Synthesis of Poly(vinyl pivalate) by Atom Transfer Radical Polymerization in Supercritical Carbon Dioxide

European Polymer Journal 61 (2014) 93–104

Contents lists available at ScienceDirect

European Polymer Journal

journal homepage: www.elsevier .com/locate /europol j

Synthesis of poly(vinyl pivalate) by atom transfer radicalpolymerization in supercritical carbon dioxide

http://dx.doi.org/10.1016/j.eurpolymj.2014.09.0030014-3057/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding author.E-mail address: [email protected] (J.-J. Shim).

Muhammad Naoshad Islam a, Yuvaraj Haldorai a, Van Hoa Nguyen a,b, Jae-Jin Shim a,⇑a School of Chemical Engineering, Yeungnam University, Gyeongsan, Gyeoungbuk 712-749, Republic of Koreab Department of Chemistry, Nha Trang University, Nha Trang, Viet Nam

a r t i c l e i n f o

Article history:Received 25 May 2014Received in revised form 9 August 2014Accepted 6 September 2014Available online 16 September 2014

Keywords:Atom transfer radical polymerization (ATRP)Poly(vinyl pivalate)Supercritical carbon dioxideCopper halideTerpyridinePoly(vinyl alcohol)

a b s t r a c t

Atom transfer radical polymerization of vinyl pivalate was carried out in supercritical car-bon dioxide using CuBr or CuCl/terpyridine (tpy) complex system and ethyl 2-bromoisobu-tyrate as an initiator. The reaction kinetics of the two different catalytic systems (CuBr/tpyand CuCl/tpy) was investigated. In addition, the effects of temperature, pressure, and cat-alyst/ligand concentration were examined systematically to obtain an acceptable rate ofpolymerization and control over the number-average molecular weight (Mn) and polydis-persity index (PDI). The result showed that relatively low PDIs were obtained and polymer-ization rates were enhanced at higher pressures. The CuCl/tpy complex exhibited themaximum conversion of 90% with a Mn and PDI of 60.2 kg/mol and 1.29, respectively.The CuCl/tpy catalyst system showed better agreement between theoretical and experi-mental Mn compared to the CuBr/tpy. The living character of poly(vinyl pivalate) (PVPi)was proven by 1H NMR spectrum and chain extension reaction. The formation of poly(vinylalcohol) via saponification of the resulting PVPi was also carried out. The structures of bothPVPi and PVA were examined by 1H NMR and FTIR spectra.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

The rapid technological revolution over the past fewdecades has provoked several complex ecological issues.The polymer industries consume huge amounts of volatileorganic solvents (VOC) as media or solvents, which areenvironmentally unfavorable. To deal with these environ-mental concerns, a number of studies have been performedon polymerization in supercritical fluids [1–3], ionic liq-uids [4–6], and aqueous media [7–9]. Among them, super-critical carbon dioxide (scCO2) has many commercial andindustrial applications because it is non-toxic, non-flam-mable, inexpensive, environmentally benign, and inert[10]. The excellent transport properties of scCO2, such asgas-like diffusivity and liquid-like density, are the most

promising features, making scCO2 more suitable for manymore reactions than environmentally hazardous VOCs.The properties of scCO2 can be tuned to a significant extentabove the critical conditions by changing the temperatureand pressure [11]. Additionally, the critical condition(31.1 �C and 7.39 MPa) of CO2 is reached readily, and CO2

can be separated easily from the product by depressurizingit to a lower pressure [12]. Consequently, scCO2 has drawnconsiderable interests to be used as a medium for polymer-ization. In recent years, dispersion, solution, emulsion, andprecipitation radical polymerizations have been success-fully performed in scCO2 [2]. Owing to the growingdemand for polymers with controlled molecular weightdistribution (MWD) or functional end groups, the use ofcontrolled radical polymerization (CRP) in scCO2 hasstarted to attract attention. The term ‘CRP’ was used todescribe polymerization techniques that enable the forma-tion of polymers with predetermined molecular weights

94 M.N. Islam et al. / European Polymer Journal 61 (2014) 93–104

and controlled MWD. Recently, several methods of CRPsuch as nitroxide-mediated polymerization (NMP) [13],atom transfer radical polymerization (ATRP) [14], andreversible addition-fragmentation chain transfer (RAFT)polymerization [15] have been performed in scCO2 thatoffer much better control over polymer microstructure.Among the three methods, ATRP has been extremely suc-cessful in synthesizing well-defined polymers with a con-trolled polydispersity index (PDI) and novel polymerarchitecture [16,17]. It can also produce polymers withhigh chain end-functionality that may serve as macroiniti-ators in the synthesis of block copolymers and participatein various post-polymerization modifications [18,19].Therefore, ATRP is considered the most promising amongall of the CRP methods.

Poly(vinyl esters) are used selectively as precursors toprepare poly(vinyl alcohol) (PVA). PVA is a hydrophilicand water soluble polymer, and is used widely in a rangeof applications, such as membranes, clothes, binders, films,and medicines for drug delivery system. These esters alsohave very good physical properties, such as high tensileand compressive strengths, high tensile modulus, and goodabrasion resistance, owing to its high crystalline latticemodulus. The physical properties of PVA are controlledby the molecular weight and stereoregularity. PVA can beobtained easily from poly(vinyl pivalate) (PVPi) via asaponification reaction [20]. PVA synthesized from PVPihas the highest syndiotacticity among the (PVA)s obtainedby radical polymerization [21]. Vinyl pivalate (VPi), whichhas a similar structure to that of vinyl acetate, is one of themost representative monomers that can be polymerizedonly by a free-radical polymerization [20,22–24]. On theother hand, controlling the radical polymerization of VPiis a challenge because unlike methacrylates and styrene,VPi lacks a conjugating substituent and its propagatingradicals are quite reactive, less stable, and tend to undergochain transfer and termination reactions. Living radicalpolymerization methods have been investigated to controlthe polymerization of VPi, such as the reversible addition-fragmentation chain-transfer (RAFT) polymerization [15]and telomerization of VPi followed by block co-polymeri-zation with styrene [25]. Recently our group reported thedispersion polymerization of VPi and CRP of vinyl acetate

Fig. 1. Schematic diagram of the experimen

in scCO2 [26,27]. Controlled synthesis of PVPi homopoly-mer [15], PVAc-s-PVPi [15,28] random copolymer, PVPi-b-PVAc and PVAc-b-PVPi block copolymer [28] by RAFTpolymerization in scCO2 were reported elsewhere.However, to the best of the authors’ knowledge, this isthe first effort of synthesizing PVPi by ATRP in scCO2.

In this study, we examined the optimal reaction condi-tions to synthesize PVPi in scCO2 with a controlled num-ber-average molecular weight (Mn) and PDI. The aims ofthis study were to (i) determine the optimal reaction con-ditions to obtain polymers with a controlled Mn and MWD,(ii) extend the chain length of the PVPi macroinitiator and(iii) saponification of PVPi.

2. Experimental

2.1. Materials

Vinyl pivalate (99%, Aldrich) was passed through an alu-mina column to remove the inhibitor and dried over cal-cium hydride (CaH2). Subsequently, a vacuum distillationwas performed and the purified monomer was storedunder nitrogen atmosphere in a freezer before using inthe polymerization. Ethyl 2-bromoisobutyrate (EBiB,98%), CuBr (99.99%), CuCl (99.99%) and 2,2:60;200-terpyri-dine (tpy, 98%) were purchased from Aldrich and used asreceived. Ultra-high purity CO2 gas (99.999%) was pur-chased from Deokyang Energy Corp.

2.2. Polymerization of vinyl pivalate in scCO2

The schematic diagram of the experimental setup forthe ATRP of VPi in scCO2 is shown in Fig. 1. The polymeri-zation reaction was carried out in a 20 mL stainless steelreactor, which was deoxygenated inside a glove box beforestarting the experiment. The reactor was then chargedwith a stirring bar, a catalyst, either CuBr (0.105 mmol or0.015 g) or CuCl (0.105 mmol or 0.0105 g), and tpy(0.105 mmol or 0.0245 g). The monomer, VPi (52.3 mmolor 7.7 ml), was bubbled with nitrogen for approximately30 min to remove the dissolved oxygen and added to thereactor. Subsequently, the reactor was purged with6.9 MPa CO2 and the reaction mixture was stirred for

tal setup for the ATRP of VPi in scCO2.

Scheme 1. Proposed mechanism of ATRP of VPi in scCO2.

M.N. Islam et al. / European Polymer Journal 61 (2014) 93–104 95

approximately 30 min to allow proper complex formation.The reactor was again purged with CO2 gas at a low flowrate of 0.5–0.6 mL/min to remove the oxygen remainingin the reactor. Finally, the initiator (EBiB, 0.105 mmol or0.015 mL) was added to the reactor, which was thenheated to 70 �C in a water bath. CO2 was supplied to thereactor until the pressure reached a desired value usingan ISCO syringe pump, while stirring with a magnetic stir-rer. The temperature of the water bath was kept constantwithin ±0.01 �C using a thermostat. The reaction wasstopped after 24 h by cooling the reactor in an ice bathand CO2 gas was vented slowly. For the kinetic studies,the polymerizations were stopped after designated timesto obtain the samples after quenching.

After the polymerization, the polymer product wasretrieved as a viscous liquid. The polymer remaining inthe reactor was recovered by dissolving in THF, followedby precipitating in a methanol/water (4/1 V/V) mixture.The polymer product was dried overnight in a vacuumoven to remove the remaining solvents. The experimentaluncertainty was approximately 2–3%.

2.3. Chain extension of poly(vinyl pivalate)

The PVPi macroinitiator was prepared by the ATRP ofVPi, as described above, at a feed ratio of [VPi]:[EBiB]:[CuCl]:[tpy] = 125:1:0.25:0.25. Polymerization was stopped

Table 1Reaction kinetics of PVPi synthesized by ATRP in scCO2.

Catalyst system Entrya Time (h) Conversionb (

CuBr/tpy 1 3 82 6 183 12 354 18 505 24 606 36 727 48 818 60 88

CuCl/tpy 1 3 112 6 223 12 414 18 605 24 706 36 847 48 90

a Feed mole ratio = [VPi]:[EBiB]:[CuX]:[tpy] = 500:1:1:1, pressure = 34.5 MPa,b Determined gravimetrically.c Mn,th = ([M]0/[I]0) �Mw,VPi � conversion.

after 10 h at a conversion of 30% (Mn = 6.9 kg/mol,PDI = 1.21). For the self-chain extension reaction, the PVPimacroinitiator (0.105 mmol or 0.724 g), CuCl (0.105 mmolor 0.0105 g), tpy (0.105 mmol or 0.0245 g), and VPi(52.3 mmol or 7.7 mL) were charged into the reactor,which was then agitated with a magnetic stirrer. Subse-quently, polymerization was conducted in scCO2 at 70 �Cand 34.5 MPa for 12 h. When polymerization reachedapproximately 31% conversion, the polymer product wascollected and purified. Later, PVPi-b-PVAc block copolymerwas prepared by the ATRP of VAc, as described above,using the PVPi macroinitiator (Mn = 6.9 kg/mol,PDI = 1.21). For synthesizing PVPi-b-PVAc block copolymer,the PVPi macroinitiator (0.105 mmol or 0.725 g), CuCl(0.105 mmol or 0.0105 g), tpy (0.105 mmol or 0.0245 g),and VAc (52.3 mmol or 4.8 mL) were charged into the reac-tor, which was then agitated with a magnetic stirrer.Subsequently, polymerization was conducted in scCO2 at60 �C and 34.5 MPa for 10 h and the polymer productwas collected and purified. Finally, the polymer samplewas analyzed by GPC to determine the Mn and PDI.

2.4. Saponification of PVPi

In a typical experiment, 2.0 g of PVPi was dissolvedin 200 mL of THF in a three-neck round-bottom flaskequipped with a reflux condenser and dropping funnel.

%) Mn,GPC (kg/mol) Mn,thc (kg/mol) PDI

13.3 5.1 1.9018.2 11.6 1.8729.1 22.5 1.8042.6 32.1 1.7048.1 38.5 1.6654.2 46.2 1.6160.6 52.0 1.5567.3 56.5 1.47

11.1 7.2 1.7118.1 14.2 1.6629.3 26.4 1.5740.6 38.6 1.5046.4 44.9 1.4355.2 53.9 1.3360.2 57.8 1.29

temperature = 70 �C.

Table 2Effect of pressure on the ATRP of VPi in scCO2.

Entrya Pressure (MPa) Conversionb (%) Mn,GPC (kg/mol) Mn,thc (kg/mol) PDI

1 13.8 25 27.3 16.1 1.772 20.7 38 31.1 24.4 1.713 27.6 51 37.5 32.7 1.614 31.0 60 41.2 38.5 1.455 34.5 70 46.4 44.9 1.43

a Feed ratio = [VPi]:[EBiB]:[CuCl]:[tpy] = 500:1:1:1; temperature = 70 �C, time = 24 h.b Determined gravimetrically.c Mn,th = ([M]0/[I]0) �Mw,VPi � conversion.

0 10 20 30 40 50 600.0

0.5

1.0

1.5

2.0

2.5(a) CuCl

CuBr

ln([M

] o/[M

])

Time (h)

Fig. 2. Plot of ln([M0]/[M]) as a function of time (a) and Mn,GPC and PDI as a function of conversion (b) for the ATRP of VPi. Feedratio = [VPi]:[EBiB]:[CuCl]:[tpy] = 500:1:1:1; pressure = 34.5 MPa, temperature = 70 �C.

12 15 18 21 24Elution time (min)

6 h (Mn,GPC = 18.1 kg/mol, PDI = 1.66)

12 h (Mn,GPC = 29.3 kg/mol, PDI = 1.57)

36 h (Mn,GPC = 55.2 kg/mol, PDI = 1.33)48 h (Mn,GPC = 60.2 kg/mol, PDI = 1.29)

24 h (Mn,GPC = 46.4 kg/mol, PDI = 1.43)

3 h (Mn,GPC = 11.1 kg/mol, PDI = 1.71)

Fig. 3. GPC curves of PVPi synthesized at different time intervals usingthe CuCl/tpy complex as the catalyst. Feed ratio = [VPi]:[EBiB]:[CuCl]:[tpy] = 500:1:1:1; pressure = 34.5 MPa, temperature = 70 �C.

96 M.N. Islam et al. / European Polymer Journal 61 (2014) 93–104

A 15 wt% NaOH solution was prepared in a methanol/water solution. Both solutions were deoxygenated by bub-bling with nitrogen. The alkaline solution was added drop-wise to the PVPi solution with constant stirring at 55 �C.After 3 h, the saponification reaction was completed,resulting in a light yellow colored PVA product. The prod-uct was filtered, washed several times with methanol, anddried in a vacuum oven at 40 �C.

2.5. Characterization

The Mn and PDI were estimated by gel permeation chro-matography (GPC) equipped with an isocratic pump(Waters 1515) and a refractive index (RI) detector (Waters2414). Two Waters Styragel columns (HR4E and HR5E) wereused in series with THF (HPLC grade) as the eluent and cali-brated using polystyrene standards (Mn = 3.07 � 103 �2.75 � 105 and PDI = 1.01–1.15). Calibration and analysiswere carried out at 35 �C and a carrier flow rate of1 mL/min. The nuclear magnetic resonance (NMR, Bruker,DPX-300) spectra of PVPi and PVA were recorded in CDCl3

and DMSO-d6 as solvents at 300 MHz. The monomerconversion was quantified gravimetrically. The Mn and PDIwere obtained by GPC. The following equation was used todetermine the degree of polymerization and the molecularweight of PVA [21]:

½g� ¼ 3:79� 10�3½Pn�0:84 ðin DMSO at 30 �CÞ ð1Þ

where [g] is the intrinsic viscosity and Pn is the numberaverage degree of polymerization. The syndiotactic-diad(S-diad) content and degree of saponification of PVA wascalculated from the NMR spectrum.

3. Results and discussion

ATRP of VPi was performed in scCO2 using a CuCl (orCuBr)/tpy complex system as a catalyst and EBiB as an

15 20 25 30 3520

30

40

50

60

70 (a)

Con

vers

ion

(%)

Pressure (MPa)10 15 20 25 30 35

25

30

35

40

45

50Mn,GPCPDI

Pressure (MPa)

Mn,

GPC

(kg/

mol

)

1.0

1.2

1.4

1.6

1.8

2.0(b)

PDI

Fig. 4. Conversion (a) and Mn,GPC and PDI (b) as a function of pressure for the ATRP of VPi. Feed ratio = [VPi]:[EBiB]:[CuCl]:[tpy] = 500:1:1:1;temperature = 70 �C, time = 24 h.

M.N. Islam et al. / European Polymer Journal 61 (2014) 93–104 97

initiator. Scheme 1 shows the mechanism of ATRP of VPi inscCO2. ATRP is a catalytic process that is controlled by theequilibrium between the two oxidation states of the tran-sition metal complexes. The radicals are generated by thehomolytic cleavage of an alkyl halogen bond (R–X) on thetransition metal catalyst, CuIX/tpy (activator) (governedby the activation rate constant, kact). These activated radi-cals do the propagation reaction with the monomers (VPi),followed by deactivation to produce the correspondingdormant alkyl halide (Pn–X) by the metal complex of ele-vated oxidation state, CuIIXBr/tpy (deactivator) (governedby the deactivation rate parameter, kdeact) [19]. AlthoughscCO2 is a non-polar solvent, the characteristic of ATRPmay not be altered significantly because the catalyst isnot bound to the growing chain. Therefore, the overall dif-ference in the catalytic activity is not too strong [29]. Thespeculated structures of transition metal catalysts of tpywith copper halide are also shown in Scheme 1.

A series of ATRP experiments were performed to obtainthe optimal reaction conditions that can affect the conver-sion, Mn, and PDI of the synthesized PVPi. The effects of thecatalyst, ligand concentration, reaction pressure, and tem-perature were measured experimentally.

3.1. Reaction kinetics

The reaction kinetics were examined for the two differ-ent catalytic systems (CuCl/tpy and CuBr/tpy) at 70 �Cand 34.5 MPa, with a feed ratio of [VPi]:[EBiB]:[CuX]:[tpy] = 500:1:1:1. To determine the reaction kinet-ics, a series of experiments were performed for differentreaction times (Table 1). Fig. 2a shows that the polymeriza-tion reactions exhibited first order kinetics in both cases, asshown in Eqs. (2) and (3)

Rp ¼ �d½M�=dt ¼ kp½P�n�½M� ¼ kappp ½M� ð2Þ

ln ½M�o=½M� ¼ kappp t ð3Þ

where Rp = the rate of polymerization, [M] = monomerconcentration, ½P�n� = the concentration of radical,

kp = propagation rate constant, and kpapp = apparent rate

constant. The apparent rate constant of the polymerization,kp

app, was calculated from the slopes in Fig. 2a. The value ofkp

app was 1.37 � 10�5 s�1 and 9.75 � 10�6 s�1 for CuCl andCuBr, respectively. Fig. 2a shows that CuCl/tpy has a stee-per slope than CuBr/tpy, indicating an approximately 40%higher polymerization rate.

For both polymerizations with the CuBr/tpy and CuCl/tpy catalytic systems, Mn increased linearly from 13.3 to67.3 and 11.1 to 60.2 kg/mol with conversion respectively(Fig. 2b). PDI decreased from 1.90 to 1.47 for the CuBr/tpycatalyst, whereas it decreased from 1.71 to 1.29 for theCuCl/tpy catalyst (Fig. 3b). The CuBr/tpy catalyst producedpolymers with broader PDIs and larger deviations from theestimated molecular weights. On the other hand, the con-trollability and initiator efficiency were better for the CuCl/tpy system. CuCl/tpy catalyst with the EBiB initiatorshowed faster initiation and slower propagation, which isessential for synthesizing a polymer with narrow polydis-persity [30]. Therefore, an efficient halide exchange reac-tion could take place between CuCl and EBiB. During thepolymerization reaction, the dormant polymer chain wasCl-capped instead of Br-capped due to the exchange ofhalide. The C–Cl bond energy was higher than that of C–Br. Therefore, the concentrations of the active propagatingradicals become lower than those in the C–Br/Cu–Br sys-tem, indicating a slower propagation rate. Initiation wasfaster in R–Br/CuCl compared to R–Cl/CuCl due to theweaker C–X bond in the former [31]. Overall, the EBiB/CuClsystem allowed better controllability over Mn and a fasterpolymerization rate. A controlled PDI was obtained athigher conversion in both cases [32,33].

PDI ¼ 1þ kp½RX�0kdeact ½CuIIX�

!2P� 1

� �ð4Þ

where kp is the propagation rate constant, kdeact is thedeactivation rate constant, [RX]o is the initial initiator con-centration, and p is the conversion. The other factors wereconstant except for the conversion for a fixed reaction con-dition, i.e. temperature, pressure, monomer, initiator, and

Table 3Effect of temperature on the ATRP of VPi in scCO2.

Entrya Temperature (�C) Conversionb (%) Mn,GPC (kg/mol) Mn,thc (kg/mol) PDI

1 50 55 37.3 35.3 1.402 60 64 42.3 41.1 1.413 65 68 44.9 43.7 1.414 70 70 46.4 44.9 1.435 75 70 48.2 44.9 1.51

a Feed ratio = [VPi]:[EBiB]:[CuCl]:[tpy] = 500:1:1:1; pressure = 34.5 MPa, time = 24 h.b Determined gravimetrically.c Mn,th = ([M]0/[I]0) �Mw,VPi � conversion.

45 50 55 60 65 70 7545

50

55

60

65

70(a)

Con

vers

ion

(%)

Temperature (°C)50 55 60 65 70 75

30

35

40

45

50

Temperature (°C)

Mn,

GPC

(kg/

mol

)

1.3

1.4

1.5

1.6

1.7

PDI

(b) Mn,GPC PDI

Fig. 5. Conversion (a) and Mn,GPC and PDI (b) as a function of temperature for the ATRP of VPi. Feed ratio = [VPi]:[EBiB]:[CuCl]:[tpy] = 500:1:1:1;pressure = 34.5 MPa, time = 24 h.

Table 4Effect of temperature on the ATRP of VPi in scCO2.

Entrya Temperature (�C) Conversionb (%) Mn,GPC (kg/mol) Mn,thc (kg/mol) PDI

1 65 79 13.9 12.7 1.212 70 85 15.2 13.6 1.233 75 88 16.1 14.1 1.21

a Feed ratio = [VPi]:[EBiB]:[CuCl]:[tpy] = 125:1:0.25:0.25; pressure = 34.5 MPa, time = 24 h.b Determined gravimetrically.c Mn,th = ([M]0/[I]0) �Mw,VPi � conversion.

98 M.N. Islam et al. / European Polymer Journal 61 (2014) 93–104

catalyst concentration. This means that a more uniformpolymer can be obtained at higher conversions. At the ini-tial stage of polymerization, the slow shifting rate of theprimary radicals into dormant species might explain thebroader polydispersity [34]. In the case of the CuBr/tpy cat-alyst system, a larger deviation between Mn,GPC and Mn,th

compared to CuCl/tpy was observed. This is because a con-siderable amount of termination reaction might have takenplace because the C–Br bond is weak. This means that atthe early stage of polymerization, more terminationoccurred via primary radicals, resulting in larger deviationbetween Mn,GPC and Mn,th [30].

Fig. 3 shows the GPC traces of PVPi obtained at differentreaction times, which shows the evolution of Mn and MWDwith time. The GPC traces were all mono-modal andmoved completely toward the higher molecular weight

region (Mn = 11.1–60.2 kg/mol) and the MWD became nar-rower (Mw/Mn = 1.71–1.29) with time (3–48 h). Birkin et al.[15] also reported the synthesis of PVPi homopolymer ofvarying Mn (5.0–20.5 kg/mol) with good control overMWD (PDI = 1.3–1.5), which is quite consistent with thePDI data at maximum conversion of this work, by employ-ing RAFT technique in bulk.

3.2. Effect of pressure

To examine the effects of pressure, a constant molarfeed ratio of [VPi]:[EBiB]:[CuCl]:[tpy] = 500:1:1:1 wasemployed at 70 �C. Five individual reactions were per-formed at pressures ranging from 13.8 to 34.5 MPa, eachfor 24 h. Fig. 4a shows that the conversion increased line-arly with pressure. The solubility of monomer, initiator

Table 5Effect of catalyst to ligand ratio on the ATRP of VPi in scCO2.

Entrya Catalyst/Ligand molar ratio Conversionb (%) Mn,GPC (kg/mol) Mn,thc (kg/mol) PDI

1 0.25:0.25 29 30.8 18.6 1.612 0.5:0.5 45 36.3 28.9 1.533 0.75:0.75 56 41.3 36.0 1.504 1:1 70 46.4 44.9 1.435 2:2 55 41.6 35.3 1.41

a [VPi]:[EBiB] = 500:1, pressure = 34.5 MPa, temperature = 70 �C, time = 24 h.b Determined gravimetrically.c Mn,th = ([M]0/[I]0) �Mw,VPi � conversion.

M.N. Islam et al. / European Polymer Journal 61 (2014) 93–104 99

and catalyst/ligand complex increased significantly withpressure. This might explain the enhanced polymerizationrate with increasing pressure. The increase in the polymer-ization rate might also be caused by the increase in theequilibrium constant (KATRP) and the propagation rate con-stant (kp) [35], which is related to the increase in the con-centration of radicals [36]. KATRP can be defined as the ratioof activation constant (kact) and deactivation constant(kdeact).

Fig. 4b shows an almost linear increase in Mn with con-version (assuming the relationship of conversion and pres-sure). The agreement between the theoretical Mn,th andexperimental Mn,GPC was better at higher pressures, proba-bly due to the higher initiator efficiency (>90%) [36,37]. PDIdecreased from 1.77 to 1.43, whereas the conversionincreased from 25% to 70%. The PDI decreased with pres-sure, resulting in a PDI of 1.43 at the highest pressure of34.5 MPa. This might be due to the increase in the propaga-tion rate and the decrease in the termination rate withincreasing pressure. Propagation is a chemically controlledbimolecular reaction, which has a negative volume of acti-vation and is accelerated at higher pressures [38]. On theother hand, termination is a diffusion-controlled reactionthat has a positive activation volume and is dropped offat higher pressures [38]. Moreover, under higher pressures,the total living chains became significantly higher com-pared to the dead chains, which might also lead to lowpolydispersity [39]. Similar results were also reported else-where [35,36,40].

3.3. Effect of temperature

To examine the effects of temperature on the polymer-ization, a series of experiments were performed at differ-ent temperatures (50, 60, 65, 70, and 75 �C). Table 3 liststhe conversion, experimental Mn, PDI, and theoretical Mn

Table 6Effect of the ligand concentration on the ATRP of VPi in scCO2.

Entrya Ligand concentration (molar) Conversionb (%)

1 0.25 252 0.5 453 1 704 2 60

a [VPi]:[EBiB]:[CuCl] = 500:1:1, pressure = 34.5 MPa, temperature = 70 �C, timeb Determined gravimetrically.c Mn,th = ([M]0/[I]0) �Mw,VPi � conversion.

data. Fig. 5a shows that the polymerization rate increasedalmost linearly at lower temperatures (50, 60 and 65 �C).On the other hand, above 65 �C, the increasing rate withtemperature was insignificant. Although the activation rateis higher at higher temperatures [41], the solvent strengthof scCO2 was reduced (constant pressure). Therefore, whenthe temperature was increased from 70 to 75 �C, the con-version remained nearly constant.

In Fig. 5b, Mn increased linearly with the level of conver-sion due to the controlled polymerization. A suddenincrease in PDI at 75 �C showed that the polymerizationlost control at high temperatures. At higher temperature,more precipitation of the polymer might have occurredat earlier times due to the smaller density of CO2, leadingthe larger PDI.

To see the effect of density of CO2, thus the solubility ofpolymer, three experiments with a smaller feed ratio([VPi]:[EBiB] = 125:1, with monomer only 25% of those inTable 3) were conducted at 65, 70 and 75 �C (Table 4).The conversions were 10–20% higher than those in Table 3,but Mn was much lower (approximately 14–16 kg/mol,about 1/3 of those in Table 3) and PDI was also quite lower(near 1.2) than those shown in Table 3. The conversion aswell as Mn increased with temperature but PDI remainedalmost constant. As the solubility of polymers increasedwith the density of CO2, the polymers with much lowerMn were not precipitated and continued the polymeriza-tion until the monomers almost used up. Therefore, wecan say that the polymers in Table 3 were precipitatedout due to the reduced density at higher temperature(75 �C) and the PDI increased. A faster propagation andchain transfer may also have resulted partly in less accu-rate control of polymerization. At a high temperature of75 �C, the difference between the calculated Mn,th andexperimental Mn,GPC was significant [42]. This might bepossible because a significant amount of initiator was

Mn,GPC (kg/mol) Mn,thc (kg/mol) PDI

24.5 16.1 1.6831.9 28.9 1.5346.4 44.9 1.4342.4 38.5 1.43

= 24 h.

Table 7Effects of the catalyst concentration on the ATRP of VPi in scCO2.

Entrya Catalyst concentration (molar) Conversionb (%) Mn,GPC (kg/mol) Mn,thc (kg/mol) PDI

1 0.25 30 27.9 19.3 1.802 0.5 40 35.1 25.7 1.663 1 70 46.4 44.9 1.434 2 69 45.3 44.3 1.455 3 69 45.6 44.3 1.43

a [VPi]:[EBiB]:[tpy] = 500:1:1, pressure = 34.5 MPa, temperature = 70 �C, time = 24 h.b Determined gravimetrically.c Mn,th = ([M]0/[I]0) �Mw,VPi � conversion.

Fig. 6. 1H NMR (300 MHz) spectrum of PVPi synthesized in scCO2 with CDCl3 as solvent. Polymerization condition: [VPi]:[EBiB]:[CuCl]:[tpy] = 125:1:0.25:0.25; pressure = 34.5 MPa, temperature = 70 �C, time = 24 h.

100 M.N. Islam et al. / European Polymer Journal 61 (2014) 93–104

consumed during the formation of XMtn+1/L, thus molecular

weights increased linearly with conversion but werehigher than the Mn,th, i.e., the initiator efficiencies were lessthan unity; such behavior were also reported by Qiu et al.[43] for the reverse ATRP of butyl methacrylate.

3.4. Effect of the catalyst/ligand ratio

A series of polymerization reactions were carried out todetermine the optimal catalyst to ligand ratio. The catalystto ligand molar ratio (CuCl/tpy) was varied with the otherconditions left unchanged to identify the best equilibriumconditions between the dormant and active species. Table 5lists the experimental data.

The reaction rate increased with increasing CuCl/tpyratio, as expected. Owing to the high catalyst concentra-tion, the deactivation and the activation rates wereincreased, leading to a better MWD [40]. Subsequently,the polymerization rate decreased rapidly with increasingCuCl:tpy concentration from 1:1 to 2:2 (entry 4,5 andTable 5). One of the most promising factors for the catalystperformance is its solubility in the reaction mixture. Therequired concentration of both activator and deactivator

was determined by the solubility [30]. The tpy ligand hasa bulky and rigid structure, which retards the solubilityof the complex formed, if large amounts are used [32]. Thismight be the prime reason for the decrease in reaction rate.On the other hand, Kickelbick and Matyjaszewski [30]demonstrated that the coordination around the copper inCuII complex was trigonal bipyramidal with the same bondlength for both Cu–Br. The CuII complex can be forced intoa distorted trigonal–bipyramidal structure or a squarepyramidal structure due to the presence of excess ligand[32].

For further investigation, only the ligand concentrationwas varied, while keeping other factors constant. A similarphenomenon for catalyst to ligand ratio was observed. Thepolymerization rate decreased when the ligand to catalystconcentration ratio (molar) was increased two fold (entry4, Table 6). The solubility of the tpy ligand decreased athigher concentrations. The formation of the complex can-not be explained by an excess concentration of tpy ligand(2-fold with respect to the catalyst).

A ligand to catalyst concentration ratio of 1:1 (entry 4,Table 6) was found to be more favorable due to the geom-etry and structure of the complex formed [30,44]. The

12 15 18 21 24 27

Mn,GPC = 27.2 kg/mol, PDI = 1.40

Elution time (min)

Mn,GPC = 6.9 kg/mol, PDI = 1.21

(a)(b)

Fig. 7. GPC curves of the (a) PVPi macroinitiator (short-dashed line) and(b) chain-extended PVPi (solid line). Polymerization conditions: (a)[VPi]:[EBiB]:[CuCl]:[tpy] = 125:1:0.25:0.25, P = 34.5 MPa, T = 70 �C,time = 10 h, conversion = 30%; (b) [VPi]:[PVPi-Br]:[CuCl]:[tpy] =500:1:1:1, P = 34.5 MPa, T = 70 �C, time = 12 h, conversion = 31%.

12 15 18 21 24 27Elution time (min)

Mn,GPC = 6.9 kg/mol, PDI = 1.21

(b) (a)

Mn,GPC = 23.5 kg/mol, PDI = 1.60

Fig. 8. GPC curves of the (a) PVPi macroinitiator and (b) PVPi-b-PVAcblock copolymer. Polymerization conditions: (a) [VPi]:[EBiB]:[CuCl]:[tpy] = 125:1:0.25:0.25, P = 34.5 MPa, T = 70 �C, time = 10 h, con-version = 30%; (b) [VAc]:[PVPi-Br]:[CuCl]:[tpy] = 500:1:1:1, P = 34.5 MPa,T = 60 �C, time = 10 h, conversion = 21%.

M.N. Islam et al. / European Polymer Journal 61 (2014) 93–104 101

residual ligand might have been unreacted and remainedinsoluble in the reaction mixture, adversely affecting thereaction. The reduction of CuII–CuI might occur in the pres-ence of excess ligand, leading to an increase in the deacti-vation rate because the ligand might have served as areducing agent.

Furthermore, to reconfirm this phenomenon, the cata-lyst concentration was varied with the other factors leftunchanged (Table 7). When small amounts of catalyst wereused, the reaction rate and controllability over both Mn andMWD were quite poor. At a higher catalyst concentrations

(entry 4 and 5, Table 7), the change in reaction rate wasnegligible. Although higher concentrations of CuCl havebeen used, it does not adversely affect the results due tothe good solubility in the reaction mixture. A completelydifferent trend was observed in this case compared to theeffect of the ligand on the polymerization reaction.

3.5. End group analysis

The structure of PVPi synthesized in scCO2 was exam-ined by 1H NMR (Fig. 6). The peak (d) at 1.0–1.2 ppm wasattributed to the methyl (–CH3) protons of the polymer.The peaks at 4.75 (c) and at 1.7 ppm (e) were assigned tothe methine (>CH–) and the methylene (–CH2–) protonsof the VPi repeating units in the backbone of the polymer,respectively [6]. The peak (f) at 4.05 ppm was assigned tothe methylene protons of the ethyl ester derived fromthe initiator EBiB [45,46]. The small peak (g) at 0.8 ppmwas attributed to the methyl protons of the initiator. Thesignal (b) at 2.1 ppm corresponded to the methylene pro-tons of the monomer unit [25]. The resonance (a) atapproximately 5.85 ppm attributed to the methine protonadjacent to terminal bromine at the x-end, which indicatesthe preservation of end group [46].

3.6. Chain extension of PVPi

To confirm the ‘‘livingness’’ of the system, chain exten-sion polymerization of VPi was performed using a PVPimacroinitiator and CuCl/tpy as the catalyst in scCO2 at70 �C and 34.5 MPa. The PVPi macroinitiator was preparedby conducting ATRP of VPi with a feed ratio of [VPi]:[E-BiB]:[CuCl]:[tpy] = 125:1:0.25:0.25. Polymerization wasstopped after 10 h at a conversion of 30%. The Mn andPDI of the PVPi macroinitiator were 6.9 kg/mol and 1.21,respectively. This PVPi macroinitiator was used in the freshfeed to synthesize the chain extended PVPi in scCO2. Poly-merization was stopped after reaching approximately 31%conversion (12 h). Fig. 7 shows the GPC curves of the PVPimacroinitiator and the resulting chain extended PVPi. Theelution peak of the macroinitiator clearly shifted to ahigher molecular weight (Mn = 27.2 kg/mol, PDI = 1.40).

A block copolymer PVPi-b-PVAc was synthesized by theATRP of VAc in scCO2 at 60 �C and 34.5 MPa by using thePVPi macroinitiator (Mn = 6.9 kg/mol, PDI = 1.21). The poly-merization was stopped after 10 h and the conversion wasmeasured. The Mn and PDI of the resulting block copolymerwere 23.5 kg/mole and 1.60, respectively. The formation ofthe block copolymer was confirmed by a clear shift in thechromatogram towards higher molecular weight afterchain extension (Fig. 8). Therefore, the livingness of thePVPi macroinitiator was further verified by the successfulchain-extension reaction. A series of random copolymersand block copolymers of different molecular weights weresynthesized by RAFT polymerization in toluene [15] andscCO2 [28], respectively. The PDIs of their block copoly-mers are ranged between 1.4 and 1.6, which are similarto what we have obtained in this study by ATRP. Later,these copolymers were employed as stabilizers in the dis-persion polymerization of N-vinyl pyrrolidone, whichyielded polymers with high conversion and molecular

0..51.01.52.02.53.0 3.54.04.5 5.05.56.0 6.57.07.58.0ppm

(a ) (b )

(a )

(b )

(c )

(c )(d )

(d )

Fig. 9. 1H NMR (300 MHz) spectrum of PVPi-b-PVAc synthesized in scCO2 with CDCl3 as solvent. Polymerization condition: [VAc]:[PVPi-Br]:[CuCl]:[tpy] = 500:1:1:1; pressure = 34.5 MPa, temperature = 60 �C, time = 10 h.

Fig. 10. 1H NMR (300 MHz) spectrum of PVA obtained from PVPisynthesized in scCO2 with DMSO as solvent. Polymerization condition(PVPi): [VPi]:[EBiB]:[CuCl]:[tpy] = 500:1:1:1;pressure = 34.5 MPa, tem-perature = 70 �C, time = 24 h. (mm: isotactic triad, mr: atactic triad andrr: syndiotactic triad).

O−C(1280)

(b)

Wavenumber (cm-1)

PVPi(a)

tert-butyl (1365)

3500 3000 2500 2000 1500 1000

C−O (1150)

Tran

smitt

ance

(%)

PVA

−OH (3300)

C=O(1730)

Fig. 11. FTIR spectra of (a) PVPi and (b) PVA saponified from the PVPi.

102 M.N. Islam et al. / European Polymer Journal 61 (2014) 93–104

weight [15,28]. In addition, Li et al. [25,47] reported thesynthesis of poly(VPi-b-St) by ATRP using VPi telomer asa macroinitiator, where the telomer was prepared by rad-ical telomerization of VPi using carbon tetrachloride.

The structure of PVPi-b-PVAc synthesized in scCO2 wasconfirmed by 1H NMR (Fig. 9). Typical segments of methy-lene (–CH2–) and methine (>CH–) protons of the VPi andVAc repeating units were clearly observed at (a)1.75 ppm and (b) 4.85 ppm, respectively. The peak (c) at1.0–1.2 ppm was attributed to the methyl (–CH3) protonsof the polymer. The peak (c) at about 1.20 ppm and peak(d) at 2.0 ppm were assigned to the methyl (–CH3) protonsof the PVPi and PVAc, respectively [46].

3.7. Saponification of PVPi

Saponification of a PVPi sample (entry 5, Table 2) wasperformed in a 15 wt% NaOH solution, resulting in PVAwith a Pn of 335 (experimental). The degree of saponifica-tion (DS), which is >99%, was determined by the ratio ofthe methylene proton peak area at 1.2–1.6 ppm to thetert-butyl proton peak area at 1.1 ppm (Fig. 10.). Thetriad tacticities calculated from the three OH peaks at4.1–4.7 ppm were 3.4, 71.14 and 25.46% for the isotactic,atactic and syndiotactic sequences, respectively [6]. Thesyndiotacticity (S-diad content) was calculated to be 61%by implementing Eq. (5), which is symbolized by the ratio

M.N. Islam et al. / European Polymer Journal 61 (2014) 93–104 103

of the racemic diad (r) content to the total racemic andmeso (m) content.

rð%Þ ¼ rr þ ðmr=2Þrr þmr þmm

� 100% ð5Þ

where mm, mr and rr are the peak areas corresponding tothe isotactic, atactic and syndiotactic triads.

FTIR spectra of the polymers (PVPi and PVA) were exam-ined to confirm the structures as well as the conversion ofPVPi to PVA. The IR spectrum of PVPi (Fig. 11) shows the typ-ical carbonyl peak at 1730 cm�1, C–O and O–C (two modesof the same bond) peak in the ester groups about 1150 cm�1

and 1280 cm�1, respectively, and the tert-butyl peak atapproximately 1365 cm�1 and 1395 cm�1 [25]. The PVAspectrum (Fig. 11) shows a broad peak of hydroxyl (–OH)groups at about 3300 cm�1 [46]. The absence of the car-bonyl peak, C–O peak in the ester group and the tert-butylpeak in the PVA spectrum after saponification indicates thatthe PVPi was almost completely converted to PVA.

4. Conclusions

PVPi with decent controllability over Mn and PDI weresynthesized by ATRP using scCO2 as the reaction medium.This is the first report of the controlled/living polymeriza-tion of VPi using a copper-based catalyst CuX/2,2:60;200ter-pyridine (CuX/tPy, X = Br or Cl). The living character ofPVPi was confirmed from 1H NMR spectrum and chain-extension polymerization. Polymerization at higherpressures yielded better controllability over PDI and ahigher reaction rate. On the other hand, at higher temper-atures, the rate of conversion increased but with inade-quate control. The kinetic plot showed that the CuCl/tpycatalytic system was approximately 40% steeper than thatof CuBr/tpy, indicating a higher polymerization rate,whereas Mn increased almost linearly with conversion. Theoptimal concentration of [catalyst]:[ligand] for producingwell-defined PVPi was found to be 1:1. CuCl was foundto be a more efficient catalyst for this system than CuBr.

Acknowledgement

This study was supported by DG Economic Circle Lead-ing Industry R&D Program of the Ministry of the Knowl-edge and Economy (MOKE) (R0001657).

References

[1] Grignard B, Jérôme C, Calberg CD, Jérôme R, Wang W, Howdle SM,et al. Dispersion atom transfer radical polymerization of vinylmonomers in supercritical carbon dioxide. Macromolecules2008;41(22):8575–83.

[2] Kendall JL, Canelas DA, Young JL, DeSimone JM. Polymerizations insupercritical carbon dioxide. Chem Rev 1999;99(2):543–63.

[3] Shim J-J, Yates MZ, Johnston KP. Polymer coatings by rapidexpansion of suspensions in supercritical carbon dioxide. Ind EngChem Res 1999;38(10):3655–62.

[4] Jelicic A, García N, L}ohmannsr}uben H-G, Beuermann S. Prediction ofthe ionic liquid influence on propagation rate coefficients in methylmethacrylate radical polymerizations based on Kamlet�Taftsolvatochromic parameters. Macromolecules 2009;42(22):8801–8.

[5] Qi M, Wu G, Sha M, Liu Y. Radiation induced polymerization of MMAin imidazolium ionic liquids and their mixed solutions with organicsolvents. Radiat Phys Chem 2008;77(10–12):1248–52.

[6] Nguyen VH, Haldorai Y, Pham QL, Noh SK, Lyoo WS, Shim J-J.Preparation of poly(vinyl pivalate) microspheres by dispersionpolymerization in an ionic liquid and saponification for thepreparation of poly(vinyl alcohol) with high syndiotacticity. EurPolym J 2010;46(11):2190–8.

[7] Simakova A, Averick SE, Konkolewicz D, Matyjaszewski K. AqueousARGET ATRP. Macromolecules 2012;45(16):6371–9.

[8] Percec V, Popov AV, Ramirez-Castillo E, Monteiro M, Barboiu B,Weichold O, et al. Aqueous room temperature metal-catalyzed livingradical polymerization of vinyl chloride. J Am Chem Soc2002;124(18):4940–1.

[9] Tsarevsky NV, Matyjaszewski K. ‘‘Green’’ atom transfer radicalpolymerization: from process design to preparation of well-definedenvironmentally friendly polymeric materials. Chem Rev2007;107(6):2270–99.

[10] Wells SL, DeSimone J. CO2 technology platform: an important toolfor environmental problem solving. Angew Chem – Int Ed2001;40(3):519–27.

[11] Leitner W. Supercritical carbon dioxide as a green reaction mediumfor catalysis. Acc Chem Res 2002;35(9):746–56.

[12] Cooper AI. Polymer synthesis and processing using supercriticalcarbon dioxide. J Mater Chem 2000;10(2):207–34.

[13] Aldabbagh F, Zetterlund PB, Okubo M. Improved control innitroxide-mediated radical polymerization using supercriticalcarbon dioxide. Macromolecules 2008;41(7):2732–4.

[14] Xia J, Johnson T, Gaynor SG, Matyjaszewski K, DeSimone J. Atomtransfer radical polymerization in supercritical carbon dioxide.Macromolecules 1999;32(15):4802–5.

[15] Birkin NA, Arrowsmith NJ, Park EJ, Richez AP, Howdle SM. Synthesisand application of new CO2-soluble vinyl pivalate hydrocarbonstabilisers via RAFT polymerisation. Polym Chem 2011;2(6):1293.

[16] Xia JH, Johnson T, Gaynor SG, Matyjaszewski K, DeSimone J. Atomtransfer radical polymerization in supercritical carbon dioxide.Macromolecules 1999;32(15):4802–5.

[17] Braunecker WA, Matyjaszewski K. Controlled/living radicalpolymerization: features, developments, and perspectives. ProgPolym Sci 2007;32(1):93–146.

[18] Queffelec J, Gaynor SG, Matyjaszewski K. Optimization of atomtransfer radical polymerization using Cu(I)/tris(2-(dimethylamino)ethyl)amine as a catalyst. Macromolecules 2000;33(23):8629–39.

[19] Matyjaszewski K. Atom transfer radical polymerization (ATRP):current status and future perspectives. Macromolecules2012;45(10):4015–39.

[20] Song DH, Lyoo WS. Preparation of high molecular weight poly(vinylpivalate) with high yield and high molecular weight poly(vinylalcohol) by emulsion polymerization of vinyl pivalate andsaponification. J Appl Polym Sci 2007;104(1):410–4.

[21] Kwark Y-J, Lyoo WS, Ha WS. Kinetics of emulsion copolymerizationof vinyl acetate and vinyl pivalate. Polymer 1996;28(10):851–9.

[22] Park J-Y, Shim J-J. Emulsion stability of PMMA particles formed bydispersion polymerization of methyl methacrylate in supercriticalcarbon dioxide. J Supercrit Fluids 2003;27(3):297–307.

[23] Lyoo WS, Kwak JW, Yeum JH, Ji BC, Lee CJ, Noh SK. Preparation ofultra high-molecular-weight syndiotactic poly(vinyl pivalate)monodisperse microspheres by low-temperature suspensionpolymerization of vinyl pivalate. J Polym Sci Part A: Polym Chem2005;43(4):789–800.

[24] Lyoo WS, Ha WS. Preparation of syndiotacticity-rich high molecularweight poly(vinyl alcohol) microfibrillar fiber by photoinitiated bulkpolymerization and saponification. J Polym Sci Part A: Polym Chem1997;35:55–67.

[25] Li GH, Cho CG. Synthesis and characterization of poly(vinyl alcohol-b-styrene)via atom transfer radical polymerization andsaponification. Macromol Res 2002;10(6):339–44.

[26] Pham QL, Nguyen VH, Haldorai Y, Shim J-J. Polymerization of vinylpivalate in supercritical carbon dioxide and the saponification forthe preparation of syndiotacticity-rich poly (vinyl alcohol). Korean JChem Eng 2013;30(5):1153–61.

[27] Islam MT, Haldorai Y, Nguyen VH, Islam MN, Ra CS, Shim J-J.Controlled radical polymerization of vinyl acetate in supercriticalCO2 catalyzed by CuBr/terpyridine. Korean J Chem Eng2014;31(6):1088–94.

[28] Birkin NA, Wildig OJ, Howdle SM. Effects of poly (vinyl pivalate)-based stabiliser architecture on CO2-solubility and stabilising abilityin dispersion polymerisation of N-vinyl pyrrolidone. Polym Chem2013;4(13):3791–9.

[29] Patten TE, Matyjaszewski K. Atom transfer radical polymerizationand the synthesis of polymeric materials. Adv Mater1998;10(12):901–15.

104 M.N. Islam et al. / European Polymer Journal 61 (2014) 93–104

[30] Kickelbick G, Matyjaszewski K. 4,40 ,400-tris(5-nonyl)-2,20: 60 ,200-terpyridine as ligand in atom transfer radical polymerization(ATRP). Macromol Rapid Comm 1999;20(6):341–6.

[31] Matyjaszewski K, Shipp DA, Wang JL, Grimaud T, Patten TE. Utilizinghalide exchange to improve control of atom transfer radicalpolymerization. Macromolecules 1998;31(20):6836–40.

[32] Matyjaszewski K, Gobelt B, Paik HJ, Horwitz CP. Tridentate nitrogen-based ligands in Cu-based ATRP: a structure-activity study.Macromolecules 2001;34(3):430–40.

[33] Semsarzadeh MA, Mirzaei A, Vasheghani-Farahani E, NekoomaneshHaghighi M. Atom transfer radical polymerization of(meth)acrylates and their novel block copolymers with vinylacetate. Eur Polym J 2003;39(11):2193–201.

[34] Ma J, Chen H, Liu D, Ji N, Zong G. Synthesis of polyacrylonitrile usingAGET-ATRP in emulsion. Mater Sci Eng 2013;33(1):570–4.

[35] Rzayev J, Penelle J. HP-RAFT: a free-radical polymerization techniquefor obtaining living polymers of ultrahigh molecular weights. AngewChem Int Ed 2004;43(13):1691–4.

[36] Morick J, Buback M, Matyjaszewski K. Effect of pressure onactivation–deactivation equilibrium constants for ATRP of methylmethacrylate. Macromol Chem Phys 2012;213(21):2287–92.

[37] Mueller L, Jakubowski W, Matyjaszewski K, Pietrasik J, KwiatkowskiP, Chaladaj W, et al. Synthesis of high molecular weightpolystyrene using AGET ATRP under high pressure. Eur Polym J2011;47(4):730–4.

[38] Arita T, Buback M, Janssen O, Vana P. RAFT-polymerization ofstyrene up to high pressure: rate enhancement and improvedcontrol. Macromol Rapid Comm 2004;25(15):1376–81.

[39] Arita T, Kayama Y, Ohno K, Tsujii Y, Fukuda T. High-pressure atomtransfer radical polymerization of methyl methacrylate for well-defined ultrahigh molecular-weight polymers. Polymer2008;49(10):2426–9.

[40] Rzayev J, Penelle J. Controlled/living free-radical polymerizationunder very high pressure. Macromolecules 2002;35(5):1489–90.

[41] Seeliger F, Matyjaszewski K. Temperature effect on activation rateconstants in ATRP: new mechanistic insights into the activationprocess. Macromolecules 2009;42(16):6050–5.

[42] Semsarzadeh MA, Daronkola MRR. Atom transfer radicalpolymerization of methyl acrylate with PVAc-CCl3 macroinitiatorand a novel synthesis of block terpolymers of PVAc-b-poly(MA-co-MMA). Iran Polym J 2006;15(10):829–39.

[43] Qiu J, Gaynor SG, Matyjaszewski K. Emulsion polymerization of n-butyl methacrylate by reverse atom transfer radical polymerization.Macromolecules 1999;32(9):2872–5.

[44] Pintauer T, Matyjaszewski K. Structural aspects of copper catalyzedatom transfer radical polymerization. Coord Chem Rev2005;249(11–12):1155–84.

[45] Yu Y-H, Liu X-H, Jia D, Cheng B-W, Zhang F-J, Chen P, et al. CuBr2/Me6TREN-mediated living radical polymerization of methylmethacrylate at ambient temperature. Polymer 2013;54(1):148–54.

[46] Tang H, Radosz M, Shen Y. Atom transfer radical polymerization andcopolymerization of vinyl acetate catalyzed by copper halide/terpyridine. AIChE J 2009;55(3):737–46.

[47] Li G-H, Cho C-G. Synthesis and micellar behavior of poly (vinylalcohol-b-styrene) copolymers containing PVA blocks with differentsyndiotacticity. Colloid Polym Sci 2005;283(9):946–53.