Effects of wall contamination on consecutive plasma processes

Abstract

Effect of substrate on the step coverage of plasma-enhanced chemical-vapor deposited tetraethylorthosilicate films J.Low temperature diamond growth using CO 2 /CH 4 plasmas: Molecular beam mass spectrometry and computer simulation investigations Plasma processes often go beyond the primary objectives focused on the substrate, or targeted materials. For instance, sputtered materials deposit on surfaces other than the substrate, and plasma deposition extends to the walls of the reactor. In the process of plasma polymerization, or plasma chemical vapor deposition ͑PCVD͒, every surface ͑not just the substrate surface͒ participates in the overall plasma deposition process. Consequently the chemical and physical natures of all surfaces within a reactor are very important factors that determine the fate of the PCVD process. The materials deposited on the wall surface ͑wall contaminants͒ are created in the previous run in a batch operation of PCVD. In a sequential plasma process, where plasma polymerization of trimethylsilane ͑TMS͒ was followed by plasma polymerization of hexafluoroethane ͑HFE͒, F-containing oligomers ͑low molecular weight compounds͒, created during the plasma polymerization of HFE in the previous run, remain on surfaces in the reactor. The wall contaminants were found to migrate to the new substrate ͑aluminum alloy͒ surface in the subsequent run upon the evacuation of the reactor. If an O 2 plasma treatment is applied, F-containing organic compounds chemisorbed on the new substrate surface are converted to F-containing inorganic compounds, which decreases the plasma-ablatable F on the surface. If no O 2 plasma treatment is applied, the F-containing organic compounds are exposed to the environment of the TMS plasma. From the viewpoint of the sequence of plasma processes, a new HFE/TMS sequence is created without the O 2 plasma treatment. The HFE/TMS system ͑reversed order to the normal cycle͒ causes adhesion failure at the interface between the plasma polymers and the aluminum alloys, whereas the TMS/HFE system yields good adhesion of plasma deposited layers to the substrate and provides superior adhesion of a primer applied on the plasma polymer coating. This difference was created by the difference in handling of the wall contaminants.

Influence of contaminations on the performance of thin-film silicon solar cells prepared after in situ reactor plasma cleaning

“…The fluorine atoms could be detached by the interaction with Si species produced in the plasma. Stable species with Si-F bonds are formed, which will either participate in the deposition process or be pumped out [3]. We also cannot exclude the contamination of chamber walls by F-containing oligomers which are mobile enough [3].…”

“…One can see that Ar purging and a long period of pumping are inefficient in reducing the CME. Also ineffective are O 2 and Ar plasma treatment despite of the fact that oxygen plasma treatments have been reported to reduce the fluorine concentration on the reactor walls [3]. The hydrogen treatment significantly reduced the fill factor drop observed for the "first" cell.…”

“…Stable species with Si-F bonds are formed, which will either participate in the deposition process or be pumped out [3]. We also cannot exclude the contamination of chamber walls by F-containing oligomers which are mobile enough [3]. There is a high probability that fluorine etches contaminants from the walls and that the resulting compounds migrate to the Si deposition zone ("re-etching" effect, see Section 3.2) [14].…”

“…Unfortunately, the etching reaction products often remain in the reactor and gas out continuously, thus leading to their incorporation into the silicon films prepared subsequently [1][2][3][4][5].…”