Germanium is an attractive channel material for MOSFETs because of its higher mobility than silicon. Ge02 has been investigated as an interfacial layer for high-K gate stacks on germanium. Thermally grown Ge02 layers have been prepared at 550°C to minimize GeO volatilization. Ge02 growth has been performed in both pure O2 ambient and O2 diluted with N2. Ge02 thickness has been scaled down to approximately 3 nm. This is the first time ever Ge02 has been scaled using dilution method. MOS capacitors have been fabricated using Ge02 grown in various ambient and different thicknesses with a standard high-K dielectric on top. Electrical properties and thermal stability have been tested up to at least 350°C. The K value of Ge02 was experimentally determined to be 4.5. Interface state densities (Did of less than 1012 cm-2eV-1 have been extracted for all devices using the conductance method.

No description available.

Conduction current measurements have been carried out in a variety of highly degassed hydrocarbon liquids using plane and spherical electrodes, at field strengths between 50kV/cm and breakdown. These liquids have been analysed by means of a high resolution chromatography apparatus and a marked dependence of the reproducibility of the results and the purity of the liquid was found. Steady currents in the region of 10 -6 A have been recorded, these being greater by several orders of magnitude than those found by other investigators under similar conditions. This high value of current has been established to be due to the high degree of degassing of the liquids used in this project. Measurements were made on the instability of the current, fast pulses in the microsecond region, particle movement and breakdown. No relationship was found between any of the electrical properties of the gap and the state of the electrode surfaces or the material used, although there was a marked change in the maximum value of the conduction current and some variation in the pattern of the instability when different liquids were tested. When using 3mm diameter, stainless steel, parallel plane electrodes the breakdown strength was found to be approximately 400kV/cm and independent of the liquid used. A mechanism has been proposed, based on realignment of the liquid molecules under the influence of the high electric field, and appears to be well able to explain the majority of the experimental observations.

Observations of enhanced conductivity in ferroelectric domain walls have led to many new electronic applications being envisioned, such as domain wall memistors. A crucial prerequisite to realising such devices is the control over the nucleation and position of the domain walls. This thesis presents different methods of exerting this control in mesoscale ferroelectric capacitors. Investigations 'Were carried out using thin single-crystal slices of the uniaxial ferroelectric KTiOP04 (KTP), embedded into a co-planar capacitor structure. Piezoresponse Force Microscopy (PFM) showed the domain nucleation positions in these capacitors were unpredictable. As ferroelectric domains switch polarisation with an applied electric field, inhomogeneous electric field distributions were created by design in KTP lamellae using Focused Ion Beam (FIB) milled holes and patterned electrode geometries. This inhomogeneity produced localised regions of enhanced field (hot-spots) and diminished field (cold-spots). Investigations showed that when a switching field was applied to these samples, nucleation initialised at the hot-spot locations, injecting domain wall pairs. This meant that the ferroelectric could be engineered to allow site-specific injection of domain walls. These results were extended by creating different magnitudes of hot-spots, allowing the sequential injection of domain walls, to provide multiple domain states required for a domain wall memristor. Further investigation showed that hot-spots Increased domain wall mobility whereas cold-spots could pin domain walls. This effect was utilised to create asymmetric domain wall mobilities and periodic pinning sites, prerequisites for domain wall ratchet devices. Finally, investigations using variations in the ferroelectric thickness were carried out. Switching occurred in smaller thicknesses first, with positive thickness gradients yielding decreased domain wall mobility. A ferroelectric with a ramped topography, then allowed the makings of the first ferroelectric domain wall diode.

Experimental and analytical investigations of a relatively new magnetic levitation device ere described. This device uses an electromagnet, which is the inductive port of a resonant circuit. Due to the effects of resonance the electromagnet current is strongly dependent on the position of the suspended object. If the circuit is properly tuned , the average magnetic force acting on the suspended object would be a restoring force similar to that exerted by a spring. This implies that static stability is attainable. The peculiar dynamic instabilities of the device have been reported by others and by the author. In order to understand these instabilities and to help with the design of auxiliary stabilization loops, a method is needed which enables quick analysis. In the systems described in this thesis, the frequency of the supply is much higher than that associated with the mechanical vibrations, and as a result, a direct digital simulation of the equations is inefficient. It is possible to overcome this difficulty by using an approximation of slow varying quantities. Due to the omission of the detailed high frequency oscillations from this new representation, it has been possible to obtain a solution of an example by using only 225 steps of integration. Using the original equations, about a hundred times this number of steps would be necessary. The results obtained by this method agree with those observed in an experimental model. They also agree with analogue simulations. It is known that static characteristics such as the variations of force and stiffness with position can be evalunted by using steady state impedances. It is, however, interesting to note that when the frequency of the supply is much higher than that of the mechanical" oscillations, then the equivalent electromechanical spring constant does not contribute solely to the understanding of static characteristics; it is also associated with the dynamic behaviour of the suspended object. A tuned circuit levitator is a member of a broader family of devices, i.e. "Parametric Electromechanical Systems". Comparison of a levitator to other members of this family enables one to obtain a physical understanding of its dynamic characteristics. Further physical insight is made possible by using a calculation similar to that usually employed when pure electronic parametric amplifiers are analyzed. The results obtained in this way also agree with previous results. Part of this work is concerned with the development of models , This is important for the investigation of technological problems. It is shown that dynamic stabilization of the device is possible by employing relatively simple electronic means. Consequently, tuned circuit levitators are probably more reliable than similar servo suspension systems. Matters related to the constructional geometry of these devices are also discussed.

This thesis examines the interaction between an RF SQUID ring coupled to two high Q resonant circuit measurement systems, operating at 24 MHz and 1 GHz at 4.2K. Energy is coupled to the system via a microwave source and is used to modulate the screening current of the ring and also to excite transitions between the underlying eigenenergy bands of the ring via non-adiabatic interaction. We study the current literature on such systems, used as qubits and as the building blocks for quantum computing and information processing elements. We discuss the relevance ofour system to those reviewed and suggest useful future direction for this work. We show that the point contact RF-SQUID coupled to multiple resonant circuits can be used as a model for studying the interaction between flux qubits and electromagnetic modes and provides insight in designing future practical qubit systems. In this work a new method of data acquisition is introduced, providing a much improved visualisation of the phase space. A 1 GHz measurement system is used to provide evidence of energy down-conversion from the input electromagnetic mode for a range of frequencies, via the SQUID ring. Previous work has focused on high ratio down conversion of up to 18,806: 1. Our data now shows a ratio of 1.67:1 and allows our fully quantum· mechanical theoretical model to be more usefully applied to give further insight about the system. As well as the 1 GHz measurement system, we also consider measurements from 24 MHz readout for a range of point contact weak links. We conclude the experimental results by showing down conversion taking place to both tank circuits simultaneously from the input mode, operating at 3.9 GHz. Finally we present simulations using the fully quantum model ofthe system, discuss the limitations and draw useful conclusions about the quantum nature of the down-conversion mechanism.

The spatial and spectral content of electromagnetic fields produced by charged sources in prescribed motion, moving in an infinitely long cylindrical beam pipe with perfect electrically conducting boundary conditions, is investigated in order to understand dynamic field effects. In the ultra-relativistic limit analytic field solutions to Maxwell's equations are determined in both the time and frequency domains, where the fields are explored for various non-stochastic and stochastic multiple pulse source configurations. Their frequency behaviour is shown to be bounded by envelopes and various features are directly related to the source configuration structure. The spectral energy density and spectral transverse field strength are found to have similar frequency response characteristics which differ from those of spectral power. The fields are found numerically for single pulse sources, which are not necessarily ultra-relativistic, for both constant velocity and hyperbolic accelerating sources. The equations are expressed in a dimensionless form to aid numerical efficiency. In the ultra-relativistic limit it is seen that the longitudinal fields are negligible, in contrast with the large slug of transverse field which is co-moving with the source and exists over its entire longitudinal spatial extent. Away from the ultra-relativistic limit all the fields spread out longitudinally beyond the spatial extent of the source and also the longitudinal field is no longer insignificant; all these fields exhibit complicated oscillatory dynamics. The model is based on theory derived from Maxwell's equations, expressed as a differential system on Minkowski spacetime using the language of differential forms. Along with boundary conditions, the mode structure of the linear system of field equations is developed, culminating in a telegraph-type equation with source, whose solution enables the system of equations for the transverse and longitudinal magnetic and electric fields in the beam pipe to be found for a source with a general charge profile.

Strained silicon (Si) and silicon-germanium (SiGe) devices have long been recognised for their enhanced mobility and higher on-state current compared with bulk-Si transistors. However, the performance and reliability of dielectrics on strained Si/strained SiGe is usually not same as for bulk-Si. Epitaxial growth of strained Si/SiGe can induce surface roughness. The typical scale of surface roughness is generally higher than bulk-Si and can exceed the device size. Surface roughness has previously been shown to impact the electrical properties of the gate dielectric. Conventional macroscopic characterisation techniques are not capable of studying localised electrical behaviour, and thus prevent an understanding of the influence of large scale surface roughness. However scanning probe microscopy (SPM) techniques are capable of simultaneously imaging material and electrical properties. This thesis focuses on understanding the relationship between substrate induced surface roughness and the electrical performance of the overlying dielectric in high mobility strained Si/SiGe devices. SPM techniques including conductive atomic force microscopy (C-AFM) and scanning capacitance microscopy (SCM) have been applied to tensile strained Si and compressively strained SiGe materials and devices, suitable for enhancing electron and hole mobility, respectively. Gate leakage current, interface trap density, breakdown behaviour and dielectric thickness uniformity have been studied at the nanoscale. Data obtained by SPM has been compared with macroscopic electrical data from the same devices and found to be in good agreement. For strained Si devices exhibiting the typical crosshatch morphology, the electrical performance and reliability of the dielectric is strongly influenced by the roughness. Troughs and slopes of the crosshatch morphology lead to degraded gate leakage and trapped charge at the interface compared with peaks on the crosshatch undulations. Tensile strained Si material which does not exhibit the crosshatch undulation exhibits improved uniformity in dielectric properties. Quantitative agreement has been found for leakage at a device-level and nanoscale, when accounting for the tip area. The techniques developed can be used to study individual defects or regions on dielectrics whether grown or deposited (including high-κ) and on different substrates including strained Si on insulator (SSOI), strained Ge on insulator (SGOI), strained Ge, silicon carbide (SiC) and graphene. Strained SiGe samples with Ge content varying from 0 to 65% have also been studied. The increase in leakage and trapped charge density with increasing Ge extracted from SPM data is in good agreement with theory and macroscopic data. The techniques appear to be very sensitive, with SCM analysis detecting other dielectric related defects on a 20% Ge sample and the effects of the 65% Ge later exceeding the critical thickness (increased defects and variability in characteristics). Further applications and work to advance the use of electrical SPM techniques are also discussed. These include anti-reflective coatings, synthetic chrysotile nanotubes and sensitivity studies.

This thesis describes work carried out in the Department of Applied Physics, Durham University, from October 1964 to September 1967. The work was a continuation of investigations, under the supervision of Dr. M.J. Mo rant, into the electrical properties of dielectric liquids. The work of Sletten in 1959 and that of Morant and Kahan in 1964 has shown that dissolved oxygen has a marked effect on the breakdown strength and the conductivity of a typical dielectric liquid, n-hexane. The present investigation shows that the reduced breakdown strength and increased conductivity observed on de-gassing hexane may not be explained in terms of an increase in carrier mobility. The mobility of photo-injected carriers in highly de-gassed hexane is found to be no greater than the mobility of the same carriers in air-saturated liquid. Further, no increase of mobility with applied field is observed in the de-gassed liquid for fields up to 140 KV/cm; nor is there any indication, under these conditions, of the existence of an additional 'fast' carrier. The variation of the injected current with both the applied field and the degree of oxidation of the cathode, however, indicates that the 'oxygen effect' referred to above is an electrode, surface phenomenon and not a property of the liquid itself. The current transients from which the mobility determinations were made are not exactly of the form predicted by simple theory. This and other anomalous observations indicate a strong interaction between the charge-carriers and the neutral liquid molecules. Such an interaction is consistent with self-trapping of the injected carriers by polarization of the surrounding medium. Suggestions for further work on the identification of the charge-carriers in dielectric liquids and on the factors affecting breakdown and conductivity are made in the text.

This work was aimed at finding suitable conditions for increased charge injection into liquid hexane. Experiments have been carried out on charge injection by both photoelectric and high-field processes. Photoelectric injection in very pure and highly degassed hexane depends mainly on the condition of the photocathode. Immediately after the preparation of the photocathode, the current increases continuously with applied voltage and it is shown to be mainly due to the emission of photo- electrons. Both cathode and liquid deteriorate over several days and a larger proportion of the current becomes due to bulk ionisation. Simultaneous d.c. and pulse measurements are used to prove that the initial photocurrent not only originates at the cathode but that its magnitude is also limited by the rate of emission of photoelectrons which subsequently form heavy negative carriers. The field dependence of the photocurrent has been explained in terms of the probability of heavy carrier formation by the electrons passing over an image barrier. The observed carrier mobility is in close agreement with that of other workers. Field-injection from razor blade electrodes resulted in high conduction currents at comparatively low applied voltages. The current is only slightly higher when the blade ^s negative. In pure hexane the current is found to be independent of bulk movement of the liquid so that it must be limited by electrode processes which are probably electron emission and electron extraction for negative and positive polarities respectively. In commercial and water or alcohol doped hexane, the current increases with liquid flow when the blade is negative but decreases when it is positive. These effects are explained in terms of field dissociation of the impurities and the formation of space charge at the high-field electrode. The present measurements suggest that the high injection currents reported elsewhere are mainly due to field ionisation of impurities rather than electron emission.