On the lifetime of the first mirrors in the diagnostic systems of the international thermonuclear experimental reactor

Plasma diagnostic systems will be necessary tools for the future success of the International
Thermonuclear Experimental Reactor (ITER) both to better understand the physics involved in
magnetically confined burning plasma and for the protection of the device in case of disruptions
etc. In contrast to conditions in today’s tokamaks, a high level of radiation and neutrons is
expected in ITER. To reduce the extent of the possible neutron leakage and to protect the
optical components (windows, fibres) from the radiations it was proposed that the light of the
plasma should be transmitted by mirror optics to diagnostics through a labyrinth embedded in
shielding material.
The first elements of the plasma diagnostic systems in ITER will therefore be metallic
mirrors called “First Mirrors”. Being the closest element to the hot confined plasma they will
suffer from intense radiation, from bombardment by energetic particles and possible deposition
of impurities eroded from the plasma-facing components. They will have to maintain the
required optical properties despite these extreme conditions. The question of the lifetime of
these first mirrors (i.e. how long will they maintain their optical properties) is thus of the
highest importance because any change in the reflectivity of the first mirror will affect the
reliability of the spectroscopic or laser signal and thus the reliability of the diagnostic system.
The objective of this thesis was to improve the understanding of the effects of the plasmawall
interactions in a tokamak (material erosion, migration, and redeposition) on the optical
properties of in-vessel metallic mirrors, with a view towards the parameters which may be
optimized to extend their lifetime. This was achieved by a thorough participation in the mirror
experiments carried out in several tokamaks (Tore Supra, TCV, TEXTOR, DIII-D and JET)
through detailed optical and surface characterizations of the mirrors before and after their
exposure, and by dedicated laboratory experiments.
Several important conclusions have been derived from these experiments. Under erosion
conditions, the progressive increase in the surface roughness (due to the appearance of the
crystallographic grains) results in a progressive decrease of the mirror reflectivity. To prevent
such effects, the mirror crystallography should ensure that the roughness will always remain
negligible in comparison with the wavelength of the light. This may be achieved in two different
ways: either by using single crystal mirrors or by manufacturing the mirror in the form of a
nanocrystalline coating (Rh or Mo) on a polished metallic substrate.
For mirrors located in deposition dominated areas (in the divertor for example), deposition
of impurities on the mirror surface will lead to drastic changes of the reflectivity. The carbon
deposition rate observed on the mirrors exposed in the DIII-D (∼2 nm·s−1) divertor gives an
idea of the extent of the problem. Carbon is expected to be the main impurity deposited on
mirrors located in areas remote from the plasma due to its long range migration. Mitigation
of the carbon deposition has been achieved by heating the mirrors to about 200◦C. This is a
very promising result because such temperature is relatively moderate. However for mirrors in
direct line-of-sight from the plasma, deposition of beryllium should also be taken into account.
According to the experiments made in the PISCES-B linear device, deposition of Be containing
layer will significantly modify the mirror reflectivity. Moreover, even if deposition of carbon
can be mitigated during the co-deposition of Be and C, deposition of Be is not affected by the
mirror temperature. The possible diffusion of beryllium in the mirror material (enhanced by
the temperature) may complicate the possible in-situ cleaning of the mirrors.
Results from experiments made in Tore Supra, TCV, and in a laboratory stand in Basel
have shown that the substrate material plays a role in determining the importance of the
erosion/deposition mechanisms affecting the mirror reflectivity. In TCV, the deposition rate
of carbon was found to be lower on a high-Z material (Mo) than on a low-Z material (Si).
This may be explained by enhanced re-sputtering of the deposited carbon due to a higher
particle reflection coefficient on Mo. From experiments made in Tore Supra and in Basel, it
was observed that under simultaneous bombardment with deuterium and carbon, the sputtering
of copper mirrors was enhanced by the presence of carbon. This leads to an anomalous effective
sputtering yield for copper. Numerical simulations with the Monte Carlo code TRIDYN have
shed some light on the results observed for molybdenum, stainless steel and silicon samples.
However, such an approach failed to reproduce the phenomena observed for copper. It seems
therefore quite likely that the chemistry of copper towards carbon plays a role in the observed
increase erosion when carbon is present in the plasma.