Kihyun Kim

Standoff laser absorption spectroscopy (LAS) has attracted considerable interest across many applications for environmental safety. Herein, we propose an anodic aluminum oxide (AAO) microcantilever LAS combined with machine learning (ML) for sensitive and selective standoff discrimination of explosive residues. A nanoporous AAO microcantilever with a thickness of <1 μm was fabricated using a micromachining process; its spring constant (18.95 mN/m) was approximately one-third of that of a typical Si microcantilever (53.41 mN/m) with the same dimensions. The standoff infrared (IR) spectra of pentaerythritol tetranitrate, cyclotrimethylene trinitramine, and trinitrotoluene were measured using our AAO microcantilever LAS over a wide range of wavelengths, and they closely matched the spectra obtained using standard Fourier transform infrared spectroscopy. The standoff IR spectra were fed into ML models, such as kernel extreme learning machines (KELMs), support vector machines (SVMs), random forest (RF), and backpropagation neural networks (BPNNs). Among these four ML models, the kernel-based ML models (KELM and SVM) were found to be efficient learning models able to satisfy both a high prediction accuracy (KELM: 94.4%, SVM: 95.8%) and short hyperparameter optimization time (KELM: 5.9 s, SVM: 7.6 s). Thus, the AAO microcantilever LAS with kernel-based learners could emerge as an efficient sensing method for safety monitoring.

Although the high incidence and recurrence rates of urothelial cancer of the bladder (UCB) are heavy burdens, a noninvasive tool for effectively detecting UCB as an alternative to voided urine cytology, which has low sensitivity, is yet to be reported. Herein, we propose an intelligent discrimination method between normal (SV-HUC-1) and cancerous (TCCSUP) urothelial cells by using a combination of micro-dimensional electrochemical impedance spectroscopy (µEIS) with machine learning (ML) for a noninvasive and high-accuracy UCB diagnostic tool. We developed a unique valved flow cytometry, equipped with a pneumatic valve to increase sensitivity without cell clogging. Since contact between a cell and electrodes is tight with a high volume fraction, the electric field can be effectively confined to the cell. This enables the proposed sensor to highly discriminate different cell types at frequencies of 10, 50, 100, 500 kHz, and 1 MHz. A total of 236 impedance spectra were applied to six ML models, and systematic comparisons of the ML models were carried out. The hyperparameters were estimated by conducting a grid search or Bayesian optimization. Among the ML models, random forest strongly discriminated between SV-HUC-1 and TCCSUP, with an accuracy of 91.7%, sensitivity of 92.9%, precision of 92.9%, specificity of 90%, and F1-score of 93.8%.