Investigation of the fs-micromachining process in GaN and diamond

• Autor(a) principal: Nolasco, Lucas Konaka
• Data de Publicação: 2021
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Biblioteca Digital de Teses e Dissertações da USP
• Texto Completo: https://www.teses.usp.br/teses/disponiveis/18/18158/tde-22042021-165900/
• topic: CVD diamond; Diamante CVD; Efeito de incubação; Fs-micromachining; GaN; GaN; Incubation effect; Microfabricação; Nonlinear optics; Óptica não-linear; Semicondutores de banda larga; Wide bandgap semiconductor

Semiconductor materials are essential for novel technology development in optoelectronic and photonic devices. Among these types of materials, diamond and GaN stand out given their high bandgap and optimal electronic and thermal properties. Amidst the many material processing techniques, the femtosecond laser pulses micromachining is employed due to its high precision and ability to create microstructures either on the surface or bulk of a material. Thus, in this dissertation, we have studied the fundamentals of fs-micromachining process in both GaN and CVD diamond at 343, 515 and 1030 nm. More specifically, we have examined the incubation effect: the damage threshold fluence (minimal fluence necessary to produce damage in the material) decreases with the number of femtosecond pulses applied per sample spot. The threshold fluence was determined through the zero damage method, which consists of using Gaussian intensity profiles with distinct energies in a material surface. Hence, the incubation curves were obtained for both materials and were fitted by an exponential defect model, which correctly predicts the saturation of the threshold fluence observed in the high-pulse superposition region. The model also indicates through its incubation parameter (k) the efficiency by which the fluence reaches the saturation: the higher its value, the less pulses are necessary. For the GaN sample, k = (0.4 ± 0.2) for 343 nm, k = (0.07 ± 0.01) for the green (515 nm) and k = (0.02 ± 0.01) for the IR case (1030 nm). In CVD diamond: k = (0.13 ± 0.04) for the UV, k = (0.3 ± 0.1) for 515 nm and k = (0.14 ± 0.03) for the 1030 nm excitation wavelength. A theoretical model, which assumes that only multiphoton and avalanche absorption are present, was used to compare its results of the single pulse damage threshold fluence to our experimental data. Through a numerical simulation based on this model, we determined the threshold fluences for GaN at all wavelengths, resulting in a satisfactory agreement to most experimental data, except at 1030 nm. This discrepancy was explained by determining the Keldysh parameter, which indicated that both multiphoton and tunneling absorption are present at this wavelength in GaN – a process that is not considered by the model. As for the CVD diamond, we also established good results for the UV (343 nm) and green (515 nm) excitation wavelengths, but due to the lack of absorption cross-section data at 1030 nm (required for the simulation), we resorted to an alternative method using the same model to determine the five-photon absorption cross-section (σ5), resulting in σ5 = 5×10-170 m10 s 4 photon-4, which is in good agreement with other five-photon absorption cross-sections of other materials in the literature. Hence, this study could prove important in improving the femtosecond micromachining processing technique in both GaN and diamond.