Impact of Plasma Electron Flux on Plasma Damage‐Free Sputtering of Ultrathin Tin‐Doped Indium Oxide Contact Layer on p‐GaN for InGaN/GaN Light‐Emitting Diodes

Abstract The origin of plasma‐induced damage on a p‐type wide‐bandgap layer during the sputtering of tin‐doped indium oxide (ITO) contact layers by using radiofrequency‐superimposed direct current (DC) sputtering and its effects on the forward voltage and light output power (LOP) of light‐emitting diodes (LEDs) with sputtered ITO transparent conductive electrodes (TCE) is systematically studied. Changing the DC power voltage from negative to positive bias reduces the forward voltages and enhances the LOP of the LEDs. The positive DC power drastically decreases the electron flux in the plasma obtained by plasma diagnostics using a cutoff probe and a Langmuir probe, suggesting that the repulsion of plasma electrons from the p‐GaN surface can reduce plasma‐induced damage to the p‐GaN. Furthermore, electron‐beam irradiation on p‐GaN prior to ITO deposition significantly increases the forward voltages, showing that the plasma electrons play an important role in plasma‐induced damage to the p‐GaN. The plasma electrons can increase the effective barrier height at the ITO/deep‐level defect (DLD) band of p‐GaN by compensating DLDs, resulting in the deterioration of the forward voltage and LOP. Finally, the plasma damage‐free sputtered‐ITO TCE enhances the LOP of the LEDs by 20% with a low forward voltage of 2.9 V at 20 mA compared to LEDs with conventional e‐beam‐evaporated ITO TCE.

. Measured plasma discharge parameters from plasma diagnosis using cutoff probe and Langmuir probe. In this study, as shown in Figure S1(a), RF-superimposed DC (RF+DC) sputtering was used for the direct sputtering of the ITO layer on p-GaN in the LED structures, which can significantly change the distribution of plasma potential according to the DC power, as schematically shown in Figure S1(b). In the RF superimposed DC sputter system used in this study, RF power was matched with the Tune variable capacitor (VC) and Load VC, and it was superimposed on the DC power. Then, the RF superimposed DC power was applied to the electrode of ITO target. The inductor and capacitor in the RF Block Filter was designed to prevent the leakage of RF power.
For characterizing the plasma potential distribution, plasma discharge parameters were measured using a cutoff probe and a Langmuir probe, as summarized in Table S1.
Furthermore, based upon the measurement of the plasma discharge parameters, a schematic illustration of the distribution of plasma potential was shown in Figure S1(b).
As the DC power in the RF-superimposed DC sputtering of ITO is applied from negative to positive power, plasma potential near to ITO target increased rapidly, which accelerate plasma electrons having negative charge from the p-GaN surface to the ITO target, followed by nearly complete repulsion of plasma electrons from the p-GaN surface.

S2. Measurement of the electron density of the plasma electron during the RF sputtering of ITO and electron-beam irradiation using the cut-off probe
In order to compare the electron-beam energy to the plasma electron during the RF sputtering of ITO, we measured the electron density of the plasma electron during the RF sputtering of ITO as well as electron-beam irradiation using the cut-off probe. As shown in the following Figure S2, the density of plasma electron during the RF sputtering of ITO increased gradually from 1.3x10 9 cm -3 to 2.3x10 9 cm -3 as the RF power increased from 40 W to 160W. The density of plasma electron for the RF power of 120 W in this study was measured as 1.9x10 9 cm -3 . The electron density during the electron-beam irradiation as a function of DC power (RF power was kept as 150 W) in the electron-beam gun was also measured. As shown in the Figure S2, the electron density increased from 4.0x10 8 cm -3 to 1.6x10 9 cm -3 as the DC power in the electronbeam gun increased from 50 V to 1500 V. The electron density for the DC power of 1000 W in this study was measured as 1.4x10 9 cm -3 . These results show that the electron density of the electron-beam is similar to that of the RF sputtering of ITO in this study.

S3. Electron-beam irradiation on p-GaN prior to e-beam evaporated ITO pelectrodes
We also examined the electron beam irradiation on p-GaN prior to the deposition of the e-beam evaporated ITO electrode. A 2-inch electron-beam gun in the sputtering system, as shown in the inset of Fig. 5(a), was used in which the electron-beam was generated with an RF power of 150 W and a DC power of 1 kV. For comparison, the e-beam evaporated ITO electrode was deposited on p-GaN without the electron-beam irradiation from the 2-inch diameter electron-beam gun. As shown in Fig. S3(a), the forward voltage at 20 mA significantly increased to over 8 V for the LEDs with the electron beam irradiation on p-GaN prior to deposition of the e-beam-evaporated ITO electrode. As shown in Figs. S3(b) and (c), The LOP and EL intensity of the LEDs with the electron-beam irradiation on p-GaN prior to the deposition of the e-beam-evaporated ITO electrode were reduced, due to the non-uniform distribution of light intensity as shown in Fig. S3(d). These clearly demonstrate that electron-beam irradiation caused severe plasma-induced damage on p-GaN, and suggest that removal of electron flux during sputtering of ITO on p-GaN is essential to obtain plasma-induced damage-free sputtered ITO electrode on p-GaN.

S4. Fabrication of sputtered ITO Schottky diodes on p-GaN in the LED structure
We fabricated the Schottky diodes on p-GaN with the RF-sputtered ITO, and twostep-sputtered ITO with first contact layer (10 nm) using RF+DC (80 W+40 W) followed by the second RF-sputtered layer (50 nm) which was developed in this study.
Four different diameters (60, 80, 100, 120 m) of the sputtered ITO Schottky diodes were fabricated on p-GaN in the LED structures, where the Ni/Ag (2 nm/200 nm) was used as an ohmic contact on p-GaN, as shown in the following Figure. As shown in the I-V characteristic of the Ni/Ag contact on p-GaN showed a linear curve, indicating the formation of good ohmic contact, and the contact resistivity obtained from the TLM was 1x10 -3 -cm 2 .  Figure S5. a) A low-magnification cross-sectional TEM image of the as-deposited e-beam evaporated ITO on p-GaN, b) high-magnification TEM image at the as-deposited e-beam evaporated ITO film, showing an amorphous crystal microstructure, and c) high-magnification TEM image at the interface between the as-deposited e-beam-evaporated ITO film and p-GaN, showing an amorphous crystal microstructure at the e-beam-evaporated ITO/p-GaN interface.