Free Radical Copolymerization Kinetics of γ-Methyl-α-methylene-γ

ARTICLE pubs.acs.org/Biomac

Free Radical Copolymerization Kinetics of γ-Methyl-r-methylene-γ-butyrolactone (MeMBL) Robert A. Cockburn,† Rebekka Siegmann,‡ Kevin A. Payne,† Sabine Beuermann,‡ Timothy F. L. McKenna,† and Robin A. Hutchinson*,† † ‡

Department of Chemical Engineering, Dupuis Hall, Queen's University, Kingston, Ontario K7L 3N6, Canada Institute of Chemistry, University of Potsdam, Karl-Liebknecht Str. 24-25, 14476 Potsdam/Golm, Germany

bS Supporting Information ABSTRACT: The propagation kinetics and copolymerization behavior of the biorenewable monomer γ-methyl-R-methylene-γ-butyrolactone (MeMBL) are studied using the pulsed laser polymerization (PLP)/size exclusion chromatography (SEC) technique. The propagation rate coefficient for MeMBL is 15% higher than that of its structural analogue, methyl methacrylate (MMA), with a similar activation energy of 21.8 kJ 3 mol
1. When compared to MMA, MeMBL is preferentially incorporated into copolymers when reacted with styrene (ST), MMA, and n-butyl acrylate (BA); the monomer reactivity ratios fit from bulk MeMBL/ST, MeMBL/ MMA, and MeMBL/BA copolymerizations are rMeMBL = 0.80 ( 0.04 and rST = 0.34 ( 0.04, rMeMBL = 3.0 ( 0.3 and rMMA = 0.33 ( 0.01, and rMeMBL = 7.0 ( 2.0 and rBA = 0.16 ( 0.03, respectively. In all cases, no significant variation with temperature was found between 50 and 90 °C. The implicit penultimate unit effect (IPUE) model was found to adequately fit the composition-averaged copolymerization propagation rate coefficient, kp,cop, for the three systems.

’ INTRODUCTION Recent investigations on the use of gamma lactone polymers,1
4 such as R-methylene-γ-butyrolactone (R-MBL, also known as tulipalin-A), as natural alternatives to petroleum sourced plastics are part of a broader focus on renewable polymers.5,6 R-MBL and two other closely related gamma lactones, γ-methyl-R-methyleneγ-butyrolactone (MeMBL) and β-methyl-R-methylene-γ-butyrolactone (MMBL), are cyclic analogues of methyl methacrylate (MMA), a vinyl monomer commonly used in automotive coatings, acrylic glasses, and medical technologies. The chemical structures of these monomers are shown in Figure 1. Although methods to isolate lactone monomers were known since the early 1950s, the first studies comparing the free-radical polymerization behavior of R-MBL and like monomers to MMA were not conducted until the late 1970s and early 1980s.7
13 Compared to poly(MMA), the R-MBL homopolymer was found to have improved solvent resistance and a significantly higher glass transition temperature (Tg of 195 vs 105 °C), while MeMBL monomer produces a polymer with an even higher Tg than that of R-MBL (210
220 °C).7,10,12 Molded gamma lactone polymers were clear, hard, and brittle, suggesting potential applications similar to those of poly(MMA), including dental resins, acrylic glasses, and coatings.7,10,12,14 Some copolymerizations of R-MBL with both styrene (ST) and MMA were also investigated at that time;8,12 however, largely due to the lack of a large scale and low cost method to synthesize the lactones,9,15 r 2011 American Chemical Society

little further research on the R-MBL family of monomers took place for a period of about 20 years. In 2004, DuPont developed a commercially viable catalytic two-step process to produce MeMBL from levulinic acid, a chemical intermediate derived from biomass.16 This development has rekindled interest in the potential uses of this biorenewable monomer. Information on the copolymerization kinetics of gamma lactone monomers is limited in the open literature, with research having been primarily focused on copolymers of R-MBL and MMBL with styrene.3,8,12 Recently Mosn
acek et al. used atom transfer radical polymerization to produce triblock polymers consisting of n-butyl acrylate (BA) soft segments and MMA/ R-MBL copolymers, but the kinetic behavior was not investigated.17 Qi et al. have examined the copolymerization of MeMBL and styrene (ST) at 70 °C,4 and we have recently studied the monomer reactivities and copolymerization propagation kinetics of MeMBL/ST using the pulsed laser polymerization/size exclusion chromatography (PLP-SEC) technique.1 MeMBL/ MMA monomer reactivity ratios were also determined, but the poor solubility of MeMBL-rich copolymer in tetrahydrofuran (THF) precluded the thorough investigation of propagation kinetics.1 This study combines PLP with SEC analyses conducted Received: March 23, 2011 Revised: April 20, 2011 Published: April 26, 2011 2319

dx.doi.org/10.1021/bm200400s | Biomacromolecules 2011, 12, 2319–2326

Biomacromolecules

ARTICLE

apparent reactivity ratios (r i) and the apparent homopropagation rate coefficients (kii) in a two-monomer system: kp, cop ¼ Figure 1. Structures of MMA and gamma lactone monomers R-MBL, MeMBL, and MMBL. R-MBL is a cyclic analogue to MMA, while MeMBL and MMBL differ from R-MBL only by the positioning of their exocyclic methylene groups.

using two solvents, N,N-dimethylacetamide (DMAc) and THF, to complete our description of MeMBL copolymerization behavior with ST and MMA, and to study for the first time the copolymerization of MeMBL with a monomer from the acrylate family, BA. The SEC calibration established for poly(MeMBL) in DMAc also allows the first determination of Arrhenius parameters for MeMBL homopropagation. PLP-SEC is a specialized technique to determine propagation rate coefficients through analysis of the molecular mass distributions (MMD) of polymers produced in a pulsed-laser setup. It has been used extensively to study homopropagation and copolymerization propagation kinetics of styrene with various methacrylates over a range of compositions and temperatures.18
26 Typical PLP experiments use a photoinitiator dissolved in a (co)monomer solution that forms radicals instantaneously upon exposure to UV light in a laser operated at a set pulse repetition rate. Those radicals that survive the dark period between laser flashes are typically terminated by the next pulse to form a significant population of dead polymer chains with a characteristic chain length of Lo. Lo ¼ kp ½Mto

ð1Þ

With the total monomer concentration [M] and dark time between two pulses to known, the copolymer-averaged propagation rate coefficient kp,cop is calculated from the experimental determination of Lo by SEC analysis of the polymer MMD. Features at 2Lo and sometimes 3Lo are also observed in the MMD, indicative of chains that are terminated at some other later flash; for example, at 2to and 3to. An overview of this technique, which has greatly improved knowledge of free radical polymerization kinetics (e.g., refs 18
26) is found in the review by Beuermann and Buback.27 The terminal model is often used to represent the composition of copolymer formed in copolymerization systems. Assuming that radical selectivity depends only on the terminal unit of the growing polymer chain, the instantaneous mole fraction of monomer-1 incorporated into the copolymer (F1inst) is only a function of monomer mole fractions (f1 and f2) and the monomer reactivity ratios: F1inst ¼

r1 f1 2 þ f1 f2 r1 f1 2 þ 2f1 f2 þ r2 f2 2

ð2Þ

where r1 = kp11/kp12, r2 = kp22/kp21, and kpij is the propagation rate coefficient for addition of monomer j to radical i. While the terminal model can typically describe copolymer composition, its ability to predict the copolymer-averaged propagation rate coefficient (kp,cop) as a function of monomer composition is often unsatisfactory, especially for systems containing ST.21
26 To rectify this, the penultimate unit representation was originally developed by Merz et al.28 to describe kp,cop, in which the identity of the penultimate unit on the polymer chain affects both the

r 1 f1 2 þ 2f1 f2 þ r 2 f2 2 ðr 1 f =k11 Þ þ ðr 2 f =k22 Þ

ð3Þ

For systems where the terminal model adequately represents copolymer composition, the penultimate unit is assumed to have a negligible effect on the selectivity of the radicals (r 1 = r1 and r 2 = r2), leading to the implicit penultimate unit effect (IPUE) model presented by Fukuda et al.29 In the IPUE, the effect of the penultimate unit on radical reactivity is captured through the introduction of radical reactivity ratios, defined by s1 = kp211/kp111 and s2 = kp122/ kp222, with k11 and k22 expressed as functions of monomer fraction, k11 ¼

kp111 ðr1 f1 þ f2 Þ r1 f1 þ ðf2 =s1 Þ

k22 ¼

kp222 ðr2 f2 þ f1 Þ r2 f2 þ ðf1 =s2 Þ

ð4Þ

where kp111 and kp222 are homopolymerization propagation rate coefficients, and kpijk is the propagation rate coefficient for the addition of monomer k to a growing radical j with unit i in the penultimate position. Note that the IPUE model reduces to the terminal model prediction for kp,cop when s1 = s2 = 1. In this study, values of the radical reactivity ratios are estimated by fitting the IPUE model to experimental kp,cop versus monomer composition data measured using the PLP-SEC technique. There is some debate in the literature regarding the use of the IPUE to describe propagation in copolymerization systems, as other models can also adequately represent the data.30,31 Indeed, it has even been proposed that monomer reactivity ratios (r values) should only be regarded as adjustable parameters, as they do not adequately reflect the fundamental mechanisms of copolymerization.32 However, Fischer and Radom have demonstrated that relative addition rates of monomer to monomeric radicals provide a good prediction of relative rates of monomer addition during polymerization; that is, copolymerization reactivity ratios can be regarded as fundamental parameters.33 In addition, ab initio calculations of r values show reasonable agreement with experimentally determined values.34,35 As shown below, these widely accepted models of copolymerization provide a good fit to our data for MeMBL copolymerization. Previous work found that MeMBL is significantly more reactive than both styrene and MMA,1,4,8,12 resulting in its preferential incorporation into the polymer and compositional drift over the course of a batch polymerization. However, the homopropagation kinetics of MeMBL have not been previously studied. The present work provides this missing information, as well as characterizing the copolymerization kinetics of MeMBL with BA and updating the copolymerization propagation behavior of MeMBL with MMA and styrene at 50 and 90 °C. The ultimate goal of this examination is to use the acquired knowledge to reduce compositional drift and produce copolymers of MeMBL with well-defined characteristics and tailored physical properties, such that MeMBL might displace some volume of petroleum-sourced monomers as well as extend the operating range and utility of acrylic glasses. In addition, the transesterification and ring-opening of pMeMBL has potential applications in drug delivery or pH-dependent substance release.3,4

’ EXPERIMENTAL SECTION Monomers used in this study include MMA (99% purity) inhibited with 10
100 ppm of monomethyl ether hydroquinone (MEHQ), 2320

dx.doi.org/10.1021/bm200400s |Biomacromolecules 2011, 12, 2319–2326

Biomacromolecules styrene (99% purity) inhibited with 10
15 ppm of 4-tert-butylcatechol, and BA (99% purity) inhibited with 10
60 ppm MEHQ, all purchased from Sigma Aldrich and used as received. MeMBL was obtained from DuPont Central Research Laboratories, inhibited with 50 ppm of hydroquinone (97.5% purity, major impurity γ-valerolactone, GVL) and used as received. The photoinitiator DMPA (2,2-dimethoxy-2phenylacetophenone, 99% purity) and chloroform-d with 99.8 atom % D were also obtained from Sigma Aldrich and used as received. A pulsed laser setup consisting of a Spectra-Physics Quanta-Ray 100 Hz Nd:YAG laser, capable of producing a 355 nm laser pulse of duration 7
10 ns with energy of 1
50 mJ per pulse was used to prepare low conversion homo- and copolymers of MeMBL with ST, BA, and MMA. Further specifics of the apparatus are described by Wang and Hutchinson.25 A Hellma QS100 optical sample cell was used as the PLP reactor, with bulk monomer mixtures containing 5 mmol 3 L
1 DMPA photoinitiator held at the desired temperature (controlled to (1.0 °C) during pulsing by a circulating oil bath. Experiments were conducted at 50 and 90 °C at varying frequencies (typically 33 or 50 Hz), with the MeMBL mole fraction in the comonomer mixtures systematically varied between 0 and 100%. Monomer conversion was controlled below 3% to avoid significant compositional drift. The molar mass distributions (MMD) of the PLP-generated polymer samples were measured by SEC after precipitating the polymers from bulk by the addition of methanol. Solid polymers were isolated by drying under a stream of forced air and were further dried in a vacuum oven at 65 °C for 24 h before being redissolved in a solvent, either tetrahydrofuran (THF) or N,N-dimethylacetamide (DMAc) depending on the SEC setup used. MeMBL/MMA and MeMBL/BA copolymers with greater than 50 and 60 mol % MeMBL, respectively, are not THFsoluble but are soluble in DMAc. Thus, samples spanning the complete composition range (including MeMBL homopolymer) were analyzed in a DMAc SEC setup at the University of Potsdam. A Waters 2960 separation module with Styragel packed columns HR 0.5, HR 1, HR 3, HR 4, and HR 5E (Waters Division Millipore) was used for THF SEC analyses at 35 °C. THF was used as the eluent at a flow rate of 1 mL 3 min
1 and detection was provided by a Waters 410 differential refractometer (DRI detector) and a Wyatt Instruments Dawn EOS 690 nm laser photometer multiangle light scattering (LS) unit.25,26 RI detector calibration was established with eight linear polystyrene standards of narrow dispersity covering a molecular weight range of 890 to 3.55  105 g 3 mol
1. The refractive index (dn/dc) of the polymer in THF is required to process the data from the LS detector and was measured as described previously.1 DMAc SEC analyses were performed using an Agilent 1200 isocratic pump, an Agilent 1200 differential refractive index detector, a WEG Dr. Bures ETA 2010 online viscosity detector, and three PSS analytical GRAM columns (8  300 mm, particle size 10 μm, pore sizes 100 Å and 2  3000 Å). DMAc containing 0.1% LiBr was used as eluent at 45 °C and a flow rate of 1 mL 3 min
1. The SEC setup was calibrated against 11 poly(ST) standards of narrow dispersity with molecular weights between 500 and 1  106 g 3 mol
1. A universal calibration curve was created using the online viscosity detector data. Proton NMR in deuterated chloroform was used to analyze the composition of polymers produced by PLP. Peak assignments for MeMBL/ST and MeMBL/MMA copolymers was discussed in Cockburn et al.1 To determine the composition of the MeMBL/BA copolymers, the characteristic peak of MeMBL (from the proton at the gamma carbon of MeMBL in the region of 4.4
4.8 ppm) was ratioed to the nbutyl acrylate peak at 4.1 ppm, representative of CH2 in the O
CH2 moiety of the butyl group. To analyze the outputs from the DMAc SEC, Mark
Houwink (MH) parameters are needed to transform the data from the relative polystyrene calibration. While MH parameters for ST and MMA are easily known from MH plots derived from coupled RI and online viscosity

ARTICLE

Figure 2. Plot of log of viscosity vs log of MW for MeMBL polymers in DMAc. Data are shown for MeMBL polymer produced via PLP (upper) and precipitation (lower) polymerization techniques. measurements of many narrow polymer standards, no narrow polymer standards for BA and MeMBL are commercially available. To rectify this, broad homopolymer samples of BA and MeMBL were analyzed by viscometry in DMAc to determine the appropriate constants. Samples analyzed to determine the Mark
Houwink parameters for MeMBL included polymer produced via PLP at 50 and 70 °C and by precipitation polymerization in organic media at 70 °C.36 Polymer samples for BA were produced by 100 Hz PLP experiments at 10 °C with 1 vol % n-dodecyl mercaptan as CTA to reduce branching. The samples were purified thrice by dissolving the polymer in acetone and precipitation in methanol. Three samples of the polymer produced at each condition were each injected twice at least. For each polymer sample the variation of intrinsic viscosity ([η]) with the elution volume, Ve, is derived on the basis of the precisely known polymer concentration. By using the aforementioned universal calibration curve and the measured intrinsic viscosity the absolute molecular weight at every elution volume and thus the MH plot are available, [η] = KMa. Taking the slopes and y-intercepts of the individual lines log(intrinsic viscosity) versus log(M) in Figure 2 provides an estimate for the a and K Mark
Houwink parameters, respectively; final estimates were obtained by averaging the values for each respective condition. When MeMBL homopolymer kp values were calculated with the various sets of Mark
Houwink parameters estimated, it was found that the difference between the highest values predicted (precipitation polymerization MH constants) and the lowest values predicted (50 °C PLP MH constants) was 12
13%, a typical uncertainty range for SEC. The kp estimates calculated using various MH parameters for MeMBL in DMAc were compared with poly(MeMBL/ST) THF results,1 and the data sets were found to be most consistent using the calibration from poly(MeMBL) produced by precipitation polymerization; this calibration was used for all of the MMD analyses in DMAc solvent in this work. All parameters required to estimate kp,cop from SEC data are summarized in Table 1. Monomer densities are calculated as a function of temperature. It was assumed that MeMBL density varied with temperature to the same extent as MMA, although the absolute value at 25 °C differs.1 A density function for BA was estimated assuming a linear dependency on temperature and density values at 20 and 60 °C.37 Outputs from RI detectors are transformed to absolute MW values assuming universal calibration and using a composition weighted average of the two homopolymer calibrations.21,24
26 THF LS results are normalized based on known polystyrene calibration standards. Thus, 2321

dx.doi.org/10.1021/bm200400s |Biomacromolecules 2011, 12, 2319–2326

Biomacromolecules

ARTICLE

Table 1. Parameters Required for the Calculation of kp,cop from the SEC Analysis of PLP-Generated Copolymer Samples of MeMBL with Styrene, BA, and MMA polymer Mark
Houwink parameters in THF
1

density F (g 3 mL ) monomer

polymer dn/dc in

K (dL 3 g
1) 

THF (mL 3 g
1)

at 25 °C

density function

in DMAc

10

K (dL 3 g
1) 

4

104

a

24

a

styrene

0.9193
0.000665T/°C

0.903

0.180

1.14

0.716

1.22

0.678

MMA

0.9569
1.2129  10
3 T/°C þ 1.6813  10
6 T2/°C þ 1.0164  10
8 T3/°C32

0.928

0.08938

0.94437

0.71937

1.44

0.663

MeMBL

1.2235
1.2129  10
3 T/°C þ 1.6813 

1.194

0.0981

0.5001

0.7191

2.63

0.579

0.894

0.06938

1.2237

0.70037

27.08

0.367

26


6

10 BA


8

T /°C þ 1.0164  10 2

0.9217
0.0011T/°C31

24

24

3

T /°C

Table 2. Monomer Reactivity Ratios for Copolymerization of MeMBL with Styrene, BA, and MMA Compared with MMA Copolymerization with the Same Monomersa r1

monomer 1|monomer 2 21

r2

MMA|ST

0.46

0.52

MeMBL|ST1 MMA|BA39

0.80 ( 0.04 2.55 ( 0.35

0.34 ( 0.04 0.36 ( 0.08

MeMBL|BA

7.3 ( 2.1

0.16 ( 0.03

MMA|MMA

1.0

1.0

MeMBL|MMA1

3.0 ( 0.3

0.33 ( 0.01

a

r1 refers to either MeMBL or MMA and r2 to the second monomer listed.

Figure 3. Mole fraction in copolymer (Fmonomer1) as a function of mole fraction in the monomer phase (fmonomer1) for low-conversion MeMBL/ BA (O,


), MeMBL/ST (0, 3 3 3 ) and MeMBL/MMA (4,
3
) bulk copolymerizations between 50 and 90 °C. Additionally, curves for MMA/MMA (—), MMA/ST (
3 3
), and MMA/BA (— —) are shown. All curves are calculated by the terminal model with associated reactivity ratios noted in Table 2. Monomer1 refers to either MeMBL or MMA and monomer2 to the second monomer listed. there are as many as three independent estimates of kp for some samples, from DMAc RI analysis and from THF analysis with RI and LS detectors.

’ RESULTS AND DISCUSSION The reactivity ratios for the MeMBL/BA system are estimated by fitting the copolymer versus comonomer composition data set covering the complete composition range at temperatures of 50, 70, and 90 °C. As found for MeMBL/ST and MeMBL/MMA,1 there was no systematic variation in composition with temperature in this range; thus the results are plotted in Figure 3 with error bars indicating the standard deviation from the averaged values measured at different temperatures. The terminal model fits copolymer composition for the MeMBL/BA system well, with rMeMBL = 7.0 ( 2 and rBA = 0.16 ( 0.03 estimated using nonlinear parameter estimation. Figure 3 also presents composition curves for MeMBL copolymerized with ST and MMA.1 The copolymerization of MeMBL with these three monomers is compared to that of MMA in Figure 3, with reactivity ratios

summarized in Table 2. In all cases, MeMBL is incorporated into the copolymer to a greater extent than MMA; that is, the curves for MeMBL copolymers lie above the corresponding MMA copolymer curves. The higher reactivity of MeMBL compared with MMA is also reflected in the reactivity ratios (Table 2). The most probable explanation for the higher reactivity, first reported by Pittman et al.3 in their copolymerization studies of MMBL, is that the ring structure of the gamma lactone monomer affects propagation kinetics through resonance stabilization and conjugation effects. The implication of these findings is that significant compositional drift will occur if MeMBL is copolymerized in a batch system and that a fed-batch system must be used to produce copolymers with constant composition.

’ HOMO- AND CO-POLYMERIZATION PROPAGATION KINETICS When well-structured MMDs (such as those in Figure 4) produced by the PLP-SEC technique are analyzed, kp,cop can be determined based on the first inflection point of the MMD according to: kp, cop ðL 3 mol
1 3 s
1 Þ ¼

MW o 1000Fto

ð5Þ

MWo is the polymer molecular weight at the first inflection point, and F (g 3 mL
1) is the density of monomer mixture, calculated from the known initial monomer composition assuming volume additivity and negligible conversion. A complete list of the experimental conditions and results of the PLP-SEC studies 2322

dx.doi.org/10.1021/bm200400s |Biomacromolecules 2011, 12, 2319–2326

Biomacromolecules

ARTICLE

Figure 4. MMDs (top) and corresponding first derivative (bottom) plots obtained for MeMBL/BA copolymers produced by PLP at 50 °C and 33 Hz, as measured by RI detector with DMAc eluent. Monomer compositions are given as mole fraction MeMBL (fMeMBL), with increasing fMeMBL values accompanied by a shift of the MMDs and first derivative curves to the right.

carried out from 50 to 90 °C is summarized in the Supporting Information tables, Table S1, Table S2, and Table S3 for the MeMBL/ST, MeMBL/MMA, and MeMBL/BA copolymer systems, respectively. The MeMBL/ST and MeMBL/MMA data analyzed by THF SEC were contained in our previous publication1 but are included for completeness. For MMA and ST copolymerized with MeMBL, most experiments were conducted at 33 Hz, and some experiments were run with pulse repetition rates of 50 and 20 Hz to check the consistency of the kp,cop estimates. (The PLP-SEC technique also provides a self-consistency check, as the MW value at the second inflection point should occur at a value double that of MWo.) MeMBL/BA experiments were typically run at 50 Hz due to the much higher reactivity of BA, with some lower temperature experiments run at 33 Hz. The kp,cop values obtained at different repetition rates were in good agreement. Typical polymer MMDs and corresponding first derivative curves obtained from PLP experiments conducted at 33 Hz and 50 °C for varying MeMBL/BA compositions are shown in Figure 4, as measured by the DMAc RI detector. As the mole fraction of MeMBL in the comonomer mixture increases, the MMDs shift to the right, and the corresponding MWo values increase. Similar well-structured MMDs with clear primary and secondary inflection points were obtained under all conditions examined. The previous estimates for the homopolymer kp values of MeMBL were obtained by extrapolating kp,cop results for MeMBL/ST (available to 80 mol % MeMBL).1 With poly(MeMBL) fully soluble in DMAc, it was possible to analyze the polymer produced by PLP over a range of temperatures (22 to 120 °C) and determine homopolymer propagation coefficients using the newly obtained Mark
Houwink parameters.

Figure 5 shows the MeMBL homopolymer kp values measured, along with the IUPAC recommended curves for MMA,19 ST,20 and BA;40 while the latter relationship is based on data obtained at lower temperatures (to 20 °C), it has recently been verified to 70 °C by application of a 500 Hz laser.41 The kp values for MeMBL are similar in magnitude to MMA, while the kp for ST is significantly lower with a higher activation energy and that for BA is much higher (by a factor of 40 at 50 °C) with a lower activation energy. On the basis of the best fit to the MeMBL data shown in Figure 5, the activation energy (Ea) and pre-exponential factor (A) are estimated as 21.77 kJ 3 mol
1 and 2.3  106 L 3 mol
1 3 s
1, respectively. These Arrhenius parameters are compared to those of ST, MMA, and BA in Table 3. Figure 5 and Table 3 show that the propagation kinetics of MeMBL and MMA monomers are very similar, with the activation energies comparable to many other methacrylates.18,27 The homopolymer MeMBL kp values determined are used in subsequent analysis of the MeMBL/ST, MeMBL/MMA, and MeMBL/BA copolymer data sets. At 50 °C, the homopolymerization kp,cop values of MeMBL, MMA and ST are 764, 650, and 237 L 3 mol
1 3 s
1, respectively, compared to a value of over 28 600 L 3 mol
1 3 s
1 for BA at the same temperature.40,41 Figure 6, which plots the kp,cop values for the MeMBL/BA system at 50 °C and the IPUE and terminal model fits to the data, shows that the kp,cop values of the copolymerization system are much closer to the homopolymerization of MeMBL than of BA. The higher reactivity of BA as well as the side reaction of intramolecular chain transfer40 presents difficulties in obtaining good PLP structure for fMeMBL < 0.1. It is clear, however, that the terminal model underrepresents the experimental data by greater than 50% at high concentrations of BA and does not adequately fit the data. Using the nonlinear 2323

dx.doi.org/10.1021/bm200400s |Biomacromolecules 2011, 12, 2319–2326

Biomacromolecules

ARTICLE

Figure 5. Arrhenius plot of ln(kp) vs inverse temperature for MeMBL compared to IUPAC-recommended Arrhenius equations for MMA19 (— —), ST20 (
3 3
), and BA40 (
3
). The line for MeMBL (), —) is the best fit for the data points shown.

Table 3. Arrhenius Parameters for MeMBL, MMA, BA, and ST Homopropagation Kinetics Ea

A

kp (50 °C)

monomer

(kJ 3 mol
1)

(L 3 mol
1 3 s
1)

(L 3 mol
1 3 s
1)

MeMBLthis work

21.77

2.29  106

764

MMA BA40

22.36 17.90

2.69  106 2.24  107

650 28 650

ST20

32.50

4.27  107

237

19

parameter estimation capabilities of Predici (based on a Gauss
Newton technique and estimation of 95% confidence intervals from the variance-covariance matrix),42,43 it was determined that the shape of the IPUE curve is insensitive to the value of sBA, with the confidence interval encompassing unity, as also was found for MMA/BA copolymerization.37 The radical reactivity ratios used to generate the IPUE curve in Figure 6 were sMeMBL = 6 and sBA = 1; the improved fit of the model to the data suggests that the presence of BA in the penultimate position increases the addition rate of MeMBL to a MeMBL radical (sMeMBL > 1), as also found for MMA/BA copolymerization.37 Figure 7 presents the kp,cop values for the MeMBL/MMA system at 50 °C and compares the IPUE model fit to the data with the terminal model prediction. The plot combines results measured in THF for monomer mixtures with less than 50 mol % MeMBL reported previously1 with new results over the complete composition estimated from SEC analysis in DMAc, with good agreement between the two analyses. The kp,cop data show a slight decrease as MeMBL is added to MMA, then climb to the higher kp values measured for MeMBL. The overall variation is small as the two monomers differ in kp values by only approximately 15%. The terminal model provides a reasonable prediction of the slope of the kp,cop curve as a function of fMeMBL but underpredicts slightly the experimental data by about 10%. The data can be better fit using the penultimate model with sMeMBL = 2.5 and sMMA = 0.6. However, given the scatter in the data, these estimates have a large uncertainty, and the confidence intervals encompass unity (equivalent to terminal model values). While the actual difference between the IPUE and terminal model predictions is of the order of the scatter of the data, the addition of the radical reactivity parameters does improve the data fit.

Figure 6. Copolymer propagation rate coefficients (kp,cop) data for the MeMBL/BA system vs MeMBL monomer mole fraction, fMeMBL, as measured by PLP-SEC at 50 °C and 33 Hz. Terminal model predictions (
3
) and IPUE model fit (—) are shown. kp,cop data are estimated from THF SEC analysis with RI (0) and LS (4) detectors, as well as by DMAc SEC analysis ()).

Figure 7. Copolymer propagation rate coefficients (kp,cop) data for the MeMBL/MMA system vs MeMBL monomer mole fraction, fMeMBL, as measured by PLP-SEC at 50 °C and 33 Hz. Terminal model predictions (
3
) and IPUE model fit (—) are shown. kp,cop data are estimated from THF SEC analysis with RI (0) and LS (4) detectors, as well as by DMAc SEC analysis ()).

The MeMBL/ST data set previously analyzed1 has also been updated using the improved measurements of MeMBL 2324

dx.doi.org/10.1021/bm200400s |Biomacromolecules 2011, 12, 2319–2326

Biomacromolecules

Figure 8. Copolymer propagation rate coefficients (kp,cop) data for the MeMBL/ST system vs MeMBL monomer mole fraction, fMeMBL, as measured by PLP-SEC at 50 °C and 33 Hz. Terminal model predictions (
3
) and IPUE model fit (—) are shown. kp,cop data are estimated from THF SEC analysis with RI (0) and LS (4) detectors, while the MeMBL homopolymer kp value measured in DMAc ()) is also shown.

ARTICLE

’ CONCLUSIONS Free radical bulk copolymerization kinetics of the biorenewable monomer MeMBL with three monomers commonly used in industry—styrene, methyl methacrylate, and butyl acrylate— have been successfully investigated by the PLP-SEC technique. Analyses were conducted using both THF and DMAc eluents for SEC, as poly(MeMBL) is fully soluble in the latter and copolymer solubility in THF varies with the mole fraction of MeMBL. The investigation of the complete copolymer composition range for each system and of the MeMBL homopolymer provides insight into the reactivity of lactone monomers relative to other common monomers. The homopolymer propagation kinetics of MeMBL are quite similar to those of MMA, as reflected in the small difference between their homopolymer kp values (approximately 15%), and an Arrhenius expression for MeMBL has been determined for the first time. Copolymer composition data, as measured using proton NMR, are well fit by the terminal model in all cases. MeMBL preferentially is incorporated into copolymer systems with ST, BA, and MMA; this compositional drift will need to be considered when conducting copolymerizations in batch. However, the terminal model, using the monomer reactivity ratios estimated from copolymer composition data, does not adequately represent the copolymerization kp,cop data for the MeMBL/BA and MeMBL/ST systems. The PLP-SEC data, however, could be fit by using the implicit penultimate unit effect (IPUE) model. The kinetic values determined in this paper will be used to design a polymerization system to produce MeMBL copolymers with controlled composition, to systematically investigate physical properties of the copolymers created, and to evaluate the potential use of this biorenewable monomer. ’ ASSOCIATED CONTENT

Figure 9. Copolymer propagation rate coefficients (kp,cop) data and IPUE model predictions for the MeMBL/ST (0,
3 3
), MeMBL/ MMA (O, — —) and MeMBL/BA (), —) systems vs MeMBL monomer mole fraction, fMeMBL, as measured by PLP-SEC at 90 °C. kp,cop data are estimated from THF SEC analysis with RI (gray symbols) and LS (black symbols) detectors, as well as by DMAc SEC analysis (unfilled symbols).

bS

Supporting Information. Table S1
S3: detailed PLP/ SEC and NMR results for ST/MeMBL MMA/MeMBL and BA/ MeMBL copolymerizations. Table S4 summarizes MeMBL homopolymerization PLP studies. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION homopolymer kp values, with results plotted in Figure 8. The terminal model significantly overpredicts the kp,cop data. The radical reactivity ratios for the IPUE model fit to the MeMBL/ST system have been recalculated after adding the newly determined MeMBL homopolymer kp value to the data sets. The nonlinear parameter estimates are sMeMBL = 1.01 ( 0.21 and sST = 0.63 ( 0.06, within the confidence intervals of the values published previously.1 Figure 9 plots the copolymer kp,cop data measured at 90 °C, the maximum temperature used in our synthesis study of MeMBL/ MMA copolymers,36 and the IPUE model predictions for all three systems. The trends at 90 °C are as found at 50 °C, with kp,cop relatively constant for the MeMBL/MMA system (between 1550 and 1950 L 3 mol
1 3 s
1) and decreasing smoothly for the MeMBL/ST system with decreasing fMeMBL (from 1950 to 800 L 3 mol
1 3 s
1). For MeMBL/BA kp,cop also decreases with decreasing fMeMBL, then starts to rise steeply toward the much higher BA value for fMeMBL < 0.2. Despite the uncertainty in the s parameters estimated at 50 °C, the 90 °C results are well-represented by the IPUE model for all cases.

Corresponding Author

*E-mail: [email protected].

’ ACKNOWLEDGMENT We thank E. I. du Pont de Nemours and Co. for providing the MeMBL monomer and the Natural Sciences and Engineering Research Council of Canada and the European Union and the state of Brandenburg for financial support of this work. ’ REFERENCES (1) Cockburn, R. A.; McKenna, T. F. L.; Hutchinson, R. A. Macromol. Chem. Phys. 2010, 211, 501–509. (2) Mosn
acek, J.; Matyjaszewski, K. Macromolecules 2008, 41, 5509–5511. (3) Pittman, C. U., Jr.; Lee, H. J. Polym. Sci., Part A: Polym. Chem. 2003, 41, 1759–1777. (4) Qi, G. G.; Nolan, M.; Schork, F. J.; Jones, C. W. J. Polym. Sci., Part A: Polym. Chem. 2008, 46, 5929–5944. 2325

dx.doi.org/10.1021/bm200400s |Biomacromolecules 2011, 12, 2319–2326

Biomacromolecules (5) Meier, M. A. R.; Metzger, J. O.; Schubert, U. S. Chem. Soc. Rev. 2007, 36, 1788–1802. (6) Williams, C. K.; Hillmyer, M. A. Polym. Rev. 2008, 48, 1–10. (7) Akkapeddi, M. K. Macromolecules 1979, 12, 546–551. (8) Akkapeddi, M. K. Polymer 1979, 20, 1215–1216. (9) McGraw, W. J. Lactone derivatives and method of making. U.S. Patent 2,624,723, 1953. (10) Suenaga, J.; Sutherlin, D. M.; Stille, J. K. Macromolecules 1984, 17, 2913–2916. (11) Ueda, M.; Suzuki, T.; Takahashi, M.; Li, Z. B.; Koyama, K.; Pittman, C. U., Jr. Macromolecules 1986, 19, 558–565. (12) Ueda, M.; Takahashi, M.; Imai, Y.; Pittman, C. U., Jr. J. Polym. Sci., Part A: Polym. Chem. 1982, 20, 2819–2828. (13) Ueda, M.; Takahashi, M.; Imai, Y.; Pittman, C. U., Jr. Macromolecules 1983, 16, 1300–1305. (14) Stansbury, J. W.; Antonucci, J. M. Dent. Mat. 1992, 8, 270–273. (15) Martin, J.; Watts, P. C.; Johnson, F. J. Chem. Soc. D: Chem. Commun. 1970, 27. (16) Manzer, L. E. Appl. Catal., A 2004, 272, 249–256. (17) Mosn
acek, J.; Yoon, J. A.; Juhari, A.; Koynov, K.; Matyjaszewski, K. Polymer 2009, 50, 2087–2094. (18) Beuermann, S.; Buback, M.; Davis, T. P.; Gilbert, R. G.; Hutchinson, R. A.; Kajiwara, A.; Klumperman, B.; Russell, G. T. Macromol. Chem. Phys. 2000, 201, 1355–1364. (19) Beuermann, S.; Buback, M.; Davis, T. P.; Gilbert, R. G.; Hutchinson, R. A.; Olaj, O. F.; Russell, G. T.; Schweer, J.; van Herk, A. M. Macromol. Chem. Phys. 1997, 198, 1545–1560. (20) 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. Macromol. Chem. Phys. 1995, 196, 3267–3280. (21) Coote, M. L.; Johnston, L. P. M.; Davis, T. P. Macromolecules 1997, 30, 8191–8204. (22) Davis, T. P.; O'Driscoll, K. F.; Piton, M. C.; Winnik, M. J. Polym. Sci., Polym. Lett. 1989, 27, 181–185. (23) Davis, T. P.; O'Driscoll, K. F.; Piton, M. C.; Winnik, M. A. Macromolecules 1990, 23, 2113–2119. (24) Li, D.; Li, N.; Hutchinson, R. A. Macromolecules 2006, 39, 4366–4373. (25) Wang, W.; Hutchinson, R. A. Macromolecules 2008, 41, 9011– 9018. (26) Liang, K.; Dossi, M.; Moscatelli, D.; Hutchinson, R. A. Macromolecules 2009, 42, 7736–7744. (27) Beuermann, S.; Buback, M. Prog. Polym. Sci. 2002, 27, 191–254. (28) Merz, E.; Alfrey, T.; Goldfinger, G. J. Polym. Sci. 1946, 1, 75–82. (29) Fukuda, T.; Ma, Y. D.; Inagaki, H. Macromolecules 1985, 18, 17–26. (30) Coote, M. L.; Davis, T. P.; Klumperman, B.; Monteiro, M. J. Rev. Macromol. Chem. Phys. 1998, 38, 567–593. (31) Harwood, H. J. Makromol. Chem., Macromol. Symp. 1987, 10
11, 331–354. (32) Coote, M. L.; Davis, T. P. Prog. Polym. Sci. 1999, 24, 1217–1251. (33) Fischer, H.; Radom, L. Angew. Chem., Int. Ed. 2001, 40, 1340–1371. (34) Dossi, M.; Liang, K.; Hutchinson, R. A.; Moscatelli, D. J. Phys. Chem. B 2010, 114, 4213–4222. (35) Yu, X.; Levine, S. E.; Broadbelt, L. J. Macromolecules 2008, 41, 8242–8251. (36) Cockburn, R. A.; McKenna, T. F. L.; Hutchinson, R. A. Macromol. React. Eng., submitted for publication. (37) Hutchinson, R. A.; McMinn, J. H.; Paquet, D. A., Jr.; Beuermann, S.; Jackson, C. Ind. Eng. Chem. Res. 1997, 36, 1103–1113. (38) Hutchinson, R. A.; Paquet, D. A., Jr.; McMinn, J. H.; Beuermann, S.; Fuller, R. E.; Jackson, C. In 5th International Workshop on Polymer Reaction Engineering, Berlin, 1995; Reichert, K. H., Moritz, H. U., Eds.; VCH Verlag: Berlin, 1995; pp 467
492. (39) Buback, M.; Feldermann, A.; Barner-Kowollik, C.; Lacík, I. Macromolecules 2001, 34, 5439–5448. (40) Asua, J. M.; Beuermann, S.; Buback, M.; Castignolles, P.; Charleux, B.; Gilbert, R. G.; Hutchinson, R. A.; Leiza, J. R.; Nikitin,

ARTICLE

A. N.; Vairon, J.-P.; van Herk, A. M. Macromol. Chem. Phys. 2004, 205, 2151–2160. (41) Barner-Kowollik, C.; G€unzler, F.; Junkers, T. Macromolecules 2008, 41, 8971–8973. (42) Wulkow, M. Predici 6.37.8; Computing In Technology GMbH: Rastede, Germany, 2003. (43) Wulkow, M. Macromol. React. Eng. 2008, 2, 461–494.

2326

dx.doi.org/10.1021/bm200400s |Biomacromolecules 2011, 12, 2319–2326