Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
2 BACKGROUND Relatively unhindered rotation about the single bonds in the backbone of most macromolecules means that a random trajectory for the polymer chain is the most common polymer conformation. Features of some special monomers (resonance, stereochemistry, etc.), however, restrict internal degrees of freedom in the polymers such that the mainchain is extended in space along an almost linear trajectory. Extended polymer chains or chain segments can, through excluded volume interactions, lead to long-range orientational ordering of the macromolecules- liquid crystallinity in concentrated solution or in the melt. Liquid crystals (LCs), sometimes also called mesoPhases. were first recognized in low-molar-mass compounds a century ago, and they end oy widespread technological applications because of their unique electro-optical properties. Low-molar-mass LCs are highly anisotropic fluids that exist between the boundaries of the solid state and the conventional isotropic liquid state and exhibit features or both states. ~ macromolecules, orientational ordering of extended polymer chains is sufficient to impart some crystal-like orientational ordering to their fluid phases melts or polymer solutions. Although this orientation is very subtle on a local scale (it is masked by the rapid and complex molecular dynamics characteristic of all fluid phases), time-averaged attributes of these fluids are anisotropic and therefore dramatically different from those same attributes 1: =~ ~ ~ To Ah" ~= c" of in ordinary isotropic liquids. Figure 2.1 exaggerates the "~ between the melt of a conventional random coil polymer (top) and that of a liquid crystalline polymer (LCP) (bottom) in order to pictorially show the long range order in a fluid phase of the LCP. In LOP melts or solutions this average anisotropy has dramatic consequences. When macroscopic uniform alignment of local directors exists, such fluids exhibit bulk anisotropic dielectric, magnetic, optical, transport, etc., properties. Materials formed from polymers that are orientationally organized in the fluid state retain this anisotropy in the solid state and frequently exhibit ultrahigh strength and stiffness (modulus) along the machine direction (parallel to the extended chains) because the organization and confirmational preferences of the chain promote an extended-chain crystal habit. For these 13
Semi | flexible Ado - chain ~ mer += FIGURE 2.1 Schematic indication of the differences between an isotropic (top) and a liquid crystalline (bottom) polymer fluid.
15 reasons, LCPs are being increasingly utilized in specialty and high- performance applications. Historical reviews (Economy, 1989; Jackson, 1989; Samulski, 1985; and Dobb and McIntyre, 1984), general introductions (Brows tow, 1988; Chapoy, 1985; Finkelmann, 1987; and S~mulski, 1982) , and contemporary reviews of LOP properties (Calundann et al., 1988; and G. E. Williams, 1987) abound. LCs and LCPs may be divided into two broad categories, according to the principal means of achieving fluidity. Lyotropic LCPs result from the action of a solvent and hence are multicomponent polymer solutions polymer plus solvents). Thermotropic LCPs are produced by heat and may be single (neat polymer) or multicomponent melts. Within each category, three distinctive supr~molecular organizational or structural classes of LCs have been identified: the nematic, smectic, and twisted nematic or cholesteric phases. Structural differentiation of these phases is related to the packing aspect and dimensionality of the translational organization of the molecules. In the examples of these phases we limit consideration to mainchain LCPs. Nematic LCs are distinguished by a unique director (optic axis) in the fluid; the nematic director is established by the parallelism of the long axes of molecules ([average] polymer chain axes). There is no translational order in this nominally uniaxial fluid (Figure 2.2~. Chain parallelism also characterizes the smectic phase, but translational order is also present in the form of long-range stratification normal to the chain axes (Figure 2.39. Mobility of the entire chain within the smectic layers is possible, although this increased translational organization lowers chain mobility relative to nematic phases. (In low-molar-mass LCs bulk fluidity in the smectic structure involves the layers gliding past one another; such a transport mechanism would be sharply attenuated in polymeric analogs, wherein a single semiflexible chain traverses more than one layer; smectics formed from rigid rod polymers with the layer spacing equal to the rod length might exist.) (Wen et al., 1989) In nematic phases chain ends (defects) are randomly distributed in the ordered fluid; there may be a tendency for such defects to segregate between layers in smectic fluids composed of semiflexible chains. Cholesteric LCs are similar to nematics in organization, with the additional feature of a cumulative twist between molecules, a result of the asymmetry of intermolecular forces. This asymmetry is due to the presence of chiral centers in these mesogens. As a result, the local nematic director twists into an inherently biaxial, helicoidal supramolecular structure (Figure 2.4~. Although there is a large number of applications of cholesterics in low-molar-mass LCs that exploit their optical properties (temperature sensors, notch filters, etc.), there is little evidence of widespread use of this biaxial superstructure in LCPs.
16 FIGURE 2.2 An absence of translational order in the idealized nematic. FIGURE 2.3 Smectic stratification (lateral registration) in a polymer mesophase.
FIGURE 2.4 Helicoidal cholesteric structure in a mainchain LOP.
18 MACROMOLECULAR DESIGN AND SYNTHESIS Thermotropics Mainchain (linear) LCPs are generally synthesized by condensation polymerization involving transesterification (Jackson, 1989) (Figure 2.5~. The growth step is an ac~dolysis reaction yielding an ester connecting link accompanied by the loss of acetic acid. The polymerization is usually conducted in the melt, although in some cases the use of an inert suspending medium is reported. The reaction in the melt is carried out either to completion or first to low molecular-weight-oligomer followed by solid state polymerization to high molecular weight. This general approach is employed for the important high-temperature all-aromatic polymers such as Xydar~ (Amoco), Spectral (Hoechst-Celanese), Victrex~ SRP (ICI), and presumably the recent polymers announced by Du Pont and Granmont/Montedison. j CHi O~O 6 CH, + CHIN O¢6 OlI ~ HO 6 4~ 6 0lI | T0~0346U' ¢~t Polyester FIGURE 2.5 Condensation polymerization involving acidolysis. In one commercial case of an aliphatic- aromatic thermotropic LOP- the lower-use-temperature polymer X7G (Eastman) - this condensation step is preceded by the reaction of acetoxybenzoic acid with polytethylene terephthalate) (PET) in an acidolysis reaction in which the PET chain is cleaved. This results in one chain capped with a carbophenyl carboxylic acid moiety and the other capped with an acetoxyphenyl moiety. This is then followed by the acidolysis reaction of the acetoxy and carboxyl end groups, with loss of acetic acid accompanied by other acidolysis reactions. This sequence in short rebuilds the molecular weight with accompanying insertion of oxyphenylcarbonyl mer sequences.
19 Typical copolyester LCPs are shown in Figure 2.6. Patented Aromatic LOP Polyesters Amoco (Carbons) t A <3~ fo~o] J ~ 3~ . touch o o . ·~C , ~_ Hoec~Cehne~e t ~ _ ~ c: to - }~44~c) _fo~] .~ DuPont - J hobo-a¢83 lo~~3} . - -3} ~ 0 0 Jeff> I is ~ 5 r°~°~ r OnzmonMo~e~bon t~ ~ ~ j- ~ o ~ :- ~ 04 l J ~ _ FIGURE 2.6 Representative potential thermotropic copolyesters.
20 LvotroDics The lyotropic LCPs are prepared by solution polycondensation (Figure 2.7~. For many of the extended chain polyamides, solution polymerization in amide solvents is the preferred method. For example, poly~p-phenylene terephthalamide) is synthesized by reacting the appropriate aromatic diamine, p-phenylenediamine in this case, with terephthaloyl chloride in an amphoteric solvent, such as N-methyl-2-pyrrolidone containing a solubility-enhancing salt. Lyotropic LC polyamides, depending on their composition, may be spun directly from their reaction media, or they may be isolated, redissolved, and spun from solutions containing strongly interacting acids, such as sulfuric acid, oleum, etc. The use of a phosphorylation method for the preparation of aromatic polyamides involves the direct condensation of aromatic amino acids or aromatic diamines with aromatic diacids in the presence of an aryl phosphite and organic base. Typical unit structures that yield polyamide lyotropic LCPs are shown in Figure 2.8. C16 ¢6C1 . = H2N~NH2 ~ . ~ iN¢N-~ ¢8 . FIGURE 2.7 Synthesis of polyarylamides.
21 N~C HAN\ HE \, A\ ASH N C~C\ FIGURE 2 . 8 Lyotropic polyamide unit structures . b~c\ of \ C1 N~N~ HE \ H O ~-C V ~ H \ OH N ':~C H: ~ ~ C Con ~0 C Poly(p-phenylene-2, 6-benzobisthiazole) (PBZT) (Figure 2.9) and poly(p- phenylene-2,6-benzobisoxazole) (PBO) were initially prepared at the Air Force Materials Laboratory at Wright-Patterson Air Force Base (see Wolfe, 1988, for a review). PBZT was prepared by the reaction of 2,5-diamino-1,4-benzene- dithiol dihydrochloride with terephthalic acid in polyphosphoric acid. PBO is similarly derived from 4,6-diamino-1,3-benzenediol dihydrochloride. The procedure also works well for A-B type monomers, such as 3-amino-4- mercaptobenzoic acid and 3-~mino-4-hydroxybenzoic acids (Chow et al., 19899. The preparation of "molecular complexes" from blends of polymers derived from A-A+B-B monomers and polymers derived from A-B types has also been reported (Wolfe, 1988~.
22 HX~HH2 H2~-XH r ~ H064/ \~6oH t............ ,X~N~ i~v l N~X~ -n X = 0,S Poly(p phenylene-2 ,6-benzobis " X "azole) rPBX] ~. FIGURE 2.9 Synthesis of PBX polymers. Sidechain Thermotropic LCPs While most of the interest in LCPs is focused on mainchain a-- important class of LCPs contains the mesogenic groups as an appendage (meso~enic core) on the polymer sidechain LCPs (see Attard and Williams, . ~ ~ ~ TV ~ ~w . Synthetic routes to sidechain LCPs have traditionally +~ ~ I ^~.rN polymers, an an appendage ~ car 1986, ~ ., involved polymerization of a vinyl monomer (e.g., acry~ate or mernacry~are under free radical conditions in solution (Figure 2.10~. (The only real difficulty with this reaction is encountered when radical reactive groups are located elsewhere in the monomer.) It is also possible to produce sidechain LCPs through polycondensation. For example, malonate monomers can be converted to polyesters in a polyesterification reaction (Figure 2.11~. This reaction is of special interest for radical reactive groups such as nitroaromatics and stilbenes, which have application in nonlinear optics. Polycondensation of combined sidechain and mainchain LCPs can also be utilized. r - -~
23 ] mesogea spacer AIBN O=< _> ~solvent £~ FIGURE 2.10 Preparation of sidechain LCPs by free radical polymerization. litOgOlit ~ HO _~OH ID \_~oL IS ( 0~)4 L ~ FIGURE 2.11 Preparation of sidechain LCPs by polyesterification. - ~s
24 Reactions on preformed polymer chains allow derivatization of a reactive polymer by attachment of a spacer and a mesogen (Figure 2.12~. ~ Ale Pt catalyst He ~ i~i-ot+=t ~ i2-°t PolysilDsane ~ N /\ S C1 C1 Pol~hosphazene - FIGURE 2.12 Examples of preparation of sidechain LCPs by derivatization of preformed polymers. Ringsdorf and co-workers have described a variety of sidechain LCPs (and combination polymers, i.e., sidechain/mainchain mesogens together) (Ringsdorf et al., 1989 , and references cited therein.) Moreover, these workers have considered ablate as well as prolate mesogenic cores. Thus far we have only considered the latter, which form conventional calamitic phases. Oblate mesogenic cores, on the other hand, form columnar phases wherein the columns may pack on a hexagonal lattice. Such lattice organizations are stabilized in the polymeric forms of these ablate mesogens. These ordered arrays of columns present an interesting and unique state of order in the LC and solid state of organic polymers. Sidechain Lvotropic LCPs For completeness we note that there are examples of sidechain lyotropic LCPs wherein the mesogenic unit is ~mphiphilic (Finklemann, 1987~. There are also recent proposed structures of combination LCPs that are derivative of lyotropic (non-=mphiphilic) rigid-rod polymers (Dowel!, 1989~.
25 UNDERSTANDING AND THEORY Order-D~sorder Transitions in Polymer Fluids For polymers, the influence of molecular anisometry on the propensity of mesophase formation has been explained mainly on the basis of intermolecular and intersegmental repulsion i.e., steric or excluded volume effects (Flory, 1984; Matheson and Flory, 1981~. For the case where no attractions are acting between chains and where the chains are perfectly rigid rods, a statistical thermodynamical theory was formulated (Onsager, 1949; Flory, 1956) more than 40 years ago that laid the foundation for much experimental work in lyotropic systems and for more advanced theories. Today, this is known: For rigid noninteracting rod-like chains in solution, there exists an isotropic phase at low concentrations and an anisotropic phase at high concentrations, with a concentration range of stable biphasic equilibrium in between. The threshold volume fraction vp* for the appearance of an anisotropic phase is approximated by vp* ~ (8/x) [1- (2/x) ] where x is the aspect ratio of the chain. (The aspect ratio is proportional to the molecular weight and can be calculated from estimates of the bulk density and the length and orientation of the constitutional repeat unit. In fact, it is the chain's persistence length that is relevant for the computation of x.) Interacting rigid rod-like chains behave somewhat differently, but this difference becomes gradually insignificant with increasing molecular weights (Flory and Ronca, 1979~. Semiflexible chains are highly extended molecules with significant flexibility that nevertheless impart, unlike their flexible random coiling counterparts, the potential for the formation of anisotropic phases to their solutions and melts. For polymers with a uniform (and small) degree of flexibility along the chain, the "worm- like chain" model is appropriate; thermotropic and lyotropic systems can be interpreted with this model. Polymers that consist of rigid segments joined by flexible spacers can be treated by appropriate modification of the individual rigid rod molecule approach (Boehm et al., 1986), and, if the spacers are sufficiently flexible joints and the rods sufficiently long, they behave roughly as if they were independent (Flory, 19841. Mesomorohic Textures and Structures Mainchain LCPs contain microstructures in their melts and solutions on both a micrometer and a submicrometer scale. The most prominent feature, undoubtedly because it can be observed using optical microscopy, is the micrometer-scale ''domain-like" structure, which is a manifestation of high local molecular orientation. These optically defined domains observed in LCPs are regions of nearly uniform director orientation. The domain- like texture is unlikely to represent an equilibrium texture and is not predicted by any available theory. The character of the subm~crometer boundaries between these apparent domains has not been well characterized; they may consist of strained nematic fluid with some disclinations (the analogue of dislocation faults in a
26 crystal; see, for example Thomas and Wood, 1986) or of more mobile fluid made up of chain ends or of isotropic fluid of different composition from the bulk (or both) (Nicely et al., 1987; Amundson et al., 1987~. Heterogeneity on macroscopic levels is also observed. Lyotropic systems have a biphasic concentration range leading to bulk separated isotropic and LC phases. Although most thermotropic melts that have been studied appear to be uniformly mesogenic on all scales, there are exceptions. The copolyester of 60 mole percent p-hydroxybenzoic acid and 40 mole percent polytethylene terephthalate) has been widely studied and is multiphasic with an isotropic component of different composition from the mesophase. Terminology derived from semi-crystalline polymers - degree of liquid crystallinity - is sometimes used to describe such heterogeneity. However, this terminology is complicated. For example, some thermotropic melts that have been characterized theologically also appear to contain a small fraction of actual crystallites with dimensions of the order of 100 A at the temperatures at which processing is normally carried out (Amundson, 1989; Kalika et al., 1990; Lin and Winter, 19889. These crystallites affect the flow, and they are one likely cause of the observed sensitivity of melt rheology to thermal history. (Large thermal history effects on rheology may also be a consequence of annealing textures [reorganization of domains] or, in some cases, chemical reaction occurring during physical characterization, including decomposition, polymerization, and transesterification.) Rheolo~v The flow of LCPs is affected by the textures described above in a variety of ways. Moreover, the flow itself changes the microstructure. Nevertheless, idealized descriptions of (assumed) homogeneous mesophases are useful to gain insights into flow behavior. Continuum theories of LC rheology developed for low-molar-mass nematics relate the stress at a point to the deformation and director field and contain a description of the dynamics of the director field in terms of the director, stress, and deformation field (Ericksen, 1977 ; Leslie, 19791. Shear flow appears to have little effect on macroscopic director orientation, either experimentally or in the context of continuum theories for anisotropic liquids, but flows with an elongational component are very effective. This is the reason unidirectional orientation (fiber formation) is achieved easily in LCPs. Typical flow in molds contains regions of extensional deformation (in the "fountain flow" near the moving front, for example), so high degrees of orientation are achieved in some "skin" regions of a molded part. The high degree of orientation translates into anisotropic mechanical properties. Flow in molding can lead to regions of rapid change in macroscop ic director orientation including high transverse orientation in the core (Field et al., 1988~. The effect can be analogous to layering of unidirectional sheets . Often the " self - adhesion" of these sheets is poor, resulting in poor lateral strength (delamination).
27 Theory In molecules of low molar mass with mesogenic properties, the molecular anisometry must be appreciable; this asymmetry of shape i s a prerequisite for substances to exhibit liquid crystallinity and has been convincingly interpreted (Flory, 1984; Chandrasekhar, 1977; and De Gennes, 1971~. In addition, intermolecular attractions, especially those strongly dependent on the mutual orientation of the molecules, play a role, but they usually contribute less toward the development of a liquid crystalline phase. Rod-like polymers, or chains with very limited flexibility and a more- or-less linear trajectory, pervade much more of space than their flexible random traj ectory counterparts; i.e., they have much larger radii of gyration. The radius of gyration for a random coil with molecular weight M varies as M° ~ at the theta point and as M06 in a good solvent; it is proportional to M for a rigid rod. As a result, rigid rod-like chains strongly interact with each other at much lower concentrations than flexible polymers do. This is reflected in, for instance, by the large dilute solution viscosity, the large relaxation times, and the small diffusion coefficients exhibited by rod-like chains. Above a threshold concentration, the strong interactions produce a mutual orientation about a common local director; this in turn allows for easy orientation of the rod-like molecules by external influences (e.g., an electric or magnetic field, a shear field), which facilitate processing of such materials into products of well-defined orientation--one of the prime attractions of LCPs. The physical and the transport property differences between random coil and rigid rod polymers increase with increasing concentration, but so do the theoretical difficulties associated with describing their behavior. PROCESSING Solid structures that benefit most from the rod-like molecular structure of LCPs are uniaxial. Easy orientation in the flow field and resistance to chain folding on solidification lead to fibers with outstanding tensile properties. LOP fibers have a very high specific strength and modulus. In addition, some of the LOP fibers exhibit the outstanding temperature performance required for aerospace applications. In this section we consider processing routes to solid LCPs via both thermotropic and lyotropic phases. Thermotronic LCPs The commercial utilization of thermotropic LCPs is a direct function of the ability to process these polymers into cost- effective solid parts . Paramount for achieving this is the understanding and reduction to practice of methods to control molecular orientation in three dimensions. Historically, orientation in polymers has been introduced mechanically. Except for the production of uniaxial shapes with at least one thin dimension (about 40 Am or less), orientation is difficult to control in LCPs. Injection molding, while successful in producing useful and even unique parts, does not allow
28 exploitation of the LOP mechanical property potential. Weak weld lines permeate molded parts. They are due to the lack of molecular mixing across flow-induced interfaces separating director fields with di fferent relative orientations. These "grain" or domain boundaries present a significant problem with complex molded shapes . Orientation- inducing processing techniques other than mechanical (for example, the use of surface epitaxy, electromagnetic fields, or "crystallization templates") may prove to be effective new directions for the production of LOP parts and devices with controlled orientation and stronger weld lines. The potential of new approaches such as cold-forming processes (see Chapter 4) needs to be assessed both in terms of technical feasibility and added manufacturing costs. Lvotropic LCPs Although the formation of liquid crystalline solutions by extended- chain, rod-like polymers has been noted in at least ten classes of polymers (Kwolek et al., 1987), only aramids and ordered aromatic heterocyclics have been extensively examined with respect to LCP-induced, unique solid-state properties. Here we focus on these two classes of polymers as fiber- and film-forming polymers. Kevlar~ aramid polymer has been manufactured by the Du Pant Company since the early 1970s (Kwolek, 1980), while rigid aromatic polybenzobisoxazoles (PBO) and polybenzobisbisthiazoles (PBZT) are currently under consideration for commercialization. Fibers The ability of extended-chain, rod-like polymers to form lyotropic (nemat~c) mesophases under certain conditions (molecular weight, solvent, concentration, and temperature) is critical to their processing into solid parts. Uniaxial solid fibers dominate the commercial utilization of lyotropic LCPs. For example, spinning lyo tropic solutions can result in fibers with very high strength and very high modulus and with nearly perfect orientation without subsequent mechanical processing, i.e., without stretching the extruded and solidified fibers. With these otherwise intractable polymers, such fiber properties have not been achieved by other means. Selection of a spinning method (dry, wet , or dry jet-wet) is determined primarily by the polymer-solvent system and economic considerations. For example , Kevlar~ aramid fiber utilizes a dry jet-wet spinning process. The highly ordered nematic phase in a solution with greater than 18 weight percent of high-molecular-weight poly-p-phenyleneterephthalamide (PPTA) in approximately 100 percent sulfuric acid is retained and perfected by elongational forces in the air gap, and by further drawdown in the coagulating medium. The high (as-spun) chain orientation is directly related to the high fiber tenacity (> 20 gpd) and high (approaching theoretical values after heat treatment) modulus. Spinning conditions for lyotropic aramids and structure- property relationships of the resulting fibers have been studied extensively and are described in patents and literature reviews (see, for example, Schaefgen, 1983; Jaffe and Jones, 1985; Schaefgen et al., 1979; Prevorsek, 1982; Dobb, 1985; and Kwolek et al., 1988~. Subsequent heat treatment of
29 these fibers under tension (but with practically no draw) at 150 to 550°C can result in further improvement of tensile properties and orientation. The heat treatment response is dependent on the structure and molecular orientation of the as-spun fibers and on the method of spinning. Since thermal polycondensation of aromatic diamines and aromatic diacids is at best very slow, the increase in tenacity and modulus upon heat treatment is attributed to increased orientation and crystallization of the polymer chains. Similar findings apply to solid fibers fabricated from polymeric aromatic heterocycles. Since PBZT and PBO are soluble only in strongly interacting acids, fiber formation involves removal of the acid solvent and rapid coagulation of the polymer by a nonsolvent, such as water. As with Kevlar@, the optically anisotropic PBZT or PBO-polyphosphoric acid solutions are dry jet-wet spun into fibers with ultrahigh tenacity and ultrahigh modulus. Polymer preparation in polyphosphoric acid, spinning conditions, and fiber properties of these extended-chain rigid heterocyclic polymers are described in a number of patents and literature reviews (Wolfe, 1988, and references cited therein). Heat treatment is similar to that of aramids except that the temperature range is from 375 to 690°C. There is a wide range of properties, depending on processing conditions. Films The very limited published material on films from fabricated lyotropic polymers may indicate that wet extrusion does not lend itself easily to film formation. Films produced from anisotropic PPTA-sulfuric acid solutions (by wet or dry jet-wet extrusion with uniaxial drawdown) exhibit polymer orientation in the machine direction and highly anisotropic mechanical properties. PPTA films produced with a lubricated conical mandrel between the die and the coagulating bath exhibit equal biaxial orientation and balanced properties. Although heat treatment at 350°C of the latter films results in a significant enhancement of tensile properties and a reduction in voids, these improved properties are very inferior to those of the corresponding fiber (Bodhagi et al., 1984~. Flood and Fellers (1987) have used mandrels of conical, hyperbolic and ogival shapes to obtain a high degree of biaxial chain orientation in PPTA films. Molecular orientation as well as mechanical properties were found to be dependent on mandrel shape and presence. Tensile strength and Youngs modulus were on the order of 30 Kpsi and 1 Mpsi (0.21 GPa and 6.9 GPa), respectively versus 400 Kpsi and 9 Mpsi (2.8 GPa and 62 GPa) for Ke~lar~ 29 fiber. Biaxial films appeared to be more homogeneous compared to uniaxially drawn films. Films prepared from PBZT-polyphosphoric acid solutions by dry jet-wet extrusion also have high axial orientation and high unidirectional properties (tensile strength and modulus of 1.5 and 240 GPa, respectively, after heat treatment) but poor properties in the transverse direction (Feldman et al., 1985~. Also, like other anisotropic films, they exhibit axial splitting. More recently, biaxially oriented films of PBT have been prepared by the use of a specially designed die that allows for biaxial orientation to occur during extrusion of the PBZT-polyphosphoric acid solution (R. Lusignea, presentation to the committee).
30 Solid-State Formine Novel processing of thermoplastics can be defined in many ways but for the purposes of this report will be considered to mean any forming method conducted at temperatures at or below the melting point (solid-to-mesomorphic transition) of the LOP. While such processes are widely used for many semicrystalline polymers, they account for far less volume than traditional melt-processing. In the case of thermotropic LCPs, the sintering process was one of the first to be applied to the hydroxybenzoic acid polymers introduced by the Carborundum Company (Economy et al., 1970a and 1970b; Economy, 19891. However, the exploitation of these techniques has not been noticeable against the rapid growth in, for example, the injection-molding area. The explanation for this is probably not in the behavior of the LOP material itself, relative to random-coil, semicrystalltne polymers. However, relatively little information is available to support the hypothesis that the LCPs can be solid-state formed with the same ease as conventional polymers. The empirical criterion of Aharoni and Sibilia (1978) that a crystal-crystal transition must exist may not apply in the case of rigid rod molecules, although such transitions are thought to be quite common (Field et al., 1988; Hsiao et al., 19881. It should be noted that transitions that could lead to mobility in the crystals may be induced at high hydrostatic pressure; thus some forming processes could succeed while others may fail. Table 2.l illustrates the more common solid-state forming methods (Straw, 1980~. It should be remembered, on scanning this table, that many commercial fabrication processes involve two or more of these, either in sequence or, in effect, simultaneously. For example, the combination of coining and extrusion will yield a decorated container in one step. Also, the mandrel technique has been examined but the results have not been outstanding; modulus values are one to two orders of magnitude below Kevlar~ 29 (see previous section). Many of the advantages of solid-state form' ng result from the avoidance of the penalties of heating a polymer to a high temperature. These penalties include energy costs, time required for cooling, large volume changes, and degradation. For many high-temperature LCPs, the latter may be the most compelling. In addition, it Is possible to introduce a high degree of orientation via solid-state forming, and thereby a substantial improvement in strength. The high optical clarity of cold-formed semicrystalline polymers is often cited as an advantage, but it is not known if this will prevail for LCPs. MECHANICAL PROPERTIES Solid State Morpholo~v The focus herein is on fibers because morphological data on more complex geometries are incomplete. When one tries to analyze the molecular criteria for a perfectly aligned uniaxial system, it is not directly applicable to fibers because of a fiber's fundamental morphological characteristics.
31 TABLE 2.1 Cold-Forming Methods for Polymers Typi Cal C] osest S i nip] e Process Products Deformation Schematic l old drawing tFapbeesr,s, extension rods Bending channel s Pure shear 4~- ~ Cold extrusion Rods, Uniaxial ~| tapes extension ~\q ~ . - ~ - ~ - ~ N~ Hydrostatic Rods, Uniaxial ~ _ | extrusion tapes extension ~| .
32 TABLE 2.1 Cold-Foz-=ing Methods for Polymers (continued) Typical Closest Simple Process Products Deformation Schematic Rolling Sheet, Pure shear tapes eep draw ng Cups Pure shear draw form ng tamp ng) ;= Upsetti ng Knobs, Compressi on (cold heating) nails Matched-die Various Pure shear (stamp) ng) Coining Raised ~ ~ ~ Letters __-- ~ _ at- . · · I I · - ~ ~ · .
33 TABLE 2.1 Cold-Forming Methods for Polymers (continued) Typical Closest Simple Process Products Deformation Schematic Hydroforming Cups extension Fluid sheet Rubber pad Cups [I Stretch forming Cups Biaxial ~ Ad extens i on ~,~ Forging, cold Genesis =~ (compression ~3 Conventional fibers derived from ordinary semicrystalline polymers are a complex aggregate of strong, highly ordered microfibrils and macrofibrils that are usually separated by a weak boundary (Peterlin, 1979) (Figure 2.13~. Therefore, polymeric fibers tend to fibrillate or split into these subelements when bent. Although there are many ways to produce fibers, the final products always exhibit a well-developed fibrillar substructure. Moreover, this tendency toward fibrillation appears to be exacerbated in fiber derived from LCPs, even though extended-chain crystal habits as opposed to thin lamellar crystallites (see inserts in Figure 2.13) prevail in these new polymers (Sawyer and Jaffe, 1986~.
Sam - core - macrofibrils 4~O.S~m skin extended LOP chain crystal habit 50~ OS- By_ ~- 1 / microfibrits semi.- \ crystalline lamellae polymer chain FIGURE 2.13 Hierarchical morphology in fibers.
35 Since the interfibrillar domains are weaker than the fibrils, it is necessary to determine the consequences of fibrillar structure and interfibrillar strength on tensile and compressive properties. An extreme case where the interf~brillar domains are very weak and their volume fraction is approximately 0.25 illustrates the role of interfibrillar strength in tension and compression. Note that, regardless of the lateral dimensions of the microfibrils, the properties in tension would in this case amount to approximately 75 percent of the fibrillar strength. The compressive strength, on the other hand, depends on the cross-sectional area of the microfibril and approaches zero as the macrofibril gets thinner. In tension, the most important factors controlling the strength are the strength of the weakest bond of the molecule and its cross-sectional area. The stronger the individual polymer chain and the smaller its cross-sectional area, the higher is the theoretical strength of the ensemble of perfectly aligned molecules. The interactions between the molecules are of secondary importance and affect only the molecular weight dependence of strength (Termonia and Smith, 1988; Prevorsek, 19881. In compression, the key factors are the torsional and bending rigidity of the molecule and the interactions between adjacent molecules (DeTeresa et al., 19851. The effect of the cross-sectional area in compression is opposite to the effect in tension. Molecules with large cross-sectional areas usually offer a higher potential for compressive strength. Consequently, a polymer molecule designed for optimal performance in tension will generally exhibit low compressive strength, and vice versa. When the molecular criteria are taken into consideration, the following ranking of fibers with respect to their increasing compressive strength potential is obtained: PE < thermotropic polyesters < rigid rod polymers (PBZT, PBO) ~ aramids ~ P\7A < carbon fibers Note that polyethylene (PE), with the highest achieved strength in tension, is the lowest on this compressive strength scale, and that the aramids are above thermotropic and rigid- rod polymers because of their proclivity for interchain hydrogen bonding. Polyvinyl alcohol) (PVA) fibers are a step beyond aramids because of the higher hydrogen bond density. The compressive strength of carbon fibers is high because of their large cross- sectional area and sheet-like structure of the macromolecule. Very recent work on compressive strength of ultra-strong PE fibers and their composites, along with the work on strain rate dependence of PE fiber properties raises the question whether the reported very low compressive strength should be attributed to a specific PE fiber or whether it reflects a true inherent characteristic of uniaxial fibrous PK. The uncertainties exist for the following reasons: The commercially available ultra-strong PE fibers have irregular, frequently flattened cross-section
36 2. On a close inspection, these fibers already show before testing, some compressive damage (kink bands), and 3. Poor adhesion between the PE fiber and various rigid matrices. The first two issues preclude a straightforward determination of the compressive strength for fibers, and the latter from composites. It has been observed that even with the most effective treatments, the fracture path in PE fiber composites never traverses the fibers as is the case in Kevlar~ fiber composites, but follows the fiber-matrix interface. In addition, PE exhibits well known strain rate dependence of mechanical properties. While the creep phenomena of PE were well researched and documented, the strain rate effects at very high rates of deformation were discussed only very recently. Recent studies (Prevorsek, 1989) confirmed the expectation that compressive and tensile properties of PE increased with increasing rate of deformation. This includes the compressive strength and modulus that at the rate of ballistic events and high speed impact appear to exceed the properties of aramids. BLENDS AND COMPOSITES The intimate mixing of two or more polymers to form a new material with a unique property set has emerged as a desirable route to new product development. The resulting blend or alloy (the terms are usually equivalent), if consisting of commercially available polymers, greatly reduces the time and costs associated with new materials development while offering the possibility of a low-cost product with tailored properties and/or improved processibility. Miscible (thermodynamically compatible) blends form a single homogeneous phase and offer the potential for control of transition and processing temperatures. Properties tend to follow the "rule of mixtures I' but, synergies are not uncommon. One significant commercial example of a miscible polymer blend is the polystyrene-poly~phenylene oxide) system developed by GE. Phase-separated (thermodynamically incompatible) blends are much more common, with "rubber-toughened" thermoplastics the most common and successful in this class of materials. Phase-separated blend technology is an effective method for modifying a key resin property (most often toughness) while leaving the majority of resin properties unchanged. The efficiency of property modif ication is a function of the nature of the dispers ion of one polymer in the other (size, geometry) and the degree of adhesion between the phases. If there is sufficient interaction between the blended polymers to cause transition temperatures of the polyphasic system to shift and/or the adhesion between the polymers to improve, the polymers are "partially compatible. n So-called compatibilizing agents are third components, often of low molecular weight, added to immiscible blends to augment dispers ion of the included phase or adhesion between the phases. (Note the term "compatibilize" is often loosely employed by technologists to denote "fine dispersion" of one polymer in another one brought about by adding a third component.)
37 Interpenetrating network (IPN) is the term used to define co-continnous phase morphology (often intimately mixed thermosets). If a thermoses is cross-linked in the presence of a thermoplastic, the system is described as a semi-IPN. If one phase orients during processing such that the matrix resin is reinforced, the system may be described as an "in situ composite" or a ''self-reinforcing polymer." Molecular composite" is the te'm used to describe a molecular dispersion (ideally) of rigid rod polymers (e.g., PBZT) in a conventional polymer matrix (Wang et al., 1988~. Molecular Composites The concept of molecular composites may be appreciated in the evaluation by Flory of ternary systems consisting of (a) the polymer rod (single macromolecule of the PBX type, aramids, copolyesters, etc.), (b) the conventional (random coil) polymer, and (c) the solvent. The evaluation predicts that a critical region will exist wherein there is a single isotropic phase consisting of rods randomly dispersed in the coils. This region is very narrow in its stability boundaries, and the retention of this structure in the solid state depends on "beating the kinetics." Hence, if phase separation can be avoided on solidification, given the extraordinarily high mechanical properties of the individual macromolecular rods, products (including fibers) with excellent tensile and compressive properties should result (a significant fraction [1/33 of rods in such a solid is always in tension). If the composite material could be fabricated into three-dimensional parts, these parts would likely possess the high level of specific mechanical properties currently achievable with LCP-derived fibers and would be extremely attractive for aerospace applications. Many of the motivations for blending LCPs with conventional polymers or with other LCPs are the same that generally make blending an attractive polymer modification option. These motivations include cost reduction, property tailoring, accelerated new-product development, and improved processibility. (Improved processibility focuses on utilizing the low viscosity of the LCP to improve the processibility of highly viscous conventional resins.) The cost-reduction objective is to~provide an LCP-like property set at appreciably less than "pure" LCP property prices. Property tailoring is attractive from two points of view: First, with conventional polymers, LCPs can function as a high-modulus fibrous (macroscopic) reinforcement, and, second, with other LCPs or at relatively low levels of conventional polymer addition, the objective is to mitigate LCP problems such as poor weld line strength or high anisotropy of properties. If a commercially useful family of LCP-containing resins can be defined, the rapid increase of new LCP-containing products will naturally follow. Finally, the blending of LCPs with other-LCPs provides useful data for studying the nature of the structure, morphology, and chain-to-chain interactions in these new LCP materials, while offering the opportunity of improved property sets (DeMeusse and Jaffe, 1988~.
38 ELECTRO-OPTICAL PROPERTIES Herein the focus is on nonlinear optical (NL0) properties of organic systems involving LCPs. There are, however, a number of recent phenomena (generally linear electro-optic) that employ conventional polymers or LCPs together with low-molar-mass LCs. The latter (guest) is dispersed in the host polymer matrix as a microemulsion and the director responds to an applied field thereby changing the refractive index difference between the guest LC and polymer (LCP) host. The phenomena may be adapted to light attenuation and optical switching (roan et al., 1986; Drzaic, 1986~. In the cases of NL0 phenomena, readers may readily sense the intensity of interest in organic polymeric systems, in general (Boyd, 1989), and in LCPs (Williams, 1987), in particular. Relevant aspects of second- and third-order NL0 processes are reviewed here so that readers may consider the potential of LCPs in this active research area, one that is anticipated to yield technologically important advances In the future (Williams, 1987~. Second-Order NLO Processes There are basically two categories of second-order NL0 processes: (a) the linear electro-optic or Pockels effect and (b) parametric processes such as second-harmonic generation (SHG) and sum or difference frequency generation. In the former a d.c. electric field is applied to a medium, which responds by altering its refractive index in proportion to the applied field. In the latter the electric field associated with incident light produces polarization components at other frequencies, which can act as a source of electromagnetic radiation at those frequencies. For a material to exhibit significant second-order NLO responses it must have a noncentrosymmetric structure. In the case of polymers this implies that a polar symmetry axis must be introduced into a medium that would otherwise be nonpolar because of orientational averaging. Electric field poling of thermoplastic polymers at elevated temperatures (above the glass transition temperature, Tg) leads to the introduction of a polar axis (by biasing molecular dipoles in the direction of the applied field) (Meredith et al., 1982; Le Barny et al., 1987; DeMartino, 1988~. This induced polarity can be retained by cooling the polymer to well below its To. The main advantage of introducing liquid crystallinity into polymers for second-order NL0 applications is the enhancement in the degree of polar molecular alignment it can provide; up to - factor of 5 under certain processing conditions. The origin of this enhancement is the effect of the local anisotropic potential associated with the liquid crystalline director on the orientational distribution function of the nonlinear chromophore. Enhanced alignment translates into up to a factor of 5 larger nonlinear coefficient, which in turn can increase the efficiency of processes such as second-harmonic generation by over an order of magnitude For third-order NL0 materials uniaxial alignment associated with the liquid . crystalline director can have a similar enhancement in the coefficient by removing the spatial averaging effects of an isotropic environment on the direction of largest nonlinearity in the chromophore. For commercial applications the stability of retained alignment is of primary concern since
39 critical device characteristics are determined by the stability of parameters related to alignment. Liquid crystallinity can assist in the retention of local alignment because of its highly anisotropic contribution to the local orientational potential energy (D. J. Williams, 1987~. An alternative approach to inducing noncentrosymmetry might be to introduce chirality and the other structural requirements of the ferroelectric smectic c* mesophase into a polymeric structure containing chromophores capable of producing a nonlinear response (Goodby and Leslie, 1984~. Because of its complexity, this approach has not yet been fully examined or exploited, but it may be fruitful for achieving high degrees of order with excellent stability. Most devices designed to take advantage of the Pockels effect operate by retarding the phase of light propagating through the medium with the field- induced refractive index change. One figure of merit (FOM) quantifying the suitability of a material for a particular electro-optic or integrated optic application (Alferness, 1982) is FOM ~ Xf21/n(£ + 1) where x`2, is the first nonlinear coefficient, n the refractive index, and £ the dielectric constant of the electro-optical material. This FOM determines the trade-off between the electric field and path length required to achieve a particular degree of phase retardation. For very-high-speed devices where electric fields are applied via microwave transmission lines, an additional factor emerges. Here the velocity of the microwave pulse Vm = C/E~ must match that of light, Ve = C/n, in the medium over the interaction length required for a given amount of phase retardation to occur (DeMartino et al., 1987~. From these considerations a large refractive index would favor phase retardation. For high-bandwidth devices operating in the traveling wave mode, n2 approximately equals £ of the electrodes at optical frequenc ies . For electro-optic materials such as LiNbO3, n2 and E are very different, so that velocity matching can only be achieved over short distances (Lytel et al., 1988~. From a device design point of view, if the phase retardation must be achieved over short distances, much higher voltages are required. A thorough discussion of device-dependent requirements for electro-optic materials is beyond the scope of this report, but a general list of requirements and desirable characteristics compared to the properties of currently existing materials is presented in Chapter 3. For second-harmonic generation, a separate FOM is required that leads to additional desirable material characteristics. Consider the fraction of power converted to the harmonic frequency, P(2~/P(~), over a certain region in a crystal. It is proportional to the following factors: P¢2 '/p`~' ~ ~x`2, /n3~/~3/2] · P(~)L · f(AkL/2) where n, A, and £ are, respectively, the refractive index, dipole moment, and dielectric constant of the nonlinear materials, and f(AkL/2) is a phase mismatch factor that is periodic in character and whose amplitude is reduced by increasing periodicity. The periodicity of this function is determined by
40 the mismatch in momentum Ak of the fundamental and harmonic waves. A FOM for harmonic conversion is given by FOM = (X ~ ~ /n (~/e ~ I Here it is clear that low-refractive- index materials have a considerable advantage relative to high-index materials, and the quadratic dependence of power conversion efficiency on xt2, puts a tremendous premium on that factor. The input power P(~) and interaction length L are parameters to be played off against the material FOM. Other factors such as birefringence and geometrical optical factors limit the length over which the interaction can be maintained. In waveguided structures designed for optimized propagation of fundamental and harmonic fields, the dependence on interaction length L becomes quadratic, and optical fields can be propagated over long distances, leading to high conversion efficiencies (Zyss and Chemla, 1987~. Because of the lack of materials with suitable properties, as well as processes to fabricate them into suitable waveguides, the potential technological advantages of waveguide SHG have yet to be realized. Third-Order NLO Processes _ Third-order nonlinear processes arise from the nonlinearity in the polarization response of all dielectric media, including conj ugated organic systems . The x-electrons in conjugated organic systems , being loosely bound, contribute much more strongly to the nonlinear response than the more tightly bound core electrons (Rustagi and Ducuing , 1974~. Third-order processes fall into two basic categories. The first is analogous to the Pockels effect, where the refractive index change is quadratically dependent on the applied field, which can be at d.c. or optical frequencies. This can lead to a variety of interesting effects that are manifested in various device designs, including bistable switches, power limiters, and optically driven modulators (Stegeman et al., 19883. The second category of processes involves the interaction of optical fields at different frequencies, where energy can be exchanged between field components in a manner similar to second-order parametric processes. The fields can all be at the same frequency (in contrast to second-order processes, where one of the fields must be at the harmonic, sum, or difference frequency) or at different frequencies. Third- harmonic generation, degenerate four-wave mixing, and real-time holography are examples of such effects (Shen, 1984~. If one of the frequencies or any combination of them matches a resonant process in the molecule or medium, large enhancements in nonlinear response can be achieved. In this case dissipation of thermal energy and the temporal response of the resonant process place constraints on the utility of the process. Momentum conservation must be maintained and can be controlled by the interaction geometry. There is no symmetry restriction for third-order processes, unlike the case for second-order processes, so they are exhibited by all media. In conjugated polymers, where electron oscillations are much larger in the chain direction than perpendicular to it, nonlinear responses are extremely large.
41 Nonresonant third-order nonlinearities are larger in polymers such as polyacetylene (Sinclair et al., 1987), polydiacetylene (Sauteret et al., 1976), and other conjugated polymers than in any other class of materials, including inorganic semiconductors. The response parallel to the chain direction suggests that macroscopic orientation of the polymer should result in a considerably larger response than in an isotropic system. The enhancement factor can be shown to be a factor of five. Conjugated liquid crystalline maincha~n polymers are also known to exhibit high degrees of shear-induced uniaxial alignment, and the expected increases in nonlinear coefficients in appropriate directions have been observed (Rao et al., 1986~. All optical signal processing applications based on third-order NLO fall into two basic categories: parallel and serial. There are two approaches to parallel processing that enable the massive parallelism and interconnectivity of optics to contribute to optical computing and information processing (Gibbs, 19861. The first of these involves the use of simple spatial patterns combined with the switching behavior of nonlinear etalon devices to perform computational functions; these devices are simply miniature resonant cavities where the thickness, refractive index, and reflectivity of the internal surfaces are chosen to provide a destructive interference condition and therefore low transmission through the device. The nonlinear contribution to the refractive index of the medium, as illustrated by the equations below, causes the transmission characteristics to be light-intensity-dependent and capable of exhibiting bistable behavior. The second approach to parallel processing involves the formation of transient holograms generated by two counter-propagating beams in a bulk nonlinear medium to alter the information content of a third beam interacting with the grating thereby producing a new fourth beam. An example of information processing by this method is associative memory (Yariv and Kwong, 1986~. Here an optical mode of a resonator containing a hologram with many messages and a nonlinear medium can selectively amplify a particular message, given only partial information from an input beam. For these applications the primary requirement is for large X(3), where xt3~/~ - n2nO~OC2/3 in MKS units and n = nO ~ n2I where nO is the material refractive index, n2 is the light-intensity-dependent refractive index, and I is the light intensity and ~ is the absorption coefficient. A large value of n2 maximizes the response of the material to small amounts of energy. Because of the inherent parallelism, high degrees of information throughput are generated, and response times in the nanosecond to millisecond time frame are useful. Because of their extremely large resonant nonlinearities, GaAlAs multiple quantum wells and photorefractive crystals such as BaTiO3 have a tremendous functional advantage for this class of applications. Investigations of NLO
42 properties of organic systems, being in a much earlier stage, have not yet demonstrated similar advantage. Serial applications of third-order nonlinear optics involve the use of the intensity-dependent refractive index in waveguided structures to perform rapid switching of bit streams between two or more optical channels. The importance of such devices can be appreciated by realizing that optical fibers can be utilized to transmit information at rates approaching 1 THz. This rate is beyond the capabilities of known light detectors and will require all- optical signal processing before light signals are converted to electronic ones. In waveguides, long interaction lengths can be used to generate phase retardations needed for switching applications so that the premium for materials is put on the speed of the nonlinear response and the optical losses in the material caused by linear optical processes. The decay times of nonlinear responses in semiconductors and quantum well structures, as well as the response time in photorefractive materials, are in general too slow for this type of application. Polymeric materials exhibiting extremely large nonlinearities (albeit much smaller than the resonant nonlinear responses of the inorganic materials that makes them more suitable for parallel processing) may offer the best hope for this important class of applications '£ additional gains can be made in their properties (Stegeman et al., 1987~. REFERENCES Alferness, R. C. 1982. Waveguide electrooptic modulators. IEEE Trans. MTT 30 ( 8 ): 1121- 1137 . Admundson, K. R., D. S. Kalika, M. R. Shen, X. M. Yu, M. M. Denn, and J. A. Reimer. 1987. Influence of degree of polymerization on phase separation and rheology of a thermotropic liquid crystal polymer. Mol. Cryst. Liq. Cryst. A153:271-280. Admundson, K. R. 1989. Investigation of the morphology of liquid crystalline polymers using nuclear magnetic resonance spectroscopy . PhD . dissertation, U. of Cali£., Berkeley. Attard, G. S. and G. Williams. 1986. Liquid-crystalline side-chain polymers. Chemistry in Britain 22~10~:919-924. Bodaghi, H., T. Kitao, J. E. Flood, J. F. Fellers, and J. L. White. 1984. Poly~p-phenylene terephthalamide) films formed from extrusion and coagulation of liquid crystalline sulphuric acid solutions. Characterization of orientation and void structure, annealing, and upgrading of film mechanical propert~es. Polym. Eng. and Sci. 24~4~:242- 251. Boyd, G. T. 1989. Applications requirements for nonlinear-optical devices and the status of organic materials. J. Opt. Soc. Am. B6~4~:685-692.
43 Browstow, W. 1988. Polymer liquid-crystals and their blends. Kunststoffe (German Plastics) 78~5~:411-419. Calundann, G., M. Jaffe, R. S. Jones, and H. Yoon. 1988. Fibre Reinforcements for Composite Materials, A. R. Bunsell, ed. New York: Elsevier . Chandrasekhar, S. 1977. Liquid Crystals. New York: Cambridge Univ. Press. Chapoy, L. L. 1985. Recent Advances in Liquid Crystal Polymers. New York: Elsevier Appl. Sci. Pub. Chow, A. W., S. P. Bitler, P. E. Penwell, D. J. Osbourne and J. F. Wolfe. 1989. Synthesis and solution properties of extended chain poly(2,6- benzothiazole) and poly(2,5-benzoxazole). Macromolecules 22~99:3S14- 3520. De Gennes, P. G. 1971. The Physics of Liquid Crystals. Oxford: Clarendon Pres s . DeMartino, R. N. 1988. U. S. Patent 4,779,961. DeMartino, R. N., D. Haas, G. Khanarian, T. Leslie, H. T. Man, J. Riggs, M. Sansone, J. Stamatoff, C. Teng, and H. Yoon. 1987. P. 65 in Nonlinear Optical Properties of Polymers: Materials Research Society Proceedings, Vol. 109, A. J. Heeger, J. Ornstein, and D. R. Ulrich, eds. Pittsburgh: Material Research Society. DeTeresa, S. J., R. S. Porter, and R. J. Farris. 1985. A model for the compressive buckling of extended chain polymers. J. Mater. Sci. 20(5):1645-1659. Doane, J. W., N. A. Vaz, B. G. Wu, and S. Zumer. 1986. Field controlled light scattering from nematic microdroplets. Appl. Phys. Lett. 48(4):269-271. Dobb, M. G. and J. E. McIntyre. 1984. Properties and applications of liquid- crystalline main-chain polymers. Pp. 61-98 in Liquid Crystal Polymers II/III, M. Gordon, ed. Berlin: Springer-Verlag. Dobb, M. G. 1985. The production, properties and structure of high- performance poly~p-phenylene terephthal~ide) fibres. Pp. 673-704 in Handbook of Composites, Vol. 1: Strong Fibers, W. Watt and B. V. Petrov, eds. New York: Elsevier Sci. Pub. Dowell, F. 1989. Prediction and design of 1st super-strong liquid- crystalline polymers. J. Chem. Phys. 91~2) :1326-1338. Drzaic, P. S. 1986. Polymer dispersed nematic liquid-crystal for large area displays and light valves. J. Appl. Phys. 60~6~: 2142-2148.
44 Economy, J . 1989 . Aromatic polyesters of p-hydroxybenzoic acid. Mo1. Cryst. Liq. Cryst. 169:1-22. Ericksen, J. L. 1977. In the Mechanics of Viscoelastic Fluids, AND Vol. 22. New York: ASME. Field, N ~ D ., R. Baldwin, R. Layton, P . Prayer, and F. Scardiglia . 1988 . Polymorphism in a liquid crystalline polyester based on 4, 4 prime- biphenol, terephthalic acid, and p-hydroxybenzoic acid (1:1:2~. Macromolecules 21:2155-2160. Feldman, L., R. J. Farris, and E. L. Thomas. 1985. Mechanical-properties of as-extruded and heat-treated poly-(para-phenylene benzobisthiazole) films. J. Mater. Sci. 20~8~:2719-2726. Finkelmann, H. 1987. Liquid-crystalline polymers. Angew. Chem. Int. Ed. 26~9~:816-824. Flood, J. E. and J. F. Fellers. 1987. J. Int. Polym. Process . 1: 107. Flory, P.J. 1984. Molecular theory of liquid- crystals. Adv. Polym. Sc'. 59:1-36. Gibbs, H. M. 1986. Workshop on Photonic Logic and Information Processing, Vol. 769, G. M. Bowden and J. G. Duthie, eds. Bellingham, Washington: SPIE . Goodby, J. W. and T. M. Leslie. 1984. Ferroelectric liquid- crystals Structure and design. Mol. Cryst. Liq. Cryst. 110~1-4~:175- 203. Jackson, W. J., Jr. 1989. Liquid-crystal polymers .11. liquid-crystal aromatic polyesters early history and future-trends. Mol. Cryst. Liq. Cryst. 169:23-49. Jaffe, M. and R. S. Jones. 1985. High performance aramid fibers. Pp. 349- 392 in Handbook of Fiber Science and Technology, Vol. 3: High Technology Fibers, Part A, M. Lewin and J. Preston, eds. New York: Marcel Dekker. Kalika, D. S., D. W. Giles, and M. M. Denn. J. Rheology, in press. Kwolek, S. L. 1980. Liquid crystalline polyamides. The Chemist (Washington, D.C.)57(11):9-12. Kwolek, S. L., P. W. Morgan, and J. R. Schaefgen. 1987. Liquid crystalline polymers. Pp. 1-61 in Encyclopedia of Polymer Science and Engineering, 2nd Ed., H. F. Mark, N. M. Bikales, C. G. Overberger, G. Menges, and J. I. Kroschwitz, eds. New York: John Wiley & Sons.
45 Kwolek, S. L., W. Memeger, and J. E. Van Trump. 1988. Liquid crystalline pare aromatic polyamides. Pp. 421-454 in Polymers for Advanced Technologies, IUPAC International Symposium, M. Lewin, ed. New York: VCH Publishers, Inc. Le Barny, P., G. Ravaux, J. C. Dubois, J. P. Parneix, R. Rjeumo, C. Legrand, and A. M. Levelut. 1987. Some new side-chain liquid crystalline polymers for nonlinear optics. Proc. SPIE Int. Soc. Opt. Eng. 682:56- 64. Leslie, F. M. 1979. Theory of flow phenomena in liquid crystals. Adv. Liq. Cryst. 4:1-81. Lin, Y. G. and H. H. Winter. 1988. Formation of a high melting crystal in a thermotropic aromatic copolyester. Macromolecules 21~8~:2439-2443. Lytel, R., G. F. Lipscomb, M. Stiller, J. I. Thackera, and A. J. Ticknor. 1988. Organic electrooptic waveguide modulators and switches. Proc. SPIE Int. Soc. Opt. Eng. 971:218-229. Meredith, G. R., J. G. VanDus en, and D. J. Williams. 1982. Optical and non- linear optical characterization of molecularly doped thermotropic liquid- crystalline polymers. Macromolecules 15~59:1385-1389. Nicely, V. A., J. T. Dougherty, and L. W. Renfro. 1987. Sequence distributions and a phase-diagram for copolymers made from poly~ethylene- terephthalate) and para-acetoxybenzoic acid. Macromolecules 20~3~:573- 578. Peterlin, A. 1979. Mechanical properties of fibrous structure. P. 279 in Ultra High Modulus Polymers, A. Ciferri and I. Ward, eds. London: Applied Science Publishers. Prevorsek, D. C. 1982. Recent advances in high-strength fibers and molecular composites. Pp. 329-376 in Polymer Liquid Crystals, A. C;ferri, W. R. Krigbaum, and R. B. Meyer, eds. New York: Material Science Series, Academic Press. Prevorsek, D. C. 1988. Polymers for Advanced Technologies, M. Levin, ed. New York: VCH Publishers, Inc. Rao, D . N., J . Swiatkiewicz, P. Chopra, S. K. Ghoshal, and P. N. Prasad. 1986. Third-order nonlinear optical interactions in thin-films of poly-p- phenylenebenzobisthiazole polymer investigated by picosecond and subpicosecond degenerate four wave mixing. Appl Phys. Lett. 48~18~:1187- 1189. Ringsdorf, H., R. Wustefeld, E. Zerta, M. Ebert, and J. H. Wendorff. 1989. Induction of liquid-crystall~ne phases formation of discotic systems by doping amorphous polymers with electron-acceptors. Angew. Chemie 28~79:914-918.
46 Rustagi ~ K. C. and J. Ducuing. 1974. Third-order optical polarizability of conjugated organic molecules. Optics Commun. 10~3) :258-261. Samulski, E. T. 1985. Macromolecular structure and liquid crystallinity. Faraday Discuss. Chem. Soc. 79:7-20. Samulski, E. T. 1982. Polymeric liquid-crystals. Physics Today 35~5) :40- 46. Sauteret, C ., J . P . Hermann, R. Frey, F . Pradere, J . Ducuing, R. H . Baughman, and R. R. Chance. 1976. Optical nonlinearities in one-dimensional- conjugated polymer crystals. Phys. Rev. Lett. 36~16) :956-959. Sawyer, L. C. and M. Jaffe. 1986. The structure of thermotropic copolyesters. J. Matl. Sci. 21~6) :1897-1913. Schaefgen, J. R. 1983. Aramid fibers: structure, properties and applications. Pp. 327-355 in The Strength and Stiffness of Polymers, A. E. Zachariades and R. S. Porter, eds. New York: Marcel Dekker. Schaefgen, J. R., T. I. Bair, J. W. Ballou, S. L. Kwolek, P. W. Morgan, M. Panar, and J. Zimmerman. 1979. Rigid chain polymers: properties of solutions and fibers. Pp. 173-201 in Ultra High Modulus Polymers, A. Ciferri and I. M. Ward, eds. London: Applied Science Publishers. Shen, Y. R. 1984. The Principals of Nonlinear Optics. New York: John Wiley and Sons. (This text describes the physical basis for most of the nonlinear opt~cal effects of interest. ~ Sinclair, M., D. Moses, K. Akagi, and A. Heeger. 1987. P. 205 in Nonlinear Optical Properties of Polymers: Materials Research Society Proceedings, Vol. 109, A. J. Heeger, J. Ornstein, and D. R. Ulrich, eds. Pittsburgh: Materials Research Society. Stegeman, G. S., R. Zanoni, and C. R. Seaton. 1988. P. 53 in Nonlinear Optical Properties of Polymers: Materials Research Society Proceedings, Vol. 109, A. J. Heeger, J. Ornstein, and D. R. Ulrich, eds. Pittsburgh: Materials Research Society. Termonia, Y. and P. Smith. 1988. P. 321 in High Modulus Polymers, A. E. Zachariades and R. S. Porter, eds. New York: Marcel Dekker. Wang, C. S., I. J. Goldfarb, and T. E. Helminiak. 1988. Processing of bulk (p -phenylene benzobi sthiazole) nylon- 6 , 6 molecular composites Polymer 29(5):825-828. Wen, X, R. B. Meyer, and D. L. D. Caspar. 1989. Observation of smectic-A ordering in a solution of rigid-rod-like particles. Phys. Rev. Lett. 63:(25)2760-2763. Williams, G. E. 1987. Thermotropic liquid-crystal polyesters. Kunststoffe- German Plastics 77~109:1032-1035.
47 Williams, D. J. 1987. P. 405 in Nonlinear Optical Properties of Organic Molecules and Crystals, Vol. 1, D. S. Chemla and J. Zyss, eds. Orlando Academic Press. Wolfe, J. F. 1988. Polybenzothiazoles and polybenzoxazoles. Pp. 601-635 in Encyclopedia of Polymer Science and Engineering, Vol. 11, 2nd Ed., H. F. Mark, N. M. Bikales, C. G. Overberger, G. Menges, and J. I. Kroschwitz eds. New York: John Wiley & Sons. Wood, B. A. and E. L. Thomas. 1986. Are domains in liquid-crystalline polymers arrays of disclinations. Nature 324~6098~:655-657. Yariv, A. and S. K. Kwong. 1986. Associative memories based on message- bearing optical modes in phase-conjugate resonators. Opt. Lett. 11(3):186-188. Zyss, J. and D. S. Chemla. 1987. P. 173 in Nonlinear Optical Properties of Organic Molecules and Crystals, Vol. 1, D. S. Chemla, and J. Zyss, eds. Orlando: Academic Press.