Efficient deep-UV light source has potential for food safety and counterfeit detection

Ultraviolet light has many industrial applications, because it is one of the best ways to kill harmful bacteria. The energetic wavelength of the band of light known as UV-C or deep UV (200 to 280nm) can penetrate the cell walls of bacteria, viruses and moulds, attacking their DNA and killing them. But producing these wavelengths is an awkward business. Lamps that emit deep-UV are based on light emission from xenon gas or mercury vapour, which is toxic; moreover, they are bulky and energy-inefficient.

However, such is the usefulness of UV for disinfecting water supplies, preserving food and checking for counterfeit currency (banknote designers are increasingly using ink visible under deep-UV to mark out legitimate currency) that many research groups are looking for better ways of producing UV-C. A group at Cornell University in New York State has now achieved the lowest wavelength yet using thin films of semiconductors rather than mercury vapour as the emitter.

In the journal Applied Physics Letters, a team led by Huili (Grace) Xing and Debdeep Jena explain how they made a deep-UV emitting LED by depositing monolayers of gallium nitride and aluminium nitride using plasma-assisted molecular beam epitaxy to produce quantum structures that act as UV emitters. Previous attempts at this type of emission have used aluminium gallium nitride instead of gallium nitride, but the team found that the substitution improved two important properties of the emitter: the proportion of all electrons in the active region that produce UV light and the proportion of photons generated in the active region that can be extracted from the device.

The new LED achieved a UV wavelength of 232nm, beating the previous record holder for a gallium nitride device of 239nm. The team is now working on ways packaging the LED into a device that could go on to the market. “In terms of quantifying the efficiency, we do want to package it within the next few months and test it as if it was a product, and try to benchmark it against a product with one of the available technologies,” Jena said.