Effects of high mechanical treatment and long-term annealing on crystal structure and thermal stability of Ti₂O₃ nanocrystals

Abstract

The effect of high-energy milling and long-term annealing on the stability of Ti₂O₃ nanocrystals was studied using a magnetic susceptibility method. In situ temperature dependences revealed that the crystal size greatly affects the magnetic susceptibility value. According to XRD, SEM and TEM data, Magnéli phases Ti₉O₁₀, Ti₄O₇, Ti₇O₁₉ and Ti₃O₅ are formed.

---

The titanium–oxygen (Ti–O) system is currently important from both scientific and applied viewpoints.¹ Titanium compounds depending on nonstoichiometry^2–5 are a promising functional material for a wide range of applications, such as a capacitors, photocatalysts for the degradation of organic pollutants in air and water, promising functional materials for renewable energy sources (e.g., solar batteries, photochemical water decomposition and hydrogen generation devices), efficient photoelectric converters, memristor memory elements, etc.^6–12 Titanium oxide Ti₂O₃ with unusual properties has a narrow homogeneity region from TiO_1.49 to TiO_1.51.¹³ There is a d-metal–semiconductor electronic phase transition at 420–550 K without any change in the symmetry and crystal structure. The strong deformation of the crystal lattice with increasing temperature caused the band gap closure.^14–16 Also, Ti₂O₃ exhibits thermoelectrical properties,¹⁷ and doped Ti₂O₃ films have high negative magnetoresistance.¹⁸

The aim of this work is to study the effect of mechanical treatment (high-energy milling) and subsequent high-temperature annealing in a vacuum on the crystal structure and thermal stability of Ti₂O₃ nanocrystals in the temperature range from 300 to 1200 K with the use of a magnetic susceptibility method.

The initial titanium(III) oxide Ti₂O₃ microcrystals with corundum structure (sp. gr. R3̄c) with an average size of about 25 μm were synthesized by solid phase sintering from a mixture of Ti and TiO₂ powders in a vacuum of 10⁻³ Pa at 1770 K. Titanium(III) oxide Ti₂O₃ nanocrystals have been obtained by high-energy milling of Ti₂O₃ microcrystals in a Retsch PM 200 planetary ball mill. The mass ratio of grinding balls made of zirconium dioxide (ZrO₂) stabilized with yttrium oxide (Y₂O₃) to Ti₂O₃ powder in the experiment was 10 : 1. Isopropyl alcohol was used as a grinding liquid. The following milling regime was employed to produce nanopowders: the duration of milling was 15, 30, 60, 120, 240 and 480 min, reversal of rotation direction was each 15 min, the interval between rotation direction reversal was 5 s, the rotation velocity of milling pot backing plate was 500 rpm.

X-ray phase analysis of all powders was performed in CuKα_1,2-radiation on a Shimadzu XRD-7000 diffractometer in Bragg–Brentano geometry in stepwise scanning mode with Δ(2θ) = 0.02° in 2θ angle interval from 10 to 120°. To identify phases, the powder diffraction database ICDD, USA, Release 2016 was used. The phases were analyzed with the use of the Powder Cell 2.4 program. For a full-profile description of X-ray diffraction reflections, the pseudo-Voigt function was used. The diameter of nanocrystals D was determined from diffraction reflection broadening with the Williamson–Hall method.^19,20 The CSR size was obtained by extrapolating the dependence β*(s) to the value s = 0, and the magnitude of microstrain was determined from the slope of this dependence.^21,22

The microstructure of nanocrystals was studied with the high-resolution scanning electron microscopy (SEM) on a ZEISS Ultra 55 microscope. The working distance (WD) was 3.9–4.3 mm, the electron high tension (EHT) was 3–5 keV, and the beam width ranged from 2 to 6 μm depending on magnification. In order to avoid excessive electrization of powder during electron microscope imaging, the examined powder was deposited on a conducting adhesive tape and then was covered with a chromium layer of about 2 to 4 nm thick; the chromium coating did not affect the quality of visualization of powder morphology.

The structure of titanium oxide nanocrystals was determined by using the high-resolution transmission electron microscopy (HRTEM) on a JEM 2010 electron microscope (JEOL, Japan) with accelerating voltage of 200 kV and ultimate lattice resolution of 140 pm. Imaging was performed by means of CCD matrix of Soft Imaging System (Germany). The device was equipped with a Phoenix (EDAX, USA) energy-dispersive characteristic X-ray radiation (EDX) spectrometer with a semiconducting Si(Li)-detector with energy resolution of 130 eV. Ti₂O₃ particles were placed into alcohol and were further deposited on perforated carbon substrates (diameter of holes of about 1 μm) fixed on copper grids. Particles were deposited with the use of a UZD-1UCh2 ultrasonic disperser, which allowed uniform particle distribution on the substrate surface. After the grids were extracted from alcohol, the alcohol evaporated.

Thermal stability and phase transformations with long-term exposure for complete running of processes in the system and structure stabilization were studied with the analysis of magnetic susceptibility variation by using the Faraday method on a pendulum magnetic Domenikalli-type balance in vacuum of about 10⁻³ Pa.²³ Magnetic susceptibility χ of Ti₂O₃ was measured in the temperature interval from 300 to 1200 K in magnetic fields with intensity from 7.2 to 8.8 kOe. The heating and cooling rate of the samples during susceptibility measurements was about 1 K min⁻¹. The powder mass and the crystal structure of Ti₂O₃ were controlled before and after χ measurements. The accuracy of χ measurements was about ±0.05 × 10⁻⁶ emu g⁻¹. The absence of ferromagnetic impurities in the powders was confirmed by measurements at different magnetic field values.

Detailed analysis of X-ray diffraction patterns of ball milled Ti₂O₃ (Fig. 1a) showed that the crystal structure of nanopowder coincides with the crystal structure of microcrystal (Fig. 1a), i.e. the structure of Ti₂O₃ is highly stable with respect to high-energy milling, and fragmentation does not lead to changing of crystal symmetry. After high-energy milling, broadening of reflections is observed on the X-ray diffraction patterns, which is related to a small grain size and the presence of microstrains in the system due to high-energy milling. The full-profile analysis of X-ray diffraction reflections showed a decrease in coherent scattering region (CSR) from 25 μm to 10 nm with an increase of milling time from 15 to 480 minutes, while the microstrain increase up to 0.26%.

Fig. 1 XRD pattern of powders at room temperature: (a) as prepared microcrystals and ball milled nanocrystals; after magnetic susceptibility measurements: (b) at 673 K; (c) at 873 K; (d) at 1200 K.

Analysis of SEM micrographs of ball milled Ti₂O₃ nanocrystals revealed that the particles are uniformly distributed in the bulk and have a granular or rounded shape (Fig. 2). It was found that small granules of about 10–30 nm in size adhere to each other and form agglomerates. Fig. 3 displays a HRTEM image of the ball milled Ti₂O₃ nanocrystal. According to HRTEM data, the sample consists mainly of crystallites of sizes from 50 to 500 nm (Fig. 3a).

Fig. 2 SEM image of ball milled Ti₂O₃. Small granules of about 10 to 30 nm in size adhere to each other and form agglomerates.

Fig. 3 HRTEM images of Ti₂O₃ nanocrystals: (a) the sample consists mainly of crystallites of sizes from 50 to 500 nm; (b) morphology of milled nanocrystalline Ti₂O₃, the measured d_hkl correspond to the Ti₂O₃ phase (PDF 00-010-0063).

According to the observed d_hkl (Fig. 3b), this phase corresponds to Ti₂O₃ (PDF 00-010-0063). Thus, the size of particles found with SEM and HRTEM methods coincide, which indicates that the size of CSR of particles determined from XRD data is accurate enough.

The experimental data obtained earlier show that the magnetic susceptibility value of nanocrystals is twice as small in absolute magnitude as that of microcrystals.²⁴ The crystal size affects greatly not only the value of magnetic susceptibility, but also the temperature behaviour of magnetic susceptibility. The temperature dependence of magnetic susceptibility for Ti₂O₃ microcrystals has a classical form of s-shaped curve typical of first-order phase transitions without considerable hysteresis. Fig. 4 shows in situ temperature dependences of magnetic susceptibility for Ti₂O₃ nanocrystals. In the temperature region from 300 to 400 K, the structure of initial Ti₂O₃ nanocrystal remains trigonal (sp. gr. R3̄c); the temperature dependence exhibits a reverse behavior of magnetic susceptibility, which shows that the state of nanocrystals in this temperature region is metastable. Annealing of Ti₂O₃ nanocrystals at temperature above 400 K leads to phase transformations and, as a result, to magnetic susceptibility enhancement. X-ray diffraction analysis shows that after annealing to 673 K the powder contains additional phases of Ti₉O₁₀ (sp. gr. Immm) – 3 mass% and Ti₉O₁₇ (sp. gr. I1̄) – 11 mass% alongside with Ti₂O₃ phase (sp. gr. R3̄c) in the amount of 86 mass% (Fig. 1b). Further annealing and increasing annealing temperature to 873 K result in variation of phases and phase proportions. X-ray diffraction analysis (Fig. 1c) also shows that the powder contains Ti₂O₃ (sp. gr. R3̄c) – 38 mass%, Ti₉O₁₀ (sp. gr. Immm) – 2 mass%, Ti₄O₇ (sp. gr. A1̄) – 60 mass%.

Fig. 4 In situ temperature dependences of magnetic susceptibility for Ti₂O₃ nanocrystals. The left-right arrow indicates the reverse behavior of magnetic susceptibility. The figures denote the magnetic susceptibility values at room temperature in different states: I – initial state; after magnetic susceptibility measurements: II – at 673 K, III – at 873 K, IV – at 1200 K.

It is noteworthy that the Ti₉O₁₀ phase (sp. gr. Immm), which was first obtained during annealing of nonstoichiometric TiO_y nanocrystal and annealing of TiO_y/HAP nanocomposite,^25–27 is formed during annealing of only titanium oxide nanocrystals with superstoichiometric composition. According to quantum chemical calculations, the Ti₉O₁₀ phase (sp. gr. Immm) in a microcrystalline state is unfavorable in comparison with the disordered cubic phase of the same TiO_9/10 composition, therefore its formation is due to size effects.^28,29

When the annealing temperature rises, the stable Ti₃O₅ with monoclinic structure (sp. gr. I2/c) is formed (Fig. 1d). Judging by the temperature dependence of magnetic susceptibility (Fig. 4), the formation of the Ti₃O₅ phase occurs at about 1073 K. At further heating from 300 to 1200 K and cooling from 1200 to 300 K, a reverse behavior of magnetic susceptibility is observed, which indicates that the system reaches the equilibrium state. Thus, the experimental results show that the phase stability and phase transitions in Ti₂O₃ are greatly affected above all by the crystal sizes.

Fig. 5 demonstrates the HRTEM micrographs of nanocrystalline powder after annealing experiments. The sample consists of plate-like particles 10 nm to 1 μm in size (Fig. 5a). Regions with well crystallized structure are observed where the dimensions of blocks are 100 nm. According to the observed interplanar spacings, the phase formed corresponds to the Ti₃O₅ phase (Fig. 5b) with monoclinic structure (sp. gr. I2/c) (PDF 01-076-1066).

Fig. 5 HRTEM images of Ti₃O₅ nanocrystals: (a) the sample consists of plate-like particles 10 nm to 1 μm in size; (b) well crystallized structure, dimensions of blocks 100 nm, the observed interplanar spacings d_hkl correspond to Ti₃O₅ phase (PDF 01-076-1066).

Thus, Ti₂O₃, subjected to mechanical treatment (high-energy milling) and heat treatment in the temperature range from 300 to 1200 K in vacuum, is an unstable phase since it changes during thermal treatment. According to the X-ray diffraction analysis, SEM and HRTEM data, the Ti₉O₁₀ phase and Magnelli phases Ti₄O₇, Ti₇O₁₉ and Ti₃O₅ are formed depending on the annealing temperature. In addition, the in situ temperature dependences of magnetic susceptibility showed that the crystal size greatly affects the value of magnetic susceptibility; the magnetic susceptibility value of nanocrystals is twice as small in absolute magnitude as that of microcrystals.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The reported study was funded by RFBR according to the research project no. 19-03-00051a.

References

1. X. Chen and S. S. Mao, Chem. Rev., 2007, 107, 2891.
2. A. A. Valeeva, S. Z. Nazarova and A. A. Rempel, Phys. Solid State, 2016, 58(4), 771.
3. A. A. Valeeva and A. A. Rempel, Mendeleev Commun., 2010, 20, 101.
4. A. A. Valeeva, A. A. Rempel and A. I. Gusev, Dokl. Phys., 2002, 47, 39.
5. M. G. Kostenko, A. A. Valeeva and A. A. Rempel, Mendeleev Commun., 2012, 22, 245.
6. V. Schöllmann, J. Johansson, K. Andersen and D. V. Haviland, J. Appl. Phys., 2000, 88, 6549.
7. A. A. Rempel, E. A. Kozlova, T. I. Gorbunova, S. V. Cherepanova, E. Y. Gerasimov, N. S. Kozhevnikova, A. A. Valeeva, E. Y. Korovin, V. V. Kaichev and Y. A. Shchipunov, Catal. Commun., 2015, 68, 61.
8. A. A. Valeeva, E. A. Kozlova, A. S. Vokhmintsev, R. V. Kamalov, I. B. Dorosheva, A. A. Saraev, I. A. Weinstein and A. A. Rempel, Sci. Rep., 2018, 8, 9607.
9. A. A. Valeeva, I. B. Dorosheva, E. A. Kozlova, R. V. Kamalov, A. S. Vokhmintsev, D. S. Selishchev, A. A. Saraev, E. Y. Gerasimov, I. A. Weinstein and A. A. Rempel, J. Alloys Compd., 2019, 796, 293.
10. D. Recatalá, R. Llusar, A. L. Gushchin, E. A. Kozlova, Y. A. Laricheva, P. A. Abramov, M. N. Sokolov, R. Gómez and T. Lana-Villarreal, ChemSusChem, 2015, 8, 148.
11. N. S. Kolobov, D. A. Svintsitskiy, E. A. Kozlova, D. S. Selishchev and D. V. Kozlov, Eng. J., 2017, 314, 600.
12. A. S. Vokhmintsev, I. A. Weinstein, R. V. Kamalov and I. B. Dorosheva, Bull. Russ. Acad. Sci.: Phys., 2014, 78, 932.
13. S. Andersson, B. Collén, U. Kuylenstierna and A. Magnéli, Acta Chem. Scand., 1957, 11, 1641.
14. H. Fan, M. Wang, Z. Yang, X. Ren, M. Yin and S. Liu, Appl. Phys. A: Mater. Sci. Process., 2016, 122, 964.
15. J. Wang, Y. Li, L. Deng, N. Wei, Y. Weng, S. Dong, D. Qi, J. Qiu, X. Chen and T. Wu, Adv. Mater., 2016, 29, 1603730.
16. A. I. Poteryaev, A. I. Lichtenstein and G. Kotliar, Phys. Rev. Lett., 2004, 93, 086401.
17. I. Veremchuk, I. Antonyshyn, C. Candolfi, X. Feng, U. Burkhardt, M. Baitinger, J. T. Zhao and Y. Grin, Inorg. Chem., 2013, 52, 4458.
18. Z. Wang, Y. Hong, J. Tang, C. Radu, Y. Chen, L. Spinu, W. Zhou and L. D. Tung, Appl. Phys. Lett., 2004, 86, 7384.
19. W. H. Hall, Proc. Phys. Soc., London, Sect. A, 1949, 62, 741.
20. G. K. Williamson and W. H. Hall, Acta Metall., 1953, 1, 22.
21. A. A. Valeeva, H. Schroettner and A. A. Rempel, Russ. Chem. Bull., 2014, 63, 2729.
22. A. A. Valeeva, S. Z. Nazarova and A. A. Rempel, Phys. Status Solidi B, 2016, 253, 392.
23. A. A. Valeeva, S. Z. Nazarova and A. A. Rempel, JETP Lett., 2015, 101, 258.
24. A. A. Valeeva, S. Z. Nazarova and A. A. Rempel, J. Alloys Compd., 2020, 817, 153215.
25. S. V. Rempel, A. A. Valeeva, E. A. Bogdanova, H. Schroettner, N. A. Sabirzyanov and A. A. Rempel, Mendeleev Commun., 2016, 26, 543.
26. S. V. Rempel, D. A. Eselevich, E. Y. Gerasimov and A. A. Valeeva, J. Alloys Compd., 2019, 800, 412.
27. S. V. Rempel, E. A. Bogdanova, A. A. Valeeva, H. Schroettner, N. A. Sabirzyanov and A. A. Rempel, Inorg. Mater., 2016, 52, 476.
28. A. A. Valeeva, M. G. Kostenko, S. Z. Nazarova, E. Y. Gerasimov and A. A. Rempel, Inorg. Mater., 2018, 54, 568.
29. M. G. Kostenko and A. A. Valeeva, Mendeleev Commun., 2019, 29, 405.

---