Dynamic Changes Of Souble Sugar And Organic Acid Content During Development Of Citrus Fruits And The Construction Of Non-Destructive Determination Model By Near Infrared Spectroscopy

Sugar and acid contents is one of the most important qualities of citrus fruit. Their dynamic change during fruit growth has great referntial value for the researches of forming mechanisms of fruit flavor and relevant qualities. Meanwhile, it can be used to judge the maturity and confirm the picking time.In this study, conventional techniques and gas chromatography (GC) analysis were applied to detect content of several sugar and acid in four citrus varieties during different development stages of two growth seasons. The detected items contained total soluble solid (TSS), titrable acidity (TA), soluble solid content (SSC), organic acids contents such as citric acid, quinic acid and malic acid, and soluble sugar such as glucose, fructose and sucrose. The four varieties were Cara Cara navel orange (Citrus sinensis L. Osbeck cv. Cara Cara navel orange), Seike navel orange (C. sinensis L. Osbeck cv. Seike navel orange), Ponkan mandarin (C. reticulata Blanco cv. Ponkan), and Owari Satsuma mandarin (C. unshiu Marc. cv. Owari) By combing the above results with corresponding change of near-infrared spectroscopy value, non-destructive detection model for common qualities was established. The main results are as follows:1. Analysis and evaluation on common qualities of different citrus varietiesTSS, TA and SSC content in the four citrus varieties all increased gradually with fruit development. Comparison result between the year 2008 and 2009 indicated that the SSC of Immature-green (IG) stage in 2008 was significantly lower than that of 2009; SSC in Seika navel orange was highest at mature-green (MG) and mature stages of the two years, while soluble sugar content decreased after the breaker (B) stage in 2009; TA content decreased with fruit development, although in Owari Satsuma mandarin there was a fluctuation within a narrow range before B stage; TA content of Ponkan mandarin in 2009 was higher than the other three varieties. 2. Analysis and evaluation of different citrus soluble sugar and organic acidsGC detection result of soluble sugars and organic acids content in juice sacs indicated that soluble sugar content all tended to increase with fruit developmrnt in the four citrus varieties though there was a mild decrease at the initial stage; Sucrose content was higher than fructose and glucose content. During fruit development, Ponkan mandarin, Seika navel orange and Cara Cara navel orange accumulated mainly hexose, while Owari Satsuma mandarin accumulated sucrose.The organic acids content decreased with fruit development, while an ascent appeared in the early stage. Quinate acid was the main acid in flavedo and albedo of ponkan and mandarin. However, in the two navel oranges, quinate acid was only dominate at the early-middle stage. Citric acid became the main acid after B stage, when the fruit began coloring. The result suggested that different regulatory mechanism on organic metabolism existed between mandarin and navel oranges.3. Evaluation of non-destructive model about different citrus varietiesFirst-derivative and Norris denoise were applied to pre-treat original spectroscopy. Partial least squares (PLS) algorithm was used to construct the near infrared spectroscopy based non-destructive determination model, which was applicable to detect common fruit qualities like SSC, TA and soluble sugar content of Ponkan mandarin, Owari Satsuma mandarin, Seika navel orange and Cara Cara navel orange. As displayed in the models, for TSS content of the four citrus varieties, coefficient of determination (r) values were 0.9747/0.9735/0.9635/0.9438, the root mean square error of calibration (RMSEC) values were 0.49/0.55/0.74/0.79, the root mean square error of prediction (RMSEP) values were 0.60/0.41/0.72/0.62; for TA content, r values were 0.8969/0.9513/0.9343/ 0.9680, RMSEC values were 0.34/0.23/0.10/0.11, RMSEP values were 0.31/0.23/0.11/ 0.11; for SSC, r values were 0.8282/0.7733/0.8542/0.8971, RMSEC values were 0.14/ 0.12/0.18/0.21; RMSEP values were 0.17/0.03/0.15/0.16. Models for TSS and TA content gave lower RMSEC、RMESCV and RMSEP values and better prediction results would be produced accordingly. R value of models for SSC was somewhat low, while RMSEC, RMSEP and RMESCV values were all low, suggesting the applicability of this model in the range of allowable deviation.