Optimizing nitroxide-mediated miniemulsion polymerization processes.
Keywords: Living radical polymerization, nitroxide, miniemulsion
Increasingly stringent regulations concerning VOC emissions are a driving force for improvements in coating technologies, especially resin design. Various approaches are being pursued, including development of increasingly high-solids solventborne systems, powder coatings, and aqueous-based coatings. Improvements in resin design target properties such as the composition, copolymer (or terpolymer, etc.) composition distribution, and the molecular weight distribution. While advances are being made using conventional radical polymerization to optimize resin properties, the inherently stochastic nature of radical polymerizations limits what can be achieved. Recently developed living radical polymerization techniques, however, offer superior control of polymer chain architecture and microstructure, including the ability to synthesize narrow molecular weight distribution polymers, functionalized polymers, block copolymers, gradient polymers, and multi-arm polymers.
Living radical polymerizations are characterized by polymeric chains that are reversibly terminated by a suitable controlling agent (e.g., nitroxide), resulting in a low rate of irreversible biradical termination (Scheme 1). The reversible termination step is an equilibrium process that strongly favors the dormant or inactive state so that most of the growing chains are dormant at any time. Living polymer chains retain the ability to add monomer units throughout the polymerization, thereby enabling the synthesis of polymers with complex microstructures such as diblock and triblock polymers. The potential to control microstructure using a simple and relatively inexpensive process has generated extensive interest. Studies have been published using various types of living/controlled radical polymerization, including nitroxide-mediated radical polymerization (NMRP), atom transfer radical polymerization (ATRP), and reversible addition fragmentation transfer (RAFT). Most studies have used either bulk or solution polymerization, while few groups have examined the behavior of living radical systems in aqueous-based polymerizations.
Early attempts to conduct NMRP in waterborne systems used emulsion polymerization, (1-3) but persistent colloidal stability issues prevented successful development. However, miniemulsion polymerization has been shown to be a preferable route to making NMRP polymers, exhibiting both the essential characteristics of a living radical system and excellent colloidal stability. (4-11) Colloidal instabilities in NMRP emulsion polymerization are believed to be related to the particle nucleation process. In miniemulsions, it is the monomer droplets that are nucleated to form polymer particles, and therefore the particle nucleation process is largely eliminated (some homogeneous nucleation likely also exists). Cunningham (12) and Qiu et al. (13) have recently reviewed living/controlled radical polymerizations in dispersed media.
Benzoyl peroxide (BPO), sodium dodecylbenzenesulfonate (SDBS), hexadecane, and 2,2,6,6 tetramethyl-1-piperidinyloxy (TEMPO) (Sigma Aldrich Ltd.) were used as received. Dowfax 8390 (Dow Chemical Ltd.), a disulfonated alkyl diphenyl oxide sodium salt, was used as received. Styrene (Sigma Aldrich Ltd.) was washed three times with a 2% wt NaOH solution to remove the inhibitor and then was washed three times with distilled water. Washed styrene was dried on calcium chloride overnight, distilled under vacuum, and refrigerated prior to use.
To synthesize the alkoxyamine BST (which consists of one benzyloxy unit, one styrene unit, and one TEMPO), styrene (300 ml) was added to benzoyl peroxide (0.116 mol) followed by TEMPO (0.154 mol). The system was purged under nitrogen (300 kPa) and heated to 135[degrees]C. Reaction progress was monitored using thin layer chromatography and reactions were typically run for 20-30 min. The crude product was then purified by hexane extraction. The solvent was removed and the resulting material was then dissolved in dichloromethane and passed through a column containing silica gel 60 (mesh 35-70, particle size 0.5-0.5 mm) using dichloromethane as the eluent. Fractions from the column were analyzed for BST using TLC. The solvent was evaporated and the BST recrystallized twice using isopropanol. Confirmation of the structure was done by proton NMR analysis.
TEMPO-Terminated Oligomeric PolyStyrene (TTOPS) Synthesis
The reactor was charged with styrene (45.3 ml), benzoyl peroxide (BPO, 1.80 mmol) and TEMPO (2.25 mmol), and purged with nitrogen. The mixture was heated to 135[degrees]C for 1 hr, and then cooled. This mixture was used as the organic phase in subsequent miniemulsion polymerizations using TTOPS as the initiating species.
USING BST AS INITIATOR: The aqueous phase consisted of SDBS (0.88 g, 2.52 mmol) and deionized water (120 mL) and the organic phase consisted of styrene (33 mL), hexadecane (5.4 g), BPO, and nitroxide (see Table 1). The two phases were mixed using a stirring rod, and then passed twice through a Microfluidizer-110S (Microflu-idics International Corp.) operating at 40 psi inlet pressure. The mini-emulsion was then poured into a 300 mL-auto-clave reactor. The reactor was purged six times with nitrogen. The polymerization was run at 135[degrees]C for six hours under approximately 300 kPa pressure. Samples were withdrawn every 90 min.
USING TTOPS AS INITIATOR: The same procedure was used as for the BST, except that the organic phase was the partially polymerized TTOPS-in-styrene mixture described previously, and no hexadecane was used.
Conversion was determined gravimetrically. Molecular weight and polydispersity were determined by size exclusion chromatography with a Waters 2690 Separations Module equipped with a Waters 410 Differential Refractometer, Waters Styragel columns, and an on-line degasser. Data analysis was performed using Millennium 2010 software. A calibration curve was constructed from monodisperse polystyrene standards ranging from 8.7 X [10.sup.2] to 2.8 X [10.sup.6] AMU. The eluant was tetrahydrofuran flowing at 1 mL/min and at temperature of 35[degrees]C. A Malvern Mastersizer 2000 (which uses laser diffraction) was used to measure particle size distributions.
RESULTS AND DISCUSSION
In developing an NMRP miniemulsion polymerization process suitable for industrial application, it would be desirable to have: (1) a high degree of chain livingness, to enable synthesis of complex microstructures; (2) a high final monomer conversion (>95%) to minimize the cost of monomer stripping; (3) a reasonably short total process time to maximize reactor throughput; and (4) low polymer polydispersity. In addition, it would be highly desirable to eliminate the need for a costabilizer such as hexadecane, which is commonly used in miniemulsion formulations and is very difficult to remove from the final product. For the applications of most interest (i.e., polymers with controlled microstructure), a high degree of chain livingness (the fraction of chains that are still "living" at the end of the polymerization) is the most important of these objectives. High reaction rates lower the overall cost, but as will be discussed in subsequent sections, are also important for maintaining a high degree of livingness. Low polydispersity will usually be achieved if livingness is high, although in itself it may not be overly important for most applications.
Dead chains in NMRP are generated primarily from two mechanisms: irreversible termination and disproportionation. Irreversible termination of two polymeric radicals gives dead polymer and also results in the accumulation of free nitroxide in the system, which in turn suppresses the reaction rate because of the equilibrium nature of the activation/deactivation step. The second important mechanism that creates dead chains is disproportionation of the inactive polymeric chain to yield a dead polymer chain with terminal unsaturation and a hydroxylamine (Scheme 2). It is unclear if the disproportionation step is a unimolecular or bimolecular process, although, as Fischer has shown, the two routes are kinetically indistinguishable. (14)
The current status of nitroxide-mediated miniemulsion polymerization is moving toward commercial feasibility but several obstacles remain. Typically conversions no higher than 80-90% can be attained, and this usually requires several hours. It is also not uncommon to see the rate decrease dramatically after ~50-60% conversion, probably due to free nitroxide accumulation in the system. The presence of hexadecane in the formulation remains an issue. The question of preserving a high degree of livingness while achieving fast rates and near-complete conversions is of paramount importance, but we currently lack a reliable method to quantify livingness. (Chain extension experiments are useful but only semiquantitative.)
Experimental and modeling studies conducted in our laboratory have led us to a better understanding of the critical issues in designing miniemulsion NMRP processes. (15-18) For example, Figure 1 shows monomer conversion versus time of the miniemulsion polymerization of styrene (at 135[degrees]C) initiated by the alkoxyamine BST (Scheme 3) at different BST concentrations. A summary of the experimental conditions used in these experiments is given in Table 1. Our mathematical model accounts for polymerization kinetics in both the aqueous and particle phases, phase partitioning, as well as radical entry and exit from particles. In addition to simulating standard measures such as conversion, Mn, and PDI (polydispersity index), the model was also used to calculate the number of living and dead chains and their chain length distribution, as well as how the dead chains were terminated (biradical termination, disproportionation, or transfer to monomer). The interested reader is referred to reference 15 for details. The three lines for predicted conversion in Figure 1 are virtually coincidental, in reasonably good agreement with the experimental data, although the last data points (12 hr) are below predicted values. Although the 14 mM BST runs appear to have lower conversions than the other runs at 7 mM and 21 mM BST, these values likely fall within experimental error. The polymerization rate of styrene NMRP processes is known to be insensitive to the alkoxyamine concentration, and is instead determined by the thermal initiation rate of the styrene. This phenomenon occurs because the level of free nitroxide in the system is governed by the rate at which thermal radicals are generated. Termination reactions, which occur at a low but nonetheless finite rate, result in the release of nitroxide thereby increasing its concentration while thermally generated radicals act to reduce the nitroxide concentration. A pseudo steady-state is reached in which the generation rate of thermal radicals is approximately balanced by the loss of radicals due to termination. Figure 1 shows that the rate decreases significantly at moderate conversions, finally leveling off at ~60% conversion. This rate decrease can be attributed to the accumulation of nitroxide, which drives the equilibrium toward the dormant state. Figure 2 shows the number average molecular weight Mn as a function of conversion. Mn grows linearly with conversion, as expected for a living radical polymerization, with higher BST loadings giving lower molecular weights.
Ma et al. (15) calculated the fate of growing chains as the polymerization progresses under different conditions for BST-initiated polymerizations. Figure 3 shows the relative concentrations of different types of chains versus conversion calculated from the simulations. IGR refers to initiator-generated radicals (derived from BST) while TGR refers to thermally generated radicals. The active radical concentration is very low, as is expected from the equilibrium relationship between active and dormant chains. The capped IGR concentration (living, dormant chains) decreases dramatically, and is seen from the large increase in the population of dead chains. The degree of livingness (defined as the fraction of total chains that are living) is reduced to ~50% at 60% monomer conversion. The importance of the various mechanisms in contributing to dead chain formation is shown in Figure 4. Transfer to monomer and irreversible termination have relatively minor roles, while it is disproportionation that is the most significant contributor to loss of livingness at these conditions. The disproportionation rate is a function of temperature (greater disproportionation occurs at higher temperatures) and the nature of the specific nitroxide. TEMPO is particularly prone to disproportionation, and development of any commercial process would need to consider the advantages of TEMPO (availability and low cost) with its inherent disadvantages.
[FIGURE 1 OMITTED]
[FIGURE 2 OMITTED]
Preserving livingness is critical for the synthesis of polymers with controlled microstructure, but is complicated by the conflicting constraints posed by minimizing dead chain formation due to irreversible termination versus disproportionation. Minimizing termination requires maintaining a low active radical concentration, or equivalently a high free nitroxide concentration, which also results in low reaction rates. However, the long reaction times required will result in significant dead chain formation due to disproportionation. (Disproportionation is a first order reaction in the dormant chain concentration, and therefore continues throughout the reaction at an appreciable rate.) Minimizing dead chain formation due to disproportionation requires reduction of the overall reaction time (higher reaction rates), an objective in direct opposition to that for minimizing termination. An optimized balance must therefore be obtained between losses due to disproportionation and biradical termination. Examining the kinetic data in Figure 1, it is clear reaction times should be shortened while achieving much higher conversions if a high degree of livingness is to be maintained.
[FIGURE 3 OMITTED]
In a recent publication by Keoshkerian et al., (10) high conversions were achieved in less than six hours using a modified miniemulsion process that did not employ a traditional costabilizer such as hexadecane. Keoshkerian et al. first polymerized styrene in bulk to a low conversion to give a nitroxide-terminated (TEMPO) oligomer, which then served the dual role of polymeric costabilizer and alkoxyamine in a subsequent miniemulsion polymerization. Their findings are consistent with the earlier published results of Reimers and Schork, (19) who showed that miniemulsions can be stabilized by polymeric hydrophobes, and that the use of small molecule costabilizers such as hexadecane is not required to produce stable miniemulsions. Use of polymeric hydrophobes yields miniemulsions that do not have as long a stable shelf-life as when hexadecane is used, but the miniemulsions are sufficiently stable during the time required to polymerize the monomer. Adopting this approach, we have further explored the hexadecane-free process.
Hexadecane-free miniemulsion polymerizations were run using a TEMPO-terminated polystyrene chain (TTOPS (6)) as both the initiating species and a polymeric hydrophobe that provided the required costabilization. The TTOPS was produced using a bulk NMRP polymerization carried out to a low conversion (see Experimental section). This partially polymerized mixture was then used as the organic phase for the miniemulsion polymerization. No additional nitroxide was added. The first factor we explored was the effect of removing hexadecane. If hexadecane is not used in polymerizations initiated by BST, colloidal failure occurs. Therefore we used TTOPS-initiated runs to examine the effect of the hexadecane concentration. At our standard SDBS concentration of 0.021 M, a small increase in conversion was observed when using a hexadecane-free formulation with TTOPS compared to a TTOPS formulation containing hexadecane (Figure 5). This modest increase in conversion was consistent with our expectations. Eliminating the hexadecane increases the rate since it (1) results in a small increase in monomer concentration, which directly results in higher rate, and (2) leads to an appreciable increase in the thermal initiation rate (and hence the active radical concentration) because of the third order dependence of the thermal polymerization rate on monomer concentration. (20) It is interesting to note that although TTOPS with M=19k may not be expected to act effectively as a costablilizer given its high molecular weight, it did yield stable latexes with similar particle sizes to those using hexadecane and lower molecular weight TTOPS.
[FIGURE 4 OMITTED]
Our observed results were different from those observed by Keoshkerian in that we did not obtain the high conversions or rates that they did. A difference in experimental conditions between their experiments and ours was the surfactant concentration, a factor we originally thought to be of negligible importance in affecting rate, as found by Pan et al. (11) using Dowfax 8390. However, when we increased the SDBS concentration from 0.021 M to 0.089 M, a dramatic increase in rate was observed, as seen in Figure 5. Conversions of >95% were obtained in 2-3 hr, while maintaining narrow polydispersities (~1.3). The final particle size distributions were nearly identical at the low and high SDBS concentrations, exhibiting a mean diameter ~150 nm. Negligible coagulum was formed in both cases. The insensitivity of particle size to the SDBS concentration is contrary to the findings of Pan et al. (11) who observed a decrease in diameter with increasing concentration of Dowfax 8390. We may already be at a high enough SDBS concentration that adding more surfactant does not further reduce the miniemulsion droplets produced in the microfluidization step.
It has not been previously recognized that [SDBS] can affect the rate in these NMRP miniemulsions. Furthermore the observed rates at the high [SDBS] are quite fast compared to conventional living radical polymerizations, and yet low polydispersities are still obtained. These results seem anomalous in that higher rates are expected to result in increased irreversible termination, which should broaden the molecular weight distribution as well as result in a significant loss of livingness.
[FIGURE 5a OMITTED]
It is unclear what causes the dependence of rate on [SDBS]. The effect may be attributable to either the SDBS itself or impurities (e.g., sulfonic acid). The most common effect of rate-enhancing additives in NMRP, usually strong acids, is to consume nitroxide. It is possible that impurities in the SDBS, which is only available in a technical grade, are responsible for the observed effect on rate. Significantly we found that a different batch of SDBS yielded lower rates at the same concentration, suggesting batch-to-batch variations are a concern.
Based on our findings using SDBS, we conducted experiments using Dowfax 8390, the surfactant used by El-Aasser's group at Lehigh University. The experiments used TTOPS as the initiating species and did not contain hexadecane. Two widely different Dowfax concentrations were used (0.021 M and 0.12 M) to determine if the concentration affected the rate as was observed for polymerizations stabilized by SDBS. (A lower molecular weight TTOPS was used for the Dowfax runs compared to the SDBS runs, but previous experimentation has shown the TTOPS molecular weight has a relatively minor effect on rate, provided the number of chains is the same.) Figure 6a shows that the large variation in Dowfax 8390 concentration did not affect the rate. The Dowfax stabilized runs all had lower conversion than the SDBS stabilized runs, yielding ~45% conversion in six hours. The Mn versus conversion data indicates a linear relationship close to the theoretical value (Figure 6b), although the actual molecular weights were slightly higher than expected, indicating fewer chains. The polydispersities for Dowfax (~1.15-1.20) were somewhat lower than those observed using SDBS (~1.15-1.35), although conversions were also lower.
[FIGURE 5b OMITTED]
[FIGURE 6a OMITTED]
Clearly there are differences in the behavior of polymerizations stabilized by the different surfactants, although the reasons are unknown at this time. The Dowfax 8390 is a purer product than the SDBS, and it is possible that impurities in the SDBS act as rate enhancers, perhaps by consuming nitroxide. Particle size should not affect rate since compartmentalization is not expected to be a factor in NMRP miniemulsions/emulsions, (21) and therefore it is surprising that surfactant affects the reaction so significantly. These effects are currently being further investigated.
Although the high rates and conversions obtained in our runs are cause for optimism, the livingness of the system at such high rates is a concern. High reaction rates imply high radical concentrations and therefore high termination rates. Although polydispersity remains low, polydispersity is itself not a sensitive measure of livingness. Chain extensions experiments are complicated by the additional loss of chains during the actual chain extension process, and therefore provide only a semi-quantitative measure of livingness. Work is currently in progress to quantify the livingness of these samples using the recently published technique developed by Scott et al. (22)
[FIGURE 6b OMITTED]
Preserving livingness in NMRP systems is a primary objective. In the TEMPO-mediated miniemulsion polymerization of styrene, reaction times of 8-12 hr at 135[degrees]C were calculated to result in a loss of about half of the polymer livingness. Hexadecane-free polymerization initiated (and costabilized) by TEMPO-terminated polystyrene oligomer yield stable latexes and modestly increased reaction rates in comparison to formulations containing hexadecane. Higher [SDBS] can dramatically increase the rate, giving conversions of >95% in 2-3 hr with relatively low polydispersities (~1.3). Batch-to-batch variation effects on rate have been observed with SDBS. The surfactant Dowfax 8390 gives slower polymerizations than SDBS with somewhat improved polydispersities. Varying the concentration of Dowfax 8390 does not affect the rate.
Table 1--Recipe for BST-Initiated Miniemulsion Polymerization of Styrene at 135[degrees]C. Aqueous Phase Water 120 g SDBS 0.88 g Organic Phase Styrene 30 g Hexadecane 4.7 g BST 0.1, 0.2, 0.3 g (7, 14, 21 mM)
Financial support from the Natural Sciences and Engineering Research Council of Canada, the Xerox Research Centre of Canada and the Canada Foundation for Innovation is gratefully acknowledged.
(1) Bon, S.A.F., Bosveld, M., Klumperman, B., and German, A.L., "Controlled Radical Polymerization in Emulsion," Macromolecules, 30, 324 (1997).
(2) Marestin, C., Noel, C., Guyot, A., and Claverie, J., "Nitroxide Mediated Living Radical Polymerization of Styrene in Emulsion," Macromolecules, 31, 4041 (1998).
(3) Farcet, C., Lansalot, M., Charleux, B., Pirri, R., and Vairon, J-P., "Mechanistic Aspects of Nitroxide-Mediated Controlled Radical Polymerization of Styrene in Miniemulsion, Using a Water-Soluble Radical Initiator," Macromolecules, 33, 8559 (2000).
(4) Charleux, B., "Theoretical Aspects of Controlled Radical Polymerization in a Dispersed Medium," Macromolecules, 33, 5358 (2000).
(5) Prodpran, T., Dimonie, V.L., Sudol, E.D., and El-Aasser, M.S., "Nitroxide-Mediated Living Free-Radical Miniemulsion Polymerization of Styrene," Macromol. Symp., 155, 1 (2000).
(6) Pan, G., Sudol, E.D., Dimonie, V.L., and El-Aasser, M.S., "Nitroxide-Mediated Living Free-Radical Miniemulsion Polymerization of Styrene," Macromolecules, 34, 481 (2001).
(7) Ma, J.W., Cunningham, M.F., McAuley, K.B., Keoshkerian, B., and Georges, M.K., "Nitroxide Partitioning Between Styrene and Water," J. Polymer Sci. A: Polymer Chemistry, 39, 1081 (2001).
(8) Cunningham, M.F., Xie, M., McAuley, K.B., Keoshkerian, B., and Georges, M.K., "Nitroxide Mediated Styrene Miniemulsion Polymerization," Macromolecules, 35, 59 (2002).
(9) Tortosa, K., Cunningham, M.F., Keoshkerian, B., and Georges, M.K., "Influence of Camphorsulfonic Acid in Nitroxide-Mediated Styrene Miniemulsion Polymerization," J. Polymer Sci., (Polymer Chemistry Ed.), 40, 2828 (2002).
(10) Keoshkerian, B., Szkurhan, A.R., and Georges, M.K., "Block Copolymer Synthesis by a Miniemulsion Stable Free Radical Polymerization Process," Macromolecules, 34, 6531 (2001).
(11) Pan, G., Sudol, E.D., Dimonie, V.L., and EL-Aasser, M.S., "Surfactant Concentration Effects on Nitroxide-Mediated Living Free Radical Miniemulsion Polymerization of Styrene," Macromolecules, 35, 6915 (2002).
(12) Cunningham, M.F., "Living/Controlled Radical Polymerizations in Dispersed Phase Systems," Prog. Poly. Sci., 27, 1039 (2002).
(13) Qiu, J., Charleux, B., and Matyjaszewski, K., "Controlled/Living Radical Polymerization in Aqueous Media: Homogeneous and Heterogeneous Systems," Prog. Poly. Sci., 26, 2083 (2001).
(14) Souaille, M. and Fischer, H., "Living Free Radical Polymerizations Mediated by the Reversible Combination of Transient Propagat-ing and Persistent Nitroxide Radicals. The Role of Hydroxylamine in Alkene Formation," Macromolecules, 34, 2830 (2001).
(15) Ma, J., Smith, J.-A., McAuley, K., Keoshkerian, B., Georges, M.K., and Cunningham M.F., "Living Radical Polymerization of Styrene in Miniemulsion. Part 1: Model Studies of Alkoxyamine Initiated Systems," Chem. Engineering Sci., (2002, in press)
(16) Ma, J., Smith, J.-A., McAuley, K., Keoshkerian, B., Georges, M.K., and Cunningham, M.F., "Living Radical Polymerization of Styrene in Miniemulsion. Part 1: Model Studies of Persulfate-Initiated Systems," Chem. Engineering Sci., (2002b, in press).
(17) Ma, J., Smith, J.-A., McAuley, K., Keoshkerian, B., Georges, M.K., and Cunningham, M.F., "Interfacial Mass Transfer in Nitroxide-Mediated Miniemulsion Polymerization," Macromolecular Theory and Simulations, (2002a, in press).
(18) Ma, J., Smith, J.-A., McAuley, K., Keoshkerian, B., Georges, M.K., and Cunningham, M.F., "Model Studies of Nitroxide-Mediated Styrene Miniemulsion Polymerization--Opportunities for Process Improvement," Macromolecular Theory and Simulations (2002b, in press).
(19) Reimers, J. and Schork, F.J., "Robust Nucleation in Polymer-Stabilized Miniemulsion Polymerization," J. Appl. Poly. Sci., 59, 1833 (1996).
(20) Hui, A.W. and Hamielec A.E., "Thermal Polymerization of Styrene at High Conversions and Temperatures. An Experimental Study," J. Appl. Poly. Sci., 16, 749 (1972).
(21) Butte, A., Storti, G., and Morbidelli, M., "Miniemulsion Living Free Radical Polymerization of Styrene," Macromolecules, 33, 3485 (2000).
(22) Scott, M.E., Parent, J.S., Hennigar, S.L., Whitney, R.A., and Cunningham M.F., "Determination of Alkoxyamine Concentration in Nitroxyl-Mediated Styrene Polymerization Products," Macromolecules, 35, 7628 (2002).
Michael F. Cunningham** and Marcus Lin--Queen's University*
Barkev Keoshkerian--Xerox Research Centre of Canada[dagger]
* Dept. of Chemical Engineering, Kingston, Ont., K7L 3N6, Canada.
[dagger] 2660 Speakman Dr., Mississauga, Ont., L5K 2L1, Canada.
** Corresponding author: email: email@example.com; 613.533.2782; fax: 613.533.6637.
|Printer friendly Cite/link Email Feedback|
|Date:||Jan 1, 2004|
|Previous Article:||Miniemulsions: overview of research and applications.|
|Next Article:||Controlled coagulation of emulsion polymers.|