3.1. Evidence of Success in Synthesizing Low (Half Critical) Molecular Weight High Ion Content Zn-salt SCA Ionomer
It is seen from
Table 1 that the weight-average molecular weight (
) and polydispersity index (
) of the product resin of the suspension free-radical copolymerization, presumably an SCA, were 21,000 g mol
−1 and 1.8 respectively, which contrast sharply with the much higher
(180,000 g mol
−1) and larger
(3.2) of the commercial PS (GPPS 666D). Thus, the
of the potential SCA was tailored to approximately half of the critical
, i.e., 50,000 g mol
−1 [
29], of PS with respect to its
–
Tg relationship. Such a low (i.e., half critical)
might maximally improve the processability (i.e., melt fluidity) of the potential SCA, despite the possible CA introduction of interchain hydrogen bond cross-links, while essentially maintaining its highest possible
Tg. Usually, this is because the ultimate
Tg (i.e.,
) at ca. the critical
of polymers decreases little with a reduction in their
by half.
The CA contents,
cas and
cCA, of the presumable as-synthesized and purified SCA resins, respectively, were evaluated from Equation (1) by the acid–base titration to be 13.6 and 10.8 mol % (
Table 2). Subject to the reactivity ratios of styrene and CA monomers, the product resin might primarily comprise one of three possible species, PS/poly(cinnamic acid) (PCA) blend, PS, and SCA. From the resin’s purified titration sample by reprecipitation from xylene into ethanol (cf.
Section 2.4), free (i.e., unpolymerized) CA and/or PCA molecules, if any, effectively were eliminated by the dissolving ethanol. Therefore, if it had been a PS/PCA blend or PS, the resin would have presented a near zero
cCA in its purified sample (basically a PS), which obviously was not the case. In other words, upon its purification (i.e., removal of any free CA from it), the resin actually showed a significant
cCA of 10.8 mol % (
Table 2(b)) considering the CA-comonomer feed ratio of 20 wt % (approximately 15.0 mol %), which preliminarily indicates success in the copolymerization trial of CA into PS to form an SCA. It is also noteworthy that the potential as-synthesized SCA had a
cas of 13.6 mol % (
Table 2(a)), suggesting that the residual, free CA molecules present in it accounted for approximately 2.8 mol %.
According to
Section 2.5, supposedly an SCA–Zn was then prepared by melt neutralization of the presumable as-synthesized SCA with an excess (2.5 times the stoichiometric amount) of ZnO. Shown in
Figure 1a are the FTIR absorption spectra of the potential (2) SCA and (3) SCA–Zn against (1) the PS (GPPS 666D). Note that, for PS-based chain structural verification purposes, all of the three FTIR resin samples used were purified versions (cf.
Section 2.6) to remove therefrom impurities, CA, PCA, their Zn salts, etc. species. In Trace 1 of the PS, the bands at 1601, 1583, 1493, and 1452 cm
−1 were attributed to the stretches of the phenyl C–C bonds [
30,
31]. Trace 2 for the potential SCA exhibited, besides all of the above bands, new peaks at 1745, 1701, and 1270 cm
−1, of which the former two arose from the stretches of free and (cyclic) dimeric carboxyl C = O bonds, respectively [
7], and the latter arose from the stretch of carboxyl C–O bonds [
32,
33]. This reveals success in synthesizing SCA resin by the suspension free-radical copolymerization of styrene and CA. Further, Trace 3 of the potential SCA–Zn showed, apart from all of the Trace 2 features, new bands at 1723, 1641, and 1415 cm
−1, of which the former was ascribed to the stretch of noncyclic dimeric carboxyl C = O bonds [
7], and the latter two, respectively, were ascribed to the antisymmetric and symmetric stretches of the C–O bonds of carboxylate groups (i.e., –COO
− anions coordinating metal cations) [
34,
35]. This discloses that the potential SCA successfully was melt neutralized partly to produce an SCA–Zn. As the 1641 and 1415 cm
−1 characteristic absorptions were relatively strong, the degree of neutralization (and thus the effective ion content), although difficult to quantify due to the presence of residual, unreacted ZnO in the purified sample, appeared to be rather high in the potential SCA–Zn as a result of the excess ZnO feed.
Likewise, the same PS, potential SCA, and SCA–Zn samples purified were used for
1H NMR spectroscopy (cf.
Section 2.7) and TGA (cf.
Section 2.8) to further verify their chain structures.
Figure 1b shows the
1H NMR spectra of the potential (2) SCA and (3) SCA–Zn relative to (1) the PS. Numbered in
Figure 1c are the supposed H atoms of the three resins, to which their
1H peak chemical shift (
δ) values in
Figure 1b were assigned as given in
Table 3. Note that in
Figure 1b(1)–(3), there were 1.49 and/or 7.28 ppm peak, respectively, of the residual H
2O and/or CHCl
3 in the chloroform-
d solvent [
36]. In
Figure 1c(1) versus 1b(1) of the PS, the peaks of the H atoms of its CH
2 (no. 1), aliphatic CH (no. 2), phenyl ortho-(no. 3), meta- and para- (no. 4) CH, respectively, appeared at
δ values of 1.45, 1.83, 6.45 and 6.56, and 7.07 and 7.12 ppm [
37]. Compared with the PS, the potential SCA in
Figure 1c(2) versus 1b(2) did not exhibit any new peak(s) for its carboxyl hydrogens (no. 5) as well as their neighboring α-H atoms (no. 6), which presumably was due to the high activity of the carboxyl groups with respect to H atoms [
38]. However, contrasted with the potential SCA, the presumable SCA–Zn demonstrated, for H-atoms no. 7, an additional new peak at 1.62 ppm (cf.
Figure 1c(3) versus 1b(3)). This had to be owing to the Zn electrophilic substitution of part of the carboxyls at their H atoms during the neutralization, which deactivated some of the α-H atoms (no. 6) into those (no. 7) that displayed their peak at 1.62 ppm [
39]. The full reconciliation of
Figure 1b,c in
Table 3, discussed above, verifies that the SCA–Zn synthetic trial was successful. Since the 1.62 ppm peak was pronounced enough in
Figure 1b(3), the (effective) ion content, dictated by the degree of neutralization, of the potential SCA–Zn should be rather high, especially considering the 10.8 mol % of CA partly neutralized by the excess ZnO (2.5 times the stoichiometric amount).
Figure 1d gives the TGA thermograms of the three resins, indicating the thermal decomposition behaviors of the potential (2) SCA and (3) SCA–Zn as opposed to (1) the PS. It is seen that compared with the PS
Td of 392 °C, the potential SCA
Td sharply was reduced to 315 °C, which was likely due to its much lower (i.e., half critical)
(cf. Paragraph 1,
Section 3.1) as well as susceptibility to decarboxylation with CO
2 evolution [
40]. Nevertheless, upon the melt neutralization, the
Td of the potential SCA–Zn rebounded greatly to 379 °C approximating to that of the PS, which was probably thanks to a significant decrease in the vulnerable carboxyls fraction and hence dense ionic cross-linking resembling an
increase effect. The above TGA discussions further prove that a high ion content SCA–Zn, having a half critical
inherited from its SCA precursor, successfully was synthesized.
Finally, the same probable SCA and SCA–Zn samples purified were subjected to intrinsic viscosity ([
η]) measurement to confirm the success in the SCA–Zn synthesis; the [
η] of the PS was not tested for the comparative studies due to its much higher (i.e., commercial)
. It was found from
Table 4 that upon the melt neutralization, the [
η] of the likely SCA–Zn was reduced considerably until 0.162 dL g
−1 from that (0.197 dL g
−1) of the likely SCA. Considering the same degree of polymerization (i.e., chain length) of the two resins, this had to be ascribed to a remarkable occurrence of intramolecular ionic cross-links through the significant introduction of Zn-carboxylate ionic groups, which shrank a single coiled SCA–Zn macromolecular chain in radial size to a large extent. All of the above contexts, including the GPC, titration, FTIR, NMR, TGA, and [
η] results, corroborate conclusively that a low (i.e., half critical)
, high ion content SCA–Zn ionomer, indeed, was synthesized successfully by the styrene–CA suspension free-radical copolymerization followed by the excess ZnO melt neutralization.
3.2. Minimization of a Processability Mismatch between PS and the SCA Zn Salt during their Melt Blending
Figure 2 illustrates the complex viscosity magnitudes (|
η*|’s) at 220 °C, during angular frequency (
ω) sweep tests, of the (2) SCA and (3) SCA–Zn against those of (1) the PS. In the light of the Cox–Merz rule [
41] shown below by Equation (4),
the obtained |
η*|–
ω traces in the oscillatory shear mode are reduced to flow curves of shear viscosity (
η) versus shear (strain) rate (
) in the steady-state shear mode. It is observed that across the range of (10
−1–10
2 s
−1)
values investigated, the
η values (Trace 2) of the SCA all fell below those (Trace 1) of the PS, which apparently resulted from competition between two effects. One effect, positive, was the sharp
lowering of the SCA until the half critical
of PS (
Table 1), which dramatically reduced its melt
η from that of the (commercial) PS. While the other effect, negative, should be the dense hydrogen bond cross-linking along with the rigid, polar copolymerization of the SCA by 10.8 mol % of CA (
Table 2), which significantly increased its melt
η. Obviously, the positive effect predominated over the negative to give a very low
η precursor, the half critical
SCA, to its high ion content ionomer. In addition, note that the SCA (Trace 2) presented a more distinct shear thinning behavior than the PS (Trace 1), which was presumably a consequence of progressive rupture of its hydrogen bond cross-links with increasing
.
Upon the melt neutralization, the
η values (Trace 3) of the high ion content SCA–Zn greatly were enhanced by ca. 2.5–3.0 orders of magnitude in contrast to those (Trace 2) of the SCA. This essentially was due to the introduction, into the SCA–Zn, of dense, ionic Zn-carboxylate triplet and aggregate cross-links, which are much higher in bond strength than the hydrogen bond cross-links in the SCA. As a result, the SCA–Zn’s
η values (Trace 3) in turn lay well (ca. 1.5–2.0 orders of magnitude) above those (Trace 1) of the PS at all of the
values studied (
Figure 2). Nevertheless, the
η differences between the SCA–Zn and PS already were strikingly smaller than if the SCA precursor had been of commercial
, when its
η–
trace would have risen above that (Trace 1) of the PS and become near parallel to that (Trace 2) of the half critical
SCA with possible
η differences of at least 2.0 orders of magnitude. In other words, tailoring the SCA to its half critical
and thereby retaining its
Tg close to
minimized the
η differences (i.e., processability mismatches) between the SCA–Zn (Trace 3) and PS (Trace 1) while, basically, it did not compromise the former’s potential to improve the latter’s heat resistance during their melt blending. It is particularly worth noting that the SCA–Zn (Trace 3) displayed milder shear thinning than the SCA (Trace 2), which was likely due to its introduction of stronger ionic cross-links that, unlike hydrogen bond cross-links, were not prone to irreversible disruption at all with increasing
. Regardless of this, the
η difference at a terminal
of 10
2 s
−1, approximate to the average
during the plasticorder melt blending (
Section 2.10), was merely approximately 1.5 orders of magnitude between the half critical
, high ion content SCA–Zn (Trace 3) and PS (Trace 1), which made the melt blending of PS and SCA–Zn possible.
In the above context, the PS and SCA–Zn were then melt-blended, at 220 °C and an 80 rpm rotor speed, at five decreasing mass ratios of 100/0 (i.e., plasticated PS), 95/5, 90/10, 80/20, and 60/40 (
Section 2.10). To examine the effect of the processability mismatch minimization, the morphologies of the PS/SCA–Zn blends were observed at 220 °C by optical microscopy, as illustrated in
Figure 3, where the dark domains represented the SCA–Zn discrete phase while the bright background represented the PS matrix in each of the images. Apparently, all of the blends, except for the neat PS (
Figure 3a), had a fine two-phase morphology of SCA–Zn particles delicately dispersed in a PS matrix with very blurred phase boundaries. This indicates that the PS–(SCA–Zn) processability mismatch presumably was minimized to result in acceptable interfacial compatibility and hence interfacial adhesion of the blends. Scrutinization of
Figure 3b–e further reveals that, dramatically, the particle size of the SCA–Zn became steadily larger with a monotonic increase in its content (5–40 wt %), suggesting that the interfacial compatibility might deteriorate significantly with increasing the SCA–Zn content of the blends.
More quantitatively, the effect of the processability mismatch minimization was probed by a rheological oscillatory shear approach. Shown in
Figure 4 are the Han (i.e., log
G′ versus log
G″) plots (Traces 2–5) of the PS/SCA–Zn blends against that (Trace 1) of the (neat) PS at 220 °C. It is observed that, similar to the PS (Trace 1), the Han plots of the 95/5 (Trace 2) and 90/10 (Trace 3) blends both demonstrated a slope of approximately 2 in the lower
ω (i.e., lower left) terminal region, which was indicative of their almost homogeneous (i.e., single) phase morphology [
42]. This discloses that instead of their apparent two-phase morphologies shown in
Figure 3b,c respectively, both of the 95/5 and 90/10 blends actually were near miscible at 220 °C, which constitutes strong evidence that the PS–(SCA–Zn) processability mismatch indeed was minimized using the half critical
SCA precursor. However, as the SCA–Zn content was raised further beyond 10 wt % more and more, the resulting 80/20 (Trace 4) and 60/40 (Trace 5) blends had Han-plot terminal slopes increasingly divergent from (i.e., smaller than) 2, revealing their gradually heterogeneous (i.e., two-phased) morphologies [
43]. On the whole, with a monotonous rise in their SCA–Zn content, the miscibility of the blends progressively was reduced, which corresponds to the steady decrease in their interfacial compatibility apparently observed from
Figure 3.
3.3. An Effective Improvement in the Heat Resistance of PS upon its Blending Modification with a Significant Fraction of the SCA Zn Salt
Figure 5a and
Table 5(a–c), respectively, give the
Tg and
VST values of the PS, SCA, and SCA–Zn. It is seen that the
Tg and
VST (109.2 and 108.2 °C), respectively, of the SCA were approximately 10 °C higher than those (99.7 and 97.7 °C) of the PS. This primarily was from the chain-segmental hydrogen bond cross-linking as well as rigid, polar copolymerization of the SCA by 10.8 mol % of CA, since the SCA’s heat resistance should little be reduced, from that of the (commercial) PS, by its half critical
. More dramatically, the
Tg and
VST of the SCA–Zn were enhanced to 146.2 and 153.1 °C, respectively, from those of the SCA by 37 and 45 °C, which obviously was due to its significant ionization to form dense, ionic (i.e., Zn-carboxylate) chain-segmental cross-links much stronger in strength than hydrogen bond cross-links. Interestingly, upon the 10 N loading and slower (50 °C h
−1) heating (cf.
Section 2.14), while the PS and SCA displayed their
VSTs (97.7 and 108.2 °C), respectively, which were slightly lower than the
Tg values (99.7 and 109.2 °C), the SCA–Zn showed an opposite trend; its
VST (153.1 °C) was considerably higher than
Tg (146.2 °C). This infers that the high ion content SCA–Zn likely had, according to the multiplet-cluster model [
44], a two-phase morphology of matrix phase (
Tg = 146.2 °C) and stiffer cluster phase whose higher
Tg was difficult to discern by DSC. As such, the
VST, a morphological behavior of the SCA–Zn should be intermediate between the
Tg values of its matrix and cluster phases, i.e., higher than its matrix
Tg.
Figure 5b and
Table 5(a,d–g), respectively, show the
Tg and
VST values of the PS/SCA–Zn melt blends ranging from 100/0 (i.e., the PS) to 60/40. It is found from
Figure 5b that the four blends (Traces 4–7) all exhibited almost the same
Tg as the PS (Trace 1), approximately 100 °C, which arose typically from the phase behavior nature of glass transition. In the near miscible (cf.
Figure 4) 95/5 and 90/10 blends, the SCA–Zn chain-segmental ionic cross-links essentially were dissolved (i.e., destructed) by the major PS chains, making the minor SCA–Zn rigidification of the PS negligible (Traces 4 and 5,
Figure 5b); even though the two blends were phase-separated near 100 °C (their near miscibility was observed from
Figure 4 merely at 220 °C), a single
Tg probably occurred for the PS matrix phase with the matrix
Tg of the minor SCA–Zn phase indiscernible. Likewise, for the apparently two-phased (cf.
Figure 4) 80/20 and 60/40 blends, there appeared only one
Tg of the PS matrix phase despite the larger SCA–Zn fractions (Traces 6 and 7,
Figure 5b), which was presumably due to the significantly weaker glass-transition step (Trace 3,
Figure 5a) of the SCA–Zn’s matrix phase than that (Trace 1,
Figure 5a) of the PS. Altogether, the two-phased PS/SCA–Zn blends (i.e., Traces 4–7 or 6 and 7,
Figure 5b) invariably demonstrated a PS-matrix
Tg little shifting from that of the neat PS (Trace 1,
Figure 5b) toward a higher temperature, suggesting that their claimed acceptable interfacial compatibility at 220 °C (
Figure 3) actually was not good enough at approximately 100 °C to give any convergence of the two components’
Tg values (99.7 and 146.2 °C).
Nevertheless, similar to that (
Table 5(c)) of the SCA–Zn, the
VST data (
Table 5(a,d–g)) of the blends had to be interpreted in the context of the morphological behavior character of softening. That is, contingent upon their miscibility and composition, the single
VST of the blends should be dictated by a weighted combination of the
VST values, 97.7 and 153.1 °C respectively, of their components, PS and SCA–Zn. To elucidate this straightforwardly, the heat resistance (i.e.,
VST) was plotted against the SCA–Zn content of the blends as shown in
Figure 6, using the
Table 5(a,c–g) data. It is apparent that, with increasing the SCA–Zn content up to 20 wt %, the
VST was raised satisfactorily linearly, as well as quite slightly given the much steeper slope (i.e., dashed line) between the PS and SCA–Zn
VSTs (97.7 and 153.1 °C respectively). This likely infers that the PS/SCA–Zn (95/5, 90/10, and 80/20) blends all were more or less near miscible at approximately 100 °C, since miscible blends frequently feature a linear composition–property empirical relationship [
45] and that only when the blends basically were miscible could the strong, dense ionic cross-links of the minor SCA–Zn thoroughly be dissolved (i.e., destructed) by the major PS to pose the markedly milder linearity slope (i.e., dotted line) than the dashed line. In other words, the SCA–Zn rigidification of the PS was rather weak in the miscible blends where there primarily remained a slightly stiffer CA and CA–Zn copolymerization effect due to the C–C bond internal rotation limitation by the steric hindrance of carboxyl and Zn-carboxylate groups.
As the SCA–Zn content further was enhanced to 40 wt %, the situation of the blend changed in that its
VST conspicuously diverged, positively (i.e., upwards), from the dotted-line linearity until approximately 111.1 °C (
Figure 6 and
Table 5(g)), which was approximately 13.5 °C higher than that (97.7 °C) of the PS. This indicates that the PS/SCA–Zn (60/40) blend distinctly was phase-separated at approximately 110 °C into a two-phase morphology, probably, of rigid (
VST = 153.1 °C) SCA–Zn particles dispersed in a far more flexible (
VST = 97.7 °C) PS matrix. Meanwhile, the 111.1 °C
VST data point fell below the dashed-line linearity, revealing that the blend’s interfacial compatibility was unsatisfactory at approximately 110 °C [
46], which agrees with the DSC observation of its not good enough interfacial compatibility at approximately 100 °C (Trace 7,
Figure 5b). Generally, the
VSTs data (
Figure 6) demonstrated stepwise reduced miscibility with raising the SCA–Zn content of the blends, which, for the most part, is consistent with the microscopic and rheological results observed from
Figure 3 and
Figure 4, respectively. In summary, across the SCA–Zn contents (0–40 wt %) studied, not until it was as significant at up to 40 wt % did the PS/SCA–Zn blend evolve a typical two-phase morphology despite insufficient interphase compatibility, in which the strong, dense ionic Zn-carboxylate cross-links of the dispersed SCA–Zn phase contributed primarily to the effective improvement in the blend’s heat resistance (i.e.,
VST increase by approximately 13.5 °C).
3.4. Deteriorations in the Mechanical Strengths of PS upon its Incorporation of the SCA Zn Salt
While the half critical
, densely ionically cross-linked SCA–Zn, at a 40 wt % fraction, minimized its processability mismatch with the PS and significantly enhanced the heat resistance of the phase-separated PS/SCA–Zn blend, it was imperative to examine the mechanical and rheological properties of the blend against the PS from a polymer engineering viewpoint.
Table 6 lists the mechanical properties, i.e., Young’s moduli (
E), tensile strengths (
σ), elongations at break (
ε), and Charpy notched impact strengths (
αcN), of the PS/SCA–Zn blends having the increasing SCA–Zn contents (0–40 wt %), from which the composition–property relationships were plotted as illustrated in
Figure 7. It is seen from
Figure 7a that the
E heightened first linearly and moderately with increasing the SCA–Zn content up to 20 wt %, and then it heightened nonlinearly and sharply as the SCA–Zn content was raised further to 40 wt %. This behavior, analogous to that of the
VST change with the SCA–Zn content shown in
Figure 6, seems to verify that for similar reasons, the blends were near miscible at the SCA–Zn contents of ≤ 20 wt %, whereas they were distinctly phase-separated with insufficient interfacial compatibility at 40 wt % at RT. However, the
σ,
ε, and
αcN values all had a general tendency to decrease steadily with a monotonic rise in the SCA–Zn content of the blends (
Figure 7b–d, respectively). This possibly was owing to the mechanical weakness and brittleness of the dissolved low-
SCA–Zn without ionic cross-links or polar chain-segmental interactions in the miscible blends (95/5, 90/10, and 80/20), as well as to primarily the unsatisfactory interfacial compatibility that caused insufficient interphase adhesion in the two-phased blend (60/40). Note that for the miscible blends without interfaces, while the
E was enhanced primarily from the rigid CA and CA–Zn copolymerization on a mer scale, the
σ,
ε, and
αcN were decreased due essentially to the SCA–Zn’s low
(i.e., plasticization and embrittlement) on a wholly macromolecular scale. Further, for the 60/40 phase-separated blend with the strong ionic cross-links, unlike the
E insusceptibility to interfacial softening upon small strain and stress, all of the
σ,
ε, and
αcN values were quite vulnerable to a deterioration from the low-strength interphase upon much larger strain and stress. Consequently, compared with the
E,
σ,
ε, and
αcN (472.6 MPa, 34.9 MPa, 5.9%, and 3.2 kJ m
−2 respectively) of the PS, the
E was enhanced until 884.6 MPa by 87.2%, while the
σ,
ε, and
αcN were decreased to 21.0 MPa, 3.4%, and 1.7 kJ m
−2, respectively, by 39.8%, 42.4%, and 46.9% for the PS/SCA–Zn (60/40) blend (cf.
Table 6). These results show that the SCA–Zn 40 wt % incorporation appreciably improved the heat resistance of the PS/SCA–Zn blend essentially at the expense of its mechanical (tensile and impact) properties due to the two phases insufficiently adhering to each other.