Identification of low-energy peaks in electron emission spectroscopy of InGaN/GaN light-emitting diodes

Abstract

The measurement of the energy distribution of vacuum emitted electrons from InGaN/GaN light-emitting diodes (LEDs) has proven essential in understanding the efficiency loss mechanism known as droop. We report on the measurement and identification of a new low-energy feature in addition to the previously measured three peaks present in the electron emission spectrum from a forward biased LED. Photoemission measurements show that the two low-energy peaks correspond to photoemitted electrons from each of the p-contact metals, palladium and gold. We confirm that the mid and high-energy peaks are due to electrons which have transited the p-type region of the device and have been emitted from the semiconductor surface from the bulk Γ-valley or a high-energy side valley.

Direct measurement of hot-carrier generation in a semiconductor barrier heterostructure: Identification of the dominant mechanism for thermal droop

“…The EE spectra from the LED without an AlGaN EBL (Fig. 4a) showed the expected four peaks, two low-energy peaks due to photoemission from the p-contact metals, a mid-energy peak that aligned with the expected Γ-valley position, and a high energy peak generated by interband Auger recombination in the LED active region [9,17]. Examination of the integrated intensity of the semiconductor emitted peaks (Figure 4b), the mid-energy, Γvalley peak intensity remained relatively constant at temperatures up to 75 °C.…”

“…EES measurements from nitride LEDs typically show four distinct peaks. Two lowenergy peaks are associated with the light produced by the LED causing photoemission of electrons from the p-contact metals [17]. A mid-energy peak and a high energy peak are associated with electrons generated within the device which have been emitted through the ptype semiconductor surface.…”

“…The mid-energy peak corresponds to electrons that have either overshot or escaped confinement in the active region or have relaxed from higher energy states into the Γ-valley conduction band minimum before being emitted. The high energy peak is due to hot-electron generation by the interband Auger mechanism in the LEDs active region [9,17,18] and subsequent thermalization into a conduction band side valley (SV) situated 1 eV above the Γ-valley [19][20][21], before being emitted from the surface.…”

“…The total injection area defined by the p-contact is ≈ 2.2×10 -3 cm. Further details of the device design and measurements can be found in Refs [17,22].…”

“…Because the energy reference is the Fermi level of the p-contact, the current dependent ohmic voltage drop across the metal-semiconductor junction increases the Γ and SV peaks energies as current increases. For this reason, increasing the current through the device will also increase the measured energy of electrons emitted from the semiconductor surface whereas the low-energy photoemission peaks (which correspond to electrons photoemitted from the contact metals) do not change with increased injection current [17].…”