Reliable solid-state nanopore sensing requires stable, low-noise ionic current baselines. Existing stability evaluation methods often rely on subjective visual inspection. Here, we propose a robust quantitative framework that assesses nanopore wettedness and stability indicators through power spectral density (PSD) analysis of the noise floor in the measured ionic current and uses the resulting features to predict wettedness. A systematic comparison of multiple PSD fitting models and weighting strategies identified a five-component, five-parameter (5C5P) model coupled with high-frequency, low-PSD (HFLS) weighting as the most accurate and consistent approach for characterizing noise in the measured ionic current. Utilizing the noise coefficients from this optimized noise PSD fit, we applied logistic regression to predict nanopore wettedness. Among these coefficients, the 1/f noise coefficient and the white-noise coefficient, together with a compact measure of overall low-frequency noise and the applied voltage, formed a compact and interpretable feature set for distinguishing wetted and unwetted pores. A logistic regression classifier trained on these features achieved high performance across a mixed-voltage dataset (median F1-score = 98%) and generalized effectively to unseen applied voltages and different pore dimensions. Segment-wise evaluation on one-second windows that mimic real-time operation demonstrated that the classifier reliably distinguished wetted and unwetted pores across all applied voltages. This physics-informed, data-driven framework enables automated wettedness prediction and stability assessment, indicating a pathway toward reliable real-time quality control for multiplexed, high-throughput solid-state nanopore sensing platforms.
Upretee et al. (Tue,) studied this question.