Halide Vapor Phase Epitaxy of Ge from an Elemental Source

Halide vapor phase epitaxy (HVPE) shows promise for low cost III-V device manufacturing because of its high growth rates compared to other growth methods and the use of lower cost elemental precursors [1]. In HVPE, the elemental sources react <i>in situ</i> with anhydrous HCl to form volatile metal chloride growth species. Germanium (Ge) is one useful semiconductor material that is typically grown using GeCl₄ or GeH₄ but has never been grown using an elemental Ge source by flowing anhydrous HCl over Ge metal to form a volatile chloride, as is done for Ga and In in HVPE. This would be a useful advancement for HVPE growth toward the development of, for example, an epitaxial bottom junction of an inverted triple junction solar cell, which helps enable substrate reuse, or as a sacrificial layer for epitaxial lift off or controlled spalling. Also, recycled Ge substrates can be used as source material, conserving this critical material. Therefore, our goal is to develop and understand the growth of Ge from an elemental Ge source by HVPE. Thermodynamic calculations suggest the driving force for GeCl₂ generation from HCl and Ge is high in an inert N₂ambient, but the driving force for growth is low. In contrast, little GeCl₂ forms <i>in situ</i> from HCl and solid Ge in an H₂ carrier, although the driving force for growth from an externally injected GeCl_x source is high, presenting a challenge to control the ambient in the source and deposition regions. Here, we show that it is possible to deposit Ge by generating GeCl₂ from solid Ge and HCl in a N₂ ambient and then supplying a source of active hydrogen locally at the substrate to drive growth.<br/>We grew Ge at 350 Torr and 800 °C, using elemental Ge also held at 800 °C, in a N₂ ambient. The source region where the GeCl₂ forms is contained within a quartz boat that is physically separated from the deposition region. To achieve Ge growth, we inject hydride gases, AsH₃ and PH₃, as sources of active hydrogen, to the growth surface. We do not observe Ge growth unless a supply of active hydrogen is added to the system, consistent with thermodynamic calculations. For a given HCl(Ge) flow rate, the Ge growth rate initially increases with increasing hydride flow, but then begins to decrease. The maximum growth rate reached increases with increasing HCl(Ge), and the peak growth rate also shifts to higher hydride flow rates with higher HCl(Ge), which likely indicates a competition between growth and etching processes because AsH₃ is known to etch Ge surfaces [2]. Ge growth from a chloride likely features a kinetic barrier analogous to that observed for GaAs growth, namely the saturation of the surface with Cl atoms. We speculate that the addition of active hydrogen to the surface, in this case in the form of pyrolyzed hydrides, reduces the Cl from the surface and allows for continued Ge growth. One drawback of using group V hydrides to drive the Ge reaction is that the group Vs are dopants in Ge. We observed As or P concentrations in the Ge films, using secondary ion mass spectrometry, in the range of 4x10¹⁷– 1x10¹⁸ atoms/cm³, concentrations that can drastically influence device characteristics. However, we note that there are numerous other “helper molecule” options that can provide active hydrogen without doping or etching the material. This work provides a path forward for the deposition of high-quality Ge for optoelectronic devices from an elemental source.<br/><br/>[1] J. Simon, <i>et.al</i>. <b>III-V-Based Optoelectronics with Low-Cost Dynamic Hydride Vapor Phase Epitaxy</b>. <i>Crystals</i> 9, 3, (2019).<br/><br/>[2] W.E. McMahon and J.M. Olson. <b>An STM Survey of As/Ge(mnn) and P/Ge (mnn) Surfaces</b>. NCPV Program Review Meeting, 121-122 (2000).