Removal of the codeposited carbon layer using He–O glow discharge

Abstract

In this study we examine the combination of a He–O glow discharge with heating as a possible technique to remove deuterium from TFTR tiles. Samples were cut from a relatively large area containing a uniform codeposited layer of deuterium and carbon. Auger/SEM was used to generate micrographs of each of the samples. The samples were also examined using Rutherford backscattering to determine the near surface composition. Individual samples were then exposed to a He–O glow discharge while being heated. After the exposure, the samples were returned for Auger/SEM and RBS of the same areas examined prior to the exposure. Comparing the samples before and after exposure revealed that the amount of the codeposited layer removed was significantly less than 1 μm. Removal rates this low would suggest that He–O glow discharge with heating is insufficient to remove the thick layers predicted for ITER in a timely fashion.

Introduction

The problem of tritium codeposition with carbon to yield excessive tritium inventories presents the ITER fusion reactor with one of its greatest challenges. To mitigate the destructive nature of disruptions and Type 1 ELMs on divertor materials, graphite or carbon composites must be placed in the bottom of the tokamak divertor. With carbon present, nearby surfaces where the rate of deposition is greater than the erosion rate will have codeposited layers of carbon and hydrogen isotopes growing indefinitely. Since ITER will be operating with a mixture of tritium and deuterium, the codeposited layer could potentially constrain the operation of ITER by retaining quantities of tritium approaching the safety limit. If we combine the high tritium inventory with the fact that the a-C:H codeposited layer is not stable at elevated temperatures in the presence of air [1], we have a potential environmental hazard in the event of an accidental vacuum loss when the tokamak vessel is hot.

Several techniques have been examined as ways of removing the codeposited layer. These techniques include heating in air or oxygen [1], [2], [3], [4], [5], [6], laser heating [7], [8], flash lamps [9], and He–O glow discharge [10]. While each of these techniques has shown some success in reducing the quantity of tritium in the codeposited layer, no technique has been identified as the solution to the problem. As an example, heating in air will completely remove the codeposited layer if the temperature is sufficiently high. Unfortunately, ITER will be limited to an upper bake out temperature of approximately 500 K. At this temperature, only about 30% to 50% of the film can be removed during a several hour bake in air or oxygen.

He–O glow discharge at room temperature has been tried in the past as a removal technique for the codeposited layer. Hsu [10] compared the glow discharge removal rate of a codeposited layer using several different types of gases. Of nitrogen, hydrogen, helium, and oxygen, only oxygen (in the form of He–O) was found to have a measurable removal rate. Hsu determined an effective removal rate of approximately 5 atoms of carbon for each oxygen ion striking the layer. While the film was produced by the plasma decomposition of methane, and was therefore a ‘soft film’ with a significant fraction of weakly bound atoms, this result certainly suggested that He–O might present a reasonable removal process for the codeposited layer. In unpublished experiments [11], Cowgill used a He–O discharge to remove a codeposited layer from a tile taken from the TFTR reactor. These experiments demonstrated a removal rate of ∼1.2 μm/r (about 2.5 carbon atom/O ion) during the initial stage of the experiment, but noted that the removal rate decreased with time. The decrease was attributed to surface texturing. He–O glow discharge was also used directly in the TFTR reactor. For the same experiments, Skinner [12] reported the process to release 50 Ci/h and to be constant with time. This value should be compared to an initial removal rate of 170 Ci/h for deuterium glow discharge, but a steady state release rate of only 10 Ci/h. In a somewhat related series of experiments, Jacob et al. [13] performed a systematic study of the removal of a-C:H layers using electron cyclotron resonance discharges (ECR). Several different species were used for the ECR low-pressure discharges (O₂, D₂, H₂, H₂O, and O₂/H₂), but oxygen always showed the highest removal rates. They noted a codeposited removal rate as high as 1.7 μm/h at 300 K. The authors noted increased yield with increased voltage or temperature, but found the two were not additive.

The essential difference between most the earlier studies and that reported here is that heating during the discharge has been added. An area of net deposition on a graphite tile removed from TFTR prior to the DT campaign was used to provide a relevant codeposited layer. Small samples cut from these tiles were examined prior to and after exposure to a He–O at temperatures from 373 K up to 513 K for 1–4 h. Changes in the layer thickness and near surface deuterium content were measured.