Crude oils are complex mixtures of thousands of organic compounds that differ significantly in relative molecular mass， volatility， content， and polarity. Traditional methods for analyzing crude oil often involve complicated steps， consume large amounts of organic solvents， and require long sample-preparation times. These limitations lead to inefficient and time-consuming analysis processes. Crude oil is commonly analyzed by gas chromatography-mass spectrometry （GC-MS）. However， this technique is incapable of effectively separating complex crude-oil components owing to its low resolution and peak capacity， resulting in overlapping peaks that can lead to inaccurate compound identification and quantification. These challenges highlight the need for advanced analytical techniques. Comprehensive two-dimensional gas chromatography （GC×GC） is a novel separation technique that has been widely used to analyze complex samples， such as food， environmental samples， natural products， and crude oil. GC×GC has several advantages over traditional GC. Firstly， it offers higher resolution and peak capacity， thereby improving separation efficiency. Secondly， its high separation power reduces the need for complex sample pretreatment. Thirdly， the ordered separation and “tile effect” in a GC×GC chromatogram facilitate easier compound identification and quantification in complex mixtures.

In this study， we developed a gas purge microsyringe extraction （GPMSE） method for the rapid pretreatment of crude-oil samples. This method reduces sample processing time to only 10 min while minimizing organic solvent consumption. The chemical compositions of 45 crude oil samples were analyzed using GC×GC-time-of-flight mass spectrometry （GC×GC-TOFMS）， which helped to establish detailed chemical fingerprints for each sample. The GC×GC-TOFMS data were processed using multivariate statistical methods， including redundancy analysis （RDA） and Monte Carlo permutation testing， which identified 36 biomarkers that are strongly associated with the origin of the crude oil （p<0.05）. A classification model was constructed using a training set of 28 samples. Four single-source and 13 mixed-source samples were used to validate the model. The GPMSE-GC×GC-TOFMS method was demonstrated to be highly efficient and accurate. A discrimination accuracy of 97.8% was achieved during the identification of crude-oil sources. The developed method not only provides a powerful tool for tracing crude oil but also has broad applications potential， including for the detection of adulterated crude oil， tracking oil-spill sources， and monitoring oilfield development. This study offers several significant benefits. For example， it helps to address crude-oil trade fraud and supports national energy security. Additionally， it provides scientific support in relation to crude-oil quality control and risk assessment. The developed method is fast， reliable， and environmentally friendly； hence， it is expected to be a valuable tool for use in the oil industry. The GPMSE-GC×GC-TOFMS method is cost-effective and requires minimal solvent； consequently， it is an attractive option for reducing environmental impacts in laboratory and industrial settings. Furthermore， the high throughput and accuracy of the developed method make it suitable for large-scale analyses. In conclusion， this study demonstrated the effectiveness of combining GPMSE with GC×GC-TOFMS for analyzing crude oil； the ability of the method to identify biomarkers and classify crude-oil sources in a highly accurate manner represents a significant advancement in the field. Future studies are expected to further explore its applications in related areas， such as oil refining and environmental monitoring.