Dagmar D'hooge

Model-based design of the polymer microstructure: bridging the gap between polymer chemistry and engineering

The performance of polymeric materials depends strongly on the control over the polymer microstructure during the synthesis step. In this review, attention is paid to the potential of microkinetic modelling to facilitate the identification of optimal reactants and reaction conditions to design the polymer microstructure in bulk and solution (post)polymerisation processes. Focus is on living polymerization, reversible deactivation radical polymerization (RDRP) and “click” chemistry techniques, covering both batch and continuous synthesis approaches. A description according to the increasing level of macromolecular detail and thus modelling complexity is provided, including not only the characterization of the polymer microstructure via a finite number of variates (e.g. chain length, overall composition and branching content) but also its explicit visualisation by the simulation of the architecture and monomer sequences of a representative number of individual polymer chains. Several complementary case studies are included to demonstrate the high relevance of model-based design for the development of improved and novel synthetic protocols for precision control.

Improved Livingness and Control over Branching in RAFT Polymerization of Acrylates: Could Microflow Synthesis Make the Difference?

The superior capabilities of structured microreactors over batch reactors are demonstrated for reversible addition-fragmentation chain transfer (RAFT) solution polymerization of n-butyl acrylate with the aid of simulations, explicitly accounting for the chain length distribution of all macrospecies types. Since perfect isothermicity can be established in a microreactor, less side products due to backbiting and β-scission are formed compared to the batch operation in which ineffective heat removal leads to an undesirable temperature spike. For a given RAFT chain transfer agent (CTA), additional microstructural control results under microflow conditions by optimizing the reaction temperature, lowering the dilution degree, or decreasing the initial molar ratio of monomer to RAFT CTA.

An alternative method to estimate the bulk backbiting rate coefficient in acrylate radical polymerization

Based on a pulsed laser polymerization-size exclusion chromatography (PLP-SEC) analysis, an alternative method to estimate the bulk backbiting rate coefficient kbb in acrylate radical polymerization is presented. For different solvent volume fractions (0–0.75), using the unsaturated analogue of the monomer as a solvent to rule out solvent effects, regression analysis is applied to inflection point data in the low frequency range (<ca. 100 s−1) only, which can be scanned with less expensive PLP equipment. Variation of the solvent volume fraction allows the independent alteration of the average mid-chain radical lifetime and improvement in the sensitivity of the method to estimate kbb confidently. The robustness of the method is verified considering in silico generated data including large artificial errors. The method is applied to experimental data of 2,2-dimethoxy-2-phenylacetophenone (DMPA) initiated PLP of n-butyl acrylate, taking butyl propionate as the solvent. A kbb value of 171 ± 21 s−1 (303 K) is found, in good agreement with literature data. The method presents a generic approach for the estimation of kbb for other acrylate monomers, allows a complete statistical analysis, and can be used as a complementary tool for existing methods. In the long term, the method can even be extended for the simultaneous estimation of the bulk kbb and mid-chain radical propagation rate coefficient kp,m.

Ab initio based kinetic Monte Carlo analysis to unravel the propagation kinetics in vinyl acetate pulsed laser polymerization

The radical propagation kinetics of vinyl acetate (VAc) in pulsed laser polymerization (PLP) is studied by combining ab initio calculated rate coefficients for propagation of head, tail and mid-chain radicals, and backbiting reactions with kinetic Monte Carlo modeling of PLP spectra. The intriguing laser pulse frequency dependency of the propagation kinetics is shown to be mainly caused by the formation of stabilized mid-chain radicals via backbiting of tail radicals, originating from head-to-head propagation. These mid-chain radicals are approximately 35 times less reactive towards propagation at 323 K which, in agreement with experimental observations, results in a 15% increase of the observed propagation rate coefficient if the laser pulse frequency is increased from low (25–100 s−1) to high (300–500 s−1) values. Under typical PLP conditions, only tail radicals are reactive towards backbiting while this reaction is energetically unfavorable for head radicals. Tail-to-tail propagation of the radicals formed by head-to-head propagation is not sufficiently slow to fully explain the observed frequency dependence. The effect of chain length dependent propagation remains limited but can no longer be neglected at frequencies above 500 s−1.

A detailed mechanistic study of bulk MADIX of styrene and its chain extension

The microstructural evolution of individual macrospecies during bulk macromolecular design by interchange of xanthates (MADIX) of styrene with (O-ethyl xanthate)-2-ethyl propionate as an initial agent (R0X) and its chain extension with fresh styrene or n-butyl acrylate (nBuA) is visualized in silico, allowing an unbiased (co)polymer product quality labelling according to monomer sequences and end-groups. Degenerative transfer coefficients for both exchange with R0X (Ctr,0) and macro-RAFT agent (Ctr) are reported (Ctr,0 = 0.80 ± 0.02; Ctr: 0.44 ± 0.07) by applying multi-response regression analysis to the experimental data on the RAFT agent and styrene conversion, number and mass average molar masses, and end-group functionality (EGF). The EGF data are obtained by combining dialysis to remove residual R0X species and elemental analysis. It is shown that the MADIX mechanism can be properly understood only by explicitly acknowledging the differences in exchange reactivities and that the macroradical homopolymer CLD follows a Flory–Schulz distribution, which is an exception for controlled reversible addition–fragmentation chain transfer polymerization. Moreover, for the selected monomer conversion ranges, both “blocks” of the chain extension are formed through a single exchange.

Estimating the photodissociation quantum yield from PLP-SEC peak heights

A fast method for the reliable estimation of the photodissociation quantum yield Φdiss is presented. Pulsed laser polymerization (PLP) experiments are performed at various pulse energies (1.5–6 mJ) and regression analysis is performed to the ratio of the peak heights identified in the size exclusion chromatography (SEC) trace. The high accuracy of the method is demonstrated for PLP initiated by 2,2-dimethoxy-2-phenylacetophenone (DMPA), considering in silico generated data including large theoretical errors (up to 20%). The method has also been successfully applied to experimental data of DMPA based isothermal PLP of n-butyl acrylate at 306 K, with an estimated Φdiss of 0.42 ± 0.04. In the long term, the method will facilitate the evaluation of current and the design of new highly efficient photoinitiators.

Thiol-Michael addition in polar aprotic solvents: nucleophilic initiation or base catalysis?

The thiol-Michael addition of ethanethiol to ethyl acrylate, methyl vinylsulfone and maleimide initiated by ethyl-, diethyl-, triethylamine and triethylphosphine in tetrahydrofuran (THF) is investigated at room temperature for concentrations ranging from 0.5 to 2 mol L−1 for the reactants and 0.03 to 0.3 mol L−1 for the initiators. Rate coefficients for all elementary steps in a reaction scheme consisting of both the base catalyzed and the nucleophile initiated mechanism are calculated using CBS-QB3 corrected for solvation with COSMO-RS. Diffusional limitations are taken into account using the coupled encounter pair model. The ab initio apparent kinetic parameters are used in a microkinetic model and simulated conversions agree well with experimental data. Competition with the aza-Michael addition is shown to be insignificant. Regardless of the choice of ene or catalyst, conversion is governed by an anionic cycle in which first an addition from the thiolate to the ene occurs, followed by a rate-controlling proton transfer to the obtained Michael adduct anion from another thiol. For acrylates and vinylsulfones, the addition of the thiolate to the ene is quasi-equilibrated, while for maleimides this elementary reaction has a positive affinity, explaining their large reactivity. The choice of catalyst or ene strongly affects the initiation mechanism. Using tertiary phosphines only nucleophilic initiation takes place while with tertiary amines, only base catalysis occurs. For primary and secondary amines both initiation mechanisms contribute. The presented kinetic parameters and the insights on diffusional limitations are critical for the further optimization of thiol-Michael additions for polymer conjugation.

A novel method for the measurement of degenerative chain transfer coefficients: proof of concept and experimental validation

A novel method is presented to determine transfer coefficients in degenerative reversible addition fragmentation chain transfer (RAFT) polymerization from experimental dispersity data. Both the exchange reactivity with the small RAFT agent (CTA; Ctr,0) and the macro-RAFT agent (Ctr) can be measured at specific monomer conversions. The method is able to capture for the first time a possible intrinsic or apparent chain length dependency of Ctr(0) or thus the RAFT addition rate coefficient, an outstanding challenge in the mechanistic understanding of RAFT polymerization up to high monomer conversion. In its development stage, the method has been theoretically evaluated via the simulation of test cases in the absence and presence of artificial error. Experimental data at 353 K for azobis(isobutyronitrile) (AIBN) initiated RAFT polymerization of methyl methacrylate (MMA) with cyano-2-propyl dithiobenzoate (CPDB) as a small RAFT agent are successfully analyzed with the novel method. A highly accurate value of 20 is obtained for the exchange with CPDB. For the RAFT exchange with the macro-RAFT agent, an approximate but higher value of 76 is obtained and no chain length dependency is detected. The method is implemented in a spreadsheet, including a step by step explanation of the input parameters.

Ab‐Initio‐Based Kinetic Modeling to Understand RAFT Exchange: The Case of 2‐Cyano‐2‐Propyl Dodecyl Trithiocarbonate and Styrene

Ab-initio-calculated rate coefficients for addition and fragmentation in reversible-addition fragmentation chain transfer (RAFT) polymerization of styrene with 2-cyano-2-propyl dodecyl trithiocarbonate initiated by azobisisobutyronitrile allow the reliable simulation of the experimentally observed conversion, number average chain length, and dispersity. The rate coefficient for addition of a macroradical Ri to the macroRAFT agent Ri X at 333 K (6.8 104 L mol-1 s-1 ) is significantly lower than to the initial RAFT agent R0 X (3.2 106 L mol-1 s-1 ), mainly due to a difference in activation energy (15.4 vs 3.0 kJ mol-1 ), which causes the dispersity to spike in the beginning of the polymerization.

Simulation of the Degradation of Cyclic Ketene Acetal and Vinyl-Based Copolymers Synthesized via a Radical Process: Influence of the Reactivity Ratios on the Degradability Properties

The radical copolymerization of vinyl and cyclic ketene acetal (CKA) monomers is a promising way to prepare degradable vinyl polymers. The reactivity of the comonomer pair is known to be dependent of the vinyl monomer structure that requires to play with experimental conditions (feed ratio, overall monomer conversion, etc.) to target a desired cumulative (average) copolymer composition. Even if the materials are completely degradable, there is no information about the homogeneity of the degraded products. This theoretical study, using kinetic Monte Carlo simulations, allows simulating degradation at the molecular level. It is shown that disparate reactivity ratios (styrene/CKA, etc.) and also a composition drift at high conversion can lead to an inhomogeneous degraded product compared to systems with similar reactivity ratios (vinyl ether/CKA, etc.). The use of reversible deactivation radical polymerization techniques does not influence the final degraded products and is only useful for the design of advanced macromolecular architectures before degradation.