Ag/Ag₂O as a Co-Catalyst in TiO₂ Photocatalysis: Effect of the Co-Catalyst/Photocatalyst Mass Ratio

Abstract

Mixtures and composites of Ag/Ag₂O and TiO₂ (P25) with varying mass ratios of Ag/Ag₂O were prepared, employing two methods. Mechanical mixtures (TM) were obtained by the sonication of a suspension containing TiO₂ and Ag/Ag₂O. Composites (TC) were prepared by a precipitation method employing TiO₂ and AgNO₃. Powder X-ray diffraction (XRD) and X-ray photoelectron spectroscopy (XPS) confirmed the presence of Ag(0) and Ag₂O. The activity of the materials was determined employing methylene blue (MB) as the probe compound. Bleaching of MB was observed in the presence of all materials. The bleaching rate was found to increase with increasing amounts of TiO₂ under UV/vis light. In contrast, the MB bleaching rate decreased with increasing TiO₂ content upon visible light illumination. XRD and XPS data indicate that Ag₂O acts as an electron acceptor in the light-induced reaction of MB and is transformed by reduction of Ag⁺, yielding Ag(0). As a second light-induced reaction, the evolution of molecular hydrogen from aqueous methanol was investigated. Significant H₂ evolution rates were only determined in the presence of materials containing more than 50 mass% of TiO₂. The experimental results suggest that Ag/Ag₂O is not stable under the experimental conditions. Therefore, to address Ag/Ag₂O as a (photo)catalytically active material does not seem appropriate.

1. Introduction

Environmental problems related to water and air contamination, due to increasing world population and the resulting tremendous growth of industry and fuel combustion, have become a major concern of advanced science. In order to deal with this important problem, photocatalytic processes with employment of semiconductors are the most conventional approaches for water and air purification, along with alternative energy storage (e.g., H₂) [1,2,3,4].

To date, different semiconductor nanoparticles such as TiO₂, ZnO, Fe₂O₃, niobates, tantalates, and metal sulfides, and their underlying working mechanisms, have been investigated with the aim of increasing their photocatalytic activity. It is well known that, besides the ability to decontaminate polluted air and water, a photocatalyst should meet certain requirements such as cost efficiency, stability, non-toxicity, and broad range response towards incident light. TiO₂ is reported as the most durable photocatalyst, responding to all the above-mentioned requirements apart from broad range response to incident solar light due to its wide bandgap energy, (3.2 eV for anatase, 3.0 eV for rutile) which accounts for no more than 5% of the entire solar spectrum [1]. This lack of photocatalytic activity under visible light illumination allows the use of TiO₂ as a UV blocker in sunscreens [5]. The tremendous interest in modification of titanium dioxide with different metals and oxides, to enable absorption of lower energy states and increase stability, has been rising over the last 20 years. Nonetheless, the range of visible-light photocatalysts is still restricted. Thus, it is essential to discover new and efficient photocatalytic materials that are sensitive to visible light.

Ag₂O nanoparticles have been broadly utilized in various manufacturing areas as stabilizers, cleaning agents, electrode supplies, dyes, antioxidants, and catalysts for alkane activation and olefin [6,7]. Several papers have been published reporting the photocatalytic activity of Ag₂O, Ag/Ag₂O, Ag₂O/semiconductors, and Ag/Ag₂O/semiconductor composites, and some reviews are available [8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33]. Ag₂O is reported to be a visible light active photocatalyst. However, due to its photosensitive and labile properties under incident light illumination, Ag₂O is infrequently employed alone as a main photocatalyst rather than as a co-catalyst [8].

Wang et al. investigated the photocatalytic performance of Ag₂O on the photocatalytic decolorization of methyl orange, rhodamine B, and phenol solution under fluorescent light irradiation, and concluded that the stability and high photocatalytic activity of Ag₂O is maintained by the partial formation of metallic Ag on its surface during the photodecomposition of organic compounds [9]. Jiang et al. also reported the decomposition of methyl orange under visible light, ultraviolet light, near-infrared (NIR) light, and sunlight irradiation, using silver oxide nanoparticle aggregation. The superb photo-oxidation performance of Ag₂O is kept almost constant after repeated exposure to light due to its narrow band gap, high surface area, and numerous crystal boundaries supplied by Ag₂O quantum dots [13]. Several authors have claimed that an Ag/Ag₂O structure exhibits ‘self-stability’ [9,10] during a photocatalytic run, due to rapid electron transfer from the excited Ag₂O to Ag(0) [12,20].

Visible light active nanocomposites of Ag/Ag₂O/TiO₂ have been synthesized using different methods, such as a microwave-assisted method [28], a low-temperature hydrothermal method [32], a one-step solution reduction process in the presence of potassium borohydride [22], a simple pH-mediated precipitation [23], and a sol-gel method [27]. Moreover, Su et al. developed a novel multilayer photocatalytic membrane, consisting of an Ag₂O/TiO₂ layer stacked on a chitosan sub-layer immobilized onto a polypropylene [31]. Light-induced hydrogen production via photoreforming of aqueous glycerol has been scrutinized, employing Ag₂O/TiO₂ catalysts prepared by a sol-gel method with varying content of Ag₂O (0.72–6.75 wt %) [30]. Hao et al. have reported that TiO₂/Ag₂O nanowire arrays forming a p-n heterojunction are applicable for enhanced photo-electrochemical water splitting [33]. Hu et al. reported the photocatalytic degradation of tetracycline under UV, visible, NIR, and simulated solar light irradiation with the Z-scheme between visible/NIR light activated Ag₂O and UV light activated TiO₂, using reduced graphene oxide as the electron mediator. They also investigated the stability of Ag₂O, Ag₂O/TiO₂, and Ag₂O/TiO₂ in combination with reduced graphene oxide as an electron mediator. A large amount of Ag(0) was formed into Ag₂O and Ag₂O/TiO₂ after four cycles of tetracycline photodegradation under UV, visible, and NIR illumination [23]. Ren et al. also observed the light-induced reduction of Ag₂O during dye degradation in Ag₂O/TiO₂ suspensions. The authors suggested that the formation of Ag(0) contributed to the high stability of their photocatalyst [29]. The stabilization of Ag₂O/TiO₂ photocatalysts by Ag(0) formed at an initial stage of an experimental run has already been proposed earlier [11]. The photocatalytic stability of Ag-bridged Ag₂O nanowire networks/TiO₂ nanotubes, which were fabricated by a simple electrochemical method, revealed only an insignificant loss in performance, with respect to photocatalytic degradation of the dye acid orange 7, under simulated solar light [15]. On the other hand, Kaur et al. reported a decrease of the degradation efficiency from 81% to 54%, after the third experimental run employing AgO₂/TiO₂ as the photocatalyst and the drug levofloxacin as the probe compound [24]. Very recently, Mandari et al. synthesized plasmonic Ag₂O/TiO₂ photocatalysts, which could absorb visible light by the resonant oscillation of the conduction band electrons under visible light illumination. With this method, they were able to improve the efficiency of TiO₂ as a photocatalyst for hydrogen production by H₂O splitting under natural solar light. The authors observed the formation of Ag(0) by light-induced reduction of Ag₂O [26]. Light-induced reduction of Ag(I) to Ag(0) has also been reported for an Ag(0)/Ag(I) co-doped TiO₂ photocatalyst [34].

The preceding discussion of published experimental results provoked doubt on the stability of Ag₂O-containing photocatalysts under UV/vis illumination. Therefore, visible light harvesting Ag/Ag₂O ⁄⁄ TiO₂ photocatalysts for water treatment and photocatalytic hydrogen generation were synthesized. To the best of our knowledge, physical Ag/Ag₂O ⁄⁄ TiO₂ mixtures synthesized by the sonication of a suspension containing TiO₂ (P25) and a self-prepared Ag/Ag₂O were investigated for the first time. Ag/Ag₂O ⁄⁄ TiO₂ composites, prepared in situ by a simple precipitation method employing TiO₂ and AgNO₃, were also prepared, in order to evaluate the effect of the synthesis method on the photocatalytic activity. Additionally, the effect of the mass ratio of Ag/Ag₂O was studied. The as-prepared mixtures and composites showed improved visible light activity for methylene blue (MB) bleaching, compared to blank TiO₂, and high photocatalytic H₂ production from a methanol-water mixture under artificial solar light illumination.

2. Results

2.1. Characterization of the Prepared Materials

The powder X-ray diffraction (XRD) patterns of Ag/Ag₂O, physical mixtures of Ag/Ag₂O ⁄⁄ TiO₂ with increasing amounts of TiO₂ (20 mass% (TM 41), 50% (TM 11), and 80% (TM 14)), and in situ prepared Ag/Ag₂O ⁄⁄ TiO₂ composites (20 mass% TiO₂ (TC 41), 50% (TC 11), and 80% (TC 14)) are shown in Figure 1. The XRD peaks for Ag/Ag₂O at 26.7°, 32.8°, 38.1°, 54.9°, 65.4°, and 68.8° perfectly correlate to the (110), (111), (200), (220), (311), and (222) crystal planes of cubic Ag₂O (JCPDS 41–1104). The three peaks at 44.3°, 64.7°, and 77.5° are indexed to the (200), (200), and (311) crystal planes of cubic Ag(0), respectively (JCPDS 04-0783) [35,36].

The TiO₂ containing mixtures (TM) and composites (TC) exhibit diffraction peaks at 25°, 38°, 48°, 54°, 55°, 63°, 69°, 71°, 75°, and 83°, which are attributed to the tetragonal phase of anatase TiO₂, whereas one peak at 27.8° corresponds to the tetragonal phase of rutile TiO₂. Figure 1a presents the patterns of the TM mixtures, where two phases of titania were present. The two strongest peaks of Ag₂O become more prominent, with the Ag₂O mass ratio increasing from TM 14 to TM 41. The small diffraction peaks situated at 44.4°, 64.2°, and 77.5° are indexed to the (200), (200), and (311) plane of metallic Ag(0) (JCPDS 04-0783) [20]. The strongest peak of Ag(111) might likely be masked by the TiO₂ peak at 2θ = 38°. The diffraction peaks in the TM mixture patterns correspond to the cubic structure of Ag₂O and the cubic structure of Ag [35,36]. Figure 1b illustrates the XRD patterns of the TC composites. As the figure shows, no significant difference between the two preparations methods was observed, except that in TiO₂-rich composites TC 11 and TC 14 no Ag₂O diffraction peaks were observed, suggesting a complete reduction of Ag₂O to metallic silver Ag(0) during the preparation of these composites. The XRD pattern of TiO₂ is presented for comparison. The diffractogram clearly indicates the presence of two TiO₂ phases with predominance of the anatase phase (JCPDS 21–1272).

In order to investigate the oxidation states of the silver species present on the materials, X-ray photoelectron spectroscopy (XPS) was performed. The results of the XPS analysis for all samples are shown in Figure S3. The deconvolution of the high-resolution spectra for Ag 3d reveals that silver was present in more than one oxidation state in all samples. The binding energies of Ag 3d at 367.5 and 373.5 eV are assigned to the Ag 3d_5/2 and Ag 3d_3/2 photoelectrons respectively, indicating the presence of silver in the +1 oxidation state. The other two peaks of Ag 3d_5/2 and Ag 3d_3/2, at 368.3 and 374.3 eV respectively, confirm the existence of silver in the Ag(0) state. These binding energies are in good agreement with the values reported for Ag(I) in Ag₂O and Ag(0) [16,37,38]. The peaks for O 1s, located in the ranges of 528.9–530.1 eV and 530.5–531.2 eV, are ascribed to O²⁻ in Ag₂O and TiO₂ respectively (Figure S3). From the Ti 2p core-level spectrum, two peaks at about 464.3 and 458.7 eV can be assigned to the Ti 2p_1/2 and Ti 2p_3/2 spin–orbital components respectively, which correspond to the characteristic peaks of Ti⁴⁺.

The SEM images of blank TiO₂, Ag/Ag₂O, TM mixtures, and TC composites are presented in Figure 2. Ag/Ag₂O showed well-defined particles with particle sizes ranging from 100 nm to 500 nm (Figure 2a). The small particles that contrast as white spots correspond to the metallic silver Ag(0) distributed on the surface of silver oxide, which is in agreement with the XRD results. The EDX reveals that the sample contained Ag and O without any other impurities (Figure S1).

Figure 2b–d shows SEM images of the physical mixtures of Ag/Ag₂O with TiO₂. It becomes obvious from these images that Ag/Ag₂O changed its shape during preparation of the mixtures by sonification of aqueous suspensions of the oxides. The increasing loading of the Ag₂O platelets with TiO₂ is also clearly recognizable in these figures. In the Ag/Ag₂O ⁄⁄ TiO₂ mixture with the highest mass fraction of TiO₂ (TM 14), the appearance was apparently determined by the titanium dioxide distributed over the underlying surface of the Ag₂O platelets (Figure 2d). This was also reflected in the specific surface area (SSA) of the materials. The TiO₂ (P25) used in this work is known to have an average diameter and specific surface area of 21 nm and about 50 m² g⁻¹, respectively [39]. The specific surface area of the Ag/Ag₂O synthesized in this work was determined to be 2.7 m² g⁻¹. As expected, the specific surface area of the Ag/Ag₂O ⁄⁄ TiO₂ mixtures was found to increase with increasing TiO₂ content (

Table 1. Brunauer-Emmet-Teller (BET) surface area, initial rates of methylene blue (MB) bleaching and H₂ generation in the presence of Ag/Ag₂O, TiO₂, the TM mixtures and the TC composites.

Sample	Preparation Method	Composition	SSA m2 g−1	r0 (MB) UV/vis mg L−1 min−1	r0 (MB) vis mg L−1 min−1	r (H2) UV/vis μmol h−1
Photolysis	-	-	-	0.08	0.05	-
Ag/Ag2O	in situ	Ag/Ag2O	2.7	2.64	1.17	-
TM 41	mechanical mixture	Ag/Ag2O ⁄⁄ TiO2 (20% TiO2)	9.7	0.12	0.17	-
TM 11	mechanical mixture	Ag/Ag2O ⁄⁄ TiO2 (50% TiO2)	22.6	0.09	0.03	9
TM 14	mechanical mixture	Ag/Ag2O ⁄⁄ TiO2 (80% TiO2)	38.5	0.55	0.03	17
TiO2	-	TiO2	50	3.08	0.03	5
TC 41	in situ	Ag/Ag2O ⁄⁄ TiO2 (20% TiO2)	8.4	0.81	0.61	3
TC 11	in situ	Ag/Ag2O ⁄⁄ TiO2 (50% TiO2)	20.1	2.03	0.27	28
TC 14	in situ	Ag/Ag2O ⁄⁄ TiO2 (80% TiO2)	22.1	1.00	0.05	8

), resulting in a SSA of 38.5 m² g⁻¹ for TM 14.

SEM images of the TC composites are presented in Figure 2f–h. The image of the TiO₂-poor composite TC 41 clearly shows the large Ag/Ag₂O particles covered with TiO₂ (Figure 2f). The specific surface area of this composite was determined to be 8.4 m² g⁻¹, thus being equal within the limits of the experimental error to the surface area of the corresponding physical mixture TC 41 (SSA = 9.7 m² g⁻¹). The images of the composites richer in TiO₂ (TC 11 and TC 14) seemed to be dominated by aggregates or agglomerates of small TiO₂ particles.

The optical properties of TiO₂ and the as-prepared Ag-containing mixtures and composites were investigated by UV/vis diffuse reflectance spectroscopy (Figure 3). Ag/Ag₂O, as well as the TM, and TC materials, had a dark brown to black color. They displayed strong absorption over the whole UV and visible range (200 nm–800 nm). TiO₂ showed only the absorption band below 405 nm, which matches the band gap energy of 3.06 eV calculated from the formula λ = 1239.8/E_bg due to the charge transfer from O (valence band) to Ti (conduction band).

Ag/Ag₂O exhibited a band gap energy < 1.5 eV, which is in agreement with the reported value of 1.3 ± 0.3 eV [40]. The scattering of the reported values might be due to different particle diameters, as shown for TiO₂ [41]. Electrochemical measurements in suspensions yielded flat band potentials of −0.4 V and +0.3 V vs. NHE for TiO₂ and Ag₂O, respectively. The value measured here for the flat band potential of Ag₂O is also in reasonably good agreement with published values [42,43].

2.2. Photocatalytic Performance of the Materials

The photocatalytic activity of all materials described above was investigated, employing methylene blue (MB) as the probe compound. The materials in aqueous suspensions were excited by the full output of a xenon lamp (UV/vis illumination), and by Xe light after passing a UV cut-off filter (≥410 nm, vis illumination). Figure 4 illustrates the bleaching of an aqueous solution of MB and the MB-containing suspensions. Photolysis of MB (initiated by the direct excitation of the probe compound) was observed under both UV/vis and visible light illumination. The bleaching of MB was significantly accelerated by the presence of Ag/Ag₂O. Under UV/vis illumination, Ag/Ag₂O was found to be nearly as active as TiO₂ (P25), which is well known to be a very efficient photocatalyst suitable to degrade MB [44] (Figure 4a). In the presence of Ag/Ag₂O, MB was bleached very rapidly even when exposed to visible light. As expected, TiO₂, having a bandgap energy of 3.1 eV, was found to be inactive under vis illumination (Figure 4c).

In the presence of mixtures of Ag/Ag₂O with TiO₂, MB was bleached under UV/vis illumination only in the presence of the TiO₂-rich TM 14, with a significantly increased rate compared to the rate of MB photolysis. In suspensions containing TM 41 and TM 11, the rate of bleaching was almost the same as the rate of photolysis (Figure 4a). Exposure to visible light in the presence of the Ag/Ag₂O-rich TM 41 resulted in bleaching of MB with a slightly increased rate. In contrast, the TiO₂-rich mixtures TM 11 and TM 14 were virtually inactive under this illumination condition (Figure 4c).

In the presence of the composites TC, MB was bleached with significantly faster reaction rates than the rate of photolysis when exposed to UV/vis and visible light. The rates were, however, lower than the rate of bleaching in the presence of the bare TiO₂ (Figure 4b,d). Interestingly, while increasing the amount of TiO₂ in the TC composites, the visible light activity of the materials seemed to decrease, thus confirming the essential influence of Ag/Ag₂O on MB bleaching under illumination with wavelengths ≥ 410 nm.

As a second test reaction for the activity of the materials, the UV/vis light-induced evolution of molecular hydrogen by reforming of aqueous methanol was used. Figure 5 shows the amount of H₂ vs. illumination time in the presence of TiO₂, Ag/Ag₂O, and the prepared mixtures and composites. No H₂ evolution was observed in the presence of Ag/Ag₂O and the Ag/Ag₂O-rich TM 41. In the presence of all other materials, the evolution of H₂ was detected. However, large amounts of H₂ were only evolved with the materials TM 14 (104 µmol/6 h) and TC 11 (174 µmol/6 h).

Many authors have reported that the kinetic behavior of photocatalytic reactions can be described by a Langmuir–Hinshelwood rate law, with the two limiting cases of zero-order and first-order kinetics [45,46]. To calculate the initial rates r₀ of the bleaching of methylene blue, first-order kinetics have been assumed (r₀ = kC₀). To determine the rate constant k, the data given in Figure 4 have therefore been fitted with C = C₀ exp(−kt). The initial rates are given in Table 1.

3. Discussion

3.1. The Photocatalytic Activity of Ag/Ag₂O

It is well known that methylene blue is photocatalytically oxidized in the presence of TiO₂ under illumination with photons having an energy equal to or larger than the bandgap energy of the semiconductor. The photocatalytic degradation of methylene blue in the presence of molecular oxygen is reported to follow Equation (1) [44].

$${\mathrm{C}}_{16}{\mathrm{H}}_{18}{\mathrm{N}}_3\mathrm{SCl}+25.5{\mathrm{O}}_2\stackrel{{\mathrm{TiO}}_2+\mathrm{h}\mathsf{\nu }\ge 3.2\mathrm{eV}}{\to }\mathrm{HCl}+{\mathrm{H}}_2{\mathrm{SO}}_4+3{\mathrm{HNO}}_3+16{\mathrm{CO}}_2+6{\mathrm{H}}_2\mathrm{O}$$

(1)

The energetic positions of the valence and conduction bands of TiO₂ and Ag₂O, and the reduction potentials of some species (possibly) present in the surrounding electrolyte are shown in Figure 6. As becomes obvious from this Figure, the conduction band electrons generated by UV illumination of TiO₂ are able to reduce O₂ adsorbed at the semiconductor surface. From a thermodynamic point of view, valence band holes at the TiO₂ surface have an energy suitable to oxidize H₂O/OH⁻, yielding OH radicals. These OH radicals are generally assumed to be the oxidizing species in photocatalytic MB degradation.

With the assumption that the flat band potential of Ag₂O, which has been determined to be + 0.3 V vs. NHE at pH 7, was equal to the conduction band edge of this semiconductor, and a bandgap energy E_g = 1.5 eV, the valence band position was calculated to be +2.0 V vs. NHE. Xu and Schoonen reported a value of +0.2 V vs. NHE for the energy of the Ag₂O conduction band [49]. As becomes obvious from Figure 6, excited Ag₂O was neither able to reduce O₂ nor to oxidize H₂O/OH⁻. Consequently, the mechanism of MB bleaching observed in the presence of Ag/Ag₂O (Figure 4 and Table 1) was different from the MB degradation mechanism in the presence of TiO₂. A possible explanation for the decolorization of MB in the presence of Ag/Ag₂O is that MB is excited by light of suitable wavelength (Equation (2), MB* = MB^S and or MB^T), which is subsequently followed by electron injection into the conduction band of Ag₂O (Equation (3)).

$$\mathrm{MB}\stackrel{\mathrm{h}\mathsf{\upsilon }}{\to }\mathrm{MB}\ast$$

(2)

$$\mathrm{MB}\ast +{\mathrm{Ag}}_2\mathrm{O}\to {\mathrm{MB}}^{+•}+{\mathrm{Ag}}_2\mathrm{O}\left\{{\mathrm{e}}^{-}\right\}$$

(3)

As an alternative to these reactions, the direct oxidation of MB by valence band holes according to

$${\mathrm{Ag}}_2\mathrm{O}\to {\mathrm{Ag}}_2\mathrm{O}\{{\mathrm{h}}^{+}+{\mathrm{e}}^{-}\}$$

(4)

$${\mathrm{Ag}}_2\mathrm{O}\{{\mathrm{h}}^{+}+{\mathrm{e}}^{-}\}+\mathrm{MB}\to {\mathrm{Ag}}_2\mathrm{O}\left\{{\mathrm{e}}^{-}\right\}+{\mathrm{MB}}^{+•}$$

(5)

has to be considered. Both mechanisms require an electron transfer between Ag₂O and MB. Despite the low surface area available for this reaction, the electron transfer between the solid and the probe compound appears to be very efficient.

It is well known that Ag₂O is sensitive to light and decomposes under illumination. However, it has been suggested that Ag(0) being present in Ag/Ag₂O acts as an electron sink and accepts the conduction band electron of Ag₂O, thus inhibiting the reduction of Ag⁺ and stabilizing the Ag₂O [9,10,12,20]. However, the possibility cannot be excluded that Ag⁺ is reduced during the processes given in the Equations (2)−(5), yielding Ag(0), since no other suitable electron acceptor is available. Regardless of whether the electrons reduce Ag⁺ or become stored in Ag(0), Ag/Ag₂O is not acting as a photocatalyst, because the material changes irreversibly during the reaction.

The potential of the Ag₂O conduction band electron is more positive than the reduction potential of the H⁺/H₂ couple (Figure 6). Consequently, light-induced proton reduction yielding H₂ is thermodynamically impossible in suspensions containing only Ag/Ag₂O. This is in accordance with the experimental results reported in Section 2.2.

3.2. The Photocatalytic Activity of Physical Ag/Ag₂O ⁄⁄ TiO₂ Mixtures

3.2.1. Bleaching of Methylene Blue

When irradiated with light at wavelengths ≥ 410 nm, methylene blue was found to be bleached in the presence of Ag/Ag₂O, and mixtures of this material with TiO₂. The rate of MB bleaching decreased with increasing amounts of TiO₂. Of course, TiO₂ itself was found to be photocatalytically inactive, since it was not excited under this illumination condition (Figure 4c and Table 1). The electron transfer reaction resulting in the observed bleaching of the MB solution occurred at the surface of the Ag₂O, as discussed in Section 3.1. According to the SEM images (Figure 2a–d), the surface of the Ag₂O was increasingly covered by TiO₂ as the content of this oxide in the mixture increased. The interfacial electron transfer was inhibited by this TiO₂ layer (Figure 7). The reaction rates suggest that this inhibition increased with increasing amounts of TiO₂ on the Ag/Ag₂O surface. Consequently, the TiO₂-rich mixtures TM 11 and TM 14 exhibited rates of bleaching almost the same as the rate of photolysis in homogeneous solution (Table 1). Interfacial electron transfer from excited MB to TiO₂ (which is thermodynamically possible; cf. Figure 6) obviously did not contribute significantly, since no MB bleaching was observed under visible light illumination of suspensions containing only this photocatalyst.

The situation was different when the TM mixtures were illuminated with UV/vis light. The rate of MB bleaching in the presence of the Ag/Ag₂O ⁄⁄ TiO₂ mixtures was found to increase with increasing TiO₂ content. However, the rates were always lower than the rates determined for suspensions containing only Ag/Ag₂O or bare TiO₂ (Figure 4a and Table 1). These rates cannot be explained solely by the optical properties of the suspensions. Of course, as the Ag/Ag₂O content increases, more UV photons are absorbed by Ag₂O. They are thus no longer available for the excitation of the TiO₂ that results in decreasing amounts of charge carriers in the TiO₂ and, consequently, decreasing rates of MB degradation. However, the MB bleaching rate calculated for the TiO₂-rich TM 14 mixture suggests that not all photogenerated charge carriers were used in the desired MB bleaching reaction, but some were lost by reactions between excited TiO₂ and Ag/Ag₂O, resulting in the reduction of Ag⁺.

XRD measurements revealed the reduction of Ag⁺ during the light-induced bleaching of MB under UV/vis illumination. The ratios of the peak intensities corresponding to Ag₂O and TiO₂ of the mixture TM 41 and the composite TC 41 were significantly lower after two experimental runs than before illumination (Figure 8). On the other hand, the ratios of the peak intensities attributed to metallic Ag and TiO₂ obviously increased. In the case of the Ag/Ag₂O ⁄⁄ TiO₂ mixture TM 11, apart from the TiO₂ peaks, the only visible XRD peaks could be assigned to AgCl and Ag(0) after illumination of a suspension containing MB (Figure S2). The new peaks in the diffractogram, which are indexed to AgCl, were possibly formed by a reaction between Ag⁺ and Cl⁻ known to be present at the surface of TiO₂ P25 [39]. This reaction certainly explains the decrease of the Ag₂O peaks in the diffractogram. However, this explanation does not exclude that Ag₂O is also transformed by a light-induced reduction reaction, yielding Ag(0).

The conclusion from the XRD data, that Ag(I) was reduced yielding Ag(0) during the light-induced bleaching of MB in the presence of the mixture TM 41, is supported by the results of the analysis of XPS data taken before and after two experimental runs (Figure 9a,b and Figure S3). It becomes obvious from Figure 9a that the Ag 3d_5/2 and Ag 3d_3/2 peaks of Ag₂O in the mixture TM 41 decreased in intensity and broadened, while the Ag(0) 3d_5/2 and Ag(0) 3d_3/2 peaks increased in intensity after two photocatalytic reactions. Furthermore, the deconvolution of the O 1s peaks denotes that the peak corresponding to the Ag-O bond had a lower intensity compared to the same peak observed before the reaction, indicating significant changes occurred during the light-induced MB bleaching reaction (Figure 9b). These changes were mainly due to the light-induced reduction of Ag⁺ yielding Ag(0). Again, the condition of stability of a catalyst was not satisfied.

3.2.2. Light-Induced Hydrogen Evolution

From a thermodynamic point of view, excited TiO₂ is able to transfer a conduction band electron to a proton present at the photocatalyst surface (Figure 6). This electron transfer is, however, known to be a kinetically inhibited process. Therefore, it is necessary to deposit an electrocatalyst at the TiO₂ surface, which accelerates the interfacial electron transfer. Ag(0) is known to be a suitable, though relatively inactive, electrocatalyst [50,51]. In this work as well, pure TiO₂ showed only a very low photocatalytic activity with regard to H₂ evolution from aqueous methanol. When using the TM materials, a significant increase in the amount of H₂ evolved (consequently corresponding with an increase in the reaction rate) during six hours of illumination of the mixture was observed with increasing TiO₂ content (Figure 5a and Table 1). On the one hand, this can be explained by the fact that a significant portion of the UV photons was absorbed by Ag₂O being inactive under this illumination condition, and thus was not available for the desired H₂ evolution reaction. However, this portion decreased with increasing TiO₂ amount of the mixture. On the other hand, some of the TiO₂ conduction band electrons were transferred to the Ag₂O, where they were consumed to reduce Ag⁺ to Ag(0). These electrons were therefore also not available for the desired reaction. Obviously, these undesired electron losses are lower the higher the mass fraction of TiO₂ in the physical mixture, resulting in increasing H₂ evolution rates with increasing mass fraction of TiO₂.

3.3. The Photocatalytic Activity of Ag/Ag₂O ⁄⁄ TiO₂ Composites

3.3.1. Bleaching of Methylene Blue

When irradiated with light at wavelengths ≥ 410 nm, methylene blue was found to be bleached in the presence of the three TC composites (Figure 4d and Table 1). All TC composites exhibited a higher activity than the corresponding TM mixtures. As in the case of the TM materials, the rate of MB bleaching decreased with increasing amounts of TiO₂. The increased reaction rates for MB bleaching in the presence of Ag₂O containing solids, compared to the rate of photolysis under visible light illumination, were explained in Section 3.2.1 with an interfacial electron transfer from (excited) MB to Ag₂O (cf. Figure 7). However, the experimental result is surprising when it is considered that the surfaces of the composites were smaller than the surfaces of the corresponding TM mixtures. A possible explanation may be due to the preparation method. For the TC materials, the Ag/Ag₂O was prepared in a TiO₂ suspension. Therefore, the Ag/Ag₂O was attached on the surface of the TiO₂ particles. In contrast, in the TM mixtures large Ag/Ag₂O particles were covered by TiO₂, hindering the electron transfer from excited MB to the Ag₂O, as discussed in Section 3.2.2.

The rate of MB bleaching in the presence of TC composite was significantly higher under UV/vis than under visible light illumination. As observed for the TM materials, the bleaching rates were lower in suspensions containing the composites than in suspensions containing only Ag/Ag₂O or TiO₂ (Figure 4b and Table 1).

XRD and XPS data indicate that Ag(I) was reduced, yielding Ag(0), during the light-induced bleaching reaction of MB in the presence of the composite TC 41. A stabilization of Ag₂O by metallic silver, as claimed by several authors [9,10,11,12,20,29], was not observed. No XRD peaks that can be attributed Ag₂O, were observed after two experimental runs of the composite. However, the ratios of the peak intensities due to metallic Ag and TiO₂ obviously increased (Figure 8b). No Ag 3d_5/2 and Ag 3d_3/2 peaks, which can be attributed to Ag(I), were present either in the deconvoluted XPS spectra obtained after two experimental runs (Figure 9c and Figure S3). The XPS peak, which was attributed to the presence of Ag-O, also disappeared during the light-induced reaction (Figure 9d and Figure S3).

These observations support the statement made above that Ag/Ag₂O cannot be called a photocatalyst. The XRD pattern shown in Figure 8b as well as the XPS data presented in the Figure 9c,d clearly evince that the Ag:Ag₂O ratio changed during the light-induced bleaching of MB. Thus, the condition for a catalyst to exit a chemical reaction unchanged is not satisfied.

3.3.2. Light-Induced Hydrogen Evolution

The three TC composites were found to be able to promote light-induced H₂ evolution from aqueous methanol. The calculated reaction rates were significantly larger than those of the corresponding TM mixtures. The highest H₂ evolution rate was observed in the presence of TC 11 (Figure 5b and Table 1), which was also characterized by a high MB bleaching rate under UV/vis illumination. A possible mechanistic explanation for the high activity of the TC 11 composite is based on the assumption of synergistic effects, due to the presence of both Ag(0) and Ag₂O at the TiO₂ surface (Figure 10). TiO₂ is excited by UV photons. The photogenerated conduction band electrons migrated to the Ag(0) attached to the TiO₂ surface. In a subsequent step, interfacial electron transfer from Ag(0) to protons present in the surrounding electrolyte occurred, thus yielding molecular hydrogen. The valence band hole inside the TiO₂ particle was filled by an electron from an attached Ag₂O particle. Methanol was oxidized by this hole in the valence band of the Ag₂O. According to this mechanism, Ag(0) acts as an electron sink, thus decreasing the electron-hole recombination, and as electrocatalyst for the hydrogen evolution reaction, while Ag₂O is an electrocatalyst for the oxidation reaction of methanol yielding methanal. The supposition made here, that the methanol oxidation occurs at the Ag₂O surface via electron transfer to the valence band of the excited TiO₂, has already been proclaimed earlier [16,19,23,26]. It should be emphasized again that the energy of an electron in the conduction band of the Ag₂O employed in this study is insufficient to reduce a proton (Figure 6). Consequently, excitation of TiO₂ is a prerequisite for photocatalytic reforming of methanol. TiO₂ is known to be a relatively inactive material for the photocatalytic reduction of protons. High evolution rates of molecular hydrogen are observed only in the presence of a co-catalyst. Ag₂O was found here to be an unsuitable co-catalyst for the hydrogen evolution reaction, since electron transfer from the excited TiO₂ can only occur into the conduction band of this material. The photocatalytic activities of the composites and mixtures discussed here are thus determined to a considerable extent by the competition between interfacial electron transfer to protons in the surrounding electrolyte, and to silver ions in Ag₂O. The mechanism of the photocatalytic hydrogen evolution by reforming of organic compounds in the presence of the mixtures and composites employed in this study does not contradict the mechanism discussed for Ag/Ag₂O ⁄⁄ TiO₂ samples, which contain Ag₂O with a significantly more negative conduction band energy than TiO₂ [17,24,26,33].

Changes in the respective mass fractions of TiO₂, Ag, and Ag₂O at constant total mass of the solid in suspension may have several impacts on the rate of hydrogen evolution. Increasing mass fractions of UV absorbing and scattering Ag and Ag₂O reduces the number of photons to be absorbed by the TiO₂, thus reducing the H₂ evolution rate. A reduction of the mass fraction of metallic Ag may possibly slow down the interfacial electron transfer to the proton, while a reduction of the mass fraction of Ag₂O might negatively affect the oxidation reaction. It should also be noted that Ag₂O can act as a sink for a TiO₂ conduction band electron (cf. Figure 6). These partially opposing effects may be responsible for the observed differences in the H₂ evolution rates in the presence of the various TC composites (and TM mixtures).

4. Experimental Section

4.1. Materials

Titania P25 (TiO₂) with a mixture of anatase (80%) and rutile (20%) crystal phase, and a specific surface area of 50.1 m² g⁻¹, was kindly provided by Evonik, Essen, Germany. Silver nitrate (99%, Sigma Aldrich Chemie GmbH, München, Germany), sodium hydroxide pellets (99%, Carl Roth, Karlsruhe, Germany), methanol (99.9%, Carl Roth), and methylene blue (Sigma Aldrich) were used without further purification. Deionized water with a resistivity of 18.2 MΩ·cm was obtained from a Sartorius Arium 611 device (Sartorius Göttingen, Germany) and used for the preparation of all aqueous solutions and suspensions.

4.2. Synthetic Methods

4.2.1. Preparation of Ag/Ag₂O

An amount of AgNO₃ was dissolved in 50 mL of distilled water. The obtained solution was stirred for 30 min. Subsequently, 50 mL NaOH (0.2 M) was added dropwise. The resulting suspension was stirred for another 30 min to promote hydrolysis, and centrifuged, washed with distilled water three times, and dried at 70 °C for 24 h.

4.2.2. Preparation of TM Mixtures

The samples were obtained by mixing the self-prepared Ag₂O with TiO₂ at mass ratios of 4:1 (20 mass% TiO₂), 1:1 (50 mass% TiO₂), and 1:4 (20 mass% TiO₂) with water. The suspensions were sonicated for 1.5 h and dried at 70 °C for 24 h. The Ag/Ag₂O ⁄⁄ TiO₂ with 20%, 50%, and 80% of TiO₂ were nominated as TM 41, TM 11, and TM 14, respectively. For purpose of comparison, a TiO₂ sample was prepared by the same procedure without the addition of Ag/Ag₂O.

4.2.3. Preparation of TC Composites

The TC composites were prepared by a published precipitation method [25,29]. A measured amount of TiO₂ was suspended in 50 mL of distilled water, and the calculated amount of AgNO₃ corresponding to the desired mass ratio of Ag₂O was added to the solution. The obtained suspension was stirred for 30 min. A volume of 50 mL 0.2 M NaOH was added dropwise. The resulting suspension was stirred for another 30 min to promote hydrolysis and centrifuged, washed with distilled water three times and dried at 70 °C for 24 h. The Ag/Ag₂O ⁄⁄ TiO₂ with 20 mass%, 50 mass%, and 80 mass% of TiO₂ were denoted as TC 41, TC 11, and TC 14, respectively.

4.3. Characterization of the Materials

The crystalline structure of the catalysts was measured by powder X-ray diffraction XRD (D8 Advance system, Bruker, Billerica, MA, USA), using a Cu Kα radiation source with a wavelength of λ = 0.154178 Å over a 2θ range from 20° to 100°, with a 0.011° step width. The morphology of the prepared materials was determined using a scanning electron microscope (SEM), employing a JEOL JSM-6700F field emission instrument (Tokyo, Japan) with a resolution of 100 nm and 1 µm using an EDXS detector. Measurements of X-ray photoelectron spectra were carried out using a Leybold Heraeus (Cologne, Germany) with X-ray source, Mg & Al anode, nonmonochromatic, hemispherical analyzer, 100 mm radius. Data analysis was performed using XPSPEAK 4.1 software (Hong Kong, China). The energy of the C1s-line was set to 284.8 eV and used as reference for the data correction. Diffuse reflectance UV–Vis spectroscopy was employed using a spectrophotometer (Varian Spectrophotometer Cary-100 Bio, Agilent technologies, Santa Clara, CA, USA) at room temperature. Barium sulfate was used as a standard for 100% reflection. The specific surface area (SSA) of the samples was calculated by N₂ adsorption–desorption measurements, employing the Brunauer-Emmet-Teller (BET) method using a FlowSorb II 2300 apparatus from Micromeritics Instrument Company (Corp., Norcross, GA, USA). Prior to these measurements, the samples were evacuated at 180 °C for 1 h. Measurements of photocurrents and flat band potentials were performed with an electrochemical analyzer using three electrodes employing an Iviumstat potentiostat (Ivium Technologies bv, Eindhoven, The Netherlands). Films of the samples were used as the working electrode, after being coated on cleaned fluorine doped tin oxide (FTO) coated glass using the doctor blade method and calcinated at 400 °C for 2 h. These working electrodes were prepared by grinding 100 mg of the photocatalysts and 50 mg polyethylene glycol with one drop of Triton, followed by addition of 200 μL of deionized water and a sufficient amount of ethanol. An Ag/AgCl electrode (3 M NaCl, +209 mV vs. NHE) and a platinum coil were used as the reference electrode and the counter electrode, respectively. Potassium nitrate aqueous solution (0.1 M) was used as the electrolyte. The impedance spectra were recorded in the range between the chosen potential from −1 V to +1 V at frequencies of 10, 100, and 1000 Hz with 20 mV amplitude vs. Ag/AgCl. The capacitance was plotted against V, and the flat band was calculated from the intercept of the plot. (i.e., a plot of C⁻² vs. V, where C was the capacitance and V was the potential across the space charge layer).

4.4. Photocatalytic Measurements

4.4.1. Methylene Blue Degradation

The apparatus used for carrying out of the photocatalytic degradation reactions consisted of a double jacket cylindrical reactor with a 230 mL volume, which circulated with cold water to maintain the ambient reaction. A volume of 200 mL of aqueous solution of methylene blue (MB, 10 mg L⁻¹) and 200 mg of photocatalysts were used for each reaction experiment. A 300 W Xenon arc lamp (Müller Electronik-Optik, Moosinning, Germany) was used both as the UV/vis light source and as the vis light source by placing a UV cut-off filter (≥410 nm) in the light path. The lamp was started 30 min before the degradation experiments to ensure maximum emission. Aliquots (1.5 mL) of the suspensions were collected at given time intervals (0, 2, 4, 6, 8, 10, 15, and 30 min), centrifuged to remove the solid, and analyzed immediately with the UV–Vis spectrophotometer.

4.4.2. Photocatalytic Hydrogen Formation

The photocatalytic H₂ generation experiments were conducted in quartz vials (capacity of 10 mL) under illumination with a 1000 W Xenon lamp (Hönle UV Technology, Gräfelfing, Germany; Sol 1200 solar). An amount of 6 mg of the photocatalyst was suspended in 6 mL aqueous methanol (10 vol% methanol). The suspension was purged with argon for 20 min to remove the air, and the quartz vial was sealed with a specially made rubber septum degassed for sampling. The amount of H₂ gas evolved during the photocatalytic reaction was quantified every two hours using a gas chromatograph (Shimadzu GC-8A, Shimadzu Deutschland GmbH, Duisburg, Germany) equipped with thermal conductivity detector (TCD) and 60/80 molecular sieve 5 Å column.

5. Conclusions

Ag/Ag₂O was found to enhance the rate of light-induced bleaching of aqueous methylene blue under both UV/vis and vis illumination, in comparison to the bleaching in homogeneous solution. Even in suspensions containing mixtures and composites of Ag/Ag₂O with TiO₂ (P25), with varying mass ratios of Ag/Ag₂O (20%, 50%, and 80%), the reaction rate was slightly increased under these illumination conditions. However, the bleaching rate of methylene blue was lower in the presence of the composites and mixtures than the rate measured for bare Ag/Ag₂O. It is therefore suggested that the bleaching of methylene blue is initiated by an interfacial electron transfer from the excited organic probe compound to Ag₂O. TiO₂ layers covering the Ag₂O seem to inhibit this electron transfer. Since Ag₂O can transfer an electron neither to dissolved molecular oxygen nor to a proton for thermodynamic reasons, it is assumed that Ag⁺ is reduced to Ag(0) in the processes investigated here. Results of XRD and XPS measurements support this assumption, and indicate that Ag/Ag₂O is not stable under the experimental conditions employed in this study. A stabilization of Ag₂O by metallic silver, as occasionally claimed, was not observed. Therefore, to address Ag/Ag₂O as a (photo)catalytically active material does not seem appropriate.