A novel approach to the tailoring of polymers for advanced composites and optical applications, involving the synthesis of liquid crystalline epoxy resins.

INTRODUCTION

In recent years, intense research and patent activities have addressed the field of liquid crystalline epoxy resins (LCER) (1-5). This class of materials seems to be very promising for the following reasons:

* The presence of chemical crosslinks between adjacent molecular segments can improve the relatively poor mechanical properties in the direction transverse to the molecular orientation, an inherent characteristic of thermoplastic liquid crystalline polymers. In fact, the strong anisotropy of the mechanical properties of uni-directionally oriented liquid crystalline thermoplastic polymers determines outstanding properties along the molecular direction, but only poor properties in directions orthogonal to the average molecular orientation.

* The low viscosity of the mixture of reagents before the crosslinking reaction permits easy flow during filling of suitable molds and a prompt orientation of the molecules in a preferred direction, if an external field is applied. Conversely, the high viscosity of polymers in the molten state prevents achievement of a uniform molecular orientation due to applied electromagnetic fields. Intense elongational and/or shearing flows are necessary to orient polymeric samples in the liquid crystalline state.

Moreover, because of the peculiar molecular structure of the liquid crystalline phase, they can offer improved toughness with respect to conventional thermosets used for advanced composites (3, 6)

New emphasis has been recently given to research in the field of liquid crystalline thermosets because of their potential optical and electro-optical applications (7). It is possible to take advantage of the anisotropy of liquid crystalline phases and exploit the unique molecular features of these phases in the production of optical and electro-optical devices. This can be achieved by orienting the material under the influence of an electric, magnetic, or mechanical field, or as an effect of surface interactions (4, 5, 8). With regard to this, LCER have been found to be very promising in this field: they can be oriented more easily before the crosslinking is performed than both ordinary and liquid crystalline thermoplastic polymers. Moreover, they can retain the orientation above [T.sub.g] because of the existence of chemical crosslinks between the chains. Many experimental and theoretical investigations concerning the orientation and relaxation processes of these materials with different procedures have been reported in the scientific and patent literature (9-11).

In previous papers (3, 6, 8, 12-14) we reported the synthesis and the properties of several liquid crystalline thermosetting epoxies whose mesogenic portions lie in the main chain. For this reason the overall mechanical properties of the final thermosets are overwhelmed by strong chemical bonds constituting the rigid primary chemical structure.

Where optical and electro-optical applications are concerned, main chain LCER exhibit some advantages: first, because of their low monomeric viscosity, the curing resin can be easily oriented before gelation; then the reduced system mobility prevents the chain relaxation. This procedure causes good control of the system morphology and structure before completion of the curing process. Of course, in order to obtain a uniformly oriented system without defects, as it is required for the previously proposed applications, crosslinking should not perturb the orientation achieved prior to the gel point.

The material must exhibit good chemical and physical stability up to high temperatures. This can be obtained by means of a proper selection of molecular parameters such as molecular shape and rigidity, nature of chemical bonds, absence of pendant reactive groups, etc.

Because of their optical properties, different applications can be planned for oriented liquid crystalline main chain thermosets in the field of optical and electro-optical displays. They can be used as waveguides, filters, second harmonic generation (SHG) devices, matrices for Polymer Dispersed Liquid Crystals (PDLC), as well as optical switches (7, 11). Of course, for these purposes some requirements must be satisfied. The materials under investigation must exhibit high optical and electrical anisotropy. The achieved orientation in the fully cured material must be very homogeneous, i.e., the presence of defects must be carefully minimized; the LC thermosets need to have a very high dimensional stability in order to prevent internal stresses and increase durability; and, since the macroscopic anisotropy of optical and electrical properties is the essential feature required for such applications, it is crucial that the molecular orientation is also very stable over a wide temperature range.

In order to attain this target, several points must be considered. First, the epoxy monomers should be properly selected if a liquid crystalline thermoset is to be obtained. The stability of a mesophase is in fact the result of different factors, such as the monomeric shape, its polarizability, the presence of highly polarizable groups, and the existence of noncovalent interactions such as hydrogen bonds. All these parameters must be taken into account if a liquid crystalline phase is to develop and be stabilized during the cure.

As already pointed out, the minimization of defects is absolutely necessary in order to get a uniformly aligned network that is suitable for optical and electro-optical applications. On the other hand, it is also necessary that chain mobility in the cured network is hindered enough for the mesogen orientation to be retained. In order to balance both the objectives, it is crucial to control the crosslink density carefully. In fact, if this structural parameter has too high a value, this promotes the presence of defects in the mesophase because of density change during the reaction and because of the distortion of molecular arrangement in the liquid crystalline phase. However, too low a crosslink density cannot guarantee that the final material will retain its orientation prior to the gel point.

Furthermore, it must be possible to get a good orientation of the liquid crystalline resin before gelation drastically reduces the chain mobility. Therefore, it is also necessary to control the reaction rate by a proper selection of parameters such as temperature, purity of reagents, presence of inhibitors, or photoreactive groups. It was reported by the authors (12, 13) that the onset of the liquid crystalline phase during the curing reaction determines a change of curing kinetics with respect to commercial isotropic epoxy resins. This is another aspect that must be taken into account when careful control of the reaction rate is desired for such systems.

EXPERIMENTAL

For the detailed syntheses of monomers the authors refer to the previously published papers (3, 6, 8, 12-14). In a typical functionalization with epoxy groups, the compounds were reacted with epichlorohydrin in the presence of sodium hydroxide according to the procedure described in the literature (3). DAF, DDS, t-CN, DAT, o-tolidine and the aliphatic acids were purchased from Aldrich and their purity was checked by NMR.

Samples were prepared by mixing the monomers in stoichiometric ratio and by subsequently curing them at the desired temperature in a test tube, DSC pan or between two glass slides previously treated with a surfactant agent (Surfasil[R], Pierce).

The phase transition of the specimens and the chemical reactions were monitored with a differential scanning calorimeter (DSC) DuPont Mod. 2910. Nitrogen was used as purge gas. In dynamic experiments a heating rate of 10 [degrees] C/min was adopted.

The textures of the monomers and cured resins were observed with an optical microscope Reichert-Jung, mod. Polyvar in transmitted light, equipped with crossed polarizers. The temperature was controlled by means of a Linkam mod.TH600 hot stage.

The X-ray diffraction spectra were obtained by the photographic method using a Rigaku mod. III/D max generator. A Ni-filtered Cu-[K.sub.alpha] radiation was used. Fracture toughness was evaluated according to ASTM E-399 test procedure using 1[inch] x 1[inch] x 1/8[inch] compact tension specimens.

Flexural properties of the cured resins were measured according to ASTM D 790. Coefficients of thermal expansion (CTE) were evaluated by means of a thermomechanical analyzer (TMA) DuPont Mod. 2940.

The length of the rigid rod molecules was calculated after minimizing energy by means of the program Desktop Molecular Modeller.

Sorption tests were performed by immersion of the specimens respectively in methylene chloride at room temperature for 17 days and in distilled water at 100 [degrees] C for 17 days and then by monitoring the weight increase.

DISCUSSION

Physical Properties of Resins Cured With Different Curing Agents

In our previous investigations (3) we demonstrated that the use of a bifunctional epoxy molecule with a stiff and elongated structure is the main factor in determining the features of the resulting liquid crystalline network. The epoxy compounds do not necessarily need to exhibit a liquid crystalline phase as a pure material. In fact, the lack of long and flexible tails attached to the rigid core leads to a high melting temperature for the epoxy compounds, thus hindering the liquid crystalline phase. The liquid crystalline phase appears at an early stage during the curing process. The mechanism of this transition for the curing thermosets can be attributed to the development of weak interactions which contribute to the stabilization of the liquid crystalline phase. For completely cured systems, the liquid crystalline order of the molecules is stabilized over a temperature range depending on the overall molecular mobility, which in turn is mainly affected by the choice of the crosslinking agent. In the case of a low crosslinking density, the transition from a liquid crystalline to isotropic state can be observed as the temperature increases. When [TABULAR DATA FOR TABLE 1 OMITTED] the crosslinking density is higher, the mesophase develops during the initial curing step when the epoxy monomers and the curing agent react and gives rise to a linear prepolymer not yet crosslinked. The liquid crystalline phase appears and is furtherly stabilized during the gelation of the thermoset and can be destroyed only upon thermal decomposition of the network. High crosslinking density thermosets can be obtained by reacting the rigid-rod epoxy-terminated molecules with polyfunctional amines. In this case, the selection of the curing temperature turns out to be crucial. In fact, if the curing reaction is carried out at a temperature higher than the isotropization of the growing prepolymer, crosslinking takes place before the chain extension has developed enough to stabilize the liquid crystalline structure and the system cures in the isotropic state.

We selected and synthesized a number of rigid-rod epoxy-terminated compounds that could give rise to a liquid crystalline phase upon curing. The molecular formulas of such compounds are reported in Fig. 1. These compounds were reacted with different polyfunctional amines [ILLUSTRATION FOR FIGURE 2 OMITTED] or with carboxylic acids of general formula HOOC-[(C[H.sub.2]).sub.n]-COOH, with n ranging from 2 to 8. In the case of reaction with amines, all the cured systems but the ones obtained with 2,6(2,3-epoxypropoxy)-naphtalene (NAFT) gave a liquid crystalline structure upon curing. In the case of NAFT, the different structure of the cured thermoset can be ascribed to the lower length of the rigid part of this epoxy molecule, which is too short (4.7 [Angstrom]) with respect to the other considered compounds, whose rigid block lengths range between 7 and 13.5 [Angstrom]. This feature reduces the nematogenic character of the epoxy monomer, as predicted by Flory (15), and allows only isotropic networks to be obtained.

In order to assess the effect of the relative amount of the curing agent on the properties of the crosslinked thermosets, p-(2,3-epoxypropoxy)-diphenyl (DIF) was cured with different amounts of p-aminoacetophenone azine (NA2). Table 1 reports the properties of the cured resins obtained by reacting DIF with 85%, 100%, and 115% of the stoichiometric amount of NA2 respectively at different temperatures. All the materials obtained have liquid crystalline character. Regarding the fracture toughness, the best results were obtained for the resin cured with an excess of amine; however, this positive result is negatively balanced by the lower modulus and the lower [T.sub.g]. Higher [T.sub.g]s were reached for resins cured with less than stoichiometric amount of amine. However, a lightly more brittle material resulted in this ease. For each stoichiometric composition studied, a higher curing temperature causes a lower fracture toughness, while [T.sub.g] and modulus are not strongly affected by curing temperature. Since the most critical parameter to be investigated is fracture toughness, the effect of the state of order of the cured system was related to this parameter (3). The lower fracture toughness at higher temperatures is due to the progressive reduction of the extent of nematic portions and the progressive increase of the isotropic parts of the cured resin, as revealed by optical microscopy. The density of the cured resin progressively decreases with increasing curing temperature, indicating that at higher temperatures the extent of ordered domains, acting as nodules around which crack can deviate, decreases. In other terms, the microstructure of LCER is predominant but consists of anisotropic domains with properties, such as strength. that are different along and across their molecular orientations. This results in the deviation of crack propagation from a straight line and suggests that inhomogeneities and localized anisotropy of the nematic structure is the main reason for the fracture toughness increase in the liquid crystalline material.

Table 2 reports the mechanical properties of DOMS cured with NA2 at two different temperatures. The system cured at lower temperature appeared to be opaque and exhibited a LC structure, while the one obtained in the other case was transparent and showed an isotropic structure. The nematic resins exhibited a fracture toughness nearly 50% higher without affecting any value of flexural mechanical properties. The water and methylene chloride uptake also remained unchanged. In the same Table, data are reported for nematic samples containing glass fibers as a reinforcing agent. The cured samples were opaque and had the same nematic textures as the unreinforced samples. The addition of glass fibers resulted in a significant improvement of modulus without reduction of toughness. The combination of high modulus and high toughness demonstrates property balance which can be attractive for applications in the area of high performance composites. The DOMS-NA2 resin exhibited excellent solvent resistance. Weight gain after immersion in methylene chloride for two weeks was only 1.0%. However, this system had a low water resistance - its weight gain after immersion in water at 100 [degrees] C for two weeks was 7.5%. Such high water uptake could be caused by the particular chemical structure of the curing agent containing nitrogen atoms, which are very sensitive to water. For this reason, NA2 was replaced by a different curing agent, o-tolidine [ILLUSTRATION FOR FIGURE 2 OMITTED]. The DOMS/o-tolidine system was cured [TABULAR DATA FOR TABLE 2 OMITTED] at 150 [degrees] C for one hour then at 175 [degrees] C for two hours. The cured panel was transparent and therefore exhibited no liquid crystalline character. It had both low water (2.5%) and low methylene chloride (1%) uptake, suggesting that with appropriate choice of chemical structure, the rod-like thermosets can have both good solvent and water resistance.

Another unusual property that sets apart rod-like thermosets from traditional flexible chain thermosets is their ability to maintain a high modulus up to very high temperatures. For example, the DOMS o-tolidine resin exhibits no significant drop in modulus up to 300 [degrees] C. Therefore, rod-like thermosets have possible applications where high temperature resins are required especially for areas such as surface mount technology. In conclusion, this example demonstrates the validity of the concept that a rigid rod epoxy thermoset material can have properties or property balance beyond the scope of existing flexible chain epoxy systems, and therefore, these materials can have numerous market opportunities.

In the case of thermosets obtained by reacting the same epoxy molecules with acids, liquid crystalline behavior was obtained in all cases. In the case of NAFT, an extra driving force, apart from the anisotropy of the molecule, must be claimed to explain this point owing to the low value of the aspect ratio of this molecule. It is known that hydrogen bonds play a positive role as driving force for liquid crystalline ordering and several authors have reposed on systems where hydrogen bonding play the major part in mesophase formation (16-19). Moreover, a recently appearing theory predicts that hydrogen bonds alone may cause liquid crystalline ordering in otherwise isotropic liquids (20). In our case, it may be supposed that the presence of a long and flexible curing agent, such as an acid, can provide enough conformational freedom to the polymeric chains which can then align under the influence of hydrogen bonding between the hydroxyl groups, formed as an effect of the epoxy ring opening during the cure, and the carboxyl groups of the chains. This could be, therefore, the driving force for the onset of a mesophase in the case of a very short rigid core in the epoxy molecule, such as NAFT.

The reaction with acids seems to favor a more ordered liquid crystalline phase compared to the amines. In the former case, with the exception of succinic acid HOOC-[(C[H.sub.2]).sub.2]COOH, thermosets with a smectic structure were obtained, while in the latter case only networks having a nematic structure could be obtained. It was shown (21) that in the case of side chain LC polymers, the flexible segments mainly act as spacers between the main chain and the side mesogenic groups, thus allowing the assembling of the material into an anisotropic structure. A similar role is also played by the flexible segments connecting the mesogenic groups in the case of a main chain LC polymer. Theoretical calculations (22, 23) showed that in low molecular weight liquid crystals the flexible segments are not simply space fillers, but take part in the anisotropy of the phase. Therefore, it is not surprising that also in the case of a poorly crosslinked LC network, the aliphatic portion plays an active role in the development and stabilization of the mesophase. In the case of reaction with amines, the high crosslinking density produces too many constraints and does not allow the mesogenic portions to organic in a mesophase more ordered than the nematic.

Effect of the Crosslinking Density

In order to investigate the effect of crosslinking density on the presence of defects, and therefore on the optical homogeneity of oriented samples, the DOMS monomer was selected and reacted with different curing agents whose structures are reported in Fig. 2.

In our first investigation, reaction was performed with stoichiometric amounts of 2, 4-diaminotoluene (DAT) at 80 [degrees] C after melting of the mixture at 130 [degrees] C. The molten mixture is in the isotropic state at the beginning, when the fractional conversion is virtually zero. As the reaction starts and proceeds further on, the fractional conversion and the average molecular weight of the epoxy resin increases. The reaction of the primary amine is favored with respect to the secondary one, leading to the formation of linear or slightly branched oligomers. The length of the oligomers increases and eventually reaches the limit beyond which the liquid crystalline phase is stabilized by intermolecular forces. As a result of formation of oligomers, the initially isotropic state turns into nematic, as it can be observed from the appearance of birefringence by optical microscopy between crossed polarizers. As the curing reaction proceeds, crosslinking irreversibly locks in the nematic structure. However, the low viscosity of the epoxy/curing agent mixture during the early stage of the reaction allows the molecules to be unidirectionally oriented before the chemical crosslinks hinder the molecular realignment process and finally lock the structure. This procedure can be used to produce highly aligned samples whose orientation is stable over a wide temperature range. In our case, a uniform molecular alignment in the liquid crystalline material was achieved by liquid crystal-surface interactions (8). The DOMS-DAT mixture was reacted between glass substrates whose surface had been previously spin-coated with polyimide and then unidirectionally robbed. When the conversion of the monomers reached a threshold value, the sample became liquid crystalline. Because of the surface-liquid crystal interactions, the molecules in the nematic phase became uniformly oriented. The fully cured sample behaved as a uniaxially optical plate if rotated between crossed polarizers. However, the final texture of the epoxy network was found to be not completely uniform due to the presence of a number of defects. In fact, the reaction of the secondary amine and the formation of crosslinks create distortions of the molecular order in the nematic phase, thus generating such defects.

In order to reduce the formation of defects in the cured sample, a less densely crosslinked network was prepared by considering that the distance between contiguous crosslinks in the network is related to the length of the monomers. Therefore, DOMS molecules were replaced by linear oligomers obtained by reacting DOMS with aniline (ANi) with a stoichiometric imbalance of 2 to 1 mole. In this way, an epoxy-terminated dimer was obtained. In order to minimize secondary reactions which lead to the formation of branching and crosslinking using difunctional amines, the reaction was not driven to completion. By back titration of epoxy groups, a value of 550 g/mol was determined for the molecular weight of the so-obtained oligomers (oli-DOMS). The synthesis of the epoxy terminated oligomers and the preparation of the reacting mixture by addition of the stoichiometric equivalent of DAT in a second step ensured a control of crosslink density along the chemical backbone. The oli-DOMS/DAT mixture was melted at 130 [degrees] C and subsequently cured at 100 [degrees] C between glass substrates coated with unidirectionally aligned layers.

The curing reaction and formation of the nematic phase for oli-DOMS/DAT is analogous to what has been already described for the DOMS/DAT system, as it can be easily predicted from the close chemical similarities of the two systems.

Comparison of the DOMS/DAT and oli-DOMS/DAT samples between crossed polars at the end of the curing reaction showed remarkable differences between the highly crosslinked and the less densely crosslinked networks. Decreasing the crosslink density resulted in a reduction of defects in the sample. Disclinations and a focal conic-like texture were present in DOMS/DAT film. X-ray diffraction measurements performed on bulk samples were in agreement with a nematic order of the structure, nevertheless the high periodicity of the crosslinks may have induced a smectic-like ordering of the phase. This periodicity was disrupted in the sample prepared with oli-DOMS/DAT mixture. As a consequence, the nematic texture was distorted to a lesser extent during the formation of the tridimensional network. This procedure yields a sample which is uniformly aligned along the robbing direction of the polyimide aligning layers. The uniform molecular alignment is thermally and mechanically stable - it can be destroyed only upon a decomposition process at high temperatures. This makes this preparation procedure very attractive for optical and electro-optical applications.

Another attempt to obtain a less densely crosslinked liquid crystalline network was done by reacting DOMS with aliphatic acids. In this case, as it was explained before, a smectic phase was obtained in the cured thermoset; the cured systems still display a transition from LC to isotropic phase. The glass transition temperatures of these systems were quite low, i.e. they typically ranged between 26 [degrees] C and 49 [degrees] C. This result led us to investigate the possibility of orienting such systems in bulk under the influence of mechanical stress. Our attention was focused on the system obtained upon reaction of DOMS with decanedioic acid (SA) at 180 [degrees] C. This sample exhibits a [T.sub.g] of 27 [degrees] C and a smectic-isotropic transition temperature of 81 [degrees] C, as inferred from differential scanning calorimetry analyse (DSC) and optical microscopic observation.

A DOMS-SA film obtained by curing the monomers under the previously described conditions was stretched at 50 [degrees] C, while it was in its rubbery state, up to 160% strain at 0.2 mm/min, then quenched down to the glassy state at room temperature. The X-ray patterns of the unstretched and stretched sample are reposed in Fig. 3. From a macroscopic point of view, the sample turned from turbid and white to almost completely transparent. The X-ray pattern at room temperature of the stretched sample shows a broadened wide-angle reflection at the equator. The smectic layer reflections can be observed at the meridian, thus indicating a perpendicular orientation of the smectic layers to the stress direction. The order parameter S was calculated from the X-ray pattern at room temperature for this oriented system and from IR circular dichroism measurements performed at 60 [degrees] C during the stretching experiment. In both cases calculations gave a value of S equal to 0.78 for the stretched sample. No remarkable variation in the S value was observed after annealing the sample at room temperature for several months. This evidence indicates that very good orientation can be achieved for this material, and makes it very promising for optical applications. However, a serious limit to its applicability resides in the low value of [T.sub.g], which suggests that a quite fast relaxation to the unoriented state will occur if the operating temperature of the material is slightly higher than room temperature.

Effect of the Liquid Crystalline Phase on the Crosslinking Process

The possibility to predict the overall chemical and physical behavior of potentially mesogenic reacting systems is crucial. The extent of reaction achieved, the tridimensional network structure, the orientation of molecular segments along preferred directions all can affect the macroscopically measured properties. From the practical point of view, it is quite clear how the thorough understanding of the curing mechanism of reactive resins and the prediction of bulk properties are necessary for the implementation of TLCP's in industrial applications. If the mesogens exhibit monotropic liquid crystalline behavior, or if the temperature stability range of the mesophase is narrow, the final desired molecular arrangement can be achieved only by modeling the phase transition coupled with the extent of reaction. However, the microscopic molecular features are governed to different extents by the curing conditions and the physical state developed during the curing reaction.

The literature concerned with the modeling of reactive systems suitable to form liquid crystalline phases is quite recent, and only few attempts have been made to develop a phenomenological model suitable to take into account the formation of the liquid crystalline phase during the curing reaction (12, 24).

Despite the complexity of the reaction between epoxy and amine groups, the overall reaction rate can be expressed by means of relatively simple kinetic equations (25, 26). A practical method for characterizing the reaction rate of exothermal reaction is to make use of calorimetric analysis.

Under the basic assumption that the heat flow relative to the DSC instrumental baseline is proportional to the reaction rate, fractional conversion [Alpha] is expressed by the following equation:

[[Alpha].sub.(t)] = [integral of] [Theta]dt between limit t and 0/[Delta]H

where [Theta] is the DSC heat flux, and [Delta]H = [Theta]dt is the enthalpy of complete reaction.

Typical DSC signal vs. time plots are given in Fig. 4. Increasing the temperature results in an increased reaction rate. This affects the DSC data, reducing the time for completion of the reaction and increasing the value of peak maxima. An interesting feature is the presence of double peaks in the DSC curves for these resins. Similar behavior was previously reported for liquid crystalline thermosets, but to our knowledge it has never been observed for conventional crosslinking reactions of simple mixtures. Double peaks have been related to the formation of liquid crystalline phases during the crosslinking reaction. Formation of a liquid crystalline phase during the crosslinking reaction can be monitored by optical microscopy. An isotropic phase results after melting of the two components. At some intermediate level of crosslinking, the liquid crystalline phase transition is observed. It is interesting that the second peaks have never been observed before birefringence occurs, suggesting that such second peaks are related to different reaction kinetics in the liquid crystalline state.

A phenomenological model has been proposed to account for the reaction kinetics in the isotropic and in the liquid crystalline phases (12, 24, 27). Different parameters can be obtained by fitting the experimental curves in the stable regions of the different phases. The mechanism of phase transition as a function of the fractional conversion measured during the reaction is also implemented in the overall kinetic model.

The agreement of the experimental results with the proposed model is excellent over a wide range of experimental conditions, confirming the assumption that different kinetic models do apply in the isotropic and in the liquid crystalline phases. A full treatment of the model can be found in reference (27), as its complete explanation is beyond the scope of this paper. Research is still ongoing in order to extend the results to liquid crystalline systems with afferent chemical compositions.

CONCLUSIONS

In conclusion, new thermosets can be obtained by curing rigid-rod epoxy monomers. The properties of the resulting resin depend on the length of the epoxy molecule, as well as on the nature of the curing agent. It was demonstrated that amines produce thermoses having high [T.sub.g]s and high fracture toughness. In particular, the enhancement of this last parameter is very promising if a new generation of matrices is to be sought for manufacturing of composites. As a comparison, if a toughness improvement of 50% is attempted by blending conventional epoxies with rubbers, a considerable depression of [T.sub.g] and modulus result. In the case of liquid crystalline epoxy resins, the higher fracture toughness is achieved without sacrificing [T.sub.g] or modulus. Moreover, if a modulus increase is also desired, composites with glass fibers can be prepared.

If the curing agent is an acid, thermosets having lower [T.sub.g]s and more ordered mesophases can be obtained. In this case, LC elastomers can be prepared that can find applications in the optical industry as waveguides or electro-optical devices. In this case, extra effort must be devoted to the synthesis of monomers having chiral properties, which can therefore give rise to sinectic [C.sup.*] phase. In fact, only in this case is the polarizability of the molecules great enough to have high values of electrical susceptibility, which is one of the requirements to produce polymers for electro-optical applications.

REFERENCES

1. Ger. 3622610 (1986) Bayer AG, invs.: R. Dhein, H. P. Mueller, R. Gipp, H. Heine; Chem. Abstr., 109, 38835h (1988).

2. Ger. 3622613 (1988/86) Bayer AG, invs.:R. Dhein, H. P. Mueller, R. Gipp, H. M. Meier: Chem. Abstr., 109, 7136b (1988)

3. M. Giamberini, E. Amendola, C. Carfagna, MoL Cryst. Liq. Cryst., 266, 9 (1995).

4. G. G. Barclay, C. K. Ober, K. I. Papathomas, D. W. Wang, J. Polym. Sci., Part A: Polym. Chem., 30, 1831 (1992).

5. G. G. Barclay, S. G. McNamee, C. K. Ober, K. I. Papathomas, D. W. Wang, J. Polym. Sci., Part A: Polym. Chem., 80, 1845 (1992).

6. C, Carfagna, E. Amendola, M. Giamberini, Composite Structures, 27, 37 (1994).

7. H. Finkelmann, W. Meier, H. Scheuermann, in Liquid Crystals: Applications and Uses, B. Bahadur, ed., World Scientific, Singapore (1992).

8. E. Amendola, C. Carfagna, M. Giamberini, L. Komitov Liq. Cryst., 21(3), 317 (1996).

9. W. Gleim and H. Finkelmann, in Side Chain Liquid Crystal Polymers, C. B. McArdle, ed., Chapman and Hall, New York (1989).

10. B. A, Rozenberg and L. L. Gur'eva, in Liquid Crystalline Polymer Systems, A. I. Isayev, T. Kyu and S. Z. D. Cheng, eds., American Chemical Society, Washington. D, C. (1996).

11. J. L. West, in Liquid Crystalline Polymers, ACS Symposium Series, 435, 491 (1981).

12. E. Amendola, C. Carfagna, M. Giamberini, G. Pisaniello, Macromol. Chem. Phys., 196, 1577 (1995).

13. C. Carfagna, E. Amendola, M. Giamberini, Liq. Cryst., 13(4), 571 (1993).

14. M. Giamberini, E. Amendola, C. Carfagna, Macromol. Chem. Rapid Commun., 16, 97 (1995).

15. P. J. Flory, Principles of Polymer Chemistry, Cornell University Press, Ithaca, New York (1955).

16. G. W. Gray, Molecular Structure and the Properties of Liquid Crystals, Academic Press, New York (1972).

17. L. J. Yu, Liq. Cryst., 14, 1311 (1993).

18. L. J. Yu, J. S. Pan, Liq. Cryst., 14, 829 (1993).

19. A. Mori, H. Takeshita, R. Nimura, M. Isobe, Liq. Cryst., 14, 821 (1993).

20. B. A. Veytsman, Liq. Cryst., 18(4), 595 (1995).

21. H. Finkelmann, J. H. Ringsdorf, H. Wendorff, Makromol. Chem. 179, 273 (1978).

22. S. Marcelja, J. Chem. Phys., 60, 3599 (1974).

23, D. A. Pink, J. Chem. Phys., 63, 2533 (1975).

24. J. Liu, C. Wang, G. A. Campbell, J. D. Earls, R. D. Priester Jr., The Polymer Processing Society, Twelfth International Annual Meeting, Sorrento, Italy (May 27-31, 1996).

25. J. M. Barton, in Advances in Polymer Science, Springer-Verlag, Berlin (1985).

26. B. A. Rozemberg, in Advances in Polymer Science, Springer-Verlag, Berlin (1986).

27. G. Micco, M. Giamberini, E. Amendola, C. Carfagna, and G. Astarita, I & EC Research, Special Issue, American Chemical Society, 36(8), 2976 (1997).