Light conversion materials for solar cells by atomic layer deposition

Abstract

There are many different solar cell technologies which aim at producing electricity from sunlight cheap and/or efficient. As the efficiency of silicon cells is slowly but continuously climbing, price plummeting, and production skyrocketing, there is in my opinion little room for other technologies unless they can beat silicon on efficiency. Commercial cells at over 20 % efficiency are available and lab cells have been reported at above 27 %. Comparing this to the theoretical maximum efficiency of a single junction cell which is just above 30 % shows that we are able to produce close-to-perfect silicon cells. Any technology which wants to beat this has to aim at an efficiency higher than 30 %. There are not per today many potential candidates for this. One aspect that often seems to be overlooked is what exactly 20 % efficiency means for a solar cell. For regular off-the-shelf silicon cells, this efficiency is a combination of zero efficiency at > 1100 nm, more than 70 % in the 700 – 1000 nm range and steadily decreasing efficiency towards the UV. That means that if we could convert all the solar energy to 1000 nm light, the efficiency of solar cells would be drastically increased without changing the cell itself. Unfortunately we do not know how we would do this today, but we know a few steps on the way there. Down and up conversion aims at splitting one UV photon into two lower energy photons and merging two low energy photons into one medium energy one, respectively. This would in theory double the solar cells efficiency in the UV range and enable the cell to utilize the > 1100 nm light. This work is part of work package 4 New materials for next generation solar cells (WP4) in The Norwegian Research Centre for Solar Cell Technology (FME-Sol). The objective of this thesis is to build competence in the field of light conversion and to attempt at making an efficient down conversion film material by atomic layer deposition (ALD). This has resulted in four papers, two which are yet to be published. In addition, a significant part of the work has been devoted to popularization of science through lectures to non-scientific audiences. The potential down conversion materials that exists in the literature usually depends on the interaction between several different types of atoms and often with the host material itself as the UV absorbing material. As ALD grows the film one sub-monolayer at a time, it can give some quite unique control over the atomic distribution throughout the film. It is relatively easy to switch between several different cation cycles at will through the deposition which enables mixing of atoms that would separate or form precipitates under other conditions in addition to the ability to have some control of the nextneighbor distribution around each type of atom. Europium titanium oxides were chosen as the model system for this investigation. This system has the characteristic luminescence of Eu3+ and strong UV absorption of TiO2. Both binary oxides are relatively easy to synthesize by ALD. Thin films of both Eu3+ doped anatase and amorphous EuxTiyOz was deposited, while crystalline Eu2Ti2O7 was obtained through annealing. In addition to homogeneous mixing, sandwich structures of separated Eu2O3 and TiO2 layers were deposited. Thus, this system provides a good opportunity to investigate the relationship between the luminescence of the material and the concentration, local symmetry and interatomic arrangement and distances. The final stage of this thesis was to attempt to make a down conversion material by replacing Eu3+ with Yb3+/Ln3+. These lanthanide pairs have been reported in literature to split one high energy excited state into two lower energy excited states. In this work, energy transfer and luminescence was observed, but efficient down conversion was unfortunately not obtained. However, ALD was shown to enable some control of the arrangement of the cations which could lead to down conversion in other material systems which are not easily obtainable by other routes.