1. Introduction
Humic acids (HAs), a major fraction of natural organic matter (NOM) in surface water, are notoriously known for their detrimental effects in water [1], including turning the water into brown [2], generating complex metals and organic pollutants [3], and reacting with active chlorine during water disinfection, resulting in the formation of trihalomethanes [4]. Therefore, HAs removal is an important task during water treatment.
Nowadays, several techniques have been developed for the treatment of HAs [1]. Removal of HAs by various adsorbents has been studied [5]. However, many of these adsorbents need to be either disposed of or regenerated via a high temperature/acid addition process before reuse, creating a substantial amount of waste containing HAs and spent adsorbent in high concentrations [6]. Recently, removal of NOM from drinking water by advanced oxidation processes (AOPs) has been studied [7]. Extensive research has been devoted to the removal of HAs by photocatalyzed oxidation processes using semiconducting metal oxide nanoparticles like TiO2 [7]. The research efforts in this area are mainly based on assessing the factors that describe the kinetics and mechanistic pathways of photocatalytic removal of HAs.
However, TiO2 has relatively wide band gap of 3.2 eV, which limits the efficiency of photocatalytic reactions, due to the high recombination rate of photogenerated electrons (e−) and holes (h+) [8]. Various metal ions (Zn, Cu, Cr, Mn, Fe, and so on) have been doped into TiO2 to narrow the Eg of TiO2 [7]. In addition, it has been reported that simultaneously doping of two dopants could enhance the photocatalytic activity of TiO2 as compared to the enhancement through single element doping [9, 10]. Studies were focused on the rare-earth metals, such as Sr, Er and Ce [8, 10, 11], which are harmful for human bodies. The metals of Fe and Zn are more suitable to be used in organic matter photodegradation [12].
In addition, the efficiency of the photocatalytic oxidation of water can be influenced by many parameters. Radical scavenging and catalyst inactivation induced by inorganic substances present in water can affect the photocatalytic oxidation process [13].
The main goal of this paper is to optimize the preparation method of Fe-Zn co-doped TiO2 catalyst, as well as to study the effect of the following parameters on the photocatalytic oxidation: (a) pH, (b) some natural anions including HCO3−, CO32−, SO42− and Cl−, (c) some cations including Mg2+ and Ca2+, and (d) the effect of tap water compared to Milli-Q water as solvent.
2. Methods
2.1. Equipment
The photocatalytic reactor was a cylindrical glass column with a diameter and height of 60 and 850 mm, respectively. A 37 W low-pressure mercury vapor lamp (Φ22 mm × 790 mm; Beijing Haili Lighting Equipment Company, China) with UV wavelength of about 254 nm and UV radiation intensity of about 110 μW/cm2 was found inside the column. The O2 gas, which was prepared by an oxygenerator (Beijing North Star Yaao Scitech Co., Ltd., China), was fed into the reactor through a porous glass core aeration plate (pore size 4–7 μm).
2.2. Preparation of TiO2
Tetrabutyl orthotitanate (analytical reagent, Tianjin) of 68 mL was dissolved in 120 mL ethylalcohol, and the obtained mixed solution was stirred for 2 h. Afterward, a series of FeCl3 or ZnCl2 solutions with different concentrations (atomic ratios of Fe/Zn to Ti of 0.0000% – 0.0100%), glacial acetic acid (80 mL), ethylalcohol (120 mL) and Milli-Q water (30 mL) were mixed together and added to the mixed solution in drops. The final mixed solution was placed at 30°C for 7 d, dried at 105°C and heated at 400–600°C. The nanometer-sized TiO2 was obtained after adequate mulling.
2.3. Characterization of TiO2
Scanning electron microscopic (SEM) images were obtained with an EVO18 electron microscope (Carl Zeiss, Germany). X-ray diffraction (XRD) patterns were collected in a Rigaku Dmax-RB diffractometer (Tokyo, Japan). The BET surface areas (SBET) of the samples were determined by Quadrasorb SI-MP apparatus (Quantachrome Instrument, USA). Diffuse reflectance spectroscopy (DRS) was performed using a HITACHI U-3010 UV-Vis scanning spectrophotometer (Tokyo, Japan).
2.4. Photocatalytic Activity Tests
The aqueous slurry of 1,500 mL (pH 7), with 50 mg/L HAs (chemical purity, Tianjin) and 0.1 g/L TiO2, was stirred and bubbled with O2 (1.0 L/min) for 30 min prior to irradiation using the UV lamp. The reaction was stopped at 5 min intervals, and clear solution was obtained after centrifugation at 2,000 r/min. The HAs concentrations were determined using a UV-Vis spectrophotometer (Hach DR5000, USA) at 254 nm. The tests were repeated 3 times to minimize the errors.
2.5. Effects of the pH Value, the Inorganic Ions and the NOM
In a supplementary experiment, the initial solution pH value was adjusted at the beginning of the oxidation by addition of HCl or NaOH to the HAs solution. The HAs removal efficiencies for the initial pH value of 2, 4, 6, 8, 10 and 12 in the presence of the optimized catalyst were determined.
The influence of inorganic anions, including HCO3−, CO32−, SO42− and Cl−, and cations including Mg2+ and Ca2+, were also investigated using the optimized catalyst. Different concentrations (0.5, 1.0, 1.5, 2.0 and 2.5 mmol/L) of NaHCO3, Na2CO3, Na2SO4, NaCl, MgCl2 and CaCl2, similar with the concentration level in natural water, were added into the water with the initial pH value of 7.0, and the HAs concentrations were determined every 5 min.
In order to obtain the NOM removal efficiency, the tap water with UV254 of 0.036 /cm and total organic carbon (TOC) of 5.5 mg/L was used. HAs of 50 mg/L was added in to the tap water, and the UV254 and TOC were measured to be 1.272 /cm and 16.5 mg/L, respectively. TOC measurements were performed by the spectrophotometer using low-range TOC ampoules (Hach Chemical, USA).
3. Results and Discussion
3.1. HAs Removal
3.1.1. Single ion-doped TiO2
Fig. 1 shows the influence of Fe or Zn dopants in the catalyst on the degradation of HAs. It can be seen that Fe doping increased the photocatalytic activity of TiO2. That was because the ionic radii of Fe3+ (55 pm) was smaller than that of Ti4+ (60.5 pm); therefore, Fe3+ could replace Ti4+, and highly disperse in TiO2 lattice, resulting in the improvement of the photocatalytic activity [7].
For Zn-doped TiO2, Zn2+ ions (ionic radii of 74 pm, a little larger than that of Ti4+) were partially dispersed in the bulk of TiO2; ZnO might be formed on the surface of TiO2, and inhibiting the growth of TiO2 [14]. In addition, the e− and h+ transfer occur between the conduction band of ZnO and the conduction band of TiO2. This efficient charge separation increases the photocatalytic activity of TiO2/ZnO composite [15].
The optimal calcination temperature was 500°C for both Fe-doped and Zn-doped TiO2, and the best photocatalytic activities were achieved when the Fe3+ or Zn2+ doping amount was 0.0010%, corresponding to the HAs removal efficiencies of 57.4% and 53.7%, respectively. The photocatalytic activities of Fe-doped or Zn-doped TiO2 depended on the calcination temperature and the dopant concentration which affected the SBET, Eg and crystallinity of the catalysts [10]. The SEM morphology of un-doped TiO2 calcined at 500°C is shown in Fig. 2, and the characteristics of the catalysts are shown in Table 1.
The SEM morphology of un-doped TiO2 calcined at 500°C indicates that the TiO2 nanoparticles are globular and agglomerated. The diameter of TiO2 was in the range of 10–50 nm.
The SBET of the TiO2 decreased as the calcination temperature increase because of the aggregation of the nanoparticles (Table 1). In addition, the doping amount can also affect the SBET values of the ion-doped TiO2. For the catalysts calcined at 400–500°C, the SBET increased with an increase in the content of the dopant. This can be attributed to the segregation of the dopant cations at the grain boundary, which inhibits grain growth by restricting the coalescence of some smaller neighboring grains [16]. For the catalysts calcined at 550–600°C, when the doping amount increased, the SBET increased at first, and then decreased. The reduction in the SBET may be due to blocking of fine capillaries of parent TiO2 surface by metal film islands [17].
A larger SBET indicates a larger amount of active sites on the catalyst surface. Therefore, the photocalalyst with lower calcination temperature and larger dopant amount may have a better catalytic activity when only SBET is considered.
The TiO2 calcined at higher temperatures showed a tendency to red shift, which resulted from the narrower Eg of rutile TiO2 compared with that of anatase TiO2. By increasing the content of the dopants, the Eg were hence decreased because of a charge-transfer transition between the new dopant level near the valence band and the conduction band of the TiO2 [17]. The reduction in the Eg of TiO2 indicated that more light energy could be used during the photocatalytic degradation process, and therefore the photocatalytic activity of the catalyst was enhanced.
As it is known that TiO2 in anatase phase shows better photocatalytic activity, the XRD analysis was conducted. With increasing doping amount of Fe3+ or Zn2+, the diffraction peaks belonging to the (101) peak of anatase phase (2θ at about 25.3°) in Fe/Zn-doped TiO2 powder increased at first, and then decreased; meanwhile, the intensity of the (110) peak of rutile phase (2θ at about 27.4°) increased after doping. This was because the phase transformation temperature reduced after doping [18]. In addition, with the increase of the calcination temperature, the diffraction peak of anatase appeared stronger and sharper, indicating that better crystallites were formed, and the intensity of the rutile phase increased when the calcination temperature increased because of the transformation of TiO2 from anatase phase to rutile phase.
Moreover, the crystallites sizes of the TiO2 catalysts can be calculated by Scherrer Formula [19] according to the XRD results. The particle diameters of the (101) peak of anatase phase for ions-doped TiO2 were in the range of 10–50 nm, which can be confirmed by the SEM analysis (Fig. 2). Differences of the crystallites sizes among the catalysts with diffident dopant amount were not obvious. The most important factor for the crystallites sizes was the calcinations temperature. When the calcination temperature and the dopant amount changed, the variation tendency of the particle diameters for the catalysts was same with that of the SBET.
According to the abovementioned results, the calcination temperature, as well as the dopant amount, can affect the photocatalytic activity of the catalyst by change the characteristics of TiO2. Although the 0.0010% ions-doped TiO2 calcined at 500°C had neither the largest SBET, nor the narrowest Eg, it had a higher catalytic activity for HAs photocatalytic oxidation compared with the other catalysts because of the combined action of the crystallinity, SBET and Eg.
3.1.2. Fe-Zn co-doped TiO2
The HAs removal efficiencies in the presence of the Fe-Zn co-doped TiO2 were higher than those in the presence of single ion-doped TiO2, which could be attributed to the synergetic co-doping effects of Fe and Zn into TiO2. Firstly, the SBET for the 0.0010% Fe-Zn co-doped TiO2 were larger than those of the Fe-doped or Zn-doped TiO2. Doping of a second metal can change the catalyst surface structure because of the conservation of a large number of micropores, and therefore enhances the SBET of the co-doped catalyst [7]. Secondly, the Eg of the co-doped TiO2 was between those of the Fe-doped and Zn-doped TiO2, and was not considerably wider than that of the Fe-doped one. Thirdly, for the Fe-Zn co-doped TiO2, the effect of the dopant type on the anatase peak intensities was not significant when the total doping concentration was conserved; only the anatase phase was found in the co-doped TiO2. Therefore, higher photocatalytic activity was achieved.
The 500°C calcined 0.0010% Fe-Zn co-doped TiO2 with the Fe:Zn ratioratio of 3:2 showed the highest catalytic activity, with the HAs removal efficiency of 65.7%. It was reported by Yang and Lee that 65% of HAs (10 mg/L) was removed after 120 min using the UV/TiO2 system with the TiO2 amount of 2 g/L [20]. Although there was a lower HAs concentration and a higher TiO2 amount, the HAs removal efficiency in this research was a little lower, indicating that higher photocatalytic activity could be achieved when Fe3+ and Zn2+ ions were doped in TiO2. Also, about 50% of HAs can be decomposed by the same process, except that Al3+-Fe3+ co-doped TiO2 was used as the catalysts [9], showing that the Fe3+-Zn2+ TiO2 synthetized by sol-gel method is better for the catalytic reaction.
High removal efficiency of the pollutant through photocatalytic oxidation process can be obtained because of the generation of H2O2 and ·OH [7]. In addition, Fe3+ and Zn2+ in the co-doped TiO2 act as shallow trapping sites for charge carriers (e− and h+) (Eq. (1) and Eq. (2)), and then react with H2O2 and OH− in water to generate ·OH (Eqs. (3)-(6)) [21]. Zn2+ can react with Fe3+ in water, and therefore reduce the recombination rate of photogenerated e− and h+.
where Xn+ represents Fe3+ or Zn2+.
3.2. Effects of Inorganic Ions in Aqueous Solutions
In this study, the optimized photocatalyst (500°C calcined 0.0010% Fe-Zn co-doped TiO2 with the Fe:Zn ratioratio of 3:2) was used, and the HAs removal efficiencies at different pH value, cation or anion concentrations were detected.
3.2.1. Initial solution pH value
It can be seen from Fig. 3 that the HAs removal efficiencies for the initial pH of 2, 4, 6, 8, 10 and 12 were 83.7%, 75.8%, 64.4%, 71.6%, 77.6% and 79.7%, respectively. At acidic conditions (pH increased from 2 to 6), the HAs photocatalytic oxidation efficiency decreased as the pH increased. However, the HAs removal efficiency increased with the solution pH increased from 8 to 12).
Changes in the pH values can affect the photocatalytic degradation rates of HAs because the initial pH value influences the zero-point charge (ZPC) of TiO2. The pHZPC for TiO2 is widely reported between 6 and 7 [22]. Because the doping amount of metal ions was extremely low, the pHZPC for Fe-Zn co-doped TiO2 was detected to be about 6, similar with that for the pure TiO2. When the pH value of the solution was close to 6, the aggregation of the co-doped TiO2 occurred, resulting in the reduction of the SBET of the catalysts. Therefore, the HAs removal for the initial pH of 6 was the lowest.
The surface of the Fe-Zn co-doped TiO2 tends to be positive charged for pH < pHZPC, and becomes negatively charged for pH > pHZPC [23]. The HAs in solution is negatively charged. Due to the electrostatic attraction, the acidic solution favors adsorption of HAs onto TiO2 surface, and the HAs removal efficiency increases accordingly. Besides, at acidic condition, H+ could react with O2 more easily, and more H2O2 would form to promote the photocatalytic oxidation of HAs.
At alkaline condition, OH− can be adsorbed on the surface of the catalysts, and react with the transitional metals (Fe4+ and Zn3+), resulting in the production of ·OH (Eq. (4)).
3.2.2. Cations
Fig. 4 shows that a beneficial effect of cations in water on the degradation of HAs was achieved. At time of 120 min of irradiation, 72.1%, 75.5%, 78.1%, 81.9% and 83.5% of degradations with Fe-Zn co-doped TiO2 were observed for 0.5, 1.0, 1.5, 2.0 and 2.5 mmol/L of Ca2+ in neutral medium, whereas 65.7% degradations occurred in medium without Ca2+. The HAs removal efficiencies were 70.4%, 73.0%, 74.9%, 76.7% and 79.5% for the medium with Mg2+ concentration of 0.5, 1.0, 1.5, 2.0 and 2.5 mmol/L, respectively. The photocatalytic oxidation of HAs was enhanced with the increase of the Ca2+ or Mg2+ concentration.
The enhancement of degradation by addition of these ions is due to increase of charge separation by accepting the conduction band electron [23]. In addition, the addition of Ca2+ or Mg2+ in the solution can affect the adsorption of HAs on the surface of TiO2. For solution without Ca2+ or Mg2+, the pH value of the solution is 7, close to the pHZPC of TiO2 (pH of 6). The adsorption of OH− is a little stronger than that of H+, and the adsorption of HAs with negatively charged is weak. When Ca2+ or Mg2+ ions were added, reactions in Eqs. (7)-(10) occurred.
where M2+ represents Ca2+ or Mg2+.
Ca2+ or Mg2+ shows bridging behavior between TiO2 and HAs (Eq. (9)). Besides, the cations can react with A− ions, resulting in the growing of HAs molecular chain. The longer molecular chain makes it easier for HAs to be adsorbed on the surface of TiO2, and the HAs removal efficiency increases.
3.2.3. Anions
Inorganic anions, including HCO3−, CO32−, SO42− and Cl−, are common in drinking water. The effects of these inorganic anions on HAs photocatalytic oxidation were determined (Fig. 5). When HCO3− with the concentrations of 0.5, 1.0, 1.5, 2.0 and 2.5 mmol/L were added in the solution, the HAs removal efficiencies were 63.3%, 61.8%, 59.8%, 58.0% and 55.8%, respectively. The figures were 61.4% (64.8%), 58.2% (64.1%), 54.5% (64.0%), 52.1% (63.5%) and 48.2% (63.1%) when CO32− (Cl−) with the abovementioned concentrations were added. The photocatalytic activities were inhibited when these anions were added, and the order of inhibition were CO32− > HCO3− > Cl−. However, when SO42− was added at the concentrations of 0.5, 1.0, 1.5, 2.0 and 2.5 mmol/L, a slight increase for HAs removal was observed, and the HAs removal efficiencies were 66.8%, 67.6%, 68.2%, 69.0% and 69.6%, respectively.
The CO32−, HCO3− and Cl− can be adsorbed by the catalyst and result in TiO2 contamination [24]. The inhibition of CO32−, HCO3− and Cl− on photocatalytic activity can be explained with the scavenging of h+ and ·OH radicals by ions [25]. The reactions of the ions with h+ and ·OH are given in Eqs. (11)-(12) [13, 25].
where Yn− represents CO32−, HCO3−, Cl− or SO42−, and ·Y(n−1)− represents ·CO3−, ·HCO3, ·Cl or ·SO4−, respectively.
Although ·CO3−, ·HCO3 and ·Cl can react with HAs, their oxidiz-abilities are limited. As the concentrations of CO32−, HCO3− or Cl− increase, the effect becomes increasingly significant.
CO32− has a higher inhibiting capacity of TiO2 than HCO3− in the photocatalytic degradation of organic compounds. However, the pH value of the 2.5 mmol/L Na2CO3 solution was 11.0, while that of the 2.5 mmol/L NaHCO3 solution was 8.5; it is shown that for alkaline solution, the HAs removal efficiency increased with the solution pH increased. The lower removal efficiency in the presence CO32− was attributed to the weaker oxidizing ability of ·CO3− than that of ·HCO3.
The HAs removal efficiency in the presence of Cl− was not decreased significantly, because Ti-O− is predominant in neutral water, and the adsorption of Cl− can be ignored [24].
3.3. HAs Removal in Tap Water
TOC removal was used to evaluate the actual HAs mineralization degree in tap water caused by the 500°C calcined 0.0010% Fe-Zn (3:2) co-doped TiO2 (Fig. 6). The UV254 removal efficiency increased rapidly during the first 100 min and then became stable because of the degradation of HAs into intermediate molecules during the photocatalysis. Meanwhile, the TOC rapidly decreased all along the 2 h.
For the HAs solution, the characteristic absorption peaks at 254 nm and 400 nm are attributed to benzene components and chromophore groups, respectively [3]. The absorbances of the original HAs at 254 and 400 nm were 0.735 and 0.093, respectively. The HAs absorbance at 254 nm (400 nm) after treatment decreased by 42.8% (40.9%). The TiO2 photocatalytic oxidation effectively destroyed both the benzene components and the chromophore groups. However, the benzene components were a little more difficult to destroy than the chromophore groups.
The TOC removal rate was 49.9%, which indicates that 49.9% of HAs was mineralized into CO2 and H2O. The UV254 removal efficiency (57.2%) was higher than that of TOC because of the generation of intermediates that could significantly reduce the UV254, but not the TOC.
4. Conclusions
HAs were effectively degraded in the presence of Fe-doped, Zn-doped or Fe-Zn co-doped TiO2 under irradiation with UV light. The HAs removal efficiencies were affected by the SBET, Eg and crystallinity of the catalysts. Highest HAs removal efficiency was achieved in the presence of 500°C calcined 0.0010% Fe-Zn co-doped TiO2 with the Fe:Zn ratioratio of 3:2.
Changes in the initial solution pH values can affect the photocatalytic degradation rates of HAs. The HAs removal for the initial pH of 2 was the highest, and for the pH of 6 was the lowest.
The inorganic cations and anions also affect the photocatalytic activity of the co-doped TiO2. The photocatalytic oxidation of HAs was enhanced with the addition Ca2+ or Mg2+ ions. The photocatalytic activities were inhibited when CO32−, HCO3− or Cl− was added, and the order of inhibition was CO32− > HCO3− > Cl−; but the removal of HAs was slightly promoted when SO42− was added.
For tap water added with HAs, the UV254 removal efficiency was higher than that of TOC because of the generation of intermediates that could significantly reduce the UV254, but not the TOC.