Titanium dioxide (P25) was obtained from Evonik Degussa Company. Silver nitrate (purity > 99.0%) was from Sigma-Aldrich. Distilled water was used throughout the experiments. Methylene blue (purity > 99.9%) was purchased from Alfa Aesar.

Ag/TiO₂ catalysts were prepared by photo-reduction method. TiO₂ was added into a water bath containing various amounts of silver nitrate. It was illuminated under UVC light with stir for 36 h. The catalyst was filter, then dried under vacuum.

In this study, the decomposition of methylene blue in aqueous solution under both UV light and visible light irradiation was used as the standard testing reaction. The detail has been described in previous literature [5] [6] [7] [8]. For the photocatalytic activity reaction, the catalyst was activated for 12 h by four pieces of 9 W UV lamps (wavelength = 254 nm (UVC), TUV PL-L 18W/4P 1CT/25, Philips). This procedure can photoreduce AgO to Ag metal completely. 40 mL of methylene blue aqueous solution with the concentration of 10 mg/L was held in quartz cell. The catalyst was added into the solution. Before the photocatalytic reaction was conducted, the reactor was held in the dark place for 1 h for the saturation adsorption of MB. The catalyst was then illuminated by four 9 W UV lamps (254 nm wavelength, TUV PL-L 18W/4P 1CT/25, Philips) or four 9 W visible light lamps (Master PL-L 18W/840/4P, Philips). The distance from the lamps to the surface of the solution was 10 cm, and the concentration of MB in aqueous solution was determined every 15 min for UV light reaction or every 1 h for visible light reaction by using UV-vis spectroscopy (Jasco V-670). The concentration of MB was determined by the intensity of 663 nm adsorption peak, which is the characteristic UV-vis peak of MB.

The crystalline structures of TiO₂ and Ag/TiO₂ were characterized by XRD. All of the diffraction peaks of the samples, as shown in Figure 1, are located at the same positions [5] [6] [7] [8] [35] - [45]. The diffraction peaks located at 2θ = 25.31˚, 37.80˚, 48.05˚, 53.89˚, 55.06˚, 62.69˚, 68.76˚, 75.03˚ corresponding to the anatase form of (101), (004), (200), (105), (211), (204), (116) and (215) (JCPDS 21-1272). The diffraction peaks located at 2θ = 27.44˚, 36.09˚, 54.32˚, 56.63˚, 64.06˚, 69.01˚, 69.80˚ corresponding to the rutile form of (110), (101), (211), (220), (310), (301) and (112) (JCPDS 34-0180). P25 TiO₂ from Evonik Degussa contains both anatase and rutile crystalline phases, addition of Ag did not change bulk crystalline structure, as expected. The peak of silver was not observed, implying that the particle size of silver was very small. The crystallite size

and d-spacing of TiO₂ were calculated by using the Debye-Scherrer equation and the Bragg’s law, respectively. The results are listed in Table 1. It should be noted that silver addition did not change the crystallite size of TiO₂, indicating the distortion of the crystal lattice by incorporation of Ag ions was only into surface of TiO₂. It did not change the bulk structure of TiO₂.

The morphology of the as-prepared catalyst was investigated by HR-TEM. Figure 2 shows that AgO particles were spherical shape, and the lattice spacing was 0.241 nm, in consistent with (111) plane of silver oxide.

Figure 3 shows the HRTEM of all the Ag/TiO₂ samples. It clearly shows that AgO was deposited on the surface of TiO₂. The particle of TiO₂ from Evonik Degussa is cubic and the size is about 5 nm. The size of AgO should be less than 5 nm. Some of AgO agglomerated on the surface of TiO₂. When Ag⁺ was deposited on TiO₂ after photoreduction preparation process, Ag⁺ was photoreduced to Ag⁰. The sample was exposed to air after preparation. Ag⁰ would be oxidized to AgO. When the sample was in HRTEM chamber, even under high vacuum, it was still in AgO state. The HRTEM shows the AgO spacing in all the samples.

The XPS spectra in the Ti 2p region of the samples are shown in Figure 4. The

peaks located at binding energy of 458.9 - 459.0 eV corresponds to Ti 2p_3/2 state and at 464.5 - 464.7 eV corresponds to Ti 2p_1/2 state [5] - [10]. It is attributed to the Ti⁴⁺ in TiO₂ lattice. It shows that the Ti 2p_3/2 peaks were broadened with increasing silver content of the sample. The results indicate that the valence state of Ti cation was reduced from Ti⁴⁺ to Ti³⁺ by charge compensation and the silver could trap the electrons, which decreased the recombination rate of electrons and holes and convert more Ti⁴⁺ ions to Ti³⁺ ions [33] [34] [35] [36]. The generation of Ti³⁺ ions could induce visible light response by the creation of oxygen vacancies [38] - [45].

Table 2 shows the fractions of Ti³⁺ and Ti⁴⁺, respectively, in each sample. Pure TiO₂ is fully crystallized, and the fraction of Ti³⁺ is zero. After loading Ag, the fraction of Ti³⁺ was not zero. Since the Ag content was very low, and only part of

Ag was incorporated into surface structure of TiO₂. The fraction of Ti³⁺ was low. Nevertheless, it clears shows that the fraction of Ti³⁺ increased after loading with Ag. The 2Ag/TiO₂ sample had lower fraction of Ti³⁺ compared with 1Ag/TiO₂, possibly because Ag was agglomerated to form big particles, resulting in small amount of Ag was incorporated into surface structure of TiO₂. Further studies are needed to investigate this phenomenon.

Figure 6 and Figure 7 display the XPS spectra of the O 1s region of Ag/TiO₂ catalysts. Two O 1s peaks appeared in both samples, which is ascribed to lattice

oxygen for Ti-O bond and hydroxyl groups on the surface [7] - [13] [38] - [45], respectively. The Ag/TiO₂ could generate more active hydroxyl radicals (HO·), which is the major species for degradation of methylene blue [38] - [45].

Figure 8 and Figure 9 show the XPS spectra of Ag 3d. The XPS curve can be well fitted by two peaks centered at 367.8 eV and 373.8 eV corresponding to silver oxide, respectively [41] [42] [43] [44] [45]. Before reaction, the catalyst was irradiated by UV light overnight. Under such condition, AgO was photoreduced to Ag metal. It is known that Ag nanoparticles could improve the photocatalytic efficiency via the following ways:

Ag 0 + e − → Ag − (1)

Ag − + O 2 → O 2 · − + Ag 0 (2)

The Ag nanoparticles were used as surface traps which can capture the electrons from the conduction of TiO₂ and inhibit the electron/hole pair recombination [38] - [44]. These electrons further move from conduction of TiO₂ to the Ag nanoparticles and react with the dissolve oxygen to generate the superoxide

radical anions ( O 2 · − ).

Figure 10 shows the adsorption amount of methylene blue on the catalyst. The amount of methylene blue adsorption increased after loading silver concentration. Multilayer methylene blue was adsorbed on the surface of catalyst, indication that adsorption of methylene blue is not the rate-determining step.

The results of photocatalytic activity under UV light irradiation are shown in Figure 11. It can be seen that the 1Ag/TiO₂ had the highest activity among all of the samples under UV light illumination. The photocatalytic activity of Ag/TiO₂ decreased when the silver content was more than 1 wt%. As silver concentration is more than the optimum amount, Ag becomes the recombination center of the photo-induced electrons and holes, which is harmful for photocatalytic reactions.

The photodegradation of organic dyes is suited to the Langmuir–Hinshelwood model. The slope of ln(C₀/C) plotted versus irradiation time (min or h) demonstrates the rate constant of the reaction of the sample, as shown in Figure 12. The rate constants of catalysts for destructure of methylene blue are calculated and listed in Table 4. The reaction rate constant of 1Ag/TiO₂ was 0.64964 h⁻¹ which was about four times of that of pure TiO₂.

Figure 13 shows the results of photocatalytic activity under visible light illumination. The photocatalytic activities of all of the Ag/TiO₂ had improvements compared to the pure P25 powder since adding Ag could have plasmon effect.

Consequently, the synthesized powders can have higher photocatalytic activity under visible light illumination [38] [39] [40]. In addition, the samples would decrease Ti⁴⁺ ions to Ti³⁺ ions by charge compensation and the generation of Ti³⁺ ions could induce visible light response by the creation of oxygen vacancies which were consistent with the XPS spectra analysis for Ti 2p region [29] [30] [38] [39] [40]. As shown in Figure 14 and Table 5, the photocatalytic activity of Ag/TiO₂ was the highest among all catalysts under visible light irradiation. The rate constant of the reaction was calculated to be 0.059911 h⁻¹ which had the most hydroxyl groups in all samples.

1Ag/TiO₂ showed the highest photocatalytic activity among all catalysts under UV light and visible light illumination since degradation by TiO₂ was conducted prominent via ·OH radicals generated in the positive holes, while Ag/TiO₂ catalysts generated ·OH radicals through the transformation of O 2 · − radicals [29] [30].

It has been reported [13] [14] [15] that the Fermi level of TiO₂ is higher than that of doped silver. Consequently, silver deposits behave like accumulation sites for photogenerated electrons moved from TiO₂. As the number of silver cluster is small and the electrons efficiently move to silver cluster, better separation of electrons and holes would be complied with increment of silver doping to the optimum content. The electrons could react with surface Ti⁴⁺ or adsorbed oxygen molecular to produce reactive center surface Ti³⁺ and reactive species

O 2 · − , respectively, the deal of recombination center of inner Ti³⁺ reduced at the same time. Furthermore, the doped silver particle can move an electron to adsorbed oxygen molecular to produce O 2 · − or to the TiO₂ surface Ti⁴⁺ to produce surface Ti³⁺. This indicated that the recombination was slowed and the production of O 2 · − and surface Ti³⁺ was speeded up. The productivity of h⁺ would also be raised. The reactions are shown as follows:

e − + Ag n → Ag n − (3)

Ag n − + O 2 → Ag n + O 2 · − (4)

Ag n − + Ti 4 + → Ag n + Ti surface 3 + (5)

With increment of silver doping, the quantity and size of silver cluster became bigger progressively and the energetic properties of the doped silver may be closed to that of bulk silver making the silver sites turn into recombination center of electrons and holes.

Ag n − + h + → Ag n (6)

Moreover, higher amount of silver would enshroud more TiO₂ surface and prevent the contact between organics and TiO₂, which would decrease lots of received photons and increase diffuse distance.