Ind. Eng. Chem. Res. 2008, 47, 9361–9369
9361
Rheological Study of Epoxy Systems Blended with Poly(ether sulfone) of Different Molecular Weights Yingfeng Yu,*,†,‡ Minghai Wang,† David Foix,‡ and Shanjun Li† The Key Laboratory of Molecular Engineering of Polymers, Ministry of Education, Department of Macromolecular Science, Fudan UniVersity, Shanghai, 200433, China, and Departament de Quı´mica Analı´tica i Quı´mica Orga`nica, UniVersitat RoVira i Virgili, Macel.li Domingo s/n, 43007 Tarragona, Spain
In this work, the rheological behavior and structural transitions in poly(ether sulfone) (PES)-modified epoxy systems during phase separation were studied by rheometry instrument, scanning electronic microscopy, timeresolved light scattering, and differential scanning calorimetry. The molecular weight and concentration of PES have profound effects on rheological behavior of modified epoxy systems. It was found that the evolution of complex viscosity is closely connected to the molecular weight of PES. Blends with higher molecular weight PES tend to have a lower conversion at the onset of phase separation and show quicker increase of viscosity with curing process. The complex viscosity exhibits an exponential growing process during phase separation at various temperatures. The relaxation time of rheological growth depends on both molecular weight of PES and test frequency. 1. Introduction Thermosetting networks tend to have a characteristic low resistance to brittle fracture; therefore, the modification of these materials with rubbers,1-3 thermoplastics,4-8 core-shell particles,9-12 block copolymers,13-17 and liquid crystals18-23 has been a significant challenge in the past few decades. These modifiers are initially miscible with the uncured thermosetting precursors, while phase separation and/or self-organization usually take place during the curing process of thermosetting resins. Among all of the above-mentioned methods, thermoplastics are one of the most widely used modifiers in practical material toughening or process controlling in thermosetting resins applications. For example, epoxy resins modified with poly(ether sulfone) (PES), poly(ether imide), or polypropyloxide are widely used as engineering materials for aircraft, automobiles, and electronics usages. In these thermoplastic/thermosetting resin blends, the polymerization-induced phase separation during the curing process has been well reported.24-26 During isothermal curing of thermoplastic-modified thermosetting resin blends, some structural transformations are expected to occur: phase separation, vitrification of thermoplastic-rich phase, chemical gelation of thermoset, and vitrification of thermoset-rich phase. Unlike thermoplastic/thermoplastic blends, phase separation is induced by the increasing molecular weight of thermosetting resins in the thermosetting/thermoplastic blends even if the temperature is kept constant. After the occurrence of phase separation, thermosetting monomers diffuse from thermoplastic-rich phase into thermosetting-rich phase and result in the gradual increase of glass transition temperature (Tg) of thermoplastic-rich phase. When the Tg of thermoplastic-rich phase reaches the curing temperature, the vitrification of thermoplastic-rich phase occurs and the phase structure is fixed. Therefore, study on the rheological characterization of thermoplastic-modified thermoset is of crucial importance in industry * To whom correspondence should be addressed. Tel.: +86-2165642865. Fax: +86-21-65640293. E-mail:
[email protected]. † Fudan University. ‡ Universitat Rovira i Virgili.
applications, since it can allow the estimation of the optimal process according to which the blends should be heated or molded. The polymerization-induced phase separation affects the rheological behaviors in great deal. In blends with thermoplasticdispersed structure, Kim and Char27 noted that there is a fluctuation in viscosity just before the abrupt viscosity increase, which is believed to be due to the phase separation of thermoplastic from the thermoset matrix. Bonnet et al.28 showed large interdependence between morphology and viscosity; a gradual increase in viscosity was observed in inverse phase structure systems at higher thermoplastic concentrations. Tercjak et al.29 investigated the influence of the syndiotactic polystyrene concentration on the dynamic rheological properties of the samples. Meanwhile, the interplay between phase separation and gelation is an important parameter controlling the final morphology and material performance.30,31 In our previous studies, we found that phase separation has an important effect on their rheological behaviors, as both the structural transitions during the curing process and viscosity evolutions connected closely to phase separation.32,33 While it is well known that commercial thermosetting systems in industrial applications are always modified with thermoplastics of different molecular weights (and polydispersity), and the thermosetting resins are also with different molecular weight or curing conversion, for example, epoxy prepregs are always used as B-stage epoxy resin blends. Therefore, one would expect large differences in processing or structure controlling in these systems because of the diversion of thermoplastics and thermosetting resins. However, even though thermoplastic-modified epoxy resin is one of the well-studied systems in both morphology and curing kinetics, the influences of molecular weight on the rheological behaviors were only studied in some specific systems; for example, Recca et al. studied the effect of molecular weight of block copolymers on rheological behaviors.34 Our work in this article will be focused on the influences of molecular weight on the rheological behaviors and gelation transitions in a PES-modified epoxy system during phase separation, and the rheological measurements were combined
10.1021/ie800845p CCC: $40.75 2008 American Chemical Society Published on Web 11/01/2008
9362 Ind. Eng. Chem. Res., Vol. 47, No. 23, 2008 Table 1. Compositions of Epoxy/Poly(ether sulfone) Blends blends neat epoxy PES (0.22)-14% PES (0.36)-14% PES (0.53)-14% PES (0.22)-20% PES (0.36)-20% PES (0.53)-20%
PES (0.22)
PES (0.36)
PES (0.53)
30 30 30 45 45 45
DER331
MTHPA
100 100 100 100 100 100 100
80 80 80 80 80 80 80
with time-resolved light scattering (TRLS) and differential scanning calorimetry (DSC).
blends were tested under a parallel plate mode with a controlled strain of 1% to ensure that measurements were performed under linear viscoelastic conditions. The test’s multifrequencies were set at 1, 2, 5, 10, and 20 rad/s. 2.5. Time-Resolved Light Scattering. The phase separation process during isothermal curing reaction was observed in situ by a self-made TRLS with controllable hot chamber. The changes of the light scattering profiles were recorded at appropriate time intervals during isothermal curing. The blend of epoxy monomer with PES for TRLS observation was prepared by melt-pressing the film. 3. Results and Discussion
2. Experimental Section 2.1. Materials and Samples Preparation. A low molecular weight liquid diglycidyl ether of bisphenol A (DGEBA) type epoxy resin DER 331 was provided by Dow Chemical Co., and the epoxide equivalent of DER 331 is 182-192 g/equiv. The curing agent is methyl tetrahydrophthalic anhydride (MTHPA), LHY 908, provided by Ciba-Geigy Co. Three kinds of PES with some hydroxyl end groups as specified by the manufacturer, supplied by Jilin University (Jilin, China), were used [intrinsic viscosity ) 0.22 dm3/kg (weight-average molecular weight ) 2.3 × 104, polydispersity ) 2.5), intrinsic viscosity ) 0.36 dm3/ kg (weight-average molecular weight ) 4.5 × 104, polydispersity ) 2.4), and intrinsic viscosity ) 0.53 dm3/kg (weightaverage molecular weight ) 6.7 × 104, polydispersity ) 2.2)]. The intrinsic viscosity was determined in N,N-dimethylformamide at 25 °C. The molecular weight of PES was measured with a PerkinElmer S-250 gel permeation chromatograph; polystyrene standards (Showa Denko, Ltd.) were employed to make the calibration curve. The homogeneous mixture of PES/DGEBA was prepared by adding PES to the stirring epoxy monomer at 150 °C for 4 h under nitrogen gas. After the mixture had cooled to 80 °C, MTHPA was added and stirred vigorously for 2 min until MTHPA was completely dissolved. The samples were degassed under vacuum for a few minutes. The weight composition and the code names of the blends in this article are listed in Table 1, in which the intrinsic viscosities of PES are in parentheses after PES, and the weight percentages of PES in blends are listed last. For example, PES (0.22)-14% means the PES was used with an intrinsic viscosity of 0.22, and the weight percentage of PES was 14.286% in fact, and herein is denoted as 14%. The cured samples with different curing time were dissolved in methylene chloride (CH2Cl2). The chemical gel time, tcg, was the time when the presence of an insoluble fraction of the blends was first observed.35 2.2. Scanning Electronic Microscopes. Philip XL 30 scanning electronic microscope (SEM) was employed to examine the morphology of the fracture surface of cured specimens. 2.3. Differential Scanning Calorimetry. A PerkinElmer Pyris 1 DSC instrument was used for the study of the curing reaction. The isothermal curing conversion was calculated from residual exotherms observed in scans in the temperature range of 50-350 °C, with heating rates of 10 °C/min and normalized by the total exotherms for uncured samples. 2.4. Rheological Measurements. The melt viscosity variations of the blends during curing reaction were recorded on an ARES-9A rheometry instrument. About 1 g of the blend was sandwiched between two round plates with diameter of 40 mm and softened at 80 °C for 2 min. The plate distance was then adjusted to 1.0 mm, and the temperature was raised quickly at a rate of 100 °C/min to the preset curing temperature. All the
3.1. Relationship between Phase Separation and Molecular Weight. Morphology. It is well known that processing, mechanical, thermal, and other properties of blends are closely related to their morphology. Therefore, the fracture surfaces of the blends were examined by SEM to evaluate the phase morphologies of the blends cured at 150 °C for 5 h. The morphology changed drastically depending on the PES content and molecular weight. Figure 1a-c show the typical morphologies of blends with different molecular weight PES at 14 wt %. For low molecular weight PES, the blend PES (0.22)-14% showed spherical domains around 1 µm in diameter dispersed in the epoxy-rich matrix. A small number of PES particles broke off the matrix because only a small amount of the hydroxyl groups in PES had reacted with the epoxy/ anhydride and there were few chemical links between the two phases. In the case of intermediate molecular weight PES, the morphology of the blend PES (0.36)-14% was a bicontinuous phase structure. Small PES particles could be seen in the epoxy continuous phase; meanwhile, spherical particles of epoxy could also be found dispersed in the PES continuous phase. However, the size distribution of the epoxy particles in the PES continuous phase was not uniform, probably because of the high viscosity at the late stage of the curing reaction, which restricted the further development of the microstructure. Moreover, for high molecular weight PES, the blend PES (0.53)-14% appears more cocontinuous in nature, rather than a phase-inverted structure. Meanwhile, the PES phase formed the matrix, and epoxy appeared as big, interconnected spherical domains ca. 10 µm in diameter. Figure 1d-f show the effects of molecular weight on the morphologies of blends with PES at 20 wt %. All three blends show a phase inversion structure. PES (0.22)-20%, similar to PES(0.53)-14%, forms a spongelike (bicontinuous) phase structure, although large epoxy spheres with diameters about 10-30 µm were found in the PES-rich membranelike structures. For PES (0.36)-20% and PES (0.53)-20%, totally phase-inverted structures were observed and the diameter of epoxy-rich globules decreased with the increase of PES molecular weight. This could be attributed to strong viscoelastic effects, and we will discuss it in detail later in the part of rheological study. Curing Conversion and Phase Separation. The curing study would provide useful information for understanding the phase separation. Isothermal DSC curves of neat epoxy DER331/ MTHPA and blends with different molecular weight and weight percentage (for comparison, we show PES (0.53)-20% as an example here) of PES cured at 150 °C are shown in Figure 2. Neat epoxy-anhydride systems react very slowly, and reaction occurs due to partial hydrolysis of anhydride leading to an epoxy-acid stepwise reaction, in which the mole ratio of anhydride to epoxy group needed for fully curing is usually ca. 0.85; for comparison, we just use a ratio of 1 here in case of
Ind. Eng. Chem. Res., Vol. 47, No. 23, 2008 9363
Figure 1. Morphologies of the modified epoxy systems with different molecular weights and contents of PES after being cured at 150 °C for 5 h. (a) PES (0.22)-14%. (b) PES (0.36)-14%. (c) PES (0.53)-14%. (d) PES (0.22)-20%. (e) PES (0.36)-20%. (f) PES (0.53)-20%.
any effects of PES on the curing. The introduction of PES increased sharply the curing rate and final conversion. The role of the hydroxyl-terminated PES, according to several researchers, is to co-initiate the chainwise reaction, thus increasing the curing rate compared to that of the neat epoxy/anhydride stepwise reaction.36,37 In addition, the effect of PES molecular weight on the curing process of the epoxy/PES blends was studied. Figure 2 also shows that the curing conversion of epoxy was affected by the PES molecular weight. However, the weight percentage and molecular weight of PES has little effect on the curing rate and conversion. The curing rate of the blend with lower molecular weight PES was slightly slower than the blend with a higher one. The reason might be that the phase separation of the blend with higher molecular weight PES was more pronounced; that is, in comparison to the blend with low molecular weight PES, the PES concentration in the epoxy-rich phase was lower in
the blend with higher molecular weight PES-modified systems during phase separation, and this reduced the dilution effect.38 As proved in our previous work, the epoxy/PES system follows viscoelastic phase separation behavior. Phase separation processes of all the blends were traced in situ by TRLS. Changes in the light scattering profiles of the blends with various PES contents were recorded at appropriate time intervals during isothermal curing. TRLS was generally used for observing phase separation via spinodal decomposition (SD) mechanism, in which profiles of ring pattern formed by light scattering through the regular phase structure, while for nuclear and growth (ND) phase separation, only light intensity slowly changing was observed by TRLS because of the irregular sphere structures from ND mechanism. The results of TRLS clearly displayed that the phase separation of all the epoxy/PES 20 wt % systems took place according to an SD mechanism. Figure 3 shows the typical examples of the
9364 Ind. Eng. Chem. Res., Vol. 47, No. 23, 2008 Table 2. Curing Conversions of Modified Epoxy Systems with Different Molecular Weight PES during Phase Separation and Gelation at 150 °Ca blends PES PES PES PES PES PES
(0.22)-14% (0.36)-14% (0.53)-14% (0.22)-20% (0.36)-20% (0.53)-20%
x-CP
x-end
x-gel
0.17 ( 0.02 0.13 ( 0.02 0.08 ( 0.01 0.18 ( 0.02 0.12 ( 0.02 0.09 ( 0.01
0.33 ( 0.03 0.29 ( 0.02 0.15 ( 0.01 0.34 ( 0.03 0.30 ( 0.02 0.16 ( 0.01
0.54 ( 0.02 0.52 ( 0.02 0.53 ( 0.02 0.56 ( 0.02 0.54 ( 0.02 0.52 ( 0.02
a x-CP: curing conversion at the onset of phase separation observed by TRLS. x-End: curing conversion when qm remains constant observed by TRLS.
Figure 2. Curing conversion versus curing time for PES/epoxy blends cured at 150 °C with PES samples of different molecular weights and different concentrations.
of the curing reaction (normally, the cure conversion was less than 0.2), it can be concluded that the low conversion had few effects on the phase separation. For the step-growth polymerization reaction, the systems could be simplified as a pseudobinary blend: PES and growing thermoset. At low conversion (before gelation), the molecular weight of the growing thermoset determined by GPC followed a Poisson distribution and increased with conversion.39 As a result, the growing thermoset could be simplified as a polymeric component with an average polymerization degree of ngt and a volume fraction of φgt in terms of the Flory-Huggins theory,40 and the mixing free energy F of the blend can be given by: F ) kT
r
∑ 1
φi ln φi + ni
r
r
i