Application Of The Metallic Nanoparticles Composites Modified Electrodes Array For Identifying The Marked Ages,Brands And Geographical Origins Of Rice Wines

Chinese rice wine enjoys "national wine" reputation because of its long history,rich nutrition and healthy function.With more and more people like drinking rice wine,many manufacturers joined the market for opportunities and profits.In contrast to the broad prospects,the approach of quality evaluation is so limited that some illegal actions such as fabricating the marked ages,faking the famous brands and geographical origins and so on often happen.These behaviors seriously violated the interests or rights of the producers and consumers.Modified electrodes have high sensitivity and selectivity because the surface of electrodes can be covered with nanometer materials and polymer of excellent chemical properties.Based on that advantage,modified electrodes have been applied to identify the marked ages,brands and geographical origins of rice wines recently.In this paper,we accomplished the classification and prediction of different rice wines based on marked age,brand and geographical origin successfully combined with pattern recognition methods by making use of the self-developed nanometer metal composites modified electrodes array.Metallic nanoparticles composites modified electrodes array provides a new and reliable solution to supervise the rice wine market.The main research contents,methods and the conclusions of this paper are as follows:(1)The polymer acid chrome blue K(PACBK)/AuNPs/GCE,polymer sulfanilic acid(PABSA)/AuNPs/GCE and polymer aspartic acid(PASP)/PtNPs/GCE were first fabricated using electrochemical methods for classifying and predicting the wine ages of 3,5,8,10,15,and 20 years produced from Guyuelongshan Company based on pattern recognition methods.The PACBK/AuNPs/GCE,PABSA/AuNPs/GCE and PASP/PtNPs/GCE are sensitive to Vitamin C(VC,sour),Tyrosine(Tyr,astringent)and Glucose(Glu,sweet)which have high correlations with the wine ages respectively.And the electrochemical behavior parameters such as pH,scan rate,accumulation potential and time about the composites modified electrodes were optimized by cyclic voltammetry(CV).Under the optimal conditions,we realized the quantitative determination of VC,Tyr and Glu solutions by using the self-made modified electrodes with the methods of linear sweep voltammetry(LS V),differential pulse voltammetry(DPV)and potentiostatic technique.What's more,the concentrations of VC,Tyr and Glu in rice wines were larger than the corresponding detection limit of the electrodes.This indicated the modified electrodes array was feasible to identify the rice wines.The respondent signals of current was obtained when the modified electrodes were applied to the rice wines with excitation waveform of multi-frequency and potential steps(STEP),and the feature data were extracted by 'area method'.Based on the feature data,PCA,LPP and LDA were applied for the classification of rice wine samples,and LDA performed best;LSSVM and LIBSVM were applied for the prediction of the wine ages,and LIBSVM worked better than LSSVM,the correlations based on the training and testing dataset were R~2 = 0.9999 and R~2 = 0.9998.(2)The PACBK/AuNPs/GCE,PABSA/AuNPs/GCE and polymer glutamate(PGA)/CuNPs/GCE were developed independently using electrochemical methods for classifying and predicting the wine brands of Guyuelongshan,Tapai and Kuaijishan(produced in Shaoxing)combined with pattern recognition methods.The PACBK/AuNPs/GCE,PABSA/AuNPs/GCE and PGA/PtNPs/GCE are sensitive to VC(sour),Tyr(astringent)and Gallic acid(GA,bitter)whose content varies greatly from brand to brand.As a result,we optimized the electrochemical parameters such as pH,scan rate,accumulation potential and time about the composite modified electrodes by CV and LSV.Then a series of concentrations of VC,Tyr and GA solutions were quantitatively determined by LSV and DPV.Finally we proved the validity of composites modified electrodes come from the fact that the concentrations of VC,Tyr and GA in rice wines were larger than the corresponding detection limit of the electrodes.The original responses was recorded by amperometric i-t curve when the modified electrodes were applied to the rice wines with excitation waveform of STEP,and the feature data were extracted by 'area method'.Based on the feature data,we applied PCA,LPP,LDA,ELM and LIBSVM to the classification of rice wine samples,and LIBSVM exhibited the most clear result with the classification accuracy of 100%(training set)and 99.05%(testing set);Similarly,we used ELM,LIBSVM and BPNN to predict the brands of rice wines.Compared with ELM and LIBSVM,BPNN had a higher prediction correct rate of 97.14%.(3)The carboxylic multi-walled carbon nanotubes(SMWCNT)/AuNPs/GCE,PABSA/AuNPs/GCE and PGA/CuNPs/GCE were developed independently using drop-coating and electrochemical methods to classify and predict the geographical origins of rice wines combined with pattern recognition methods.There were five rice wines which included Jiangsudanyang(Zhenjiang city),Qingdaojimo(Qingdao city),Zhejiangfenhu(Jiaxing city),Zhejiangtongkang(Taizhou city)and Guyuelongshan(Shaoxing city).The SMWCNT/AuNPs/GCE,PABSA/AuNPs/GCE and PGA/PtNPs/GCE have selective responses to 5'-GMP(umami),Tyr(astringent)and GA(bitter)whose content varies greatly from region to region.Then the electrochemical behavior parameters such as pH and scan rate about the composites modified electrodes were optimized by CV and LSV.Under the best detection parameters,a series of concentrations of 5'-GMP,Tyr and GA solutions were quantitatively determined by LSV and DPV.In order to verify the validity of composite modified electrodes,we calculated the detection limit of each electrode and found they were less than the practical contents of 5'-GMP,Tyr and GA in rice wines.The current responding curves were obtained when the modified electrodes were put into the rice wines with excitation waveform of STEP,and the feature data were extracted by 'area method' from the i-t curve.Based on the feature data,we applied PCA,LPP,LDA,ELM and LIBSVM to the classification of rice wine samples.Both ELM and LIBSVM gave a satisfactory classification result while PCA,LPP and LDA had the problem of little distances between classes;ELM and LIBSVM were also used to predict the geographical origins of rice wines.Both of them performed well,especially the corresponding coefficients of determination(R~2)were good.