Solid-State UV Photodetectors Search for a Niche

Fred Perry

Commercial solid-state UV photodetectors did not exist until around 1996 when Cree Research in Durham, N.C., began production of silicon carbide photodiodes. Since then, additional materials have debuted commercially, including GaN, AlGaN and TiO₂. Because of their relative newness, these devices are — like lasers were in the 1960s — solutions in search of a problem.

Potential applications of solid-state UV sensors are flame sensing, lamp control in UV sterilization and curing-dose regulation, personal tanning monitors, analytical instruments, and welding-goggle automatic dimming controls. Of these applications, lamp control in water purification systems is the largest current market. Because of its extreme robustness under high UV doses from mercury lamps, SiC dominates in this application. In contrast, silicon photodiodes suffer from significant degradation after a few hundred hours of exposure to 10 mW/cm² at 254 nm.

To measure only the UV signal at 340 nm with a silicon-based detector, a 340-nm interference filter must block the radiation from the visible to beyond 1000 nm. But if a photodiode such as SiC, GaN or TiO₂ were used, the filter would have to block only to beyond 425 nm.

General Electric and Honeywell both have flame sensors based on SiC, but this constitutes a smaller market than UV lamp monitoring. Compared with silicon, UV-sensitive materials offer longer life under UV flux. They also are intrinsically more stable at high temperatures because of their wide bandgaps. Therefore, there may be no need to isolate flame sensors thermally from their targets if UV is used rather than the legacy IR sensors and UV phototubes now in use.

Another potential application for sensors with spectral response that is limited to the UV is in nondispersive UV analytical instruments. Today, a large segment of the optical market uses interference filters and silicon photodiodes to measure concentrations of chemicals in liquid samples such as blood and urine. However, silicon’s photoresponse extends from the UV to beyond 1000 nm in the IR. To measure only the UV signal at 340 nm, a 340-nm interference filter must block the radiation all the way through the visible to beyond 1000 nm.

If, on the other hand, a photodiode such as SiC, GaN or TiO₂ were used, the filter would have to block only to beyond 425 nm. The implications of this for filters is significant, because it could allow them to be lower-cost, thinner and of higher transmittance.

Military applications

UV photodetectors also may find a niche in the military market. Solar energy does not reach the Earth’s surface at wavelengths below about 280 nm because of absorption by ozone in the upper atmosphere. However, there are many sources of interest at the Earth’s surface, such as extremely hot missile exhaust plumes. Because there is no natural solar illumination to produce background clutter to obscure the hot target, a UV imager could see a missile unambiguously. Visible or IR imagers, in contrast, must distinguish the missile and exhaust from a background cluttered with many other features.

New materials for UV sensors are emerging as well. Diamond is one, but there are still issues related to finding dopants that will allow the creation of PN junctions. The promise offered with diamond, though, is the capability to produce a detector that is intrinsically solar-blind, so that it could provide the target-only images that the military needs.

As costs drop for future UV sensors, other applications will appear. For instance, it may ultimately be cost-efficient to have a solar dose sensor in every wristwatch that could set off an alarm whenever a safe dose limit is reached within 24 hours.

Meet the author

Fred Perry is president of Boston Electronics Corp. in Brookline, Mass.