Objectives This study was to establish a non-destructive method for rapid detection of nitrogen contents in tea leaves using electronic nose and spectrophotometer.
Methods The 3rd or 4th healthy leaves below the top bud of tea tree were used as the tested materials. In preliminary experiment, the parameters of gas collector volume, headspace preheating temperature and headspace time were optimized. The first step of the main experiment used the built-in Winmuster software of electronic nose, and principal component analysis (PCA), linear discriminant analysis (LDA) and loadings analysis (LA) were carried out with characteristic values responded to sensors after optimization, thus the most sensitive sensors were chosen. Meanwhile, the chromatic values were determined by spectrophotometer. The measured area was mesophyll with 20 replications. The chromatic values included L (represents black or white), a (represents red or green) and b (represents yellow or blue). The one variant linear regression analysis to L, a and b values was conducted by Origin 8.0 software, and one way Anova analysis was made with SPSS 16.0 software and eventually the t test was done. Then chromatic aberration values were screened and parameter with the highest correlation coefficient was acquired. And nitrogen content was detected with Kjeldahl method. In the second step of main experiment based on the electronic nose and spectrophotometric colorimeter data under different nitrogen contents, the prediction models of nitrogen content by odor, by color and by odor combining with color were established and compared.
Results In preliminary experiment, the electronic nose detection system was established: with 50 mL gas collector volume, 30℃ headspace temperature and 30 min headspace time. The nitrogen contents of tea leaves could be distinguished by electronic nose. In the first step of main experiment, the sensors sensitive to oxynitride (S2), methane (S6), sulfur compounds (S7), alcohols, aldehydes and ketones (S8) and organic sulfur compounds (S9) were selected in the process of parameter optimization. According to L, a, b chromametric system, the b value showed a significant linear correlation with the total nitrogen content, so that it could be used as a marker to judge leaf nitrogen contents. In the second step of main experiment based on the odor, color and odor combined with color data, the prediction models for nitrogen contents were established. Among them, the prediction model based on odor combined with color had the highest accuracy, reaching 90%.
Conclusions The prediction model in predicting nitrogen contents in tea leaves by odor combining with color had high accuracy, hence could be used in practical operation.
DownLoad: