Chen, Kejian. / Thesis (Ph.D.)--Chinese University of Hong Kong, 2009. / Includes bibliographical references (leaves 156-164). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstracts in English and Chinese.

O titânio é considerado um biomaterial (material biocompatível) pela baixa reatividade e pela grande estabilidade da sua camada de óxido superficial. Pelas suas propriedades mecânicas ele é usualmente utilizado na fabricação de substitutos ósseos e implantes dentários. Visando contribuir para o desenvolvimento de novas superfícies osteointegráveis, neste trabalho foram estudadas duas superfícies de titânio modificadas com implantação iônica de P e BF2 com energia de 30 e 77keV, respectivamente. As superfícies foram submetidas a ensaios in vitro de cultura celular de células osteoblásticas (MG-63) e os resultados foram comparados com duas superfícies comerciais: poliestireno tratado para meio de cultura celular (PC) e titânio comum utilizado em implante dentário (TI). A atividade enzimática da fosfatase alcalina (FA) associadas à formação de biofilme mineralizado pelas células diferenciadas foram avaliadas. Concentrações de cálcio e fósforo no biofilme foram determinadas por espectroscopia de raios-X induzida por prótons, PIXE. Imagens superficiais com microscopia eletrônica de varredura serviram para verificar a homogeneidade e integridade dos biofilmes formados. Medidas de ângulo de contato foram realizadas para determinar o caráter hidrofílico das superfícies e do biofilme formado. As superfícies PC e TI se apresentaram hidrofílicas com formação de biofilme menos homogêneo que as apresentadas pelas superfícies PT e BFT, ambas hidrofóbicas. As concentrações de FA medidas no meio de cultura indicaram consumo maior que a secreção em PC e TI. Em PT e BFT foram observados acúmulos de FA no meio de cultura. Concentrações maiores de cálcio e fósforo foram observadas em PC e TI. Destas observações, as quatro superfícies podem ser consideradas biocompatíveis (não citotóxicas) pela formação de biofilme mineralizado e pelas medidas de atividade enzimática. / Titanium is considered a biomaterial (biocompatible material) due to its low reactivity and the great stability of its surface native oxide layer. With excellent mechanical properties, titanium is widely used as bone substitutes and dental implants. Aiming to contribute for the development of new surfaces to promote osteoblastic cell grow, two ion-beam-modified titanium samples were compared with two commercially available surfaces: polystyrene treated for cellular cultures (PC), and regular titanium used in dental implants (TI). The modified titanium samples were ion beam implanted with 30 and 77 keV phosphorous and BF2, respectively. All surfaces were employed for in vitro assays using cultures of osteoblastic cells (MG-63). The Phosphatase Alkaline Enzymatic Activity (ALP) was measured and associated with the mineralized bio-film formed by the differentiated cells. The concentration of calcium and phosphorus in the mineralized bio-film was measured by Proton Induced X-ray Emission, PIXE. Scanning Electron Microscope images were used to access the bio-film homogeneity and integrity. Contact angle measurements were used to characterize the wettability of the surfaces and the resulting bio-film. In spite of being more hydrophilic, the surfaces PC and TI formed of a less homogeneous bio-film than the PT and BFT surfaces, which were less hydrophilic. The ALP in the cell culture medium indicated greater consumption than secretion for PC and TI, while for PT and BFT accumulation of ALP was observed. Concentrations of calcium and phosphorus were greater for PC and TI samples. The four surfaces can be considered biocompatible (not cytotoxic) due to mineral bio-film formation and enzymatic activity measurements.

In this research, selective and bottom-up manufacturing of silica nanowires on silicon (Si) and its applications has been investigated. Localized synthesis of these nanowires on Si was achieved by metal thin film catalysis and metal ion implantation based seeding approach. The growth mechanism of the nanowires followed a vapor-liquid-solid (VLS) mechanism. Mass manufacturing aspects such as growth rate, re-usability of the substrate and experimental growth model were also investigated. Further, silica nanowires were explored as surface enhanced Raman (SER) substrate and immunoassay templates towards optical and electrochemical detection of cancer biomarkers respectively. Investigating their use in photonic applications, optically active silica nanowires were synthesized by erbium implantation after nanowire growth and implantation of erbium as a metal catalyst in Si to seed the nanowires. Ion implantation of Pd in Si and subsequent annealing in Ar at 1100° C for 60 mins in an open tube furnace resulted in silica nanowires of diameters ranging from 15 to 90 nm. Similarly, Pt was sputtered on to Si and further annealed to obtain silica nanowires of diameters ranging from 50 to 500 nm. Transmission electron microscopy studies revealed the amorphous nature of the wires. In addition, nano-sized Pd catalyst was found along the body of the nanowires seeded by Pd implantation into Si. After functionalization of the wires with 3-AminoPropylTriMethoxySilane (APTMS), the Pd decorated silica nanowires served as an SER substrate exhibiting a sensitivity of 107 towards the detection of interleukin-10 (IL-10, a cancer biomarker) with higher spatial resolution. Voltammetric detection of IL-10 involved silica nanowires synthesized by Pd thin film catalysis on Si as an immunoassay template. Using the electrochemical scheme, the presence of IL-10 was detected down to 1fg/mL in ideal pure solution and 1 pg/mL in clinically relevant samples. Time resolved photoluminescence (PL) results from the Er doped silica nanowires indicate a sharp emission around 1.54 µm representative of the I13/2 to I15/2 transition in Erbium. Also, a five-fold increase in the PL intensity and 30% augment in luminescence life time have been observed in nanowires when compared to fused silica sample prepared under similar conditions. The experimental results indicate the potential of silica nanowires in a wide variety of applications such as the development of orthogonal biosensors, fabrication of metallic nanowires, and environmental sensing probes.

Doctor of Philosophy (PhD) / This thesis investigates ion energy distributions (IEDs) during plasma immersion ion implantation (PIII). PIII is a surface modification technique where an object is placed in a plasma and pulse biased with large negative voltages. The energy distribution of implanted ions is important in determining the extent of surface modifications. IED measurements were made during PIII using a pulse biased retarding field energy analyser (RFEA) in a capacitive RF plasma. Experimental results were compared with those obtained from a two dimensional numerical simulation to help explain the origins of features in the IEDs. Time resolved IED measurements were made during PIII of metal and insulator materials and investigated the effects of the use of a metal mesh over the surface and the effects of insulator surface charging. When the pulse was applied to the RFEA, the ion flux rapidly increased above the pulse-off value and then slowly decreased during the pulse. The ion density during the pulse decreased below values measured when no pulse was applied to the RFEA. This indicates that the depletion of ions by the pulsed RFEA is greater than the generation of ions in the plasma. IEDs measured during pulse biasing showed a peak close to the maximum sheath potential energy and a spread of ions with energies between zero and the maximum ion energy. Simulations showed that the peak is produced by ions from the sheath edge directly above the RFEA inlet and that the spread of ions is produced by ions which collide in the sheath and/or arrive at the RFEA with trajectories not perpendicular to the RFEA front surface. The RFEA discriminates ions based only on the component of their velocity perpendicular to the RFEA front surface. To minimise the effects of surface charging during PIII of an insulator, a metal mesh can be placed over the insulator and pulse biased together with the object. Measurements were made with metal mesh cylinders fixed to the metal RFEA front surface. The use of a mesh gave a larger ion flux compared to the use of no mesh. The larger ion flux is attributed to the larger plasma-sheath surface area around the mesh. The measured IEDs showed a low, medium and high energy peak. Simulation results show that the high energy peak is produced by ions from the sheath above the mesh top. The low energy peak is produced by ions trapped by the space charge potential hump which forms inside the mesh. The medium energy peak is produced by ions from the sheath above the mesh corners. Simulations showed that the IED is dependent on measurement position under the mesh. To investigate the effects of insulator surface charging during PIII, IED measurements were made through an orifice cut into a Mylar insulator on the RFEA front surface. With no mesh, during the pulse, an increasing number of lower energy ions were measured. Simulation results show that this is due to the increase in the curvature of the sheath over the orifice region as the insulator potential increases due to surface charging. The surface charging observed at the insulator would reduce the average energy of ions implanted into the insulator during the pulse. Compared to the case with no mesh, the use of a mesh increases the total ion flux and the ion flux during the early stages of the pulse but does not eliminate surface charging. During the pulse, compared to the no mesh case, a larger number of lower energy ions are measured. Simulation results show that this is caused by the potential in the mesh region which affects the trajectories of ions from the sheaths above the mesh top and corners and results in more ions being measured with trajectories less than ninety degrees to the RFEA front surface.

The deformation mechanisms and contact response of TiN-based thin films deposited onto a soft substrate using a physical vapour deposition (PVD) technique is still an area of both technological importance and considerable discussion. These coatings are commonly applied to various kinds of steel cutting tools, creating surfaces with enhanced tribological properties. However, no extensive systematic study of the deformation mechanisms in these thin film systems has been performed to date. In the present study, the effect of the coating microstructure, indenter geometry, coating thickness and substrate hardness on the deformation mechanisms in both TiN and TiAlN coatings of varying thickness deposited onto ductile steel substrates has been investigated using a combination of nanoindentation and microstructural analysis, including focused ion beam (FIB) milling and transmission electron microscopy (TEM). Different modes of cracking, such as columnar and transverse cracking, as well as shear steps at the coating/substrate interface, were observed. The microstructure of the TiN coatings was found to be very important in controlling their modes of deformation. Thicker coatings were seen to contain more equiaxed grains, so less columnar shearing occurred and inclined cracks were found to be a more dominant fracture type in the thicker coating. Also, it was found that soft substrates absorbed most of the energy from indentation by plastic deformation. It was found that both the TiN and TiAlN/TiN dual-layer coatings exhibited broadly similar mechanisms of deformation. The epitaxial interface between the TiAlN and TiN in the dual-layer coating did not appear to affect the deformation behaviour. As a further investigation of the overall deformation behaviour for the coating/substrate systems studied, a DualBeam FIB was used to generate three dimensional images of the indented regions which provided additional information on the crack morphology. For the first time, a systematic study of the deformation behaviour of TiN and TiAlN coatings upon indentation has been carried out. FIB milling was demonstrated to be a highly appropriate technique for characterization of the deformation behaviour of these coatings, allowing detailed, high resolution microstructural investigations to be performed in both two and three dimensions.

Group-III-nitride semiconductors (GaN, InGaN, and AlGaN) are important for the fabrication of a range of optoelectronic devices (such as blue-green light emitting diodes, laser diodes, and UV detectors) as well as devices for high-temperature/high-power electronics. In the fabrication of these devices, ion bombardment represents a very attractive technological tool. However, a successful application of ion implantation depends on an understanding of the effects of radiation damage. Hence, this thesis explores a number of fundamental aspects of radiation effects in wurtzite III-nitrides. Emphasis is given to an understanding of (i) the evolution of defect structures in III-nitrides during ion irradiation and (ii) the influence of ion bombardment on structural, mechanical, optical, and electrical properties of these materials. ¶ Structural characteristics of GaN bombarded with keV ions are studied by Rutherford backscattering/channeling (RBS/C) spectrometry and transmission electron microscopy (TEM). Results show that strong dynamic annealing leads to a complex dependence of the damage buildup on ion species with preferential surface disordering. Such preferential surface disordering is due to the formation of surface amorphous layers, attributed to the trapping of mobile point defects by the GaN surface. Planar defects are formed for a wide range of implant conditions during bombardment. For some irradiation regimes, bulk disorder saturates below the amorphization level, and, with increasing ion dose, amorphization proceeds layer-by-layer only from the GaN surface. In the case of light ions, chemical effects of implanted species can strongly affect damage buildup. For heavier ions, an increase in the density of collision cascades strongly increases the level of stable implantation-produced lattice disorder. Physical mechanisms of surface and bulk amorphization and various defect interaction processes in GaN are discussed. ¶ Structural studies by RBS/C, TEM, and atomic force microscopy (AFM) reveal anomalous swelling of implanted regions as a result of the formation of a porous structure of amorphous GaN. Results suggest that such a porous structure consists of N$_{2}$ gas bubbles embedded into a highly N-deficient amorphous GaN matrix. The evolution of the porous structure appears to be a result of stoichiometric imbalance, where N- and Ga-rich regions are produced by ion bombardment. Prior to amorphization, ion bombardment does not produce a porous structure due to efficient dynamic annealing in the crystalline phase. ¶ The influence of In and Al content on the accumulation of structural damage in InGaN and AlGaN under heavy-ion bombardment is studied by RBS/C and TEM. Results show that an increase in In concentration strongly suppresses dynamic annealing processes, while an increase in Al content dramatically enhances dynamic annealing. Lattice amorphization in AlN is not observed even for very large doses of keV heavy ions at -196 C. In contrast to the case of GaN, no preferential surface disordering is observed in InGaN, AlGaN, and AlN. Similar implantation-produced defect structures are revealed by TEM in GaN, InGaN, AlGaN, and AlN. ¶ The deformation behavior of GaN modified by ion bombardment is studied by spherical nanoindentation. Results show that implantation disorder significantly changes the mechanical properties of GaN. In particular, amorphous GaN exhibits plastic deformation even for very low loads with dramatically reduced values of hardness and Young's modulus compared to the values of as-grown GaN. Moreover, implantation-produced defects in crystalline GaN suppress the plastic component of deformation. ¶ The influence of ion-beam-produced lattice defects as well as a range of implanted species on the luminescence properties of GaN is studied by cathodoluminescence (CL). Results indicate that intrinsic lattice defects mainly act as nonradiative recombination centers and do not give rise to yellow luminescence (YL). Even relatively low dose keV light-ion bombardment results in a dramatic quenching of visible CL emission. Postimplantation annealing at temperatures up to 1050 C generally causes a partial recovery of measured CL intensities. However, CL depth profiles indicate that, in most cases, such a recovery results from CL emission from virgin GaN, beyond the implanted layer, due to a reduction in the extent of light absorption within the implanted layer. Experimental data also shows that H, C, and O are involved in the formation of YL. The chemical origin of YL is discussed based on experimental data. ¶ Finally, the evolution of sheet resistance of GaN epilayers irradiated with MeV light ions is studied {\it in-situ}. Results show that the threshold dose of electrical isolation linearly depends on the original free electron concentration and is inversely proportional to the number of atomic displacements produced by the ion beam. Furthermore, such isolation is stable to rapid thermal annealing at temperatures up to 900 C. Results also show that both implantation temperature and ion beam flux can affect the process of electrical isolation. This behavior is consistent with significant dynamic annealing, which suggests a scenario where the centers responsible for electrical isolation are defect clusters and/or antisite-related defects. A qualitative model is proposed to explain temperature and flux effects. ¶ The work presented in this thesis has resulted in the identification and understanding of a number of both fundamental and technologically important ion-beam processes in III-nitrides. Most of the phenomena investigated are related to the nature and effects of implantation damage, such as lattice amorphization, formation of planar defects, preferential surface disordering, porosity, decomposition, and quenching of CL. These effects are often technologically undesirable, and the work of this thesis has indicated, in some cases, how such effects can be minimized or controlled. However, the thesis has also investigated one example where irradiation-produced defects can be successfully applied for a technological benefit, namely for electrical isolation of GaN-based devices. Finally, results of this thesis will clearly stimulate further research both to probe some of the mechanisms for unusual ion-induced effects and also to develop processes to avoid or repair unwanted lattice damage produced by ion bombardment.

This thesis examines the optical properties of nanoscale silicon and the sensitization of Er with Si. In this context, it predominantly investigates the role of defects in limiting the luminescence of Si nanocrystals, and the removal of these defects by hydrogen passivation. The kinetics of the defect passivation process, for both molecular and atomic hydrogen, are studied in detail. Moreover, the optical absorption of Si nanocrystals and the effect of annealing environment (during nanocrystal synthesis) on the luminescence are investigated. The effect of annealing temperature and hydrogen passivation on the coupling (energy transfer) of Si nanocrystals to optically active centres (Er) is also examined.¶ The electronic structure of silicon-implanted silica slides is investigated through optical absorption measurements. Before and after annealing to form Si nanocrystals, optical absorption spectra from these samples show considerable structure that is characteristic of the particular implant fluence. This structure is shown to correlate with the transmittance of the samples as calculated from the modified refractive index profile for each implant. Due to the high absorption coefficient of Si at short wavelengths, extinction at these wavelengths is found to be dominated by absorption. As such, scattering losses are surprisingly insignificant. To eliminate interference effects, photothermal deflection spectroscopy is used to obtain data on the band structure of Si in these samples. This data shows little variance from bulk Si structure and thus little effect of quantum confinement. This is attributed to the dominance of large nanocrystals in the absorption measurements.¶ The effect of annealing environment on the photoluminescence (PL) from silicon nanocrystals synthesized in fused silica by ion implantation and thermal annealing is studied as a function of annealing temperature and time. Interestingly, the choice of annealing environment (Ar, N2, or 5 % H2 in N2) is found to affect the shape and intensity of luminescence emission spectra, an effect that is attributed both to variations in nanocrystal size and the density of defect states at the nanocrystal/oxide interface.¶ The passivation kinetics of luminescence-quenching defects, associated with Si nanocrystals in SiO2, during isothermal and isochronal annealing in molecular hydrogen are studied by time-resolved PL. The passivation of these defects is modeled using the Generalized Simple Thermal model of simultaneous passivation and desorption, proposed by Stesmans. Values for the reaction-rate parameters are determined for the first time and found to be in excellent agreement with values previously determined for paramagnetic Si dangling-bond defects (Pb type centers) found at planar Si/SiO2 interfaces; supporting the view that non-radiative recombination in Si nanocrystals is dominated by such defects.¶ The passivation kinetics of luminescence-quenching defects during isothermal and isochronal annealing in atomic hydrogen are studied by continuous and time-resolved PL. The kinetics are compared to those for standard passivation in molecular hydrogen and found to be significantly different. Atomic hydrogen is generated using the alneal process, through reactions between a deposited Al layer and H2O or –OH radicals in the SiO2. The passivation and desorption kinetics are shown to be consistent with the existence of two classes of nonradiative defects: one that reacts with both atomic and molecular hydrogen, and the other that reacts only with atomic hydrogen. A model incorporating a Gaussian spread in activation energies is presented that adequately describes the kinetics of atomic hydrogen passivation and dissociation for the samples.¶ The effect of annealing temperature and hydrogen passivation on the excitation cross-section and PL of erbium in silicon-rich silica is studied. Samples are prepared by co-implantation of Si and Er into SiO2 followed by a single thermal anneal at temperatures ranging from 800 to 1100 degrees C, and with or without hydrogen passivation performed at 500 degrees C. Using time-resolved PL, the effective erbium excitation cross-section is shown to increase by a factor of 3, while the number of optically active erbium ions decreases by a factor of 4 with increasing annealing temperature. Hydrogen passivation is shown to increase the luminescence intensity and to shorten the luminescence lifetime at 1.54 micron only in the presence of Si nanocrystals. The implications of these results for realizing a silicon-based optical amplifier are also discussed.

In the context of transistor size miniaturization the motivation of this work was focused on the fabrication process of planar double gate devices. We proposed in this work three process flows based on the use of buried mask which could allow the fabrication of self-aligned planar double gate transistors. The novel concept of buried mask consists into modifying the etch rate of a buried polysilicon or oxide layer. This etch rate modification being defined by ion implantation, etch stop or scacrificial zones aligned with the implantation mask can thus be fabricated. This technique solve the alignment of the front and back gate. Ion implantation causes damages to the implanted target, and is used to dope semiconductor material. If the implanted atoms have a small radii they can induce stress to the implanted lattice. These three consequences of ion implantation, damage, doping and stress are used to modify the etch rate of oxide and polysilicon. High etching selectivity are reached, which allow the fabrication of a localized buried sacrificial or etch stop zone, called buried mask. The definition of the buried mask being done by ion implantation, it opens the possibility to fabricate a buried mask aligned with the implantation mask. Although some more work has to be invested to fabricate planar double gate MOS using buried mask in polysilicon, this concept of buried mask, which could also be called anisotropic wet and vapor etching, is foreseen as a very promising technique in MEMS micromachining and for bio sensor applications.

This thesis details the design and development of a solid-state, high power modulator for driving plasma ion implantation systems. A plurality of modulators can be stacked in a Marx geometry to allow complete voltage (implantation energy) scalability. Unlike a classic Marx modulator, the design employs actively controlled charging and discharging paths. This allows maximum modulation flexibility and efficiency. A hybrid Marx bank - pulse transformer configuration was commissioned in a 20keV 12A plasma ion implantation system for the purpose of photonics research. <p>The design portion of this work is accompanied by an investigation, extension and discretization of the Lieberman analytical model of plasma ion implantation dynamics. The model predicts final implantation concentrations as well as system operational limits in specific plasma conditions. A new extension to the model accounts for subtle time-of-flight effects on accelerating ions. Agreement between modeled and measured ion currents is good.<p>Finally, a collection of material processing experiments conducted with the plasma ion implantation system since its inauguration in February 2006 is briefly presented. In it, a new silicon-based light emitting diode is introduced.

This thesis describes the fabrication and testing of electroluminescent diodes made from silicon subjected to plasma ion implantation. A silicon-compatible, electrically driven light source is desired to increase the speed and efficiency of short-range data transfer in the communications and computing industries. As it is an indirect band gap material, ordinary silicon is too inefficient a light source to be useful for these applications. Past experiments have demonstrated that modifying the structural properties of the crystal can enhance its luminescence properties, and that light ion implantation is capable of achieving this effect. This research investigates the relationship between the ion implantation processing parameters, the post-implantation annealing temperature, and the observable electroluminescence from the resulting silicon diodes.<p> Prior to the creation of electroluminescent devices, much work was done to improve the efficiency and reliability of the fabrication procedure. A numerical algorithm was devised to analyze Langmuir probe data in order to improve estimates of implanted ion fluence. A new sweeping power supply to drive current to the probe was designed, built, and tested. A custom software package was developed to improve the speed and reliability of plasma ion implantation experiments, and another piece of software was made to facilitate the viewing and analysis of spectra measured from the finished silicon LEDs.<p> Several dozen silicon diodes were produced from wafers implanted with hydrogen, helium, and deuterium, using a variety of implanted ion doses and post-implantation annealing conditions. One additional device was fabricated out of unimplanted, unannealed silicon. Most devices, including the unimplanted device, were electroluminescent at visible wavelengths to some degree. The intensity and spectrum of light emission from each device were measured. The results suggest that the observed luminescence originated from the native oxide layer on the surface of the ion-implanted silicon, but that the intensity of luminescence could be enhanced with a carefully chosen ion implantation and annealing procedure.</p>