A novel technique for degenerate p-type doping of germanium

• We present a novel technique for low-cost, degenerate doping of Ge.
• We present carrier concentrations ≫5 × 1019 cm−3, as verified using SIMS.
• A simple analytical model is derived for the dopant profile after “drive-in”.
• Enthalpies of activation and preexponential factors are extracted.
• We find these values in agreement with published work.

Spin-on dopants offer an expedient, straight forward, and low cost method for doping semiconductors. These sol–gel dopant films contain sufficient impurity content to obtain surface concentrations well above the solubility limit of most semiconductors (e.g., Si, Ge). While much has been published about spin-on dopants for degenerate doping of Si, there is to the authors’ knowledge no report in the literature of such films being used to achieve degenerate p-type doping of Ge, as there are a number of technical challenges associated with this technique. In this work, we present a novel technique for degenerate p-type doping of Ge using Ga spin-on dopant films in a regular horizontal tube furnace. This technique gives both excellent uniformity, higher doping concentrations and better potential for ultra-shallow junctions than diffusion from solid sources, is much preferred to rapid melt/alloyed junctions (especially for optoelectronic applications), and is readily applicable to rapid thermal processing. In this preliminary investigation, we report doping concentrations exceeding 1020cm-3 which are shown to be fully electrically activated. We report on the use of a traditional “pre-dep”/“drive-in” approach, which could be used to give shallow dopant profiles, or tailored dopant concentrations, for a wide variety of electronic and optoelectronic applications, and to form back surface fields in photovoltaic and thermo-photovoltaic devices. Values of 3.4 ± 0.2 eV and 3.2 ± 0.4 eV are extracted for the activation enthalpies of the diffusion processes into p-type 〈1 0 0〉 and n-type 〈2 1 1〉 Ge substrates, respectively, which are shown to be in good agreement with previously published data [1] and [2]. We conclude that Ga is a more suitable dopant for reliable, low-cost, degenerate p-type doping of Ge, rather than B (also investigated in this work), which is known to be problematic.