AVS 50th International Symposium
    Plasma Science and Technology Monday Sessions
       Session PS-MoM

Paper PS-MoM3
Chemical Topography Analyses of Photoresist Patterns Exposed to HBr/O2 and Cl2/O2 Trimming Plasma Processes

Monday, November 3, 2003, 9:00 am, Room 315

Session: Critical Dimension Etching
Presenter: E. Pargon, CNRS/LTM, France
Authors: E. Pargon, CNRS/LTM, France
O. Joubert, CNRS/LTM, France
L. Vallier, CNRS/LTM, France
S. Xu, Applied Materials
Correspondent: Click to Email
Nowadays, a way to bypass the lithography limitation in typical gate etch processes is to introduce a step of "resist trimming" prior all the other classical etch steps. Resist trimming induces a lateral erosion of the photoresist mask to reach a range of dimension smaller than the resolution of the lithography. To better understand the mechanisms involved in this process, an experimental procedure based on XPS has been developed to determine the chemical composition and thickness of the reactive layers formed both on top and sidewalls of the resist features during the process. The processes are performed in a high density plasma source (ICP) and two trim chemistries are investigated: HBr/O@sub2@ and Cl@sub2@/O@sub2@. The XPS analyses show that the transformations occurring on the resist sidewalls can well explain the faster trim rate obtained with a HBr/O@sub2@ chemistry. Indeed, the XPS results reveal that HBr/O@sub2@ is a very reactive chemistry leading to the formation of very thin (0.5 @nm@) reactive layers on the resist sidewalls, while when using Cl@sub2@/O@sub2@, there is a competition for the adsorption sites between atomic Chlorine and Oxygen leading in this case to thick (1.5 @nm@) carbon rich chlorine reactive layers on the resist sidewalls. Other plasma parameters (pressure, bias and source powers) have also been studied and we have obtained good correlations between the trim rate and the modifications measured by XPS on the resist sidewalls. In most cases, with the Cl@sub2@/O@sub2@ chemistry, a decrease in trim rate is well correlated with an increase in reactive layer thickness on the resist sidewalls and with a decrease of the O/Cl ratio in the reactive layer. Finally, this XPS experimental procedure enables us to better understand the mechanisms involved in resist triming processes and to determine the key plasma parameters that drive such processes.