Application of Mid-Infrared Photoacoustic Spectroscopy in the Prediction of Soil Properties.

Photoacoustic spectroscopy (PAS) is based on the absorption of electromagnetic radiation by analyte molecules, and this technique has emerged as a valuable tool for the study of materials like biological, chemical and geological samples. In this paper, FTIR-PAS spectroscopy was used in the soil identification as well as the prediction of soil properties. Firstly, three typical soil types in China (red soil in the south, fluvo-aquic soil in the north, and the rice soil in the east) were used to do the soil identification using infrared photoacoustic spectroscopy for qualitative analysis; Fourier Transform mid-Infrared Photoacoustic Spectra (FTIR-PAS) of these soil samples were recorded, and the soil spectra indicated more abundant soil information comparing with infrared transmittance and reflectance spectroscopy, and the soil type could be fast identified with these spectra information. Secondly, air-dried soil samples (n = 56) with varied fertility from Fengqiu Ecology experimental Station Chinese Academy of Sciences were involved in this experiment for the purpose of quantitative analysis; overlapping of absorption made it difficult to direct measurement of soil properties. Partial Least Square (PLS) and Artificial Neural Networks (ANN) models based on soil FTIR-PAS spectra were developed to predict soil available N, P, K and organic matter. Calibration error, calibration coefficient, validation error and RPD (Ratio of standard deviation and standard error in validation) were obtained in each calibration. The calibration statistics showed that these two models were very suitable to use in the fast prediction of soil available N, P, K and organic matter. This prediction technique was non-destructive, no sample pretreatment was needed, and online monitoring could be realized, which made FTIR-PAS a very promising method for fast evaluation of soil properties and soil quality.