Thin-film microcells: a new generation of photovoltaic devices

One way to increase the efficiency of a solar cell is to increase the power density of the light incident upon it. Indeed, if light is concentrated on a solar cell (e.g., via a lens), its voltage, and thus its efficiency, increases logarithmically. There are, however, limits to this efficiency increase; for example, the temperature of a device under intense light can become high enough to decrease the efficiency and eventually destroy it. Additionally, the output power for large current densities is limited by series resistance.

Detrimental effects such as these are manageable on commercial concentrator cells based on high-quality crystalline materials, but it has not been possible to use high light concentrations of >50 suns (where one sun is 1kW/m²) efficiently on thin-film solar cells. Indeed, most such cells are grown on glass substrates, which are poor thermal conductors, while thin-film semiconductive layers—especially the top, window, layers—are not sufficiently conductive for high-current-density operation.

Overcoming these limitations is a particularly attractive prospect because thin-film solar cells can be rapidly deposited on large areas, through the self-assembly of microstructures, at a cost lower than for the current concentrator cells (which are based on epitaxial growth of single crystals). However, thin film solar cells such as copper indium gallium diselenide (Cu(In,Ga)Se₂), cadmium telluride (CdTe), and gallium arsenide (GaAs) are based on elements that are not abundant on earth (rare earths). Due to their scarcity, a reduction in the required quantity of these materials could lead to cheaper cells. We have designed a new solar cell architecture that fulfills both the requirements.

To make high-concentration possible on thin films, we aimed to miniaturize the cells for efficiency and affordability. A device used under concentration requires a solar panel to be only fractionally covered by solar cells—see Figure 1—which reduces the amount of material required. Additionally, the reduced volume-to-surface area ratio and shorter current flow path in microscale solar cells leads to diminished heat elevation and reduced resistive losses, respectively.

Figure 1. In this schematic of a photovoltaic device, light passing through a microlens array is concentrated and focused onto miniaturized solar cells.

Figure 2. A second-generation copper-indium-gallium-diselenide microcell. The absorber layer is removed from the microcells by chemical etching, leaving the localized absorber layer covering only required areas.

For proof of concept, we fabricated thin-film microcells based on the Cu(In,Ga)Se₂technology. These microstructures were obtained by photolithography. We worked on two generations of prototype. The first was obtained by creating a microstructure on top of the absorber layer, with the latter kept intact over the whole substrate. The second generation was obtained by a chemical etch of the entire solar cell stack, except in defined micro-areas (i.e., the microcells): see Figure 2. Because of this top-down process, the absorber layer is present only where it is required (a 'localized absorber').

To measure the efficiency increase and to gain further insight into the behavior of Cu(In,Ga)Se₂ devices under extremely intense light fluxes, we tested the cells under both laser light and concentrated sunlight. We found that the efficiency of a first-generation 50μm wide Cu(In,Ga)Se₂ microcell increased significantly, from an unconcentrated (1 sun) efficiency of 16%, to a maximum efficiency of 21.3% at a concentration of 475 suns.¹ It should also be noted that the efficiency of the microcells under standard solar illumination (AM1.5G) is the same as their macro counterparts. This is particularly of interest because state-of-the-art crystalline microcells in Si or GaAs^2,3 show a decreased efficiency compared to their large-area solar cell counterparts.

In the second generation of devices, performances were stable even down to areas of 5x10⁻³mm². For devices made with other materials, such as GaAs, the solar cell performance generally decreases when areas are reduced to below 1mm².⁴ The impact of the edge becomes important in miniaturized devices due to the increase of the perimeter-to-surface area ratio, making good passivation crucial for small devices. The edge surfaces are very well passivated in Cu(In,Ga)Se₂ devices, which lowers the recombination at the edges and enables efficiency in extremely small devices.

In summary, we have fabricated solar microstructures using thin-film microcells and proven that concentrated applications are possible. We have also shown that Cu(In,Ga)Se₂ cells can be miniaturized to tens of microns wide and still be very efficient. In future devices, we hope to use bottom-up approaches to directly deposit localized absorbers and reduce material usage. Our next step is to develop up-scalable fabrication processes and concentrating optics that could take advantage of the miniaturization of the cells by, for example, being more compact.