Sol-gel powder synthesis in the TiO₂-TeO₂-ZnO system: Structural characterization and properties

2.1 Samples preparation

A combination of Te(VI) acid (Aldrich) along with Ti butoxide (Fluka AG), Zn acetate dihydrate (Merck) precursors dissolved in ethylene glycol (C₂H₆O₂) (99% Aldrich) were used. Bearing in mind the high hydrolysis rate of tellurium(VI) alkoxide, the telluric acid (H₆TeO₆) is selected in order to overcome that problem (Weng et al., 2004; Hodgson and Weng, 2006; Lecomte et al., 2007; Weng and Hodgson, 2001). The precursor solutions were subjected to 5–30 min intensive stirring at room temperature in order to achieve complete dissolution. No direct addition of water was made to the precursors’ solutions. Sol-gel hydrolysis reaction was acquired from absorbed atmospheric moisture. The adjusting of the pH during the sol-gel processes is very important. Generally, this parameter influences the rate of hydrolysis and condensation processes. At strongly acidic values (pH up to 5) a slow gelation occurred which favors the formation of extended network while basic values of pH (up to 9) induce a fast gelation which is necessary for the formation of nanoparticles (Wright and Sommerdijk, 2000). In our experiments, the measured pH was 4–5 depending on composition. The gelation time for the investigated compositions differed varying from minutes and hours to days. The aging of gels was performed in air in order to achieve completeness of the processes. The obtained gels were subjected to the stepwise heating from 200 to 700 °C for one hour exposure time in air. The heat treatment at 200 °C was performed in air in order to hydrolyze any unreacted - OR groups as well as to decrease their content. Further increases of temperature (300–700 °C) were important to verify the phase and structural transformations of the gels.

2.2 Samples characterization

The thermal stability of selected gels (Table 2) was determined by differential thermal analysis (LABSYSTM EVO apparatus) with Pt–Pt/Rh thermocouple at a heating rate of 10 K/min in air flow, using Al₂O₃ as a reference material. The accuracy of the temperature was ±5 °C. The heating of the samples was limited up to 600 °C. Powder XRD patterns were registered at room temperature with a Bruker D8 Advance X-ray powder diffractometer with a Cu Kα radiation (λ = 1.54056 Å) with a LynxEye solid position sensitive detector and X-ray tube operated at 40 kV and 40 mA. X-ray diffraction patterns were recorded in the range of 5.3–80° 2θ with a step of 0.02° 2θ. The optical absorption spectra of the powdered samples in the wavelength range 200–1000 nm were recorded by a UV–VIS diffused reflectance Spectrophotometer “Evolution 300″ using a magnesium oxide reflectance standard as the baseline. The absorption edge and the optical band gap were determined following Dharma et al. instructions (Stoyanova et al., 2012). The band gap energies (E_g) of the samples were calculated by the Planck’s equation: ${\mathrm{E}}_g=\frac{h·c}{\lambda }=\frac{1240}{\lambda }$ , where Eg is the band gap energy (eV), h is the Planck’s constant, c is the light velocity (m/s), and λ is the wavelength (nm). The infrared spectra were registered in the range 1600–400 cm⁻¹ using the KBr pellet technique on a Nicolet-320 FTIR spectrometer with 64 scans and a resolution of ±1 cm⁻¹.

X-ray photoelectron spectroscopy (XPS) measurements were carried out on an AXIS Supra electron spectrometer (Kratos Analytical Ltd.) with base vacuum in the analysis chamber in the order of 10⁻⁹ Torr. The samples were irradiated with Al Kα photons with energy of 1486.6 eV. The photoemitted electrons were separated, according to their kinetic energy, by a 180° – hemispherical analyzer with a total instrumental resolution of 1.0 eV (as measured by the FWHM of Ag 3d_5/2 line) at constant pass energy of 20 eV. Due to a charging effect the energy calibration was performed by normalizing the C 1s line of adventitious adsorbed hydrocarbons on silver folio to 284.8 eV. The concentrations (in at.%) of the observed chemical elements were calculated by normalizing the areas of their most intense photoelectron peaks to their relative sensitivity factors using the commercial software of the spectrometer.

3.1 Thermal stability and phase transformations of the gels

All gels prepared at room temperature were transparent and monolithic. A wide gel formation region was determined consisting of several subregions depending on the gelation time of the compositions (Fig. 1). As one can see, the gel formation region is expanded toward ZnO corner and the obtained gels are situated between 20 and 100 mol % TiO₂, up to 65 mol % TeO₂ and 90 mol % ZnO. The first gelation region (A) comprises samples for which the gelation occurred immediately after mixing of the components. For the other two regions (B and C) this process took longer time, from hours and days to one month, respectively. In the binary ZnO-TeO₂ system, gels were not obtained using H₆TeO₆ and Zn acetate as a combination of precursors. Table 1 shows all investigated compositions in the TiO₂-TeO₂-ZnO system.

Close

Fig. 1

Gel formation region in the ternary TiO₂-TeO₂-ZnO system.

Export to PPT

The thermal decomposition data in air of selected gels for which the time for gelation was different are summarized in Table 2. Generally, several decomposition steps were observed and bearing in mind the similarity of the DTA/TG curves of the samples, only two selected gels TTZ1 and TTZ5 are presented (Fig. 2a, b). As is seen from the figure some peaks are well evident for both samples. For the interpretation of these results, DTA data for pure Ti (IV) n-butoxide and Zn acetate precursor discussed in the literature (Velasco et al., 1999; Yordanov et al., 2017; Arii and Kishi, 2003; McAdie, 1966) have been kept in mind. As is seen from Fig. 2 both gels are characterized by endothermic peaks in the range 100–200 °C with weight loss between 8 and 18% depending on composition. The mass loss is mainly related to the local elimination of water bonded with the surface of the materials as well as to the evaporation of organic solvent. Looking at the DTA/TG curves, several exothermic peaks are observed in the range 220–440 °C. The first one is observed about at 230 and at 300 °C for gels TTZ1 and TTZ2, respectively. The observed peak is sharper and more intensive for the sample containing the higher TeO₂ amount (above 40 mol %, gels TTZ1 and TTZ3) and it could be related to the intense decomposition of both Ti(IV) butoxide and Zn acetate. Obviously, in the sample containing the higher TiO₂ amount (TTZ2) the combustion of organics occurred at a higher temperature (300 °C). In both cases this peak is accompanied by a strong weight loss about 20% which according to the TG curve finished at nearly 450 °C. The next exothermic peak is registered at a higher temperature of about 440 °C in both samples. The last small exothermic effect at 440 °C and it’s presence in gel TTZ1 could be related to the oxidation of Te to TeO₂ but in the other sample TTZ2 it could be attributed to the crystallization of TiO₂ (anatase). Additionally, that peak is accompanied by a very low weight loss (2%) but the calculated total weight loss for the selected samples is about 40%.

Close

Fig. 2

DTA/TG curves of two representative gels: (a) TTZ1 and (b) TTZ5.

Export to PPT

X-ray diffraction (XRD) patterns of as-prepared ternary gels with different nominal compositions are shown in Fig. 3a and it is evident that they are amorphous. The investigated samples contain a different TiO₂ amount aiming to verify it’s role in the formation of the amorphous network as well as in the phase formation with variation of temperature. As is seen from the figure the samples exhibit different behavior depending on the temperature (Fig. 3b). The XRD patterns of samples obtained below 200 °C are complicated due to the presence of organics and formation of intermediate metal-organic complexes, which are not considered in this paper. Generally, in the temperature range 200–400 °C composite materials containing an amorphous phase and different crystalline phases {metallic Te (JCPDS 78-2312), α-TeO₂ (JCPDS 42-1365), anatase (JCPDS 78-2486)} are prepared. In contrast to the samples TTZ5, TTZ3 and TTZ2 where at 400 °C TiO₂ (anatase) or α-TeO₂ were observed, in the last sample TTZ1 only metallic tellurium has been found. Despite the low TiO₂ content in the later sample, the amorphous state is more stable due to presence of higher amount of both TeO₂ and ZnO (40 mol %). The oxidation of tellurium starts at about 400 °C and it is fully oxidized to TeO₂ about 500 °C (samples TTZ1 and TTZ2). In the samples TTZ3 and TTZ5, anatase appears at 400 °C and a small amount is converted to rutile at 500 °C, while in the other two samples, anatase did not appeared in the XRD patterns. The average crystallite size (calculated using Sherrer’s equation) of TiO₂ (anatase) in the powdered sample TTZ5 is about 10 nm (Fig. 3b) at 400 °C. At higher temperatures (500–700 °C) only crystalline phases (TiO₂ – rutile, α-TeO₂ and ZnTeO₃) were found depending on the composition of the prepared composite materials (Fig. 3b). These investigations are upheld by the previous ones where TiO₂/TeO₂/ZnO composite powders containing constant TiO₂ amount (40 mol%) have been prepared (Gegova et al., 2015).

Close

Fig. 3a

XRD patterns of the as prepared gels.

Export to PPT

Close

Fig. 3b

XRD patterns of the investigated samples: (■) Te, (♦) α-TeO₂, (▴) TiO₂-anatase, ( ) TiO₂-rutile, (○) ZnTeO₃.

Export to PPT

Close

Fig. 3b

XRD patterns of the investigated samples: (■) Te, (♦) α-TeO₂, (▴) TiO₂-anatase, ( ) TiO₂-rutile, (○) ZnTeO₃.

Export to PPT

3.2 Optical properties and structural transformations

The diffuse reflectance absorption spectra of representative gels (aged at room temperature) as well as of pure Ti butoxide gel are illustrated in Fig. 4. The observed absorption edges and calculated optical band gap values are pointed out in Table 3. It is evident that the TTZ5 gel containing higher TiO₂ content (80 mol%) exhibited the highest absorption in the UV region in comparison to the other ternary gels as well as pure Ti(IV) butoxide gel. Another peculiarity is the appearance of two absorption bands below and above 300 nm, the so called charge transfer bands. The optical properties of various ternary gels from the systems TiO₂-TeO₂-B₂O₃ and TiO₂-TeO₂-SeO₂ were discussed in our previous papers (Iordanova et al., 2015, 2016). By analogy it could be accepted that for the isolated TiO₄ units, the ligand to metal charge transfer band is in the region 200–260 nm, while in a titania network (anatase) the charge transfer in TiO₆ is above 300 nm (Singh and Singh, 2001; Clarck, 1968; Klein et al., 1996). As is known the unhydrolyzed Ti butoxide consists mainly of TiO₄ groups as main building units but after polymerization the coordination geometry changed to TiO₆ as a result of formation of Ti-O-Ti links between TiO₆ units (Klein et al., 1996). As can be seen from the figure, for gels TTZ5, TTZ3 and TTZ2 the more intensive absorption peak is that about 300–320 nm which is associated with an increase of the polymerization degree of Ti atoms in comparison to pure Ti butoxide gel. The UV–Vis spectrum of TTZ1 gel differs from the other ternary ones and pure Ti(IV) butoxide. One of the reasons could be that the lowest TiO₂ content (20 mol%) is not sufficient for the fast formation of the amorphous network and resulted in a very slow gelation of that sample. On the other hand, the charge transfer of O → Ti is in the range 260–300 nm but it is known that the exciton absorption peak of bulk ZnO exists in the same region (300–360 nm, Eg ∼ 3.24 eV) (Bahnemann et al., 1987). Due to the overlapping of these absorption bands it is difficult to obtain more precise information for the influence of ZnO on the polymerization ability of Ti. The absorption edge of the investigated gels varies from 330 nm to 364 mm and the increase in TiO₂ content caused a red shifting of the cut-off (Table 3). Among the ternary gels the TTZ5 (80 mol% TiO₂) one exhibited the highest cut-off value 363.82 nm. Several authors (Kabalci et al., 2011; Ghribi et al., 2015) claimed that TiO₂/TeO₂/ZnO glasses (containing up to 35 mol % ZnO) exhibited an optical band gap value of about 2.94–3.00 eV. Our values for the investigated gels are higher (3.41–3.75 eV) than in the literature (Niu et al., 2018) for pure TiO₂ (3.2 eV) and ZnO (3.37 eV) but lower than the value of melted bulk tellurite glass (∼3.79 eV). Obviously, the Eg of the obtained products is influenced by the simultaneous presence of TiO₂, TeO₂ and ZnO.

Close

Fig. 4

UV–Vis spectra of selected ternary TiO₂/TeO₂/ZnO gels compared to pure Ti(IV) n-butoxide gel.

Export to PPT

The IR spectroscopy was used for verifying the phase transformations during the heat treatment (in the temperature range 200–700 °C) as well as to evaluate the completeness of the hydrolysis – condensation reaction (Fig. 5). The assignments of the vibrational bands of the separate structural units are made on the basis of well-known spectral data of the precursors (Ti(IV) n-butoxide, H₆TeO₆ and Zn acetate – Fig. SI.1) and crystalline phases existing in the system (Iordanova et al., 2013, 2015, 2016; Yordanov et al., 2017). The measured by us spectra of the precursors were used only as supporting information. These spectra have been discussed in details elsewhere (Iordanova et al., 2016; Gegova et al., 2015a, 2015b).

Close

Fig. 5

IR spectra of the representative samples heat treated in the range 200–700 °C.

Export to PPT

Close

Fig. 5

IR spectra of the representative samples heat treated in the range 200–700 °C.

Export to PPT

For evaluation of the completeness of the hydrolysis-condensation reactions, it is necessary to follow the changes of several bands in the range 1400–1000 cm⁻¹ (Iordanova et al., 2016; Gegova et al., 2015a, 2015b). Moreover, it is known that the band at 1130 cm⁻¹ is characteristic for the stretching vibrations of Ti—O—C (Iordanova et al., 2016), while those at 1100 and 1040 cm⁻¹ are assigned to the vibrations of terminal and bridging C—O bonds in butoxy ligands (Iordanova et al., 2016). On the other hand, ethylene glycol possesses asymmetric and symmetric stretching vibrations of C—O bonds in CH₂—OH group at 1090 and 1040 cm⁻¹, while in the spectrum of butanol they appeared at 1110, 1050 cm⁻¹ along with a very intensive band at 1070 cm⁻¹ (Iordanova et al., 2016).

As it is seen from Fig. 5, the IR spectra of all gels and heat treated at 200 °C samples are very similar and they are characterized with very intensive bands in the 1120–1040 cm⁻¹ region. The comparison between the IR spectra of the investigated gels showed differences regarding the intensity of the bands at 1080 and 1040–1030 cm⁻¹ which could be related mainly to the C—O bonds in CH₂—OH group. The band at 1040–1030 cm⁻¹ is stronger in the spectra of TTZ1, TTZ2 and TTZ3 gels, while in TTZ5 this band exhibited lower intensity. Additionally, the other band at 1080 cm⁻¹ possesses lowest intensity only in the IR spectrum of TTZ1 gel. Considering the experimental facts obtained, it could be suggested that a greater degree of hydrolysis and condensation processes is achieved in TTZ5, TTZ3 and TTZ2 samples containing 80, 50 and 40 mol% TiO₂, respectively. Thus, despite the higher ZnO content (40 mol%), the lower amount of TiO₂ (20 mol%) is not enough to complete these processes in TTZ1 gel. The band at 1120–110 cm⁻¹ could be ascribed to Ti—O—C bonds and it has to be noted that this band did not change it’s intensity.

The other intensive band is at 890–880 cm⁻¹ which is assigned to C—O or C—H stretching vibrations of the residual solvent ethylene glycol (Iordanova et al., 2016; Gegova et al., 2015a, 2015b; Zubkova and Shabadash, 1971; Kato et al., 1996). The bands located in the range of 1500–1300 cm⁻¹ are assigned to the bending vibrations of CH₃ and CH₂ organic groups. Additionally, bands below 900 cm⁻¹ which are typical for the inorganic units were registered as well up to 200 °C. Accordingly, we can suggest that at lower temperatures (200 °C) the amorphous phases consist both of organic and inorganic building units. The bands decreased with increasing the temperature but are still observed up to 300 °C then completely disappeared about 400 °C. This means that the hydrolysis-condensation processes are not completed at room temperature and continue in a wide temperature range along with the decomposition of the organic groups (Hayakawa et al., 2012). All bands above 1000 cm⁻¹, characteristic for the organic groups are not visible above 300 °C and the spectra in the temperature range 300–500 °C are characterized mainly with bands below 900 cm⁻¹. These bands have low intensity and are also broadened which are a peculiarity of the disordered systems. The obtained IR results give us a reason to suggest that the amorphous phase in the temperature range 400–700 °C consists only of inorganic building units: TeO₄ (780–770 cm⁻¹; 630–620 cm⁻¹), TiO₆ and ZnO₄ (630–620 cm⁻¹, 480 cm⁻¹, 410 cm⁻¹) (Murashkevich et al., 2008; Yurchenko et al., 1981; Yamaguchi et al., 1987; Shabalin, 1982; Mancheva et al., 2011; Andres-Verges and Martinez – Gailego, 1992; Dimitriev et al., 2009). As it has been mentioned above at this temperature range only crystalline phases TiO₂ – rutile, α-TeO₂ and ZnTeO₃ were detected in the prepared composite materials (Figs. 3). It has to be noted that it is difficult to distinguish the vibrations of TiO₆ and ZnO₄ structural units due to their strong overlapping. In the IR spectra above 400 °C, vibrations were found which are characteristic of the inorganic structural units building up the crystalline phases (TeO₂, TiO₂ (anatase), TiO₂ (rutile) and ZnTiO₃. The obtained IR spectra proved the phase transition established by XRD (Figs. 3). Based on the IR results it is difficult to generalize which is the better network former TiO₂, TeO₂ or ZnO. Obviously, under the experimental conditions provided and the combination of precursors used in this study, the amorphous network is influenced by the simultaneous presence of all oxides. On the other hand, bearing in mind the peculiarity of the IR spectra of the TTZ1 gel and the lower intensity of the bands at 1110, 1080 and 1030 cm⁻¹ it could be suggested that TiO₂ behaves as a better non-traditional amorphous network former.

The chemical states of the main elements (C, Ti, Te and Zn) were determined by X-ray photoelectron spectroscopy (XPS) (Figs. 6–11) of selected gels containing different TiO₂ amount. The concentrations of the observed chemical elements and the binding energies of their different valence states are listed in Table 4.

Close

Fig. 6

C 1s-photoelectron region of the studied B samples. The peak contributions to the C 1s peak structure and the sum of deconvolution peaks are marked in blue and red, respectively.

Export to PPT

Close

Fig. 7

Te 3d-photoelectron region of the studied gel and annealed B samples. The peak contributions to the Te 3d doublet and the sum of deconvolution peaks are marked in blue and red, respectively.

Export to PPT

Close

Fig. 8

Ti 2p spectral regions of (a) anatase and (b) Ti butoxide. The curves in blue are the fit peak contributions and their sum is marked in red.

Export to PPT

Close

Fig. 9

Ti 2p-photoelectron region of the studied gel and annealed B samples. The peak contributions to the Ti 2p doublet and the sum of deconvolution peaks are marked in blue and red, respectively. The peaks, colored in green, correspond to the contributions of tetrahedral coordinated Ti⁴⁺ ions.

Export to PPT

Close

Fig. 10

Zn 2p-photoelectron region of the studied gels and annealed at 200 °C samples.

Export to PPT

Close

Fig. 11

O 1s spectra, measured for the studied samples. The peak contributions and their sum are marked in blue and red, respectively.

Export to PPT

For all samples studied, the C1s spectral regions show broad features which indicate the presence of organic components in the gels as well as in the samples annealed at 200 °C (Fig. 6). These spectra can be fitted with three peak contributions. The first component at 284.8 eV is ascribed to C—C and C—H bonds (Lu et al., 2013), while the second contribution at 285.9–286.2 eV can be attributed to C—OH bonds (Ohtsu et al., 2009; Zhong et al., 2010). The high binding – energy component at 287.8–288.5 eV can be related to —O—C bonds to the metal ions, for example Ti—O—C, which presence is an indication for the uncompleted hydrolysis at these experimental conditions (Dai et al., 2013; Ren et al., 2007; Stankovich et al., 2007). As can be seen from Table 4, the concentration of carbon atoms is highest relative to the other chemical elements found in the samples. For gels, it varies in the range of 47.3–48.8 at%. When heated to 200 °C, the carbon concentration decreases slightly to 39.0–43.8 at%. However, within each sample, the concentration ratio of the three carbon – containing species (suggested above) remains insignificantly changed.

Similarly, the Te 3d spectral regions also exhibit complex peak structures which can be fitted with several chemically shifted spin-orbital 3d_5/2 and 3d_3/2 doublets (Fig. 7). The chemical shifts between their binding-energy positions imply the presence of several valence states of tellurium atoms in the surface layers of the samples. For their identification the XPS spectrum of the H₆TeO₆ has been used (Gegova et al., 2015). It was already found that in H₆TeO₆ the tellurium has 6+ valence state (octahedral coordinated) and the Te 3d5/2 binding energy has been measured at 577.4 eV. In all samples a high binding-energy peak contribution in the range of 576.4–576.9 eV is observed which can be attributed to Te⁶⁺ ions in octahedral chemical environment (gels and annealed samples) (Fig. 7). As is seen the intensity of the characteristic for the Te⁶⁺ peak did not decrease in the heated samples (Fig. 7), which probably is proof of the higher stability of the amorphous state and the decomposition of the orthotelluric acid (H₆TeO₆) is postponed (Remy, 1963). For the TTZ3 and TTZ5 sample a relatively intense second peak with maximum at 572.8–573.4 eV is observed which could be associated with the presence of Te²⁺ ions on the sample surfaces (Häggblad et al., 2008) and/or of metallic tellurium (Te⁰) (Chang et al., 2014). As can be seen from Fig. 7, tellurium atoms in these two valence states exist not only in gels but also in the annealed samples. While it can be argued for the TTZ3 sample that heating to 200 °C causes a decrease in their concentration, in TTZ5 the reverse trend-increase may be observed (Fig. 7). Significantly lower concentrations of Te⁰/Te²⁺ valence states are observed for the TTZ1 sample compared to that of Te⁶⁺.

In order to obtain optimal deconvolutions of the Te 3d spectra of TTZ3 and TTZ5 samples, a peak contribution with binding energy in the range of 574.4–575.1 eV (Fig. 7) must be allowed. According to our previous studies of Te - containing systems (Bachvarova-Nedelcheva et al., 2014), this contribution may be due to the existence of Te³⁺ valence state.

The XPS spectra of investigated by us gels and annealed samples show Ti 2p_3/2 peaks in the range of 458.5–458.9 eV (Fig. 8). It should be noted that these values are located between the Ti 2p_3/2 binding energies characteristic for anatase and Ti butoxide, where the Ti⁴⁺ ions are octa- and tetrahedral coordinated, respectively. For anatase, containing TiO₆ structural units, we measured Ti 2p_3/2 binding energy at 458.6 eV (Iordanova et al., 2016; Sachse et al., 2015) which is lower than those of pure Ti butoxide (gel and annealed at 200 °C) (459.1 eV). Fig. 8 shows the corresponding Ti 2p spectral regions for both pure compounds. As can be seen, the spectra can be very well fitted with two pairs of Ti 2p doublets with spin-orbital splitting of 5.7 eV and two times higher 2p_3/2 peak area than that of the 2p_1/2 peak, according to the electron population of both core levels. The results show that for homogeneous substances (anatase and Ti butoxide) the Ti 2p_3/2 binding-energies differ modestly by 0.5 eV. However, unlike anatase, the Ti 2p_3/2 the binding energy of tetrahedral coordinated Ti⁴⁺ varies greatly with decrease of it’s relative concentration reaching high values of 460.5 eV (at below 1 at%) (Sachse et al., 2015). These considerations are very useful in analyzing the Ti 2p regions of the gels we studied (Fig. 9). The wide and somewhat asymmetrical peaks can be very well fitted with two pairs of Ti 2p_3/2+1/2 doublets. The highly intense Ti 2p doublet has a 2p_3/2 binding-energy of 458.4–458.6 eV, indicating that octahedrally coordinated Ti⁴⁺ ions predominate. The low intense Ti 2p doublet (colored in green) with a 2p_3/2 binding-energies in the range of 459.6–460.5 eV corresponds to tetrahedral coordinated Ti⁴⁺ ions. According to the above considerations, these high binding-energies indicate that the concentrations of TiO₄ units containing tetrahedral coordinated Ti⁴⁺ are isolated or form very small clusters. Indeed, the relative areas of the Ti 2p doublets allow one to estimate the concentrations of tetrahedral Ti⁴⁺ ions in the gels and they are in the range of 0.4–1.2 at %. Annealing at 200 °C leads to the increase of the total amount of Ti which is valid for both types of octa- and tetrahedral coordinated Ti⁴⁺ ions. These results sit well with the data discussed above, obtained by UV–Vis spectroscopy.

As a summary of the Ti 2p considerations it can be concluded that the studied samples contain predominantly TiO₆ units. This can be as a result of a higher degree of polycondensation as suggested by Barlier et al. (2008) studying various Ti alkoxides. The authors related the slight shifting toward higher binding energies with uncompleted condensation reactions. In the present study our analysis suggests that the Ti 2p binding-energy shift to higher energies is due to the presence of isolated TiO₄ units (or their small clusters) containing tetrahedral coordinated Ti⁴⁺ ions. Uncompleted condensation was another suggestion by us for the annealed binary 80TiO₂.20SeO₂ and 80TiO₂.20TeO₂ compositions (Iordanova et al., 2016).

The Zn 2p photoelectron regions are characterized by single 2p_3/2 and 2p_1/2 peaks [Fig. 10]. The binding energy (BE) of Zn 2p_3/2 core – level is found in the region of 1021.9 ± 0.2 eV. It is worth to be mentioned that this value corresponds well to the binding energy of the Zn 2p3/2 core-level electrons in pure and Al,N co-doped ZnO films (Gao et al., 2013; Levtushenko et al., 2017). In order to determine the Zn valence state, an Auger Zn LMM region has been also measured. For all samples the kinetic energies (KE) of this Auger peak has been found to be 988.3 ± 0.2 eV which gives a modified Auger parameter of 2010.2 eV (defines as a sum of BE of Zn 2p_3/2 and KE of Zn LMM Auger peak (Wagner et al., 1979). This value is characteristic for the Zn²⁺ valence state (Biesinger et al., 2010).

In the O 1s spectral regions of all samples, a wide asymmetric peak between 529 and 534 eV is observed, indicating the presence of oxygen atoms involved in different chemical bonds (Fig. 11). The lack of clear features of these peaks makes their deconvolutions difficult and respectively, the identification of various chemical bonds involving oxygen. One possible way out is to use the location of the O 1s peaks of appropriate standards. The Te 3d_5/2-, Ti 2p_3/2- and Zn 2p_3/2- spectral regions discussed above (Figs. 7, 9, 10 and Table 2) indicate that the appropriate standards are anatase (standard for Ti⁴⁺), H₆TeO₆ (standard for Te⁶⁺), and ZnO (standard for Zn²⁺). However, these substances are insulators and are charged due to the electron photoemission in the XPS experiments, and therefore the absolute value of the O 1s binding – energy depends on the energy calibration. This problem is avoided, as Perez-Gonzales and Tomas proposed (Perez-Gonzales and Tomas, in press), if a relative difference Δ between the binding energies of the most intense photoelectron peaks and the O 1s energy is used. For our study these differences are: Δ(Te 3d_5/2-O 1s) = 577.4–532.2 = 45.2 eV for the Te—O—H bonds (Iordanova et al., 2016), Δ(Te 3d_5/2-O 1s) = 576.8–531.3 = 45.5 eV for Te—O—Te bonds (Bachvarova-Nedelcheva et al., 2014), Δ(Ti 2p_3/2-O 1s) = 458.6–529.8 = −71.8 eV for the Ti—O—Ti bonds (Sachse et al., 2015) and Δ(Zn 2p_3/2-O 1s) = 492.1 eV for ZnO (Perez-Gonzales and Tomas, in press).

In Fig. 11 the O 1s spectra as well as their deconvolutions of the gels and the heated samples are presented. These spectra can only be fitted with peak contributions with binding energies exceeding 530 eV. According to the research of Mekki et al. (2005); this is an indication of the existence of bridging oxygen bonds, for example Me—O—Me (Me = Te, Ti, Zn). In agreement with our considerations above, only the Te—O—Te bonds have clearly different binding energies at about 531.3 eV, while the Ti—O—Ti and Zn—O—Zn have similar energies close to 530 eV. The adsorbed OH groups are expected to contribute to the highest binding-energy peak, above 532 eV.

Indeed, according to these assumptions the O 1s spectra are well fitted with three peak contributions (Fig. 11). As can be seen from the figure, a broad and intense contribution exists at about 530.4–530 6 eV, which could be related to bridging oxygen bonds involving Ti and Zn. This assumption is also confirmed by the total concentrations of Ti and Zn (Table 4) relative to that of Zn in the various samples. Their ratio is in keeping with the relative peak area at 530.4–530.6 eV relative to the contribution area at about 531.4–531.7 eV, which we attribute to the Te—O—Te bridge connection. The peak contribution with the highest binding energy of about 532.5–533.0 eV is associated with the presence of adsorbed OH groups on the sample surface.

Considering the XPS results, especially for Te 3d (Fig. 7), O 1s and Table 4, one can see their similarity for the TTZ3 and TTZ5 samples, for which the ratio of the total concentration of Ti and Zn relative to that of Te is significantly more than 1. For the TTZ1 sample this ratio is almost 1 (Table 4). Completing the results of XPS analysis, it should be noted that the photoelectron spectroscopy is a surface-sensitive method, and therefore its results apply to the surface layers of samples with a thickness of about 6–8 nm.