Etching of InP surface oxide with atomic hydrogen produced by electron cyclotron resonance

Abstract

The plasma output of an electron cyclotron resonance source operating with hydrogen has been characterized using optical emission spectroscopy to observe neutral excited atomic hydrogen H* and a Langmuir probe to observe H+ as a function of currents to the two electromagnets producing the magnetic field. The rate at which the hydrogen plasma etches the surface oxide on InP substrates has been measured as a function of substrate temperature over the range 250–485 °C for two operating conditions, one that maximizes H* and the other for which a combination of high H+ and minimal H* occurs. The results indicate that the plasma stream with maximum H* etches approximately five times faster than the plasma stream with high H+ and low H* content. However, the activation energy of the etching reaction, Ea∼0.45 eV, is the same for the two plasma conditions. Considering the oxide as InPO4, produced by exposing the substrates to ozone generated by ultraviolet light, the proposed reaction for the etching is InPO4+7.3H→2H2O+0.5PH2 +0.5PH3+0.8InH+0.2In.

Ohmic Contact of Cadmium Oxide, a Transparent Conducting Oxide, to n-type Indium Phosphide

“…If the Schottky behavior is caused by the interfacial oxide layer, the removal or sufficient reduction of the oxide should permit ohmic contact (Figure c); when the barrier thickness W is thin enough, electrons can tunnel through the barrier giving ohmic contact. Frequently, oxide-layers can be reduced by an H 2 plasma. − Both CdO and CdO:Sn films were deposited on H 2 plasma cleaned InP and contacts formed. The current−voltage curves for contacts with a 60 μm gap distance are shown in panel (c) of Figure ; it is easily seen that samples with H 2 plasma pretreatment achieve ohmic contact.…”

Evaluation of H-plasma and thermal oxide desorption procedures for MBE regrowth of an InP/InGaAs/InP quantum well

“…These defects could be native point defects (vacancies, interstitials or antisites) or their complexes created at the InP surface or in a few monolayers below the surface. The fact that the plasma desorption procedure creates more of these defects than the thermal procedure is likely due to the fundamentally different chemistry between the two oxide etching reactions [7,12]. Additional insights to the identification of the defects introduced by the regrowth procedure is provided by the results of a hybrid regrowth procedure (figure 4(e)) in which the sample was cleaned with a 5 min H-plasma exposure followed by a 10 min anneal at 520 • C. This was done in order to study the influence of sample temperature during the oxide desorption on the defects evidenced by the PL line at 15 meV below the intrinsic quantum well PL line.…”

“…In the case of InP, thermal desorption involves controlled ramping of the temperature to above 500 • C [2][3][4], considerably higher than the congruent sublimation temperature of 360 • C. The use of an ECR H-plasma to remove the oxide on InP has been shown to produce interfaces as good as thermal desorption in an InP(p)/InP(p) regrowth structure, but at a somewhat lower treatment temperature [5,6]. The effects of the PH 3 /H 2 flux ratio and various plasma modes have also been investigated, the activation energy of the oxide etching reaction has been established and a model for the oxide removal has been proposed [6,7].…”