| Accurate, stable, and low-cost soil macronutrient detection technology has become an important direction for domestic and foreign scholars to explore and research precision agriculture. Posessing grater superiority in convenient operation, easy intergration and low cost over other methods, ion selective electrode(ISE) was commonly used in soil macronutrient rapid detection for the grassroots application. Aiming at settling the existing problems of conventional ion selective electrode(ISE) in sample contamination and microminiaturization, solid-state nitrate, potassium and phosphate ISEs were developed on the basis of previous literature research and feasibility analysis. The electrochemical performances of these self-developed ISEs were systematically analyzed in standard solutions as well as soil solutions. The dynamic response mechanism of the solid-state ISE was demonstrated with the model theory. Improvements were made on the existing detection platform. The suitable soil pretreatment parameters for ISE method were optimized, and the effects of soil extracts on ISE performance were discussed. The main findings are summarized as follows:(1) Novel solid-state nitrate and potassium ISEs, using graphene as the solid-contact layer, glassy carbon electrode as the substrate, nitrate doped polypyrrole and valinomycin as sensitive materials for nitrate and valinomycin respectively, were developed. Solid-state phosphate ISE was fabricated with nano cobalt particles. Performance testing results of these three ISEs in standard solutions indicted that:a Nernstia response of (-54.1＋1.1) mV/decade over the linear response rang of10-4～10-1mol/L and the low detection limit of10-4.7mol/L were obtained for nitrate ISE. Its response time was20s. Potassium ISE exhibited a linear response rang from10’5to10-1mol/L with the Nernstia response slope of (57.9＋0.8) mV/decade. The low detection limit was10-56mol/L. Equilibrium respones in solutions could be arrived within25s. Phosphate ISE has a linear response rang of10’5-10-2mol/L and Nernstia response slope of (-27.4＋2.8) mV/decade. The response time was100s. The sensitivity and stability of nitrate and potassium ISE could meet the demand of clear soil extracts analysis of nitrate and available potassium. However, further research should be conducted on the performance improvement of phosphate ISE.(2) Further theoretical analysis was made on the dynamic response mechanism of the solid-state ISE with the Phase Boundary Model and Nernst-Planck-Poisson Model. Using Phase Boundary model, the distribution characteristics of the preferred and discriminated ion in both diffusion layer and ISE membrane were simulated under the steady-state conditions. The influence of interfering ion and membrane thickness on the ISE performance, and the preferred ion concent at the inner side of the ISE membrane on the low detection limit were discussed. Using Nernst-Planck-Poisson model, the distribution of the preferred and discriminated ion in both diffusion layer and ISE membrane were analyzed in space and time domains. The influence of model parameters, including diffusion layer thickness, diffusion coefficients of the involved membrane species, permittivity, and film thickness, on ISE performance were simulated. Simulation results of the two models provided mechanism analysis for the effects of nitrate ion and pyrrole monomer concentration in polyelectrolyte on the sensitivity and precision of the nitrate doped polypyrrole ISE made by the constant-current method. Thus, the relevant dynamic modeling theory could provide theoretical foundations for the ISE preparation process and its detection mechanism analysis.(3) Improvements were achieved on the soil macronutrient detection platform. Firstly, the existing multi-channel data acquisition analysis processor was modified to obtain a better anti-jamming performance. Then, an automatic fluidic transfer system was developed. At last, the PC monitoring software was consummated by adding more functions. The experimental results showed that the improved platform produced a significantly reduced indication error than the original system. The zero drift problem was also ameliorated effectively. The measurement accuracy was greatly improved. Different channels of the platform possessed good consistency and self-clean ability.(4) Feasibility verification was conducted on the ISE-based soil nitrate and available potassium detection system with a large number of soil testing experiments. The process and parameters of soil pretreatment applicable to the ISE method were determined. Extractant of distilled deionized water, soil-to-solution ratio of1:2.5, shaking time of15minutes without stirring and soil flocculant were recommended for soil NO3-N extraction. Extractant of Kelowna solution, soil-to-solution ratio of1:10, and shaking time of15minutes without stirring were recommended for soil potassium extraction. Comparisons were made on the diminishment of environmental disturbance applying the classical Nernst model, BP-ANN model and RBF-ANN model data processing algorithm, and RBF-ANN model was determined to be used for the coexistence interference inhibition in soil testing. In order to obtain the best prediction performance, the model parameters were optimized. A large number of soil testing experiments were carried out, the results showed that the absolute errors between the NO3-N results analyzed by the detection platform and conventional UV spectrophotometry ranged from0.04to18.1mg/L with the RMSE of4.4mg/L when soil NO3-N content ranged from4.0to168.4mg/L. When soil available potassium content ranged from2.6to365.2mg/L, the absolute errors between the potassium results tested by the detection platform and conventional flame photometry ranged from0.4to7.7mg/L mg/L with the RMSE of3.7mg/L. Good consistency and correlation was existed between the results determined by the detection platform and conventional methods. It indicated that the developed soil macronutrient detection platform could meet the basic needs of soil testing and fertilizer recommendation project. |
PDF Full Text Request