Development Of Intelligent Packaging Labels For Non-destructively Monitoring Freshness Of Typical Prepared Fresh Foods

Fresh foods are perishable due to high water content and rich nutrition.Minimal processing like cutting operation causes tissue damage,increases the possibility of microbial infection,and accelerates quality deterioration.Traditional methods have been successfully applied for evaluating freshness of prepared fresh foods.However,they are expensive,sample-destructive and time-consuming.This is why the development of novel techniques to rapidly and non-destructively assess freshness of fresh food is much needed.The objective of this study was to design,apply and evaluate the effectiveness of pH-sensitive indicators as on-package colorimetric indicator label for real-time monitoring freshness of typical prepared fresh foods during cold storage.That is,the aim of this work was to study the applicability of CO₂-sensitive indicator labels for monitoring freshness of fresh-cut green bell pepper and broccoli as well as the applicability of total volatile basic nitrogen?TVB-N?-sensitive indicator labels for lean pork and bighead carp combined with chemometric methods.Firstly,evaluation of freshness of prepared fresh foods?i.e.,fresh-cut green bell pepper,lean pork and bighead carp?during cold storage by traditional quality indices and electronic nose?e-nose?,which provided the foundation for design of intelligent packaging labels.Results of multivariate statistical analysis of traditional quality indices showed that the freshness levels of fresh-cut green bell pepper,lean pork and bighead carp could be defined as fresh,medium fresh and spoiled.Fisher discriminant models based on traditional quality indices could differentiate freshness for fresh-cut green bell pepper,lean pork and bighead carp.The discriminating rates of prediction sets were 93.8%,81.5%and 90.0%,respectively.Moreover,results of evaluation of the freshness for prepared fresh foods mentioned above using electronic nose were consistent with the results of traditional quality indices.Secondly,preparation of CO₂-sensitive indicator labels and TVB-N-sensitive indicator labels based on synthetic dyes such as bromothymol blue,methyl red,bromocresol purple and mixed ones.Results showed that the type and concentration of the indicator as well as pH of the indicator solution had an impact on color response of the indicator film.The greater the hydrophilicity of the plasticizer,the better the water absorption of the indicator film,and the faster the color response of the indicator film.?1?High and significant correlations?p<0.01?were found between CO₂ concentrations in packaging and aerobic plate counts as well as sensory scores for fresh-cut green bell pepper during storage.CO₂-sensitive indicator label made by a mixture of methyl red and bromothymol blue at 3:2 proportion with an initial pH of 7.1?MB2 formula?could more clearly monitor pepper decay in cold storage at 7±1?,where indicator label of MB2 type changed from yellow-green to orange.A Fisher discriminant model based on color coordinates of MB2 indicator label could distinguish freshness for fresh-cut green bell pepper,and the discriminating rate of prediction set was87.5%.?2?The TVB-N-sensitive indicator label made by a mixture of bromothymol blue and methyl red at 3:2 proportion?at an initial pH of 5.0?was able to discriminate fresh?red?,medium fresh?goldenrod?,and spoiled?green?pork in cold storage at 5±1?.A Fisher discriminant model based on color coordinates of the indicator label for freshness identification of pork was established and the discriminating rate of prediction set was 95.8%.The least squares-support vector machine?LS-SVM?model based on color coordinates of the indicator label exhibited the best result for predicting TVB-N contents of pork.The determination coefficient of prediction?R_p²?was 0.9028 and the residual predictive deviation?RPD?was 3.1201.While the partial least squares regression?PLSR?model showed the best result for predicting aerobic plate counts of pork?R_p²=0.8802,RPD=2.9515?.?3?The TVB-N-sensitive indicator label made by bromocresol,purple glycerin and methyl cellulose?at an initial pH of 4.0?was used to non-destructively and rapidly detect fresh?green?,medium fresh?blue-green?,and spoiled?blue or blue-violet?bighead carp.The time of color response was 10 min.A Fisher discriminant model based on color coordinates of the indicator label for freshness identification of bighead carp was established and the discriminating rate of prediction set was 90.5%.The LS-SVM model based on color coordinates of the indicator label exhibited the best result for predicting TVB-N contents of bighead carp?R_p²=0.9240,RPD=3.2196?.While the multivariable linear regression?MLR?model showed the best result for predicting aerobic plate counts of bighead carp?R_p²=0.8191,RPD=2.4089?.Thirdly,combination of on-package colorimetric indicator label made by synthetic dyes and packaging containing silicon gum film window and then applicability of this colorimetric indicator label for real-time detection of fresh-cut broccoli freshness.Results showed that the optimal size of silicon gum film window of the package for 150±10 g fresh-cut broccoli stored at 4±1?was 0.9 cm².A Fisher discriminant model based on traditional quality indices could differentiate fresh,medium fresh and spoiled fresh-cut broccoli,and the discriminating rate of prediction set was 95.2%.Result of evaluation of fresh-cut broccoli freshness using electronic nose was consistent with the result of traditional quality indices.According to correlation analysis of freshness and CO₂ concentrations in packaging for fresh-cut broccoli during cold storage,results showed that MB2 indicator label could more clearly monitor broccoli decay in cold storage at 4±1?,where indicator label of MB2 type changed from yellow-green to orange.A Fisher discriminant model based on color coordinates of MB2indicator label could distinguish freshness for fresh-cut broccoli,and the discriminating rate of prediction set was 83.3%.While the MLR model based on color coordinates showed the best result for predicting aerobic plate counts of broccoli?R_p²=0.8622,RPD=2.7320?.Moreover,indicator label made by bromothymol blue with an initial pH of 7.1?B1 formula?could also detect broccoli decay in cold storage,where indicator label of B1 type changed from green to yellow-green.The discriminating rate of prediction set for Fisher discriminant model based on color coordinates of B1 indicator label was 83.3%.While the MLR model based on color coordinates showed the best result for predicting aerobic plate counts of broccoli?R_p²=0.8496,RPD=2.6152?.The intelligent refrigerator for online freshness detection of fresh-cut broccoli was initially developed.The photoelectric sensors in the refrigerator can detect the color change of the above-mentioned MB2 type and B1 type label and then freshness level of the fresh-cut broccoli was outputted.Finally,application of natural dyes instead of synthetic dyes in intelligent packaging labels for real-time detection of freshness of bighead carp heads.That is,the objective was to study the applicability of TVB-N-sensitive indicator labels incorporated with curcumin,anthocyanins and their mixture.The results showed that the stability of the indicator film incorporated with anthocyanins was worst,while the stability of the indicator film incorporated with curcumin was best,and the stability of the indicator film incorporated with the mixed dyes was moderate.Curcumin-anthocyanins/starch/polyvinyl alcohol/glycerol film?No.4 film?could more clearly divide freshness of bighead carp heads into fresh,medium fresh and spoiled groups.While curcumin/starch/polyvinyl alcohol/glycerol film?No.2 film?was able to discriminate fresh and spoiled ones.The anthocyanins/starch/polyvinyl alcohol/glycerol film?No.3 film?had the widest color change response to freshness.Fisher discriminant models based on color coordinates of the indicator films could differentiate freshness for bighead carp heads.The discriminating rates of prediction sets were 93.3%?No.2 film?,90.0%?No.3 film?and 96.7%?No.4 film?,respectively.The multivariate calibration models based on color coordinates of No.3 film exhibited the best results for predicting TVB-N contents and aerobic plate counts of bighead carp heads.While the multivariate calibration models based on color coordinates of No.4 film and No.2 film exhibited the worst results and moderate results,respectively.