On the Origin of Nanochessboard Superlattices in A-Site-Deficient Ca

Subscriber access provided by UNIV LIBRE DE BRUXELLES

Article

On the origin of nano-chessboard superlattices in A-site deficient Ca-stabilized Nd TiO 2/3

3

Feridoon Azough, Despoina Maria (Demie) Kepaptsoglou, Quentin M. Ramasse, Bernhard Schaffer, and Robert Freer Chem. Mater., Just Accepted Manuscript • DOI: 10.1021/cm5036985 • Publication Date (Web): 22 Dec 2014 Downloaded from http://pubs.acs.org on December 28, 2014

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Chemistry of Materials is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Chemistry of Materials

On the origin of nano-chessboard super-lattices in A-site deficient Ca-stabilized Nd2/3TiO3 Feridoon Azough†, Demie KepaptsoglouØ,*, Quentin M. RamasseØ, Berhnard SchafferØ and Robert Freer† †

School of Materials, Materials Science Centre, University of Manchester, Manchester M1 7HS, United Kindgom. Ø

SuperSTEM Laboratory, STFC Daresbury Campus, Keckwick Lane, Warrington WA4 4AD, United Kingdom.

ABSTRACT A-site deficient Nd2/3TiO3 ceramics stabilized with CaTiO3, with an overall composition of 0.9 Nd2/3TiO3 – 0.1 CaTiO3, were synthesized by the mixed oxide route. Synchrotron X-ray diffraction was used to identify the basic perovskite structure and revealed cross-type superlattice reflections. An incommensurate super-lattice structure with dimensions of a≈b≈20ap and c=2ap (where ap is the cell parameter for the parent perovskite phase) was identified, giving rise to contrast features resembling a nano-chessboard pattern in electron microscopy images. The super-lattice was further characterized by aberration-corrected scanning transmission electron microscopy (STEM): atomically-resolved lattice images were obtained along orientations to visualize the A-site (Ca, Nd and vacancies) and B-site (Ti) cation column intensities, in correlation with observations of the nano-chessboard super-lattice. Electron energy loss spectroscopy (EELS) was used to precisely determine the distribution of Nd and Ca across the structure, confirming the absence of long range elemental segregation or phase separation across the nano-chessboard super-structure. Closer inspection of the chemical maps in two orthogonal

1 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

directions suggests however the presence of localized ordering of cations and vacancies. The chessboard pattern super-lattice is thus likely to be caused by periodic octahedral tilt distortions of the O sub-lattice, possibly induced by these short-range chemical variations, as a result of a complex interplay between cation and vacancy ordering in three dimensions. KEYWORDS: A-site deficient perovskites, nano-chessboard, diamond frame contrast, cation-vacancy ordering, scanning transmission electron microscopy, electron energy loss spectroscopy


Page 2 of 44

Page 3 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1. Introduction Many classes of electronic oxide ceramics are dominated by ABO3 perovskite and perovskiterelated compounds. Among those are the A-site deficient perovskites. Their crystal structure and cation – vacancy ordering are of interest for potentially attractive properties of ionic conductivity1, dielectric behaviour2 and transport properties3. The Nd2/3TiO3 solid solution with CaTiO3 is a candidate for microwave applications4. In addition, we have recently reported that it can be considered as a thermoelectric ceramic for high temperature applications due to its high Seebeck coefficient and very low, temperature-stable thermal conductivity5. Investigations of such materials are starting to reveal interesting and important features at the atomic scale. Importantly, a number of these ceramics show ‘cross-type’ satellite reflections in their [001] diffraction patterns indicating the presence of a two-dimensional superstructure, which is likely to contribute to the macroscopic properties of the bulk material. Indeed, in recent studies it has been demonstrated that self-assembled Zn(Mn, Ga)O4 with a two dimensional nano-chessboard superstructure can effectively reduce thermal conductivity6. Since the 1930’s this type of superstructure has been reported to occur in intermetallic compounds. Using electron microscopy, it has been demonstrated that the formation of the cross-like satellite reflections is due to long range ordering and is associated with the formation of nano-size domains in the microstructure7. The electron microscopy studies of this type of ordering have been summarized by Amelinckx et al.8. For oxide ceramics, this type of superstructure was first reported for Th0.25NbO39; a series of seminal papers linked the observed cross type splitting of the satellite reflections to a micro domain model, comprising of a system of periodically tilted oxygen octahedra9–11.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

More recently, a family of A-site deficient perovskites showing cross type satellite reflection, namely Ln2/3-xLi3x□1/3-xTiO3 (Ln=La or Nd and where □ denotes A-site vacancies), has been attracting attention with extensive studies by advanced electron microscopy and radiation techniques. Fourquet et al.1 examined the La-based compound for 0.06≤ x ≤0.14 and reported weak cross-type satellite reflections for the composition x=0.08 although the initial high resolution transmission electron microscopy (HRTEM) images did not reveal the presence of a chessboard-type super-structure. More recently, however, Guiton et al. investigated the A-site deficient (Nd2/3-xLi3x)TiO3 system by HRTEM imaging and nm-resolution chemically-sensitive Z-contrast imaging in STEM and they were the first to report a ‘nano-chessboard’ arrangement in [001] HRTEM images, the sides of the pattern being parallel to the (100) and (010) lattice directions12–14. They also observed within the chessboard blocks a diamond-shaped contrast feature with edges parallel to (110) and (1-10) directions. Intriguingly, in Z-contrast images, the diamond-shaped areas exhibited a brighter intensity compared to the surrounding square-shaped domains. They therefore concluded that the nano-chessboard pattern is a result of phase separation into Li-rich square domains, separated by Li-poor boundary regions. Using the relative size of the diamond-shaped areas and of the square-shaped bocks and the overall composition of their material, but without chemical analysis via spectroscopy, they proposed compositions equivalent to Li1/2Nd1/2TiO3 and Nd2/3TiO3 for the Li-rich and Li-poor regions, respectively. They suggested that the two-dimensional structural modulation and the resulting modulation in the lattice parameter would lead to a two-dimensional strain modulation which could cause the light and dark chessboard contrast in TEM. It was also proposed that this strain modulation could likely be responsible for the occurrence of what was termed by the authors “a periodic octahedral tilt twinning” in this system. Following this pioneering contribution to the


Page 4 of 44

Page 5 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


understanding of the crystal structure of this family of perovskites12–14, Withers et al.15 very recently turned their attention to Li0.5-3xNd0.5+xTiO3 with x=0.067. Using Z-contrast STEM lattice images along [001] they also proposed a number of likely structural models, concluding in relative agreement with the earlier work of Guiton et al.14, that the chessboard super-structure must be the result of Li phase separation giving rise to a combination of a Li ion sub-lattice and an octahedral tilt-twinned sub-structure. A small controversy surrounding the interpretation of the nano-chessboard patterns observed in this family of materials became apparent however when Abakumov et al.16–20 reported their study of the super-structure of Li3xNd2/3-xTiO3 with x=0.05. Using annular dark field STEM imaging, synchrotron X-ray and neutron powder diffraction, as well as ab initio structure relaxation, their particularly detailed crystallographic analysis very convincingly suggests that the chemical phase separation proposed by Guiton et al.14 and apparently confirmed in part by Withers et al.15, does not occur in this system and that the super-structure must instead be understood in terms of displacement modulations related to an intricate octahedral tilting pattern mainly affecting the O sub-lattice. Although none of the samples studied by these different groups were entirely equivalent and the possibility therefore exists that different mechanisms are at play to explain the observed nanochessboard contrast, the very strong similarities between the systems would suggest a common structure and a common source to the observed contrast. It is therefore essential to carry out further analysis to elucidate the origin of the nano-chessboard patterns. The recent, apparently contradictory reports on Li3xNd2/3-xTiO3 were arguably limited by several factors. Firstly, focusing on a Li-containing compound makes any atomic-resolution electron microscopy study particularly difficult: Li-based materials are notoriously electron beam sensitive21 and as a result 5 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

any quantitative image contrast and/or spectroscopic data analysis risks being limited by sample damage. Additionally, Li is particularly difficult to visualize in high angle annular dark field imaging due to its very low atomic number. Furthermore, none of the aforementioned studies was able to include atomic level chemical analysis using electron energy loss or energy dispersive X-ray spectroscopy, as both techniques are limited in their Li detection ability. Finally, most of the reported observations were often limited to only one major zone axis: for such a complex system, any proposed three-dimensional crystal structure should ideally take into account structural and chemical information obtained along all major zone axes. In the light of these important issues, and in order to clarify the occurrence of the phase separation which has been the main subject of discussion in the recent publications, we report here on complementary observations of a nano-chessboard structure in A-site deficient Nd2/3TiO3 stabilized with CaTiO3. The Ca-containing counterpart to the Nd2/3-xLi3xTiO3 structure that has proved so difficult to elucidate was here found to be particularly stable under the electron beam, while all its chemical constituents lent themselves perfectly to atomic-level analytical electron microscopy investigations. We are therefore able to provide a full characterization using synchrotron radiation source data, as well as advanced electron microscopy which includes for the first time atomic level electron energy loss spectroscopy to analyze the distribution of cations and vacancies along all lattice directions.

2. Experimental Nd0.6Ca0.1□0.3TiO3 (where □ denotes A-site vacancies) ceramics were prepared by the conventional mixed oxide route. Starting materials were high purity powders of CaCO3 (Solvay, 99.5%), Nd2O3 (AMR Limited, 99.9%) and TiO2 (Tioxide, 99.9%). The Nd2O3 powder was dried 6 ACS Paragon Plus Environment

Page 6 of 44

Page 7 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


at 900°C for 6 hours prior to use. The powders were weighed in batches according to the required formulation and wet-milled for 24 hours in a vibratory mill using zirconia balls and propan-2-ol. The powders were then dried at 85oC for 24 hours and calcinated at 1100oC for 4 hours. Powders were uniaxially compacted into pellets of 20mm diameter and 15mm thickness at a pressure of 50MPa prior to sintering at 1450oC for 4 hours in air. The heating and cooling rates were set at 180oC/hr.; samples were sintered in alumina crucibles with sacrificial powder below the sample to prevent any contamination from the substrate. High resolution X-ray data were collected at the Diamond Light Source. The data were collected in flat-plate geometry with a fixed 5° angle (α) of the sample face with respect to the incident beam. An X-ray wavelength of 0.827442Å was used. Microstructures were initially examined by scanning electron microscopy (Philips FEGSEM XL30). Samples were ground using 400, 800 and 1200 grade SiC and polished using 6µm and 1µm diamond pastes. The final polishing stage employed an Oxide Polishing Suspension (OPS). Samples for TEM and STEM investigations were prepared by both ion beam thinning and crushing techniques. For ion beam-thinning, specimens were first ground on 1200 grade SiC to reduce the thickness to ~300µm. They were ultrasonically cut into 3mm diameter disks (model KT150; Kerry Ultrasonic Ltd.) then dimpled (model D500; VCR Group) to reduce the thickness of the center of the disk to 30µm. Finally, the disks were ion beam thinned (using a Gatan precision ion polishing system model 691; PIPSTM) operating at 4-6kV. For the crushing method, the sintered disks were crushed into a powder using an agate mortar and pestle. Grains of individual powders were dispersed in chloroform, dropped onto a copper grid covered with a holey carbon film, and then dried. Structures were initially investigated using selected area electron diffraction (SAED) and high-resolution transmission electron microscopy (HRTEM) 7 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

techniques using an FEI FEGTEM (Tecnai G2) operating at 300kV. Subsequently, atomicresolution structural characterization was carried out using a Nion UltraSTEMTM100 aberrationcorrected dedicated STEM, equipped with a Gatan Enfina spectrometer. The microscope was operated at 100kV acceleration voltage and the probe-forming optics were configured to form a ~0.9Å probe (full width at half-maximum) with a convergence angle of 30mrad and a probe current of 120pA. The semi-angular ranges of high angle annular dark field and medium angle annular dark field detectors were 86-190 and 40-86mrad, respectively. The native energy spread of the electron probe was 0.35eV and the collection semi-angle for the electron energy loss spectroscopy measurements was 36mrad.

Figure 1. SEM image of the Nd0.6Ca0.1□0.3TiO3 ceramic sintered at 1450oC. The grains in are general equiaxed in shape, and no evidence of a second phase can be seen.


Page 8 of 44

Page 9 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 2. X-ray diffraction pattern of the NT-CT sample: while the refinement of the data does not converge with acceptable enough confidence, the best fit is obtained using Pmmm as the spacegroup. The intensity doublets between 20-21 degrees (marked by the dashed line in the plot and shown as insert), are consistently observed in other nano-chessboard compounds such as Nd2/3-xLi3xTiO3.

3. Results & Discussion The Nd0.6Ca0.1□0.3TiO3 ceramics (NT-CT) sintered at 1450oC attained at least 95% of their theoretical density. A typical SEM micrograph of the ceramic microstructure is shown in Figure 1. There is no evidence of any secondary phase, while porosity is located at the grain boundaries or trapped as large pores within the grains. The grains in are general equiaxed in shape, with a bimodal grain size distribution; most of the individual grains are in the 5-40µm size



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 44

range. Twin boundaries, arising from a phase transformation from the high temperature form to the low temperature form of the ceramic, can be clearly seen within the grains. An X-ray diffraction pattern of the NT-CT sample is shown in Figure 2. Consistent with the SEM microstructure observations there is no evidence of a secondary phase. According to previous investigations employing a range of radiation techniques, the majority of A-site cationdeficient perovskites, regardless of vacancy content and of the number of different elements populating the A-site [for example Ln2/3TiO3 where Ln=(La, Pr and Nd), Th0.25NbO3, La1/3NbO3, and Ln2/3-xA3xO3 where (Ln=La or Nd, A=Na or Li)], tend to adopt one of four types of crystal structure for their basic unit cell3,9–11,22–31. These are namely the orthorombic Pmmm, tetragonal P4/mmm, orthorhombic Cmmm and monoclinic C2/m space groups. Both Pmmm (a≈b≈ap, c≈2ap) and Cmmm (a≈b≈c≈2ap) structures have been reported for Nd2/3TiO3 stabilized with additives, or prepared in a reducing atmosphere24,25,32, while there is only one report for NdTiO3-NdAlO3 adopting the monoclinic C2/m crystal structure25. Although a full refinement of our X-ray data is beyond the scope of this paper (work to that end is currently ongoing and along with complementary neutron diffraction data will be the subject of a following publication), it is interesting to report that initial refinements based on a simple perovskite structure for both Pmmm (a≈b≈ap, c≈2ap) and Cmmm (a≈b≈c≈2ap) space groups did not give acceptable R-values, although a slightly better fit was obtained in the case of Pmmm, yielding lattice parameters a=3.835311Å, b=3.837193 Å and c=7.721441 Å (obtained through a Le Bail fit). This is hardly surprising, as previous work on very similar compounds12,16 has shown that in order to refine the X-ray data with an acceptable tolerance much more complicated descriptions of the structure have to be considered. It is useful however to note the presence of two doublets at 2θ values between 20 to 21 degrees (Figure 2 and inset graph within): this specific feature has been 10 ACS Paragon Plus Environment

Page 11 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


previously observed in very similar structures such as (Nd2/3-xLi3x)TiO3 and is believed to be closely associated with intricately modulated octahedral tilting distortions of the O sub-lattice16.

Figure 3. HRTEM image along the [001] zone axis, showing the chessboard nano-structure and SAED pattern showing the ‘cross-type’ satellite reflections (inset). The nano-chessboard size is approximately 8 x 8 nm. A high resolution transmission electron microscopy (HRTEM) image taken with the electron beam parallel to the [001] zone axis is shown in Figure 3. Nanometer-scale contrast can be seen in the HRTEM image, where a combination of bright and dimmer square blocks, with edges parallel to the (010) and (100) directions, forms a chessboard pattern with a periodicity of approximately 8 by 8 nanometers. This type of image was observed in all the grains examined and the super-cell size was the same in various areas of the grains. No apparent variation in the size or shape of the super-cell was observed. The uniformity of such microstructures throughout was also evident in the report by Guiton et al.14 for the Nd2/3-xLi3xTiO3 system. In the work by 11 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 44

Abakumov et al.16 the size of the nano-chessboard domains appears to vary along the adirection, while Withers et al.15 described a Nd2/3-xLi3xTiO3 structure that exhibited some ‘striped’ domains in addition to the nano-chessboard pattern. In the corresponding SAED patterns shown in the inset of Figure 3, it is possible to observe strong so-called ‘cross-type’ satellite reflections

9,10

(around the , spots and the ½ , h,k odd, positions)

corresponding to the nano-chessboard modulations in the NT-CT sample. The separation of these ‘cross-type ‘satellite reflections is the same in both directions and corresponds to about 20ap-20ap or an 80Å by 80Å super-lattice, consistent with the (8nm x 8nm) size of chessboard pattern observed in the HRTEM images. The absence of any reflection conditions in the [001] electron diffraction pattern (and in accordance with X-ray data) suggests a Pmmm space group as the basic structure. These initial X-ray, SAED and HREM results thus indicate that our NT-CT compound is closely related and indeed exhibits a near identical average structure to the Nd2/3xLi3xTiO3 compositions

described previously14–16.

A series of atomically resolved STEM images recorded along the NT-CT [001] zone axis is presented in Figure 4. The nano-chessboard pattern is again clearly observed in the bright field (BF) images (Figure 4a, b), whose contrast is closely related to that of HREM micrographs such as the one shown in Figure 3. Simultaneously acquired high angle annular dark field (HAADF) images (Figure 4c,d) also reveal associated intensity modulations, although as discussed by other authors in previous work this diamond-shaped pattern is much fainter in Z-contrast images than in BF images and its visibility diminishes with decreasing sample thickness. Figure 4d for instance corresponds to a reasonably thick part of the sample (estimated as ~45nm using low loss


Page 13 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 4. (a, b) BF STEM images of [001] NT-CT showing a nano-chessboard contrast; (c, d) simultaneously-acquired HAADF images (a-c, b-d, respectively), showing a faint wire frame contrast modulation; (e) HAADF image of [001] NT-CT acquired at ~60nm under-focus showing enhanced wireframe contrast; (f) MAADF STEM image (acquired at 0nm defocus) of the same area as (e) showing a bright diamond wire frame pattern contrast. electron energy loss spectroscopy33) and the intensity modulations, while still present, are very difficult to distinguish with the naked eye. This contrast, akin to a diamond wireframe pattern can be significantly enhanced all the way to the edge of the sample when the HAADF images are acquired at relatively large defocus values for which the fine features in the images are smeared out and the probe channeling conditions are not met: see for example an HAADF image acquired at 60nm of under-focus in Figure 4e of the very same area shown in Figure 4e. More



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 44

interestingly, the diamond-shaped pattern is particularly evident when using an annular detector with a much smaller inner angle: see Figure 4f, recorded on the exact same area as Figure 4e. Diffraction effects contribute strongly to the contrast formation mechanism of these medium angle annular dark field images (MAADF), and they are thus very sensitive to strain34,35, which, as was suggested earlier16,18,19 must play a role in the appearance of the nano-chessboard patterns. The precise origin of the nano-chessboard or diamond frame patterns in A-site deficient perovskites has been the object of debate among authors. Despite the modulation having been observed in systems with a single A-site cation10,36, several authors14,15,37 attribute the nature of the contrast observed in images of mixed A-site cation compounds fully, or partly, to a phase separation process. Guiton et al.14, basing their argument on the intensity of HAADF images of their Li2/3-xNd3xTiO3 compounds, suggested the presence of Li-poor and Li-rich phases in the dark diamond wire frame and brighter areas observed in their samples, respectively. In a similar fashion, Garcia-Martin et al.37 describe the nano-chessboard contrast they observed in KLaMnWO6 double-ordered perovskites

(and nano-stripes

contrast in NaLaMnWO6

perovskites38) in terms of a compositional modulation and phase separation process within the A cation sub-lattice; in their proposed model, two domains with different cation contents (for La and K) are periodically arranged in a nano-chessboard pattern. In complete contrast, however, Abakumov et al.16, using very complex X-ray refinement analysis as well as image simulations and a series of annular dark field images acquired with different detector annular ranges, counter-argued that the faint contrast variation observed their HAADF images cannot be reproduced using the previously proposed phase-separated structures. They demonstrate that the


Page 15 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 5. (a) MAADF STEM image of [001] NT-CT. The area selected for 2D EELS mapping is marked with a yellow square; (b) MAADF signal during the EELS acquisitions and maps of Ca, Ti and Nd; (c) integrated spectrum from the spectrum image. diamond wire frame contrast disappears almost entirely in the experimental images acquired at higher annular ranges and in thin regions of the sample, conditions for which the image contrast scales, to a good approximation, as the square of the atomic number Z (the actual dependence is closer to Zn, n=1.6-1.8, here denoted as ~Z2 for simplicity) and these images are therefore often denoted as Z-contrast images. Additionally, depending on the annular ranges used for imaging, diffraction or strain contrast39 can contribute to the image formation and give the impression of chemical variation. Moreover, they argue that the wireframe or chessboard patterns reported in Refs.

14,37

are visible mostly in relatively thick areas of samples, where the HAADF image

intensity should not anymore be expected follow a relatively pure ~Z2 dependence40. The observations and arguments of Abakumov et al.16 are in good agreement with our own experimental observations (Figure 4), although we should point out that with Ca (ZCa=20) being 15 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 44

about 6.5 times heavier than Li (ZLi=3), the approximate Ca/Nd (ZNd=60), HAADF intensity ratio is only (ZNd/ZCa)2=9, compared to (ZNd/ZLi)2=400 for the Li-containing compounds (after Abakumov et al.16). Hence, any possible phase separation in Nd0.6Ca0.1□0.3TiO3 is potentially more difficult to confirm from imaging alone. In order to further investigate the possibility of a nano-domain phase separation in the NT-CT compound, we therefore turn to electron energy loss spectroscopy (EELS) chemical mapping. Large scale chemical maps were acquired in multiple areas of the sample varying in thickness, by rastering the electron probe serially across a defined region and collecting an EEL spectrum at each point. Chemical maps are created by integrating at each point of these spectrum images the spectrum intensity over an ~40eV window above the Ca L2,3, Ti L2,3, O K and Nd M4,5 EELS edge onsets after background subtraction using a power law model. As the intensity of the EELS signal is not exempt from multiple scattering, channeling or diffraction effects, great care should be taken when treating or interpreting the collected data. In order to minimize the effect of elastic scattering in the EELS signal the spectra were acquired at a relatively large collection angle (36 mrad) and the effect of multiple scattering were removed through the Fourier ratio method33 using low loss EELS spectrum images of identical dimension, acquired subsequently in the exact same areas as the core-loss data. Figure 5 is representative of data acquired in a region where the diamond wireframe shape is relatively hard to distinguish in HAADF images but can still be clearly seen in the more strain-sensitive MAADF images (Figure 5a, b). The sample relative thickness was estimated as 80% of the inelastic mean free path λ of the 100keV electrons within the material (0.8t/λ, corresponding approximately to 45nm). In the resulting atomically-resolved chemical maps no appreciable modulation in intensity is visible throughout the structure (for reading clarity, the map intensities are also plotted as integrated line profiles in Figure S1 of the 16 ACS Paragon Plus Environment

Page 17 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Supplementary Information provided). Similar results were systematically obtained across the sample, in both thin and thicker regions, when in the latter a simple Z-contrast interpretation would have suggested a very strong Nd/Ca content variation. These observations thus provide clear evidence that no nano-domain phase separation is present in our sample. This result is in complete disagreement with previous phase separation models, and in particular with La N4,5 jump-ratio-images, which appear to show a periodic modulation of the La content in a KLaMnWO6 compound along the diamond wire frame contrast pattern37. Although this earlier work dealt with a different (albeit related) compound which may well exhibit true chemical modulations, we note that the energy filtered images presented in the work by Garcia-Martin et al.37 were acquired in rather thick areas of their sample and were not treated for multiple scattering effects. This in combination with the collection conditions used (in particular the use of a low energy edge, which can still be heavily affected by elastic scattering) strongly suggests that the observed contrast in their La maps could have been affected by the elastic signal and thus possibly be an artefact. We therefore suggest that the absence of any chemically separated domains in our NT-CT sample supports the argument that the nano-chessboard or wire frame contrast observed in this (and likely in several other systems such as Li2/3-xNd3xTiO3) originates in fact from other sources, such as strain induced by the oxygen sub-lattice octahedral tilting and distortions16. Interestingly however, the cross-type satellite reflections characteristic of the nano-chessboard modulations are preserved in the Fourier transform diffractogram of the [001] HAADF images obtained in extremely thin areas of the sample and with very high detector inner angle (Figure 6b). The structural features giving rise to the cross-type reflections must be preserved in the HAADF images and as the O atomic positions are not typically visible in Z-contrast images the ‘cross17 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 44

type’ satellite reflections in HAADF diffractograms (and a chessboard-like pattern when the FFT is inversed using only those reflections, Figure 6c) are therefore likely to be associated with modulations of the cation column intensities, as well as with any potential shifts in the cation positions. Although it was not possible to reliably measure any such shifts directly in our images such an explanation would of course still be perfectly consistent with the incommensurate O sublattice tilting modulation, the former possibly being the result (and/or the cause) of the latter. Alario-Franco et al.9 suggest for instance that vacancies and cation displacements could follow a sinusoidal or incommensurate periodicity, while Abakumov et al.

26

in their model predict both

A-site cation shifts as well as second order B-site displacements in the form of second order Jahn-Teller distortions.


Page 19 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 6. (a, d-f): HAADF STEM images of [001] NT-CT showing an intensity variation of the A-site columns in a and b directions; (b) Fourier transform diffractogram of (a) showing the cross-type super-lattice reflections (encircled); (c) inverse Fourier transform of (b) using only cross-type super-lattice reflections. Indeed, closer inspection of the atomically resolved HAADF images of very thin areas of the sample provides some further clues about the structure. In the HAADF images shown in Figure 6d, f the brighter spots correspond to the A-site atomic columns (Nd with Z=60, Ca with Z=20 and vacancies) and the darker spots are the B-site atomic columns (Ti with Z=22) of the perovskite structure. Although no obvious contrast from the nano-chessboard or diamond frame patterns is visible in the images with the naked eye, the intensity of the A-site columns shows variations in both a and b directions. The structure consists of few very dark A-site columns and 19 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 44

short segments of brighter A-site columns that are arranged in a ‘zig-zag’ manner, with alternating segments of slightly darker A-site columns. This intensity variation between neighboring columns is indicative of some degree of local (short range) compositional ordering on the A-site. Indeed the arrangement is reminiscent of the "irregular tweed pattern" described by Labeau et al.10 for Th0.25NbO3 double perovskites and the nano-chessboard patterns in the NTCT sample could thus relate, as in Th0.25NbO3, to A-site cation-vacancy ordering.

Figure 7. (a): HAADF STEM image of [001] NT-CT, showing the outline of the area where a 2D EELS spectrum image was acquired (marked with a square); (b) HAADF signal during the EELS acquisition and atomically resolved maps of Ca (c), Ti (d) and Nd (e). Atomically-resolved EELS chemical maps of the thinnest possible areas of the sample (Figure 7) can help investigate this hypothesis. As the thickness of the sample did not exceed ~0.3t/λ (17nm) consecutive acquisitions in the same areas were avoided to prevent any beam-induced damage (as a mere precaution as no damage was visibly observed in any of the experiments described here) and as a result the data were not further treated for multiple scattering effects which are far less prominent for such thin regions. The chemical maps reveal that all A-sites contain Nd (Figure 7e). As expected, the brighter columns observed are Nd rich, while the darker 20 ACS Paragon Plus Environment

Page 21 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


columns appear to contain significantly less Nd, with the lowest map intensity coinciding with the darker columns of the HAADF images. Similarly, the Ca L2,3 map (Figure 7c) shows a nonuniform distribution throughout the lattice, but while there are a number of lattice sites with increased Ca concentration, the overall Ca distribution in the lattice does not appear to correlate directly (in phase or anti-phase) to the Nd distribution. Furthermore, this compositional inhomogeneity does not appear to exceed more than 2-3 atomic distances and there is no indication of long-range lateral correlation. This is a strong indication that the observed A-site intensity variation in the [001] HAADF images is not only caused by a difference in the total cation content of the A-site columns but perhaps also by the cation/vacancy ratio therein a possibility that has so far not been systematically considered by other authors who focused on the effects of long range occupancy/compositional modulations clearly ruled out by our large scale EELS data. The occurrence of a degree of local cation/vacancy ordering can be confirmed by considering the chemical distribution along the [010] zone axis of the structure, for which the two independent A-sites, A1 and A2, lie in projection in different columns. We must emphasize that the term "ordering" is used here to denote a very local effect rather than a systematic crystallographic arrangement. In [010] selected area electron diffraction (SAED) patterns (Figures 8a, b) the presence of reflections at ½(001)p suggests a doubling of the c-axis periodicity. Earlier radiation technique-based investigations22,23 proposed that such a doubling of the c-axis in A-site deficient perovskites originates in vacancy-cation ordering as their common crystal structural feature41,42. In addition to the ½(001)p reflections, satellite reflections along the a-direction can be seen in Figure 8b. The separation of the satellite reflections in real space is about 40Å, which corresponds perfectly to the length of each domain in the chessboard pattern (e.g. Figure 3). This 21 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 44

one dimensional modulation can if fact be observed as ~4nm stripes in the corresponding HRTEM images (see Figure S2 in Supplementary Information provided). Furthermore, very weak streaking and very weak diffraction spots at ½(100)p can be seen Figure 8b. These weak features in the [010] diffraction patterns have previously been attributed to clustering and pairing of vacancies or pairing of fully occupied A-sites43,44, supporting further our hypothesis of some degree of cation-vacancy ordering in NT-CT. We should note that a similar doubling of the caxis periodicity in A-site deficient perovskites has been reported in many systems: in La2/3TiO3 stabilized by SrTiO343 or for Sr1-3/2xLaxTiO344, where Sr also segregates mostly to the partially occupied A2-site. Very few of these systems exhibit a nano-chessboard contrast, however.

Figure 8. (a-b): [010] SAED diffraction patterns showing ½(001)p reflections and satellite reflections along the a-direction; (c-f) [010] HAADF STEM images.


Page 23 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


HAADF images acquired in the [010] orientation (Figure 8c-f) confirm the presence of two distinct A-sites with strikingly different column intensities, in agreement with the doubling of the c-axis observed in the SAED patterns. From the image intensity alone, the A1-sites appear to be fully occupied (by Nd and/or Ca), while the A2-sites show a variable intensity; some A2-site columns have a similar intensity to the A1-sites, while others are comparatively dark, which again is suggestive of a shared occupancy of cations with vacancies. Atomically-resolved STEM-EELS maps of these features (Figure 9) show that the A1-sites are consistently occupied by Nd with almost constant occupancy, while the Nd content of the A2-sites is quite variable. A2-site columns appearing brighter in the simultaneously acquired HAADF image contain more Nd (Figure 9b) and darker columns contain less Nd, with some A2-site columns appearing to be almost entirely vacant. Ca on the other hand is found almost exclusively on the A2-sites. As with the maps acquired in the [001] orientation, there does not appear to be any direct correlation to the distribution of Nd: a Nd-poor A2-site does not systematically have more Ca. This means that the vacancy content is not uniform among A2-sites, and the observed column intensities in the images must therefore be a complex function of the Nd, Ca and vacancy content therein rather than a simple Nd vs. Ca phase separation (for reading clarity, the map intensities are also plotted as integrated line profiles in Figure S3 of the Supplementary Information provided).



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 9. (a): HAADF STEM image of [010] NT-CT, showing the outline of the area where a 2D EELS spectrum image was acquired (marked with a square); (b) HAADF signal during the EELS acquisition and atomically resolved maps of Ca, Ti and Nd; (c) integrated spectrum from the spectrum image. As an additional and perhaps more subtle effect observed specifically in this case where a nanochessboard structure is stabilized, fully occupied A2-sites are paired along the c-direction, creating 'segments' of brighter intensity whose length varies between 7Å (2ap) to 64Å (16ap): see dashed yellow boxes on Figure 8d. Similarly, some clustering of dark columns or vacancies is also evident in the sample (Figure 8d, red dashed boxes). This clustering or pairing causes a localized doubling of the a- and b-axes and is the likely reason for the weak diffraction spots at ½(100)p. It manifests itself in the orthogonal [001] projection as the irregular tweed pattern mentioned earlier which is visible in the thinnest areas of the sample (the segments are only a few nm long) and is a continuation of the nano-chessboard pattern clearly observed in thicker 24 ACS Paragon Plus Environment

Page 24 of 44

Page 25 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


areas (see Figure 6). Although the ½(100)p spots associated with this pairing are weak and diffuse in the [010] SAED patterns, suggesting short correlation-length features, their common periodicity with the nano-chessboard pattern suggests that in thicker areas of the sample the strain fields associated with such clustering could in turn contribute to the nano-chessboard pattern itself. Although the stabilization mechanism for nano-chessboard structures is not yet known, our data and the observation that such additional pairing in the orthogonal direction seems to only occurs in compositions exhibiting a nano-chessboard structure, suggest that in Asite deficient perovskites, in addition to O octahedral tilting, it likely involves a complex interplay between the segregation of the additive element to the partially occupied sites (Ca cations here, or Li in Li2/3-xNd3xTiO3 samples) and, importantly, vacancy ordering. The idea of cation-vacancy ordering is not new in A-site deficient perovskites. Labau et al.9,10 have proposed micro domain models of cation/vacancy ordering for Th0.25NbO3, along with octahedral tilting in (001)p layers, to explain certain observed super-lattice reflections. Due to the extreme complexity of the systems, this possibility was not systematically considered by previous authors for Nd2/3TiO3 type compounds, although Guiton et al.14, for example, study a range of compositions with vacancies from ~4 to ~23% of the A-sites. Similarly in the compositions studied by Withers et al.15 and Abakumov et al.16, about ~17% and ~23% of the Asites are vacant, respectively. Since all the vacancies are located in the (001)1/2 atomic plane therefore the actual number of vacancies in (001)1/2 plane is around 22% and 30% for the two above mentioned references respectively. Therefore due to this high content, it is likely an ordered arrangement between A-site cations and vacancies will occur. Furthermore, there is some evidence that in the (Li, Nd, Ti, O) compositions studied by Withers et al.15 and Abakumov et al.16, the formation of cross-like satellite reflection is dependent on the vacancy content; for 25 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 44

example in Li(0.5-3x)Nd(0.5+x)TiO3 ceramics the split cross satellite reflections when x is greater than 0.05 [unpublished data].

Figure 10. (a): HAADF STEM image of [010] NT-CT, showing the outline of the area where a 2D EELS spectrum image was acquired (marked with a square). (b): HAADF signal during the EELS acquisition; (c) Ti L2,3 and (d) O K spectra from selected B-site columns. The local cation vacancy ordering, evidenced by presence of strings of mostly vacant or mostly occupied A2-sites ,suggests that associated local distortion of the oxygen sub-lattice should be created, which either add to or form part of the tilting distortion system derived by Abakumov et al.16 The octahedral distortion and change in the Ti bonding environment depending on the occupancy of the A2-site is in fact evidenced by subtle changes in the electron energy loss. This is most clearly demonstrated by comparing the fine structure of the Ti L2,3 edge of B-site columns, adjacent to some mostly vacant (Ti3 and Ti4 in Figure 10 b,c), or mostly occupied A2sites (Ti1 and Ti2, in Figure 10d), as well as the associated O K fine structure. 45 The Ti L2,3 edge (Figure 10c) arises from electron excitations from Ti 2p3/2 and Ti 2p1/2 into Ti 3d states (L3 and L2 peaks respectively) and typically exhibits a splitting of the L3 and L2 EELS peaks into t2g and eg sub-peaks due to the crystal field splitting

46–50

. We observe a small increase in the L2/L3


Page 27 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


intensity ratio from the mostly full A2 column accompanied by a decrease in the eg /t2g ratio (Figure 10b) compared to the mostly vacant position, indicating a weaker crystal field splitting, consistent with longer Ti-O bond lengths as a result of a distortion in the TiO6 octahedron46,48–50. While this trend is consistently observed between some of the mostly vacant and mostly full A2sites (see additional data in Figure S4 of the Supplementary information provided), the degree of this change is not uniform throughout; for example positions marked as Ti3 and Ti4 in Figure 10b located on opposite sides of the same A2 ‘dark’ column are not entirely identical. This is of course in good agreement with the localized nature of the variation in the vacancy/cation content of the A2 sites, which should consequentially affect differently the local bonding environment of different B-sites and hence the degree of distortion of TiO6 octahedra. The corresponding O K spectra (Figure 10d) show two distinct peaks. Peak A at ~ 3eV above the edge onset is attributed to hybridized O 2p states with Ti 3d states

45,51

and peak B at ~8eV above the onset arises from

contributing Ti 3d states and states from the A-site cations (Nd 5d and/or Ca 4d) hybridized with O 2p states

45,46,49,52,53

. The relative increase of the intensity of peak B (relative to peak A) is

systematically observed close to the full A2-sites, consistent with a stronger contribution from Nd (and/or Ca) columns, when by contrast the contribution is weaker when there are neighboring vacant A2-sites

25,45,54

. These complementary observations are very much consistent with the

concept of a network of incommensurately tilted TiO6 octahedra proposed by Abakumov et al.16 and competing second order Jahn-Teller distortions of the TiO6 octahedra, which appear to be localized around the vacant A2-sites. 4. Conclusions Using state of the art analytical scanning transmission electron microscopy techniques, we have demonstrated beyond any doubt that the nano-chessboard or wire frame contrast observed in Ca27 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 44

stabilized A-site deficient titanates does not originate from chemical phase separation into nanodomains. Instead, closer inspection of the images and atomically resolved EELS chemical maps in two orthogonal directions suggest that in the Nd0.6Ca0.1□0.3TiO3 system, Ca predominantly occupies Nd-vacancy shared sites, creating locally a higher occupation of the site and thus promoting vacancy-cation ordering in both a and b lattice directions. These observations corroborate previous studies, which suggest that the observed contrast in electron micrographs is a result of strain originating in intricately-modulated octahedral tilting distortions of the O sublattice combined with local cation-vacancy pairing. These detailed microstructure results could be beneficial to understand electrical properties of materials with nano-chessboard superstructure and to design the assembly of nano-scale functional ceramics.

Acknowledgements We gratefully acknowledge the financial support of the Engineering and Physical Sciences Research Council (EPSRC) through awards EP/H043462, EP/I036230 and grant R113738. The SuperSTEM Laboratory is the U.K. National Facility for Aberration-Corrected STEM, supported by the EPSRC. The authors greatly acknowledge Prof. C. Tang for organizing the collection of high resolution X-ray data in station I11 of Diamond light source, as well as Dr. Rolf Erni (EMPA - Swiss Federal Laboratories for Materials Science and Technology, Switzerland) and Prof. Rik Brydson (University of Leeds, U.K.) for very open and fruitful discussions on the results and the manuscript.


Page 29 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


ASSOCIATED CONTENT

Supporting information available: MAADF STEM survey image and line integrated MAADF and EES signal intensities; MAADF STEM survey images and line integrated MAADF and EELS signal intensities, HRTEM images. This material is available free of charge via the internet at http://pubs.acs.org.

AUTHOR INFORMATION Corresponding author *SuperSTEM Laboratory, STFC Daresbury Campus, Keckwick Lane, WA4 4AD Daresbury, United Kingdom. Email: [email protected]; Tel: +44 (0)1925 864 902 Notes The authors declare no competing financial interest.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

REFERENCES (1)

Fourquet, J. L.; Duroy, H.; Crosnier-Lopez, M. P. J Solid State Chem 1996, 127, 283–294.

(2)

Suvorov, D.; Valant, M.; Skapin, S.; Kolar, D. J Mater Sci 1998, 33, 85–89.

(3)

Kim, I.-S.; Nakamura, T.; Inaguma, Y.; Itoh, M. J Solid State Chem 1994, 113, 281–288.

(4)

Yoshida, M.; Hara, N.; Takada, T.; Seki, A. Jpn J Appl Phys 1997, 36, 6818 LP – 6823.

(5)

Jackson, S.; Azough, F.; Freer, R. J Electron Mater 2014, 43, 2331–2336.

(6)

Kosuga, A.; Kurosaki, K.; Yubuta, K.; Charoenphakdee, A.; Yamanaka, S.; Funahashi, R. Jpn J Appl Phys 2009, 48, 10201.

(7)

Terasaki, O.; Watanabe, D. Jpn J Appl Phys 1981, 20, L381.

(8)

Amelinckx, S.; Van Tendeloo, G.; Van Landuyt, J. Bull Mater Sci 1984, 6, 417–452.

(9)

Alario-Franco, M. A.; Grey, I. E.; Joubert, J. C.; Vincent, H.; Labeau, M. Acta Crystallogr Sect A 1982, 38, 177–186.

(10)

Labeau, M.; Grey, I. E.; Joubert, J. C.; Vincent, H.; Alario-Franco, M. A. Acta Crystallogr Sect A 1982, 38, 753–761.

(11)

Chakhmouradian, A. R.; Mitchell, R. H.; Burns, P. C. J Alloys Compd 2000, 307, 149– 156.

(12)

Guiton, B. S.; Wu, H.; Davies, P. K. Chem Mater 2008, 20, 2860–2862.

(13)

Guiton, B. S.; Davies, P. K. J Am Chem Soc 2008, 130, 17168–17173.

(14)

Guiton, B. S.; Davies, P. K. Nat Mater 2007, 6, 586–591.

(15)

Withers, R. L.; Bourgeois, L.; Snashall, A.; Liu, Y.; Norén, L.; Dwyer, C.; Etheridge, J. Chem Mater 2013, 25, 190–201.

(16)

Abakumov, A. M.; Erni, R.; Tsirlin, A. A.; Rossell, M. D.; Batuk, D.; Nénert, G.; Tendeloo, G. Van. Chem Mater 2013, 25, 2670–2683.

(17)

García-Martín, S.; Urones-Garrote, E.; King, G.; Woodward, P. Chem Mater 2013, 26, 1286–1287.

(18)

Abakumov, A. M.; Erni, R.; Tsirlin, A. A. Chem Mater 2014, 26, 1288.


Page 30 of 44

Page 31 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(19)

Erni, R.; Abakumov, A. M.; Rossell, M. D.; Batuk, D.; Tsirlin, A. A.; Nenert, G.; Van Tendeloo, G. Nat Mater 2014, 13, 216–217.

(20)

Davies, P. K.; Guiton, B. S. Nat Mater 2014, 13, 217–218.

(21)

Rossell, M. D.; Erni, R.; Asta, M.; Radmilovic, V.; Dahmen, U. Phys Rev B 2009, 80, 24110.

(22)

Ali, R.; Izumi, F.; Yashima, M. J Am Ceram Soc 2006, 89, 3805–3811.

(23)

Howard, C. J.; Zhang, Z. J Phys Condens Matter 2003, 15, 4543.

(24)

Howard, C. J.; Knight, K. S.; Zhang, Z.; Lumpkin, G. R. Acta Crystallogr Sect B 2006, 62, 60–67.

(25)

Sefat, A. S.; Amow, G.; Wu, M.-Y.; Botton, G. A.; Greedan, J. E. J Solid State Chem 2005, 178, 1008–1016.

(26)

Abakumov, A. M.; Shpanchenko, R. V.; Antipov, E. V. Mater Res Bull 1995, 30, 97–103.

(27)

Carrillo, L.; Villafuerte-Castrejón, M. E.; González, G.; Sansores, L. E.; Bucio, L.; Duque, J.; Pomés, R. J Mater Sci 2000, 35, 3047–3052.

(28)

Kennedy, B. J.; Howard, C. J.; Kubota, Y.; Kato, K. J Solid State Chem 2004, 177, 4552– 4556.

(29)

García-Martín, S.; Alario-Franco, M. A. J Solid State Chem 1999, 148, 93–99.

(30)

Robertson, A. D.; Martin, S. G.; Coats, A.; West, A. R. J Mater Chem 1995, 5, 1405– 1412.

(31)

Smith, R. I.; Skakle, J. M. S.; Mather, G. C.; Morales, M.; West, A. R. Mater Sci Forum 1996, 228-231, 701–704.

(32)

Jung, W.-H. Mater Lett 2005, 59, 2408–2411.

(33)

Egerton, R. F. Reports Prog Phys 2009, 72, 016502.

(34)

Perovic, D. D.; Rossouw, C. J.; Howie, A. Ultramicroscopy 1993, 52, 353–359.

(35)

Rossell, M. D.; Abakumov, A. M.; Ramasse, Q. M.; Erni, R. ACS Nano 2013, 7, 3078– 3085.

(36)

Zhou, Q.; Saines, P. J.; Sharma, N.; Ting, J.; Kennedy, B. J.; Zhang, Z.; Withers, R. L.; Wallwork, K. S. Chem Mater 2008, 20, 6666–6676.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(37)

Garcia-Martin, S.; King, G.; Urones-Garrote, E.; Nénert, G.; Woodward, P. M. Chem Mater 2011, 23, 163–170.

(38)

García-Martín, S.; Urones-Garrote, E.; Knapp, M. C.; King, G.; Woodward, P. M. J Am Chem Soc 2008, 130, 15028–15037.

(39)

Hillyard, S.; Silcox, J. Ultramicroscopy 1995, 58, 6–17.

(40)

Hartel, P.; Rose, H.; Dinges, C. Ultramicroscopy 1996, 63, 93–114.

(41)

King, G.; Garcia-Martin, S.; Woodward, P. M. Acta Crystallogr Sect B 2009, 65, 676– 683.

(42)

Azough, F.; Freer, R.; Schaffer, B. J Am Ceram Soc 2010, 93, 1237–1240.

(43)

Kepaptsoglou, D. M.; Azough, F.; Ramasse, Q.; Schaffer, B.; Sarahan, M.; Jackson, S.; Freer, R. Microsc Microanal 2013, 19, 1918–1919.

(44)

Battle, P. D.; Bennett, J. E.; Sloan, J.; Tilley, R. J. D.; Vente, J. F. J Solid State Chem 2000, 149, 360–369.

(45)

De Groot, F. M. F.; Grioni, M.; Fuggle, J. C.; Ghijsen, J.; Sawatzky, G. A.; Petersen, H. Phys Rev B 1989, 40, 5715–5723.

(46)

Eberg, E.; van Helvoort, A. T. J.; Takahashi, R.; Gass, M.; Mendis, B.; Bleloch, A.; Holmestad, R.; Tybell, T. J Appl Phys 2011, 109, 034104.

(47)

Prabhumirashi, P.; Dravid, V. P.; Lupini, A. R.; Chisholm, M. F.; Pennycook, S. J. Appl Phys Lett 2005, 87, 121917.

(48)

Zhang, Z.; Sigle, W.; Rühle, M. Phys Rev B 2002, 66, 94108.

(49)

Stoyanov, E.; Langenhorst, F.; Steinle-Neumann, G. Am Mineral 2007, 92, 577–586.

(50)

De Groot, F. M. F.; Fuggle, J. C.; Thole, B. T.; Sawatzky, G. A. Phys Rev B 1990, 41, 928–937.

(51)

Bugnet, M.; Radtke, G.; Botton, G. A. Phys Rev B 2013, 88, 201107.

(52)

Guo, X. .; Chen, X. .; Lu, W. Solid State Commun 2003, 126, 441–446.

(53)

Radtke, G.; Botton, G. A. Microsc Microanal 2004, 10, 852–853.

(54)

Brydson, R.; Sauer, H.; Engel, W.; Hofer, F. J Phys Condens Matter 1992, 4, 3429.


Page 32 of 44

Page 33 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


TOC GRAPHIC



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

SEM image of the NdCa0.1TiO3 ceramic sintered at 1450oC. The grains in are general equiaxed in shape, and no evidence of a second phase can be seen. 99x65mm (300 x 300 DPI)

ACS Paragon Plus Environment

Page 34 of 44

Page 35 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


X-ray diffraction pattern of the NT-CT sample: while the refinement of the data does not converge with acceptable enough confidence, the best fit is obtained for Pmmm. Intensity doublets between 20-21 degrees (marked by the dashed line in the plot and shown as insert), are consistent with two-dimensional periodic octahedral tilt twinning. 75x53mm (300 x 300 DPI)



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

HRTEM image along the [001] zone axis, showing the chessboard nano-structure and SAED pattern showing the ‘cross-type’ satellite reflections (inset). The nano-chessboard size is approximately 8 x 8 nm. 75x75mm (300 x 300 DPI)


Page 36 of 44

Page 37 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(a, b) BF STEM images of [001] NT-CT showing a nano-chessboard contrast; (c, d) simultaneously-acquired HAADF images (a-c, b-d, respectively), showing a faint wire frame contrast modulation; (e) HAADF image of [001] NT-CT acquired at ~60nm under-focus showing enhanced wireframe contrast; (f) MAADF STEM image (acquired at 0nm defocus) of the same area as (e) showing a bright diamond wire frame pattern contrast. 152x101mm (300 x 300 DPI)



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(a) MAADF STEM image of [001] NT-CT. The area selected for 2D EELS mapping is marked with a yellow square; (b) MAADF signal during the EELS acquisitions and maps of Ca, Ti and Nd; (c) integrated spectrum from the spectrum image. 225x111mm (300 x 300 DPI)


Page 38 of 44

Page 39 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(a, d-f): HAADF STEM images of [001] NT-CT showing an intensity variation of the A-site columns in a and b directions; (b) Fourier transform diffractogram of (a) showing the cross-type super-lattice reflections (encircled); (c) inverse Fourier transform of (b) using only cross-type super-lattice reflections. 120x80mm (300 x 300 DPI)



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(a): HAADF STEM image of [001] NT-CT, showing the outline of the area where a 2D EELS spectrum image was acquired (marked with a square); (b) HAADF signal during the EELS acquisition and atomically resolved maps of Ca (c), Ti (d) and Nd (e). 258x86mm (300 x 300 DPI)


Page 40 of 44

Page 41 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(a-b): [010] SAED diffraction patterns showing ½(001)p reflections and satellite reflections along the adirection; (c-f) [010] HAADF STEM images. 242x161mm (299 x 299 DPI)



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(a): HAADF STEM image of [010] NT-CT, showing the outline of the area where a 2D EELS spectrum image was acquired (marked with a square); (b) HAADF signal during the EELS acquisition and atomically resolved maps of Ca, Ti and Nd; (c) integrated spectrum from the spectrum image. 217x117mm (300 x 300 DPI)


Page 42 of 44

Page 43 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(a): HAADF STEM image of [010] NT-CT, showing the outline of the area where a 2D EELS spectrum image was acquired (marked with a square). (b): HAADF signal during the EELS acquisition; (c) Ti L2,3 and (d) O K spectra from selected B-site columns. 175x63mm (300 x 300 DPI)



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

171x110mm (299 x 299 DPI)


Page 44 of 44

On the Origin of Nanochessboard Superlattices in A-Site-Deficient Ca

Recommend Documents