Post-illumination activity of SnO2 nanoparticle-decorated Cu2O nanocubes by H2O2 production in dark from photocatalytic “memory”

Most photocatalysts only function under illumination, while many potential applications require continuous activities in dark. Thus, novel photocatalysts should be developed, which could store part of their photoactivity in “memory” under illumination and then be active from this “memory” after the illumination is turned off for an extended period of time. Here a novel composite photocatalyst of SnO2 nanoparticle-decorated Cu2O nanocubes is developed. Their large conduction band potential difference and the inner electrostatic field formed in the p-n heterojunction provide a strong driving force for photogenerated electrons to move from Cu2O to SnO2 under visible light illumination, which could then be released to react with O2 in dark to produce H2O2 for its post-illumination activity. This work demonstrates that the selection of decoration components for photocatalysts with the post-illumination photocatalytic “memory” could be largely expanded to semiconductors with conduction band potentials less positive than the two-electron reduction potential of O2.

Scientific RepoRts | 6:20878 | DOI: 10.1038/srep20878 produce • OH. Due to the gradual release of trapped photogenerated electrons, their activity in dark could last for more than 10 h, which is desirable for the construction of continuous solar-powered photocatalytic disinfection/ degradation systems effective for both daytime and at night.
It is generally believed that the redox ability of photogenerated electrons and holes highly relied on the conduction and valence band potentials of the photocatalyst 20 . The one-, two-, and four-electron reduction potentials of O 2 could be expressed as reactions (1) to (3) as following 8 : 2H O 3 2 2 Till now, photocatalysts with the post-illumination photocatalytic "memory" effect required that photogenerated electrons were trapped on decoration components with the conduction band potential negative than the one-electron reduction potential of O 2 (− 0.05 V vs NHE (Normal Hydrogen Electrode)) to react with O 2 in the dark to produce • O 2 − and subsequently • OH, which largely limited the selection of decoration components. Their relatively more negative conduction band potentials also lowered the potential difference between their conduction bands and that of the light absorber components. Thus, it would be interesting to examine if a decoration component with the conduction band potential less positive than the two-electron reduction potential of O 2 (0.68 V vs NHE) could be effective to generate activity from the production of H 2 O 2 in the dark, which could not only largely expand the the selection of potential decoration components but also increase the conduction band potential difference to enhance the driving force for the photogenerated electrons to be injected from the light absorber component's conduction band to that of the decoration component for their better transfer, trapping and subsequent release.
As an n-type, wide band gap semiconductor with interesting chemical, physical and mechanical properties, tin dioxide (SnO 2 ) had been extensively studied for applications in gas sensors, dye-based solar cells, transparent conducting electrodes, and catalyst supports 21,22 . The chemical state of Sn could exchange between Sn 2+ and Sn 4+ by trapping and release electrons, while it has a conduction band potential (0.4 V vs NHE) less positive than the two-electron reduction potential of O 2 22,23 . Thus, it could have the potential to serve as the decoration component in a composite photocatalyst system to trap the photogenerated electrons injected from the light absorber component, and release them in the dark by the reaction with O 2 to produce active H 2 O 2 to possess the post-illumination photocatalytic "memory" effect. In this work, we designed a novel Cu 2 O/SnO 2 composite photocatalyst composed of Cu 2 O nanocubes decorated with SnO 2 nanoparticles (Cu 2 O/SnO 2 ), in which Cu 2 O nanocubes served as the main light absorption component for a good visible light absorption capability while SnO 2 nanoparticles formed p-n heterojunctions of good contact with Cu 2 O nanocubes to serve as the decoration component. The large potential difference (~1.5 eV) between the conduction bands of Cu 2 O and SnO 2 23 , combined with the inner electrostatic field ξ formed in the p-n heterojunction, provided a strong driving force for the photogenerated electrons to move from Cu 2 O to SnO 2 through the heterojunction, which resulted in the enhanced photocatalytic performance under visible light illumination from better charge-carrier separation. The post-illumination photocatalytic "memory" effect was observed as expected for this composite Cu 2 O/SnO 2 photocatalyst, and the working mechanism was verified as the production of H 2 O 2 by the release of trapped photogenerated electrons from SnO 2 to react with O 2 in the dark.

Results
The formation and morphology of SnO 2 nanoparticle-decorated Cu 2 O nanocubes. The electron-hole pair recombination could be largely reduced in single crystal photocatalysts because they have much fewer defects compared with their polycrystalline counterparts, where the electron-hole pair recombination tends to occur 24 . Figure 1a shows the TEM image of the as-prepared Cu 2 O sample and the insert in Fig. 1a shows the corresponding selected area electron diffraction (SAED) pattern. It clearly demonstrated that the sample was composed of desirable single crystal nanocubes, which would favor the transportation of photogenerated electrons/holes and was ideal for constructing heterojunctions with SnO 2 nanoparticles in our material design. The average edge length of these Cu 2 O nanocubes was ~70 nm, and all their six exposed surfaces were {100} facets. Their fine nanosize could largely increase their specific surface area compared with their counterparts with submicron sizes 25 , beneficial to their contact efficiency with pollutants in water.
The control of the hydrolytic speed of tin precursors was critical for the formation of uniformly dispersed SnO 2 nanoparticles on these Cu 2 O nanocubes. In our approach, ethyl acetate (C 4 H 8 O 2 ) was chosen as the hydrolysis agent to get a good dispersion of SnO 2 nanoparticles onto the Cu 2 O nanocube surface. As shown in Fig. 1b, the Cu 2 O nanocube morphology was well preserved after the deposition and subsequent hydrothermal process to decorate SnO 2 nanoparticles onto the Cu 2 O nanocube surface and their crystallization. The surfaces of Cu 2 O nanocubes became relatively rough after the SnO 2 nanoparticle decoration. Figure 1c shows the TEM image of SnO 2 nanoparticle-decorated Cu 2 O nanocubes with a higher magnification. It demonstrated clearly that fine SnO 2 nanoparticles distributed uniformly on surfaces of Cu 2 O nanocubes, and their average size was ~5 nm. Figure 1d shows a representative HRTEM image of the Cu 2 O/SnO 2 interface area on these SnO 2 nanoparticle-decorated Cu 2 O nanocubes. The HRTEM image of the SnO 2 nanoparticle area verified their highly crystallized structure. One set of lattice planes could be clearly observed with the d-spacing at ~0.34 nm, corresponding to the (101) plane of the tetragonal rutile structure of SnO 2 phase. The HRTEM image of the Cu 2 O nanocube area also verified its highly crystallized structure. The electron beam was aligned along [001] direction, two sets of lattice planes could be clearly observed with the d-spacing at ~0.30 nm and ~0.21 nm, respectively, and their separation angle was ~45 o , corresponding to the (110) and (100) planes of the fcc Cu 2 O phase. The good crystallization of both Cu 2 O nanocubes and SnO 2 nanoparticles was beneficial to a good photocatalytic performance due to their lack of defects. The observed Cu 2 O/SnO 2 interface indicated that these SnO 2 nanoparticles grew on Cu 2 O nanocubes through our synthesis approach and p-n heterojunctions were formed with good contact between p-type Cu 2 O and n-type SnO 2 , beneficial to the photoexcited electron transfer between them.
Crystal structure and chemical composition of SnO 2 nanoparticle-decorated Cu 2 O nanocubes. Figure 2a shows the X-ray diffraction pattern of as-synthesized Cu 2 O nanocubes, compared with that of SnO 2 nanoparticle-decorated Cu 2 O nanocubes. For both samples, no diffraction peaks of CuO or Cu could be detected. All diffraction peaks in curve a belonged to the fcc Cu 2 O phase (PDF Card No. 05-0667), and the strong and sharp peaks indicated that these Cu 2 O nanocubes had a high degree of crystallinity. After the decoration with SnO 2 nanoparticles, several new diffraction peaks emerged in curve b, which could be readily indexed to tetragonal rutile structure of SnO 2 (PDF Card No. 41-1445). These peaks had relatively weak intensities due to the much smaller size of SnO 2 nanoparticles, compared with Cu 2 O nanocubes. No other diffraction peak could be observed, which confirmed that the final product was composed of Cu 2 O and SnO 2 . The amount of SnO 2 nanoparticles in the Cu 2 O/SnO 2 sample was measured by the sodium diethydlthiocabamate spectrophotometric method, and SnO 2 :Cu 2 O molar ratio was determined at ~0.15: 1.
The chemical composition and element valence states in SnO 2 nanoparticle-decorated Cu 2 O nanocubes were investigated by X-ray photoelectron spectroscopy (XPS). Figure 2b shows the XPS survey spectrum of the Cu 2 O/ SnO 2 sample, which demonstrated clearly the existence of Sn, O, and Cu in the sample. Due to the widespread presence of carbon in the environment, C 1s peak could also be observed in the XPS survey spectrum. Figure 2c shows the high resolution XPS spectrum over Cu 2p 3/2 peak. The main peak located at 932.7 eV could be attributed to the Cu + 2p 3/2 orbitals 25 . No obvious shake-up satellite peaks on the higher binding energy side could observed, which confirmed no existence of Cu 2+ on the sample surface 26,27 . Figure 2d shows the high-resolution XPS spectrum over Sn 3d 5/2 peak. It could be best fitted by the combination of two peaks centered at 486.5 eV and 485.5 eV, which could be assigned to Sn 4+ 3d 5/2 peak and Sn 2+ 3d 5/2 peak, respectively 28,29 . Thus, a small portion of Sn 4+ on the SnO 2 nanoparticle surface was reduced to Sn 2+ during the sample synthesis and storage under normal ambient condition. The Sn 2+ percentage was determined to be ~36%, while no SnO could be distinguished either in TEM or XRD analysis results. As a surface characterization technique, XPS could determine the surface composition within a very shallow depth. Thus, the existence of Sn 2+ must be on the very surface of SnO 2 nanoparticles, while the dominant Sn species in the sample existed as Sn 4+ . It had been well reported in literature that Sn 2+ state could be detected on the surface of SnO 2 nanoparticles due to the oxygen deficiency at the surface of SnO 2 28,29 . Optical properties of SnO 2 nanoparticle-decorated Cu 2 O nanocubes. The optical properties of SnO 2 nanoparticle-decorated Cu 2 O nanocubes were investigated by measuring their diffuse reflectance spectrum. From the reflectance data, optical absorbance could be approximated by the Kubelka-Munk function, as given by Eq. (4): where R is the diffuse reflectance 30 . Figure 3a shows the light absorbance (in term of Kubelka-Munk equivalent absorbance units) of the Cu 2 O/SnO 2 sample, compared with that of the as-synthesized Cu 2 O nanocubes and SnO 2 nanoparticles. SnO 2 nanoparticles demonstrated the characteristic spectrum with the fundamental absorbance stopping edge at ~350 nm, so most of their adsorption was within the UV light region 23 34 . This observation further

Photocatalytic disinfection of Staphylococcus aureus bacteria under visible light illumination.
The superior photocatalytic performance of SnO 2 nanoparticle-decorated Cu 2 O nanocubes was demonstrated by their photocatalytic disinfection effect on the viability of S. aureus cells, which is a common pathogenic coccus that could cause nonspecific infection and nosocomial infection 13,36 . The photocatalytic disinfection was conducted by exposing S. aureus cells suspended in 0.9% NaCl solution with the photocatalyst under visible light illumination for varying time intervals. The survival ratio of S. aureus was determined by the ratio of N t /N 0 , where N 0 and N t were the numbers of colony-forming units at the initial and each following time interval, respectively.   It must be pointed out that the total amount of H 2 O 2 produced should be much higher than the measured equilibrium concentration value because it was consumed in situ in the system with its generation. Figure 5b shows the H 2 O 2 concentrations in the test solution in the dark for an extended period of time up to 24 h after the visible light illumination was turned off by the Cu 2 O/SnO 2 sample, the as-synthesized Cu 2 O nanocubes, and SnO 2 nanoparticles, respectively. For the as-synthesized Cu 2 O nanocubes and SnO 2 nanoparticles, their production of H 2 O 2 was limited and the H 2 O 2 concentrations dropped quickly within the first 30 min in the dark. This observation was similar to the previous reports on Se 17 and Bi 18 , in which very limited charge carriers could remain at their surfaces, produce ROSs after the illumination was shut off, and be consumed quickly.   Figure 5c shows the high resolution XPS scan over Sn 3d peaks under visible light illumination. The Sn 4+ / Sn 2+ ratio was determined to be ~27:73. Compared to that without illumination as shown in Fig. 2d, a large part of Sn 4+ was reduced to Sn 2+ , which came from the transfer of photogenerated electrons from Cu 2 O to SnO 2 under visible light illumination and the subsequent trapping of part of these electrons by SnO 2 . After the light illumination was shut off, these trapped electrons could be gradually released and react with O 2 to produce H 2 O 2 as shown in Fig. 5b. To further confirm the production of H 2 O 2 and its major contribution to the disinfection of S. aureus cells by the Cu 2 O/SnO 2 sample in the dark after the illumination was shut off, a H 2 O 2 scavenger, EDTA-Fe(II), was used to examine if its existence could affect the survival ratio of S. aureus cells 38 . 0.1 mM EDTA-Fe(II) was added into the S. aureus cell suspension, and the pre-illuminated Cu 2 O/SnO 2 sample was used to conduct the disinfection experiment in the dark. As shown in Fig. 5b, the presence of EDTA-Fe(II) largely enhanced the survival ratio of S. aureus cells, which was very close to that treated by the Cu 2 O/SnO 2 sample without pre-illumination in the dark. This observation further confirmed that H 2 O 2 was the dominant ROS involved in the photocatalytic "memory" disinfection of S. aureus cells by the pre-illuminated Cu 2 O/SnO 2 sample in the dark. Figure 6 shows the proposed energy band structure of the Cu 2 O/SnO 2 p-n heterojunction, the photocatalytic activity enhancement mechanism under visible light illumination 39,40 , and the post-illumination photocatalytic "memory" mechanism in the dark. When p-type Cu 2 O and n-type SnO 2 formed a heterojunction, charge carrier concentration gradient occurred at the interface. Thus, the diffusion of electrons from SnO 2 to Cu 2 O and the diffusion of holes with the opposite direction happened until reaching the equilibrium, and an inner electric field (ξ ) was built at the interface as demonstrated in Fig. 6. Under visible light illumination, only Cu 2 O was excited to produce electron-hole pairs. The combined effect from both the large conduction band potential difference (~1.5 eV) and the inner electric field (ξ ) provided a strong driving force for photogenerated electrons to transfer from the conduction band of Cu 2 O to that of SnO 2 and be trapped there as verified by experimental evidences of the H 2 O 2 production and XPS analysis on Sn chemical status change from Sn 4+ to Sn 2+ . Thus, the photogenerated electron-hole pairs were separated effectively, and a largely enhanced photocatalytic performance was observed on the Cu 2 O/SnO 2 sample for its degradation of SMX and disinfection of S. aureus cells, compared with pure Cu 2 O nanocubes, which was very similar to our previous report on Cu 2 O-NS/TiO 2 -NI photocatalyst system 16 . When the visible light illumination was shut off, the trapped electrons could be released from SnO 2 and the two-electron reduction of O 2 could happen by its reaction with these released electrons due to their matched reduction potentials, which was verified by the continuous production of H 2 O 2 in the dark for more than 24 h. So the Cu 2 O/SnO 2 sample could demonstrate the post-illumination photocatalytic "memory" disinfection of S. aureus cells in the dark after the illumination was shut off.

Discussion
The ROS production in the dark by the Cu 2 O/SnO 2 sample developed in this study was quite different with that of previous reported photocatalysts with the post-illumination photocatalytic "memory" effect [14][15][16]19 . For TiON/PdO 14,15 , Cu 2 O-NS/TiO 2 -NI 16 , and I-TiO 2 19 , photogenerated electrons were released from the decoration components in the dark and one-electron reduction of O 2 happened to produce • O 2 − and subsequently • OH as ROSs, which required the decoration components had the conduction band potential negative than the one-electron reduction potential of O 2 . For the Cu 2 O/SnO 2 sample, however, photogenerated electrons were released from SnO 2 in the dark and two-electron reduction of O 2 happened to produce H 2 O 2 as ROS. Thus, this work demonstrated that a decoration component with the conduction band potential less positive than the two-electron reduction potential of O 2 could also be effective to generate the post-illumination photocatalytic "memory" effect from the production of H 2 O 2 in the dark.
This finding suggested that the selection of potential decoration components to construct photocatalyst systems with the post-illumination photocatalytic "memory" effect could be largely expanded to more semiconductors, such as SnO 2 , WO 3 , CuWO 4 , BiWO 6 , CeO 2 , etc. Although they do not have conduction band potentials negative than the one-electron reduction potential of O 2 , so trapped electrons released by them could not reduce O 2 in the dark to produce • O 2 − and subsequently • OH. However, electrons trapped on them could be released and then reduce O 2 in the dark to produce the reactive oxygen species of H 2 O 2 because their conduction band Figure 6. The proposed energy band structure of the Cu 2 O/SnO 2 p-n heterojunction, the photocatalytic activity enhancement mechanism under visible light illumination, and the post-illumination photocatalytic "memory" mechanism in the dark. potentials are less positive than the two-electron reduction potential of O 2 . Thus, novel photocatalyst systems with the post-illumination photocatalytic "memory" effect could be designed based on these decoration components paired with light absorber components of proper conduction band potentials in which photogenerated electrons could transfer from the light absorber component to the decoration component for subsequent trapping under light illumination and release after the illumination was shut off. Furthermore, different photocatalyst systems with the post-illumination photocatalytic "memory" effect could be designed by modulating the conduction band potential of the decoration component to produce different kinds of ROSs for the optimized performance for various applications.
In summary, a novel composite photocatalyst composed of Cu 2 O nanocubes decorated with SnO 2 nanoparticles was successfully created, in which Cu 2 O served as the main visible light absorber, while SnO 2 nanoparticle decoration formed p-n heterojunction of good contact with Cu 2 O nanocubes. The combined effect from both their large conduction band potential difference and the inner electric field provided a strong driving force for photogenerated electrons to transfer from the conduction band of Cu 2 O to that of SnO 2 and be trapped there under visible light illumination. Thus, a largely enhanced photocatalytic performance was observed on these Cu 2 O/SnO 2 photocatalysts as demonstrated by its disinfection of S. aureus cells and degradation of SMX, compared with pure Cu 2 O nanocubes. When the visible light illumination was turned off, trapped electrons could be released from SnO 2 and react with O 2 to produce H 2 O 2 in the dark for more than 24 h, and the Cu 2 O/SnO 2 sample demonstrated a strong post-illumination photocatalytic "memory" disinfection of S. aureus cells in the dark. This work demonstrated that the selection of potential decoration components to construct photocatalyst systems with the post-illumination photocatalytic "memory" effect could be largely expanded to semiconductors with conduction band potentials less positive than the two-electron reduction potential of O 2 . With high efficiency and relatively low cost/energy consumption, photocatalysts with the post-illumination photocatalytic "memory" effect could have the potential for a broad range of environmental applications which require the continuous activity in the dark for an extended period of time.
Synthesis of Cu 2 O nanocubes. Cu 2 O nanocubes were synthesized by a modified process based on a previous report 41 . In a typical experiment, 0.3 g PVP was first dissolved in 270 mL DI water, and 30 mL of 0.02 M CuCl 2 •2H 2 O solution was added into the PVP solution. Then, 3.6 mL of 0.6 M NaOH was added drop wise (1 drop/s) into the above mixture solution with continual stirring. Finally, 4 mL of 0.3 M L-ascorbic acid was added drop wise into the mixture solution and it was further stirred for 5 min before being centrifuged at 9,500 rpm for 5 min. All of these procedures were carried out in a water bath at 35 °C. The obtained yellow precipitates were washed with excessive DI water and ethanol for several times to remove unreacted chemicals and PVP surfactants.
Preparation of SnO 2 nanoparticle-decorated Cu 2 O nanocubes. The obtained Cu 2 O nanocubes were dispersed in a mixture solvent consisting of 15 mL of DI water and 10 mL of absolute ethanol with the aid of ultrasonication for 15 min. Then, 2.5 mL of 0.01 M potassium stannate trihydrate (K 2 SnO 3 ·3H 2 O) solution was slowly dropped into the Cu 2 O suspension and stirred for 10 min. After thorough mixing, 0.2 mL of ethyl acetate (C 4 H 8 O 2 ) was added drop wise into the mixture under vigorous stirring for 1 h. Finally, the suspension was transferred into a 50 mL Teflon-lined stainless steel autoclave, and heated at 170 o C for 6 h in an oven. After the reaction, the products were collected, went through several rinse-centrifugation cycles with DI water and ethanol separately, and then dried at 40 o C for 12 h in a vacuum oven. Bare SnO 2 nanoparticles were also prepared via the same hydrothermal process without the adding of Cu 2 O nanocubes and were used as a reference material for the photocatalytic testing. F20 transmission electron microscope (FEI, Acht, The Netherlands) was used to obtain high-resolution TEM (HRTEM) images of samples. X-ray photoelectron spectroscopy (XPS) measurements were conducted using an ESCALAB 250 X-ray photoelectron spectrometer (Thermo Fisher Scientific Inc., Waltham, MA, U. S. A.) with an Al Kα anode (1486.6 eV photon energy, 300 W). The UV-vis spectra of samples and concentration of SMX were measured on a UV-2550 spectrophotometer (Shimadzu Corporation, Kyoto, Japan).

Photocatalytic disinfection of Staphylococcus aureus (S. aureus) bacteria under visible light illumination.
Wild-type S. aureus (CMCC(B)26003, China national standard material network, P. R. China) were used for photocatalytic disinfection experiments. After overnight culture, cells were diluted to a cell suspension (ca.10 7 cfu/mL) in 0.9% NaCl solution prior to the use for photocatalytic disinfection experiments. All solid or liquid materials had been autoclaved for 30 min at 121 o C before use. The same visible light source was used as in the photocatalytic degradation of SMX. In the photocatalytic disinfection of S. aureus bacteria experiment, aliquot of 10 mL S. aureus cell suspension was pipetted onto a sterile 50 × 10 mm petri dish with the photocatalytst sample, which was first spin coated at the bottom of the dish. A fixed concentration of ~0.2 mg photocatalyst/mL S. aureus solution was used in this experiment. At regular time intervals, 100 μ L of aliquots of the powder-treated cell suspensions were withdrawn in sequence. After appropriate dilutions in 0.9% NaCl solution, aliquot of 100 μ L was spread onto an agar medium plate and incubated at 37 o C for 15 h. The number of viable cells in terms of colony-forming units was counted. Tests were also performed in the dark in the presence of the photocatalyst for comparison. Analyses were in triplicate, and control runs were carried out each time under the same experiment conditions, but without any photocatalytic materials.
Photocatalytic "memory" disinfection of Staphylococcus aureus (S. aureus) bacteria in the dark. For S. aureus bacteria disinfection under dark environment, the Cu 2 O/SnO 2 sample was firstly illuminated by the same lamp for ~3 h. Then, the lamp was shut off and they were used to conduct disinfection experiments in the dark over fresh S. aureus cell suspensions (ca. 10 7 cfu/mL) either immediately or after being kept in dark for 3, 8 and 24 h. In some experiments, Fe(II)-EDTA (0.1 mM) was added for the removal of H 2 O 2 to examine the reactive oxygen species 37,38 . All experimental conditions were the same as that for the photocatalytic disinfection of S. aureus bacteria, but without the visible light illumination. . 0.1 g N,N-diethyl-p-phenylenediammonium sulfate was first dissolved in 10 mL of 0.1 M H 2 SO 4 solution, and 10 mg POD was dissolved in 10 mL DI water. Both DPD and POD solutions were stored in the dark at 4 °C in a refrigerator and replaced with fresh solutions at weekly intervals. For the detection of H 2 O 2 concentration, 5 mL aliquot of the test solution was pipetted into a 10 mL test tube, and mixed with 0.5 mL phosphate buffer solution (0.5 M KH 2 PO 4 and 0.5 M K 2 HPO 4 ) to yield a pH of ~6.0. 50 μ L DPD solution was then added into the mixture solution, followed by the addition of 50 μ L POD solution with shaking for 10 sec. The solution was then settled for 30 sec before the UV-vis spectrum measurement. The H 2 O 2 concentration could be quantified by the UV-2550 spectrophotometer monitoring the absorption maximum at λ max of 551 nm.