Next-generation displays demand efficient, low cost RGB pixels that liquid crystal (LCD) and organic LED (OLED) platforms are unable to achieve. Alternatively, III-nitride microLEDs are well-poised to replace LCDs and OLEDs due to the nitride’s superior efficiency, high brightness and long lifetimes. The individual microLED pixels must have lateral dimensions below 5 μm—causing significant nonradiative surface recombination losses at mesa sidewalls—either to minimize material costs or meet pixel density requirements. Dielectric passivation and chemical treatments have been able lower surface recombination losses, but these studies have largely focused on devices with lateral dimensions at or above 10 μm due to fabrication and characterization challenges. Furthermore, producing high quality material with indium compositions needed for full red-green-blue color tuning is presently very challenging. This requires incorporating large amounts of indium into InGaN, which leads to highly strained material.
The goals of my research are to (i) mitigate surface recombination losses and (ii) use nanostructuring to relieve strain in InGaN to enable red emission. This talk will first highlight our recent work on fabricating microLEDs down to 2 μm diameters. Chemical etching and dielectric passivation were used to minimize nonradiative sidewall defects in InGaN/GaN microLEDs. The measured external quantum efficiency (EQE) increased as the diameter was decreased—behavior that is atypical of microLED EQEs. Analysis of the EQE versus current density behavior determined enhanced backside light extraction efficiency (LEE), which was confirmed by ray tracing analysis.
Large internal electric fields are also present in InGaN quantum wells due to piezoelectric polarization at the InGaN/GaN interfaces. This bends the energy bands in the active region, red-shifts the emission, and ultimately lowers the radiative efficiency. Since surface recombination losses may be effectively eliminated through proper passivation, nanopatterning the active region itself is a potential route to solve the aforementioned problems. X-ray diffraction (XRD) determined the presence of strain relaxation which increased as the nanopatterned diameter was decreased. Additional photoluminescence characterization found evidence of a reduced piezoelectric field as a result of strain relaxation. Current and future work is also geared at developing a full device process for nanoscale LEDs in order to better understand the effects on device performance.