Shock synthesis and characterization of titanium dioxide with α-PbO₂ structure

Titanium dioxide and related materials have attracted considerable attention worldwide due to its specific commercial application in environmental photocatalysis, solar cells, gas sensors, etc [1–4]. In addition, the study of high-pressure phases of titanium dioxide is of great geophysical significance [5–8]. The experimental and theoretical studies revealed that there could exist many different structured polymorphs of titanium dioxide under different pressure and temperature conditions [9]. At ambient condition, titanium dioxide exists mainly in the form of anatase, brookite and rutile, which were reported to be the most stable and abundant in nature [5, 10]. Under certain high-pressure high-temperature conditions, various types of high-pressure modifications with excellent electronic and mechanical properties may occur [10–12], for example α-PbO₂, baddeleyite, OI, cotunnite, fluorite or distorted fluorite-like cubic phase [10, 12–20] and other post-cotunnite phases [21]. However most of the high-pressure polymorphs reverse to α-PbO₂ phase during the decompression process. α-PbO₂ phase is one of the high-pressure polymorph which can be quenched at ambient condition in addition to the baddeleyite phase TiO₂ reported by Goresy et al [8].

As the common phases of titanium dioxide, the powder or single crystals of anatase or rutile TiO₂ were often selected as the starting materials to study the phase transition behavior of titanium dioxide under static or shock compression conditions. Recently, Nishio-Hamane et al [18] revealed that the phase transition sequence under static high pressure was rutile  →  α-PbO₂ phase  →  baddeleyite phase  →  OI phase  →  cotunnite phase with increasing pressure. This is consistent with the sequence reported by Dubrovinskaia et al [20], and they also suggested that the cotunnite phase was the most stable phase under high pressures ranging from 40 GPa to at least 70 GPa. The cotunnite phase was reported to be formed at pressures above 60 GPa and temperatures above 1000 K under static compression condition by Dubrovinsky et al [12] and Ahuja et al [22], and then the cotunnite phase TiO₂ was quenched below 170 K. The mechanical properties measured indicate that the cotunnite phase TiO₂ is the hardest dioxide so far discovered. However, the results conflicted with the results reported by Nishio-Hamane et al [18], so the hardness of the cotunnite phase TiO₂ is still controversial. In 2004, Mattesini et al [17] reported the coexistence of the known cotunnite phase and a new cubic phase TiO₂ under static compression conditions. However the new cubic phase was hard to quench during the decompression process, and they also suggested that high temperature was required to synthesize the new cubic phase TiO₂. In 2011, Dekura et al [21] reported the discovery of a new post-cotunnite phase of TiO₂ both by theoretical and experimental studies under high pressure, but when the pressure was released, the new phase converted to the normal phase. Despite the fact that the static compression conditions are easy to control and the loading time can vary from microseconds to several days, the amount of the sample treated is too little, and the high-pressure modifications are always hard to quench due to the low cooling rate.

The shock compression technology has attracted considerable attention worldwide and has been successfully used to synthesize various new materials due to its unique advantages, such as high pressure, high temperature, high quenching rate and a larger sample volume compared with static compression method. In the past few decades, some techniques have been developed to control the pressures and temperatures during shock compression to synthesize cubic diamond, cubic born nitride and cubic silicon nitride [23–26]. In 1967, the shock Hugoniot data for c-axis direction were measured by the flash-gap method by McQueen et al [16], and ones for unknown axis done by the pin contactor up to  >200 GPa by Alt'shuler et al [27]. They found the phase transition to α-PbO₂ phase at about 33 GPa. Linde and DeCarli [28] confirmed this transition by the pressure measurement using manganin-gauge method and the recovery experiment. X-ray diffraction (XRD) patterns of the recovered samples demonstrated the presence of α-PbO₂ structured high pressure polymorph, and the maximum transformation rate for rutile-type single crystal was about 90%. For the powder samples, the transformation rate was no more than 30%. Anisotropic phase transitions were first measured for 〈1 0 0〉, 〈1 1 0〉 and 〈0 0 1〉 axis directions by the electro-magnetic-gauge method by Mashimo et al [29]. Then, Syono et al [30] measured them up to  >100 GPa by the inclined-mirror method. They revealed that the phase transition pressure of rutile single crystal to α-PbO₂ phase TiO₂ depended strongly on the shock wave propagation direction, e.g. 13.7 GPa for [1 0 0], 16.9 GPa for [1 1 0] and 33.8 GPa for [0 0 1], respectively. They also observed two new phases, which were named as the HPP-I and HPP-II, for pressures from 70 GPa to 100 GPa. In this work, they did not perform the recovery experiment except the Hugoniot measurement experiment in this paper. Later, they performed the recovery experiments of single crystals [14], and the maximum yield of α-PbO₂ phase TiO₂ was about 70% for rutile single crystal parallel to [1 0 0] direction. In 1984, Morosin et al used a mixture of anatase and rutile TiO₂ powder as the starting material to study the phase transition behavior under shock compression conditions for pressure range from 20 GPa to 27 GPa, but none of high pressure polymorphs was observed [31]. Compared with considerable work conducted for rutile TiO₂, few literatures can be found for shock compression of anatase TiO₂. Molodets et al [32] and Shul'ga et al [33] selected pure anatase TiO₂ as the starting material to study the phase transition behavior under shock compression, and found that the anatase particles would either transform into α-PbO₂ phase or exhibit amorphization. The occurrence of α-PbO₂ phase TiO₂ in nature due to the high-pressure endogenic geological processes or bolide impact events was reported by Hwang et al [34] for the first time, and later by Goresy et al [13], Chen et al [6] and Shen et al [7] and recently by Smith et al [35]. In addition to α-PbO₂ phase TiO₂, Goresy et al [8] reported the discovery of another high-pressure phase of TiO₂ in the shocked rocks with cell parameters of a  =  4.606(2) Å, b  =  4.986(3) Å, c  =  4.933(3) Å, which can be attributed to the baddeleyite-type TiO₂.

Despite many attempts to synthesize various kinds of high pressure polymorphs of titanium dioxide, the transformation rate of α-PbO₂ phase TiO₂ is not very high and pure α-PbO₂ phase TiO₂ is hard to obtain no matter by static or shock-compression method up to date. This is mainly due to the vulnerable reverse phase transformation of α-PbO₂ phase to rutile TiO₂ induced by the high residual temperature after shock unloading. In addition, the α-PbO₂ phase TiO₂ obtained through static or shock induced phase transition has not been well characterized. In this work, the phase transformation behavior of commercial anatase and rutile TiO₂ powder under shock compression is reported, and the structure and thermal stability of the obtained pure α-PbO₂ phase TiO₂ are characterized.

A planar shock compression assembly was used to generate shock pressures with different magnitude. Figure 1 illustrates the shock compression setup. A copper flyer with a diameter of 40 mm and a thickness of 2 mm is accelerated to a high velocity by the detonation of the main explosive charge of nitromethane (CH₃NO₂), initiated by a booster charge of 8701 explosive. The impact velocity calculated according to the Gurney equation [36] approaches 1.7 km s⁻¹ for each shot. Then the copper flyer impacts on the copper sample container, and the generated shock waves act on the powder sample stored in the sample container.

Zoom In Zoom Out Reset image size

Commercial anatase and rutile TiO₂ powder with particle sizes of 60 nm and150 nm were chosen as the starting material. To increase the shock pressure and reduce the shock and residual temperatures, copper powder was mixed with the TiO₂ powder due to its high shock impedance and thermal conductance. To further reduce the porosity and lower the shock temperature and residual temperature, a small amount of paraffin was mixed with the TiO₂ and copper powder. Paraffin can also act as a lubricant to reduce the heat generated by particle friction during shock compression. Then the powder mixture was pressed into the sample container with a pressing pressure of 150 MPa. During the pressing process, excessive paraffin was extruded out and nearly all the interparticle pores between copper and/or TiO₂ particles can be filled by paraffin. A mere mixture of TiO₂ and copper powder without addition of paraffin was also tested for comparison. The equation of state (EOS) of the porous powder mixtures was calculated using the Mie–Grüneisen EOS of the components and a mixture method [36, 37]. The impedance match method [36, 38] was used to calculate the pressure generated in the powder sample. The shock temperature and residual temperature were calculated with consideration of reverberation effects with surrounding copper. In the calculation, the TiO₂ particles were treated as thin layers surrounded by the mixture of copper and paraffin. At the upper and lower interface between the TiO₂ thin layer and surrounding copper mixture, the shock wave reflects and increases the pressure. After multiple reflections, the sample pressure reaches the pressure at the interface of mixture and Cu flyer. The final shock temperature and residual temperature in TiO₂ were approximated by the first wave in the multiple reflection, because the Hugoniot energy increases of the following shock wave are very small. Figure 2 shows the impedance matching technique for determining the first shock-loading states in rutile TiO₂. The shock temperature and the residual temperature after shock unloading were estimated by the thermodynamic method [36], given by

$$\begin{align} \newcommand{\e}{{\rm e}} \displaystyle {{T}_{1}}=~&{{T}_{0}}\exp \left[ \left(\frac{{{\gamma }_{0}}}{{{V}_{0}}}\right)\left( {{V}_{0}}-V \right) \right]+ \frac{\left( {{V}_{0}}-V \right)}{2{{C}_{{\rm v}}}}P+\frac{{\rm exp}\left[ \left( -{{\gamma }_{0}}/{{V}_{0}} \right)V \right]}{2{{C}_{{\rm v}}}} \nonumber \\ &\int_{V_0}^{V}{P\cdot \exp }\left[ \left( {{\gamma }_{0}}/{{V}_{0}} \right)V \right]\left[ 2-\frac{{{\gamma }_{0}}}{{{V}_{0}}}\left( {{V}_{0}}-V \right) \right]{\rm d}V\nonumber \end{align} \tag{ 1 }$$

$$\begin{align} \newcommand{\e}{{\rm e}} \displaystyle {{T}_{2}}={{T}_{1}}{\rm exp}\left[ -\frac{{{\gamma }_{0}}}{{{V}_{0}}}\left( {{V}_{0}}-V \right) \right].\nonumber \end{align} \tag{ 2 }$$

Zoom In Zoom Out Reset image size

Where T₀ is the initial temperature; ${{T}_{1}}$ and ${{T}_{2}}$ are shock temperature and residual temperature, respectively; ${{V}_{0}}$ and V are initial specific volume and specific volume corresponding to the determined shock pressure P, respectively; ${{\gamma }_{0}}$ is the Gruneisen coefficient and ${{C}_{{\rm v}}}$ is the specific heat at constant volume. Table 1 shows the detailed experimental conditions for all the tests, including shock pressure, shock temperature and residue temperature. After the shock experiment, the samples were taken out of the container and immersed into nitric acid to remove the copper powder. To remove the paraffin, the sample was further extracted with low boiling point petroleum ether under reflux. Finally, the purified sample was dried for various characterization.

XRD patterns of all the recovered samples were recorded by operating a Rigaku D/MAX-2500 at a working voltage of 40 kV and a working current of 200 mA with Cu Kα radiation. According to XRD analysis, sample 1324 was selected to further characterize the structure and thermal stability of pure α-PbO₂ phase TiO₂. Raman spectra were recorded on a LabRAM Aramis Raman spectrometer with a He–Ne laser at an excitation wavelength of 514.5 nm. The microstructures were characterized by using a transmission electron microscope (TEM, JEM-2010) with an accelerating voltage of 200 kV. High temperature x-ray diffraction (HTXRD), thermogravimetric analysis (TG) and differential scanning calorimetry (DSC) were used to study its thermal stability. HTXRD analysis were conducted at temperatures from 300 °C to 900 °C under atmospheric pressure and preserved for half an hour at each temperature points. The TG-DSC data were recorded from room temperature to 1200 °C.

3.1. X-ray diffraction (XRD)

Figure 3 shows the XRD patterns of the five recovered samples. The XRD patterns of the original anatase and rutile TiO₂ are given for comparison. For all the recovered samples, the diffraction peaks at approximately 31.4° are assigned to the (1 1 1) plane of α-PbO₂ phase TiO₂ (JCPDS 21-1236), demonstrating the phase transformation of anatase and rutile TiO₂ to α-PbO₂ phase TiO₂ under shock compression. The characteristic peaks of rutile, anatase and α-PbO₂ phase TiO₂ are marked with various symbols respectively. By carefully analyzing the diffraction patterns, the phase contents were calculated according to a formula given in [39], as listed in table 1.

Zoom In Zoom Out Reset image size

The XRD pattern of the sample for anatase TiO₂ and copper powder mixture without the addition of paraffin (shot 373, figure 3(e)) shows that the original anatase TiO₂ is completely transformed to rutile and α-PbO₂ phase TiO₂, and the content of α-PbO₂ phase TiO₂ is only 37.61%. The XRD patterns of the samples for anatase and copper powder mixtures with the addition of paraffin (shots 1322 and 1324, figures 3(f) and (g)) show that all the diffraction peaks can be assigned to α-PbO₂ phase TiO₂, demonstrating that anatase TiO₂ is completely transformed into α-PbO₂ phase TiO₂ and pure α-PbO₂ phase TiO₂ is obtained. The cell parameters calculated according to the XRD patterns are a  =  4.547 Å, b  =  5.507 Å and c  =  4.893 Å, which is in agreement with the results reported by Goresy et al [13]. These results clearly show that addition of a small amount of paraffin plays an important role to quench α-PbO₂ phase TiO₂ during shock compression. Figures 3(c) and (d) show the XRD patterns of shocked rutile and copper powder mixture with addition of paraffin and the content of α-PbO₂ phase TiO₂ is approximately 90%. The transformation rate of α-PbO₂ phase TiO₂ for rutile powder is close to that for rutile single crystal (90%) and much higher than that of the powder samples (30%) reported by Linde et al [28]. Considering the transformation rate of 100 % for anatase powder and 90% for rutile powder under similar shock conditions, we can conclude that anatase TiO₂ is easier to transform to α-PbO₂ phase TiO₂ than rutile TiO₂ under shock compression. If we compare the transformation rates of α-PbO₂ phase TiO₂ for rutile powder with two particle sizes of 60 nm and 150 nm (91.71% for shot 1323 and 90.90% for shot 1321) and complete transformation of α-PbO₂ phase TiO₂ for anatase powder with two particle sizes of 60 nm and 150 nm, we can conclude that the influence of the particle size on the yield of α-PbO₂ phase is not obvious.

3.2. Raman spectra

Figure 4 shows the Raman spectra of α-PbO₂ phase TiO₂ obtained from shot 1324. The Raman spectra of the starting anatase TiO₂ is given for comparison. The Raman spectrum of the anatase TiO₂ (figure 4(a)) displays well defined peaks at 143, 197, 396, 518, 639 cm⁻¹, respectively, which is consistent with the data reported in literatures [33, 40, 41]. While the Raman spectrum of shock synthesized α-PbO₂ phase TiO₂ displays ten Raman bands (figure 4(b)) at the wave numbers of 150, 176, 289, 316, 339, 354, 428, 533, 570 and 612 cm⁻¹, respectively. The wave numbers of these bands are consistent with the data measured for α-PbO₂ TiO₂ obtained by static experiments [40] and that discovered in nature due to impact events (152, 175, 285, 315, 340, 358, 428, 532, 575 and 610 cm⁻¹) [35]. No peaks corresponding to anatase or rutile TiO₂ were observed. Raman spectra analysis confirms that the sample obtained from shot 1324 is pure α-PbO₂ phase TiO₂.

Zoom In Zoom Out Reset image size

3.3. Transmission electron microscopy

Figure 5 shows typical high-resolution TEM images of α-PbO₂ phase TiO₂ obtained from shot 1324. The TiO₂ particles after shock compression are ellipsoidal in shape with good crystallinity. The interplanar spacings of 0.354 nm and 0.286 nm corresponding to the (1 1 0) and (1 1 1) crystal planes of α-PbO₂ phase TiO₂ can be clearly discerned. The average particle size is approximately 14 nm, which is far smaller than the original 150 nm. This demonstrates that particle size refinement occurs during shock compression. The interpalanar spacings of α-PbO₂ phase TiO₂ were calculated through the SAED pattern (figure 5(a)), as shown in table 2. The standard values of α-PbO₂ phase (JCPDS 21-1236) are also given for comparison. The d-spacings of the α-PbO₂ TiO₂ obtained in this paper are in good agreement with the standard values of α-PbO₂ phase.

Table 2. Interplanar spacings calculated by SAED patterns for shock synthesized α-PbO₂ TiO₂ obtained from shot 1324.

h k l	d-spacings (d(h k l)) (Å)
	Sample 1324	α-PbO2 phase
1 1 0	3.54	3.50
1 1 1	2.82	2.84
0 0 2	2.46	2.45
1 0 2	2.17	2.15
2 0 2	1.66	1.66
2 2 2	1.41	1.42

Zoom In Zoom Out Reset image size

3.4. Thermal stability

α-PbO₂ phase TiO₂ is a thermal metastable phase. It may transform to rutile phase when heated. The reported onset reverse transformation temperature of the shock synthesized α-PbO₂ phase TiO₂ under atmospheric pressure is 440 °C by Linde et al [28] and above 450 °C by McQueen et al [16]. For the α-PbO₂ phase TiO₂ discovered in nature, the reported onset reverse transformation temperature under atmospheric pressure is 500 °C by Goresy et al [13], above 300 °C–900 °C by Chen et al [6].

Figure 6 shows the high temperature XRD patterns of pure α-PbO₂ phase TiO₂ obtained from shot 1324. For temperature below 500 °C, the diffraction peaks of α-PbO₂ phase TiO₂ do not show any noticeable change (figures 6(a)–(c)), demonstrating that the α-PbO₂ phase TiO₂ is rather stable at temperatures below 500 °C under atmospheric pressure. When the temperature is increased to 600 °C, a diffraction peak around 27.5° (figure 6(d)) assigned to (1 1 0) plane of rutile TiO₂ (JCPDS 65-1119) can be observed, implying the occurrence of reverse transformation of α-PbO₂ phase TiO₂ to rutile TiO₂. When the temperature is further increased to 800 °C and above, all the diffraction peaks assigned to α-PbO₂ phase TiO₂ disappear and all the diffraction peaks with enhanced intensities can be assigned to rutile TiO₂.

Zoom In Zoom Out Reset image size

Figure 7 shows the TG-DSC curves for pure α-PbO₂ phase TiO₂ obtained from shot 1324. The TG curve shows that the mass loss is approximately 8.9 wt% when temperature is increased to 560 °C, which can be attributed to the evaporation of residual solvents and absorbed water. The DSC curve suggests the onset of heat release at 560 °C, demonstrating that the onset reverse transition temperature is about 560 °C, which is consistent with the result of high temperature XRD patterns. When the temperature is above 600 °C, little weight loss (about 2.3 wt%) is observed and the DSC curve shows a smooth downward trend due to continuous reverse transformation of α-PbO₂ phaseTiO₂ to rutile TiO₂.

Zoom In Zoom Out Reset image size

3.5. Mechanism of the phase transition

The phase transformation behavior of anatase and rutile TiO₂ with different particle sizes was investigated under shock compression. Paraffin was added to the mixture of titanium dioxide and copper powder to fill the pores among the particles to reduce the residual temperatures of the shocked samples after shock unloading. The XRD analysis indicate that all the samples recovered under shock compression undergo a phase transition to a certain extent. Both anatase and rutile TiO₂ can undergo phase transformation to α-PbO₂ phase TiO₂ under shock compression. The transformation rate of α-PbO₂ phase TiO₂ for anatase TiO₂ precursor can reach 100%, while the transformation rate for rutile TiO₂ is about 90%. High resolution TEM examination and Raman spectra analysis also verify that pure α-PbO₂ phase TiO₂ was obtained. The yield of α-PbO₂ phase TiO₂ is largely affected by the structure of the starting TiO₂ powder, and the anatase TiO₂ powder is easier to transform to α-PbO₂ phase TiO₂ than rutile TiO₂ powder. The influence of the particle size on the yield of α-PbO₂ phase TiO₂ is not obvious. High temperature XRD analysis together with TG-DSC examination suggest that α-PbO₂ phase TiO₂ is a thermally metastable phase, and can convert to rutile phase at temperature above 560 °C. The residual temperatures in the shocked powder samples is crucial for the preservation of the shock formed metastable α-PbO₂ phase TiO₂. Adding a small amount of paraffin to the powder mixture to reduce the porosity is proven to be effective to lower the residual temperature of the shocked powder and preserve the metastable α-PbO₂ phase TiO₂. The anatase and rutile TiO₂ are assumed to convert to α-PbO₂ phase directly under shock compression, and then α-PbO₂ phase TiO₂ may reverse transform to rutile phase, depending on the cooling rate and residual temperature after shock unloading.

This work was financially supported by the National Natural Science Foundation of China under Grants 11521062 and 10972039 and the Project of State Key Laboratory of Explosion Science and Technology, Beijing Institute of Technology under Grant ZDKT18-01.