3.1. RAFT Mediated Miniemulsion Polymerization
DBTTC is a symmetrical RAFT agent that has two homolytic leaving R groups, which enable the polymerization of ABA triblock copolymers in two sequential monomer additional steps as shown in
Figure 1.
When the initial monomer is added it inserts in between the leaving groups and –S(C=S)- moiety. The first polystyrene (A) block is grown this way and is extended by the second monomer addition (2EHA, B block), which accommodates in the middle and shifts the A block to the outside of the polymer chain.
Polystyrene A blocks with targeted
Mn of 30,000 (pSt30) and 50,000 g/mol (pSt50) were synthesized using the following equation:
where
represents the predicted molecular weight,
represent the molecular weight of the monomer,
is the molecular weight of the RAFT agent, x is the conversion and
and
are the initial moles of the monomer and the RAFT agent, respectively. The amount of initiator was not considered in the calculation of the theoretical
Mn due to the relatively small amount used compared to the RAFT agent. This approach was also used by de Brouwer [
23], who disregards the chains derived from the initiator in the calculation, due to the low efficiency of initiator systems such as AIBN in heterogeneous polymerizations. The amount of co-stabilizer used was also not included in the calculation of the targeted
Mn. SA is an extremely hydrophobic monomer that was used to prevent the Oswald ripening in the miniemulsions [
24,
25]. It incorporates in the polymer backbone upon polymerization and stays in the film upon film formation, unlike the widely used hexadecane co-stabilizer, which evaporates into the atmosphere and increases the volatile organic content of the latex.
Prior to polymerization, the stability of the miniemulsion was investigated by Turbiscan measurements and the results showed excellent stability during 6 h at 60 °C (see
Figure S2 in Supporting Material). Polystyrene initial blocks were then synthesized using a molar ratio of RAFT:AIBN = 5:1, at a reaction temperature of 70 °C and with a reaction time of 6 h. The evolution of the droplet-particle size is shown in
Figure 2a. It can be seen from the graph that dp initially increases in the first 2 h with the progress of the Ostwald ripening process, and ultimately a relatively stationary particle size is achieved. The particle size at the end of the reactions was 30 nm higher than the droplet size for both pSt30 and pSt50.
The conversion versus time plot is shown in
Figure 2b and one can observe that the conversion increment with time was not linear. In fact, if the ln(1/(1 − x)) vs time plot is plotted (see
Figure S3), it can be observed that the reaction did not follow first order kinetics with respect to styrene monomer, most likely due to a non-constant concentration of radicals. After 6 h of reaction 80% conversion was obtained for both pS30 and pSt50. The reaction was stopped at 6 h even if full conversion was not reached, since a living system was desired for the extension of the initial block with the soft monomer.
Mn for both pSt30 and pSt50 increased with conversion (
Figure 3) following the theoretical prediction showing that the polymerization proceeded in well controlled manner. Moreover, the whole MWD shifted to higher molecular weight by increasing the conversion (see
Figure S4a,b). Therefore, it could be considered that the polymerization was living and well controlled.
3.2. ABA Triblock Copolymer Latex
Once the initial pSt blocks were formed with 30,000 and 50,000 g/mol of targeted molecular weight (pSt30 and pSt50), the extension with the second soft monomer proceeded. The 2EHA monomer was fed as a pre-emulsion for three hours with additional amount of initiator added. Four block copolymers were synthesized having different molecular weights, named through the text as: p(St25-2EHA50-St25) as an extension of pSt50 and p(St15-2EHA50-St15), p(St15-2EHA70-St15) and p(St15-2EHA100-St15) as extensions of pSt30. The results of initial and the final block copolymers are shown in
Table 1. The numbers after the letters indicate the targeted molecular weight. The second monomer was fed without any removal of the unreacted monomer (styrene) from the first step. Koiry and Singha studied the copper mediated controlled radical copolymerization of styrene and 2-ethylhexyl acrylate and determined their reactivity ratios to be r
1 (styrene) = 1.29 and r
2 (2EHA) = 0.73 [
26]. Considering these reactivity ratios, the middle soft block is expected not to be pure poly(2EHA), but rather a gradient polymer that will still incorporate some styrene at the beginning of the polymerization of the second block. However, for simplicity reasons, this middle gradient block will be named as soft poly(2EHA) block throughout the text. Regarding the particle size, it can be seen that the particle size increased from the pSt seed to the block copolymers, indicating the growing of the particles by the fed 2EHA.
The evolution of the
Mn versus total monomer conversion (determined by solids content measurements) for block copolymers with different
Mn prepared from pSt30 seed is shown in
Figure 4.
Mn grows linearly with conversion, however there is a visible negative deviation after the addition of the second monomer. The deviation gets more pronounced with the increase of the polymerized middle soft block. This is most likely due to the fact that the
Mn’s obtained from GPC were based on pSt standards and the Mark-Houwink constant of p2EHA was not taken into consideration. Equation (2) presents how the molecular weights of two different polymers can be related using the Mark-Houwink equation. In the equation K
1, α
1 and K
2, α
2 represent the Mark-Houwink parameters for polymer 1 and 2, respectively, and M
1 and M
2 are the molecular weight of the polymers. Thus, as seen from Equation (2)—taking the constants for pSt (K = 0.000158 and α= 0.704) and p2EHA (K = 0.000124 and α= 0.667) into consideration—we were underestimating the real
Mn.
Furthermore, it can be seen (
Table 1) that all the final block copolymers have an overall conversion above 80% and the polydispersity index increased as the length of middle block was increased. Nevertheless, it should be noted that triblock copolymers were successfully formed, as is evident from the GPC curves (
Figure 5), where the MWD of the final block copolymer moved to higher molecular weights compared to initial pSt block.
Widening of the MWD in the region of high molecular weight was observed when a higher molecular weight was targeted (for p(St15-2EHA70-St15) and p(St15-2EHA100-St15)
Figure 5b,c respectively). This is most likely due to branching reactions, characteristic for acrylates [
27,
28] occurring as a result of combination of intermolecular chain transfer to polymer, which leads to long chain branches in the polymer, and termination by combination of branched growing radicals, leading to network formation and eventually gel polymer.
3.3. Thermal Properties of the Initial Homopolymers and Final Block Copolymers
The glass transition temperatures of the pSt homopolymers and p(St-2EHA-pSt) block copolymers were determined by DSC analysis from the samples dried at room temperature (
Table 2).
Polystyrene initial blocks having different molecular weights showed a single
Tg with very similar values (58 °C for pSt30 and 54 °C for pSt50). The
Tg values obtained are much lower than the
Tg of pSt reported in the literature (100–107 °C) [
29]. This comes from the fact that the reactive co-stabilizer stearyl acrylate was used to prevent Oswald ripening in the miniemulsion. When stearyl acrylate is polymerized, a semi-crystalline polymer is obtained with a low
Tg of −100 °C and a melting temperature of 50 °C [
30]. Thus, when styrene copolymerizes with stearyl acrylate, its
Tg is reduced. Another possible reason for the low
Tg is the fact that the initial pSt blocks are of low molecular weight. Fox and Flory [
31] studied the dependence of second-order transition temperature of polystyrene on the molecular weight and concluded that the
Tg for high molecular weight polystyrene is 100 °C and decreases linearly with 1/
Mn [
31]. Therefore, both the addition of reactive co-stabilizer SA in the miniemulsion and the relatively low molecular weight pSt block can explain the
Tgs around 54–58 °C that have been found for the first block.
The triblock copolymers on the other hand exhibited two distinct transition temperatures—a lower and an upper one (as seen from
Figure S5 in the Supplementary Material). The lower transition temperature ranges from −43 to −60 °C, depending on the composition of the block copolymers, and it is associated with the soft middle block. The difference in the values obtained for the lower
Tg is due to the difference in the composition of the soft middle block. The block copolymers with higher overall composition of pSt contain soft middle block chains that bear higher amount of St units (unreacted from the first step). As a result, the
Tg of the soft domains shifts to higher temperatures closer to the
Tg of the pSt. On the contrary, block copolymers with higher overall amounts of p2EHA are composed of middle soft chains bearing more 2EHA units. Thus, the lower
Tg of this block copolymer will get closer to the
Tg of p2EHA (−60 °C) [
26]. The upper transition temperature ranges from 55 to 63 °C and corresponds to the
Tg of the pSt initial block. Therefore, there is a clear indication of a two-phase system.
Furthermore, the thermal stability of the copolymers was studied by TGA in nitrogen atmosphere and the results are summarized in
Table 3 (see also
Figure S6 in the Supplementary Material). The temperature at 10% weight loss was taken as
Tonset.
Tmax represents the maximum degradation temperature at which polymer back-bone starts degrading and forming a lot of volatile decomposition products and was determined from the TGS thermograms at the maximum of the derivative curve. It can be seen from the results that as the styrene content in the block copolymers increased, there was a slight increase in
Tonset.
Tmax on the other hand decreased as the length of the middle block increased. Moreover, DBTTC degradation started at 240 °C, and at 300 °C it was completely degraded. All the block copolymers showed a single step decomposition at rather high temperatures well above 200 °C, which is comparable to similar block ABA acrylic block compolymers synthesized by RAFT miniemulsion polymerization [
22].
3.4. Morphology of the Block Copolymers
Block copolymer morphology was studied both by TEM and AFM. Due to the complexity of the system, i.e., mixed monomer sequence arising from the presence of unreacted monomer (initial block: styrene + stearyl acrylate and middle block: styrene + 2-ethylhexylacrylate), it was not possible to exactly determine the interaction parameter χ and also not possible to precisely determine the volume fraction of each microphase. Thus, it was not possible to precisely predict the equilibrium morphology of the block copolymers. However, considering the publication by Lee et al. [
32] and Wang and Robertson [
33], an interaction parameter of 0.025 for polystyrene and p(2EHA) can be estimated. Considering the smallest block copolymer p(St15-2EHA50-St15), its N was around 450 (156 (pSt + pSA) + 294 p2EHA). Therefore, χN was 11.25, which is above the phase separation limit (10.5). This is an indication that phase separation for all block copolymers can be expected. An estimate of the volume fraction of each component in the block copolymers was made based on the compositional information derived from NMR (
Figure S1 in the Supplementary Material). The assumption that the weight fraction of monomers is equal or close enough to the volume fraction was made and the results are shown in
Table 4. As stated in the review article of Mai and Eisenberg [
34], the morphology of AB block copolymers changes when increasing the volume fraction f
A of the A block at a fixed χN > 10.5. When the volume fraction of a hard domain (such as polystyrene) is small (<20 vol %) spheres of polystyrene dispersed in an elastic matrix are formed, which then change to cylinders or gyroids as the pSt content increases. When the volume fractions of both components are about equal (40–60 vol %), the two-component form alternating lamellae. The prediction of the theoretical morphologies of the synthesized block copolymers was made based on the volume fractions calculated from NMR and the morphology diagram by Bates at al. [
35], see
Table 4.
The morphology of the block copolymer with composition p(St15-2EHA50-St15) was analyzed by AFM and the results are shown in
Figure 6. The top surface of the polymer film prepared by drop casting of the latex at room temperature (
Figure 6a) reveals the presence of spherical particles with hard shell (shown as bright regions), which are smaller than the particle size obtained by DLS (187 nm), dispersed in a nanophase separated continuous matrix, which most likely formed as a result of particle coalescence. The cross section of the AFM latex film dried at room temperature, (
Figure 6b) showed no presence of particles—instead, an extended nanophase separated pattern was visible. The most likely reason for observing a regular structure already at room temperature is the fact that this sample contained a substantial amount of unreacted monomer (as seen in
Table 1), acting as a plasticizer and causing particles to coalesce already at room temperature. On the other hand, the sample annealed at 100 °C (
Figure 6c) showed even higher ordering compared to the sample obtained at room temperature. Nevertheless, no significant difference in morphology was seen between the films annealed at higher temperature and the ones cast from THF solution, indicating that equilibrium morphology was already reached by thermal treatment of the films. According to the diagram for AB block copolymers, the sample p(St15-2EHA50-St15), based on its volume fraction, should phase separate either in cylinder or gyroid structure depending on the interaction parameters. If we take into account
Figure 6b–d, it seems that the gyroid morphology is the predominant one for this block copolymer.
The self-assembly of the block copolymer with composition p(St15-2EHA70-St15) was studied by AFM (
Figure 7) and compared to TEM images (
Figure 8). The top surface of the polymer film (
Figure 7a) studied by AFM shows the existence of two different populations of particles having lower and higher amount of styrene—visible as white dots. The AFM imaging performed on a cross-section of a latex film dried at room temperature reveals the presence of spherical particles inside of which “onion-ring” lamellar morphology can be distinguished (
Figure 7b).
These findings were also confirmed by TEM (
Figure 8b) where it is clearly visible that almost all the particles show the same structure. In the TEM images, the styrene rich phase appears dark due to the RuO
4 staining. Moreover, from the TEM images (
Figure 8a,b) it is seen that the pSt rings are perforated, which could be the reason for observing the white dots in the particles in AFM (
Figure 7a). Although measurements derived from AFM are only an approximation, the size of the spherical objects shown in
Figure 7b is in good agreement with the average particle size measured by DLS (189 nm). The thermal treatment of the polymer film caused complete particle coalescence and transformed the “onion-ring” structure into more classical lamellar morphology. To completely erase the impact of the emulsion polymerization process and thermal history, the latex film was dissolved in THF (
Figure 7d). This yielded a morphology as close as possible to equilibrium morphology and the obtained structure resembled the one obtained at 100 °C, although the presence of onion-ring structures was not completely erased in the 100 °C annealed film, as it was after the THF treatment.
Based on the general diagram, block copolymers with a volume fraction of hard domain in the range of 20% should phase separate as either spheres or cylinders. On the contrary, the microscopic data clearly demonstrates that p(St15-2EHA70-St15) block copolymer phase separates into lamellar structure. However, as Matsen and Thompson stated in their article [
36], ABA block copolymers phase separate slightly different than AB block copolymers. They predicted that the lamellar region for ABA block copolymers was reached for lower f
A compared to AB block copolymers, which is exactly what has been observed for p(St15-2EHA70-St15) in this study.
In order to confirm the lamellar structure, SAXS analysis of p(St15-2EHA70-St15) sample was carried out (
Figure 9).
As can be seen, the spectra present three main bands at 0.02, 0.2 and 0.5 Å
−1. The last two bands have been attributed to pSt (0.2 Å
−1 and p2EHA 0.5 Å
−1), but the one at 0.209 Å
−1 can be attributed to the interlamellar distance between soft and hard phases. This q would mean an average interlayer distance of 30 nm. The interlamellar spacing is less developed in the sample dried at room temperature, but it is more clearly seen when the sample is annealed at 100 °C (maximum phase separation between soft and hard phases). The interlayer distance of 30 nm fits with the interlayer distances that can be measured in TEM images (
Figure 8c).
The sample p(St15-2EHA100-St15) with the longest soft block was very sticky, thus it was not possible to analyze by microscopy. The sample with highest pSt content p(St25-2EHA50-St25) showed hard spherical particles with no particular outer morphology as evident form the AFM image of the top surface of the polymer film (
Figure 10a). AFM images of the cross section of the film dried at room temperature on the other hand showed single or bilayer particle morphology (
Figure 10b). At this stage, the differences observed for the top and cross-sections of the films dried at room temperature have to be pointed out. In all three latexes analyzed so far, the differences have been significant, but they are very clear here. The top view only shows the surface of the particles, whereas the cross-section shows their interior. If only the top view would have been considered, no phase separation would have been envisaged from
Figure 10a. On the other hand, annealing of the film led to complete coalescence of the hard particles and homogeneous distribution of the p2EHA domains through the hard pSt matrix (
Figure 10c). Presence of small voids was also visible in annealed films, which could not be seen in the film obtained from THF solution (
Figure 10d). According to the general diagram based on the volume fraction either lamellar or gyroid structure should be expected for this sample. The images obtained, especially from the film cast from THF-solution 10d, suggest that we had a gyroid structure.
3.5. Viscoelastic Properties of the ABA Hard-Soft-Hard Block Copolymers
To investigate whether the obtained morphological changes observed during annealing influenced the viscoelastic properties of films, the hard-soft-hard block copolymers were investigated by DMTA and the results are presented in
Figure 11.
The solid line presents the influence of block copolymer composition or annealing temperature on storage modulus and the dashed line the effect on tan δ. Viscoelastic properties of the films cast at room temperature presented in
Figure 11a indicate that the elastic modulus decreased as the length of the middle soft p2EHA block increased. The elastic modulus decreased significantly in the rubbery plateau, especially in the temperature range between 0 and 50 °C. If the AFM images of
Figure 6 (p(St15-2EHA50-St15)) and
Figure 7 (p(St15-2EHA70-St15)) are observed, it is clear that the periodic size of the 2EHA domains is much smaller for the first case (around 400 nm for p(St15-2EHA50-St15) block copolymer) than for the second one (more than 1 µm for (p(St15-2EHA70-St15)). Therefore, the higher modulus of p(St15-2EHA50-St15) can be attributed to smaller soft 2EHA domains. Moreover, liquid-like behavior induced by further temperature increase was evident from the significant drop of the modulus and abrupt increase in tan δ at temperatures close to the
Tg of the hard domains. The viscoelastic properties of p(St25-2EHA50-St25) were not possible to be measured, because the sample was too brittle and handling was very difficult.
Additionally, in
Figure 11b–d the viscoelastic properties of the block copolymers latex films cast at room temperature were compared with the ones of the latex films annealed at 100 °C. Shifting the
Tg (of soft domains) to lower temperatures, an elastic modulus increase in the plateau region was clearly visible in all the samples upon annealing, irrespective of the composition of the block copolymers. When films were annealed, complete particle coalescence occurred as evidenced from the AFM images, thus pSt domains were able to move and uniformly distribute throughout the elastic matrix, which led to its reinforcement. As a result, hard thermoplastic rubber-like materials were obtained with increased elastic modulus in the plateau region. When the films were dried at room temperature the contact between blocks was most likely higher, and thus their influence on each other was higher. As a result, the lower and upper
Tg’s approached each other. On the other hand, when they were annealed, less contact was achieved and the influence of the blocks on each other was lower. As a result, the two
Tg’s got apart from each other. In addition, the upper
Tg of the hard domain got more pronounced, most likely due to the fact that the domains got bigger by annealing and the interface region between the microdomains in which segments of both blocks mix got narrower.