Supplementary Materials for
Antisite defects in La0.7Sr0.3MnO3 and La0.7Sr0.3FeO3
Meng Gu1,2, Zhiguo Wang3, Michael D. Biegalski4, Hans M. Christen4, Yayoi Takamura1 and Nigel D. Browning1, 5, 6
1Department of Chemical Engineering and Materials Science, University of California-Davis, Davis, California, 95616
2Present address: Environmental Molecular Sciences Laboratory, Pacific Northwest National Laboratory, RichlandWA99354
3Department of Applied Physics, University of Electronic Science and Technology of China, Chengdu, 610054, P.R. China
4Center for Nanophase Materials Science, Oak Ridge National Laboratory, Oak Ridge, Tennessee, 37831; 5Department of Molecular and Cellular Biology, University of California-Davis, Davis, CA, 95616, 95616;
6 Present address: Fundamental & Computational Sciences Directorate, Pacific Northwest National Laboratory, RichlandWA99354
This supplementary material section provides details regarding to Z-contrast imaging, electron energy loss (EEL) spectroscopy, and the density functional theory (DFT) calculations in this study.
The KrF laser (248 nm) was operated at 10 Hz with an energy density of ~ 1.5 J/cm2. The substrate was heated to 7700C and the chamber pressure was held at 200 mTorr O2. The growth of these superlattices was monitored in-situ using reflection high energy electron diffraction. After the deposition, the film was cooled to room temperature in 200 Torr of O2. The samples for STEM were polished to electron transparency using the wedge polishing method using a Multi-prep machine from Allied High Tech Products, Inc. A Fischione ion milling system was used for low-voltage (2kV) Ar cleaning for about 30 minutes to remove the mechanically damaged surface layer. The images shown in Fig. 1(a,b) and Fig. 2 were obtained using a probe aberration-corrected JEM2100F STEM operating at 200 kV with a convergence angle of 18 mrad and collection semi-angle of around 110-300 mrad at UC Davis. The images in Fig. S2 and EEL spectra were acquired using the TEAM 0.5 microscope operating at 80 kV to avoid beam damage at the NationalCenter for Electron Microscopy at Lawrence Berkeley National Laboratory. The Z-contrast images acquired using TEAM 0.5 had a convergence angle of 29 mrad and collection semi-angle of 90-390 mrad and the EELS inner collection angle was 100 mrad.
The first-principles calculations are performed based on DFT using projector augmented wave potentials (PAW)1. The exchange correlation is described in the generalized gradient approximations (GGA)2 using the PW91 functional as implemented in the VASP code3. PAW potentials with valence configurations of 5s25p65d16s2, 3p63d64s2, 3d64s1, and 2s22p4 are used to describe the La, Fe, Mn and O atoms, respectively. Defect calculations are simulated with 240-atom (4a0×4a0×3a0; a0 is the lattice constant of a cubic LaMnO3 (LMO) or LaFeO3 (LFO) perovskite) supercell. A plane-wave expansion of the wave functions up to a kinetic-energy cutoff of 400 eV has been employed and the Brillouin zone was sampled by a Monkhorst-Pack k-point mesh of 2×2×2.
The defect formation energy of a native defect is simply the reaction energy to create the defect from ideal matter. The formation energy is not a constant but depends on the growth or annealing conditions4. The formation energy of a defect in charge state is defined as4,
(1)
where is the total energy of a supercell containing a relaxed defect (vacancy or antisite) with charge state q, is the total energy of the supercell without defects. ni indicates the number of atoms of type i atoms that have been added to (ni>0) or removed from (ni<0) the supercell. EF is the Fermi level referenced to the valence band maximum (EVBM) in the bulk materials. are the corresponding chemical potentials of these species.
The ranges of the chemical potentials, , and are defined by:
(i)the values that prevent the formation of pure metals (La, Fe, Mn) and O2 gas. In thermodynamic equilibrium the component chemical potential cannot be higher than the energy of bulk materials of this species. For example, if we tried to push the higher than we would no longer be growing LaMnO3, but instead precipitating metal Mn:
; ; ; and (2)
(ii)the values that prevent the formation of binary oxides:
, , (3)
(iii)the values that maintain a stable LMO or LFO compounds:
, (4)
These conditions are applied to the ternary LMO and LFO system and allow determining the chemical potential from the various equilibrium conditions. Fig. 4(a) shows the phase diagram of the ternary La-Mn-O (La-Fe-O) system, which is derived from the combination of the chemical potentials of Mn (Fe), Mn2O3 (Fe2O3), La, La2O3, and O2. The vertices of the triangle correspond to three dominant component chemical potentials and the compounds are placed on the edge according to their stoichiometric proportion. Points A–E denoted the five possible chemical-potential combinations and the intersections of these five points define the thermodynamically equilibrium region of the ternary system under different chemical potentials. At points A and B, the chemical potentials for O are similar to a strongly oxidation environment like those LMO (LFO) sample grown in an O2 rich state. On the contrary, at point D the system is in a strongly reducing environment. Under the assumption that LMO (LFO) is always stable, the atomic chemical potentials for La, Mn (Fe), and O are constrained by the following equilibrium equations, we take LMO as an example:
Point A: , (5)
Point B: , (6)
Point C: , (7)
Point D: , (8)
Point E: , (9)
The chemical potential of O was calculated from the total energy of O2 molecule, the bulk materials La (FCC), α-Mn, Fe (FCC), and La2O3 (P-3m1), Mn2O3 (Pcab), and Fe2O3 (R-3c) are used to calculate their chemical potential in the compound.
Fig.4(b) and (c) shows the formation energies of various defects along the vertices (A→B→C→D→E) defined in Fig.4(a). It can be seen the defect formation energies show great dependence on the chemical potentials. Under oxidation conditions, the vacancies of La, Mn and Fe exhibit negative formation energy. The large negative energies suggest formations of cation vacancies are thermodynamically favorable under O2-rich conditions. Formation of cation vacancies (oxygen overstoichiometry) will persist until the cation vacancy concentration is high enough that the cation vacancy interactions compensate for the negative cation vacancy formation energy. In contrast, under reducing conditions the formation energies of VFe, VMn, LaFe, LaMn and VO show the lower formation energy, indicating that LaFe and LaMn antisites occur easily under reducing condition companied with amount of oxygen and Mn (Fe) vacancies, and in this case La will be overstoichiometry.
The EEL spectra shown in Fig. 3 are raw spectra without normalization. The EEL spectra were background subtracted before the Mn/Fe L edge and overlaid together for comparison in Fig. 3. The signal for Mn/Fe (B site species) in the defect region is lower than normal region, while the signal of La (A site species) in the defect region is higher than the normal region. (The O K edge spectra corresponding to these spectra are overlaid for comparison in Fig. S3.) The thickness effect on the EEL signal can be excluded considering the fact that thickness variations will only result in coherent changes in the core edges of Mn/Fe and La. Therefore, we conclude that the EEL spectra provide important clues on the formation of these antisite defects.
Fig S1. Overall view Z-contrast (a) and bright field image (b) of the film and STO substrate region. Note: the defect contours are circled in red in the bright field image and no such cluster-shaped defect contours are observed in the STO substrate region.
Fig S2 (a) Z-contrast image where some regions with enhanced contrast can be observed. (b) Integrated Mn (Fe) L edge divided by O K edge intensity maps from the red rectangular region shown in (a). (c) Integrated Mn (Fe) L edge divided by O K edge intensity line scans across the whole film from the film surface to the interface marked by the white line in (a).
Fig S3. Raw O K edge comparison of the defect and normal regions at LSMO layer (a) and LSFO layer (b); (c) plot of the normalized O K edges normalized to the middle peak of O K edge at LSFO layer
References:
1.G. Kresse and D. Joubert, Phys. Rev. B 59 (3), 1758-1775 (1999).
2.J. P. Perdew and W. Yue, Phys. Rev. B 33 (12), 8800-8802 (1986).
3.G. Kresse and J. Furthmuller, Comput. Mater. Sci. 6 (1), 15-50 (1996).
4.C. G. Van de Walle and J. Neugebauer, J. Appl. Phys. 95 (8), 3851-3879 (2004).