SAN polymer blends compatibilized with linear ABC triblock terpolymers

Polymer 80 (2015) 52e63

Contents lists available at ScienceDirect

Polymer journal homepage: www.elsevier.com/locate/polymer

Micromechanics of “raspberry” morphology in PPE/SAN polymer blends compatibilized with linear ABC triblock terpolymers € bling b, Holger Schmalz b, Axel H.E. Müller b, c, Ronak Bahrami a, Tina I. Lo €dt a, * Volker Altsta a b c

Faculty of Engineering Science, University of Bayreuth, 95440 Bayreuth, Germany Macromolecular Chemistry II, University of Bayreuth, 95440 Bayreuth, Germany Institute of Organic Chemistry, Johannes Gutenberg University, 55099 Mainz, Germany

a r t i c l e i n f o

a b s t r a c t

Article history: Received 16 July 2015 Received in revised form 15 September 2015 Accepted 20 October 2015 Available online 23 October 2015

The effect of compatibilization with a symmetrical polystyrene-block-polybutadiene-block-poly(methyl methacrylate) (SBM) triblock terpolymer on the morphological and mechanical properties (specifically toughness) of immiscible poly(2,6-dimethyl-1,4-phenylene ether)/poly(styrene-co-acrylonitrile) (PPE/ SAN) blends with different blend (w/w) ratios is investigated. We study the effect of blend viscosity on the localization of the compatibilizer at the blend interface, influencing the mechanical properties of the macroscopic material. The impact of the specific morphology of the blends, known as “raspberry morphology”, on the final material will be explained using thermomechanical analysis and revealing relevant deformation mechanisms. The correlations between domain size, toughness of the blends and the corresponding toughening mechanisms are explained. After compatibilization, the fracture toughness is increased due to better bonding and stress transfer between the phases and smaller domain sizes with strong interface impose shear yielding of the matrix. Very small inter-domain distances hinder matrix deformations, which forces debonding as a “weaker” main deformation mechanism. It is found that a PPE/SAN ratio of 60/40 (w/w) has the optimum viscosity during processing to control the morphology. At this blend ratio the dispersed PPE domains have a low packing density, allowing shear deformations of the matrix. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Polymer blends Compatibilization Micromechanics Fracture toughness Raspberry morphology

1. Introduction Polymer blending is one of the most versatile and economical methods to produce new materials, combining properties from each blend components [1,2]. However, most polymers are immiscible and tend to phase separate from each other during processing. In order to prevent this macrophase separation, different compatibilization methods are frequently used such as grafting the components with similar or miscible polymers [3], using graft copolymers as compatibilizers [4], adding low molecular weight chemicals [4,5], and reactive blending [6e8]. Poly(2,6dimethyl-1,4-phenylene ether) (PPE) (also referred to as PPO (2,6dimethyl-1,4-phenylene oxide) in the American-Pacific region) has a high heat resistance and dimensional stability making this compound a very interesting material. However, due to its high

* Corresponding author. €dt). E-mail address: [email protected] (V. Altsta http://dx.doi.org/10.1016/j.polymer.2015.10.039 0032-3861/© 2015 Elsevier Ltd. All rights reserved.

viscosity and low processability it is commonly blended with other polymers to overcome this drawback. Poly(styrene-co-acrylonitrile) (SAN) is one candidate to blend with PPE as it is easy to process and its high stiffness exerts an synergistic effect on the mechanical properties of the final blend. However, studies show that the miscibility of PPE/SAN blends depends on the acrylonitrile (AN) content of SAN [9,10]. Common commercially available SAN copolymers with AN contents between 19 and 35 wt.% are immiscible with PPE. Block copolymers consisting of two covalently linked polymer blocks are suitable candidates to compatibilize immiscible polymer blends. The chemistry of the two blocks can be synthetically tailored to match the desired blend system and, hence, adsorption to the blend interface is favored [11e14]. Triblock copolymers with an elastomeric middle block have shown to improve the compatibility as well as the mechanical properties such as toughness [14e17]. In case of PPE/SAN blends in particular, polystyrene-block-polybutadiene-block-poly(methyl methacrylate)

R. Bahrami et al. / Polymer 80 (2015) 52e63

(PS-b-PB-b-PMMA, SBM) triblock terpolymers could promote formation of a specific structure called raspberry morphology [14], which is known to improve the toughness of the material without sacrificing its modulus. A common strategy for improving toughness is the addition of impact modifiers (rubber particles, core shell particles, etc.) to the polymer. Since these elastomeric materials have lower modulus and stiffness compared to thermoplastics, their addition decreases these properties in the final compounds accordingly. The effect of SBM concentration and molecular weight of each block on the compatibilization efficiency and localization at the interface have been previously investigated [18,19]. The PS block is thereby miscible with PPE, and the PMMA block with SAN, while the immiscible elastomeric PB middle block forms discontinuous patches at the PPE/SAN interface. The interface and blend morphology are among the most important parameters controlling the blend properties (especially mechanical properties). Recent studies can predict the compatibilization efficiency of the system by molecular dynamics simulation using the Flory-Huggins interaction parameter (c) [17]. There are also experimental studies correlating the size of the dispersed phase in the blends, thickness of the interphase, and interfacial tension to the FloryeHuggins interaction parameter [20]. However, in order to be able to increase the compatibilization efficiency, the search for more efficient compatibilizers such as Janus nanoparticles [21,22], graphene oxide sheets [23], nanoparticles [24], carbon nanotubes [3,25] and other novel functionalized materials [26] is in progress. The effect of compatibilization with block copolymers on the mechanical properties, especially tensile strength and modulus, has been previously investigated [27e29]. In most cases a decrease in the tensile strength and modulus is observed after compatibilization, since the tensile strength of the admixed rubbery block is much lower than those of the blend components (due to the fact that elastomers have a glass transition temperature (Tg) below room temperature). Stadler [30] and later Kirschnik [12] suggested that the special geometry of the raspberry morphology with a discontinuous elastomeric phase at the blend interface allows optimum stress transfer between the phases and the modulus is expected to stay constant. However, the underlying mechanism is still not completely understood. Even though fracture toughness measurements are able to precisely correlate the morphological features and the microstructure with the macroscopic mechanical properties, only few studies have so far focused on these correlations in case of polymer blends [29,31e33]. By studying the fractured surfaces of polyacrylonitrile-polybutadiene-polystyrene (ABS) copolymers toughened with coreeshell particles, Michler [31] suggested core-fibrillation mechanisms for the first time, consisting of fibrillation at the craze interface during craze thickening in glassy polymers. Tiejune et al. [32] investigated complex shear band formation mechanisms combined with rubber cavitation in polycarbonate/ABS blends. Handge et al. [29,33] investigated the micromechanical deformations of polyamide 6/SAN blends, which were compatibilized with maleic anhydride grafted poly(styrene-co-acrylonitrile) (SANMA), via in-situ tensile tests on semi-thin TEM specimens. Here, in the blend with a ductile matrix and rigid particles, local failure is initiated by rupture and crazing of the interface between the constituents. They concluded that the mechanical properties of the SANMA compatibilized PA6/ABS system increase due to improved interfacial adhesion between the blend phases. This effect was very pronounced and exceeded the influence of the particle size on the mechanical properties. In complex systems like SBM compatibilized PPE/SAN blends, it is important to distinguish different effects influencing the micromechanics, considering all factors such as domain size, interfacial adhesion and structure of the raspberry morphology. Hence, the

53

current study thoroughly investigates the morphological features and their effect on the micromechanics and correlates them to the macro-mechanical properties of the blends. In particular, the effect of PPE domain size on micromechanics with special emphasis on the fracture toughness was characterized. PPE/SAN blends with different blend ratios were compatibilized with 10 wt.% SBM triblock terpolymer and the properties of the neat and compatibilized blends were compared. 2. Experimental 2.1. Materials Poly(2,6-dimethyl-1,4-phenylene ether), PPE, was obtained as powder from Mitsubishi Engineering Plastics Europe (commercial grade PX100F), Düsseldorf, Germany with Mw ¼ 12.9 kg/mol. It is important to mention that PPE is pure and without any PS addition and therefore has a high viscosity. The commercial poly(styrene-coacrylonitrile), SAN, has an acrylonitrile content of 19 wt.%, and was obtained as pellets from BASF AG, Ludwigshafen, Germany (SAN VLL 19100, Mw ¼ 97.1 kg/mol). The low acrylonitrile content of the used SAN ensures its miscibility with the PMMA block of the SBM triblock terpolymer at the relevant processing conditions [34,35]. A 2:1 mixture of Irganox 1010 and Irgafos 168 (BASF, Germany), 0.1 wt.% in total, were used as stabilizers during compounding. 2.2. Synthesis of the SBM triblock terpolymer The polystyrene-block-polybutadiene-block-poly(methyl methacrylate) (SBM) triblock terpolymer was synthesized via sequential living anionic polymerization as reported elsewhere in detail [36]. The used S32B36M32 triblock terpolymer (subscripts denote the weight fraction of the respective block) has a number-averaged molecular weight of Mn ¼ 93 kg/mol and a molar-mass dispersity of ÐM ¼ 1.04. 2.3. Melt processing of the blends Before melt blending the polymers, the PPE powder and the SAN granulates were dried at 80  C for at least 12 h under vacuum. For SBM, a lower temperature of 40  C was chosen due to the sensitivity of the PB middle block to crosslink at higher temperatures. The PPE, SAN, and SBM particles were dry blended using powder mixers prior to melt blending. The PPE/SAN weight ratio in the blends was changed from 50/50 to 60/40 and 70/30, while for each blend the amount of SBM compatibilizer was kept constant at an optimum concentration of 10 wt.% according to earlier studies [16]. Melt blending of the compounds was performed in a co-rotating twin-screw extruder (Brabender DSE 20/40) with L/D ¼ 30. The maximum barrel and nozzle temperatures were fixed at 250 and 245  C, respectively, and the screw speed was kept constant at 85 rpm with constant throughput of 1 kg/h. Therefore, the mean residence time of the blends in the extruder was approximately 5 min. In order to prepare specimens for shear rheology and dynamic mechanical analysis (DMA), the compounded granulates were dried at 80  C in a vacuum oven for at least 4 h and then compression molded under vacuum using a hot press for 6 min with 100 kN at 260  C, and later cooled down in a cold press with 30 kN compression load. Specimens were compression molded to eliminate the strong effect of any orientation of the PPE particles during the injection molding process, based on previous studies [16].

54

R. Bahrami et al. / Polymer 80 (2015) 52e63

2.4. Characterization of the blend systems 2.4.1. Morphological characterization Transmission Electron Microscopy (TEM): Ultrathin sections (ca. 50 nm) from the blended materials after extrusion were cut at room temperature using an ultra-microtome (Reichert-Jung Ultracut E) equipped with a diamond knife. The ultrathin sections were stained with OsO4 vapor for 2 h at ambient pressure to enhance the contrast between the phases [14]. Due to this particular staining method, SAN and PPE appear as the lighter and the darker phase, respectively, whereas the PB block in the SBM triblock terpolymer is selectively stained and appears black. The TEM images were taken using a 922 OMEGA EFTEM electron microscope (Carl Zeiss, Jena, Germany) operated at an acceleration voltage of 200 kV. Field Emission Scanning Electron microscopy (FESEM): The fractured surface of the blends after mechanical testing (fracture toughness) was analyzed via bright field emission scanning electron microscopy (Leo 1530 Gemini from Zeiss) using a secondary electron detector and an acceleration voltage of 15 kV. The samples were sputtered with a 1.3 nm thick Platinum layer. 2.4.2. Thermal analysis The dynamic mechanical thermal analysis (DMA) of the blend systems was performed in the dual-cantilever mode on hotpressed rectangular specimens with dimensions of 25  6  1 mm3, using a Mettler Toledo DMA/SDTA 821e. The frequency of the measurement was constant at 1 Hz and the test setup applied tensile forces to the specimens. The applied strain was kept small enough to ensure linear-elastic behavior of all systems. The samples were heated from 100  C (after establishment of equilibrium) to 230  C at a constant heating rate of 2 K/min. The same device was used to investigate the non-linear behavior of the blends with the Payne test. A strain sweep test under tension load was performed with constant frequency of 1 Hz at 150  C. The onset of the decrease in modulus was defined as the intersection of the tangents on the traces. 2.4.3. Rheological investigations Rheological properties were investigated by a stress controlled dynamic-mechanical rheometer (RDA III from Rheometric Scientific) with plateeplate geometry under nitrogen atmosphere. The pressed samples had a diameter of 25 mm and a thickness of 1.5 mm. The storage modulus, the loss modulus, and the complex viscosity of blend systems were measured as a function of frequency within the range of 0.01e500 rad/s at 260  C. Prior to each measurement, the linear viscoelastic region was determined by an amplitude sweep in the deformation range of 0.1e100%, at frequencies of 1 and 50 rad/s. Subsequently, the deformation during the frequency sweeps was set to be within the linear viscoelastic region. The rheological measurements of neat PPE and SAN were performed on samples, which were prepared by extrusion applying the same condition as for the blends. Each measurement was repeated at least three times to minimize the experimental errors. 2.4.4. Mechanical characterization Fracture toughness measurements were performed according to the standard test method ISO13586 to obtain the mode I critical stress intensity factor (KIC) of the polymer blends at 23  C. For each sample, tapping a new razor blade into the machined V-notch generated a sharp crack. Samples were afterwards loaded under tension mode so the crack can grow until the end of the specimen. The crack opening displacement during crack growth is measured using a clip-on extensometer. At least 5 notched, compact tension specimens were tested at a strain rate of 10 mm/min. The thickness of the specimens was 2 mm. The tests were carried out on a Zwick

BZ2.5/TN1S universal testing machine. The critical stress intensity factor was calculated using the following equation:

KIC ¼

  F pffiffiffiffi*f a= w B* w

(1)

Where F represents the force required for the crack to start propagation, B and w are the thickness and width of the specimen respectively, a is the initial crack length, and f is the geometrical term.

3. Results and discussions 3.1. Morphological characterization The morphology of an immiscible polymer blend depends strongly on the rheological properties of the polymer components. In order to understand and explain the morphologies in the blend, the rheological properties of the neat polymers need to be discussed first. In our case, it is expected that the high viscosity difference between the PPE and SAN polymers (shown in Fig. 1) shifts the phase inversion region for having a PPE matrix far from the expected 50/50 blend ratio [37]. The phase inversion of this system has been predicted with the aid of two models proposed by Chen [38] (Eq. (2)) and Utracki [39] (Eq. (3)). The models calculate a threshold value of the viscosity ratio, above which PPE can no longer form the continuous matrix, although PPE is the main component in the blend.

0:3  fPPE h ¼ 1:2 PPE fSAN hSAN P¼

hPPE ¼ hSAN

  fm  fSAN h*fm fm  fPPE

(2)

(3)

Where P ¼ hhPPE is the viscosity ratio, fm is the maximum packing SAN volume fraction equal to 0.84 for most polymer blends [2], fSAN and fPPE are the SAN and PPE weight fractions respectively, and h represents the corresponding viscosities. In case of 50/50, 60/40 and 70/30 (w/w) PPE/SAN blends, the predicted threshold values are shown in Table 1. A comparison between these predicted values with the measured viscosity ratios (Fig. 1) suggests a continuous SAN phase with PPE droplets for all blend ratios. The viscosity ratio of the blend at high frequencies (simulating the high shear rates in the extruder) is around 12. According to these values, a PPE content above 70 wt.% is necessary to achieve a continuous PPE phase with dispersed SAN droplets. However, this blend recipe is not applicable for melt processing due to the high viscosity of PPE. Having the same morphologies in all blends, i.e., a SAN matrix with dispersed PPE droplets, facilitates the direct comparison of the micromechanical properties for all blend ratios. Fig. 2 shows TEM micrographs of the neat and SBM compatibilized PPE/SAN blends at different blend ratios. The brighter matrix phase is SAN, and the PPE phase appears as the dark phase. Using SBM triblock terpolymers, the blend morphologies become more homogeneous, and the PPE phase forms droplets instead of a semi-continuous structure which is consistent with prior studies [22]. It is expected that the PPE droplet sizes decrease after compatibilization, as the interfacial energy between the blend components decreases. In case of the 50/50 blend (Fig. 2a and b), the droplets have a much more homogenous shape after compatibilization, however, their sizes are not significantly reduced. This may be due to SBM micelle formation within the PPE phase resulting from a slight preferential interaction of the PS block with PPE

R. Bahrami et al. / Polymer 80 (2015) 52e63

55

Fig. 1. a) Absolute viscosity values of the neat PPE and SAN (blend components) at 260  C. b) Viscosity ratio, P, in dependence of frequency. The dashed line at P ¼ 10.7 represents the lowest viscosity ratio of PPE and SAN.

Table 1 Viscosity ratio (P ¼

50/50 60/40 70/30

hPPE hSAN )

of the blends calculated by different models. Chen's model [38]

Utracki's model [39]

0.5 2.1 9.2

1 2.6 8.6

(FloryeHuggins segmentesegment interaction parameter cPS/

PPE ¼ 0.044) compared to PMMA/SAN (cPS/PPE ¼ 0.008) [40]. This

has been observed before [36], but also the blend viscosity and shear forces during compounding play a significant role in determining the final blend morphology and formation of micelles. The TEM micrograph in Fig. 3a, which was acquired at a higher magnification, clearly shows SBM micelles in the PPE phase as well as SBM triblock terpolymer chains located at the PPE/SAN interface. The latter appear as black dots at the PPE/SAN interface, due to the selectively stained PB block of the SBM. The core of the micelles consists of PMMA and PB, and the PS shell points to the PPE. The different possible micelle formation mechanisms were thoroughly discussed previously [36]. At high SAN contents the blend viscosity is comparably low, and the initially formed smaller PPE droplets can coalesce and from larger PPE droplets. Consequently, there is excess SBM that cannot assemble at the interface and thus forms micelles. Additionally, SBM located at the interface of smaller PPE droplets may be trapped inside larger PPE domains as a micelle after coalescence. This extensive micelle formation reduces the compatibilizer efficiency, as the amount of effective SBM triblock terpolymer chains at the interface is reduced (the SBM micelles can be counted as ineffective compatibilizer). Preventing coalescence in blends of low viscosities by either higher shear forces, or more efficient compatibilizer with higher surface activity like Janus particles [22], would lead to smaller PPE droplets without SBM micelle formation. It is shown that the amount of micelles in the PPE phase decreases as the PPE content and the viscosity of the system increases (from Fig. 3a and b). In the 60/40 blends (Figs. 2c, d and 3b), the PPE domains are more homogeneous in their appearance with less micelle formation after compatibilization as compared to the 50/50 blend. The higher viscosity of the 60/40 blend (due to its higher PPE content) reduces SBM mobility and droplet coalescence rate during extrusion and less micelles are trapped within the PPE domains. In case of the 70/30 blends (Figs. 2e, f and 3c), the viscosity is even higher and micelles in the PPE phase are almost absent and the

triblock terpolymer chains are exclusively located at the interface between the blend phases. However, due to the high PPE fraction, the SBM amount is probably not high enough to sufficiently cover all of the PPE domains. Surprisingly, the PPE domain size decreases with increasing PPE content (in both neat and compatibilized blends). This may be explained by the addition of the much higher viscous PPE to the system. Higher internal shear forces are produced during the extrusion process, leading to higher break up rates and, hence, smaller PPE droplet sizes. 3.2. Rheological characterization In order to confirm the above-mentioned theories, shear rheological measurements of polymer blends were performed. Rheology is an important tool to compare the interfacial adhesion between the phases after compatibilization. The viscosity of a blend system depends on the viscosities of the components and their weight fractions, as well as the behavior of the interface between them. If the interfacial adhesion is strong, the stress can be transferred from one phase to the other upon applying shear forces. Hence, the higher viscous phase (which is attached to the lower viscous phase) hinders its flowability and increases the viscosity of the system. In our case, the more viscous PPE phase hinders the deformation of the SAN matrix and, thus, increases the viscosity of the blends. Fig. 4 shows the shear viscosity of the neat and compatibilized blend systems. As expected, increasing the PPE amount from 50 to 70 wt.% leads to an increase in viscosity of the neat blends (without compatibilizer). In addition, after compatibilization with SBM, an increase in the viscosity relative to the neat blends is also observed, indicating the presence of the triblock terpolymer chains at the interface and better stress transfer between the phases. The increase in the viscosity after compatibilization is more pronounced in the 60/40 and 70/30 blends at lower frequencies. This is due to the decrease in PPE droplet size, which results in a higher amount of PPE droplets. Consequently, there is a considerable increase in interfacial area, which strongly influences the viscosity of the blends especially at low frequencies. 3.3. Dynamic mechanical analysis (DMA) Dynamic mechanical analysis is an efficient way to investigate

56

R. Bahrami et al. / Polymer 80 (2015) 52e63

Fig. 2. TEM micrographs of the neat and SBM (10 wt.%) compatibilized PPE/SAN blends after extrusion: a) neat 50/50, b) SBM compatibilized 50/50, c) neat 60/40, d) SBM compatibilized 60/40, e) neat 70/30, f) SBM compatibilized 70/30.

Fig. 3. TEM micrographs of the SBM compatibilized PPE/SAN blends with: a) 50/50, b) 60/40, and c) 70/30 blend ratio. SBM micelles in the PPE phase are marked with orange arrows. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

R. Bahrami et al. / Polymer 80 (2015) 52e63

57

Fig. 4. Absolute shear viscosity of the a) neat and b) SBM compatibilized PPE/SAN blends.

the mechanical properties, namely the complex modulus, of polymer blends in a wide temperature range. Fig. 5 shows the complex modulus of neat and compatibilized blends in dependence of temperature. In the DMA curves of the neat blends, two sharp and distinct steps are visible indicating the glass transition temperatures (Tg) of SAN and PPE, respectively. The discontinuous distribution of the elastic PB phase around the PPE domains in the compatibilized blends (raspberry morphology) prevents reduction of the modulus and stiffness of the material after toughness modification [12]. Therefore, below the glass transition temperature of SAN, the complex modulus of the neat and compatibilized blends remains almost constant. Right above the glass transition of SAN, the complex modulus shows a pronounced drop as the temperature increases, which is in accordance with the morphological results having SAN as the continuous phase. The second reduction step in the complex modulus curves correlates to the Tg of the PPE phase. After compatibilization, a shift in Tg of the PPE to lower temperatures is observed for all blend compositions. This may be explained by the miscibility of PS/PPE and the high difference between the Tg of both polymers (ca. 103 and 210  C, respectively). In addition, SBM micelles located in the PPE domains would further contribute to the reduction of its Tg. In contrast, SAN and PMMA have a similar Tg (ca. 110 and 100  C, respectively) and, hence, no significant change in Tg is expected. The trend of the DMA trace above the Tg of the SAN matrix is a direct indication of the blend morphology. In case of pseudo cocontinuity (quasi co-continuous structure) in the neat 50/50 blend, a pronounced second plateau is visible around 250 MPa. After compatibilization with SBM, the morphology changes into droplets dispersed in the SAN phase (which is only 50 wt.-% of the specimen), hence a significant drop and continuous decrease of the modulus is observed until the Tg of the PPE is reached. Therefore, the difference between the complex modulus of the neat and SBM compatibilized blends at 50/50 weight ratio is very large between the two Tg values (Fig. 5a). In case of the neat 60/40 blend the quasi co-continuous morphology [22] is preserved and visible in the DMA curve as plateau above the Tg of SAN at around 800 MPa. After compatibilization, the TEM (and also SEM images in the next section) show a morphology consisting of PPE droplets dispersed in the SAN matrix. However, the DMA curve of the compatibilized blend still shows a plateau at around 600 MPa instead of the expected fast decrease like in the compatibilized 50/50 blend. This phenomenon can be

explained by the very large difference in the Tg values of the blend components, as well as the bonding at the interface via SBM. After softening of the SAN matrix, the PPE droplets, which represent the main fraction in the blend, are still in their glassy state and are able to hold the specimen together. In addition, as the droplet size decreases, the interfacial area between the phases increases. Hence, there is a strong linkage at the interface attaching the small glassy PPE domains to the SAN matrix and compared to the neat blend, only a small decrease in the plateau after compatibilization is observed (Fig. 5b). In the neat 70/30 blend, even though the morphology shows no co-continuity, there is a plateau visible above the Tg of SAN at a large value of 1000 MPa. The reason may be the high fraction of the PPE phase, resulting in densely packed small PPE particles within the SAN matrix. The hard glassy PPE domains in this region have an effect comparable to that of fillers in a highly filled polymer composite. The PPE droplets form a network providing structural viscosity to the blend, preventing its collapse (since the SAN phase has already softened) and holding the specimen together. After compatibilization, the same phenomenon as described above is observed. In comparison to the other blend ratios, the difference between the neat and compatibilized 70/30 blends is rather small (both have values around 1000 MPa), which is due to the large effect of the higher PPE fraction (Fig. 5c). The increase in the plateau values (from 600 to 1000 MPa) with increasing PPE content confirms that the PPE droplets act as glassy fillers in the SAN matrix. In conclusion, the high difference between the Tg of the blend components, sizes of the droplets in the matrix (which influences the interfacial area), and the quality/strength of the interface play important roles in determining the properties of the blend materials in DMA analysis. Here, the quality of the interface has been deliberately kept constant (always the same type of SBM triblock terpolymer) and the effect of the droplet size has been investigated. Smaller, densely packed droplets seem to build a strong network holding the matrix together, and after compatibilization they effectively attach to the matrix polymer via a stronger interface mediated by the SBM chains. Therefore, finer morphologies can improve the stability and modulus of the blend at higher temperatures. In order to further confirm the quasi co-continuity of the blend structures before compatibilization and the change in morphology on a macro scale after compatibilization, Payne tests were performed on all blends. The Payne effect [41] is mostly defined for filled rubber systems in which the filler particles form clusters and

58

R. Bahrami et al. / Polymer 80 (2015) 52e63

certain point during the strain sweep [42]. This analogy can be applied to our blend systems with their unique structures. In case of the neat blends, the quasi co-continuous structure cannot be destroyed easily via strain, as it is partially connected through the elongated PPE domains. However, after compatibilization, the dispersed PPE particles resemble fillers that form a network in the blend system. This network is destroyed earlier upon increasing strain (Payne strain) compared to the neat blends. This behavior is shown in Fig. 6, where Payne tests of each blend system performed at 150  C are shown. This temperature is chosen to be within the second plateau above the Tg of SAN. The sharp decrease in modulus, which is usually expected in a Payne test, is not visible here, since the thermoplastic blends used are rather brittle compared to the elastomers. Before entering the higher deformations, all samples broke and the test was stopped. Fig. 6a shows the test results for 50/ 50 neat and compatibilized blends. The decrease in the modulus starts for the neat blend at a Payne strain value of around 20%, whereas the network of PPE particles in the compatibilized blend starts breaking up at values around 10%. The quasi co-continuous structure of the neat blend is responsible for the higher Payne strain value. A similar trend is visible for the 60/40 blends (Fig. 6b), where the decrease in modulus appears at lower strains for the SBM compatibilized blends. Here the difference between the moduli of the neat and compatibilized blends is smaller as discussed previously in Fig. 5. In case of the 70/30 blends, Fig. 6c, there is also a difference visible between the Payne strain of the neat and compatibilized blends. This could be related to the compatibilization, as Wang et al. [43] reported that compatibilization with block copolymers resulted in larger reductions of the storage modulus in the Payne test. 3.4. Fracture toughness characterizations Fracture toughness measurements can provide important information about the toughness of materials. In addition, by investigating the fractured surfaces different toughening mechanisms for each blend component can be identified. In order to be able to correlate the micro/nano-structure to the measured macro-mechanical properties of polymer blends with non-trivial morphologies, it is important to understand the micromechanical deformation mechanisms taking place in each phase. Any polymer material can withstand crack tip stresses up to a critical value of stress intensity (critical stress intensity factor, KIC). Beyond this point, the crack propagates rapidly in the sample. The critical stress intensity factor is a measure of the material toughness. Fig. 7 shows the critical stress intensity factor (KIC) value of the neat and compatibilized blends as a function of their PPE contents.

Fig. 5. Temperature-dependent complex modulus of the neat and SBM compatibilized PPE/SAN blends with a) 50/50, b) 60/40, c) 70/30 (w/w) ratios.

interact in a filler network. In such systems, the storage modulus depends on the amplitude of the applied strain. After applying high deformations (increasing the strain) the filler network can be disturbed and destroyed, and a decrease in the modulus occurs at a

3.4.1. Neat blends In case of the neat blends, it is expected that increasing the PPE content (as the more ductile phase) should result in higher toughness values. This is valid when moving from the 50/50 to the 60/40 blend. However, a further increase of the PPE content (as in the case of the 70/30 blend) results in a significant reduction of toughness as compared to the 60/40 blend. In order to understand and explain this behavior of the neat blends, morphologies of the fractured surfaces (after KIC experiments) were investigated (Fig. 8). The larger PPE domains in the neat 50/50 and 60/40 blends (Fig. 8a and c) have formed some fiber-like structures inside the SAN matrix, probably due to their quasi co-continuous structure before compatibilization, as explained in detail previously [22]. It is known that elongated particles, which are parallel or perpendicular to the crack growth direction, are capable of affecting the crack more pronounced than spherical particles [44]. These elongated particles can plastically deform during crack growth and

R. Bahrami et al. / Polymer 80 (2015) 52e63

59

elongated ductile particles, which can improve the toughness of the SAN matrix. In addition, since the interface between two immiscible polymers is weak, debonded PPE domains are visible in the fractured surface of the 50/50 (Fig. 8b) and more pronounced in the 60/40 (Fig. 8d) blends. This debonding leads to the pull out of PPE domains from the matrix, therefore contributing to plastification in the material as a toughening mechanism. In case of the 50/50 and 60/40 blends, both mechanisms (deformation of non-spherical PPE domains, and pull out of these domains from the matrix) consume large amounts of energy during fracture, leading to an increase in the toughness of the material. In case of the 70/30 blend, the PPE domains are more spherical and smaller in size (Fig. 8e and f), despite of their higher fraction in the blend. These spherical PPE particles are more stable compared to the elongated ones in the 50/ 50 and 60/40 blends, as they have a smaller surface area to volume ratio. In other words, the size of the dispersed particles has a dramatic effect on their deformation behavior in blend systems [27,46], and finer dispersions are more stable and more resistant to deformations due to lower resultant stress concentrations [45,46]. Furthermore, in the 70/30 blend, the interparticle distance between the PPE domains is so small that the deformation of the SAN matrix is prevented. Therefore, there are only debonded PPE domains visible due to lack of compatibilization. Hence, due to absence of the aforementioned toughening mechanisms (PPE domain plastic deformation or pull out, and SAN matrix deformations), the KIC value of the 70/30 blend decreases in comparison with blend systems with lower PPE contents.

Fig. 6. Storage modulus of the neat and SBM compatibilized PPE/SAN blends in dependence of the applied strain at 150  C (Payne test): a) 50/50, b) 60/40, c) 70/30 (w/ w) ratios.

sometimes even hinder the crack propagation. In general, if the interface is strong enough, the fiber-like phase can also promote plastic deformations of the surrounding matrix. However, this is not the case here, since the neat blends have week interfaces. The PPE domains in the neat 50/50 and 60/40 blends can be assumed as

3.4.2. Compatibilized blends After effective compatibilization it is expected that there is an enhanced stress transfer from one phase to the other resulting in higher toughness values. According to Fig. 7, in case of the 50/50 blend, the KIC values remain almost constant after compatibilization. The fractured surface in Fig. 9a shows that the addition of SBM homogenizes the morphology and produces spherical PPE domains (oppose to the random non-spherical ones in the neat blend) as seen in Figs. 2 and 3. These PPE domains, containing SBM triblock terpolymers at their interface with SAN, show other toughening mechanisms (Fig. 9a). The PPE domains are partially embedded in the SAN matrix (due to partial SBM coverage) and are, to some extent, capable of transferring the stress to the more brittle SAN at the points where SBM connects the two phases. At a higher magnification (Fig. 9b), one can see the partial coverage of the PPE domains with SBM (white points representing SBM micelles). When the SBM connects the phases, some matrix deformations are visible in forms of crazes. However, the deformations of the spherical PPE domains in the compatibilized 50/50 blend are not as effective as that of the anisotropic ones in the neat blend [44]. In addition, due to the different morphology (larger spherical PPE domains instead of smaller anisotropic ones), the interfacial area between PPE and SAN is smaller. Thus, the new toughening mechanisms provided by the SBM compatibilized interface are not sufficient to result in better KIC values. In addition, the stronger adhesion between the phases after compatibilization prevents pull out of the PPE droplets from the SAN matrix that would consume a lot of energy and contribute to the toughening of the system as well. The 60/40 blend shows an increase in the KIC value after compatibilization. According to Fig. 9c and d, the stress is transferred from the PPE phase to the SAN phase more effectively compared to the 50/50 blend, and some shear bands are visible in the SAN matrix (white branched strips in the SEM images marked with arrows). Due to the location of SBM triblock terpolymer chains at the interface, the stress can be easily transferred from one phase to the other leading to more shear bands in the SAN matrix. In addition,

Fig. 7. a) Critical stress intensity factor (KIC) of neat and SBM compatibilized blends and b) schematic illustration of the compact specimen with force direction (right).

Fig. 8. SEM images of the fractured surface of the neat a, b) 50/50; c, d) 60/40 and e, f) 70/30 blends after KIC test.

R. Bahrami et al. / Polymer 80 (2015) 52e63

61

Fig. 9. SEM images of the fractured surface of compatibilized a, b) 50/50; c, d) 60/40 (orange arrows pointing to plastification of SAN phase via shear yielding) and e, f) 70/30 blends after KIC test. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

the smaller sizes of the PPE domains after compatibilization play an important role during plastification, as the size of the dispersed phase determines the main deformation mechanism in the material. Due to the higher weight fraction and the smaller PPE droplet size in the sample, the overall interfacial area between PPE and SAN is significantly larger as compared to the 50/50 blend. Soft PPE domains promote crazing in the less ductile SAN matrix and craze formation happens at the early stage of deformation. At higher magnification, the PPE domains with SBM (white patches) and SBM fibrils that attach the phases to each other are visible. However, the increase of toughness after compatibilization in the 60/40 blend is not so profound, which again may be explained by the change in the geometries of PPE domains in the compatibilized blend (spherical) compared to the neat blend (anisotropic elongated structure), which can effect the plastic deformation mechanisms of

the PPE particles. In case of the 70/30 blend, the increase in toughness after compatibilization is much more significant. This is expected since entire localization of the SBM chains at the interface causes a more effective stress transfer to the SAN matrix and the formation of small PPE droplets is observed. Shear yielding deformations are visible only in some regions of the SAN matrix (Fig. 9e and f), probably due to the fact that small interparticle distances between PPE particles hinder SAN matrix deformations. These shear bands can form a shear yielding network all over the sample, which can result in PPE droplets cavitation from the SAN matrix and contribute to the toughening of the polymers via debonding mechanisms. Hence, the increase in toughness of the 70/30 blend is much larger compared to the other blends, due to multiple toughening mechanisms.

62

R. Bahrami et al. / Polymer 80 (2015) 52e63

It has to be mentioned that conclusions on the effect of domain size after compatibilization on the toughness of the polymer blends can only be drawn if the fractured surfaces show similar toughening mechanisms. In other cases, fractured surfaces need to be carefully investigated as each material can act differently. 4. Conclusion Immiscible PPE/SAN blends at blend ratios close to the phase inversion region were successfully compatibilized with a SBM triblock terpolymer. The morphologies of the blends, before and after compatibilization, with a special focus on the interface, were studied in detail. Micelle formation is observed in the blend systems with lower viscosities, whereas higher viscous blends show smaller PPE droplets in the blend morphologies with SBM triblock terpolymers exclusively located at the interface. DMA analysis revealed that the blend systems show a plateau even above the Tg of SAN. This is explained by the strong difference between the Tg values of the components, the high viscosity mismatch of the system, and the strong linkage between the PPE and SAN phases at the interface mediated by the SBM triblock terpolymer. The dispersed PPE particles in their glassy state build a network structure that hold the softened SAN matrix together and can be analogically compared to a highly filled composite system. The quasi cocontinuous structure of the neat blends and its transformation to fully dispersed PPE droplets in SBM compatibilized blends was proven by Payne tests. Characterization of the fractured surfaces showed complex and multiple deformation mechanisms strongly depending on the size of the dispersed PPE droplets. It is concluded that smaller PPE domain sizes after compatibilization only improve the toughness of the material when the main deformation mechanisms stay constant for both the neat and compatibilized blend. Among the studied blend compositions, the compatibilized 60/40 (PPE/SAN w/w) blend combines excellent toughness with good processability (low viscosity) and, thus, may be suited with regard to possible industrial applications. Acknowledgments The German Research Foundation (DFG) has supported this work within the AL 474/21-1 and Mu 896/39-1 grants. Authors would like to thank Anneliese Lang and Melanie Müller for SEM and TEM measurements. The authors acknowledge BASF for donating SAN and Mitsubishi for PPE polymer. References [1] D.R. Paul, C.B. Bucknall, Polymer Blends, Wiley, New York, 2000. [2] L.A. Utracki, Commercial Polymer Blends, first ed., Chapman and Hall, London, 1998. [3] B. Du, U.A. Handge, S. Majeed, V. Abetz, Localization of functionalized MWCNT in SAN/PPE blends and their influence on rheological properties, Polymer 53 (24) (2012) 5491e5501. [4] D.J. Lohse, S. Datta, E.N. Kresge, Graft copolymer compatibilizers for blends of polypropylene and ethylene-propylene copolymers, Macromolecules 24 (2) (1991) 561e566. [5] C. Koning, M. van Duin, C. Pagnoulle, R. Jerome, Strategies for compatibilization of polymer blends, Prog. Polym. Sci. 23 (97) (1998) 707e757. [6] B.J. Kim, J. Bang, C.J. Hawker, J.J. Chiu, D.J. Pine, S.G. Jang, S.-M. Yang, E.J. Kramer, Creating surfactant nanoparticles for block copolymer composites through surface chemistry, Langmuir 23 (25) (2007) 12693e12703. rard, N. Mignard, Reactive compatibiliza[7] F. Teyssandier, P. Cassagnau, J.F. Ge tion of PA12/plasticized starch blends: towards improved mechanical properties, Eur. Polym. J. 47 (12) (2011) 2361e2371. [8] H. Pernot, M. Baumert, F. Court, L. Leibler, Design and properties of cocontinuous nanostructured polymers by reactive blending, Nat. Mater. 1 (54) (2002) 54e58. [9] C.L. DeLeo, S.S. Velankar, Morphology and rheology of compatibilized polymer blends: diblock compatibilizers vs crosslinked reactive compatibilizers, J. Rheol. 52 (6) (2008) 1385.

[10] J. Kressler, H.W. Kammer, Miscibility behaviour of poly(2,6- dimethylphenylene oxide) and poly(styrene-co-acrylonitrile), Acta Polym. 38 (11) (1987) 600e602. [11] Y.S. Soh, Compatibility in blends of poly (2,6-dimethyl phenylene ether)/ styrene-acrylonitrile copolymer, J. Appl. Polym. Sci. 44 (2) (1992) 371e375. [12] T. Kirschnick, A. Gottschalk, H. Ott, V. Abetz, J. Puskas, V. Altst€ adt, Melt processed blends of poly(styrene-co-acrylonitrile) and poly(phenylene ether) compatibilized with polystyrene-b-polybutadiene-b-poly(methyl methacrylate) triblock terpolymers, Polymer 45 (16) (2004) 5653e5660. €dt, T. Kirschnick, H. Ott, R. Stadler, [13] R. Lach, W. Grellmann, R. Weidisch, V. Altsta C. Mehler, Poly (styrene-block-butadiene-block-styrene-block-butadiene) and poly (styrene-block-butadiene-block-methyl methacrylate) copolymers as compatibilizers in PPO e SAN blends. I. Morphology and fracture behavior, J. Appl. Polym. Sci. 78 (11) (2000) 2037e2045. [14] C. Auschra, R. Stadler, Polymer alloys based on poly(2,6-dimethyl-l,4phenylene ether) and poly(styrene-co-acrylonitrile) using poly(styrene-b(ethylene-co-butylene)-b-methyl metacryrate) triblock copolymers as compatibilizers, Macromolecules 26 (24) (1993) 6364e6377. [15] C.W. Macosko, P. Gue, A.K. Khandpur, A. Nakayama, P. Marechal, T. Inoue, Compatibilizers for melt blending: premade block copolymers, Macromolecules 29 (17) (1996) 5590e5598. €dt, H. Schmalz, V. Abetz, A.H.E. Müller, [16] H. Ruckd€ aschel, J.K.W. Sandler, V. Altsta Toughening of immiscible PPE/SAN blends by triblock terpolymers, Polymer 48 (9) (2007) 2700e2719. [17] Q. Zhuang, Z. Xue, Y. Yuan, Z. Han, Molecular simulation of miscibility of poly(2, 6-dimethyl-1, 4-phenylene ether)/poly(styrene-co-acrylonitrile) blend with the compatibilizer triblock terpolymer SBM, Polym. Compos. 32 (10) (2011) 1671e1680. [18] C. Auschra, R. Stadler, I.G. Voigt-Martin, Poly(styrene-b-methyl methacrylate) block copolymers as compatibilizing agents in blends of poly(styrene-coacrylonitrile) and poly(2,6-dimethyl-1,4-phenylene ether): 2. Influence of concentration and molecular weight of symmetric block copolymers, Polymer 34 (10) (1993) 2094e2110. [19] C. Auschra, R. Stadler, I.G. Voigt-Martin, Poly(styrene-b-methyl methacrylate) block copolymers as compatibilizing agents in blends of poly(styrene-coacrylonitrile) and poly(2,6-dimethyl-1,4-phenylene ether): 1. Location of block copolymers in ternary blends d compatibilization versus micelle formation, Polymer 34 (10) (1993) 2081e2093. [20] B. Imre, K. Renner, B. Puk anszky, Interactions, structure and properties in poly (lactic acid)/thermoplastic polymer blends, Express Polym. Lett. 8 (1) (2014) 2e14. [21] A. Walther, K. Matussek, A.H.E. Müller, Engineering nanostructured polymer blends with controlled nanoparticle location using Janus particles, ACS Nano 2 (6) (2008) 1167e1178. €bling, A.H. Gro €schel, H. Schmalz, A.H.E. Müller, V. Altsta €dt, [22] R. Bahrami, T.I. Lo The impact of Janus nanoparticles on the compatibilization of immiscible polymer blends under technologically relevant conditions, ACS Nano 8 (10) (2014) 10048e10056. [23] Y. Cao, J. Zhang, J. Feng, P. Wu, Compatibilization of immiscible polymer blends using graphene oxide sheets, ACS Nano 5 (7) (2011) 5920e5927. [24] X. Cai, B. Li, Y. Pan, G. Wu, Morphology evolution of immiscible polymer blends as directed by nanoparticle self-agglomeration, Polymer 53 (1) (2012) 259e266. €schel, T.I. Lo € bling, M. Drechsler, S. Ehlert, S. Fo € rster, [25] T. Gegenhuber, A.H. Gro H. Schmalz, Noncovalent grafting of carbon nanotubes with triblock terpolymers: Toward patchy 1D hybrids, Macromolecules 48 (6) (2015) 1767e1776. €la €, Compatibilization of poly[26] U. Hippi, J. Mattila, M. Korhonen, J. Seppa ethylene/aluminum hydroxide (PE/ATH) and polyethylene/magnesium hydroxide (PE/MH) composites with functionalized polyethylenes, Polymer 44 (4) (2003) 1193e1201. [27] J.C. Angola, Y. Fujita, T. Sakai, T. Inoue, Compatibilizer-aided toughening in polymer blends consisting of brittle polymer particles dispersed in a ductile polymer matrix, J. Polym. Sci. Part B Polym. Phys. 26 (4) (1988) 807e816. [28] B. Chen, X. Li, S. Xu, T. Tang, B. Zhou, B. Huang, Compatibilization effects of block copolymers in high density polyethylene/syndiotactic polystyrene blends, Polymer 43 (3) (2002) 953e961. [29] U.A. Handge, C. Sailer, H. Steininger, M. Weber, S. Scholtyssek, V. Seydewitz, G.H. Michler, Micromechanical processes and failure phenomena in reactively compatibilized blends of polyamide 6 and styrenic polymers. II. Polyamide 6/ styrene-acrylonitrile copolymer blends, J. Appl. Polym. Sci. 115 (5) (2010) 2529e2539. [30] H.R. Brown, U. Krappe, R. Stadler, Effect of ABC triblock copolymers with an elastomeric midblock on the adhesion between immiscible polymers, Macromolecules 29 (95) (1996) 6582e6588. [31] G.H. Michler, Microstructural construction of polymers with improved mechanical properties, Polym. Adv. Technol. 9 (10) (1998) 812e822. [32] W. Tiejun, K. Kishimoto, M. Notomi, Effect of triaxial stress constraint on the deformation and fracture of polymers, Acta Mech. Sin. 18 (5) (2002) 480e493. €tz, F. Fischer, G.T. Lim, [33] U.A. Handge, A. Galeski, S.C. Kim, D.J. Dijkstra, C. Go V. Altst€ adt, C. Gabriel, M. Weber, H. Steininger, Melt processing, mechanical, and fatigue crack propagation properties of reactively compatibilized blends of polyamide 6 and acrylonitrile-butadiene-styrene copolymer, J. Appl. Polym. Sci. 124 (1) (2012) 740e754. [34] M.E. Fowler, J.W. Barlow, D.R. Paul, Kinetics of adhesion development at

R. Bahrami et al. / Polymer 80 (2015) 52e63 PMMA-SAN interfaces, Polymer 28 (12) (1987) 2145e2150. [35] M. Suess, J. Kressler, H.W. Kammer, The miscibility window of poly(methylmethacrylate)/poly(styrene-co-acrylonitrile) blends, Polymer 28 (6) (1987) 957e960. €schel, J.K.W. Sandler, V. Altsta €dt, C. Rettig, H. Schmalz, V. Abetz, [36] H. Ruckda A.H.E. Müller, Compatibilisation of PPE/SAN blends by triblock terpolymers: Correlation between block terpolymer composition, morphology and properties, Polymer 47 (8) (2006) 2772e2790. [37] J.J.M. Janssen, A. Boon, W.G.M. Agterof, Influence of dynamic interfacial properties on droplet breakup in plane hyperbolic flow, AIChE J. 43 (6) (1997) 1436e1447. [38] T.H. Chen, A.C. Su, Morphology of poly(p-phenylene sulfide)/polyethylene blends, Polymer 34 (23) (1993) 4826e4831. [39] L.A. Utracki, On the viscosity-concentration dependence of immiscible polymer blends, J. Rheol. 35 (8) (1991) 1615e1638. [40] G.D. Merfeld, A. Karim, B. Majumdar, S.K. Satija, D.R. Paul, Interfacial thickness in bilayers of poly(phenylene oxide) and styrenic copolymers, J. Polym. Sci.

63

Part B Polym. Phys. 36 (17) (1998) 3115e3125. [41] A.R. Payne, Effect of dispersion on the dynamic properties of filler-loaded rubbers, J. Appl. Polym. Sci. 9 (6) (1965) 2273e2284. [42] G. Heinrich, T.A. Vilgis, A statistical mechanical approach to the Payne effect in filled rubbers, Express Polym. Lett. 9 (3) (2015) 291e299. [43] Z. Wang, S. Li, D. Wei, J. Zhao, Mechanical properties, Payne effect, and Mullins effect of thermoplastic vulcanizates based on high-impact polystyrene and styreneebutadiene rubber compatibilized by styreneebutadieneestyrene block copolymer, J. Thermoplast. Compos. Mater. 28 (8) (2015) 1154e1172. [44] P. Lipetzky, S. Schmauder, Crack-particle interaction in two-phase composites part I: Particle shape effects, Int. J. Fract. 65 (4) (1994) 345e358. [45] W.L. Nachlis, R.P. Kambour, W.J. MacKnight, In situ polymerization of bisphenol-A-carbonate cyclic oligomers in miscible blends with a styreneacrylonitrile copolymer: Phase separation dynamics and the influence of phase dispersion on ductility, Polymer 35 (17) (1994) 3643e3657. [46] S.C. Tjong, S.A. Xu, Ternary polymer composites: PA6,6/maleated SEBS/glass beads, J. Appl. Polym. Sci. 81 (13) (2001) 3231e3237.