Growth of strontium ruthenate films by hybrid molecular beam epitaxy

The discovery of unusual ground states in the layered ruthenates, such as p-wave superconductivity^1,2 and nematic order,³ has led to significant interest in the growth of high-quality thin films of these materials. For example, Sr₂RuO₄ is an unconventional, spin triplet superconductor,⁴ whose superconducting transition temperature is highly sensitive to strain.⁵ Thin films of Sr₂RuO₄ would allow for advancing the understanding of unconventional forms of superconductivity through the application of epitaxial strain⁶ and are needed for applications in topological and hybrid superconducting devices. A remarkable success of pulsed laser deposition (PLD) is the growth of superconducting Sr₂RuO₄ films^7,8 because even minute defect concentrations destroy the superconductivity completely.⁹ 

Molecular beam epitaxy (MBE) has been key in the development of high quality semiconductor thin film heterostructures¹⁰ and has produced high performance complex oxide thin films.¹¹ MBE of the ruthenates is, however, challenging because of the extremely low volatility of Ru,¹² requiring the use of electron beam assisted evaporation.^13,14 Electron beam sources suffer from inherent flux variations and require in situ monitoring and real time feedback to continuously adjust the flux,¹⁴ which makes stoichiometry control difficult. Other challenges include practical limits on the oxygen pressure/fluxes that can be used and flux instabilities in the presence of oxygen.¹⁵ 

An alternative approach is hybrid MBE, which uses a combination of solid and gas sources to supply the metals.^16,17 In particular, metalorganic and other gaseous sources that have much higher vapor pressures and already contain metals bonded to oxygen, and whose fluxes can be precisely controlled, have been shown to successfully address all of the challenges in MBE of perovskite titanates,^17–19 zirconates,²⁰ and vanadates,²¹ resulting in high quality thin films having record properties.^22,23

Here, we demonstrate the growth of epitaxial Sr₂RuO₄ films using hybrid MBE. Strontium is evaporated from an elemental metal source while a volatile ruthenium tetroxide (RuO₄)-containing precursor is used to supply pre-oxidized Ru.²⁴ The RuO₄ precursor has been used to synthesize Ru, RuO₂, and SrRuO₃ films^24–26 but not yet in MBE. The resulting films are shown to be of excellent structural quality. The approach should be suitable for other ruthenate systems as well, thus opening up a viable MBE approach for thin films of a very exciting class of materials.

Sr₂RuO₄ films were grown on (001) (LaAlO₃)_0.3(Sr₂AlTaO₆)_0.7 (LSAT) substrates in an oxide MBE system (GEN 930, Veeco Instruments). The in-plane lattice parameters of Sr₂RuO₄ and LSAT are (nearly) perfectly matched (3.87 Å^27,28), making LSAT an excellent substrate for the epitaxial growth of strain-free Sr₂RuO₄. Substrates were etched in a 3:1 solution of HCl:HNO₃ and backed with 350 nm Ta to facilitate the radiative heating from the substrate heater during growth. Films were grown by co-deposition. Strontium was supplied from a high-purity elemental source contained in a low-temperature effusion cell. Ruthenium was supplied in the form of the RuO₄-containing precursor.²⁴ This precursor was packed with fluorinated ethers to stabilize the highly reactive RuO₄ molecules, which comprised about 10% by weight of the mixture. The precursor cylinder was at room temperature and no carrier gas was used, as the vapor pressure of RuO₄ is sufficiently high for delivery to the growth chamber. The precursor mixture was delivered to the growth chamber via a gas inlet system using a capacitive manometer and a linear leak valve to monitor and control the flow rate. It was introduced to the growth chamber through a gas injector heated to 50 °C. Additional oxygen was supplied using a plasma source operated at a power of 300 W with a background pressure of ∼3 × 10⁻⁶ Torr during growth. Film growth and structure were monitored in situ using reflection high-energy electron diffraction (RHEED) and ex situ via x-ray diffraction (XRD), atomic-force microscopy (AFM), and transmission electron microscopy (TEM). Room temperature electrical measurement was performed in four point probe (van der Pauw) geometry using Au/Ti ohmic contacts.

The film stoichiometry was varied by changing the Sr flux via the effusion cell temperature and keeping all the other fluxes constant. Similar to hybrid MBE of SrTiO₃,²⁹ post-growth RHEED patterns (Fig. 1) were found to be sensitive to the film stoichiometry. For Sr-deficient films, the RHEED pattern was noticeable spotty, especially along the [110] direction, while chevron patterns became visible in Sr-rich films. The streaky post-growth RHEED patterns along the [100] and [110] directions for the stoichiometric films indicated a smooth film surface.

The stoichiometry was also evident in the electrical properties, as the room temperature resistivities of the Sr-deficient, stoichiometric, and Sr-rich films were 509 μΩ cm, 230 μΩ cm, and 643 μΩ cm, respectively. In contrast, the out-of-plane lattice parameter (12.74 Å) was found to be independent of stoichiometry. Thus, the best indicators of stoichiometric Sr₂RuO₄ films were high intensity and streakiness of the RHEED pattern, prominent thickness fringes in XRD, and low resistivity.

In addition to the flux ratios during growth, high substrate temperatures were essential. Films grown at substrate temperatures of less than 920 °C (thermocouple reading) showed no oscillations in RHEED during growth and were highly resistive. The use of higher substrate temperatures during growth also resulted in a substantial decrease in the room temperature resistivity, from 1100 μΩ cm for films grown at 920 °C to 230 μΩ cm for films grown at 950 °C. Furthermore, the highest substrate temperatures (950 °C) used in this study were found to promote layer-by-layer growth (see below). All results shown here are from films grown at 950 °C.

RHEED intensity oscillations, shown in Fig. 2(a), indicated the layer-by-layer growth mode. This growth mode allows for excellent control over layer thickness and is a signature of sufficiently high adatom mobility on the growing film surface. The initial behavior of the RHEED intensity may be a result of the change in symmetry from the cubic substrate to the tetragonal film, as described previously for PLD-grown films.⁷ The period of the intensity oscillations corresponded to a growth rate of 16 nm/h, with two intensity oscillations per Sr₂RuO₄ unit cell, as has also been found by others.³⁰ AFM [Fig. 2(b)] confirmed the smooth film surface with features that are on the order of the unit cell height (12.74 Å).

A wide angle XRD scan is shown in Fig. 3(a). The 002, 004, 006, 008, and 0010 Bragg reflections of Sr₂RuO₄ are all present. The lattice spacing is consistent with an out-of-plane lattice constant of 12.74 Å, matching the known value for the tetragonal unit cell of Sr₂RuO₄ (space group I4/mmm³¹). Thus, the films are phase-pure Sr₂RuO₄. This is significant, as mixed-phase films, containing perovskite SrRuO₃ and the n = 2 Ruddlesden-Popper phase Sr₃Ru₂O₇, are also possible.³² Figure 3(b) shows a high-resolution scan around the 006 reflection of Sr₂RuO₄ and the 002 reflection of the LSAT substrate. The presence of prominent thickness fringes indicates very smooth interfaces and surfaces. The spacing of the fringes corresponds to a film thickness of 29 nm, in good agreement with the thickness derived from the period of the RHEED oscillations. The rocking curve full width at half maximum of the 006 Sr₂RuO₄ peak was 0.075°, confirming a high crystalline quality of the film [see Fig. 3(c)]. Epitaxial growth of phase-pure Sr₂RuO₄ was also confirmed by cross section high-angle annular dark-field (HAADF) imaging in scanning TEM (STEM), as shown in Fig. 4. The films consist of only the n = 1 Ruddlesden-Popper layered crystal structure of Sr₂RuO₄. The images show abrupt interfaces and no extended defects.

To summarize, we have demonstrated the growth of phase pure thin films of Sr₂RuO₄ with high structural quality using a hybrid molecular beam epitaxy approach that utilizes a highly volatile precursor containing reactive RuO₄ molecules. The method allows for control over the stoichiometry and layer-by-layer growth. Future studies will focus on the characterization of the electrical properties as a function of temperature. The excellent structural quality of the Sr₂RuO₄ film points towards this precursor as a promising route to the development of other materials in the strontium ruthenate family.

The work was supported by a MURI funded by the U.S. Army Research Office (Grant No. W911NF-16-1-0361) and by FAME, one of six centers of STARnet, a Semiconductor Research Corporation program sponsored by MARCO and DARPA. The microscopy work was supported by the U.S. Department of Energy (Grant No. DEFG02-02ER45994).