Binary oxide mixtures as a keystone for new coated components in the UV: Multiscale study of nanosecond laser-induced damage

Abstract

Summary form only given. The development of new, more powerful laser systems in the last decades has been accompanied by an extensive research in the development of more laser-resistant coated components. Abrupt interfaces between deposited layers of such components have been shown to be the main cause of low laser-damage resistance. Indeed, defects and impurities concentrate at interfaces, and efforts have been concentrated on the development of new designs with smother interfaces. To perform those designs, like in rugate filters [1], binary oxide mixtures have been employed and the resulting components have shown a higher laser resistance in the infra-red range (IR). Moreover, oxide mixtures show interesting physical because in some cases, their properties are not simply a linear combination of both pure deposited materials [2], but the whole structure behaves as a new material with its own physical properties. In the Ultra-Violet (UV) range, where photon energies are particularly high, the lack of laser resistant optical components is critical, and oxide mixtures are seen as good potential candidates to increase the performances of optical components in terms of laser-resistance. In laser induced damage investigations, the beam diameter is known to play an important role in the damage probability [3,4]. In this present work, the impact of three beam waists (15μm, 50 μm and 170 μm) has been tested on a series of six monolayers deposited on superpolished fused silica substrates. The set of components is composed of three pure materials (HfO2, Al2O3 and SiO2) and of their binary mixtures. All components have been tested in the nanosecond regime (pulse duration of 8ns), in multiple pulse mode (S-on-1 testing) at 355nm and with a repetition rate of 50Hz. The installed setup includes a dynamic in-situ damage detection using a fast camera, which allows to study the so-called fatigue effects of the tested layers. For data reduction, a thermal model with nanometric inclusions was assumed leading to a precursor density versus fluence relation that follows a two-parameter power-law [5]. Adjustment of these parameters provides a fit of the damage probability (Fig. 1). The resulting precursor density and the damage threshold fluence are drawn from the fit and can be compared from one component to the next.