Tonio Buonassisi, Department of Mechanical Engineering, MIT
Predictive Defect Engineering for Scalable Photovoltaics at $1/Wp
At $1 per peak watt installed cost, solar photovoltaics is projected to be cost-competitive with traditional fossil fuels in many markets across the United States. To meet this cost target and ensure scalable production, thin low-cost materials must be used. Herein lies an important trade-off: Low-cost materials are typically defect-rich, and defects impede electronic transport and photoconversion efficiency. Since efficiency and cost are inversely related, defect-rich materials have until recently resulted in poor-quality, economically uncompetitive solar cells.
In this presentation, we explore a path towards low-cost, high-performance, and scalable photovoltaic absorbers. We introduce the concept of “defect engineering,” the science of controlling defects to engineer desired material properties. We review recent successful applications of defect-engineering technologies to ingot multicrystalline silicon grown from “upgraded metallurgical silicon,” achieving cell efficiencies above 16%. Accurate identification of performance-limiting defects requires multiscale characterization, evaluating cm-size devices and probing down to the nanometer scale for defect recognition. We will review recent advances in macroscopic CCD-based PV device characterization tools, and elucidate how these can be coupled to synchrotron-based nanoprobe techniques to characterize chemical natures and distributions of performance-limiting defects less than 20 nm in diameter. Once the natures and underlying physical behavior of these defects are known, an opportunity exists to engineer these defects – aided by predictive modeling – to enhance solar cell performance.
We then consider candidate PV materials with cost-reduction and scaling potential to support $1/Wp installed costs, yet which are currently defect-limited: Hyperdoped silicon, with a modified silicon band structure and potential to exceed the Shockley-Queisser efficiency limit; tin-sulfide (SnS), a potential Earth-abundant replacement for CIGS; and cuprous oxide (Cu2O), a potential absorber material for high-efficiency multijunction devices.
