Photosensitization of TiO2 nanofibers by Ag2S with the synergistic effect of excess surface Ti3+ states for enhanced photocatalytic activity under simulated sunlight

TiO2 nanofibers, with mean diameter ~200 nm, were fabricated by electrospinning and successfully photosensitized with low bandgap Ag2S nanoparticles of 11, 17, 23 and 40 nm mean sizes, with corresponding loading of 4, 10, 18 and 29 wt.% Ag2S, respectively. 17 nm Ag2S@TiO2 nanofibers exhibited optimal activity in the photodegradation of methylene blue under simulated sunlight with pseudo-first order rate constant of 0.019 min−1 compared to 0.009 min−1 for pure TiO2 nanofibers. In spite of greater visible-light absorption and reduced bandgap, larger than 17 nm Ag2S nanoparticles exhibited sluggish photodegradation kinetics probably due to less photo-induced carriers generation in TiO2 and reduced electron injection rates from the larger sized Ag2S into TiO2. Furthermore, a UV-O3 surface treatment induced excess Ti3+ surface states and oxygen vacancies which synergistically enhanced the photodegradation rate constant to 0.030 min−1 for 17 nm Ag2S@TiO2 sample which is ~70% better than the previously reported for Ag2S/TiO2 hierarchical spheres. This was attributed to the efficient charge separation and transfer driven by increased visible-light absorption, bandgap narrowing and reduced electron-hole recombination rates. The present study demonstrate the potential utilization of Ag2S@TiO2 nanofibers in filtration membranes for removal of organic pollutants from wastewater.

enhanced photocatalytic activity of electrospun mesoporous TiO 2 nanofibers was demonstrated by CO 2 reduction to renewable hydrocarbon fuels 11 .
Many attempts have been made to enhance the photocatalytic activity of TiO 2 under illumination conditions similar to solar light 12,13 . These include doping transition metal ions such as Fe 3+ 14, 15 , Ni 2+ & Ti 3+ 16 , Ti 3+ -doped TiO 2 17, 18 or nonmetal atoms such as N, F and S [19][20][21] , dye-sensitization [22][23][24] , and coupling with a photosensitizer such as g-C 3 N 4 25 , MFe 2 O 4 26 , α-Fe 2 O 3 , Au, Pt and Au/Pt co-catalysts [27][28][29] , CdS 30, 31 , MoS 2 @zeolite 32 , Ag 33 and Ag 3 PO 4 34, 35 . The purpose of coupling with a photosensitizer is to enhance the visible-light harvesting and promote charge separation by effective interface charge transfer between the photosensitizer and TiO 2 for which the relative position of the energy band levels are critical. Recently, Ag 2 S has gained considerable attention as a photosensitizer for TiO 2 [36][37][38][39][40][41] . Ag 2 S is a direct low bandgap semiconductor material (E g~1 .0 eV) which responds well in the whole solar energy spectrum. The conduction (CB) and valence (VB) bands of Ag 2 S are −0.3 eV and +0.7 eV, respectively, which is very desirable for photosensitization of TiO 2 with corresponding CB and VB of −0.1 eV and +3.1 eV, respectively 36  Immobilization of Ag 2 S NPs on TiO 2 nanofibers (NF) is expected to give higher number of active sites. Multiple factors contribute to the photosensitization effect of Ag 2 S on TiO 2 nanofibers such as (a) Effect of Ag 2 S mean size and loading amount on the injection rate of photo-induced electrons into TiO 2 as well as photo-induced carriers generation rate in TiO 2 , (b) Effect of Ag 2 S mean size, loading and coverage on the light harvesting capability, (c) Effect of chemical interaction between Ag 2 S and TiO 2 on possible lowering of the energy bandgap and the recombination rate, and (d) Effect of changing the surface chemical states and defects on TiO 2 (e.g. Ti 3+ and oxygen vacancies) on the carrier separation and transport. Li et al. 43 deposited different amounts of Ag 2 S NPs on TiO 2 nanorod arrays by varying the number of cycles in successive ionic layer absorption and reaction (SILAR) method. They reported an optimum 25 cycle deposition for maximum photocatalytic degradation rates of methylene orange (MO) under visible-light. However, other factors as noted above were not investigated in detail.
In this report, the combined effect of all these factors for photosensitization of TiO 2 nanofibers has been experimentally investigated and optimized. We demonstrate optimal simulated solar light driven photocatalytic activity for TiO 2 NFs sensitized by ~17 nm Ag 2 S NPs, loaded at 10 wt.%, with the synergistic effect of enhanced Ti 3+ chemical states and oxygen vacancies induced by a facile UV-O 3 surface treatment. Moreover, these composite nanofibers were stable over multiple cycles of photodegradation and demonstrate their potential use for water remediation.

Results and Discussion
Morphology, structure and chemical composition. The morphology with mean sizes and distribution of Ag 2 S@TiO 2 NFs are shown in Fig. 1. The uncoated TiO 2 NFs ( Fig. 1(a) inset) have a mean diameter of 195 nm and interstices of ~1 μm size, formed by the interpenetrating nanofibers network. The color of the TiO 2 NFs changed instantaneously as they were dipped in the Tollens reagent with addition of 0.05 M dextrose solution indicating the formation of Ag NPs followed by washing with distilled water and ethanol. The mean size of the Ag NPs was controlled by varying the dipping time in the solution from 0.5-4 min. Figure 1(a-d) shows the successful deposition of Ag 2 S NPs by nucleation and growth of Ag followed by sulfurization without any structural damage to TiO 2 NFs. The phase and chemical composition of the Ag 2 S NPs was confirmed by XPS and XRD as discussed later. The sizes of Ag 2 S NPs showed characteristically log-normal distribution with mean number-averaged size of 11, 17, 23 and 40 nm for dipping times of 0.5, 1, 2 and 4 min, respectively. The number-averaged sizes were statistically measured by tracing at least 100 particles, computing their areas, A, and calculating the corresponding area-equivalent diameters, D, assuming spherical shape π = D A ( 4 / ). The particles remain distinct and approximately spherical for dipping times up to 2 min but most of them coalesce into bigger and mostly irregular particles after 4 min dipping. In this study, the samples will be labelled as x nm-Ag 2 S@TiO 2 where 'x' is the mean size of the Ag 2 S NPs. The weight percent loading of Ag 2 S in the composite nanofibers was determined by inductively coupled plasma optical emission spectroscopy. A standard AgNO 3 salt solution of 1-10 ppm was used to calibrate the emission intensity of Ag which fitted linearly with R 2 value of 0.999. The weight fraction of 11, 17, 23 and 40 nm Ag 2 S@TiO 2 composite nanofibers was found to be 4, 10, 18 and 29%, respectively, assuming that all the Ag converted to Ag 2 S. TEM images of 11 nm-and 23 nm-Ag 2 S@TiO 2 NFs in Fig. 2(a and b) also show uniform deposition of Ag 2 S NPs on the NF surfaces. Their sizes are comparable to those measured from the SEM images. The insets show HR-TEM images of Ag 2 S NPs near the NF edges. The lattices identified in 11 nm-Ag 2 S@TiO 2 correspond to (031) and (112) crystallographic planes with d-spacing of 0.212 and 0.178, respectively, whereas the lattice in 23 nm-Ag 2 S@TiO 2 correspond to (121) crystallographic plane with d-spacing of 0.196 nm. These were also confirmed by the powder XRD scans of Ag 2 S@TiO 2 NF samples shown in the Fig. 3(a), where (112), (121) and (031) 44 signals were easily detected (JCPDS # 14-0072, Ag 2 S, monoclinic). The most intense peak at 2θ = 25.3° corresponds to the (101) crystallographic plane and confirms the formation of pure Anatase TiO 2 phase in all the samples 45  adsorbed surface contaminations. The detailed XPS analysis, discussed next, confirmed the presence of Ag 2 S on the NF surfaces.
Adsorption-desorption isotherm of pure TiO 2 NFs and 17 nm-Ag 2 S@TiO 2 composite NFs in Fig. 4(a) closely resembles a type-II isotherm which is typical for non-porous materials due to multilayer formation. The specific surface area of pure TiO 2 NFs and 17 nm-Ag 2 S@TiO 2 composite NFs was 7.1 and 6.6 m 2 /g, respectively, which is consistent with the reported value of 13.3 m 2 /g for control TiO 2 nanofibers 46 . The pore size distribution in Fig. 4(b) was calculated by NLDFT method. Absence of hysteresis on desorption branch rules out the presence of mesopores, however, high intake of nitrogen at P/P 0 > 0.8 could be attributed to the presence of macropores. Contribution of micropores to the total surface area as calculated by the t-plot method was found to be almost zero.   48 . Ti 2p signal was deconvoluted into four peaks which showed 94.3% Ti 4+ and 5.7% Ti 3+ oxidation state on the TiO 2 NFs surface. The Ti 3+ oxidation state increased to 6.7% after coating with 17 nm Ag 2 S NPs which confirms the lower electron density of Ti atoms in Ag 2 S@TiO 2 NFs. Additionally, the Ti 2p peak positions shifted to higher binding energy indicating the withdrawal of valence electron charge (i.e. oxidation) owing to the strong interaction between Ag 2 S and TiO 2 NFs 49 . The UV-O 3 treatment further shifted the Ti 2p peaks towards higher binding energies with increase in Ti 3+ surface state to 8.2% which indicates the removal of the lattice oxygen. This increase in surface Ti 3+ had a significant effect on the photocatalytic activity as discussed later.
Deconvolution of O 1 s spectra of TiO 2 NFs gave three peaks at 529.7, 531.8 and 533.6 eV, as shown in the Fig. 5(b), which are associated with O 2− species in the lattice (TiO 2 ), oxygen vacancies or defects (Ti 3+ or Ti 2 O 3 ) and chemisorbed or adsorbed oxygen species, respectively 50 . Similar to Ti 2p, the peaks shifted to higher binding energies after coating with Ag 2 S NPs and UV-O 3 treatment. The Ti 2 O 3 content increased from 27.4 to 30.2% after   spheres 39 . This was attributed to the generation of excess Ti 3+ surface states and oxygen vacancies, as shown by the XPS results, which facilitated the charge separation process and reduced the electron-hole recombination. We previously reported enhanced photocatalytic activity under UV irradiation for TiO 2 NPs (P25 Degussa) loaded polyacrylonitrile NFs due to excess Ti 3+ surface defects and enhanced hydrophilicity induced by a UV-O 3 surface treatment 55 . These NFs showed stable performance when tested over 5 cycles of photodegradation as shown in the Fig. 6(b). Please note that the slight decrease in performance is attributed to ~10% material weight loss during the repeated separation and re-use over 5 cycles. The stability of the samples after photodegradation was investigated by SEM and XRD. The phase, size and loading remained stable with no signs of leaching as shown in the Supplementary Information Fig. 1s. The radical trapping experiments were carried out to identify the main active oxidation species for photodegradation of MB. As shown in the Fig. 6(c), the photocatalytic efficiencies of the composite nanofibers significantly decreased from 97% to 54%, 62%, 65%, and 73% after the addition of ascorbic acid (as O 2 −• scavenger), sodium sulfate (as e − scavenger), sodium oxalate (as h + scavenger), and isopropanol (as • OH scavenger), respectively. This indicates that the photoinduced electrons react with dissolved O 2 molecules to yield O 2 −• radicals, while the positively charged holes (h + ) react with OH − derived from H 2 O to form • OH radicals. Thus, all these play a significant role in the photodegradation of MB.
Optical properties of Ag 2 S@TiO 2 nanofibers. To investigate the optical response, the absorption spectra of the Ag 2 S@TiO 2 NFs and their corresponding Kulbelka-Munk transformation of the reflectance spectra are shown in Fig. 7(a and b), respectively. Pure TiO 2 showed the expected bandgap of 3.2 eV and very weak visible-light absorption. However, the visible-light absorption increased and the bandgap decreased, monotonically, by coating with Ag 2 S NPs of increasing size due to the enhanced interfacial chemical interaction and coverage of TiO 2 by the smaller bandgap (E g~1 eV) Ag 2 S generating new midgap energy levels. However, the increased size of Ag 2 S will have a negative impact on the photo-induced carriers generation rate in TiO 2 and injection rate of electrons from the conduction band of Ag 2 S into TiO 2 , thus, increasing their recombination probability. This was shown by the 23 nm-and 40 nm-Ag 2 S@TiO 2 samples which, despite having a lower bandgap (E g~2 .75 eV),  showed even less photocatalytic activity then pure TiO 2 NFs with rate constants of 0.014 and 0.009 min −1 , respectively, as shown in the Fig. 8(a). To investigate the promoted charge separation, the samples were excited by 325 nm laser which generates electron-hole pairs. The recombination of these photoexcited carriers gives photoluminescence (PL) signal with an intensity directly proportional to the electron-hole recombination process. Thus, a lower PL is desirable as it indicates a lower charge recombination which will result in a higher photocatalytic activity. As shown in the Fig. 7(c), the pure TiO 2 NF showed a high PL intensity around 425 nm which significantly diminished for Ag 2 S photosensitized TiO 2 NFs. Similar PL results are reported for Ag/TiO 2 56 and Ag 2 O/TiO 2 57 systems.
The synergistic effect of photosensitization and surface Ti 3+ states. The enhanced photocatalytic performance of 17 nm-Ag 2 S@TiO 2 + UV-O 3 sample is attributed to the synergistic effect of photosensitization by Ag 2 S NPs and excess surface Ti 3+ states and oxygen vacancies. The proposed effective charge transfer process between Ag 2 S and TiO 2 due to the preferred conduction band (CB) and valence band (VB) positions of Ag 2 S, which are lower than corresponding bands in TiO 2 , is schematically depicted in the Fig. 8(b). This facilitated the generated electrons transfer from CB of Ag 2 S to the CB of TiO 2 to effectively inhibit the electron-hole recombination. This photosensitization effect is further supplemented by the high Ti 2 O 3 (Ti 3+ state and oxygen vacancies) near the surface which leads to enhanced absorption of incident solar light below the bandgap of TiO 2 (Ti 4+ state) as shown in the Fig. 8(b) and reported in ref. 58. The midgap state is close to the CB which facilitates the injection of photogenerated electrons into TiO 2 under simulated solar light. Moreover, the Ti 3+ defects and oxygen vacancies slows down the electron-hole recombination kinetics by acting as hole traps, thus, prolonging the lifetime of the charges 59 .
In summary, photosensitization of electrospun TiO 2 nanofibers was done by Ag 2 S nanoparticles with mean sizes of 11-40 nm range for enhanced simulated solar light driven photocatalytic performance. Detailed morphological and structural characterization by SEM, HR-TEM & XRD and chemical states identification by XPS confirmed the growth of Ag 2 S nanoparticles. The key characteristics and parameters of all the samples are shown in the Table 1. The size and coverage of Ag 2 S had significant effect on the photocatalytic activity with 17 nm-Ag 2 S@ TiO 2 giving the optimal activity. This was further enhanced by a surface UV-O 3 treatment which introduced excess Ti 3+ surface defects and oxygen vacancies and reduced the bandgap. This synergistic effect of Ag 2 S sensitization and UV-O 3 treatment is attributed to the reduced recombination effect due to efficient charge transfer. The prepared composite NFs were stable in performance when tested over 5 cycles with a maximum photodegradation rate constant of 0.03 min −1 for the UV-O 3 treated 17 nm-Ag 2 S@TiO 2 nanofibers which is ~70% better than the previously reported value for Ag 2 S@TiO 2 hierarchical spheres. These results demonstrate the potential of using these composite nanofibers for water remediation under sunlight.

Fabrication of TiO 2 nanofibers.
To fabricate TiO 2 NFs by electrospinning, 1.5 g polyvinylpyrrolidone (PVP, M w = 1,300,000 g/mol) was mixed in 10 g ethanol and 4 g acetic acid under vigorous stirring for 1 hr. Then 4 g Titanium(IV) n-butoxide (TNBT, 97%) was added in the solution and further stirred for 1 hr to generate a homogeneous electrospinning solution. This precursor solution was electrospun at 18 kV electrical potential with tip to collector distance of 18 cm with pumping rate of 0.8 mL/hr. The non-woven NF mat was left on the aluminum wired collector overnight for complete hydrolysis and annealed in air at 500 °C for 2 hr with a heating rate of 5 °C/min to obtain pure Anatase TiO 2 NFs. Characterization. The size and morphology of Ag 2 S NPs and TiO 2 NFs was characterized by field emission scanning electron microscopy (FE-SEM, FEI 450 Nova NanoSEM) and high resolution transmission electron microscopy (HR-TEM, Jeol JEM-2100F operating at 200 kV). The weight fraction of Ag 2 S in the composite nanofibers was determined by inductively coupled plasma -optical emission spectroscopy (ICP-OES, Perkin Elmer Optima 2100 DV). 3 mg of the sample was digested in 5 mL conc. HNO 3 and then diluted up to 50 ml with deionized water. The surface chemical composition and oxidation states of Ti, O, Ag and S were quantified by X-ray photoelectron spectroscopy (XPS, PHI-1600) equipped with Al Kα radiation. The peak positions were calibrated to C 1 s (284.8 eV) and plotted using the CasaXPS software. The photoluminescence spectra of the composite nanofibers were recorded by exciting the samples by 325 nm wavelength laser. The crystalline phases were determined by powder X-ray diffraction (XRD, Bruker D2 Phaser). Diffuse reflectance spectra (DRS) in 200-800 nm range were taken by UV-vis-NIR spectrophotometer (Jasco, V-770) to study the light harvesting capabilities of the samples and estimate their bandgaps using Kubelka-Munk function. The specific surface area was measured using Quantachrome (NOVA 2200 e) surface area and pore size analyzer. Silica kernel with cylindrical pore morphology and NLDFT (equilibrium mode) were used for the calculation of pore volume and pore size. Surface area was measured from nitrogen adsorption-desorption isotherm by applying B.E.T. plot in relative pressure range of 0.05-0.30 P/P o . Photocatalytic performance. The photocatalytic performance of the Ag 2 S@TiO 2 NFs were evaluated by photodegradation of methylene blue (MB) solution in water. In a typical experiment, 5 mg of the sample was dispersed in a 10 mL solution of 10 ppm MB and kept in dark for 30 min under continuous stirring to attain a complete adsorption-desorption equilibrium on the surface of the composite NFs. Photocatalytic degradation was carried out by illuminating the suspension with simulated solar light using a commercial 100 W Xenon lamp. The analytical sample from the photocatalytic reaction was collected at regular time intervals and centrifuged. The resulting solution was analyzed by recording the absorbance of the residual MB solution with UV-vis spectrophotometer at a wavelength of 662 nm. To identify the active oxidation species responsible for the photocatalytic degradation of MB, a series of radical trapping experiments were carried out. Isopropanol as • OH scavenger (10 mM), ascorbic acid as O 2 −• scavenger (0.1 mM), sodium sulfate as e − scavenger (10 mM) and sodium oxalate as h + scavenger (10 mM) were added directly into the MB solution containing the composite photocatalysts prior to irradiation by simulated solar light.