*2.4. Antibacterial Mechanism*

Figure 7 depicts the possible 980 nm NIR light-driven antibacterial mechanism of the UCNPs@R-TiO2 nanocomposites. After the 980 nm NIR light irradiation, the 2F7/2 state electrons of Yb3+ would be promoted into the 2F5/2 excited state band of Yb3+, then the Yb3+ 2F5/2 state would be relaxed by energy transfer to a neighboring Er3+ ion. The energy promotes the valence band (4I15/2) electrons of Er3+ into the excited state band (2H11/2, 4S3/2, or 4F9/2). The electrons in the states H, S, or F in Er3+ are unstable and would be relaxed to the 4I15/2 state by releasing energy, emitting at 523 nm, 542 nm, and 658 nm, respectively [21,35,36]. The visible light energy would be further absorbed by the valence band electrons of neighboring R-TiO2 nanoparticles. The high energy valence electrons would jump into the stable conduction band, producing electron-hole pairs.

The visible light (red and green light) emitted by UCNPs is absorbed by the R-TiO2 to produce strongly reductive electrons (e−) and oxidative holes (h+). The valid h+/e<sup>−</sup> pairs afterwards could react with H2O and O2 in an aqueous solution (Equations (1)–(7)) to produce reactive species. As it is known, the generated hydroxyl radical (*·*OH) species can be used as a strong oxidizer for the non-selective killing of bacteria [37]. The amount of *·*OH is detected by the fluorescent intensity of 2-hydroxyterephthalic acid (λ = 420 nm) which is a product of the reaction of *·*OH with terephthalic acid [38]. As shown in Figure 7, the UCNPs@R-TiO2 nanocomposite has the highest fluorescent intensity among these materials, indicating the amount of OH that was generated. These results show that the visible light emitted by UCNPs can be effectively absorbed by R-TiO2 nanoparticles. The UCNPs@R-TiO2 composites possess excellent photocatalytic performance and the specific process of the NIR photocatalytic sterilization of *E. coli* is summarized using the following reactions:

UNCPs + NIR light → Visible light, (1)

$$\text{IR-TiO}\_2 + \text{Visible light} \rightarrow \text{h}^+ + \text{e}^- \tag{2}$$

$$\text{O}\_2 + \text{e}^- \rightarrow \text{\textquotedblleft O}\_2\text{\textquotedblright},\tag{3}$$

$$\rm 2\cdot O\_2^- + 2H^+ \to H\_2O\_2 + O\_{2'} \tag{4}$$

$$\cdot \text{O}\_2\text{}^- + \text{H}\_2\text{O}\_2 \rightarrow \cdot \text{OH} + \text{OH}^- + \text{O}\_2 \tag{5}$$

H2O+h<sup>+</sup> <sup>→</sup> *·*OH + H+, (6)

*·*OH + *E. coli* → Inactivated *E. coli*. (7)

**Figure 7.** Photoluminescence spectra of UCNPs (**a**), R-TiO2 (**b**), and UCNPs@R-TiO2 nanocomposite (**c**), respectively, measured under 980 nm NIR illumination for 20 min.

### **3. Experimental Designs**

### *3.1. Reagents and Materials*

All of the chemical reagents—at analytical grade, unless otherwise noted—were used without further purification. Ytterbium(III) chloride hexahydrate (YbCl3·6H2O), Yttrium(III) chloride hexahydrate (YCl3·6H2O), Gadolinium(III) chloride (GdCl3·6H2O), and Erbium(III) chloride (ErCl3) were purchased from Alfa Aesar (Shanghai, China). Titanium isopropoxide, sodium hydroxide (NaOH, 96%), oleic acid (OA), sodium citrate, hydrofluoric acid, acetonitrile, ethanol, and chloroform of analytical grade were purchased from Sinopharm Chemical Reagent Co., Ltd. from Shanghai, China. NH4F and Thiazolyl Blue Tetrazolium Bromide (MTT) were obtained from Aladdin (Hang Kong, China) and Bomei Biotechnology (Hefei, China), respectively. Ultrapure water (18.2 MΩ·cm, Mili-Q, Millipore, Burlington, MA, USA) was used throughout the experiment.
