Standard semiconductor photodetector structures have been fabricated with such layers. The Al content of such layers was tailored to get a photoresponse adequate to monitor the solar UV-B radiation. The performances of such photodetectors have been determined and compared. Photoconductive detectors have shown not to be adequate for such application, due to their low UV/visible contrast, non linear behaviour, very slow speed and their temperature sensitivity. Schottky photodiodes, metal-semiconductor-metal and p-n junction photodiodes have all shown to have good responsivities, a UV/visible contrast higher than 103, linear response with incident power, no practical temperature drift, a sensitivity in the range of 1 pW/cm2, and a time response always below 1 5s. All of them are thus potential candidates to monitor the UV-B band.

In order to monitor the UV-B band, ALDUV has considered three approaches, each one requiring slightly different detector parameters: biological-action solar UV monitoring, radiometric UV-B monitoring, and narrow band data for solar UV-B monitoring.

For biological-action UV monitoring, and considering the CIE erythema response as the reference, we have developed Schottky barriers with an Al mole fraction of 28% which provides an excellent correlation with such biological action.

For radiometric UV-B monitoring, we concluded that metal-semiconductor-metal photodiodes are adequate, giving UV/visible contrasts higher than 104, with sharp cut-off at 320 nm, for an Al mole fraction of 0.25.

For narrow-band solar UV-B monitoring the number of narrow bands and their detection bandwidth is mainly an economic decision, linked to the reproducibility and yield of the detector fabrication. For this application, photodiode structures (Schottky, MSM and p-n junctions) are all adequate. For a reasonable cost, we can consider 10 nm width bands, which would require a system of four AlGaN-based photodiodes.

As a general conclusion, once this AlGaN photodetector technology has been developed, the improvements in commercially available substrates will lead to a better quality of the AlxGa1-xN epitaxial layers. Thus, an increase of the AlGaN photodiode responsivity is foreseen in a near future, simplifying even more the signal conditioning electronics of the detector head. In this direction, ELOG susbtrates were developed and used by the consortium. Besides, sapphire substrates but with a GaN layer on top (GaN templates, grown by MOVPE, and ready for homoepitaxial growth) have recently been offered by various US companies. This is a very convenient solution for a quick technology transfer to a detector manufacturing to fabricate high performance AlGaN photodiodes for the UV-B band.