Abstract: The accurate, sensitive, and selective on-site detection of volatile aldehyde biomarkers associated with lung cancer holds significant importance for early cancer diagnosis and effective intervention. However, employing surface-enhanced Raman scattering (SERS) for gas sensing poses challenges due to substrate specificity issues in enriching gaseous molecules, as well as substrate instability and reproducibility concerns in environments with numerous interferences. Additionally, the modest Raman cross sections of most gaseous molecules make them susceptible to interference from other components in exhaled breath, complicating analysis. To address these challenges, we developed an innovative SERS sensing substrate. This substrate was created by in-situ growth of a two-dimensional covalent organic framework (COF) on three-dimensional reduced graphene oxide (rGO) as a surface template, followed by modification with Au nanoparticles (Au NPs). This design not only enhances the substrate's specificity for capturing target analytes but also facilitates the diffusion of gaseous molecules, as both ends of the membrane remain open for gas molecule. Utilizing benzaldehyde as a representative gas marker for lung cancer, our sensor leverages the Schiff base reaction with a Raman-active probe molecule, 4-aminothiophene (4-ATP), pre-immobilized on Au NPs for trace-level and multicomponent detection. Furthermore, the integration of machine learning algorithms with Raman spectroscopy eliminates subjective analysis of gaseous aldehyde species based on single feature peaks, enabling more precise identification. Overall, this membrane-based sensor offers a promising approach for developing a tabletop SERS analytical system tailored for point-of-care testing (POCT) in lung cancer diagnosis.

lung cancer; volatile organic compounds; SERS