(226b) Optimal Operating Policies for Producing Styrene and Methyl Methacrylate Copolymers Via RAFT Polymerization

Controlled radical polymerization (CRP) techniques offer the possibility of creating new materials with pre-specified properties and controlled molecular structures in mild operating conditions. These advantageous characteristics have given rise to an increasing interest in these synthetic methods, which is reflected in the numerous patents applications, journal articles and commercial products already been produced.^{[1, 2]}

The polymerization control is achieved through the establishment of an equilibrium between active and inactive chains which is strongly shifted toward the dormant species. The active radical concentration reduction leads to a decrease in the total effect of termination, which results in a uniform growth of the polymer chains. Moreover, the lifetime of each chain is increased from a fraction of a second to hours, allowing experimental and industrial practitioners to manipulate the polymer structure through changes in operating conditions, such as reactants feeding policies.^{[3, 4]} In this way it is possible to obtain copolymers with controlled molecular weights, molecular weight distribution (MWD), composition and chain architecture.^[5]

Three CRP approaches have been more extensively studied: atom transfer radical polymerization (ATRP), nitroxide mediated polymerization (NMP), and reversible addition-fragmentation chain transfer (RAFT). The two former techniques rely upon adding an agent that reversibly terminates the radicals. Instead, RAFT polymerization is controlled through the addition of a chain transfer agent (CTA) that transfers the active radical to a large number of chains. RAFT polymerization is one of the most versatile approaches due to the large range of applicable monomers. Unlike ATRP, it does not require the use of transition metals. However, RAFT polymerization products have some potential for odor and discoloration if the dithioester group that acts as CTA leaches from the polymer chain.^{[4, 6]}

Through a copolymerization process different comonomers can be combined into a single material in many diverse structures. By changing chain length, composition or sequence length distribution, unique products can be obtained. The resulting phase behavior influences the physical properties and ultimate applications.^[7] Polystyrene-block-poly(methyl methacrylate), or pSt-b-pMMA, is of special interest for nanoscale patterning thanks to its ability to produce diverse morphological patterns and the ease of removal of pMMA domains by UV exposure.^[8] To this end, block chain length, chain length ratios, polydispersity index, and block composition greatly impact the final nanoscale morphology.^[9] On the other hand, gradient copolymers of the same comonomers (pSt-g-MMA) are hoped to be a superior class of blend stabilizers than their block counterpart due to the formation of broad interphase regions.^[10]

Simulation of RAFT systems is particularly difficult due to the distinct features of its kinetic mechanism. The polymerization proceeds through the formation of a two-arm intermediate adduct which makes necessary the computing of a bivariate MWD even for homopolymerization reactions.^[11] In previous works we presented a mathematical model able to predict the molecular structure of styrene and methyl methacrylate copolymerization via RAFT.^[12] Average properties, such as molecular weights and global composition, were modeled using the well-known method of double index moments. In addition, the full bivariate MWD of the copolymer was obtained by means of 2-D probability generating functions (pgf). This technique proved to be capable of dealing with the complex kinetic mechanisms and obtain the MWD without simplifying assumptions or any a priori knowledge of the distribution shape.

In this work, the developed model is used to seek optimal operating policies to produce styrene and methyl methacrylate RAFT copolymers with controlled structure and pre-specified properties. The difference between the obtained properties and the desired ones is minimized by changing design variables (such as reactor type or lateral feeds position) and operating policies. For semibatch reactors, operating conditions such as initial charge and feeding policy were optimized. On the other hand, tubular reactor optimization space included the flowrate of the main feed, and the position, flowrate and composition of the lateral feeds. The optimization problem was formulated and solved in gPROMS (Process Systems Enterprise, Ltd.). Optimal operating policies were obtained for producing copolymers with controlled polydispersity and pre-determined copolymer structure: either total molecular weight and chain length ratios for block copolymers, or pre-specified compositional gradients for gradient copolymers. Modeling studies like the one presented in this work are of great help to find operating conditions that allow manufacturing unique structures designed beforehand. In this way, experimental trial and error procedures can be minimized, resulting in resource savings and facilitation of the scaling up of CRP techniques to an industrial level.

[1]        W. Jakubowski, N. V. Tsarevsky, P. McCarthy, K. Matyjaszewski, Materials Matters 2010, 5 (1), 16.

[2]        K. Matyjaszewski, J. Spanswick, Materials Today 2005, 8, 26.

[3]        M. Zhang, W. H. Ray, J. Appl. Polym. Sci. 2002, 86, 1630.

[4]        W. A. Braunecker, K. Matyjaszewski, Prog. Polym. Sci. 2007, 32, 93.

[5]        K. Matyjaszewski, K. A. Davis, Statistical, Gradient, Block and Graft Copolymers by Controlled/Living Radical Polymerizations, Springer-Verlag, Berlin, Germany, 1^st edition, 2002.

[6]        S. Grajales, Controlled Radical Polymerization Guide, Sigma-Aldrich Co. LLC. Materials Science, 2012.

[7]        F. S. Bates, Science 1991, 251, 898.

[8]        G. S. W. Craig, C. J. Thode, M. Serdar Onses, P. F. Nealey, Materials Matters 2011, 6 (3), 82.

[9]        W. Jakubowski, A. Juhari, A. Best, K. Koynov, T. Pakula, K. Matyjaszewski, Polymer 2008, 49, 1567.

[10]      U. Beginn, Colloid. Polym. Sci. 2008, 286, 1465.

[11]      C. Barner-Kowollik, Handbook of RAFT Polymerization, Wiley-VCH, Weinheim, Germany, 1^st edition, 2008.

[12]      C. Fortunatti, C. Sarmoria, A. Brandolin, M. Asteasuain, Computer Aided Process Engineering 2014, June 15-18 Budapest, Hungary. Accepted 6 pages paper.