Ferroelectric liquid crystal over silicon devices

Over the past 20 years, ferroelectric liquid crystal over silicon (FLCOS) devices have made a wide impact on applications as diverse as optical correlation and holographic projection. To cover the entire gamut of this technology would be difficult and long winded; hence, this paper describes the significant developments of FLCOS within the Engineering Department at the University of Cambridge.

The purpose of this paper is to highlight the key issues in fabricating silicon backplane spatial light modulators (SLMs) and to indicate ways in which the technology can be fabricated using cheap, low-density production and manufacturability. Three main devices have been fabricated as part of several research programmes and are documented in this paper. The fast bitplane SLM and the reconfigurable optical switches for aerospace and telecommunications systems (ROSES) SLM will form the basis of a case study to outline the overall processes involved. There is a great deal of commonality in the fabrication processes for all three devices, which indicates their potential strength and demonstrates that these processes can be made independent of the SLMs that are being assembled. What is described is a generic process that can be applied to any silicon backplane SLM on a die-by-die basis. There are hundreds of factors that can affect the yield in a manufacturing process and the purpose of a good process design procedure is to minimise these factors. One of the most important features in designing a process is fabrication experience, as so many of the lessons in this business can only be learned this way. We are working with the advantage of knowing the mistakes already made in the flat panel display industry, but we are also faced with the fact that those mistakes took many years and many millions of dollars to make.

The fabrication process developed here originates and adapts earlier processes from various groups around the world. There are also a few totally new processes that have now been adopted by others in the field. Many, such as the gluing process, are still on-going and have to be worked on more before they will fully suit ‘manufacturability’.     

Time-dependent simulation of Czochralski silicon crystal growth

We have developed a detailed mathematical model and numerical simulation tools based on the streamline upwind/Petrov-Galerkin (SUPG) finite element formulation for the Czochralski silicon crystal growth. In this paper we consider the mathematical modeling and numerical simulation of the time-dependent melt flow and temperature field in a rotationally symmetric crystal growth environment. Heat inside the Czochralski furnace is transferred by conduction, convection and radiation, Radiating surfaces are assumed to be opaque, diffuse and gray. Hence the radiative heat exchange can be modeled with a non-local boundary condition on the radiating part of the surface. The position of the crystal-melt interface is solved by the enthalpy method. The melt flow is assumed to be laminar and governed by the cylindrically symmetric and incompressible Navier-Stokes equations coupled with the calculation of temperature.     

X-ray diffraction studies of protein crystal disorder

Protein crystals contain many kinds of disorder, but only a small fraction of these are likely to be important in limiting the diffraction properties of interest to crystallographers. X-ray topography, high-angular-resolution reciprocal space measurements, and standard crystallographic data collection have been used to probe three factors that may produce diffraction-limiting disorder: (1) solution variations during crystal growth, (2) macromolecular impurities, and (3) post-growth crystal treatments. Variations in solution conditions that occur in widely used growth methods may lead to variations in equilibrium protein conformation and crystal packing as a crystal grows, and these may introduce appreciable disorder for sensitive proteins. Tetragonal lysozyme crystals subjected to abrupt changes in temperature, pH, or salt concentration during growth show increased disorder, consistent with this mechanism. Macromolecular impurities can have profound effects on protein crystal quality. A combination of diffraction measurements provides insight into the mechanisms by which particular impurities create disorder, and this insight leads to a simple approach for reducing this disorder. Substantial degradation of diffraction properties due to conformation and lattice constant changes can occur during post-growth crystal treatments such as heavy-atom compound and drug binding. Measurements of the time evolution of crystal disorder during controlled crystal dehydration – a simple model for such treatments – suggest that structural metastability conferred by the constraints of the crystal lattice plays an important role in determining the extent to which the diffraction properties degrade.     

Investigation of the thermal conditions during silicon carbide crystal growth

In the present work an analysis of the thermal conditions during silicon carbide crystal growth from the vapour phase by the sublimation method is carried out. On the basis of the obtained results from the calculation of the temperature distribution along the length of the growing crystal it was drawn a conclusion that in order to decrease the density of dislocations in the growing crystals it is necessary to decrease the temperature gradients in the crucible for growing.     

Use of an inhomogenous magnetic field for silicon crystal growth

The flow of liquid silicon and oxygen transfer during crystal growth under three different types of cusp-shaped magnetic field were clarified using numerical simulation, flow visualization, and infrared measurement of oxygen concentration in grown crystals. Velocity vectors obtained from numerical simulation are almost parallel to cusp-shaped magnetic fields since flow parallel to a magnetic field does not produce a Lorentz force. This parallel flow enhances homogenization of oxygen concentration along the radial direction in grown crystals. Cusp-shaped magnetic fields can control the flow velocity at the top of the melt. Since melt with a low concentration of oxygen at the top of the melt transfers directly from the free surface to the solid-liquid interface, a low concentration of oxygen in crystals can be achieved. Separation of fluid flow between the near surface and bulk can produce a spatial distribution of the concentration in the melt, and therefore a low oxygen concentration can be obtained in grown crystals.     

Reflection image enhancement for photonic crystal devices by using dielectrics

Abstract

We investigated the optical properties of photonic crystal devices (PCDs) using dielectrics. Different dielectrics were injected into a cell gap of the PCDs as a swelling solvent. It is evident that the PCDs reflected a deep blue color when two different materials, chiral ionic liquid (CIL) and photosensitive small molecules, were introduced. To compare the reflection images according to the different dielectrics, a well-known ionic liquid (IL) was used as a control sample. A thinner polymer layer induced a shorter wavelength, which created a strong blue shift phenomenon with a larger refraction index and a larger dielectric constant. In this paper, we obtained 12?nm of reflection enhancement with an applied voltage of 2V using an IL and a deep-blue color image by using the effects of the molecular structure of the CIL and photosensitive materials.     

Mechanism for generating interstitial atoms by thermal stress during silicon crystal growth

It has been known that, in growing silicon from melts, vacancies (Vs) predominantly exist in crystals obtained by high-rate growth, while interstitial atoms (Is) predominantly exist in crystals obtained by low-rate growth. To reveal the cause, the temperature distributions in growing crystal surfaces were measured. From this result, it was presumed that the high-rate growth causes a small temperature gradient between the growth interface and the interior of the crystal; in contrast, the low-rate growth causes a large temperature gradient between the growth interface and the interior of the crystal. However, this presumption is opposite to the commonly-accepted notion in melt growth. In order to experimentally demonstrate that the low-rate growth increases the temperature gradient and consequently generates Is, crystals were filled with vacancies by the high-rate growth, and then the pulling was stopped as the extreme condition of the low-rate growth. Nevertheless, the crystals continued to grow spontaneously after the pulling was stopped. Hence, simultaneously with the pulling-stop, the temperature of the melts was increased to melt the spontaneously grown portions, so that the diameters were restored to sizes at the moment of pulling-stop. Then, the crystals were cooled as the cooling time elapsed, and the temperature gradient in the crystals was increased. By using X-ray topographs before and after oxygen precipitation in combination with a minority carrier lifetime distribution, a time-dependent change in the defect type distribution was successfully observed in a three-dimensional manner from the growth interface to the low-temperature portion where the cooling progressed. This result revealed that Vs are uniformly introduced in a grown crystal regardless of the pulling rate as long as the growth continues, and the Vs agglomerate as a void and remain in the crystal, unless recombined with Is. On the other hand, Is are generated only in a region where the temperature gradient is large by low-rate growth. In particular, the generation starts near the peripheral portion in the vicinity of the solid–liquid interface. First, the generated Is are recombined with Vs introduced into the growth interface, so that a recombination region is always formed which is regarded as substantially defect free. Excessively generated Is after the recombination agglomerate and form a dislocation loop region. Unlike conventional Voronkov's diffusion model, Is hardly diffuse over a long distance. Is are generated by re-heating after growth.[In a steady state, the crystal growth rate is synonymous with the pulling rate. Meanwhile, when an atypical operation is performed, the pulling rate is specifically used.]This review on point defects formation intends to contribute further silicon crystals development, because electronic devices are aimed to have finer structures, and there is a demand for more perfect crystals with controlled point defects.     

Estimated effects of silicone glue on protein crystal growth

Silicone glue (modified silicone polymer) is widely used for both experiments involving inorganic crystal growth and those involving organic materials like proteins. This material is very useful for building a hand-made experiment setup or for fixing protein crystals to specific locations. Though silicone glue is regarded as harmful to proteins, no systematic verification was performed to investigate its impurity effects on protein crystal growth. We focused on and estimated the impurity effects of silicone glue on protein crystal growth.     

Numerical investigation of carbon and silicon carbide contamination during the melting process of the Czochralski silicon crystal growth

Carbon contamination in single crystalline silicon is detrimental to the minority carrier lifetime, one of the critical parameters for electronic wafers. In order to study the generation and accumulation of carbon contamination, transient global modeling of heat and mass transport was performed for the melting process of the Czochralski silicon crystal growth. Carbon contamination, caused by the presence of carbon monoxide in argon gas and silicon carbide in the silicon feedstock, was predicted by the fully coupled chemical model; the model included six reactions taking place in the chamber. A simplified model for silicon carbide generation by the reaction between carbon monoxide and solid silicon was proposed using the closest packing assumption for the blocky silicon feedstock. The accumulation of carbon in the melted silicon feedstock during the melting and stabilization stages was predicted. Owing to this initial carbon content in the melt, controlling carbon contamination before the growth stage becomes crucial for reducing the carbon incorporation in a growing crystal.     

Combined effects of crucible geometry and Marangoni convection on silicon Czochralski crystal growth

In order to understand the influence of crucible geometry combined with natural convection and Marangoni convection on melt flow pattern, temperature and pressure fields in silicon Czochralski crystal growth process, a set of numerical simulations was conducted. We carry out calculation enable us to determine temperature, pressure and velocity fields in function of Grashof and Marangoni numbers. The essential results show that the hemispherical geometry of crucible seems to be adapted for the growth of a good quality crystal and the pressure field is strongly affected by natural and Marangoni convection and it is more sensitive than temperature. (© 2009 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)     

Kinetic studies on the influence of temperature and growth rate history on crystal growth

Crystallization experiments of sucrose were performed in a batch crystallizer to study the effect of temperature and growth rate history on the crystal growth kinetics. In one of the growth methods adopted, the isothermal volumetric growth rate (R_V) is determined as a function of supersaturation (S) at 35, 40 and 45 ºC. In the other, crystals are allowed to grow at constant supersaturation by automatically controlling the solution temperature as the solute concentration decreased. Using the latter method R_V is calculated as the solution is cooled. The obtained results are interpreted using empirical, engineering and fundamental perspectives of crystal growth. Firstly, the overall activation energy (E_A) is determined from the empirical growth constants obtained in the isothermal method. The concept of falsified kinetics, widely used in chemical reaction engineering, is then extended to the crystal growth of sucrose in order to estimate the true activation energy (E_T) from the diffusion‐affected constant, E_A. The differences found in the isothermal and constant supersaturation methods are explained from the viewpoint of the spiral nucleation mechanism, taking into account different crystal surface properties caused by the growth rate history in each method. Finally, the crystal growth curve obtained in the batch crystallizer at 40 ºC is compared with the one obtained in a fluidized bed crystallizer at the same temperature. Apparently divergent results are explained by the effects of crystal size, hydrodynamic conditions and growth rate history on the crystallization kinetics of sucrose. (© 2008 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)     

Centrifugal separation of primary silicon crystal in solvent refining of silicon using Al‐30% Si alloy

This study describes the formation of a primary silicon network during separation of an Al‐30% Si melt and the process conditions to make bigger primary silicon crystals. As the crucible rotates in the centrifugal separation, the unfrozen aluminum‐rich phase and small silicon particles are pushed outside through the openings in the silicon network. As a result, primary silicon crystals are separated in the form of a foam after centrifugation. The recovery of the silicon ranged from 13 to 18% depending on the location in the crucible. The size of the primary silicon achieved by changing the cooling rate and quenching temperature during solidification is also measured using a quenching furnace. Primary silicon particles exhibit a coarse, plate‐like morphology, although small star‐like silicon particles are also found in the aluminum‐rich matrix. The fraction of plate silicon decreases, while the fraction of small globular silicon increases with an increasing cooling rate. The thickness of the primary silicon plate also decreases with an increasing cooling rate in the samples quenched at various temperatures during solidification.     

Single crystal growth of copper phthalocyanine using exaltation–evaporation growth method

A new method of growing single crystal for β-form copper phthalocyanine (CuPc) is presented in this paper. Melted anthracene was used as solvent of CuPc. The method, vaporizing the solvent using an automatic exaltation machine, was employed to grow CuPc single crystals. The needle-like single crystals of CuPc up to 11.6 mm in length were obtained by applying this method. The influences of different temperatures, exaltation speeds and concentrations on the single crystals growth were also discussed. The method was called exaltation–evaporation growth method.     

Crystalline growth of germanium thin films on single crystal silicon substrates by solid phase crystallization

We have studied the epitaxial-like growth of germanium (Ge), due to solid phase crystallization (SPC) from amorphous Ge (a-Ge) deposited on single crystal silicon (Si) substrate. The crystalline growth of Ge following the orientation of Si substrates was successfully obtained by the SPC at 400 °C or higher. The preferential growth on Si (111) substrates continues up to 10,000 Å. Different orientations from the substrate orientation in XRD patterns are slightly observed in the growth on Si (100) substrates at 450 °C, but the preferential growth of (100) orientation continued in the whole film thickness in TEM images. The epitaxial-like growth of Ge may be more preferable on the Si (111) substrate than the (100) one.     

Global simulation of a Czochralski furnace for silicon crystal growth against the assumed thermophysical properties

In order to understand the effects of the thermophysical properties of the melt on the transport phenomena in the Czochralski (Cz) furnace for the single crystal growth of silicon, a set of global analyses of momentum, heat and mass transfer in small Cz furnace (crucible diameter: 7.2 cm, crystal diameter: 3.5 cm, operated in a 10 Torr argon flow environment) was carried out using the finite‐element method. The global analysis assumed a pseudosteady axisymmetric state with laminar flow. The results show that different thermophysical properties will bring different variations of the heater power, the deflection of the melt/crystal interface, the axial temperature gradient in the crystal on the center of the melt/crystal interface and the average oxygen concentration along the melt/crystal interface. The application of the axial magnetic field is insensitive to this effect. This analysis reveals the importance of the determination of the thermophysical property in numerical simulation. (© 2006 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)     

Applicability of LES turbulence modeling for CZ silicon crystal growth systems with traveling magnetic field

To examine the applicability of LES turbulence modeling for CZ silicon crystal growth systems with traveling magnetic fields, LES calculations with Smagorinsky–Lilly turbulence model and van Driest damping at the solid walls are carried out. The program package for the calculations was developed on the basis of the open-source code library OpenFOAM^®${\mathrm{OpenFOAM}}^{\circledR }$. A previously published laboratory model with low temperature melt InGaSn, a 20” crucible, and process parameters corresponding to industrial Czochralski silicon systems is considered. Flow regimes with two crystal and crucible rotation rates and with different strengths of the traveling magnetic field “down” are analyzed. The calculated distributions of averaged temperature and standard temperature deviations are compared with measured ones in the laboratory system, and a relatively good agreement between them is shown. The influence of chosen time steps and grid sizes is analyzed by comparing Fourier spectra of temperature time-autocorrelation functions and temperature spatial distributions, and it is shown that the used moderate meshes of few hundred thousand cells can be applied for practical calculations.     

Investigation of GaN crystal quality on silicon substrates using GaN/AlN superlattice structures

The crystal quality of GaN thin film on silicon using GaN/AlN superlattice structures was investigated. The growth was carried out on Si(111) for GaN(0001) in a metal‐organic vapor phase epitaxy system. Various GaN/AlN superlattice intermediate layers have been designed to decrease the dislocation density. The results showed that the etch pit density could be greatly reduced by one order of magnitude. Cross‐sectional transmission electron microscopy (XTEM) study confirmed the efficiency of GaN/AlN superlattice in blocking threading dislocation propagation in GaN crystal. The design of nine period GaN/AlN (20nm/2nm) superlattice has been evidenced to be effective in reducing the dislocation density and improving the crystal quality. In addition, the dislocation bending in GaN/AlN interface and dislocation merging is investigated. (© 2007 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)     

Nonlinear dynamics of layer growth and consequences for protein crystal perfection

In situ high-resolution optical interferometry of lysozyme crystal growth reveals that under steady external conditions, the local growth rate R, vicinal slope p and step velocity are not steady but fluctuate by several times their average values. The variations in p, which is proportional to the local step density, indicate that these fluctuations occur through the dynamic formation of step bunches. Our previous work with unstirred solutions has shown that the fluctuation amplitude of R increases with supersaturation and crystal size (Vekilov et al., Phys. Rev. E 54 (1996) 6650). Based on scaling arguments and numerical simulations, we have argued that the fluctuations are the response of the coupled bulk transport and nonlinear interface kinetics to finite amplitude perturbations provided by the intrinsically unsteady step generation. In this paper, we emphasize the recently discovered spatio-temporal correlation between the sequence of moving step bunches and striations (compositional variations) in the crystal, visualized by polarized-light microscopy. Hence, these unsteady kinetics have detrimental effects on the perfection of the crystals, and means to reduce and eliminate them should be sought. To this end, based on the above conclusion as to the mechanism of the kinetic unsteadiness, we accelerated the bulk transport towards the interface by forced solution flow. We found that this results in lower fluctuation amplitudes. This observation confirms that the system-dependent kinetic Peclet number, Pe_k, i.e., the relative weight of bulk transport and interface kinetics in the control of the growth process, governs the step bunching dynamics. Since Pe_k can be modified by either forced solution flow or suppression of buoyancy-driven convection under reduced gravity, this model provides a rationale for the choice of specific transport conditions to minimize the formation of compositional inhomogeneities. Interestingly, on further increase of the solution flow velocities >500 μm/s, the fluctuation amplitudes in R increased again, while the average growth rate decreased. At low supersaturations, this leads to growth cessation. The growth instability, deceleration and cessation were immediately reversible upon reduction of the flow velocity. When solutions, intentionally contaminated with ∼1% of covalent lysozyme dimer were used, these undesirable phenomena occurred at about half the flow rates required in pure solutions. Thus, we conclude that enhanced convective supply of impurities to the interface causes an increase in step-bunching related defects, growth deceleration and, in some cases, cessation. Finally, we correlate the “slow protein crystal growth” to step bunch formation. We show that in the absence of significant step density variations, the kinetic coefficient for step propagation is as high as 4×10⁻³ cm/s, which is 1–2 orders of magnitude higher than the previously determined, apparent values for any protein.