Mattu M J, Small G W
Department of Chemistry, Clippinger Laboratories, Ohio University, Athens 45701-2979, USA.
Anal Chem. 1995 Jul 1;67(13):2269-78. doi: 10.1021/ac00109a055.
The feasibility of performing quantitative analysis with short segments of bandpass-filtered Fourier transform infrared (FT-IR) interferograms is demonstrated. The protocol developed in this work addresses four limitations that hinder the use of FT-IR spectroscopy in nonlaboratory applications: (1) the need for a rugged, low-cost, and reliable spectrometer, (2) the lack of representative background spectra for use in acquiring absorbance spectra of the target analyte, (3) the presence of overlapping spectral bands that interfere with the analyte determination, and (4) the difficulty in obtaining useful information from data collected near the limit of detection. In this work, spectral information pertaining to a specific analyte band of interest is isolated directly from a short interferogram segment by the application of narrow-bandpass digital filters. When processed in this way, the filtered interferogram segments contain compound-specific information that can be utilized for quantitative analysis. Successful use of a univariate calibration procedure with filtered interferogram data of benzene and nitrobenzene of varying concentrations is demonstrated. Calibrations based on filtered interferogram segment magnitudes vs concentration yield models with values of R2 in excess of 99%. These results are obtained without the use of a separate background or reference interferogram. This interferogram-based analysis is shown to perform analogously to a conventional spectral-based analysis, with the interferogram method being more efficient in terms of data collection and computational requirements.
本文证明了使用带通滤波傅里叶变换红外（FT-IR）干涉图的短片段进行定量分析的可行性。本研究中开发的方案解决了阻碍FT-IR光谱在非实验室应用中使用的四个限制：（1）需要坚固、低成本且可靠的光谱仪；（2）缺乏用于获取目标分析物吸收光谱的代表性背景光谱；（3）存在干扰分析物测定的重叠光谱带；（4）难以从检测限附近收集的数据中获取有用信息。在本研究中，通过应用窄带通数字滤波器，直接从短干涉图片段中分离出与特定感兴趣分析物谱带相关的光谱信息。以这种方式处理时，滤波后的干涉图片段包含可用于定量分析的化合物特异性信息。证明了对不同浓度的苯和硝基苯的滤波干涉图数据成功使用单变量校准程序。基于滤波干涉图片段幅度与浓度的校准产生的模型，其R2值超过99%。这些结果是在不使用单独的背景或参考干涉图的情况下获得的。结果表明，这种基于干涉图的分析与传统的基于光谱的分析类似，干涉图方法在数据收集和计算要求方面更有效。
