J. Mater. Sci. Technol. ›› 2026, Vol. 241: 107-113.DOI: 10.1016/j.jmst.2025.03.073

• Research Article • Previous Articles     Next Articles

Principal component analysis of normalized SERS spectra for trace-level analyte quantification

Xin Xie1, Yuanhao Zheng1, Fengtong Zhao, Weipeng Wang*, Wangyang Fu, Yunhan Ling, Zhengjun Zhang*   

  1. Key Laboratory of Advanced Materials (MOE), School of Materials Science and Engineering, Tsinghua University, Beijing 100084, China
  • Received:2024-11-25 Revised:2025-02-24 Accepted:2025-03-08 Published:2026-01-10 Online:2025-05-12
  • Contact: *E-mail addresses: wpwang@tsinghua.edu.cn (W. Wang), zjzhang@tsinghua.edu.cn (Z. Zhang)
  • About author:1These authors contributed equally to this work.

Abstract: Development of approaches for reliable quantification of trace-level analytes by surface enhanced Raman scattering (SERS) is of great significance due to its potential applications in a diversity of fields. We developed an approach by principal component analysis (PCA) of the normalized SERS spectrum and the relative scattering ability (RSA) between molecules to quantify trace-level analytes. It was found that both the normalized SERS spectra and the RSA between molecules on a SERS substrate can be experimentally measured. Due to their material-determined nature, once measured they can serve as a database for quantification by SERS on substrates of the same material. PCA can greatly reduce the dimension of spectral data and list the few non-zero data in the spectra in sequence first. Thus, it makes SERS quantification simple and automated, with less error. This approach enables SERS to be an applicable technique in trace-level identifications in a diversity of fields.

Key words: Normalized SERS spectrum, Principal component analysis, Scattering ability, Quantitative analysis by SERS