| Abstract [eng] |
Intense laser radiation is controlled using optical elements. When it exceeds the optical resistance limit (LIDT - Laser-Induced Damage Threshold) the element is damaged. In developing high-power laser systems, it is, therefore, of crucial importance to understand thoroughly the physics of optical resistance and accurately assess LIDT of optical elements. In the nanosecond pulse regime, LIDT is mostly determined by laser light-absorbing nano-size defects inherently occurring on the surfaces and in the bulk of transparent materials in the process of manufacturing optical elements. Thus, the aim of the thesis is to advance optical resistance metrology towards better quantification of laser damage precursors. By employing Monte Carlo method, it has been shown that maximum likelihood based evaluation procedure, which takes into consideration the binomial nature of damage probability measurement, helps to improve the repeatability of determined LIDT values, as compared with standard approach based on the least squares method. The accuracy of the determined results increases when the fluence fluctuations are taken into account in the damage probability model. These improvements have been tested experimentally by making 1-on-1 LIDT measurements in the nanosecond pulse regime and at a wavelength of 355 nm. In order to assess the true distributions of defect ensembles of damage precursors, a direct comparison of two methods, i.e. raster scan and damage probability measurements, has been performed for the first time. The research has found that the results obtained using the raster scan method are affected due to surface contamination by ablation products, and the quantity of defects might be misinterpreted owing to defect clusters. It has been demonstrated that in order to interpret defect density dependence on peak laser fluence correctly, spatial intensity distribution should be taken into account. In an attempt to understand the role of defects in multilayer coatings better, optical resistance metrology has been expanded to include interference phenomena and volumetric defect ensembles. A new statistical tool for interpreting data on damage probability measurements has been created, which allows for a better understanding of the properties of surface defects occurring in the process of polishing when they are additionally deposited with transparent layers. Also, it has been revealed that the LIDT of identical layers forming multilayer highly reflective HfO2-SiO2 mirror coating depends on a specific depth of a layer in terms of the surface. |
