Multi-platform chemometric analysis of volatile and non-volatile components for quality evaluation of base liquor: A case study on Qingxiangxing Baijiu

Rapid and accurate evaluation of newly distilled base liquor is essential for ensuring product consistency and quality. To address this challenge, a multi-platform analytical and chemometric approach to assess the quality of different grades Qingxiangxing crude Baijiu (QXCBJ) produced continuously was employing. Volatile organic components of QXCBJ were profiled using gas chromatography-flame ionization detector/mass spectrometry (GC-FID/GC-MS), and non-volatile organic components (NVOCs) were characterized using liquid chromatography-high resolution mass spectrometry (LC-HRMS). In total, 136 VOCs and 270 NVOCs were identified, and 36 of them were screened as potential quality markers. In the GC-MS analysis, isoamyl lactate, ethyl phenylacetate, 4-ethyl-2-methoxyphenol, ethyl cinnamate, and 2-methoxy-4-vinylphenol showed significant differences among different QXCBJ samples, contributing to sample differentiation. Metabolomic profiling via LC-HRMS revealed differential accumulation of gustatory components (malic acid, vanillin acid, mannitol, betaine, azelaic acid, and lysine) across QXCBJ samples. Multifactor analysis (MFA) demonstrated that the integrated multi-platform strategy significantly improved sample discrimination compared to single-method approaches. This comprehensive methodology not only establishes a scientific basis for the quality control of Qingxiangxing Baijiu (QXBJ) but also demonstrates the potential of this multi-platform chemometric strategy for evaluating other Chinese spirits with similar complexity. • Comprehensive evaluation reveals the complex sensory profile of crude Baijiu. • A comprehensive profiling of 136 VOCs and 270 NVOCs in crude Baijiu was achieved using a multi-platform analysis. • Non-volatile organic compounds were applied to the discovery of sensory grade-differentiating components. • Thorty potential quality control markers were screened through multivariate statistical analysis.