*3.3. Influence of the Nature of the Solution for Suspension*

The results obtained from the stirring experiment are presented in Figure 4. For TiO2 samples, both inactivation curves consist of two steps: the first with a very low inactivation rate followed by the second with a higher inactivation rate. In addition, the second rate appears to be the same for both distilled water and 1/500 NB. It is likely that 1/500 NB acts as a retarding agent of the photocatalytic disinfection process.

**Figure 4.** Survival of *E. coli* cells *vs.* irradiation time at ~5 W/m2 with the standard error of three experiments. Mean ± SE, *n* = 3.

The present finding is in full agreement with the inhibitory effect of various ions and organic compounds on photocatalytic disinfection, which is widely reported in the literature [34,37,56,65,66]. The presence of ions and organic compounds can reduce the efficiency in different ways:


According to Dunlop *et al.* [56] and Sunada *et al.* [42], the low rate of inactivation in the first step may be due to the preliminary attack of the outer membrane of cells by ROS.

During this first step, the damage sustained by the outer membrane may be insufficient to kill bacteria: they can recover from the damage and re-grow once they are plated in agar media [42,56]. After some time, degradation of the outer membrane enables reactive species to penetrate, which induces damage, leading to the death of the bacterial cells (second, higher rates on the curves of Figure 4). This hypothesis has also been considered by other researchers [58,68]. Mitoraj *et al.* [68] explained this "incubation period" as the time for the concentration of photogenerated ROS to increase to a level that is harmful to bacteria.

Another possible explanation for the first step with the low inactivation rate is proposed by Gogniat *et al.* [38]. In their works, they observed the two-stage curve only in a sodium phosphate solution and not in a NaCl-KCl solution. They hypothesized that the change of adsorption properties of TiO2 when illuminated led to a photo-desorption of ions previously adsorbed on its surface. Thus, the time taken for the photo-desorption process explains the low inactivation rate observed during the first minutes of the experiment [38].

Interestingly, the third step observed in earlier studies [58,68–70] and consisting of strong attenuation of the bacterial inactivation was not observed here. One of the hypotheses suggested is that photocatalytic inactivation is built up by bacterial growth after a certain period of time [69]. It can be supposed that bacterial growth in pure water is slowed down or stopped. Further investigations in 1/500 NB after longer times could show similar attenuation of the inactivation rate.

Some authors have compared efficiencies between scattered and fixed TiO2 [71–73]. Pablos *et al.* [72] observed a higher inactivation rate at the beginning of the reaction with fixed TiO2. They suggested that damage was uniformly distributed over the whole cell wall in slurries, whereas it was more concentrated on small areas with fixed TiO2, requiring smaller amounts of radicals to achieve inactivation. However, they observed similar times for total inactivation of bacteria (*E. coli*) for both implementations (fixed and scattered). On the other hand, Gumy *et al.* [40] found higher inactivation efficiency with suspended TiO2 than with TiO2 coated on a fibrous web and suggested that particles dispersed in slurry would provide more surfaces for the adsorption of bacteria. In addition, inactivation of bacteria has been observed in the presence of TiO2 nanoparticles in the dark, suggesting that phenomena other than photocatalytic processes can explain inactivation [50,61]. Although the complete process is not perfectly understood yet, the overall literature points to the importance of the contact between bacteria and TiO2 for improving disinfection efficiency, suggesting both chemical and physical influences.

In their work, Gomes *et al.* [71] also reported a higher inactivation rate in slurry than with TiO2 supported on Ahlstrom paper. They suggested that such results could be explained by competitive reactions of TiO2 with the organic matter released by the paper during the experiment. Accordingly, the presence of ions and/or organic compounds in the slurry/inoculum considerably reduced the efficiency by reacting with ROS and being adsorbed on TiO2 in place of bacterial cells [38,65]. These works also raised the problem of TiO2 coatings in which the organic matter from the binder can monopolize photogenerated radicals and, thus, lead to a decrease in disinfection efficiency.
