads to a high concentration of
30, and 32) would be exactly the same. The only propagating macroradicals and hence favors their ir-
difference is in the way of writing the activation- reversible self-termination. Consequently, a complete
deactivation equilibrium constants of eqs 2 and 3, where disappearance of the active centers can occur within a
actually [N] should be replaced by [deactivator]/[activa- very short time, and only very low limit conversions can
tor]. be achieved. As shown by Benoit et al.,34 the copoly-
3. Discussion about 〈K〉 and 〈kp〉〈K〉. The IPUE merization of methyl methacrylate (MMA) with styrene
model better describes the copolymerization kinetics (S) in the presence of TIPNO is, however, much better
than the terminal model, whereas the latter allows only controlled than in the presence of TEMPO; this was
a good description of the copolymer composition.22-24 demonstrated by the low polydispersity indexes ob-
Consequently, 〈K〉 and 〈kp〉〈K〉 are also better expressed served for the copolymers, until 85 mol % of methyl
with the IPUE model (eqs 30 and 32). With the recent methacrylate. The authors, however, did not provide any
development of the pulsed-laser polymerization to de- kinetic data (conversion vs time and average molar mass
termine the absolute rate constants of (co)propagation, vs conversion) allowing to draw conclusion about the
the r and s reactivity ratios have been now established applicability of the controlled polymerization over a wide
for several comonomer couples,24 which thus allows conversion range.
quite a good prediction of 〈K〉 and 〈kp〉〈K〉. Only when 2. Results and Discussion. The conclusion of those
the s values are not known can the TM equations be various seminal works concerning the polymerization
applied, to get an approximate trend. of methacrylic esters in the presence of an alicyclic
The equation giving 〈K〉 exhibits interesting fea- nitroxide such as SG1 is that the best way to reach a
tures: it depends not only on the individual activation- high level of control would be to significantly reduce the
deactivation equilibrium constants but also on the activation-deactivation equilibrium constant. This would
reactivity ratios and the homopropagation rate con- have the consequence of reducing the instantaneous
stants (i.e., on the cross-propagation rate constants concentration of propagating radicals and hence of
expressed by the ratio of kp over the corresponding decreasing the contribution of the irreversible termina-
monomer reactivity ratio). Additionally, 〈K〉 is a function tion. A possibility, purely based on the kinetic behavior,
of the comonomer mixture composition. Consequently, is to copolymerize methyl methacrylate with a small
〈K〉 changes with conversion throughout the copoly- amount of an appropriate comonomer, in the presence
merization, in agreement with the continuous change of SG1. According to the eqs 15 and 30 giving 〈K〉, such
in fA and fB. It appears also that a comonomer with a comonomer should exhibit a much smaller activation-
very high K, along with a high kp, would not have a deactivation equilibrium constant than MMA along with
great influence on 〈K〉, over a broad composition range. a low cross-propagation rate constant (kp/reactivity
This is, in particular, the point we want to illustrate in ratio). The latter point is important because a slow
the following example. cross-propagation of the comonomer-based macroradical
will ensure a longer lifetime, which is favorable to its
Experimental Application to the efficient deactivation by the free nitroxide. In addition,
Nitroxide-Mediated Polymerization of Methyl the comonomer should also present a good control in
Methacrylate the SG1-mediated polymerization; i.e., K should not be
1. Current Situation. Nitroxide-mediated polymer- too low. Styrene actually fulfills all the criteria, in
ization does not allow to control the family of meth- particular at the optimal temperature of 90 °C, as
acrylic esters.25-28 In the case of TEMPO (2,2,6,6- illustrated by the kinetic and thermodynamic param-
tetramethylpiperidinyl-1-oxy), the disproportionation eters given in Table 1. Unfortunately, the reactivity
reaction between the nitroxide and the growing radical, ratios are not available at 90 °C, and we considered here
which yields an alkene chain-end along with a hydroxy- the data given at 57.2 °C in the literature,36 assuming
lamine, dominates over the reversible combination with little temperature effect. Figure 1 gives the variation
Macromolecules, Vol. 38, No. 13, 2005 Polymerization of Methyl Methacrylate 5489

Table 1. Kinetic Paramaters for the SG1-Mediated Polymerization of Styrene (S) and Methyl Methacrylate (M)
kinetic parameter value ref
mol-1
kp,S at 90 °C (L s-1) propagation rate constant of styrene 900 24
kp,M at 90 °C (L mol-1 s-1) propagation rate constant of methyl methacrylate 1640 24
rS (av value in the 20-60 °C range) monomer reactivity ratio for styrene 0.4890 36
rM (av value in the 20-60 °C range) monomer reactivity ratio for methyl methacrylate 0.4929 36
sS at 57.2 °C radical reactivity ratio for styrene 0.3615 36
sM at 57.2 °C radical reactivity ratio for methyl methacrylate 0.6014 36
KS at 90 °C (mol L-1) activation-deactivation equilibrium constant for the SG1-mediated 4 × 10-10 33
polymerization of styrene
KM at 90 °C (mol L-1) activation-deactivation equilibrium constant for the SG1-mediated 1 × 10-7 35
polymerization of methyl methacrylate

Table 2. Predicted Values of the Average Activation-Deactivation Equilibrium Constant 〈K〉 and of 〈kp〉〈K〉, with 〈kp〉 the
Average Propagation Rate Constant, for the Copolymerization of Methyl Methacrylate and Styrene at 90 °C Based on
the Kinetic Parameters of Table 1, Using Either the Terminal Model (TM) or the Implicit Penultimate Unit Effect Model
(IPUE)

molar fraction of styrene in 〈K〉 (mol L-1) 〈kp〉〈K〉 (s-1)

the comonomer mixture (fS) TM (from eq 15) IPUE (from eq 30) TM (from eq 17) IPUE (from eq 32)
0 1.0 × 10-7 1.0 × 10-7 1.6 × 10-4 1.6 × 10-4
0.025 8.3 × 10-9 3.5 × 10-9 1.4 × 10-5 5.4 × 10-6
0.05 4.4 × 10-9 2.0 × 10-9 7.6 × 10-6 2.9 × 10-6
0.10 2.3 × 10-9 1.2 × 10-9 4.2 × 10-6 1.6 × 10-6
1 4.0 × 10-10 4.0 × 10-10 3.6 × 10-7 3.6 × 10-7

of 〈K〉 in the copolymerization of methyl methacrylate Scheme 1. Structure of the SG1 Nitroxide
Deactivator and of the SG1-Based Alkoxyamine
with increasing proportions of styrene from 0 to 10 mol Initiator Used in This Work
%. Only below 3-4 mol % of styrene is 〈K〉 increasing
very significantly. Above this fraction, the value of 〈K〉
is close to that found for the styrene-SG1 system (Table
2). Consequently, at best, the system might exhibit the
same quality of control as that found in the SG1-
mediated homopolymerization of styrene. Another mono-

mer with a lower K than styrene might be used instead.

Considering only K, n-butyl acrylate would be a good
choice since K ) 2 × 10-10 at 120 °C.33 However, the
very high kp value24,37 does not allow a sufficient
decrease of 〈K〉, which remains above 5 × 10-9 for
n-butyl acrylate contents below 75 mol %. For such
comonomer, the cross-propagation of the n-butyl acryl-
ate-terminated macroradical is much faster than its
deactivation by SG1.
We first intended to check the quality of control in
the copolymerization of MMA with various ratios of
styrene, from 0 to 8.8 mol %, expecting to find a similar
control as in the homopolymerization of styrene for the
largest ratios. To reduce the rate of irreversible termi-
nation from the early polymerization stages, we decided
to introduce 10 mol % of free SG1, based on the initiator,
at the beginning of the polymerization. To optimize the
control, the polymerizations were performed at 90 °C
with an alkoxyamine initiator able to dissociate at a
sufficiently low temperature (owing to the presence of
a tertiary carbon radical attached to the nitroxide; see
Scheme 1 and the experimental part)38 to induce a fast
initiation process. The results are shown in Figure 2
for experiments 1-4 summarized in Table 3. In the
absence of styrene, the homopolymerization of MMA
stopped at low conversion within a very short time,
indicating the prevalence of the irreversible termination.
The formed poly(methyl methacrylate) had a number-
Figure 1. Predicted values of the average activation- average molar mass in rather good agreement with the
deactivation equilibrium constant 〈K〉 and of 〈kp〉〈K〉, with 〈kp〉 theoretical value, but a high polydispersity index (1.8).
the average propagation rate constant, as a function of the With the addition of 2.1 mol % of styrene, the situation
molar fraction of styrene in the comonomer mixture, for the
copolymerization of methyl methacrylate and styrene at 90 °C
significantly improved, but a limit conversion of ≈50%
based on the kinetic parameters of Table 1, using either the was reached within 1 h, indicating the strong decrease
terminal model (TM) or the implicit penultimate unit effect of the propagating radicals. The quality of control was
model (IPUE). also improved, but the polydispersity indexes increased
5490 Charleux et al. Macromolecules, Vol. 38, No. 13, 2005

Figure 3. Size exclusion chromatograms recorded at various

conversions for the bulk copolymerization of methyl methacryl-
ate (8.48 mol L-1) and styrene (0.82 mol L-1; fS0 ) 0.088) at
90 °C, carried out in the presence of 2.73 × 10-2 mol L-1 of
alkoxyamine initiator and 10 mol % of added free SG1 based
on the initiator (experiment 1) (see Table 3).

tal error of the SEC measurements; see Experimental

Part). Such a result would indicate that the added 10
mol % (based on the initiator) of free SG1 was not
sufficient to significantly reduce the early irreversible
terminations. Nevertheless, the polydispersity indexes
continuously decreased throughout the polymerization
to reach values as low as 1.2-1.3. Moreover, the
polymerization rate slightly decreased with the increase
Figure 2. Bulk polymerizations of methyl methacrylate at
90 °C, in the presence of 2.73 × 10-2 mol L-1 of alkoxyamine in the proportion of styrene, in agreement with the
initiator, 10 mol % of added free SG1 based on the initiator, small decrease in 〈kp〉〈K〉 (Table 2). Figure 3 shows the
and various initial molar fractions of styrene (fS0) (see Table SEC chromatograms at increasing conversions for the
3): 9, experiment 1 (fS0 ) 0.088); [, experiment 2 (fS0 ) 0.044); copolymerization with 8.8 mol % of styrene (experiment
4, experiment 3 (fS0 ) 0.021); 0, experiment 4 (fS0 ) 0). (a) 1): the complete shift of the RI traces toward higher
ln[1/(1 - conv)] vs time (conv ) overall conversion); (b) molar masses proves that the polymerization proceeded
number-average molar mass, Mn, and polydispersity index, Mw/
Mn, vs conversion; the full line represents the theoretical Mn, via an activation-deactivation mechanism that allowed
whereas the dotted line is the linear regression for the all the chains to grow simultaneously. The tailing on
experimental data. the low molar mass side, however, might be reasonably
assigned to the contribution of irreversible terminations.
with conversion, beyond 40% conversion. In contrast, It should be mentioned that, based on the reactivity
at and above 4.4 mol % of styrene, the polymerization ratios, at low molar fraction, styrene is consumed faster
exhibited all the features of a well-defined system, that methyl methacrylate (1H NMR gave 4 mol % of
allowing to reach high conversions while keeping a good styrene in the copolymer isolated at 52% conversion for
control over the molar mass and molar mass distribu- the experiment with fS0 ) 0.021; 7 mol % of styrene at
tion. The experimental Mn’s increased with conversion, 58% conversion for fS0 ) 0.044; and 12 mol % of styrene
in good agreement with the predicted values; they were, at 64% conversion for fS0 ) 0.088), leading to a continu-
however, systematically above the theoretical line, ous decrease of fS throughout the copolymerization, and
indicating an incomplete initiation (≈75% initiating hence a continuous increase in 〈K〉. Consequently, the
efficiency; see Figure 2) due to irreversible termination quality of control should decline with conversion, leading
of primary radicals and/or oligoradicals, leading to short to enhanced irreversible termination. Nevertheless, this
dead chains not taken into consideration in the SEC trend should be counterbalanced by a decrease of the
chromatograms (care should however be taken in quan- rate constant of termination with the increase in mac-
tifying the initiating efficiency owing to the experimen- roradical chain length.39

Table 3. Experimental Conditions for the SG1-Mediated Bulk (Co)polymerizations of Methyl Methacrylate (M) and
Styrene (S) at 90 °C
initial molar fraction initial concn of alkoxyamine initial concn of SG1
expt [M] (mol L-1) [S] (mol L-1) of styrene (fS0) (mol L-1) nitroxidea (mol L-1)
1 8.48 0.82 0.088 2.73 × 10-2 2.74 × 10-3
2 8.92 0.41 0.044 2.72 × 10-2 2.75 × 10-3
3 9.14 0.20 0.021 2.73 × 10-2 2.74 × 10-3
4 9.36 0 0 2.73 × 10-2 2.74 × 10-3
5 8.49 0.82 0.088 5.46 × 10-2 5.45 × 10-3
6 8.49 0.82 0.088 1.37 × 10-2 1.36 × 10-3
a [SG1]0/[alkoxyamine initiator]0 ) 0.10; structure of the SG1 nitroxide and of the alkoxyamine is shown in Scheme 1.
Macromolecules, Vol. 38, No. 13, 2005 Polymerization of Methyl Methacrylate 5491

Conclusion
The average activation-deactivation equilibrium con-
stant, 〈K〉, was determined on a theoretical basis for
controlled free-radical copolymerization operating via
a reversible termination mechanism, using the terminal
model for the activation-deactivation equilibrium and
the terminal model or the implicit penultimate unit
effect model for the propagation reaction. The form of
the equation shows that 〈K〉 remains close to the lowest
activation-deactivation equilibrium constant in a co-
polymerization system where a monomer with a very
large K is copolymerized with a monomer with a low K,
even if the latter is added in very small proportion. The
reactivity ratios and the homopropagation rate con-
stants have also a nonnegligible influence (direct effect
of the cross-propagation rate). Consequently, the addi-
tion of a small fraction of an appropriate comonomer
can be performed in order to enhance the control of a
monomer that exhibits a too high K, detrimental to a
good control. The method was indeed successfully ap-
plied to the nitroxide-mediated polymerization of methyl
methacrylate at 90 °C, with SG1 as a nitroxide. By
adding 4.4 or 8.8 mol % of styrene, the polymerization
could be carried out to large conversions, while exhibit-
ing all the features of a controlled system. In this
copolymerization system, the reactivity ratios and the
composition parameter fS are such that the chains
contain isolate styrene subunits. The existence of styrene-
terminated macroradicals favors the deactivation reac-
tion, which would not otherwise efficiently take place
with a methyl methacrylate terminal unit.
A further step of this work is to analyze the chain
Figure 4. Bulk copolymerizations of methyl methacrylate and
styrene (fS0 ) 0.088) at 90 °C in the presence of 10 mol % of
ends (overall alkoxyamine end-functionality and nature
added free SG1 based on the initiator, with various initiator of the terminal monomer unit attached to the SG1) in
concentrations (see Table 3): b, experiment 6 (1.37 × 10-2 mol order to draw conclusions about the true livingness of
L-1); 9, experiment 1 (2.73 × 10-2 mol L-1); 2, experiment 5 the system. Nevertheless, the method allows to build
(5.46 × 10-2 mol L-1). (a) ln[1/(1 - conv)] vs time (conv ) controlled methacrylic ester-rich macromolecular archi-
overall conversion); (b) number-average molar mass, Mn, and tectures via SG1-mediated free-radical polymerization.
polydispersity index, Mw/Mn, vs conversion; the full lines
represent the theoretical Mn’s, whereas the dotted lines are
The important point is that it enables high conversions
the linear regressions for the experimental data. to be reached, which represents a very important
breakthrough for the technique.

Experimental Part
The alkoxyamine initiator concentration was then
1. Materials. Methyl methacrylate (MMA or M, Aldrich,
varied for experiments 1, 5, and 6 with 8.8 mol % of 99%) and styrene (S, Aldrich, 99%) were distilled under
styrene based on the monomers and 10 mol % of free reduced pressure before use. The SG1-based alkoxyamine
SG1 based on the alkoxyamine to target different molar initiator with the 2-(hydroxycarbonyl)prop-2-yl radical38 (Bloc-
masses. In all three experiments, well-defined polymers Builder, Arkema, 99%, see Scheme 1) and the N-tert-butyl-N-
were obtained, with a high level of control (Figure 4): (1-diethylphosphono-2,2-dimethylpropyl) nitroxide (SG1, 86%,
see Scheme 1) were kindly supplied by Arkema.
Mn followed the predicted values, even though they
2. Bulk Polymerizations. In a typical experiment (experi-
remained slightly above, and the polydispersity indexes
ment 1, see Table 3), a mixture of methyl methacrylate (80.0
were as low as 1.3 at large conversions. Unexpectedly, g, 0.80 mol), styrene (8.0 g, 0.77 mol), and SG1 (0.076 g, 0.26
the polymerization rate was different for all three mmol) was deoxygenated by nitrogen bubbling for 20 min at
experiments despite a similar C0/[N]0 initial ratio. This room temperature. The alkoxyamine initiator (0.980 g, 2.57
result is in agreement with the previously mentioned mmol) was added, and nitrogen bubbling was carried out for
indication that the added free SG1 was not in sufficient 10 min. The mixture was then introduced into a 250 mL three-
neck round-bottom flask heated at 90 °C, immersed in a
concentration to avoid the existence of irreversible thermostated oil bath, and equipped with a reflux condenser,
terminations. New free SG1 is released in the early a nitrogen inlet, and a thermometer. The time zero of the
stage of the polymerization, in larger proportion when reaction was triggered when temperature in the reactor
the initial concentration of alkoxyamine is lower, lead- reached 85 °C. Samples of about 5 mL were periodically
ing to a decreased polymerization rate when C0 is withdrawn to follow the monomer conversion by gravimetry;
decreased. This trend is in good agreement with the they were dried in a ventilated oven thermostated at 70 °C
under vacuum, until constant weight. Conditions for the
equations describing the persistent radical effect.21 The various bulk polymerizations are summarized in Table 3.
existence of early termination reactions is also con- 3. Analytical Techniques. Size exclusion chromatography
firmed by the incomplete initiator efficiency (experi- (SEC) was performed at 40 °C with two columns (PSS SDV,
mental Mn’s above the theoretical line; see Figure 4). linear MU, 8 mm × 300 mm; bead diameter: 5 µm; separation
5492 Charleux et al. Macromolecules, Vol. 38, No. 13, 2005

limits: 400-2 × 106 g mol-1). The eluent was tetrahydrofuran (13) Heuts, J. P. A.; Davis, T. P. Macromol. Rapid Commun. 1998,
at a flow rate of 1 mL min-1. A differential refractive index 19, 371-375.
(RI) detector (LDC Analytical refractoMonitor IV) was used, (14) Arehart, S. V.; Matyjaszewski, K. Macromolecules 1999, 32,
and molar mass distributions were derived from a calibration 2221-2231.
curve based on polystyrene (PS) standards from Polymer (15) Zaremski, M. Y.; Plutalova, A. V.; Lachinov, M. B.; Golubev,
V. B. Macromolecules 2000, 33, 4365-4372.
Standards Service. The polystyrene calibration is not ap-
(16) Bisht, H. S.; Ray, S. S.; Pandey, D.; Sharma, C. D.; Chatterjee,
propriate for poly(methyl methacrylate) (PMMA) samples. A. K. J. Polym. Sci., Part A: Polym. Chem. 2002, 40, 1818-
Thus, the Mn values obtained from polystyrene conventional 1830.
calibration were converted using the Mark-Houwink-Sakura- (17) Skeist, I. J. Am. Chem. Soc. 1946, 68, 1781-1784.
da (MHS) relationship between PS and PMMA with MHS (18) Klumperman, B.; Chambard, G.; Brinkhuis, R. H. G. ACS
parameters determined at 35 °C: KPS ) 15.8 × 10-5 dL g-1, Symp. Ser. 2003, 854, 180-192.
RPS ) 0.704, KPMMA ) 12.2 × 10-5 dL g-1, and RPMMA ) 0.69.40 (19) Matyjaszewski, K. Macromolecules 2002, 35, 6773-6781.
In all the figures representing the experimental Mn as a (20) Fischer, H. Chem. Rev. 2001, 101, 3581-3610.
function of monomer conversion, the displayed full line cor- (21) Goto, A.; Fukuda, T. Prog. Polym. Sci. 2004, 29, 329-385.
responds to the theoretical evolution of Mn calculated by (22) Fukuda, T.; Ma, Y.-D.; Inagaki, H. Macromolecules 1985, 18,
17-26.
Mn ) MW(alkoxyamine initiator) + (23) Coote, M. L.; Davis, T. P. Prog. Polym. Sci. 1999, 24, 1217-
1251.
initial mass of monomers (24) Beuermann, S.; Buback, M. Prog. Polym. Sci. 2002, 27, 191-
× conversion
initial mol number of alkoxyamine 254.
(25) Moad, G.; Anderson, A. G.; Ercole, F.; Johnson, C. H. J.;
The final copolymer compositions were analyzed by 1H NMR Krstina, J.; Moad, C. L.; Rizzardo, E.; Spurling, T. H.; Thang,
spectroscopy (250 MHz) in CDCl3 solution, at 25 °C, in 5 mm S. H. ACS Symp. Ser. 1998, 685, 332-360.
tubes, using a AC250 Bruker spectrometer. (26) Burguière, C.; Dourges, M.-A.; Charleux, B.; Vairon, J.-P.
Macromolecules 1999, 32, 3883-3890.
(27) Ananchenko, G. S.; Fischer, H. J. Polym. Sci., Part A: Polym.
Acknowledgment. The authors thank Stéphanie Chem. 2001, 39, 3604-3621.
Magnet and Jean-Luc Couturier from ARKEMA for (28) Souaille, M.; Fischer, H. Macromolecules 2001, 34, 2830-
fruitful discussions and for kindly providing the SG1 2838.
and the alkoxyamine. (29) Hawker, C. J.; Elce, E.; Dao, J.; Volksen, W.; Russell, T. P.;
Barclay, G. G. Macromolecules 1996, 29, 2686-2688.
(30) Butz, S.; Baethge, H.; Schmidt-Naake, G. Macromol. Rapid
References and Notes Commun. 1997, 18, 1049-1055.
(1) (a) Controlled Radical Polymerization; Matyjaszewski, K., (31) Miura, Y.; Nakamura, N.; Taniguchi, I.; Ichikawa, A. Polymer
Ed.; ACS Symp. Ser. 1998, 685. (b) Controlled/Living Radical 2003, 44, 3461-3467.
Polymerization: Progress in ATRP, NMP, and RAFT; Maty- (32) Grimaldi, S.; Finet, J. P.; Le Moigne, F.; Zeghdaoui, A.; Tordo.
jaszewski, K., Ed.; ACS Symp. Ser. 2000, 768. (c) Advances P.; Benoit, D.; Fontanille, M.; Gnanou, Y. Macromolecules
in Controlled/Living Radical Polymerization; Matyjaszewski, 2000, 33, 1141-1147.
K., Ed.; ACS Symp. Ser. 2003, 854. (33) Benoit, D.; Grimaldi, S.; Robin, S.; Finet, J. P.; Tordo, P.;
(2) Hawker, C. J.; Bosman, A. W.; Harth, E. Chem. Rev. 2001, Gnanou,Y. J. Am. Chem. Soc. 2000, 122, 5929-5939.
101, 3661-3688. (34) Benoit, D.; Chaplinski, V.; Braslau, R.; Hawker, C. J. J. Am.
(3) Matyjaszewski, K.; Xia, J. Chem. Rev. 2001, 101, 2921-2990. Chem. Soc. 1999, 121, 3904-3920.
(4) Kamigaito, M.; Ando, T.; Sawamoto, M. Chem. Rev. 2001, 101, (35) Ananchenko, G. S.; Souaille, M.; Fischer, H.; Le Mercier, C.;
3689-3745. Tordo, P. J. Polym. Sci., Part A: Polym. Chem. 2002, 40,
(5) Madruga, E. L. Prog. Polym. Sci. 2002, 27, 1879-1924 and 3264-3283.
references therein. (36) Coote, M. L.; Johnston, L. P. M.; Davis, M. P. Macromolecules
(6) Farcet, C.; Charleux, B.; Pirri, R. Macromol. Symp. 2002, 182, 1997, 30, 8191-8204.
249-260. (37) Asua, J. M.; Beuermann, S.; Buback, M.; Castignolles, P.;
(7) Couvreur, L.; Charleux, B.; Guerret, O.; Magnet, S. Macro- Charleux, B.; Gilbert, R. G.; Hutchinson, R. A.; Leiza, J. R.;
mol. Chem. Phys. 2003, 204, 2055-2063. Nikitin, A. N.; Vairon, J.-P.; van Herk, A. M. Macromol.
(8) Gray, M. K.; Zhou, H.; Nguyen, S. T. Torkelson, J. M. Chem. Phys. 2004, 205, 2151-2160.
Macromolecules 2004, 37, 5586-5595. (38) Couturier, J. L.; Guerret, O.; Bertin, D.; Gigmes, D.; Marque,
(9) Gray, M. K.; Zhou, H.; Nguyen, S. T. Torkelson, J. M. Polymer S.; Tordo, P.; Chauvin, F.; Dufils, P. E. WO 2004/014926.
2004, 45, 4777-4786. (39) Buback, M.; Egorov, M.; Gilbert, R. G.; Kaminsky, V.; Olaj,
(10) Ziegler M. J.; Matyjaszewski, K. Macromolecules 2001, 34, O. F.; Russell, G. T.; Vana, P.; Zifferer, G. Macromol. Chem.
415-424. Phys. 2002, 203, 2570-2582.
(11) Matyjaszewski, K.; Ziegler, M. J.; Arehart, S. V.; Greszta, D.; (40) Mori, S.; Barth, H. G. Size Exclusion Chromatography;
Pakula, T. J. Phys. Org. Chem. 2000, 13, 775-786. Springer-Verlag: Berlin, 1999; p 201.
(12) Butz, S.; Baethge, H.; Schmidt-Naake, G. Angew. Macromol.
Chem. 1999, 270, 42-48. MA050087E