Nitroxide mediated living radical polymerization of styrene in miniemulsion—modelling persulfate-initiated systems

Chemical Engineering Science 58 (2003) 1177–1190www.elsevier.com/locate/ces

Nitroxide mediated living radical polymerization of styrenein miniemulsion—modelling persulfate-initiated systems

John W. Maa, Michael F. Cunninghama ;∗, Kim B. McAuleya,Barkev Keoshkerianb, Michael Georgesc

aDepartment of Chemical Engineering, Queen’s University, Kingston, Ont., Canada, K7L 3N6bXerox Research Centre of Canada, Mississauga, Ont., Canada, L5K 2L1

cUniversity of Toronto, Toronto, Ont., Canada, L5L 1C6

Received 29 January 2002; accepted 11 June 2002

Abstract

Recently we have constructed a mechanistic model describing the nitroxide mediated miniemulsion polymerization (NMMP) of styrene at135◦C, using alkoxyamine initiators to control polymer growth (Nitroxide-Mediated Polymerization of Styrene in Miniemulsion. ModelingStudies of Alkoxyamine-Initiated Systems, 2001b). The model has since been expanded to describe styrene NMMP at 135◦C usingTEMPO and the free radical initiator, potassium persulfate (KPS). The model includes mechanisms describing reactions in the aqueousand organic phases, particle nucleation, the entry and exit of oligomeric radicals, and the partitioning of nitroxide and styrene between theaqueous and organic phases. Predicted monomer conversions, number average molecular weights and polydispersities were in agreementwith experimentally measured values. Model simulations revealed that for systems employing high ratios of TEMPO:KPS, the consumptionof TEMPO by polymer radicals derived from KPS decomposition and styrene thermal initiation (using the accepted literature kineticrates) is not su=cient to lower TEMPO concentrations to levels where polymer growth can occur. By accounting for the consumption ofTEMPO by acid-catalyzed disproportionation, TEMPO concentrations are signi>cantly reduced, allowing for accurate model predictionsof monomer conversion, number average molecular weight and polydispersity at every experimental condition considered.? 2003 Elsevier Science Ltd. All rights reserved.

Keywords: Emulsion; Mathematical; Modelling; Multiphase reactions; Polymers; Polymerization

1. Introduction

Several “living” radical polymerization processes (e.g.,nitroxide-mediated polymerization, ATRP, RAFT) havebeen developed in recent years that show characteristics ofliving systems, e.g., narrow molecular weight distributions(MWDs) and active polymer end-groups. Of particularinterest to this study is nitroxide-mediated polymerization(NMP). Originally developed by Solomon, Rizzardo, andCacioli (1985) and later advanced by Georges, Veregin,Kazmaier, and Hamer (1993), NMP employs a nitroxide,such as 2,2,6,6-tetramethyl-1-piperidinyloxyl (TEMPO)to reversibly deactivate/trap growing polymer radicals toform dormant alkoxyamines (see Scheme 1). At 125◦C, theequilibrium coe=cient for radical deactivation, Keq = ka=kd,for TEMPO and polystyrene radicals greatly favors the

∗ Corresponding author. Tel.: +1-613-533-2782;fax: +1-613-533-6637.

E-mail address: [email protected] (M. F. Cunningham).

formation of the dormant alkoxyamine (Keq = 2:1 ×10−11 M−1, Fukuda et al., 1996).As a result, polymer radicals are predominantly dormant

in NMP, which reduces the likelihood of bimolecular ter-mination. Because monomer can only add to the growingpolymer radical between successive activation and deacti-vation reactions, the rate of polymerization is slower thanin conventional free radical polymerization. By initiatingpolymer radicals at nearly the same instant, polymers withnarrow MWDs can be synthesized by NMP. Because the>nal product consists of nitroxide-trapped polymer radi-cals, polymers made by NMP can be used to initiate subse-quent homo/copolymerizations to make well-de>ned blockcopolymers and polymers with complex architectures.Being a free radical process, NMP is relatively insensi-

tive to impurities and can be performed without the needto purify solvents or reagents, unlike traditional livingpolymerization processes (e.g., ionic polymerization andgroup transfer polymerization). Additionally, NMP can beperformed in commonly used heterogeneous systems, such

0009-2509/03/$ - see front matter ? 2003 Elsevier Science Ltd. All rights reserved.PII: S0009 -2509(02)00557 -2

mailto:[email protected]

1178 J. W. Ma et al. / Chemical Engineering Science 58 (2003) 1177–1190

+ka

kd

polystyreneradical

TEMPOalkoxyamine(dormant polymer radical)

NO•

NO

n

•CH

n

Scheme 1.

as suspension (Keoshkerian, Georges, & Boils-Boissier,1995), dispersion (Gabaston, Jackson, & Armes, 1998),emulsion (Bon, Bosveld, Klumperman, & German, 1997;Marestin, Noel, Guyot, & Claverie, 1998; Cao, He, Li,& Yang, 2001) and miniemulsion (MacLeod, Keoshke-rian, Odell, & Georgeo 1999; Prodpran, Dimonie, Sudol, &El-Aasser, 1999, 2000; Lansalot, Charleaux, Valron, Pirri,& Tordo, 1999; Pan, Sudol, Dimonie, & El-Aasser, 2001;Farcet, Lansalot, Charleux, Pirri, & Vairon, 2000; Farcet,Charleux, & Pirri, 2001; Cunningham, Xie, McAuley,Keoshkerian, & Georges, 2002; Keoshkerian, MacLeod,& Georges, 2001). However, relatively little it understoodabout the mechanisms involved in these processes due tothe complex interactions between reactions in the aqueousand organic phases.To help better understand these systems, in particular

nitroxide-mediated miniemulsion polymerization (NMMP)systems, we have recently measured the styrene-waterpartition coe=cients of TEMPO, 4-hydroxyl-TEMPO,and 4-amino-TEMPO (nitroxides with varying degreesof water solubility) at temperatures ranging from 25–135◦C, and studied the eNects of polystyrene, surfactantand hexadecane (costabilizer) on the partitioning prop-erties of these nitroxides (Ma, Cunningham, McAuley,Keoshkerian, & Georges, 2001). The hydrophilicityof the nitroxides was found to increase in the orderof TEMPO¡ 4-amino-TEMPO¡ 4-hydroxyl-TEMPO.Polystyrene, surfactant and hexadecane did not have a sig-ni>cant inOuence on the measured partitioning coe=cients.The results from the partitioning studies were used

in the construction of a mathematical model describingalkoxyamine-initiated styrene NMMP at 135◦C (Ma et al.,2003). The model included mechanisms to describe NMPreactions in the aqueous and organic phases, the entry andexit of polymer radicals, and the partitioning of nitroxideand styrene between the aqueous and organic phases. In themodel study, we simulated experimental systems initiatedby the alkoxyamines BST and hydroxyl-BST to examinethe eNects of nitroxide partitioning on NMMP kinetics; BSTand hydroxyl-BST are benzolystyryl radicals terminated byTEMPO and 4-hydroxyl-TEMPO respectively. The modelwas able to predict the monomer conversion, number av-erage molecular weight ( PMn) and polydispersity (PD) ateach experimental condition.

Complex interactions among chemical reaction equilib-rium, phase equilibrium, radical termination and thermal ini-tiation were found to inOuence the organic-phase concentra-tion of active polymer radicals in such a way that the diNerentpartitioning properties of TEMPO and 4-hydroxy-TEMPOdid not have a signi>cant inOuence on the rate of polymer-ization. Although only ∼ 40% of 4-hydroxy-TEMPO re-sides in the particle phase, compared to ∼ 98% of TEMPOin the particle phase (for typical recipes), the dominant roleplayed by thermal polymerization with styrene in determin-ing radical concentrations negates the large diNerences inpartitioning properties. (Note that for other monomers suchas acrylates this is not true, and nitroxide partitioning doessigni>cantly aNect the kinetics (Tortosa, Smith, Cunning-ham, 2001). Additionally, from quantitative estimates of thepopulation of living and dead polymer chains, the degree ofpolymer livingness was predicted to decrease continuouslyin each simulation, primarily due to radical termination byalkoxyamine disproportionation.The use of pre-made alkoxyamines to initiate large-scale

polymerization processes may not desirable because of thecost of synthesizing these compounds. Therefore, it wouldbe bene>cial if we could model NMMP systems initiatedby conventional initiators (e.g., KPS, benzoyl peroxide,azodiisobutyronitrile) to further our under understandingof these processes and to help select experimental condi-tions. In this article, we describe a mathematical model de-signed to predict the behavior of TEMPO-mediated styreneNMMP at 135◦C, initiated by KPS. The model equationswere solved using the simulation software PREDICI?.This report focuses mainly on the construction of the modeland builds upon our previous modeling eNorts concerningalkoxyamine-initiated NMMP (Ma et al., 2003). From thismodel-based study, we seek to identify whether or not ourcurrent understanding of the mechanisms involved in KPS-initiated NMMP is consistent with experimental evidence.

2. Model construction

2.1. System characteristics

The systems considered in this study are based on theTEMPO-mediated miniemulsion polymerization recipes

J. W. Ma et al. / Chemical Engineering Science 58 (2003) 1177–1190 1179

Table 1Styrene nitroxide-mediated miniemulsion polymerization system characteristics

Aqueous phase Organic phaseWater (123 g) Styrene (30 g)SDBS (surfactant) (0:88 g) Hexadecane (see below)KPS (see below) TEMPO (see below)

System KPS TEMPO TEMPO:KPS Hexadecane(mmol) (mmol) (mole ratio) (g)

1 0.37 0.63 1.7 5.42 0.37 1.5 4.0 5.43 1.1 1.9 1.7 5.44 1.1 4.4 4.0 5.4

5 0.37 0.63 1.7 3.46 0.37 1.5 4.0 3.47 1.1 1.9 1.7 3.48 1.1 4.4 4.0 3.49 0.74 2.2 3.0 4.4

Based on recipes used Cunningham et al. (2002). Temperature = 135◦C; reaction time = 6 h.

employed by Cunningham et al. (2002). The character-istics of each system are summarized in Table 1. Detailsconcerning the experimental work are discussed elsewhere(Cunningham et al., 2002).NMMP systems typically do not exhibit compartmental-

ization eNects because the average number of active polymerradicals per particle, Pn, is controlled by the rate of radicaltrapping, and not by the rate of radical entry and termina-tion (ButtRe, Storti, & MorbidelliRe, 2000; Charleux, 2000).Consequently, NMMP does not need to be modeled on adroplet-by-droplet basis. In our model, the aqueous phase iscomprised of water, KPS, TEMPO, styrene and styrenic rad-icals, and the organic phase is comprised of styrene, hexade-cane, TEMPO and polymer. KPS is assumed to be solubleonly in the aqueous phase. Polymerization reactions occur inboth the aqueous and organic phases, while TEMPO, styreneand styrenic radicals partition between the two phases tomaintain phase equilibrium.

2.2. Kinetic mechanism

2.2.1. Primary reactionsThe primary reactions involved in styrene NMMP are

summarized in Scheme 2, and with the exception of reac-tions 1 and 2, the reactions are the same as those used pre-viously to model alkoxyamine-initiated NMMP (Ma et al.,2003). All reactions are assumed to be independent of con-version and are uninOuenced by the solvent. The values ofall rate coe=cients are summarized in Table 2. A 15-minheating period is included in the model to reOect the actualtime required to raise the temperature of the polymerizationmixture from 25◦C to 135◦C. The temperature was assumedto increase linearly from 25 to 135◦C.Reaction 1 shows the decomposition of the persulfate ini-

tiator, S2O2−8 , to form two sulfate radicals, SO− ·4 , which Scheme 2.


Table 2Reaction rate coe=cients

Coe=cient Value/expression References/comments

ka 3× 1013 s−1 exp(−124 kJ mol−1=RT ) Fukuda et al., (2000)kad 5:7× 1014 s−1 exp(−153 kJ mol−1=RT ) Ohno, Tsujii, and Fukuda (1997)kd 8× 107 M−1 s−1 Bowry and Ingold (1992)kdecomp 8× 1015 s−1 exp(−135 kJ mol−1=RT ) Behrman and Edwards (1980)kinit 1× 109 M−1 s−1 Ingold (1973)kox 70 M−1 s−1 He et al. (2000)kp 107:63 M−1 s−1 exp(−32:51 kJ mol−1=RT ) Buback et al. (1995)kp1 4kp Deady, Mau, Moad, and Spuling (1993)kp2 3kp Deady et al. (1993)kp3 2kp Deady et al. (1993)kt0 1× 109 M−1 s−1 DiNusion limitktc 4× 108 M−1 s−1 Beuermann and Buback (2002)ktd 0:01ktc Chen, Wu, and Kuo (1987)kthermal 1:964× 105 M−2 s−1 exp(−10; 040K=T ) Hui and Hamielec (1972)ktr;M =kp 10−0:658 M−1 s−1 exp(−23:4 kJ mol−1=RT ) Tobolsky and ONenbach (1955)Keq 2:1× 10−11 M Fukuda et al. (2000)

can subsequently react with styrene,M , in the aqueous phaseto initiate polymerization (reaction 2) or terminate with an-other radical (reaction 11). For the moment, we wish tokeep the reaction mechanism as simple as possible and havechosen to neglect a number of known side reactions thattake place during KPS decomposition (KolthoN and Miller,1951; House, 1962; Behrman and Edwards, 1980). We havechosen to model the initiation step without the use of anempirical initiator e=ciency factor because the inOuence ofTEMPO on initiator e=ciency is unknown at this time. Thefate of primary radicals is thus determined by rate equationsfor propagation, termination, or deactivation by TEMPO toyield a living chain.

2.2.2. Radical entryEntry of polymer radicals into the organic phase

is modeled according to the mechanism proposed byMaxwell, Morrison, Napper, and Gilbert (1991), wherethe rate-determining step is the growth of polymer radicalsin the aqueous phase to a critical degree of polymeriza-tion, z, which renders the polymer radicals surface active.Surface-active radicals are then assumed to irreversiblyenter the monomer droplets without reacting further in theaqueous phase. Results from experimental and model stud-ies indicate that z=2 or 3 at 50◦C for sulfate-functionalizedpolystyrene radicals (Maxwell et al., 1991; Morrisonet al., 1993). Maxwell et al. (1991) developed the followingcorrelation to determine z for styrenic radicals with SO−

4end-groups:

z = 1 +−23 kJ mol−1RT ln[M ]aq;sat

: (16)

In Eq. (16), [M ]aq;sat is the saturation concentration ofstyrene in water at the temperature T , and R is the idealgas constant. At 135◦C, [M ]aq;sat = 1:6 × 10−2 M−1

(Ma et al., 2001), giving z = 3 at 135◦C from Eq. (16);

Scheme 3.

hence, we assume that z = 3 for sulfate-functionalizedpolystyrene radicals (P∗

i ).Polymer radicals derived from transfer to monomer (p∗

iand m∗

i ) and thermal initiation (M∗i ) are assumed to be-

come surface active upon growing to a length of two styreneunits; i.e., z = 2. We have assumed a lower value of z forpolymer radicals created by transfer to monomer and ther-mal initiation because these polymer radicals lack a neg-atively charged SO−

4 end-group, and as such, should beless hydrophilic than corresponding sulfate-functionalizedoligomers. We also assume that TEMPO is similar to onestyrene unit; hence, the radicals P∗

2 , p∗1 , m

∗1 and M

∗1 be-

come surface active when coupled with TEMPO. Reactionsincluded in the model to describe irreversible radical entryinto the organic phase are shown in Scheme 3.

2.2.3. Phase partitioningChemical species that are soluble in both styrene and wa-

ter will partition between the aqueous and organic phasesto maintain phase equilibrium. Modeling studies of NMMPsystems indicate that the rate of interfacial mass transferis su=ciently fast under typical polymerization conditionssuch that phase equilibrium can be assumed at all timesfor low molecular weight chemical species; e.g., TEMPO,styrene and styrenic radicals (Ma, 2001). We modeled thepartitioning of TEMPO, styrene and styrenic radicals in


PREDICI? using a series of equilibrium reactions (seeScheme 4). Phase equilibrium is maintained by assigninglarge values to the forward and reverse diNusion rate co-e=cients (¿ 108 s−1) in appropriate ratios to ensure thatthe desired phase partitioning is achieved. Aqueous- andorganic-phase equilibrium concentrations of styrene andTEMPO were calculated from recently reported nitroxidepartition coe=cients and water-solubility data (Ma et al.,2001). Styrenic radicals are assumed to have the samepartition coe=cient as styrene.

3. Results and discussion

The model was used to simulate 9 h of polymerizationtime. The initial conditions used in each simulation are sum-marized in Table 3.Aqueous-phase concentrations are eval-uated with respect to the volume of water. Organic phaseconcentrations are evaluated with respect to the total volumeof the organic phase.

3.1. Evaluation of model performance

Preliminary simulations were performed for each of theexperimental conditions outlined in Table 3. The predictedmonomer conversions were compared with experimentallymeasured values to assess the performance of the model.Results from these simulations are shown in Fig. 1.Fig. 1a shows that for systems 1 and 3, the predicted

monomer conversions agree reasonably well with the ex-perimental values. However, in systems 3 and 7, extensiveinductions periods are predicted at the start of the polymer-ization, which signi>cantly oNset the predicted monomer

Scheme 4.

Table 3Initial conditions used in model simulations

System [KPS]aq;0 [T · ]aq;0 [T · ]org;0 [M ]aq;0 [M ]org;0 [HD]org;0(mM) (mM) (mM) (mM) (M) (M)

1 2.8 0.15 13.5 15.4 6.39 0.532 2.8 0.36 31.8 15.4 6.39 0.533 8.5 0.46 40.6 15.4 6.39 0.534 8.5 1.1 95.5 15.4 6.39 0.53

5 2.8 0.16 14.4 15.4 6.82 0.366 2.8 0.37 33.9 15.4 6.82 0.367 8.5 0.47 43.3 15.4 6.82 0.368 8.5 1.1 102 15.4 6.82 0.369 5.7 0.54 49.3 15.4 6.60 0.45

conversions from the experimental values (see Fig. 1c). Theworst predictions occur in systems 2, 4, 6, 8 and 9 (seeFigs. 1b, d and e), where the induction periods are predictedto last for the duration of each simulation. The disparitiesbetween the predicted and measured monomer conversionscan be attributed to the predicted levels of free TEMPO ineach system.In systems 1, 3, 5, and 7, TEMPO is employed in a ratio

of 1.7:1 with respect to the initial number of mole of KPS.Since every molecule of KPS dissociates into two primaryradicals, there is a slight de>ciency of TEMPO with respectto primary radicals in these systems. Consequently, polymerradicals initiated from KPS decomposition are able to con-sume the majority of free TEMPO in these systems, allowingfor polymer growth. On the other hand, in systems 2, 4, 6 and8 (TEMPO : KPS=4 : 1) and system 9 (TEMPO : KPS=3 :1), TEMPO is employed in excess of what is needed to trapall polymer radicals initiated from KPS decomposition. Thepresence of excess TEMPO in NMMP systems inhibits poly-mer growth by forcing the equilibrium shown in Scheme1 towards the dormant alkoxyamine species. For polymergrowth to occur, the level of excess free TEMPO must bereduced su=ciently so that polymer propagation can com-pete with radical deactivation. In styrene NMMP, thermallygenerated radicals help to consume excess free nitroxide andallow polymer growth to take place at a sustainable rate.However, in systems 2, 4, 6, 8 and 9, our model predictsthat not enough thermally generated radicals are formed af-ter 9 h to su=ciently lower the free nitroxide concentrationto levels where polymer growth can occur.Fig. 2 shows the predicted fractions of free TEMPO

(T=T0) remaining in systems 1–4 and system 9 (the be-havior of T=T0 in systems 5–8 is similar to the behaviorexhibited by systems 1, 2, 3 and 4, respectively, and are notshown for this reason). Because of the low TEMPO:KPSratios employed in systems 1 and 3 (TEMPO : KPS= 1:7),the model predicts that polymer radicals initiated from KPSconsume nearly all the free TEMPO within the >rst 0:3 hof polymerization. Afterwards, T=T0 rebounds to ¡ 10%due to the release of free TEMPO by terminated polymer


0.0

0.5

1.0

Mon

omer

con

vers

ion

System 3 (expt'l)System 7 (expt'l)System 3 (model)System 7 (model)

——- - - -

0.0

0.5

1.0

0 3 6 9

Time (hours)

0 3 6 9

0 3 6 9

Time (hours)

Mon

omer

con

vers

ion ——

- - - -


0.0

0.5

1.0

0 3 6 9Time (hours)

Mon

omer

con

vers

ion


——- - - -

0.0

0.5

1.0

0 6 93

Time (hours)M

onom

er c

onve

rsio

n


——- - - -

(a) (b)

(c) (d)

0.0

0.5

1.0

Time (hours)

Monom

er

convers

ion

System 9 (expt'l)

System 9 (model)——

(e)

Fig. 1. Preliminary model predictions of monomer conversion. Predicted monomer conversions are compared with experimentally measured values.

0.0

0.5

1.0

0 3 6 9Time (hours)

T /

T0

system 1system 3

system 4

system 2

system 9

Fig. 2. Model predictions of the fraction of free TEMPO (T=T0) remainingin systems 1, 2, 3, 4 and 9.

radicals. Eventually, thermally generated radicals reduceT=T0 to¡ 1% in these systems, allowing polymer growth tooccur. Conversely, in systems 2, 4 and 9, systems employing

an excess of TEMPO, more than 33% of the initial amountTEMPO is predicted to remain in these systems followinginitiation. Even after 9 h, ¿ 20% of the initial amount ofTEMPO is still predicted to remain in these systems despitethe consumption of free TEMPO by thermally generated rad-icals. Consequently, the model predicts that polymer growthin inhibited for the duration of these simulations.Induction periods are commonly observed in NMMP sys-

tems (Lansalot et al., 1999; Farcet et al., 2000; Cunninghamet al., 2002) and exist when the free nitroxide concentra-tion makes radical trapping more favorable than polymerpropagation. The consumption of free nitroxide by ther-mally generated radicals is thought to reduce the concentra-tion of free nitroxide to levels where polymer growth canoccur. However, if the consumption of free nitroxide bythermally generated radicals is, by itself, responsible for re-ducing free nitroxide concentrations in NMMP, our modelsimulations should have predicted signi>cantly lower levelsof free TEMPO at each experimental condition and hence,shorter induction periods. The fact that our model is ableto accurately predict the behavior of alkoxyamine-initiated


N+

O

N

OH

2N

O•

+kdisprop

H ++

oxoammoniumion

hydroxylamine

Scheme 5.

N

O

+ H+N

+

OH

N

O

+N

+

OH

N+

O

N

OH

+

Scheme 6.

styrene NMMP systems (Ma et al., 2003) suggests that ni-troxide consumption by thermally generated radicals maynot be the only mechanism responsible for reducing free ni-troxide levels in KPS-initiated systems.One mechanism that may be responsible for consuming

free TEMPO in KPS-initiated NMMP is the acid-catalyzeddisproportionation of TEMPO, where two molecules ofTEMPO react in the presence of a strong acid to forma hydroxylamine and an oxoammonium ion (Scheme 5)(Aurich, 1989):The acid-catalyzed disproportionation of TEMPO likely

does not occur by a single reaction step, but may involvethe protonation of TEMPO, followed by the oxidation of asecond TEMPO molecule (Scheme 6) (Aurich, 1989):Acid-catalyzed disproportionation is thought to be re-

sponsible for consuming free nitroxide in systems employ-ing organic acids (e.g., camphorsulfonic acid and ascorbicacid) as rate enhancing agents (Veregin, Odell, Michalak, &Georges, 1996). Additionally, Farcet et al. (2000) reportedthat the pH of the aqueous phase has a signi>cant inOuenceon the rate of polymerization in NMMP systems employingthe phosphonylated nitroxide, SG1, due to the rapid con-sumption of the nitroxide by acid-catalyzed side reactions.Because HSO−

4 (a strong acid) is a bi-product of KPSdecomposition, it is reasonable that acid-catalyzed dispro-portionation could consume free TEMPO in the aqueousphase of KPS-initiated systems. The pH in our experimentsdecreased to 2–3 by the time the temperature had reached135◦C, and remained relatively constant for the durationof the experiment. The consumption of free TEMPO inthe aqueous phase would cause TEMPO to diNuse out ofthe organic phase to maintain phase equilibrium, thereby

signi>cantly reducing the amount of free TEMPO in organicphase. Therefore, the model was modi>ed to account for theformation of HSO−

4 and the consumption of free TEMPOby acid-catalyzed disproportionation to investigate if reduc-ing the level of free TEMPO by additional side reactionsimproves the model’s ability to predict the behavior of theexperimental systems. In the following section we discussthe modi>cations made to the model and the results obtainedfrom simulations using the modi>ed reaction mechanism.Consumption of TEMPO by acid is not the only possible

explanation for the unexpectedly short induction periods.The consumption of nitroxide at rates faster than would beexpected due to generation of thermally initiated radicalsin the thermal polymerization of styrene was >rst notedby Moad, Rizzardo, and Solomon (1982). Thermal styrenepolymerization in the presence of nitroxide was also ex-amined by Connolly and Scaiano (1997), Devonport et al.(1997), and Boutevin and Bertin (1999). Connolly andScaiano (1997) reported that TEMPO is not inert under typ-ical polymerization conditions, and may react with styreneand polystyrene by abstracting benzylic hydrogens. Devon-port et al. (1997) conducted thermal styrene polymerizationsat various styrene:TEMPO ratios, and observed inductionperiods that increased with decreasing styrene:TEMPO.However these induction periods are much less than whatis expected for the consumption of TEMPO using the ac-cepted thermal initiation rate equation of Hui and Hamielec(1972), in agreement with the observations of Moadet al. (1982). Zhang and Ray (2002), in their modeling ofnitroxide-mediated polymerizations, found that they hadto increase the thermal initiation rate by at least a factorof ten during the induction period to accurately reproduce


experimentally observed induction periods from other labo-ratories. Boutevin and Bertin (1999) showed in their kineticstudies that TEMPO inOuences the rate of Diels-Alder rad-ical formation, and that in the presence of excess TEMPOa transfer reaction involving the Diels-Alder dimer andTEMPO to yield an active radical and an inactive hy-droxylamine becomes signi>cant. This reaction thereforeconsumes a TEMPO molecule and yields an active radical,which can then be “capped” by another TEMPO to give aliving chain. In our miniemulsion system, this reaction isalso expected to occur. However this reaction increases thenumber of living chains (one new living chain for everytwo excess TEMPO molecules) and therefore decreasesmolecular weight. In contrast, TEMPO consumption byacid does not increase the chain number. If the reactionsproposed by Boutevin and Bertin (1999) were dominant inconsuming the excess TEMPO in our system, molecularweights would be signi>cantly lower than our observedexperimental values.

3.2. Model modi;cations

3.2.1. Supplemental reactionsThe reactions shown in Scheme 7 were added to the

model to account for the generation of HSO−4 and the

acid-catalyzed disproportionation of TEMPO:Reaction 24 shows the hydrolysis of SO− ·

4 to formHSO−

4 and a hydroxyl radical, HO· . The reported values for

khydrolysis, the rate coe=cient for SO− ·4 hydrolysis, range

from 6 × 104 to 1 × 107 M−1 s−1 (Neta, Huie, & Ross,1988; Moad and Solomon, 1989). A value of khydrolysis =2×105 M−1 s−1 was found to provide reasonable predictionsof monomer conversion at 135◦C.Reaction 25 shows the reversible dissociation of HSO−

4 .The acid dissociation coe=cient, Ka, was calculated forHSO−

4 from the following correlation (Matsushima andOkuwaki, 1988):

logKa = a+ b log(T ) + c=T + dT + eT 2: (28)

In Eq. (28), T is the temperature in Kelvin; a = 577:214;b = −246:01; c = −12717 K; d = 0:283133 K−1; and e =−1:37566 × 10−4 K−2 (Matsushima and Okuwaki, 1988).From Eq. (28),Ka was calculated to be 2:59×10−4 at 135◦C.HO· can initiate polymerization by reacting with styrene

in the aqueous phase (reaction 26). The reaction betweenHO· and styrene generates a hydroxyl-functionalized poly-mer radical, r ·1 , and occurs at rates near the diNusion limit,kinit=1×109 M−1 s−1 (Ingold, 1973). Details regarding thepartitioning and entry of hydroxyl-functionalized oligomersinto the organic phase are discussed in the next section.The acid-catalyzed disproportionation of TEMPO is

represented by reaction 27 and has been modeled as athird-order reaction using a single reaction step, where twoTEMPO molecules react in the presence of a hydroniumion, H+, to form a hydroxylamine molecule, TH , and anoxoammonium ion, TO. Modeling acid-catalyzed dispro-

Scheme 7.

Scheme 8.

portionation in this manner enabled us to minimize thenumber of unknown rate coe=cients used in the model.The rate coe=cient, kdisprop, was obtained by >tting thepredicted monomer conversions to experimental results. Avalue of kdisprop =1×107 M−2 s−1 provided the best modelpredictions. We would like to stress that the purpose ofincluding acid-catalyzed disproportionation in the model isto demonstrate that in addition to nitroxide consumption bythermally generated radicals, other side reactions, of whichacid-catalyzed disproportionation is a possibility, may act toconsume free TEMPO. Therefore, the value of kdisprop usedin the model should not be taken as an accurate estimateof its true value without further experimental validation.Experimental studies are required to study the kineticsof acid-catalyzed disproportionation and to determine theextent to which this reaction takes place in NMMP.Propagation, termination, transfer to monomer, and

reversible radical deactivation reactions involving OH-functionalized polymer radicals are treated according toreactions 3–15.

3.3. Radical entry and oligomer partitioning

Initiation of polymer radicals by HO· creates a species ofuncharged radicals (r ·i ) that possess a hydroxyl end-group.These hydroxyl-functionalized oligomer radicals are likelyto be signi>cantly less hydrophilic than correspondingsulfate-functionalized oligomers (P ·i ), but slightly morehydrophilic than corresponding oligomers formed by ther-mal initiation and transfer to monomer (p·i andm·i ). Whenconsidering the entry of these hydroxyl-functionalizedpolymer radicals into the organic phase, we assume that


0.0

0.5

1.0

0 3 6 9

Time (hours)0 3 6 9

Time (hours)

0 3 6 9Time (hours)

0 3 6 9Time (hours)

0 3 6 9Time (hours)

Mon

omer

con

vers

ion

0.0

0.5

1.0

Mon

omer

con

vers

ion


——- - - -

0 0

0.5

1.0

Mon

omer

con

vers

ion


——- - - -


——- - - -

0 0

0.5

1.0

Mon

omer

con

vers

ion


——- - - -

0.0

0.5

1.0

Fra

ctio

nal C

onve

rsio

n

System 9 (expt'l)

System 9 (model)——

(a) (b)

(c) (d)

(e)

Fig. 3. Comparison of predicted and measured monomer conversions. Predicted monomer conversions were obtained with additional nitroxide-consumingside reactions.

z = 2 (reaction 29). Additionally, the radical, r ·1 , is as-sumed to be surface active when coupled with TEMPO(reaction 30).We also assume that the partitioning behavior of

hydroxyl-functionalized unimers (r ·1 ) is similar to that ofstyrene (see reaction 31). This validity of this assumptionis uncertain due to a lack of data pertaining to the partition-ing of hydroxide-functionalized oligomer radicals betweenstyrene and water (Scheme 8).

3.4. Model performance revisited

3.4.1. Conversion versus timeSimulations were performed again for the experimental

conditions outlined in Table 3. The predicted monomer con-versions are compared with the experimentally measuredvalues in Fig. 3. With the addition of reactions to accountfor the consumption of free TEMPO by acid-catalyzed dis-

proportionation, the induction periods are signi>cantly re-duced in each system and the predicted monomer conver-sions agree reasonably well with the measured values. How-ever, there are notable diNerences in the predicted behaviorat the start of the polymerization depending on the ratio ofTEMPO:KPS employed in each system.Due to the rapid decomposition of KPS and the rapid

growth of polymer radicals at the polymerization temper-ature, the model predicts that free TEMPO is consumedrapidly by radical coupling and acid-catalyzed dispropor-tionation at the start of the polymerization. In systems 1,3,5 and 7—systems employing a TEMPO:KPS ratio of 1.7:1(a stoichiometric de>ciency of TEMPO with respect toprimary radicals)—the consumption of free TEMPO byacid-catalyzed disproportionation exacerbates the de>ciencyof free TEMPO in these systems, allowing a signi>cantnumber of polymer radicals to grow unmediated, whichleads to a short period of poorly controlled polymer growth.


0

25

50

0.0 0.5 1.0

Monomer conversion

0.0 0.5 1.0

Monomer conversion

Mn (

×10

3 g m

ol-1

)


——- - - -

0

25

50

Mn (

×10

3 g m

ol-1

)


——- - - -

0

25

50

0.0 0.5 1.0

Monomer conversion

Mn (

×10

3 g m

ol-1

)


——- - - -

0

25

50

0.0 0.5 1.0

Monomer conversion

Mn (

×10

3 g m

ol-1

)


——- - - -

0

25

50

0 0.5 1.0

Monomer conversion

Mn (

×10

3 g m

ol-1

)

——System 9 (expt'l)System 9 (model)

(a)

(c) (d)

(b)

(e)

Fig. 4. Comparison of predicted and measured number average molecular weights. Predicted number average molecular weights were obtained withadditional nitroxide-consuming side reactions.

This is can be seen in Figs. 3a and c by the sharp increasesin the predicted monomer conversions after ∼ 0:3 h. Fromthe simulation output, the ratio of the rate of polymer prop-agation, RP;org, to the rate of radical deactivation, Rd;org, inthe organic phase was calculated from the following:

RP;orgRd;org

=kp[M ]org[R

· ]orgkd[T · ]org[R· ]org

; (32)

where [M ]org is the organic-phase concentration of styrene;[R· ]org is the organic-phase concentration of active poly-mer radicals; and [T · ]org is the organic-phase concentrationof free TEMPO. In systems 1, 3, 5 and 7, the rate of poly-mer propagation was found to be up to 300 times faster thanthe rate of radical deactivation during the periods of poorlycontrolled polymer growth. Whereas at all other times dur-ing the polymerization, the rate of polymer propagation wasfound to be less than or slightly greater than the rate of radi-cal deactivation in these systems (0:4¡RP;org=Rd;org¡ 10).

In contrast, in systems 2, 4, 6, 8 and 9 (systems employ-ing a stoichiometric excess of free TEMPO) due to the ex-cess amounts of free TEMPO, induction periods are pre-dicted at the start of polymerization despite the consumptionof free TEMPO by acid-catalyzed disproportionation. Con-sequently, the model predicts that polymer growth is wellcontrolled in these systems. From the simulation results, therate of polymer propagation was calculated to be¡ 5 timesfaster than the rate of radical deactivation in the organicphase at anytime in these systems.

3.4.2. Number average molecular weightPredicted number average molecular weights were found

to agree reasonably well with the experimentally measuredvalues (see Fig. 4). There are minor discrepancies betweenthe predicted and measured values of PMn for system 8 (Fig.4d), but it is uncertain whether the diNerences are due to ex-perimental or simulation errors because of the scarce num-


1.0

1.5

2.0

2.5

0.0 0.5 1.0

Monomer conversion

PD


——- - - -

1.0

1.5

2.0

2.5

0.0 0.5 1.0

Monomer conversion

PD


——- - - -

1.0

1.5

2.0

2.5

0.0 0.5 1.0

Monomer conversion

PD


——- - - -

1.0

1.5

2.0

2.5

0.0 0.5 1.0

Monomer conversion

PD


——- - - -

1.0

1.5

2.0

2.5

0.0 0.5 1.0

Monomer conversion

PD

——System 9 (expt'l)System 9 (model)

(a)

(c) (d)

(b)

(e)

Fig. 5. Comparison of predicted and measured polydispersities. Predicted polydispersities were obtained with additional nitroxide-consuming side reactions.

ber of measurements taken. Additionally, the model predictsPMn to be independent of the amount of hexadecane, as ob-served by Cunningham et al. (2002). The good agreementwith PMn and therefore the number of chains supports theconsumption of excess TEMPO by a mechanism such asacid-catalyzed disproportionation that does not increase thenumber of chains.In systems 1, 3, 5 and 7 (systems employing a

TEMPO:KPS ratio of 1.7) the predicted values of PMn gothrough a maximum at low conversions before increasinglinearly with conversion (see Figs. 4a and c). The maximain the predicted MWs are caused by the de>ciency of freeTEMPO in these systems, which allow polymer radicals togrow rapidly at the start of the polymerization, resultingin poorly controlled polymer growth and signi>cantly highMWs. Unfortunately, no measurements of PMn are availableat very low monomer conversions that could validate thepredicted behavior; however Cao et al. (2001) have identi->ed minor peaks in GPC traces that correspond to the for-mation of high molecular weight polymer at low monomer

conversions in NMP systems performed in emulsion,and attributed this behavior to poorly controlled polymergrowth.PMn is predicted to increase linearly with conversion

throughout each simulation in systems 2, 4, 6, 8 and 9 dueto the excess amounts of free TEMPO employed in thesesystems (TEMPO : KPS = 4 : 1 and 3:1), which ensuresthat polymer growth is well controlled.

3.4.3. PolydispersityThe predicted polydispersities are compared with the

experimentally measured values in Fig. 5. Overall, thereis good agreement between the predicted and measuredpolydispersities. In systems 1, 2, 5, 6 and 9, PD is pre-dicted to increase at conversions ¿ 50%, which is inagreement with the measured values (see Figs. 5a, b ande). There are diNerences between the predicted and mea-sured polydispersities for systems 4 and 8 (see Fig. 5d),but it is di=cult to identify the cause of the discrepancies


because of the scarce number of experimental datapoints.High polydispersities (PD¿ 2) are predicted at low con-

versions in systems 1, 3, 5 and 7 (systems employing aTEMPO:KPS ratio of 1.7:1). In these systems, the de>-ciency of free TEMPO increases the likelihood that polymerradicals will terminate by combination and disproportion-ation, and allows polymer radicals to grow rapidly at thestart of the polymerization; both of which are factors thatcause the MWD to broaden. Signi>cantly lower polydisper-sities (PD¡ 1:4) are predicted in systems 2, 4, 6, 8 and 9(systems employing a TEMPO:KPS ratio of 4:1 and 3:1)because polymer growth is predicted to be well controlleddue to the excess amounts of TEMPO employed in thesesystems. Other investigators have also predicted polydisper-sities ¿ 2 at low conversions. These values are typicallynot seen experimentally, likely because the low molecularweight portion of the distribution responsible for the broad-ness lies below the detection limit of most gel permeationchromatographs.

4. Conclusions

In this article, we have described construction of amathematical model designed to predict the behavior ofstyrene NMMP at 135◦C, initiated by KPS. The resultsfrom our model simulations indicate that the mechanismsof KPS-initiated styrene NMMP are more complex than>rst anticipated. In systems employing excess amounts ofTEMPO with respect to KPS, our model predicts that theconsumption of excess free TEMPO by thermally gener-ated radicals alone is not su=cient to lower free TEMPOconcentrations to levels where polymer growth can takeplace. The addition of acid-catalyzed disproportionationwas found to signi>cantly reduce the level of free TEMPO,allowing for reasonable predictions of monomer conver-sion, PMn and PD at each of the experimental conditionsconsidered.Although we have suggested acid-catalyzed dispropor-

tionation as a plausible mechanism by which TEMPO maybe consumed in KPS-initiated systems and have approxi-mated the rates at which this reactions need to occur toexplain the experimental observations, experimental stud-ies are required to test whether or not the actual rate ofacid-catalyzed disproportionation is consistent with the ratesused in our model. Additionally, experimental studies arerequired to determine the extent to which acid-catalyzeddisproportionation takes place in KPS-initiated NMMP andto identify any other signi>cant nitroxide-consuming sidereactions in these systems. After further validation, we an-ticipate that the model could be used to identify promisingexperimental conditions that will lead to optimal processperformance and polymer properties.

Notation

Di dead polymer chain (bimolecular termination)Da;i dead polymer chain (alkoxyamine dispropor-

tionation)Dh;i dead polymer chain (hydroxylamine dispropor-

tionation)Dt; i dead polymer chain (transfer to monomer)H+ hydronium ionHO· hydroxyl radicalHSO−

4 hydrogen sulfate acidH2O water[HD]org;0 initial organic-phase concentration of hexade-

caneKPS potassium persulfate[KPS]aq;0 initial aqueous phase concentration of potassium

persulfatem·i active polymer radical formed when thermally

generated radicals undergo transfer to monomerM styrene molecule[M ]aq aqueous phase concentration of styrene[M ]aq;0 initial aqueous-phase concentration of styrene[M ]org organic-phase concentration of styrene[M ]org;0 initial organic-phase concentration of styreneM ·i active polymer radical derived from styrene

thermal initiationNMP nitroxide mediated free radical polymerizationNMMP nitroxide-mediated miniemulsion polymeriza-

tionp·i active polymeric radical formed when

initiator-derived radicals undergo transfer tomonomer

P ·i active polymer radical derived fromalkoxyamine initiator

r ·i hydroxyl-functionalized polymer radicalR·i various polymer radicals[R· ]org organic-phase active radical concentrationS2O2−8 persulfate ionSO− ·

4 sulfate radicalSDBS sodium dodecylbenzenesulfonic acid (surfac-

tant)T · TEMPO radical[T · ]aq aqueous-phase TEMPO concentration[T · ]aq;0 initial aqueous-phase TEMPO concentration[T · ]org organic-phase TEMPO concentration[T · ]org;0 initial organic-phase TEMPO concentrationTH hydroxylamine moleculeTO oxoammonium ionTRi nitroxide-capped polymer radical (alkoxya-

mine)

Acknowledgements

We wish to thank the government of Ontario, the Xe-rox Research Centre of Canada, the Natural Sciences and


Engineering Research Council of Canada (NSERC), andQueen’s University for their >nancial support of this project.We would also like to thank Dr. Erwin Buncel for insightfuldiscussions regarding nitroxide chemistry and Dr. MichaelWulkow for his assistance with PREDICI?.

References

Aurich, H. G. (1989). Nitroxides. In S. Patai & Z. Rappoport (Eds.),Nitrones, nitronates and nitroxides. New York: Wiley.

Behrman, E. J., & Edwards, J. O. (1980). The thermal decomposition ofperoxidisulphate ions. Reviews in Inorganic Chemistry, 2, 179.

Beuermann, S., & Buback, M. (2002). Rate coe=cients of free-radicalpolymerization deduced From pulsed laser experiments. Progress inPolymer Science, 27, 191–254.

Bon, S. A. F., Bosveld, M., Klumperman, B., & German, A. L. (1997).Controlled radical polymerization in emulsion. Macromolecules, 30,324–326.

Boutevin, B., & Bertin, D. (1999). Controlled free radical polymerizationof styrene in the presence of nitroxide radicals. I. Thermal initiation.European Polymer Journal, 35, 815–825.

Bowry, V. W., & Ingold, K. W. (1992). Kinetics of nitroxide radicaltrapping. 2. Structural eNects. Journal of the American ChemicalSociety, 114, 4992–4996.

Buback, M., Gilbert, R. G., Hutchinson, R. A., Klumperman, B.,Kuchta, F. D., Manders, B. G., O’Driscoll, K. F., Russell, G. T., &Schweer, J. (1995). Critically evaluated rate coe=cients for free-radicalpolymerization. Macromolecular Chemistry and Physics, 196, 3267–3280.

ButtRe, A., Storti, G., & Morbidelli, M. (2000). Miniemulsion living freeradical polymerization of styrene. Macromolecules, 33, 3485–3487.

Cao, J., He, J., Li, C., & Yang, Y. (2001). Nitroxide-mediated radicalpolymerization of styrene in emulsion. Polymer Journal, 33(1), 75–80.

Charleux, B. (2000). Theoretical aspects of controlled radicalpolymerization in a dispersed medium. Macromolecules, 33(15),5358–5365.

Chen, C. -Y., Wu, Z. -Z., & Kuo, J. -F. (1987). Determination of the modeof termination in free-radical copolymerization. Polymer Engineeringand Science, 27, 553–557.

Connolly, T. J., & Scaiano, J. C. (1997). Reactions of the “stable”nitroxide radical TEMPO. Relevance to “living” free radicalpolymerizations and autopolymerization of styrene. TetrahedronLetters, 38, 1133–1136.

Cunningham, M. F., Xie, M., McAuley, K. B., Keoshkerian, D., &Georges, M. K. (2002). Nitroxide-mediated styrene miniemulsionpolymerization. Macromolecules, 35, 59–66.

Deady, M., Mau, A. W. H., Moad, G., & Spuling, T. H. (1993). Evaluationof the kinetic parameters for styrene polymerization and their chainlength dependence by kinetic simulation and pulsed laser photolysis.Makromolekulare Chemie, 194, 1691–1705.

Devonport, W., Michalak, L., Malmstrom, E., Mate, M., Kurdi, B.,Hawker, C. J., Barclay, G. C., & Sinta, R. (1997). “Living”fee-radical polymerizations in the absence of initiators: Controlledautopolymerization. Macromolecules, 30, 1929–1934.

Farcet, C., Charleux, B., & Pirri, R. (2001).Poly(n-butyl-actylate) homopolymer and poly[n-butylacrylate-b-(n-butyl acrylate-co-styrene)] block copolymer preparedvia nitroxide-mediated living/controlled radical polymerization inminiemulsion. Macromolecules, 34, 3823–3826.

Farcet, C., Lansalot, M., Charleux, B., Pirri, R., & Vairon, J. P.(2000). Mechanistic aspects of nitroxide-mediated controlled radicalpolymerization of styrene in miniemulsion using a water-soluble radicalinitiator. Macromolecules, 33, 8559–8570.

Fukuda, T., Goto, A., Ohno, K. (2000). Mechanisms and kineticsof living radical polymerizations. Makromolekulare Chemie: RapidCommunications, 21, 151–165.

Fukuda, T., Terauchi, T., Goto, A., Ohno, K., Tsujii, Y., Miyamoto,T., Kobatake, S., & Yamada, B. (1996). Mechanisms and kinetics ofnitroxide-controlled free radical polymerization. Macromolecules, 29,6393–6398.

Gabaston, L. I., Jackson, R. A., & Armes, S. P. (1998). Living free-radicaldispersion polymerization of styrene.Macromolecules, 31, 2883–2888.

Georges, M. K., Veregin, R. P. N., Kazmaier, P. M., & Hamer,G. K. (1993). Narrow molecular weight resins by a free-radicalpolymerization process. Macromolecules, 26, 2987–2988.

He, J., Li, L., Yang, Y. (2000). ENects of hydrogen transferreaction on kinetics of nitroxide-mediated free-radical polymerization.Macromolecules, 33, 2286–2289.

House, D. A. (1962). Kinetics and mechanism of oxidations byperoxydisulfate. Chemistry Reviews, 62, 185–203.

Hui, A. W., & Hamielec, A. E. (1972). Thermal polymerization of styreneat high conversions and temperatures. An experimental study. Journalof Applied Polymer Science, 16, 749–769.

Ingold, K. U. (1973). Rate constants for free radical reactions in solution.In J. K. Kochi (Ed.), Free radicals. New York: Wiley.

Keoshkerian, B., Georges, M. K., & Boils-Boissier, D. (1995). Livingfree-radical aqueous polymerization.Macromolecules, 28, 6381–6383.

Keoshkerian, B., MacLeod, P. J., & Georges, M. K. (2001).Block copolymer synthesis by a miniemulsion stable free radicalpolymerization process. Macromolecules, 34, 3594–3599.

KolthoN, I. M., & Miller, I. K. (1951). The chemistry of persulfate. I.The kinetics and mechanism of the decomposition of the persulfateion in aqueous media. Journal of the American Chemical Society,73, 3055–3059.

Lansalot, M., Charleaux, B., Valron, J. -P., Pirri, R., & Tordo, P. (1999).Nitroxide-mediated controlled free-radical polymerization of styrene.Polymer Preprints, 40(2), 317–318.

Ma, J. W. (2001). Understanding the nitroxide-mediated polymerizationof styrene in miniemulsion through modeling studies. Doctoral thesis.

Ma, J. W., Cunningham, M. F., McAuley, K. B., Keoshkerian, B., &Georges, M. K. (2001). Nitroxide partitioning between styrene andwater. Journal of Polymer Science, Part A: Polymer Chemistry, 39,1081–1089.

Ma, J. W., Smith, J. A., Cunningham, M. F., McAuley, K. B.,Keoshkerian, B., & Georges, M. K. (2003). Nitroxide-mediatedpolymerization of styrene in miniemulsion. Modeling studies ofalkoxyamine-initiated systems. Chemical Engineering Science, doi:10.1016/S0009-2509(02)00427-X.

MacLeod, P. J., Keoshkerian, B., Odell, P., Georgeo, M. K. (1999).Stable free radical miniemulsion polymerization. Proceedings of theACS Division of Polymeric Materials, Science and Engineering, SanFrancisco, CA, USA, 80, 539–545.

Marestin, C., Noel, C., Guyot, A., & Claverie, J. (1998). Nitroxidemediated living radical polymerization of styrene in emulsion.Macromolecules, 31, 4041–4044.

Matsushima, Y., & Okuwaki, A. (1988). The second dissociationconstant of sulfuric acid at elevated temperatures from potentiometricmeasurements. Bulletin of the Chemical Society of Japan, 61,3344–3346.

Maxwell, I. A., Morrison, B. R., Napper, D. H., & Gilbert, R. G. (1991).Entry of free radicals into latex particles in emulsion polymerization.Macromolecules, 24, 1629–1640.

Moad, G., Rizzardo, E., & Solomon, D. H. (1982). A product studyof the nitroxide inhibited thermal polymerization of styrene. PolymerBulletin, 6, 589–593.

Moad, G., & Solomon, D. H. (1989). Azo and peroxy initiators. In G. C.Eastmond, A. Ledwith, S. Russo, & P. Sigwalt (Eds.), Comprehensivepolymer science. Chain polymerization, Vol. 3. Oxford: PergamonPress.


Morrison, B. R., Maxwell, I. A., Napper, D. H., Gilbert, R. G.,AmmerdorNer, J. L., & German, A. L. (1993). Characterization ofwater-soluble oligomers formed during the emulsion polymerizationof styrene by means of isotachophoresis. Journal of Polymer Science,Part A: Polymer Chemistry, 31, 467–483.

Neta, P., Huie, R. E., & Ross, A. B. (1988). Rate constants for reactionsof inorganic radicals in aqueous solution. Journal of Physical andChemical Reference Data, 17(3), 1027–1284.

Ohno, K., Tsujii, Y., & Fukuda, T. (1997). Mechanisms andkinetics of nitroxide-controlled free radical polymerization. ThermalDecomposition of 2,2,6,6-tetramethyl-1-polystyroxypiperidines.Macromolecules, 30, 2503–2506.

Pan, G., Sudol, E. D., Dimonie, V. L., & El-Aasser, M. S. (2001).Nitroxide-mediated living free-radical miniemulsion polymerization ofstyrene. Macromolecules, 34, 481–488.

Prodpran, T., Dimonie, V. L., Sudol, E. D., & El-Aasser, M. S. (1999).Nitroxide-mediated living free radical miniemulsion polymerization ofstyrene. Proceedings of the American Chemical Society, Division ofPolymeric Materials: Science and Engineering, 80, 534–535.

Prodpran, T., Dimonie, V. L., Sudol, E. D., & El-Aasser, M. S. (2000).Nitroxide-mediated living free radical miniemulsion polymerization ofstyrene. Macromolecular Symposia, 155, 1–14.

Solomon, D. H., Rizzardo, E., & Cacioli, P. (1985). Polymerizationprocess and polymers produced thereby. US Patent 4,581,429.

Tobolsky, A. V., & ONenbach, J. (1955). Kinetic constants for styrenepolymerization. Journal of Polymer Science, 16, 311–314.

Tortosa, K., Smith, J.-A., & Cunningham, M. F. (2001). Synthesisof polystyrene-block-poly(butyl acrylate) copolymers using nitroxide-mediated living radical polymerization in miniemulsion.Macromolecular Rapid Communications, 22, 957–961.

Veregin, R. P. N., Odell, P. G., Michalak, L. M., & Georges,M. K. (1996). Mechanism of rate enhancement using organicacids in nitroxide-mediated living free-radical polymerizations.Macromolecules, 29, 4161–4163.

Zhang, M., & Ray, W.H. (2002). Modeling of “living” free radicalpolymerization processes I. Batch, semibatch, and continuous tankreactors, Journal of Applied Polymer Science, in press.

Nitroxide mediated living radical polymerization of styrene in miniemulsion—modelling persulfate-initiated systems

Documents