CHE 543: Polymer Production: Polymer Reaction Engineering
Estimated study time: 9 minutes
Table of contents
Sources and References
Primary texts — Odian, G., Principles of Polymerization, 4th ed., Wiley, 2004; Meyer, T. and Keurentjes, J. (eds.), Handbook of Polymer Reaction Engineering, Wiley-VCH, 2005.
Supplementary texts — Dotson, N.A. et al., Polymerization Process Modeling, Wiley-VCH, 1996; Rudin, A. and Choi, P., The Elements of Polymer Science and Engineering, 3rd ed., Academic Press, 2013.
Online resources — MIT OCW 10.569 “Synthesis of Polymers”; NPTEL “Polymer Reaction Engineering”; IUPAC polymer kinetics recommendations; ANSI/DIN standards on polymer characterization (public summaries).
Chapter 1: Commercial Polymerization Processes
1.1 Bulk (Mass) Polymerization
Monomer is its own solvent; only initiator (and sometimes chain transfer agent) is added. Advantages: highest product purity, simplest recovery. Disadvantages: rapid viscosity rise, heat removal difficulties, gel effect (autoacceleration) at moderate conversion. Used for PMMA casting, polystyrene, and polyolefins (melt phase). Industrial solutions include staged reactors and in-line devolatilization.
1.2 Solution Polymerization
Monomer and polymer dissolve in a compatible solvent. Lower viscosity aids heat transfer and mixing but adds downstream solvent recovery. Common for SBR, polyacrylates, some polyolefins. Solvent choice influences chain transfer (Mayo equation) and tacticity.
1.3 Suspension Polymerization
Monomer droplets (10–1000 µm) dispersed in water with stabilizers; each droplet is effectively a micro bulk reactor. Water carries the heat away, solving the heat-removal problem. Products: polystyrene beads (including expandable EPS), PVC beads, ion-exchange resins. Stabilizers (PVA, cellulose ethers, inorganic pickering stabilizers) prevent droplet coalescence.
1.4 Emulsion Polymerization
Surfactant-stabilized micelles host radical polymerization. Smith-Ewart kinetics for Case II (single radical per particle, average 1/2 radical per particle over time):
\[ R_p = k_p [M]_p \bar n \frac{N}{N_A}. \]With particle numbers \( N \sim 10^{14}-10^{16} \)/L and \( [M]_p \sim 5 \) M, rates far exceed bulk. Products: SBR, NBR, polyacrylates for coatings, PVC, VAE emulsions. Latex (direct use) or dried polymer are end products.
1.5 Gas-Phase and Slurry Polymerization
Olefin polymerization with Ziegler-Natta or metallocene catalysts uses fluidized bed (gas phase, Unipol) or slurry loop (Phillips). Catalyst particles grow into polymer granules; morphology control is crucial for flow and downstream processing. Modern single-site catalysts give tightly controlled molecular weight distribution and comonomer incorporation.
Chapter 2: Kinetic Modeling of Chain Polymerization
2.1 Free-Radical Kinetics Revisited
Balance equations for initiator, monomer, and radicals; steady-state on radicals gives
\[ R_p = k_p [M]\sqrt{\frac{f k_d [I]}{k_t}}. \]Elementary rate constants (\( k_p, k_t, k_d \)) are specific to monomer, temperature, and medium. IUPAC benchmark values from pulsed-laser polymerization define the modern data standard.
2.2 Molecular Weight Distribution Prediction
Method of moments: define \( \lambda_k = \sum_n n^k P_n \), derive ODEs for \( \lambda_0, \lambda_1, \lambda_2 \). These give instantaneous \( M_n, M_w, PDI \) without tracking the full distribution. For full distributions, method of moments is supplemented by discrete Galerkin, moment closure, or direct Monte Carlo simulation.
2.3 Gel Effect
Autoacceleration arises at moderate-to-high conversion when viscosity suppresses termination (\( k_t \) drops) but not propagation (\( k_p \) largely preserved). Rate and molecular weight rise sharply. Modeling requires a diffusion-controlled \( k_t(X) \), typically by free-volume theory (Vrentas-Duda).
2.4 Controlled Radical Polymerization Kinetics
RAFT: chain transfer agent \( CTA \) reversibly caps growing chains:
\[ P\cdot + CTA \rightleftharpoons P-CTA. \]Equilibrium favors dormant species; active chain concentration low, termination low, chains grow uniformly. Gives narrow PDI and controlled architectures. Kinetic modeling includes the exchange equilibrium plus normal FRP steps; proper parameter estimation requires dedicated experiments.
Chapter 3: Copolymerization
3.1 Reactivity Ratios
Terminal model: reactivity depends on last-added unit. Reactivity ratios \( r_A = k_{AA}/k_{AB} \) and \( r_B = k_{BB}/k_{BA} \). Mayo-Lewis equation predicts instantaneous composition.
3.2 Composition Drift
In a batch reactor, as preferred monomer is consumed, the remaining feed shifts; copolymer composition drifts with conversion. Strategies to suppress drift:
- Starved-feed semibatch: feed the more reactive monomer so that its reservoir concentration stays at the target.
- Azeotropic composition: choose \( [A]/[B] \) at the azeotrope (where \( d[A]/d[B] = [A]/[B] \)) if one exists.
- Living polymerization producing gradient or block copolymers.
3.3 Sequence Distribution
Beyond overall composition, the run-length statistics matter. For the terminal model, mean run lengths \( \bar n_A = 1 + r_A[A]/[B] \). Alternation, blockiness, or randomness influences \( T_g \), crystallinity, and phase behavior.
Chapter 4: Reactor Types and Operation
4.1 Batch vs Semibatch vs Continuous
Batch: versatile, simple, but composition drift is unavoidable. Semibatch: widely used when heat release or composition control demand it (emulsion, solution copolymerization). Continuous: CSTRs or tubular/loop for large-scale commodity polymers; narrow distributions in CSTR (steady concentrations) vs. batch (drift).
4.2 CSTR Networks
A single CSTR operates at constant low monomer concentration; multiple in series approach plug flow. For FRP, CSTR gives instantaneous distribution matching kinetic chain length—narrower than batch accumulation across conversion.
4.3 Tubular and Loop Reactors
High L/D, plug-flow approximated. Challenges: viscosity rise, heat removal (annular cooling, static mixers), pressure drop. Loop reactors circulate slurry or solution past a heat exchanger, achieving both the dynamics of a CSTR and the heat transfer of a pipe.
4.4 Heat Management
Polymerization is exothermic (\( \Delta H_{rxn} \) typically 50–100 kJ/mol monomer for vinyl monomers). Adiabatic temperature rise of bulk styrene polymerization exceeds 300 K—runaway territory. Jacketed reactors, internal coils, external loops, autorefrigeration (boiling monomer), and reduced-pressure evaporation cool the reacting mass.
Chapter 5: Branched and Crosslinked Systems
5.1 Long-Chain Branching
Chain transfer to polymer (LDPE), terminal double-bond reinsertion (polyolefins with Ziegler-Natta), and macromonomer copolymerization create long-chain branches. LCB profoundly alters rheology: shear thinning strengthens, extensional hardening appears, melt strength increases. Tailored branching yields specific processability.
5.2 Gelation
For polyfunctional monomers, \( f_w \)-average functionality > 2 leads to network formation at a critical conversion. Flory-Stockmayer:
\[ p_c = \frac{1}{\sqrt{(f_A - 1)(f_B - 1)}} \]for equimolar A\(_f\) + B\(_f\). Beyond gel point, insoluble network coexists with sol fraction; molecular weight of sol diverges at gel point.
5.3 Crosslinked Networks
Thermosets (epoxy, unsaturated polyester, phenolic), rubbers (sulfur-vulcanized, peroxide-cured), and ionomers. Network structure characterized by crosslink density and dangling-chain fraction. Cure kinetics (DSC, rheology) couple with temperature history; incomplete cure causes aging and property drift.
5.4 Cure Monitoring
Isothermal and scanning DSC reveal degree of cure vs time; rheological gel-point identification by crossover of G’, G’’; FTIR tracks functional group conversion. Model-fitting (autocatalytic, nth-order) or model-free (isoconversional) analysis extracts activation energy and cure model.
Chapter 6: Process Design and Optimization
6.1 Quality and Productivity Metrics
Number-average and weight-average molecular weight, PDI, sequence distribution for copolymers, tacticity for polyolefins, branching, residual monomer, particle size for dispersion products. Each product targets specific values; reactor design and operation must hit them reproducibly.
6.2 Parameter Estimation and Optimal Control
From pilot-plant data, estimate kinetic parameters by nonlinear regression on batch concentration/MW trajectories. Optimal control formulates a dynamic optimization: find feed trajectories, temperature profile, and initiator dosing that maximize productivity subject to quality constraints. Shrinking-horizon implementation against real-time measurements closes the loop.
6.3 Model-Based Operation
On-line measurements: reactor temperature, pressure, density, composition by NIR or Raman. Soft sensors estimate MW and composition from measurements plus the kinetic model. Model-predictive control adjusts feed rates and temperature to track reference trajectories through the batch.
6.4 Scale-Up
Heat transfer scales unfavorably with size (area/volume decreases as \( 1/L \)); mixing time scales with tip speed and impeller design. Scale-up rules vary by system: geometric similarity plus constant \( P/V \) for low-viscosity; constant tip speed for shear-sensitive; constant mixing time for composition uniformity. Pilot data and validated models guide the transition from bench to commercial.
