- Research
- SERS
- Nanoparticles
| Surface-Enhanced Raman Scattering （SERS） |
1. Chemical-assisted SERS
Molecular sensing is key for diverse applications including medical diagnostics, environmental monitoring, and food analysis. To achieve enhanced SERS detection performances, we leverage a range of chemical interactions, such as covalent bonding, hydrogen bonding and electrostatic interactions, between target analyte molecules and various receptor molecules grafted on our nanostructures. This approach offers two key benefits: (1) by bringing the analytes close to the plasmonic surface, we can increase the effective concentrations and generate higher signal responses, and (2) by generating differential interaction profiles, we can trigger analyte-specific spectral fingerprint and accurately identify analytes. In addition, to amplify spectral differences for highly similar analytes or complex, multicomponent mixtures, we have developed ‘SERS super-profiles’ which are horizontal combinations of the SERS fingerprint of multiple receptors. Our results demonstrate the effectiveness of this approach across a wide range of analytes, including toxic, polluting gases such as NO2 and SO2, flavor molecules such as menthol and limonene, as well as more complex targets such as microorganisms and breath metabolites in COVID-19 breath. Such platforms exhibit huge potential as next generation molecular nanosensors.
Relevant publications ACS Nano 2020, 14, 2, 2542–2552 ACS Appl. Mater. Interfaces, 2020, 12, 33421–33427 Nano Lett. 2021, 21, 6, 2642–2649 Angew. Chem. Int. Ed. 2022, 61, e202207447 Chem. Sci., 2022,13, 11009-11029 2. Machine learning-driven SERS
In-depth analysis of the inherently complex SERS spectral data often holds the key to elucidating the wealth of chemical information that is embedded within. Our aim is to realize the integration of machine learning algorithms for SERS based analytics in four different ways – (1) As dimensionality reduction techniques to condense complex spectral variations into principal changes, (2) As evaluators to determine which input spectral features are the most important for classification or regression, (3) As predictive modelling methods to construct accurate classification or regression models and (4) As a robust approach to guide downstream feature extraction and engineering in an ensemble analytical framework involving multiple processes or models. Our results illustrate the promising potential to translate SERS-based nanosensors for practical use in biomedical, environmental and food industries.
Relevant publications ACS Nano (2022) 16, 9, 13279-13293 Angew. Chem. Int. Ed. (2022) 134, 33, e202207447 Nanoscale Horiz. (2022) 7, 6, 626-633 ACS Nano (2021) 15, 1, 1817-1825 Nano Lett. 2021, 21, 6, 2642–2649 ACS Appl. Mater. Interfaces, 2020, 12, 33421–33427 ACS Nano 2020, 14, 2, 2542–2552 |
