3D perovskites have enormous potential for optoelectronic applications such as light-emitting devices, photodetectors and lasers, due to tunable optical properties. Achieving precise control over their characteristics, specifically color purity, can be costly to discover because of their highly nonlinear behavior.  In this work, machine learning (ML) will be employed to explore the synthesis parameter space and target perovskite films with desired RGB values. By varying the annealing time and composition of the MAPbIBr₂ perovskite while fixing other synthesis parameters the film’s optical response can be adjusted. Using Bayesian Optimization, a data-driven approach will be established based on experimental feedback for precisely tuning the perovskite. This synthesis framework is designed for easy adaptation to other synthetic spaces requiring precise material control. This research aims to accelerate ML-driven design of perovskites while enhancing our understanding of their nonlinear synthesis space.

Laboratory automation has demonstrated great potential in accelerating the discovery and optimization of new materials. However, the lack of low cost high-throughput characterization has been a limiting factor in the development of autonomous self-driving labs. To address this, we developed an open-source 3D-printable robotic framework that can be integrated with an ocean optics spectrometer probe designed to measure materials properties in a high-throughput fashion. The device is low-cost, easy to construct and fully compatible with the Opentron (OT-2) automated liquid handler. The system operates on a printer-gantry system that moves the spectrometer probe across a laboratory plate as scanning progresses. We aim to achieve scanning speeds of 1 second per well, allowing a standard 48 well laboratory plate to be completed in under 1 minute – a significant improvement over current times achieved with human testing. Additionally, we outline potential applications for the system through the characterization of perovskite semiconductors for energy-efficient lighting and discuss the challenges of fully integrating this device into a completely autonomous workflow. Despite its current limitations, by facilitating high throughput characterization through affordable, open-source technologies, this device enables materials researchers in underserved regions to accelerate progress in key areas such as green technology development. 

Developing solutions to address social risk factors (SRF) in low- and middle- income countries (LMIC) can be difficult as many SRFs are a result of lacking financial support. SRFs are adverse living conditions that may impact the physical or mental well-being of an individual or community. Addressing SRFs in LMICs can increase implementation of evidence-based practices aimed at improving mental health outcomes. In this study, we focused on who is involved in proposed solutions that address SRFs and what specific roles those individuals undertake. Understanding what persons and roles are involved in a solution can help organize and facilitate action. We conducted a secondary inductive thematic analysis on qualitative data from a parent NIMH-funded study which aimed to develop strategies to address SRFs alongside a culturally adapted form of trauma-focused cognitive behavioral therapy (TF-CBT) for children who experienced parental death in Western Kenya. In the parent study, clinical supervisors conducted a workshop and training for TF-CBT lay-counselors to co-develop strategies to address SRFs. Most suggested strategies supported economic empowerment and a worksheet was designed to aid implementation of the strategies. This study uses data from worksheets filled out at 10 different schools in which lay-counselors designed economic empowerment strategies. The primary solutions included poultry rearing, vegetable gardening, and tree nurseries. Preliminary results show that most solutions tend to require school administration, teachers, children, for whom the solutions are for, and their guardians. Administrative roles tended to supply land required for solutions, teachers and guardians mostly supplied resources, while children were tasked with implementation of the solutions. Knowing who executes what roles can help inform what resources, skills, or knowledge a person can contribute to a solution, which may facilitate transferability between solutions. This can help researchers and communities individualize strategies to address SRFs where certain persons may be unavailable.