Deep-learning framework for fully-automated recognition of TiO2 polymorphs based on Raman spectroscopy.

Emerging machine learning techniques can be applied to Raman spectroscopy measurements for the identification of minerals. In this project, we describe a deep learning-based solution for automatic identification of complex polymorph structures from their Raman signatures. We propose a new framework using Convolutional Neural Networks and Long Short-Term Memory networks for compound identification. We train and evaluate our model using the publicly-available RRUFF spectral database. For model validation purposes, we synthesized and identified different TiO2 polymorphs to evaluate the performance and accuracy of the proposed framework. TiO2 is a ubiquitous material playing a crucial role in many industrial applications. Its unique properties are currently used advantageously in several research and industrial fields including energy storage, surface modifications, optical elements, electrical insulation to microelectronic devices such as logic gates and memristors. The results show that our model correctly identifies pure Anatase and Rutile with a high degree of confidence. Moreover, it can also identify defect-rich Anatase and modified Rutile based on their modified Raman Spectra. The model can also correctly identify the key component, Anatase, from the P25 Degussa TiO2. Based on the initial results, we firmly believe that implementing this model for automatically detecting complex polymorph structures will significantly increase the throughput, while dramatically reducing costs.

Tailored Interfaces of the Bulk Silicon Nanowire/TiO2 Heterojunction Promoting Enhanced Photovoltaic Performances.

We report significantly improved silicon nanowire/TiO2 n+-n heterojunction solar cells prepared by sol-gel synthesis of TiO2 thin film atop vertically aligned silicon nanowire arrays obtained by facile metal-assisted wet electroless chemical etching of a bulk highly doped n-type silicon wafer. As we show here, chemical treatment of the nanowire arrays prior to depositing the sol-gel precursor has dramatic consequences on the device performance. While hydrofluoric treatment to remove the native oxide already improves significantly the device performances, hydrobromic (HBr) treatment consistently yields by far the best device performances with power conversion efficiencies ranging between 4.2 and 6.2% with fill factors up to 60% under AM 1.5G illumination. In addition to yield high-quality and easy to produce solar cell devices, these findings regarding the surface treatment of silicon nanowires with HBr suggest that HBr could contribute to the enhancement of the device performance not only for solar cells but also for other optoelectronics devices based on semiconductor nanostructures.