SEARCH

SEARCH BY CITATION

Keywords:

  • branching;
  • control;
  • diffusion;
  • ICAR ATRP;
  • kinetics

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Kinetic Model
  5. 3. Results and Discussion
  6. 4. Conclusion
  7. Nomenclature
  8. Acknowledgements
  9. Supporting Information
Thumbnail image of graphical abstract

The potential of initiators for continuous activator regeneration atom transfer radical polymerization (ICAR ATRP) for the synthesis of well-defined poly(n-butyl acrylate) is analyzed by means of simulations. The kinetic model accounts for reactivity differences between secondary and tertiary macrospecies and considers the possible influence of diffusional limitations. CuBr2 is used as transition metal salt and the commercially available N,N,N′,N″,N″-pentamethyldiethylenetriamine as ligand. For targeted chain lengths (TCLs) up to 1000, the ICAR ATRP can be performed relatively quickly, and with ppm levels of ATRP catalyst. For moderate TCLs, slightly higher ppm levels are required if excellent control over chain length is also desired. In all cases, limited loss of end-group functionality (EGF) results.


1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Kinetic Model
  5. 3. Results and Discussion
  6. 4. Conclusion
  7. Nomenclature
  8. Acknowledgements
  9. Supporting Information

Industrial production of polymers is often carried out by free radical polymerization (FRP) since this polymerization technique is applicable to a wide range of monomers, tolerant to impurities, and economically beneficial compared to other chain-growth polymerization techniques.1, 2 In particular, acrylate-based polymers contribute significantly to the polymer market as evidenced by their use in a wide variety of applications covering mainly paints, adhesives, textiles, and coatings.3 Typically, polyacrylates are produced via emulsion polymerization, but depending on the required production scale and the area of application, solution and bulk FRP are also of practical importance.3, 4

For some applications of polyacrylates, the production process could benefit from a better control over end-group functionality (EGF). For instance, in the coating industry, the controlled incorporation of EGF in the polymer could avoid the use of expensive functional monomers. Moreover, advanced well-defined macromolecular architectures, such as linear gradient copolymers and star-shaped polymers with controlled arm length, cannot by obtained by FRP due to the inherent difficulty to control EGF and chain length. Therefore, in the last decades a variety of so-called controlled radical polymerization (CRP) techniques5–11 have been developed at laboratory scale in which a mediating agent (e.g., a nitroxide or a transition metal complex) is added allowing control over the polymer microstructure and thus the synthesis of complex polymer topologies and compositions, involving polyacrylate segments.

Since polyacrylates are manufactured almost exclusively by radical polymerization (RP)3 a thorough understanding of the interplay of the radical reactions involved in the production process is of paramount importance. Of special interest are chain transfer to polymer and βC-scission reactions,12–15 which influence polymer properties, such as the average chain length and branching content. These properties can be directly manipulated in view of the desired application of the final polymer product, e.g., by promoting or inhibiting the occurrence of side reactions.16–18 Short chain branches (SCBs) originate after propagation of tertiary macroradicals that are generated via intramolecular chain transfer to polymer, i.e., backbiting (left reaction path in Scheme 1), which consists of self-abstraction of a hydrogen atom from the backbone of a secondary macroradical, mainly involving a cyclic six-membered transition state.16 In addition, due to the higher stability of the tertiary macroradicals a rate retardation takes place in case backbiting is sufficiently important. On the other hand, long chain branches (LCBs) typically result after propagation of tertiary macroradicals formed by intermolecular chain transfer to polymer (right reaction path in Scheme 1), in which a secondary macroradical abstracts a hydrogen atom from another polymer chain generating a tertiary macroradical and a dead polymer chain. Alternatively, at elevated temperatures LCBs can be obtained after addition of macroradicals to macromonomers that are formed by βC-scission. However, several kinetic studies have indicated that the contribution of LCBs to the total amount of branches is very low, especially at low to intermediate conversions.15, 19–23

thumbnail image

Scheme 1. Mechanism of chain transfer to polymer reactions in acrylate polymerization. End denotes the ATRP initiator fragment or a hydrogen atom. For nBuA polymerization, R corresponds to an n-butyl group; M stands for monomer.

Download figure to PowerPoint

In contrast to RP of ethylene and vinyl acetate in which the occurrence of chain transfer to polymer reactions is a long-standing fact that has been well documented since its discovery more than a half century ago,24–28 the importance of branching in acrylate RP has only emerged in the early nineties by the detection of quaternary carbons via 13C NMR spectroscopy.29–31 Interestingly, Ahmad et al.32 have recently reported that the branching level of poly(n-butyl acrylate) can be reduced significantly by performing a CRP instead of an FRP.

One of the most frequently used CRP techniques is (normal) atom transfer radical polymerization (ATRP),33, 34 the principle of which is given in Scheme 2a. During the polymerization macroradicals (Ri; i: chain length) are temporarily and catalytically deactivated by a transition metal complex (equation image) to a dormant form (RiX), which contains EGF (X). Typically, the ATRP is started with ATRP initiator (R0X) and activator in absence of deactivator (equation image). For sufficiently high deactivation rates, termination reactions can be suppressed resulting in a high EGF, i.e., almost no dead polymer molecules are formed. If the ATRP initiation is fast, a polymer characterized by a low polydispersity index (PDI) can be obtained as well.

thumbnail image

Scheme 2. (a) Principle of normal ATRP and (b) principle of ICAR ATRP; ka, kda, kp, and kt are the rate coefficients for activation, deactivation, propagation, and termination; ka,IX, kda,I, and kp,I are the rate coefficients for activation, deactivation, and propagation related to the conventional initiator; f and kdis correspond to the conventional radical initiator efficiency and the corresponding rate coefficient; i = 0 corresponds to ATRP initiator; n(+1) is the oxidation number of the transition metal complex (de)activator; X corresponds to a halogen atom, and L to ligand.

Download figure to PowerPoint

Crucial for control of the branching level in polyacrylates is the selection of the ATRP catalyst. Based on simulations, Reyes and Asua35 indicated that ATRP catalysts which can strongly deactivate macroradicals are necessary to obtain a lower branching level than in FRP, in which no mediating agent is present. Later on, Konkolewicz et al.23, 36 showed that the branching level is also influenced by tertiary activation and is to a large extent determined by the relative importance of the rate of backbiting and tertiary propagation. Only if these two rates are well-balanced from low conversion onwards, ATRP provides branching levels as high as FRP. Moreover, these authors pointed out that the branching level can be increased by increasing the targeted chain length (TCL), i.e., the initial molar ratio of monomer to ATRP initiator.

Even though selection of the appropriate ATRP catalyst enables the synthesis of well-defined polyacrylates and purification methods are available for its removal,37–41 the amount of catalyst in a normal ATRP process is too high to obtain an economic profitable process.1, 42 Therefore, in the last years, ATRP modified techniques5, 43–54 have been developed in which a low catalyst concentration (ppm level with respect to monomer) is employed. Importantly, these techniques can be applied using commercially available ligands, are more environmentally friendly and avoid the oxygen sensitivity of the activator upon storage. Furthermore, they can be carried out within polymerization times similar to industrially applied RPs (≈ ≤8 h). In contrast, normal ATRP processes typically take longer than 1 d when using low catalyst amounts.

One of the most important techniques to reduce the amount of ATRP catalyst is initiators for continuous activator regeneration (ICAR) ATRP,43 in which the polymerization is started in the presence of a conventional radical initiator (I2), ATRP initiator (R0X) and deactivator (equation image). Thermal dissociation of I2 provides a source of free radicals from which activator molecules (equation image), are continuously generated, allowing activation of the ATRP initiator and thus the occurrence of the same reactions as in normal ATRP (Scheme 2b).

Recent simulations55 have shown that in ICAR ATRP the control over polymerization rate and polymer properties can be directly manipulated by adjusting the polymerization conditions. For instance, it was demonstrated that, depending on the ATRP catalyst reactivity, step-wise addition of conventional radical initiator in the ICAR ATRP of methyl methacrylate and styrene is needed to reach high conversion while still obtaining a good livingness and control over chain length with ppm levels of ATRP catalyst. Moreover, Konkolewicz et al.56 recently reported for the first time the successful ICAR ATRP of oligo(ethylene oxide) methyl ether acrylate in water using a low (<100) ppm level of ATRP catalyst further proving the importance of this modified ATRP technique for a broad range of monomers and in particular acrylates. Additionally, the potential of ICAR ATRP for the controlled production of polyacrylates can be inferred from the new polymeric materials prepared via normal ATRP. For example, Auschra et al.4 reported the development of new pigment dispersants for the formulation of high solids and waterborne coatings using n-butyl acrylate (nBuA) and dimethylaminoethyl acrylate (DMAEA) as monomers.

Alternatively, well-defined polyacrylates can be synthesized while using low catalyst amounts through activators regenerated by electron transfer (ARGET) ATRP47, 57 or electrochemically mediated ATRP (eATRP)48 in which the activator is regenerated from a reducing agent or by reduction at an electrode. In the presence of metallic copper, a so-called supplementary activator and reducing agent (SARA) ATRP or single electron transfer living radical polymerization (SET-LRP) can also be obtained depending on comproportionation or disproportionation being the dominant side reaction path for the involved catalytic species.58–60 In particular for methyl acrylate, Chan et al.61 and Kwak et al.57 have recently successfully combined ARGET ATRP and the use of a copper wire reducing significantly the residual catalyst amount up to 10 ppm for a TCL of ca. 100.

However, both for ICAR ATRP and these alternative techniques, only a limited number of kinetic studies are available in which the influence of the TCL and catalyst amount on the polymerization rate and control over polymer properties is mapped in detail.56–58, 62–64 Such information is crucial for a comparison of these techniques to evaluate their potential industrial application, since the ultimate goal of a CRP technique is to produce a wide range of average chain lengths at acceptable polymerization rates while preserving control over PDI and EGF. In this work, such detailed kinetic modeling study is presented for the bulk ICAR ATRP of nBuA using the commercially available N,N,N′,N″,N″-pentamethyldiethylenetriamine (PMDETA) as ligand and copper (II) bromide (CuBr2) as transition metal salt. A similar modeling approach can be followed for other related CRP techniques avoiding time consuming experimental screening procedures for a given catalyst.

The activation/deactivation kinetic parameters are assessed based on the experimental data of Ahmad et al.32 for the normal bulk ATRP of nBuA taking into account the reactivity difference between secondary and tertiary macrospecies. The kinetic model also considers the possible influence of diffusional limitations on termination and deactivation. For TCLs up to thousand, it is shown that ICAR ATRP can be successfully applied to synthesize well-defined poly(nBuA) with ppm levels of ATRP catalyst within reasonable polymerization time and with limited loss of EGF. The simulation results confirm in particular the potential of CRP techniques using low amounts of copper for the controlled incorporation of EGF in polymer chains. Diffusional limitations are shown to be most important on secondary deactivation leading to a rate acceleration at high conversion.

2. Kinetic Model

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Kinetic Model
  5. 3. Results and Discussion
  6. 4. Conclusion
  7. Nomenclature
  8. Acknowledgements
  9. Supporting Information

The reaction scheme used in the kinetic modeling of the bulk ATRP of nBuA and the corresponding intrinsic kinetic parameters are summarized in Table 1. Activation, deactivation, backbiting and βC-scission are included next to typical radical polymerization (RP) reactions, i.e., propagation, chain transfer to monomer and termination. A distinction is further made between the reactivity of secondary (s) and tertiary (t) macrospecies. For simplicity, intermolecular chain transfer to polymer is neglected since literature reports indicate that LCBs barely contribute to the total branching content.15, 19–23 Similarly, termination by disproportionation65, 66 and addition reactions involving macromonomers are neglected in a first approximation.67

Table 1. Reactions involved in the bulk normal/ICAR ATRP of nBuA and their intrinsic kinetic parameters (kl,chem); i,j = chain length; Ri,t and Ri,s: tertiary and secondary macroradicals; equation image: apparent rate coefficient for the reaction step l; R0 and I: ATRP initiator and conventional initiator derived radical; T = 378 K; MM: macromonomer.
 Elementary reactionEquationkl, chemRef.
Normal ATRPActivation (a)equation image3.1 [L · mol−1 · s−1]72
equation image1.6 [L · mol−1 · s−1]This work
equation image2.6 × 101 [L · mol−1 · s−1]This work
Deactivation (da)equation image8.0 × 108 [L · mol−1 · s−1]This work
equation image4.0 × 108 [L · mol−1 · s−1]This work
equation image4.0 × 107 [L · mol−1 · s−1]This work
Propagation (p)equation image7.4 × 104 [L · mol−1 · s−1]68
equation image7.4 × 104 [L · mol−1 · s−1]68
equation image1.5 × 102 [L · mol−1 · s−1]69
Chain transfer to monomer (trM)equation image9.0 [L · mol−1 · s−1]70
equation image9.0 [L · mol−1 · s−1]70
equation image8.5 × 10−2 [L · mol−1 · s−1]70
Backbiting (bb)equation image2.0 × 103 [s−1]69
βC-scission (βC-sc)equation image2.2 [s−1]71
equation image
Termination by recombination (tc)equation image2.3 × 108 [L · mol−1 · s−1]69
equation image4.6 × 107 [L · mol−1 · s−1]69
equation image3.0 × 106 [L · mol−1 · s−1]69
Extra for ICAR ATRPDissociation (dis)equation image8 × 10−2 [s−1] with f = 0.75 [–]55, 73
Activation (a)equation image1.6 × 101 [L · mol−1 · s−1]Based on55
Deactivation (da)equation image4.0 × 107 [L · mol−1 · s−1]Based on55
Propagation (p)equation image7.4 × 105 [L · mol−1 · s−1]Based on55

The ATRP catalyst and initiator are the same as those used by Ahmad et al.32 in their experimental study of the bulk normal ATRP of nBuA at 378 K, i.e., CuBr/PMDETA and methyl 2-bromopropionate (MBrP). For the reactions common with FRP, intrinsic rate coefficients reported in literature are used.68–71 For activation of the ATRP initiator a value of 3.1 L · mol−1 · s−1 is considered based on the experimental study of Seeliger and Matyjaszewski.72 The remaining secondary and tertiary activation and deactivation intrinsic rate coefficients are adjusted according to the experimental data of Ahmad et al.32 As discussed below, the obtained values are consistent with literature values and confirm the higher stability of tertiary macrospecies.

For ICAR ATRP, dissociation of the conventional radical initiator and activation, deactivation and propagation involving conventional radical initiator fragments are also considered (Table 1). tert-Butyl peroxy-2-ethylhexanoate (Trigonox21s) is employed as conventional radical initiator, since it is particularly suited for the polymerization of acrylates in the range of 353–423 K.73, 74 For simplicity, a typical constant initiator efficiency of 0.75 is used and the intrinsic activation and deactivation rate coefficients related to conventional radical initiator fragments are taken equal to those of the secondary macrospecies. The latter approach has been selected since simulations have revealed that a typical tenfold reactivity difference has no significant influence on the results.55 The conversion profile and the evolution of the polymer properties with time are simulated using the methodology developed by D'hooge et al.,67, 75 which is based on an extension of the method of moments coupled with an application of the quasi-steady state approximation for the calculation of population weighted apparent rate coefficients using a convergence test.

CRP kinetic studies75–79 have indicated that diffusional limitations can result in a lowering of the apparent termination reactivity during the polymerization and that deactivation can become diffusion controlled at sufficiently high conversion. Therefore, in this work the possible influence of diffusional limitations on termination and deactivation is considered in agreement with literature reports.75, 80–84 For more details, the reader is referred to the Supporting Information.

3. Results and Discussion

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Kinetic Model
  5. 3. Results and Discussion
  6. 4. Conclusion
  7. Nomenclature
  8. Acknowledgements
  9. Supporting Information

3.1. Normal ATRP of nBuA

Figure 1 presents a comparison of the model simulations while accounting for possible diffusional limitations (full lines) and the experimental findings for the evolution of conversion with time and the evolution of the number average chain length (equation image), the PDI and the cumulative branching fraction (on a molar basis) with conversion. As mentioned above, the experimental data are taken from Ahmad et al.32 and correspond to a polymerization temperature of 378 K and a TCL ([M]0/[R0X]0) of 289. In the same figure, the simulated EGF and cumulative CC double bond fraction (on a molar basis) are given as a function of conversion. No experimental data are however available for the latter polymer properties.

thumbnail image

Figure 1. Comparison of experimental data and simulations for the bulk normal ATRP of nBuA; diffusional limitations considered (full lines) and neglected (dashed lines); (a) conversion profile, (b) number-average chain length (equation image), (c) PDI, (d) branching fraction (cumulative; molar), (e) EGF, and (f) CC double bond fraction (cumulative; molar) as a function of conversion; 378 K; [M]0/[R0X]0/[CuBr]0/[L]0:347/1.2/0.6/0.6; experimental data from Ahmad et al.;32 no experimental data for e) and f); for continuity equations see D'hooge et al.67

Download figure to PowerPoint

Clearly, a good description of the experimental data is obtained. The control over chain length and livingness is good, since PDI values close to 1.2 are simulated and limited loss of EGF takes place during the ATRP. In agreement with simulation results on the ATRP of isobornyl acrylate,67 the cumulative CC double bond fraction is very low and can therefore be neglected in a first approximation. At high conversion, the cumulative branching fraction amounts to 0.015 corresponding to an average of five branches per polymer chain.

The corresponding kinetic parameters for (de)activation of secondary and tertiary macrospecies are listed in Table 1. An equilibrium coefficient equation image of 4 × 10−9 results for the secondary macrospecies, whereas the equilibrium coefficient for tertiary macrospecies (equation image = 6.4 × 10−7) is about two orders of magnitude higher. In accordance with Seeliger and Matyjaszewski,72 a value of 1.6 L · mol−1 · s−1 is obtained for the secondary activation rate coefficient, which is approximately twenty times lower than the tertiary activation rate coefficient. For deactivation, the rate coefficient for tertiary macroradicals is about ten times lower than for secondary macroradicals reflecting the higher stability of these species in agreement with the theoretical kinetic study of Konkolewicz et al.23

From Figure 1 it also follows that the increase of the viscosity upon polymerization leads to a rate acceleration at high conversion while the control over polymer properties is barely affected. This rate acceleration can be explained based on the difference in reaction probabilities for secondary and tertiary macroradicals in case diffusional limitations are accounted for or neglected (Figure 2). The reaction probability of a species for a reaction l (Pl) is defined as the ratio of the rate of the reaction Rl to the summation of all reaction rates for the considered species:

  • equation image((1))
thumbnail image

Figure 2. Effect of diffusional limitations on the reaction probabilities for secondary (a–c) and tertiary (d–f) macroradicals for the bulk normal ATRP of nBuA; (a) deactivation, (b) propagation, and (c) backbiting of secondary macrospecies; (d) deactivation, (e) propagation, and (f) βC-scission of tertiary macrospecies; diffusional limitations considered (full lines) and neglected (dashed lines); 378 K; [M]0/[R0X]0/[CuBr]0/[L]0 = 347/1.2/0.6/0.6; for continuity equations see D'hooge et al.67

Download figure to PowerPoint

For secondary macroradicals a distinction is made among the reaction probability for deactivation, propagation, and backbiting while for tertiary macroradicals deactivation, propagation, and βC-scission are considered. As demonstrated in the Supporting Information, the reaction probabilities for termination are very low at high conversion and can thus be neglected to explain the rate acceleration mentioned above.

In agreement with Figure 1, diffusional limitations influence the reaction probabilities mainly at high conversion. In particular, the probability of secondary macroradicals to deactivate (Figure 2a) plummets, since the mobility of the deactivator is significantly reduced from a conversion of ca. 0.60. As a consequence, propagation of secondary macroradicals is strongly favored (Figure 2b), leading to an increase of the conversion and the probability of backbiting (Figure 2c). In contrast, a much less pronounced effect of diffusional limitations on the reaction probabilities for tertiary species is observed (Figure 2d–f). Only a slightly lower and higher reaction probability, respectively, is obtained for tertiary deactivation and propagation, accompanied by a minimal increase of the βC-scission probability that takes place at high conversion. In other words, the increase in backbiting does not induce a rate retardation due to the ca. hundred times lower propagation of tertiary macrospecies. Hence, it can be concluded that the conversion profile at high conversion is determined by diffusional limitations on secondary deactivation, which promote the observed rate acceleration.

For an amount of copper (I) bromide (CuBr) of 250 ppm (with respect to monomer), the effect of TCL on the conversion profile and control over polymer properties is shown in Figure 3 for TCLs varying between 50 and 1 000, the latter being a typical maximum TCL in experimental CRP studies.57, 85–87 For the considered range of TCLs, a good livingness and control over chain length is obtained as evidenced by the high EGF and low PDI values at high conversion. In agreement with Konkolewicz et al.,23 a higher cumulative branching fraction is obtained for higher TCLs. Note that for sufficiently low TCLs the polymerization proceeds relatively quickly.

thumbnail image

Figure 3. Effect of TCL for the bulk normal ATRP of nBuA on (a) the conversion profile, evolution of (b) number-average chain length (equation image), (c) PDI, (d) branching fraction (cumulative; molar), and (e) EGF with conversion; 378 K with 250 ppm of ATRP catalyst; [M]0/[R0X]0/[CuBr]0/[L]0:50/1/0.0125/0.0125, 100/1/0.025/0.025, 200/1/0.05/0.05, 500/1/0.125/0.125, and 1 000/1/0.25/0.25; for continuity equations see D'hooge et al.67

Download figure to PowerPoint

However, when the initial amount of CuBr is lowered further as required for industrial application,43 the considered normal ATRP process slows down considerably resulting in polymerization times that are too long from an industrial point of view, as illustrated in Figure 4 for a TCL of 500. For example, for an initial amount of CuBr of 100 ppm it takes about 2 d to reach a conversion of 0.80 and a good control over polymer properties. As will be illustrated in the next section, ICAR ATRP can resolve the former issue while preserving control over polymer properties. For the particular case mentioned above, it is shown that with ICAR ATRP only half a day is required.

thumbnail image

Figure 4. Effect of initial amount of CuBr (ppm with respect to monomer) for the bulk normal ATRP of nBuA on (a) the rate of polymerization, evolution of the (b) number-average chain length (equation image), (c) PDI, (d) branching fraction (cumulative; molar), and (e) EGF with conversion; 378 K; [M]0/[R0X]0/[CuBr]0/[L]0:500/1/y/y with y = 0.05 (100 ppm), 0.075 (150 ppm), 0.125 (250 ppm), 0.2 (400 ppm); for continuity equations see D'hooge et al.67

Download figure to PowerPoint

3.2. ICAR ATRP of nBuA

Figure 5 and 6 present the simulated trends of polymerization time, average chain length (equation image), PDI, EGF, and cumulative branching fraction related to the initial ppm level of Cu(II) and TCL at a fixed conversion of 0.80. The initial amount of Cu(II) is varied between 5 and 250 ppm, the latter being a typical value for a normal ATRP (Figure 3 and 4). For TCL, again an upper limit of 1 000 is considered.57, 85–87 For all simulations presented in Figure 5 and 6, the initial molar ratio of conventional radical initiator to ATRP initiator is fixed at 0.02.

thumbnail image

Figure 5. Polymerization time required to reach a conversion of 0.8 as a function of initial amount of Cu(II) (ppm with respect to monomer) and TCL in the bulk ICAR ATRP of nBuA at 378 K; [I2]0/[R0X]0:0.02/1; for continuity equations see D'hooge et al.67

Download figure to PowerPoint

thumbnail image

Figure 6. Polymer properties for the bulk ICAR ATRP of nBuA as function of the initial amount of Cu(II) (ppm with respect to monomer) and TCL; (a) number-average chain length (equation image), (b) PDI, (c) EGF and (d) branching fraction (cumulative; molar); 378 K; [I2]0/[R0X]0:0.02/1; conversion: 0.8; for continuity equations see D'hooge et al.67

Download figure to PowerPoint

It appears from Figure 5 that longer polymerization times are required to reach a conversion of 0.80 for higher initial ppm levels of Cu(II) and TCLs. For example, using 40 ppm of Cu(II) 1 h is required for a TCL of 50, whereas approximately 20 h are required for a TCL of 1 000. Similarly, for a TCL of 300, 4 h are necessary when using 5 ppm of Cu(II), which increases to half a day for 250 ppm. In particular, with 100 ppm Cu(II) and for a TCL of 500 only half a day is needed, which confirms the potential of ICAR ATRP in terms of short polymerization times. As explained above (Figure 4a) for the same TCL and ppm level of catalyst, normal ATRP requires approximately 2 d.

The inverse relationship between EGF and TCL, described previously by Goto and Fukuda,88 is shown in Figure 6c. Close inspection reveals that even at relatively high TCLs good livingness is still obtained. For example, for a TCL of 700 and an initial Cu(II) amount of 50 ppm, the simulated EGF is ca. 0.90. Furthermore, a similar EGF results with 5 ppm demonstrating the limited influence of the initial Cu(II) concentration on the livingness of ICAR ATRP. In agreement with the recent simulations of Zhong and Matyjaszewski,89 ICAR ATRP thus allows to use a low initial amount of Cu(II) while preserving EGF (Figure 6c) and maintaining a fast polymerization rate.

As for the normal ATRP of nBuA, the cumulative branching fraction in the ICAR ATRP process (Figure 6d) changes only as a function of TCL and is rather independent of the initial amount of Cu(II). By varying TCL e.g. from 50 to 700 the branching fraction can be increased by about 10%, independent of the initial amount of Cu(II). Comparison of Figure 3d and 6d reveals that for each TCL a similar branching fraction is simulated for normal and ICAR ATRP.

Based on Figure 5, the initial amount of Cu(II) required to reach a conversion of 0.8 in 8 and 12 h can be calculated, as shown in Figure 7(a, b). In the same figures two regions are highlighted indicating longer (t > 8 or 12 h) and shorter (t < 8 or 12 h) polymerization times. Clearly, if a polymerization time of 12 h or less is demanded, the full range of TCLs can be covered using initial amounts of Cu(II) well below 250 ppm with a broader range of ppm levels available for low TCLs. If the polymerization time demanded is shorter, for instance 8 h, a narrower range of ppm levels is available for high TCLs. Interestingly, for all TCLs, fast ICAR ATRPs can be performed using very low ppm levels, i.e., lower than 5 ppm.

thumbnail image

Figure 7. Full lines: initial amount of Cu(II) (ppm with respect to monomer) for the bulk ICAR ATRP of nBuA to obtain a conversion of 0.8 in (a) 8 h, (b) 12 h, (c) a PDI value equal to 1.2 at a conversion of 0.8 as a function of TCL; 378 K; [I2]0/[R0X]0:0.02/1; for continuity equations see D'hooge et al.67

Download figure to PowerPoint

As reported by D'hooge et al.,55 who studied the ICAR ATRP of methyl methacrylate and styrene, too high initial levels of Cu(II) lead to too high initial deactivation rates and, hence, retard the polymerization. Figure 8(a, b) shows that for the ICAR ATRP of nBuA the same conclusion can be drawn. In this figure, the initial deactivation rate of secondary macroradicals is shown as a function of conversion using 5, 50, and 250 ppm of Cu(II) for a TCL of 50 and 500. For the latter TCL, the retardation is more pronounced as a result of the decrease in initial concentration of both the ATRP and conventional initiator, since their ratio remains equal to 0.02, as mentioned above.

thumbnail image

Figure 8. Effect of initial amount of Cu(II) (ppm with respect to monomer) on secondary deactivation rate (a-b) and ATRP initiator concentration (c-d) as a function conversion in the bulk ICAR ATRP of nBuA; full line: 5 ppm; dashed line 50 ppm; dotted line: 250 ppm; in (a) and (b) time is given for a conversion of 0.8; in (c) and (d) PDI is given at a conversion of 0.8; 378 K; [I2]0/[R0X]0:0.02/1; for continuity equations see D'hooge et al.67

Download figure to PowerPoint

However, as depicted in Figure 6(a, b) a sufficiently high initial ppm level of Cu(II) is demanded for a good control over chain length, particularly for relatively low TCLs. For TCLs between 50 and 200, PDI values much higher than 1.5 result for initial amounts of Cu(II) lower than 75 ppm and a good control over chain length (PDI ≈ 1.2) is only obtained when this initial amount is higher than ca. 100 ppm. Too low initial amounts of ATRP catalyst result in a too slow activator (re)generation as evidenced by the too slow disappearance of the ATRP initiator on a conversion basis shown in Figure 8(c-d). For a good control over chain length, the ATRP initiation has to be completed at a relatively low conversion so that enough monomer is available to compensate for the chain length difference caused by the non-instantaneous ATRP initiation. Since for higher TCLs the ATRP initiation is finished at lower conversions (Figure 8d), less Cu(II) can thus be employed to obtain a PDI of 1.2 at conversion of 0.8 (Figure 7c).

Although, for TCLs higher than 600 a slight increase of the ppm level of Cu(II) is observed. As suggested by D'hooge et al.,55 well-defined ICAR ATRP-based polymers can only be prepared if the initial concentration of conventional radical initiator is sufficiently low with respect to the initial concentration of ATRP initiator. Therefore, it can be expected that for higher TCLs, the ppm level of Cu(II) shown in Figure 7c can be further decreased by lowering the initial ratio of conventional radical initiator to ATRP initiator from 0.02 to e.g. 0.01. Indeed, Figure 9a demonstrates that in case this initial ratio is reduced to 0.01 from a TCL of 700 onwards, the necessary amount of Cu(II) to reach a PDI of 1.2 (conversion of 0.8) decreases further as a function of TCL. However, as shown in Figure 9b this approach is accompanied by a significant increase in the required polymerization time.

thumbnail image

Figure 9. (a) Full line: initial amount of Cu(II) (ppm with respect to monomer) in the bulk ICAR ATRP of nBuA to obtain a PDI value equal to 1.2 at a conversion of 0.8; 378 K; [I2]0/[R0X]0:0.02/1 for TCL < 700; [I2]0/[R0X]0:0.01/1 for TCL ≥ 700. (b) Polymerization time needed to reach a conversion of 0.8 for TCLs ≥ 700; conditions as in (a); for continuity equations see D'hooge et al.67

Download figure to PowerPoint

Summarizing, the above simulations clearly indicate that ICAR ATRP can be used to prepare well-defined poly(nBuA) using low initial amounts of Cu(II). In particular, due to high EGF values it is expected that poly(nBuA) based block copolymers can be synthesized provided that appropriate conditions are selected for the addition of a second block. Moreover, the level of control over the polymer properties can be adjusted by selecting the appropriate initial ppm Cu(II) level for a given TCL and required polymerization time. This is illustrated in Figure 10, in which the polymerization time required to reach a conversion of 0.80 is plotted as a function of the initial amount of Cu(II) (1–250 ppm) and TCL (50–1 000). In the same figure, six lines are indicated to delineate regions of conditions according to constraining values for EGF and PDI, i.e., those conditions which lead to a PDI of 1.1, 1.2, and 1.3 on the one hand and to an EGF of 0.85, 0.9, and 0.95 on the other hand.

thumbnail image

Figure 10. Diagram for the bulk ICAR ATRP of nBuA illustrating control over chain length, livingness, and the required polymerization time for a conversion of 0.80 as a function of the initial amount of Cu(II) (ppm with respect to monomer) and TCL; 378 K; [I2]0/[R0X]0:0.02/1; full lines indicate the limits: PDI (white) equal to 1.1, 1.2, and 1.3, EGF (black) equal to 0.85, 0.9, and 0.95; for continuity equations see D'hooge et al.67

Download figure to PowerPoint

Overall, Figure 10 confirms the effectiveness and robustness of ICAR ATRP of acrylates and in particular of nBuA. For TCLs lower than 500, well-defined poly(nBuA) (1.1 < PDI < 1.2 and EGF > 0.9) can be obtained within relatively short polymerization times (≈12 h) employing initial amounts of Cu(II) between 60 and 250 ppm. For higher TCLs, still relatively fast polymerizations can be conducted with very low (<10) Cu(II) ppm levels at the expense of a reduction of the control over chain length (PDI > 1.3) and livingness (0.85 < EGF < 0.9). In case for these TCLs the control over chain length is important (PDI ≈ 1.1), initial amounts of Cu(II) above 65 ppm are required leading to longer polymerization times (>12 h).

4. Conclusion

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Kinetic Model
  5. 3. Results and Discussion
  6. 4. Conclusion
  7. Nomenclature
  8. Acknowledgements
  9. Supporting Information

A computational kinetic study is performed for ICAR ATRP to support the synthesis of well-defined poly(nBuA), i.e., with predetermined chain length and cumulative branching level, low PDI, and high EGF. Kinetic parameters related to activation and deactivation are assessed based on literature experimental data for the corresponding normal ATRP system. The obtained parameter values are in line with the expected higher stability of tertiary macroradicals compared to secondary species. As for normal ATRP, in ICAR ATRP higher branching levels are obtained for higher TCLs. Diffusional limitations are mainly important on secondary deactivation leading to a rate acceleration at high conversion.

The advantages of the ICAR ATRP technique are illustrated and guidelines are provided for the selection of relevant polymerization conditions, such as initial catalyst concentration and polymerization time for a given TCL. Up to TCLs of 1000, ICAR ATRP of nBuA can be performed while reaching a high conversion relatively quickly and attaining reasonable control over polymer properties at low (<50) ppm level of ATRP catalyst. Only for moderate TCLs, slightly higher ppm levels of ATRP catalyst are needed, when low PDI values are desired. In all cases, the livingness of the ICAR ATRP is sufficiently high.

Nomenclature

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Kinetic Model
  5. 3. Results and Discussion
  6. 4. Conclusion
  7. Nomenclature
  8. Acknowledgements
  9. Supporting Information
[C]0

initial concentration of C (mol · L−1)

EGF

end group functionality

i, j

chain length

kl

rate coefficient for reaction l (L · mol−1 · s−1) or (s−1)

L

ligand

M

monomer

Mt

transition metal

equation image

activator

equation image

deactivator

PDI

polydispersity index

R0

ATRP initiator radical

equation image

secondary macroradical having a chain length i

equation image

tertiary macroradical having a chain length i

R0X

ATRP initiator

equation image X, equation image X

dormant polymer molecule having a chain length i; s and t refer to the type of macroradical formed upon activation

TCL

targeted chain length; initial molar ratio of monomer to ATRP initiator

X

halogen/EGF

equation image

number average chain length

Subscripts

a

activation

da

deactivation

dis

dissociation

I

conventional radical initiator fragment

p

propagation

0

ATRP initiator

t

termination

tc

termination by recombination

Superscripts

app

apparent

chem

Intrinsic chemical

n(+1)

oxidation number

s

secondary macroradical

t

tertiary macroradical

Abbreviations

ATRP

atom transfer radical polymerization

CRP

controlled radical polymerization

FRP

free radical polymerization

ICAR

initiators for continuous activator regeneration

nBuA

n-butyl acrylate

MBrP

methyl 2-bromopropionate

PMDETA

N,N,N′,N″,N″-pentamethyldiethylenetriamine

RP

radical polymerization

Acknowledgements

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Kinetic Model
  5. 3. Results and Discussion
  6. 4. Conclusion
  7. Nomenclature
  8. Acknowledgements
  9. Supporting Information

The authors acknowledge financial support from the Long Term Structural Methusalem Funding by the Flemish Government, the Interuniversity Attraction Poles Programme – Belgian State – Belgian Science Policy and the Fund for Scientific Research Flanders (FWO). This work is performed in the framework of the project ‘Model-Based Optimization & Control for Process-Intensification in Chemical and Biopharmaceutical Systems’ (OPTICO/ G.A. No. 280813) funded by the European Commission; the contents of the publication reflects only the authors view.

Supporting Information

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Kinetic Model
  5. 3. Results and Discussion
  6. 4. Conclusion
  7. Nomenclature
  8. Acknowledgements
  9. Supporting Information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

FilenameFormatSizeDescription
mats_201200074_sm_suppdata.pdf143Ksuppdata

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.