UC Santa Barbara

Germicidal ultraviolet (GUV) light with wavelengths between 250 nm and 300 nm was first discovered to have a strong disinfectant effect in the 19th century. GUV irradiation is a chemical-free, species-agnostic disinfection technology: some example applications include disinfection of medical equipment and buildings; municipal water treatment, residential water purification, and consumer devices like water bottles; air disinfection in offices, schools, planes, cruise ships, and retail spaces, and countless others. Most importantly, a low-cost, durable, portable, and efficient solid-state GUV light source could bring these disinfection applications to low-income and developing regions where most sewage enters waterways untreated, and the lack of sterile environments and equipment worsens healthcare outcomes. The current incumbent technology for GUV light emission is the mercury lamp. This technology has remained largely unchanged for the past century—it is efficient and inexpensive, and has a large market due to its use for phosphor-converted white light (“fluorescent lighting”). On the other hand, this technology relies on toxic mercury vapor contained within a hot quartz bulb, requires complex and inefficient driving electronics, and cannot be reliably dimmed or rapidly switched on/off. To meet growing demand for disinfection technologies, and for GUV to expand into residential, consumer, and smart applications, a solid-state GUV light emitter is needed. A UV LED would have many benefits, much like the now-mature solid-state blue and white LEDs which have revolutionized the lighting and display industries. Semiconductor LEDs can be produced at smaller sizes, lower costs, higher efficiencies, and longer lifetimes than their 19th-century vapor lamp counterparts; these semiconductor devices can be dimmed or switched rapidly with no warm-up or flicker, and are readily produced in a variety of target wavelengths. Currently available UV LED technologies based on the AlGaN materials system are much less efficient (3-5%) than Hg lamps (20-30%), due to a number of challenges related to AlGaN material quality, chemical purity, process technologies, and optical properties. Therefore, there is a strong desire to improve the efficiency, cost, and lifetime of UV emitters to match their visible LED counterparts, which have efficiencies well over 50% and lifetimes over 10,000 h. In this dissertation, I discuss the ways in which AlN and AlGaN defect density on foreign substrates has been greatly reduced using improved MOCVD growth methods, achieving a dislocation density of 2x108 cm-2 on SiC. Next, I will explain why n-type and p-type doping (both of which are needed for LED operation) of AlGaN are challenging, and how both have been greatly improved using growth condition modification and polarization engineering, respectively. Using the methods outlined in this dissertation, we achieved n-type AlxGa1-xN (x>0.65) resistivities below 10 m-cm in material grown on three different substrate platforms (4H-SiC, 6H-SiC, and sapphire), and on two different MOCVD reactors. I will discuss active region engineering for increased light emission power, as well as an analysis of the various LED processing methods and packaging architectures used in the devices shown. Combining improvements in material quality, electrical conductivity, and optical quality, we have increased GUV LED external quantum efficiency from 2% to 10%.