We demonstrate a solar cell on reformed porous Ge with an efficiency of 7.7%. We generate mesopores in (100) Ge by bipolar electrochemical etching and anneal them at high temperature. The pores coalesce deep in the structure rather than at the surface as desired, although resulting in coarse superficial morphology unsuitable for device growth. To combat this issue, we developed a surface treatment involving an HBr dip, annealing at 415°C, and a postannealing ultrasonic De-ionized water dip to improve the surface structure, resulting in a smoother reformed surface on which we grow a GaInAs solar cell. The structure retains embedded pores after growth and the transitions between Ge and III–V layers are distinct. The solar cell fabricated using the improved coalescence has an efficiency of 4.5%. The efficiency improves to 7.7% by isolating the rest of the device from three limiting localized shunt areas. Protruding defects in the porous Ge and III–V layers still limit the performance, but this work establishes a step toward the technical viability of this exfoliation approach, showing decent efficiency if protruding defects can be removed or reduced.

Mesoporous Ge was generated by bipolar electrochemical etching of (100) Ge. Through high-temperature annealing experiments, we found that the mesoporous Ge microstructure produces pore coalescence deep in the structure rather than at the surface. III-V epitaxy on annealed porous Ge with subsurface coalescence and coral-like structures on the surface produced shunted photovoltaic devices. The roughness of the annealed porous Ge film was reduced by a pre-annealing HBr treatment and a post-annealing ultrasonic DI water dip. A smoother reformed surface was obtained, on which a GaInAs upright device with an efficiency of 5.7% was grown, demonstrating a first step towards technical viability.

We spalled 2-inch Ge wafers using a Ni stressor layer and grew GaAs solar cells by hydride vapor phase epitaxy on the spoiled Ge surface without other surface treatments. The controlled spoiling procedure leaves behind arrest lines, which are parallel surface striations, perpendicular to the crack front, caused by the crack propagation repeatedly halting and restarting during the fracture. We show that small arrest lines do not significantly affect device performance, but larger lines act as regions for carrier recombination. We demonstrate a 12.8% efficient single-junction device, without anti-reflection coating, grown on a spoiled surface containing arrest lines. The quantum efficiency of this device is similar to devices grown on non-spalled GaAs and Ge substrates.

Decreasing the cost of single-crystal substrates has long been sought for III-V solar cells. Controlled spalling is a possible pathway for epitaxial liftoff, which would help reduce costs, but chemo-mechanical polishing after liftoff tends to limit the impact. Growth on an unpolished spalled surface would be an additional step toward lower costs, but it is crucial to show high efficiency solar cell devices on these unprocessed substrates. In this study, we spalled 2 inch Ge wafers using a Ni stressor layer, and GaAs solar cells are grown by HVPE on the spalled Ge surface without any other surface treatment. We show a 12.8% efficient device, without anti-reflection coating, with equivalent quantum efficiency as identical devices grown by HVPE on non-spalled GaAs substrates. Demonstrating a high carrier collection on unpolished spalled wafers is the first step toward reducing substrate-related liftoff and polishing costs.

Authors: Matthew Musselman, Nooraldeen Alkurd, Brett Ley, Xena Murugappan, Jacob Feldman, Corinne E. Packard Date of creation: June 2017 Publisher: Colorado School of Mines How to Cite: M. Musselman, N. Alkurd, B. Ley, X. Murugappan, J. Feldman, C.E. Packard, “Ceramic Materials in Renewable Energy: A Module Developed for Hands-On Learning.” 2017. PDF file. Funding Information: This work was supported by NSF CAREER grant DMR 1352499. Description: This document is a lesson plan for a short module describing the use of ceramic materials in renewable energy applications for STEM outreach. Colorado State science standards addressed by the module are included. The module is readily adapted for students in grades 1-12. The module was originally drafted for presentation to children aged 7-13 at the Rocky Mountain Camp for Dyslexic Children, thus activities are largely hands-on, driven by physical concepts, and do not rely heavily on reading or writing skills.