Bronze phase titanium dioxide (TiO2-B) has an ideal open structure for applications in high-rate lithium-ion batteries, but high quality and water-free TiO2-B is difficult to synthesize since TiO2-B is energetically less stable compared to other TiO2 polymorphs. Using CaTi5O11 as a template layer can help stabilize TiO2-B phase, but it is still challenging to avoid the formation of TiO2 anatase (TiO2-A) impurity phase. Here we show the synthesis of phase pure TiO2-B films by in situ engineering of the surface quality of the buffer layer using molecular beam epitaxy (MBE). By applying surface sensitive in situ reflection high-energy electron diffraction (RHEED), the formation of the impurity TiO2 anatase phase on the surface of CaTi5O11 buffer layer can be monitored and eliminated in real time, leaving a clean template surface for the growth of phase pure TiO2-B films.
I. INTRODUCTION
Titanium dioxide has four major different polymorphs: rutile, anatase, brookite, and bronze, and they have wide applications in photocatalysis,1,2 memory devices,3,4 solar cell,5 dilute ferromagnetic oxides,6–8 and lithium-ion batteries.9,10 Especially for the bronze phase of titanium dioxide, the open structure and fast lithium-ion transport via a pseudocapacitive Faradaic process can lead to ultrahigh discharge rate compare to those of supercapacitors,11 which is promising for high performance energy applications. However, as TiO2-B is less stable compared to anatase and rutile phases, direct growth TiO2 film on SrTiO3 and other substrates typically results in anatase phase instead of TiO2-B phase. So TiO2-B was mainly synthesized in nanostructures by hydrothermal methods and products often contain unreacted precursors, other TiO2 phases, and the unavoidable presence of structural water.9,12–16 This hinders the understanding of the intrinsic properties and further applications of this compound.
Recently, Kui Zhang and co-workers reported a new path to synthesize water-free TiO2-B single-crystal films by pulsed laser deposition (PLD), which significantly enhances the sample quality and battery performance.11 The key process is to utilize CaTi5O11 as a template layer to stabilize the metastable bronze phase. The crystal structure of CaTi5O11 consists of two-dimensional TiO2-B-like slabs interleaved by a Ti-Ca layer, turning the structure into a twinned zigzag pattern.17 Compared to perovskite substrates, these TiO2-B like slabs provide a template for subsequent growth of TiO2-B film. However, CaTi5O11 is very sensitive to the exact growth condition and slight deviation from the optimal growth condition will lead to TiO2-A impurity phase. Since deposing TiO2 on this mixed surface will form TiO2-A impurity phase,18 it is still extremely challenging to synthesis phase-pure TiO2-B films.
In this letter we show that phase-pure TiO2 films can be synthesized by MBE. With the help of in situ reflection high-energy electron diffraction (RHEED), we can monitor and eliminate the impurity TiO2-A phase in real-time, leaving a pure CaTi5O11 surface as the ideal template layer for the growth of phase-pure TiO2-B films.
II. MATERIAL AND METHODS
CaTi5O11 films were grown on (001) SrTiO3 substrate by molecular beam epitaxy (MBE) with an oxidant (O2) background pressure of 4 × 10−7 Torr. The substrate temperature was changed from 840 °C to 950 °C to optimize the growth condition. TiO2-B films were grown with the optimal growth temperature of CaTi5O11. In situ RHEED was used to monitor the film growth mode and surface structure. The crystal quality of all films was observed by X-ray diffraction (XRD) analysis using Bruker D8 system. High angle annular dark field (HAADF) scanning transmission electron microscopy image were obtained on a double aberration-corrected S/TEM FEI Titan cubed 60-300 at 300kV with a field emission gun.
III. RESULTS AND DISCUSSION
High quality CaTi5O11 template layer is the key process to grow phase pure TiO2-B film, but CaTi5O11 film is very sensitive to the growth temperature and small deviation from the optimal temperature will result in large mount of TiO2-A impurity. As seen in Fig. 1, strong diffraction peaks of TiO2-A phase in XRD data with weak peaks of CaTi5O11 phase indicate the dominant of TiO2-A impurity in films grown at 950 °C and 930 °C. As the growth temperature decreases to 900 °C, strong and sharp diffraction peaks of CaTi5O11 phase with tiny TiO2-A peak indicate a relatively high quality of CaTi5O11 film. Further decreasing the growth temperature lead to the formation of TiO2-A impurity again. These result indicates the small optimal growth temperature window for CaTi5O11 phase.
2θ-θ XRD scans of CaTi5O11 films grown at different temperatures, diffraction peak of TiO2-A impurity phase is marked with *. Corresponding RHEED patterns are shown on the right side. The 1× and 1.5× diffraction stripes are corresponding to TiO2-A and CaTi5O11 phase, respectively.
2θ-θ XRD scans of CaTi5O11 films grown at different temperatures, diffraction peak of TiO2-A impurity phase is marked with *. Corresponding RHEED patterns are shown on the right side. The 1× and 1.5× diffraction stripes are corresponding to TiO2-A and CaTi5O11 phase, respectively.
RHEED patterns also show the same trend as XRD data, but has the advantage of providing realtime information during the film growth. The films with large mount of TiO2-A impurity grown at 950 °C and 930 °C show clear 1× diffraction stripes in RHEED patterns. According to pervious research,19–24 these 1× diffraction stripes are coming from the TiO2-A phase. The film with tiny TiO2-A impurity grown at 900 °C show clear 1.5× diffraction stripes without any 1× diffraction stripes, indicating these 1.5× diffraction stripes are coming from the CaTi5O11. As seen in Fig. 2(a), the period of CaTi5O11 along in-plane axises is 1.5 times larger than SrTiO3,18 which leads to a 1.5× diffraction pattern. For the films grown at 870 °C and 840 °C, the RHEED patterns exhibit both 1.5× and 1× diffraction stripes, indicating a mixed surface of CaTi5O11 with TiO2-A impurity. Surprisingly, the TiO2-A phase seems to unavoidable at the initial growth of CaTi5O11 even at the optimal growth temperature. As seen in Fig. 2(c), both 1.5× and 1× diffraction stripes exist during the initial film growth. As the film grows thicker at optimal growth condition, the intensity of 1× diffraction stripes decreases and eventually disappears, leaving a clean and pure CaTi5O11 surface [Fig. 2(d)]. HAADF image shows that the TiO2-A impurity forms at the interface and disappears after 5nm away from the interface, resulting in pure CaTi5O11 on the surface. As the growth of high quality CaTi5O11 is very sensitive to the exact growth condition, in situ RHEED is a powerful technique to monitor the surface quality. Moreover, with the help of in situ RHEED, one can even fine tune the growth condition (temperature in this case) in real time to prepare phase pure surface for the subsequent growth of TiO2-B films.
(a) Schematics of the crystal structure of grown structure. (b) HAADF image of CaTi5O11 film grown on (001) SrTiO3 substrate, showing impurity anatase phase mostly near the interface. (c) RHEED pattern taken during the initial growth. (d) RHEED pattern taken after the growth of CaTi5O11 film. The 1× and 1.5× diffraction stripes are corresponding to anatase and bronze phase, respectively.
(a) Schematics of the crystal structure of grown structure. (b) HAADF image of CaTi5O11 film grown on (001) SrTiO3 substrate, showing impurity anatase phase mostly near the interface. (c) RHEED pattern taken during the initial growth. (d) RHEED pattern taken after the growth of CaTi5O11 film. The 1× and 1.5× diffraction stripes are corresponding to anatase and bronze phase, respectively.
By applying in situ RHEED to monitor and optimize the growth condition, we prepared high quality CaTi5O11 template leaving a clean and pure CaTi5O11 surface and subsequently grew TiO2-B films. As seen in Fig. 3(a), only 1.5× diffraction stripes with bright specular spot indicate a flat and phase-pure CaTi5O11 surface. The TiO2-B film also shows clear and pure 1.5× diffraction stripes (Fig. 3). XRD data show sharp diffraction peaks of TiO2-B and CaTi5O11 phase with negligible amount of TiO2-A impurity. This result indicates the power of in situ RHEED in preparing high quality template surface and phase pure target films.
RHEED patterns taken after the growth of (a) CaTi5O11 and (b) TiO2-B films on (001) SrTiO3 substrates. (c) XRD 2θ-θ scan of the TiO2-B/CaTi5O11/SrTiO3 film.
RHEED patterns taken after the growth of (a) CaTi5O11 and (b) TiO2-B films on (001) SrTiO3 substrates. (c) XRD 2θ-θ scan of the TiO2-B/CaTi5O11/SrTiO3 film.
IV. CONCLUSIONS
In summary, we grew phase-pure TiO2-B film on CaTi5O11 buffer layer by MBE. HAADF images indicate that TiO2-A impurity mainly forms near the interface, and can be covered by CaTi5O11 with optimal growth conditions. The clear and strong relationship between the surface structure and RHEED pattern provides a path to monitor and eliminate the TiO2-A impurity during CaTi5O11 film growth, leaving an ideal template layer for the growth of phase-pure TiO2-B film. This method demonstrates the power of in situ RHEED in optimizing the film quality and should be able to be applied in the epitaxial growth of other oxide films.
ACKNOWLEDGMENTS
This work was supported by the National Basic Research Programme of China (Grant No. 2015CB654901), the National Natural Science Foundation of China (Grant Nos. 51672125, 11774153, 51772143, and 11874199).
