Investigating photovoltaic-powered light-emitting diode based disinfection of water for point-of-use application

This interdisciplinary investigation assessed photovoltaic (PV) power and light-emitting diode (LED) technologies for point-of-use (POU) drinking water disinfection. A literature review identified the prospects offered by emerging UV-C LEDs and mature UV-A/visible LEDs. Improved understanding of microbial sensitivity to different wavelengths and various engineering constraints appeared necessary to realize the POU device concept.Bench-scale POU-model experiments were conducted to determine the effectiveness of commercial LEDs (270-740nm) at inactivating E. coli K12 and E. faecalis ATCC19433 and develop action spectra. Effective >3-log10 disinfection of E. coli was achieved with 270, 365, 385, 405, 430 and 455nm arrays over periods of 3 minutes to 6 hours.Three prototype systems were then constructed, utilizing UV-C and UV-A LEDs to treat single-glass and 10-15L volumes in 15 minutes to 3 hours. Consistent with the POU models, these systems achieved 2 to 5-log10 inactivation with differences explained by suboptimal reactor hydrodynamics and increased path-length.A range of engineering issues were evaluated experimentally at bench-scale. UV-C LED lifetime testing achieved in excess of 5 000 hours operation, retaining 65% of initial flux. Solarization was not encountered. Self-heating losses of 15-24% were observed under continuous operation. Low-cost monitoring of UV-C flux using modified LEDs and fluorochromes was successfully trialled.TiO2 photocatalyst coated Raschig rings did not enhance inactivation despite literature reports. However, Chlorophyllin as a photosensitizer showed some inconsistent gains. From reviewing literature, other enhancement strategies (i.e. multiple wavelength synergies, LED pulsing and overdriving) were of limited benefit with increased cost, reduced time efficiency and LED lifetime.For comparison, photoinactivation via solar disinfection method (SODIS) was modelled using SMARTS simulated sunlight spectra and bacterial action spectra from literature. High sensitivity of SODIS to seasonality, time of day, water container material and altitude was evident, which would not affect PV-powered LED disinfection. Finally, a market survey was undertaken to assess present-day commercial LED and PV technology advances and volume pricing prospects. Costing models were developed to determine the financial feasibility of various configurations. LED costs are dominant for UV-C, although near-term reductions in price are expected. Future prospects and areas of consideration for PV-powered LED-based disinfection were identified