Towards a Rationalization of Ultrafast Laser-Induced Crystallization in Lithium Niobium Borosilicate Glasses: The Key Role of The Scanning Speed

Femtosecond (fs)-laser direct writing is a powerful technique to enable a large variety of integrated photonic functions in glass materials. One possible way to achieve functionalization is through highly localized and controlled crystallization inside the glass volume, for example by precipitating nanocrystals with second-order susceptibility (frequency converters, optical modulators), and/or with larger refractive indices with respect to their glass matrices (graded index or diffractive lenses, waveguides, gratings). In this paper, this is achieved through fs-laser-induced crystallization of LiNbO3 nonlinear crystals inside two different glass matrices: a silicate (mol%: 33Li2O-33Nb2O5-34SiO2, labeled as LNS) and a borosilicate (mol%:33Li2O-33Nb2O5-13SiO2-21B2O3, labeled as LNSB). More specifically, we investigate the effect of laser scanning speed on the crystallization kinetics, as it is a valuable parameter for glass laser processing. The impact of scanning energy and speed on the fabrication of oriented nanocrystals and nanogratings during fs-laser irradiation is studied. Fs-laser direct writing of crystallized lines in both LNS and LNSB glass is investigated using both optical and electron microscopy techniques. Among the main findings to highlight, we observed the possibility to maintain crystallization during scanning at speeds ~ 5 times higher in LNSB relative to LNS (up to ~ 600 μm/s in our experimental conditions). We found a speed regime where lines exhibited a large polarization-controlled retardance response (up to 200 nm in LNSB), which is attributed to the texturation of the crystal/glass phase separation with a low scattering level. These characteristics are regarded as assets for future elaboration methods and designs of photonic devices involving crystallization. Finally, by using temperature and irradiation time variations along the main laser parameters (pulse energy, pulse repetition rate, scanning speed), we propose an explanation on the origin of 1) crystallization limitation upon scanning speed, 2) laser track width variation with respect to scanning speed, and 3) narrowing of the nanogratings volume but not the heat-affected volume.