Interpretation of optical emission in a strongly inhomogeneous PIAD environment

Abstract

A variety of methods exists for the production of high quality optical coatings. These are for instance magnetron or ion beam sputtering, or thermal evaporation assisted by ion or plasma ion beam sources. The choice of a particular technique is due to the specific demands for tailoring the thin film properties such as homogeneity, refractive index, absorption, mechanical stress, porosity (optical shift), etc.. To allow for economic production of complex multilayer designs, the issue of reproducibility at the highest deposition rate possible has to be faced.

The plasma ion assisted deposition (PIAD) has been invented to avoid contamination of the process environment present when gridded ion sources are employed. One example for this approach is the Advanced Plasma Source (APS) which holds a considerable market share in the field of optical coatings. The APS is a hot cathode (LaB6) DC glow discharge with an auxiliary magnetic field, typically operated with argon. A high density (ne ~ 1012 cm-3) and high temperature (Te ~ 20eV) plasma is generated in the source region (V ~ 0.7l) and expands to the chamber (V ~ 103 l) which is held at high vacuum (p ~ 2 • 10-2Pa). The expansion induces a strong drop of the plasma potential Vp which results in an acceleration of the ions towards the substrates. The setup is outlined in figure 1. By varying the discharge voltage (VD=50..150 V), the magnetic field (Bmax=10..40mT) and the gas flux (GAr=2..20 sccm) different characterstics of the plasma ion beam are obtained, where typical ion energies are Ei=50..150 eV.

As is described in the references, various probe techniques were adopted to elucidate the mechanisms of plasma beam formation in this particular PIAD setup. In an earlier work the approach of optical emission spectroscopy (OES) has been pursued. This paper contains a description of the diagnostic installation allowing for tomographic reconstruction of the local optical emission near the source exit. With a simple corona model ne and Te could be estimated for an Ar/He mixture. The interpretation of emission close to the APS is hampered by the lack of detailed knowledge of neutral density and temperature in this region.

In this extended abstract the preparation of a more elaborate approach using collisional radiative modelling is sketched. The new aspect is the consideration of a global electron energy probability function (EEPF) based on the nonlocal approximation. This concept is useful for reducing the complexity of electron kinetics by coordinate transformation from the 6D phase-space to a 1D total energy space. The analysis of OES data is not merely ment as a proof of principle. The final goal is to interpret the OES data in terms of electron parameters, as a foundation of a control scheme for the APS using non-invasive optical diagnostics. Although global parameters, such as voltages and currents can be maintained accurately, drifts in the plasma parameters which are not measured routinely in the industrial application, may lead to limitations in the reproducibility of the optical coatings. The main reason is suspected to be the alteration of the electrodes of the APS during the PIAD process.