The development of a measuring technique for the UV-C distribution emitted from low pressure mercury lamps

The measurement of spectral irradiance in the ultraviolet (UV) region of the electromagnetic spectrum has higher uncertainties largely due to a low signal-to-noise ratio (SNR). UV-C radiation emitted by low pressure (LP) mercury (Hg) lamps is very important in a wide range of applications including disinfection of air and surfaces of microorganisms. This study investigates some measurement aspects to be considered when calibrating LP Hg lamps. The aim is to characterise the spectroradiometer system and improve measurement techniques for calibrating this lamp for use as a source standard for disseminating the unit of spectral irradiance in the UV-C spectral region. The research question is answered through an experiment that includes a direct measurement of a LP Hg lamp against a deuterium (D2) standard (STD) lamp. The measurements showed that the alignment of the LP Hg lamp should be a primary concern as this aspect contributed 15 % to the combined uncertainty. Ambient temperature changes are known to influence some components used in the measurement of spectral irradiance. In this regard, the study indicated that the LP Hg lamp may be operated at lower temperatures to achieve maximum light output. The lamp has a calculated temperature coefficient of -2,5 \pm 0,01 % oC at an ambient temperature of 21,3 oC. During measurements, a short-term drift of a measurand can be an obstacle to lamp spectral irradiance calibration. The cause is frequently heat build-up and temperature changes in the components of a spectroradiometer. The spectroradiometer had a higher measured short-term drift of 5 % when irradiated with a LP Hg lamp compared to 1 % when irradiated with a D2 STD lamp. Apart from a higher measured drift, the lamp’s light output was more stable over time when monitored with UV enhanced silicon (Si) detectors: 1,6 % for a LP Hg lamp and 0,11 % for a D2 STD lamp. The LP Hg lamp had a calibrated total UV irradiance of 1,12 W/m2 from 230 nm to 400 nm wavelengths. The expanded uncertainty was 8 % which is lower than 11 % uncertainty that the National Metrology Institute of South Africa (NMISA) is accredited for. This showed that even though the spectroradiometer system had a higher measured short-term drift when irradiated with a LP Hg lamp, the lamps can be used as a source standard for calibrating clients’ instruments. Further research must be conducted to identify and characterise other measurement aspects that could have significant uncertainty contributions.Dissertation (MSc)--University of Pretoria, 2019.PhysicsMScUnrestricte