Document re3ooaKBeGmo3mrn9O8OMKdpa

le) '7 == IPE CSS U ANG7 EWAUNTDTE 1]=EINW WILEY-VCH REVIEWS Nematic Liquid Crystals for Active Matrix Displays: Molecular Design and Synthesis** Peer Kirsch* and Matthias Bremer* Substances forming calamitic mesophases have been known for more than 100 years but only the recent, rapid advance in active matrix liquid crystal display (AM-LCD) technology helped these materials to achieve the crucial position in flat panel display technology they hold today. Due to their high contrast, large viewing angle, and rapid switching times, modern AM-LCDs offer a superior picture quality even compared to conventional cathode ray tubes. Their flatness, low weight, and low energy consumption render them the technology of choice for all kinds of portable devices. Some of the future promises of AM-LCD technology are centered around the development of liquid crystalline materials for the different subtypes of active matrix applications. This development is aimed, on the one hand, towards improved electrooptical and viscoelastic properties; on the other hand, the increasing performance of LCDs leads to extremely stringent reliability demands on the liquid crystals. Responding to these high standards of performance and quality, most liquid crystals for contemporary AM-LCD applications are multiply fluorinated compounds with very high purities, as is typical for materials used in the electronics industry. The synthesis of these superfluorinated materials (SFMs) often requires specialized methods, which, in several cases, had to be introduced for the first time into the canon of industrial production. The immense market pressure, as well as the rapid advance of AM-LCD technology on the side of the display manufacturers, urges an increasing pace of the materials development. This demand for new materials can no longer be fulfilled by conventional trial-and-error approaches. As in the pharmaceutical industry, in the search for new, superior liquid crystals, the purely empirical methods are increasingly supported by a rational design based on computational methods. Keywords: fluorine chemistry liquid crystals mesogens metallation molecular modeling 1. Introduction In the past few years, liquid crystal displays (LCD) have become a nearly inseparable part of our daily life. In 1999, more than two billion LCDs were produced worldwide;[1] half of them small monochrome displays for watches and video games. A smaller but rapidly growing fraction of this overall number are full-color displays capable of high-resolution graphics. Perhaps the most prominent example for this type of [*] Dr. P. Kirsch, Dr. M. Bremer Merck KGaA, Liquid Crystals Division 64271 Darmstadt (Germany) Fax: ( 49) 6151-722593 E-mail: peer.kirsch@merck.de, matthias.bremer@merck.de [**] A part of the work reviewed in this article was performed under the management of the Association of Super-Advanced Electronics Industries (ASET) in the R&D program of the Japanese Ministry of International Trade and Industry (MITI) with support from the New Energy and Industrial Development Organization (NEDO). Another part was supported by the German Bundesministerium fr Bildung und Forschung (01 B 621/1). high information content display are the desktop computer monitors (about 3.6 million units produced in 1999, most of them with a 15 (37.5 cm) diagonal screen), which have started to compete seriously with the long-established cathode ray tube (CRT) monitors. Having overcome some of the major obstacles on the way to larger panel sizes, display manufacturers are now targeting the hang-on-the-wall flat-screen TV as their next strategic goal, which will surely be accomplished within the first half of the next decade. Most of these high-end displays are based on the active matrix (AM) technology,[2, 3] where each picture element (pixel) is controlled separately by a thin film transistor (TFT) integrated on the panel glass. The principle components of an LCD are the glass substrate, covered by an indium tin oxide (ITO) layer as transparent electrode material, polarizer and birefringent compensator films, and color filters. The heart of most commercial LCDs is a 5 6 mm thick film of a nematic liquid crystal. The apparently quite small world production volume of liquid crystals (about 55 t in 1999)[4] is easily explained by the fact that most LCDs require only 0.5 Angew. Chem. Int. Ed. 2000, 39, 4216 4235 WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000 1433-7851/00/3923-4217 $ 17.50+.50/0 4217 REVIEWS 0.6 mg of liquid crystal per square centimeter. Currently around 40 % of the worldwide liquid crystal production is used for AM-LCDs. The broad application of liquid crystal displays became feasible when, in 1971, the twisted nematic (TN) cell was invented by Schadt and Helfrich.[5] In a liquid crystal cell based on the TN mode, a nematic liquid crystalline material, with positive dielectric anisotropy (De) and helically twisted by 908, is placed in an ITO lined glass cell and arranged in an homogeneously aligned layer (that is, parallel to the glass) between crossed polarizers (Figure 1).[612] The orientation of P. Kirsch and M. Bremer the liquid crystal is achieved by an alignment layer of directionally rubbed polyimide within the cell. In order to ensure a homogeneous handedness of the helical structure and, thus, to avoid the formation of domains in the display, a small amount (up to 0.1 %) of a chiral dopant is added to the liquid crystal material.[13] In the off state, the incoming light is polarized and the polarization plane of the light passing through the liquid crystal layer is rotated by 908 and is thus able to exit the second polarizer. If an electrical potential is applied, the liquid crystal helix is deformed and the incoming light cannot pass the crossed polarizers. Thus, in the off state, the cell, which is illuminated from the back, appears white; in the on state, black. A grey scale can be achieved by applying a voltage between the threshold voltage (Vth) and the saturation voltage. At the early 1970s, Gray and co-workers[14] synthesized the first chemically and photochemically stable liquid cystals that exhibited a nematic phase at room temperature (1, Scheme 1). Another commercially important new class of materials also Figure 1. Schematic representation of the working principle of a TN cell in the off (left) and on (right) states. Reproduced from ref. [2] with permission. Scheme 1. Examples of the first chemically stable liquid crystals which have a nematic phase around room temperature. Peer Kirsch was born in Herford, Germany, in 1965. He studied chemistry at Heidelberg University and received his Ph.D. early in 1993 under the supervision of Prof. Heinz A. Staab (Max Planck Institute for Medical Re- search) in physical organic chemistry. After a collabora- tion with Prof. R. Heiner Schirmer (Heidelberg University) on flavin-based enzyme inhibitors, he was from 1994 to 1995 on a Feodor Lynen fellowship of the Alexander von Humboldt Foundation and a fellowship of the Japa- nese Science and Technology Agency (STA) at the Institute of Physical and Chemical Research (RIKEN) in Wako, Saitama. There he worked with Prof. Tomoya Ogawa in M. Bremer P. Kirsch the field of carbohydrate chemistry. He joined the Liquid Crystal Division of Merck KGaA in November 1995. His main research interest is the application of fluorine chemistry for the design and synthesis of liquid crystals. Matthias Bremer received his Diplom-Chemiker from the Erlangen University in 1986 and a Ph.D. degree from the same institution in 1989. In Erlangen, his work was on physical organic chemistry under the supervision of Prof. Paul von Rague Schleyer. From late 1989 until December 1990 he was a Feodor Lynen fellow of the Alexander von Humboldt Foundation at the University of California in Berkeley, where he studied chiral organometallic compounds with Prof. Andrew Streitwieser. He joined Merck KGaA in early 1991 and has worked in the Liquid Crystal research department since. In addition to a number of patents, he has published some 30 research papers on computational chemistry, synthetic and physical organic chemistry, X-ray crystallography, and liquid crystals. 4218 Angew. Chem. Int. Ed. 2000, 39, 4216 4235 Liquid Crystals with a nematic phase around ambient temperature, the cyanophenylcyclohexanes (2), were introduced by Eidenschink[15] in 1977. By systematically utilizing the melting point depression of mixtures of these and similar materials,[1619] it was possible to produce TN-LCDs with a reasonable working temperature range and, in principle, an unlimited lifetime.[20] Soon, the range of commercially applicable liquid crystalline materials was further broadened by other cyano based compounds. The first TN-LCDs were simple, directly addressed segment displays, as still used today in wrist watches, for example. In attempting to increase the information content of the displays, by time-sequential addressing in rows and lines, the limits of the TN cell were soon met: At higher multiplex ratios,[21] a severe loss in contrast occurred due to the ever shorter addressing times. The development of the super-twisted nematic (STN) cell in 1984[22] pushed the practicable limit to higher multiplex ratios but did not lead to a general solution to the problem. Already at the end of the 1960s, the first commercial LCDs still based on the dynamic scattering mode (DSM)[23, 24] had faced the same principal problem of addressability at higher multiplex ratios. As a solution, the use of an active matrix of one thin film transistor (TFT), in combination with a voltageholding capacitor for each pixel, was proposed.[2527] In the long term, the DSM mode itself did not prove feasible but the general idea of active matrix addressing was intensively reevaluated during the 1980s for TN displays,[2] in order to precisely control the applied voltage and thus the optical transmission for each pixel separately. These efforts resulted in the first prototype of a 3 (7.5 cm diagonal) TFT display presented by Sharp in 1986. Since the manufacturing process of the TFT arrays is highly complex and expensive in terms of financial investment and human resources, the larger scale production of AM-LCDs started only 1989 for use in notebook computers.[28] The major drawback of the first AM-LCDs was the strong dependance of the contrast on the viewing angle, which resulted in grey-scale inversion and color shifts in the display when viewed from any than an approximately perpendicular direction. This problem was solved for the classic TN-TFT design by using birefringent compensator films, sometimes in combination with multidomain technology.[29] On the other hand, the last five years showed a technological diversification targeted to further improve the performance of active matrix addressed LCDs: The in-plane switching (IPS) mode[30] and the multidomain vertical alignment (MVA) LCD mode[31] increased the viewing angle to 1608 and resulted in a dramatically improved contrast ratio (up to 1:300 for the MVA-LCD). The plasma-addressed LCD (PALC)[32] uses plasma discharges instead of solid state TFTs to address the pixels and allows the design of flat TV screens up to 42 (105 cm) diagonal. 2. Physical Properties of Nematic Liquid Crystals In order to obtain an optimal performance of any LCD mode, a certain set of physicochemical requirements has to be Angew. Chem. Int. Ed. 2000, 39, 4216 4235 REVIEWS met by the liquid crystal material. These specifications include, among others, the nematic phase range, dielectric anisotropy (De), birefringence (Dn), rotational viscosity (g1), and elastic constants (K1, K2, K3).[8, 3336] The most basic prerequisite for all types of commercially significant LCDs is surely a broad nematic mesophase range of the liquid crystal. Especially for the rapidly spreading automotive uses, such as car navigation displays, the material must remain stably nematic in a wide temperature range, in many cases from 40 to 110 8C. In order to respond to an applied switching voltage, a liquid crystal has to exhibit a dielectric anisotropy (De ek ec),[8, 21, 37] defined as the difference between the dielectric constants parallel and perpendicular to the director of the nematic phase. At the molecular level, De is correlated to the square of the dipole moment m and to the angle b between the dipole moment vector and the effective orientation axis of the liquid crystal molecule.[38] In order to make Equation (1) useful for the design of materials on an atomistic molecular level and for the prediction of De by computational methods, the orientation axis is usually approximated by the long molecular axis (here defined as the rotation axis with the smallest moment of inertia) of the rodlike liquid crystals, even if this leads, in exceptional cases, to inaccurate results. NhF m2 De Da F (1 3 cos2b) S (1) e0 2 kB T For the design of liquid crystals, two implications of the Maier Meier Equation (1)[38] have to be kept in mind: Firstly, independent from the absolute value of the dipole moment m, at the magic angle b 54.78 the contribution from the molecular dipole to De always vanishes, to leave only the very small contribution from Da. Secondly, if a local dipole moment isporiented parallel to the orientation axis, it is by a factor of 2 as effective in generating absolute dielectric anisotropy (j De j ) than in the case of perpendicular orientation. The threshold voltage (Vth) of the electrooptical response of a dielectrically positive liquid crystal mixture (Figure 2)and therefore also the driving voltage of a TN displaydepends Figure 2. Function of an active matrix liquid crystal display (AM-LCD). Left: the electronic scheme; VD denotes voltage on the data line, VG voltage on the gate line, CVH voltage-holding capacitance, CLC capacitance of the liquid crystal layer. Right: the electrooptical response curve of a normally black TN cell (parallel polarizers) showing optical transmission T against the applied voltage U; Vth denotes the threshold voltage of the electrooptical response. [Eq. (2)], besides the dielectric anisotropy (De), on a set of elastic constants (K1, K2, and K3) also.[21, 37] They describe the v u uK2K s Vth pu tK1 3 4 2 p K1 (2) e0 De e0 De 4219 REVIEWS P. Kirsch and M. Bremer different types of elastic deformation of the helical liquid crystal alignment under the influence of an applied electric or magnetic field. Most decisive for TN type cells is the constant K1, the so-called splay deformation.[37] In order to obtain a good contrast ratio in a display based on the normal TN mode (908 twist, normally black mode with parallel polarizers), the relation between the birefringence Dn of the liquid crystal[37] and the cell gap d has to be chosen in a way that in the off state the transmission T, according to the Gooch Tarry Equation (3), is minimized.[21, 39, 40] p p Tu 1 sin2 2 1 u2 ; u 2 d Dn (3) 2 1 u2 l For active matrix displays, the first minimum is usually used, to lead to the best viewing angle independence of the picture quality.[41] This arrangement corresponds to a typical birefringence Dn value of 0.0837 for a cell gap of 6 mm. Recently, many display manufacturers are using birefringent compensation films to improve further the viewing angle of the panel. The optical properties of these films have also to be taken into account for the specification of the birefringence of the liquid crystal material. On the molecular level, the birefringence of a nematic liquid crystal mostly depends on the anisotropy of the polarizability (Da ak ac axx (ayy azz)/2, [Eq. (4)]; the xx axis is the molecular rotation axis with the smallest moment of inertia).[8b, 42] n2 1 N 2DaS n2 1 N DaS n2 2 n2 e a ;0 a ; n2 e 0 (4) n2 2 3e0 3 n2 2 3e0 3 3 For most AMD applications, the switching time has become a decisive factor. The time interval between two video frames at 60 Hz is 16.67 ms. Therefore, multimedia and video applications usually require a fast switching time (t ton toff; [Eq. (5)])[43] of, at least, less than 20 25 ms. g1 d2 g1 d2 ton V 2 on ; toff p2K (5) p2K1 1 1 V 2th On the materials side, the switching time predominantly depends on the rotational viscosity g1,[43] which can vary over a wide range for differently structured liquid crystals. The molecular parameter predominantly influencing g1 is the length of the liquid crystal molecule. The splay elastic constant K1 is a far less variable parameter. A theoretical model quantitatively correlating molecular parameters with the viscoelastic parameters of the nematic phase, such as g1 or K1, would be of value but is not available so far. On the side of display design, the switching time can be reduced by a smaller cell gap d. A consequence of this would be the requirement to increase the birefringence Dn of the liquid crystal, according to Equation (3), in order to obtain an optimized viewing angle independence of the contrast. So far, for the larger panel sizes used in notebook, monitor, or TV applications, the reduction of the cell gap, with sufficient accuracy, is technically feasible to about 4.5 mm. Table 1 shows the most important types of active matrix displays currently in use and their requirements for the physicochemical properties of the liquid crystalline material. The complex property profile required for the optimum performance of each display type cannot be met by currently existing single nematic liquid crystals. The initial task, just to produce a liquid crystal mixture which is nematic around room temperature,[20b] has thus evolved into a multidimensional optimization problem, which is usually approached by mixing 10 to 15 and, sometimes, even more single substances. In general, the development of nematic liquid crystal mixtures is still based on eutectic blocks[20b] of different alkyl homologues of the same classes of materials. Not all components of a liquid crystal mixture have to exhibit a thermodynamically stable nematic phase. Some commonly used materials (for example, the dialkylbicyclohexanes) show only a smectic B (SB) phase or even no mesophase at all. Nevertheless, if applied in moderate concentrations, they do not induce a smectic phase in the mixture and thus allow Table 1. Overview over the material requirements for the most important active matrix addressed LCD types. Technology Applications Material Requirements Characteristics Standard AM-LCD (5V/4V driver) Low Vth AM-LCD (3.3V/2.5V driver) Reflective TFT In-Plane Switching (IPS)[a] VA-TFT Plasma Addressed LCD (PALC) PC monitors, notebook computers Notebook computers, Personal Digital Assistant (PDA), camera view finders Video games, small notebook computers, PDA PC monitors PC monitors TV, advertising De 4 6, Dn 0.085 0.10, TNI 80 120 8C De 10 12, Dn 0.085 0.10, TNI $ 70 8C De 4 8, Dn 0.06 0.07, TNI $ 80 8C De 12 16, Dn $ 0.075, TNI 70 85 8C De $ 4.5, Dn $ 0.08, TNI $ 70 8C De < 0, Dn $ 0.08, TNI $ 70 8C Well established technology; use of compensation films for improvement of viewing angle independence of contrast Allows cheaper and more compact driving electronics than for standard AM-LCD; lower power consumption; very sensitive towards impurities of the liquid crystal material No backlight required, therefore reduction of power consumption by up to $ 90 %; relatively low brightness and contrast Very wide viewing angle and superior picture quality; nitrile based materials can also be used Very wide viewing angle and superior picture quality; very high contrast and fast switching time Large size (1 m diagonal); very good picture quality and fast switching time [a] The IPS mode is not based on a twisted nematic (TN) cell but on an interdigital electrode arrangement.[30] 4220 Angew. Chem. Int. Ed. 2000, 39, 4216 4235 Liquid Crystals utilization of their otherwise quite advantageous properties, such as low viscosity or low birefringence. Even for materials with a stable nematic phase, the comparative evaluation of their application potential is severely complicated by the strong dependence of all anisotropic parameters on the temperature and on the nematic isotropic transition point of each individual compound.[8] One method to obtain a set of comparable characteristics for any kind of material, independent of its phase sequence, is the introduction of so-called virtual parameters: nematic isotropic transition points (TNI,extr), electrooptic (De, Dn), and viscoelastic parameters (g1) are measured in a defined solution of the respective single compound in a standardized nematic host mixture. The values thus obtained are empirically corrected for the change in the order parameter induced by the addition of the single compound and extrapolated to 100 % pure, single compound. Even if, from the theoretical point of view, this method is based on some simplifying assumptions, in most cases the virtual parameters are linearly additive to a degree, sufficient for the calculation and prediction of mixture properties from these standardized single-compound properties. The practical development of commercial liquid crystal mixtures is based strongly on this method. 3. Material Requirements for Active Matrix Displays In addition to the above-mentioned physicochemical requirements, liquid crystals for display applications also have to meet a different set of specifications, which are usually collected under the term reliability. The exact definition of reliability varies for each application and also for each display manufacturer but every liquid crystal mixture, prior to commercialization, usually has to pass tests focusing on the purity, long term chemical and photochemical stability, and specific dielectric properties. In desktop computer monitors, most components of the display are permanently heated by the backlight to sometimes more than 60 8C. Here, the long term stability of the liquid crystal mixture against thermal or photochemical degradation is a decisive point for the material development. Whole classes of materials with otherwise excellent properties, such as tolane (diphenylethyne) or (E)-stilbene ((E)-diphenylethene) -based liquid crystals, are generally not used for AMLCDs due to their limited photochemical stability. High stability towards UV irradiation is especially important for materials used in projection displays with high pressure xenon or mercury lamps as the light source. In an AM-LCD, the driving voltage is applied to each pixel once per refresh cycle. Until the next cycle, the voltage has to be kept constant by the capacitance of the liquid crystal layer and by the voltage-holding capacitor integrated into each pixel (Figure 2). If there is a voltage drop, such as one due to an electric current leaking through the liquid crystal, the contrast of the display is reduced or flicker phenomena are observed.[44, 45] Reliability tests include measurements of the voltage-holding Angew. Chem. Int. Ed. 2000, 39, 4216 4235 REVIEWS ratio (VHR, the ratio of the voltages applied to a pixel at the end and beginning of a defined timespan), the specific resistivity (SR),[46] and the ion density.[47] Additional reliability problems, such as image sticking, are caused by the residual direct current (DC),[48] which is linked to an inhomogeneous distribution of charge carriers within the liquid crystal cell after application of a voltage for a prolonged period.[49, 50] Changes in the ion density, VHR, SR, or of the clearing point under the influence of elevated temperatures or UV irradiation usually indicates chemical degradation of the liquid crystal or of another peripheral material, such as the alignment layer. Insufficient reliability of a liquid crystalline material, especially an insufficiently low VHR, strongly affects the observable picture qualityand thereby also the production yieldof a display panel. Therefore, reliability has become one of the most decisive factors for the development of new materials designed for active matrix displays. Already during the initial development phase of the AMLCD technology, a body of evidence developed that indicated the cyano-based materials shown in Scheme 1 could not meet the stringent requirements on the specific resistivity and the VHR for active matrix displays. Even after extensive purification and removal of potential ionic contaminants, the VHR of most of these materials could not be raised to a sufficient level.[51] Since the early 1980s, a large number of liquid crystals carrying fluorine substituents were synthesized. The initial interest in such materials was aroused by the greater nematic phase range, often dramatically increased, of laterally fluorinated liquid crystals compared to their nonfluorinated analogues. Later, it was found that liquid crystals that derive their molecular dipole moment not from a terminal cyano group but from one or more carbon fluorine bonds the so-called superfluorinated materials (SFMs)[5254]fulfilled the reliability requirements for materials used in AM-LCD. 4. Chemistry of Highly Fluorinated Liquid Crystals Due to the high fluorination degree of most liquid cystals used in active matrix applications, a specialized methodology for the synthesis of these materials was developed. Usually the syntheses commence from aromatic building blocks already carrying the required fluorination pattern. These building blocks are either functionalized via their commercially available halogen derivatives or directly by ortho-metallation.[55] In 1989, when the first liquid crystals with an orthometallation step in their synthetic sequence were introduced to the market, it was also the first time that the necessary low temperature reactions ran on an industrial scale. The activated intermediates can be subsequently treated with a variety of organometallic CC coupling reactions[5658] to carbonyl compounds, (cyclo)alkyl bromides,[59] aromatic halides,[60] or boronic acids.[61] Other, often used, types of CC coupling reactions are the Wittig reaction[62] and the Heck coupling (Scheme 2).[63] 4221 REVIEWS P. Kirsch and M. Bremer additional lateral fluorine atoms in order to augment the molecular dipole moment.[52] As the comparison between 2, 6, and 7 illustrates (Table 2), it is generally much more difficult to achieve a high dielectric anisotropy with materials in which their polarity is based solely on carbon halogen bonds than for cyano-based materials. Table 2. Examples for the first generation of commercially used materials used in TN and AM-LCDs.[65] Scheme 2. General synthetic procedures with typical reaction conditions and yields for the important substructures of nematic liquid crystals: a) 1. ArMgBr, THF; RT, 2 h; 2. Cat. TsOH, toluene; azeotropic removal of water (50 60 %). b) H2, THF, 5 % Pd/C (25 30 % pure trans product). c) Ph3PBr2, CH3CN; 0 350 8C, 5 h (80 85 %). d) 1. Li, ZnBr2, THF, toluene; sonication at 5 15 8C; 2. ArBr, 0.05 equiv [Pd(dppf)Cl2]; RT, 48 h (50 60 % pure trans product). e) 1. nBuLi, THF; 78 8C, 30 min; 2. B(OMe)3; 78 8C 3RT; 3. 2 n HCl (75 80 %). f) ArBr, aq. Na2CO3, toluene, 0.05 equiv [Pd(PPh3)4]; 50 8C, 18 h (70 80 %). g) 1. KOtBu, THF; 10 8C 3RT; 2. H2, 5 % Pd/C, THF (80 90 %). h) 1. ArBr, 0.05 equiv (otolyl)3P, NEt3, CH3CN; reflux, 96 h; 2. H2, 5 % Pd/C, THF (85 90 %). RT room temperature, Ts 4-toluene sulfonyl, dppf 1,1'-bis(diphenylphosphinyl)ferrocene. 4.1. Materials for Standard AMD Applications Currently, most AM-LCD in use as, for example, desktop computer monitors are based on 5 or 4 V driver circuits, that require a threshold voltage Vth of around 2 V for the liquid crystal mixture. The mixtures used for this applications usually have a dielectric anisotropy De value between four and six and a birefringence Dn in the range of 0.085 to 0.10, depending on the cell gap of the display and the choice of compensation films for the improvement of brightness or viewing angle independence of the contrast. Cyano materials such as 1, 2, or 3 are not used due to their low voltage-holding ratio. Therefore, the required De value has to be achieved by other means, such as the dipole moment of terminal carbonhalogen bonds. The electronegativity differences DEN are 0.33 and 1.60 for the carbon chlorine and the carbon fluorine bond, respectively, while the lengths of the aromatic carbon halogen bonds are 170 and 136 pm for chlorine and fluorine, respectively. A phenylcyclohexanebased liquid crystal with one terminal CarF dipole[64] has a De value of only around four. One often applied method to increase the dielectric anisotropy is the introduction of 4222 No. Mesophases TNI,extr De Dn g1 1 C 23 N 35.1 I 4.2 21.6 0.237 112 2 C 31 N 54.6 I 25.7 18.0 0.125 199 3 C 56 N 72 I 49.8 16.8 0.128 4 C 65 SB (52) N 86.2 I 34.3 8.3 0.067 408 5 C 34 N (19) I 14.1 8.2[a] 0.027 6 C 32 I 12.4 4.2 0.108 62 7 C 34 I 56.9 4.0 0.075 27 8 C 6.1 I[b] 9 C 55 N 105.4 I 107.8 6.3 0.144 210 10 C 90 N 158.3 I 161.3 3.0 0.079 156 11 C 46 N 124.3 I 116.0 6.4 0.079 160 12 C 25 SB 53 N 119.1 I 111.7 5.1 0.082 229 13 C 39 N 104.3 I 105.7 5.5 0.067 247 [a] Extrapolated from the Merck mixture ZLI-2857. [b] Data cited from ref. [52]. From the data shown in Table 2, a principal disadvantage of the superfluorinated materials (SFMs) becomes obvious immediately: The clearing points (real nematic isotropic transitions as well as virtual extrapolated values) are much lower than for the cyano materials (2 36 37). On the other hand, the rotational viscosities g1 are also far lower, so that especially the fluorinated two-ring materials (7, 8) become interesting as mixture components in order to reduce the switching time of the display. Comparison of the pair 10 and 11, for example, shows that the lateral fluorination is effective for increasing the value of De by three to four units. On the other hand, the major drawback of the lateral fluorination strategy is a significant decrease of the clearing temperatures by 30 to 40 K for each lateral fluorine substituent. In order to increase De values and avoid the concomittant decrease of the clearing points, the next logical step was to make use of the cumulated carbon fluorine dipole moments of highly fluorinated terminal alkyl groups.[66] As shown in Table 3, terminal perfluoroalkyl chains are well suited to achieve a strong dielectric anisotropy compared to the terminally fluorinated materials. On the other hand, use Angew. Chem. Int. Ed. 2000, 39, 4216 4235 Liquid Crystals Table 3. Examples for liquid crystals deriving their polarity from highly fluorinated terminal alkyl chains.[65] No. Mesophases TNI,extr 14 C 23 I 71.0 15 C 43 S? 109 N 122.9 I 100.0 16 C 89 N (88.6) I 116.1 17 C 127 N (126) I 110.8 [a] Extrapolated from the Merck mixture ZLI-1132. De 9.2 9.1[a] 6.3[a] 7.5 Dn 0.086 0.103[a] 0.083[a] 0.084 of longer fluorinated side chains does not increase De but often results in higher melting temperatures and occasionally also in solubility problems. Despite their reasonably high virtual clearing temperatures, many of the highly fluorinated materials (such as 15, 16, or 17) show only a monotropic nematic phase. The synthesis of the (per)fluoroalkyl-substituted liquid crystals is based on the bromoaromatic building blocks of the type 20 (Scheme 3). They are attached to the basic mesogenic core structures by the general methods depicted in Scheme 2.[67] Scheme 3. Syntheses for highly fluorinated 4-alkyl-bromobenzenes (20), used as starting materials for the liquid crystals shown in Table 3: a) 1. nBuLi, THF; 78 8C, 30 min; 2. EtOOC-Rf (Rf CF3, C2F5); 78 8C, 2 h (70 85 %). b) Diethylaminosulfurtrifluoride (DAST), neat; 0 350 8C (Caution: sometimes decomposes explosively), 5 h (65 75 %). As an alternative concept, which makes use of the advantages of highly fluorinated alkyl chains (namely, their strong dipole moment) while improving the mesogenic properties, fluorinated alkoxy functions were introduced to the basic mesogenic core structures (see Table 4).[68, 69] In spite of their tendency to form smectic mesophases in the pure state, some of the materials based on a terminal trifluoromethoxy group (such as 23) are playing a central role in commercially used standard AMD liquid crystal Table 4. Liquid crystals with fluorinated alkoxy groups.[65] No. Mesophases TNI,extr De Dn g1 21 C 3 I 68.0 7.1 0.083 25 22 C 52 SB 69 N 173.6 I 163.2 5.2 0.086 23 C 39 SB 70 N 154.7 I 146.2 6.9 0.087 142 24 C 119 SB 152 N 168.6 I 152.1 6.5 0.088 25 C 178 I 157.3 3.9 0.074 Angew. Chem. Int. Ed. 2000, 39, 4216 4235 REVIEWS mixtures, together with the 3,4-difluorobenzene-derived materials (11 13; Table 2).[28] Compared to their structurally analogous 3,4-difluorobenzene derivatives, materials with a terminal trifluoromethoxy group possess higher clearing temperatures, which allows the operating temperature range for LCDs to be extended to higher temperatures. The synthesis of the building blocks (Scheme 4) used for this type of materials starts from 4-bromophenol 26 and involves three different types of reactions: Simple nucleophilic substitution (327; probably via a difluorocarbene Scheme 4. Syntheses for highly fluorinated 4-alkoxy-bromobenzenes used as starting materials for the liquid crystals shown in Table 4: a) KOH, CHClF2, dioxane/H2O (1/1); 40 8C, 1 3 bar, 4 h (55 %). b) CCl4, HF; 150 8C, 18 h (61 %). c) (CF3CO)2O neat; reflux, 5 h. d) SF4, HF; 1003175 8C over 10 h (70 75 %). e) NEt3, F2CCFCF3; 75 8C, 45 min (53 %). intermediate),[70] nucleophilic substitution followed by a Lewis acid catalyzed fluorination (328; Swarts-type reaction),[71] nucleophilic addition to perfluoroolefins (329),[72] or fluorination of an intermediate trifluoroacetate (330) with sulfur tetrafluoride to the pentafluoroethoxy function (331).[73] The main characteristic of the materials listed in Tables 2 4 is their strong dielectric anisotropy De. This basic property is sometimes accompanied by relatively low clearing temperatures TNI, high rotational viscosities g1, or a narrow nematic phase range. In order to improve the operating temperature range of the display and to adjust the birefringence Dn and dielectric anisotropy to the exact specifications dictated by the display design, often less polar compounds are used to fine tune the liquid crystal mixture.[28] Some commonly used examples of such materials are listed in Table 5. Compounds 35 and 36 have an olefinic side chain. Such materials often have lower rotational viscosities g1 and a more pronounced tendency to form a nematic mesophase than their saturated analogues.[74, 75] The synthetic strategy leading to alkenyl compounds, such as 36, is based on Wittig reactions, either for the homologation of carbonyl compounds by one to two carbon atoms or for the generation of the double bonds in the side chain (Scheme 5).[76] The comparison between the four-ring materials 37[77] and 38[78] demonstrates the aforementioned expansion of the nematic phase range upon lateral fluorination.[79] 4223 REVIEWS Table 5. Neutral materials used for miscellaneous purposes, such as the exact adjustment of electrooptic properties, reduction of the rotational viscosity, or increasing the clearing temperature of mixtures for LCDs.[65] P. Kirsch and M. Bremer Table 6. Materials deriving their polarity from a 3,4,5-trifluorophenyl group.[65] [a] No. Mesophases TNI,extr De Dn g1 32 C 42 N (37.6) I 28.3 0.1 0.098 44 33 C 22 SB 98 I 56.2 0.5 0.049 31 34 C 49 N 49.6 27.2 0.4 0.048 46 35 C 9 SB 52 N 63.1 I 51.8 0.3 0.054 39 36 C 54 SB 104 N 176.6 I 186.5 1.0 0.097 159 37 C 158 SB 212 SA 223 N 327 330 0.1 0.137 491 38 C 133 N 302 I 299.5 0.0 0.126 651 Scheme 5. Example synthesis of the alkenyl compound 36: a) 1. 4MePhMgBr, THF; 2. cat. TsOH, toluene; azeotropic removal of water (70 %); 3. H2, 5 % Pd/C, THF; 4. 98 % HCOOH, toluene; 5. selective crystallization of the trans isomer from n-heptane (40 60 %). b) 2-(1,3dioxolan-2-yl)ethylphosphonium bromide, KOtBu, THF; 10 8C 3RT. c) H2, 5 % Pd/C, THF (64 %). d) 1. 98 % HCOOH, toluene; 2. MePPh3Br, KOtBu, THF; 10 8C 3RT (66 %). 4.2. Materials for Displays with Low Driving Voltages Recently, on the side of LCD manufacturers there is a trend to reduce the driving voltage, especially for notebook computer displays. A reduction of the driving voltage from 4 5 V to 3.3 V or even 2.5 V allows a denser integration of the electronic components, which is, of course, an advantage for portable devices. The power consumption of the electronic components of the display is also lower, to result in a longer battery lifetime. For the liquid crystal material, a reduced driving voltage requires a lower threshold voltage Vth for the electrooptical response, which is most effectively met by increasing the dielectric anisotropy De. The consequent application of this lateral fluorination concept in order to increase De leads from the 3,4-difluorobenzene derivatives (Table 2) to the 3,4,5-trifluorobenzenebased liquid crystals listed in Table 6.[68, 80] 4224 No. Mesophases TNI,extr De Dn g1 43 C 25.6 I 103.0 6.8 0.034 49 44 C 40.7 N (33.2) I 57.8 12.8 0.137 151 45 C 64.7 N 93.7 I 74.0 8.3 0.073 171 [a] Except g1 values, all data are cited from ref. [80]. The values of the dielectric anisotropies De of this class of substance range between 7 and 13, depending on the mesogenic core structure. The rotational viscosities g1 are also reasonably low, allowing for displays with switching times compatible with video demands. Nevertheless, the lateral difluorination results in a severe drop of real as well as of virtual clearing temperatures of the two and three-ring materials. In order to address this problem, the lateral fluorination concept was combined with polar terminal groups, such as trifluoromethoxy and difluoromethoxy moieties, which are known to induce higher clearing temperatures.[53, 54] In the first generation of liquid crystal mixtures for use in AM-LCD at low driving voltages, compound 47 and its alkyl homologues played an especially dominant role. While the laterally nonfluorinated materials in Table 4 have a tendency to form smectic phases, the laterally fluorinated liquid crystals in Table 7 are all purely nematic. The syntheses of laterally fluorinated (per)fluoroalkoxy compounds are based on ortho-metallation[55, 81] of precursors such as 52.[13] The lithiated species is oxidized to the phenol via a boronic ester intermediate and, subsequently, converted to the desired liquid crystals by the general methods shown in Table 7. Materials with increased dielectric anisotropy by lateral fluorination or difluorination.[65] No. Mesophases TNI,extr De Dn g1 46 C 33 N 144.6 I 123.8 7.4 0.087 242 47 C 62 N 127.5 I 86.9 8.8 0.083 313 48 C 46 N 129.8 I 107.9 9.0 0.089 200 49 C 66 N 118.3 I 86.2 10.5 0.083 279 50 C 49 N 135.9 I 105.3 9.8 0.099 132 51 C 64 N 80.9 I 53.6 10.7 0.088 158 Angew. Chem. Int. Ed. 2000, 39, 4216 4235 Liquid Crystals REVIEWS Scheme 7. Example of the sequential synthesis of a terphenyl- based liquid crystal (58): a) 3-Fluorobromobenzene, Na2CO3, EtOH, toluene, 0.03 equiv [Pd(PPh3)4]; 50 8C, 18 h (92 %). b) 1. KOtBu, BuLi, THF; 100 8C; 2. I2; 100 8C 3RT Scheme 6. Syntheses for liquid crystals 50 and 51: a) 1. nBuLi, THF; 70 8C, 30 min. 2. B(OMe)3; 70 8C 3RT. 3. 30 % H2O2, HOAc; 30 350 8C (50 60 %). b) 1. CF3CHO, (51 %). c) 3,4,5-Trifluorophenylboronic acid, 0.03 equiv [Pd(PPh3)4], Na2CO3, EtOH, toluene; 50 8C, 18 h (98 %). THF; 5 8C, 30 min; 2. DAST, CH2Cl2; 10 8C 3RT, 18 h (35 %). c) 1. NaH, THF; 2. TsOCH2CF3; reflux, 18 h (45 %). d) LDA, THF; 70 8C 3RT (70 90 %). LDA lithium diisopropylamide. reduce the second aromatic moiety to a cyclohexene structure (59), which can also serve as a Scheme 2. The fluoroalkenyloxy derivatives 50 and 51 are synthesized as depicted in Scheme 6.[82, 83] If the lateral fluorination concept is applied to additional scaffold to attach the fluorine atom but does not contribute so much to the anisotropy of the molecular polarizability (Scheme 8).[85] aromatic rings in the mesogenic core structure, materials with dielectric anisotropies De up to about 20 (57 in Table 8) can be obtained.[84] On the other hand, the same example also illustrates the extreme effect multiple lateral fluorination has on the (virtual) clearing temperature, which has dropped below 0 8C. Table 8. Materials with increased dielectric anisotropy by lateral fluorination or difluorination on a second ring structure.[65] Scheme 8. Synthesis of the fluorocyclohexene-based liquid crystal 59: a) 1. BH3 THF; 2 8C 3RT; 2. 30 % H2O2, HOAc; RT 335 8C (66 % isomer mixture); 3. PCC, CH2Cl2; RT, 18 h (85 90 %). b) DAST, CH2Cl2; reflux (86 % crude product). c) KOtBu, THF; 70 8C, 5 h (76 %). PCC pyridiniumchlorochromate. No. Mesophases TNI,extr De Dn g1 56 C 64 I 25.0 15.2 0.135 173 57 C 123 I 9.6 20.5 0.117 58 C 54 N (35.7) I 39.7 17.6 0.219 168 59 C 46 N 69.6 I 60.8 15.1 0.093 210 60 C 65 I 25.8 13.9 0.075 508 61 C 73 N 115 I 89.4 9.8 0.082 301 Materials such as 56, 57, or 58 are usually synthesized by repetitive ortho-metallation, iodination, and subsequent palladium-catalyzed addition of the ring increments as boronic acids (Scheme 7). The introduction of a second aromatic substructure in order to attach the additional lateral fluorine atoms leads to a concomittant increase in birefringence (for example 45 344). If a lower birefringence is required, it is possible to (on paper) Angew. Chem. Int. Ed. 2000, 39, 4216 4235 A recent approach to compensate the unacceptably low clearing temperatures of highly fluorinated liquid crystals is the exchange of axial hydrogen atoms of cyclohexane substructures by fluorine. Materials such as 61 (TNI 115 8C) generally have virtual clearing points 10 15 K higher than their axially nonfluorinated analogues (45: TNI 93.7 8C).[86, 87] The reduction of the threshold voltage Vth of a liquid crystal mixture by exclusively increasing the dielectric anisotropy also increases the average dielectric constant (emean (ek 2ec)/3). Materials with a high emean value promote the dissociation of ionic trace impurities, to result in a decrease in the VHR and specific resistivity.[44, 80] Therefore, a more differentiated approach is the utilization of mixture components with a small elastic constant K1 for the splay deformation, according to Equation (2).[53] A substance class generally suitable for this strategy are fluorinated phenol 4225 REVIEWS esters, such as 67 70 (Table 9). Even if an ester-based mixture has a lower overall dielectric anisotropy, it can still have a lower threshold voltage for the electrooptical response than a conventional ester-free material. Table 9. Strongly polar, phenol ester structures.[65] P. Kirsch and M. Bremer treated with the corresponding phenolate (Scheme 9). The resulting intermediate 77 is coupled to the zinc derivative of a suitable cis-4-alkylbromocyclohexane.[89] No. Mesophases TNI,extr De Dn g1 67 C 56 N 117.2 I 110.6 11.1 0.067 175 68 C 101 N (90) I 82.8 21.4 0.129 216 69 C 52 SB 126 N 187.4 I 172.3 6.2 0.078 199 70 C 50 N 159.3 I 142.1 8.9 0.071 235 Of course, the ester structure is potentially always sensitive towards hydrolysis, thus limiting the chemical stability of the liquid crystal material. For this reason, the more inert difluorooxymethylene bridge was intensively investigated as an alternative, inherently polar linking group for liquid crystals (Table 10).[88] Table 10. Structures deriving additional polarity from a polar difluorooxymethylene link within the mesogenic core.[65] Scheme 9. Synthesis of the difluoroxymethylene-linked liquid crystal 71: a) DAST, CH2Cl2; reflux, 18 h (80 %). b) Br2, CCl4; hn, 14 d (70 80 %). c) 3,4,5-trifluorophenol, NaH, DMF; 50 60 8C, 15 h (51 %). d) 1. cis-4propylbromocyclohexane, Li, ZnBr2, THF, toluene; sonication, RT, 4 h; 2. 77, cat. [Pd(dppf)Cl2], THF, toluene; RT, 18 h (46 %). An alternative method to augment the dipole moment of a polar terminal group is the introduction of a polar ring increment into the mesogenic core structure.[90, 91] Replacement of a cyclohexane substructure by a 1,3-dioxane unit leads to a dramatic increase in the value of De without the pronounced drop of the clearing temperatures, which is observed for lateral fluorination (Table 11). The dioxanebased materials 78 80 are synthesized by acid catalyzed reaction of an aldehyde with a suitable 1,3-diol.[90] Table 11. Structures that derive additional polarity from a polar 1,3dioxane subunit within the mesogenic core.[57] No. Mesophases TNI,extr De Dn g1 71 C 57 SB 70 N 82.9 I 69.7 9.1 0.115 93 72 C 43 SB 116 I 100.2 4.4 0.105 73 C 47 SB 68 N 73.7 I 62.3 6.5 0.120 204 Use of this group, as exemplified by 71, results in a significant increase of De compared to the ethylene-linked reference compound 73, since the dipole moments of the polar terminal trifluoromethoxy group and of the polar difluorooxymethylene bridge are added. If the direction of the linking group is reversed, the dielectric anisotropy is decreased (72). Difluorooxymethylene-based liquid crystals can be synthesized from 4-bromobenzaldehyde 74, which is fluorinated with diethylaminosulfurtrifluoride (DAST), photobrominated, and 4226 No. Mesophases TNI,extr De Dn g1 78 C 20 I 97.6 15.2 0.080 79 C 57 I 88.3 16.1 0.093 80 C 74 N (51.2) I 69.7 17.0 0.068 201 The most severe drawback of liquid crystals carrying lateral fluorine atoms at aromatic substructures still are their low clearing temperatures compared to, for example, the cyano derivatives. Therefore, it is a most important target for application-oriented liquid crystal research to identify new, highly polar terminal groups, which induce reasonably high clearing points and have a minimized interaction with ionic trace impurities. The most polar liquid crystals resulting so far from this line of research are based on sulfur in its highest oxidation state (Table 12).[82, 92] A terminal pentafluorosulfuranyl group[92] has a chemical stability comparable to the trifluoromethyl group. The resulting materials are more polar than the corresponding trifluoromethyl analogues.[93] Judging from its influence on the Angew. Chem. Int. Ed. 2000, 39, 4216 4235 Liquid Crystals Table 12. Highly polar liquid crystals based on sulfur-containing terminal groups.[65] No. Mesophases TNI,extr De Dn g1 81 C 11 I 96.8 12.0 0.087 133 82 C 69 I 79.9 20.2 0.091 145 83 C 121 I 97.8 11.6 0.094 612 84 C 90 N 152.8 I 132.9 12.3 0.087 REVIEWS One possible empirical approach to materials with very low birefringence is the minimization of the number of polarizable structure elements, such as double or triple bonds within the molecule. On the other hand, the indirect contribution of the polarizability to the molecular dipole moment via induced dipole components makes it relatively difficult to obtain a large dielectric anisotropy De with purely cycloaliphatic materials. This effect is illustrated by the trifluormethyl bicyclohexane 90,[96] which has a significantly lower virtual De than its structurally similar phenylcyclohexane analogue 14,[66] that carries the same polar terminal group (Table 13). Table 13. Polar two-ring structures with very low birefringence based solely on alicyclic substructures in the mesogenic core.[65] electrooptic properties of liquid crystals, the pentafluorosul- furanyl group can be considered as a kind of super- trifluoromethyl group. The central building block 85, which carries the pentafluorosulfuranyl function, is most conven- No. Mesophases TNI,extr De Dn g1 iently prepared by direct fluorination of the disulfide 86,[94] 90 C 35 S? (33) I 20.5 5.3 0.051 99 followed by hydrogenation and the Sandmeyer reaction. The 91 C 33 N (18.2) I 13.2 6.9 0.059 89 bromide 85 can be converted to other building blocks, such as 92 C 34 N (31.0) I 15.1 5.6 0.065 65 88 or 89, via its lithiated derivative (Scheme 10). 93 C 82 SB 125 I 0.7 8.3 0.048 98 Scheme 10. Synthesis of the central building blocks 85, 88, and 89 for pentafluorosulfuranyl-based liquid crystals: a) 10 % F2 in N2, CH3CN; 10 8C (80 %). b) 1. H2, 5 % Pd/C, THF; 2. HBr, NaNO2; 5 8C; 3. CuBr; RT 380 8C (46 %). c) 1. tBuLi, Et2O, 78 8C; 2. N-formylpiperidine; 40 8C 3RT (76 %). d) 1. tBuLi, Et2O, 78 8C; 2. B(OMe)3; 70 3 20 8C; 3. 30 % H2O2, HOAc; 20 335 8C, 1 h (58 %). This difficulty was also addressed by combination of two 1,3-dioxanes within the mesogenic core structure (93).[97] Unfortunately, many of the more polar structures with low birefringence suffer from broad smectic mesophases or low solubility (93) in liquid crystal mixtures. Due to the sensitivity of many cyclohexane-derived substances (such as 91) towards base- or acid-catalyzed elimination, the introduction of the polar terminal groups must rely on a completely different methodology than for the aromatic derivatives. Thus, the trifluoromethyl function is most conveniently introduced by reaction of the corresponding transcarboxylic acid with sulfur tetrafluoride (Scheme 11).[96, 98] The aliphatic trifluoromethoxy group is available by oxidative 4.3. Materials with Very Low Birefringence The largest contribution to the overall power consumption of a LCD (about 70 90 %) is due to the backlight used to illuminate the display. Recently, many portable devices, such as small notebook computers, video games, or personal digital assistants (PDAs, such as electronic notebooks and calendars), have therefore been equipped with reflective TFT displays[95] in order to increase the battery lifetime. Since the optical path of the reflected light through a reflective display is different from a usual TN-LCD, there are other requirements for the birefringence of the liquid crystal materials. While a standard TFT display requires a Dn value of roughly around 0.1, an optimal picture quality reflective with TFT display materials need Dn values around 0.06. Angew. Chem. Int. Ed. 2000, 39, 4216 4235 Scheme 11. Synthesis of the bicyclohexane derivatives 90 92: a) SF4, HF, CH2Cl2; 120 8C, 10 h (80 %). b) 1. NaH, THF; 40 8C, 2 h; 2. CS2; RT, 30 min; 3. MeI; RT, 18 h (93 %). c) NBS, 50 % HF/pyridine (20 30 %). d) HCOOH, DCC, CH2Cl2; 0 8C 3RT (74 %). e) P(NMe2)3, CF2Br2, THF/dioxane (10/1); 0 8C 3RT (42 %). NBS N-bromosuccinimide. 4227 REVIEWS desulfurization of a methyl xanthogenate 96 in the presence of a fluoride ion source.[71c, 99] The difluorovinyloxy function is introduced by a modified Wittig-type reaction with a formate ester 97.[97, 100] The key step for the synthesis of the bis(1,3-dioxane) derivative 93 is the trimethylsilyltriflate-catalyzed ketalization[101] of the bis(trimethylsilyloxy) derivative 98 with butyraldehyde in order to avoid oligocondensation of the ambivalent (protected aldehyde and diol) intermediate 99 (Scheme 12). Scheme 12. Synthesis of the bis(1,3-dioxane) derivative 93: a) Xylene, cat. TsOH; removal of EtOH by distillation (53 %). b) 1. LiAlH4, THF; 2. Isomer separation by crystallization from CH3CN (29 %). c) Me3SiCl, NEt3, DMF; 0 8C 3RT (90 %). d) H7C3CHO, cat. Me3SiOTf, CH2Cl2; 78 8C, 30 min (21 %). OTf trifluormethane sulfonate. Since the clearing points of the two-ring materials listed in Table 13 are relatively low with regard to display applications, some additional mixture components with low birefringence Dn, high clearing temperatures, and a broader nematic phase range are required.[102] From the data in Tables 13 and 14, a common problem for the development of liquid crystal mixtures with very low birefringence becomes obvious: Nearly all materials with mesogenic core structures based only on cyclohexane have, in the best case, only very small nematic phase ranges. For this reason, in the future it will remain a very challenging task to develop mixtures for reflective TFT displays with a broad operating temperature range and low driving voltage. Table 14. Materials with very low birefringence and high clearing points.[65] No. 103 104 105 106 4228 Mesophases C 21 SB 265 I C 59 SB 154 N 190.2 I C 110 SB 212 N 325 I C 36 SB 313 N 322.5 I TNI,extr 251.7 196.9 338.1 315.1 De 0.1 1.3 0.6 0.3 Dn 0.057 0.041 0.072 0.077 P. Kirsch and M. Bremer 4.4. Dielectrically Negative Materials Recently, AM-LCDs based on dielectrically negative liquid crystals were introduced to the market. These devices are mostly used for desktop computer monitors and, compared to standard AM-LCD, they offer a superior picture quality due to their high contrast ratio, wide viewing angle, and fast switching time.[28, 103] In dielectrically negative materials, the dipole moment is ideally oriented perpendicular to the orientational axis (b 908).[38] In contrast to a standard TN cell with a positive De material, the dielectrically negative liquid crystals are oriented homeotropically (namely, perpendicularly) to the alignment layer in the cell. If a voltage is applied, the molecules experience a torque towards an orientation perpendicular to the applied electric field. In principle, there are two different ways to attach a perpendicular dipole to a typical liquid crystal structure.[19] In materials containing a cyclohexane substructure, the axial hydrogen atomsdirected in an exactly perpendicular direction to the molecular long axiscan be replaced by polar moieties, such as cyano groups[18] (such as in 5) or fluorine.[86, 104, 105] In aromatic substructures, the unilateral 2,3positions can be substituted pairwise by fluorine, in a way that the dipole vector components in the direction of the long molecular axis cancel. Materials of the latter type (Table 15)[106] are typical components of the dielectrically negative liquid crystal mixtures used, for example, in multidomain vertical alignment thin-film transistor (MVA-TFT) displays.[103] Table 15. Examples for the first generation of dielectrically negative liquid crystals suitable for AM-LCD.[65] No. Mesophases TNI,extr De[a] Dn g1 107 C 14 I 49.2 1.8 0.086 62 108 C 49 N (12.9) I 16.5 6.2 0.099 110 109 C 67 N 145.3 I 139.0 2.7 0.095 218 110 C 79 SB (78) N 184.5 I 175.4 5.9 0.096 413 111 C 74 SA 86 N 170.7 I 190.6 5.3 0.146 344 112 C 80 I 44 7.3 0.133 637 [a] Extrapolated from the Merck mixture ZLI-2857. The synthesis of this type of materials starts with 1,2difluorobenzene 113, which is derivatized in successive orthometallation steps (Scheme 13).[55] The replacement of one fluorine atom by the more polar trifluoromethyl group[19] results in a further increase of the (negative) dielectric anisotropy but also in a severe drop of the virtual clearing temperature. This very pronounced effect Angew. Chem. Int. Ed. 2000, 39, 4216 4235 Liquid Crystals REVIEWS Table 16. Dielectrically negative liquid crystals based on axially fluorinated cyclohexane subunits.[65] Scheme 13. Synthesis of the dielectrically negative materials 109 and 110: a) 1. nBuLi, THF, 70 8C; 2. B(OMe)3; 70 8C 3RT; 3. 30 % H2O2, HOAc; RT 350 8C (90 %). b) EtBr, K2CO3, acetone; reflux, 18 h (94 %). c) 1. nBuLi, THF, 70 8C; 2. MeI; 70 8C 3RT (77 %). d) 1. nBuLi, THF, 70 8C; 2. 4-propylbicyclohexan-4'-one; 3. Cat. TsOH, toluene; azeotropic removal of water; 4. H2, 5 % Pd/C, THF (109: 35 %, 110: 38 %). is presumably due to a decrease of the length-to-breadth ratio of the molecule 112 compared to its analogue 111 (see Scheme 14 for synthesis). No. Mesophases TNI,extr De[a] Dn g1 121 C 75 SB 94 I 79.6 122 C 13 SB 37 N 79.8 I 56.5 123 C 78 SB 105 I 76.5 124 C 74 SB (70) N 83.0 I 37.3 125[b] C > 120, dec. 2.2 0.054 1.9 0.057 41 4.6 0.051 198 3.8 0.061 156 [a] Extrapolated from the Merck mixture ZLI-2857. [b] Compound 125 could not be fully characterized due its insufficient solubility. Scheme 14. Synthesis of the fluorotrifluoromethylbiphenylcyclohexane derivatives 112: a) 1. nBuLi, THF; 70 8C; 2. B(OMe)3; 70 8C 3RT; 3. 30 % H2O2, HOAc; RT 350 8C; 4. EtBr, K2CO3, ethylmethylketone; reflux, 18 h (58 %). b) 1. nBuLi, THF; 70 8C; 2. I2, THF; 70 8C 3RT (76 %). c) 1. nBuLi, Et2O; 70 8C; 2. Me3SiCl, Et2O; 70 8C 3RT (70 %). d) 1. nBuLi, KOtBu, THF; 78 8C; 2. I2, THF; 78 8C 3RT (59 %). e) 1. 4-Propylcyclohexylphenylboronic acid, 0.05 equiv [Pd(PPh3)4], 2 n aq. Na2CO3, toluene; 50 8C, 18 h (60 %); 2. DMF, CsF; 80 8C, 1 h (quant.). tion of double bonds into the side chains suppresses the smectic phases and a nematic phase emerges, as demonstrated by 122 and 124.[75, 105] The synthesis of axially fluorinated cyclohexane derivatives is generally based on Wittig olefinations and a subsequent hydrofluorination step (Scheme 15). Surprisingly, the selec- A less favourable property of the materials listed in Table 15 is their relatively high rotational viscosity g1, which limits the switching time. Therefore, as an alternative approach, liquid crystals based on axially fluorinated cyclohexane structures were extensively studied (Table 16).[86, 104] The first objective, the reduction of g1, was clearly achieved. Further, as already mentioned before, the (virtual) clearing points are increased significantly compared to their axially nonfluorinated analoguesin the case of the bicyclohexane derivatives (such as 121) by 50 to 70 K. Repetition of ethylene linked ax-fluorocyclohexane units (121 3123 3125) results in a further increase of the (negative) De values. With three repetitive units, the concept meets its limitations, since 125 is not sufficiently soluble in typical liquid crystal mixtures in order to measure properties for extrapolation.[65] The doubly alkyl-substituted bicyclohexanes 121 have a strong tendency to form smectic B phases and, in higher concentrations, cause liquid crystal mixtures to become smectic on extended storage at low temperatures. Introduc- Angew. Chem. Int. Ed. 2000, 39, 4216 4235 Scheme 15. Synthesis of the liquid crystals 121 and 124: a) H7C3PPh3Br, THF; 10 8C 3RT, 2 h (80 90 %). b) 20 equiv 70 % HF/pyridine, CH2Cl2; RT, 18 h (35 %). c) 4-Vinylcyclohexanone, KOtBu, THF; 10 8C 3RT, 4 h (95 %). d) HCOOH, toluene ; RT, 18 h (73 %). e) H2CCHCH2CH2PPh3Br, THF; 10 8C 3RT, 2 h (55 %). f) 1. 4 equiv 70 % HF/pyridine, CH2Cl2; 15 8C 310 8C, 1 h; 2. Recrystallized three times from n-pentane (6 %). tivity of 70 % HF/pyridine (Olahs reagent)[107] is so high, that even tetraenes such as 131 are smoothly and selectively converted into the difluorinated bis(alkenyl) compound 124. 4229 REVIEWS 5. Tools for the Rational Design of Liquid Crystals Much of the materials research in the liquid crystal field is characterized by empirical structure property relationships. Some of those are straightforward, such as the increase in De upon the introduction of stronger polar terminal substituents or the lowering of Dn by minimizing the number of polarizable groups. Others are perhaps less obvious, such as the suppression of smectic phases by lateral fluorination or the enhancement of the nematic phase range after the replacement of fluorinated aromatics by heterocyclic analogues.[108] Other properties, such as rotational viscosities, elastic constants, or the nature and range of mesophases, are only related in an indistinct and unpredictable manner to molecular structures. The liquid crystalline state is governed by weak intermolecular forces, in the order of a few kcal mol1, and the prediction of bulk properties for a liquid crystal could rely only on a highly accurate description of these interactions. For the simulation of molecular ensembles, Monte Carlo or molecular dynamics calculations have been used, and simulations for simple model liquid crystals have been carried out successfully.[109] Unfortunately, substantial simplifications must be made at the level of the individual molecule to treat the interactions of the ensemble at all. Even with these restrictions, computer times are enormous and the methods are still far from being routine modeling tools. In addition, careful parameterization of the force-field and/or the anisotropic potentials used in the simulation is crucial for reliable results. On the other hand, classical computational chemistry[110] deals with isolated molecules in the gas phase but allows a very detailed and accurate description of their electronic and geometric structures. The methods range from simple molecular mechanics[111] and semiempirical methods[110] to ab initio[112] and density functional (DFT)[113] theory. P. Kirsch and M. Bremer Figure 3. The four lowest energy conformers (within 0.6 kcal mol1) of the bis(1,3-dioxane) derivative 93. The global minimum is shown in the upper left corner (MMFF 94 force-field calculations). recently been applied for the prediction of clearing temperatures for a number of liquid crystals.[115] However, the variation in chemical structure must not be too large and the properties of structures not radically different from those used in the training set, otherwise the properties will probably not be predicted correctly. 5.2. Molecular Modeling of Electrooptic Properties Some of the key parameters for a liquid crystal mixture are the dielectric and optical anisotropies De and Dn. The dielectric anisotropy De is related to molecular properties via the Maier Meier equation [Eq. (1)][38] and Dn can be obtained from the molecular polarizability a from the Vuks equation [Eq. (4)].[42] The molecular dipole moment m and the polarizability a can be obtained from a semiempirical, ab initio, or DFT calculation (Figure 4).[116120] 5.1. Shape Analysis Nematic liquid crystals, as they are used in most LCDs, are rodlike molecules. Deviations from a linear structure towards bent arrangements qualitatively result in a decrease of the clearing temperatures. It is therefore of interest to rapidly form an impression of the three-dimensional shape of novel materials that are candidates for chemical synthesis. In addition, in most cases, there will be conformational flexibility instead of complete rigidity; that is, in the condensed phase at room temperature, the thermal energy available will be sufficient to populate conformers higher in energy than the global minimum. Modeling at the force-field level of theory gives a quick overview of such conformers. For example, four low energy conformers of the bis(1,3-dioxane) 93 (Table 13) are shown in Figure 3. In addition to these mostly qualitative aspects, it is possible to compute geometrical and electronic structure parameters. From these molecular descriptors, quantitative structure activity relationships (QSAR)[114] can be derived. This has 4230 Figure 4. A three-dimensional representation of the liquid crystal CCP30CF3 23. The calculated (AM1) anisotropic polarizabilities are axx 255.7, ayy 153.2, and azz 159.6 a.u.; the dipole moments are mx 2.95, my 0.08, and mz 0.52 D; the angle b towards the orientational axis, approximated by the x axis of the molecule, is 10.08. The electrooptical anisotropies derived from these values are DeAM1 5.8 and DnAM1 0.099. The experimentally determined virtual parameters are 6.9 and 0.087 for De and Dn, respectively. (1 a.u. 0.14818 3.) Angew. Chem. Int. Ed. 2000, 39, 4216 4235 Liquid Crystals REVIEWS For conformationally flexible molecules, the prediction, especially that of De, is somewhat more complicated, since different conformers may possess very different dipole moments even though they are similar in energy. It is then necessary to systematically screen the conforma- tional space and perform a Boltzmann-type weighting of the calculated quantities. An extreme case in point is the polar terminal group difluoromethoxybenzene. Figure 5 shows three conformers located at the B3LYP/6-311 G**//B3LYP/6-311 G** level of theory togeth- er with their relative energies and dipole mo- ments. Although these calculations are at an adequate level, their predictive power is rather low. After all, we are dealing with the condensed Figure 6. The correlation between semiempirically calculated (AM1) versus extrapolated experimental electrooptical anisotropies (84 samples). phase where energy differences of less than 1 kcal mol1 in the gas phase are certainly insig- nificant. Fortunately, in most cases, there is only one minimum within a few kcal mol1 and the conformer problem disappears. In order to assess the value of the method, calculations were carried out on a representative test set of liquid crystals that provides a wide range of electrooptical anisotropies. The pre- dicted and experimental electrooptical anisotro- pies for this set are shown in Figure 6. Only single conformers were used for the calculations. The influence of the quantum chemical meth- od on the quality of the prediction can easily be checked by moving to higher levels of theory (Figure 7). Hartree Fock ab initio theory with the split valence 6-31G* basis set is currently the Figure 7. The correlation between ab initio (HF/6-31G*//HF/6-31G*) calculated electrooptical anisotropies versus extrapolated experimental data (26 samples). practical limit for molecules the size of liquid crystals. Calculation times for ab initio geometry optimization Although the ab initio approach is more rigorous theoret- and polarizability calculations are two to three orders of ically, there is a large systematic overestimation of De and a magnitude longer than those of the AM1 semiempirical (smaller) overestimation of Dn. The most likely explanation method. for this poor performance is the neglect of electron correla- tion and the use of a too small basis set. It is well known that Hartree Fock ab initio theory systematically overestimates molecular dipole moments, while DFT and MP2 ab initio and even AM1 reproduce experimentally determined values quite well.[121] Unfortunately, the computational problem cannot be tackled adequately at present. However, the performance of the semiempirical method (which includes some electron correlation effects via the parameterization) is quite satisfac- tory for standard organic materials, such as most liquid crystals. However, semiempirical theory understandably fails with molecules it was not parameterized for, which became evident with hypervalent sulfur fluorides.[92] Figure 5. Three low energy conformers of difluoromethoxybenzene at the B3LYP/6-311 G**//B3LYP/6-311 G** level of theory: The relative energies and corresponding dipole moments are 0.4 kcal mol1 and 0.98 D (upper left), 0.0 kcal mol1 and 1.12 D (upper right), and 0.1 kcal mol1 and 2.93 D (below). Angew. Chem. Int. Ed. 2000, 39, 4216 4235 5.3. Modeling-Assisted Prediction of Reliability Parameters One of the characteristic parameters for the reliability of a liquid crystal mixture is a high VHR value. The VHR of liquid crystal mixtures has been much improved after the introduc- 4231 REVIEWS P. Kirsch and M. Bremer tion of superfluorinated materials (SFMs)[5254] but there still remains an ob- servable leakage current, which suggests the presence of ions or contamination in the cell. Modeling the solvation of ions by liquid crystal molecules was supposed to provide a better understanding of the interactions involved and possibly allow predictions regarding the solvation potential of struc- tures not yet synthesized. Many liquid crystals are polar organic materials that possess large dipole moments and lone electron pairs which can interact with ions. In principle, the solvents macroscopic di- electric constants, dipole moments, and polarizabilities can affect solvation but only a weak correlation with experimental data was found.[122] For our modeling approach, the semiempirical AM1 method was used because of the relatively large size of liquid crystal Figure 8. Energy profile for the interaction of a positively charged sparkle with benzonitrile. DH denotes the calculated heat of interaction (AM1) and d the distance of the sparkle from the cyano nitrogen atom. molecules. The MOPAC program[120] offers positively charged artificial elements, the sparkles, which Another case in point, typical for SFM-based materials, is can be conveniently used to model, for example, alkali metal 1,2,3-trifluorobenzene. As for benzonitrile, very little change ions. Sparkles may be regarded as unpolarizable ions of a in molecular geometry is observed but the change in atomic fixed 1.4 diameter. They do not contribute to the orbital charges is now reduced. Consequently, the heat of interaction count and cannot accept or donate electrons. Since the is only 4.4 kcal mol1, which implies a smaller solvation strongest interactions are expected for positive ions (negative power. It is well known, that for active matrix applications, ions are generally larger and their charge is better delocal- nitriles are not suitable because of poor voltage-holding ized) only singly positively charged sparkles were considered. behavior. The empirical and perhaps initially puzzling finding, As a suitable measure for the ion solvation power of a that only the recently introduced SFMs are suitable for active material, we chose the calculated (AM1) heat of interaction of matrix displays, is nicely corroborated by this simple model of a liquid crystal molecule with a positive sparkle. ion solvation. Figure 8 shows the energy of a system consisting of Finally, instead of the model systems discussed so far, a set benzonitrile as a simple model liquid crystal and a sparkle of complete liquid crystals for TN, as well as AMD, as a function of distance. applications was used. Immediately, the problem of more The calculation shows that there is a steep energy minimum than one polar group in a given molecule arises. For example, at a nitrogen sparkle distance of about 3.1 and an in esters of the type shown in Figure 9, the sparkle may interact interaction energy of almost 11 kcal mol1. The steepness with the ester group (a) or with the terminal nitrile (b). of this energy curve leads to smooth convergence of the geometry optimization: Placing the sparkle in the initial configuration directly over the aromatic ring instead of in the ring plane leads to the same structure after energy minimi- zation. The geometry of the organic molecule is altered little by the presence of the sparkle; bond lengths change by less than 0.005 and bond angles by less than 0.38. The charge distribution, however, is significantly perturbed by the presence of the charged particle, although the total charge in the organic part of the complex remains zero. A simple valence-bond resonance structure for benzonitrile (Figure 8) illustrates these changes qualitatively but in a strikingly simple manner. Electron density is easily shifted in the delocalized p-system and this induces an extra dipole moment in addition to the one already present. Thus, the overall effect may best be described as a combination of ion dipole and ion induced dipole interactions. Figure 9. Two local minima for the interaction of a positive sparkle (yellow) with the liquid crystal ME2N.F (132). 4232 Angew. Chem. Int. Ed. 2000, 39, 4216 4235 Liquid Crystals Both structures are energy minima but the heat of interaction with the nitrile group is 12.6 kcal mol1, whereas, with the ester carbonyl group (a) it is only 8.0 kcal mol1. Under equilibrium conditions, only structure (b) will exist and (a) can be safely neglected. In a similar way, the structures and energies of other typical liquid crystals were calculated and the data are shown in Figure 10. Although in essence a very simple electrostatic model is used, the results are often counterintuitive: The calculated heats of interaction do not necessarily have a direct relationship with the dipole moment or the polarizability; for example, the computed dipole moment is 7.1 D for ME2N.F but only 3.9 D for PCH-3 and 3.1 D for CCP-3OCF3 23. Admittedly qualitative in nature, this simple model allows a preliminary assessment of the AMD suitability of new liquid crystals before their actual synthesis. REVIEWS modeling offers important tools for the synthetic organic chemist to guide synthetic strategies by providing information on conformational space, thermodynamic stabilities, activation energies, and the ground-state and transition structures of molecules. A decisive driving force for the future development of new AM-LCD technologies will be cost reduction in mass production. The reduction of the power consumption, such as by use of reflective displays, will be another important goal. Within the next few years, the consumer will have to choose from a broad spectrum of different devices based on AMLCDs, ranging from large screens for TV applications to very small displays with increased resolution for viewfinders, mobile phones, or virtual reality goggles. For the design of liquid crystals, the challenge to explore materials with extreme electrooptical and viscoelastic properties will con- tinue, with ever more stringent quality requirements than currently required. Received: July 16, 1999 Revised: May 4, 2000 [A 353] Figure 10. Calculated (AM1) heats of interaction (DH) and voltage-holding typical liquid crystals. 6. Conclusion and Outlook During the time span between their first application in simple LCDs and their use in modern AM-LCDs, structural as well as physicochemical characteristics of liquid crystalline materials have changed drastically. The parallel development of various types of active matrix technologies, with their sometimes very different demands with regard to the physical properties of the materials, led to a broad diversification of the chemical structures and also of the synthetic methods used for their production on an industrial scale. The increasing pace of the technological advance in the display field has created a powerful incentive for the materials manufacturers to put the design and development of new liquid crystals on a more rational basis. Although it might not be expected offhand, some important properties of the condensed liquid crystalline phase can be predicted by modeling the isolated molecules. There are relationships between molecular parameters and experimental clearing temperatures, electrooptical anisotropies can routinely be calculated for any structure, and a simple model for the solvation of ions yields results in agreement with important liquid crystal reliability parameters like VHR or residual DC. In addition to these properties specific to liquid crystals, [1] Source: Stanford Resources, Inc. Further information is available in the internet: http ://www.webcom.com/sr/. [2] S. Kobayashi, H. Hori, Y. Tanaka in Active Matrix Liquid Crystal Displays in Hand- book of Liquid Crystal Research (Eds.: P. J. Collings, J. S. Patel), Oxford University Press, New York, 1997, pp. 415 444. [3] K. Nakao, EKISHO J. Jpn. Liq. Cryst. Soc. 1999, 3, 62 63. ratios (VHR) for some [4] Merck KGaA, internal estimate. [5] M. Schadt, W. Helfrich, Appl. Phys. Lett. 1971, 18, 127. [6] F. Reinitzer, Monatsh. Chem. 1888, 9, 421. [7] O. Lehmann, Flssige Kristalle, sowie Plastizitt von Kristallen im Allgemeinen, molekulare Umlagerungen und Aggregatzustandsn- derungen, Engelmann, Leipzig, 1904. [8] a) P. G. de Gennes, J. Prost, The Physics of Liquid Crystals, Oxford University Press, London, 1993; b) W. H. de Jeu, Physical Properties of Liquid Crystalline Materials, Gordon and Breach, New York, 1980. [9] D. Demus, H. Demus, H. Zaschke, Flssige Kristalle in Tabellen, VEB Deutscher Verlag fr Grundstoffindustrie, Leipzig, 1974. [10] D. Demus, L. Richter, Textures of Liquid Crystals, Verlag Chemie, Weinheim, 1978. [11] Handbook of Liquid Crystals (Eds.: D. Demus, J. Goodby, G. W. Gray, H.-W. Spiess, V. Vill), WILEY-VCH, Weinheim, 1998. [12] V. Vill, LiqCryst 3.1Databases of Liquid Crystalline Compounds, LCI Publisher GmbH, Hamburg, 1998. [13] a) General review: G. Solladie, R. G. Zimmermann, Angew. Chem. 1984, 96, 335 349; Angew. Chem. Int. Ed. Engl. 1984, 23, 348 362; b) Technical applications: D. Pauluth, A. E. F. Wchtler in Synthesis and Application of Chiral Liquid Crystals in Chirality in Industry, Vol. II (Eds.: A. N. Collins, G. N. Sheldrake, J. Crosby), Wiley, New York, 1997, pp. 264 285. [14] a) G. W. Gray, K. J. Harrison (Ministry of Defense, Great Britain), GB-B 1433130, 1972 [Chem. Abstr. 1974, 81, 96988]; b) G. W. Gray, K. J. Harrison, J. A. Nash, Electron. Lett. 1973, 9, 130. [15] R. Eidenschink, D. Erdmann, J. Krause, L. Pohl, Angew. Chem. 1977, 89, 103; Angew. Chem. Int. Ed. Engl. 1977, 16, 100. [16] D. Demus, H.-J. Deutscher, F. Kuschel, H. Schubert (VEB Werk fr Fernsehelektronik), DE-B 2429093, 1975 [Chem. Abstr. 1977, 89, 129118] . [17] R. Eidenschink, D. Erdmann, J. Krause, L. Pohl, Angew. Chem. 1978, 90, 133; Angew. Chem. Int. Ed. Engl. 1978, 17, 133. Angew. Chem. Int. Ed. 2000, 39, 4216 4235 4233 REVIEWS [18] R. Eidenschink, G. Haas, M. Rmer, B. S. Scheuble, Angew. Chem. 1984, 96, 151; Angew. Chem. Int. Ed. Engl. 1984, 23, 147. [19] P. Kirsch, V. Reiffenrath, M. Bremer, Synlett 1999, 389 396. [20] For historical overviews of the materials development for all types of LCDs: a) T. Inukai, K. Miyazawa, EKISHO J. Jpn. Liq. Cryst. Soc. 1997, 1, 9 22; b) R. Eidenschink, Kontakte (Darmstadt) 1979, (1), 15 18. [21] a) Applications: TN, STN Displays: H. Hirschmann, V. Reiffenrath in Handbook of Liquid Crystals, Vol. 2 A (Eds.: D. Demus, J. Goodby, G. W. Gray, H.-W. Spiess, V. Vill), WILEY-VCH, Weinheim, 1998, pp. 199 229; b) B. S. Scheuble, Kontakte (Darmstadt) 1989, (1), 34 48; c) P. M. Alt, P. Pleshko, IEEE Trans. Electron Devices 1974, ED-21, 146 155. [22] T. J. Scheffer, J. Nehring, M. Kaufmann, H. Amstutz, D. Heimgart- ner, P. Eglin, SID Dig. Tech. Pap. 1985, 16, 120 123. [23] a) R. Williams, J. Chem. Phys. 1963, 39, 384; b) R. Williams (RCA), US-A 3322485, 1962 [Chem. Abstr. 1968, 68, 34254]. [24] G. H. Heilmeier, L. A. Zanoni, L. A. Barton, Appl. Phys. Lett. 1968, 13, 46. [25] B. J. Lechner, Proc. IEEE 1971, 59, 1566. [26] T. P. Brody, J. A. Asars, G. D. Dixon, IEEE Trans. Electron Devices 1973, ED-20, 995. [27] D. R. Baraff, J. R. Long, IEEE Trans. Electron Devices 1973, ED-20, 995. [28] a) U. Finkenzeller, Spektrum Wiss. 1990, (8), 55 62; b) D. Pauluth, T. Geelhaar, Nachr. Chem. Tech. Lab. 1997, 45, 9 15. [29] a) K. H. Yang, Int. Dev. Res. Cent. Techn. Rep. IDRC 1991, 68; b) Y. Iimura, S. Kobayashi, T. Sugiyama, Y. Toko, T. Hashimoto, K. Katoh, SID Dig. Tech. Pap. 1994, 915; c) T. Sugiyama, Y. Toko, T. Hashimoto, K. Kato, Y. Iimura, S. Kobayashi, SID Dig. Tech. Pap. 1994, 919; d) M. Schadt, H. Seiberle, A. Schuster, S. M. Kelly, Jpn. J. Appl. Phys. 1995, 34, 764. [30] a) R. Kiefer, B. Weber, F. Winscheid, G. Baur, Proceedings of the 12th International Display Research Conference (Hiroshima, Japan), 1992, 547 (Society of Information Display and the Institute of Television Engineers of Japan); b) M. Oh-e, K. Kondo, Liq. Cryst. 1997, 22, 379 390, and references therein. [31] A. Takeda, EKISHO J. Jpn. Liq. Cryst. Soc. 1999, 3, 117 123. [32] a) M. Matsuura, EKISHO J. Jpn. Liq. Cryst. Soc. 1997, 1, 23 32; b) M. Hayashi, A. Seki, T. Togawa, SID Dig. Tech. Pap. 1997, 383. [33] a) W. Maier, A. Saupe, Z. Naturforsch. A 1959, 14, 882; b) W. Maier, A. Saupe, Z. Naturforsch. A 1960, 15, 287. [34] S. Marcelja, J. Chem. Phys. 1974, 60, 3599. [35] J. G. J. Ypma, G. Vertogen, H. T. Koster, Mol. Cryst. Liq. Cryst. 1976, 37, 57. [36] W. H. de Jeu, J. van der Veen, Mol. Cryst. Liq. Cryst. 1977, 40, 1. [37] U. Finkenzeller, Kontakte (Darmstadt) 1988, (2), 7 14. [38] In Equation (1), kB denotes the Boltzmann constant, S the Saupe orientational order parameter of the nematic phase,[23] F the reaction 1 e1 field factor F ;f , and h the cavity 1 f a 2pe0 a32e 1 3e factor h . For e, the macroscopic dielectric constant is 2e 1 used: a) W. Maier, G. Meier, Z. Naturforsch. A 1961, 16, 262 267; b) D. Demus, G. Pelzl, Z. Chem. 1982, 21, 1; c) J. Michl, E. W. Thulstrup, Spectroscopy with Polarized Light: Solute Alignment by Photoselection, in Liquid Crystals, Polymers, and Membranes, VCH, Weinheim, 1995, pp. 171 221. [39] C. Mauguin, Bull. Soc. Fr. Mineral. 1911, 34, 71 117. [40] a) C. A. Gooch, H. A. Tarry, Electron. Lett. 1974, 10, 2 4; b) C. A. Gooch, H. A. Tarry, J. Phys. D 1975, 8, 1575 1584. [41] a) L. Pohl, R. Eidenschink, F. del Pino, G. Weber (Merck KGaA), US-A 4285829, 1979 [Chem. Abstr. 1980, 92, 207152]; b) L. Pohl, G. Weber, R. Eidenschink, G. Baur, W. Fehrenbach, Appl. Phys. Lett. 1981, 38, 497 499. [42] M. F. Vuks, Opt. Spectrosc. (Engl. Transl.) 1966, 20, 361. [43] a) J. L. Ericksen, Trans. Soc. Rheol. 1961, 5, 23; b) F. M. Leslie, Arch. Ration. Mech. Anal. 1968, 28, 265; c) A. Raviol, W. Stille, G. Strobl, J. Chem. Phys. 1995, 103, 3788 3794. [44] M. Bremer, S. Naemura, K. Tarumi, Jpn. J. Appl. Phys. 1998, 37, L88 L90. 4234 P. Kirsch and M. Bremer [45] T. Jacob, E. Bhm, K. Tarumi, SID Dig. Tech. Pap. 1996, 131. [46] a) A. Sasaki, T. Uchida, S. Miyagami, Japan Display 1986, 62; b) M. Schadt, Displays 1992, 13, 11. [47] Y. Nakazono, H. Ichinose, A. Sawada, S. Naemura, M. Bremer, K. Tarumi, Int. Dev. Res. Cent. Techn. Rep. IDRC 1997, 65. [48] S. Nagata, E. Takeda, Y. Nanno, T. Kawaguchi, Y. Mino, A. Otsuka, S. Ishihara, SID Dig. Tech. Pap. 1989, 242. [49] S. Naemura, Y. Nakazono, H. Ichinose, A. Sawada, E. Bhm, M. Bremer, K. Tarumi, SID Dig. Tech. Pap. 1997, 199. [50] N. Fukuoka, M. Okamoto, Y. Yamamoto, M. Hasegawa, Y. Tanaka, H. Hatoh, K. Mukai, AM LCD 1994, 216. [51] G. Weber, U. Finkenzeller, T. Geelhaar, H. J. Plach, B. Rieger, L. Pohl, Liq. Cryst. 1989, 5, 1381 1388. [52] Y. Goto, T. Ogawa, S. Sawada, S. Sugimori, Mol. Cryst. Liq. Cryst. 1991, 209, 1 7. [53] K. Tarumi, M. Bremer, B. Schuler, IEICE Trans. Electron. 1996, E79- C, 1035 1039. [54] K. Tarumi, M. Bremer, T. Geelhaar, Annu. Rev. Mater. Sci. 1997, 27, 423 441. [55] F. Mongin, R. Maggi, M. Schlosser, Chimia 1996, 50, 650 652, and references therein. [56] a) T. Hayashi, M. Konishi, Y. Kobori, M. Kumada, T. Higuchi, K. Hirotsu, J. Am. Chem. Soc. 1984, 106, 158; b) E. Negishi, Acc. Chem. Res. 1982, 15, 340; c) P. L. Castle, D. A. Widdowson, Tetrahedron Lett. 1986, 49, 6013; d) R. Arnek, K. Zetterberg, Organometallics 1987, 1230. [57] For reviews on various types of palladium-mediated CC coupling reactions, see the special issue: J. Organomet. Chem. 1999, 576, 1 317. [58] E. Poetsch, Kontakte (Darmstadt) 1988, (2), 15 28. [59] a) C. Petrier, J. C. de Souza Barboza, C. Dupuy, J. L. Luche, J. Org. Chem. 1985, 50, 5761; b) C. Petrier, J. L. Luche, C. Dupuy, Tetrahedron Lett. 1984, 3463. [60] a) K. Tamoa, K. Sumitani, M. Kumada, J. Am. Chem. Soc. 1972, 94, 4374; b) K. Tamoa, K. Sumitani, M. Kumada, J. Am. Chem. Soc. 1972, 94, 9268; c) R. J. P. Couriu, J. P. Masse, J. Chem. Soc. Chem. Commun. 1972, 144; d) A. Sekiya, N. Ishikawa, J. Organomet. Chem. 1976, 118, 349; e) G. W. Parshall, J. Am. Chem. Soc. 1974, 96, 2360. [61] M. Miyaura, A. Suzuki, J. Chem. Soc. Chem. Commun. 1979, 866. [62] J. March, Advanced Organic Chemistry, 3rd ed., Wiley, New York, 1985, pp. 845 854. [63] R. F. Heck, Org. React. 1982, 27, 345. [64] V. Reiffenrath, U. Finkenzeller, E. Poetsch, B. Rieger, D. Coates, SPIE Liq. Cryst. Displays Appl. 1990, 1257, 84 94. [65] The phase transition temperatures are given in 8C, the g1 values in mPa s. Phase abbreviations are C crystalline, Sx smectic X, N nematic, I isotropic. Usually, TNI,extr, De, Dn, and g1 were determined by linear extrapolation from a 10 % w/w solution in the commercially available Merck mixture ZLI-4792 (TNI 92.8 8C, De 5.3, Dn 0.0964). The extrapolated values are corrected empirically for differences in the order parameter. In some marked cases, the following mixtures were used for extrapolation: ZLI-2857 (TNI 82.3 8C, De 1.4, Dn 0.0776), ZLI-3086 (TNI 72.0 8C, De 0.1, Dn 0.1131), and ZLI-1132 (TNI 71.0 8C, De 12.8, Dn 0.1406). For the pure substances, the mesophases were identified by optical microscopy and the phase transition temperatures by differential scanning calorimetry. [66] a) E. Bartmann, Ber. Bunsenges. Phys. Chem. 1993, 97, 1349 1355; b) J. C. Liang, S. Kumar, Mol. Cryst. Liq. Cryst. 1987, 142, 77 84; c) J. C. Liang, J. O. Cross, L. Chen, Mol. Cryst. Liq. Cryst. 1989, 147, 199 206. [67] For general methods for the synthesis of similar substances, see: a) L. N. Markovskij, V. E. Pashinnik, A. V. Kirsanov, Synthesis 1973, 787 789; b) N. Yoshino, M. Kitamura, T. Seto, Y. Shibata, M. Abe, K. Ogino, Bull. Chem. Soc. Jpn. 1992, 65, 2141 2144; c) B. Betzemeier, P. Knochel, Angew. Chem. 1997, 109, 2736 2738; Angew. Chem. Int. Ed. Engl. 1997, 36, 2623 2624. [68] V. Reiffenrath, H. A. Kurmeier, E. Poetsch, H. Plach, U. Finkenzeller, E. Bartmann, J. Krause, B. Scheuble, D. Dorsch, G. Weber (Merck KGaA), EP-B 441932, 1991 [Chem. Abstr. 1996, 125, 208555] . Angew. Chem. Int. Ed. 2000, 39, 4216 4235 Liquid Crystals [69] A. I. Pavluchenko, N. I. Smirnova, V. F. Petrov, Yu. A. Fialkov, S. V. Shelyazhenko, L. M. Yagupolskii, Mol. Cryst. Liq. Cryst. 1991, 209, 225 235. [70] S. V. Shelyazhenko, Yu. A. Fialkov, L. M. Yagupolskii, J. Org. Chem. USSR (Engl. Transl.) 1992, 28, 1317 1324; Zh. Org. Khim. 1992, 28, 1652 1660. [71] a) G. A. Olah, T. Yamato, T. Hashimoto, J. G. Shih, N. Trivedi, B. P. Singh, M. Piteau, J. A. Olah, J. Am. Chem. Soc. 1987, 109, 3708 3713. Different syntheses for trifluoromethoxy derivatives are reported in: b) M. R. Pavia, S. J. Lobbestael, D. Nugiel, D. R. Mayhugh, V. E. Gregor, J. Med. Chem. 1992, 35, 4238 4248; c) M. Kuroboshi, K. Suzuki, T. Hiyama, Tetrahedron Lett. 1992, 29, 4173 4176; d) W. A. Sheppard, J. Org. Chem. 1964, 29, 1 11. [72] Analogous to: D. C. England, L. R. Melby, M. A. Dietrich, R. V. Lindsey, Jr., J. Am. Chem. Soc. 1960, 82, 5116 5122. [73] W. A. Sheppard, J. Org. Chem. 1964, 29, 1 11. [74] M. Schadt, R. Buchecker, A. Villiger, Liq. Cryst. 1990, 7, 519 536. [75] Review: S. M. Kelly, Liq. Cryst. 1996, 20, 493 515. [76] M. Petrzilka, Mol. Cryst. Liq. Cryst. 1985, 131, 109 123. [77] R. Eidenschink, Mol. Cryst. Liq. Cryst. 1983, 94, 119 129. [78] R. Eidenschink, Mol. Cryst. Liq. Cryst. 1985, 123, 57 75. [79] a) M. A. Osman, Mol. Cryst. Liq. Cryst. 1985, 128, 45 63; b) P. Balkwill, D. Bishop, A. Pearson, I. Sage, Mol. Cryst. Liq. Cryst. 1985, 123, 1 13; c) J. E. Fearon, G. W. Gray, A. D. Ifill, K. J. Toyne, Mol. Cryst. Liq. Cryst. 1985, 124, 89 103. [80] D. Demus, Y. Goto, S. Sawada, E. Nakagawa, H. Saito, R. Tarao, Mol. Cryst. Liq. Cryst. 1995, 260, 1 21. [81] G. W. Gray, M. Hird, D. Lacey, K. J. Toyne, Mol. Cryst. Liq. Cryst. 1989, 172, 165 189. [82] E. Bartmann, D. Dorsch, U. Finkenzeller, Mol. Cryst. Liq. Cryst. 1991, 204, 77 89. [83] E. Bartmann, J. Krause, K. Tarumi, 23nd Freiburger Arbeitstagung Flssigkristalle (Freiburg, Germany), 1994. [84] E. Bartmann, U. Finkenzeller, E. Poetsch, V. Reiffenrath, K. Tarumi, 22nd Freiburger Arbeitstagung Flssigkristalle (Freiburg, Germany), 1993. [85] M. Hird, K. J. Toyne, A. J. Slaney, J. W. Goodby, G. W. Gray, J. Chem. Soc. Perkin Trans. 2 1993, 2337 2349. [86] P. Kirsch, K. Tarumi, Angew. Chem. 1998, 110, 501 506; Angew. Chem. Int. Ed. 1998, 37, 484 489. [87] R. Eidenschink, L. Pohl, M. Rmer, B. Scheuble (Merck KGaA), EP-B 107759, 1984 [Chem. Abstr. 1984, 101, 63770]. [88] a) E. Bartmann, Adv. Mater. 1996, 8, 570 573; b) E. Bartmann, K. Tarumi (Merck KGaA), DE-B 19531165, 1995 [Chem. Abstr. 1996, 124, 328585]; c) T. Ando, K. Shibata, S. Matsui, K. Miyazawa, H. Takeuchi, Y. Hisatsune, F. Takeshita, E. Nakagawa, K. Kobayashi, Y. Tomi (Chisso Corp.), EP-B 0844229, 1998 [Chem. Abstr. 1998, 129, 60620] . [89] A. Haas, M. Spitzer, M. Lieb, Chem. Ber. 1988, 121, 1329 1340. [90] a) H.-M. Vorbrodt, J. Vogel, H. Zaschke, G. Pelzl, D. Demus, Mol. Cryst. Liq. Cryst. 1985, 123, 137 141; b) C. Tschierske, H. Altmann, H. Zaschke, G. Brezesinski, F. Kuschel, Mol. Cryst. Liq. Cryst. 1990, 191, 295 300. [91] V. S. Bezborodov, V. F. Petrov, V. I. Lapanik, Liq. Cryst. 1996, 20, 785 796. [92] P. Kirsch, M. Bremer, M. Heckmeier, K. Tarumi, Angew. Chem. 1999, 111, 2174 2178; Angew. Chem. Int. Ed. 1999, 38, 1989 1992. [93] a) W. A. Sheppard, J. Am. Chem. Soc. 1960, 82, 4751 4752; b) W. A. Sheppard, J. Am. Chem. Soc. 1962, 84, 3064 3071; c) W. A. Sheppard, J. Am. Chem. Soc. 1962, 84, 3072 3076. [94] M. P. Greenhall, 15th International Symposium on Fluorine Chemistry (Vancouver, Canada), 1997, presentation FRx C-2. [95] a) H. Okumura, EKISHO J. Jpn. Liq. Cryst. Soc. 1999, 3, 17 24; b) P. Raynes, 28th Freiburger Arbeitstagung Flssigkristalle (Freiburg, Germany), 1999. REVIEWS [96] [97] [98] [99] [100] [101] [102] [103] [104] [105] [106] [107] [108] [109] [110] [111] [112] [113] [114] [115] [116] [117] [118] [119] [120] [121] [122] V. Reiffenrath, R. Hittich, M. Kompter, G. Weber, H. A. Kurmeier, H. Plach, E. Poetsch (Merck KGaA), EP-B 418362, 1991 [Chem. Abstr. 1991, 114, 228424]. a) P. Kirsch, E. Poetsch, Adv. Mater. 1998, 10, 602 606; b) E. Poetsch, W. Binder, J. Krause, K. Tarumi (Merck KGaA), DE-B 19525314, 1995 [Chem. Abstr. 1997, 126, 150650]; c) R. Buchecker, G. Marck, A. Villiger (Hoffmann La Roche AG), EP-B 0732330, 1995 [Chem. Abstr. 1996, 125, 343068]. Preparation analogous to: W. Dmowski, R. A. Kolinski, Rosz. Chem. 1974, 48, 1697 1706. K. Kanie, Y. Tanaka, M. Shimizu, S. Takehara, T. Hiyama, Chem. Lett. 1997, 827 828. J. Fried, S. Kittisopikul, E. A. Hallinan, Tetrahedron Lett. 1984, 25, 4329. R. Noyori, S. Murata, M. Suzuki, Tetrahedron 1981, 37, 3899. S. Sugimori, T. Kojima (Chisso Corp.), EP-B 0084974, 1983 [Chem. Abstr. 1984, 100, 15786]. K. Ohmuro, S. Kataoka, T. Sasaki, Y. Koike, SID Dig. Tech. Pap. 1997, 845. P. Kirsch, A. Ruhl, R. Sander, J. Krause, K. Tarumi (Merck KGaA), DE-B 19714231, 1998 [Chem. Abstr. 1998, 129, 296532]. P. Kirsch, K. Tarumi, Liq. Cryst. 1999, 26, 449 452. V. Reiffenrath, J. Krause, H. J. Plach, G. Weber, Liq. Cryst. 1989, 5, 159 170. G. A. Olah, J. T. Welch, Y. D. Vankar, M. Nojima, I. Kerekes, J. A. Olah, J. Org. Chem. 1979, 44, 3872 3881. M. Bremer, Adv. Mater. 1995, 7, 867. a) For a general overview on computer simulation of liquid crystals, see: Theory of the Liquid Crystalline State: Molecular Modelling: M. R. Wilson in Handbook of Liquid Crystals, Vol. 1 (Eds.: D. Demus, J. Goodby, G. W. Gray, H.-W. Spiess, V. Vill), WILEY-VCH, Weinheim, 1998, p. 72; Computer Simulation of Liquid Crystal Phases Formed by Gay Berne Mesogens: b) M. A. Bates, G. R. Luckhurst in Liquid Crystals I (Ed.: D. M. P. Mingos), Springer, Heidelberg, 1999, p. 65. T. Clark, A Handbook of Computational Chemistry, Wiley, New York, 1985. U. Burkert, N. L. Allinger, Molecular Mechanics, ACS Monograph 177, American Chemical Society, Washington, DC, 1982. W. J. Hehre, L. Radom, P. von R. Schleyer, J. A. Pople, Ab Initio Molecular Orbital Theory, Wiley, New York, 1986. W. Kohn, A. D. Becke, R. G. Parr, J. Phys. Chem. 1996, 100, 12 974 12 980. M. Karelson, V. S. Lobanov, A. R. Katrizky, Chem. Rev. 1996, 96, 1027 1043. S. R. Johnson, P. C. Jurs, Chem. Mater. 1999, 11, 1007 1023. M. Bremer, K. Tarumi, Adv. Mater. 1993, 5, 842 848. M. Klasen, M. Bremer, A. Gtz, A. Manabe, S. Naemura, K. Tarumi, Jpn. J. Appl. Phys. 1998, 37, L945-L948. SPARTAN Version 5.1, Wavefunction Inc., Irvine, CA. Gaussian 98, Revision A.6: M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery, Jr., R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala, Q. Cui, K. Morokuma, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, J. L. Andres, C. Gonzalez, M. Head-Gordon, E. S. Replogle, J. A. Pople, Gaussian, Inc., Pittsburgh, PA, 1998. MOPAC V6.0, QCPE 455, Indiana University, Bloomington, IN. A. Klamt, V. Jonas, T. Brger, J. C. W. Lohrenz, J. Phys. Chem. 1998, 102, 5074 5085. C. Reichardt, Chem. Rev. 1994, 94, 2319 2358. Angew. Chem. Int. Ed. 2000, 39, 4216 4235 4235