New Publication: Beyond Traditional airPLS – Smarter Baseline Removal for SERS
- Yiping Zhao
- Aug 12
- 2 min read
We’re excited to share our latest article in Analytical Chemistry, titled:👉 “Beyond Traditional airPLS: Improved Baseline Removal in SERS with Parameter-Focused Optimization and Prediction” (DOI link)
This work addresses a persistent challenge in Raman and SERS data processing—accurate baseline correction. The widely used airPLS method often struggles with complex spectra, leading to inconsistent results. We present an optimized version, OP-airPLS, along with a machine learning–based parameter predictor that makes baseline correction both more precise and faster.
Why It Matters
Reliable baseline removal is crucial for extracting accurate peak information, especially in fields like chemical sensing, biomedical diagnostics, and materials characterization. Our method dramatically improves correction accuracy while reducing the need for manual tuning.
Main Contributions
OP-airPLS uses adaptive grid search to fine-tune key parameters for optimal performance.
A PCA–Random Forest predictor rapidly estimates the best parameters from spectral shape, enabling near-instantaneous correction.
Tested on 6,000 simulated spectra covering 12 peak–baseline combinations, OP-airPLS achieved ~96% improvement in baseline accuracy, with errors as low as 5.55 × 10⁻⁴.
The ML predictor processed each spectrum in just 0.038 seconds while maintaining high accuracy (~90%).
What Sets This Work Apart
This is the first systematic approach to link spectral shape with optimal baseline parameters. OP-airPLS delivers near-perfect correction in challenging conditions, and the ML predictor brings this capability into real-time workflows.
Looking Ahead
Future work will focus on validating the method on real-world spectra, integrating it into on-instrument software for live correction, and extending it to other spectroscopic techniques like FTIR and fluorescence.
Citation:
Cui, J.; Chen, X.; Zhao, Y. Beyond Traditional airPLS: Improved Baseline Removal in SERS with Parameter-Focused Optimization and Prediction. Anal. Chem. 2025, 97(30), 16211–16218.
Comments