Ahmed Waseem, Osborne Eleanor L, Veluthandath Aneesh Vincent, Senthil Murugan Ganapathy
Optoelectronics Research Centre, University of Southampton, Southampton SO17 1BJ, U.K.
Anal Chem. 2024 Nov 12;96(45):18052-18060. doi: 10.1021/acs.analchem.4c03594. Epub 2024 Oct 31.
Infrared (IR) spectroscopy is a powerful analytical technique used to identify and quantify different components within a sample. However, spectral interference from fluctuating concentrations of water vapor and CO in the measurement chamber can significantly impede the extraction of quantitative information. These temporal fluctuations cause absorption variations that interfere with the sample's spectrum, making accurate analysis challenging. While several techniques to overcome this problem exist in the literature, many are time-consuming or ineffective. We present a simple method utilizing just two sample spectra taken sequentially. The difference of these spectra, multiplied by a scaling factor, determined by minimization of the point-to-point spectral length, provides a correction spectrum. Subtracting this from the spectrum to be corrected results in a fully corrected spectrum. We demonstrate the effectiveness of this method via the improved ability to determine analyte concentration from corrected spectra over uncorrected spectra using a partial least square regression (PLSR) model. This technique therefore offers rapid, effective, and automated spectral correction, which is ideal for a nonexpert user in a clinical or industrial setting.
红外（IR）光谱法是一种强大的分析技术，用于识别和量化样品中的不同成分。然而，测量室中水蒸气和一氧化碳浓度波动产生的光谱干扰会严重阻碍定量信息的提取。这些随时间的波动会导致吸收变化，干扰样品的光谱，使得准确分析具有挑战性。虽然文献中存在几种克服此问题的技术，但许多技术既耗时又无效。我们提出了一种仅利用依次采集的两个样品光谱的简单方法。这些光谱的差异乘以一个由逐点光谱长度最小化确定的缩放因子，可得到一个校正光谱。从待校正光谱中减去此校正光谱，可得到完全校正的光谱。我们通过使用偏最小二乘回归（PLSR）模型，与未校正光谱相比，从校正光谱中确定分析物浓度的能力得到改善，证明了该方法的有效性。因此，该技术提供了快速、有效且自动化的光谱校正，这对于临床或工业环境中的非专业用户而言是理想的。
