The lighting industry is nearly 100 years old and is often characterized as mature. However, in the next few decades, general illumination technology will undergo a remarkable transformation. Just as transistors replaced vacuum tubes fifty years ago and flat panel displays are replacing cathode ray tube monitors today, solid-state lighting is poised to take the place of incandescent and fluorescent lamps for general illumination.
Although nitride-based LEDs have been successfully introduced into the marketplace, high-performance devices suitable for solid-state lighting (SSL) sources in general illumination still require significant development in substrate and emitter technology. A major limitation has been the lack of a lattice-matched substrate that addresses both cost and function concerns.
Native GaN substrates are not commercially available. Alternative substrates, such as sapphire and SiC, are frequently used to fabricate nitride-based white emitters. Although sapphire is a relatively inexpensive material and readily available, the large lattice mismatch between sapphire and GaN leads to a high defect density in GaN epi-layers. This limits any potential improvements of the internal quantum efficiency of the light emitters. SiC has a better lattice match to GaN, and high brightness light emitting diodes have been fabricated on SiC substrates. However, the cost of these substrates is very high, which limits the competition of LEDs formed on SiC with current lighting sources.
Our research focuses on using ZnO for both the base and devices, as this material is proving to be a good alternative substrate for GaN-based devices and a promising candidate for ultraviolet and blue solid-state emitters. ZnO has a wurtzite structure, which is the same crystal structure as GaN. The lattice mismatch between the two materials can be as small as 2%. Epitaxial growth of GaN-based materials on ZnO could be a promising path to achieving high-quality with low defect density. However, the growth process is slow and results in a costly product. Bulk single-crystal ZnO has been produced using sublimation and hydrothermal techniques. The latter use extremely reactive zinc-containing solutions that require platinum-lined hydrothermal vessels. Unfortunately, these crystals are costly, and their electrical conductivity is very low because of the presence of large amounts of Li.
We have developed a unique crystal-growth technology to produce high-quality single-crystal ZnO substrates.1–4 Based on a pressurized induction melting apparatus, where the melt is contained in a cooled crucible, the process begins by batching ZnO powder into crucibles ranging from two to eight inches in diameter. Radio-frequency (RF) energy is used to produce joule heating until the ZnO is molten (∼1900°C). Once a melt is formed, the ingot is slowly crystallized. This technique enables us to perform the necessary doping to adjust the electrical conductivity of ZnO. With this technology, we have developed high-quality single-crystal ZnO substrates with a maximum diameter of 50mm (see Figure 1).
Figure 1. Single crystal ZnO substrates were developed using a crystal growth technology based on a pressurized induction melting apparatus.
The LEDs made with this substrate feature a direct band gap of 3.37eV and a free exciton binding energy of 60meV at room temperature. As a result, they offer several advantages as compared to those made from other compounds, including lower threshold voltages, lower growth temperatures, wet-etch processing techniques, and a large exciton energy. Initially, further progress was hampered by difficulties encountered when growing the p-doped layers required to form the devices. In response, we developed a metal organic chemical vapor deposition (MOCVD) process to produce p-type ZnO layers with a carrier concentration of 3 × 1017cm-3 and a mobility of 4.6cm2/Vs.5,6
LEDs were fabricated based on p–n junctions formed by an MOCVD grown p-type ZnO epilayer and an n-ZnO substrate. By using multiple metal layers to form ohmic contacts to the p–n junction, we developed ZnO-based LEDs that showed rectifying function with a turn-on voltage of 3V. When driven at 140mA, these devices demonstrated strong emission at 384nm (Figure 2) due to donor-acceptor pair recombination between shallow donors and nitrogen luminescent centers.6
Figure 2. This electroluminescence spectrum shows that when driven at 140mA, ZnO based LEDs demonstrate strong emission at 384nm.
While solid-state lighting products are already demonstrating strong performance and market growth in signaling and display applications, broad commercial applications in general illumination are not projected to command a significant market share for a decade or more. The challenge is to develop high-power devices that emit high-quality white light while meeting targets for efficiency, longevity, manufacturability, and cost. High-quality single-crystal ZnO substrates offer the promise of growing GaN and InGaN with reduced defects to enhance efficiency at low cost. ZnO based LEDs also demonstrate the possibility of new sources for SSL applications. Our next steps will be to refine the epi-growth process of GaN-based devices on ZnO substrates and to develop more reliable ZnO based LEDs with higher efficiency.
The authors thank the US Department of Energy, Army Research Office, Air Force Office of Scientific Research, Office of Naval Research, and Missile Defense Agency for supporting this work.