Early examples of research into the cleaning properties of TiO2 include those carried out by Frank and Bard. They found that TiO2 powder in contaminated water were able to photocatalyze the conversion of cyanide into cyanate and thus detoxified the water [15, 16].
One of the advantages of TiO2 photocatalysis for water decontamination is that only the TiO2 photocatalyst (immobilized or suspended) and UV light, either from solar light or artificial light sources, are needed and its cost can thus be lower than other kinds of advanced oxidation techniques (UV/O3; UV = H2O2, photo-Fenton). Moreover, no toxic intermediate products are generated in the photocatalytic decontamination process. As a result, TiO2 photocatalysis has been attracting a lot of interest in the areas of water detoxification or drinking water purification [6, 9, 36].
The main potentially-polluting nutrients in relation to water are nitrogen, ammonia, phosphorus and sulphur. Ammonia is a common water contaminant that has notable effects upon the environment and human health when it is in the presence of excess amounts. They arise from the natural breakdown of crop residues and soil organic matter, rainfall, wastewater and industrial effluents…etc. Free ammonia (NH3) and ionized-ammonia (NH4+) represent two forms of reduced inorganic nitrogen. The free ammonia is a gaseous chemical, whereas the NH4+ in the reduced nitrogen form exists in ionized form and remains soluble in water . Ammonia is one of the key matters causing eutrophication pollution. Ammonia brings injurious effects on human and animal health in forms of gas and particle phase [2, 3].
In most cases, contamination of groundwater by chemicals derived from urban and industrial activities, modern agricultural practices, or waste disposal takes place almost imperceptibly. The concentration of nitrogen of the regional groundwater rises along with over fertilization in agriculture [1, 23, 45]. Solid waste produced by both municipal and industrial sectors increased the organic, heavy metals, and inorganic groundwater ions [34, 41, 49]. Excess discharge of domestic and industrial wastewater has also polluted the groundwater . In general, ammonia level in ground waters is below 0.2 mg per litre. Higher natural contents (up to 3 mg/litre) are found in strata rich in humic substances, iron or in forests . Surface waters may contain up to 12 mg/litre . Ammonia may be present in drinking-water as a result of disinfection with chloramines. The maximum limit of ammonia set by the European Association for drinking water is approximately 0.5 mg/l (AWWA, ).
Currently, the most popular treatments for ammonia removal from wastewater are air stripping, ionic exchange and nitrification denitrification techniques where however, ammonia cannot be removed completely in most cases. For this reason in the recent years advanced oxidation processes have been investigated as the alternatives to conventional water treatment.  used the photocatalytic reduction method of Pt / TiO2 nano-photocatalysts with high platinum loading of 0.5 % to achieve the best efficiency of nitrogen removal.  conducted TiO2 photocatalyst doped with metal ions for treating ammonia nitrogen wastewater from coal gasification process. The ammonia nitrogen removal rate was approximately 79 %. Songhua River water treatment using catalyzed ozonation were investigated by . In these cases, Nano TiO2 coated haydite, silica-gel and zeolite were used as the catalyst. The ammonia removal efficiency was close to 80 % after 30 min reaction.
Typically, the studies with respect to photocatalytic reactors adopted either TiO2 powder or suspended counterpart in the water being treated and then went through further process to remove the TiO2. The primary aim of doing so is to avoid the post separation difficulties associated with the powder form of the TiO2 catalyst. Immobilization of TiO2 on various substrates is an important issue in the water treatment applications associated with photocatalytic technologies . There are several advantages in these researches, including higher surface area, superior adsorption properties [24, 53], surface hydroxyl groups increase, and charge recombination reduction  in immobilized systems.
This study attempted to implant immobilized TiO2 onto the small glass beads which were placed into the photocatalytic reactor to investigate and analyze its performance of treating ammonium-rich wastewater. Moreover, we aim to develop a simple, energy efficient, less expensive to build and operate photocatalytic reactor to handle ammonium-rich wastewater. This low cost solution will provide effective and successful experiences for large-scale industrial applications in the future.
Experimental setup and procedure
Experimental apparatus is shown in Fig. 1; it mainly consists of an inner tube and outer borosilicate glass reactor of 500 mL capacity. The inner tube is equipped with a UV lamp (Xin Guang Yuan Lighting, 36 W, with an emission spectrum of 200–600 nm, and λmax at 365 nm), which was mounted axially in the reactor inside a cylindrical, double walled lamp jacket. Instead of ordinary, the glass lamp jacket is made of quartz glass in order to reduce the absorption of ultraviolet light. The outer glass reactor as a photocatalytic reactor is to accommodate the inner tube where TiO2 coated glass beads as the photocatalytic constituent, also as the aqueous samples were placed in. In this study, UV protective enclosure was used to bring out protection effect against exposure to UV radiation.
TiO2 thin film preparation
The TiO2 films were prepared via sol-gel process. Fig. 2 illustrates this procedure in detail. Tetrabutyl titanate ((C4H9O) 4Ti) was used as a precursor to prepare TiO2 sol. CH3CH2OH, ((HOCH2CH2)3 N) and distilled water were added to ((C4H9O) 4Ti) under continuous magnetic agitation at room temperature to form a mixture according to volume ratio of 3: 12: 1: 1. All chemicals used for this study were of analytical grade and from Chengdu Kelong Chemical Co., Ltd. Taking their high mechanical and chemical stability into account, this study selected the glass beads (diameter 5 mm) from Yongqing Ziguang Glass Beads Co., Ltd, as the substrate to be coated with nano-TiO2. Before coating, the glass beads were etched for 24 h in diluted hydrofluoric acid and subsequently were sonicated (DH-120DT, Shanghai Dihoo Instrument) in acetone, absolute ethanol and distilled water for 10 min in order to remove all the organic contaminants. After that, they were dried out at 100 °C by DZF-6050 vacuum dryer (Shanghai Jinghong Laboratory Instrument).
TiO2 thin films were deposited on glass slides by dip coating method at two different vertical withdrawal speeds of 1 cm/min. TiO2 films were deposited several layers on substrates. During successive coatings, each layer was applied vacuum drying process for 30 min at 100 °C. After multiple dip coating was finished, the films were annealed at various temperatures in the range of 400 ~ 650 °C for 2 h at the heating rate 6 °C/min in an atmosphere.
Determination of the analyte concentration in a solution after degradation
In this study, the Beer–Lambert law was applied to predict the linear relationship between the absorbance of the solution and the concentration of the analyte (assuming all other experimental parameters invariant). A series of ammonium chloride NH4Cl standard solutions were prepared. The absorbance of NH4Cl for various concentrations was measured with a UV-vis spectrophotometer (Shanghai Puyuan Instrument) at a wavelength 425 nm, which offered an effective way to engender a calibration curve to show the absorbance variation along with the concentration as shown in Fig. 3. The slope and the intercept of the line provided a relationship between absorbance and concentration:
Experimental procedures followed this way: the process began with degradation before 300 ml aqueous sample was added to the outer container. The UV light was then turned on to start UV photolysis for a period of time. The absorbance of the sample was re-measured when the experiment was over. The degradation rate can be calculated with:
Where η represents degradation rate; A0 represents the initial absorbance of the solution sample; A represents absorbance of the solution sample after a certain time of photocatalytic reaction;represents the initial concentration of solution sample; C represents the concentration of solution sample after degradation.
To ensure reproducibility and decrease the uncertainty, each measurement was repeated 3 times to take the average value. The nonlinear regression analysis was performed to best-fit the response function (photocatalytic degradation efficiency) with the actual experimental data.
Results and discussion
Structural properties of TiO2 thin film
To determine the intrinsic properties of the TiO2 layers, various techniques were used. X-ray diffraction (XRD, model D8 ADVANCE, Bruker. Ltd) analysis of the films on the glass bead substrate was performed to determine the crystalline phase and to estimate the amendment of TiO2 as heating temperature varied. The XRD patterns were recorded by step scanning in the 2θ scan range of 5–80° where a step size was of 0.01°. In Fig. 4, anatase phase and rutile phase are denoted as A and R, respectively.
Before heat treatment, the TiO2 thin layer consisted of amorphous or poorly crystallized oxide. Figure 4 shows the XRD patterns of the TiO2 films annealed at various temperatures. The line at 400 °C shows a prominent anatase peak at 25.4° (101) due to the transition from the amorphous phase to the anatase phase. When the calcined temperature increased to 450 °C, the crystallinity of TiO2 was improved but still stayed on the anatase phase. As the temperature increased from 400 to 600 °C, there was significant growth for the intensities peaks, which revealed the improvement of the crystallinity. And the outcomes are in agreement with the finding demonstrated by Hasan et al. . When temperature went up to 500 °C, the XRD pattern exhibited a rutile peak at 27.4° (110). At this temperature, it is thought that the transformation from anatase to rutile took place. Anatase/rutile mixture phase appeared when the TiO2 thin film was annealed at 500 °C. At 600 °C, the rutile (110) peak went higher while the counterpart of the anatase (101) dropped. As far as the line of 650 °C, the anatase peak disappeared while the the rutile peak increased considerably, indicating that a complete transformation from the anatase phase to the rutile phase occurred. This phenomenon is in compliance with the work done by [26, 51]. They revealed that TiO2 thin films can be transformed from amorphous phase into high degree crystalline anatase and then converted into rutile by temperature treatment. At above 600 °C, x-ray diffraction pattern showed a peak belonging to the rutile peak. Transformation from anatase to rutile is thought to occur at about 500 °C with the complete transformation at higher temperature. The average crystallite sizes of the samples were estimated by the Debye–Scherrer’s equation as expressed below:
where D is the diameter of crystalline in nm, λ is the X-ray wavelength of Cu Kα radiation, βhkl is broadening of the hkl diffraction peak width at half height of the maximum intensity (FWHM) in radians and θhkl is the diffraction angle (Bragg’s angle) in degrees.
For TiO2 thin films that contain anatase and rutile phase only, the percentage of mass fraction (XA) for anatase phase in the solution was calculated using the relative intensity of maximum peak of anatase phase (101) and a maximum peak of rutile phase (110). The XA is expressed by the equation below ;
Where XA, IA, IR refer to a contents of the anatase phase in the solutions, the intensity at the anatase phase peak as well as the intensity at rutile phase peak, respectively. Meanwhile, the constant 1.265 is the scattering coefficient . The rutile weight fraction XR can be calculated by the empirical equation as followed :
The influent factors associated to annealing temperature on the structure of TiO2 were summarized in Table 1. As the table has shown, annealing temperature rise results in bigger crystallite size. For instance, when the annealing temperature increased from 400 to 600 °C, the crystallite size on of anatase phase grew from 14.4 to 18.72 nm. Similarly, the crystallite size of rutile phase grew from 37.06 nm to 67.94 nm with the increasing of annealing temperature from 500 to 650 °C. It is observed that the XA of anatase phase was 86.31 % and it decreased to 26.92 % as the annealing temperature increased from 500 to 600 °C. It is noticeable that the amount of rutile phase increased with annealing temperature rise from 500 to 650 °C.
|T (°C)||Phase||K||λ (nm)||β (radian)||cosθ||hkl||D (nm)||Mass ratio (%)|
TiO2 exists in the form of three polymorphs: anatase, rutile, and brookite . Although rutile phase has less photocatalytic activity than anatase, Anatase/rutile mixture phase is known to exhibit enhanced photoactivity relative to single-phase titania [13, 21, 37, 42, 48]. It is widely recognized that this is a result of improved charge carrier separation, possibly through the trapping of electrons in rutile and the consequent reduction in electron–hole recombination [8, 21]. Anatase-to-rutile phase transformation is governed by the annealing temperature, compactness of the anatase nanocrystallites, and grain boundary defects [28, 30].
The influence of exposure time on the TiO2 films under UV irradiation
Over the trail test, TiO2 thin films were categorized into 4 groups based on their anneal temperatures of 450 °C, 500 °C, 550 °C and 600 °C under UV irradiation. Comparison among the 4 groups was conducted to check the influence of exposure time on ammonia degradation rate. Ammonia content of wastewater was about 700 mg / L. The volume of reaction solution was 300 ml. The solution was adjusted to a pH of 7. The degradation was carried out at room temperature. The photocatalytic reaction time was 30, 60, 90, 120 and 150 min, respectively. The results are shown in Fig. 5.
The results from Fig. 5 indicate that the photocatalytic activities/degradation rate of the thin TiO2 films prepared at 4 different temperatures displayed the same pattern in function of UV exposure time. As shown in Table 1, TiO2 presented photocatalytic activity on anatase phase without phase conversion at 450 °C. Anatase/rutile mixture phase was seen at 500 °C. At this temperature, anatase is converted to rutile with 9.04 % rutile phase in the mixed phase. This mixed phase has higher photocatalytic activities than single anatase phase. When the sample was prepared at 550 °C (Rutile 18.87 %, Anatase 72.88 %), which has the highest photocatalytic activity among all thin films tested. Sample annealing at 600 °C resulted in 62.98 % rutile phase as a predominantly phase in the mixed phase. At this temperature, TiO2 thin film displayed much less photocatalytic activity than other 3 samples. Ding et al  conducted the degradation of organic dye solution using mixed crystal nano TiO2 as a photocatalyst. They found the photocatalytic activity of mixed phase nano-TiO2 crystal samples was significantly higher than that of a single crystal sample or mechanical mixed crystal samples. The photocatalytic activities had the best effect when the mass fraction of rutile phase was about 20 % of mixed crystals. Ji et al  prepared anatase - rutile mixed Nano TiO2 crystal samples to test degradation effect of Methyl orange solution. They found when rutile phase is of 18.9 % of mixed crystal, the photo degradation rate of methyl orange can reach the highest 85 %. Degussa P-25 (a mixed-phase titania photocatalyst) is commercially available and was utilized as a reference material in many studies. This nanocrystalline material, formed by flame pyrolysis, consists of 80 wt% anatase and 20 wt% rutile. Our experimental results were in good agreement with other authors’ findings described above.
As can be seen from Fig. 5, over UV exposure time from 30 to 120 min, it is evident that the percentage of photo-degradation for the four samples increases along with irradiation time growth. After 120 min, the growth of the degradation rate of all samples substantially slowed down and almost ceased. This could be attributed to the following reason: in general, the photoactivity of TiO2 is determined by the processes of electron/hole pair generation, recombination, interfacial transfer and by the surface reactions of these charge carriers with the species adsorbed on the surface of the photocatalyst [[10, 17, 18, 31]]. The presence of molecular oxygen in the solution also plays a substantial role in the photo-induced processes on irradiated TiO2 surfaces because that it enables an effective charge carriers separation. Consequently, both superoxide radical anion and hydrogen peroxide as the most important products of the molecular oxygen reduction play an important role in the complex mechanism of Oxygen Species (ROS, e.g., •OH, •O2H or singlet oxygen) generation on the irradiated TiO2 surfaces [[20, 31, 35, 52] ]. Hydroxyl radicals are powerful and indiscriminate oxidizing species. The redox reactions at the surface of the particles lead to the generation of active oxygen species which can attack organic and inorganic species at or near the surface. However, the photocatalytic degradation of ammonia will lead to generate significant number of intermediates species in the reaction mixture which can’t be removed efficiently in the reactor. They gradually accumulated and eventually deposited on the TiO2 film surfaces around reaction time of 2 h which deactivate the active sites and reduce the hydroxyl radicals and active oxygen species generation on thin film surface who play an important role in oxidizing.
The influence of TiO2 layer coated times on ammonia degradation rate
In order to obtain multilayer thin films with controllable thickness, 1-10 TiO2 layers were successively coated on glass beads using dip-coating steps and then thermally treated as aforementioned process.
The photocatalytic degradation of ammonia in water was conducted respectively with the glass beads coated with 1 to 10 layers of TiO2 thin film. The ammonia concentration in wastewater was about 700 mg / L, initial pH of wastewater sample was 7; the degradation was at room temperature for 120 min.
It is clearly seen from Fig. 6 that under the same experimental conditions, the degradation efficiency depends on the film thickness. It was found that the degradation efficiency increased with both TiO2 film layers and film thickness growth but not linearly dependent. It is most likely that more titanium dioxide participated in the photocatalytic reaction. However, the growth of the degradation efficiency over 6 times of coating seemed to slow down or even stayed on the same level while the first 6 coatings greatly improved the degradation efficiency.
Figure 6 implies that we cannot achieve greater degradation efficiency by unlimited increasing coating layers or film thickness. An optimum value of TiO2 film thickness is presented. Jin et al.  showed the transmission spectra of the TiO2 films with various coating times. They demonstrated that the transmittances of the TiO2 thin films coated 5 times and 15 times were higher than that of the thin film coated 30 times in the visible range of 400 nm ~ 800 nm. Moreover, photocatalytic chemical reactions occurred on the surface of TiO2 thin films. Meanwhile, titanium dioxide in the interior region of thin films was not easy to contact the reactant and reaction products were also not easy to enter the solution. Therefore, the photocatalytic activity of titanium dioxide in the interior region of thin films was lower than that of titanium dioxide in the surface of thin films.
The surface topography and the microstructure of the coatings were examined by scanning electron microscopy (SEM). The film thickness was measured by using SEM and weight methods.
Fig. 7a showed that the glass beads surfaces presented inhomogeneous surfaces after two times coating without full coverage of nanoparticles to provide more reaction sites on the surface over the photocatalytic reaction. Nanoparticles are in irregular shapes and majority of the nanoparticles formed into a monolayer with a few bare or multilayered patches. The film thickness is ~230 nm. Fig. 7b illustrated that the layer consisting of four films yielded totally smooth surface. The full coverage of nanoparticles was observed. When the coating became thicker, it became denser. The film thickness is ~410 nm. The surface particles are in good contact with low surface roughness. Fig. 7c revealed that the surface of the six films layer appeared to be wrinkled while the good crystallinity and homogeneous nanostructure surface area were obtained after the 6 times of coating. Both the boundaries between particles and the grain sizes were clearly observed. These surface particles were in close contact and they formed the blocks with higher density where the film thickness is ~570 nm. Fig. 7c suggested a link between the grain size growth and the increase in roughness. Fig. 7d provided the image of the surface of ten layers films under higher magnification. The surface particles were piled up in layers and the grains agglomeration formed clusters. The film thickness is ~900 nm.
The influence of aqueous initial pH on degradation efficiency
The pH condition has been reported to affect the photocatalytic reaction rate. Before turning on the UV lamp, the pH of the aqueous solution was adjusted to the desired value by dropwise addition, 0.5 mol HCl solution or NaOH solution (Chengdu Kelong Chemical Co., Ltd). The concentration of ammonia was about 700 mg/L. The suspension was placed in the dark, shielded with aluminum foil and stirred until the pH stable.
The degradation efficiencies of ammonia in aqueous wastewater during UV illumination for pH from 1 to 13 were investigated as presented in Fig. 8. It was observed that the initial pH values of water samples had significant impacts on the photocatalytic degradation rate. In the pH range 1 – 3.4, the photocatalytic degradation efficiency increased with time. pH =3.4 had the maximum degradation of ammonia, removal efficiencies of ammonia in wastewater were above 70 %. When pH values exceeded 3.4, the photocatalytic degradation rate decreased dramatically. It is concluded that the overall degradation rate under strong acidic or mildly acidic were higher than those under neutral or mildly alkaline or strongly alkaline conditions. The possible reason is that the surface charge of TiO2 particles was influenced by the pH value in the solution.  already reviewed that acid-base properties of the metal oxide surfaces could considerably affected their photocatalytic activity. The point of zero charge (pzc) of the TiO2 (Degussa P25) was at pH 6.8 . Thus, the TiO2 surface was positively charged in acidic media (pH < 6.8), whereas it is negatively charged under alkaline conditions (pH > 6.8). At the acidic condition, TiO2 surface was positively charged while the wastewater contained negatively charged groups to improve the absorbance and the photocatalytic effect. In contrast, Sun et al.  suggested that the ammonia removal rate was larger in the alkaline condition than that in acidic condition. He also pointed out that an optimum removal rate was obtained at pH 9.1 when the concentration of ammonia was 546 mg/L. Murgia et al.  carried out a kinetic study of photo oxidation over TiO2 of NH3/NH4 in the high concentration range of 26 – 214 mg/l. They found that the degradation decreased from 50 % to 17 % at catalyst concentration of 0,008 % and meanwhile the pH value increasing from 10.7 to 9.5.
The influence of repetitions times on degradation rate and restoration of photocatalytic properties of TiO2
The surface modified glass beads were immersed in the TiO2 sol for 10 min. Subsequently, they were dried at 100 °C in a vacuum dryer and calcinated at 550 °C for 2 h. The procedure was repeated until the glass beads were coated with 6 layers of TiO2. The degradation was conducted at room temperature and degradation time was 120 min with ammonia content about 700 mg / L in wastewater. The same glass beads were reused for 4 cycles to check the degradation efficiencies. Restoration of photocatalytic properties of TiO2 thin film was executed as following: after reusing for 4 cycles, TiO2 film coated samples were rinsed with distilled water and dried in the vacuum oven. After that, they were placed in a resistance furnace to calcinate at 550 °C for 2 h. Under the same experimental conditions, the degradation efficiencies were re-measured. The results are presented in Fig. 9.
From the above chart, we can conclude that repetitions result in degradation efficiency reduction to some extent. After 4 cycles of usages, the degradation efficiency of the sample reclined ~34 % comparing to the original value. This implies that the TiO2 films can be used continuously for water treatment. Fortunately, this study did not find any deposited thin films peeling off from the glass beads.
This can be explained from the perspective of TiO2 photocatalytic mechanism. Oxidation and decomposition occurred on the surface of TiO2. After photocatalytic reaction, more or less residuals remained on the surface, which weakened the participation of TiO2 in the next photocatalytic reactions and therefore resulted in a steady decline in the degradation efficiency. During the restoration process, calcination at 550 °C for 2 h physically removed residuals adsorbed on TiO2 surface. Restoration of photocatalytic properties of TiO2 thin film had significant effect and ~93 % efficiency was restored in the recycled sampled in contrast to a new sample as shown on Fig. 9.
This project was undertaken to embed the TiO2 thin film on the glass beads via a deep coating in a sol–gel system and the performance of our lab-scale photocatalytic reactor where the volume of reactant solution is <1 L was evaluated on the basis on change in concentration of high ammonia content with respect to time. Immobilization of TiO2 on rigid glass beads surface is due to the main advantage that the glass substrate is the transparent even after the immobilization. This allows the penetration of light resulting in improved photocatalysis. Several suggestions on the basis of the results from the trail tests can be summarized, which contributes to the continue research in the future.
At 550 °C, TiO2 sample presented the highest photocatalytic activity among all counterparts, where the mass fraction on the rutile phase accounted for around 18.87 % in the mixed crystals; the relationship between film thickness and degradation efficiency is worth noticing. The thicker films brought out higher degradation. However, degradation went slower when TiO2 film was over 6 times of coating; it was observed that the optimum removal rate is obtained at pH 3.4 when aqueous initial pH changed from 1 to 13; the outcomes supported the finding that better degradation rate was achieved for high concentration ammonia in strong acidic or mildly acidic wastewater rather than neutral, mildly alkaline and strongly alkaline conditions; the influence of exposure time of the TiO2 films under UV irradiation was examined. Within time range of 30 ~ 120 min of UV exposure, the percentage of photo-degradation of the samples increases with irradiation time growth. After 120 min, as time proceeds, the growth of the degradation rate of all samples substantially slowed down and almost unchanged; one of the more significant findings to emerge from this study is that repetitions led to degradation efficiency reduction to some level; after 4-cycles re-utilization, the degradation efficiency reduced ~34 % compared to the original value. However, ~93 % efficiency was restored in the recycled sampled in contrast to a new sample and the TiO2 thin film could recover its photocatalytic properties effectively.
The photocatalytic degradation treatment in this study has proved to be very effective to removal of high concentration ammonia from its solutions. In addition, this photocatalytic reactor seems to be a simple, energy efficient, eco-friendly, less expensive to build and operate photocatalytic reactor to handle ammonium-rich wastewater. This study provides good reference data for the design of pilot plant-scale reactors for treating ammonium-rich wastewater.