University College London
The global energy sector has now sustainability as one of its main objectives and solar technology is amongst the most viable options for a reliable and clean energy production. Reducing the cost of solar modules is a crucial element to increase the potential of such a solution. Solar cells are based on Transparent and Conductive Oxide (TCO) films, which can be produced via a process called Aerosol-Assisted Chemical Vapour Deposition (AACVD). Concisely, this process is based on sequential stages. Firstly, aerosol is generated from a solution containing TCO precursors. Then, the aerosol is transported to a heated chamber, where the solvent evaporates and the precursors deposit, forming the desired film. Not only can TCOs be produced by this process, but also composites, powders, coatings, nanotubes, etc.
Although there is some research available in the literature reporting small scale experimental results of the AACVD technique, there is little about computational modelling. The latter would nevertheless be crucial for the design and scale-up of the AACVD. Consequently, we present an integrated model composed of aerosol generation, transport and delivery. For the aerosol generation, we predict the range of droplet sizes obtained via ultrasonic atomisation of the precursor solution. This is done using probability distributions, which become one of the inputs for the transport model. We can then predict the aerosol loss as a function of the properties of the droplets, the flow and the piping system. The model output shows the fraction of aerosol that reaches the heated chamber as well as its size range. It has been tested for a variety of scenarios, using different combinations of horizontal, inclined and vertical pipes. Finally, the solvent evaporation in the heated chamber is modelled, which depends mainly on the temperature profile of the site and the flow conditions. The precursors are now free to deposit and form the desired products. Uncertainty and sensitivity analyses are performed throughout the process.
The stochastic aspects of aerosol generation, transport and delivery were incorporated using probability distributions. Experimental results were used to validate the model predictions for droplet sizes for the generated aerosol and the loss during transport. Through the formulation of an optimisation problem, the model presented has been used to suggest possible designs for the transport system, minimising the aerosol loss. This approach is suitable for transport distances on the scale of industrial processes. Other applications based on particle atomisation and transport, such as fuel combustion and spray drying or cooling, can also use the models presented here. Finally, the knowledge acquired while handling uncertainties in the modelling of the AACVD process has been used to plan a modelling framework based on Julia that will enable us to represent uncertainties and manipulate variables with uncertain values.
Pedro received his BEng in Chemical Engineering with first-class honours in 2015 from the Federal University of Minas Gerais (UFMG). He was part of the Scientific Initiation Program in Mathematics at the same university (2011-2014). He was a fellow of the Science without Borders program at McMaster University, Canada (2012-2013). In 2015 Pedro started his PhD studies at University College London, while maintaining hobbies like football, volunteering, activities related to the work of J. R. R. Tolkien and C. S. Lewis, etc.
