Morphology and mechanical properties of epoxy resins containing functionalized perfluoroether oligomers.
Research work on epoxy resins modified with reactive elastomeric oligomers was pioneered almost 30 years ago (1-3), using primarily carboxylic acid terminated butadiene-acrylonitrile oligomers. A considerable amount of work has followed to fully evaluate the effects of all possible variants, including acrylonitrile content, molecular weight, and type of functionality (4-6). The principle underlying the toughening mechanism of such systems is now well understood and has been interpreted in terms of a spinodal decomposition of the resin solution causing the precipitation and subsequent growth of rubbery particles prior to the onset of gellation (7, 8). A relatively slow curing system is, therefore, required to allow sufficient time for molecular diffusion to take place of reactive species into the nuclei of the precipitating phase (9).
The incorporation of fluorine in epoxy resins has also been the subject of extensive research over the last 20 years or so (10, 11), aiming primarily to reduce the water absorption and to increase the durability of the cured resin at high temperatures. These monophase resins are intrinsically brittle although attempts have been made to enhance their toughness with the use of functionalized high molecular weight fluoroelastomers dissolved in solvents (12).
Another approach to the incorporation of fluorine in an epoxide network is to use fluorinated hardeners and other reactive modifiers such as elastomeric fluoroligomers as toughening agents.
Mijovic et al. (13) have grafted functional groups on a fluorocarbon elastomer to render them compatible and reactive with epoxy resins. A large increase in fracture energy was observed with the addition of 15% elastomer into the resin. The dissolution of high molecular weight polymers into a thermosetting resin invariably requires, however, the use of solvents which have to be removed by vacuum extraction prior to curing. This results in a very large increase in resin viscosity which impairs their processing characteristics. It is for this reason, therefore, that functionalized oligomeric modifiers are preferred.
Research on the use of [Alpha]-[Omega] disfunctionalized fluoroligomers has focused primarily on the production of high molecular weight linear polymers (14, 15). The first attempt to use fluoroligomers in epoxy resins was reported by Rosser et al. (16), who have incorporated a di-acid fluoride perfluoroether oligomer into an epoxy resin for use as a matrix for glass cloth composites. Despite the very high reactivity of the acid fluoride groups full compatibility could not be achieved even after pre-reacting the fluoroligomer with a large excess of epoxy resin. As a consequence the improvements in toughness were rather modest but the water absorption was reduced considerably.
In the present study fully miscible acid terminated fluoroligomers have been used to modify epoxy resins to produce both IPN's and two-phase precipitated systems from conventional resins and hardeners. These fluorinated modifiers are mixtures of telechelic hydroxy acid and di-acid oligomeric species, as illustrated in Fig. 1. Note that the central perfluoroether oligomer blocks of the precursor are highly insoluble in epoxy resins, even at 1% concentration, while the introduction of non-fluorinated segments at the chain ends renders these oligomers miscible with epoxy resins at all concentrations.
The synthesis of the central block of the perfluoroether oligomer, known commercially as Fomblin [TABULAR DATA FOR TABLE 1 OMITTED] ZDOLTX (Ausimont SpA) is described in Refs. 17 and 18, while the subsequent chain extension reactions to produce the perfluorether modifiers shown in Fig. 1 have been reported elsewhere by the authors (19, 20). These will be referred to as Prep TX.
IPN formulations were produced by first mixing a bisphenol-A epoxy resin (Epikote 828, supplied by Shell Chemical) with 80 phr hexahydrophthalic anhydride (HHPA), (supplied by Ciba-Geigy under the trade name of Hardner HT907) until a clear solution was obtained. This was followed by the addition of 1 phr benzyldimethylamine (BDMA). The two-phase precipitated formulations were prepared by first producing the epoxy-extended prepolymer, which was then mixed with the equivalent of approximately 80 phr HHPA and 1 phr BDMA.
The epoxy-extended prepolymer was produced by heating 4:1 excess epoxy resin with the perfluoroether prepolymer at 85 [degrees] C in the presence of 1 phr triphenyl phosphine (TPP) for different periods of time up to 6 h, when a cloudy product was obtained. The reaction was followed by measuring the viscosity of the solution at 60 [degrees] C using a Haake cone-and-plate apparatus. Samples taken at different times during the course of the reaction were subsequently evaluated for the production of two-phase precipitated formulations as outlined earlier. The composition of both IPN and two-phase systems is shown in Table 1.
The prepolymer Prep TX(A) was as illustrated in Fig. 1, i.e. a 1:1 mixture of the two compounds, while prepolymer Prep TX(B) was obtained by reacting Prep TX(A) with 20% of an epoxy terminated perfluoroether oligomer (shown below) for 15 h at 85 [degrees] C, i.e. until the mixture became transparent and fully miscible with the epoxy resin. Longer reaction times were not used to avoid excessive rises in viscosity.
[Chemical Expression Omitted]
where p, q, and n are as indicated in Fig. 1.
In order to explore further the possibilities of producing different morphologies for two-phase systems, additional formulations were produced by diluting the epoxy-extended prepolymer used in System A with various quantities of unreacted mixtures of epoxy resin and prepolymer prior to curing.
This procedure was expected to produce morphologies containing features of both IPN (matrix) and dispersed particles two-phase systems. For the purpose of identification these formulations are referred to as System C. For all three types of formulations the final mixtures were cast into PTFE molds to produce rectangular specimens 150 by 12 by 2.5 mm, cured for 24 h at 120 [degrees] C and post-cured for 3 h at 150 [degrees] C and 1 h at 180 [degrees] C.
The mechanical properties were assessed by means of 3-point bending tests to measure flexural strength, flexural strain at break, flexural modulus (according to ASTM D790), and critical stress intensity factor [K.sub.IC]. For the latter tests central rectangular slots, about 1 mm wide and in various lengths, were machined by means of a circular saw and a short crack was introduced at the tip with a sharp razor blade. The a/w ratio (a = crack length and w = specimen depth) was changed within the range 0.2 to 0.5. The specimens were fractured in a 3-point bending mode at 2 mm/min, using a span/width ratio of 4:1 and 8:1 respectively. The load to fracture, P, was recorded and used to calculate the [K.sub.IC] value from the slope of the plot YP vs [a.sup.-1/2] forming a straight line going through the origin, where Y = compliance calibration factor.
The critical strain energy release rate, [G.sub.C], was subsequently calculated using the formula [Mathematical Expression Omitted], where E[prime] was taken to be equal to the flexural modulus of the material. Measurements were also made on specimens aged for 3 weeks in an air-oven at 200 [degrees] C.
The morphology of the cured resins was examined by SEM using the specimens fractured in the notched flexural tests, and the Vickers microhardness for both matrix and precipitated particles was measured on the Riechert Me[F.sub.2] instrument. Dynamic mechanical spectra were also obtained from - 120 to 180 [degrees] C using the Du Pont DMA-983 instrument at 1 Hz and a heating rate of 10 [degrees] C/min; the temperature of the E[double prime] peaks was taken as a measure of the [T.sub.[Beta]] and [T.sub.g] values.
RESULTS AND DISCUSSION
The IPN formulations produced transparent castings, whereas the two-phase systems were opaque. SEM analysis revealed that the IPN formulations contained a very free layered morphology with diffused lamellar microdomains about 0.2 [[micro]meter] thick [ILLUSTRATION FOR FIGURE 2 OMITTED]. The two-phase, A and B systems, on the other hand, contained spherical particles, ranging from 1.5 to 20 [[micro]meter]; the larger particles consisting of an organized agglomeration of smaller particles about 0.5 to 1.0 [[micro]meter] in diameter [ILLUSTRATION FOR FIGURES 3 AND 4 OMITTED]. The particle size distribution for systems A and B is shown in Fig. 5. Although there was a very sharp interphase with no sign of debonding between the matrix and the dispersed particles, the outer layers of these particles were very smooth and exhibited a morphology similar to the IPN formulations. The volume fraction of the precipitated particles was about twice the amount of original perfluoroether used (i.e. about 1.5 times the amount of prepolymer) suggesting that the dispersed particles contained about 30 to 40% of resin/HHPA mixture.
Microhardness measurements showed that, at about 4% prepolymer content, the precipitated particles were much softer than the matrix, i.e. HV [approximately equal to] 11 against HV [approximately equal to] 20; the latter value being approximately equal to that exhibited by the control samples i.e. without modifier [ILLUSTRATION FOR FIGURE 6 OMITTED]. For IPN systems the micro-hardness decreased linearly with the perfluoroether content to HV [approximately equal to] 11 for prepolymer content in the region of 25% [ILLUSTRATION FOR FIGURE 6 OMITTED].(*) The DMA analysis showed that at low concentrations of Prep TX the [T.sub.g] was slightly higher (i.e. [similar to] 6 [degrees] C) than for the control samples and subsequently decreased by about 25 [degrees] C at prepolymer concentration in the region of 30%. For the two-phase systems the [T.sub.g] was about 10 [degrees] C lower than for the control specimens and at equal prepolymer concentration, the [T.sub.g] was also lower than for IPN systems.
For both systems [T.sub.g] values were considerably higher than those predicted from the Gordon-Taylor equation (21), indicating the existence of a very fine co-continuous heterogeneous morphology.
There was little difference, on the other hand, between the IPN and two-phase A and B systems with respect to [Beta] relaxations, both showing more pronounced and broader secondary relaxations than the control resin.
In Fig. 7 are shown micrographs obtained for system C containing 5.0% fluoroligomer. For the 1/1 mixtures of prepolymer/epoxy-extended prepolymer, the formulation failed to produce precipitated particles and, therefore, remained transparent [ILLUSTRATION FOR FIGURE 7A OMITTED]. The 1/3 mixtures of prepolymer/epoxy-extended pre-polymer, on the other hand, produced very large and irregular precipitated particles [ILLUSTRATION FOR FIGURE 7B OMITTED], as a result of the continuation of the growth of the particles (caused by the delay in the gellation of the resin, brought about by the dilution), forcing them to expand into the available interstices between particles.
The IPN structure of the more diluted (1:1) system, on the other hand, is believed to have been caused by the failure of the particles to nucleate from the activated species of the epoxy extended prepolymer as a result of their diffusion into the surrounding resin mixture.
The results from the flexural tests showed that both modified epoxy formulations exhibited much higher strength, fracture strain, and fracture energy ([G.sub.C]) than the control samples. While the flexural strength increased by 30 to 70%, depending on the amount of fluoroligomer modifier used, the modulus only decreased by 20 to 30% i.e. from 3.4 GPa to a minimum of 2.5 GPa, in a nonsystematic manner, although the lack of consistency may have resulted from experimental error [ILLUSTRATION FOR FIGURES 8 AND 9 OMITTED].
Within the context of particle precipitation during the curing of the two-phase systems it is important to consider the effect of reaction time for the preparation of the epoxy-extended prepolymer. For the preparation of the epoxy extended prepolymer for System A formulations, the graph in Fig. 10 reveals practically no change in viscosity up to about 4 h reaction time but a rapid decrease occurs at longer time, which is accompanied by the development of turbidity. The reduction in viscosity is, therefore, associated with the removal of higher molecular weight species from the liquid phase as a result of the precipitation of particles.
The cured formulations were found to exhibit an IPN morphology (transparent products) for reaction times up to 3 h and a two-phase microstructure for longer reaction times. At reaction times greater than 5 h for the preparation of the prepolymer, however, the precipitated particles in the cured formulations were coarse and irregular in shape, although some spherical particles were also present [ILLUSTRATION FOR FIGURE 11 OMITTED]. Furthermore these systems were rather brittle i.e. [G.sub.c] = 0.69 KJ/[m.sup.2] for epoxy-extended prepolymers produced with a 5 h reaction time, [G.sub.c] = 2.39 KJ/[m.sup.2] for a 3 h reaction, and [G.sub.c] = 0.39KJ/[m.sup.2] for the control sample.
It is thought that the prematurely precipitated coarse particles are intrinsically very brittle due to the absence of anhydride hardener, and cannot therefore provide an efficient toughening mechanism. This illustrates the importance of allowing the particles to grow within the fully formulated system, so that the hardener can distribute itself in the two phases according to its solubility.
Thermal aging brought about a reduction in flexural strength and strain at break for the two-phase formulations, but the measured values were always much greater than for the control samples. The changes in flexural strength for IPN formulations were similar to those for the two phase systems, whereas the strain at break increased considerably with aging in proportion to the concentration of prepolymer used [ILLUSTRATION FOR FIGURES 12 AND 13 OMITTED].
Thermal aging appears to have produced a coarsening of the dispersed fluoroligomer-rich lamellae ([ILLUSTRATION FOR FIGURE 2 OMITTED], micrograph at bottom), while the two-phase systems showed signs of debonding between matrix and precipitated particles and a coarsening of the microstructure of the precipitated particles [ILLUSTRATION FOR FIGURE 14 OMITTED].
The effect of thermal aging on precipitated System C formulations (1:3 dilution) showed a similar behavior to the analogous System A (undiluted), while those that failed to precipitate, i.e. remain as IPN's (1:1 dilution), not only exhibited the expected increase in strain at break but showed also an increase in flexural strength. After 3 weeks at 200 [degrees] C the strain at break increased from 4.2% to 5.1%, while the flexural strength increased from 85 to 104 MPa.
From the results of this study it can be concluded that:
1. Highly insoluble perfluoroether oligomers can be miscibilized with epoxy resins (anhydride hardener mixtures) through small extensions of the chains with acid terminated non-fluorinated species.
2. These functionalized chain extended oligomers can be used as effective modifiers for epoxy resins to produce both IPNs and precipitated particles (two-phase) systems. These exhibit substantial enhancement in toughness and resistance to high temperature degradation with only small changes in the glass transition temperature.
3. The large difference in solubility parameters between. the central blocks of the perfluoroether oligomers and epoxy resins provide a very efficient mechanism for the nucleation of particle precipitation during curing via epoxy-extension reactions, while the well matched solubility parameters of the segments at the end of the oligomer chains prevents phase separation and results in the formation of finely dispersed lamellar IPN's.
* The quantities shown for Fomblin in Figs. 6, 8, 9, 12, and 13 correspond to the quantities of perfluoroether before being chain extended to achieve solubilization in the epoxy resin.
1. R. S. Drake and A. R. Siebert, SAMPE Quart, No 6, 4 (1975).
2. F. J. McGarry and A. M. Willner, ACS Div. Org. Coat. Plast. Chem. Prepr. 28, 55 (1968)
3. A. R. Siebert, E. H. Rowe, C. K. Riew, and J. M. Lipiec, 28th Ann. Tech. Conf. RP/Composites, Inst., SPI Section 1-A (1973).
4. A. K. Kinloch and R. J. Young, Fracture Behaviour of Polymers, p. 91, Appl. Sci. Publ., London (1983).
5. L. T. Manzione, J. K. Gillham, and C. A. McPherson, J. Appl. Polym. Sci., 26, 889 (1981).
6. D. S. Kunz, P. W. R. Beaumont, and M. J. F. Ashby, J. Mater. Sci., 15, 1109 (1980).
7. R. J. J. Williams, J. Borrajo, H. E. Adabb, and A. J. Rojas, Polym. Prepr. ACS, Div. Polym. Mater. Sci. Eng., 49, 432 (1983).
8. H. S. Y. Hsich, SPE ANTEC Tech. Papers, 36, 530 (1990).
9. C. K. Riew, E. H. Rowe, A, R. Siebert, and J. H. Lipiec, 27th Ann. Tech. Conf. Reinf. Plast. Comp. Inst., SPI Section 2-D (1972).
10. J. R. Griffith, Chem Tech., 112, 290 (1982).
11. L. L. Huang, Adhesives, Sealants, Coatings for Space and Harsh Environments, pp. 45-66, Plenum Press, New York (1988).
12. T. E. Twardoski and P. H. Geil, J. Appl Polym. Sci., 42, 69 (1991).
13. J. Mijovic, E. M. Pearce, and C.-C. Foun, in Rubber Modified Thermoset Resins pp. 293-307, K. Riew and J. K. Gillham, eds., American Chemical Society Series (1984).
14. E. F. Cassidy, H. X. Xiao, and H. L. Frisch, J. Polym. Sci., Polym. Chem. Ed., 22 2667 (1984).
15. F. Pilati, M. Toselli, A. Vallieri, and C. Tonelli, Polym. Bull, 28, 151 (1992).
16. R. W. Rosser, T. S. Chen, and M. Taylor, Polym. Compos. 5, 198 (1984).
17. U.S. Patent 3,722,729 (1973).
18. G. Marchioni, G. Ajroldi, and G. Pezzin, Europ. Polym. J., 24, 1211 (1988).
19. F. Zitouni, PhD thesis, Loughborough University of Technology, U.K. (1992).
20. L. Mascia, F. Zitouni, and C. Tonelli, J. Appl. Polym. Sci (in print).
21. G. I. Taylor, Proc. R. Soc. London, 226A 34 (1954); Ibid., 138A, 41 (1932).
|Printer friendly Cite/link Email Feedback|
|Author:||Mascia, L.; Zitouni, F.; Tonelli, C.|
|Publication:||Polymer Engineering and Science|
|Date:||Jul 15, 1995|
|Previous Article:||Effect of holding pressure on orientation distribution.|
|Next Article:||Double yield points in poly(tetramethylene terephthalate) and its copolymers under tensile loading.|