Nitrogen-based fertilizer is critical to feed the growing population of humanity. However, current production of nitrogen-based fertilizer entails enormous emissions of greenhouse gasses, threatening the climate. Moreover, this nitrogen-based fertilizer represents the addition of large flows of reactive nitrogen to the planet, mostly as volatile ammonia that evaporates into the air. This addition of reactive nitrogen perturbs the global nitrogen cycle. To address these issues, the way in which nitrogen-based fertilizer is manufactured must change to reduce green-house gas emissions and nitrogen pollution. Plasmas, particularly non-equilibrium plasmas, hold great promise for nitrogen fixation via nitrogen oxyacids that can acidify and enrich biowaste. Nitrogen oxyacids prevent ammonia evaporation by transforming it into involatile ammonium while also enriching the nitrogen content of biowaste. As non-equilibrium plasma nitrogen fixation requires only energy, air, and water, this can be done renewably at the farm level. This dissertation investigates upcycling of biowaste via renewable nitrogen fixation from air using atmospheric pressure non-equilibrium plasma. In particular, the chemistry of the plasma-liquid interface, including in the presence of biowaste, and the plasma discharge behavior in the presence and absence of interfaces are investigated. To this end, this dissertation makes three distinct contributions.
The first contribution is to investigate the non-equilibrium plasma-manure interface chemistry. Non-equilibrium plasma in air generates many reactive oxygen and nitrogen species that are potent oxidizers, such as nitric oxide and hydroxyl radical. In water, these plasma-generated species are known to ultimately produce nitric acid and nitrous acid, which are themselves oxidizers. In biowaste such as dairy manure, it was not known if plasma treatment would result in nitric acid and nitrous acid, or if these reactive oxygen and nitrogen species would react with manure compounds. This research established that treating manure with non-equilibrium plasma in atmospheric air will enrich manure nitrogen content by generating nitrogen oxyacids. An increase in total nitrogen is also observed, which matches the supplied increase in nitrogen oxyacids with a yield of approximately 100%. It is seen that plasma treatment of manure will acidify manure and thereby depress ammonia volatilization. These findings lay the foundational knowledge for renewable upcycling of biowaste using air plasmas, towards green fertilizer manufacture.
The second contribution of this dissertation focuses on the non-equilibrium plasma-water interface chemistry. The physicochemical phenomena are complex, and efforts to explore the chemistry experimentally by varying processing conditions are faced by the multivariable and nonlinear nature of the dependence of plasma treatment outcomes on the process inputs. The interest is in the chemistry that enables a large quantity of fixed nitrogen and minimizes thermal effects of the liquid. Multi-objective Bayesian optimization is used to efficiently find the Pareto front that provides a systematic tradeoff between the multiple (possibly conflicting) plasma treatment objectives. It is demonstrated that there is a trade-off between nitrogen fixation and interface heating, whereas nitrite, nitrate, and hydrogen peroxide all co-maximize.
The third contribution of this dissertation focuses on the connection between the discharge behavior and the non-equilibrium plasma chemistry. A learning-based method is used to discover interpretable similarity parameters that relate system variables to dimensionless target variables, such as the dimensionless electron temperature and dimensionless electron number density. Using the Buckingham-Pi Theorem, the system variables are combined to form dimensionless numbers, which can be scale-free similarity parameters. This dimensionless discovery method allows for learning parsimonious dimensionless representations that can accurately predict the dimensionless target and can extrapolate outside of the range of measured process variables. It is shown how the discovered dimensionless relations provide physical insights into the intrinsic system behavior that may not be readily accessible from conservation laws.
