Strategies towards Compact UV Laser Sources


Wunderer, T. Strategies towards Compact UV Laser Sources. International Workshop on Nitride Semiconductors. 8/27/2014


There is increasing interest in developing compact UV laser sources that emit at wavelengths below 300 nm. This range includes the germicidal wavelengths (nominally 260 to 280 nm), which are also of strong interest for LEDs. Applications specifically requiring laser sources are predominantly for detection and identification of chemical species and bio-particles by native fluorescence or Raman spectroscopy. For point-of-need applications, the source requirements can include high output power, high power efficiency, compactness, ruggedness, and various degrees of spectral and spatial beam quality. Wide bandgap semiconductors based on the AlGaInN materials system offer the greatest promise for realizing UV laser sources that can meet these requirements. Even within the AlGaInN materials system there are several possible strategies to realize a compact UV laser. The first is direct current injection with an AlGaN-based laser diode (LD). Visible LDs have attained output powers of several Watts from a single device. These are InGaN-based devices grown on bulk GaN substrates, with a compressively strained hetero-epitaxially grown active zone. The layers are designed to confine both the electrical and optical particles and improve light-matter interaction. Growth procedures have been optimized for highest electrical-to-stimulated-emission power conversion. For emission at the shorter wavelengths in the UV band, some of the techniques used in the visible can be translated to the higher Al-containing materials and some existence-proof lasers have been demonstrated. However, the performance continually degrades as the band gap increases [1]. There are several reasons for this behavior and some are inter-correlated. For example, as the band gap increases the electrical performance degrades essentially exponentially, with higher resistance losses and reduced carrier availability, and material quality degrades with the generation of defects. PARC has pioneered the development of nitride UV lasers with demonstrations of LDs containing AlGaN quantum wells and LDs on bulk single-crystal AlN substrates [2]. More recently we have demonstrated optically pumped UV lasers with AlGaN MQWs on bulk AlN at wavelengths down to 237 nm (Fig. 1) and with record-low pump power thresholds, for example, 41 kW/cm2 at ? = 266 nm [3]. Fundamental work has included experimental observation and computational understanding of polarization switching of the laser emission at wavelengths near 250 nm [4]. Nano-structured AlGaN-based composition- and doping-modulated superlattice materials were designed, develop, tested and integrated into sub-300-nm LD heterostructures and fully processed UV laser test devices. Significant progress has been made in all aspects of LDs, with maximum current density exceeding 40 kA/cm2 and improved carrier injection [5]. Yet, challenges remain to the realization of deep UV LDs. These include optical absorption losses as well as further heterostructure optimization and device design. Alternative strategies have been proposed that are also based on AlGaInN semiconductors. One is the use of a high-energy electron beam to generate electron-hole pairs near the active zone of a UV laser heterostructure. The heterostructure itself is incorporated as the gain medium in an optical resonator for stimulated emission and amplification for direct generation of UV laser radiation. For the deep UV emission a major advantage is elimination of p-doping with the concomitant reduction of electrical resistance losses and optical absorption losses and improvement in carrier injection efficiency. The carrier generation process associated with the ballistic energy transfer is non-ideal, but has the potential to outperform other direct emission laser concepts. A practical disadvantage of the approach is the need for a vacuum environment, although there is precedent for such devices (e.g., X-ray tubes).

Read more from SRI