Journal of Colloid and Interface Science 428 (2014) 24–31

Contents lists available at ScienceDirect

Journal of Colloid and Interface Science www.elsevier.com/locate/jcis

Effects of Ag doping on the photocatalytic disinfection of E. coli in bioaerosol by Ag–TiO2/GF under visible light Thanh-Dong Pham, Byeong-Kyu Lee ⇑ Department of Civil and Environmental Engineering, University of Ulsan, Daehakro 93, Namgu, Ulsan 680-749, Republic of Korea

a r t i c l e

i n f o

Article history: Received 19 February 2014 Accepted 13 April 2014 Available online 24 April 2014 Keywords: Mechanism TiO2 Ag doped Disinfection capacity Humidity effects

a b s t r a c t Ag doped TiO2/glass fibers (Ag–TiO2/GF) were prepared and used for photocatalytic disinfection of Escherichia coli (E. coli) in an indoor air environment. The prepared photocatalysts were characterized using scanning electron microscope (SEM) for morphology, X-ray diffraction (XRD) for microstructure, UV–Visible diffuse reflectance spectra (DRS) for optical properties and X-ray photoelectron spectroscopy (XPS) to determine elemental state. The optimized weight fraction of TiO2 in the TiO2/glass fiber (TiO2/GF) was 3%. The silver content in Ag/TiO2 was altered from 1% to 10% to investigate the optimal ratio of Ag doped on the TiO2/GF for the photocatalytic disinfection of E. coli. Doped Ag enhanced the electron–hole separation as well as charge transfer efficiency between the valance band and the conduction band of TiO2. The generated electron–hole pairs reacted with water and molecular oxygen to form strong oxidative radicals, which participated in the oxidation of organic components of E. coli, resulting in bacterial death. The photocatalytic disinfection activity under visible light increased with the increase in silver content up to 7.5% and then decreased slightly with further increasing Ag content. Among the three humidity conditions used in this study (40 ± 5%, 60 ± 5%, 80 ± 5%), the highest disinfection ratio of E. coli by the photocatalytic system was observed in the intermediate humidity level followed by the high humidity level. Using the 7.5% Ag–TiO2/GF and the intermediate level of humidity (60 ± 5%), the highest disinfection ratio and disinfection capacity of E. coli were 93.53% and 26 (CFU/s cm2), respectively. Ó 2014 Elsevier Inc. All rights reserved.

1. Introduction The indoor environment is affected by many factors, including ambient temperature, relative humidity (RH), air exchange rate, air velocity, ventilation, ambient and indoor levels of air pollutants (SOx, NOx, O3, CO, VOCs, and particles), and bioaerosols [1,2]. Bioaerosols are everywhere in the environment and include viruses, bacteria, fungi, pollen, plant or animal debris, as well as fragments and products of these organisms. Recently, airborne microorganisms have received significant attention due to their potential health effects, as well as due to the threat of bioterrorism [3]. There is a growing concern that bioaerosols may be associated with various adverse health effects. Three major disease groups are associated with bioaerosol exposure and are distinguished as respiratory, cancer, and infectious diseases [4]. In addition, people with compromised immunity or with existing respiratory conditions, such as allergies and asthma, are at an increased risk of exposure to bioaerosols and their derivatives [5]. In light of the increasing occurrence of asthma and respiratory tract diseases, ⇑ Corresponding author. Fax: +82 52 259 2629. E-mail address: [email protected] (B.-K. Lee). http://dx.doi.org/10.1016/j.jcis.2014.04.030 0021-9797/Ó 2014 Elsevier Inc. All rights reserved.

and the deterioration of indoor air quality, the need for controlling bioaerosol exposures has increased recently. To reduce the health risk from bioaerosol exposures, various controlling technologies for this contaminant, including filtration, ozone generators, germicidal ultraviolet (UV), negative air ionization, and photocatalytic oxidation, have been investigated for air purification [2]. These conventional purification systems are effective in removing nonbiological airborne pollutants, but may not be very effective for bacterial disinfection. Thus, it has become necessary to develop air purification systems, devices, or methods to for the treatment of bacteria. The integration of TiO2 photocatalyst into air purifiers for the disinfection of microbial contamination has become more popular for dealing with problems or drawbacks associated with the air filtration of bioaerosols [6–9]. TiO2 has been intensively investigated as a potential photocatalyst, due to its low price, non-toxicity, physical and chemical inertness, and most importantly, high photocatalytic efficiency [10–12]. In general, a TiO2-mediated photocatalytic reaction occurs when TiO2 is exposed to UV irradiation, at which point it can absorb photons with energies equal to or larger than its band gap (3.2 eV, anatase). Subsequently, electrons (e) in the valence band are promoted to the conduction band, leaving positive holes

T.-D. Pham, B.-K. Lee / Journal of Colloid and Interface Science 428 (2014) 24–31

(h+) in the valence band [13]. The hole in the valence band reacts with water to produce hydroxyl radicals (OH), while the electron in the conduction band reacts with molecular oxygen to provide superoxide radical anions (O 2 ) [13]. When TiO2 is used as disinfectant, the oxy radical species participate in reduction and oxidation reactions to attack vital organic components of microorganisms resulting in the death of microorganisms [14–18]. However, the use of TiO2 in practical applications is limited by its low utilization efficiency of solar energy because of the restricted absorption in the UV region due to its wide band gap (3.2 eV). Hence, increasing the efficiency of visible photocatalysis is one of the most important tasks for the practical application of TiO2 in the future. To enhance the TiO2 photocatalytic activity of photocatalyst systems, the modification of pure TiO2 has been extensively investigated in various studies [19–27] Silver, in both metal and in ionic states, is particularly used as a doping agent to enhance photocatalytic activity of TiO2 applying for bioaerosol disinfection [28–33].Silver ions are reported to cause disruptions of the cell membrane and to electron transport in microorganisms leading death of bacteria [30,31]. Ag metal is of particular interest as a doping agent due to its ability to act as both an electron sink and donor [28,32]. While acting as electron sink and donor to attract photogenerated electrons, Ag promotes the electron–hole pair separation and prevents electron recombination from TiO2 after photon absorption. Ag can also modify the grain sizes of the TiO2, resulting in an increased surface area and, as a consequence, photoactivity [31]. On the other hand, the plasmon resonance in metallic silver (Ag particles) can enhance the electric field, facilitating electron– hole production [29]. Therefore, Ag is one of the most promising metals for improving the photocatalytic disinfection property of TiO2 due to its strong and broad bactericides. It is difficult to separate and regenerate powder photocatalyst from purification systems after being used as photocatalysts for water or air purification [34]. To avoid the disadvantages of TiO2 powders as photocatalysts, various methods to apply TiO2 coatings to many substrates have been developed [35]. The use of a glass fiber as a substrate for the TiO2 coating to be used as a photocatalyst has many advantages compared to TiO2 powders [36,37]. A TiO2/GF has a much higher surface area and is more suitable to be used as a photocatalyst than TiO2 powder. However, the TiO2/GF and Ag–TiO2/GF systems as germicidal photocatalysts have not been extensively investigated or established. Our previous study reported initial feasibility of TiO2/GF and Ag–TiO2/GF photocatalysts as a germicide for the removal of E. coli [38]. However, it was still vague and contradictory in some aspects. Therefore, the main purpose of this study is to investigate the chemical mechanism of Ag working as doping agent to enhance photocatalytic activity of TiO2 for E. coli disinfection. The current study is also focusing on the photocatalytic disinfection efficiency of E. coli as a function of bacterial input concentration, temperature, humidity, and reaction time.

TiO2/GF photocatalyst used for this study was 3%. To prepare the Ag-impregnated catalyst, the Ag–TiO2/GF photocatalyst, a 0.05-M AgNO3 solution was deposited on the 3% TiO2/GF photocatalyst. The added volume of AgNO3 solution that was calculated to synthesize Ag–TiO2/GF with the weight fractions of Ag in the Ag/ TiO2 was 1%, 2.5%, 5%, 7.5%, and 10%. All the photocatalyst materials were calcined at 200 °C for 2 h before being used for the bacterial disinfection process. 2.2. Photocatalyst characterization The prepared Ag–TiO2/GF photocatalysts were first analyzed by X-ray diffraction (XRD) methods using a Bruker AXN model with a Cu Ka radiation (k = 1.5418 Å) source operated at a scan rate of 0.02° s1 over a 2h range of 10–80°. A Hitachi S-4700 scanning electron microscope (SEM) was used to determine surface morphologies of the metal-doped TiO2/GF photocatalysts. A Thermo Fisher K-alpha model was used for X-ray photoelectron spectroscopy (XPS) measurements to determine the elemental states of Ag doped TiO2/GF photocatalysts. The photocatalysts were further examined by an Evolution 300 spectrophotometer (UV-1700 Shimadzu) to obtain UV–Vis absorption spectra at 300–700 nm wavelengths. 2.3. E. coli disinfection experiment For bacterial disinfection efficiency tests, we used Escherichia coli (E. coli) TOP10 as bacterial target (obtained from an environmental bioengineering laboratory, University of Ulsan, Korea). A Luria–Bertani medium containing bacto tryptone (10 g/l), bacto yeast extract (5 g/l), and NaCl (5 g/l) was utilized for the cultivation of bacterium (E. coli) in this study. Bacteria were incubated at 37 °C for 12 h on a rotary shaker. Cultures were centrifuged at 5000 rpm for 10 min. Next, the centrifuged cultures were coated on the cotton to be used as a bacterial source for the disinfection experiments. Tris-buffered saline (TBS) solution, a mixture of 0.15-M NaCl and 0.05 M Tris–HCl at pH 7.5, was used to transfer E. coli from the filter to a test solution. Then, the solution was diluted with saline and spread onto an agar plate. The agar plate was kept at 37 °C for 8 h for bacterial growth, after which the resulting bacterial colonies were counted. The count of bacteria contained in the filters located in the pipe ends was used as a reference for the disinfection efficiency evaluation of bacteria after the disinfection experiments. An experimental model, involving a bacterial source cask connected to the reaction chamber by two pipes, was designed to test the E. coli disinfection efficiency using the prepared photocatalysts (Fig. 1). In the bacterial source cask, a fan was placed on the central ceiling to create a homogeneous environment for the disinfection test. Bacteria were placed on the center of the cask

2. Materials and methods 2.1. Photocatalyst preparation Tetra isopropyl ortho titanate (TIOT) and nitric acid (HNO3) were dissolved with stirring for 8 h in distilled water to make a clear solution of TiO2. Then, the clear solution was maintained at 60 °C for 8 h to prepare a colloidal solution of titania, which was used for the TiO2 coating on the glass fiber. The glass fiber was immersed in the colloidal solution, after which the TiO2-coated glass fiber was taken out and dried at 100 °C for 8 h. The immersion-drying process was repeated several times to produce the specified TiO2 layer thickness. The weight fraction of TiO2 in the

25

Fig. 1. Schematic model for the bacterial photocatalytic disinfection system.

26

T.-D. Pham, B.-K. Lee / Journal of Colloid and Interface Science 428 (2014) 24–31

bottom to create bacterial sources for disinfection experiments. Two pipes of the same size were placed at the same height and distance from the bottom and the lateral sides of the bacterial source cask. Each pipe in the reaction chamber has two shelves inside. One pipe system includes the photocatalyst, which is placed on each shelf. The other pipe system has only glass fiber on the shelves, without photocatalyst. The reaction chamber has a dark cover to isolate light from the outside from the inside of the cask. Thus, outside light could not penetrate to the inside of the chamber, preventing interference with the photocatalytic activities. Four 20 W white-light lamps were placed at the top of the reaction chamber to generate visible light for the photocatalytic disinfection experiments. Filters were placed at the end of each pipe to collect bacteria that were still alive after the disinfection process. During the disinfection process, in which the reaction time was 1 h, the temperature inside the experimental model system was 25 °C. In this study, the photocatalytic disinfection efficiencies of E. coli using 1%, 2.5%, 5%, 7.5% and 10% Ag–TiO2/GF photocatalysts were compared in order to investigate the optimal content of Ag doped on the TiO2/GF. The identified optimal photocatalyst was used for photocatalytic disinfection of E. coli under different relative humidity (RH) conditions (40 ± 5%, 60 ± 5%, 80 ± 5%) to determine the effect of humidity on photocatalytic disinfection. This study also analyzed the effects of different input amounts of E. coli to determine disinfection capacity of the identified optimal photocatalyst while keeping the same reaction conditions of time, temperature and relative humidity. The E. coli photocatalytic disinfection ratio was calculated by Eq. (1):

Fig. 2. XRD patterns of the TiO2 and 1%, 2.5%, 5%, 7.5%, and 10% Ag–TiO2.

Photocatalytic disinfection ratio ¼

E: coli input  E: coli output  100% E: coli input

ð1Þ

The E. coli disinfection capacity was calculated using Eq. (2) with a disinfection time (t) of 3600 s and a photocatalytic area (S) of 240 cm2:

E: coli disinfection capacity ¼

E: coli input  E: coli output Photocatalyst surface area  Disinfection time

ð2Þ

3. Results and discussion 3.1. Photocatalyst characterization 3.1.1. XRD analysis In the Ag–TiO2/GF catalyst system, the intensity of the X-ray diffraction peaks of the TiO2 coated on the glass fiber was less intense for phase analysis due to the limited amount of coated TiO2 photocatalysts. To determine the crystalline phase of TiO2 and Ag deposited on TiO2 using XRD peak analysis, TiO2 and 1%, 2.5%, 5%, 7.5% and 10% Ag–TiO2 powder were prepared from the same deposition solution without the addition of the glass fiber. Only the anatase phase of TiO2 was identified in TiO2 and 1%, 2.5%, 5%, 7.5% and 10% Ag–TiO2 X-ray diffraction peaks shown in Fig. 2. This study investigated the magnified XRD patterns in the 2h range 20–30° to analyze changes in X-ray peak intensity of the anatase peak of TiO2 and 1%, 2.5%, 5%, 7.5% and 10% Ag–TiO2 as a function of the amount of doped Ag (Fig. 3). Fig. 3 shows that, the diffraction angles, occured at 26° in 2h in the XRD patterns, correspond to the anatase (1 0 1) peak of TiO2. The anatase peak intensity and shape of the TiO2 in the Ag–TiO2 samples decreased and broadened, respectively, with increasing Ag doping as compared to the pure TiO2 sample, possibly due to the fact that Ag+

Fig. 3. XRD patterns of the TiO2 and 1%, 2.5%, 5%, 7.5%, and 10% Ag–TiO2 in the 2h range of 20–30°.

ions, deposited on the surface of TiO2 particles, suppressed the crystallization of the TiO2 anatase phase [31]. In addition, the anatase (1 0 1) peaks of TiO2 in the Ag–TiO2 samples slight shifted to a smaller diffraction angle. This is because the diffusion and rearrangement of the Ti4+ and O2 ions in the anatase grain boundaries could be disturbed by the Ag+ ions being spread onto the anatase grains with such a high density, leading to distortion in the crystal lattice of TiO2 [32,39]. The radius of Ti4+ and Ag+ ions is 68 Å and 126 Å, respectively. However, silver could still defect the lattice of TiO2 because the substitution of silver ions into the lattice sites of TiO2 induced O vacancies or deficiencies of Ti4+, which results in greater peak broadening, increased shifting to lower angles and greater decreases in anatase peak intensity [39]. The magnified XRD patterns of TiO2 and 1%, 2.5%, 5%, 7.5%, and 10% Ag doped TiO2 in the 2h range of 30–70° are shown in Fig. 4. In the XRD pattern of Ag–TiO2, the Ag metallic peak was shown at 44.4°, and one more weak Ag metallic peak occurred at 64.5° when the Ag content increased up to 5%. The appearance of metallic Ag could be explained by the following reactions:

AgNO3 ! Ag2 O þ NO2 þ NO

ð3Þ

T.-D. Pham, B.-K. Lee / Journal of Colloid and Interface Science 428 (2014) 24–31

27

adsorption of Ag–TiO2/GF positive related to Ag content. It is because the plasmon resonance of Ag particles dispersing into the TiO2 layer positive related to Ag content. The increase in the plasmon resonance can facilitate electron–hole production resulting in increased visible adsorption [41]. However, the visible light adsorption of Ag–TiO2/GF began to drop when Ag content was 10%. It is because at the high silver loading, the size of the silver clusters became too large, thus blocking light from reaching the TiO2 surface resulting in the chances of visible light reaching the TiO2 layer decreased. Therefore, visible light adsorption of the photocatalyst decreased [42].

Fig. 4. XRD patterns of the TiO2 and 1%, 2.5%, 5%, 7.5%, and 10% Ag–TiO2 in the 2h range of 30–70°.

Ag2 O ! Ag þ O2

ð4Þ

When the Ag content in Ag doped TiO2 samples increased from 5% to 10%, a weak Ag2O peak was also notably observed at 33.1° in the XRD pattern of Ag–TiO2. The occurrence of a small amount of silver oxide (Ag2O) in the samples suggested that at the calcination temperature around 200 °C, the Ag2O produced from thermal decomposition of AgNO3 by reaction (3), have not been totally decomposed to Ag metallic. Even though, the Ag–O bonding is much weaker than Ag–Ag bonding, there is a possibility of small amounts of Ag2O remaining from the decomposition of AgNO3 at calcination temperature around 200 °C [32].

3.1.4. Elemental states The high-resolution XPS spectra of titanium (Ti) in TiO2/GF and 1%, 2.5%, 5%, 7.5% and 10% Ag–TiO2/GF indicated that the Ti 2p3/2 peak and the Ti 2p1/2 peak of TiO2 in the TiO2/GF samples appeared at binding energies of 459.18 eV and 465.18 eV, respectively. The narrow, sharp peak and the doublet splitting energy of Ti 2p peaks, which was 6.0 eV, of Ti 2p in XPS spectra of TiO2/GF indicated that all Ti ions in TiO2 existed in the Ti4+ state [42]. As compared to TiO2/GF photocatalyst, the Ti 2p peaks of TiO2 in Ag–TiO2/GF showed peak broadening and a shift into lower binding energy (Fig. 7). The peak broadening of Ti 2p peaks in Ag–TiO2/GF implies that the ion state of titanium in Ag–TiO2/GF was not only Ti4+. Fig. 8 shows Ti 2p peaks of TiO2 in the Ag–TiO2/GF, which were fit by using Gaussian multipeak shapes, indicating that the Ti 2p matched with Ti3+ and Ti4+ [43]. The Ti3+ might have formed as a result of the reduction of Ti4+ during the calcination process to prepare the TiO2 layer with the assistance of Ag or Ag+ ion. Ag or Ag+ ion can extract an electron from isopropyl radical (C3H7) during the calcination process, and then the extracted electron transfers to Ti4+ to form Ti3+ [40,44]. These reactions are summarized as follows:

e þ Ag ! Ag 4þ

3.1.2. Photocatalyst morphology The SEM images of 0%, 1%, 2.5%, 5%, 7.5% and 10% Ag–TiO2/GF photocatalysts are shown in Fig. 5. Fig. 5a shows that TiO2 in the TiO2/GF was flat-deposited on the surface of the glass fiber, indicating that the TiO2 had successfully coated the surface. Ag was dispersedly coated on to the TiO2 layer, displaying rough surface morphologies composed of small aggregated particles, which indicates that the specific surface area of the Ag–TiO2/GF was larger than the surface area of the pure glass fiber or TiO2/GF. Thus, the Ag–TiO2/GF had more reactive sites or surfaces that could be available for adsorption or disinfection of bacteria. The Ag particle size in Ag–TiO2/GF was positive related to Ag content. Ag clustered into big particles on the TiO2 layer at high Ag-doping. 3.1.3. UV–Visible absorption spectra The optical absorption properties of the TiO2/GF and TiO2/GF and 1%, 2.5%, 5%, 7.5% and 10% Ag–TiO2/GF are shown in Fig. 6. As compared with the TiO2/GF, the Ag–TiO2/GF exhibited a redshift of absorbance edge and a significant enhancement of light absorption in the region of visible light (400–800 nm). This property is due to the Fermi level of Ag is lower than that of TiO2. The literature value for the work function of bulk Ag is 4.64 eV, which positions within the TiO2 band gap (top of valence band 7.6 eV, bottom of conduction band 4.4 eV) [35,40]. Thus, Ag acts as intermediate agent for the transfer of photo-generated electrons from the valance band of TiO2 to an acceptor. The recombination of electrons and holes is also prevented due to the migration of photo-generated electrons to Ag clusters, hence increasing the visible light absorption efficiency of Ag–TiO2/GF. The visible

Ag þ Ti

ð5Þ 3þ

! Ag þ Ti

ð6Þ

During the calcination process of tetra isopropyl ortho titanate (TIOT), which formula is Ti(OCH(CH3)2)4, the Ti4+ and O2 would be rearranged to form TiO2. Ag+ could disturb the rearrangement process resulting in changes in the TiO2 lattice [31,39]. When an Ag+ disturbs the TiO2 formation, Ag+ could replace Ti4+ ion in a TiO2 lattice site, inducing O vacancies or deficiencies of Ti4+ in the TiO2 lattice, resulting in some changes of Ti4+ ions into Ti3+ ions [39]. Because the radius of Ag+ (126 Å) is much larger than that of Ti4+ (68 Å), the replacement of Ag+ into the TiO2 lattice requires energy and thus only a small fraction of Ti4+ is replaced by Ag+ leading to only small amounts of Ti3+. Therefore, the peak intensity of Ti4+ is much higher than that of Ti3+ or the fraction of Ti4+ in TiO2 lattice is dominant compared to Ti3+ (Fig. 8). In addition, Fig. 8 shows that the Ti4+ and Ti3+ peaks shifted into lower binding energy fields as the doped amount of silver in the TiO2 lattice increased. It is because the plasmon resonance of silver could expand the titanium ion radius resulting in electron movement far from the titanium nuclei [45]. Therefore, the peaks of Ti 2p shifted into a lower binding energy. The increase in the plasmon resonance of silver positive related to silver content in the Ag–TiO2/GF. The plasmon resonance of Ag present on the outside of the titanium ions can contribute to the expansion of the electron cloud around titanium nuclei thereby increasing the titanium radius and the further release of electrons. Therefore, electron–hole pairs of TiO2 could be easily generated when the photocatalyst is excited by light irradiation. On the other hand, the occurrence of Ti3+, which contains one more electron than Ti4+, could promote electron generation from TiO2 when the Ag–TiO2/GF photocatalysts are excited by

28

T.-D. Pham, B.-K. Lee / Journal of Colloid and Interface Science 428 (2014) 24–31

Fig. 5. SEM photographs of 0% (a), 1% (b), 2.5% (c), 5% (d), 7.5% (e) and 10% (f) Ag–TiO2/GF.

Fig. 6. UV–Visible absorption spectra of the TiO2/GF and 1%, 2.5%, 5%, 7.5% and 10% Ag–TiO2/GF.

Fig. 7. High resolution XPS spectra of Ti 2p of TiO2/GF and 1%, 2.5%, 5%, 7.5% and 10% Ag–TiO2/GF.

T.-D. Pham, B.-K. Lee / Journal of Colloid and Interface Science 428 (2014) 24–31

29

Fig. 9. XPS of Ag 3d in 7.5% Ag–TiO2/GF.

process of AgNO3 due to the reciprocal conversion between Ag0 and Ag+ by the following reaction: 200  C

4Ag þ O2 ¢ 2Ag2 O 450  C

ð7Þ

Both Ag and Ag2O were induced and deposited on the surface of TiO2/GF photocatalysts. Therefore, the photocatalytic activity of TiO2 was enhanced due to the sensitization of the plasmon resonance generating from Ag metallic as well as Ag2O [42,45]. 3.2. E. coli disinfection results

Fig. 8. High resolution XPS spectra of Ti 2p in X% Ag–TiO2/GF.

light irradiation. Hence, the generation of electron–hole pairs from Ag–TiO2/GF photocatalysts could require relatively lower energy to reach excitation for photoreaction. This is the reason that photocatalytic activity can occur even under visible light irradiation. Silver ions can make more complexes around titanium if silver content is too high, i.e. 10%. There are several silver ions existing at the opposite site of the titanium molecules, an electron cloud can develop opposite, thus withdrawing forces onto titanium, resulting in decrease in the expansion power of the titanium radius. Thus, as compared to Ti 2p peaks of TiO2 in 7.5% Ag–TiO2/GF, those peaks in 10% Ag–TiO2/GF slightly shifted into higher binding energy fields. Fig. 9 shows the high-resolution XPS spectra of Ag 3d of Ag in 7.5% Ag–TiO2/GF. The Gauss multipeak shapes were also applied to fit Ag 3d peaks of Ag in the 7.5% Ag–TiO2/GF. The XPS spectra showed the presence of two Ag 3d peaks matching for Ag in Ag2O and Ag metallic (Fig. 9). Zhang et al. also reported the simultaneous occurrence of Ag and Ag2O in Ag–TiO2 when AgNO3 experience thermal decomposition at the calcination temperature around 200 °C [42]. Ag and Ag2O were produced during calcination

3.2.1. Optimal Ag content The E. coli disinfection results at different Ag loading amounts in Ag–TiO2/GF under visible light are shown in Table 1. The Ag–TiO2/ GF showed very high photocatalytic disinfection efficiencies of E. coli, even under visible light. As compared to TiO2/GF, the great improvement of E. coli disinfection by the photocatalyst Ag–TiO2/ GF is a result of the Ag doping agent. The first reason is Ag could enhance the electron–hole separation of TiO2 resulting in improved photocatalytic activity under visible light. The second reason is due to the defective effects of Ag into TiO2 matrix, resulting in the formation of Ti3+ leading to increase the electron–hole pair generation capacity of photocatalysts. Therefore, the electron–hole pairs could be produced when the Ag–TiO2/GF was irradiated even by visible light, and then the generated electron–hole pairs reacted with water and oxygen absorbed on the surface of photocatalyst to form hydroxyl radicals (OH) and superoxide radical anions (O 2 ) [32]. These oxy radicals could decompose organic components of E. coli resulting in death of the bacteria [46]. In Ag–TiO2/GF, Ag could also act as an intermediate agent to transfer photo-generated electrons from TiO2 to an acceptor (O2 gas) to produce superoxide radicals. The mechanism of the Ag–TiO2/GF photocatalysts working as disinfection agent mainly undergoes the following reactions:

TiO2 þ hm ! e þ h

þ

þ

h þ H2 O ! Hþ þ  OH

ð8Þ ð9Þ

e þ Ag ! Ag

ð10Þ

Ag þ O2 ! Ag þ O2

ð11Þ

O2 þ e þ Hþ ! H2 O2

ð12Þ

30

T.-D. Pham, B.-K. Lee / Journal of Colloid and Interface Science 428 (2014) 24–31

Table 1 E. coli disinfection results of different Ag loadings in Ag–TiO2/GF. Ag doping in TiO2 (%)

0

1

2.5

5

7.5

10

E. coli input (CFU) E. coli output (CFU) Disinfection ratio (%)

2.14E + 07 2.08E + 07 2.80

2.32E + 07 5.20E + 06 77.59

2.33E + 07 3.58E + 06 84.64

2.36E + 07 2.21E + 06 90.64

2.38E + 07 1.54E + 06 93.53

2.40E + 07 1.96E + 06 91.83

H2 O2 !  OH

ð13Þ



ð14Þ

OH þ organic compounds of E: coli ! CO2 þ H2 O

The photocatalytic disinfection efficiencies of E. coli by the Ag–TiO2/ GF different amounts of Ag doping are also shown in Table 1. The photocatalytic disinfection efficiencies of E. coli positive related to Ag content in Ag–TiO2/GF photocatalyst. The highest disinfection efficiency of E. coli was 93.53% with Ag–TiO2/GF at 7.5% Ag loading. The disinfection efficiency began to drop when the Ag content was increased up to 10%. The obtained result coincided with the preliminary conclusion inferred from characterization methods including SEM, UV–Vis and XPS. When the silver content is increased too much, Ag on the TiO2 surface limited the amount of light, H2O and O2 that were able to reach the TiO2 layer, resulting in decreased hydroxyl radical generation, which is a main factor for E. coli disinfection [47]. The radius shrinkage of the titanium ion, was evident in the peak shift to the high binding energy of Ti 2p of TiO2 in 10% Ag–TiO2/GF, indicated that the 10% Ag–TiO2/GF has more difficultly in generating electron–hole pairs as compared to the 7.5% Ag–TiO2/GF. Therefore, the photocatalytic disinfection efficiency of E. coli by the 10% Ag–TiO2/GF will be lower than that by the 7.5% Ag–TiO2/GF. 3.2.2. Humidity effects to E. coli disinfection Table 2 shows the effects of relative humidity on photocatalytic disinfection efficiency of E. coli by 7.5% Ag–TiO2/GF. Among the three tested relative humidity (RH) conditions, the highest disinfection efficiency of E. coli, which was 95.53%, was observed at RH 60 ± 5%. The photocatalytic disinfection efficiency of E. coli was 79.49% under the dry humidity condition (40 ± 5% RH). The decrease in photocatalytic disinfection efficiency of E. coli under the dry humidity condition could be due to the lack of available H2O molecules for hydroxyl radical generation by photocatalytic reaction. Li et al. reported that hydroxyl radical generation or availability was a major factor affecting bacterial disinfection, thus, the lack of hydroxyl radical would greatly decrease disinfection efficiency [47]. The photocatalytic disinfection efficiency of E. coli was 87.67% at the humid condition (80 ± 5% RH). The photocatalytic disinfection efficiency of E. coli at the humid condition was higher than that at the dry condition, but lower than that of the intermediate humidity condition. It is because high humidity could induce reactivation of organisms, or water may occupy most of the TiO2 sites resulting in fewer sites available for disinfection or adsorption of microorganisms [48]. The oxy radical formation can even decrease at a certain humidity level because H2O could occupy the adsorption site on the TiO2 surface resulting in decreased disinfection efficiency at high humidity conditions [48]. At a certain level of relative humidity, the biopolymers of the cell wall membrane or protein structure of bacteria could be

Table 2 Effects of relative humidity on photocatalytic disinfection efficiency of E. coli. Relative humidity (%)

40 ± 5

60 ± 5

80 ± 5

E. coli input (CFU) E. coli output (CFU) Disinfection ratio (%)

2.32E + 07 4.84E + 06 79.14

2.38E + 07 1.54E + 06 93.53

2.40E + 07 2.96E + 06 87.67

changed to be stronger to protect them from oxidation of oxy radicals [49]. Therefore, bacteria can better survive when humidity is high, resulting in a slight decrease in the disinfection efficiency of E. coli. 3.2.3. E. coli disinfection capacity The disinfection capacities of E. coli by 7.5% Ag–TiO2/GF photocatalyst under visible light when the E. coli input amount ranged from 1.65  107 to 2.93  107 (CFU) are shown in Table 3. The photocatalytic disinfection capacity of E. coli was calculated in CFU/s as well as CFU/s cm2. Table 3 shows that the photocatalytic disinfection capacity increased as the E. coli input increased. The photocatalytic disinfection capacity almost reached the limit when E. coli input was higher than 2.56  107 (CFU). The photocatalytic disinfection capacity increased from 4244 to 6269 (CFU/s) when the E. coli input increased from 1.65  107 to 2.93  107 (CFU). The photocatalytic disinfection capacity of E. coli seemed to be stable around 6200 (CFU/s). Thus, there was only a slight increase in the photocatalytic disinfection efficiency accompanying the substantial increase in E. coli remaining without disinfection even with further increased in E. coli input. This is because of the limited number of available hydroxyl radicals generated per second by the photocatalyst under the excitation condition resulting in an increase in untreated E. coli, which have been not disinfected by the photocatalytic system. Although there is slight increase in photocatalytic disinfection capacity of E. coli after 6200 (CFU/s), the disinfection capacity of E. coli in CFU/s cm2 did not increase further after reaching 26 (CFU/s cm2) (Fig. 10). The amount E. coli, which was still alive and on the filter after the photocatalytic disinfection experiments, is shown in Fig. 11. When the E. coli input changed from 1.65  107 to 2.38  107 (CFU), the E. coli remaining on the filter after the photocatalytic disinfection experiments small increased from 339 to 428 (CFU). The obtained result suggests that there was a small amount of E. coli input that was not affected by hydroxyl radicals available on the surface of the photocatalyst when the E. coli input was under 2.38  107 (CFU). However, Fig. 11 shows that there was great increase in the E. coli that was still alive and on the filters after the photocatalytic disinfection as the E. coli input increased from 2.38  107 to 2.93  107 (CFU). This indicated that a large amount of E. coli was just passing through the photocatalytic pipe without being disinfected even with hydroxyl radicals generated from the surface of the photocatalyst. The obtained result confirmed that there is a limited number of generated hydroxyl radicals available for disinfection under the given photocatalytic system. Table 3 Disinfection capacity of 7.5% Ag–TiO2/GF with different E. coli inputs. E. coli input (CFU)

E. coli output (CFU)

Disinfection capacity (CFU/s)

(CFU/s cm2)

1.65E + 07 1.83E + 07 2.04E + 07 2.22E + 07 2.38E + 07 2.56E + 07 2.75E + 07 2.93E + 07

1.22E + 06 1.34E + 06 1.38E + 06 1.47E + 06 1.54E + 06 3.16E + 06 5.00E + 06 6.73E + 06

4244 4711 5283 5758 6183 6233 6250 6269

18 20 22 24 26 26 26 26

E. coli remained (CFU/s) 339 372 383 408 428 878 1389 1869

Disinfection capacity (CFU.s-1.cm-2 )

T.-D. Pham, B.-K. Lee / Journal of Colloid and Interface Science 428 (2014) 24–31

30

GF photocatalyst under intermediate humidity conditions was 6250 (CFU/s) or 26 (CFU/s cm2). Acknowledgment

25

This work was supported by the National Research Foundation of Korea (NRF) Grant funded by the Ministry of Education (2013R1A2A2A03013138). 20

References

15

1.5x107

2x107

2.5x107

3x107

E. coli iput (CFU) Fig. 10. Disinfection capacity of E. coli by 7.5% Ag–TiO2/GF photocatalyst at different E. coli inputs.

2000

E. coli remained (CFU)

31

1500

1000

500

0

1.5x107

2x107

2.5x107

3x107

E. coli input (CFU) Fig. 11. Remaining E. coli at different input amounts.

4. Conclusion The chemical mechanism of photocatalytic disinfection of E. coli using Ag–TiO2/GF under visible light was successfully investigated in the presented study. Ag doping agent enhanced the electron– hole pair separation efficiency and charge transfer efficiency of TiO2 resulting in increase photocatalytic activity of the prepared photocatalyst. Increased silver loading produced a greater number of radicals, which can increase E. coli disinfection via increased electron–hole pair separation and charge transfer efficiency of generated electrons. However, when Ag content too much increase, the photocatalytic disinfection efficiency decreased because too much Ag on the TiO2 layer blocked effective contact between TiO2 and H2O, resulting in decrease oxidative radical formation, as well as between E. coli and oxidative radicals. Optimal humidity conditions for the E. coli disinfection using Ag–TiO2/GF photocatalyst system were intermediate humidity condition (60 ± 5% RH). Both dry (40 ± 5% RH) and humid (80 ± 5% RH) conditions decreased the photocatalytic disinfection efficiency because of the lack of available H2O to produce hydroxyl radicals in the dry condition and because of the blockage of effective contact between E. coli and radicals by too much H2O in the humid condition. The 7.5% Ag loading in the Ag–TiO2/GF photocatalyst and 60 ± 5% RH were the identified optimum silver loading and humidity conditions for the disinfection of E. coli, respectively. The indentified disinfection capacity of E. coli using 7.5% Ag–TiO2/

[1] B.F. Yu, Z.B. Hu, M. Liu, H.L. Yang, Q.X. Kong, Y.H. Liu, Inter. J. Refrig. 32 (2009) 3. [2] W.P. Sung, T.T. Tsai, M.J. Wu, H.J. Wang, R.Y. Surampalli, J. Environ. Eng. 137 (2011) 163. [3] F. Wu, D. Jacobs, C. Mitchell, D. Miller, M.H. Karol, Environ. Health Perspect. 115 (2007) 953. [4] J. Douwes, P. Thorne, N. Peace, D. Heederik, Ann. Occup. Hyg. 47 (2003) 187. [5] Z.A. Nasir, I. Colbeck, S. Sultan, S. Ahmed, Environ. Pollut. 168 (2012) 15. [6] P. Chuaybamroong, R. Chotigawin, S. Supothina, P. Sribenjalux, S. Larpkiattaworn, C.Y. Wu, Indoor Air 20 (2010) 246. [7] J. Zhao, X. Yang, Build. Environ. 38 (2003) 645. [8] H. Shintani, S. Kurosu, A. Miki, F. Hayashi, S. Kato, Biocontrol Sci. 11 (2006) 17. [9] R. Portela, R.F. Tessinari, S. Suárez, S.B. Rasmussen, M.D.H. Alonso, M.C. Canela, P. Avila, B. Sánchez, Environ. Sci. Technol. 46 (2012) 5040. [10] A.L. Linsebigler, G. Lu, J.T. Yates, Chem. Rev. 95 (1995) 735. [11] F. Han, V.S.R. Kambala, M. Srinivasan, D. Rajarathnam, R. Naidu, Appl. Catal. AGen. 359 (2009) 25. [12] X. Hu, G. Li, J.C. Yu, Langmuir 26 (26) (2010) 3031. [13] A. Vohra, D.Y. Goswami, D.A. Deshpande, S.S. Block, Appl. Catal. B-Environ. 65 (2006) 57. [14] J.A. Ibanez, M.I. Litter, R.A. Pizarro, J. Photochem. Photobiol., A-Chem. 157 (2003) 81. [15] P.A. Madrid, G.V.N. Moorillon, E.O. Borunda, M.M. Yoshida, FEMS Microbiol. Lett. 211 (2002) 183. [16] J.A.H. Melian, J.M.D. Rodriguez, A.V. Suarez, E.T. Rendon, C.V.D. Campo, J. Arana, J.P. Pena, Chemosphere 41 (2000) 323. [17] T. Saito, T. Iwase, J. Horie, T. Morioka, J. Photochem. Photobiol. B-Biol. 14 (1992) 369. [18] B. Kim, D. Kim, D. Cho, S. Cho, Chemosphere 52 (2003) 277. [19] J.C. Yu, B. Cheng, X.J. Zhao, Sci. China B 46 (2003) 549. [20] C. Young, T.M. Lim, K. Chiang, J. Scott, R. Amal, Appl. Catal. B-Environ. 78 (2008) 1. [21] I. Paramasivam, J.M. Macak, P. Schmuki, Electrochem. Commun. 10 (2008) 71. [22] S.Z. Hu, A.J. Wang, X. Li, H. Lowe, J. Phys. Chem. Solid 71 (2010) 156. [23] L. Lin, W. Lin, J.L. Xie, Y.X. Zhu, B.Y. Zhao, Y.C. Xie, Appl. Catal. B 65 (2006) 57. [24] H. Park, W. Choi, J. Phys. Chem. B 108 (2004) 4086. [25] H. Yamashita, M. Anpo, Catal. Surveys Asia 8 (2004) 35. [26] E. Bae, W. Choi, J. Park, H.S. Shin, S.B. Kim, J.S. Lee, J. Phys. Chem. B 108 (2004) 14093. [27] C. Karunakaran, G. Abiramasundari, P. Gomathisankar, G. Manikandan, V. Anandi, J. Colloid Interface Sci. 352 (2010) 68. [28] D.W. Sheel, L.A. Brook, I.B. Ditta, P. Evans, H.A. Foster, A. Steele, H.M. Yates, J. Photochem. Photobio. A 187 (2007) 53. [29] J.M. Herrmann, H. Tahiri, Y. Ait-Ichou, G. Lassaletta, A.R. Gonzalez-Elipe, A. Fernandez, Appl. Catal. B-Environ. 13 (1997) 219. [30] D. Cozzoli, R. Comparelli, E. Fanizza, M.L. Curri, A. Agostiano, D. Laub, J. Am. Chem. Soc. 126 (2004) 3868. [31] H.E. Chao, Y.U. Yun, H.U. Xingfang, A. Larbot, J. Eur. Ceram. Soc. 23 (2003) 1457. [32] C. He, Y. Yu, Appl. Surf. Sci. 200 (2002) 239. [33] A. Dobosz, A. Sobczynski, Water Resour. 37 (2003) 1489. [34] L. Zhao, H. Wang, K. Huo, L. Cui, W. Zhang, H. Ni, Y. Zhang, Z. Wu, P.K. Chu, Biomaterials 32 (2011) 5706. [35] H. Kawakami, R. Ilol, L. Straka, S. Papul, J. Romu, H. Hanninen, R. Mahlberg, M. Heikkila, J. Electrochem. Soc. 155 (2008) 62. [36] H. Yu, S.C. Lee, J. Yu, C.H. Ao, J. Mol. Catal. A 246 (2006) 206. [37] D. Robert, A. Piscopo, O. Heintz, J.V. Weber, Catal. Today 54 (1999) 291. [38] T.D. Pham, B.K. Lee, N.M. Nguyen, C.H. Lee, Adv. Mater. Res. 518–523 (2012) 864. [39] Y. Liu, L.C. Yan, R.Q. Hui, Z. Zhang, W.C. Yi, Res. Chem. Intermed. 30 (2004) 569. [40] F. Menga, Z. Sun, Mater. Chem. Phys. 118 (2009) 349. [41] M.A. Behnajady, N. Modirshahla, M. Shokri, B. Rad, Global NEST J. 10 (2008) 1. [42] H. Zhang, G. Wang, D. Chen, X. Li, J. Li, Chem. Mater. 20 (2008) 6543. [43] Y. Fu, H. Du, S. Zhang, W. Huang, Mater. Sci. Eng. 403 (2005) 25. [44] S. Rengaraj, X.Z. Li, J. Mol. Catal. A-Chem. 243 (2006) 60. [45] X. Chen, S.S. Mao, Chem. Rev. 107 (2007) 2891. [46] K. Sunada, T. Watanabe, K. Hashimoto, J. Photochem. Photobiol. A-Chem. 156 (2003) 227. [47] J. Yu, J. Xiong, B. Cheng, S. Liu, Appl. Catal. B-Environ. 60 (2005) 211. [48] Y. Li, C.Y. Wu, Environ. Sci. Technol. 40 (2006) 6444. [49] J. Peccia, H.M. Werth, S. Miller, M. Hernandez, Aerosol Sci. Technol. 35 (2001) 728.