Guided-mode resonance in periodic surface textures on Si thin films induced by nanosecond laser irradiation

Laser-crystallization processes to convert amorphous-Si to poly-crystalline-Si are attracted considerable attention owing to their industrial applications, such as in active matrix organic light-emitting diode displays (AMOLEDs) and photovoltaic devices and general scientific interest. Multiple irradiation sequences were conducted with a 355 nm UV laser on amorphous-Si (40, 50, and 60 nm)/oxide(300 nm) layers on glass substrates. The resultant periodic surface textures on Si thin layers were characterized by resonance peaks in the transmittance spectra, depending on the polarization of the normal-incident light. By modeling the periodic textures as 1D linear gratings, the variations in peak intensities and positions were analyzed with respect to Si-layer thickness using the guided-mode resonance theory. In the case of transverse electric polarization, the electric field generated at resonance was entirely concentrated in the Si layer, leading to a red-shift with an increase in Si-layer thickness. Meanwhile, with transverse magnetic polarization, the magnetic field generated at resonance was mainly localized within the Si layer, but was partially distributed in the lower oxide layer when the Si-layer thickness was less than 45 nm and highly concentrated in the Si layer at Si-layer thicknesses greater than 45 nm. These results were consistent with the measured ripple positions and intensities. Thus, we propose that the observed resonance peaks can be used to characterize the surface ripples generated by laser crystallization, for e.g., in optical sensor devices on flexible substrates. •Surface ripples formation on an a-Si thin film via scanning with a laser beam having a Gaussian shape.•Modeling the periodic textures as 1D linear gratings.•The resonance peaks analyzed with respect to Si-layer thickness using guided-mode resonance theory.•Good correlations between measurement data and resonance modes predicted by simulations