NSF Convergence Accelerator Track L: Accelerating VOC Sensor Advances and Translation by Machine Learning and Bioinspiration

North Carolina State University, Raleigh NC

Investigators

Abstract

Impressive olfactory sensing systems are present in nature-born biological subjects. For instance, jewel beetles can detect a burning tree 50 miles away, and dogs can sniff out substances at concentrations of one part per trillion – orders of magnitudes more sensitive than human noses. Olfaction-based chemical sensing represents one of the most promising detection technologies that has many outstanding analytical attributes. It is noninvasive, high throughput, fast, easy for multiplexing, and relatively low cost. The past few decades have witnessed a growing amount of gas detectors (e.g., electronic nose), However, the engineered gas sensors do not match natural olfactory systems in terms of key performance attributes: sensitivity and specificity. Leveraging the investigators’ previous experience in developing volatile organic compound (VOC) sensors, this project team, consisting of engineers, chemists, material scientists, biologists, data scientists, and partners from industry and healthcare, aims to innovate and further mature two miniature VOC sensing technologies, namely, colorimetric VOC sensor arrays and wearable VOC sensor patches to the level of scaled manufacturing. These cost-effective, field-portable, and sensitive sensors may prove particularly valuable for disadvantaged and resource-limited communities to address critical challenges associated with global health and food security by improving their capability in personal health monitoring, crop protection, and environmental detection. The miniature sensor tools also provide excellent opportunities for public outreach and training the next-generation workforce. This convergence project seeks to break down the translational science barriers for olfactory sensors and accelerate the development and translation of such sensor technology into real products for addressing urgent needs in noninvasive diagnostics of human and plant diseases and environmental monitoring. The overarching goal of the project is to build a convergence framework for developing a set of affordable and accessible VOC sensors with significantly improved analytical performance by applying machine learning and bio-inspired design. Specifically, the project plan includes the following research tasks: 1) develop a machine learning prediction model for colorimetric VOC sensing dye screening using the Weaver Dye Library with 98,000 dyes and its scalable manufacturing; 2) design and optimize highly sensitive wearable VOC sensors by studying the insect-inspired wax coating as a “chemical lens” for active “focusing” of VOCs onto sensors, and 3) sensor scaling up and demonstration of exemplar applications for human, plant, and environmental detection. The convergence approach of this project relies on the merging of conventional sensor research (chemistry, materials, and electronics) with two other distinct disciplinary areas: data intelligence and sensory entomology. The project results will establish a scientific foundation in olfactory sensor design and partnership between academic research groups and industry manufacturers for sensor scaling up. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.