Rapid Determination Of The Quality Of Surimi-Based Products Using Near Infrared Spectroscopy

Near infrared spectroscopy is one of high-tech analysis which has rapid development at present. It has been developed in agriculture, pharmacy, petrochemical engineering, food and other fields, and the range of its analysis is continually expanding. The industry of surimi-based products is one of the most promising industry in aquatic products industry. The traditional methods of its quality detection exist several disadvantages, such as complex experiment, long experimental period and delayed detection results, In this paper, after the introduction of the fundamental principles, characteristic, routine testing methods and steps of analysis of the Near-Infrared (NIR) spectroscopy, and the current status of the application in the field of aquatic by Near-Infrared spectroscopy, rapid determination of the quality of surimi-based products using NIR Spectroscopy technology and chemometries knowledge is studied. The research contents and conclusions mainly including:(1) NIR spectra in the range of1000-2500nm are recorded for all the homogenized samples, including minced muscles of grass carp, chub, black carp, perch and adulteration samples with varying levels (5-30% (w/w) minced muscle of chub in the minced muscle of grass carp). Spectral information used for the identification analysis and partial least square (PLS) analysis is contained in the NIR region of1100-2400nm. Pattern recognition technique of soft independent modeling of class analogy (SIMCA) provided satisfactory classification results. The results show that both samples species and adulteration samples are differentiated in their individual two-dimensional score plot using the first and the second principal component (PC) despite overlay of several samples. Additionally, a partial least square (PLS) method is constructed to predict the value of adulteration content in the minced muscle samples of grass carp. The results show that PLS models with multiplicative signal correlation (MSC) pretreatments provided acceptable results with root mean standard error of the calibration set (RMSEC)1.63, the root mean standard error of prediction set (RMSEP)1.90and correlation coefficient value (r)0.9809. The studies demonstrate that NIR can be used to identify the minced fish-meat, adulterations and quantitative detection of the adulteration levels.(2) Near-infrared reflectance spectroscopy (NIRS) is used as a rapid and nondestructive method to determine the gel molding temperature of fish-balls.150fish-balls produced at5different temperatures(40℃50℃、70℃、80℃、90℃) are prepared for establishing the predicting models. Soft independent modeling of class analogy (SIMCA) and partial least-squares (PLS) are used to develop a calibration which is tested with a validation set. SIMCA provids satisfactory classification results about these different fish-balls. The classification accuracy for the prediction samples achieve to92%.the optimal PLS model is acquired by the fitting spectral pretreatment methods. The multiply correlation coefficient (R), root mean square error of calibration (RMSEC) and root mean square error of validation (RMSEV) are respectively0.97077,4.49and4.58in the model. A promising linear relationship is observed between NIR-predicted temperatures and the actual gel molding temperatures. The result demonstrate that NIR can effectively achieve the molding temperature of fish-balls’qualitative discrimination and quantitative prediction.(3) The study builds the textural properties and gel strength of fish-ball’s PLS model by near-infrared spectroscopy. On the bases of initial model, we analyze and compare various spectral effects of pretreatment on the optimization model to obtain the optimal model. Finally, a more ideal model including hardness, cohesiveness, chewing, recoverability, and gel strength were got. The correlation coefficient (R), RMSEC and RMSEV values of the models were:0.99635、13.5、13.1;0.90753、0.0182、0.0166;0.99154、12.3、11.9;0.94310、0.0246、0.0273;0.96919、23.2、27.4. There is no significant difference between the near-infrared model predictions and the results of texture analyzed statistically including hardness, cohesiveness, recoverability and gel strength by a t-test. The results show that it is feasible to detect the textural characteristics and gel strength of fish-balls by the NIR.(4) Extraction of the myofibrillar protein from Lepidotrigla surimi, which have been heated at different temperatures between40-80℃. The effect of treatment at different temperature on structure of myofibrillars protein is studied, including protein sulfhydryl groups, protein aggregation and the change of secondary structure content. The results showed that, as temperature increasing, the total sulfhydryl content decreased significantly. The band of myosin heavy chain (MHC) is completely disappeared when the temperature of heating reached50℃, and the actin band is disappeared at60℃. The intensity of other proteins’ band decreased gradually. Determination of protein secondary structure found that the hydrogen of protein have been destroyed by heat treatment, and the rearrangement of secondary structure is happened, leading to the reduction of α-helix and β-sheet’s content, increase of disordered structure such as random coil.