A Comparison of Solar Photocatalytic Inactivation of Waterborne E. coli Using Tris (⁠$2,2′$-bipyridine)ruthenium(II), Rose Bengal, and $TiO2$

Background. The development of alternative processes to eliminate pathogenic agents in water is a matter of growing interest. Current drinking water disinfection procedures, such as chlorination and ozonation, can generate disinfection by-products with carcinogenic and mutagenic potential and are not readily applicable in isolated rural communities of less-favored countries. Solar disinfection processes are of particular interest to water treatment in sunny regions of the Earth. Solar light may be used to activate a photocatalyst or photosensitizer that generates, in the presence of molecular oxygen dissolved in water, reactive oxygen species (ROS), such as the $HO•$ radical, singlet oxygen $(O21)$⁠, or superoxide $(O2•)$⁠, which are toxic to waterborne microorganisms. Method of Approach. Wild and collection-type Escherichia coli have been selected as model bacteria. Inactivation of such bacteria by either $TiO2$ nanoparticles, water-soluble tris(⁠$2,2′$-bipyridine)ruthenium(II) dichloride or Rose bengal (RB) subject to simulated sunlight have been compared. Although $TiO2$ is the prototypical material for heterogeneous photocatalysis, the other two dyes are known to generate significant amounts of $O21$ by photosensitization but have different chemical structures. The concentration of dye, illumination time, photostability, presence of scavengers, and post-treatment regrowth of bacteria have been investigated. Results. After $1hr$ of solar illumination the Ru(II) complex produced a strong loss of E. coli culturability monitored with solid selective agars. Both the collection- and wild-type bacteria are sensitive to the treatment with $2-10mgL−1$ of dye. This photosensitizer showed a better inactivation effect than $TiO2$ and the anionic organic dye RB due to a combination of visible light absorption, photostability, and production of $O21$ and other ROS when bound to the bacterial membrane. A complete loss of culturability was observed when the initial concentration was $103CFUmL−1$⁠, with no bacteria regrowth detected after $24hr$ of the water treatment. At higher initial microorganism levels, culturability still remains and regrowth is observed. Scavengers show that the $HO•$ radical is not involved in bacteria inactivation by photosensitization. Conclusions. A higher quantum yield of ROS generation by the sensitizing dyes compared to the semiconductor photocatalyst determines the faster sunlight-activated water disinfection of photodynamic processes. The homogeneous nature of the latter determines a more efficient interaction of the toxic intermediates with the target microorganisms. Solid supporting of the Ru(II) dye is expected to eliminate the potentials problems associated to the water-soluble dye.