Perovskite-silicon tandem solar cells are now entering into commercial production, but present several manufacturing challenges. Some of these challenges arise from the limited thermal processing budget of the metal-halide perovskite sub-cells, where the halide perovskite itself and organic charge transport layers readily undergo degradation when processed at >100 C for extended periods of time. In this talk, I discuss the possibilities enabled through the use of spatial atomic layer deposition (SALD) for depositing oxides used as buffer layers, encapsulation, and transparent conducting electrodes in perovskite-silicon tandems. 

SALD allows the growth of oxide thin films with similar conformality, uniformity, density and electronic properties as conventional ALD, but with an order of magnitude shorter processing time [1]. This is achieved through the spatial separation of the organometallic and oxidant precursors in a gas manifold, where the substrate (e.g., perovskite device) is oscillated repeated through the gas channels, and the number of oscillations controls the thickness of the film. Such a growth method does not require vacuum, and can readily be scaled-up to grow over large area (e.g., see nanoprintinnovations.com). By shortening the processing time, we demonstrate how a wider range of processing temperatures can be used to grow oxide buffer layers onto perovskite devices, even thermally-sensitive compositions. These oxides include n-type oxides for p-i-n structure solar cells, as well as p-type Cu₂O films for n-i-p structured perovskite solar cells, and we show that these oxides successfully protect the device from mechanical damage during sputter deposition of the transparent top electrode. 

Another critical layer in tandems is the recombination contact required to monolithically connect sub-cells together electrically. We demonstrate the use of SALD to grow H:In₂O₃ thin films, where we achieve 7.2 Ohm/sq sheet resistance. Commercial sputter-deposited indium tin oxide has 13 Ohm/sq sheet resistance, and a transmittance that diminishes toward the near-infrared (NIR). By contrast, H:In₂O₃ achieves low resistivity through high mobility rather than high carrier concentration, and we show that the NIR transmittance is maintained at 89%, important for minimising parasitic optical losses to the low-gap sub-cell. We use nano-SIMS, TEM-EDX, XPS and ToF-ERDA measurements to understand the doping mechanism in H:In₂O₃.

Finally, I will discuss our emerging efforts at using nanolaminate oxide layers grown by SALD for encapsulation. With a stack of 5 layers of AlO_x and ZnO grown at 100 C, a water vapour transmission rate of 10^-5 g m^-2 day^-1. Using this stack on top of a perovskite solar cell that is encapsulated with glass improves the damp-heat stability from T₈₀ after 1200 h (without nanolaminate stack) to T₁₀₀ after 1200 h (with nanolaminate stack).

Overall, this talk demonstrates the utility of SALD for perovskite-based tandem fabrication to improve efficiency and stability.