annealing assisted substrate coherency and high temperature antiferromagnetic insulating transition in epitaxial la0 67ca0 33mno3 ndgao3 001 films

Annealing assisted substrate coherency and high-temperature antiferromagnetic insulating transition in epitaxial La0.67Ca0.33MnO3/NdGaO3(001) films L F Wang, X L Tan, P F Chen, B W Zhi, B B Chen, Z Huang, G Y Gao, and W B Wu Citation: AIP Advances 3, 052106 (2013); doi: 10.1063/1.4804541 View online: http://dx.doi.org/10.1063/1.4804541 View Table of Contents: http://scitation.aip.org/content/aip/journal/adva/3/5?ver=pdfcov Published by the AIP Publishing Articles you may be interested in Enhancing the orthorhombicity and antiferromagnetic-insulating state in epitaxial La0.67Ca0.33MnO3/NdGaO3(001) films by inserting a SmFeO3 buffer layer J Appl Phys 116, 203706 (2014); 10.1063/1.4902951 Anisotropic resistivities in anisotropic-strain-controlled phase-separated La0.67Ca0.33MnO3/NdGaO3(100) films Appl Phys Lett 103, 072407 (2013); 10.1063/1.4818636 Effect of growth oxygen pressure on anisotropic-strain-induced phase separation in epitaxial La0.67Ca0.33MnO3/NdGaO3(001) films J Appl Phys 113, 203701 (2013); 10.1063/1.4807293 Phase evolution and the multiple metal-insulator transitions in epitaxially shear-strained La 0.67 Ca 0.33 MnO / NdGaO ( 001 ) films J Appl Phys 108, 083912 (2010); 10.1063/1.3499650 Magnetotransport properties in La − x Ca x MnO ( x = 0.33 , 0.5) thin films deposited on different substrates J Appl Phys 97, 083909 (2005); 10.1063/1.1870118 All article content, except where otherwise noted, is licensed under a Creative Commons Attribution 3.0 Unported license See: http://creativecommons.org/licenses/by/3.0/ Downloaded to IP: 158.42.28.33 On: Fri, 12 Dec 2014 11:06:17 AIP ADVANCES 3, 052106 (2013) Annealing assisted substrate coherency and high-temperature antiferromagnetic insulating transition in epitaxial La0.67 Ca0.33 MnO3 /NdGaO3 (001) films L F Wang, X L Tan, P F Chen, B W Zhi, B B Chen, Z Huang, G Y Gao, and W B Wua Hefei National Laboratory for Physical Sciences at Microscale, University of Science and Technology of China, and High Magnetic Field Laboratory, Chinese Academy of Sciences, Hefei 230026, People’s Republic of China (Received 25 March 2013; accepted 26 April 2013; published online May 2013) Bulk La0.67 Ca0.33 MnO3 (LCMO) and NdGaO3 (NGO) have the same Pbnm symmetry but different orthorhombic lattice distortions, yielding an anisotropic strain state in the LCMO epitaxial film grown on the NGO(001) substrate The films are optimally doped in a ferromagnetic-metal ground state, after being ex-situ annealed in oxygen atmosphere, however, they show strikingly an antiferromagnetic-insulating (AFI) transition near 250 K, leading to a phase separation state with tunable phase instability at the temperatures below To explain this drastic strain effect, the films with various thicknesses were ex-situ annealed under various annealing parameters We demonstrate that the ex-situ annealing can surprisingly improve the epitaxial quality, resulting in the films with true substrate coherency and the AFI ground state And the close linkage between the film morphology and electronic phase evolution implies that the strain-mediated octahedral deformation and rotation could be assisted by ex-situ annealing, and moreover, play a key role in controlling the properties of oxide heterostructures C 2013 Author(s) All article content, except where otherwise noted, is licensed under a Creative Commons Attribution 3.0 Unported License [http://dx.doi.org/10.1063/1.4804541] I INTRODUCTION For the past decades, researches in transition-metal perovskite oxides have flourished The interest in these oxides arises from the fact that they are based on the same simple ABO3 perovskite structure and can exhibit a wide diversity of functionalities such as the large and tunable piezoelectricity, ferroelectricity, colossal magnetoresistance, and superconductivity Furthermore, by stacking different perovskites into epitaxial films, multilayers, and superlattices, the reconstructions of charge, orbital, spin, and lattice degrees of freedom on the nanometer scale may allow not only their properties to be combined but, sometimes, also totally new phenomena to be induced at the heterointerface.1 Specifically, the reconstructions of the flexible corner-sharing BO6 networks are always intriguing and crucial in perovskite heterostructures.2 On one hand, both the experimental and theoretical results have revealed that the epitaxial strain, generated by the lattice constant mismatch at the heterointerface between two dissimilar perovskites, can directly alter the patterns and magnitudes of the octahedral rotation and deformation, then giving rise to structures that are not been found in bulk phase diagram,3 and even structural transitions as a function of the lattice mismatch, symmetry mismatch, and film thickness.4–7 More importantly, the strong electron-lattice coupling in those correlated perovskites enables the strain-coupled octahedra behaviors to effectively manipulate their properties For instance, in the (Nd,Pr)0.5 Sr0.5 MnO3 films grown on SrTiO3 substrates a Electronic mail: wuwb@ustc.edu.cn 2158-3226/2013/3(5)/052106/14 3, 052106-1 C Author(s) 2013 All article content, except where otherwise noted, is licensed under a Creative Commons Attribution 3.0 Unported license See: http://creativecommons.org/licenses/by/3.0/ Downloaded to IP: 158.42.28.33 On: Fri, 12 Dec 2014 11:06:17 052106-2 Wang et al AIP Advances 3, 052106 (2013) with various orientations, the antiferromagnetic charge-orbital-ordering (COO) transition can be either enhanced or suppressed, depending on the strain-stabilized various octahedral deformation modes.8 And Zayak et al.9 have shown that the strain induced octahedral rotation and deformation could alter the magnetic-nonmagnetic transition in CaRuO3 and SrRuO3 films In addition to the strain-mediated elastic octahedral coupling, recently it has been predicted and demonstrated that the corner-connectivity nature of octahedral networks could enable the octahedral rotation/deformation patterns in the substrate to transfer across the interface and imprint into the films, yielding interfacial layers with octahedral configurations unstable in bulk.10–12 Unlike the epitaxial strain that can be coherently maintained over tens of nanometers, the interfacial octahedral coupling only persists for several unit-cells.2, 11 Even so, this effect was reported to stabilize novel electronic properties in the interface region Examples include the anomalous dielectric behavior of BiFeO3 near the BiFeO3 /La0.7 Sr0.3 MnO3 heterointerface,12 the improper ferroelectricity in ultrashort period PbTiO3 /SrTiO3 superlattice,13 and the room-temperature magnetic insulating phase in ultrathin La0.67 Sr0.33 MnO3 /SrTiO3 (110) films.14 For the (Sr,Ca)TiO3 /LaAlO3 systems, the interfacial electronic states were also suggested to be affected by the interfacial reconstruction of octahedral patterns.15, 16 Accordingly, the strain or interface engineering of octahedral rotation and deformation patterns has provide an opportunity to stabilize familiar perovskites with new functionalities However, synthesis of epitaxial films with substrate coherency, i.e., maintaining pseudomorphic strain state and interfacial octahedral coupling, is always a non-trivial task On one hand, the control of epitaxial strain is challenged by the lattice mismatch between the film and substrate A typical instance is the La1-x Cax MnO3 films grown on SrTiO3 (001) and LaAlO3 (001) substrates, in which the large biaxialstrain could induce selective orbital occupancy and phase separation (PS) states.17 Nevertheless, since the large lattice mismatch limits the critical film thickness for coherent strain state, the strain-mediated electronic properties in these ultrathin films are unfortunately entangled with many extinct effects, including the grainy surface morphology, finite-size scaling effect, and nonuniform termination of the substrate.18 Further, as the film thickness increases the partially relaxed strain may result in the lattice disorders or defects that are usually unwanted.17, 18 On the other hand, the oxygen vacancies, interfacial atomic intermixing and cation stoichiometry errors formed during the film growth may also lead to various structure inhomogeneities, which influence the strain state as well as the octahedral behavior.19–21 In order to improve the substrate coherency, much effort has been devoted in the in-situ heteroepitaxy processes, such as developing advanced epitaxial growth techniques,22 sophisticated control of thin film growth modes,23–25 and choosing suitable substrates with small lattice and symmetry mismatch.5, 25, 26 Moreover, the ex-situ annealing in oxygen atmosphere has been widely employed to optimize the crystallinity and eliminate oxygen deficiency for complex oxide films.21, 27, 28 But it was also shown that the high temperature annealing may also facilitate the strain relaxation in large lattice-mismatched systems.17, 29 Therefore, the effect of ex-situ annealing on substrate coherency should be multiplicate, and surely needs to be further investigated In this paper, in order to shed some light on the relationships between ex-situ annealing, substrate coherency, and electronic properties in perovskite heterostructures, we choose the La0.67 Ca0.33 MnO3 (LCMO) films grown on NdGaO3 (001) [NGO(001)] substrates as a model system In bulk, the NGO and LCMO show the same orthorhombic Pbnm symmetry with GdFeO3 -type octahedral rotation pattern (tilting system a− a− c+ in Glazer notation) but different octahedral rotation angles.30, 31 And the GaO6 octahedra are nearly regular and rigid, whereas the MnO6 octahedra are slightly deformed by the JT distortion Because of the distinct octahedral configurations for these two perovskites, the coherently grown LCMO/NGO(001) film suffers a considerable in-plane anisotropic strain At the same time, the average lattice mismatch between the film and substrate is negligible, which enables the anisotropic strain state to be maintained even in thick films, easily preventing the strain relaxation More strikingly, though the bulk LCMO is optimally doped in the ferromagnetic-metal (FM) ground state, the LCMO/NGO(001) films, after being annealed in oxygen atmosphere, show an antiferromagnetic insulator phase transition at ∼250 K and the PS with the coexistence of antiferromagnetic-insulating (AFI) and FM phases at the temperature below.32–35 In the early work we have demonstrated that the ex-situ annealing process and the NGO(001) substrate stabilized All article content, except where otherwise noted, is licensed under a Creative Commons Attribution 3.0 Unported license See: http://creativecommons.org/licenses/by/3.0/ Downloaded to IP: 158.42.28.33 On: Fri, 12 Dec 2014 11:06:17 052106-3 Wang et al AIP Advances 3, 052106 (2013) in-plane anisotropic strain state are both indispensable for the formation of AFI phase.34, 35 However, the detailed evolutions of the strain state and octahedral behavior during annealing, which are central for understanding the AFI phase and PS, are still unclear Here we develop a new strategy by investigating the surface morphology and magnetotransport property in parallel from the LCMO/NGO(001) films with various thicknesses after being ex-situ annealed under various annealing parameters The surface morphology evolutions demonstrate that the ex-situ annealing can surprisingly improve the epitaxial quality and thus enhance the substrate coherency And based on observed close relationship between the epitaxial quality and phase evolution, along with the quantitative analyses of the octahedral response to the anisotropic strain, we suggest that after annealing the enhanced substrate coherency could stabilize the Jahn-Teller distortion in the LCMO films and thus trigger the AFI phase II EXPERIMENT LCMO/NGO(001) films of 5-40 nm thick were grown by the pulsed laser deposition method.29, 32 The ceramic LCMO target was prepared by standard solid state reactions During deposition, the substrate temperature and O2 pressure were 735 ◦ C and 45 Pa, and the laser energy and repetition rate were set at J/cm2 and Hz, resulting in a growth rate of ∼5 nm/min After deposition each film was in-situ annealed for 15 before being cooled down to room temperature, during this process the substrate temperature and O2 pressure remained at 735 ◦ C and 45 Pa Then the films were ex-situ annealed in flowing O2 of ambient pressure for a fixed duration time and temperature The magnetization (M) and resistivity (ρ) were measured on Quantum Design superconducting quantum interference device (SQUID) magnetometer (MPMS) and physical property measurement system (PPMS) In all the ρ measurement the applied current was fixed at μA, and the data beyond the measurement limit (∼ 10 cm) were not shown The structures of the films were checked by x-ray diffraction (XRD) including ω-2θ linear scan, ω-scan rocking curves (RCs) and reciprocal space maps (RSMs) using CuKα radiation (λ = 1.5406 Å, Panalytical X´pert) The film thicknesses were determined by analyzing the Laue fringes around LCMO(004) diffraction peak, consistent with the evaluated value from growth rate The surface morphologies of the films were measured by the atomic force microscopy (AFM, Vecco, MultiMode V) III RESULTS AND DISCUSSION The strain states of the LCMO/NGO(001) films were schematically analyzed in Fig 1(a) In Pbnm orthorhombic index the bulk NGO (LCMO) has the lattice constants a = 5.4332 Å (5.4717 Å), b = 5.5034 Å (5.4569 Å), c = 7.7155 Å (7.7112 Å).36 According to these data, although the average in-plane lattice mismatch is negligible (0.06%), the coherently grown LCMO/NGO(001) film could take an in-plane anisotropic strain: −0.70% compressive strain along [100] but 0.85% tensile strain along [010] In Fig 1(b), the AFI phase and PS evolving with film thickness and annealing parameter were briefly depicted by the temperature dependent resistivity (ρ-T) curves The h-annealed film of 40 nm thick only shows FM transition at 267 K, the same as the bulk counterpart By contrast, after being ex-situ annealed at 780 ◦ C for 10 h, the film of the same thickness shows not only a FM transition at 259 K, but also a subsequent AFI phase transition at ∼250 K (TAFI , which has been confirmed by previous magnetization measurements), and the subsequent complex hysteresis and large residual resistivity manifest the PS state with competing AFI and FM phases.34, 35 With decreasing the thickness to 20 nm, the annealed film shows a higher ρ below TAFI , indicating an enhanced AFI phase In fact, by comparing the magnetotransport properties of LCMO films grown on the NGO substrates with various orientations, we have excluded the interfacial chemical diffusion and variation of La:Ca doping ratio during ex-situ annealing, and demonstrated that the AFI phase is elastic-driven.33, 35 In Fig 1(c), the ω-2θ linear scans of the 10-h annealed films (20 and 40 nm) show sharp Laue fringes, further confirming that the LCMO films have high crystal quality and sharp interface with the underlying substrates.4 The strain states of the films were characterized by the RSMs as shown in Fig 1(d) When compared with the (116) reflection from the as-grown film, the Q[001] of the same reflection from annealed film increases, signifying that the ex-situ annealing can All article content, except where otherwise noted, is licensed under a Creative Commons Attribution 3.0 Unported license See: http://creativecommons.org/licenses/by/3.0/ Downloaded to IP: 158.42.28.33 On: Fri, 12 Dec 2014 11:06:17 052106-4 Wang et al AIP Advances 3, 052106 (2013) FIG (a) The schematic top views of LCMO(001) (green) and NGO(001) (wine) unit cells The strain state of the commensurate LCMO/NGO(001) film was analyzed via the in-plane lattice mismatches in the orthorhombic Pbnm index (b) ρ-T curves of 20 and 40 nm films annealed at 780 ◦ C for 10 h The curve of h-annealed 40 nm film was also inserted for comparison The arrows denote the thermal processes (c) ω-2θ scans from the 10 h-annealed films of 20 and 40 nm thick The sharp peaks at 2θ = 47.074◦ can be indexed as NGO(004), and the broad humps at the higher angle side stands for the LCMO(004) reflections (d) RSMs measured from the as-grown and 10 h-annealed films (40 nm) around (116) reflection The open and solid arrows denote the reflections from the films and the substrates, respectively cause oxygen incorporation and then the contraction of the out-of-plane film lattice.28 On the other hand, the reflections from the as-grown and annealed films are both sharp and concentrated, and have exactly the same in-plane lattice spaces as those of the substrates, suggesting that LCMO/NGO(001) films are coherently strained even after being ex-situ annealed.17, 26, 29, 35 That means the strain relaxation during annealing may not be the driven force for the AFI phase For these LCMO/NGO(001) thin films with negligible lattice mismatch, the local structure changes during annealing could be averaged out by the X-ray diffraction, and the AFM used for characterizing the surface morphology might be a more effective approach Based on this consideration, as presented in Fig 2, the AFM images along with zero-field ρ-T curves were correspondingly measured from a set of as-grown and annealed films with unequal thicknesses For the as-grown films, as observed in Fig 2(a)–2(e), although all the surfaces are very smooth with the root-mean-square roughness less than nm, the randomly distributed grainy structures indicate that a high density of defects accumulated during the in-situ growth, which may be attributed to 3-dimentional island film growth mode in our deposition condition.24 By contrast, after being ex-situ annealed at 780 ◦ C for h, the films of 8, 12 and 16 nm thick [Fig 2(f)–2(h)] exhibit singleunit-cell stepped terraces with sharp edges and uniform widths As the thickness increases to 24 and 40 nm [Fig 2(i) and 2(j)], the atomic terraces in the annealed film surfaces turn out to be distorted and discrete Aside from this drastic morphology change, as shown in Fig 2(k) and 2(l), the close correspondences between the surface morphologies and electronic transport properties are rather interesting The as-grown films with grainy surfaces only show bulk-like FM transition with a gradually reduced TC as the thickness decreases In contrast, for the annealed films of 8, 12 and 16 nm thick, accompanying with the improvement of surface morphology, the FM transition disappears, instead the ρ-T curves show the AFI phase transition followed by insulating behavior, signifying an AFI ground state And for those 24 and 40 nm films that display distorted terraces, the ρ-T curves show a typical PS behavior as characterized by the thermal hysteresis and multiple metal-insulator transitions All article content, except where otherwise noted, is licensed under a Creative Commons Attribution 3.0 Unported license See: http://creativecommons.org/licenses/by/3.0/ Downloaded to IP: 158.42.28.33 On: Fri, 12 Dec 2014 11:06:17 052106-5 Wang et al AIP Advances 3, 052106 (2013) FIG AFM images (1 × μm2 ) scanned from (a)-(e) as-grown and (f)-(j) h-annealed films with different thicknesses as denoted The as-grown film morphologies for all the thickness show grainy surface, whereas the annealed film morphologies show single-unit-cell stepped terraces evolving with film thickness (k) [(l)] The ZFC-ZFW ρ-T curves measured from the as-grown (annealed) films All article content, except where otherwise noted, is licensed under a Creative Commons Attribution 3.0 Unported license See: http://creativecommons.org/licenses/by/3.0/ Downloaded to IP: 158.42.28.33 On: Fri, 12 Dec 2014 11:06:17 052106-6 Wang et al AIP Advances 3, 052106 (2013) FIG (a) ZFC-FW and (c) FC-ZFW ρ-T curves measured from the LCMO/NGO(001) films with various thickness, as denoted The solid arrows denote the cooling or warming The ZFC-ZFW ρ-T curves (gray, dashed line) were also inserted in (a) for comparison In (c) the AFI reentrance temperatures TR were marked by the solid triangles (b) [(d)] The isothermal ρ-H curves measured after ZFC the 16, 24 and 40 nm films to K (130 K) The open arrows denote the cycling of H (0→ 4→ T), and the AFI phase melting (reentry) fields HM (HR ) were marked by the solid arrows In all the measurements the magnetic fields were applied along the c axis The ρ-T curves during various thermal processes and isothermal ρ-H curves were measured from the annealed films in order to characterize the phase evolution and instability in detail As shown in Fig 3(a), after zero-field-cooling (ZFC) the 40, 24, and 16 nm films to K and then during field warming (FW) at H = T, the ρ-T curves show steep drops in resistivity, which can be explained as the phase transition from ‘frozen PS’ to ‘dynamic PS’.34, 35, 37 It is clear that this phase transition shifts to higher temperatures as the film thickness decreases, indicating a more stable frozen state in a thinner film In Fig 3(b), the ρ-H curves measured after ZFC the films to K show irreversible AFI phase melting behavior, consisting with the dynamic T-H phase diagram as previously constructed.34 For the films of 16, 24 and 40 nm thick, the AFI phase can be completely melted at 5.1, 4.5 and 3.9 T (HM ), respectively The increased HM further confirms that the frozen state becomes more robust against H as the thickness decreases In Fig 3(c), after field-cooling at T (6 T for the nm film) to 10 K then during zero-field warming (FC-ZFW), the films can keep the FM phase till TD followed by a sharp increase in ρ (nearly five orders for the 16 nm film), signifying the AFI phase reentrant behavior during the phase transition from FM-dominated PS to AFI-dominated PS.34, 38 For the and 16 nm films, the sharp AFI reentrance is stabilized at ∼94 K, while for the 24 and 40 nm films the TD increases rapidly to 121 and 140 K, respectively, and the resistivity jump becomes sluggishly That means the FM phase could become energetically preferred as the film thickness increases In Fig 3(d), the ρ-H curves at 130 K show not only the AFI phase melting behavior but also the thickness dependent AFI phase reentrance.38 In the 16 nm (24 nm) film the AFI phase can reenter as H is decreased to 2.15 T (1.06 T), while no reentrant AFI phase was observed in the 40 nm film, because the TD of this film is well above 130 K Based on the All article content, except where otherwise noted, is licensed under a Creative Commons Attribution 3.0 Unported license See: http://creativecommons.org/licenses/by/3.0/ Downloaded to IP: 158.42.28.33 On: Fri, 12 Dec 2014 11:06:17 052106-7 Wang et al AIP Advances 3, 052106 (2013) FIG AFM images (1 × μm2 ) from (a)-(d) 14 nm and (e)-(h) 28 nm films during the accumulative annealing processes The surface terraces are gradually improved as the annealing temperature and duration time increases (i) [(j)] The corresponding ρ-T curves of 14 (28) nm films morphology data as well as the ρ-T and ρ-H curves presented in Fig and Fig 3, we may conclude that in the LCMO/NGO(001) epitaxial system the ex-situ annealing improved surface morphologies are closely related to the electronic phase instability: the annealed thin films that display uniform terraced surfaces tend to show stable AFI ground state; and for those thicker ones, in accord with the imperfect surface morphologies, the AFI phase is weakened and the FM phase is enhanced, thus leading to the unstable PS state In order to further elucidate the ex-situ annealing effect on the improvement of surface morphologies and the formation of the AFI phase, as shown in Fig 4, the AFM images and zero-field ρ-T curves were correspondingly measured from a set of the films after being ex-situ annealed under various annealing parameters For the films deposited and ex-situ annealed in the same run, we All article content, except where otherwise noted, is licensed under a Creative Commons Attribution 3.0 Unported license See: http://creativecommons.org/licenses/by/3.0/ Downloaded to IP: 158.42.28.33 On: Fri, 12 Dec 2014 11:06:17 052106-8 Wang et al AIP Advances 3, 052106 (2013) have confirmed the good reproducibility of ρ-T data, but the surface morphology changes could be affected by not only the annealing parameters but also other extinct factors such as the variation of miscut angle for different substrates.39 Hence, a special procedure, named accumulative annealing process, was used for the three kinds of measurements to obtain the most reliable data In detail, five commensurate films of 14 nm (28 nm) thick were deposited and then annealed accumulatively side by side for four times with progressively increased annealing temperature (duration time) At the end of each annealing procedure, the AFM images were always measured from one specified sample, while the Pt electrodes were prepared on one of the rest samples for the corresponding ρ-T measurements After these steps, the morphology and transport data of the 14 nm (28 nm) films evolving with gradually increased annealing temperature (duration time) were obtained, as presented in Fig 4(a)–4(d) [Fig 4(e)–4(h)] and Fig 4(i) [Fig 4(j)], respectively For the 14 nm film, after being annealed at 600 ◦ C for h, the grainy surface is similar to that of as-grown film, and the ρ-T curve only shows the bulk-like FM transition As the annealing temperature rises to 660 ◦ C and 720 ◦ C, the surface terraces and unit-cell-height 2D islands appear, and the coalescence of these islands and steps leads to the meandering of the terrace edges Further being annealed at 780 ◦ C, the 2D islands disappear and the uniform terraces were observed At the same time, for the samples annealed at 660, 720 and 780 ◦ C, all the ρ-T curves exhibit an evident AFI transition followed by the insulating behavior, further confirming the close linkage between the AFI phase ground state and the terraced surface morphology A similar correspondence was also observed in the 28 nm films annealed at 780 ◦ C for different duration time For the films annealed for and h, the surfaces exhibit a high density of 2D islands, and the ρ-T curves show a characterized PS behavior When the annealing duration goes up to 15 and 20 h, the films turn out to display sharp surface terraces as well as the insulating ρ-T behavior in the whole temperature range These results clearly show that both the high annealing temperature and long duration are quite necessary for inducing the AFI ground state and the terraced surface structure Especially for the 28 nm films, the required annealing duration time is much longer than that for oxygen incorporation and strain relaxation,28, 29 indicating an unusual mechanism lies behind this annealing effect In Fig 5, the annealing induced structure changes were examined by the XRD ω-2θ linear scans and RCs around the LCMO(004) reflections, from the 14 and 28 nm films during the aforementioned accumulative annealing processes During the entire annealing procedures, the film thickness, as calculated by the Laue fringes, remains unchanged And the appearance of the sharp fringes and the narrow full width at half maximum of the RCs exclude any crystal degeneration after annealing.4, 26 As the annealing temperature or the duration time increases, accompanying with the gradually improved surface morphology and enhanced AFI phase, the LCMO(004) reflections first shift to a higher Bragg angle side because of the oxygen uptake [Fig 5(b) and 5(g)], and then barely move after the annealing temperature (duration time) reaches 660 ◦ C (7 h) [Fig 5(c)–4(e) and Fig 5(h)–5(j)], reflecting an already saturated oxygen content.28 Accordingly, it may be concluded that the improvement of film morphology and the oxygen incorporation are two intrinsically different physical processes that occur during the ex-situ annealing In addition, after annealing, the Laue fringes at the higher angle side of the LCMO(004) reflection become depressed, while the ones at the lower angle side remain sharp This asymmetric change is still an open question, and we suggest that it may be related to the anisotropic strain enhanced orthorhombic distortion in the annealed films Based on the results presented above, here we can discuss the role of ex-situ annealing in improving the surface morphology and triggering the AFI phase Previously, the relationship between surface morphology and in-situ film growth mode has been widely reported Specifically, the irregular grainy surface structures manifest the 2-dementional/3-dimentional island growth mode, leading to a high density of defects and strain relaxation.18, 24, 25 On the contrary, the terraced surface structures are commonly observed in the epitaxial films grown in the well-controlled step-flow or layer-by-layer mode, which usually signifies an defect-free structure and coherent strain state.4, 16, 24, 40 Furthermore, the symmetry mismatch, which is originated from the differences in octahedral configurations between the film and substrate, has been reported to increase the surface roughness.5 These facts clearly reveal that the surface morphology could intrinsically manifest the structure coherency between film and substrate Along this line, for the LCMO/NGO(001) films, the presence of atomically flat surface All article content, except where otherwise noted, is licensed under a Creative Commons Attribution 3.0 Unported license See: http://creativecommons.org/licenses/by/3.0/ Downloaded to IP: 158.42.28.33 On: Fri, 12 Dec 2014 11:06:17 052106-9 Wang et al AIP Advances 3, 052106 (2013) FIG ω-2θ scans and RCs measure from the (a)-(e) 14 nm films and (f)-(j) 28 nm films during the accumulative annealing processes The dashed lines are guidance for eyes for the shift of LCMO(004) peaks In (b)-(e) and (g)-(j) the ω-2θ scans of the as-grown films were inserted for comparison terraces suggests that the ex-situ annealing with high temperature and long duration time may not only supply the oxygen uptake but also lead to a recrystallization which assists the true substrate coherency To support this point, we compared the surface morphologies of the LCMO film and the underlying substrate after being cut and annealed separately In detail, one as-received NGO substrate was cut into two pieces and the LCMO film (14 nm) was deposited on one of them Prior to annealing, as shown in Fig 6(a), the atomic terraces on the surface of the as-received NGO substrate are indistinct due to the mixed A-site and B-site surface termination, and the film displays the typical grainy surface (not shown) After being simultaneously annealed side by side at 780 ◦ C for 10 h, however, the substrate [Fig 6(b)] and film [Fig 6(c)] both display atomically flat surfaces with sharp terraces.41 Moreover, the terraces in the film share almost the same width and pattern as those in the annealed substrate, which strongly suggests that the recrystallization during annealing is not just limited in the surface but related to structure coherency of the entire film with the substrate If we take the anisotropic strain as the driving force of the AFI phase, the picture over annealing-assisted substrate coherency can harmoniously explain the correspondence between the surface morphology All article content, except where otherwise noted, is licensed under a Creative Commons Attribution 3.0 Unported license See: http://creativecommons.org/licenses/by/3.0/ Downloaded to IP: 158.42.28.33 On: Fri, 12 Dec 2014 11:06:17 052106-10 Wang et al AIP Advances 3, 052106 (2013) FIG The AFM images (1 × μm2 ) measured from (a) the as-received NGO(001) substrate, (b) the bare NGO(001) substrate annealed at 780 ◦ C for 10 h, and (c) the 14 nm LCMO/NGO(001) film annealed side by side with the bare substrate and electronic phase evolution For the as-grown films, the grainy structures and oxygen vacancies may act as defects and lattice disorders to disrupt the long-range anisotropic strain field, disallowing the JT distortion and facilitating the FM ground state After being annealed, for the thinner films ( a But for the unstrained LCMO films, because of the weaker octahedral tilting and the on-site deformation of MnO6 octahedron, the a axis turns out to be slightly larger than b.9, 44 Consequently, for the commensurate LCMO/NGO(001) films, the anisotropic strain rises from the orthorhombic mismatch between the film and substrate, and due to the “axis-to-axis” growth, the b axis must elongate while a needs to shrink, thus enhancing the orthorhombic lattice distortion of the films.35, 45 In the rest of this paper, we will show that the anisotropic strain mediated AFI phase can be quantitatively understood by mathematically expressing the effects of octahedra rotation/deformation on the response of the in-plane lattice constants (a and b) to the orthorhombic distortion As reported by Tamazyan et al.,46 the in-plane lattice constants of the Pbnm orthorhombic structure with a− a− c+ titling pattern are given by a = 2[r1 cos(π/4 − ϕ) + r2 cos(π/4 + ϕ) − 2r1 cos(π/4 + δ) cos(ϕ + δ)(1 − cos θ )] b = 2[r1 sin(π/4 − ϕ) + r2 sin(π/4 + ϕ)] (1) In these two mathematical expressions the three directions of the Mn-O bonds in one MnO6 octahedron were assumed to be perpendicular with each other for simplicity The r1 and r2 denote the two possible lengths of the B-O bonds that predominantly lie in the ab-plane, and the ϕ and θ describe the octahedra rotation about c axis and the tilting about the R axis which lies in the ab-plane The angle between R and b axis is δ, which is defined by r1 − r2 (2) tan δ = r1 + r2 It is calculated from the octahedral connectivity condition: r1 cos(π/4 + δ) = r2 cos(π/4 − δ) (3) For simplicity, here we just discuss two limiting cases of equation (1) All article content, except where otherwise noted, is licensed under a Creative Commons Attribution 3.0 Unported license See: http://creativecommons.org/licenses/by/3.0/ Downloaded to IP: 158.42.28.33 On: Fri, 12 Dec 2014 11:06:17 052106-11 Wang et al AIP Advances 3, 052106 (2013) First, if we forbid the tilting about the R axis (θ = 0), leading to an a0 a0 c+ rotation system, the equation (1) can be written as a = 2[r1 cos(π/4 − ϕ) + r2 cos(π/4 + ϕ)] b = 2[r1 sin(π/4 − ϕ) + r2 sin(π/4 + ϕ)] (4) When r1 = r2 , the octahedral deformation can be characterized by the Q2 mode JT distortion with the magnitude defined by = r2 – r1 (assuming r2 > r1 ).47 Then the equation (4) can be written as √ a = √2(2r0 cos ϕ − b = 2(2r0 cos ϕ + sin ϕ) , sin ϕ) (5) where the average in-plane Mn-O bond length is defined as r0 = (r1 + r2 )/2 According to equation (5), given a non-zero ϕ and fixed r0 , the b axis will expand with increasing the JT distortion , while the a axis will shrink at the same time That means the orthorhombic distortion can be enhanced by the Q2 mode JT distortion together with the octahedral rotation about c axis.48 In the second case, we limit r1 = r2 = r0 in the a− a− c+ titling system Under this condition, δ turns out to be zero according to (2), and the in-plane lattice constants are given by √ a = 2√2r0 cos ϕ cos θ b = 2r0 cos ϕ (6) In this structure, increasing r0 causes a and b axis to expand equally, while increasing the tilting angle θ only shrinks the a axis Hence, by choosing suitable values of these two parameters, the enhanced orthorhombic lattice distortion (shrinking the a axis and meanwhile elongating the b axis) could be achieved without the Q2 mode JT distortion The above analyses suggest two possible octahedral behaviors to accommodate the anisotropic strain enhanced orthorhombic lattice distortion in the LCMO/NGO(001) films In practice, considering that the in-plane octahedral tilting angle in bulk LCMO is small and the MnO6 octahedra can be easily deformed, we speculate that the JT distortion could dominate the octahedral behavior of LCMO/NGO(001) films under anisotropic strain As illustrated in Fig 7(c), the in-plane tilting for the LCMO films is similar to that in bulk LCMO, while the pronounced Q2 mode JT distortion increases the orthorhombicity and accommodates the anisotropic strain When compared with the bulk La0.5 Ca0.5 MnO3 with a simultaneous AFI phase transition at 180 K,49 the anisotropic strain stabilized JT distortion at room temperature can push the TAFI to 250 K, which is the highest AFI transition temperature for La1-x Cax MnO3 systems However, for the octahedral rotation and deformation pattern of LCMO film near the heterointerface, the octahedral proximity effect should be taken into consideration as well as the anisotropic strain.2 According to the robust in-plane octahedral tilting in NGO substrate, it is anticipated that the octahedral tilting angle of the LCMO interfacial layer will increase, and a different deformation pattern might be stabilized Since the in-plane tilting behavior only shrinks the a axis, the required octahedral deformation is expected to elongate the b axis for accommodating the anisotropic strain According to the analyses based on equation (6), as depicted in Fig 7(d), one possibility is the breathing mode deformation that can equally elongate the in-plane Mn-O bonds,50 whereas the Q2 mode JT distortion could be suppressed by this competing deformation mode The differences of octahedral patterns between the bulk of the film and interfacial layer inevitably lead to a structure gradient along the out-of-plane direction and influence the properties As shown in Fig 7(e), when the thickness is decreased down to nm, the 780 ◦ C/5 h annealed film, albeit displaying the atomically flat surface with terraces, only exhibits an insulating behavior with no sign of AFI phase transition And the ρ-H curves [Fig 7(f)] show a reversible magnetoresistance behavior instead of the AFI melting hysteresis The suppression of AFI phase in the ultrathin film further implies that the octahedral configuration in the LCMO interfacial layer should be quite different from Q2 mode JT distortion in the bulk of the film All article content, except where otherwise noted, is licensed under a Creative Commons Attribution 3.0 Unported license See: http://creativecommons.org/licenses/by/3.0/ Downloaded to IP: 158.42.28.33 On: Fri, 12 Dec 2014 11:06:17 052106-12 Wang et al AIP Advances 3, 052106 (2013) FIG The sketches for lattice distortions in LCMO/NGO(001) films are shown in (a)-(d), in which the red balls (small) stand for oxygen anions, and the green and yellow balls (bigger) stand for manganese and gallium cations respectively For clarity the A-site cations are ignored The a-b plane is indexed in Pbnm setting, as represented by the open rectangles, solid (green) for bulk LCMO and dash (black) for NGO (a) The structure top view of bulk LCMO with small octahedral rotation and tilting angles, and the on-site octahedral deformation is weak (b) The structure top view of NGO substrate, in which the rigid GaO6 octahedra exhibit evident octahedral rotation and tilting (c) Top view and side view for the structure of LCMO film with enhanced Q2 mode Jahn-Teller distortion (d) Top view and side view for a possible structure of the LCMO film interfacial layer with enhanced in-plane octahedral tilting and breathing mode octahedral deformation As marked by the dashed rectangles that represent the a-b plane of NGO substrate, the anisotropic strain can be accommodated for the both structures shown in (c) and (d) (e) The zero-field ρ-T curve and (f) the isothermal ρ-H curves at 250, 200, and 100 K measured from the nm film annealed at 780 ◦ C for h In (e) the zero-field ρ-T curve of annealed 16 nm film was inserted for comparison (the TAFI was marked by the solid arrow), and the inset of (e) shows the AFM image (1 × μm2 ) of the nm film after annealing The arrows in (f) denote the cycling of H (0→ 14→ T) IV CONCLUSIONS In summary, we have systematically investigated the surface morphology and magnetotransport in LCMO/NGO(001) films with various thicknesses after being ex-situ annealed under various annealing parameters The film morphology evolutions demonstrate that the ex-situ annealing can not only eliminate the oxygen deficiency but, more importantly, significantly improve the epitaxial All article content, except where otherwise noted, is licensed under a Creative Commons Attribution 3.0 Unported license See: http://creativecommons.org/licenses/by/3.0/ Downloaded to IP: 158.42.28.33 On: Fri, 12 Dec 2014 11:06:17 052106-13 Wang et al AIP Advances 3, 052106 (2013) quality and even the substrate coherency And the close linkage between the improvement of film morphology and the drastic change of electronic phase after the ex-situ annealing procedure suggests that the substrate coherency triggered octahedral rotations and deformations are crucial in inducing the high-temperature AFI phase in this epitaxial system Although the detailed evolution of octahedral behaviors during annealing needs to be further characterized and clarified in microscale, we believe that the ex-situ annealing process can effectively modulate the structure as well as functionalities, and the approach could be applied even in other correlated electron oxide heterostructures ACKNOWLEDGMENTS This work was supported by the NSF of China (Grant Nos 11074237 and 11274287), and the National Basic Research Program of China (Grant Nos 2009CB929502 and 2012CB927402) J Mannhart and D G Schlom, Science 327, 1607 (2010); P Zubko, S Gariglio, M Gabay, Philippe Ghosez, and J.-M Triscone, Annu Rev Condens Mater Phys 2, 141 (2011); H Y Hwang, Y Iwasa, M Kawasaki, B Keimer, N Nagaosa, and Y Tokura, Nat Mater 11, 103 (2012) J M Rondinelli, S J May, and J W Freeland, MRS Bulletin 37, 261 (2012); J M Rondinelli and N A Spaldin, Adv Mater 23, 3363 (2011); J Chakhalian, A J Millis, and J Rondinelli, Nat Mater 11, 92 (2012) H Rotella, U Lă uders, P.-E Janolin, V H Dao, D Chateigner, R Feyerherm, E Dudzik, and W Prellier, Phys Rev B 85, 18401 (2012); A Vailionis, H Boschker, W Siemons, E P Houwman, D H A Blank, G Rijnders, and G Koster, Phys Rev B 83, 064101 (2011) A J Hatt and N A Spaldin, Phys Rev B 82, 195402 (2010); S J May, J.-W Kim, J M Rondinelli, E Karapetrova, N A Spaldin, A Bhattacharya, and P J Ryan, Phys Rev B 82, 014110 (2010); A J Hatt, N A Spaldin, and C Ederer, Phys Rev B 81, 054109 (2010) K J Choi, S H Baek, H W Jang, L J Belenky, M Lyubchenko, and C.-B Eom, Adv Mater 22, 759 (2010) D L Proffit, H W Jang, S Lee, C T Nelson, X Q Pan, M S Rzchowski, and C.-B Eom, Appl Phys Lett 93, 111912 (2008) S H Chang, Y J Chang, S Y Jang, D W Jeong, C U Jung, Y.-J Kim, J.-S Chung, and T W Noh, Phys Rev B 84, 104101 (2011); D Kan and Y Shimakawa, Cryst Growth Des 11, 5483 (2011) Y Ogimoto, N Takubo, M Nakamura, H Tamaru, M Izumi, and K Miyano, Appl Phys Lett 86, 112513 (2005); M Nakamura, Y Ogimoto, H Tamaru, M Izumi, and K Miyano, Appl Phys Lett 86, 182504 (2005); Y Wakabayashi, D Bizen, Y Kubo, H Nakao, Y Murakami, M Nakamura, Y Ogimoto, K Miyano, and H Sawa, J Phys Soc Jpn 77, 014712 (2008) A T Zayak, X Huang, J B Neaton, and K M Rabe, Phys Rev B 77, 214410 (2008); 74, 094104 (2006) 10 J M Rondinelli and N A Spaldin, Phys Rev B 82, 113402 (2010) 11 J He, A Borisevich, S V Kalinin, S J Pennycook, and S T Pantelides, Phys Rev Lett 105, 227203 (2010) 12 A Y Borisevich, H J Chang, M Huijben, M P Oxley, S Okamoto, M K Niranjan, J D Burton, E Y Tsymbal, Y H Chu, P Yu, R Ramesh, S V Kalinin, and S J Pennycook, Phys Rev Lett 105, 087204 (2010) 13 E Bousquet, M Dawber, N Stucki, C Lichternsteiger, P Hermet, S Gariglio, J.-M Triscone, and P Ghosez, Nature 452, 732 (2008) 14 H Boschker, J Kautz, E P Houwman, W Siemons, D H A Blank, M Huijben, G Koster, A Vailionis, and G Rijinders, Phys Rev Lett 109, 157207 (2012) 15 C L Jia, S B Mi, M Faley, U Poppe, J Schubert, and K Urban, Phys Rev B 79, 081405(R) (2009) 16 S Y Moon, D.-H Kim, H J Chang, J K Choi, C.-Y Kang, H J Choi, S.-H Hong, S.-H Baek, J.-S Kim, and H W Jang, Appl Phys Lett 102, 012903 (2013) 17 M Ziese, H C Semmelhack, and K H Han, Phys Rev B 68, 134444 (2003); L Abad, V Laukhin, S Valencia, A Gaup, W Gudat, L Balcells, and B Mart´ınez, Adv Func Mater 17, 3918 (2007) 18 A Biswas, M Rajeswari, R C Srivastava, Y H Li, T Venkatesan, R L Greene, and A J Millis, Phys Rev B 61, 9665 (2000); M Ziese, H C Semmelhack, K H Han, S P Sena, and H J Blythe, J Appl Phys 91, 9930 (2002); M Bibes, S Valencia, L Balcells, B Mart´ınez, J Fontcuberta, M Wojcik, S Nadolski, and E Jedryka, Phys Rev B 66, 134416 (2002) 19 J Li, C K Ong, J M Liu, Q Huang, and S J Wang, Appl Phys Lett 76, 1051 (2000); J Sakai, N Ito, and S Imai, J Appl Phys 99, 08Q318 (2006) 20 L F Kourkoutis, J H Song, H Y Hwang, and D A Muller, Proc Natl Acad Sci U.S.A 107, 11682 (2010) 21 T Ohnishi, K Shibuya, T Yamamoto, and M Lippmaa, J Appl Phys 103, 103703 (2008) 22 S A Chambers, Adv Mater 22, 219 (2010) 23 G Koster, L Klein, W Siemons, G Rijnders, J S Dodge, C.-B Eom, D H A Blank,and M R Beasley, Rev Mod Phys 84, 253 (2012) 24 J Shin, S V Kalinin, A Y Borisevich, E W Plummer, and A P Baddorf, Appl Phys Lett 91, 202901 (2007) 25 R A Rao, Q Gan, and C.-B Eom, Appl Phys Lett 71, 1171 (1997) 26 S W Jin, G Y Gao, Z Z Yin, Z Huang, X Y Zhou, and W B Wu, Phys Rev B 75, 212401 (2007) 27 M P Siegal, J M Phillips, R B van Dover, T H Tiefel, and J H Marshall, J Appl Phys 68, 6353 (1990) 28 S H Seo, H C Kang, H W Jang, and D Y Noh, Phys Rev B 71, 012412 (2005); P Murugavel, J H Lee, K.-B Lee, J H Park, J.-S Chung, J.-G Yoon, and T W Noh, J Phys D: Appl Phys 35, 3166 (2002) All article content, except where otherwise noted, is licensed under a Creative Commons Attribution 3.0 Unported license See: http://creativecommons.org/licenses/by/3.0/ Downloaded to IP: 158.42.28.33 On: Fri, 12 Dec 2014 11:06:17 052106-14 Wang et al AIP Advances 3, 052106 (2013) 29 G Y Gao, S W Jin, and W B Wu, Appl Phys Lett 71, 1171 (2007) Vasylechko, L Akselrud, W Morgenroth, U Bismayer, A Matkovskii, and D Savytskii, J Alloys Compd 297, 46 (2000) 31 Q Huang, A Santoro, J W Lynn, R W Erwin, J A Borchers, J L Peng, K Ghosh, and R L Greene, Phys Rev B 58, 2684 (1998) 32 Z Huang, G Y Gao, Z Z Yin, X X Feng, Y Z Chen, X R Zhao, J R Sun, and W B Wu, J Appl Phys 105, 113919 (2009); F H Zhang, Z Huang, G Y Gao, P F Chen, L F Wang, X L Tan, and W B Wu, Appl Phys Lett 96, 062507 (2010) 33 L F Wang, Z Huang, X L Tan, P F Chen, B W Zhi, G M Li, and W B Wu, Appl Phys Lett 97, 242507 (2010) 34 Z Huang, L F Wang, X L Tan, P F Chen, G Y Gao, and W B Wu, J Appl Phys 108, 083912 (2010) 35 Z Huang, L F Wang, P F Chen, G Y Gao, X L Tan, B W Zhi, X F Xuan, and W B Wu, Phys Rev B 86, 014410 (2012) 36 Powder diffractions standards: 87-1608 and 87-1084 37 W Wu, C Israel, N Hur, S Park, S.-W Cheong, and A De Lozanne, Nat Mater 5, 881 (2006) 38 H Kuwahara, Y Moritomo, Y Tomioka, A Asamitsu, M Kasai, R Kumai, and Y Tokura, Phys Rev B 56, 9386 (1997) 39 F S´ anchez, G Herranz, I C Infante, J Fontcuberta, M V Garc´ıa-Cuenca, C Ferrater, and M Varela, Appl Phys Lett 85, 1981 (2004) 40 V A Vas’ko, C A Nordman, P A Kraus, V S Achutharaman, A R Ruosi, and A M Goldman, Appl Phys Lett 68, 2571 (1996); M Izumi, Y Konishi, T Nishihara, S Hayashi, M Shinohara, M Kawasaki, and Y Tokura, Appl Phys Lett 73, 2497 (1998) 41 T Ohnishi, K Takahashi, M Nakamura, M Kawasaki, M Yoshimoto, and H Koinuma, Appl Phys Lett 74, 2531 (1999) 42 Z H Wang, O I Lebedev, G Van Tendeloo, G Cristiani, and H.-U Habermeier, Phys Rev B 77, 115330 (2008) 43 E Dagotto, T Hotta, and A Moreo, Phys Rep 344, (2011); Y Tokura, Rep Prog Phys 69, 797 (2006) 44 J.-S Zhou and J B Goodenough, Phys Rev Lett 94, 065501 (2005) 45 A Vailionis, W Siemons, and G Koster, Appl Phys Lett 93, 051909 (2008); A Vasudevarao, S Denev, M D Biegalski, Y Li, L.-Q Chen, S Trolier-McKinstry, D G Schlom, and V Gopalan, Appl Phys Lett 92, 192902 (2008) 46 R Tamazyan and S van Smaalen, Acta Cryst B 63, 190 (2007) 47 J.-S Zhou and J B Goodenough, Phys Rev B 77, 132104 (2008) 48 M W Lufaso and P M Woodward, Acta Cryst B 60, 10 (2004) 49 P G Radaelli, D E Cox, M Marezio, S.-W Cheong, P E Schiffer, and A P Ramirez, Phys Rev Lett 75, 4488 (1995) 50 A J Millis, Phys Rev B 53, 8434 (1996); K Yamamoto, T Kimura, T Ishikawa, T Katsufuji, and Y Tokura, Phys Rev B 61, 14706 (2000) 30 L All article content, except where otherwise noted, is licensed under a Creative Commons Attribution 3.0 Unported license See: http://creativecommons.org/licenses/by/3.0/ Downloaded to IP: 158.42.28.33 On: Fri, 12 Dec 2014 11:06:17