Hideaki Ninomiya

Creation and passivation of electron traps in n-InP treated with hydrogen plasma

Studies on deep electron traps in n‐InP treated with hydrogen plasma have been performed using isothermal capacitance transient spectroscopy measurements. Five electron trap levels, E1–E5, with activation energies of 0.21, 0.51, 0.32, 0.54, and 0.63 eV, respectively, below the conduction band are detected. Only E1 traps are observed in as‐etched InP. Hydrogen‐plasma treatment leads to enhancement of the density of E1 traps and creation of E2 and E5 traps. E1 and E5 traps are annealed out at 300 °C, while E2 traps are annealed out at a temperature as low as 150 °C. On the other hand, densities of E3 and E4 traps are significantly enhanced by annealing at 350 °C. This experimental result suggests that the E3 and E4 traps are generated during hydrogen‐plasma exposure and are passivated by hydrogen atoms.

Effect of Hydrogen Plasma Treatment on n-InP Surfaces.

Surfaces of n-InP treated with remote hydrogen plasma have been analyzed in terms of X-ray photoelectron spectroscopy (XPS), Kelvin probe, current-voltage characteristics of Schottky barrier junctions and isothermal capacitance transient spectroscopy (ICTS). It is confirmed by XPS analysis that the native oxide is removed from the InP surface by the \H2-plasma treatment. Schottky junctions formed by in situ evaporation of various metals immediately after the remote \H2-plasma exposure show that the barrier height is pinned at about 0.5 eV, irrespective of Schottky metal. This value is somewhat higher than the barrier height of 0.4 eV for untreated surfaces. It is observed by Kelvin probe measurement that the Fermi level shifts to an energy around 0.53 eV below the conduction band edge upon \H2-plasma treatment from 0.39 eV for an untreated surface. Furthermore, a deep trap level with the activation energy of 0.51 eV below the conduction band edge was detected for samples treated with \H2 plasma by ICTS measurement. The pinned behavior of the Schottky barrier height is speculated to be related to the trap level generated by \H2-plasma treatment.

Effect of Phosphine on Plasma-Induced Traps in n-InP

Electron traps in n-InP generated by exposure to Ar or phosphine (PH3) plasma have been investigated using isothermal capacitance transient spectroscopy (ICTS). One electron trap (Ec-0.54 eV) is generated by Ar-plasma treatment at 250°C for 60 min. Annealing at 350°C for 3 min after Ar-plasma treatment induces another electron trap (Ec-0.32 eV) together with an increase of the (Ec-0.54 eV) trap density. In contrast to the case of Ar plasma, no traps are detected in InP treated with PH3 plasma consisting of Ar(90%) and PH3(10%) at 250°C for 60 min. Moreover, addition of PH3 to hydrogen (H2) plasma is shown to be effective in suppressing generation of the traps. Generation of these traps does not occur due to annealing after PH3-plasma treatment, while simple annealing of as-etched InP introduces these traps. It is demonstrated that PH3-plasma treatment leads to diffusion of phosphorus atoms during the process and deposition of a thin phosphorus layer at the surface of InP.

Measurement of Surface Fermi Level in Phosphidized GaAs

The location of the surface Fermi level in phosphidized GaAs has been investigated using the Kelvin probe method. The surface Fermi level changes from an energy around 0.47 eV below the conduction band edge for the as-etched surface covered with a natural oxide layer to an energy of 0.25 eV after sufficient phosphidization by phosphine (PH3)-plasma treatment. On the other hand, isothermal capacitance transient spectroscopy measurement reveals that a new electron trap with an activation energy of 0.24 eV is introduced at or near the phosphidized surface. This provides experimental evidence that the location of the surface Fermi level is closely associated with that of the electron trap induced by PH3-plasma treatment.

Characterization of GaAs surfaces treated with phosphine gas photodecomposed by an ArF excimer laser

Phosphidization of GaAs surfaces is attempted with phosphine gas photodecomposed by an ArF excimer laser. Electron traps at and near the phosphidized GaAs surfaces are characterized by isothermal capacitance transient spectroscopy measurements. Phosphidization leads to a reduction in the trap (Ec−0.81 eV) known as an EL2 center and generation of two traps (Ec−0.24 eV and Ec−0.49 eV), which are designated as NL1 and NL2, respectively. A significant metal work-function dependence of the barrier height is demonstrated for Schottky junctions formed on the GaAs surfaces phosphidized under optimum condition, suggesting that phosphidization is effective in reducing surface states of GaAs.

Characteristics of Au/ n -InP Schottky junctions formed on H 2 - and PH 3 -plasma treated surfaces

Characteristics of An/n-InP Schottky junctions formed onn-InP treated with hydrogen (H2)-and phosphine (PH3)-plasmas have been investigated. An enhancement of the barrier height up to 0.7 eV or more is observed for Schottky junctions processed sequentially with plasma treatment, laboratory air oxidation and Au evaporation. From the measurement of Schottky junctions formed by in-situ metallization immediately after H2-plasma treatment, it is found that laboratory air oxidation permits an increase in the barrier height by about 0.1 eV. The annealing experiment of Schottky junctions treated with plasma reveals that the substantial part of the barrier height enhancement is caused by release of the Fermi level pinning due to hydrogen passivation of surface defects. Although both H2- and PH3-plasmas are effective in enhancing the barrier height, PH3-plasma is preferable in respect to minimizing plasma-induced damage. In the case of H2-plasma treatment deep electron traps with activation energies of 0.21 and 0.51 eV below the conduction band are generated at and/or near the surface of InP, while these traps are not detected after PH3-plasma treatment.

Effects of Phosphine-Plasma Treatment on Characteristics of Au/n-InP Schottky Junctions

The effects of phosphine (PH3)-plasma treatment on the characteristics of Au/n-InP Schottky junctions are investigated and compared with those of hydrogen (H2)-plasma treatment. An enhancement of the barrier height of as high as 0.7 eV or more is found for Schottky junctions fabricated by the process consisting of plasma treatment, oxidation and Au evaporation. In the case of PH3-plasma treatment, no degradation of the ideality factor occurs and change in the barrier height is suppressed even after annealing at temperatures as high as 350°C. Formation of Schottky junctions is attempted by using an in situ process of H2-plasma treatment and metallization in order to reveal the effect of oxidation on variation in the Schottky barrier height. It is demonstrated that the enhancement of barrier height is due to the combined effects of H passivation of the surface defects and surface oxidation.

Characteristics of Electron Trap Induced in n-InP by Hydrogen Plasma Exposure

Behaviors of the electron trap at E c-0.51 eV in n-InP treated with hydrogen plasma have been investigated by isothermal capacitance transient spectroscopy. It is found that considerable enhancement of the trap density occurs when a reverse bias voltage of only -1 V is applied to Au Schottky junctions on n-InP treated with hydrogen plasma. This implies that the enhancement of the electron trap is due to a dissociation of plasma-induced defect-hydrogen complexes into electrically active defects and hydrogen. An isochronal annealing experiment shows that the annealing procoss follows the first-order reaction kinetics and its activation energy is 1.4 eV.

Deep electron traps in n-InP induced by plasma exposure

Deep electron traps in n-InP introduced during helium (He)- or hydrogen ( H2)-plasma exposure have been studied by means of isothermal capacitance transient spectroscopy (ICTS). Two electron traps, (E c-0.51 eV) and (E c-0.54 eV), which are designated E2 and E4, respectively, are detected at and near the surface treated with He plasma. These traps induced by He-plasma exposure can be passivated with hydrogen. When the samples are treated with H2 plasma, E2 traps are only partly active and E4 traps are not detected due to being totally passivated with hydrogen. The density of E2 traps near the sample surface treated with H2 plasma is strongly enhanced by applying reverse bias at room temperature because of dissociation and removal of passivating hydrogen. In contrast, hydrogen-passivated E4 traps become reactivated only by thermal annealing. An isochronal annealing experiment for the He-plasma-treated samples shows the first-order annealing process of E2 traps with the activation energy and the attempt-to-escape frequency of 1.5 eV and 3.2 ×1014 s-1, respectively. The thermal dissociation process of hydrogen from E4 traps follows first-order kinetics and its dissociation energy and attempt-to-escape frequency are 1.65 eV and 4.9 ×1013 s-1, respectively.

Characterization of plasma-induced traps in n-InP

Deep electron trap levels in n-InP treated with remote hydrogen and helium plasmas have been studied by isothermal capacitance transient spectroscopy (ICTS) measurement. The trap levels at E/sub c/-0.51 eV (E2) and E/sub c/-0.54 eV (E4) are induced by both plasma treatment. However, the E4 traps are related to phosphorus vacancies and are fully passivated in H/sub 2/ plasma treated sample. The existence of the E4 traps in H/sub 2/-plasma-treated sample can be revealed after 350/spl deg/C annealing. It is found that the width of the surface region in which the E2 traps are dominantly introduced becomes thicker with increasing acceleration voltage of He ions.<<ETX>>