Rapid Detection of SARS-CoV-2 in Clinical and Environmental Samples via a Resonant Cavity SERS Platform within 20 min

The coronavirus disease 2019 (COVID-19) epidemic has given a warning that it is important to explore the rapid detection of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in clinical specimens or environmental samples for public health strategies and future variants. The surface-enhanced Raman spectroscopy (SERS) technique was demonstrated to achieve this goal. However, the consistency of signals originating from the poor compatibility of virions with SERS hotspots remains a key scientific challenge for the practical applications of SERS. Herein, we develop a SERS platform for the ultrasensitive and rapid detection of SARS-CoV-2 antigen within 20 min by the combination of a highly consistent SERS substrate and a supervised deep learning algorithm. A V-shaped resonant cavity array (VRC) substrate was fabricated to trap SARS-CoV-2 virions in the periodic V cavity array and stimulate the integral SERS signal of the virus via a resonance coupling effect. Benefiting from the unique architecture of the VRC substrate, we were able to directly detect the SARS-CoV-2 virus with high sensitivity and high consistency. These excellent performances enabled us to identify five different kinds of SARS-CoV-2 variants and detect SARS-CoV-2 from clinical and environmental samples with high accuracies.