Search

The erosion and transport of fertile topsoil is a serious problem in the U.S., Australia, China and throughout Europe. It results in extensive environmental damage, reduces soil fertility and productivity, and causes significant environmental loss. It is as big a threat to the future sustainability of global populations as climate change, but receives far less attention. With both chemicals (fertilizers, pesticides, herbicides) and biological pathogens (bacteria, viruses) preferentially sorbing to silt and clay sized soil particles, estimating contaminant fluxes in eroded soil also requires predicting the transported soils particle size distribution. The Hairsine-Rose (HR) erosion model is considered in this thesis as it is one of the very few that is specifically designed to incorporate the effect of particle size distribution, and differentiates between non-cohesive previously eroded soil compared with cohesive un-eroded soil. This thesis develops a new extended erosion model that couples the HR approach with the one-dimensional St Venant equations, and an Exner bed evolution equation to allow for feedback effects from changes in the local bed slope on surface hydraulics and erosion rates to be included. The resulting system of 2I +3 (where I = number of particle size classes) nonlinear hyperbolic partial differential equations is then solved numerically using a Liska-Wendroff predictor corrector finite difference scheme. Approximate analytical solutions and series expansions are derived to overcome singularities in the numerical solutions arising from either boundary or initial conditions corresponding to a zero flow depth. Three separate practical applications of the extended HR model are then considered in this thesis, (i) flow through vegetative buffer strips, (ii) modelling discharge hysteresis loops and (iii) the growth of antidunes, transportational cyclic steps and travelling wave solutions. It is shown by comparison against published experimental flume data that predictions from the extended model are able to closely match measurements of deposited sediment distribution both upstream and within the vegetative buffer strip. The experiments were conducted with supercritical inflow to the flume which due to the increased drag from the vegetative strip, resulted in a hydraulic jump just upstream of the vegetation. As suspended sediment deposited at the jump, this resulted in the jump slowly migrating upstream. The numerical solutions were also able to predict the position and hydraulic jump and the flow depth throughout the flume, including within the vegetative strip, very well. In the second application, it is found that the extended HR model is the first one that can produce all known types of measured hysteresis loops in sediment discharge outlet data. Five main loop types occur (a) clockwise, (b) counter-clockwise, (c,d) figure 8 of both flow orientations and (e) single curve. It is clearly shown that complicated temporal rainfall patterns or bed geometry are not required to developed complicated hysteresis loops, but it is the spatial distribution of previously eroded sediment that remains for the start of a new erosion event, which primarily governs the form of the hysteresis loop. The role of the evolution of the sediment distribution in the deposited layer therefore controls loop shape and behavior. Erosion models that are based solely on suspended sediment are therefore unable to reproduce these hysteretic loops without a priori imposing a hysteretic relationship on the parameterisations of the erosion source terms. The rather surprising result that the loop shape is also dominated by the suspended concentration of the smallest particle size is shown and discussed. In the third application, a linear stability analysis shows that instabilities, antidunes, will grow and propagate upstream under supercritical flow conditions. Numerical simulations are carried out that confirm the stability analysis and show the development and movement of antidunes. For various initial parameter configurations a series of travelling antidunes, or transportational cyclic steps, separated by hydraulic jumps are shown to develop and evolve to a steady form and wave speed. Two different forms arise whereby (a) the deposited layer completely shields the underlying original cohesive soil so that the cohesive layer plays no role in the speed or shape of the wave profile or (b) the cohesive soil is exposed along the back of the wave such that both the non-cohesive and cohesive layers affect the wave profile. Under (a) the solutions are obtained up to an additive constant as the actual location of the boundary of the cohesive soil is not required, whereas for (b) this constant must be determined in order to find the location on the antidune from where the cohesive soil becomes accessible. For single size class soils the leading order travelling wave equations are fairly straightforward to obtain for both cases (a) and (b). However for multi-size class soils, this becomes much more demanding as up to 2I + 3 parameters must be found iteratively to define the solution as each size class has its own wave profile in suspension and in the antidune.

The influence of surface effects on a hysteresis loop for single domain, ferromagnetic nanoparticles was examined. Theoretical equations were derived to describe the magnetic behavior of the domains and a MATLAB program was used to solve them. M-H curves were calculated for the case when a magnetic field is applied in the favorable magnetization direction (easy axis). In contrast, the calculations show there were no hysteresis loops when the magnetic field was applied perpendicular to the easy axis. Our studies showed how parameters of the surface such as a associated with saturation magnetization near the surface of nanoparticles and Ks related to anisotropy had an impact on the hysteresis loop. The hysteresis loops were calculated for single-domain nanoparticles of MnBi, CoPt, and FePt. These materials have a wide range of values of radius R between the critical radius Rc10 for transition to the superparamagnetic phase and the critical radius Rc20 for transition to a multi-domain structure and also for high and low values of the volume anisotropy K0 which were used for analysis.The results showed that coercivity increased with increasing values of a, which is related to a decreasing interaction between magnetic moments, and thus saturation magnetization at the surface. They also showed that the coercivity increased with increasing values of Ks, which is related to the anisotropy. In contrast, the remanence decreased with increasing values of a and remained constant with increasing values of Ks. In addition, the coercivity and remanence increased with increasing values of the radius R of the single domain region. Furthermore, theoretical results showed that the area enclosed by each hysteresis loop have almost the same value of energy density for different values of a whereas, the area enclosed by each hysteresis loop has an increased value of energy density for increasing values of Ks.

Two dimensional (2D) materials have gained much attention these years for their intrinsic physical properties and promising applications in the field of electronics, optics and mechanics. Those materials, especially the 2D transition metal dichalcogenides (TMD), are regarded as a potential replacement plan for silicon in the next generation semiconductor technology. The combination of 2D semiconductors and epitaxial grown single crystalline ferroelectrics enables the development of next generation emerging ferroelectric memory devices, and supplies a polarization controlled system to explore the fancy physical phenomena in the 2D crystals under high carrier density doping at the same time. This thesis talks about the researches on two dimensional transistors with non-volatile functionalities originating from different mechanisms. Chapter 1 briefly introduced the research topics as well as motivations. Chapter 2 and 3 discussed details of two main components in the research: lead zirconate titanate (PZT) and molybdenum disulfide (MoS2) respectively. The MoS2 flakes were integrated with single crystalline PZT films directly and trap related hysteresis loops were observed in chapter 4. In chapter 5, methods to minimize the effects of interface states in the 2D ferroelectric structure were carried out. Correct handedness hysteresis loops were detected, with large loop size (>20V) and high ON/OFF ratio (>106), exhibiting the promising future of 2D ferroelectric devices. The switchable metal-insulator phase transitions of MoS2 controlled by ferroelectric polarization were also detected. In chapter 6, the clockwise hysteresis loops induced by interface states in the MoS2 2D dielectric transistor were investigated. Then a pulse gating technique with self-designed pulse sequences was developed, which effectively minimized the hysteresis loops induced by adsorbates and the intrinsic electrical properties of tested devices were eventually extracted.

In the present study, a power supply capable of converting a direct current into an alternating current was built. The frequency of oscillation of the output current could be varied from 174.8 kHz to 726.0 kHz by setting a set of capacitors in resonance. To this power supply is attached a 20-turns copper coil in the shape of a spiral. Because of the high heat generated in the coil, the latter has to be permanently water-cooled. A vacuum pump removes the air between the sample holder and the coil. A fiber optic temperature sensor with an accuracy of 0.001 K was used to measure the temperature of the nanoparticles.Four ferromagnetic nanoparticles (CoFe_2O_4, NiFe_2O_4, Ni_0.5Zn_0.5Fe_2O_4, Co_0.4Ni_0.4Zn_0.2Fe_2O_4) with different magnetic properties were subjected to heating. The heating performance is given by the specific power loss (SPL) which was calculated from the initial slope of the heating curve whereby there is minimum heat loss to the surroundings and the highest SPL was obtained when magnetic field and frequency were highest at 20.6 µT and 348.0 kHz, respectively. When the frequency was changed from 174.8 to 348.0 kHz, the SPL value was double indicating a one-to-one relationship between the frequency and the SPL. No such relationships were found when the magnetic field was increased from 13.8 µT to 20.6 µT. There is also a coercivity and remanent magnetization dependence on the SPL and both of them obey an inverse law. We found that the SPL for Ni_0.5Zn_0.5Fe_2O_4 (GPZ4) was 1.09 ± 0.02 W/g2. Moreover, the sample CoFe_2O_4 (GPZ11) with the highest power loss density (12250 J/m3) measured by the variable sample magnetometer was found to have the lowest SPL (0.135 ± 0.002 W/g2) in the lot. The magnetic field inside the coil is only a fraction the VSM, so when the low magnetic field was turned on, the hysteresis loop opened only to the extent of the minor loop.

Systems of magnetic nanowires are considered strong candidates in many technological applications as microwave filters, sensors or devices for data storage. Because of their strong potential as candidates in such applications they became lately the object of many studies. However, due to the very complicated nature of the interwire interactions, their magnetic behavior is very difficult to be interpreted. The main parameter controlling the response of magnetic nanowires assemblies is the aspect ratio of the nanowires that is defined as the ratio of the length to the wire's diameter. In our study we choose to modify the aspect ratio by keeping a constant length of nanowires and modifying the wire's diameter while keeping the same interwire distance. The samples were studied at room temperature, using vibrating sample magnetometer and X-band ferromagnetic resonance experiments. The results are explained taking into account the effects of the magnetostatic interactions and shape anisotropy.