Dynamics of Phase Separation in Poly(acrylonitrile ... - ACS Publications

J. Phys. Chem. B 2010, 114, 13271–13281

13271

Dynamics of Phase Separation in Poly(acrylonitrile-butadiene-styrene)-Modified Epoxy/DDS System: Kinetics and Viscoelastic Effects ¨ zdilek,‡ Paula Moldenaers,‡ Christophe Sinturel,§ Andreas Janke,| P. Jyotishkumar,† Ceren O | Ju¨rgen Pionteck, and Sabu Thomas*,†,⊥ School of Chemical Sciences, Mahatma Gandhi UniVersity, Priyadarshini Hills, Kottayam, Kerala 686560, India; Department of Chemical Engineering, Catholic UniVersity of LeuVen, de Croylaan, 46, B-3001, LeuVen, Belgium; Centre de Recherche sur la Matie`re DiViseé, UMR 6619 CNRS UniVersite´ d’Orleáns, 1 B Rue de la Fe´rollerie, FR 45071, Orleáns Cedex 2, France; Leibniz Institute of Polymer Research Dresden, Hohe Str 6, DE 01069, Dresden, Germany; and Centre for Nanoscience and Nanotechnology, Mahatma Gandhi UniVersity, Priyadarshini Hills, Kottayam, Kerala 686560, India ReceiVed: February 24, 2010; ReVised Manuscript ReceiVed: September 19, 2010

The dynamics of phase separation and final morphologies of poly(acrylonitrile-butadiene-styrene) (ABS)modified epoxy system based on diglycidyl ether of bisphenol A (DGEBA) cured with 4,4′-diaminodiphenylsulfone (DDS) have been monitored in situ throughout the entire curing process by using optical microscopy (OM), differential scanning calorimetry (DSC), rheometry, and small-angle laser light scattering (SALLS). The evolution of phase separation and final morphologies with substructures were explored by OM. The final morphologies of the blend cured at 150 and 165 °C are of phase-inverted type and are quite different from the final morphologies of the same blend cured at 180 °C, in which the final morphologies are cocontinuous. AFM observations of the fully cured sample confirmed the existence of three different phases, the epoxy continuous phase, SAN (styrene/acrylonitrile) continuous phase, and PB droplets at the interface, with a strong tendency to stay at SAN continuous phase. Furthermore, the continuous epoxy phase contains SAN particles and the continuous SAN phase contains epoxy particles. Cure kinetics and rheological results correspond well with the viscoelastic phase separation revealed by OM. The SALLS results display clearly that the phase separation takes place according to nucleation and growth mechanism followed by spinodal decomposition. The development of light scattering patterns during the second stage phase separation follows the Cahn-Hilliard model of spinodal demixing. Furthermore, the evolution of the scattering vector follows a Maxwell-type relaxation equation establishing the viscoelastic behavior of phase separation. The relaxation time of phase separation can be described by the Williams-Landel-Ferry equation for viscoelasticity. As a whole, the dependence of phase separation on cure temperature and the development of final morphologies and the associated mechanisms were explored in detail for the complex epoxy/ABS system. 1. Introduction The viscoelastic effects on phase separation and the evolution of phase morphology in thermoset-thermoplastic blends have attracted increased attention due to their great significance to optimize the mechanical properties. Systems such as polysulfone-modified epoxy,1 polyoxymethylene-modified epoxy,2 etc. have been studied. On mixing thermoplastic with epoxy resin, one can obtain a homogeneous or heterogeneous mixture. Upon curing, phase separation starts immediately due to the increase in molecular weight of the epoxy resin, which results in a thermodynamically unstable system.3 Tanaka and Araki reported that viscoelastic effects play a crucial role in phase separation for the blends with dynamic asymmetry.4-6 In other words, viscoelastic effects on phase separation originate from the dynamic asymmetry in the relaxation and diffusion of molecular * To whom correspondence should be addressed. Phone: +91-4812730003. Fax: +91-481-2731002. E-mail: [email protected], [email protected]. † School of Chemical Sciences, Mahatma Gandhi University. ‡ Catholic University of Leuven. § UMR 6619 CNRS Universite´ d’Orleáns. | Leibniz Institute of Polymer Research Dresden. ⊥ Centre for Nanoscience and Nanotechnology, Mahatma Gandhi University.

chains, which may be due to large difference in the Tg and molecular weight of the blend components.7,8 The phase separation may take place through two main mechanisms, which are nucleation and growth (NG) or spinodal decomposition (SD).9-14 SALLS, OM, DSC, and rheology are the most imperative techniques to explore the phase separation behavior of polymer blends.15-17 Apart from the evolution of phase separation, the curing reaction in epoxy/DDS/ABS blends is accompanied by chemical reactions and physical transformations from gelation to vitrification and hence results in a complex kinetic behavior and deserves a detailed exploration. In comparison with thermoplastics, the processing of thermosets and thermoset matrix composites is more complicated and less controlled because of their reactivity. In these processes, polymer synthesis and shaping take place in a single operation: this involves the conversion of liquid monomers or polymers into solid crosslinked polymers.18 The mechanism and kinetics of cure determine the network morphology, which in turn decides the physical and mechanical properties of the cured product. The viscosity is a very important parameter to characterize accurately the process conditions and to visualize the gelation and vitrification transition state.19 Thus, the knowledge of rheological

10.1021/jp101661t  2010 American Chemical Society Published on Web 10/06/2010

13272

J. Phys. Chem. B, Vol. 114, No. 42, 2010

and curing behavior of the resin system is particularly necessary to control the processing and subsequent engineering design.20 Moreover, the rheological analysis is indispensable to find out the relationship between conversion, phase separation, and viscosity evolution for the process simulation.19 For thermosetting systems in industrial applications, it is important to establish relations of transition phenomena to the reaction time at different cure temperature so that desired properties can be controlled.21 Furthermore, the study of mechanism of phase separation and the evolution of scattering vector (q) are very important to establish the effect of viscoelastic behavior on the resulting morphologies and properties. The cure temperature and polymer molecular weight are important factors influencing the thermodynamic and kinetic aspects of phase separation and thus the final morphology of the epoxy blends. However, the influence of cure temperature and polymer molecular weight on the phase separation has not been extensively studied to date. Moreover, the mechanism of phase separation is not fully understood for epoxy/ABS thermoplastic blends. In this paper, our focus is to understand the mechanism of phase separation in epoxy/ABS blends by considering all the above points. As shown in our previous publications, there is a high dependence of the cure kinetics and final morphologies on the thermoplastic concentration.13,22 Recently, we have also recognized that, for the epoxy blend system containing 12.9 wt % ABS, phase separation takes place by the combination of both NG and SD mechanism with a welldefined phase inversion process.14 The main objective of this work is to investigate the affect of temperature and viscoelastic effects on the dynamics of phase separation of an ABS-modified epoxy system. For that purpose, a time-resolved study of the phase separation was carried out by OM, DSC, rheology, and SALLS. 2. Experimental Section 2.1. Materials Used. The matrix material used in the experiments consists of diglycidyl ether of bisphenol A (DGEBA) (Lapox L-12, Atul Ltd., India.) and 4,4′-diaminodiphenylsulfone (DDS) (Lapox K-10, Atul Ltd., India). The epoxy content in Lapox L-12 varies between 5.25 and 5.40 eq/kg. Lapox L-12 has a viscosity of about 1.15-1.2 g/cm3. DDS is a white powder with a melting point of 178 °C. It 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. The used poly(acrylonitrile-butadiene-styrene) (ABS) is a commercially available thermoplastic polymer consisting of 70 wt % polystyrene (PS), 25 wt % acrylonitrile (AN), and 5 wt % polybutadiene (PB). The molecular weight of the soluble part of ABS was determined to be Mn ) 51 300 g/mol and Mw ) 125 200 g/mol (PDI ) 2.44, GPC, PS standard) and the density was determined to be 1.051 g/cm3 by means of an helium pycnometer. 2.2. Preparation of Blends. The epoxy/ABS blend containing 12.9 wt % ABS was prepared by mixing 20 g of ABS with 100 g of DGEBA at 180 °C under constant stirring. After proper mixing, 35 g of DDS was added to epoxy/ABS mixture corresponding to a stoichiometric epoxy:amine ratio of 2:1. The freshly prepared mixtures were immediately used for the analysis. 2.3. Characterization Techniques. 2.3.1. Optical Microscopy. A few milligrams of the epoxy/ABS system placed between two glass slides were observed by a Leitz laborlux 12 Pols optical microscope, while the sample was being cured in a Linkam CSS450 shearing cell. The Linkam cell was only used

Jyotishkumar et al. for heating; no shear was applied. Digital micrographs were taken at several times by a Hamamatsu TSU digital camera controller C4742-95. 2.3.2. Differential Scanning Calorimetry. The calorimetric measurements were performed on a Perkin-Elmer Pyris DSC 6 differential scanning calorimeter. The instrument was calibrated with indium and dry cyclohexane standards. Dry nitrogen was used as purge gas and samples of 15-20 mg were analyzed. Dynamic DSC was done at 2.5, 5, 7.5, and 10 °C/min from 20 to 350 °C for the neat epoxy resin/DDS mixture. Isothermal measurements were done at 150, 165, and 180 °C. The curing was assumed to be completed when the isothermal curve leveled off to a straight line. The areas under the peaks during the isothermal cure were used to determine the conversion R at various times. The conversion R ) ∆Ht/∆Htot, where ∆Ht is the heat of cure at time t, calculated by the integration of the DSC isothermal signal, and ∆Htot is the total heat of cure of the neat epoxy resin calculated by the integration of DSC dynamic signal. 2.3.3. Oscillatory Shear Rheology. The viscoelastic properties of the epoxy blends during isothermal curing were determined by oscillatory shear measurements performed on an ARG2 rheometer (TA Instruments). Each sample was placed on aluminum parallel plates of 25 mm in diameter, which were of disposable type. Temperature control was achieved by an ETC system (environmental testing chamber) suitable for studying the polymer melts in the temperature range -160 to 600 °C under a N2 atmosphere. The experiments were carried out with 5% strain and an angular frequency of 1 Hz and were used to determine G′, G′′, tan δ, and complex viscosity. 2.3.4. Small-Angle Laser Light Scattering (SALLS). The small-angle light scattering setup (Newport model ULM-TILT) consists of a 5 mW He-Ne laser (λ ) 638.2 nm) as the scattering light source. The sample is placed between two glass slides being cured in a Linkam cell (CSS450). The distance between the sample and screen was adjusted as 29 cm. The laser beam passing through the Linkam cell was examined by a Pulnix camera, which was used to record the change in scattering patterns. The processing of the data was performed using the special SALLS software developed in Catholic University, Leuven, Belgium. 2.3.5. Atomic Force Microscopy. The AFM measurements were done in the tapping mode by a Dimension 3100 Nanoscope V (Veeco, USA). We used silicon-SPM sensors (Budget Sensors, Bulgaria) with a spring constant of ca. 40 N/m and resonance frequency of ca. 280 kHz; the tip radius is lower than 10 nm. 2.3.6. Field Emission Scanning Electron Microscopy. The morphology of the cross-linked epoxy blend was examined using a ULTRA FESEM (model ULTRA plus, Carl Zeiss NTS GmbH, Germany). The fractured samples were smoothed with an ultramicrotome and the SAN-rubber phase (SAN ) styrene/ acrylonitrile) was etched out by immersing the cut in chloroform for 2 h. The samples were coated with platinum by vapor deposition using a SCD 500 sputter coater (BAL-TEC AG, Liechtenstein). 3. Results 3.1. Evolution of Phase Separation by OM. Figures 1, 2, and 3 demonstrate the phase separation and evolution of phase morphologies in the ABS-modified epoxy/DDS blends cured at different temperatures, illustrating different stages of the phase separation process from the initially formed matrix droplet morphology to the final complex structures. Before going into the detailed mechanism of phase separation in the ABS-modified

Phase Separation of an ABS-Modified Epoxy System

J. Phys. Chem. B, Vol. 114, No. 42, 2010 13273 TABLE 1: Time for the Generation of New Epoxy Phase Due to Complex Phase Separation

Figure 1. Evolution of phase morphologies of ABS-modified epoxy blends at 150 °C: (a) 360 s, (b) 2220 s, (c) 2420 s, (d) 2600 s, (e) 2900 s, (f) 3680 s, and (g) 9000 s (800 × 800 µm2).

Figure 2. Evolution of phase morphologies of ABS-modified epoxy blends at 165 °C: (a) 480 s, (b) 1440 s, and (c) 6000 s (800 × 800 µm2).

Figure 3. Evolution of phase morphologies of ABS-modified epoxy blends at 180 °C: (a) 480 s, (b) 960 s, and (c) 6000 s (800 × 800 µm2).

epoxy system, it is important to mention that the Tg of SAN and PB are 110 and -80 °C, respectively, while that of epoxy monomer is -20 °C and hence there is a strong dynamic asymmetry. As mentioned in the Introduction, the other important factor that contributes to the dynamic asymmetry is the difference in molecular weight. The number-average molecular weights of ABS and the epoxy resin (DGEBA) were Mn ) 51 300 and 355 g/mol, respectively. This dynamic asymmetry can play a very important role in the dynamics of phase separation and the final morphology. The final morphologies of 12.9 wt % ABS-modified epoxy blends cured at 150 and 165 °C are of phase-inverted type and are quite different from the final morphologies of ABS-modified epoxy blends cured at 180 °C, in which the final morphology

temp (°C)

time for the appearance of second increase in reaction rate (shoulder) from DSC (s)

time for the generation of new epoxy phase from OM (s)

time for the generation of new epoxy phase from rheology (s)

150 165 180

2440 1220 760

2400 1380 920

2106 1021 662

is cocontinuous. This indicates that, at temperatures higher than 165 °C, the rate of phase separation is faster and, as a result, the generated epoxy domains grow faster. In later stages, the growing epoxy domains start to agglomerate, resulting in the final cocontinuous morphology, in which both SAN phase and epoxy phase are continuous. The time at which the formation of the new epoxy phase due to phase separation could be detected and is listed in Table 1. The effect of temperature on the phase separation was also carefully examined using DSC, rheology, and SALLS and the results are elaborated in detail in the coming text. From the evolution of morphology and resultant final morphologies at different temperatures, it was obvious that the temperature plays a prominent role for the growth rate of phase-separated structures and the resulting final phase morphologies. 3.2. Calorimetric Analysis. The epoxy/ABS blend was cured isothermally at three different temperatures, i.e., 150, 165, and 180 °C. Figure 4a shows the isothermal DSC curves of neat epoxy system and the ABS-modified epoxy blend cured at 150 °C. A rapid increase in the reaction rate was observed which passes through a maximum in the exothermic heat flow. In the presence of ABS the reaction rate is reduced. However, a second increase in reaction rate was observed in the DSC curve of the ABS-containing blend. The time at which this second increment in reaction rate occurs (appearance of the shoulder) is listed in Table 1. A similar kind of acceleration phenomenon was observed by Pascault and co-workers for the blends based on polystyrene or polyetherimide in DGEBA/MCDEA (4,4′methylenebis(3-chloro-2,6-diethylaniline)) systems3 and poly(methyl methacrylate) in a DGEBA/DDS system.23 To study the effect of temperature on the rate and phase separation of the blend, a comparison of curing rate vs time in the ABS-modified epoxy cured at 150, 165, and 180 °C is shown in Figure 4b. The peak maximum and the shoulder are slightly shifted toward shorter cure times with increase in temperature. This reflects an increase in the reaction rate as well as the rate of phase separation at higher cure temperatures supporting the OM micrographs. This is because of the increased molecular mobility that favors the chemical interactions such that a higher number of chains are involved in the curing process and, as a consequence, both the rate of the reaction and rate of phase separation are increased. 3.3. Rheological Behavior. The profile of change in complex viscosity of ABS-modified epoxy blends with time at different temperatures (150-180 °C) is shown in Figure 5. In general, the rheology on polymer blends is an accepted method to provide valuable information on the structure development during curing. At the beginning of the cure, the viscosity slowly increases with time and then, at a certain point, a very rapid increase is observed. Moreover, this increase in complex viscosity curves is shifted to shorter times with increase in temperature; this is because the reactions between epoxy and DDS are thermally accelerated. The reaction between epoxy

13274


Jyotishkumar et al.

Figure 6. Storage modulus, loss modulus, and tan δ vs time curves of modified epoxy blends at 150 °C.

Figure 4. (a) Curing rate against conversion plot for neat resin and epoxy/ABS blends cured at 150 °C. (b) Curing rate against time plot for epoxy/ABS blends cured at 150, 165, and 180 °C.

Figure 5. Complex viscosity of ABS-modified epoxy blends at 150, 165, and 180 °C.

and DDS induces various chemical and physical phenomena. In fact, the resin viscosity increases and cross densification advances. The curing reaction proceeds until vitrification or until the transformation of the rubbery network into a glassy solid takes place.24-26

It is well-known that the phase separation has a profound effect on the rheological behavior of the blends. It is important to mention that phase separation is triggered by the NG mechanism, which takes place immediately after mixing the curing agent into the epoxy monomer/ABS blend.14 This means that phase separation may affect the curing reaction even at very low conversion but this may not be prominent. On the other hand, we have an increase in reaction rate of the epoxy amine system during the second stage of phase separation by spinodal decomposition. This means that the phase separation affects the reaction rate and hence may influence the gelation times of the epoxy blends. However, we have difficulties to predict the exact gelation times of the epoxy blends since they are always coupled with the phase separation process. We have investigated the evolution of rheological parameters such as tan δ, G′, and G′′ at different temperatures; the results seem to be identical irrespective of temperature. For avoiding overlapping of the results, we are giving only the representative rheological profile obtained at 150 °C. Figure 6 reveals the rheological parameter as a function of real time/conversion scale in case of epoxy/ABS-modified epoxy cured at 150 °C. The rheological curves reveal a reduction in the tan δ curve during the early stages of the reaction, which may be due to dominance of SAN-rich phase (the generation of thermoplastic-rich continuous phase) and is accompanied by a gradual increase in viscosity; hence, this drop in tan δ curve supports the results from OM micrographs. During the early stages of the reaction, G′ is below G′′, indicating the liquid nature of the material; after some time, both G′ and G′′ increase rapidly. As mentioned before, this phenomenon may be due to the formation of continuous SAN-rich phase; this is followed by a crossover point where G′ equals G′′ which means the system acts as both elastic and viscous, storing and dissipating an identical amount of energy at this point. Pascault et al. observed a similar kind of behavior for the blends based on polystyrene (PS) or polyetherimide (PEI) in DGEBA/MCDEA systems3 and poly(methyl methacrylate) (PMMA) in the DGEBA/DDS system.23 3.4. Phase Separation Dynamics by SALLS. The phase separation process of all modified systems was monitored in situ by SALLS. The change in scattered intensity and scattering vector (q) was recorded at appropriate time intervals during isothermal curing at 135, 140, 150, 160, 165, 170, and 180 °C. At all curing temperatures, the light scattering profile remains the same irrespective of the temperature.


J. Phys. Chem. B, Vol. 114, No. 42, 2010 13275

Figure 7. Scattering intensity vs scattering vector profile for the ABS-modified epoxy system cured at (a, b) 135, (c) 150, (d) 165, and (e) 180 °C.

Figure 7a shows the scattering intensity vs scattering vector (q) for with 12.9 wt % ABS-modified blends cured at 135 °C during the very early stage of phase separation. Here the scattering peaks at low q values are observable at the beginning of cure, implying the presence of microphase separation. The following factors are possible for the initiation of microphase separation: the first one could be the presence of SAN-grafted

PB particles, and the PB phase in SAN-grafted PB is crosslinked and hence they are insoluble in epoxy resin. The presence of insoluble SAN-grafted PB may initiate the phase separation process. The second reason for phase separation probably from the thermal treatment of samples at the temperature as high as 180 °C. These initial scattering patterns during early stages of phase separation exactly match with our previous observations

13276


for NG mechanism.14 Figure 7b shows the scattering intensity vs scattering vector (q) during the second-stage phase separation. Scattering peaks are at larger q values originating at the beginning of the second stage phase separation and are typical for SD. These peaks shift rapidly to lower q values due to epoxy phase coarsening. A clear shift of the main peak in the intermediate to late stages establishes an SD mechanism. Furthermore, a sharp increase in intensity of scattered light from the beginning of second-stage phase of separation was observed. Figure 7c-e shows the scattering intensity vs scattering vector during the second-stage phase separation by SD mechanism at 150, 165, and 180 °C, respectively. From the light scattering profiles at different temperatures, we observe a rapid shift in the scattering vector to lower values at 180 °C as compared to light scattering profiles at lower temperatures, establishing rapid epoxy phase coarsening by coalescence at high temperatures. This supports the facts observed by OM, DSC, and rheological measurements. Even though the final morphologies are different at different temperatures, the shape of the scattering profiles looks almost identical. 3.5. Final Morphology by AFM and FESEM. AFM investigations were made to understand the final morphology in detail. According to Magonow et al.,27 we choose the scan conditions (free amplitude >100 nm, set-point amplitude ratio 0.8) in order to get stiffness contrast in the phase image, which means bright features in the phase image are stiffer than dark areas. Parts a and b of Figure 8 show the AFM micrograph of neat ABS and completely cured ABS-modified epoxy/DDS system at 180 °C. The AFM micrograph of ABS shows that the cross-linked PB phase is dispersed in the SAN continuous phase, indicating two-phase morphology. The dispersed PB domains consist of small agglomerates of original PB (dark phase) particles containing some (bright phase) grafted SAN. On the other hand, the AFM micrographs of the completely cross-linked ABS/epoxy blend show a very interesting morphology with three different phases: two continuous phases forming a cocontinuous structure with substructures (epoxy continuous phase containing dispersed SAN particles and the SAN continuous phase in which epoxy particles are dispersed) and the most important feature, the PB phase appears as dispersed small agglomerates at the blend interface between the cocontinuous structures. The driving force for the segregation at the interface between the SAN and epoxy continuous phases is the minimization of the specific interfacial energy of the system, but there is still a tendency to locate more near to the SAN phase due to the better compatibility caused by the grafted SAN part. A minor quantity of PB is dispersed in the SAN continuous phase. For the better understanding of the final generated complex morphology of ABS-modified epoxy/DDS system at 180 °C, the blend was carefully examined by FESEM. FESEM micrograph of epoxy/ABS blend after ultramicrotome cutting followed by etching in chloroform for 2 h is shown in Figure 8c. During the etching process, the SAN phase has been dissolved away by chloroform. FESEM micrograph of the 12.9 wt % ABScontaining cross-linked epoxy specimen exhibits a cocontinuous structure in which both the epoxy phase as well as the SAN phase are continuous. In the continuous epoxy phase, SAN particles are dispersed, while in the continuous SAN phase clusters of big epoxy particles were observed. These particles result from viscoelastic phase separation as described below. 4. Discussion In general, the phase separation phenomenon in the modified epoxy blends may be interpreted as follows based on OM

Jyotishkumar et al. analysis. The system was partially miscible initially (before curing). There were only the small insoluble SAN-graft-PB particles present in the homogeneous DGEBA/DDS/SAN mixture. Because of the thermal treatment of the samples, these SAN-graft-PB particles dispersed in the homogeneous epoxy/ SAN mixture may initiate the phase separation through nucleation and growth mechanism. During the initial stages of phase separation, spherical SAN nuclei are found to develop in the epoxy matrix by the nucleation process. More and more SAN particles may form by NG as the curing time advances (Figures 1a, 2a, 3a). After a few minutes, the system transforms into a cocontinuous structure, like the one in Figure 1b. The cocontinuous structure may come from the SD mechanism, in which both SAN-rich phase with PB particles and epoxy-rich phase are continuous. After this point, the cocontinuous structure ruptures rapidly and the SAN-rich phase with PB particles becomes the continuous phase while the epoxy phase becomes the dispersed phase, indicating a clear phase inversion at which epoxy particles develop in the SAN continuous matrix as shown in Figure 1c. This transition in the phase behavior could be explained as follows. The enhancement of the concentration fluctuation makes the SAN-rich phase more elastic (due to its high molecular weight) than epoxy-rich phase such that these epoxy droplets subsequently grow by the diffusion process (Figure 1d-g). During the fast growth of the epoxy phase, there may be some miscible SAN chains get separated out of the SAN-rich phase. These original miscible SAN chains finally phase separate due to the reduced miscibility caused by the epoxy network formation resulting in the formation of SAN substructures in the epoxy-rich matrix. On the other hand, during the phase inversion process, some epoxy will not reach the growing epoxy-rich domains since the diffusion is hampered by the highly viscous SAN surrounding. This part of the epoxy resin trapped in the SAN-rich phase separates in later stages to spherical epoxy substructures in the SAN-rich continuous phase. The immiscible PB particles are driven during this process to the interface of the bulky epoxy phase and SAN phase, with a tendency to have more contacts with the SAN phase due to the better compatibility caused by the grafted SAN part. Kyu et al. reported similar kind of complex crossover behavior of reactioninduced phase separation from NG to SD.28 These observations are completely different from the simple NG mechanism advocated by Pascault et al.8,29-32 and the SD process described by Inoue.12 A schematic model describing the evolution of phase separation based on OM, AFM, and FESEM investigations is given in Figure 9. Figure 9a shows the schematic representation of immiscible SAN-graft-PB particles dispersed in a homogeneous epoxy/SAN mixture. In Figure 9b, the NG mechanism possibly initiated by SAN-graft-PB particles is depicted, which results in SAN particles with PB inclusions dispersed in the epoxy matrix. Figure 9c shows that the change from the NG mechanism to the SD mechanism results in the formation of cocontinuous structures of both the epoxy phase and the SAN phase. Obviously, the PB will be in the SAN phase. Figure 9d shows a rapid phase inversion in which the SAN phase with PB particles becomes the continuous phase and epoxy becomes the droplet. Eventually, Figure 9e represents the final morphological structures obtained at different curing temperatures. As mentioned in the Results, the final morphology of ABS/ epoxy blends cured at 150 and 165 °C is of phase-inverted type and is different from the final morphology of blends cured at 180 °C, in which the final morphology is cocontinuous. This effect of temperature on the final morphologies can be explained



Figure 8. (a) AFM micrograph of neat ABS; (b) AFM micrographs of the ABS-modified epoxy system; and (c) FESEM micrograph of the ABS-modified epoxy system.

as follows. At lower temperatures, the growth of the epoxy phase may get restricted due to the high viscosity of the system or in other words, the elastic effect of SAN may play an important role in the phase separation process. The diffusion flow of the fast dynamic epoxy phase was hampered by the slow dynamic SAN phase. This means the rate of growth of epoxy phase is controlled by the SAN elastic phase. At higher temperatures, the elastic effect of SAN will be less important and hence the epoxy domains grow faster and result in the final cocontinuous morphology, in which both the SAN phase and epoxy phase are cocontinuous. It is important to mention that as the epoxy phase begins to grow, the SAN-rich elastic phase encounters

volume shrinking. This volume shrinkage process of SAN is dominated by the transfer of the mobile epoxy phase from more elastic phase to less elastic phase. This is in agreement with the theory proposed by Tanaka et al.5,33 In other words, the SANrich phase becomes more and more viscoelastic with the escape of epoxy monomer and eventually behaves as an elastic body. It is important to mention that the mechanism of phase separation remains the same irrespective of the temperature. From the evolution of phase morphology and the resultant final morphologies at different temperatures, it was obvious that the temperature plays a prominent role in the final phase morphology. From our understanding, all dynamic asymmetric polymer

13278


Jyotishkumar et al.

Figure 9. Schematic representation of phase separation mechanism (white shade, epoxy-rich phase; yellow shade, SAN-rich phase; black shade, SAN-g-PB particles).

blend systems possibly will phase separate through the viscoelastic phase separation process as described here. The above mechanism of phase separation is further established in the coming text. The curing study by DSC proved to be a valuable tool to provide useful information on the phase separation process. The second increase in reaction rate (shoulders in Figure 4) becomes visible during the phase inversion process, where the epoxyrich phase is separated out from the continuous SAN-rich phase followed by the rapid growth of the epoxy-rich phase. The generation of new epoxy amine-rich phase with faster reaction rate followed by the subsequent generation of new phases may enhance the reaction rate. In other words, this acceleration is caused by the higher concentration of the reactive groups in the mobile epoxy amine phase that was created upon complex phase separation.34 The shoulders may also be due to a small heat effect accompanying the generation of the new epoxy phase that is expected to contribute to the heat flow signal. From the

DSC measurements, it was clear that the rate of phase separation increases with higher cure temperature. Furthermore, rheology is an important tool that provides significant information about the phase separation behavior of the blend systems. When the continuous SAN phase is generated, tan δ decreases slightly while the loss modulus and storage modulus increase. The time at which continuous SAN phase is generated is illustrated in Table 1 and is in agreement with the findings from OM and DSC. All these results clearly indicate that the rate of phase separation was influenced by temperature. For a comprehensive analysis of the second stage of phase separation by SD, blends were subjected to the Cahn-Hilliard theory for spinodal phase separation. This theory was utilized to explain the experimental observations in the early stages of SD.35 If spinodal decomposition is the mechanism of microphase separation, then the peak position qmax should remain constant in the early stages, but peak intensity should exhibit an exponential increase with time as shown in eq 1.


I(q, t) ) I(q, t ) 0) exp[2R(q)t]


(1)

In this equation, t is the phase separation time, I(q,t)0) is the scattered intensity at zero time associated with native composition fluctuations, and R(q) is the amplification rate of the composition fluctuations (rate of phase separation) and R(q) depends on q, the scattering wave vector defined by eq 2.

q ) (4Π/λ) sin(θ/2)

(2)

λ and θ are the wavelength of the light and scattering angle measured in the medium, respectively. Therefore, a plot of ln I(q) vs time should give a straight line during the early stages of SD. The amplification rate R(q) is obtained from the slope in the ln I(q) vs time graph.35,36 R(q) is described by

R(q) ) -Mq2(d2fm /dφ2 + 2kq2)

(3)

Here M, φ, k, and fm are the mobility (assumed to be q-independent), volume fraction of component of B (thermoplastic-rich phase), the gradient energy coefficient, and the mean field free energy of mixing, respectively. The above equation includes the apparent diffusion coefficient Dapp. Dapp is negative for SD, which corresponds to an uphill diffusion (diffusion from a region of lower concentration to higher concentration).

Dapp ) M(d2fm /dφ2)

(4)

Dapp can be obtained from the intercept in the R(q)/q2 vs q2 plot. Figure 10a shows the change of scattered intensity maximum (I(q)) as a function of time during the isothermal cure at 150 °C. A sharp increase in the intensity of scattered light was observed. This increase in intensity is due to spinodal phase separation. The linearized theory was applied to the early stages of SD. Figure 10b shows the plot of ln I(q) vs time which gives a nearly straight line during the early stages of phase separation. As mentioned above, the value of R(q) was calculated from the slope in the ln I(q) vs time plot. A linear relationship between R(q)/q2 and q2 as predicted from the theory of Cahn and Hilliard, has been established for this epoxy/ABS blend (Figure 10c). The data fit the linear plot, which suggests the validity of the linearized theory. The temperature-dependent diffusion parameters Dapp are given in Table 2. The magnitude of diffusion coefficient Dapp is the measure of ease with which the epoxy block diffuses: the higher the values of Dapp the less resistance the epoxy block encounters as microphase separation occurs. The negative diffusion coefficient indicates diffusion against concentration gradient; that means diffusion from a region of lower concentration to higher concentration as found in spinodal decomposition mixtures. The negative diffusion coefficient strongly supports the SD mechanism.35 It is important to mention that with increase in temperature Dapp decreases, indicating that the rate of diffusion from region of lower concentration (SAN-rich phase with epoxy minor phase) to a region of higher concentration (epoxy-rich phase) increases. The exponential decay of maximum scattering vector qm according to Maxwell-type relaxation equation was observed

Figure 10. (a) Plot of I(q) as a function of cure time for ABS/ epoxy system at 150 °C. (b) Plot of ln I(q) vs cure time for ABS/ epoxy system at 150 °C. (c) Plot of R(q)/q2 vs q2 for ABS/epoxy system at 150 °C.

upon calculation of SALLS results of ABS/epoxy blends for all experimental temperatures.17 According to this

qm(t) ) q0 + A0 exp(-t/τ)

(5)

with t ) R, qm ) q0. Here A0 is the magnifier and τ is the relaxation time of phase separation, which indicates the coarsen-

13280


Jyotishkumar et al.

TABLE 2: Apparent Diffusion Coefficients for ABS-Modified Epoxy Blends temp (°C)

Dapp (µm2 s-1)

135 140 150 160 165 170 180

-8.5 -12.7 -21.1 -24.4 -26 -29.3 -35.6

ing capability of epoxy droplets. An exponential decay of qm with respect to time at 150 °C is shown in Figure 11a. The different values of τ at different temperatures are reported in Table 3. Since the correlation coefficient R2 varied between 0.955 and 0.998, the results can be considered as realistic. The relaxation time τ indicates the growth of epoxy droplets; it may also be regarded as the viscous capability of epoxy monomer. As the temperature increases, the relaxation time decreases. This can be due to the acceleration of the relaxation moment at higher temperature or, in other words, the higher the temperature the more rapid is the epoxy/amine escape from the SAN-rich continuous phase. Therefore, the final morphology can be fixed quickly. It is important to mention that the time-dependent qm fits very well with the Maxwell-type relaxation equation for the system at all temperatures from 135 to 180 °C. This suggests that the relaxation movement of the ABS/DGEBA/DDS system is a kind of viscoelastic process.7 For further verification of viscoelastic behavior in ABSmodified epoxy/DDS system, the relaxation times vs temperature was also simulated by the Williams-Landel-Ferry (WLF) equation,37 which is given by

τ ) τ(s) exp[-ln 10 × 8.86(T - Ts)/(101.6 + (T - Ts))] (6) The WLF equation is very important for the clarification of temperature dependency in viscoelasticity. The simulation results provide a good fit to the experimental data and are shown in Figure 11b. This means that the relaxation time obeys the time-temperature superposition (TTS) principle and can be described by the WLF function. Thus, it can be suggested that the coarsening process of epoxy droplets is mainly controlled by viscoelastic flow.38 The Tg of the mixture epoxy/monomer and DDS was taken as the reference temperature (282 K). The values Ts and τ(s) obtained from the fitting are 321 K and 1.63 × 106 s for the ABS/epoxy system, respectively. The Ts may be related to the flow temperature of the epoxy polymer chains, which may also be related to the onset temperature of viscoelastic phase separation,38 which should be around 50 °C higher than the Tg of the polymer; τ(s) is the relaxation time at Ts. Thus, from the above WLF analysis, Ts and the τ(s) at different temperatures can be calculated, but it was really difficult to measure the τ(s) experimentally at low temperatures due to the poor mobility of epoxy/amine monomers (escape movement of epoxy/amine monomers), which even might be undetectable within limited observation time. Therefore, applying the above TTS principle will help us to have valuable information over the wide range of temperatures. For further verification, we have calculated the shift factor aT ) τ/τ(s) at various temperatures. 1/log(aT) vs 1/(T - Ts) has been plotted in Figure 11c. It is obvious that a linear function undoubtedly exists. This also implies that the coarsening process

Figure 11. (a) Plot of qm vs temperature for ABS/epoxy system at 150 °C; symbols correspond to the experimental data and lines correspond to the fitting with eq 5. (b) Plot of relaxation time τ (s) vs temperature for ABS/epoxy system. (c) Plot of 1/log(aT) vs 1/(T - Ts) for ABS/epoxy system.

at the late stage is mainly controlled by viscoelastic flow of epoxy monomers. 5. Conclusions The viscoelastic effects and kinetics of phase separation in 12.9 wt % ABS-modified epoxy/DDS blends at various cure temperatures were investigated in detail. The phase separation

Phase Separation of an ABS-Modified Epoxy System TABLE 3: Results Obtained from SALLS for ABS-Modified Epoxy Blends temp (°C)

τ (s)

R2

180 170 165 160 150 140 135

11.9 ( 1.4 35.5 ( 2.2 47.3 ( 2.5 49.4 ( 5.4 54.1 ( 12.2 122.9 ( 14.2 158.7 ( 15.3

0.991 0.992 0.998 0.996 0.955 0.99 0.993

from initial normal droplet matrix to final bicontinuous structure was observed with OM. A real time/conversion scale investigation was made for the blend with DSC and rheology. The results totally agree with the viscoelastic phase separation as revealed by OM. The time for the generation of new epoxy phase during phase separation process also shows a good agreement by all techniques used. From our studies, we confirm that the blend system phase separates through NG followed by SD mechanism. The second stage phase separation by SD was confirmed by Cahn-Hilliard theory for spinodal phase separation. SALLS results from monitoring the second stage phase separation show a typical decay in maximum scattering vector qm. The characteristic relaxation time of phase separation at different temperatures can be described well by the WLF equation. It is suggested that the relaxation movement during phase separation is a kind of viscoelastic flow. To conclude, a detailed investigation of the curing process, rheology, evolution of complex phase separation, and generated final morphology of the complex epoxy/ABS/DDS blends was carefully performed by employing sophisticated techniques. Acknowledgment. The authors thank Professor J. P. Pascault for enlightening discussions. References and Notes (1) Yu, Y.; Wang, M.; Gan, W.; Tao, Q.; Li, S. J. Phys. Chem. B 2004, 108, 6208–6215. (2) Goossens, S.; Goderis, B.; Groeninckx, G. Macromolecules 2006, 39, 2953–2963. (3) Bonnet, A.; Pascault, J. P.; Sautereau, H.; Taha, M.; Camberlin, Y. Macromolecules 1999, 32, 8517–8523. (4) Araki, T.; Tanaka, H. Macromolecules 2001, 34, 1953–1963. (5) Tanaka, H.; Araki, T. Phys. ReV. Lett. 1998, 81, 389–392. (6) Tanaka, H. Phys. ReV. E 1997, 56, 4451–4462. (7) Gan, W.; Yu, Y.; Wang, M.; Tao, Q.; Li, S. Macromolecules 2003, 36, 7746–7751.

J. Phys. Chem. B, Vol. 114, No. 42, 2010 13281 (8) Girard-Reydet, E.; Sautereau, H.; Pascault, J. P.; Keates, P.; Navard, P.; Thollet, G.; Vigier, G. Polymer 1998, 39, 2269–2280. (9) Ishii, Y.; Ryan, A. J. Macromolecules 2000, 33, 158–166. (10) Oyanguren, P. A.; Galante, M. J.; Andromaque, K.; Frontini, P. M.; Williams, R. J. J. Polymer 1999, 40, 5249–5255. (11) Maugey, J.; Van Nuland, T.; Navard, P. Polymer 2001, 42, 4353– 4366. (12) Inoue, T. Prog. Polym. Sci. 1995, 20, 119–153. (13) Jose, J.; Thomas, S.; Pionteck, J.; Joseph, K. J. Phys. Chem. B 2008, 112, 14793–14803. (14) Jyotishkumar, P,; Koetz, J.; Tierisch, B.; Strehmel, V.; Hassler, R.; Ozdilek, C.; Moldenaers, P.; Thomas, S. J. Phys. Chem. B 2009, 113, 5418–5430. (15) Goossens, S.; Groeninckx, G. Macromolecules 2006, 39, 8049– 8059. (16) Yu, Y.; Zhang, Z.; Gan, W.; Wang, M.; Li, S. Ind. Eng. Chem. Res. 2003, 42, 3250–3256. (17) Yu, Y.; Wang, M. H.; Foix, D.; Li, S. Ind. Eng. Chem. Res. 2008, 47, 9361–9369. (18) Thomas, R.; Durix, S.; Sinturel, C.; Omonov, T.; Goossens, S.; Groeninckx, G.; Moldenaers, P.; Thomas, S. Polymer 2007, 48, 1695–1710. (19) Bonnet, A.; Pascault, J. P.; Sautereau, H.; Camberlin, Y. Macromolecules 1999, 32, 8524–8530. (20) Izuka, A.; Winter, H. H.; Hashimoto, T. Macromolecules 2000, 6158–6165. (21) Yu, Y. F.; Wang, M. H.; Gan, W. J.; Li, S. J. Colloid Polym. Sci. 2006, 284, 1185–1190. (22) Francis, B.; Rao, V. L.; Poel, G. V.; Posada, F.; Groeninckx, G.; Ramaswamy, R.; Thomas, S. Polymer 2006, 47, 5411–5419. (23) Ritzenthaler, S.; Girard-Reydet, E.; Pascault, J. P. Polymer 2000, 41, 6375–6386. (24) Calabrese, L.; Valenza, A. Eur. Polym. J. 2003, 39, 1355–1363. (25) Ishii, Y.; Ryan, A. J. Macromolecules 2000, 33, 167–176. (26) Kim, H.; Char, K. Ind. Eng. Chem. Res. 2000, 39, 955–959. (27) Magonov, S. N.; Elings, V.; Whangbo, M.-H. Surf. Sci. 1997, 375, L385–L391. (28) Kyu, T.; Lee, J. Phys. ReV. Lett. 1996, 76, 3746–3749. (29) Oyanguren, P. A.; Frontini, P. M.; Williams, R. J. J.; Girard-Reydet, E.; Pascault, J. P. Polymer 1996, 37, 3079–3085. (30) Oyanguren, P. A.; Frontini, P. M.; Williams, R. J. J.; Vigier, G.; Pascault, J. P. Polymer 1996, 37, 3087–3092. (31) Poncet, S.; Boiteux, G.; Pascault, J. P.; Sautereau, H.; Seytre, G.; Rogozinski, J.; Kranbuehl, D. Polymer 1999, 40, 6811–6820. (32) Williams, R. J. J.; Rozenberg, B. A.; Pascault, J. P. AdV. Polym. Sci. 1997, 128, 95–156. (33) Tanaka, H. J. Phys.: Condens. Matter 2001, 13, 4637–4674. (34) Geert Vanden, P. Ph.D. Thesis, Katholieke Universiteit, Leuven, Nov 2003. (35) Boyard, N.; Vayer, M.; Sinturel, C.; Seifert, S.; Erre, R. Eur. Polym. J. 2005, 41, 1333–1341. (36) Zheng, Q.; Peng, M.; Song, Y.; Zhao, T. J. Macromolecules 2001, 34, 8483–8489. (37) Williams, M. L.; Landel, R. F.; Ferry, J. D. J. Am. Chem. Soc. 1955, 77, 3701–3707. (38) Gan, W.; Xiong, W.; Yu, Y.; Li, S. J. Appl. Polym. Sci. 2009, 114, 3158–3167.

JP101661T

Dynamics of Phase Separation in Poly(acrylonitrile ... - ACS Publications

Recommend Documents