Investigation Of Some Theory And Application In Infrared Radiometry Technique

Photoacoustic (PA) and photothermal (PT) techniques are important means of nondestructive detection. Because of their superior performance in high sensitive, broad detecting bandwidth and good adaptability, they have been used extensively in such fields as physics, chemistry, biology, materials science, medical diagnosis and aeronautic and spatial technologies.Studied in this thesis is the exploration for founding new methods to reconstruct physical parameter distribution in materials based on PT technique. We start with the research in two kinds of PT signals by both theoretical analysis and experimental validation. Then, a new way of doing depth profiling reconstruction from measured PT data is presented, and more details are summarized in following.Firstly, we make a characteristic analysis in evolution of photothermal temperature spectrum (PTS) responded with the variation of the interlayer included in a layered sample. In order to calculate PTS signals, two methods are presented to solve the heat conducting equation in frequency domain, and also theoretical analysis is given to discuss their practicality and limitations. The obtained results include: (1) the extreme values of the signal spectrum are mainly affected by the position of the defect, and without obvious variation while the width or absorption intensity of defect is changed; (2) a special curve is found from which one can predict the position of a defect by using detected signals; (3) more detailed analysis about the responding characteristic of PTS are displayed by signal ratio curves.On the basis of the previous work, the theoretical analysis and application of photothermal radiometry infrared spectrum (PRIS) have been invested in detail. The main contents include: (1) to find the theoretical model of PRIS and make numerical simulations and compare the results with experimental data; (2) to give an analysis in the responding of the PRIS with the variation of the interlayer included in a layered sample; (3) to find a special curve which can be used to predict the position of the defects in a sample from practical signals. The obtained results are not only very helpful for constructing inverse algorithm but also can be used directly to analyze thedefect hided in a solid sample and to make depth profiling image of the sample.In the last part of this thesis, we investigate the feasibility of using a new kind of artificial neural network (ANN) to reconstruct the depth profile of photo absorbing coefficients of samples from PTS or PRIS. The main contents include: (1) the structure design of a NN group which is able to treat inverse problems; (2) the training method of NN and its capability testing on prediction and association ability; (3) some factors that are possible to affect the inversing performance of the NN are discussed, which is useful to provides necessary experience for practical application of the NN. The numerical simulating results from large numbers of samples confirmed the advantages of the NN in following aspect: (1) quick calculating speed and its capacity of treating sample in batches; (2) good recognizing and associating ability, i.e., a well-trained NN can recognize not only the monotone and smooth functions but also those with discontinuous profiles; (3) high stability and validity in resisting noise disturbance. The introduced NN has potential application in the field of nondestructive detection of materials.