Development Of New Taste Sensor Array And Detection System For Traditional Brewed Food

Brewed food was a type of food that people use the fermentation to process and manufacture with the help of biological catalysts.The traditional brewed food in our country was very popular among consumers due to its unique flavor and richness in our dietary life.In recent years,people’s demand on food quality has been increasing.The quality examination of traditional brewed food consists of the quantitative determination of important components and the evaluation of flavor.At present,the quantitative determination of components mainly depends on the analysis technology of large-scale food instruments.But the sample pretreatment is often cumbersome,and the sample detection is time-consuming and laborious.The evaluation of flavor mainly depends on artificial sensory evaluation methods,which have the disadvantages of strong subjectivity and cannot be quantitatively analyzed.This research develops a new taste sensor array and detection system for traditional brewed food soybean paste,and the purpose is to promote the industrialization and scale development of traditional brewed food.The main contents of this study include:1.The overall design of a new taste detection systemThe fermentation process experienced by traditional brewed foods was relatively complicated.The small molecular substances produced by the decomposition of food during fermentation were not only diverse,but also differ in content between different components.Taste detection system was expected to be able to detect complex solution systems with sensor arrays compose of a few sensors.Moreover,taste detection system needs the advantages of high sensitivity,versatility,simplicity,robustness,etc.,so it could meet the detection requirements of traditional brewed food.After pre-research,the existing commercial products could not meet the requirements.Therefore,this research specifically developed a new taste sensor detection system for traditional brewed food detection.The overall design of the system consists of three parts:new sensor array,data acquisition system,data analysis with pattern recognition system.2.Design and construction of a new sensor arrayA sensor array was the core of the entire taste detection system.First,a sensor array was designed according to the characteristics of traditional brewed food.And then a new taste detection system was builted around the sensor array.In this study,the nano-metal modified electrode sensor were used to build the sensor array,and graphene modification was used to enhance the sensor’s responsiveness.The working electrode of the new sensor array is composed of gold,silver,zinc,copper composite graphene modified electrodes,the auxiliary electrode was platinum column electrode,and the reference electrode was silver chloride electrodes.The cyclic voltammetry method was used to screen the metal-modified materials of the sensor.Modified Hummers’s method was used to prepare graphene;and the electrodeposition method was used to prepare graphene/metal nanoparticle composite modified electrodes.The electrochemical performance of the obtained composite modified electrode was tested.The peak current of the modified electrode’s cyclic voltammetry response was3.24×10^-5 A,which was 1.33 times that before modification.The results of current showed that the electrical performance of the modified electrode was greatly improved.3.Design of data acquisition system and data analysis with pattern recognition systemThe performance of the data acquisition system and data analysis with pattern recognition system must match the new sensor array to achieve good detection results.The pulse signal generator was used as the input terminal of the data acquisition system to provide the pulse potential for the sensor.The design of constant potential circuit used a reference voltage source and a negative feedback circuit to control the potential stability of the working electrode.A program-controlled filter amplifier circuit was used to eliminate noise while amplifying useful signals.The output end of the data acquisition system was a data acquisition card,which collects the data and stores it in the computer.Because the data acquisition card only collected voltage signals,a current-voltage conversion circuit was added to the conditioning circuit to convert the current signals into voltage signals.According to the characteristics of traditional brewed food,we specially developed data analysis and pattern recognition systems.In view of the characteristics of the original data with many dimensions and large cardinality,feature value extraction and principal component analysis were used to reduce the feature information of the sample to several principal components.Pattern recognition used Partial Least Squares（PLS）and Support Vector Machine（SVM）methods to establish prediction models.4.Research on the practical verification of the new taste detection systemAs a typical traditional brewed food,soybean paste was selected as the detection object to verify the new taste detection system developed.（1）In accordance with the national standards,soybean paste samples were artificially evaluated.The results were used to build the predictive models for the flavor properties of soybean paste based on the new taste detection system,and the model used the artificial sensory evaluation results of soybean paste as the output reference of the model,and the model used the characteristic value of the response signal of the flavor sensor array as the input.In the constructed PLS prediction model,the Rp of the prediction set was 0.92327 and the RMSEP was 6.15.The Rp of the SVM model was 0.86575 and the RMSEP was 16.09.The results showed that the PLS model has high correlation and stability in the prediction of flavor quality,and the new taste detection system has realized the effective prediction of the flavor quality of soybean paste.（2）Refer to the national standards,the physicochemical index of soybean paste samples were detected.The results were used to build the predictive models for the content of ammonium salt and amino acid nitrogen of soybean paste based on the new taste detection system,and the input of the model was the characteristic value of the response signal of the taste sensor array,and the output of the model refers to the physicochemical detection results.The Rp of the PLS model of amino acid nitrogen was 0.90752,and the RMSEP was 0.0143.the Rp of the SVM model of amino acid nitrogen was 0.86928,and the RMSEP was 0.0102.For the PLS model of ammonium salt,the Rp was 0.87253 and the RMSEP was 0.0073.The SVM model of ammonium salt,the Rp was 0.90472 and the RMSEP was 0.0143.The results showed that the PLS model in the prediction model of amino acid nitrogen has a better correlation,but the SVM model is more stable.The SVM model of ammonium salts has better correlation and less stable than the PLS model.In general,the new taste detection system could effectively predict the content of soybean paste components.The results proved that the new sensor array and taste detection system developed have a better detection effect on traditional brewed food represented by soybean paste.The results obtained can be popularized for the rapid detection of traditional or similar traditional brewed food,thereby promoting the rapid and stable development of the industrialization of the traditional brewed food industry.