ANCHORING OF LIQUID CRYSTAL AND DYNAMICS OF MOLECULAR EXCHANGE BETWEEN ADSORBED LC FILM AND THE BULK

Transcription

1 ANCHORING OF LIQUID CRYSTAL AND DYNAMICS OF MOLECULAR EXCHANGE BETWEEN ADSORBED LC FILM AND THE BULK A dissertation submitted to Kent State University in partial fulfillment of the requirements for the degree of Doctor of Philosophy By Rui Guo August 2008

2 Dissertation written by Rui Guo B. Sc., Jilin University, China, 1998 M. Sc., Shanghai University, China, 2001 Ph.D., Kent State University, 2008 Approved by, Chair, Doctoral Dissertation Committee Satyen Kumar, Physics department, Co-chair, Doctoral Dissertation Committee Yuriy Reznikov, department of Physics of Crystals, Institute of Physics, National Academy of Sciences of Ukraine, Members, Doctoral Dissertation Committee David Allender, Physics department, Bryon Anderson, Physics department, Shan-hu Lee, Chemistry department, Qihuo Wei, Chemical Physics Interdisciplinary Program Accepted by, Chair, Department of Physics Bryon Anderson,, Dean, College of Arts and Sciences John R.D. Stalvey ii

3 TABLE OF CONTENTS TABLE OF CONTENTS...iii LIST OF FIGURES. v LIST OF TABLES..ix ACKNOWLEDGMENTS...xi Chapter 1 Introduction Liquid Crystals Surface Anchoring of Nematic Liquid Crystals Overview of the Dissertation..8 Chapter 2 Liquid Crystal Anchoring Theory of Anchoring Drift of the Easy Axis Adsorbed LC film 19 Chapter 3 Anchoring Energy, Angular Distribution and Drift of Easy Axis Experimental Details Sample Preparation Measurement of Twist Angle Dependence of the Twist Angle on Thermal Annealing Time 35 iii

4 3.2 Theoretical Model and Discussion Conclusions..47 Chapter 4 Dynamics of Molecular Exchange between Adsorbed LC Film and the Bulk Experiment Details Sample preparation Measurement with magnetic null method Experimental Method and Results Theoretical Model and Discussion Conclusions...66 Chapter 5 Summary...69 References iv

5 LIST OF FIGURES Figure 1: (a) Sodium palmitate C 15 H 31 COONa molecules assemble into micelles, shown in (b), by shelding the hydrophobic part from the solvent, water. (b) shows a spherical micelle. However, many other surfactants form rod- and disc-like micelles which can undergo orientational and positional order. (c) An orientationally order lyotropic nematic phase of disc-like micelles is shown as an illustration. The straight lines represent orientationally ordered symmetry axis of the micelles along the director n....3 Figure 2: Illustration of the arrangement of molecules (rod-like) in liquid crystal different phases.. 4 Figure 3: The anchoring energy is responsible for aligning the director n along the preferred orientation (Z-axis) at a substrate. This schematic depicts n when it has been moved away from the Z-axis (θ=φ=0) to angles θ and φ, in polar and azimuthal direction, respectively Figure 4: Apparent twist angle in the LC cell, ϕ test, is smaller than the angle ϕ 0 set by the preferred alignment at the two cell substrates. The preferred direction at the lower substrate is shown by the dashed lines. The upper substrate with strong azimuthal anchoring dictates the resultant director orientation at the weakly azimuthal anchoring lower surface Figure 5: Chemical structures of pentyl-cyanobiphenyl (5CB) and PI v

6 Figure 6: Schematic view of preparation of an adsorbed LC layer on ITO coated glass substrates Figure 7: Orientation of LC induced by magnetic field. (a) Alignment direction parallel to the polarizer axis of the microscope. (b) Alignment direction at 45 to the polarizer. A and P are analyzer and polarizer directions of the microscope, n is the director alignment direction Figure 8: Schematic representation of the hybrid test cell prepared to measure the twist angle.. 30 Figure 9: homogeneous twist structure of the combined cell...32 Figure 10: Experimental set up for measuring twist angle. The source of the laser beam is HeNe laser operating at 10mw of power and wavelength of λ=633nm. The laser beam passes through a polarizer, the test cell, and analyzer before reaching a Premier laser diode stand detector Figure 11: The dependence of transmission on angle of rotation the analyzer. When the axis of the analyzer was parallel to the easy axis on the test substrate, the detector showed the largest transmission. In this case, the twist angle between the easy axes on the two cell substrates is the angle between the polarizer and analyzer at maximum transmission. Twist angle was 42 close to the intended value of Figure 12: Dependence of the twist angle on thermal annealing time at different temperatures Figure 13: Schematic representation of the hybrid test cell used to confirm a decrease of the anchoring energy on the test substrates 37 vi

7 Figure 14: These two pictures show perfect planar alignment and prove that the easy axis on the test substrate did not drift after thermal annealing and the decrease of the angle ϕ test after annealing is caused by a decrease of the anchoring energy on the test surface. (a) Easy axis on the reference substrate in the cell is parallel to the polarizer axis. (b) Easy axis on the reference substrate makes an angle of 45 to the polarizer. A and P are analyzer and polarizer of the microscope, n is the easy axis on the reference substrate Figure 15: Dependence of parameter λ on the nematic order parameter S Figure 16: Dependence of the twist angle and best fits (solid line) on thermal annealing time. The probability of desorption is the fitting parameter..48 Figure17. Dependence of the probability of desorption on the temperature of thermal annealing Figure 18: Dependence of ln on 1/T which yields activation energy Δ E = 0.55 A ev Figure 19: Chemical structures of (a) poly(amic acid), (b) polyimide and (c) Rf8MPD..54 Figure 20: Illustration of pre-tilt angle measurement using the magnetic null method. The source of the laser beam is JDS HeNe laser operating at 10mw of power and a wavelength of λ = 633nm. The laser beam, perpendicular to the magnetic field, passes through a polarizer, Soleil-Babinet optical compensator, the test cell, and analyzer before reaching a Premier laser diode detector. Polarizer and analyzer directions were oriented at ±45 to the Y-axis. A Soleil-Babinet compensator was placed between the sample and the polarizer with its slow axis parallel to the Y-axis of the cell.55 vii

8 Figure 21: The test cells were made with anti-parallel arrangement of two rubbed polyimide Rf8MPD coated substrates. Both 5CB and ZLI-4792 have homogeneous alignment. Pre-tilt angle of 5CB and ZLI-4792 on Rf8MPD alignment layer is 2 and 32, respectively Figure 22: The pretilt angles in the cells refilled with LC-1, θ 12 approached the value θ 1, 1 and the pretilt in the cells refilled with LC-2, the value θ 2, 1 evolved towards θ 2, 2 with time...60 Figure 23: Clearing temperatures of LC mixtures. α m is the mole fraction of 5CB in mixtures..61 Figure 24: Calculated time dependences θ ( t) and ( t) 12 θ after refilling the cells Figure 25: Time dependence of the pretilt angle of the mixture of 5CB and ZLI-4792 after the cell is refilled for different molar ratios α m of the two compounds viii

9 LIST OF TABLES Table 1 Physical properties of NLCs used Table 2 Clearing temperature, before and after refilling, of LC ix

10 To my family x

11 ACKNOWLEDGEMENTS This work is dedicated to my husband Dr. Fenghua Li and our parents. Without their ceaseless encouragement and support, the dissertation would not have been possible. I also feel deeply grateful to Dr. Satyen Kumar who has been my advisor on the dissertation work. His knowledge and enthusiasm have been a constant source of motivation to me during this endeavor. His wisdom has guided me through one of the most important stages of my life. Being always very considerate and helpful, Dr. Kumar has given me the most support, not only in my research but also in many other ways. I would like to give special thanks to my co-advisor Dr. Yuriy Reznikov, who has given me precious advice on so many things. I consider myself very lucky to have the opportunity to work with these top scientists in their fields, and to have learned so much from them. Special thanks I want give to Kostyantyn Slyusarenko who I worked with for the theoretical calculations. Without him I couldn t finish it. I also want to thank all members of my group at the physics department for their kind help and valuable discussions. My committee members deserve special thanks for their willingness to participate, and for their valuable insights. This work was supported by a grant (DMR ) from the Solid State Chemistry program in the Dinsion of Materials Research of the National Science Foundation. xi

12 CHAPTER 1 Introduction 1.1 Liquid crystals When Reinizer discovered the liquid crystalline phase in the late 19 th century, he probably could not have imagined the popularity liquid crystals enjoy today. In fact, liquid crystalline materials did not attract much attention until after liquid crystal displays were invented in Kent in the last century. Today, liquid crystals have been an active area of research because of their fascinating phase behavior, rich scientific phenomenology, and their technological application. Many organic materials exhibit phases of intermediate order called mesophases upon heating from the crystalline to the isotropic liquid phase. [1] In the crystal phase, the ordering of the component atoms or molecules is long-range positional order in three dimensions, and atoms/molecules are arranged on a periodic lattice. Long-range orientational order also is present if the components are anisotropic in shape. While isotropic liquids possess no orientational or positional order, liquid crystals possess order intermediate these two extremes, with orientational and one- or two-dimensional positional order. Liquid crystalline materials are separated into two categories: lyotropic and thermotropic, which are distinguished by the primary physical parameters that can be 1

13 2 varied to bring about a change in their phase. For lyotropic liquid crystals, parameters are the concentration of the components of a mixture, whereas it is temperature for thermotropics. One example of lyotropic liquid crystals is surfactant, which have a hydrophobic tail group at one end and with a hydrophilic ionic group at the other end. [2] Such amphiphilic molecules form ordered structures in both polar and non-polar solvents. For example, as can been seen from figure 1, soaps have a polar head group attached to a hydrocarbon tail. When dissolved in a polar solvent such as water, the surfactant molecules assemble to shield the hydrophobic tail from water and present the hydrophilic head to the solvent. If these amphiphilic molecules are mixed with a nonpolar solvent such as hexane, similar structures form but now the polar heads assemble in such a way that the non-polar tails are in contact with the solvent. These are called reversed micelles. Systems with rod- or disk-like micelles with a symmetry axis from lyotropic nematic phase as shown in figure 1 (c). Thermotropic liquid crystals are mostly organic molecules which change their organization with temperature to exhibit a variety of different phases, such as: the smectic, nematic, cholesteric and columnar phases, etc. Many organic compounds exhibit one or more liquid crystalline phases. A comparison between crystalline, smectic A, nematic, and isotropic phase is shown in figure 2. [2] In the isotropic phase, all molecules can randomly orient and move around and the liquid has no order. In the nematic phase, the molecules possess orientational long-range order. The preferred direction of alignment is called the director and is labeled as unit

14 3 + Na O O C C 15 H 31 (a) (b) n (c) Fig. 1: (a) Sodium palmitate C 15 H 31 COONa molecules assemble into micelles, shown in (b), by shelding the hydrophobic part from the solvent, water. (b) shows a spherical micelle. However, many other surfactants form rod- and disc-like micelles which can undergo orientational and positional order. (c) An orientationally order lyotropic nematic phase of disc-like micelles is shown as an illustration. The straight lines represent orientationally order symmetry axis of the micelles along the director n.

15 4 Crystalline phase Smectic A phase Isotropic phase Nematic phase Fig. 2: Illustration of the arrangement of molecules (rod-like) in liquid crystal different phases.

16 5 vector n. The orientational order parameter, S, describes the degree of order of the molecular long axes with respect to n, and is defined in terms of the angle θ made by individual molecules with the director. S is written in terms of a distribution function f ( θ ): cosθ = a n = f ( θ )cosθ dω with S = cos θ = f ( θ ) (3cos 2 θ 1) dω A typical nematic liquid crystal molecule has a rigid elongated core and one or two flexible tails. In a smectic phase, the molecular centers of mass (CM) are arranged, on the average, in layers. The deviations of an individual molecule s CM from the equilibrium position can be fairly large in these phases. The molecules possess liquid-like order within the 2D smectic layers. If the alignment of the molecules is perpendicular to the layers, they form the Smectic-A phase. Several smectic phases possess a layered structure with liquid-like in-plane order but with tilting of the long axis away from the layer normal (Smectic-C, Smectic-C*, -Cα,-C 2, etc.) Due to thermal fluctuations there is no true long-range order in these one dimensionally ordered phases. In the crystal phase, each molecule has a fixed position in a periodic three dimensional array and exhibits genuine three-dimensional long-range positional order.

17 6 1.2 Surface anchoring of nematic liquid crystals Since nematic LC phases are anisotropic, any investigation of their physical properties or electro-optical applications requires unidirectional macroscopic alignment of the director. Alignment phenomena of liquid crystal phases is critically important in understanding the essential features and behavior of liquid crystals and their performance in liquid crystal devices. External (magnetic or electric) fields, mechanical shear, and anisotropic interactions offered by treated surfaces are commonly used for liquid crystal alignment. For external electric or magnetic fields, the anisotropy in LC s response function (mainly corresponding to the anisotropy in their susceptibility) is responsible for the alignment and reorientation of the director. When a LC is placed between two substrates and one substrate is moved sideways with respect to the other, the resulting mechanical shear causes the nematic director to align along the direction of shear flow. Since Mauguin successfully aligned liquid crystals on a glass surface by rubbing the glass with a piece of paper in the early 20 th century, [3] the use of rubbed polymer films has become the most common, empirical but reliable method to obtain unidirectional alignment of LCs. It is believed that the anisotropic surface potential at the LC-substrate interface aligns the LCs. This alignment propagates into the bulk due to the long range nematic elasticity. However, the physical mechanism responsible for the alignment of LC molecules at the interface has not been fully understood. Several mechanisms have been proposed to explain this effect. The generation of microgrooves or scratches on the polymer surface led Berreman [4] to suggest that the LC director adapts a configuration to minimize the total elastic distortion energy, resulting in director orientation along the

18 7 direction of the microgrooves. Another argument put forth relies on the physico-chemical interactions such as Van der Waals interactions, steric interactions, and dipole-dipole interactions between the LC and the polymer molecules to align the director [5]. Several other studies [6-8] have suggested that both of the two effects act simultaneously to align the LCs. Since Berreman s work, significant research has been done to explain the mechanism of LC alignment on the solid substrates. Dubois, et al., [9] obtained a homogeneous alignment of LC on a microgroove birefringence, Miyano [10] found that the ordering of LCs at the polymer-lc interface was better than that in the bulk. Ishihara, et al., [11] observed that homogeneous alignment of nematic LC could be obtained by rubbing a polystyrene film, but the orientation of LC director was perpendicular to the direction of the microgrooves. The results of the second harmonic generation experiments [6] showed that a short-range molecular interaction was responsible for the alignment of the first monolayer of liquid crystals on a rubbed polymer surfaces. Grazing angle incidence x-ray reflectivity experiments on the rubbed polymer films suggested strong orientation of polymer segments happened in the vicinity of the air-film interface. [12] Kikuchi, et al., [13], using atomic force microscopy (AFM), concluded that the presence of microgrooves on a rubbed polyimide surface was not necessary for the alignment of nematic LC. More recently, a polyimide with a fluorine unit incorporated in a side chain has been reported to show alignment of LC director perpendicular to the rubbing direction. [14] A recent comprehensive study of variously treated surfaces [15] by X-ray reflectivity and AFM techniques has revealed that surface morphology plays the most

19 8 important role in determining the direction of LC alignment. However, how the pretilt depends on the surface structure remains to be understood. All these studies with varied result show that we do not fully understand everything about LC alignment. At present mechanically rubbed polyimide films are most widely used in commercial preparation of liquid crystal displays because of its simplicity and high thermal stability of the resultant alignment. After decades of empirical work, this technology is perfect now. No other technique has been able to replace it especially for large area displays. Meanwhile, non-contact alignment methods are being intensely pursued by researchers. Photo-induced and external field induced alignment are two of the non-contact methods. However, our understanding of physical or chemical interactions between a treated (e.g., exposed to UV) substrate and liquid crystal molecules is still limited. These factors become very important for systems with no surface treatment, such as field aligned, which are the subject of this dissertation. 1.3 Overview of the dissertation The research conducted over the past 50 years has resulted in a good phenomenological description of the bulk properties of the nematic phase. However, the understanding of interaction with and the influence of surface properties on alignment of the nematic phase is unclear. As in any other field of condensed matter physics, the surface and interfacial properties of nematic liquid crystals are rather complex. It is important to understand the alignment of a nematic liquid crystal when it comes in

20 9 contact with a solid substrate. The precise nature and the origin of anchoring of nematic liquid crystals cannot be considered a solved problem. Recent measurements show that an adsorbed LC film can be invoked to give an explanation to the anchoring energy of nematic liquid crystals. The purpose of this dissertation research is to investigate the role of an adsorbed LC layer and to develop a theoretical treatment for adsorption-desorption (AD) phenomena and how they affect the anchoring energy of nematic liquid crystals. In the following chapters, we will first review the literature and summarize the status of current understanding of the alignment of LCs. In chapter 3, a study of the anchoring energy and the drift of the easy axis of an adsorbed liquid crystal (5CB) film will be presented. We will concentrate on the case where one substrate has strong anchoring and at the other substrate, director anchoring is determined by the process of adsorption/desorption of LC molecules. Using the Rapini-Papoular type surface potential for interaction between LC molecules in the bulk and adsorbed LC molecules, we will build a simple model to investigate the relationship between the angular distribution of the adsorbed molecules and the changes in it due to AD-processes and how the values of the LC anchoring energy and easy axis direction change. In chapter 4, the dynamics of molecular exchange between the aligning adsorbed film of LC and the bulk will be presented. The characteristic times of the changes in the pretilt angle we have found to depend on the coefficients of desorption of liquid crystal molecules from the adsorbed layer. The experiments and the model presented here provide an important insight into the mechanism of LC alignment and anchoring.

21 CHAPTER 2 Liquid Crystal Anchoring 2.1 Theory of Anchoring The bulk properties of liquid crystals depend strongly on the molecular structure and on intermolecular interactions. When the nematic phase comes in contact with a solid surface, the orientation of n in its proximity is primarily determined by its interaction with and the structure of the substrate. Anchoring may be defined as the phenomenon of orientation of a liquid crystal by a surface. The anchoring of liquid crystals is very important, from the fundamental point of view, for understanding the anchoring mechanism and for improving the performance of the liquid crystal devices. A surface normally imposes a preferred direction on the director, called the anchoring direction, or, simply, easy direction. The easy axis is then the direction of spontaneous orientation of n at the surface, in the absence of any external field/force. The energy of the interfacial region, between the LC and the substrate, depends on the orientation of the director relative to the easy direction. The origin of the surface alignment can be discussed in terms of anisotropic torques acting on the director and arising from either physico-chemical or geometrical factors. These torques deform the nematic director configuration which costs (elastic) energy. The equilibrium director 10

22 11 configuration can be calculated by minimizing the free energy (describing the bulk distortion) density while taking into account the configuration of the surface free energy, f s. The bulk part of the total energy is well known, the surface energy, hand, does not have a universally accepted and well defined form. f s, on the other The term anchoring energy was introduced to describe the contribution due to deviation of the director at the surface from its preferred orientation determined by elastic torques in a constrained liquid crystal. The omission of the variation in the value of the order parameter can be justified on grounds similar to those used in the case of Frank elastic theory. Orientational anchoring can be classified as homeotropic, tilted, or planar depending on whether the preferred direction is perpendicular, tilted, or parallel to the surface. Furthermore, the last two cases can be monostable, bistable, multistable, or degenerate depending on the nature of the surface. At a clean and smooth surface, the only unique direction is the surface normal, which we denote by e. The surface energy, f s, is then expected to depend only on the deviation of n from the normal orientation, or, on n e. The director orientation at the interface between bulk NLC and a solid is defined in terms of a polar surface angle, θ s, and of an azimuthal angle, ϕ s, of the director. Since the anisotropic part of the surface energy has to depend on the director orientation at the interface, one can write f s ( Θ, Φ) + W ( θ Θ, ϕ Φ), = f s s s where, Θ and Φ correspond to the values of the polar and azimuthal angles which minimize the surface energy (and define the easy direction). The positive function, W, is

23 12 the anchoring energy. As follows from the definition of easy direction, at equilibrium and in the absence of an external torque, θ = Θ and ϕ = Φ. The anchoring energy can be s physically interpreted as the work that must be done to rotate the director away from the easy direction, e.g. from being parallel to the Z-axis to the angles θ s and φ s in polar and azimuthal directions (figure 3), respectively. It is then possible to define the polar and azimuthal torques, per unit area, respectively, in the following way: s f τ s θ = and θ s f τ s ϕ =. ϕ s If the deviations from equilibrium are small, these torques can be approximatly given by τ θ 2 Wθ ( θ Θ) and τ W ( ϕ Φ) = s Here, we have introduced two important quantities: ϕ = 2 ϕ s, W θ = θs f s 2 Θ, Φ i.e., polar anchoring coefficient or polar anchoring energy and W ϕ 2 1 = 2 ϕs f s 2 Θ, Φ known as azimuthal anchoring coefficient or azimuthal anchoring energy. Typical experimental values of W and W vary from 10-4 to 10-1 erg/cm 2. θ ϕ Usually one can write the anchoring energy function in Rapini-Papoular form W ( θ, ϕ) = Wθ sin θ + Wϕ sin ϕ

24 13 Z, e n θ s Y φ s X Fig. 3: The anchoring energy is responsible for aligning the director n along the preferred orientation (Z-axis) at a substrate. This schematic depicts n when it has been moved away from the Z-axis (θ=φ=0) to angles θ s and φ s, in polar and azimuthal direction, respectively.

25 14 where, W θ and W ϕ are the corresponding polar and azimuthal anchoring energies. The values of W θ and W ϕ are very important characteristics of LC-surface interaction. A variety of techniques have been employed to determine W θ, [16-20] most of which use elastic deformations created by an external field. The most commonly used method, especially for strongly anchoring cells of industrial interest, is the so-called highelectric-field (HEF) technique [19] developed by Yokoyama and van Sprang. [17] It requires simultaneous measurements of the capacitance C and the optical phase retardation R of a nematic cell as a function of applied voltage V. A very attractive feature of this technique is that in a certain range (V min, V max ) of applied voltages, R is a linear function of the reciprocal electric displacement (~1/CV) and W (normalized by an elastic constant) W θ can be simply deduced from the intercept of this linear dependence on R. However, this technique has its limitations. The choice of the proper range (V V max ) presents a problem. According to Ref. 7, the range (V min, V max ) is determined by the following considerations. First, V min should be well above the Fredericks threshold [V min 6 V th ] to assure that the director in the middle of the cell is parallel to the field. This is why the method is often referred to as the high-electric-field technique. Second, the field-induced director deviations from the easy axis should be sufficiently small so that the energetic cost of surface director deviations can be approximated by the Rapini- Papoular potential; this requirement limits V max. As a result, the choice of V max is ambiguous because V max depends on W, and W is an unknown. min,

26 15 The azimuthal anchoring energy, W ϕ0, can be measured using two substrates, one coated with a rubbed polyimide layer which provides strong anchoring, and the second one with a weakly aligning layer (for example, UV-treated film). The two substrates are assembled with a twisted nematic configuration (figure 4). The value of the apparent twist angle in the LC cell is lower than that set by the easy directions of the two substrates (shown in figure 4 by solid and dashed lines, respectively), because the strongazimuthally anchoring substrate affects the director alignment on the weakly anchoring surface. The azimuthal anchoring energy W ϕ0 of the weak-aligning surface can be calculated from the torque balance equation in the LC cell, provided that the anchoring energy of the rubbed surface is infinite: W [21,22] = 2K22ϕtest / d sin 0 [ 2( ϕ ϕ )] ϕ0 test where, ϕ 0, ϕ test, K 22, and d are the twist angle set by the easy directions at the two LC cell substrates, apparent twist angle, twist elastic constant of the liquid crystal, and cell gap, respectively. 2.2 Drift of the easy axis Surface dynamics of liquid crystal molecules at solid substrates has been an interesting subject for applied and fundamental researchers over the last decade. Traditional description of the reorientation of a LC director in electric or magnetic fields assumes a fixed position of the easy axis, e, on the aligning surfaces of a cell. At the

27 16 Strong anchoring ϕ test ϕ 0 Weak anchoring Fig. 4: Apparent twist angle in the LC cell, ϕ test, is smaller than the angle ϕ 0 set by the easy directions at the two cell substrates. The preferred direction at the lower substrate is shown by the dashed lines. The (upper) substrate with strong azimuthal anchoring dictates the resultant director orientation at the weakly azimuthal anchoring (lower) surface.

28 17 same time, field-induced deviation of the director n near the surface, away from the direction e due to a finite anchoring strength may result in a gradual drift of the easy axis away of its initial position. This is called the gliding effect. [23-28] The adsorption/desorption of LC molecules on/from the aligning surface and cooperative reorientation of polymer fragments and LC molecules are considered as possible mechanisms for gliding. The adsorption/desorption (AD) mechanism was first proposed by Vetter et. al., [23] for describing of the gliding effect on polyvinyl-alcohol surface. To the best of our knowledge, it was the first report of the drift of an easy axis. According to Vetter, the drift of the easy axis is caused by the rotation of the symmetry axis of the distribution function of adsorbed LC molecules under the influence of a reorientation torque. It is suggested that adsorbed molecules are preferentially oriented along the initial direction of director in the cell. Application of a torque reorients the director near the surface and leads to the adsorption of molecules along this new direction. As a consequence, the symmetry axes of the angular distribution function of the adsorbed molecules reorients as well as the associated easy axis. The model of cooperative reorientation of the director and polymer fragments was proposed first by Kurioz et. al., [24] as explanation of unexpectedly slow relaxation of the director in the zenithal plane after the application of electric field to the cell with a soft polymer surface. It was suggested that due to weak anchoring, the electric field reorients the director on the polymer surface, which, in turn, drags the flexible polymer fragments. As a result, the applied electric field orients both LC molecules and flexible fragments and causes a drift of the easy axis. Later on, Janossy [25] explained the drift of the easy

29 18 axis in the azimuthal plane over a soft surface in a similar way. His interpretation of the azimuthal gliding is based on the assumption that the polymer main chains can go through conformational transitions under the influence of an anisotropic potential of the liquid crystal. The change of the director orientation at the surface initiates a conformational change in the polymer and, as a result, the easy axis rotates in response to the external field, causing the drift of the easy axis. The drift speeds up when the temperature of the polymer s glass transition is approached. A microscopic description of the easy-axis gliding effect in terms of the rotation of the angular distribution function of the LC molecules adsorbed on the surface was developed by A. Romanenko. [26,27] It should be noted that both the model of adsorption/desorption and the cooperative model are described by the same equations in Refs. 26 & 27. The difference lies in the physical meanings of these equations and the values of microscopic parameters. Physical interpretation becomes even more complicated since a drift of the easy axis can also be observed on rigid inorganic surfaces with no flexible surface fragments as well as on soft polymer surfaces. [28] One can suggest that only the adsorption/desorption mechanism governs the drift of the easy axis on rigid surfaces but both mechanisms may contribute to the gliding effect on a soft polymer surface.

30 Adsorbed LC film The surfaces of real solids, even if cleaned and polished, are usually covered with adsorption sites, to which particles of the fluid and gas may attach themselves. [29] This phenomenon is due to interatomic forces acting between the adsorbed particles and the surface, and may occur in a broad variety of systems. [30] The phenomena of adsorption have attracted attention since the first decades of the last century. The name of Irving Langmuir is associated with the research in this area, because he made extensive pioneering studies of the adsorption of gases onto metal surfaces. It is important to distinguish between absorption, in which the molecules go to the interior of a substance, and adsorption, where the molecules are stuck to the surface. Desorption is the reverse of the adsorption and requires that the adsorbed particle gains enough energy to become free of the surface. A molecule may bind to the surface as a chemisorbed or physisorbed molecule. In the case of chemisorption, a chemical bond is formed between the particle being adsorbed (from the gas or the liquid state) and the substrate. In the second (physisorption) case, the particles are bound to the substrate by physical forces (like Van der Waals, dispersion, and coulomb forces), retaining their separate chemical identity. The energy involved in the process is called the adsorption energy. Some of the particles bombarding the surface will bounce back if they recoil with energy larger than the adsorption energy. At any particular density of particles, a certain fraction will remain bound. The surface coverage by adsorbed particles is defined as [31] :

31 20 σ σ R = = Number of surface sites occupied / total number of surface sites. σ 0 The rate at which the first adsorbed layer builds up decreases as the layer reaches a saturation value in the filling process. Here, we are concerned with the adsorption of particles coming from the liquid phase. The coverage of a surface when dynamic equilibrium is reached depends on the density of the phase. The variation of σ R at a given temperature is described by the adsorption isotherm. The simplest isotherm is the Langmuir isotherm, which gives the relation between the coverage of the first layer and the density at a particular temperature. This isotherm is constructed on three basic assumptions: 1. The adsorption occurs only in the first layer ( monolayer coverage); 2. All adsorbing sites are equivalent, and the surface is smooth ( it is perfectly flat on a microscopic scale); 3. The adsorption energy of one site is independent of the occupancy of neighboring sites. The equation is derived starting from equilibrium between empty surface sites, particles, and filled particle sites. The expression S * + P SP depicts equilibrium between empty surface sites S * and particles P and filled surface sites SP. The rate of adsorption is proportional to the density of ρ, and the number of adsorbing sites at the surface, i.e.,

32 21 dσ dt κ ρ( σ σ ) = a 0 (2.1) where κ a is the rate constant for adsorption, ρ is the bulk density of adsorbate just in front of the adsorbing surface, and σ 0 σ is the total number of free sites. The rate of desorption is proportional to the number of adsorbed species, i.e., dσ = κ dσ (2.2) dt where κd is the rate constant for desorption. Let us introduce the reduced quantities σ R = σ /σ 0 and ρ R = ρ / ρ0, where ρ0 is the bulk density of the adsorbate, in the absence of adsorption. At equilibrium, the net rate of adsorption is zero, implying the equality between the absolute values of (2.1) and (2.2); solving forσ, one obtains the Langmuir isotherm: R σ R αρr = 1+ ρ R where, α is Langmuir adsorption constant which governs the steady state and is expressed as κτρ0 α = σ 0 if we define τ = 1/κ d and κ = κaσ 0. The first parameter represents a characteristic time associated with the desorption process. The second parameter is connected with the adsorption phenomenon.

33 22 In this dissertation, adsorption/desorption processes of the LC molecules on the ITO and polyimide coated surfaces are investigated and attempts are made to understand these through a simple model which incorporates the easy axis and the anchoring energy.

34 CHAPTER 3 Anchoring Energy, Angular Distribution, and Drift of Easy Axis Traditional description of liquid crystal (LC) alignment on a solid surface suggests that the easy orientation axis is given by direct anisotropic interaction between a physically anisotropic surface and LC molecules, whose translation mobility diffusion does not differ much from the mobility of the molecules in bulk LC. Since the first observation of the surface memory effect, [32-34] it became understandable that adsorption of LC molecules on the aligning surface can yield an anisotropic layer which itself aligns the LC. Subsequent studies confirmed that the layer formed of adsorbed molecules plays an important role in orienting the bulk LC that comes in contact with it. For example, when a cell filled with LC is cooled from the isotropic phase to a nematic phase in the presence of a magnetic field, homogeneous LC orientation can be attained without surface treatment. [33-35] Light-induced adsorption of dye molecules on an isotropic polymer surface from the LC bulk also causes homogeneous alignment of the LC. [36] Investigations of the drift of the easy axis (or, the gliding effect [35-39] ) with time and temperature in electric, magnetic, and optical fields showed that this effect is caused by processes of adsorption and desorption of LC molecules on the aligning surface. A layer of adsorbed molecules should be considered as a mobile system wherein an active 23

35 24 exchange between bulk and surface molecules occurs via the processes of adsorption and desorption (AD-processes) of the LC and dopant molecules. Macroscopic characteristics of alignment given by the adsorbed layer are determined by the angular distribution of the adsorbed molecules. Recently Reshetnyak, et. al, developed a theory that describes director gliding in terms of changes in the angular distribution function and rotation of the adsorbed molecules due to a bulk torque. [26,27] This theory connects macroscopic characteristics of gliding with microscopic properties of adsorbed LC molecules. For instance, the maximum of the angular distribution function of the adsorbed molecules specifies the direction of the easy axis, e, and the width of the distribution is inversely related to the anchoring energy, W. The drift of the easy axis due to adsorption-desorption process of LC molecules on traditional polymer surfaces is believed to be accompanied by collective reorientation of flexible polymer fragments and LC molecules. [40-42] The last process results in the drift of the easy axis along with AD-process. Both mechanisms of gliding are described by the same equations in Refs. 25 and 26. The difference between them is just in the physical meaning of these equations. The determination, of which process plays the key role here, is difficult. This chapter reports experimental studies of anchoring properties of a layer of anisotropically adsorbed LC molecules on a rigid substrate obtained by cooling the LC from its isotropic phase in the presence of a magnetic field. Use of a rigid isotropic ITO surface and formation of a reliable adsorbed layer of LC molecules under a strong magnetic field enabled me to avoid interference due to the effects of rearrangement of the

36 25 aligning surface and non-uniformity of the angular distribution of the adsorbed molecules. The observed drift of the easy axis with time in the nematic phase and upon annealing in the isotropic phase and associated changes in the anchoring energy are explained in the frame of the Reshetnyak s model. [26,27] It allowed me to establish the relationship between the angular distribution of adsorbed molecules and the changes due to ADprocess with the value of the LC anchoring energy and the direction of the easy axis. Experimental data reveal essential insight into the interactions between LC molecules and the substrate surface. 3.1 Experimental details Sample preparation A layer of adsorbed LC molecules of pentyl-cyanobiphenyl (5CB, from Merck) was formed on the ITO coated glass substrates. Chemical structure of 5CB is shown as figure 5, and its phase sequence is: Crystalline 24.0ºC Nematic 35.3 ºC Isotropic. An empty symmetric 5 µm thick cell made with ITO coated glass plates was heated to 70ºC in a hot stage and placed between poles of an electromagnet so that the lines of magnetic field, H 0.8T, were in the plane of the cell (figure 6). The cell was then filled with 5CB that was preheated to 70ºC. The direction of filling was kept either parallel or perpendicular to the field to see if it affected the alignment through flow effect. After filling, the 5CB was cooled down to the nematic phase in 15 minutes in the presence of magnetic field. Then, the field was lowered to zero strength and the cell was removed from the hot stage. A

37 26 (a) pentyl-cyanobiphenyl (5CB) O O O N N O O O n (b) C 29 H 14 N 2 O 6 (PI 2555) Fig. 5: Chemical structures of pentyl-cyanobiphenyl (5CB) and PI 2555.

38 27 Magnet LC Cell N S Hot stage H Fig. 6: Schematic view of preparation of an adsorbed LC layer on ITO coated glass substrates.

39 28 A n P (a) A n P (b) Fig. 7: Orientation of LC induced by magnetic field. (a) Alignment direction parallel to the polarizer axis of the microscope. (b) Alignment direction at 45 to the polarizer. A and P are analyzer and polarizer directions of the microscope, n is the director alignment direction.

40 29 satisfactory uniform planar orientation of LC in these cells was visually confirmed under a polarizing microscope (figure 7). The LC alignment direction in the cell was along the direction of H and independent of the filling direction. This suggested that the magnetic field strength was sufficient to achieve uniform planar orientation of LC in the bulk and at the surfaces of the cell Measurement of twist angle To determine the azimuthal anchoring energy at the adsorbed alignment layer, the symmetric cells were disassembled. These substrates with adsorbed LC layers were assembled into cells (figure 8) in which an ITO-glass covered with a rubbed polyimide (PI2555) film were used as second substrate. The chemical structure of PI 2555 was shown in figure 5. The LC coated substrates were our test surfaces while the PI coated substrates served as the reference substrate. A solution of PI2555 at a concentration of 1.3% (wt) was prepared. The solution was spin-coated on ITO/glass substrates at 3000 rpm for 30 second. The PI coated substrates were first soft-baked at 90 C for 3 minute to evaporate the solvent. Then, they were hard-baked at 250 C for 2 hour to finalize the imidization reaction to close polyimide ring. This reference substrate provided negligibly small pretilt and good quality planar alignment of 5CB. Therefore, the direction of the easy axis on the reference surface did not change during the experiment. Combined cells were made in a way that the angle, ϕ 0, between the rubbing direction of the reference substrate and the alignment direction on the test substrate, was 45. Since the thickness of

41 30 e ref n test Reference Substrate ϕ test ϕ 0 e test Test Substrate Fig. 8: Schematic representation of the hybrid test cell prepared to measure the twist angle.

42 31 the combined cell, L = 10 µm, was larger than that of the symmetric cell and some of the LC was lost during its disassembly, additional 5CB was filled in the isotropic phase in vacuum. Observation under a polarizing microscope showed a uniformly aligned twist structure (figure 9), which confirmed that the anchoring energy on the test surface was high enough to support twist deformation. According to Mauguin, for light propagation along the twist axis, the polarization of both ordinary and extraordinary waves follow the local director when the pitch P of the helicoidal twist is much larger than the wavelength λ of light (the so-called Mauguin regime [43] ). The Mauguin regime for the propagation of light was valid in our case. Therefore, the probe beam followed the rotation of the director and its polarisation on the output test surface was parallel to the director on the test surface. It allowed the measurement of the twist angle, ϕ test, through the angle between polarizer and analyzer. [26] The experimental set up to measure the twist angle is shown as figure 10. Test cells were placed between the polarizer and analyzer with the polarization direction of the incident beam parallel to the easy axis on the rubbed polyimide coated reference substrate. Transmission was recorded as the analyzer was rotated. When the polarization direction of the analyzer became parallel to the easy axis on the test substrate, maximum light was transmitted and the detector showed the largest reading. At this moment, the twist angle between the easy axes on the two cell substrates was the angle between the polarization directions of the polarizer and analyzer. Average twist angle measured at room temperature was 42 o as shown in figure 11.

43 Fig. 9: homogeneous twist structure of the combined cell. 32

44 33 Analyzer Sample Polarizer Detector Laser Fig. 10: Experimental set up for measuring twist angle. The source of the laser beam is HeNe laser operating at 10mw of power and wavelength of λ = 633nm. The laser beam passes through a polarizer, the test cell, and analyzer before reaching a Premier laser diode stand detector.

45 34 9 twist angle is 42 degree 8 Transmission Angle between the polarization directions of the polarizer and analyzer (deg.) Fig. 11: The dependence of transmission changed on the analyzer angle of rotation. When the axis of the analyzer was parallel to the easy axis on the test substrate, the detector showed the largest transmission. In this case, the twist angle between the easy axes on the two cell substrates is the angle between polarizer and analyzer at maximum transmission. Twist angle was 42 close to the intended value of 45.

46 Dependence of the twist angle on thermal annealing time A very slow decrease in the measured ϕ test with time was observed at room temperature. This result was mainly caused by the drift of the easy axis of the test substrate toward the easy axis of the reference surface. [26,27,39,43] However, a rapid change in ϕ test was measured when the cell was annealed at an elevated temperature T > T NI. The decrease was faster at a higher annealing temperature (figure 12). To check if this decrease was caused by a drift of the easy axis or a decrease in the anchoring energy of the test surface, a 45 o -twist combined cell was annealed in the isotropic phase (T = 90 o C) for 12h. After cooling to room temperature, this cell revealed significant decrease of the twist angle (ϕ test = 22 o ). This could have happened only if either (a) the anchoring energy decreased significantly, or (b) if the easy axis on the test substrate drifted towards the easy axis of the reference substrate. To determine which mechanism was at play, the cell was disassembled and the easy axis on the test substrate rotated by 45 o toward the rubbing axis on the reference substrate as shown in figure 13. In this geometry, the cell demonstrated a perfect planar alignment (figure 14) with zero twist that denotes no change of the easy axis on the test surface. Thus, the decrease of the angle ϕ test after annealing was caused by a decrease of the anchoring energy on the test surface. If it had been caused by rotation of the easy axis, nonzero twist would have been measured.

47 36 Twist Angle (deg.) o C 80 o C 85 o C 90 o C 95 o C 100 o C 120 o C Time (hours) Fig. 12: Dependence of the twist angle on thermal annealing time at different temperatures.

48 37 Reference Substrate e ref e test Test Substrate Fig. 13: Schematic representation of the hybrid test cell used to confirm a decrease of the anchoring energy on the test substrate.

49 38 A n P (a) A n P (b) Fig. 14: These two pictures show perfect planar alignment and prove that the easy axis on the test substrate did not drift after thermal annealing and the decrease of the angle ϕ test after annealing is caused by a decrease of the anchoring energy on the test surface. (a) Easy axis on the reference substrate in the cell is parallel to the polarizer axis. (b) Easy axis on the reference substrate makes an angle of 45 to the polarizer. A and P are analyzer and polarizer of the microscope, n is the easy axis on the reference substrate.

50 Theoretical model and discussion Below, a model is described which yields the direction of the easy axis and the value of the anchoring energy from the angular distribution function of the molecules adsorbed on the substrate. The changes in the angular distribution of the adsorbed molecules arise from AD-processes mentioned above. First, consider the formation of the adsorbed layer in a symmetric cell. During cooling, just after transition from the isotropic phase to the nematic phase, molecules in the bulk of LC are preferentially oriented along the direction determined by H. In this case, the adsorbed molecules, which mimic the angular distribution of the molecules in the bulk, cause the easy axis, e, to be parallel to H. Assuming that the angular distribution of the long axes of adsorbed molecules F S ( s ϕ ) is determined as a projection of the bulk distribution FV ( ΩV ) onto the surface: F S π ( ϕ ) = F ( Ω ) S 0 V V sinθ dθ Here, θ V is the polar angle between the long axis of molecule and the normal to the substrate, ϕ = ϕ is the azimuth angle between the long axis of molecules and the easy S V V V axis, e, Ω is the solid angle θ, ϕ ) in the polar reference frame. The angular V ( V V distribution function of LC molecules in the bulk in the Maier-Saupe approach is given by: [34] * Theoretical calculation was cooperatived with Kostyantyn Slyusarenko.