分别利用超临界CO2、亚临界CO2与石油醚萃取茶树花精油，采用GC-MS分离和鉴定这3种萃取方法得到的茶树花精油的挥发性成分，再运用主成分分析程序分析和比较这三个样品的主要挥发性物质。结果表明：茶树花萃取物中共有58种挥发性成分，其中超临界CO2、亚临界CO2与石油醚的萃取样品各有55、50和54种成分。有17种脂肪烃与芳香烃，4种萜烯、6种萜烯醇与萜烯酮、4种芳香醇与芳香酮、2种脂肪醇、6种脂肪酮、2种醛、5种酯、7种酸。茶树花的主要挥发性成分有苯乙酮（超临界、亚临界与石油醚萃取样品中分别含有11.04、10.23、8.41 μg/g。以下也按此顺序表示）、苯乙醇（11.49、9.98、8.52 μg/g）、α-苯乙醇（13.91、12.30、9.72 μg/g）、氧化芳樟醇（5.34、6.35、4.38 μg/g）、苯甲醇（6.60、5.34、4.43 μg/g）、1, 8-二甲基-4-异丙基-8, 9环氧-茶螺酮（3.68、2.37、1.61 μg/g）、β-紫罗兰酮（2.67、1.76、1.16 μg/g）。主成分分析结果显示萃取物样品的挥发性成分共有2种主成分。超临界萃取样品的香气成分比亚临界与石油醚萃取的更丰富。

The volatile components of tea blossom were extracted by supercritical CO2, subcritical CO2, and petroleum ether medium extractions. Gas chromatography-mass spectrometry was applied to isolate and identify the volatile components in the three extracts. Principal component analysis, using the SPSS software, was adopted to analyze and compare the main flavors of the three extracts. Results showed that there were, in total, 58 volatile components in tea blossom. Of these, 55, 50, and 54 components were identified in the supercritical CO2, subcritical CO2, and petroleum ether extracts, respectively. They included 17 fatty hydrocarbons and aromatic hydrocarbons, 4 terpenes, 6 terpenols and terpenones, 4 aromatic alcohols and aromatic ketones, 2 aliphatic alcohols, 6 aliphatic ketones, 2 aldehydes, 5 esters, and 7 acids. The major volatile compounds of tea blossom included 2-hydroxy-2-phenylacetophenone (11.04, 10.23, and 8.41 μg/g, respectively, in the supercritical CO2, subcritical CO2, and petroleum ether extracts; same below), phenethylol (11.49, 9.98, and 8.52 μg/g), α-phenethylol (13.91, 12.30, 9.72 μg/g), linalool oxide (5.34, 6.35, and 4.38 μg/g), benzyl alcohol (6.60, 5.34, and 4.43 μg/g), 1,8-dimethyl-4-isopropyl-8,9- epoxy theaspirone (3.68, 2.37, and 1.61 μg/g), and β-ionone (2.67, 1.76, and 1.16 μg/g). The two principal components were deduced based on an analysis by SPSS. The fragrances of the supercritical CO2 extract were found to be richer than those of the subcritical CO2 and petroleum ether extracts

The National Natural Science Foundation of China (Grant No. 31301530); The Fundamental Research Funds for the Central Universities (Grant No.2013ZM0063); Department of Science and Technology of Guangdong Province (Grant No. 2012B020311004, 2010B020312005) and the Science and Information Technology of Guangzhou, China (Grant No. 2010Z1-E221)