AVS 61st International Symposium & Exhibition
    Plasma Science and Technology Tuesday Sessions
       Session PS-TuM

Paper PS-TuM5
Study of Plasma-Surface Interaction in HBr/Cl2/O2 ICP

Tuesday, November 11, 2014, 9:20 am, Room 308

Session: Plasma Surface Interactions I
Presenter: Ashutosh Srivastava, University of Houston
Authors: A.K. Srivastava, University of Houston
T. Ohashi, Hitachi High-Technologies
V.M. Donnelly, University of Houston
Correspondent: Click to Email

HBr plasmas are used to etch Si, as well as GaN, PZT, InP, Indium Zinc Oxide and other materials. In Si etching, HBr plasmas create better anisotropic profiles than Cl2 plasmas, with better selectivity toward SiO2. Selectivity can be further improved by adding oxygen to the plasma. The feed gas composition of HBr/Cl2/O2 plasmas is optimized to best meet the needs of the particular application. Keeping such a complex process stable over time requires tight control over all plasma parameters, including reactor wall conditions. Here, we have studied the interaction of HBr/Cl2/O2 inductively-coupled plasmas (ICP) with the etching products-coated reactor walls, using the “Spinning Wall” technique. Surface composition was determined by Auger electron spectroscopy, and species desorbing from the surface were detected by quadrupole mass spectrometry. It was observed that in a pure HBr ICP with no Si etching, an oxygen plasma-conditioned surface SiOx layer remained unaltered, indicating the inability of Br-containing species to break the Si-O bond. When 5-50% oxygen was added to HBr ICPs, large fluxes of H2O were found to desorb from the spinning wall surface, peaking at ~20% O2. The H2O signal dependence on spinning wall rotation frequency indicates that the reaction time, and/or water retention time on the surface is of the order of 10’s of ms. After turning off the plasma, signal could be detected for several minutes. This is likely because of the highly disorder surface, leading to a stretched exponential decay of water desorption. Addition of Cl2 to an HBr ICP (1:1 ratio) resulted in small amount (<3%) of Cl on the surface but no Br. This Cl was not removed by a pure HBr ICP, but was in an O2 ICP. When a self-bias of -120Vdc was applied to the Si substrate in an HBr ICP, a Si:O:Br layer (32:7:60) was deposited on the spinning wall surface. As O2 was added, the surface composition changed from a Br-rich layer at 6% O2 to an O-rich layer at 20% O2. Si etching in a Cl:HBr ICP (1:1) resulted in equal amounts of Cl and Br on the surface. Again, when O2 is added the surface transitions from Cl-rich at 6% O2 to Cl/O-rich at 20% O2 and finally to O-rich at 40% O2.