Characteristics and local structure of hafnia-silicate-zirconate ceramic nanomixtures

Zirconate systems having the composition 3HfO₂·15SiO₂·xY₂O₃·(82 − x)ZrO₂, where x = 2, 7 and 12 mol% Y₂O₃, were synthesized by a sol-gel method. The analysis of X-ray diffraction data showed the presence of the t-ZrO₂, m-ZrO₂, m-HfO₂, Y₂SiO₅ and Y₂Si₂O₇ crystalline phases in a ceramic nanomixture. Spectroscopic data show that the increase of the Y₂O₃ content of samples determines the increase of the t-ZrO₂, m-HfO₂ and silicate crystalline phases. Gap energy values decrease almost linearly with increasing Y₂O₃ content of samples. A detailed study of XANES data does not show a significant difference with increasing Y₂O₃ content of the samples suggesting an appreciable stability of the hafnium ions +4 oxidation state and their microvicinity. EXAFS results show that the local structure around the Hf cation is similar to that from the monoclinic crystalline HfO₂ where the Hf–O coordination number tends to 7. The bond lengths of Hf–O shells show small deviations from ∼2.12 Å and the Hf–metal paths become more structured by increasing the Y₂O₃ content of the samples.

1. Introduction

Zirconium oxide based systems have been drawing increasing attention of researchers due to their unique properties that are promising for biomedical applications. Thus, ceramics based on ZrO₂ are well known for their low thermal conductivity, high fracture toughness and biocompatibility in comparison with other materials (Ponnilavan et al., 2015).

As has been observed up to now, pure zirconia (ZrO₂) possesses three polymorph structures: cubic (c-ZrO₂), tetragonal (t-ZrO₂) and monoclinic (m-ZrO₂). The t-ZrO₂ and m-ZrO₂ phase transformation has important technological consequences, since the extensive microcracking caused by the anisotropic volume expansion accompanying the reaction leads to the loss of mechanical integrity in sintered bodies. When a metal oxide with a lower valence cation is incorporated into cubic ZrO₂, the fluorite-type cubic phase is stabilized. One of the common `stabilizers' is yttrium oxide since it has good physical and chemical stability (Torres & Llopis, 2009; Qin et al., 2012; Singh et al., 2015). The substitution of this lower valence cation (+3) in the Zr⁺⁴ sublattice sites creates oxygen vacancies. The Y₂O₃ can stabilize the fluorite-type cubic structure and hence avoid the problem of the t-ZrO₂ and m-ZrO₂ transformation by eliminating the t-phase (Wu et al., 1992; Chu et al., 2016).

Further, the dual role of zirconia in glass matrices was confirmed as a network former and/or as a network modifier. Thus, zirconia playing the network former role leads to a network consisting of ZrO₄ tetrahedra while playing the network modifier role leads to a network consisting of ZrO₆ octahedra with four bridging and two non-bridging oxygen atoms (Glazneva et al., 2012). Addition of HfO₂ to ZrO₂ enhances the chemical stability and thermal-shock resistance of the host ceramic (Zhen et al., 2007; Karunarathne et al., 2017). Note that zirconia containing silicon oxide has attracted attention too because of its excellent resistance to alkali corrosion and its low thermal expansion (Simhan, 1983; Kwon et al., 2017). In such a material, hafnium oxide or zirconium oxide were the best metal oxides to replace the silicon dioxide.

Note that fundamental studies on the crystalline structure and microstructure of monoclinic ZrO₂ ceramics with or without dopants have rarely been reported (Kwon et al., 2017; Rada, Culea et al., 2018). The research on monoclinic ZrO₂ ceramics can be associated with difficulties in the manufacturing process and the mechanical instability at temperature variations.

The ZrO₂–Y₂O₃ compositions are candidates for the transmutation of radionuclides in reactors and as accelerator-based neutron sources due to their high radiation resistance (Moll et al., 2009).

Taking into account the important properties of the oxides presented above, we prepared some new ceramics containing all of these oxides using the sol-gel method. Considering the development of new materials with mixed compositions, the experimental conditions can be adjusted in order to obtain stable or metastable compounds. The physical and chemical properties of mixed oxides prepared by the sol-gel method are of great interest due to their thermal and chemical properties and because they are highly acidic due to the presence of OH groups on the surface.

Bearing in mind the above-mentioned aspects, in this study we wanted to prepare some new zirconate materials with potential biomedical and optical applications. Thus, new zirconate nanostructures were prepared in the 3HfO₂·15SiO₂·xY₂O₃·(82 − x)ZrO₂ system where x = 2, 7 and 12 mol% Y₂O₃ via the sol-gel method. To the best of our knowledge, this system has not been reported in the scientific literature until now.

The investigation of the 3HfO₂·15SiO₂·xY₂O₃·(82 − x)ZrO₂ zirconate materials was carried out by X-ray diffraction (XRD) and scanning electron microscope (SEM) analysis, Fourier transform infrared (FT-IR), UV-Vis, photoluminescence (PL), Raman and X-ray absorption spectroscopies. X-ray absorption spectroscopy (XAS) with its two amendments, X-ray absorption near-edge structure (XANES) and extended X-ray absorption fine structure (EXAFS), represents an important investigation technique of the local structure around ions of interest. Thus, EXAFS analysis was used to study the structure of the Hf atoms' microvicinity, their coordination number and the length of the bonds involving Hf atoms.

First, one of the main objectives of the present work was to perform the structural characterization of the studied new zirconate materials. Our interest was focused on obtaining structural information (via XRD, SEM analysis, FTIR and X-ray absorption spectroscopies) concerning the characteristic crystalline phases and structural units that occur in the studied materials. Second, some spectroscopic properties of the studied zirconate samples were investigated (by using FT-IR, UV-Vis, PL and Raman spectroscopies) to find out their potential applications for optical devices. The electronic transitions that take place were studied. Finally, the influence of the addition of Y₂O₃ was followed, since dopants can be useful tools to improve the structural and/or behavioral properties of the host materials.

2. Experimental data

Tetraethoxysilane and high purity ZrO₂, Y₂O₃ and HfO₂ powders were used as starting materials. Tetraethoxysilane was diluted with ethanol. Solutions of Zr(NO₃)₄, Y(NO₃)₃ and Hf(NO₃)₄ were prepared by dissolving ZrO₂, Y₂O₃ and HfO₂ into HNO₃ solutions of pH = 2. Suitable volumes of tetraethoxysilane, Zr(NO₃)₄, Y(NO₃)₃ and Hf(NO₃)₄ were mixed and stirred at 70°C to obtain homogeneous solutions. The resulting solutions were stirred at 100°C until a dry powder was obtained after the evaporation of H₂O. These dried powders were uniaxially compacted under pressure in the form of disks (2 mm thick, 9 mm diameter). The disks were calcined in alumina crucibles using an electric furnace heated at 1400°C for two hours. The disks were put into the electric furnace direct at this temperature. After two hours, the melted disk materials were quenched at room temperature by pouring onto a stainless-steel plate.

The samples were characterized by several methods: XRD, SEM analysis, FTIR, UV-Vis, PL, Raman and XAS spectroscopy that were previously described by Rada et al. (2013, 2017, 2019), Rada, Zhang et al. (2018) and Dehelean et al. (2016).

The changes of the microstructure of the samples by doping were characterized by analysing the profile of the diffraction peaks and SEM micrographs. FTIR spectroscopy was recorded to determine the presence of the specific structural units, such as zirconia or silicate network in the prepared samples. UV-Vis data were used to determine the gap energy. The UV-Vis, PL and Raman spectroscopy were also used to show the presence of the zirconia polymorphs. The XRD peaks of monoclinic zirconia and monoclinic hafnia are very similar because they arise at nearly the same angles. XANES and EXAFS were performed in order to indicate the formation of the HfO₂ crystalline phase with monoclinic structure.

3. Results and discussion

3.1. XRD data

X-ray diffraction was used to identify the main crystalline phases in the studied samples. Fig. 1 shows the XRD patterns of the samples from the 3HfO₂·15SiO₂·xY₂O₃·(82 − x)ZrO₂ system where x = 2, 7 and 12 mol% Y₂O₃. All diffractograms show peaks characteristic of crystalline phases. In order to identify the crystalline phases that appear in the diffraction patterns we followed a standard procedure, namely computer analysis based on the powder diffraction file (PDF).

Figure 1 X-ray diffraction patterns for the 3HfO2·15SiO2·xY2O3·(82 − x)ZrO2 system.

This analysis permitted the identification of five crystalline phases present in the samples. Two of these phases belong to the ZrO₂ crystalline phase with monoclinic (PDF No. 03-065-1024, monoclinic lattice with a = 5.14608 Å, b = 5.21177 Å, c = 5.31302 Å and α = 99.222° cell parameter) and tetragonal (PDF No. 00-050-1089, a = 3.5984 Å, c = 5.152 Å) symmetry. The HfO₂ crystalline phase with monoclinic structure (PDF No. 00-034-0104, monoclinic lattice with a = 5.2851, b = 5.1819, c = 5.1157 Å cell parameter), the yttrium silicate (Y₂SiO₅, PDF No. 00-021-1456) and yttrium disilicate (Y₂Si₂O₇, PDF No. 00-021-1459) crystalline phases were also shown in the studied samples. The assignments of the most important crystalline peaks are shown in Fig. 1. The X-ray diffractograms from Fig. 1 show that by increasing the yttrium content of samples, x, the relative intensity of the peaks attributed to the monoclinic ZrO₂ decreases, while the relative intensity of peaks assigned to the tetragonal ZrO₂ phase increase. Bearing in mind different potential applications, this is an important evolution since the tetragonal ZrO₂ is more stable that the monoclinic one.

The () and (111) reflection peaks situated at about 28.2° and 31.4° of the monoclinic ZrO₂ crystalline phase (Rada, Culea et al., 2018) did not appear to be significantly modified by increasing the Y₂O₃ content of the samples. Analysis of the reflections in the XRD patterns of the monoclinic ZrO₂ phase and monoclinic HfO₂ crystalline phase shows the same satellite peaks. Accordingly, it is difficult to follow the effect of the Y₂O₃ content on the amount of monoclinic ZrO₂ phase in the studied samples.

A small amount of Y₂Si₂O₅ crystalline phase was detected in the sample with x = 2 mol% Y₂O₃. By increasing the Y₂O₃ content up to x = 7 mol%, the intensity of the characteristic peaks of the Y₂SiO₅ crystalline phase increases. After that, at a higher Y₂O₃ content, x = 12 mol%, a new diffraction peak centered at about 22.03° can be seen suggesting the presence of the Y₂Si₂O₇ crystalline phase. This compositional evolution shows that by increasing the yttrium ions content of samples the zirconia host matrix was converted into a yttrium–silicate–zirconia matrix. The excess of oxygen atoms can be accommodated in the host matrix by the formation of yttrium silicates.

3.2. FT-IR data

The FT-IR spectra of the 3HfO₂·15SiO₂·xY₂O₃·(82 − x)ZrO₂ samples with x = 2, 7 and 12 mol% are shown in Fig. 2. Each FT-IR absorption band is related to some vibration type specific of the structural groups that build up the investigated samples. The positions and assignments of the most important absorption bands shown in the FT-IR spectra are discussed. The assignments were made based on data from the literature concerning some vitreous or crystalline compounds related to the system investigated in this work (Lee & Condrate, 1988; Choi & Choy, 2016; Vitanov et al., 2014; Ortiz et al., 2005; Khomenkova et al., 2014; Najafinezhad et al., 2017; Ferdov et al., 2012).

Figure 2 Infrared spectra of the 3HfO2·15SiO2·xY2O3·(82 − x)ZrO2 system.

According to Lee & Condrate (1988), the wavenumber and/or intensity of vibrational modes involving Zr—O bonds will vary depending upon the arrangement of the oxygen atoms around the zirconium atoms. Thus, several zirconium–oxygen structural units can be considered with respect to the Zr—O stretching modes: ZrO₄ units, ZrO₆ units, monoclinic zirconia-type ZrO₇ units, tetragonal zirconia-type ZrO₈ units and cubic zirconia-type ZrO₈ units.

The absorption bands around 436, 386 and 310 cm⁻¹ are present in all the studied samples (Fig. 2). These bands can be assigned to the symmetric and/or asymmetric stretching vibrations of chemical bonds in which zirconium and hafnium ions are involved.

The FT-IR features situated at about 525 cm⁻¹, 580 cm⁻¹ and 745 cm⁻¹ are due to Zr—O—Zr asymmetric and Zr—O stretching modes and confirm the presence of monoclinic ZrO₂ phases in the studied system. The intensity of the mentioned vibrational features becomes less prominent with increasing the Y₂O₃ content of samples. In addition, a new weak IR feature associated with the Zr—O vibrations in tetragonal ZrO₂ crystalline structure appears around 600 cm⁻¹ for x = 12 mol% Y₂O₃. Note that this crystalline phase was detected by XRD analysis, too.

The IR bands centered at about 870, 960 and 1020 cm⁻¹ can be associated with the symmetric and antisymmetric stretching vibrations of the silicate units. By increasing the Y₂O₃ dopant level up to 12 mol%, the intensities of these bands increase, suggesting the formation of higher amounts of yttrium silicates in the host matrix.

The structural changes observed by the FT-IR investigation suggest that the yttrium ions play a network modifier role in the studied 3HfO₂·15SiO₂·xY₂O₃·(82 − x)ZrO₂ system. These structural changes, determined by increasing the Y₂O₃ content of the samples, consist of the increase of the tetragonal ZrO₂ and silicate crystalline phases. This assumption is in good agreement with the results of the XRD investigation. The observed compositional evolution shows that the Y₂O₃ content can be used as an efficient tool to improve the stability of the studied system.

3.3. UV-Vis data

The UV-Vis spectra of the 3HfO₂·15SiO₂·xY₂O₃·(82 − x)ZrO₂ zirconate nanostructures where x = 2, 7 and 12 mol% are given in Fig. 3. The samples show a very intense absorption in the UV region and a moderate absorption in the visible region. The bands from 225 nm, 265 nm and 350 nm correspond to the metal–ligand charge-transfer (oxygen electron from the free orbital of the metal ion). The absorption band from 225 nm is the highest one and is assigned to the charge transfers (i) from the oxygen anions to the zirconium cations (O²⁻ → Zr⁴⁺) in ZrO₂ and (ii) from the oxygen anions to the hafnium cations (O²⁻ → Hf⁴⁺) in HfO₂ (Fan et al., 2017; Qi & Zhou, 2015).

Figure 3 UV-Vis spectra of the 3HfO2·15SiO2·xY2O3·(82 − x)ZrO2 system.

The appearance of the UV absorption band centered at about 265 nm indicates the formation of tetragonal ZrO₂ phase (Balaji et al., 2017). By increasing the Y₂O₃ content, the intensity of this band increases.

Villabona-Leal et al. (2014) consider that the UV-Vis absorption bands around 250 nm and 220 nm are due to different ZrO₂ crystalline phases, namely the monoclinic (the 250 nm band) and tetragonal (the 220 nm) ones.

The UV-Vis feature at 350 nm is intense for the lowest content of Y₂O₃, flattened for the middle content and then disappears at the maximum yttrium content. This evolution can be assigned to the structural changes that occur in the samples. Thus, according to the XRD and FT-IR data, it is possible to associate the compositional evolution of this feature with the increase of the tetragonal ZrO₂ crystalline phase and the decrease of the monoclinic ZrO₂ crystalline phase, respectively.

Using the UV-Vis spectra of the 3HfO₂·15SiO₂·xY₂O₃·(82 − x)ZrO₂ system, we determined the gap energy of the samples. The calculated gap energy values are presented in Fig. 4. The values for the direct gap energy are between 4.87 eV and 5.05 eV, whereas those for the indirect gap energy are between 5.12 and 5.14 eV. Both direct and indirect gap energy values decrease almost linearly with increasing the Y₂O₃ content of the samples. Differences in the band gaps suggest different distributions and densities of the electronic states. The smaller values of the gap energy for the ceramic nanomixtures indicate smaller average crystallite sizes, in agreement with the SEM data.

Figure 4 (αhν)^1/2 and (αhν)^2 as a function of photon energy for the 3HfO2·15SiO2·xY2O3·(82 − x)ZrO2 system; and the dependence of the optical band gap for the 3HfO2·15SiO2·xY2O3·(82 − x)ZrO2 system.

3.4. PL spectra

Pure zirconium and hafnium oxide samples containing only monoclinic phases show broadband PL emissions situated at about 490 nm. The microscopic origin of these PL bands is still disputable (Kiisk et al., 2016).

The PL spectra of the 3HfO₂·15SiO₂·xY₂O₃·(82 − x)ZrO₂ samples with x = 2, 7 and 12 mol% Y₂O₃ are shown in Fig. 5. These spectra exhibit a broad band situated between 300 and 750 nm which is indicative of a white emission. This broad band consists of four PL bands of higher emission centered at about 335, 410, 450 and 470 nm and of two weak PL bands located at about 530 and 650 nm. The above-mentioned emission bands can be associated with structural defects such as oxygen vacancies in monoclinic ZrO₂ and monoclinic HfO₂ structures.

Figure 5 Photoluminescence spectra of the samples in the 3HfO2·15SiO2·xY2O3·(82 − x)ZrO2 composition where x = 2, 7 and 12 mol% Y2O3.

The high-intensity PL band located at about 410 nm and the bands centered at about 530 and 650 nm can be caused by interstitial defects such as oxygen vacancies in HfO₂ and ZrO₂ crystalline phases with monoclinic structure when Hf and Zr atoms are seven-coordinated. The O atoms can occupy the threefold and fourfold positions and the average Hf—O bond length is 2.08 and 2.20 Å, respectively (Chuang et al., 2012). Therefore, the fourfold-coordinated O atom around the four adjacent Hf atoms is more susceptible to forming a vacancy. The stronger PL emission situated at ∼530 nm can be due to a fourfold-coordinated O atom in the HfO₂ structure.

As mentioned, the PL spectra of the 3HfO₂·15SiO₂·xY₂O₃·(82 − x)ZrO₂ samples are related to structural oxygen vacancies. The intensity of the PL emission bands increases with increasing Y₂O₃ content of the samples. This behavior can be explained by considering that the small particle size was the main reason for the broad photoluminescence band.

3.5. XANES data

XANES measurements in fluorescence mode were performed using the X-ray energy of the incident flux in the range 9400–10400 eV. The L₃-edge of hafnium is located at about 9562 eV.

Zirconium and hafnium oxides are known to produce similar phase transitions under temperature and pressure variation. Under normal conditions of temperature and pressure, both oxides are in the monoclinic phase because the zirconium and hafnium cations have the same valence configuration with very similar ionic radii (0.78 Å for Zr⁺⁴ and 0.76 Å for Hf⁺⁴). Since the crystallographic structures of the monoclinic phase of zirconium and hafnium are very close, many of the XRD peaks arise at nearly the same angles (Benyagoub, 2012). The transition temperature from the tetragonal to the monoclinic phase is higher in the case of HfO₂ (∼1700°C) than for ZrO₂ (∼1200°C). Note also that the Y⁺³ ions have radii (0.96 Å) larger than those of Zr⁺⁴ or Hf⁺⁴ ions and thus their insertion in the host matrix will produce lattice distortion and the elongation of the bonds with the oxygen ions.

Fig. 6 shows the XANES spectra for the Hf L₃-edge from 9525 to 9700 eV for the zirconate nanostructures in the 3HfO₂·15SiO₂·xY₂O₃·(82 − x)ZrO₂ composition with x = 2, 7 and 12 mol% Y₂O₃. The second derivative Hf L₃-edge XANES spectra are also presented in the lower graph of Fig. 6 and permit the variation in the energy to be distinguished.

Figure 6 Hf L3-edge XANES spectra of the studied samples in the 3HfO2·15SiO2·xY2O3·(82 − x)ZrO2 system where x = 2, 7 and 12 mol% Y2O3. The lower graph shows the second derivative Hf L3-edge XANES spectra.

The XANES features at the Hf L₃-edge of the studied composites show a strong main peak centered at about 9562 eV (feature A) corresponding to the Hf⁺⁴ ions. In the post-edge region, a small peak located at about 9578 eV (feature B) and a broad peak situated around 9604 eV (feature C) were found.

The analysis of the XANES and second derivative spectra of the Hf L₃-edge indicates that the Y₂O₃ content of samples do not affect the edge position of the peaks that remains the same for all the studied ceramics but influences the peaks' intensity. Thus, the intensity of the main peak (feature A) was the highest for sample x = 2, decreases for x = 7 mol% Y₂O₃ and suffers a new increase for x = 12 mol% Y₂O₃. This variation of the peak intensity can be related to the different oxygen contents of the samples. The above-mentioned XANES feature (feature A) is located at about 9562 eV and corresponds to a dipole 2p → 5d transition due to the empty 5d states and the first EXAFS oscillation. The intensity of this feature depends on the filling of the d band of hafnium ions. In this view, note that the Hf⁺⁴ ions (5d⁰6s⁰) have an empty 5d state while the Hf⁺³ ions (5d⁰6s¹) are partially filled. For the sample with x = 7 mol% Y₂O₃, the peak energy and intensity of the white lines are smaller than those in the case of the samples with x = 2 and 12 mol% Y₂O₃, suggesting a smaller number of 5d valences.

In conclusion, the Hf XANES data do not show significant differences with increasing Y₂O₃ content of the samples suggesting an appreciable stability of the hafnium ions' +4 oxidation state and their microvicinity.

3.6. EXAFS

To verify the origin of the XANES features at the Hf edge, the spectra corresponding to the monoclinic HfO₂ structure were simulated.

The potential structural modifications and local structural changes caused to a host matrix by the addition of a dopant can be studied by EXAFS analysis and theoretical fitting procedures (Rada et al., 2016). The EXAFS function was derived from the absorption data via the pre-edge and post-edge subtraction. The magnitude of the Fourier transform of EXAFS oscillations at the Hf L₃-edge of the studied ceramics is shown in Fig. 7. The EXAFS investigation detected three coordination shells around the hafnium ions for all the samples, namely Hf—O, Hf—Hf and a second Hf—Hf coordination shell at farther distances. The coordination numbers (N), average inter-atomic distances (R) and Debye–Waller disorder factors (σ²) of the first coordination shell (Hf—O) around the hafnium ions are shown in Table 1.

Table 1
EXAFS structural parameters for hafnium in the studied samples

N: model coordination number of the first coordinate sphere; R: the average Hf—O interatomic distance; σ²: parameter of Debye–Waller thermal disorder (mean squared displacement); ΔE: edge energy; ΔR: shift of bonding distance between atoms.

Sample	N	R (Å)	σ^2	ΔE (eV)	ΔR (Å)
x = 2%Y2O3	7	2.13041	0.00794	7.506	0.03041
x = 7%Y2O3	7	2.12626	0.00799	7.154	0.02626
x = 12%Y2O3	7	2.12059	0.00792	6.809	0.02059

Figure 7 Magnitude of the Hf L3-edge Fourier transform of the EXAFS oscillation for the studied samples in the 3HfO2·15SiO2·xY2O3·(82 − x)ZrO2 composition where x = 2, 7 and 12 mol% Y2O3.

The Fourier transform of the EXAFS signals at the Hf L₃-edge of all samples shows two major oscillations. The first oscillation situated in the 0.8–2.2 Å range corresponds to the nearest neighbors around the hafnium and oxygen atoms. This main peak centered at about 1.7 Å indicates the distributions of oxygen atoms around hafnium atoms and is representative of the local structure order of Hf—O bonds in the first shell. The first coordination shell is almost uninfluenced by varying the Y₂O₃ content from x = 2 to 7 mol% suggesting that the local order of Hf—O bonds is similar in both samples. After that, the increase of the Y₂O₃ content up to x = 12 mol% leads to the increase of the peak related to the Hf—O bonds suggesting an increase of the average oxidation number of hafnium (Cho et al., 2017).

The coordination number and the bond length of the first Hf—O shell are 8 and 2.22 Å, respectively, for the cubic HfO₂ structure (Qi et al., 2013). The EXAFS results presented in Table 1 show that in our samples the coordination numbers are around 7 and the bond lengths are ∼2.12 Å that correspond to the monoclinic HfO₂ structure. The lower coordination number and the bond length are due the charge-compensating oxygen vacancies induced around the Hf atoms by addition of higher Y₂O₃ contents.

The local structural parameters presented in Table 1 show a weak compositional variation. By increasing the Y₂O₃ content of the samples the EXAFS first coordination shell distances (Hf—O distances) decrease from 2.13041 to 2.12059 Å. This evolution suggests that in sample x = 2 mol% Y₂O₃ the local environment around the hafnium ions in the first coordination shell presents a lower order in comparison with the other samples.

The second major oscillation is situated in the 2.5–3.7 Å range and is related to the next nearest neighbors such as Hf—Hf or Hf–metals paths (Sharma et al., 2017). The other higher oscillations are due to the complex O–metal–O kinds of multiple scattering paths. Note that there is a significant difference in the second shell features. Thus, the fit of the second oscillation of the EXAFS signal shows that this oscillation consists of two path contributions with different lengths situated at ∼2.9 and ∼3.05 Å. The first contribution is related to the Hf—Zr (2.89 Å) and Hf—Hf (2.88 Å) bonding distances while the second contribution is associated with the Hf—Y bonding distance (3.06 Å). With increasing the yttrium content up to x = 12 mol% Y₂O₃, the last peak gradually intensifies and shifts towards higher R values. This evolution depicts an increase in the Hf–metal bond length due to the enhancement of the crystallization process of the samples.

These compositional evolutions of the EXAFS data can be regarded as an indicator of the degree of crystallization at the hafnium local environment. For the sample with x = 1 mol% Y₂O₃, the contributions from external shells are more structured in comparison with those of the samples with x = 7 and 2 mol% Y₂O₃. The differences observed for the EXAFS second coordination shell can be caused by the partial replacements of hafnium atoms by another metal atom as the second nearest neighbor of a hafnium atom, such as zirconium and yttrium atoms.

In conclusion, the EXAFS data suggest that the coordination number of the hafnium ions is around 7 and the Hf—O bond length is ∼2.12 Å, that are characteristic of the monoclinic HfO₂ structure. The contributions from the second coordination shells attributed to the Hf–metal paths become more structured by increasing the Y₂O₃ content in the host ceramic matrix.

3.7. SEM images

SEM micrographs of the samples in the 3HfO₂·15SiO₂·xY₂O₃·(82 − x)ZrO₂ system with x = 2, 7 and 12 mol% Y₂O₃ are shown in Fig. 8. Due to the increase of the Y₂O₃ content, the samples x = 12 mol% will reach a more dense microstructure with smaller crystallite sizes. The high-magnification SEM image reveals that the structure of this sample consists of many domains with small and fine grains. The structural evolution of samples determined by increasing their Y₂O₃ content is in agreement with the XRD data. As was previously shown, the increase of the yttrium content determines the occurrence of yttrium–silicates precipitates (Y₂SiO₅ and Y₂Si₂O₇ crystalline phases) that are distributed on the surface of zirconate samples.

Figure 8 SEM images of the studied samples in the 3HfO2·15SiO2·xY2O3·(82 − x)ZrO2 composition with x = 2, 7 and 12 mol% Y2O3.

3.8. Raman spectroscopy

Characterization of zirconia systems by XRD may produce difficulties concerning the presence or absence of certain crystalline phases due to the highly broadened X-ray peaks. The Raman spectroscopy technique can be useful in order to help the identification of zirconia polymorphs (Ghosh et al., 2006; Hartmanova et al., 1998).

The Raman spectra of the 3HfO₂·15SiO₂·xY₂O₃·(82 − x)ZrO₂ samples where x = 2, 7 and 12 mol% Y₂O₃ are shown in Fig. 9. The Raman bands characteristic of the monoclinic ZrO₂ phase are those located at 180, 220, 225, 300, 335, 380, 470, 535, 555, 617 and 635 cm⁻¹ (Suya Prem Anand et al., 2018). The tetragonal ZrO₂ phase is characterized by Raman bands centered at about 150, 255, 295, 306 and 430 cm⁻¹. Thus, the Raman spectroscopic study of the studied zirconate samples offers a clear evidence of the presence of both monoclinic and tetragonal ZrO₂ structure in the prepared samples. The intensity of the Raman band centered at about 255 cm⁻¹ increases in the sample with x = 12 mol% Y₂O₃ suggesting the monoclinic → tetragonal phase transformation determined by the successive addition of Y₂O₃ to the host matrix of samples.

Figure 9 Raman spectra of the 3HfO2·15SiO2·xY2O3·(82 − x)ZrO2 samples with x = 2, 7 and 12 mol% Y2O3.

The complex investigations of the ceramic nanomixtures (composites) were useful tools for distinguishing the zirconia and hafnia phases and the microstructural modifications at high dopant levels. The experiments reported in this paper demonstrate the structural modifications occurring in samples by doping.

The results show that yttrium as a dopant can be successfully used in order to modify structural and behavioral properties in a desired way. The studied materials can be used for optical applications after their gap band energies are lowered below 3 eV by appropriate procedures.

In order to show the applicability of these materials in biomedicine, investigations concerning their mechanical properties and chemical stability must complete the study.

4. Conclusions

Zirconate nanostructured samples of 3HfO₂·15SiO₂·xY₂O₃·(82 − x)ZrO₂ composition where x = 2, 7 and 12 mol% were prepared and studied by means of X-ray diffraction, X-ray absorption spectroscopy, SEM analysis, photoluminescence, FT-IR, Raman and UV-Vis spectroscopy. XRD and XAS techniques were used to identify the crystalline phases from the studied samples. Thus, the presence of the following crystalline phases was observed: t-ZrO₂, m-ZrO₂, m-HfO₂, Y₂SiO₅ and Y₂SiO₅. SEM investigation highlighted the presence of Y₂SiO₅ and Y₂Si₂O₇ crystalline phases, in agreement with XRD data.

The photoluminescence emission bands are associated with structural defects such as the oxygen vacancies in monoclinic ZrO₂ and monoclinic HfO₂ structures. Based on the UV-Vis spectroscopy data, the gap energy of the samples was determined. The compositional evolution of gap energy values shows that the increase of the xY₂O₃ content of the samples led to the enhancement of the crystallization process in agreement with the XRD data. Smaller values of the gap energy can also be related to smaller average crystallite sizes, in agreement with the SEM data.

FT-IR and Raman spectroscopies were useful tools for distinguishing the zirconia phases. Thus, it was observed that the increase of the Y₂O₃ content of samples led to an increase of the t-ZrO₂ and silicate crystalline phases. The structural changes observed by the FT-IR and Raman investigations suggest that the yttrium ions play a network modifier role in the studied system.

The local structure around the Hf cations from the studied ceramic nanomixtures was studied by analysing the Hf L₃-edge EXAFS spectra obtained at room temperature. The EXAFS analysis suggests smaller deviations in bond length sets of Hf—O shells with increasing Y₂O₃ content in the host ceramic matrix. The contributions from the second coordination shells attributed to the Hf–metal paths are more structured by increasing the Y₂O₃ content in studied ceramics.