Cure and Reaction Kinetics of an Anhydride-Cured Epoxy Resin Catalyzed by N-benzylpyrazinium Salts Using Near-Infrared Spectroscopy.
GUN-HO KWAK [+]
MASAO SUMITA [+]
The cure behavior and the reaction kinetics of an anhydride-cured epoxy resin catalyzed by cationic thermal latent initiator, N-benzylpyrazinium salts, have been Investigated by a near-infrared (NIR) spectroscopy. The spectral changes are interpreted in terms of the mechanism of cure. NIR absorption bands are due to protons connected to carbon, nitrogen, and oxygen. In this work, the homogeneous model system involves a simple addition reaction mechanism leading to an exothermic reaction between epoxide and anhydride activated groups. A comprehensive account of the origin, location, and shifts during reaction of all major NIR absorption peaks in the spectral range from 4000 to 7100 [cm.sup.1] is provided. The extent of reaction is calculated from NIR absorption band at 4530 [cm.sup.-1], which depends on epoxide concentration. The utility of NIR spectroscopy to study the kinetics of epoxy cure reaction has been established. Consequently, absorbencies in the NIR spectra suitable for quantitative studies of epoxy resin reaction kinetics have been identified.
Epoxy resins are used extensively as high performance polymers in composite materials, protective coatings, and encapsulation materials in the electronic industries . Usually, amines or diacid anhydrides serve as hardeners, yielding mostly networks of alternating copolymers. Anhydride-cured epoxies are one of the thermosetting resins that are used as the matrix for many advanced materials used in the aerospace and electronic industries, and generally exhibit more improved high-temperature stability compared as amines and good physical and electrical properties .
However, it is known that in the absence of catalyst the anhydrides do not react directly with the epoxy group. Accordingly, tertiary amines have been widely used as catalyst with glycidyl ether resins, ring-saturated epoxy resins, and resins with internal epoxy groups (1). So, anhydrides react with a hydroxyl forming a carboxylic acid, which in turn reacts with an oxirane, re-forming the hydroxyl moiety (3).
Recently, N-benzylpyrazinium salts have been shown to be excellent thermal latent initiators for epoxy resins (4). These initiators are not hygroscopic, dissolve readily in epoxy resins and exhibit a long pot-life than the more commonly used [BF.sub.3]-4-methoxyaniline complex (5). In this system, the initiator activity of N-benzylpyrazinium salts can be enhanced by decreasing the nucleophilicity of the counterion, [SbF.sub.6], introducing an electron-donating substituent on the phenyl ring of the benzyl group (6).
Generally, the mechanism and kinetics for curing anhydride-epoxy resins with catalyst are complex due to involve epoxy-catalyst, anhydride-catalyst, hydroxylanhydride. hydroxyl-epoxy and side reactions (7). These reactions occurring during the crosslinking of the resin change the properties of the epoxy network (8).
The mechanism governing the cure and the kinetics have been investigated by numerous techniques such as FT-IR and NMR (9), thermal analysis (10). dynamic mechanical analysis (11), and dielectric and rheological monitoring (12). FT-JR spectroscopy is an important tool for monitoring cure reactions of epoxy resins with hardeners. The mid-infrared (MIR) region of the spectrum typically has been utilized in such studies (13), but sometimes the sample preparation may be relatively time consuming, in particular when the thickness of the sample has to be reduced, or when special reflection accessories are used. FT-IR analysis of polymeric materials provides highly precise measurements that are widely interpretable in terms of chemical structure. Yet, the occurrence of overlapping bands for epoxy resin renders the quantitative analysis in the region MIR particularly difficult (9).
Near-infrared spectroscopy (NIR) provides sample analysis with minimal technical effort and is a simple and reliable routine on-line measurement (14). The spectra often have well-isolated absorption bands for some functional groups. This is a consequence of the overtone and combination bands observed in near infrared region. These absorption bands have different intensities depending on the anharmonization of vibrations due to hydrogen atoms including C-H, N-H, C-O and O-H (7, 15). NIR spectroscopy presents several potential merits over MIR region for characterizing the cure behaviors of epoxy systems, such as higher energy sources, enhanced resolution, and lower oscillator strength (16). For the process and quality control measurements, NIR spectroscopy, with the development of new, rugged, reliable techniques, gives the possibility of rapid, non-destructive multicomponent analysis. NIR spectroscopy is therefore a useful technique to investigate polymeric materials and to study the kinetics and cure behavio rs of various epoxy resins cured with anhydride hardener (17).
In this paper, cure behavior and kinetics for epoxy resins cured with an anhydride hardener and catalyzed by N-benzylpyrazinium salts are studied using NIR spectroscopy.
Materials and Sample Preparation
In the system studied, the epoxy system consisted of commercial diglycidyl ether of bisphenol A (DGEBA, YD-128 supplied by Kukdo Chem. Co. of Korea, E.E.W = 187 g/mol) with methylbicyclo-[2,2,l]heptene-2,3-dicarboxylic acid anhydride (NMA, supplied by Aldrich Chem. Co.) as a curing agent. N-benzylpyrazinium hexafluoroantimonate (BPH) for the thermal latent catalyst was synthesized throughout the recent work (4). Curing agents and catalyst are shown in Fig. 1.
The 1 wt% catalyst was added to DGEBA and mixed with NMA at levels of 25, 45, 65, 85, and 105 wt%, where the stoichiometric ratio is 45 wt%, as listed in Table 1. These mixtures were stirred continuously until clear and were degassed under a vacuum for 1 h to remove voids and residual organic solvents.
Thermal Latent Properties
Latent properties of this system were performed by the investigation of non-isothermal DSC and conversion as a function of reaction time using isothermal DSC at 200[degrees]C, 120[degrees]C, 110[degrees]C, and 50[degrees]C. Thermal latent behaviors were carried out using DuPont DSC910 supported by a DuPont thermal analyzer. The heating rate was fixed to 10[degrees]C/min.
In-situ NIR Spectroscopy
The instrument of choice was a Perstorp Analytical NIR System 6500 infrared spectrometer equipped with a lead sulfide detector. The NIR spectra of liquid state sample were collected using an optic fiber probe. An optic fiber bundle consisting of 316 silica fibers was used to the sample preparation. The spectrometer was operated in the NIR region from 4000 to 7100 [cm.sup.-1].
Kinetic characterization of the reactive resin depends on the cure temperature. So, a direct comparison of the extent of reaction at ll0[degrees]C and 120[degrees]C for 110 min was measured by NIR spectra (18). On the other hand, the cure cycle of a fully cured epoxy system as a reference heat of reaction was 70[degrees]C for 30 min, 140[degrees]C for 2 h, and finally postcured at 200[degrees]C for 1 h.
RESULTS AND DISCUSSION
Addition of BPH used as a catalyst in this system shows a fast reverse reaction: therefore, the low stability of benzyl cation limits the initiator activity in some extents of initiation temperature (19). Consequently, these reasons explain the latent properties of this epoxy system.
Figure 2 shows the non-isothermal DSC experiments as a reference peak for the calculation of conversion. The conversion as a function of the curing time for the different temperatures of the system made with 1 wt% BPH catalyst by isothermal DSC is shown in Fig. 3. As the curing time proceeds, the conversion increases with increasing reaction temperature in the case of 110[degrees]C, 120[degrees]C, and 200[degrees]C (20). Meanwhile, the conversion in the 50[degrees]C reaction temperature shows no significant change as the reaction time increases or as the NMA composition increases. Therefore, this is due to the low temperature that has become a limiting factor prohibiting the activity of catalyst. This indicates that the systems including BPH have good thermostable latent properties at a given temperature condition in spite of the presence of a limited external heat stimulation.
Monitoring DGEBA/NMA/BPH Cure
It is generally accepted that the curing of thermosetting resins involves physical and chemical changes leading to a network formation. Although many approaches have been used to study the cured resin, the characterization of a network formation is not fully understood. We expect that NIR spectroscopy can contribute to the understanding of epoxy cure behaviors.
The NIR spectrum contains information on constitution, conformation, crystallinity, intermolecular and intramolecular interactions, as well as thermal and mechanical treatments of the polymers. Pointing out the entire interpretation of MR spectra, the correlation with chemical species is very complex and difficult. Nevertheless, some absorption bands are correlated with functional groups. This allows us to measure the appearance or disappearance of species during the cure.
NIR major absorption peaks observed in this work, their molecular origin, and some relevant comments regarding their utility are listed in Table 2. The assignments of the bands generally agree with those earlier reported in the literature (21-23).
Figure 4 shows the spectrum of uncured DGEBA/NMA/BPH resins with different anhydride ratios, and Figs. 5 and 6 illustrate the selected NIR spectra of DGEBA/NMA/BPH systems in the stoichiometric ratio during 110[degrees]C and 120[degrees]C cure, respectively.
As experimental results, the region from 4000 to 4900 [cm.sup.-1] is the fingerprint section that contains specific information about the chemical structure of the resin: therefore the section is useful for identification purposes. This zone contains the conjugated epoxy [CH.sub.2] deformation band at 4530 [cm.sup.1], whose absorption intensity decreases systematically during reaction and can therefore be used in the kinetic studied (24), as shown in Figs. 5 and 6.
The combination bands of aromatic at 4623-4675 [cm.sup.-1], which are believed not to change during the cure reaction, therefore can be chosen as internal reference bands (18, 21). The absorption at 4850 [cm.sup.-1] is a representation of the first C=O stretching overtone of anhydride, which peak increases with increasing anhydride content, as shown in Fig. 4.
In this work, three important absorption bands at 5000-6000 [cm.sup.-1] can be considered. The absorption at 5225 [cm.sup.-1] is a representation of hydroxyl-water interactions and can be used for the evaluation of moisture content (25). As the content of anhydrides increases, the peak systematically decreases, as shown in Fig. 3. The water absorption characteristics of epoxy resins have been used as an indicator of the degree of structural packing of the epoxy network on the basis of the reported water absorption mechanism (26). Another interesting peak is due to the second C=O overtone of anhydride, which is found in the range between 5300 and 5500 [cm.sup.-1] The overtone bands are not changed during cure, and therefore the peaks depend only on the content of anhydrides that do not react. Thus, the peaks gradually increase as a function of anhydride composition, as shown in Fig. 4, but on the other hand, there is no change in the case of Figs. 5 and 6. The first aromatic C-H overtone band is located at 5980 [cm.sup.-1]. This band also does not involve the cure reaction as we confirmed in NIR spectra.
An important peak appears at 6075 [cm.sup.-1] It has been assigned to the first overtone of terminal methylene fundamental stretching vibration.
Kinetic characterization of the reactive resin is important for better understanding of the structureproperty relationships and is fundamental for optimizing the process conditions.
The use of NIR's diffuse reflectance for quantitative analysis is now widely accepted (25-27). Various mathematical models of calibration standards have been developed to correlate concentration to spectroscopic data, usually expressed in the common logarithm of the inversion of reflectance (log 1/R), for each wavelength studied (28). If the sample absorbs at the same wavelengths, log (1/R) appears to be the best method to relate reflectance to concentration (29).
R = [I.sub.s]/[I.sub.r] (1)
where R is the reflectance, [I.sub.s] and [I.sub.r] the intensity of sample and reference, respectively.
Terminal epoxide has a strong NIR absorbance band at 4530 [cm.sup.-1], which has successfully quantified epoxy resin cure reactions (24).
Figures 5 and 6 show the disappearance of the peak as reaction time increases. The first overtone of the hydroxyl group stretching is found at 7000 [cm.sup.-1]. As shown below, the use of this band to monitor reaction kinetics in terms of the appearance of hydroxyl group has successfully quantified epoxy cure reaction (27).
Figures 7 and 8 show the conversion of epoxy and hydroxyl groups were monitored by changes in the intensity of each band at 4530 [cm.sup.-1] and 7000 [cm.sup.-1], respectively measured at 110[degrees]C and 120[degrees]C. The time dependence of the band intensity for both systems exhibits a sigmoidal-shape consistent with the autocatalytic reaction mechanism (28). The hydroxyl groups are found as a result of the cleavage of the oxirane rings by the catalyst or activated sites. In addition to the hydroxyl groups attached to the backbone of the epoxy oligomer, more hydroxyl groups are produced by ring opening during the cure. Thus, a number of hydroxyl groups undergo nucleophilic addition reactions with epoxide and anhydride, resulting in ring cleavage and the formation of increased crosslinking networks (26-30). However, quantitative analysis of the hydroxyl absorption band is complicated by the formation of hydrogen bonding, which occurs both intermolecularly and intramolecularly (28).
The role played by hydrogen bonds is important for physical, chemical, and mechanical properties of polymers containing hydroxyl, urethane or amide functional groups. It is also well known that hydrogen bond shifts the baseline significantly (31). Figures 9 and 10 show the baseline shift of this system in the stoichiometric ratio, measured at- 110[degrees]C and 120[degrees]C, respectively. The intensities and integrated area of the epoxy hydroxyl absorption are determined to follow the mechanisms of the reaction and epoxy conversions (17, 18, 23). Thus, kinetics of epoxy cure reactions are widely studied using NIR spectroscopy (25, 28).
As demonstrated in the reference above, the band at 4530 [cm.sup.-1] is due to the combination of conjugated epoxy-[CH.sub.2] deformation band, and the decrease in intensity is due to the reaction of epoxide, which is the dominant reaction in the early stages of the cure process (32). If the decrease in the integrated area under the band is followed with reaction time, the degree of conversion of epoxide over time can be determined using Eq 2.
[[alpha].sub.EP] = [A.sub.o] - [A.sub.t]/[A.sub.o] - [A.sub.[infinity]] (2)
where [[alpha].sub.EP] is the degree of conversion and [A.sub.o], [A.sub.t],and [A.sub.[infinity]] the areas at time zero, t and after completion of the cure cycle, respectively.
The extent of reaction at 110[degrees]C and 120[degrees]C calculated from Eq 2 is plotted as a function of reaction time, as seen in Fig. 11. When the kinetic results of this system are compared, the kinetics of reaction group in NIR spectra are found to different conversions and in reproducible trends. The two curves of extent of reaction for a given temperature appear horizontally shifted with respect to one another, but retain similar slopes between approximately 20% and 50% conversions. These results show good agreement of kinetic behavior measured with the isothermal DSC scan. It is seen that the autocatalytic nature of the two techniques is apparent. As a result, we know that the cure reactions of epoxy resins are highly exothermic in nature and the cure kinetics is dependent on the temperature, resulting in the activity of catalyst.
The current research provides a comprehensive, kinetic analysis of an anhydride-epoxy resin constrained to intermolecular reactions. The kinetic reaction analysis incorporates the reaction mechanism described by Antoon and Koenig (33) and yields theoretical descriptions for the rate of monomer decay and the distribution of polymeric molecules at the time observation.
The reaction involves the attack of hydroxyl groups, which occurs as substituent on a fraction of the epoxy resin, on the anhydride molecules to form ester linkage and carboxyl group. The latter then reacts with the epoxide ring, again forming an ester linkage and another hydroxyl group.
In this work, cure reaction for epoxy resin cured with an anhydride hardener and catalyzed by a BPH are studied using the NIR reflection spectroscopy.
The latent properties of this system are evaluated by using non-isothermal and isothermal DSC. For DGEBA/NMA/BPH system, the cure process can be monitored by following the changes in the absorption intensity of the terminal epoxy groups at 4530 [cm.sup.-1] Characteristic NIR band assignments are identified and used in evaluation of reaction kinetics. NIR spectroscopy is shown to be an attractive method for qualitative and quantitative evaluations of the epoxy cure behaviors.
(+.) Department of Organic and Polymer Materials Tokyo Institute of Technology 2-12-1 Ookayama, Meguro-ku, Tokyo 152, Japan
(*.) Corresponding author. E-mail: email@example.com
(1.) R. S. Bauer, Epoxy Resin Chemistry, American Chemical Society, Washington, D.C. (1979).
(2.) H. Lee and K. Neville, Handbook of Epoxy Resins, McGraw-Hill, New York (1967).
(3.) J. A. Nielsen, S. J. Chen, and D. C. Timm, Macromolecules, 26, 1369 (1993).
(4.) Y. C. Kim, S. J. Park, and J. R. Lee, Polym. J., 29, 759 (1997).
(5.) K. Morio, H. Murase, H. Tsuchiya, and T. Endo, J. Appl. Polym. Chem., 32, 5727 (1985).
(6.) J. A. Mcgowen and L. J. Mathias, Polym. Compos., 18, 348 (1997).
(7.) G. Lachenal, A. Pierre, and N. Poisson, Micron, 27, 1329 (1996).
(8.) M. Ochi, N. Tsuyuno, K. Sakaga, Y. Nakanishi, and Y. Murata, J. Appl. Polym. Sci., 56, 1161 (1995).
(9.) M. C. Finzel, J. Delong, and M. C. Hawley, J. Polym. Sci., Polym. Chem., 33, 673 (1995).
(10.) J. P. Eloundou, M. Feve, D. Harrau, and J. P. Pascault, Angew. Macromol Chem., 230, 13 (1995).
(11.) F. F. Nograro, P. Guerrero, M. Corcuera, and I. Mondragon, J. Appl. Polym. Sci, 56, 177 (1995).
(12.) J. O. Simpson and S. A. Bidstrup, J. Polym. Sci., Polym. Phys., 33, 55 (1995).
(13.) R. J. Morgan and E. T. Mones, J. Appl. Polym. Sci., 33, 999 (1987).
(14.) G. Lachenol, Vib. Spectrosc., 9, 93 (1995).
(15.) J. H. Fu and J. R. Schlup, J. Appl. Polym. Sci., 49, 219 (1993).
(16.) E. Stark. K. Luchter, and M. Margoshes, Appl. Spectrosc. Rev., 22, 333 (1986).
(17.) L. Xu and J. R. Schlup, J. Adv. Mater., 28, 47 (1997).
(18.) J. Mijovic and S. Andjelic, Macromolecules, 28, 2787 (1995).
(19.) K. Takuma, T. Takata, and T. Endo. J. Photo Polym. Sci. Tech., 6, 67 (1993).
(20.) B. S. Kim and T. Inoue, Polymer, 36, 1985 (1995).
(21.) V. Strenmel and T. Scherzer, Eur. Polym. J., 30, 361 (1994).
(22.) B. G. Min, Z. H. Stachurski, J. H. Hodgkin, and G. R. Heathy, Polymer, 34, 3620 (1993).
(23.) N. A. John and G. A. Georges, Polymer, 33, 2679 (1992).
(24.) R. F. Goddu and D. A. Delker, Anal. Chem., 30, 2013 (1958).
(25.) B. G. Min, D. K. Shin, Z. H. Stachurski, and J. H. Hodgkin, Polym. Bull., 33, 465 (1994).
(26.) V. B. Gupta and C. B. Waran, Polymer, 32, 1875 (1991).
(27.) L. K. England, M. Fritzsche, and W. A. P. Luck, J. Mol. Struct., 175, 277 (1988).
(28.) G. A. George, P. Cole-clarke, N. John, and G. Friend, J. Appl. Polym. Sci., 42, 643 (1991).
(29.) J. M. Olinger and P. R. Griffiths, Anal. Chem., 60, 2427 (1988).
(30.) J. B. Enns and J. K. Gillham, J. Appl. Polym. Sci., 28, 2831 (1983).
(31.) L. G. Weyer, J. Appl. Polym. Sci., 31, 2417 (1986).
(32.) A. Gupta, M. Cizmecioglu, D. Coulter, R. H. Liang, A. Yaurouain, F. D. Tsay, and J. J. Moacanin, J. Appl. Polym. Sci., 28, 1011 (1983).
(33.) M. K. Antoon and J. L. Koenig, J. Polym. Sci., 19, 549 (1981).
Composition of DGEBA/NMA/BPH System. Mixing Ratio (r) DGEBA/NAM/BPH (wt%) 0.25 100/25/1 0.45 100/45/1 0.65 100/65/1 0.85 100/85/1 1.05 100/105/1 Major Assignments for the Characteristic NIR Bands in This System Between 4000 and 7100 [cm.sup.-1]. Wavenumber Assignments ([cm.sup.-1]) 7000 -OH overtone and combination bands 6075 first overtone of terminal methylene fundamental stretching vibration of epoxide 5980 aromatic C-H stretching first overtone brands 5500-5300 second C=O stretching overtone of anhydride 5225 -OH stretching combination bands due to moisture 4850 first C=O stretching overtone of anhydride 4675-4623 Combination band of the aromatic conjugated C=C stretching with the aromatic -CH fundamental stretching 4530 Conjugated epoxy [CH.sub.2] deformation band with the aromatic CH fundamental stretching
|Printer friendly Cite/link Email Feedback|
|Author:||PARK, SOO-JIN; KWAK, GUN-HO; SUMITA, MASAO; LEE, JAE-ROCK|
|Publication:||Polymer Engineering and Science|
|Article Type:||Statistical Data Included|
|Date:||Dec 1, 2000|
|Previous Article:||Effect of Strain Rate on the Fracture Toughness of Some Ductile Polymers Using the Essential Work of Fracture (EWF) Approach.|
|Next Article:||Micro-Mechanical Deformation Mechanisms in the Fracture of Hybrid-Particulate Composites Based on Glass Beads, Rubber and Epoxies.|