The effect of silver oxidation on the photocatalytic activity of Ag/ZnO hybrid plasmonic/metal-oxide nanostructures under visible light and in the dark

A new synergetic hybrid Ag/ZnO nanostructure was fabricated which is able to cause photocatalytic degradation (in high yields) of methylene blue under visible light as well as in the dark. In this nanostructure, ZnO was synthesized using the arc discharge method in water and was coupled with Ag via a chemical reduction method. X-ray photoelectron spectroscopy (XPS) and photoluminescence spectroscopy results confirmed the existence of defects in ZnO in the hybrid nanostructures; these defects act as electron traps and inhibit the recombination of electron-hole pairs. The absorption edge of the hybrid nanostructure shifts toward the visible region of the spectrum due to a combination of the Ag plasmonic effect and the defects in ZnO. Band-edge tuning causes effective visible light absorption and enhances the dye degradation efficiency of Ag/ZnO nanostructures. Silver oxidation in the hetero-structure changed the metal-semiconductor interface and suppressed the plasmonic enhancement. Nevertheless, the synthesized Ag/ZnO decomposed methylene blue in visible light, and the silver oxidation only affected the catalytic activity in the dark. This work provides insight into the design of a new and durable plasmonic-metal oxide nanocomposite with efficient dye degradation even without light illumination.


Results and Discussion
Morphology and chemical composition characterization. TEM images of ZnO and the as-synthesized Ag/ZnO HNS are shown in Fig. 1. Figure 1a illustrates pristine ZnO nanoparticles with spherical shape and Fig. 1b depicts aggregation of nanoparticles with addition of Ag in fresh hybrid sample. To study the chemical composition and distribution of Ag in the hybrid nanostructure, EDX analysis was performed. Figure 2 illustrates an HR-SEM image of the as-synthesized Ag/ZnO HNS and its corresponding EDX maps. The EDX maps clearly show very similar distribution of Zn and O atoms while the Ag atoms distribution is different. The areas that are enclosed by yellow and red boxes in Fig. 2, show the most Ag-concentrated regions in fresh Ag/ZnO HNS. In this research, aged Ag/ZnO HNS refers to the sample after being applied in photocatalytic degradation process, which there is AgO phase. A full discussion about different crystallographic phases in hybrid nanostructures is in the XRD section. Figure 3 shows high angular annular dark field (HAADF)-scanning TEM (STEM) image of aged Ag/ZnO HNS and the corresponding elemental mapping. Figure 3 shows very similar distribution of Zn and www.nature.com/scientificreports www.nature.com/scientificreports/ Ag atoms in the aged sample while the O atoms distributed almost everywhere. To investigate oxidation of silver atoms and the atomic structure in the aged Ag/ZnO HNS the high resolution TEM (HRTEM) analysis was performed. Figure 4a shows the HRTEM of an aged Ag/ZnO HNS, which shows silver and silver oxide nanoparticles decorating the ZnO nanostructure. Figure 4b is the HAADF-STEM image of Fig. 4a with the corresponding EDX

Ag O Zn
Electron image   Figure 4c,d,e shows the magnified view of the three regions in Fig. 4  XRD characterization. The crystallinity, phase and composition of the nanoparticles were determined with XRD analysis. Figure 5 shows the XRD patterns of pure ZnO, the as-synthesized, and aged Ag/ZnO HNS for 2θ from 20 to 80 degrees. Eight diffraction peaks of wurtzite ZnO were observed and indexed in Fig. 5. These peaks are in both pristine ZnO and in Ag/ZnO hybrid nanostructures without any shifts, thus suggesting that the hybridization with Ag nano-colloids did not alter the crystal structure of the ZnO domains. In Fig. 5b   XPS analysis. XPS measurements were used to identify the elemental composition and surface chemical state of the Ag/ZnO HNS. The peak positions were calibrated based on the carbon C 1 s peak, centered at binding energy of 284.8 eV. In Fig. 6a, a wide XPS spectrum displays the Zn 2p, O 1 s, Ag 3d and C 1 s peaks. No other elements were identified. All experimental data in the high-resolution XPS spectra were deconvoluted and fitted with Voigt profiles. Figure 6b reveals that the O 1 s peak in Ag/ZnO HNS is composed of two peaks at 530.6 and 532.6 eV. The peak with a lower binding energy is attributed to the O 2− lattice oxygen in the ZnO wurtzite crystal structure 31 . The peak near 532.6 eV is ascribed to the adsorbed oxygen species. This peak may be related to the adsorbed H 2 O, − OH, O − , O 2− , -CO ions that are on the surface 21,32-38 . In Fig. 6c, two characteristic peaks of Zn 2p 1/2 and 2p 3/2 are observed at 1044.7 eV and 1021.6 eV respectively. These two peaks are related to Zn 2+ due to the presence of Zn-O bonds 39 . These asymmetric peaks have shoulders in lower binding energies, and they are located at 1042.1 eV and 1019.0 eV. The shoulders with lower binding energies indicate that Zn is in a lower oxidation state than in ZnO. In Fig. 6d, the spectrum collected on the energy region that is typical for Ag 3d peaks is reported. Two sets of peaks are clearly visible: one has Ag 3d 3/2 and Ag 3d 5/2 peaks at 374.3 eV and 368.2 eV, which are assigned to the presence of metallic Ag; the other has Ag 3d peaks at 373.3 eV and 367.3 eV, which are linked to the presence of AgO. These findings are in agreement with previously reported results 40 .
Optical properties. To investigate the effects of silver coupling on the optical absorption and band edge energy of ZnO nanoparticles, UV-vis spectroscopy was performed. Figure 7 shows the absorption spectra of pure ZnO and Ag/ZnO HNS. ZnO shows a high absorption in UV region of the spectrum and a band-edge near 400 nm. It can be observed that the optical absorption increases significantly in the visible region of the spectrum by coupling the silver to the ZnO nanoparticles. The main feature of the optical absorption of Ag/ZnO HNS is a broad peak at 446 nm, which appears to be due to the LSPR of silver. The other peak, which is observed at a wavelength of 353 nm, is the exciton peak of ZnO nanoparticles 41,42 . In aged Ag/ZnO HNS, both plasmonic and excitonic peaks were damped. Furthermore, the plasmonic peak shifted to 428 nm due to the oxidation of the www.nature.com/scientificreports www.nature.com/scientificreports/ silver nanoparticles 43 . High visible light absorption and effective charge carrier separation is required for efficient dye degradation. If the photo-generated electrons recombine very quickly, no electron-hole pairs would remain to carry out photocatalytic reactions. A less effective dye degradation is predicted in the case of aged hybrid samples due to the fact that they absorb visible light less effectively than the fresh samples.
To study the electronic structure of the nanoparticles, a PL analysis was carried out. The room temperature PL spectra were recorded with an excitation wavelength of 320 nm. Figure 8 illustrates the normalized PL spectra of the nanostructures. Normalizations were calculated according to the absorption spectra of different samples at 320 nm. ZnO and Ag/ZnO HNS both exhibit two distinct UV emissions. These two UV peaks appear as a result of near band edge emissions (NBE), but they are attributed to different effects in the nanostructures. The emission peak at 364 nm originates from a direct recombination of free excitons (FEs), while the emission peak at 378 nm is related to bound excitons (BEs) 44 . This recombination feature, which is caused by BEs, has been observed previously but its origin is still unknown [45][46][47][48] . It has been ascribed to different BEs from crystal defects such as donor BEs 49,50 , acceptor BEs 51 , and excitons bound to extended structural defects 46,52 . As Fig. 6 depicts, the dominant emission in ZnO is from the BEs, while the FEs show a more intense emission in the Ag/ZnO samples. This means that the Ag coupling, increased free exciton emissions and quenched the bound exciton emissions in the hybrid nanostructures. Other features of the PL analysis include several emissions in the visible region of the spectra from 500 nm to 600 nm. Figure 8 shows that there are several green and yellow emission peaks in both the ZnO and Ag/ZnO HNS at 505 nm, 523 nm, 545 nm, and 595 nm. These peaks are due to deep level emissions (DLE) from different crystal defects 53,54 . The reason for the green emission in ZnO is not yet clear; it is, in fact, a highly controversial topic. Some calculations suggest that oxygen vacancies cause the green emission 53,55 , while others conclude that it is due to oxygen antisite defects 56,57 , and a few reports ascribe it to a zinc vacancy 58   www.nature.com/scientificreports www.nature.com/scientificreports/ suggested that the yellow emission near 598 nm originates from oxygen interstitial defects 59,60 . In Fig. 8, the intensities of these visible emissions in hybrid Ag/ZnO are quite similar to that of pristine ZnO. Generally, in doped Ag/ZnO nanostructures, Ag ions may place in substitutional Zn 2+ or in interstitial positions, and this creates more lattice defects. This leads to more electron-hole separations, and it quenches the visible emissions 61 . In this research, we did not observe any significant changes in visible emissions with Ag coupling. However, any change of the PL emissions in Ag/ZnO HNS depends on the band alignment and Schottcky barrier properties.
Photocatalytic activity. The photocatalytic activity of pure ZnO and Ag/ZnO HNS was evaluated by MB dye degradation in the dark and under solar irradiation. At low concentrations of the dye, the degradation kinetics of the MB can be described using the first-order Langmuir equation 62 : in which t is the time of the reaction, C 0 is the initial concentration of MB, C is the concentration at different times after exposure, and k is the first order rate constant of the degradation reaction. The photocatalytic activity and MB degradation kinetics for Ag/ZnO HNS is presented in Fig. 9. Figure 9a,b shows the absorption spectra of MB during degradation process by fresh and aged Ag/ZnO HNS respectively. It should be noticed that the degradation process by the fresh sample was carried out without light illumination and was completed after 30 minutes. For aged Ag/ZnO HNS the degradation was carried out under solar simulator and it took 120 minutes to remove the dye. Figure 9c shows C/C 0 plot versus irradiation time for ZnO and Ag/ZnO HNS. From Fig. 9c it can be clearly seen that ZnO shows highly efficient visible light degradation that might be due to the presence of defects as well as to its oxygen rich surface. Coupling ZnO with Ag further enhanced the efficiency of the dye degradation in the dark, and this could be as a result of the electron trap states that were introduced by Ag. In addition, the combination of defective ZnO and LSPR of Ag nanoparticles makes the Ag/ZnO HNS an interesting candidate for dye degradation in visible light and dark. However, aged Ag/ZnO HNS shows less degradation efficiency due to oxidation of silver and less visible light absorption. Figure 9d shows the degradation reaction kinetics, the reaction rate constant was measured from the slopes of the −ln(C/C 0 ) plots versus time. The calculated k values obtained 5.09 × 10 −2 , 12.78 × 10 −2 , and 2.10 × 10 −2 min −1 for ZnO, fresh Ag/ZnO and aged Ag/ZnO respectively. The mechanism behind the photocatalytic degradation activity of hybrid Ag/ZnO nanostructures can be explained as follows. For the fresh sample, there are three possible electron transfers at the metal-semiconductor interface, which is explained schematically in Fig. 10a (the energy levels are taken from ref. 63 ). The first one is a direct electron transfer from Ag to ZnO: the electrons in the conduction band of the Ag nanoparticles are excited to the SPR state, then the excited electrons will transfer to the ZnO conduction band. In second mechanism, the electrons in the ZnO conduction band may transfer to the Fermi state of the Ag nanoparticles. In the third mechanism, the electron transfer depends on the position of energy levels of different defects in ZnO, and the electrons in the Fermi state of the Ag nanoparticles may transfer to the defect levels of ZnO. Electron transfer through these mechanisms will inhibit the recombination of charge carrier in fresh Ag/ZnO HNS. However, the electron transfer in aged Ag/ZnO HNS is more complicated due to the oxidation of silver. In this case, there are several metal semiconductor interfaces between ZnO, Ag, AgO and Zn. From the XRD results, it was found that the aged Ag/ZnO HNS are mostly composed of ZnO and AgO. Therefore the ZnO/AgO semiconductor/semiconductor interface is likely the dominant one. Figure 10b shows the schematic band diagram of AgO/ZnO (the AgO energy levels were taken from ref. 64 ). AgO is a p-type semiconductor with a 1.7 eV band gap energy 64,65 . When silver is oxidized in the HNS, an AgO/ZnO p-n nano heterojunction is formed, and the electrons will transfer from ZnO to AgO. In the p-n junction, the electrons transfer from an n-type ZnO region to a p-type AgO region in order to reach a thermal equilibrium. Therefore, ZnO is positively charged, while AgO has a negative charge, and an electric field is generated at the junction. As is shown in Fig. 10b, the excited electrons in the AgO conduction band may transfer to the ZnO conduction band, and holes can transfer to the AgO valence band due to the formation of an inner electric field at the p-n junction.
Degradation mechanism. To investigate the exact mechanism of the MB dye removing in the dark and to make sure whether it is surface adsorption or degradation, series of experiments were carried out. In order to study desorption process of MB from the surface of the fresh Ag/ZnO, we used combination of thermal and chemical regeneration technique [66][67][68][69] . In this regeneration method, the sample is heated after the dye adsorption and the desorption process performed using ethanol or acetonitrile 70 . Typically, samples with high surface area exhibit a high adsorption capacity of the dye. In our case, the surface area of the samples obtained 32.3, 18.7, and 28.0 m 2 /g for ZnO, fresh Ag/ZnO, and aged Ag/ZnO (BET, N 2 ) respectively. As the BET surface area values are not high we do not expect high adsorption capacity of the MB by the samples.
In the desorption experiment, after complete dye removing the photocatalyst was separated from the dye and heated at 160 °C for 1 hour. Then the powder redispersed in ethanol and stirred for 2 hours. Every 15 minutes 1.5 ml of the solution collected, centrifuged and the supernatant used for the spectroscopy. The optical absorption spectrum of the desorbed MB did not change during 2 hours of the measurement. Figure 11 a shows the optical absorption spectra of the initial MB before adsorption and the desorbed dye from the sample. From Fig. 11 a, one can conclude that the concentration of the desorbed MB from the sample is very low which did not change after 2 hours. Hence, the ethanol contains very little amount of desorbed MB from the sample while Fig. 9 a shows that the MB was completely removed after 30 minutes without light illumination and the color of the nano-powder did not changed. Therefore, in this experiment adsorption should not be the dominant effect in the MB removing.
In situ capture experiments also carried out to evaluate the MB degradation under visible light and without illumination. Different scavengers were used to determine the active species generated during the degradation process. Isopropyl alcohol (IPA), methanol, and AgNO 3 were used to capture the generated OH radicals, holes, and electrons during the reaction. Dye degradation experiment was performed in the presence of scavengers with 10 mM concentration. As shown in Fig. 11b, by using scavengers, degradation of the MB completely quenched in the dark. This result reveals that all the generated reactive species play an important role in the MB degradation activity and capturing each of them will eliminate the dye degradation. The use of hole and hydroxyl radical scavenger, also hindered the photocatalytic degradation under visible light. In case of AgNO 3 , as an electron scavenger www.nature.com/scientificreports www.nature.com/scientificreports/ under the visible light illumination, 40% of the dye was degraded and it is probably due to reducing of silver nitrate under illumination and loading some amount of metallic silver which may cause regeneration of surface electrons and increasing the degradation activity.
Performing these batch experiments, we can conclude that the dye degradation in the dark is due to generation of different active species on the surface of the fresh Ag/ZnO sample which lead to a high degradation efficiency.

Conclusion
We synthesized ZnO nanoparticles using an arc discharge method in water, and successfully coupled them with Ag nanoparticles by means of a chemical reduction synthesis. XRD results revealed that fresh Ag/ZnO HNS were composed of wurtzite ZnO and metallic silver, but this composition changed as the sample aged. In addition to the ZnO and Ag peaks, a metallic Zn peak and two other peaks ascribed to silver oxide were also observed in the XRD pattern of aged Ag/ZnO HNS. Ag/ZnO HNS efficiently degraded the dye both under visible light irradiation and in the dark, without the involvement of illumination. Ag plasmonic effects, together with defects in the ZnO, lead to a high yield photocatalytic activity in degradation of MB. Under visible light irradiation, the enhanced degradation activity can be attributed to the higher visible light absorption of Ag/ZnO HNS. However, the defect levels in the ZnO and metallic Ag electron traps are responsible for the effective charge separation and photocatalytic activity in the dark. The present work provides a method for synthesis a plasmonic/metal-oxide nanostructure with efficient dye degradation in visible light and dark.

Methods
Synthesis of ZnO nanoparticles. Zinc rods (99.9%), trisodium citrate dehydrate (>98%) and silver nitrate (>99%) were purchased from sigma-aldrich. ZnO nanoparticles were prepared using a DC arc discharge method in water. The complete synthesis and oxidation process is discussed elsewhere 71 . Two zinc rods, which were used as electrodes, were immersed in 200 ml of deionized water. After applying a 100 A current to the electrodes, the electric discharge created a plasma in the short distance between the electrodes. The generated plasma caused the zinc electrodes to evaporate, creating ZnO nanoparticles. After complete oxidation in water, a white ZnO precipitate was collected. This precipitate was washed with ethanol and acetone several times before finally being dried at 40 °C for 3 hours.

Synthesis of Ag/ZnO hybrid nanostructures.
For the synthesis of Ag/ZnO HNS, colloidal Ag nanoparticles were prepared using a chemical reduction method. In this procedure, 36 mg of AgNO 3 was dissolved in 200 ml of deionized water. This solution was then heated to 100 °C. Next, 4 ml of trisodium citrate 1 wt. % was added as a reducing agent, and the temperature of the solution was kept at 100 °C for one hour. The solution was then stirred and cooled down to room temperature, when it turned to a yellowish color. After preparing Ag colloidal nanoparticles, 20 mg of ZnO nanoparticles was dispersed in 20 ml of the prepared Ag colloid via a sonication process for 30 minutes. The solution was stirred for 24 hours at room temperature to obtain a brown precipitate. Finally, the precipitate was centrifuged, collected, and washed with ethanol and deionized water several times.
Photocatalytic activity. The photocatalytic activity of the Ag/ZnO hybrid nanoparticles was determined by measuring the degradation of the MB dye under solar irradiation. 3 mg/L of an aqueous dye solution was prepared in deionized water, and this was used for all degradation measurements. In each test, 3 mg of the photocatalyst was added to 30 ml of the MB solution, and the solution was stirred in the dark for 60 minutes to reach an adsorption-desorption equilibrium. After keeping the solution in the dark for 60 minutes, irradiation was performed with visible light under a 100 W xenon lamp with a power density of 100 mW cm −2 . Every 15 minutes, 3 ml of the solution was removed and centrifuged at 10000 rpm for 5 minutes, and UV-vis spectroscopy was recorded for the supernatant.