Preparation and characterization of nanostructured S and Fe co-doped TiO2 thin film by ultrasonic-assisted spray pyrolysis method

INTRODUCTION

With advantage of having long-term stability, cost-effectiveness, strong catalytic activity, non-toxicity and biocompatibility, TiO₂ has received considerable attention in recent decades [1-3]. This semiconductor is used in various applications like paints, cosmetics, antibacterial and self-cleaning surfaces, water purification, cancer treatment [4]. Recently, most of the researches are focused on improving the optical properties of TiO₂ for renewable energy applications [3]. With a band gap (Eg) ranging from 3.0 to 3.2 eV (i.e., from 413 to 387 nm), TiO₂ has low photosensitivity to visible radiation. Therefore, for efficiently harvesting solar energy by TiO₂, its band gap must be reduced which can be achieved by tailoring its morphology and crystal phase, and by introducing of doping elements [5]. The electron-hole recombination rate is another important factor which has considerable effect on the visible light photoactivity of the TiO₂ and any factor that reduce the electron–hole recombination rate will enhance the photocatalytic activity [6]. To reduce the electron–hole recombination rate and the wide band-gap of TiO₂, a variety of methods such as, incorporation of quantum dots (QDs) [7-9], sensitizing with dyes [10], doping and codoping with metal and non-metal elements [11, 12], and decorating with graphene nanosheets [13], have been extensively studied in recent years. Among them, nonmetal and metal doping has been proved to be an efficient approach [14].

Sulfur, unlike other non-metal dopants, can present in more than one oxidation state such as S^2-, S⁴⁺ or S⁶⁺ in TiO₂, depending on the synthesis conditions and the precursor type [15]. Ti could be replaced with cationic forms of S [16], and O is substituted with anionic S [17]. In sulfur doping, the S atoms can occupy the substitutional and interstitial sites in the TiO₂ lattice and the mixing of S 3p states with valance band of TiO₂ can cause the band gap narrowing [18]. Furthermore, the doping of S element reduces the charge-carrier recombination rate, and increases the number of photo-generated electrons and holes on TiO₂ surface [19]. Another advantage of S doping in TiO₂ is the improvement in thermal stability of anatase phase up to 900 °C [7].

Among various transition metals for TiO₂ doping, Fe seems to be the most appropriate one owing to its unique half-filled electronic configuration, which can reduce the band gap energy through the formation of new intermediate energy levels [20]. Furthermore, ionic radius of Fe³⁺ (0.64 Å) ion is nearly same as that of Ti⁴⁺ (0.68 Å) and this ion can easily substitute for Ti⁴⁺ in the crystal structure of TiO₂ [21]. Moreover, Fe³⁺ ions in the TiO₂ lattice can act as traps to capture the photogenerated electrons which result in the reduction of the recombination rate of electron–hole pairs [22].

To avoid the expensive and complex separation of the photocatalyst after degradation proses, TiO₂ in the form of thin film on a substrate are preferred to the TiO₂ nanoparticle suspensions. Semiconductor-based photoactive films are deposited on various substrates of metals, ceramics, silicon wafer, and glass in widespread applications such as solar cells, self-cleaning surfaces, hydrogen production, photocatalysis and air purification [23]. Among these substrates, glass substrate is the most appropriate one because of its excellent chemical stability, high transmittance, and low cost. Transparency is an important requirement in consideration of the potential photocatalytic application of coatings on the buildings and vehicles glass, therefore in recent years, there is considerable interest on the preparation of transparent photoactive films on glass substrate [24, 25].

Over the past decade, many studies have been conducted on the simultaneous doping of TiO₂ with metal and non-metal elements. In this regard, in the present work, S and Fe co-doped TiO₂ thin film with high monodispersity and transparency were deposited on glass substrate through ultrasonic-assisted spray pyrolysis technique, for the first time.

MATERIAL AND METHODS

Materials

The following chemicals were used in this work without further purification: titanium(IV) butoxide (≥97.0% from Sigma-Aldrich), Iron(III) chloride hexahydrate (FeCl₃·۶H₂O, Merck),thiourea (≥99.0% from Sigma-Aldrich), absolute ethanol (99.5% v/v, Merck) and concentrated hydrochloric acid (37 wt %, Merck).

Preparation of the thin films

A schematic diagram of the deposition set-up is shown in Fig. 1. TiO₂ sol solution contaning thiourea (as sulfur source) or Iron(III) chloride was prepared by addition of 40 ml the HCl ethanolic solution (0.01M) into the 30 ml of ethanolic solution of titanium (IV) butoxide (0.71M) in dropwise manner. Amount of thiourea and Iron(III) chloride in TiO₂ sol solution are 61.5 mg, and 171 mg respectively. The pH of the final solutions were adjusted to 2.1 by addition appropriate amount of the HCl ethanolic solution. After about 1 hours stirring, each of the thiourea and iron containing TiO₂ solution were poured into the separate flasks and were sprayed with aid of a floated piezoelectric ultrasonic transducer and were carried by N₂ gas (with flow rate of 300 ml/min). The migrated droplets were thermally decomposed at the surface of the hot glass substrate at 400°C, and the S and Fe co-doped TiO₂ thin film was formed after about 30 min. The presence of sedimentation flask in the deposition set-up is for the improvement of the monodispersity of the thin films by sedimentation of large droplets in its inside.

Characterization

The surface of the prepared thin films were examined using MIRA3 TESCAN field emission scanning electron microscope (FESEM) (Czech Republic). X-ray diffraction (XRD) patterns of the prepared thin films have been recorded by a STOE Stadi diffractometer equipped with Cu-Kα irradiation (λ=1.54018 Å). Photoluminescence (PL) emission spectra of the prepared samples were determined using Cary Eclipse fluorescence spectrophotometer (Varian, Inc., USA) at room temperature, with an excitation wavelength of 300 nm. XPS spectra of the PT0.8 sample were obtained by using a Gammadata-Scienta Esca 200 (Uppsala, Sweden) hemispherical analyzer equipped with an Al Kα (1486.6 eV) X-ray source. All binding energy values were calibrated by using the value of the C 1s peak at 284.6 eV as a reference. Ultraviolet–visible (UV–vis) absorption and transmission spectra of the samples were recorded using Cary 100 Bio spectrophotometer (Varian, Inc., USA). Raman spectra of the samples were obtained using T64000-HORIABA Jobin Yvon Raman spectrometer equipped with an argon ion laser (514 nm).

RESULTS AND DISCUSSION

XRD

The crystalline structure of the undoped and doped TiO₂ thin films were analyzed from the XRD patterns (Fig. 2). In XRD patterns of all samples, the observed diffraction peaks at 2θ of about 25.3°, 37.7°, 47.9°, 54.1°, 55.1°, and 62.8°, indexed as (101), (004), (200), (105), (211), and (204) lattice planes, respectively, assigned to the anatase phase of TiO₂ (JCPDS card No. 21-1272). Therfore, all of the prepared thin films have anatase crystalline phase and presence of doped elements don’t interfere in the crystal phase of TiO₂. TiO₂ sample shows sharper XRD peaks (indicating a higher crystallinity) than doped TiO₂ thin films. As shown in inset of Fig. 2, compared to the undoped TiO₂ thin film, there is a slight shift to higher diffraction angle and broadening in the (101) XRD peaks of the doped TiO₂ samples which largest shift is observed for the S and Fe co-doped TiO₂ sample. Presence of dopant elements in the lattice of TiO₂ and substitution of Ti or O atoms with these dopants may disrupt the crystalline structure of TiO₂ which results in the defect formation and low crystallization in the doped TiO₂ [26, 27].

Raman spectroscopy

Raman spectroscopy was used as an additional characterization tool to precisely confirm the presence of the trace amounts of the crystal phases of TiO₂ [28]. Room temperature Raman spectra of the prepared samples are shown in Fig. 3. The anatase phase of TiO₂ exhibits five Raman active modes at Eg (144, 197 and 639 cm⁻¹), B1g (395 cm⁻¹), and A1g (513 cm⁻¹) while the rutile phase consists of four Raman-active modes at 143, 233, 447, and 610 cm⁻¹ [7, 29]. In the Raman spectra of the pure and doped TiO₂ samples, all of the peaks are assigned to the anatase TiO₂, and there is no peak related to the rutile phase of TiO₂. Hence, the prepared thin films have anatase crystal phase, which are consistent with XRD results. Presence of the oxygen vacancies and formation of defects states can create shift in Raman peaks, as shown in the inset of Fig. 3, there are some shifts in Raman peaks of the S and Fe co-doped and S doped TiO₂ samples, indicating defect formation in these thin films [22, 30, 31].

SEM

Surface morphology of the synthesized thin films was characterized by FESEM and Fig. 4 shows FESEM images of the samples at different magnifications. As shown in this figure all of the prepared doped and undoped TiO₂ thin films have nanostructured features, however, compared to the pure TiO₂ thin film, the doped TiO₂ thin films have smaller particles in their structures. Therefore, presence of S and Fe dopants in the TiO₂ lattice hinders further growth of the film particles by disrupting the crystalline structure of TiO₂ as evidenced in XRD and Raman analyzes. Morphologies of the S and Fe co-doped and undoped TiO₂ thin film are irregular shaped nanoparticle, however, some particles with cubic morphology are clearly seen in the structure of S doped and Fe doped TiO₂ thin films. Because of the special design of the deposition set-up, all of the prepared thin films have high monodispersity.

XPS

Fig. 5 depict the S 2p and Fe 2p core-level XPS spectra of the prepared thin films. S 2p core-level XPS spectra of the S and Fe co-doped and S doped TiO₂ samples be deconvoluted into two Gaussian peaks located at the binding energies of about 167.4 and 168.3 eV. According to the literature, the XPS peaks at 167.4 eV depicts the presence of S⁴⁺ ion in these samples, and the XPS peaks at 168.3 eV the presence of S⁶⁺ ion in these thin films [7, 18]. Therefore, in these samples, sulfur was doped in the forms of S⁴⁺ and S⁶⁺ cations. In Fe 2p core-level XPS spectra of the S and Fe co-doped and Fe doped TiO₂ thin films, there are two peaks at about 712 and 725 eV which belong to the binding energies of Fe 2p_3/2 and Fe 2p_1/2, respectively. The difference of 13 eV between binding energies of Fe 2p_3/2 and Fe 2p_1/2 peaks demonstrates that the oxidation state of Fe in these thin films is +3 [32].

Photoluminescence (PL)

The photocatalytic activity of semiconductor photocatalyst greatly depends on the separation and transfer ability for photoinduced electron–hole pairs. For an irradiated semiconductor sample, the PL intensity is directly related to the electron–hole recombination rate and the high electron–hole recombination rate results in the high PL intensity, and the low photocatalytic performance [33]. Therefore, the efficiency of charge carrier trapping in theprepared samples were investigated by PL spectroscopy. As shown in Fig. 6, all of the doped TiO₂ thin films have lower PL intensity than undoped TiO₂ suggesting an enhanced charge transfer and effective separation of electron-hole pairs in these samples. Among doped samples, the S and Fe co-doped TiO₂ thin film has the lowest PL intensity, therefore simultaneous presence of S and Fe dopants in the structure of TiO₂ has better effect on charge carriers’ separation. Both of oxygen vacancies and lattice defects in the structure of the doped TiO₂ sample can act as charge carrier trapping center and thereby decrease the photoinduced electron–hole recombination rate [34, 35].

Optical properties

Fig. 7(a) depicts the UV–visible absorption spectra of the prepared thin films. Compared with undoped TiO₂, it is obvious that the visible light absorbance intensity of the doped TiO₂ thin films increased. Furthermore, there are noticeable shifts in the absorption edge of the doped TiO₂ samples towards longer wavelengths (red shift), indicating the narrowing of the band gap energy of TiO₂by doping with S and Fe. The precise value of the band-gap energy (Eg) for TiO₂-based semiconductors can be calculated from the intercept of the tangent line to the plots of (αhν)^1/2 versus hν (Tauc plots) [36, 37] as shown in Fig. 7(b). The estimated band gap energies of the undoped, S doped, Fe doped, and S and Fe co-doped TiO₂ thin films are 3.16, 2.53, 2.78 and 2.26 eV, respectively. The band gap narrowing of TiO₂ by iron doping can be attributed to (1) a charge transfer transition between TiO₂ valence or conduction band and d electrons of iron or (2) d–d transition in the crystal field depending on the energy levels [30, 38, 39]. According to the previous first-principle calculation studies, introduction of S element to the lattice of TiO₂ induces band gap narrowing by mixing of the S 3p states with the valence band and localization of S 3p states in the band gap [40, 41]. In the case of S and Fe co-doped TiO₂ thin film, synergic effects of the simultaneous presence of both S and Fe dopants in TiO₂ cause remarkable reduction in the band gap energy of TiO₂.

Transparency of the prepared thin films samples, were examined via UV–Vis transmittance spectroscopy, the results are shown in Fig. 8. As shown in this figure, the transparency of TiO₂, S doped, Fe doped, S and Fe co-doped TiO₂ thin films and glass substrate in visible light region (wavelength of 400-900 nm) are about 79.71 %, 71.62 %, 73.20 %, 71.14 % and 93.40 %, respectively. Therefore, because of the high uniformity and monodispersity of the prepared thin films, they have high transparency in visible light region.

CONCLUSION

In summary, at the present work nanostructured TiO₂ and S and Fe co-doped TiO₂ thin films were deposited on glass substrate through ultrasonic-assisted spray pyrolysis technique. Because of the high uniformity and monodispersity of the prepared thin films, they have high visible light transparency. Both of the undoped and doped TiO₂ have anatase crystal structure, however there are some defect and oxygen vacancies in the doped sample. According to the XPS results, S was doped in the forms of ‎S⁺⁴ and ‎S⁺⁶ cations and oxidation state of Fe in TiO₂ lattice is ‎Fe⁺³. Based on the PL spectroscopy results, because of the defect formation in the doped TiO₂ thin film, this sample has lower PL intensity and consequently lower electron–hole recombination rate than TiO₂ thin film. Furthermore, co-presence of the S and Fe dopants in the structure of TiO₂ results in the remarkable narrowing of the band gap energy of TiO₂.

CONFLICTS OF INTEREST

The authors declare that there are no conflicts of interest regarding the publication of this manuscript.