Janis Zideluns, PhD

Accueil › Animation Scientifique › Janis Zideluns, PhD

Janis Zideluns soutiendra sa thèse intitulée "Optical monitoring methods and strategies for magnetron sputtered thin-film filters" le Mercredi 28 Septembre à 10h00 en salle Pierre Cotton de l’Institut Fresnel.

Cette thèse a été co-encadrée par by Julien LUMEAU et Fabien LEMARCHAND de l’équipe RCMO.

Résumé :
In this thesis, we describe the optical monitoring methods and strategies for magnetron sputtered thin-film optical filters. As with any technologies, the requirements for the performance of the thin-film filters increases, the number of layers in the designs grows and therefore the accurate monitoring of the deposited thicknesses becomes more and more crucial. Therefore, the methods of the thickness control have evolved as well, and, as of today, optical monitoring methods are the most precise ones. The quality of the filter, however, does not rely on the method that is used for the thickness monitoring, but rather on how the method is used. The how is what we call the monitoring strategy. To create a successful monitoring strategy, deep knowledge about the monitoring method and the thin film filter itself is required. In this thesis we investigate the three established optical thin film monitoring methods.
Turning point optical monitoring is arguably the first commercially available optical monitoring method. Although it is the least accurate optical monitoring method when it comes to thickness errors in monitored layers, it is still widely used, especially when it comes to bandpass filter deposition as it benefits from strong error self-compensation. The use of this method, however, is limited to so called quarter-wave designs and good results are usually obtained in limited wavelength range.
Monochromatic monitoring is probably the most popular monitoring method as of today, as it can be used to monitor almost any design with transparent layers. The difficulty of this method is the determination of the monitoring strategy, as the monitoring wavelength(s) has to selected for each design. The error self-compensation is less effective compared to the turning point monitoring, however good spectral match between experiment and theory is expected over wide wavelength range, if the strategy is chosen wisely.
Broadband optical monitoring is lately gaining popularity, similarly to monochromatic monitoring it can be used for various types of designs. Although this method is not associated with error self-compensation and the spectral resolution of the broadband systems is inferior to monochromatic systems, filters with very low thickness errors can be produced, what is crucial if indirect monitoring strategy is selected.
Each of these monitoring methods have their technical limits (for example spectral resolution and signal to noise ratio) that are either fixed or adjustable. The designs, that we want to deposit, can have spectral regions that are not suited for optical monitoring (for example the transmittance can be too low for measurement) that can vary from layer to layer. To create the
monitoring strategy, a good balance between the technical possibilities of the motoring system and the spectral behavior of the layer to be monitored must be found.

 Pour les personnes souhaitant suivre la thèse en visio, un lien ZOOM sera communiqué très prochainement