AVS 62nd International Symposium & Exhibition
    Electronic Materials and Processing Monday Sessions
       Session EM+NS+PS-MoM

Paper EM+NS+PS-MoM2
Selective Wet Etching of III-V Semiconductors with HCl and H2O2

Monday, October 19, 2015, 8:40 am, Room 210E

Session: More Moore! Materials and Processes to Extend CMOS Another Decade
Presenter: Pablo Mancheno-Posso, University of Arizona
Authors: L. Mancheno-Posso, University of Arizona
. Jain, University of Arizona
A.J. Muscat, University of Arizona
Correspondent: Click to Email
The etching of III-V semiconductors is needed to insert these materials into current device flows to extend CMOS transistor technology. III-V oxides are detrimental to electrical performance and must be removed, because they adopt different oxidation states and can be soluble in water. Plasma etching to create profiles can damage and change the stoichiometry of the surface. Wet etching of these oxides can control the roughness and chemical termination of the surface by choice of oxidant and etchant, concentration, and pH. Wet etching of III-V semiconductors is accomplished by oxidizing acid and base chemistries that can preferentially remove group III or V atoms. In new 3 D transistor architectures, the formation of the channel fin requires a low etching rate to ensure a smooth surface and a highly selective etching bath with respect to other materials or crystal faces that are exposed. In this work, we varied the group III and V atoms across five binaries (GaAs, InAs, InP, GaSb, and InSb) and measured etching rates. These materials were etched using mixtures of HCl (0.01 M) and H2O2 (0.0001-5 M). The etching rate was measured using profilometry on wafers patterned with conventional photolithography. The chemical composition was monitored using X-ray photoelectron spectroscopy (XPS). The etching rate of GaAs and InAs (same group V atom) exhibited a volcano-shaped dependence on H2O2 concentration. At H2O2 concentrations of 5 to 100 mM, the etching rate increased linearly from 0.08±0.03 to 1.1±0.1 nm/s for GaAs and from 0.06±0.04 to 0.9±0.3 nm/s for InAs. The rate decreased to 0.04±0.01 nm/s for GaAs and 0.26±0.13 nm/s for InAs at 1 M H2O2. InP, which is often exposed during etching of another III-V, showed a linear dependence on H2O2 concentration (0.01 to 5 M), increasing from 0.003±0.001 to 0.012±0.009 nm/s. The selectivity of etching GaAs to InP at three points along the volcano was about 55, 140, and 4 at H2O2 concentrations of 0.01, 0.1, and 1 M. Like the arsenides, the antimonides etched at about the same rate, but the volcano dependence moved to lower peroxide concentrations. The etching rate of GaSb increased from 0.07±0.04 to 0.21±0.04 nm/s and InSb from 0.09±0.03 to 0.38±0.09 nm/s for H2O2 concentrations from 0.1 to 1 mM. The group V atom determined the etching rate and is involved in the rate determining step in the reaction. The presence of As-Cl bonds on the surface after etching GaAs in HCl was confirmed by temperature programmed desorption (TPD) experiments after immersion in 1.7 M HCl. The mechanism for etching III-V semiconductors will be discussed based on the etching rate data and chemical composition of the surface.