Hyperdoping of semiconductors can be used to shift the optical band gap towards lower energies and thereby extend the optical response of a material like Si further into the infrared. Therefore, the approach is attractive for opto-electronic applications, e.g. photodiodes or intermediate band solar cells. However, the high impurity concentration results in an elevated recombination rate in the hyperdoped layer which poses a severe challenge when aiming for a high carrier collection efficiency. Here, we realize sulfur-hyperdoping by irradiating Si with ultrashort laser pulses in an SF6 atmosphere. This involves rapid melting/resolidification cycles and high stresses in the substrate, which gives rise to additional recombination channels in the form of laser-induced defects.
The goals of our study are i) to distinguish between the respective impact of the two types of recombination on the overall effective minority carrier lifetime, ii) identify the major limitations and iii) find a path to optimize the material system for its designated application.
We therefore prepare sulfur-hyperdoped samples that have been laser-processed in SF6 as well as in atmosphere with otherwise identical parameters (pulse density and laser fluence). The latter shall serve as reference for laser-induced defects, whereas the hyperdoped samples come with both recombination channels. The laser process conditions are optimized for a high absorptance of 90%abs in the sub-bandgap spectral region between 1200 and 2500 µm. The microstructured surface morphology of the wafer causes the absorptance to approach 100 %abs in the spectral range between 300 and 1200 nm (“black Si”). We then etch the samples for different durations in a diluted isotropic Si etch bath at room temperature to successively remove the defect-rich surface layers “top-down”. This gives us a set of samples with different concentrations of sulfur and laser-induced defects. After the etch step, the samples are cleaned and receive an AlOx passivation layer (by ALD) on both surfaces.
We characterize the set of samples by measuring the effective carrier lifetime, the absorptance, Raman crystallinity, and, for specific samples, the sulfur concentration depth profile by dynamic SIMS. In addition, we examine the surface-morphology by SEM. As the measured effective carrier lifetime refers to the whole sample, i.e. the high-lifetime substrate and the defect-rich and hence low-lifetime surface layer, its interpretation is not straight-forward. We therefore conduct one-dimensional numerical modelling with the software PC1D to interpret the evolution of the effective lifetime and identify its practical upper limits.