Research On The Compensation Method To The Error Arisen From The Constant Optical Path Length In Detection Of Pesticide Residue Based On The Microfluidic Chip

At present,most of the detection equipment applied in the pesticide residue detection has the defects of large size,low detection accuracy and poor stability.Therefore,a portable pesticide residue detection device is constructed based on microfluidic chip.However,there are two challenges remain to the portable pesticide residue detection device:First,there exists the error arisen from the constant optical path length;Second,there is no corresponding hardware or software error compensation model.In view of the above problems,the research work is carried out in the following aspects:(1)the relational model between the error arisen from the constant optical path length a and the optical path l is established according to the Lambert-Beer's law and there is also an amendment for the model from the point of noises;(2)The CorelDraw is used to design the centrifugal and CD-like microfluidic chip and the simulation is finished by COMSOL,then the chip materials and production methods are determined,a pesticide residue photometric detection device is built and three parameters are optimized,including experimental temperature,light source intensity and wavelength.With the help of the device,the simulation result of solution homogeneous index is validated and the hardware compensation model caused by the error arisen from the constant optical path length is built;(3)By using the three algorithms of the least square method,BP neural network and support vector machine,the corresponding error compensation models are are established,and the optimal model is selected as the final software model of the error;(4)A portable pesticide residue detection device is built based on microfluidic chip,the hardware circuit design of the portable device is introduced and three parameters of substrate concentration,enzyme inhibition time and pH value of the solution are optimized.With the help of the portable device and the established compensation models,there is a detection of trichlorfon,carbofuran and monocrotophos in strawberry and tomatoes to verify the correctness of the hardware and software compensation models.The experimental results show that the root mean square error(RMSE)are 0.3901 and 0.0883 after the the hardware and software compensation models are carried out respectively,so the compensation to the error arisen from the constant optical path length is realized basically.Compared with the traditional photometric detection method based on microfluidic chip,the errors of the detection of organophosphorus and carbamate pesticides including trichlorfon,carbofuran,monocrotophos in strawberry and tomatoes are reduced by more than 24.30%after the compensation models are carried out.The accurate detection of pesticide residues is realized basically.The relative errors between the detection values of pesticide residues and the standard values are 3.99%;4.58%,4.03%respectively after the compensation models are carried out.So,the hardware and software compensation model are basically correct.