Complex Phase Separation in Poly(acrylonitrile−butadiene−styrene

Mar 30, 2009 - Jyotishkumar Parameswaran Pillai , Jürgen Pionteck , Rüdiger Häßler , Christophe Sinturel , Viju Susan Mathew , and Sabu Thomas. In...
0 downloads 0 Views 4MB Size
5418

J. Phys. Chem. B 2009, 113, 5418–5430

Complex Phase Separation in Poly(acrylonitrile-butadiene-styrene)-Modified Epoxy/ 4,4′-Diaminodiphenyl Sulfone Blends: Generation of New Micro- and Nanosubstructures P. Jyotishkumar,† Joachim Koetz,‡ Brigitte Tiersch,‡ Veronika Strehmel,‡ Ceren Özdilek,§ Paula Moldenaers,§ Rudiger Hässler,| and Sabu Thomas*,† School of Chemical Sciences, Mahatma Gandhi UniVersity, Priyadarshini Hills, Kottayam, Kerala 686560, India, Institute of Chemistry, UniVersity of Potsdam, Karl-Liebknecht-Strasse 24-25, D-14476 Potsdam-Golm, Germany, Department of Chemical Engineering, Catholic UniVersity of LeuVen, de Croylaan 46, B-3001 LeuVen, Belgium, and Leibniz-Institute for Polymer Research Dresden, Hohe Strasse 6, 01069 Dresden, Germany ReceiVed: March 25, 2008; ReVised Manuscript ReceiVed: January 31, 2009

The epoxy system containing diglycidyl ether of bisphenol A and 4,4′-diaminodiphenyl sulfone is modified with poly(acrylonitrile-butadiene-styrene) (ABS) to explore the effects of the ABS content on the phase morphology, mechanism of phase separation, and viscoelastic properties. The amount of ABS in the blends was 5, 10, 15, and 20 parts per hundred of epoxy resin (phr). Transmission electron microscopy (TEM) and scanning electron microscopy (SEM) were employed to investigate the final morphology of ABS-modified epoxy blends. Scanning electron microscopic studies of 15 phr ABS-modified epoxy blends reveal a bicontinuous structure in which both epoxy and ABS are continuous, with substructures of the ABS phase dispersed in the continuous epoxy phase and substructures of the epoxy phase dispersed in the continuous ABS phase. TEM micrographs of 15 phr ABS-modified epoxy blends confirm the results observed by SEM. TEM micrographs reveal the existence of nanosubstructures of ABS in 20 phr ABS-modified epoxy blends. To the best of our knowledge, to date, nanosubstructures have never been reported in any epoxy/thermoplastic blends. The influence of the concentration of the thermoplastic on the generated morphology as analyzed by SEM and TEM was explained in detail. The evolution and mechanism of phase separation was investigated in detail by optical microscopy (OM) and small-angle laser light scattering (SALLS). At concentrations lower than 10 phr the system phase separates through nucleation and growth (NG). However, at higher concentrations, 15 and 20 phr, the blends phase separate through both NG and spinodal decomposition mechanisms. On the basis of OM and SALLS, we conclude that the phenomenon of complex substructure formation in dynamic asymmetric blends is due to the combined effect of hydrodynamics and viscoelasticity. Additionally, dynamic mechanical analysis was carried out to evaluate the viscoelastic behavior of the cross-linked epoxy/ABS blends. Finally, apparent weight fractions of epoxy and ABS components in epoxy- and ABS-rich phases were evaluated from Tg analysis. Introduction Thermosetting resins are often used in engineering applications because of their high modulus and easy processability. Among the various thermosetting resins, epoxy resins (ERs) are widely used as matrixes for high-performance composites in the aerospace and automotive industry due to their high tensile strength and modulus, easy processability, good thermal and chemical resistance, and low shrinkage on curing.1 The structure of epoxy networks is influenced by several factors, such as the molar ratio of epoxy groups to amino hydrogen, the cross-linking temperature, and the content of accelerators.2-6 However, cured epoxy resins have low toughness and poor crack resistance that prevent even their wider applications. The most common method to increase the fracture toughness is to incorporate rubber into epoxy resin. Rubber toughening is effective in diglycidyl ether of bisphenol A (DGEBA) resins. However, it is ineffective in * To whom correspondence should be addressed. Phone: +91-4812730003, +91-481-27310376. Fax: +91-481-2731002. E-mail: sabut552001@ yahoo.com. † Mahatma Gandhi University. ‡ University of Potsdam. § Catholic University of Leuven. | Leibniz-Institute for Polymer Research Dresden.

highly cross-linked epoxy resins such as (tetraglycidyldiaminodiphenyl)methane (TGDDM) due to the high cross-link density of TGDDM by which shear deformation of the matrix was prevented and, apart from that, incorporation of rubber into epoxy, which will affect the thermal stability and mechanical properties of the cross-linked system. To improve the toughness of epoxy networks without sacrificing mechanical and thermal properties, engineering thermoplastics are preferred over elastomers. Thermoplastics such as poly(ether sulfone)s (PESs),7 poly(etherimide)s(PEIs),8-10 poly(acrylonitrile-butadiene-styrene) (ABS),11-13 etc. are widely used as toughening agents. Studies clearly indicate that the morphologies of these systems have a direct influence on the properties of the modified epoxy resins. Phase separation may occur by spinodal decomposition (SD) or by nucleation and growth (NG) depending on the composition and curing temperature. SD takes place very close to the critical composition and is initiated by small concentration fluctuations, which result in a bicontinuous morphology. NG usually occurs at off critical compositions, and here large concentration fluctuations are necessary for phase separation, normally dispersed phase morphology was obtained by NG. According to Flory-Huggins mean field theory, as the average molecular

10.1021/jp8094566 CCC: $40.75  2009 American Chemical Society Published on Web 03/30/2009

Phase Separation in ABS-Modified Epoxy/DDS Blends

J. Phys. Chem. B, Vol. 113, No. 16, 2009 5419

Figure 1. Chemical structures of epoxy resin, DDS, and ABS.

weight of epoxy resin is reached, the epoxy/thermoplastic system will no longer be homogeneous. Different phase morphologies can be obtained, depending on the composition, thermodynamics, kinetics of phase separation, and chemical reaction in the binary mixture. These factors are strongly influenced by the curing condition and molecular weight of the toughener.14 Tanaka15-18 reported that phase separation was caused by asymmetric molecular dynamics. The origin of dynamic asymmetry is due to a glass transition temperature (Tg) difference between the components in polymer blends or due to the size difference in component molecules of the mixture. Viscoelastic phase separation has been known since 1980. Tanaka and Araki19,20 proposed that the quick hydrodynamic reduction of the interfacial area might spontaneously destabilize the phase separated macroscopic domains and induce viscoelastic phase separation. Zhou et al.21 recently reported volume shrinkage, phase inversion, and final complex phase morphologies are caused by diffusional asymmetry of the consistent molecules. It was reported that complex phase separation may take place due to unusual fast growth of domains in the matrix and chemical reaction between the binary mixtures.22,23 Gan et al.24 proved experimentally that the coarsening process of epoxy droplets and the final morphologies obtained in PEI/epoxy systems are affected by viscoelastic behavior. The viscoelastic phase separation process in epoxy-based blends has not been analyzed in detail. Very recently, the viscoelastic phase separation and the consequent volume shrinkage in epoxy/SAN blends have been reported from our laboratory.25 The prime aim of the present work is to investigate the phase morphology and complex phase separation process of ABSmodified epoxy/4,4′-diaminodiphenyl sulfone (DDS) system by scanning electron microscopy (SEM), transmission electron microscopy (TEM), optical microscopy (OM), small-angle laser light scattering (SALLS), and dynamic mechanical analysis (DMA). To the best of our knowledge, the viscoelastic phase separation and complex phase morphologies of the epoxy/ABS system have not been analyzed yet. SALLS and OM confirm the NG mechanism for the phase separation of 5 and 10 phr ABS-modified epoxy blends and a combination of NG and SD mechanisms for the phase separation of 15 and 20 phr epoxy blends. Finally, attempts have been made to explain the formation of micro- and nanosubstructures in detail. Experimental Section Materials Used. The matrix material used in the experiments consists of DGEBA (Lapox L-12, Atul Ltd., India) and DDS

(Lapox K-10, AtuI Ltd., India). The epoxy content in Lapox L-12 varies between 5.25 and 5.40 equiv/kg. Lapox L-12 has a viscosity of about 1.15-1.2 g/cm3. DDS, a white powder with a melting point of 178 °C, has a pot life of about 75-115 min at 25 °C. The toughener ABS (Poly lac PA-757K) was manufactured by Chi Mei Corp., Taiwan. ABS used is a commercially available thermoplastic polymer with 25 wt % acrylonitrile (AN), 5 wt % polybutadiene (PB), and 70 wt % polystyrene (PS). The chemical composition of ABS was determined from CHN analysis and FTIR spectroscopy. The chemical structures of DGEBA, DDS, and ABS are given in Figure 1. Preparation of Blends. Blends of epoxy resin/ABS containing 5, 10, 15, and 20 phr ABS were prepared using the melt mixing technique. It was not possible to add more quantities of ABS (above 20 phr) due to the high viscosity of the system. ABS was mixed with epoxy resin at 180 °C under constant stirring. After proper mixing, DDS was added to the epoxy/ ABS mixture with an epoxy:amine ratio of 1:1. The solution was evacuated if necessary and transferred to the open mold. The blends were cured in the air oven at 180 °C for 3 h and then postcured at 200 °C for a further 2 h. The blends were then allowed to cool slowly to room temperature. Characterization. Scanning Electron Microscopy. Each sample was cooled and fractured in liquid nitrogen at atmospheric pressure and etched with CHCl3 for 24 h to remove the thermoplastic phase. After that, the specimens were dried at 50 °C in a vacuum oven to remove the solvent and sputtered with gold in the SCD 050 sputter coater. Finally, the sputtered samples were analyzed by using a Hitachi S-4800 highresolution scanning electron microscope. Transmission Electron Microscopy. Due to the lower penetration power of electrons, it was necessary to mount objects for examination in the electron microscope on very thin films. Therefore, the samples were cut by using an ultramicrotome (Ultracut E from Reichert-Jung) into 50-80 nm thick ultrathin films. After treatment in the vapor of OsO4, the stained samples were examined in the EM 902 transmission electron microscope (Zeiss, Germany) with an accelerating voltage of 80 kV. Optical Microscopy. A few milligrams of the epoxy/ABS system placed between two glass slides was viewed through a Leitz Laborlux 12 Pols optical microscope, while the sample was being cured in a Linkam CSS450 shearing cell. The Linkam cell was used only for heating; no shear was applied. Digital

5420 J. Phys. Chem. B, Vol. 113, No. 16, 2009

Jyotishkumar et al.

Figure 2. SEM micrographs of cryogenically fractured surfaces of (a) 100/0, (b) 95/5, (c) 90/10, (d, e) 85/15, and (f, g) 80/20 cross-linked ER/ABS blends. The ER/ABS blends in (d) and (f) were not extracted with chloroform, while the remaining cross-linked ER/ABS blends in (a), (b), (c), (e), and (g) were extracted with chloroform for 24 h.

micrographs were taken at several times by a Hamamatsu TSU digital camera controller, C4742-95. Small-Angle Laser Light Scattering. The SALLS setup (Newport model ULM-TILT) consisted of a 5 mW He-Ne laser (λ ) 638.2nm). The sample was placed between two glass slides being cured in a Linkam cell, CSS450. The laser beam passed through the Linkam cell was examined by a Pulnix camera and was used to record the change in scattering patterns. The processing of the data was performed using software developed in the Catholic University, Leuven, Belgium. Dynamic Mechanical Analysis. The viscoelastic properties of ABS, neat epoxy, and epoxy blends were measured using a TA Instruments DMA 2980 dynamic mechanical thermal analyzer (DMTA). Rectangular specimens of 60 × 10 × 3 mm3 were used for the analysis. The analysis was done in single-cantilever mode at a frequency of 1 Hz, and the samples were scanned from -100 to +300 °C at a heating rate of 1 °C/min. Results and Discussion Scanning Electron Microscopy. Scanning electron microscopy is an excellent method for characterizing the morphol-

ogy of polymer blends. SEM micrographs were taken for all blends. Figure 2a shows the SEM micrograph of the neat epoxy system, which reveals a single phase. The fracture surface was typically flat and smooth, and the cracks propagated uninterrupted, indicating the typical brittle nature of the surface. In the SEM micrographs in Figure 2b,c, the spheres correspond to the ABS phase and the holes correspond to the regions from where the ABS phase has been preferentially dissolved out from the cross-linked matrix by using chloroform. Figure 2b shows the SEM micrograph of 5 phr ABS-modified epoxy. The 5 phr ABS-modified epoxy network shows a normal dispersed phase morphology in which ABS particles approximately 500 nm in diameter were uniformly dispersed in a continuous epoxy phase. Figure 2c shows the SEM micrograph of 10 phr ABS-modified epoxy. The 10 phr ABS-modified epoxy network exhibits a similar morphology with ABS particles of about 800 nm, which were dispersed in the continuous epoxy phase. When the ABS concentration is increased to 15 phr, the morphology revealed by SEM (without an extraction process) can be subdivided into two separate regions (Figure 2d), i.e., a continuous epoxy phase in which ABS domains are

Phase Separation in ABS-Modified Epoxy/DDS Blends TABLE 1: Number-Average Diameter (Dn), Weight-Average Diameter (Dw), Polydispersity Index (PDI), Interparticle Distance, and Interfacial Area per Unit Volume of Cross-Linked Epoxy/ABS Blends ABS content (phr)

Dn (µm)

Dw (µm)

PDI

interparticle distance (µm)

interfacial area per unit volume (µm-1)

5 10

0.52 0.8

0.59 0.88

1.13 1.11

0.59 0.58

0.62 0.77

dispersed and a continuous ABS phase in which epoxy particles are dispersed. The relative dark region is the continuous epoxy phase in which ABS domains are dispersed. In fact, this region shows the brittle fracture characteristic of the epoxy matrix. The brighter region exhibits ductile shear yielding, which is a characteristic feature of ABS thermoplastic; therefore, this phase is obviously the ABS phase, and the spherical particles inside the ABS phase are obviously epoxy domains. A careful examination of the SEM micrograph of the ABS-modified epoxy sample revealed that there exists good adhesion between the epoxy phase and ABS phase. From the SEM micrographs of the unextracted samples of 15 phr ABS/epoxy system, one can conclude that the thermoplastic-modified epoxy shows a complex bicontinuous phase separation with a continuous epoxy phase in which substructures of the ABS phase are dispersed and a continuous ABS phase in which substructures of the epoxy phase are dispersed. This can be further understood from the SEM micrographs of the extracted samples. Figure 2e shows the SEM micrograph of 15 phr ABSmodified epoxy after treatment with CHCl3 which means after removal of the ABS phase from the epoxy network. In contrast to Figure 2d, the brighter continuous ABS phase was absent, and the former embedded epoxy particles remain in the form of a particle cluster on the surface. Obviously, the spherical particles in the brighter continuous phase were of thermosetting nature since the CHCl3 did not dissolve the cross-linked epoxy amine matrix. Thus, the SEM results of the CHCl3-treated samples confirm the existence of bicontinuous phase separation with an epoxy continuous phase in which ABS particles are dispersed and an ABS continues phase in which epoxy particles are dispersed. A similar type of morphology was explained by Oyanguran et al.,26 Min et al.,27 Woo et al.,28 and Guo et al.29 A quite similar bicontinuous phase separation was observed for the 20 phr ABS-modified epoxy system (Figure 2f). The SEM micrograph (without any extraction) shows a continuous brighter area of the ABS phase in which epoxy domains are dispersed. On the other hand, the continuous dark region might be identified as the epoxy phase, since it reveals a brittle fracture characteristic of epoxy in the micrographs. It has to be mentioned here that the spherical particles in the continuous epoxy phase were obviously ABS domains. Figure 2g shows the micrograph at 20 phr ABS concentration after etching with CHCl3 to remove the ABS phase. The absence of the brighter continuous phase indicates that the continuous thermoplastic phase was easily etched away by the CHCl3. Obviously the spherical particles in the brighter continuous phase reveal the thermosetting nature since they were not affected by the etching procedure. Thus SEM results of the CHCl3-etched samples confirm the existence of a bicontinuous phase separated structure with a continuous epoxy phase in which ABS particles are dispersed and a continuous ABS phase in which epoxy particles are dispersed.

J. Phys. Chem. B, Vol. 113, No. 16, 2009 5421 For a more comprehensive consideration, the number-average and weight-average domain diameters and polydispersity index (PDI) of the 5 and 10 phr blends were calculated using the following equations:30

number-average diameter, Dn )

∑ nidi ∑ ni

(1)

weight-average diameter, Dw )

∑ nidi2 ∑ nidi

(2)

Dw Dn

(3)

polydispersity index, PDI )

where ni are the number of domains having diameter di. The interparticle distance and interfacial area per unit volume were calculated using the following equations:30

[( )

interparticle distance ) dTP

π 6φTP

1/3

interfacial area per unit volume )

]

-1

3φTP r

(4)

(5)

where dTP is the number-average diameter of the domains, r is the number-average radius of the domains, and 3φTP is the volume fraction of the dispersed phase. The domain diameter and the other parameters calculated from the above-mentioned equations are summarized in Table 1. From the parameters given in Table 1, one can conclude that the domain diameter increased with increasing ABS content while the polydispersity index remained constant, indicating the uniform particle size distribution. The interparticle distance decreased whereas the interfacial area per unit volume increased with increasing ABS content in the blends. All these factors favor the improvement of the fracture toughness. Transmission Electron Microscopy. In addition, TEM investigations were done to provide supporting evidence to the above results obtained by SEM. Parts a and b of Figure 3 show the TEM micrographs of cross-linked epoxy and neat ABS, respectively. In Figure 3b butadiene particles are dispersed in the SAN matrix. The TEM micrograph of the 5 phr ABSmodified epoxy sample (Figure 3c) shows a dispersed phase morphology in which dark domains, i.e., ABS particles about 500 nm in size, are uniformly dispersed in the continuous epoxy matrix. These results are in full agreement with our SEM micrographs for the 5 phr modified epoxy blends. The TEM micrograph of the 10 phr ABS-modified epoxy blend (Figure 3d) shows a normal dispersed phase morphology in which ABS particles about 800 nm in size are dispersed in the continuous epoxy phase, supporting the results of the SEM investigations. When the ABS concentration is increased to 15 phr, the TEM micrographs indicate a bicontinuous phase separated structure in which both epoxy and ABS are continuous. In the continuous epoxy phase, ABS particles are dispersed (Figure 3e), and in the continuous ABS phase, epoxy particles are dispersed (Figure 3f); it has to be mentioned here that these results are in full agreement with our SEM micrographs.

5422 J. Phys. Chem. B, Vol. 113, No. 16, 2009

Jyotishkumar et al.

Figure 3. Transmission electron micrographs of fractured surface of (a) cross-linked ER, (b) neat ABS, and (c) 95/5, (d) 90/10, (e, f) 85/15, and (g) 80/20 cross-linked ER/ABS blends .

Phase Separation in ABS-Modified Epoxy/DDS Blends

J. Phys. Chem. B, Vol. 113, No. 16, 2009 5423

TABLE 2: ABS and Epoxy Domain Sizes Measured by TEM ABS content (phr) 5 10 15 20

ABS domain size (µm) 0.48 0.79 0.48

epoxy domain size (µm)

1.5 2.05

ABS subinclusion size in epoxy domains (µm)

0.1

Figure 3g shows the TEM micrographs of the 20 phr ABSmodified epoxy blend. It consists of a continuous ABS phase, and it is to be noted that in the ABS phase large domains of epoxy particles are dispersed. A more careful examination of the inner epoxy phase reveals ABS subinclusions approximately 100 nm in diameter. This phenomenon may be due to the segregation of ABS when the conversion is increased. Because of the high viscosity of the ABS-rich phase, the segregated ABS phase does not merge with the matrix and therefore gets trapped as subinclusions in the epoxy-rich domains. The presence of substructures indicates that the process evolves out of equilibrium because of thermosetting reactions (gelation and vitrification), which block the component diffusion and the purification of the components. The approximate sizes of the ABS and epoxy domains for all the blends measured by TEM are given in Table 2. Mechanism of Phase Separation. Evolution of phase separation in epoxy/ABS blends was investigated in detail by OM and SALLS; Figure 4 displays the optical micrographs taken at different time intervals for the 10 phr ABS-modified epoxy blend cured at 180 °C. These optical micrographs reveal the evolution of morphology in 10 phr ABS-modified epoxy blends. It is important to mention that ABS is a blend of SAN and SAN-graft-PB. Since PB is immiscible with epoxy resin, it is reasonable to believe that there may be some SAN-graftPB micelles in the nanoscale dispersed in the epoxy monomers at the beginning of the reaction. These micelles can initiate the NG process. Since the PB content is very small, we cannot detect these micelles using TEM and SALLS. However, experiments are progressing in this direction using synchrotron radiation and neutron scattering. After 660 s a continuous structure of epoxy-rich phase is observed in which a few ABS rich particles are dispersed. With an increase in time, more and more ABS-rich droplets of larger size appeared in the epoxy-rich matrix, supporting the NG mechanism for the phase separation. Finally, we have a large number of ABS-rich droplets dispersed in the epoxy-rich matrix. The optical micrographs support the facts revealed by the SEM and TEM micrographs for the 10 phr ABS-modified epoxy system. The evolution of the morphology in 5 phr ABS-modified epoxy blends also follows a similar trend. The results obtained here exactly follow the classical theory in which the NG mechanism is behind the phase separation process for epoxy/ thermoplastic blends having a matrix/droplet morphology. The evolution of phase separation for the 20 phr ABSmodified epoxy blend is displayed in Figure 5. Figure 5a reveals OM micrographs taken at 420 s. Here ABS-rich droplets dispersed in an epoxy-rich matrix are observed; with an increase in time more and more ABS-rich particles are found to be dispersed in epoxy rich matrix typical for NG mechanism for phase separation. Figure 5b reveals a bicontinuous-like structure in which both the epoxy-rich phase and ABS-rich phase are continuous and was taken at 660 s; this region of phase separation is considered as the beginning of SD. After a few

Figure 4. Evolution of the morphologies in 10 phr ABS-modified epoxy blends: (a) 660 s, (b) 1020 s, and (c) 5400 s.

seconds, the bicontinuous structure ruptures; rapidly the ABSrich phase becomes the continuous phase and the epoxy-rich becomes the dispersed phase, which means a clear phase inversion took place (shown in Figure 5c), which is typical for viscoelastic phase separation. This phenomenon is due to the high molecular weight of ABS compared to epoxy resin during the initial stages of curing kinetics. Within seconds, the epoxyrich droplets grow very rapidly, resulting in a bicontinuous structure in which both the epoxy-rich phase and ABS-rich phase are continuous (Figure 5d). Figure 5e taken after 6000 s reveals

5424 J. Phys. Chem. B, Vol. 113, No. 16, 2009

Jyotishkumar et al.

Figure 5. Evolution of the morphologies in 20 phr ABS-modified epoxy blends: (a) 420 s, (b) 660 s, (c) 720 s, (d) 780 s, (e) 6000 s.

a large number of ABS-rich particles dispersed in the epoxyrich continuous phase and a large no of epoxy-rich particles dispersed in the continuous ABS-rich phase, supporting the facts observed by SEM and TEM. The 15 phr ABS-modified epoxy blend also has a similar kind of phase separation mechanism. The observed phenomenon could be subdivided into two stages: initial NG and final SD. In the initial stages of phase separation, ABS droplets develop in the epoxy matrix by NG initiated by some SAN-graft-PB micelles (immiscible in the epoxy phase) immediately after the addition of DDS. The ABSrich particles grow in size because of diffusion. The final stage (SD) could be explained in terms of viscoelastic and hydrodynamic effects. The viscosity of the ABS-rich phase is much higher than that of the epoxy-rich phase due to the high molecular weight of ABS; hence, the ABS-rich phase acts as the slow dynamic phase (elastic phase), and the epoxy/DDSrich phase acts as the fast dynamic phase (viscous phase). In the initial stages of the SD phenomenon, because of dynamic

asymmetry, the viscoelastic effect of the ABS-rich phase prevails over the hydrodynamic effect of the epoxy-rich phase and hence results in the formation of a major ABS-rich phase with epoxyrich droplets. However, after a few minutes, the hydrodynamic effect of the epoxy-rich phase comes into play, which results in the sudden growth of some of the epoxy-rich phase. The elastic effect of the ABS-rich phase cannot stop the fast growth of the epoxy-rich phase. During the fast growth of the epoxy phase there may be some miscible ABS chains that also separate out of the ABS-rich phase. These ABS chains phase separate from the growing epoxy network, resulting in the formation of ABS substructures in the epoxy-rich matrix. In other words the hydrodynamic effect prevails over the viscoelastic effect of the ABS-rich phase in the later stages of phase separation, hence resulting in complex substructures.31-40 This kind of behavior is usually seen in less viscous systems where the effect of the thermoplastic elastic phase on the growth of the epoxy-rich viscous phase is negligible due to hydrodynamicity of the epoxy-

Phase Separation in ABS-Modified Epoxy/DDS Blends

J. Phys. Chem. B, Vol. 113, No. 16, 2009 5425

Figure 6. Development of SALLS patterns for 10 phr ABS-modified epoxy blends: (a) 50 s, (b) 100 s, (c) 200 s, (d) 2400 s.

Figure 7. Scattering intensity vs scattering vector plot for 10 phr ABSmodified epoxy blends.

rich phase.41 The spherically shaped epoxy and ABS phases possess a minimum interfacial area, which supports the hydrodynamic effect in phase separation, while the occurrence of interconnected networklike structures supports the viscoelastic effect of phase separation. It is important to mention that, apart from the hydrodynamic and viscoelstic effects, the phenomenon “phase inversion” also plays an important role in the formation of substructures, because some epoxy-rich particles are buried in the ABS-rich matrix during phase inversion, and these epoxy subparticles have difficulties diffusing and reaching the epoxyrich continuous phase due to the high viscosity of the ABSrich phase. Therefore, they get trapped as epoxy substructures in the ABS-rich continuous phase. Eventually, in some cases,

some ABS-rich nanosubparticles are observed at the end of the reaction, inside the epoxy-rich substructure. This phenomenon may be due to a secondary phase separation process inside the epoxy-rich substructure. For a more comprehensive understanding of the phase separation mechanism, SALLS investigations were performed. Figure 6 shows the evolution of light scattering patterns during the phase separation of epoxy/ABS blends (10 phr ABS) cured at 180 °C. In the case of 10 phr ABS-modified epoxy blends, it is worth mentioning that concentric rings like pattern appears in the beginning of the phase separation after a few seconds of curing; this may be due to the formation of ABS spherical particles in the epoxy matrix, supporting the phase separation by the NG mechanism.42 There was a slight increase in scattering intensity observed upon curing; this phenomenon may be due to the increase in size of the particles. In the later stages of phase separation, the concentric ring disappears while the intensity of the scattering pattern increases slightly. This may due to the growth of the ABS domains by diffusion, supporting the OM micrographs. We have similar concentric ringlike patterns for 5 phr ABS-modified epoxy blends cured at 180 °C, confirming the phase separation by NG. It is important to mention that ringlike patterns are seen even in the very early stages of curing (after a few seconds); this means that phase separation by nucleation may start immediately after the addition of a curing agent. For more detailed and quantitative analysis of the structural development, the scattering intensity versus scattering vector was plotted, given in Figure 7. Here the scattering peak appears at lower angles, while the scattering intensity increases with time initially; this may be due to the growth of the ABS-rich particles, which scatter more as they increase in size. These results totally agree with the optical micrographs.

5426 J. Phys. Chem. B, Vol. 113, No. 16, 2009

Jyotishkumar et al.

Figure 8. Development of SALLS patterns for 20 phr ABS-modified epoxy blends: (a) 600 s, (b) 610 s, (c) 660 s, (d) 720 s, (e) 2360 s.

Figure 9. Scattering intensity vs scattering vector plot for 20 phr ABSmodified epoxy blends.

Figure 8 shows the evolution of light scattering patterns during the phase separation of the 20 phr ABS-modified epoxy system. A spinodal ring appears after 10 min; the spinodal ring was less intense and was far away from the beam stop. With an increase in time the spinodal ring becomes smaller and sharper, comes closer to the beam stop, and finally merges in

the beam stop. The spinodal ring indicates a correlated structure supporting the SEM, TEM, and OM micrographs. A similar light scattering pattern appeared in the case of 15 phr ABSmodified epoxy blends. This behavior is very typical for the systems having SD.43 It is well-known that SD is a fast process, so what was happening until 600 s cannot be explained by SALLS. However, from OM it is understandable that the initial NG mechanism of phase separation results in the formation of ABS-rich particles dispersed in the epoxy-rich matrix. This means that we have no concentric rings in the 20 and 15 phr ABS-modified epoxy systems in the beginning of phase separation, or in other words, the initial NG mechanism was not visible by SALLS for our systems having both NG and SD phenomena. For more detailed analysis, the scattering intensity versus scattering vector was plotted, given in Figure 9. The scattering peak appearing after 10 min supports the fact that nucleation is not visible by SALLS; only the second-stage phase separation by SD is visible. Scattering peaks at larger angles appear at the beginning of the phase separation typical for SD and shift rapidly to lower angles due to phase coarsening. The intensity of scattered light increased continually from the beginning of phase separation. A clear shift in the main peak in the intermediate to late stages establishes an SD mechanism. A similar behavior was observed for 15 phr epoxy/ABS blends. In summary, we have an NG mechanism for phase separation in 5 and 10 phr ABS-modified epoxy blends. On the other hand, a combination

Phase Separation in ABS-Modified Epoxy/DDS Blends

J. Phys. Chem. B, Vol. 113, No. 16, 2009 5427

Figure 10. Schematic representation of the cross-linked epoxy amine/ABS system according to TEM and SEM micrographs: (a) 95/5 cross-linked epoxy/ABS blends (ABS is dispersed in continuous epoxy), (b) 90/10 cross-linked epoxy/ABS blends (ABS is dispersed in continuous epoxy), (c) 85/15 cross-linked epoxy/ABS blends (bicontinuous morphology with microsubstructures), (d) 80/20 cross-linked epoxy/ABS blends (bicontinuous morphology with micro- and nanosubstructures).

of both NG and SD mechanisms operates for 15 and 20 phr modified epoxy blends. It is important to point out that the SALLS investigational results totally support the facts observed by SEM, TEM, and OM. A schematic model describing the different morphologies of the ABS-modified epoxy systems based on the OM, SEM, and TEM investigations is given in Figure 10. Figure 10a represents a model for the 5 phr ABS-modified epoxy/DDS blends, where the ABS particles are dispersed in a continuous epoxy matrix. Figure 10b represents the 10 phr ABS-modified epoxy/DDS blends with significantly larger ABS droplets. Figure 10c demonstrates the 15 phr modified epoxy blends with a bicontinuous structure in which the both epoxy phase and ABS phase are continuous. In the continuous epoxy phase ABS particles are dispersed, and in the continuous ABS phase epoxy particles are dispersed. Figure 10d represents the scheme of the 20 phr ABS-modified epoxy system, indicating nanosubstructures, which means a continuous phase of epoxy in which ABS particles are dispersed and a continuous phase of ABS in which epoxy microsubstructures are dispersed. Inside the epoxy microsubstructures ABS nanosubsubstructures exist. Dynamic Mechanical Analysis. The viscoelastic properties of the blends were studied using DMA. The plot of tan δ against temperature is shown for DGEBA/ABS blends in Figure 11. A single relaxation was observed for the unmodified epoxy network. In contrast to this, two relaxation peaks were observed in the dynamic mechanical spectrum of the blends as in Figure 11a. This reveals the heterogeneous nature of epoxy blends, which supports the results of SEM and TEM. The relaxation peak around 210 °C corresponds to the Tg of the epoxy-rich phase, and the relaxation peak near 110 °C corresponds to that

of the SAN-rich phase of ABS in the blends. Careful examination of the tan δ curve of the unmodified epoxy network and the cross-linked epoxy blends reveals a relaxation peak of very low amplitude at around -65 °C (called β relaxation). This β relaxation is attributed to the motions of glycidyl units in the network.29 Figure 11b shows the tan δ versus temperature plot for the neat ABS system. Two sharp relaxation peaks were observed in the dynamic mechanical spectrum; the relaxation peak around -88 °C corresponds to the Tg of the polybutadiene phase and that at 107 °C to that of the SAN-rich phase. As mentioned earlier, for the unmodified epoxy network, there exists a well-defined sharp relaxation peak centered at 216 °C, which is ascribed to the Tg of the amine-cured epoxy resin. This value is comparable to the Tg of the cross-linked epoxy/DDS system documented in the literature.44 It is interesting to note that the Tg of the epoxy-rich phase slightly shifts toward the low-temperature side with the addition of the thermoplastic. The decrease in Tg of the epoxy-rich phase can be explained on the basis of three reasons. The first one could be the decrease in the cross-link density of the cured blends. This is because the addition of thermoplastic material raises the viscosity of the system, which may result in an incomplete cross-linking due to the kinetic factors. The second reason may be the dilution effect by the addition of the thermoplastic, which may also result in an incomplete cross-linking. The third reason may be the formation of substructures, which results in subinclusions of the ABS-rich phase in the continuous cross-linked epoxy region, leading to increased miscibility between both the epoxy and ABS phases. The Tg of the SAN phase of ABS remains the same for 5 and 10 phr ABS/epoxy blends. However, for 15 and 20 phr ABS-modified epoxy blends, Tg of the SAN phase

5428 J. Phys. Chem. B, Vol. 113, No. 16, 2009

Jyotishkumar et al.

Figure 11. (a) tan δ vs temperature plot for different cross-linked DGEBA/ABS blends. (b) tan δ vs temperature plot for ABS. (c) Storage modulus vs temperature plot for different cross-linked DGEBA/ABS blends. (d) Loss modulus vs temperature plot for ABS, unmodified crosslinked epoxy, and different cross-linked DGEBA/ABS blends.

TABLE 3: Glass Transition Temperatures (Tg) of Polybutadiene, Styrene-Acrylonitrile, and Epoxy Phases in Cross-Linked Epoxy/ABS Blends Measured by DMA ABS content (phr) 0 5 10 15 20 100

Tg (polybutadiene phase) (°C)

Tg (styreneacrylonitrile phase) (°C)

-88

105 105 112 113 105

Tg (epoxy phase) (°C) 216 216 214 213 208

TABLE 4: tan δ Peak Heights, Peak Widths at Half-Height, and Peak Areas of Different Cross-Linked DGEBA/ABS Blends ABS content (phr)

tan δ peak height

tan δ peak width at half-height

tan δ peak area

0 5 10 15 20

0.73 0.73 0.76 0.85 0.81

19 19 19 18 19

16.36 16.14 16.44 17.33 16.59

increases to a higher temperature. This is because during phase separation some cross-linked epoxy-rich phase gets trapped in the ABS-rich phase. With the progress of the curing reaction, the trapped cross-linked epoxy-rich phase remains in the ABSrich phase. The increase in peak height of the SAN phase

becomes more prominent at high ABS content. The glass transition temperatures of polybutadiene, SAN, and the epoxy phase for all the blends are listed in Table 3. It is important to add that no transitions corresponding to the PB phase of ABS can be obtained from DMA of the blends. The tan δ peak heights, peak widths at half-height, and peak areas are summarized in Table 4. The tan δ peak height for cross-linked epoxy/ABS blends was found to be higher than for the unmodified amine-cross-linked epoxy resin. The increase in the height of the tan δ peak of epoxy/ABS blends is associated with changes in the cross-link density. The addition of the thermoplastic to the epoxy resin increases the viscosity of the material, and hence, the cross-link density decreases upon curing, which results in higher segmental mobility; hence, the peak height increases.45 This may also due to a small amount of ABS molecules dissolved in the epoxy system. The peak areas and peak widths of the epoxy blends are comparable to those of the neat cross-linked resin. Variation of the storage modulus with temperature for neat epoxy resin and blends is shown in Figure 11c. The unmodified cross-linked epoxy resin shows only one inflection point, but the cross-linked epoxy blend shows two inflection points, one at the Tg of the SAN phase of ABS and the other at the Tg of the cross-linked epoxy resin. The storage modulus decreases with an increase in temperature. The storage modulus of the blends is higher than that of the unmodified epoxy network for the entire temperature range except for 15 and 20 phr ABSmodified epoxy blends. The storage modulus of 15 and 20 phr

Phase Separation in ABS-Modified Epoxy/DDS Blends

J. Phys. Chem. B, Vol. 113, No. 16, 2009 5429

TABLE 5: Apparent Weight Fractions (ω) of the Epoxy and ABS Components in the Epoxy-Rich Phase and the ABS-Rich Phase Calculated According to the Woods Equationa apparent weight apparent weight apparent weight apparent weight cross-linked styrene-acrylonitrile fraction of epoxy fraction of ABS in fraction of epoxy fraction of ABS ABS content epoxy phase in the epoxy-rich the epoxy-rich phase, in the ABS-rich phase, in the ABS-rich phase, phase Tg2 (K) (phr) Tg1 (K) phase, ω1′ ω2′ ω1′′ ω2′′ 0 5 10 15 20 100

489 489 487 486 481

380 380 385 386 380

1 0.98 0.97 0.93

0 0.02 0.03 0.07

0 0 0.05 0.06

1 1 0.95 0.94

a The single prime and double prime denote the epoxy-rich phase and ABS-rich phase, respectively, and the subscripts 1 and 2 denote epoxy and ABS components.

ABS-modified cross-linked epoxy blends is higher than that of the unmodified epoxy network up to 110 °C; at higher temperatures the storage modulus is lower than that of the unmodified epoxy network. At around 110 °C, the decrease in storage modulus to some extent is due to the bicontinuous nature of the blends. Higher values of the storage modulus of epoxy /ABS blends compared to the unmodified epoxy network reveal better load-bearing capacity of the blends. The higher values of the storage modulus of the cross-linked blends show that the blends have better interaction between the phases. A sharp decrease in the storage modulus was observed for all blends near the glass transition of the epoxy network; it thereafter remained constant in the rubbery plateau region, which is typical for cross-linked polymers. The storage modulus values of 5 and 10 phr ABS-modified epoxy resin are higher than that of neat epoxy in the rubbery plateau region, while the storage modulus of 15 and 20 phr ABS-modified epoxy network in the rubber plateau region is lower than that of unmodified epoxy networks due to the continuous nature of the ABS phase. The plot of the loss modulus as a function of temperature is shown in Figure 11d. The loss modulus of cross-linked blends is higher than that of the unmodified network. The higher values of the loss modulus of the cross-linked blends show that the blends have better interaction between phases. Three relaxation peaks are observed in the loss modulus curves. The peak at around 110 °C is due to the thermoplastic phase. The increase in peak height of the SAN phase becomes more prominent at higher ABS content. The peak at around 210 °C is due to the epoxy-rich phase, and a β relaxation peak at around -65 °C supports the results of the tan δ curve. Careful examination of loss modulus curves reveals a peak around 60 °C called the ω relaxation peak due to the lower cross-link density sites in the epoxy network or due to the β-relaxation overtones in the regions of higher cross-link density matrix which are occluded in the lower cross-link density matrix.30 Miscibility Behavior of Cross-Linked Epoxy/ABS Polymer Blends. The miscibility of blends of epoxy/ABS was examined by determining experimentally the glass transition temperature by DMA. Two glass transition regions, which we designated as Tg associated with the cross-linked epoxyrich phase and Tg associated with the ABS-rich phase, are evident from the DMA results. The apparent weight fraction of epoxy and ABS in the epoxy-rich phase and in ABS-rich phase was determined using the Woods equation.46-48 The Woods equation is often used to determine the dependence of Tg on the composition in random copolymers and plasticized systems.

According to Woods

Tg ) ω1Tg1 + ω2Tg2

(6)

where Tg is the observed Tg of the blend, ω1 is the weight fraction of homopolymer 1 having Tg1, and ω2 is the weight fraction of homopolymer 2 having Tg2. Equation 6 can be rearranged49 to

ω1′ ) Tg1,b - Tg2/Tg1 - Tg2

(7)

where ω1′ is the apparent weight fraction of polymer 1 in the polymer 1-rich phase, Tg1,b is the observed Tg of polymer 1 in blends, and Tg1 and Tg2 are the Tg values of homopolymers 1 and 2, respectively. The apparent weight fractions of epoxy and ABS in the epoxy-rich phase and ABS-rich phase were calculated by Woods equations, and the results are summarized in Table 5. From the data obtained, we can conclude that epoxy/ ABS blends are partially miscible, especially 15 and 20 phr ABS/epoxy blends. The partial miscibility in 15 and 20 phr ABS/epoxy blends may be due to development of substructures in bicontinuous blends. Conclusion New micro- and nanosubstructures in cross-linked epoxy/ABS blends have been investigated as a function of the concentration of ABS in the epoxy system for the first time. At lower concentrations of ABS (5 and 10 phr), ABS particles were dispersed in the continuous epoxy matrix. Very complex morphologies have been obtained for 15 and 20 phr ABSmodified epoxy blends. SEM reveals a bicontinuous phase morphology, with substructures of the ABS phase dispersed in the epoxy phase and substructures of the epoxy phase dispersed in the continuous ABS phase for the blends with 15 and 20 phr ABS-modified cross-linked epoxy. However, TEM micrographs of 20 phr ABS showed nanosubstructures of ABS dispersed in the microsubstructure of the epoxy phase which is dispersed in the continuous ABS phase. From optical and SALLS measurements we confirm that the phase separation took place by the NG mechanism for 5 and 10 phr blends and a combined effect of both the NG and SD mechanisms for 15 and 20 phr blends. We believe that in all these blends the phase separation must have been initiated by the immiscible SAN-graft-PB micelles present in the blend system. Since the size of the micelles is very small (nanostructured), we cannot detect them using TEM and SALLS; however, experiments are progressing using synchrotron radiation and neutron scattering. We have estab-

5430 J. Phys. Chem. B, Vol. 113, No. 16, 2009 lished that the phenomenon of micro- and nanosubstructures is due to the combined effect of hydrodynamics and viscoelastic effects acting on the system followed by phase inversion. It is important to mention that a bicontinuous morphology was conserved only for near-critical compositions; for off-critical compositions a dispersed phase morphology is observed. The DMA results fully support the facts observed by SEM, TEM, OM, and SALLS. Finally, on the basis of the Woods approach, the apparent weight fractions of epoxy and ABS in the epoxyrich phase and ABS-rich phase have been estimated. Acknowledgment. The contributions of Prof. Andre´ Laschewsky, Dr. Sabine Kosmella, and Mrs. Sibylle Rusting, all from the Institute of Chemistry, University of Potsdam, are gratefully acknowledged. We thank Prof. J. P. Pascault for enlightening discussions. References and Notes (1) Strehmel, V.; Krukowka, L.; Fedtke, M. Chem. Tech. (Leipzig) 1990, 42, 416–419. (2) Strehmel, V.; Zimmermann, E.; Ha¨usler, K.-G.; Fedtke, M. Prog. Colloid Polym. Sci. 1992, 90, 206–208. (3) Scherzer, T.; Strehmel, V.; Ta¨nzer, W.; Wartewig, S. Prog. Colloid Polym. Sci. 1992, 90, 202–205. (4) Fryauf, K.; Strehmel, V.; Fedtke, M. Polym. Bull. 1993, 31, 183– 190. (5) Strehmel, V.; Scherzer, T. Eur. Polym. J. 1994, 30, 361–368. (6) Strehmel, V.; Deltschewa, K.; Ha¨usler, K. G.; Schro¨ter, K. Angew. Makromol. Chem. 1994, 220, 99–109. (7) Kim, B. S.; Chiba, T.; Inoue, T. Polymer 1995, 36, 43–47. (8) Girard-Reydet, E.; Riccardi, C. C.; Sautereau, H.; Pascault, J. P. Macromolecules 1995, 28, 7608–7611. (9) Girard-Reydet, E.; Sautereau, H.; Pascault, J. P.; Keates, P.; Navard, P.; Thollet, G.; Vigier, G. Polymer 1998, 39, 2269–2280. (10) Girard-Reydet, E.; Vicard, V.; Pascault, J. P.; Sautereau, H. J. Appl. Polym. Sci. 1997, 65, 2433–2445. (11) Abad, M. J.; Barral, C.; Cano, J.; Lopez, J.; Nagueira, P.; Ramirez, C.; Torres, A. Eur. Polym. J. 2001, 37, 1613–1623. (12) Cano, J.; Torres, A.; Abad, M. J.; Barral, L.; Diez, F. J.; Lopez, J. Polym. Int. 2002, 51, 1268–1276. (13) Lopez, J.; Ramirez, C.; Abad, M. J.; Barral, L.; Cano, J.; Diez, F. J. J. Appl. Polym. Sci. 2002, 85, 1277–1286. (14) Cui, J.; Chen, W.; Zhang, Z.; Li, S. Macromol. Chem. Phys. 1992, 198, 1865–1872. (15) Tanaka, H. Macromolecules 1992, 25, 6377–6380. (16) Tanaka, H. J. Chem. Phys. 1995, 103, 2361–2364. (17) Tanaka, H. J. Chem. Phys. 1996, 105, 10099–10114.

Jyotishkumar et al. (18) Tanaka, H. J. Phys.: Condens. Matter 2000, 12, R207-R264. (19) Tanaka, H.; Araki, T. Phys. ReV. Lett. 1998, 81, 389–392. (20) Tanaka, H. J. Phys.: Condens. Matter 2001, 13, 4637–4674. (21) Zhou, D.; Zhang, P.; Weinan, E. Phys. ReV. E 2006, 73, No. 061801 (1-10). (22) Siggia, E. D. Phys. ReV. A 1979, 20, 595–605. (23) Huo, Y.; Jiang, X.; Zhang, H.; Yang, Y. J. Chem. Phys. 2003, 118, 9830–9837. (24) Gan, W.; Yu, Y.; Wang, M.; Tao, Q.; Li, S. Macromolecules 2003, 36, 7746–7751. (25) Jose, J.; Thomas, S.; Pionteck, J.; Joseph, K. J. Phys. Chem. B 2008, 112, 14793–14803. (26) Oyanguren, P A.; Riccardi, C. C.; Williams, R. J.; Mondragon, I. J. Polym. Sci., Part B: Polym. Phys. 1998, 36, 1349–1359. (27) Min, B. G.; Stachurski, Z. H.; Hodgkin, J. H. J. Appl. Polym. Sci. 1993, 50, 1511–1518. (28) Woo, E. M.; Bravence, L. D.; Seferis, J. C. Polym. Eng. Sci. 1994, 34, 1664–1673. (29) Guo, Q.; Figueiredo, P.; Thomann, R.; Gronski, W. Polymer 2001, 42, 10101–10110. (30) Francis, B.; Thomas, S.; Asari, G V.; Ramaswamy, R.; Jose, S.; Rao, V. L. J. Polym. Sci., Part B: Polym. Phys. 2006, 44, 541–556. (31) Taniguchi, T.; Onuki, A. Phys. ReV. Lett. 1996, 77, 4910–4913. (32) Goossens, S.; Goderis, B.; Groeninckx, G. Macromolecules 2006, 39, 2953–2963. (33) Tanaka, H. Phys. ReV. Lett. 1996, 76, 787–790. (34) Aoki, Y.; Tanaka, T. Macromolecules. 1999, 32, 8560–8565. (35) Tanaka, H. Phys. ReV. Lett. 1994, 72, 3690–3693. (36) Clarke, N.; McLeish, T. C. B.; Jenkins, S. D. Macromolecules 1995, 28, 4650–4659. (37) Zhang, J. ; Zhang, Z. , Zhang, H.; Yang, Y. Phys. ReV. E 2001, 64, No. 051510 (1-10). (38) Tanaka, H. Phys. ReV. E 1997, 56, 4451–4462. (39) Witzius, P.; Cumming, A. Phys. ReV. Lett. 1991, 66, 3000–3003. (40) Tegze, G.; Pustai, T.; Granasy, L. Mater. Sci. Eng. 2005, 413, 418– 422. (41) Yu, Y.; Wang, M.; Gan, W.; Tao, Q.; Li, S. J. Phys. Chem. B 2004, 108, 6208–6215. (42) Kulkarni, A. S.; Beaucage, G. Polymer 2005, 46, 4454–4464. (43) Yang, K.; Yang, Q.; Li, G.; Sun, Y.; Mao, Y. Mater. Lett. 2006, 60, 589–593. (44) Grillet, A. C.; Galy, J.; Pascault, J. P.; Bardin, I. Polymer 1989, 30, 2094–2103. (45) Rao, B. S.; Reddy, K. R.; Pathak, S. K.; Pasala, A. R. Polym. Int. 2005, 54, 1371–1376. (46) Perrin, P.; Prud’homme, R. E. Acta Polym. 1993, 44, 307–312. (47) Kim, W. N.; Burns, C. M. Macromolecules 1987, 20, 1876–1882. (48) Wood, L. A. J. Polym. Sci. 1958, 28, 319–330. (49) Stoelting, J.; Karasz, F. E.; MacKnight, W. J. Polym. Eng. Sci. 1970, 10, 133–138.

JP8094566