Traditional LED has revolutionized the field of lighting and display due to their superior performance in terms of efficiency, stability and device size. LEDs are typically stacks of thin semiconductor films with lateral dimensions of millimeters, much smaller than traditional devices such as incandescent bulbs and cathode tubes. However, emerging optoelectronic applications, such as virtual and augmented reality, require LEDs in the size of microns or less. The hope is that micro – or submicron scale LED (µleds) continue to have many of the superior qualities that traditional leds already have, such as highly stable emission, high efficiency and brightness, ultra-low power consumption, and full-color emission, while being about a million times smaller in area, allowing for more compact displays. Such led chips could also pave the way for more powerful photonic circuits if they can be grown single-chip on Si and integrated with complementary metal oxide semiconductor (CMOS) electronics.
However, so far, such µleds have remained elusive, especially in the green to red emission wavelength range. The traditional led µ-led approach is a top-down process in which InGaN quantum well (QW) films are etched into micro-scale devices through an etching process. While thin-film InGaN QW-based tio2 µleds have attracted a lot of attention due to many of InGaN’s excellent properties, such as efficient carrier transport and wavelength tunability throughout the visible range, until now they have been plagued by issues such as side-wall corrosion damage that worsens as device size shrinks. In addition, due to the existence of polarization fields, they have wavelength/color instability. For this problem, non-polar and semi-polar InGaN and photonic crystal cavity solutions have been proposed, but they are not satisfactory at present.
In a new paper published in Light Science and Applications, researchers led by Zetian Mi, a professor at the University of Michigan, Annabel, have developed a submicron scale green LED iii – nitride that overcomes these obstacles once and for all. These µleds were synthesized by selective regional plasma-assisted molecular beam epitaxy. In stark contrast to the traditional top-down approach, the µled here consists of an array of nanowires, each only 100 to 200 nm in diameter, separated by tens of nanometers. This bottom-up approach essentially avoids lateral wall corrosion damage.
The light-emitting part of the device, also known as the active region, is composed of core-shell multiple quantum well (MQW) structures characterized by nanowire morphology. In particular, the MQW consists of the InGaN well and the AlGaN barrier. Due to differences in adsorbed atom migration of the Group III elements indium, gallium and aluminum on the side walls, we found that indium was missing on the side walls of the nanowires, where the GaN/AlGaN shell wrapped the MQW core like a burrito. The researchers found that the Al content of this GaN/AlGaN shell decreased gradually from the electron injection side of the nanowires to the hole injection side. Due to the difference in the internal polarization fields of GaN and AlN, such volume gradient of Al content in AlGaN layer induces free electrons, which are easy to flow into the MQW core and alleviate the color instability by reducing the polarization field.
In fact, the researchers have found that for devices less than one micron in diameter, the peak wavelength of electroluminescence, or current-induced light emission, remains constant on an order of magnitude of the change in current injection. In addition, Professor Mi’s team has previously developed a method for growing high-quality GaN coatings on silicon to grow nanowire leds on silicon. Thus, a µled sits on a Si substrate ready for integration with other CMOS electronics.
This µled easily has many potential applications. The device platform will become more robust as the emission wavelength of the integrated RGB display on the chip expands to red.