Colloids and Surfaces B: Biointerfaces 128 (2015) 211–218

Contents lists available at ScienceDirect

Colloids and Surfaces B: Biointerfaces journal homepage: www.elsevier.com/locate/colsurfb

Bactericidal mechanisms of Au@TNBs under visible light irradiation Lingqiao Guo a,b , Chao Shan b , Jialiang Liang b , Jinren Ni b , Meiping Tong a,b,∗ a Key Laboratory for Heavy Metal Pollution Control and Reutilization, School of Environment and Energy, Shenzhen Graduate School of Peking University, Shenzhen 518055, PR China b The Key Laboratory of Water and Sediment Sciences, Ministry of Education, College of Environmental Sciences and Engineering, Peking University, Beijing 100871, PR China

a r t i c l e

i n f o

Article history: Received 22 August 2014 Received in revised form 6 January 2015 Accepted 8 January 2015 Available online 16 January 2015 Keywords: Au@TNBs Disinfection activity Active species Gold nanoparticles Reusability

a b s t r a c t Au@TNBs nanocomposites were synthesized by depositing Au nanoparticles onto the surfaces of TiO2 nanobelts (TNBs). The disinfection activities of Au@TNBs on model cell type, Gram-negative Escherichia coli (E. coli), were examined under visible light irradiation conditions. Au@TNBs exhibited stronger bactericidal properties toward E. coli than those of TNBs and Au NPs under visible light irradiation. The bactericidal mechanisms of Au@TNBs under light conditions were explored, specifically, the specific active species controlling the inactivation of bacteria were determined. Active species (H2 O2 , diffusing • OH, • O2 − , 1 O2 , and e− ) generated by Au@TNBs were found to play important roles on the inactivation of bacteria. Moreover, the concentrations of H2 O2 , • OH, • O2 − , and 1 O2 generated in the antimicrobial system were estimated. Without the presence of active species, the direct contact of Au@TNBs with bacterial cells was found to have no bactericidal effect. The reusability of Au@TNBs were also determined. Au@TNBs exhibited strong antibacterial activity toward E. coli even in five consecutively reused cycles. This study indicated that the fabricated Au@TNBs could be potentially utilized to inactivate bacteria in water. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Pathogenic microbe contamination of water has been the source of numerous disease outbreaks worldwide especially in developing countries [1]. Drinking water safety thus has attracted great concern over past decades. Chemicals such as chlorine, chloramines, and ozone are commonly utilized to disinfect microbe in water. However, harmful disinfection byproducts (DBPs), many of which are carcinogens, could be generated in these traditional disinfection processes [2,3]. Therefore, great efforts have been recently devoted to developing new techniques to inactivation of pathogenic microbes in water. Since they can produce reactive oxygen species (ROS) to inactivate microbe, TiO2 nanomaterials have recently attracted increasing attention in bacteria decontamination in water [4,5]. Due to their wide band gap, low electron transfer rate to oxygen, and the high electron–hole recombination rate, TiO2 nanomaterials could not generate ROS efficiently under visible light irradiation conditions [6,7]. To enhance the generation of ROS

∗ Corresponding author at: The Key Laboratory of Water and Sediment Sciences, Ministry of Education, College of Environmental Sciences and Engineering, Peking University, Beijing 100871, PR China. Tel.: +86 10 6275 6491; fax: +86 10 6275 6526. E-mail address: [email protected] (M. Tong). http://dx.doi.org/10.1016/j.colsurfb.2015.01.013 0927-7765/© 2015 Elsevier B.V. All rights reserved.

under visible light conditions, noble metals such as Pd [8], Pt [9], Ag [10], and Au [11–13] have been utilized to dope TiO2 particles. Among these noble metals, Previous studies [14,15] showed that Au nanoparticles (NPs) could also generate ROS under light conditions. Thereby, Au NPs have recently attracted great attention for doping TiO2 to inactivate bacteria. The bactericidal effects of Au NPs doped TiO2 are expected to be greatly improved under light irradiation conditions [12,13]. It is well known that the photocatalytic activity of TiO2 is affected by the size, morphology, crystalline structure, and surface structure [16,17]. One-dimensional TiO2 nanobelts (TNBs) has been found to have higher photocatalytic activity relative to TiO2 nanoparticles due to the enhanced visible-light harvesting capabilities, less grain boundaries, and lower e− −h+ recombination rate [18]. The amount of ROS generated by TNBs would be greater than TiO2 nanospheres under visible light conditions. Thus, the antibacterial effects of TNBs are expected to be more significant. By acting as “transition metal impurities”, Au NPs depositing on the surfaces of TNBs could stimulate the generation of active species [19]. Thereby, Au NPs doped TNBs (Au@TNBs) are expected to have improved bactericidal property under visible light irradiation. However, to date, the disinfection activities of Au@TNBs have not been investigated and thus require examination. Although the antibacterial activities of Au@TiO2 have been investigated [12,13],

212

L. Guo et al. / Colloids and Surfaces B: Biointerfaces 128 (2015) 211–218

the mechanisms involved in the photocatalytic disinfection processes of Au NPs loaded TiO2 nanospheres yet have not been addressed. The possible bactericidal mechanisms driving to the bactericidal activities of Au@TNBs thus require systematical investigation. Hence, the objective of this manuscript is to synthesize the Au NPs deposited TNBs nanocomposites (Au@TNBs) and systematically investigate their bactericidal mechanisms. The disinfection effects of Au@TNBs for Escherichia coli (E. coli) under visible light conditions were determined. The bactericidal mechanisms were systematically discussed, and the major active species contributing to antimicrobial activity were proposed. In addition, the reusability of Au@TNBs was also explored. 2. Materials and methods 2.1. Materials TiO2 nanoparticles (diameter of 25–35 nm) were purchased from Degussa (Hanau, Germany). Potassium iodide (KI), 4-hydroxy2,2,6,6-tetramethylpiperidine 1-oxyl (TEMPOL), isopropanol, potassium dichromate (K2 Cr2 O7 ), furfuryl alcohol (FFA), 2, (2-methoxy-4-nitro-5-sulfophehyl)-2H-tetrazolium-53-bis carboxanilide (XTT), and para-chlorobenzoic acid (pBCA) were all purchased from Sigma–Aldrich (USA). All other chemicals were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China) and Xilong Chemical Group (Shantou, Guangdong, China). All the chemicals were analytical grade and used without further purification. 2.2. Synthesis and characterization of Au@TNBs TNBs were synthesized by a hydrothermal process [20] with slight modification, and then Au NPs were deposited on the surface of TNBs via a self-assembly procedure similar to that described in a previous study [21]. The detailed fabrication protocols for TNBs, Au NPs, as well as Au@TNBs fabrication are provided in the Supplementary Information. Powder X-ray diffraction (XRD), Energy dispersive X-ray spectroscopy (EDS) and X-ray photoelectron spectroscopy (XPS), high resolution transmission electron microscope (HRTEM), and inductively coupled plasma optical emission spectroscopy (ICP-OES), were employed to characterize the fabricated materials. Detailed information can be found in Supplementary Information. 2.3. Bacteria preparation Gram-negative Escherichia coli ATCC15597 (E. coli), has been employed as model cell type in previous study [20], was used to determine the bactericidal activities of the Au@TNBs. The detailed bacterial growth and stock suspension preparation protocols can be found in previous study [20] as well as in Supplementary Information. The prepared bacterial stock concentration was typically 109 –1010 colony forming unit (CFU) per milliliter, which was diluted to a concentration of 6.5 × 107 CFU mL−1 prior to addition to the Au@TNBs suspension. 2.4. Bactericidal experiments The photocatalytic disinfection was carried out by using the setup described in Fig. S1. Specifically, a xenon lamp (300 W, Osram Instruments, USA) with a UV cutoff filter ( ≥ 400 nm) was used as a light source, which was put 15 cm above the reaction system. A double-walled beaker with volume of 100 mL, which filled with constant-temperature water between the double walls, was employed as reaction container for disinfection experiments. The

reaction suspensions were well-mixed by stirring and maintained at 25 ◦ C for all experiments. Specifically, a constant-temperature water circulator was employed to keep the temperature of the suspension maintaining at 25 ◦ C. The light intensity at the surface of the reaction solution was measured to be 30 mW/cm2 by a solar power meter (TM-207, TENMARS, Taiwan). The irradiance spectrum (Fig. S2) of the light at the position of the surface of the reaction solution in the wavelength range from 290 to 700 nm was measured with a spectrometer (HR4000CG-UV-NIR, Ocean Optics, FL, USA) calibrated with a radiometric reference light source (HL2000-CAL, Ocean Optics, FL, USA). The filter completely cut off the UV portion of the irradiation, and the visible light portion absorbed by the reaction solution was calculated to be 4.0 mW/cm2 according to the method described in previous studies [22,23]. Thus, the power in the visible light region absorbed by the whole reaction solution was 60.8 mW. Prior to the photocatalytic disinfection experiments, all glass apparatuses used in the experiments were autoclaved at 121 ◦ C for 20 min to ensure sterility. Following that, 5 mg of Au@TNBs was placed into 45 mL sterilized water and sonicated for 5 min to fully disperse the nanomaterials in solution. After that, 5 mL of bacterial stock solution (6.5 × 107 CFU mL−1 ) was added into the sonicated mixture to obtain the target initial cell concentration of 6.5 × 106 CFU mL−1 (pH = 7.0 ± 0.1, ionic strength = 3.0 mM). At different time intervals, 0.5 mL of the reaction solution was sampled and serially diluted with sterilized water. Then, 0.1 mL of the diluted samples were immediately streaked on nutrient agar plates and incubated at 37 ◦ C for 24 h. The number of colonies formed was counted to determine the number of viable cells. Each set of experiments that utilized the newly cultured bacteria was performed in triplicate at pH 7. FEI Quanta 200 FEG environmental scanning electron microscopic (ESEM) was employed to observe the morphological change of the bacterial during the disinfection process. The bacterial pretreatment protocol prior to ESEM characterization can be found in Supplementary Information. Blank control experiment, bacterial solution without Au@TNBs, under visible light irradiation was also conducted. Moreover, the bactericidal activity of Au NPs (3–7 nm, 5 mg L−1 , similar size and concentration of Au particles on 0.1 g L−1 Au@TNBs), TNBs (95 mg L−1 , similar concentration of TNBs on 0.1 g L−1 Au@TNBs), and Au NPs (5 mg L−1 ) + TNBs (95 mg L−1 ) mixture on E. coli were also investigated with visible light irradiation. To clarify the antibacterial effects of different active species that generated by Au@TNBs on the inactivation of bacterial cells, experiments with addition of various scavengers to remove the corresponding active species were performed under visible light irradiation. Specifically, KI was used to remove h+ and surface bounded • OH [24]. Isopropanol, Cr (VI), and 4-hydroxy-2,2,6,6tetramethylpiperidin 1-oxyl (TEMPOL) were used to remove diffusing • OH, e− , and • O2 − in the solution bulk, respectively [25]. Fe (II)-EDTA was used as an enhancer of the bulk phase • OH (Fe (II) could react with H O (Fenton reaction) to generate 2 2 bulk phase • OH) to indirectly verify the existence of H2 O2 [26]. To further elucidate the bactericidal activities of active species, experiments conducted in a partition system were implemented. The detailed information for partition system experiments can be found in previous study [20], as well as in Supplementary Information. To further explore the roles of e− in the disinfection system, and investigate whether 1 O2 had disinfection effect on bacterial cells, disinfection experiments were conducted under anaerobic conditions to avoid the generation of oxidative radicals (such as H2 O2 , 1 O , and • O − ). The water utilized in anaerobic experiments was 2 2 boiled for 30 min before use. High purity nitrogen (>99.999%) was continuously purged during the disinfection process to ensure the absence of O2 in the inactivation system.

L. Guo et al. / Colloids and Surfaces B: Biointerfaces 128 (2015) 211–218

213

Fig. 1. HRTEM images of Au@TNBs in bright (a) and dark field (b).

7

The concentration of Au@TNBs used for measurement of ROS was the same as that for bactericidal experiment. Previous studies [25–30] found that pCBA, FFA, and XTT could be utilized to estimation of generation of • OH, 1 O2 , and • O2 − , respectively. These chemicals (20 ␮M pCBA, 850 ␮M FFA, and 100 ␮M XTT) thus were employed to determine the concentrations of • OH, 1 O2 , and • O2 − in our study. The detailed information about these chemicals as probes for • OH, 1 O2 , and • O2 − , as well as their detection protocols were presented in the Supplementary Information. The concentrations of H2 O2 were determined by a spectrophotometer at 352 nm after the reaction of water sample with a mixture of 0.1 M potassium iodide and 0.01 M ammonium heptamolybdate tetrahydrate (H24 MO7 N6 O24 • 4H2 O) [27]. The detailed measurement process can be found in the Supplementary Information. Control experiments without Au@TNBs were also performed under visible light irradiation. Three types of ROS stoichiometricaly react with their corresponding indicators in a mole ratio of 1:1 [28], and degradation of ROS induced by other side reactions could be negligible. It should be pointed out that bacteria could consume the ROS generated by Au@TNBs and would interfere ROS detection. Thereby, the detection of ROS generated by Au@TNBs was performed without bacteria.

6

Cell density (log10CFU/mL)

2.5. Detection of 1 O2 , • OH, • O2 − , and H2 O2

5 4 3 Control 100 mg/L Au@TNBs 95 mg/L TNBs 5 mg/L Au NPs 5 mg/L Au NPs + 95 mg/L TNBs

2 1 0

0

30

60

90

120 150 180 Time (min)

210 240

Fig. 2. Disinfection efficiencies of E. coli by Au@TNBs, Au NPs, TNBs, and Au NPs + TNBs mixture under visible light irradiation. Error bars represent standard deviations from triplicate experiments (n = 3).

3. Results and discussion

peaks were not observed in the XRD spectra, which was presumably due to the low Au content [30]. Whereas, the elemental Au peaks were observed in the EDS spectra (Fig. S4), indicating Au has been successfully anchored onto the surface of TiO2 . To further identify the composition of the prepared material, the XPS spectra of Au 4f was obtained and presented in Fig. S5. The two symmetric peaks at 83.0 and 86.9 eV (corresponding to the Au4f7/2 and Au4f5/2 peaks, respectively) can be assigned to the characteristic doublets of Au◦ loaded on TNBs [31], which suggested that only elemental Au was formed on the surface of TNBs. According to the XPS hand books and the previous studies [32,33], the double peaks for bulk metallic Au were centered at 84.0 and 87.7 eV, respectively. The slight shift of Au 4f peaks of Au@TNBs toward lower binding energies indicated the formation of a strong metal–support (Au-TNBs) interaction [34]. The contents of Au in the TNBs is found to be 4.53 wt%. HRTEM images of Au@TNBs at bright field (Fig. 1a) and dark field (Fig. 1b) showed that spherical Au NPs with size of 3–7 nm were uniformly deposited on the surface of TNBs. The width and length of the TNBs ranged from 20 to 30 nm and from 40 nm to100 nm, respectively.

3.1. Characterization of Au@TNBs

3.2. Antibacterial activity of Au@TNBs

X-ray diffraction (XRD) patterns (Fig. S3) of the prepared material were recorded and revealed that the prepared material contained anatase phase of TiO2 , which had a higher photocatalytic activity than rutile phase [29]. Typical crystalline gold diffraction

The antibacterial activities of Au@TNBs to Gram-negative E. coli, model cell widely employed in disinfection studies [35,36], were investigated and the results were presented in Fig. 2. Experiment without fabricated material in sterilized water under visible

2.6. The recycle and reuse of the materials To investigate the reusability of the Au@TNBs for inactivation of E. coli under visible light irradiation, the nanocomposites were recovered by filtration after each bactericidal cycle, and then were soaked in 75% of ethanol for 10 min. After that, the harvested nanocomposites were washed thoroughly with sterile water to remove the residual ethanol, and dried at 60 ◦ C. To avoid possible microbe contamination during drying process at 60 ◦ C, the nanocomposites were irradiated with UV light for 30 min prior to the following cycle.

L. Guo et al. / Colloids and Surfaces B: Biointerfaces 128 (2015) 211–218

light irradiation was also conducted as blank control. Moreover, as comparison, the antimicrobial effects of Au NPs, TNBs, and the mixture of Au NPs and TNBs under visible light irradiation were also explored. In the control experiment, viable bacterial population remained unchanged during the experimental process, indicating visible light irradiation had no toxic effect on the bacterial cells in sterilized water during the treatment process. The introduction of Au NPs into bacterial suspension caused 1.15-log decrease of viable E. coli within 4 h treatment, indicating that Au NPs contained bactericidal effect under visible light irradiation. The antibacterial properties of Au NPs have also been reported previously [37]. The introduction of TNBs killed 0.86-log viable bacterial population within 4 h under visible light irradiation, indicating that TNBs contained bactericidal effect under light conditions. The antibacterial property of TiO2 nanobelts under light conditions has also been reported previously [20]. When exposing viable E. coli cells to the mixture of Au NPs and TNBs, the cell density of viable bacterial declined 3.21-log within 4 h treatment. Compared with Au NPs, TNBs, and the mixture of Au NPS and TNBs, Au@TNBs yet could completely inactivate E. coli within 4 h under visible light conditions. The observation demonstrated that Au@TNBs were more effective on bacterial disinfection than Au NPs, TNBs, and the mixture of Au NPS and TNBs. Clearly, the deposition of Au NPs onto TNBs significantly enhanced the photocatalytic bactericidal property of TNBs. Compared with the mixture of Au NPs and TNBs (3.21-log cell decrease within 4 h), Au@TNBs nanocomposites fabricated in present study could produce a synergistic antibacterial effect and thus lead to the increased disinfection effect. Previous studies revealed that the SPR effect of Au NPs could enhance the photocatalytic activity of TiO2 [38]. 3.3. Disinfection mechanism Active species (including • OH, H2 O2 , • O2 − , 1 O2 , h+ , and e− ) generated by NPs under light conditions have been demonstrated as the major mechanisms driving to the inactivation of cells [39]. Since both Au NPs [14,15] and TiO2 [4] could generate active species, the decreases of cell densities might be attributed to the active species produced by Au NPs and TNBs in Au@TNBs disinfection system. However, although antibacterial effects were observed when cells were exposure to solely Au NPs (1.15-log) and TNBs (0.86-log), greater bactericidal effect was observed in Au@TNBs disinfection system (complete inactivation) relative to those of TNBs and Au NPs under light conditions. Moreover, the bactericidal effect of Au@TNBs was also greater relative to those obtained when Au NPs and TNBs were copresent in disinfection systems (3.21-log). Clearly, the amounts of active species produced by Au@TNBs might be greater than the sum of active species produced by solely TNBs and Au NPs. Nel et al. [19] has found that noble metal on the surface of TiO2 could work as “transition metal impurities” and stimulate the generation of active species by TiO2 . Au NPs present on the surfaces of Au@TNBs nanohybrids clearly would also stimulate the generation of active species by TNBs, leading to the greater inactivation rate. 3.3.1. Quenching experiments To examine the significance of each active species on the inactivation of bacterial cells in Au@TNBs system, experiments were performed by adding different scavengers to remove the corresponding active species under visible light irradiation and the results were shown in Fig. 3. Control experiments were performed by only adding scavengers into bacterial solutions (Fig. S6). In all control experiments, the number of viable cells did not change with light irradiation during the treatment process, indicating the scavengers did not have obvious disinfection effect on E. coli cells within the reaction period. As shown in Fig. 3, when KI (scavenger of h+ and

7 Cell density (log10CFU/mL)

214

6 5 4 Control No scavenger 0.05 mM KI 0.20 mM TEMPOL 0.01 mM Cr (VI) 2.00 mM Fe (II)-EDTA 0.05 mM Isopropanol

3 2 1 0

0

30

60

90 120 150 Time (min)

180

210

240

Fig. 3. Disinfection efficiency of E. coli by Au@TNBs in the presence of different scavengers (KI, Isopropanol, Cr (VI), Fe (II)-EDTA, TEMPOL) under visible light irradiation. Error bars represent standard deviations from triplicate experiments (n = 3).

surface • OH) was added into the disinfection system, the antibacterial efficiency was almost the same as that without introduction of any scavengers. The observation showed that neither surface • OH nor h+ on the surface of Au@TNBs played a role in disinfection process. The insignificance of surface • OH and h+ during the disinfection process has also been reported for sphalerite [25] and Ag2 O/TNBs nanocomposites [20]. With the addition of isopropanol (scavenger of diffusing • OH), the viable cell density decreased 2.43log, which was lower than that observed without adding of any scavengers. The observation showed that the diffusing • OH played an important role in the inactivation reaction. Du et al. [40] showed that although its lifetime was short, • OH was a highly strong and nonselective oxidant that could damage bacterial cell. It has been proved that Fe (II) complex could react with H2 O2 (Fenton reaction) to generate • OH [41]. If H2 O2 was present in the disinfection system, the introduction of Fe (II)-EDTA into the disinfection system could generate • OH by reacting with H2 O2 and thus would improve the inactivation efficiency [40]. When Fe (II)EDTA was added into the treatment system, complete inactivation of E. coli was achieved within 180 min, which was faster than that without addition of any scavengers (240 min). The observation indicated that H2 O2 was present in the reaction system and it reacted with Fe (II) to generate • OH (more effective ROS), leading to the faster bactericidal effect. When TEMPOL (scavenger of • O2 − ) was included in the disinfection system, the viable cell density declined 2.33-log, which was also lower than the complete inactivation acquired without addition of any scavengers. The observation demonstrated that • O2 − was also involved in the bacterial disinfection process. Rosen et al. [42] revealed that • O2 − could penetrate biomembrane into the interior of the cell and damage the enzymes and DNA by acting as both reductant and oxidant. Moreover, • O2 − could react with H2 O2 (the Haber–Weiss reaction) and generated highly reactive • OH inside the cell, leading to the inactivation of bacteria [43]. When Cr (VI) was added to remove e− , the antimicrobial efficiency was greatly retarded and the viable cells only decreased 1.73-log, indicating e− strongly contributed to the antimicrobial efficiency of Au@TNBs. The above results showed that surface • OH and h+ did not contribute to the antibacterial effect of Au@TNBs, whereas H2 O2 , diffusing • OH, • O2 − , and e− were strongly involved in the disinfection system. 3.3.2. Partition system experiments To further elucidate the role of active species (diffusing • OH and H O ) in the inactivation system, partition sys2 2 tem experiments, during which the bacterial suspension was

7

7

6

6

Cell density (log10CFU/mL)

Cell density (log10CFU/mL)

L. Guo et al. / Colloids and Surfaces B: Biointerfaces 128 (2015) 211–218

5 4 3 Control Au@TNBs 0.05 mM Isopropanol+Au@TNBs 0.01 mM Fe(II)-EDTA + Au@TNBs

2 1 0

0

30

60

90

120 150 Time (min)

180

210

240

215

5 4 3 No oxygen without Au@TNBs No oxygen with Au@TNBs No oxygen + Cr (VI) with Au@TNBs

2 1 0

0

30

60

90

120 150 Time (min)

180

210

240

Fig. 4. Disinfection efficiency of E. coli by Au@TNBs under visible light irradiation in a partition system. Error bars represent standard deviations from replicate experiments (n = 3).

Fig. 5. Disinfection efficiency of E. coli by Au@TNBs with visible light irradiation under anaerobic conditions. Error bars represent standard deviations from triplicate experiments (n = 3).

placed inside the semipermeable membrane compartment and the Au@TNBs was suspended outside the compartment, were also employed in present study. The semipermeable membrane only allowed small molecules to freely pass through, whereas the Au@TNBs (several dozens of nm in size) and E. coli cells (1.57 ± 0.06 ␮m × 0.7 ± 0.05 ␮m) could not go through the semipermeable membrane. In the partition system, bacterial cells were separated from Au@TNBs, therefore, the active species on the surface of Au@TNBs, such as h+ , e− , surface • OH, • O2 − , and 1 O , could not be involved in the inactivation process, only dif2 fusing active species, such as diffusing • OH and H2 O2 , could pass through the semipermeable membrane and react with the bacterial cells inside the membrane container. The viable cell population did not change within 4 h when Au@TNBs was not included in the partition system (Fig. 4), indicating that neither the visible light nor the semipermeable membrane was toxic to the model cells. In contrast, the cell density decreased 1.0-log with the addition of Au@TNBs outside the membrane compartment (Fig. 4). The observation suggested that the diffusing active species (• OH and H2 O2 ), which could pass through the semipermeable membrane, inactivated bacteria inside the membrane compartment. To figure out which diffusing reactive species (• OH, H2 O2 , or both) are involved during the disinfection process in the partition system, isopropanol was added into the outside of the membrane container to remove diffusing • OH. Similar as that without introduction of any scavengers outside compartment, the cell population also decreased ∼1.0-log with the addition of isopropanol, indicating that diffusing • OH did not have contribution to bacterial disinfection in partition system. Moreover, when Fe (II)-EDTA (could remove H2 O2 by Fenton reaction and enhance the production of diffusing • OH when H2 O2 was present) was added into the outside compartment, the viable bacterial density did not change with increasing reaction duration, which was contradicted with the increased inactivation efficiency observed in disinfection system without partition membrane when Fe (II)-EDTA was added. Clearly, enhanced generation of • OH (removal of H2 O2 ) did not have contribution to inactivation of bacteria in partition system. Du et al. [40] revealed that only the • OH produced within the cells or near the cell surface was effective in causing damage to the cell. Since they were produced in the outside compartment in the partition system, most of • OH would be annihilated before they can pass through the membrane due to the short life time (only 10 ␮s in natural water) [44]. As a result, the antimicrobial effect of • OH was not observed in the partition system. In contrast with • OH, H2 O2 was relatively much more stable [36] and thus could pass through the semimembrane, leading to the inactivation of bactericidal cells in the partition system. The

observations further confirmed that H2 O2 had contribution to the bactericidal effect of Au@TNBs. 3.3.3. Anaerobic system experiments To further investigate the effect of e− in the disinfection system, additional experiments were performed in an anaerobic system under visible light irradiation and the results were shown in Fig. 5. To obtain anaerobic condition, nitrogen was used to expel oxygen out of the system. Control experiment was also carried out and showed that the viable bacteria population did not change under anaerobic conditions. It should be noted that under anaerobic conditions, oxidative radicals such as 1 O2 , • O2 − and H2 O2 would not be generated [45], whereas, h+ and e− could be produced in the disinfection reactions. When Cr (VI) (scavenger of e− ) was added to remove e− , no decrease of viable bacterial population was observed with the presence of only h+ in the system, indicating that h+ did not have toxic effect on bacteria. This observation was consistent with the results observed in aerobic conditions (Fig. 3). Moreover, without addition of any scavengers (presence of h+ and e− ), Au@TNBs did not change the cell density in the anaerobic system (Fig. 5). The observation implied that e− had no toxic effect on bacteria under anaerobic conditions. However, as discussed above, e− had great contribution to the disinfection effect of Au@TNBs in aerobic system (Fig. 3). Thus, e− might indirectly inactivate bacteria via generation of other ROS through a series of reactions in the presence of oxygen [46]. Under visible light irradiation, the conduction electrons of the Au NPs absorbed the light energy and then migrated to the conduction band (CB) of the TiO2 , leaving positive charged holes (h+ ) on the Au NPs [47]. The negative charged bacteria contacted with Au NPs through electrostatic interaction and acted as electron donor in the Au@TNBs disinfection system. The conduction-band e− (come from Au NPs) reacted with O2 absorbed on the surface of TNBs to produce • O2 − (Eq. (1)) [46], which could subsequently react with e− and H+ to produce H2 O2 (Eq. (2)). The • O2 − also could react with H2 O2 to produce the • OH (Haber–Weiss reaction, Eq. (3)). The produced H2 O2 , • OH, and • O2 − have strong oxidative capacity and could directly oxidize and damage the cell structure, which were confirmed in the scavenger experiments in aerobic system (Fig. 3). The removal of e− with Cr (VI) therefore indirectly inhibited the generation of H2 O2 , • O2 − , and • OH, and thus the bactericidal effect of Au@TNBs was inhibited (Fig. 3). The indirectly effect of e− in the disinfection process has also been reported previously [20]. e− + O2 → • O2 −

(1)

O2 − + e− + 2H+ → H2 O2

(2)

216

L. Guo et al. / Colloids and Surfaces B: Biointerfaces 128 (2015) 211–218

(a) 240 min 180 min 120 min 90 min 60 min 30 min 0 min

Absorbance

0.08 0.06 0.04 0.02 300

400 500 600 700 Wavelength (nm)

0.9 0.8 0.7 pCBA photolysis Au@TNBs under visible light irradiation

0.5 0

800

(c)

1.00

30

60

90 120 150 180 210 240 Time (min)

(d)

5

FFA (Ct / Co)

H2O2 Concentration (mM)

(b)

0.6

0.00

6

1.0

pCBA (Ct / Co)

0.10

4 3 2

0.96 0.94 0.92

1 0 0

0.98

30

60

90 120 150 180 210 240 Time (min)

0.90 0

FFA photolysis Au@TNBs under visible light irradiation

30

60

90 120 150 180 210 240 Time (min)

Fig. 6. • O2 − generation kinetics as indicated by the reduction of 100 ␮M XTT (a), • OH generation kinetics as indicated by the degradation of 20 ␮M pCBA (b), concentrations of H2 O2 generated during the disinfection process (c), and 1 O2 generation kinetics as indicated by the degradation of 0.85 mM FFA (d). Error bars represent standard deviations from triplicate experiments (n = 3).

O2 − + H2 O2 → • OH + OH− + O2

(3)

It is worth pointing out that the results above indicated that e− is the precursor of H2 O2 , • OH, and • O2 − in this disinfection system. Thus, the absence of e− should result in no decrease of cell density during the inactive process. However, the cell density of E. coli decreased 1.73-log when e− was removed by Cr (VI) in the aerobic experiment (Fig. 3). The observation indicated that there must be other active species generated and involved in the disinfection process under aerobic condition. Misawa et al. [15] and Zhang et al. [14] revealed that the SPR effect of Au NPs induced a strong absorption of the incident photon energy, which can be transferred to O2 absorbed on the material and lead to the generation of 1 O2 . Previous studies [15,27] have also shown that 1 O2 was an important disinfectant to inactivate bacteriophage and bacterial cells. Therefore, 1 O2 might be generated and had inactivation effect in Au@TNBs disinfection process. The presence of 1 O2 in the photocatalytic disinfection system was further verified in the section below. Moreover, as stated above, no oxidative radicals except e− and h+ were produced in the disinfection reactions under anaerobic conditions. The aerobic system experiments have shown that h+ did not have toxic effect on bacteria. Thus, the addition of Cr (VI) (scavenger of e− ) in anaerobic system therefore would remove all the active species. The decreases of viable cell population observed during the treatment process under anaerobic conditions thus could be attributed to the direct contact of cells with Au@TNBs. However, viable cell density did not decrease during the reaction duration with the removal of e− under anaerobic conditions. Moreover, even without the addition of Cr (VI), Au@TNBs did not induce the decrease of viable cell density in the anaerobic system. The observation implied that direct contact of Au@TNBs with bacterial cells had no bactericidal effect when oxidative radicals were absent. It should be noted that without the visible light irradiation (under dark conditions, the generation of ROS could be reduced), viable cell density did not obviously change during the treatment process in aerobic system (Fig. S7). The observation further demonstrated that without ROS, direct contact of Au@TNBs with bacterial cells did not physically inactivate the cells.

3.4. The detection of ROS Since oxidative radicals such as 1 O2 , H2 O2 , • OH, and • O2 − played important roles during the disinfection process, the existence of these ROS generated by Au@TNBs under visible light irradiation was further confirmed through photometric method and chromatographic technology. 20 ␮M pCBA, 850 ␮M FFA, and 100 ␮M XTT were used as indicators for • OH, 1 O2 , and • O2 − , respectively. H2 O2 was measured through iodide method. Fig. 6a shows the generation kinetics of • O2 − during 4 h visible light irradiation. A control experiment was conducted in sterilized water in the absence of the fabricated material under visible light irradiation, and no absorption peaks were detected (Fig. S8). The observation demonstrated that the side reaction and the photolysis of XTT could be ignored during the measurement period. In the presence of Au@TNBs, there was an increasing absorption intensity at 470 nm during 4 h with visible light irradiation. The observation indicated that the XTT transformed into XTT-formazan due to the generation of • O2 − . The total concentration of • O2 − produced during 4 h treatment was 1.2 ± 0.6 ␮M, indicating that • O2 − did exist in the disinfection system. Since the reaction rate between XTT and • O2 − (8.6 × 104 M−1 S−1 ) [48] was lower than that of TEMPOL and • O2 − ((6.9 ± 2.9) × 105 M−1 S−1 ) [49] at the same reaction conditions, the amount of • O2 − could be trapped by XTT (Fig. 3) thus was lower than that of trapped by TEMPOL. The concentration of • O2 − in the disinfection system therefore would be higher than the determined concentration of 1.2 ± 0.6 ␮M. Moreover, • O2 − was very unstable in water, it would subsequently undergo facile disproportionation to produce other ROS including H2 O2 and • OH [50]. Thereby, the actual concentration of • O2 − in disinfection system would also be greater than the determined concentration. Fig. 6b presented the • OH production kinetics of Au@TNBs within 4 h under visible light irradiation. A control test was also performed in the absence of Au@TNBs and showed that the concentration of pCBA decreased only 0.3% during the reaction duration. The observation indicated that the photolysis of pCBA could be negligible. In contrast, with the presence of Au@TNBs, the concentration of pCBA decreased 31 ± 0.4% within 4 h treatment. The calculated total concentration of • OH produced in the disinfection

L. Guo et al. / Colloids and Surfaces B: Biointerfaces 128 (2015) 211–218

217

Fig. 7. ESEM images of E. coli cells at different photocatalytic disinfection time, 0 h (a), 2 h (b), and 4 h (c).

system was found to be 5.73 ± 0.27 ␮M. Li et al. [51] also revealed that the average molar concentration of • OH produced by ZnO nanoparticles (with the dosage of 10 mg L−1 ) in deionized water under 4 h UV-365 light irradiation was 5.2 ± 0.2 ␮M. Fig. 6c showed the H2 O2 production kinetics of Au@TNBs under visible light irradiation. It could be seen from Fig. 6c that the concentration of H2 O2 reached a maximum of 5.23 ± 0.45 ␮M in 60 min and slightly decreased afterwards. Similar trend has also been previously reported [52]. Chen et al. [25] showed that H2 O2 even on the level of several micrometers could kill E. coli cells by damaging the antioxidative enzyme of bacteria. Moreover, it is worth pointing out that the production and decomposition of H2 O2 is synchronous and continuous in the presence of bacteria. The total amounts of H2 O2 produced in the system thus would be much greater than the measured values. Clearly, the H2 O2 generated by Au@TNBs played an important role in the disinfection process under visible light conditions. Fig. 6d illustrated a significant degradation of FFA, which indicated that 1 O2 was generated in Au@TNBs suspensions under visible light irradiation. The total concentration of 1 O2 generated by Au@TNBs was 81 ± 4.34 ␮M within 4 h visible light irradiation. Badireddy et al. [53] found that 1 O2 contributed to inactivation of MS-2 bacteriophage by fullerol. The above observations clearly demonstrated that oxidative radicals 1 O2 , H2 O2 , • OH, and • O2 − were generated in Au@TNBs disinfection system, resulting in the inactivation of bacteria.

could be inactivated by Au@TNBs within 4 h under visible light irradiation. The results indicated that Au@TNBs exhibited good potential for repeated use. 4. Conclusions Au@TNBs nanocomposites were synthesized and exhibited stronger bactericidal properties toward E. coli than those of TNBs and Au NPs under visible light irradiation. Doping Au NPs on the surfaces of TNBs produced synergistic antibacterial effect. Active species including H2 O2 , diffusing • OH, • O2 − , 1 O2 , and e− generated by Au@TNBs were found to play important roles on the inactivation of bacteria under light conditions. Specifically, H2 O2 , diffusing • OH, • O − , and 1 O directly disinfect bacteria, whereas, e− indirectly 2 2 inactivate bacterial cells through reactions to generate H2 O2 , • O2 − , and • OH. The presence of H2 O2 , • OH, • O2 − , 1 O2 in the antimicrobial system were further confirmed through photometric method and chromatographic technology. The direct contact of Au@TNBs with bacteria cells was demonstrated to have no bactericidal effect when oxidative radicals were removed from the disinfection systems. In addition, Au@TNBs exhibited strong antibacterial activity toward E. coli even in five consecutively reused cycles. The fabricated Au@TNBs could be potentially utilized to inactivate bacteria in water. Acknowledgements

3.5. Scanning electron microscopy analysis ESEM was used to explore the damage effect of Au@TNBs on bacteria. The morphology changes of E. coli cell during the inactivation process were recorded and presented in Fig. 7. The bacteria contained a smooth and intact structure prior to the disinfection reaction (Fig. 7a). Many pits could be observed on the cell wall and the cell was deformed to be cataplastic by ROS after 2 h treatment (Fig. 7b). The cells were ruptured completely after 4 h treatment (Fig. 7c), indicating that ROS produced by Au@TNBs destroyed the membrane integrity and led to the leakage of cytoplasm. The death of bacteria thus was obtained. Similar observation has also been previously reported [35].

This work was supported by the National Natural Science Foundation of China under grant Nos. 21177002 and 21377006, and also by the program for New Century Excellent Talents in University under grant No. NCET-13-0010. We acknowledge the editor and three reviewers for their very helpful comments. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.colsurfb. 2015.01.013. References

3.6. The reusability of Au@TNBs The recovery and reuse of Au@TNBs for cell inactivation throughout five consecutive cycles was also investigated and the results were provided in Fig. S9. The bactericidal efficiencies of Au@TNBs slightly decreased with the increase of reused cycles. However, even in the fifth cycle, greater than 99.68% of bacteria

[1] M.A. Montgomery, M. Elimelech, Environ. Sci. Technol. 41 (2007) 17–24. [2] S.W. Krasner, H.S. Weinberg, S.D. Richardson, S.J. Pastor, R. Chinn, M.J. Sclimenti, G.D. Onstad, A.D. Thruston Jr., Environ. Sci. Technol. 40 (2006) 7175–7185. [3] Q. Li, S. Mahendra, D.Y. Lyon, L. Brunet, M.V. Liga, D. Li, P.J.J. Alvarez, Water Res. 42 (2008) 4591–4602. [4] S. Gelover, L.A. Gomez, K. Reyes, M.T. Leal, Water Res. 40 (2006) 3274–3280. [5] L. Rizzo, A. Della Sala, A. Fiorentino, G.L. Puma, Water Res. 53 (2014) 145–152. [6] M.M. Ali, K.Y. Sandhya, Carbon 70 (2014) 249–257.

218

L. Guo et al. / Colloids and Surfaces B: Biointerfaces 128 (2015) 211–218

[7] C. Zhao, M. Pelaez, D.D. Dionysiou, S.C. Pillai, J.A. Byrne, K.E. O’Shea, Catal. Today 224 (2014) 70–76. [8] R.-J. Wu, Y.-S. Liu, H.-F. Lai, J.-H. Wang, M. Chavali, J. Nanosci. Nanotechnol. 14 (2014) 6792–6799. [9] D. Yuan, Y. Gao, H. Wu, T. Xiao, Y. Wang, B. Wang, Z. Zhang, RSC Adv. 4 (2014) 19533–19537. [10] R. Goei, T.-T. Lim, Water Res. 59 (2014) 207–218. [11] E. Kowalska, O.O.P. Mahaney, R. Abe, B. Ohtani, Phys. Chem. Chem. Phys. 12 (2010) 2344–2355. [12] L. Armelao, D. Barreca, G. Bottaro, A. Gasparotto, C. Maccato, C. Maragno, E. Tondello, U.L. Stangar, M. Bergant, D. Mahne, Nanotechnology 18 (2007) 375709. [13] G.F. Fu, P.S. Vary, C.T. Lin, J. Phys. Chem. B 109 (2005) 8889–8898. [14] W. Zhang, Y. Li, J. Niu, Y. Chen, Langmuir 29 (2013) 4647–4651. [15] M. Misawa, J. Takahashi, Nanomed. Nanotechnol. Biol. Med. 7 (2011) 604–614. [16] C.H. Cho, D.K. Kim, D.H. Kim, J. Am. Ceram. Soc. 86 (2003) 1138–1145. [17] A. Fujishima, X. Zhang, D.A. Tryk, Surf. Sci. Rep. 63 (2008) 515–582. [18] N. Wu, J. Wang, D.N. Tafen, H. Wang, J.-G. Zheng, J.P. Lewis, X. Liu, S.S. Leonard, A. Manivannan, J. Am. Chem. Soc. 132 (2010) 6679–6685. [19] A. Nel, T. Xia, L. Mädler, N. Li, Science 311 (2006) 622–627. [20] Y. Jin, Z. Dai, F. Liu, H. Kim, M. Tong, Y. Hou, Water Res. 47 (2013) 1837–1847. [21] W. Hu, B. Liu, Q. Wang, Y. Liu, Y. Liu, P. Jing, S. Yu, L. Liu, J. Zhang, Chem. Commun. 49 (2013) 7596–7598. [22] T. Kohn, K.L. Nelson, Environ. Sci. Technol. 41 (2007) 192–197. [23] T. Kohn, M. Grandbois, K. McNeill, K.L. Nelson, Environ. Sci. Technol. 41 (2007) 4626–4632. [24] T.M. El-Morsi, W.R. Budakowski, A.S. Abd-El-Aziz, K.J. Friesen, Environ. Sci. Technol. 34 (2000) 1018–1022. [25] Y. Chen, A. Lu, Y. Li, L. Zhang, H.Y. Yip, H. Zhao, T. An, P.-K. Wong, Environ. Sci. Technol. 45 (2011) 5689–5695. [26] L.-S. Zhang, K.-H. Wong, H.-Y. Yip, C. Hu, J.C. Yu, C.-Y. Chan, P.-K. Wong, Environ. Sci. Technol. 44 (2010) 1392–1398. [27] O.C. Romero, A.P. Straub, T. Kohn, T.H. Nguyen, Environ. Sci. Technol. 45 (2011) 10385–10393. [28] Y. Li, W. Zhang, J. Niu, Y. Chen, ACS Nano 6 (2012) 5164–5173. [29] A. Sclafani, L. Palmisano, M. Schiavello, J. Phys. Chem. A 94 (1990) 829–832.

[30] M.V. Dozzi, L. Prati, P. Canton, E. Selli, Phys. Chem. Chem. Phys. 11 (2009) 7171–7180. [31] M. Ye, H. Zhou, T. Zhang, Y. Zhang, Y. Shao, Chem. Eng. J. 226 (2013) 209–216. [32] Y.C. Liu, L.C. Juang, Langmuir 20 (2004) 6951–6955. [33] M.C. Henry, C.C. Hsueh, B.P. Timko, M.S. Freund, J. Electrochem. Soc. 148 (2001) D155–D162. [34] J. Huang, W.-L. Dai, H. Li, K. Fan, J. Catal. 252 (2007) 69–76. [35] K. Pathakoti, S. Morrow, C. Han, M. Pelaez, X. He, D.D. Dionysiou, H.-M. Hwang, Environ. Sci. Technol. 47 (2013) 9988–9996. [36] K. Sunada, Y. Kikuchi, K. Hashimoto, A. Fujishima, Environ. Sci. Technol. 32 (1998) 726–728. [37] Y. Zhao, Y. Tian, Y. Cui, W. Liu, W. Ma, X. Jiang, J. Am. Chem. Soc. 132 (2010) 12349–12356. [38] A. Primo, A. Corma, H. Garcia, Phys. Chem. Chem. Phys. 13 (2011) 886–910. [39] R.J. Watts, D. Washington, J. Howsawkeng, F.J. Loge, A.L. Teel, Adv. Environ. Res. 7 (2003) 961–968. [40] J. Du, J.M. Gebicki, Int. J. Biochem. Cell Biol. 36 (2004) 2334–2343. [41] L.S. Zhang, J.L. Li, Z.G. Chen, Y.W. Tang, Y. Yu, Appl. Catal. A: Gen. 299 (2006) 292–297. [42] H. Rosen, S.J. Klebanoff, J. Exp. Med. 149 (1979) 27–39. [43] J.P. Kehrer, Toxicology 149 (2000) 43–50. [44] J. Hoigne, Water Sci. Technol. 35 (1997) 1–8. [45] I.K. Konstantinou, T.A. Albanis, Appl. Catal. B Environ. 49 (2004) 1–14. [46] M. Cho, H. Chung, W. Choi, J. Yoon, Water Res. 38 (2004) 1069–1077. [47] D. Tsukamoto, Y. Shiraishi, Y. Sugano, S. Ichikawa, S. Tanaka, T. Hirai, J. Am. Chem. Soc. 134 (2012) 6309–6315. [48] G. Bartosz, Clin. Chim. Acta 368 (2006) 53–76. [49] M.C. Krishna, A. Russo, J.B. Mitchell, S. Goldstein, H. Dafni, A. Samuni, J. Biol. Chem. 271 (1996) 26026–26031. [50] W.H. Ma, Y.P. Huang, J. Li, M.M. Cheng, W.J. Song, J.C. Zhao, Chem. Commun. (2003) 1582–1583. [51] Y. Li, J. Niu, W. Zhang, L. Zhang, E. Shang, Langmuir 30 (2014) 2852–2862. [52] C. Kormann, D.W. Bahnemann, M.R. Hoffmann, Environ. Sci. Technol. 22 (1988) 798–806. [53] A.R. Badireddy, E.M. Hotze, S. Chellam, P. Alvarez, M.R. Wiesner, Environ. Sci. Technol. 41 (2007) 6627–6632.