Ultraviolet–Visible Light–Sensitive High Surface Area Phosphorous-Fluorine–Co-Doped TiO 2 Nanoparticles for the Degradation of Atrazine in Water

Preparation of reference TiO₂, P-TiO₂, F-TiO₂, and PF-TiO₂ nanoparticles

The reference TiO₂ (undoped TiO₂), P-TiO₂, F-TiO₂, and PF-TiO₂ nanoparticles were prepared by a modified sol-gel method. Titanium (IV) isopropoxide (TTIP, 97%; Sigma-Aldrich), H₃PO₄ (99.99%; Sigma-Aldrich), and ammonium fluoride (NH₄F, 96%; Fisher Scientific) were used as Ti, P, and F precursors, respectively. The nonionic surfactant polyoxyethylene (80) sorbitan monooleate (Tween 80; Sigma-Aldrich) was used as a pore directing agent in the sol-gel synthesis. Initially, Tween 80 was taken in a beaker and then isopropyl alcohol (i-PrOH, 99.8%; Pharmco) was added dropwise under vigorous stirring until Tween 80 was completely dissolved. After the addition of i-PrOH, acetic acid (AcOH; Fisher) was added immediately to the solution to maintain a low pH and for in situ formation of water through its esterification reaction with i-PrOH. Before the addition of P and F precursors, TTIP was introduced dropwise into the solution. The solution was kept under vigorous stirring for about 30 min. Finally, H₃PO₄, NH₄F, or both were added to the resulting solution for the synthesis of P-TiO₂, F-TiO_2, and PF-TiO₂ nanoparticles, respectively. In case of using both, that is, H₃PO₄ and NH₄F, NH₄F was added after 8 h of stirring of H₃PO₄ solution to obtain a homogenous solution. The resulting solutions were stirred for 24 h at room temperature. The transparent, homogenous, and stable solutions were obtained after 24 h of aging. The molar ratio of ingredients was Tween 80:TTIP:i-PrOH:AcOH:H₃PO₄:NH₄F=3.6:10:137:25:1:1 for PF-TiO₂. P-TiO₂ and F-TiO₂ were synthesized without the addition of NH₄F or H₃PO₄, respectively, keeping the same molar ratio of the rest of the ingredients. Reference TiO₂ was synthesized by the above mentioned procedure without the addition of P and F precursors. After 24 h stirring, the sol was dried in borosilicate glass dishes at room temperature for 12 h and then calcined in a multi-segment programmable high temperature furnace (Paragon Model HT-22-D, Thermcraft Inc.) to prepare the catalysts. The sols were calcined at 400°C for 1 h, the temperature was initially increased from room temperature to 400°C with a ramp rate of 60°C/h, and then up at 580°C for 2 h (at a ramp rate of 80°C/h). The samples were then cooled down naturally. The anatase crystal phase of the nanoparticles was maintained at this temperature (Baiju et al., 2005; Lv et al., 2009). During the calcination, all the organic compounds were completely removed from the resultant nanoparticles. The calcined samples were scratched from the dishes and then ground into fine powder form.

Characterization of reference TiO₂, P-TiO₂, F-TiO, and PF-TiO₂ nanoparticles

Structural properties of the reference TiO₂, P-TiO₂, F-TiO₂, and PF-TiO₂ nanoparticles were determined using X-ray powder diffraction (X'Pert PRO; Philips) with Cu K (λ=0.15406 nm) radiation together with Raman spectroscopy. Raman spectra were measured in backscattering configuration on a Renishaw inVia Reflex microscope using a high power near infrared diode laser (λ=785 nm) and an Ar⁺ ion laser (λ=514.5 nm) as excitation sources. The laser beam was focused onto the samples by means of a 5× objective (low magnification, to avoid possible inhomogeneity problems and have more representative data), while the laser power density was kept at low levels (<0.1 mW/μm²) to avoid any local heating and phase transformation of the titania samples. A Tristar 3000 (Micromeritics) porosimeter analyzer was used for the determination of Brunauer–Emmett–Teller (BET) surface area, Barrett–Joyner–Halenda (BJH) pore size, pore volume, pore size distribution, and porosity. The powders were purged with nitrogen gas for 2 h at 150°C using Flow prep 060 (Micromeritics) prior to the BET analysis by a Tristar porosimeter analyzer for degas. The light absorption properties and optical band gap of the synthesized TiO₂ nanopowders were investigated from UV-vis absorption spectra obtained with UV-vis spectrophotometer (2501 PC, Shimadzu) equipped with an integrated sphere accessory (ISR 1200) employing barium sulfate (BaSO₄) as a reference standard material. The information about crystal size and crystal structure was obtained with A JEM-2010F (Jeol) HR-TEM with field emission gun at 200 kV. The samples were prepared by dispersing the nanoparticles in methanol (high-performance liquid chromatograph [HPLC] grade; Pharmco) by ultrasonication (2510R-DH; Bransonic) over 30 min. A few drops of the solution were then fixed on a carbon-coated copper grid (LC325-Cu, EMS) and dried with an infrared lamp. XPS (SSX-100 XPS) was employed to reveal the chemical composition and valence state of the dopants in the nanoparticle powders. The instrument was fitted with monochromatic Al Kα radiation (E=1486.6 eV) and 600 μm nominal X-ray beam diameter. The absolute binding energies were calibrated using C1s level peak at 284.6 eV. To determine the presence of P and F in the synthesized photocatalysts, FTIR was performed using Schimadzue, IR Prestige-21, FTIR-8400S spectrometer.

Evaluation of photocatalytic activity

The photocatalytic activity of the as-prepared nanoparticles was evaluated in terms of atrazine (99.9%; Sigma Aldrich) degradation under UV-visible light illumination. The photocatalytic reactor consists of a borosilicate glass Petri dish (diameter 10 cm). Milli-Q grade water (resistivity of 18.2 MΩ·cm) was used for all experiments. Prior to the addition of atrazine, the pH of water was adjusted at 3.0 with 0.1 N HCl solutions. The low pH was selected due to the high oxidation properties of TiO₂ in the acidic region and high redox potential of ^•OH at acidic pH (i.e., 2.7 in acidic versus 1.8 in neutral solution) (Buxton et al., 1988; Chen et al., 2012). Then, the pH-adjusted aqueous solutions were spiked with an aliquot of atrazine solution to achieve the initial concentration of 2.32 μM. To obtain uniform suspension of TiO₂ photocatalysts, the solution containing TiO₂ nanoparticles was sonicated for 12 h using an ultrasonicator (2510R-DH; Bransonic). The photocatalyst suspension was then added to the reactor containing atrazine solution at 0.5 g/L as initial catalyst loading and 10 mL as final reaction solution volume. The reactor was completely sealed with parafilm and a fan was located behind the reactor to avoid any loss of solution through evaporation. A magnetic stirrer was used for continuously mixing the suspension in the reactor during the reaction under UV-visible light illumination. Two 15 W fluorescent lamps (Cole-Parmer) were employed as UV-visible light source (315–700 nm; Supplementary Fig. S1). The UV-visible light intensity was determined with a radiant power meter (Newport Corp.) and was found as 9.50×10⁻² mW/cm². Sampling was performed at specific time intervals, that is, at 0, 0.5, 1, 2, 3, 4, and 6 h by taking 150 μL of sample solution and quenched with 150 μL of methanol to avoid any reaction. The collected samples were filtered using syringeless filters (pore size: 0.2 μm; Whatman) to separate out the catalyst particles. The concentration of atrazine in the sample was analyzed using a HPLC (Agilent Series 1100) with a similar method reported in our previous publication (Khan et al., 2013). Moreover, reaction intermediates were identified using a combination of liquid chomatography/quadrupole time-of-flight (an Agilent G6540A quadrupole time-of-flight mass spectrometer) with a same method reported by Andersen et al. (2013). An Agilent Eclipse XDB-C18 column (2.1 mm×50 mm, 3.5 μm) was employed and the flow rate was 0.2 mL/min.

Characterization of as-prepared nanoparticles

UV–vis absorption spectra of the nanoparticles

The UV-vis absorption spectra of the P-TiO₂, F-TiO₂, and PF-TiO₂ nanoparticles in comparison with the reference TiO₂ nanoparticles are shown in Fig. 1a. To estimate the effective band gap energy for the synthesized samples, (αhν)^1/2 was plotted as a function of the photon energy (hν), assuming indirect allowed transitions (Gonzalez and Santiago, 2007). The values of the indirect band gap for PF-TiO₂, P-TiO₂, F-TiO₂, and reference TiO₂ were determined by the linear extrapolation of the corresponding curves as shown in Fig. 1b. The value was 2.70 eV for PF-TiO₂, 2.85 eV for P-TiO₂, and F-TiO₂, and 3.08 eV for reference TiO₂. These results suggest that the strongest absorbance in the visible light region is exhibited by PF-TiO₂ and hence, such material is anticipated to have the highest photocatalytic activity under visible light irradiation. In addition, a red absorption shift was also observed for PF-TiO₂ (Fig. 1a), which further strengthens our hypothesis that co-doping of TiO₂ with P and F is highly effective in extending the optical response of TiO₂ to the visible light region.

OPEN IN VIEWER

FIG. 1.

Optical properties of the different as-prepared TiO₂ nanoparticles: (a) UV–vis absorption spectra of the PF-TiO₂, P-TiO₂, F-TiO₂, and reference TiO₂ samples; (b) calculation of band gap energy for indirect allowed transitions. UV, ultraviolet; vis, visible light.

Besides, it has been reported that doping of F into TiO₂ crystal lattice contributed to the creation of two types of surface oxygen vacancies, two trapped electrons (F center) and one trapped electron (F⁺ center) oxygen vacancies (Li et al., 2005), both contributing to the photocatalytic activity of F-TiO₂ under visible light (Li et al., 2005). Similarly, Yang et al. (2007) reported that the enhanced photocatalytic activity of P-TiO₂ could be attributed to the large surface area and high degree of crystallinity of P-TiO₂. These observations suggest that P-TiO₂, F-TiO₂, and PF-TiO₂ nanoparticles would be ideal photocatalysts under UV-visible/visible light illumination.

Structural analysis of reference TiO₂, P-TiO₂, F-TiO₂, and PF-TiO₂ nanoparticles

The XRD patterns of the as-prepared nanoparticles are shown in Fig. 2. It can be seen from the characteristic diffraction peak at about 2θ=25.5° corresponding to the (101) anatase crystal plane, that all the nanoparticles exhibit well-crystallized anatase structures with no observation of peaks corresponding to the rutile phase. These observations indicated that the main phase composition of the TiO₂ particles has not been drastically influenced by the process of doping and co-doping, within the resolution of powder X-ray diffraction. It can be seen from Table 1 that no significant variations in the (101) plane spaces [D (101)] were found between the modified and reference TiO₂ nanoparticles. These results suggest that the average unit cell dimensions of the anatase crystal lattices were not significantly affected by the incorporation of P and/or F. This may possibly be due to the low doping percentages of the P and F in the TiO₂ crystal lattice. The most intense diffraction peak of anatase, that is, (101) plane was used to calculate the average crystallite size using the Scherrer formula (Zhang et al., 2010a): OPEN IN VIEWER

FIG. 2.

XRD pattern of PF-TiO₂, P-TiO₂, F-TiO₂, and reference TiO₂ samples. XRD, X-ray diffraction analysis.

\documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}D = \frac { K \lambda } { \beta \cos \theta } \tag { 1 } \end{align*} \end{document}

OPEN IN VIEWER

Table 1.

Structural Characteristics of PF-TiO₂, P-TiO₂, F-TiO₂, and Reference TiO₂ Nanoparticles

Catalyst type	Crystal phase	S BET (m2/g)	Crystal size (nm) a	Particle size (nm) b	Total pore volume (cm3/g)	Porosity (%) c	D (101) (Å) d	k obs of atrazine (min−1)
PF-TiO2	Anatase	212.0	5.9	7.4	0.152	36.5	3.50	4.95×10−3
P-TiO2	Anatase	175.0	9.9	9.1	0.133	33.5	3.52	3.82×10−3
F-TiO2	Anatase	88.8	14.6	17.8	0.095	26.5	3.53	1.82×10−3
Reference TiO2	Anatase	79.7	16.4	19.9	0.087	24.8	3.53	1.13×10−3

a

Calculated from XRD, using Scherrer equation: D=0.89 λ/(β×cosθ).

b

Calculated from BET, using D=6000/(ρ×BET surface area), where ρ=3.79 g/cm³ of anatase density.

c

Porosity (%)=pore volume (cm³/g)/(pore volume [cm³/g]+solid catalyst volume without pore [cm³/g])×100; the solid catalyst volume (cm³/g) without pore was calculated from the density of anatase (3.79 g/cm³) and is equal to 1/density of the solid catalyst volume without pore.

d

Based on Bragg's Law: d=λ/(2×sin θ), where λ=1.54 Å.

BET, Brunauer-Emmett-Teller; XRD, X-ray diffraction analysis.

where D is the average crystallite size, K is a constant equal to 0.89, λ is the wavelength of the X-ray radiation (0.154 nm), β is the full width at half-maximum (FWHM) of 101 peak, and θ is the diffraction angle. The average crystallite sizes were calculated to be 5.9, 9.9, 14.6, and 16.4 nm for PF-TiO₂, P-TiO₂, F-TiO_2, and reference TiO₂ nanoparticles, respectively (Table 1).

The structural properties and phase composition of the modified TiO₂ nanomaterials were further investigated by Raman spectroscopy. Figure 3 compares the Raman spectra of the reference TiO₂, P-TiO₂, F-TiO₂, and PF-TiO₂ nanoparticles at 785 nm laser excitations (identical results were obtained at 514.5 nm). All samples exhibit the characteristic Raman-active modes of the anatase TiO₂ phase (Yang et al., 2007), consistently with the materials phase composition determined by XRD. Moreover, analysis of the Raman spectra revealed significant intensity reduction together with distinct shifts and broadening of the anatase Raman bands, especially for the P-TiO₂ and PF-TiO₂ nanoparticles compared with F-TiO₂ and reference TiO₂. This variation is characteristic of size effects due to optical phonon confinement, resulting in both, broadening and blue shift of the Raman modes (Kelly et al., 1997) in addition to the reduction of the Raman scattering efficiency in nanosized materials. Specifically, the peak position and FWHM (in parentheses) of the lowest frequency and most intense (E_g) anatase mode, frequently used to quantify phonon confinement size effects in anatase (Han et al., 2011; Liu et al., 2013), were determined to be at 143.4 (9.5), 143.8 (10), 147 (17), and 147.5 (25) cm⁻¹ for the reference TiO₂, F-TiO₂, P-TiO₂, and PF-TiO₂, respectively. Projecting the E_g peak positions to the correlation plots of the anatase E_g frequency versus crystallite size predicted by the phonon confinement model (Pottier et al., 2003), indicates that the anatase crystallite size decreases to 17, 15, 8.0, and 6.5 nm for the reference TiO₂, F-TiO₂, P-TiO₂, and PF-TiO₂, respectively, in close agreement with the XRD results (Table 1). While the same model could also justify the FWHM values of the anatase E_g mode for the reference TiO₂ and F-TiO₂ samples, excess broadening was noted for P-TiO₂ and particularly for PF-TiO₂, deviating from the predicted scaling of the Raman shift versus FWHM (Kelly et al., 1997). This implies the presence of additional effects related to lattice strain and structural defects that could shift and/or broaden the anatase modes for P-TiO₂ and PF-TiO₂.

OPEN IN VIEWER

FIG. 3.

Raman spectra for the reference, F-TiO₂, P-TiO₂, and PF-TiO₂ at 785 nm. The inset compares in detail the variation of the lowest frequency E_g anatase mode for the different TiO₂ nanoparticles.

Besides, the intense anatase Raman modes, 2 weak bands at 450 and 610 cm⁻¹ could be identified for the reference TiO₂ (Fig. 3), characteristic of small amounts of the rutile phase that is beyond the detection limit of powder XRD. The formation of the rutile phase that can be justified by the relatively high calcination temperature (580°C) of the bare TiO₂ nanoparticles was, however, inhibited by the presence of H₃PO₄ and NH₄F reagents for the doped-TiO₂ nanomaterials. On the other hand, two additional Raman bands emerged at 246 and ∼360 cm⁻¹ for P-TiO₂ and PF-TiO₂, indicating the formation of a small fraction of brookite TiO₂ nanocrystals (Han et al., 2011), similar to other anion doped TiO₂ nanomaterials prepared by the surfactant assisted sol-gel method (Likodimos et al., 2013).

BET surface area analysis

The N₂ adsorption-desorption isotherms and BJH pore size distributions of the reference TiO₂, P-TiO₂, F-TiO₂, and PF-TiO₂ nanoparticles are shown in Fig. 4. The Halsey equation was used for calculating the pore size distribution from the N₂ desorption isotherm by applying the BJH method. According to IUPAC classification, the N₂ adsorption-desorption isotherms of Fig. 4a can be classified as type IV on the basis of hysteresis loops (Sing et al., 1985). This type of isotherm is characteristic of mesoporous materials showing that the as-prepared catalysts are mesoporous in nature (Sing et al., 1985). Figure 4b reveals that the pore size distributions significantly varied with doping of the TiO₂ crystal lattice with F and especially P. The pore diameter for the reference TiO₂, F-TiO₂, P-TiO₂, and PF-TiO₂ were determined to be about 4.0, 3.8, 2.5, and 2.5 nm, respectively, while the width of the derived pore size distributions were narrow, especially for the P-TiO₂ and PF-TiO₂ particles, indicating good homogeneity of the pores. The results obtained by analyzing the prepared nanoparticles for BET surface area are presented in Table 1. It can be seen that simultaneous doping of P and F into TiO₂ noticeably increased the surface area, that is, the surface for area PF-TiO₂ is 212.0 m²/g, which is about 2.7 times larger than that of reference TiO₂ (79.7 m²/g). However, the results showed that P doping resulted in modified TiO₂ catalysts with much higher surface area than those obtained with F doping, with 175.0 for P-TiO₂ and 88.8 m²/g for F-TiO₂. The pore volume followed similar trend to that of surface area, that is, the pore volume decreases for photocatalysts in the following order: PF-TiO₂>P-TiO₂>F-TiO₂>reference TiO₂.

OPEN IN VIEWER

FIG. 4.

(a) N₂ adsorption–desorption isotherms of various as-prepared TiO₂ nanoparticles. (b) Pore size distribution of different as-prepared TiO₂ nanoparticles.

The average particle size was further estimated from the BET surface area by using the following equation (Pelaez et al., 2009): \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}D = \frac { 6000 } { ( \rho \times { \rm BETsurface } \ { \rm area } ) } \tag { 2 } \end{align*} \end{document}

where D is average particle size and ρ is density of the catalyst (3.79 g/cm³ for anatase form) (Prasai et al., 2012). Based on the above equation, the average particle sizes for PF-TiO₂, P-TiO₂, F-TiO₂, and reference TiO₂ were calculated as 7.4, 9.1, 17.8, and 19.9 nm, respectively, complying with the results of XRD and Raman analyses and the inverse relationship between particle size and surface area (Pelaez et al., 2009).

HR-TEM studies

To further investigate the crystal size and morphology of the reference TiO₂, P-TiO₂, F-TiO₂, and PF-TiO₂ nanoparticles, HR-TEM and fast Fourier transform (FFT) were performed and the results are presented in Fig. 5a–d. The average crystal size evidenced from HR-TEM images were 5.5±1.0, 6.1±0.9, 8.8±0.9, and 11.2±1.6 nm for PF-TiO₂, P-TiO₂, F-TiO₂, and reference TiO₂, respectively. The results of the crystallite sizes for the samples were consistent with the nanoparticle size evolution derived by XRD, Raman, and BET analysis. A clear ring can be revealed from FFT patterns, showing that all TiO₂ catalysts have one dominant crystal phase. In addition, clear lattice fringes with an interlayer distance in the range of 3.50–3.53 Å for the (101) planes of anatase were invariably resolved in HR-TEM images for all prepared nanoparticles, in close agreement to the 3.53 Å lattice spacing of the reference anatase (101) TiO₂ planes, as shown in Table 1. Energy-dispersive X-ray spectroscopy (EDX) analysis was also performed for the determination of elemental composition of the as-prepared nanoparticles. The EDX results are shown in Supplementary Figs. S2–S5. It can easily be seen from these figures that phosphorous is efficiently doped into TiO₂ crystals in P-TiO₂ and PF-TiO₂ (peak at 2.0 keV). However, due to the overlapping of the fluorine peak with the oxygen one, fluorine could not be identified in the EDX spectrums of F-TiO₂ and PF-TiO₂ nanoparticles.

OPEN IN VIEWER

FIG. 5.

High-resolution transmission electron microscopy images of TiO₂ nanoparticles: (a) PF-TiO₂, (b) P-TiO₂, (c) F-TiO₂, and (d) reference TiO₂ samples.

XPS analysis

XPS analysis was carried out for the identification of incorporated P and F elements in the doped TiO₂ crystal lattices. The XPS survey scan spectrum of PF-TiO₂ is shown in Fig. 6a. High-resolution XPS spectra of P 2p and F 1s are displayed in Fig. 6b and c, respectively. Figure 6a shows that the elements detected in PF-TiO₂ nanoparticles are P, F, Ti, O, and C. The carbon peak at 284.5 eV for C1s core level is expected to be an impurity peak that could be assigned to the adsorption of adventitious carbon in TiO₂ (Pelaez et al., 2009). The peak at 134 eV representing P 2p is attributed to the pentavalent oxidation state (P⁵⁺) of phosphorous ions. The absence of a characteristic peak at 129 eV in the XPS spectra of P-TiO₂ indicated that Ti atoms are unlikely to bond P atoms for the formation of Ti-P bonds (Lv et al., 2009). It has been reported that the presence of P⁵⁺ ions suggest that Ti⁴⁺ are partially replaced by P⁵⁺ ions in the crystal lattice of TiO₂. These concomitant charge imbalances could decrease the recombination rate of electron-hole pairs (Lv et al., 2009) and hence lead to enhanced photocatalytic activities. Furthermore, the XPS spectrum showed a characteristic peak for Ti 2p3/2 at 458.5 eV, indicating the presence of Ti⁴⁺ ions in the TiO₂ crystal lattice (Yang et al., 2006; Xu et al., 2008b). The peak at 529.5 eV in the survey spectra of the nanoparticles has been assigned to the O 1s core level in Ti-O bonds, that is, O₂⁻ in TiO₂ crystal lattice (Han et al., 2013). The peak located at 688.2 eV is assigned to F 1s and was attributed for substitutional F-atoms in TiO₂ crystal lattice (Pelaez et al., 2009). It has been reported that the recombination rate of photo-generated electrons and holes could be reduced by F substitution into TiO₂ crystal lattice. As a result, oxygen vacancies are created in TiO₂ lattice that can act as active sites for oxidizing species formation and subsequently leading to enhanced photocatalytic activity (Pelaez et al., 2009).

OPEN IN VIEWER

FIG. 6.

(a) XPS survey scan spectrum for PF-TiO₂, (b) high resolution XPS of P 2p, and (c) F 1s core levels and reference TiO₂. XPS, X-ray photoelectron spectroscopy.

FTIR analysis

The FTIR results of doped and reference TiO₂ nanomaterials are presented in Fig. 7. The hydroxylation state of TiO₂ nanoparticles is strongly dependent on the conditions and methods of their preparation. It has been observed that the presence of free hydroxyl groups on the surface of TiO₂ plays an important role in the catalyst activity due to their ability of transferring the photogenerated charge species during photocatalytic activity (Xu et al., 2008a). The broad peak at 3420 cm⁻¹ is attributed to the stretching vibration of hydroxyl groups adsorbed on the surface of titania, whereas the peak at 1640 cm⁻¹ is considered as a characteristic peak of bending vibration of surface-adsorbed free water molecules (Xu et al., 2008a). The stronger vibrations were detected in P-TiO₂ and PF-TiO₂ nanoparticles compared with reference TiO₂ and F-TiO₂ nanoparticles suggesting larger adsorption of water and hydroxyl groups on the surface of P-TiO₂ and PF-TiO₂ catalysts. The higher adsorption of water and hydroxyl groups positively affects the photocatalytic activities of these nanoparticles (Lv et al., 2009). Additionally, an absorption peak at 1045 cm⁻¹ was observed in P-TiO₂ and PF-TiO₂ nanoparticles but not in reference TiO₂ and F-TiO₂ samples. This peak is actually a characteristic frequency of Ti–O–P vibrations suggesting the successful doping of P atom into TiO₂ lattice, in good agreement with the EDX results (Supplementary Data) (Zheng et al., 2008; Lv et al., 2009). An absorption peak at 839 cm⁻¹ can be seen in the FTIR spectra of F-TiO₂ and PF-TiO₂, which is ascribed to the vibration of Ti-F bond (Vijayabalan et al., 2009).

OPEN IN VIEWER

FIG. 7.

Fourier transform infrared (FTIR) spectra of PF-TiO₂, P-TiO₂, F-TiO₂, and reference TiO₂ samples.

Photocatalytic activity

Photocatalytic activity of the prepared doped and reference TiO₂ samples was tested for the degradation of atrazine in aqueous solutions, under UV-visible light illumination. Results obtained from photocatalytic degradation study are presented in Fig. 8a and b. Figure 8a showed that no significant photolytic degradation of atrazine under UV-visible light was observed after 6 h irradiation, implying its persistent nature to UV-visible light under the mentioned experimental conditions. In addition, the adsorption of atrazine on PF-TiO₂ (Fig. 8) and also on reference TiO₂, F-TiO₂, and P-TiO₂ (data not shown) is negligible. Under light irradiation, F-TiO₂, P-TiO₂, and PF-TiO₂ showed high pollutant removal efficiency compared with reference TiO₂. The observed degradation rate constant (k_obs) for atrazine was calculated as 1.1×10⁻³, 1.8×10⁻³, 3.8×10⁻³, and 5.0×10⁻³ min⁻¹ for reference TiO₂, F-TiO₂, P-TiO₂, and PF-TiO₂, respectively. These results confirmed that high pollutant removal rate can be obtained when both P and F were doped into TiO₂ crystal lattice confirming the synergistic effect of these dopants. The higher k_obs for PF-TiO₂ can be attributed to the higher amounts of surface-adsorbed free water and hydroxyl groups in addition to the relatively small crystal size of the PF-TiO₂ nanoparticles. It has been found that the photoinduced holes can attack surface-adsorbed water molecules and hydroxyl groups, which results in the formation of hydroxyl radicals (Ding et al., 2000; Yu et al., 2010; Elghniji et al., 2012). The photodegradation mechanism of atrazine is supposed to take place in accordance with a well-established mechanism. The photocatalytic activation of semiconductor catalysts is due to the formation of positive holes in the valence band along with the corresponding excited electrons in the conduction band as can be shown by the following reaction (Yao et al., 2010): \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}{ \rm TiO}_2 + { \rm hv} \rightarrow { \rm e}_{{ \rm cb}}^ - + { \rm h}_{{ \rm vb}}^ + \tag{3}\end{align*} \end{document}

OPEN IN VIEWER

FIG. 8.

(a) Degradation of atrazine under UV-visible light irradiation and (b) degradation rate of atrazine at 6 h. Experimental conditions: [ATZ]₀=2.32 μM (0.5 mg/L), [catalyst]₀=0.5 g/L, pH=3.0.

where e⁻_cb and h⁺_vb are the electrons in the conduction band and positive holes in the valence band, respectively.

Both these photoinduced species have the potential to initiate several redox reactions with other entities present on the surface of photocatalyst. The various chemical reactions in UV-visible light illuminated atrazine (ATZ) solutions containing photocatalyst are summarized below (Xiao et al., 2008; Vijayabalan et al., 2009; Yu et al., 2010): \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}{ \rm H}_2{ \rm O}_{{ \rm ads}} + { \rm h}_{{ \rm vb}}^ + \rightarrow \ ^ \bullet { \rm OH} + { \rm H} ^ + \tag{4}\end{align*} \end{document} \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}{ \rm O}_2 + { \rm e}_{{ \rm cb}}^ - \rightarrow { \rm O}_2^{ - \bullet} \tag{5}\end{align*} \end{document} \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}{ \rm O}_2^{ - \bullet} + 2{ \rm H}^ + + { \rm e}_{{ \rm cb}}^ - \rightarrow { \rm H}_2 { \rm O}_2 \tag{6}\end{align*} \end{document} \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}{ \rm H}_2{ \rm O}_2 + { \rm O}_2^{ - \bullet} \rightarrow \ ^ \bullet { \rm OH} +^ - { \rm OH} + { \rm O}_2 \tag{7}\end{align*} \end{document} \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}{ \rm ATZ} + ^ \bullet { \rm OH} ( { \rm or \ e}^ - _{{ \rm cb}} ) \rightarrow { \rm Intermediates} \tag{8}\end{align*} \end{document} \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} \begin{split}{ \rm Intermediates} + ^ \bullet { \rm OH} ( { \rm or \ e}_{{ \rm cb}}^ - ) \rightarrow { \rm CO}_2 + { \rm H}_2 { \rm O} \\ \quad + { \rm NO}_3^ - + { \rm Cl}^ -\end{split} \tag{9}\end{align*} \end{document} \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}{ \rm h}_{{ \rm vb}}^ + + { \rm e}_{{ \rm cb}}^ - \rightarrow { \rm TiO}_2 \tag{10}\end{align*} \end{document}

The electron-hole production [Eq. (3)] is a reversible process and the photoinduced species may recombine following Equation (10). However, Equation (4) can slow down the recombination Equation (10). Therefore, surface-adsorbed water molecules and hydroxyl groups are considered to play a key role in the photocatalytic degradation of water pollutants by promoting the formation of hydroxyl radicals through Equations (4) and (7). The high redox potential of hydroxyl radicals (2.8 V) is responsible for their high oxidation capability toward organic pollutants, in this case atrazine (Mehrdad et al., 2011). The small crystallite size, greater surface area, more surface adsorbed free water and hydroxyl groups, and large pore volume of P-TiO₂ and PF-TiO₂ nanoparticles speed up the transfer of surface charge carriers, decreasing recombination of electron-hole pairs, and consequently enhancing the photocatalytic activity compared with reference TiO₂ and F-TiO₂ nanoparticles (Hoffmann et al., 1995; Elghniji et al., 2012).

In addition to the investigation of photocatalytic activity of the synthesized catalyst, the reaction intermediates were determined after 10 h of atrazine degradation using PF-TiO₂ under UV-visible light illumination. During the photocatalysis of atrazine, one new intermediate (m/z=200.0705, C₇H₁₀ClN₅, degree of unsaturation [DU]=5; see the tandem mass spectrometry spectra shown in Supplementary Fig. S6) and 10 intermediates reported by Andersen et al. (2013), Chan and Chu (2006), and Chen et al. (2009) were observed. We also identified the intermediate (m/z=214.0856), which was recently found by Andersen et al. (2013). However, any dechlorination and deamination products were not observed during PF-TiO₂ photocatalysis and complete mineralization of atrazine did not occur. The end product of atrazine by an attack of hydroxyl radical in TiO₂ photocatalysis was cyanuric acid [C₃N₃(OH)₃] in previous reports (Pelizzetti et al., 1990; Parra et al., 2004). Table 2 summarizes the observed reaction intermediates and their m/z values. The possible decomposition pathway is provided in Fig. 9. Other possible structures of m/z=232.0960 and 248.0909 are provided in Supplementary Figs. S7 and S8, respectively.

OPEN IN VIEWER

FIG. 9.

Possible decomposition pathway of atrazine by UV-visible light-induced PF-TiO₂ photocatalysis. *Andersen et al. (2013), ^#Chan and Chu (2006), or ^$Chen et al. (2009) also reported the same reaction intermediates of atrazine which were observed in this study.