Formation of oxygen vacancies and Ti(3+) state in TiO2 thin film and enhanced optical properties by air plasma treatment — TEI-tagged view

Bandna BhartipersName, Santosh KumarpersName, Heung-No LeepersName, Rajesh KumarpersName (2016date). “Formation of oxygen vacancies and Ti(3+) state in TiO2 thin film and enhanced optical properties by air plasma treatment.” Scientific Reportstitle·j.

Object of study — air plasma treatmentterm·mth of Fe- and Co-doped TiO2term·key thin films and their optical properties

Research question — how to raise the light harvesting of TiO2term·key films without losing transparency

Contribution — moderate dopingterm·key combined with low-cost air plasma treatmentterm·mth that creates Ti3+ and oxygen vacanciesterm·dom

Full text — tagged

The reference list is not part of the text we hold for this paper. Our copy was extracted from the publisher's page and the bibliography did not survive that extraction, so the article below ends before its references. Nothing has been removed by us.

Formation of oxygen vacanciesterm·dom and Ti(3+) state in TiO2term·key thin film and enhanced optical properties by air plasma treatmentterm·mth Abstract This is the first time we report that simply air plasma treatmentterm·mth can also enhances the optical absorbanceterm·dom and absorption region of titanium oxide (TiO 2 ) films, while keeping them transparent. TiO 2 thin films having moderate dopingterm·key of Fe and Co exhibit significant enhancement in the aforementioned optical properties upon air plasma treatmentterm·mth. The moderate dopingterm·key could facilitate the formation of charge trap centersterm·dom or avoid the formation of charge recombination centersterm·dom. Variation in surface species viz. Ti 3+ , Ti 4+ , O 2− , oxygen vacanciesterm·dom, OH group and optical properties was studied using X-ray photon spectroscopyterm·mth (XPSterm·mth) and UV-Vis spectroscopy. The air plasma treatmentterm·mth caused enhanced optical absorbanceterm·dom and optical absorption region as revealed by the formation of Ti 3+ and oxygen vacanciesterm·dom in the band gapterm·dom of TiO 2 films. The samples were treated in plasma with varying treatment time from 0 to 60 seconds. With the increasing treatment time, Ti 3+ and oxygen vacanciesterm·dom increased in the Fe and Co doped TiO 2 films leading to increased absorbance; however, the increase in optical absorption region/red shiftterm·key (from 3.22 to 3.00 eV) was observed in Fe doped TiO 2 films, on the contrary Co doped TiO 2 films exhibited blue shiftterm·key (from 3.36 to 3.62 eV) due to Burstein Moss shift.

Among various metal oxide semiconductors, TiO 2 is considered as a prime candidate due to its many peculiar properties 1 2 for diverse applications. It is the most suitable candidate for photocatalytic applications due to its biological and chemical inertness, strong oxidizing power, non-toxicity and long term stabilization against photo and chemical corrosion 3 . The films of TiO 2 have valuable applications in LEDs, gas sensors, heat reflectors, transparent electrodes, thin film photo-anode to develop new photovoltaic, photo-electrochemical cells, solar cells and water splitting 4 5 6 7 8 9 10 . In anodic applications, it is a preferred material because of its low density/molar mass and structural integrity over many charge and discharge cycles 11 . However, the efficiency of pure TiO 2 is substantially low because of its wide band gapterm·dom and fast recombination of photo-generated electrons and holes. The key issue to improve the performance of TiO 2 relies on efficient light harvesting, including the increase of its photo-efficiency and expansion of photo-response region, and to ensure efficient number of photo-generated electrons and holes reaching to the surface before their recombination. In order to meet these desired performances the bands structure modification of TiO 2 is preferred.

Generally, three fundamental approaches are implemented for band structure modification viz. dopingterm·key with metallic/non-metallic elements or co-doping of metallic and non-metallic elements 1 12 13 14 , modification via introducing defects such as oxygen vacanciesterm·dom and Ti 3+ in the band gapterm·dom 15 16 , and surface modification by treatment methods 11 17 18 19 . In metallic dopingterm·key, among the range of dopants such as Ni, Mn, Cr, Cu, Fe etc. 3 20 21 22 23 , the Fe is found most suitable due to its half filled electronic configuration. Similarly, from non-metallic dopants S, C, F, N etc. 24 25 26 27 , the N is preferred. In the case of metallic dopants, there are some contradictory reports that show disadvantages of thermal and chemical instability of TiO 2 . Also, their high dopingterm·mth although enhances the band gapterm·dom but at the same time reduces optical/photocatalytic activityterm·dom because of increasing carrier recombination centersterm·dom 28 29 30 31 . What is the mechanism of observed photo-response of doped/modified TiO 2 ; it is still a question, however a generally accepted concern states that the photo absorption of a material is explained better by introducing the defects in the lattice of TiO 2 . For example, Ti 3+ and oxygen vacanciesterm·dom 32 create trap centersterm·dom, rather than the recombination centersterm·dom unlike the high dopingterm·mth case, and results in the variation of band gapterm·dom of pristine TiO 2 .

On the other hand, surface modification methods including surface hydrogenation 33 , vacuum activation 32 and plasma treatment 34 are also practiced. In the hydrogenation method, the surface of TiO 2 is terminated with hydrogen leading to an enhanced photocatalytic activityterm·dom 35 in visible region; however, it is still unknown that how does the hydrogenation modify a surface to enhance its optical performance (photocatalytic activityterm·dom) 36 . The drawback of the hydrogenation method is that it requires high temperature and the obtained TiO 2 sample/film are black 35 , which makes the films unable for many optoelectronic applications, such as a transparent electrode in optoelectronic devices. Both the vacuum activation and plasma treatment methods create highly stable Ti 3+ and oxygen vacanciesterm·dom 32 34 . In vacuum activation method, the sample may exhibit higher absorption intensity but it appears brown in color 35 , that makes it unable for transparent electrode applications. Finally, in case of plasma treatment methods, generally hydrogen gas is used to create Ti 3+ and oxygen vacanciesterm·dom in TiO 2 , but it is always avoidable to use such a hazardous and expensive gas. Except hydrogen there are few reports on the use of argon 37 , oxygen 38 and nitrogen plasma 39 for surface modification of TiO 2 . We know that the implementation of gas in the treatment chamber may be hazardous and cost effective; therefore, it is always required to avoid the use of hazardous gas, and to implement a simple and low cost approach to meet the requirements. In this regard, treatment by air plasmaterm·key may be an effective approach. However, to the best of our knowledge there is no report on the application of air plasmaterm·key for the surface modification of TiO 2 film.

In this report, the band structure modification of thin transparent films of TiO 2 was done by implementing simply the air plasmaterm·key and thus creating Ti 3+ and oxygen vacanciesterm·dom in TiO 2 films. The effect of air plasma treatmentterm·mth was studied in conjunction with metallic dopingterm·key. First, Fe and Co doped TiO 2 thin films were formed on glass substrate, which were subsequently treated in air plasmaterm·key. Considering the drawback of high metallic dopingterm·key (formation of recombination centersterm·dom), in this study, a moderate amount of dopants were used to enhance the optical properties of TiO 2 thin film and thereafter the air plasmaterm·key was applied to enhance them further. The moderate amount of metallic dopant not only favors the separation of electrons and holes but also narrows the band gapterm·dom of TiO 2 3 . We observed that simultaneous effect of the joint approaches increases photo absorbance as well as expends photo response region of the films towards both the visible and UV spectrum. The doped films of TiO 2 were treated in plasma with varying treatment time. The moderate dopingterm·key of Fe and Co elements reduces band gapterm·dom minutely in both the cases, but when treated with air plasmaterm·key a significant change in the optical properties was observed due to the formation of Ti 3+ and oxygen vacanciesterm·dom in the band gapterm·dom.

Results and Discussion After fabricating, the thin films of pure TiO 2 , Fe and Co doped TiO 2 were treated in air plasmaterm·key for 0, 10, 30 and 60 seconds, which were analyzed for surface morphology and crystal structure variations using SEM (see Supplementary Information ; Figure S1 ) and XRD. Here we show XRD pattern of doped thin films for extreme treatment time 0 and 60 seconds (for XRD spectra of samples treated at other treatment time, please see Supplementary Information ; Figure S2 ). Figure 1(a,b) represents XRD pattern of Fe doped, and Fig. 1(c,d) represents XRD pattern of Co doped TiO 2 thin films for 0 (untreated) and 60 seconds of plasma treatment time. Since there is no detection of Fe and Co signals, it indicates that all the Fe and Co ions in the respective samples gets incorporated into the structure of TiO 2 by replacing some of Ti ion, and occupying the interstitial sites 40 .

Absence of sharp peak in XRD patterns represents amorphous phase of TiO 2 thin films 41 . After plasma treatment 2 θ angle and FWHM of the peaks remain almost unchanged, indicating negligible effect on the film structure. XRD indicates that plasma treatment does not create any change in the crystal structure of Fe and Co doped TiO 2 thin films. The obtained low signal-to-noise ratio in the above XRD spectra is due to the low crystallinity of the films and small crystallite size; such observations have been reported by others as well 42 .

The presence of atomic percentage of the dopants in TiO 2 thin films was detected by EDX signals (see Supplementary Information ; Figure S3 ). The EDX of Fe doped TiO 2 film shows the atomic percentage of Fe, Ti and O as 1.66%, 12.93% and 85.41%, respectively, which closely matches to the stoichiometry of elements in Ti 0.95 Fe 0.05 O 2 . Similarly, in case of Co doped TiO 2 , the obtained atomic percentage of Co, Ti and O in EDX are 1.33%, 23.33% and 75.35%, respectively, which confirms the stoichiometry of elements of Ti 0.95 Co 0.05 O 2 thin film.

Variation in optical properties of TiO 2 thin films by dopingterm·key and subsequent air plasma treatmentterm·mth was analyzed by UV-Vis spectrophotometer. The change in absorption edgeterm·key and corresponding band gapterm·dom is mentioned in Table 1 . Pure TiO 2 film (undoped and untreated) showed absorption edgeterm·key at 367 nm and band gapterm·dom 3.37 eV, whereas Fe doped TiO 2 film showed a shift in the absorption edgeterm·key to 385 nm, with a decreasing in the band gapterm·dom to 3.22 eV. Similarly, Co dopingterm·key shifts the absorption edgeterm·key from 367 nm to 369 nm with a reduction in the band gapterm·dom to 3.36 eV. The observed red shiftterm·key in absorption edgeterm·key and narrowing band gapterm·dom in both dopants cases is similar to other reports on metallic dopingterm·key 3 . In both the cases, samples were doped with a moderate (5%) concentration of Fe and Co forming Ti 0.95 Fe 0.05 O 2 and Ti 0.05 Co 0.05 O 2 , respectively. We could have tuned the optical properties further by increasing the dopant concentration but that would form recombination centersterm·dom 28 ; therefore, to avoid the formation of recombination centersterm·dom, a further tuning in the optical properties was done by treating these moderately doped TiO 2 films in air plasmaterm·key. The films were treated in air plasmaterm·key for treatment time (0, 10, 30 and 60 seconds), and investigated for the shift in absorption edgeterm·key and band gapterm·dom variation. With increasing treatment time, the absorption edgeterm·key of Fe doped TiO 2 films shifts continuously from 385 nm (for 0 seconds treatment time) to 413 nm (for 60 seconds treatment time), with a corresponding band gapterm·dom change from 3.22 eV to 3.00 eV, showing a significant increase in the absorption region. In case of Co doped TiO 2 films, the absorption edgeterm·key shifts from 369 nm to 342 nm (for 60 seconds treatment time) with a corresponding band gapterm·dom change from 3.36 to 3.62 eV, which shows an increase in the optical band gapterm·dom/UV absorption region probably due to the Burstein-Moss effectterm·dom 43 , explained latter.

From the Table, it is observed that the change in optical properties of TiO 2 films appears at two levels; first by the dopingterm·key of Fe and Co, and then by plasma treatment. However, here it should be noted that the change in the band gapterm·dom due to the dopingterm·key is smaller as compared to the subsequent band gapterm·dom change by plasma treatment. While discussing the effect of dopingterm·key on the change of band gapterm·dom, we know that the reduction may take place due to either by the increasing grain size of highly crystalline sample 44 or the formation of electronic energy levels within energy band gapterm·dom 45 . In our study, since the XRD results showed the samples to be amorphous, thus the first reason can be discarded. Therefore, Fe 3+ and Co 2+ ions substitute Ti 4+ ions in TiO 2 matrix and cause a change in the band gapterm·dom by forming their mid gap energy levels in the respective samples along with the formation of Ti 3+ and oxygen vacanciesterm·dom. The electronic transition from valance band to dopant level and then from dopant level to conduction band, and/or from valance band to oxygen level and then form oxygen level to Ti 3+ level/dopant level effectively cause a red shiftterm·key in the absorption edgeterm·key, showing reduced band gapterm·dom 46 47 48 . In many cases, the localized level of t 2g state of the dopingterm·key element even lies in the middle of band gapterm·dom (in case of, Cr, Mn or Fe as the dopingterm·key materials), and at the top of the valance band (when Co is used as a dopant) 49 . Next, the variation in the absorption edgeterm·key/band gapterm·dom with plasma treatment time is due to the increase of Ti 3+ and oxygen vacanciesterm·dom, detailed discussion is given under XPSterm·mth studies in the following section.

Figure 2 shows variation in the absorption spectra of Fe doped TiO 2 thin film treated for 60 seconds of time ( Fig. 2(b) ) with respect to untreated one ( Fig. 2(a) ) (to see the increase in the absorption edgeterm·key and reduction in band gapterm·dom, please refer to Supplementary Information ; Figure S4 ). There is a continuous change in the absorbance, absorption edgeterm·key and band gapterm·dom of the films with plasma treatment time. The absorbance of the film increased from 60% (untreated film) to 87% (treated for 60 seconds) along with a red shiftterm·key in the absorption edgeterm·key and band gapterm·dom narrowing by 0.22 eV (Tauc plot shown in the inset of Fig. 2(b) ). The band gapterm·dom and absorption edgeterm·key were estimated using the following equations 50 : where α is absorption coefficient and E g is band gapterm·dom energy.

Similarly, the variation in absorption spectra of Co doped TiO 2 thin film treated for 0 and 60 seconds is shown in Fig. 3(a,b) (details of other samples is given in Supplementary Information ; Figure S5 ). In this case, dopingterm·key shows a red shiftterm·key due to the presence of Co levels in the energy gap of TiO 2 , whereas after plasma treatment the film shows continuous blue shiftterm·key with increasing treatment time. This overall shift (due to treatment in plasma for 60 seconds) in the band gapterm·dom is 0.26 eV. The observed blue shiftterm·key can be explained by Burstein-Moss effectterm·dom 43 , resulted by the change in the position of Fermi level into the conduction band. General equation representing enhancement in the band gapterm·dom energy is given by: where and are the effective mass of hole and electron in the respective bands, and K F is Fermi wave vector. In our case, the shift of Fermi level into the conduction band leads to the energy band widening. Absorption edgeterm·key shifts to shorter wavelength region due to the increase in the carrier concentration, which is discussed in XPSterm·mth studies section.

The overall variation in the absorption edgeterm·key and band gapterm·dom of TiO 2 thin film due to the dopingterm·key (Fe and Co) and air plasma treatmentterm·mth is plotted in Fig. 4 . In the plasma treatment region, a remarkable change in the band gapterm·dom values can be observed with treatment time.

XPSterm·mth study In order to understand the mechanism resulting the change in the band gapterm·dom of Fe and Co doped TiO 2 films with plasma treatment time, the films were investigated by XPSterm·mth. The XPSterm·mth being surface sensitive technique provides information about the change in chemical state of film constituting species. Here, the variation in the chemical state of elements ‘O’ and ‘Ti’ with plasma treatment time was analyzed in detail to correlate it with the observed variations in the band gapterm·dom of the films. Figure 5(a,b) shows XPSterm·mth survey spectra of untreated and plasma treated Fe and Co doped TiO 2 thin films, respectively. In these spectra, C1s is probably an instrumental impurity. The intensities of O1s and Ti2p peaks increase with the increasing plasma treatment time, indicating an increase in these states with treatment time.

Figure 6(a) shows high resolution XPSterm·mth spectrum of pure TiO 2 film. In this spectrum, the doublet ‘Ti2p 3/2 (binding energy 458.6 eV) and Ti2p 1/2 (binding energy 464.4 eV)’ arises from spin orbit-splitting. These peaks are consistent with Ti 4+ in TiO 2 lattice 51 . Also, the shoulder Ti2p 1/2 at binding energy 460.2 eV is corresponding to Ti 3+ 52 in Ti 2 O 3 . This indicates that both TiO 2 and Ti 2 O 3 are formed in the film (Without deconvolution, the XPSterm·mth spectra are shown in Supplementary Figure S6 ). After dopingterm·key with Fe, the high resolution XPSterm·mth spectrum ( Fig. 6(b) ) shows a slight shift in the position along with a variation in the area of the original peaks. The peaks in the Fe doped samples are now located at binding energies 458.4 (Ti2p 3/2 ), 464.3 eV (Ti2p 1/2 ) and 459.0 eV (Ti2p 1/2 ), respectively (see Supplementary Information ; Table S1 ). The shift in the position of these peaks indicates an influence of Fe addition on the electronic state of Ti element; probably some of the Ti ions get substituted with Fe ions in the lattices. After dopingterm·key, the area of Ti 3+ peak increased by 81% and that of the peak Ti 4+ decreased by 19%. The increase in the area of Ti 3+ peak indicates that either Ti 2 O 3 is formed in large amount or some mixed oxide structure with Fe (having oxidation state Ti 3+ ) is formed after dopingterm·key. Meanwhile, the decreasing area of Ti 4+ indicates a reduction of TiO 2 in the sample, and probably formation of Ti-O-Fe structure in the TiO 2 lattice through the substitution of transition metal ions. Observed shift in the peaks also indicates interaction between Ti and Fe atoms and an overlapping of their 3d orbital 53 . This causes an electronic excitation from Fe to Ti in the optical absorption experimentterm·mth, which shows a reduction in the band gapterm·dom of Fe doped TiO 2 film (as observed in the optical analysis).

After dopingterm·key, the film was treated in air plasmaterm·key. In the XPSterm·mth results, only the sample which was treated for 60 seconds in plasma is demonstrated. The XPSterm·mth shows a further increase in the peak corresponding to Ti 3+ at 459.0 eV ( Fig. 6(c) ) and a decrease in the peak area of Ti 4+ . The change in stoichiometry was estimated by the change in the area of relative peaks. The peak area of Ti 3+ increases by 20% and that of Ti 4+ decreases by 12%. The increase in the peak area of Ti 3+ indicates that after plasma treatment there is removal of oxygen from the lattice, which shows a relative increase in Ti 3+ in the XPSterm·mth spectrum. On the other hand decreasing peak area of Ti 4+ is inferred due to the reaction of Ti 4+ with electrons coming either from plasma or due to the formation of oxygen vacanciesterm·dom in the surface layer generated by the plasma treatment 41 . Now, as observed in optical analysis, the band gapterm·dom of Fe doped films (3.22 eV) decreased to 3.00 eV (for 60 seconds of treatment time), this is correlated with the increasing career/electrons density due to plasma treatment. As we know that in the doped samples, the possible reasons of red shiftterm·key/decreasing band gapterm·dom is the introduction of donor states in the energy gap (here oxygen vacanciesterm·dom and Ti 3+ , Table 1 ). In the present case, the band gapterm·dom decreases further with increasing treatment time, while the concentration of the dopant was kept constant, which is due to the change in the surface states of the constituents i.e. Ti element and oxygen vacanciesterm·dom.

Next, the O1s spectrum of pure TiO 2 thin film is shown in Fig. 6(d) , which is fitted with three peaks. The peaks at binding energies 529.9 eV, 530.3 eV and 531.3 eV are attributed to lattice oxygen, Ti 2 O 3 and non-lattice oxygen 54 55 . Similarly, for the doped sample, O1s spectrum of Fe doped TiO 2 thin film fitted with two peaks is shown in Fig. 6(e) . In this spectrum, only two peaks at binding energies 529.8 eV, and 531.9 eV are observed which are attributed to lattice oxygen and surface adsorbed OH group, whereas the peak 530.3 eV corresponding to Ti 2 O 3 , disappears. This indicates that in the dopingterm·key process TiO 2 is formed along with some mixed oxide. Again, the change in stoichiometry was estimated by the change in area of relative peaks. In case of Fe doped TiO 2 film, the area of the peak at 529.7 increases by 64% and that of the peak at 531.5 eV increases by 54%.

After plasma treatment, the binding energy of lattice oxygen (O in TiO 2 ) shifts slightly from 529.8 eV to 529.7 eV ( Fig. 6(f) ), whereas its area increases by 35%. Also, the area of the peak at 531.5 eV (non-lattice oxygen/OH) increases by 15% (see Supplementary Information ; Table S1 ). The increase in the area of non-lattice oxygen indicates the formation of oxygen vacanciesterm·dom in the lattice. This result is analogues to the XPSterm·mth spectrum of Ti2p ( Fig. 6(c) ).

Fe dopingterm·key results in a minor shift in the binding energy, indicating that Fe ions are better dispersed in the substitutional sites of TiO 2 lattice and produce more mixed oxide structure, probably Fe-O-Ti. Figure 7(a) shows high resolution XPSterm·mth spectrum (for Fe2p 3/2 ) of Fe doped TiO 2 film. After plasma treatment, the high resolution XPSterm·mth spectrum of Fe2p 3/2 is shown in Fig. 7(b) . These spectra are fitted with Gauss–peak shapes as shown in Fig. 7(c,d) . The deconvoluted XPSterm·mth spectrum of Fe2p 3/2 ( Fig. 7(c,d) ) contains main peaks at 710.1 eV and 724.6.1 eV corresponding to Fe2p 3/2 and Fe2p 1/2 , respectively (see Supplementary Information ; Table S2 ). The appearance of these peaks supports the presence of Fe in Fe 3+ ionic state 55 . Further, after plasma treatment the shift in the binding energy of Fe2p 3/2 from 710.1 eV to 711.3 eV also indicates the presence of Fe 3+ species, irrespective of the particular oxide (i.e., Fe 2 O 3 , Fe 3 O 4 , and FeOOH). Shake up satellite at 716.9 eV also supports that Fe is presented in Fe 3+ state (oxide) 56 . These shake-up satellites are associated with Fe3d-O2p hybridization. Thus XPSterm·mth analysis confirmed that Fe ions are doped into TiO 2 matrix in the form of Fe-O-Ti. From the XPSterm·mth analysis, we confirmed that by increasing the plasma treatment time the concentration of Ti 3+ and oxygen vacanciesterm·dom also increases.

The Co doped samples after treating in plasma show adverse effect on the band gapterm·dom of the doped TiO 2 film. In this case, band gapterm·dom increases with the increasing treatment time as observed in optical studies. To investigate this divergent behavior, the samples were analyzed via XPSterm·mth, Fig. 8 shows high resolution spectra. Figure 8(a) shows the XPSterm·mth spectrum of pure TiO 2 , and Fig. 8(b) shows XPSterm·mth for Co doped sample. As discussed above in the case of Fe doped sample, the XPSterm·mth of pure TiO 2 is also fitted with three peaks corresponding to titanium dioxide (Ti 4+ ) and titanium sub oxide (Ti 3+ ) in Ti2p 1/2 and Ti2p 3/2 , respectively. These peaks are fitted as Ti 4+ 2p 1/2 at 464.4 eV, Ti 4+ 2p 3/2 at 458.6 eV, and Ti 3+ 2p 3/2 at 460.2 eV. The line separation between Ti2p1/2 and Ti2p3/2 is 5.8 eV, which is consistent with the standard binding energy of TiO 2 51 . However, in this case the Ti2p spectrum ( Fig. 8(b) ) is fitted with four peaks as 464.4 for Ti 4+ 2p 1/2 , 458.6 eV for Ti 4+ 2p 3/2 , 460.4 for Ti 3+ 2p 3/2 and 457.9 eV for Ti 3+ 2p 1/2 57 , respectively (see Supplementary Information ; Table S1 ). In comparison to the pure TiO 2 , the area of Ti 3+ peak in Co doped TiO 2 increases by 26%, while that of the peak Ti 4+ decreases by 7%, indicating a reduction in the formation of TiO 2 , which is similar to the case of Fe doped samples.

After the plasma treatment ( Fig. 8(c) ), binding energies of the mentioned peaks are shifted slightly to the positions such as 464.3 eV (Ti 4+ 2p 1/2 ), 458.5 eV (Ti 4+ 2p 3/2 ), 460.6 eV (Ti 3+ 2p 3/2 ) and 457.4 eV (Ti 3+ 2p 1/2 ), respectively. The change in stoichiometry was estimated by the change in peak area of respective peaks.

After plasma treatment, while investigating for peak area, we observed that the peak area of Ti 3+ increases by 30%, whereas the peak area of Ti 4+ decreases by 12%. Again, this is expected due to the reaction of Ti 4+ with the electrons coming either from plasma or due to the formation of oxygen vacanciesterm·dom in the surface layer by the plasma treatment. Further, the high resolution O1s XPSterm·mth spectrum obtained for Co doped sample is shown in Fig. 8(d–f) . The spectrum is fitted with three peaks i.e. 529.9 eV, 530.3 eV and 531.6 eV that correspond to lattice oxygen of TiO 2 , oxygen in Ti 2 O 3 and non-lattice oxygen, respectively.

The change in stoichiometry was estimated by change in the peak area of relative peaks. With the dopingterm·key of Co, the lattice oxygen (corresponding to TiO 2 ) peak at 529.9 shifts to the position 530.3 eV, and the area of the peaks at 530.3 eV and 531.6 eV increases by 51% and 24%, respectively. The original peak at 530.3 eV ( Fig. 8(d) ) corresponding to Ti 2 O 3 disappears after dopingterm·key ( Fig. 8(e) ), which is due to the formation of mixed oxide structure. Further, with the increasing treatment time, the areas of the peaks at 530.3 eV and 531.6 eV (( Fig. 8(f) ) also increases by 24% and 25%, respectively. (To explain in a more quantitative manner we have tabulated all the data in a table by comparing all the peaks at different plasma treatments time, see Supplementary Information ; Table S1 ).

Next, Fig. 9(a) corresponds to high resolution XPSterm·mth spectra of Co2p region of Co doped TiO 2 thin films and Fig. 9(b) shows high resolution XPSterm·mth spectra with plasma treatment. Figure 9(c,d) represent deconvoluted XPSterm·mth spectra of doped TiO 2 and plasma treated TiO 2 thin films, respectively. The core level binding energies of peaks Co2p 1/2 and Co2p 3/2 are 796.9 eV and 781.0 eV, respectively. The satellite peaks at 787 eV and 802 eV reveal high spin Co(II) state with complex transitions 58 . These results are an indication that Co does not precipitate as metallic Co on the film surface. After plasma treatment, the satellites peaks shifts slightly to the 785.3 eV and 802.3 eV. Also, the binding energies of Co2p 1/2 and Co2p 3/2 are shifted to 796.6 eV and 781.2 eV, respectively (see Supplementary Information ; Table S1 ). These spectra are typical of compounds containing high-spin Co 2+ ions 59 60 , reveling the presence of CoO(Co 2+ ), CoTiO 3 (Co 2+ ), Co 2 O 3 (Co 3+ ) or mixed valence Co 3 O 4 (Co 2+ and Co 3+ ) in the surface. The presence of strong satellites indicates that Co atoms in the doped TiO 2 film are in 2+ oxidation state, referring the possible formation of CoO or CoTiO 3 inside the film.

Now we discuss the probable reason of band gapterm·dom narrowing in TiO 2 film with Fe dopingterm·key, and widening in the case of Co dopingterm·key after plasma treatment. As reported, the iron dopant acts as an acceptor impurity in TiO 2 lattice 61 . Thus when the TiO 2 film is doped with Fe, the acceptor levels of Fe along with oxygen vacanciesterm·dom are created in the band gapterm·dom of TiO 2 62 . In our case, as discussed above Ti 3+ is also formed which creates energy level in the band gapterm·dom, contributing to the reduction of band gapterm·dom. Next, when this Fe doped TiO 2 film was treated in air plasmaterm·key, the Ti 3+ levels and oxygen vacanciesterm·dom increases further with the treatment time, whereas no change in the dopant levels occurs as the dopant concentration was kept constant. The increase in Ti 3+ levels and oxygen vacanciesterm·dom would further reduce the band gapterm·dom of Fe doped TiO 2 film. In case of Co dopingterm·key, there is a formation of Co acceptor levels along with Ti 3+ and oxygen vacanciesterm·dom levels in the band gapterm·dom which reduces the band gapterm·dom of Co doped TiO 2 film. But when the film was treated with plasma we observed continuous widening in the band gapterm·dom with treatment time. The observed increase in the band gapterm·dom can be explained by Burstein-Moss effectterm·dom. The probable reason for Burstein-Moss shift in this case is that with the treatment time the Ti 3+ levels and oxygen vacanciesterm·dom increases more as compared to Fe doped case. By plasma treatment for 60 seconds the Ti 3+ increases by 20%, oxygen vacanciesterm·dom increases by 15% in case of Fe doped TiO 2 , whereas Co doped TiO 2 Ti 3+ increases 30%, oxygen vacanciesterm·dom increases 25%. These created levels donate more electrons and thus shift the Fermi level to the conduction band, which increases the band gapterm·dom of Co doped TiO 2 film. The exact reason for this divergent behavior is unclear as of now but the most appropriate reason seems to us is, the on-site coulomb interaction/repulsion that are occurring only in case of Co doped TiO 2 films 63 . When Co 2+ ion substitutes Ti 4+ ions, the imbalance positive charge inside the lattice is compensated by the formation of oxygen vacanciesterm·dom located near Co ion. The formation of oxygen vacanciesterm·dom is equivalent to the addition of two electrons per Co ion 64 65 . The oxygen vacanciesterm·dom produced in case of Co doped TiO 2 thin films are higher as compared to Fe doped TiO 2 films as observed by XPSterm·mth. Suppose both Fe and Co doped films increase by same values of Ti 3+ levels and oxygen vacanciesterm·dom, but due to Columbian interactions, which are only in case of Co doped TiO 2 64 65 , the optical transition results in the blue shiftterm·key of the absorption spectra. The proposed mechanism for both the Fe and Co doped TiO 2 is illustrated in Fig. 10 .

Conclusion Treatment by air plasmaterm·key leads to significant change in the optical properties of TiO 2 thin films. Unlike other treatment methods, in this approach the transparency of TiO 2 thin film remains invariant. The charge separation centers i.e. oxygen vacanciesterm·dom and Ti 3+ is created with the dopingterm·key of metallic Fe and Co elements; however, they are significantly enhanced by the air plasma treatmentterm·mth. In Fe doped TiO 2 thin film, the formation of oxygen vacanciesterm·dom and Ti 3+ causes enhances absorbance and red shiftterm·key due to the formation of energy levels in the band gapterm·dom, whereas in Co doped TiO 2 the Burstein-Moss shift is effective to make blue shiftterm·key in the absorption spectra. Conclusively, we can say that the joint approaches i.e. low level/moderate dopingterm·key and safe and low cost air plasma treatmentterm·mth resulted in enhanced optical properties of transparent TiO 2 thin films, making them efficient candidate for transparent electrode applications.

Experimental Methods Thin films of TiO 2 , Fe doped TiO 2 (Ti 0.95 Fe 0.05 O 2 ) and Co doped TiO 2 (Ti 0.95 Co 0.05 O 2 ) were fabricated on glass substrate using dip-coating method. Titanium (IV) isopropoxide (TTIP, Ti[OCH (CH 3 ) 2 ] 4 , 97%, Aldrich) was used as precursor solution. First of all triethanolamine C 6 H 15 NO 3 , a stabilization agent was dissolved in C 2 H 5 OH, which resulted in a colorless solution. In this solution, the precursor solution Ti[OCH (CH 3 ) 2 ] 4 was added dropwise to form a pale yellow solution with a continuous stirring. To avoid the precipitation of TiO 2 , C 2 H 5 OH and H 2 O was added in a ratio 9:1. Now during the sol gel synthesis the solutions of ferric nitrate (Fe (NO 3 ) 3 .9H 2 O), and cobalt nitrate (Co (NO 3 ) 2 .6H 2 O) were added separately as the dopant in TiO 2 . These solutions were stirred for two hours and allowed for ageing overnight. Then glass substrates cleaned with H 2 O, detergent, C 3 H 6 O and C 2 H 5 OH were coated with the aged solution. Coated films were dried and annealed at 400 °C to form transparent thin films. The fabricated films were treated in air plasmaterm·key, generated in a vacuum coating unit (Hindhivac model: 12A4D), for varying treatment time; 0, 10, 30, and 60 seconds, respectively. The air plasmaterm·key was generated at reduced pressure of 10 −3 mbar in the vacuum chamber. During the treatment process the applied bias voltage was 30 volts with a power of 22.7 watt. After treating in plasma, the samples were analyzed for optical, structural, morphological and surface properties.

Materials Characterization The optical (absorbance, shift in absorption edgeterm·key and band gapterm·dom) properties of the films were studied by UV-Vis spectrophotometer (Perkin-Elmer Lambda 750). The band gapterm·dom of Fe and Co doped thin films was calculated by using the absorbance spectra by plotting ( αhv ) 1/2 against hv , where hv being incident photon energy. Surface morphology was studied using scanning electron microscopy (SEM), and elemental confirmation was done using energy dispersive X-ray (EDX). The structural analysis of the samples was done using X-ray diffractometer (XRD) (company name Rigaku, with Cu kα radiation, λ = 1.5406 Å), and to observe the effect of plasma treatment on surfaces states, X-ray photoelectron spectroscopy (XPSterm·mth: VG Multilab 2000, Thermo electron corporation, UK) studies were performed.

Additional Information How to cite this article : BhartipersName, B. et al . Formation of oxygen vacanciesterm·dom and Ti 3+ state in TiO 2 thin film and enhanced optical properties by air plasma treatmentterm·mth. Sci. Rep. 6 , 32355; doi: 10.1038/srep32355 (2016).

Supplementary Material