Separation of Gelation from Vitrification in Curing of a Fiber-Reinforced Epoxy Composite
Total Page:16
File Type:pdf, Size:1020Kb
Separation of Gelation From Vitrification in Curing of a Fiber-Reinforced Epoxy Composite BRYAN BILYEU and WITOLD BROSTOW Laboratory of Advanced Polymers and Optimized Materials (LAPOM) Department of Materials Science Unwerslty of North Texas P. 0. Box 3053 1 0 Denton, ?x 76203-531 0 and KEVIN P. MENARD Perkin Elmer Instruments 761 MainAve. F71 Nowalk, CTO6987 Prepregs of a mixture of the tetrafunctional epoxy tetraglycidyl 4.4-diamin- odiphenyl methane (TGDDM) and the tetrafunctional amine 4.4'-diaminodiphenyl- sulfone (DDS) were characterized with temperature-modulated DSC (TMDSC) as well as dynamic mechanical analysis (DMA). The baseline shift of the glass transi- tion was separated from the curing exotherm by using temperature-modulated and step scan DSC temperature scans. Likewise, the baseline shift in heat capacity due to vitrification was isolated using TMDSC isotherms. Using the TMDSC glass tran- sition temperature, degree of conversion, and vitrification results, combined with the gelation data generated from DMA, a time-temperature-transformation (TIT)di- agram was constructed, providing information necessary for optimization of indus- trial processing of the epoxy prepreg. Thus, effects of storage, preprocessing, and postprocessing on the overall curing process are taken into account. 1. INTRODUCTION appears as a change in physical properties, it is con- veniently determined by DMA. Vitrification a rub- iber-reinforced epoxy prepregs are commonly as ber to in processed using isothermal curing. Although pro- glass transition appears both DMA and DSC. F Although vitrification is commonly determined by cessing appears simple, the development of properties DMA, DSC offers increased temperature accuracy and during curing is a complex multistep process. In addi- control. However, since vitrification occurs before tion, the relationship between the curing temperature complete conversion, the is usually masked by the and time to a specific conversion value is not linear. Tg curing exotherm in traditional DSC. Since vitrification Industrial processing control is simplified using a time-temperature-transformation (ITT) diagram. and curing are thermodynamically different effects, one should be able to separate the signals by However, construction of such a TIT diagram requires two knowledge of the rate and degree of conversion as well using either temperature modulated DSC (TMDSC) or as the time to reach gelation and vitrification at each step scan DSC. isotherm. The rate and degree of conversion is deter- mined either with a series of DSC isotherms, which 2. EPOXY CURE CHARACTERIZATION show the change of enthalpy in time or with the OPTIONS change in the glass transition temperature Tg during To obtain the most useful and reliable characteriza- the isothermal cure. Accuracy and resolution of Tg tion of the curing process we need to consider in turn measurements by DSC are improved by using heat the options available. Characterization of the epoxy cur- capacity C, rather than enthalpy. Since gelation ing reaction and generation of a ??T diagram requires POLYMER COMPOSITES, DECEMBER 2002, Vol. 23, No. 6 1111 Bryan Bilyeu, Witold Brostow, and Kevin P. Menard knowledge of the rate and degree of cure as a function a C, value calculated as the ratio of the heat flow am- of time and temperature as well as the conditions plitude to the heating rate amplitude, the complex which produce gelation, vitrification, full cure and heat capacity (C,’ = Ah/&). The complex heat capaci- degradation. The rate and degree of cure is typically ty (C,’ ) is separated into two components, real and tracked by a change in a cure-dependent property, imaginq: such as the glass transition temperature Tg Gelation C,’ = C,’ ic,” can be determined by examining the molecular + weight-dependent properties. Since vitrification is a with C,’ representing the real, in-phase component, rubber-to-glass transition, it results in a thermody- termed storage heat capacity, and C; representing namic as well as a physical change. A detailed review the imaginary, out-of-phase component, termed loss of thermal analysis techniques applicable to epoxy heat capacity. The two values are calculated using the characterization is available (1)as well as an extensive phase angle, cp, between the signal and response: study using a variety of techniques to fully character- = c,*c0scp ize an epoxy system (2). c,’ Many thermosetting polymer systems exhibit a rela- cp” = C,*Sincp tionship between the Tg and the degree of chemical (4) conversion. Most epoxy-amine systems exhibit a lin- As in DMA measurements, the storage signal will ear relationship, which implies that the change in include the elastic or in-phase response of the materi- molecular structure with conversion is independent of al, which in this case represent the molecular level re- the cure temperature (3). sponses, including glass transitions and melting. The The most convenient and generally most accurate loss signal represents viscous or out-of-phase events, method for determining the Tg of polymers is differ- which are the kinetic effects, such as stress relaxation ential scanning calorimetry (DSC). The Tg is taken as and curing. The total heat capacity is the non-sepa- the temperature at the inflection point (peak of deriva- rated average signal, which is equivalent to the heat tive curve) of the baseline shift in heat flow or as the capacity signal produced by traditional DSC. In epoxy temperature at the half height shift in baseline heat Tg shifts, this allows separation of the Tg from the flow. The shift in baseline heat flow associated with exotherm. the glass transition is a result of the difference in heat Gelation represents a change in mechanical proper- capacity between the rubber and the glass. Since this ties, but typically not a change in conversion rate. shift is an effect of the heat capacity change, resolu- Thus, gelation does not appear in calorimetric mea- tion of the glass transition can be increased by calcu- surements. However, as already pointed out by one of lating and plotting the constant pressure heat capaci- us (8).it does appear prominently in DMA. ty, c,. Although vitrification is a thermal transition from a The versatility of DSC to measure both exotherms rubber to a glass and does appear in DSC measure- and Tgs is also a limitation; when measuring an un- ments (9).the determination of the point and quantifi- cured or a partially cured thermoset, a residual cation of the shift in baseline heat flow or C, usually exotherm follows the Tg being measured, sometimes occurs around the end of the curing and as such is even overlapping. Accurate Tg calculations require usually masked by the curing reaction exotherm. This stable baselines before and after the transition and is one of the clearest applications of TMDSC since the the curing exotherm interferes with the upper base- curing exotherm appears in the loss C, and the vitrifi- line. In these cases, the Tg can only be determined as cation appears in the storage C, (10-12). the onset. The first epoxy ‘IT diagram, proposed by Gillham One alternative to the measurement of C, is temper- and Ems (13). was constructed from torsional braid ature-modulated DSC (TMDSC). TMDSC utilizes a analysis (TBA), a torsional DMA measurement. DMA modulated temperature ramp. The basis for the mod- has been used extensively to investigate the vitrifica- ulation signals and evaluation, including the phase tion point, and continues to be the most common lag, is derived from electrical signal modulation in the method. However, as determined earlier by one of us electronics and telecommunications field. Analogous (141, DMA produces higher Tg values than DSC be- to dynamic mechanical analysis (DMA), TMDSC cause of the measurement of extrinsic mechanical mathematically deconvolutes the response into two properties rather than intrinsic heat capacity and the types of signals, an in-phase and an out-of-phase re- poorer temperature control of the instrument. DSC sponse to the modulations, as well as producing an could produce more accurate and meaningful Tg and average heat flow. vitrification points-if these transitions were separat- A Fourier transform deconvolutes the signal to pro- ed from the curing exotherm. Thus, we turn to duce an average heating rate (%”). an average heat TMDSC. Can TMDSC isolate the Tg and vitrification flow (h& an amplitude of heating rate (4).an ampli- points? tude of heat flow (AJ and a phase angle between the In other words, on one hand we have several as- heating rate and heat flow (cp)(4). Schawe’s technique pects of curing that all must be determined. On the (5-7)uses a linear response approach, which begins other hand, we have several techniques at our dispos- with the total heat capacity (CpT= hav/G)and includes al. We need to define a combination of techniques that 1112 POLYMER COMPOSITES, DECEMBER 2002, Vol. 23, No. 6 Separation of Gelation From Vitn@ation will provide sufficient, reliable and if possible fast 4. GLASS TRANSITION TEMPERATURES characterization of the curing process. The Tg shift in time for specific isothermal curing temperatures was determined by TMDSC (Fig. I), Step 3. EXPERTMENTAL Scan DSC (Fig. 2), standard DSC and DMA. TMDSC Hercules (Hexcel) 8552 neat resin and glass fiber- and Step Scan DSC produces similar values of glass reinforced prepregs, which are a mixture of the tetra- transition temperature as a function of curing time functional epoxy tetraglycidyl 4,4-diaminodiphenyl and temperature. We chose to use TMDSC rather methane (TGDDM) and the tetrafunctional amine 4,4'- than Step Scan DSC because TMDSC was also used diaminodiphenylsulfone (DDS), along with an ionic for the pseudo-isothermal measurements of vitrifica- initiator/accelerator and a thermoplastic modifier tion. These values were consistent with standard DSC were studied. The fiber-reinforced prepregs contain 66 and lower than DMA results. However, the DSC val- weight percent unidirectional glass fiber.