No description available.

No description available.

Developing resilient sensor systems for deployment in extreme environments is a challenge which silicon carbide, along with other wide band gap materials, stands to play a major role in. However, any system developed will be hindered in its usefulness unless the problem of providing a power supply in these extreme conditions is addressed. This work addresses this need; a wireless sensor node conceived of standard o the shelf components was first developed and used as the basis for the design considerations required for a silicon carbide sensor node. The silicon system developed uses a piezoelectric energy harvester for the power supply and exhibits favourable operating characteristics for low vibration environments. It is capable of continuous operation at 120 mg (1.177 ms⁻²) and at 40 mg operates with a system duty cycle of 0.05. PZT, a standard piezoelectric energy harvesting material, was characterised to 300°C to test its resilience to the conditions found in hostile environments. The material degrades considerably with temperature, with a decrease in Youngs modulus from 66 GPa at room temperature to 8.16 GPa at 300 C. The room temperature value is repeatable once cooled with an observed hysteresis in the upper temperature range. The peak output voltage at resonance also varies with temperature, resulting in an 11.6% decrease in room temperature voltage once the device is heated to 300°C. The output voltage at 300°C is found to be 2.05 V, a considerable decrease from the initial 11.1 V output at room temperature. The decrease in voltage with temperature is not monotonic as maybe expected, the data showing that at 473 K there is an increase in output voltage which is caused by a decrease in mechanical damping. SiC pin diodes were fabricated with wide drift regions to promote a large depletion width, in order to maximise the capture cross section of incident light on the devices. The large drift region produces a high series resistance. However, ll factors above 0.7 show that the device is not signi cantly a ected. SiC is shown to be an e ective UV harvester with an observed increase in output power from 0.17 mWcm⁻² at room temperature to 0.32 mWcm⁻² at 600 K. Fill factor also remains stable with temperature, indicating that the device is not a ected by variation in parameters such as shunt and series resistances or the ideality factor. There are current technological di culties which preclude the manufacture of large area silicon carbide solar cells and as such, an alternative networking solution is presented as a way to increase the output power of the devices. Given that these devices would be subject to long term high temperature exposure, a 700 hour thermal stress test is carried out at 450°C to explore the failure mechanism of the devices. There is an observed decrease in device ll factor which indicates that the device su ers increasing degradation. The data shows that this is caused by increasing series resistance, which reduces the devices ability to output power. SEM imaging and SIMS analysis show this is likely caused by signifcant metal diusion in the contact stack which could potentially be overcome by the addition ofan alternative di usion barrier. Once energy is generated by an energy harvester is must be stored so that it can be used when required. To this end both substrate and on chip storage technologies are discussed in the forms of AlN and HfO₂ metal insulator metal (MIM) capacitors. To test the feasibility of both solutions, AlN and HfO₂ MIM capacitors were characterised to 300°C. The HfO₂ device leakage has a strong temperature dependence as observed in the IV characteristics and the capacitance density does not scale according to parallel plate theory. However, the devices can be e ectively networked and their leakage reduced with series connection. The internal voltage decay of the device is reduced with series connection, due to the di er-ing work functions of the metal-insulator contacts. The alternative AlN solution exhibits substantially weaker temperature dependance and signi cantly improved lm quality. The data shows no existence of a barrier at the insulator - metal interface, as observed in the HfO2 device IV characteristics. The extracted activation energy is stable with temperature at 1.26 +/- 0.15 eV indicating a trap assisted leakage mechanism. This method is more suitable to fabrication of large area storage as it can be fabricated o chip on a less expensive substrate and the devices fabricated exhibit a higher yield than the HfO₂ devices.

An attractive application for the solid oxide cell (SOC) technology is to enable reversible energy conversion. SOCs can efficiently convert chemical energy into electricity. Equally, SOCs can utilise electrical energy to upgrade low energy fuels into higher energy mixtures. A key step in the development of the reversible SOC is mitigating carbon deposition to allow stable long term operation in the presence of hydrocarbons. The work presented in this thesis describes comprehensive analysis of the performance of current state-of-the-art materials followed by the development of a copper-based electrode with both technologically relevant performance and no carbon formation during electrolysis of CO2 in the presence of methane. For current state-of-the-art nickel materials, positive bias was found to suppress carbon deposits in pure CO, while negative bias facilitated cell degradation through increased carbon formation and subsequent electrode delamination. In situ and ex situ Raman analysis of the conventional cermet Ni/Gd0.1Ce0.9O2-δ (Ni/CGO) electrodes revealed differences in amount, location and type of carbon formed during CO - CO2 electrolysis. The rate of carbon deposition was also greatly increased in the presence of H2. The positive effect of a CGO interlayer on reducing carbon formation for nickel was demonstrated. Building upon the result of the conventional electrodes' testing, a new batch of electrodes was prepared through a low temperature metal infiltration technique. Low temperature processing allowed Cu/CGO electrodes to be fabricated, which are difficult to produce with conventional techniques because of the low melting point of copper. It was demonstrated that the electrochemical performance of Cu/CGO electrodes is equivalent to Ni/CGO electrodes, whilst carbon formation is fully suppressed. This observation suggests that the electrocatalytic activity of these electrodes is dependent on CGO rather than metal (Cu or Ni), and moreover that the electrodes are ideally suited to biogas upgrade applications without deleterious carbon deposition.

Practical vibration energy harvesting tends to be employed where the frequency of the vibration does not vary with time. This is a result of energy harvesting devices resonating at a specific frequency. If the frequency of the vibration changes even slightly from the harvester's natural resonant frequency, then the power generated drops off significantly. Various methods have been proposed to apply kinetic energy harvesting over wider frequency bands, for example by active mechanical tuning of the structure to the vibration, or by adjusting the electrical load. This thesis explores electronic methods of optimally controlling the power electronics that extracts power from the harvester and charges a storage capacitor. The focus is on power levels below lmW, where the power budget for control circuitry is limited to lOs of microwatts or lower. Low-power techniques are investigated that maximise power by optimal resistive loading of the harvester, over a given excitation frequency bandwidth. This includes an experimental investigation of switching-converter control strategies. A low-power power sensing method is presented, which adapts to the available power level to maintain high sensitivity over a wide power range. A discrete component implementation of the method consumes 7.5 /!W and is demonstrated to be effective over a power range of 390 /!W to 750 /!W. A frequency detection method is presented which shows lower power consumption compared to the typically used method. To address the problem of variable frequency excitation, a feed-forward maximum power point tracking control strategy is presented, which demonstrates fast response to changes in the excitation frequency and significant improvement of the performance over the standard maximum power point tracking control. The method relies on an analytical model of the system and the measurement of the excitation frequency. These low-power techniques are implemented in hardware, as a complete self-contained and self-starting energy harvesting system. Depending on the amount of stored energy available, the system moves automatically between passive and active conversion modes, and a mode with optimised feed -forward maximum power point tracking. A minimum operating power of 26 /! W for the active conversion mode has been demonstrated.

The penetration of unscheduleable generation will increase due to legislation and eventually saving on fuel cost. This will cause an increase in uncertainty of power-flow and drive up balancing market costs, the safety margin for N-1 will have to increase. i.e. N-1 will not accurately represent the state of the system. A security assessment scheme (SAS) that considers probabilistic uncertainty could give financial savings and/or better security of supply. In other words a power system with a high penetration of renewables is likely to require a new type of security assessment scheme. Before that is done we must be able to compare and evaluate existing and proposed schemes. This thesis has two goals. Firstly, to be able to compare two security assessment schemes to determine which is better for the current system. The work details a computer program that combines a two stage Monte Carlo Sampler and a power system simulator to generate a level of security. The number of simulations that fail to converge within limits in N-1 and N-2 was compared to the calculated level of security and found to not be a good predictor. The second goal is to see how the level of security changes as the uncertainties of renewable generation get added into a given power system. In doing this, the effect of adding renewables can be quantified. The work found that if 15% of the generation power comes from generators the are unscheduleable or stochastic the security of supply does not greatly change. Whereas if the penetration is increased to 30% the security level become significantly worse in almost all tested scenarios.

As the developed world seeks to reduce the dependence upon fossil fuels, amine boranes have received increased interest as a clean, stable hydrogen source. This thesis sought to scrutinise the mechanism of amine borane dehydrocoupling proposed by Hill and provide quantitative evidence for the dependence of dehydrocoupling activity upon identity of the metal centre. This work led to a refined mechanism, incorporating proton-assisted β-hydride elimination steps, occurring via potentially concerted processes. A pathway to account for the previously unreported reactivity of [R2N-BH2]2 to ultimately afford [HB(NR2)2] is also included. A dependence upon the identity of the metal centre on the dehydrocoupling activity was found, identifying a decrease in reactivity with increasing ionic radius and decreasing cation charge density as each group is descended. This thesis showed that the mechanism of secondary amine borane dehydrocoupling by d0 reagents is complex, suggesting that a generic mechanism which fully describes dehydrocoupling of all amine boranes would be too much of an oversimplification.

Climate change, pollution, unprecedented population growth, geopolitical tensions and rapid technological development are intrinsically connected to the nature, level and availability of global energy, which shapes present and future aspects of human society. Particularly, in a society where global energetic demand is continuously rising and the awareness of the negative impact of fossil fuels on the environment is becoming widespread, the exploitation of renewable sources for the generation of sustainable energy is highly needed. In this regard one key requirement for an effective deployment and expansion of renewable energy in the global energy market is represented by its ability to conveniently convert and store the energy derived from intermittent sources, in order to guarantee a constant supply to the electric grid. The technologies for the energy conversion and storage present various degrees of maturity, each one having specific advantages and disadvantages depending on the type of application and energetic source. This thesis aims to give a tiny contribution to the complex problem of energy conversion and storage, through the design, characterisation and testing of electrocatalytic materials for water electrolysis, photoelectrochemical water splitting and direct methanol fuel cell. It is expected that the first two processes will play an important role in the future as convenient technologies for the conversion of solar and wind power into chemical energy in the form of hydrogen. The third process is regarded as promising way to convert the renewable chemical energy in the form of methanol into electrical energy. At the core of the research lies the design and development of electrocatalysts, which are directly responsible for lowering the reaction overpotentials and ultimately increasing the overall efficiency of the processes. As such, in this thesis three materials were synthesised using straightforward methodologies and evaluated as electrocatalysts for the alkaline hydrogen evolution, the photoelectrochemical oxygen evolution and the alkaline methanol oxidation. Their performances were directly linked to the morphological and structural properties which in turn significantly affected the nature of active sites. For the first work reported in Chapter 3, a material based on a mixed cobalt nickel sulphide nanoparticles supported onto Ni foam showed high activity toward the hydrogen evolution reaction, with a required small overpotentials of 163 mV at a current density of 10 mA/cm2 in 1.0 M KOH electrolyte. This value compares well with the best existing hydrogen evolution reaction electrocatalysts based on non-noble elements. Moreover the catalyst showed good durability which was tested under chronoamperometric conditions, maintaining an optimal performance for 72 hours. The origin of such high activity was attributed to the existence of an optimal nickel-cobalt sulphide ratio at the surface of the electrode, which was obtained by selecting the appropriate temperature and time of thermal annealing of the material. This optimal presence of the biphasic nickel-cobalt sulphide nanoparticles led to high electrochemically active surface area and small charge transfer resistance, as evidenced by the extensive characterisation analysis carried out on these materials. For the second work reported in Chapter 4, a WO3/Co3O4 photoanode was successfully synthesised via a facile sol-gel method and tested for the photoelectrochemical oxygen evolution. It was found that the degree of crystallinity of the cocatalyst influenced heavily the photoelectrochemical activity towards the oxygen evolution. In particular, a poorly crystalline structure of Co3O4 led to an improvement of up to 40% in photocurrent generation compared to the bare WO3. Interestingly, the highly crystalline Co3O4 significantly suppressed the photocurrent generation, as a result of the creation of an unfavourable band alignment, with a dramatic increase in the charge recombination at the interface. Finally, for the third and last work reported in Chapter 5, ultra-small Pt nanoparticles embedded on a 3D structure composed of CeO2, NiO and Ni foam was synthesised and tested as electrocatalyst for the alkaline methanol oxidation reaction. The generated catalyst showed extremely high activity for the alkaline methanol oxidation, with mass and geometric current density values of 1160 mA/mgPt and 202 mA/cm2, whose values are among the highest ever reported for Pt-based materials. It was demonstrated that the unique morphological architecture and existence of a synergistic effect between Pt and adjacent CeO2 nanoparticles contributed decisively to the observed high performance. Particularly the presence of defective and poorly crystalline CeO2 nanoparticles was beneficial to the efficient oxidative removal of the CO from the Pt active sites which resulted in a higher durability of the electrocatalyst. Moreover, the concomitant presence of the superficial Ni(OH)2 was thought to contribute to the supply of OH species to the Pt, which act as reactants for the CO removal. The most active electrocatalyst was subjected to stability test, retaining 40 % of the initial geometric current density after 6 hours, and quite surprisingly the activity could be totally restored through straightforward CV scans in 1.0 M NaOH electrolyte.

Pervasive networks of wi reless sensors and communication nodes have been developed during past decades and most of them are powered by fixed-energy sources, e.g., wiring power and batteries. However, these traditional energy sources are impractical for powering wireless devices due to their inherent limitations, e.g., the high setup cost of wiring power and the finite life span of batteries. In view of these facts, more attentions have been drawn on vibration energy sources existing in ambient environment where sensors operate. Dozens of different types of vibrat ion energy scavenging devices have been developed, which are mainly consisted of mechanical systems coupled with transduction mechanisms. The linear mechanical system has been used in most of existing vibration generators. The main drawback of such system is that it has a rather narrow bandwidth, meaning that the device can only effectively harvest vibration energy when the resonance frequency of the system coincides with the excitation frequency. Various methods have been proposed recently 10 overcome the drawback including utili sing mechan ical systems with non-linear mechanisms in order to increase the bandwidth of vibration energy scavenging devices. The dissertation is intended 10 fi nd a practically effective non-linear mechanical system with desirab le dynamic behaviours for vibration energy scavenging devices. To do so, we first examined three non-linear mechanisms numerically to find the most desirable one based on the corresponding typical mechanical systems. The Numerical Simulation (NS) method built in Matlab has been used to explore the static and dynamic characteristics of these systems with hardening and softening mechanisms

Energy Harvesting is a key enabling technology for highly distributed electronic systems such as wireless sensor networks, One of the most commonly described harvesting techniques is vibration harvesting, where a base-excited resonant mass/spring system is damped via an electromechanical transducer. A significant drawback of the resonant mechanical system, required to amplify low-level base vibrations, is the narrow bandwidth over which the system can operate, This thesis describes work carried out towards a method of increasing the bandwidth of resonant vibration harvesters by synthesising a variable complex load impedance using highly efficient switched-mode converters: loading the harvester with an optimum complex impedance effectively tunes the complete electromechanical system to the excitation frequency. This tuning effect is described analytically and demonstrated in practice with linear impedance emulating circuits. To benchmark the electrically tuned system against a more conventional converter, a micro power resistance emulator was developed, featuring an average quiescent power consumption of 56.6uW and a peak efficiency of 85.4%. The prototype harvester for which this tuning system was developed generates approximately 20mW, presenting a challenge for • converter design. This is further exacerbated when synthesizing large reactive loads at tuning frequencies far from mechanical resonance where the apparent power is much larger than the real power delivered from the harvester. To achieve the desired performance a custom micropower VSC power stage was implemented using discrete components, having an average quiescent power consumption of 454uW. The complete switched-mode tuning system requires current and voltage sensing, control compensation and PWM generation. A range of analogue and miJ5'ed signal implementations of these subsystems were investigated and it was concluded that at this power level the most significant challenge arises from balancing gate drive losses against the requirement for ultra low conduction losses in the power devices. Experimental results demonstrate that electrical tuning can extend the bandwidth over which a resonant vibration harvester can provide useful DC power, to almost three times that achieved with a simple unity power factor converter synthesising a fixed load resistance.