High Voltage Multijunction Photovoltaic Devices

Abstract

A next step in the energy transition is to produce green fuels from sunlight. One way to achieve this is via high voltage photovoltaic (PV) devices which can directly drive electrolysis reactions to produce hydrogen or other chemical fuels. Group IV based multijunction (MJ) PV devices are a candidate to function as these high voltage devices.
This work is focused on improving the efficiency of crystalline silicon/nanocrystalline silicon/amorphous silicon (c-Si/nc-Si/a-Si) triple junction (3J) solar cells through better current matching. On top of that, the low bandgap material germanium tin (GeSn:H) is characterized as it shows potential for use as bottom junction in multijunction photovoltaic devices.

Several methods were tried to improve current matching in the 3J PV devices. Varying the nc-Si middle absorber thickness was successful at improving the shortcircuit current density (J 𝑠𝑐 ) in the middle junction, which was the current limiting junction. A range between 35μm ncSi absorber thickness is effective at keeping the middle junction J𝑠𝑐 high while keeping opencircuit voltage (V𝑜𝑐 ) and fill factor (FF) losses at an acceptable level. Varying the n-type nanocrystalline silicon oxide (n-nc-SiO𝑥) intermediate reflective layer (IRL) thickness was also effective at improving the J𝑠𝑐 of the middle junction. Introducing a silver IRL between the n-layer of the middle junction and the player of the bottom junction was ineffective, as was introducing various transparent conductive oxides (TCOs). The J𝑠𝑐 of the middle junction did not increase with the introduction of these layers. In the case of the TCOs, the shunt resistance decreased drastically resulting in decreased V𝑜𝑐 s and FFs. The best performing device in this thesis was manufactured using a 400nm a-Si:H absorber, a 4.5μm i-nc-Si:H absorber and a 60nm n-nc-SiO𝑥 intermediate reflective layer. This device had a V𝑜𝑐 of 1.947V, FF of 0.789, a J𝑠𝑐 of 9.51mA/cm2 and an efficiency of 14.6%, which, to the best of our knowledge, is the highest conversion efficiency achieved to date with this device architecture.

Introducing tetramethyltin (TMT) to process GeSn:H films resulted in carbonization of the films. Carbon passivates Ge dangling bonds better than hydrogen. This reduces the defect density, leading to higher activation energy (E𝑎𝑐𝑡) with high TMT injection rates. A high germane flow rate resulted in dense, relatively defectfree films with a high photoconductivity, low bandgap and near intrinsic semiconductor properties. A high germane flow of around 2 sccm is therefore recommended. Varying the pressure in the deposition chamber offers some control over postprocessing oxidation, where higher pressure results in less oxidation. Changing the radiofrequency power offers a tradeoff between film growth rate and film quality.