Computer simulations for designing green energy solutions

Reducing the CO2 emissions and mitigating climate change is the grand challenge facing humanity in the next couple of decades. Hydrogen has emerged as an economically viable alternative to fossil carbon based fuels in various industries, such as steel production, transportation and even biochemical food production. In order to effectively reduce the carbon emissions the hydrogen must be produced via electrolysis from water, as the prevailing standard method of steam reforming extracts hydrogen from fossil hydrocarbons, producing carbon emissions as byproduct. The fundamental theme in reducing carbon emissions is renewable energy production. This thesis studies how computer simulations can be used in designing and improving different solutions tapping into the vast amount of energy available in everyday environment. The paradigm behind this thesis work is that in order to optimize any design, we must first understand how the laws of physics influence the operation of the device to be optimized, while ultimately pursuing the ability to predict the effects of design changes. This thesis covers two separate cases of developing computer models explaining the operation of renewable energy devices. The first case is electricity production from streaming current in nanoscale porous structures, driven by constant evaporation of water. The porous structure presents a multiphase fluid dynamics problem, where the flow of water generating the electricity depends on various material properties and ambient conditions. The work in this thesis demonstrates how a carefully designed multiphysics simulation can not only reproduce experimental results known when building the model, but also produce emergent behaviour that can be experimentally verified. The second case is photoelectrochemical water splitting, converting sunlight to hydrogen with a single, monolithic semiconductor device without the need for electrical grid connections. This case is first simulated using a semi-classical theory based on the Boltzmann transport equation. However, as the design of new materials and electrode geometry increasingly utilizes nanoscale features, the classical approximations behind the Boltzmann equation are no longer accurate. Another method to simulate the semiconductor electrodes is the quantum mechanical formalism of nonequilibrium Green's functions (NEGF), which has been widely utilized in simulating solar cells and transistors. This work demonstrates that NEGF can be applied to modeling photoelectrochemical devices, offering significant improvement.The results cover how computational models can help not only understanding the behaviour of nanoscale solutions for renewable energy, but also provide critical ability to optimize device design to maximize the efficiency.