ARTICLE https://doi.org/10.1038/s41467-021-22916-4 OPEN A high-entropy manganite in an ordered nanocomposite for long-term application in solid oxide cells F. Baiutti 1 ✉, F. Chiabrera 1, M. Acosta 2, D. Diercks 3, D. Parfitt4, J. Santiso 5, X. Wang6, A. Cavallaro 7, A. Morata1, H. Wang 6, A. Chroneos4, J. MacManus-Driscoll 2 & A. Tarancon 1,8 ✉ 1234567890():,; The implementation of nano-engineered composite oxides opens up the way towards the development of a novel class of functional materials with enhanced electrochemical prop- erties. Here we report on the realization of vertically aligned nanocomposites of lanthanum strontium manganite and doped ceria with straight applicability as functional layers in high- temperature energy conversion devices. By a detailed analysis using complementary state-of- the-art techniques, which include atom-probe tomography combined with oxygen isotopic exchange, we assess the local structural and electrochemical functionalities and we allow direct observation of local fast oxygen diffusion pathways. The resulting ordered mesos- tructure, which is characterized by a coherent, dense array of vertical interfaces, shows high electrochemically activity and suppressed dopant segregation. The latter is ascribed to spontaneous cationic intermixing enabling lattice stabilization, according to density functional theory calculations. This work highlights the relevance of local disorder and long-range arrangements for functional oxides nano-engineering and introduces an advanced method for the local analysis of mass transport phenomena. 1 Catalonia Institute for Energy Research (IREC), Jardins de Les Dones de Negre 1, Sant Adrià del Besòs, Barcelona, Spain. 2 Department of Materials Science and Metallurgy, University of Cambridge, Cambridge, UK. 3 Department of Metallurgical and Materials Engineering, Colorado School of Mines, Golden, CO, USA. 4 Faculty of Engineering, Environment and Computing, Coventry University, Coventry, UK. 5 Catalan Institute of Nanoscience and Nanotechnology, ICN2, CSIC and The Barcelona Institute of Science and Technology (BIST), Campus UAB, Bellaterra, Barcelona, Spain. 6 School of Materials Engineering, Purdue University, West Lafayette, IN, USA. 7 Department of Materials, Imperial College London, London, UK. 8 ICREA, 23 Passeig Lluís Companys, Barcelona, Spain. ✉email: fbaiutti@irec.cat; atarancon@irec.cat NATURE COMMUNICATIONS | (2021)12:2660 | https://doi.org/10.1038/s41467-021-22916-4 | www.nature.com/naturecommunications 1 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-22916-4 S olid Oxide Cells (SOCs), which are capable of reversibly electrochemical applications: Yoon et al31. and Cho et al32. pro- converting chemical to electrical energy with high theore- posed similar solutions, based on self-assembled MIEC-fluorite tical efficiency and a minimized footprint1,2, can play a thin film vertical nanocomposites, for application as functional pivotal role in the widespread implementation of next-generation layers in fuel cells, while Ma et al33. studied the fundamental energy devices. By means of modern fabrication technologies, properties of Co-based VANs, highlighting enhanced ORR SOCs can take advantage of beyond-state-of-the-art concepts for kinetics. Although such works demonstrate the great potential of materials’ nanoengineering, this way achieving enhanced per- the VAN approach, to the best of the authors’ knowledge, so far formance and further downscaling3. To date, however, sluggish inert model substrates had to be employed for the realization of electrode catalytic activity, poor thermochemical stability, and ideal structures, while the use of “non-ideal” supports such as real limitations in the effective implementation of advanced nanoscale electrolyte layers irremediably led to the formation of arrange- materials, have hindered the scalability of SOCs technology and ments with higher disorder (i.e., incoherent interfaces and low its viability as a portable energy source4. In order to improve the interface density), possibly inhibiting the full exploitations of performance of SOC air electrode without using expensive and VANs in real electrochemical devices. Notably, the possibility of rare noble metals, the development of functional ceramic fabricating superior MIEC nanostructures by design goes hand- nanostructures based on mixed ionic-electronic conducting oxi- in-hand with the development of techniques which are capable of des (MIEC) represents a promising approach5. In MIEC materials capturing nanoscale phenomena with higher accuracy. (e.g., doped perovskite manganites, cobaltites, ferrites), a good In the present work, we describe the fabrication of a self- trade-off between the free energy of the oxygen redox cycle and assembled thin film VAN nanocomposite as an ad hoc- the enthalpy of oxide reduction can be achieved, ensuring high engineered MIEC. The nominal constituting phases are a electrocatalytic activity on the whole free surface together with mainly electronic conductor such as La0.8Sr0.2MnO3 (LSM) and fast oxygen ion transport6. Most importantly, when MIECs are an ionic conductor, namely Ce0.8Sm0.2O2 (SDC). The structure employed in the form of thin films, the dense microstructure possesses long-range order, out-of-plane epitaxy and intimate allows for excellent in-plane percolation, while the possibility of phase alternation (column width ≈10 nm) for maximized triple- nanostructuring via cutting-edge thin film techniques opens up phase boundary density (>106 cm cm−2). The fabrication uses an the path toward the exploitation of nanoscale effects such as local yttria-stabilized zirconia (YSZ) electrolyte as a support—we stress fast oxygen reduction kinetics and transport7. Although the here that such a support is considered as an ideal platform for the implementation of MIEC-based thin films for the fabrication of widespread implementation of functional oxide nanomaterials in SOC functional layers is of great potential e.g., for avoiding the devices34—and results in the combination of key features for occurrence of areas with high electronic current density or high high-temperature electrochemical applications: accelerated ORR oxygen chemical potential gradients which may potentially lead kinetics and, most importantly, high stability against to device failure8, intrinsic limitations still occur especially in temperature-driven cationic migration. The nanostructure has relation to a poor thermal stability owing to dopant segregation therefore straight applicability as a SOC functional electrode toward the surface. Limiting cationic migration is of paramount layer. The system is also ideal for implementing complementary importance in order to avoid the possible formation of insulating state-of-the art and novel techniques for locally assessing struc- secondary phases and to maintain the surface activity over tural and functional aspects. Direct observation of a local fast time9,10. State-of-the-art strategies for enhancing MIEC thin film mass transport pathway along the SDC matrix (<10 nm wide) has performance comprise strain engineering via epitaxial been obtained by isotopic oxygen labelling combined with atom- growth11,12, chemical doping13, introduction of an interface (so- probe tomography (APT), achieving superior resolution for called heterogeneous doping)14–16, or surface decoration17. Such describing oxygen diffusion. methods, however, are often limited to model systems with no technological applicability, and very few of them have proven to be successful in simultaneously improving the initial oxygen Results and discussion reduction reaction (ORR) kinetics while mitigating detrimental Structural characterization highlights long-range order. A surface cationic segregation18. sketch of the LSM-SDC VAN mesostructure, composed by LSM In order to enhance the mass transport kinetics at intermediate pillars in an SDC continuous matrix, is depicted in Fig. 1 (panel temperatures, MIEC are combined to pure oxygen conductors to a), together scanning transmission electron microscopy (STEM— form (nano)composites, which find wide application not just as panels b–e) and high-resolution x-ray diffraction (XRD—panel f). SOCs electrodes, but also as dual phase membranes for oxygen In the lower-magnification high-angle angular dark field separation owing to the ability of transporting both ionic and (HAADF) STEM image of a thin cross-sectional lamella (Fig. 1b), electronic species while maintaining structural stability19–21. one can appreciate the thin columnar, long-range ordered, Classical approaches for composite fabrication lead however to mesostructure for large portions of the film (column width is as the formation of disordered structures, which suffer from thermal narrow as ≈5 nm, length ≈100 nm, no tortuosity). One can also instabilities and limited percolation for mass and charge see that no porosity is present. The precise sequence of alternating transport22,23. A highly potential alternative for the fabrication of (La, Sr, Mn)- and Ce-based phases can be clearly identified in nanocomposite electrode functional layers is represented by Fig. 1c, where a HAADF image and an Energy-dispersive X-ray vertically aligned nanostructures (VANs), which are characterized spectroscopy (EDS) color maps of the area highlighted in Fig. 1b by an off-plane, nanoscale phase alternation24. Nano-engineered are shown. In the HAADF image, owing to the higher Z numbers VANs have been extensively investigated in the last years: Model of Ce and Sm, SDC columns have higher contrast than those of systems, generally characterized by ‘nanocolumn in matrix’ LSM. Interestingly, the color map underlines not only the clear mesostructures (domain-matching epitaxy)25,26, have revealed separation between the LSM columns (dominated by Sr, Mn) and unexpected properties including e.g., enhanced ionic the SDC matrix (Ce, Sm), but also a certain degree of cationic conductivity27, controlled resistive switching28 and improved intermixing especially between La, Sm and Ce. A quantitative and photocatalytic activity29. Such functionalities have been ascribed theoretical analysis on cationic redistribution will be discussed in to the entanglement of interface effects including elastic strain detail in later paragraphs. In Fig. 1c, it is also interesting to notice and ionic or electronic redistribution30. Some reports also deal that the substrate-film interface is characterized by an even more with effective nanocomposite layers or interlayers for pronounced intermixing and that no phase separation is 2 NATURE COMMUNICATIONS | (2021)12:2660 | https://doi.org/10.1038/s41467-021-22916-4 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-22916-4 ARTICLE a c 40 nm b Film 100 nm Substrate d e f 5 nm 5 nm Fig. 1 Structural investigation of LSM-SDC. a Sketch of the LSM-SDC VAN structure. b Low-magnification image of the film, highlighting the long-range order of the pillar structure. The thickness decrease on the right side should be ascribed to TEM lamella preparation. c HAADF image and EDS colour maps for a representative columnar region of the film. d HR-TEM of the alternating SDC and LSM columnar structure. e IFFT of (d). The atomic ordering and the epitaxial relationship between the LSM (out-of-plane [110], zone axis [1-10]) and the SDC (out-of-plane [001], zone axis [100]) regions is clearly visible, as schematically sketched by the bottom inset. In the top inset, the original FFT and the filtered reflections (highlighted by the red circles) are shown. f RSM of the out-of-plane matching (001)-oriented SDC domains (404 and 204 SDC reflections) and (110)-orientedLSM domains (222 and 221 reflections). The 404 and 202 YSZ substrate reflections are also visible. The dotted horizontal line passes through the maxima of the LSM and SDC reflections, highlighting perfect out-of- plane lattice match between the two phases. Please note that the structural investigation was carried out after thermal ageing (700 °C, 100 h). apparent. We ascribed such a finding to the structural (lattice Fig. 1c), in agreement with previous reports and as also confirmed match) and chemical (wettability) properties of the fluorite YSZ by the simulations presented later in the text (see below DFT substrate which favors the formation of a single, Ce-rich, phase section)24. Please refer to Supplementary Note 1 for the complete during the first stages of film growth. XRD characterization of the films. The HR-TEM of a representative columnar region is shown in Fig. 1d, where one can notice a perfect local order. The atomic arrangement can be further analyzed in Fig. 1e (inverse filtered Observing oxygen diffusion pathways and cationic intermixing fast Fourier transform (IFFT) of the region in Fig. 1d): A by APT. Complementary to HR-TEM, the use of Atom Probe structurally coherent relationship is established out-of-plane (z- Tomography (APT) allowed us to obtain meaningful insights on direction) between the two phases. Also, no evidence of other the structure chemistry and on the oxygen kinetics, as summar- phases or of spurious orientations are detected in these regions. ized in Fig. 2. The technique, which is based on the Time-of- The in-plane and out-of-plane orientation relationships between Flight (ToF) analysis of atoms evaporated through a pulsed laser the film phases and the substrate in such highly ordered majority impinging a highly biased sample tip, possesses unique cap- portions are LSM (001) || SDC (100) || YSZ (100) and LSM (110) abilities for 3D imaging and precise quantitative chemical com- || SDC (001) || YSZ (001), respectively, as sketched in the bottom- position analysis with sub-nm resolution35. In Fig. 2a, b, left inset of panel e. A dedicated XRD analysis by means of projections of the spatial distribution of Mn (red—representative reciprocal space mapping (XRD-RSM) of the matching planes in of LSM) and Ce (light blue—for SDC) over a sample tip the z-direction i.e., (110) for LSM and (001) for SDC is shown in (HAADF-STEM in Supplementary Fig. 2) are reported. Con- Fig. 1f. Here, the alignment of the LSM and SDC reflection in z sistently with the TEM analysis (cf. Fig. 1), one can observe the confirms the lattice match epitaxy in the out-of-plane direction at presence of well-defined LSM pillars, immersed in an SDC the global scale. Interestingly, the resulting lattice spacing for matrix. The composition of the two phases has been carefully LSM (dx = 3.874 Å and dz = 2.739 Å for the in-plane and for the analyzed along a region crossing multiple LSM and SDC domains z-direction, respectively, vs d = 3.906 Å for bulk LSM in (blue cylinder in Fig. 2a, b). The resulting elemental distribution pseudocubic notation) indicates an isotropic compressive defor- is reported in Fig. 2c and confirms the modulation of Mn, Sr and mation of the cell (−0.82%). (Please note here that dz refers to the Ce when crossing each phase. One challenge in APT analysis of (110) direction and the corresponding out-of-plane lattice multiphase materials is that differences in the evaporation fields parameter is dz ∙ √2 = 3.874 Å.) The reduced lattice parameter is of the materials can result in changes in the local radius of cur- consistent with cationic substitution (Ce and Sm replacing La—cf. vature near the interfaces of the phases, which produce ion NATURE COMMUNICATIONS | (2021)12:2660 | https://doi.org/10.1038/s41467-021-22916-4 | www.nature.com/naturecommunications 3 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-22916-4 Fig. 2 APT on LSM-SDC. Lateral (a) and bottom view (b) 3D renderings, highlighting the areas of LSM (red surface) and SDC (light blue dots) using a 30 at% La+Sr+Mn isoconcentration surface. c Cationic atomic fraction along the linescan indicated by the blue cylinder in (a) and (b). d Concentration color maps for Ce, Mn, 18O and 16O from the slice indicated by the gray dotted line in (b). e Top: Experimental (blue—obtained upon integration of the area in the rectangle in (d) and simulated (red) 18O fraction profiles. Bottom: Mn concentration profile. Table 1 Elemental atomic fractions in LSM and SDC. LSM SDC Counts Cat.% Counts Cat.% A-site LSM Mn 22790 43.5 ± 0.5 La 50705 20.9 ± 0.7 Ce 8227 15.7 ± 0.6 Mn 12420 5.1 ± 0.7 La 11374 21.7 ± 0.6 Sr 2610 1.1 ± 0.2 Sr 6588 12.6 ± 0.5 Ce 150310 61.9 ± 0.8 Sm 3389 6.5 ± 0.4 Sm 26978 11.1 ± 0.4 These were calculated from the core regions of each phase, excluding the interfacial regions. The uncertainty values are taken from dividing the LSM phase into 7 regions and the SDC phase into 30 regions and taking the standard deviations of the compositions from each of the individual regions. trajectories that deviate from that presumed in generating the more details on the method. In Supplementary Table 1, we also reconstruction of the analyzed volume36. The result is apparent report the results on the quantification of atomic fractions for intermixing of the higher evaporation field phase (SDC in this single phase LSM and SDC films that we performed in order to case) into the lower evaporation field phase (LSM). Previous assess the accuracy of the APT method under different mea- research has indicated that the extent of this overlap is generally surement (laser energy) conditions. We note here that, while up to about 2 nm around the phase interfaces and that for regions TEM imaging and spectroscopy result in a direct phase/compo- larger than 2 nm, the “core” of the lower evaporation field phase sition analysis on large portions of a cross-sectional specimen, that excludes the outer 2 nm of material retains the accurate APT is able to provide complementary local information via a 3- composition of that phase36,37. Even with this complication, the dimensional reconstruction, with enhanced ability to quantita- non-negligible intermixing of the other cations is clear (in tively assess local elemental distributions. agreement with the TEM observations—cf. Fig. 1) and results in While the observed level of intermixing is expected to have the substitution of La by Ce and Sm in LSM (15.7 ± 0.6 and 6.5 ± only a limited impact on the oxygen mobility in SDC owing to 0.4 cat. %, respectively), while a large La content is found in SDC dopant excess (≈32 at.% Sm+La)38, the replacement of La in LSM (20.9 ± 0.7 cat. %). The measured elemental atomic fraction is by Ce and Sm plays a key role, according to our theoretical reported in Table 1. Please refer to Supplementary Note 2 for calculations, in the oustanding thermal stability of the structure as 4 NATURE COMMUNICATIONS | (2021)12:2660 | https://doi.org/10.1038/s41467-021-22916-4 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-22916-4 ARTICLE discussed later (see below DFT). As reported previously and as at the LSM-SDC interface), as a direct consequence of lateral resulting from steric considerations, Ce is expected to be mostly oxygen diffusion from the SDC, while vertical transport in LSM is present as an isovalent substitutional dopant for La in LSM, thus negligible (cf. also Supplementary Note 3). Similar results on a not direcly affecting the electrochemical transport properties39. different 2D cross-section are reported in Supplementary Fig. 4. Taking advantage of the highly ordered and dense micro- The resulting surface exchange coefficient of the VAN is structure of our VANs, we employed APT as a techinque for remarkably higher than the one of LSM (k*~1*10−11 cm2*s−1) analyzing the electrochemical functionalites of the VAN func- and also than the one reported for related LSM-YSZ dense tional layer at the nanoscale (oxygen mass transport properties) composites (k*~2*10−10 cm2*s−1) at the same temperature45,46. via direct observation of oxygen diffusion pathways in VANs As far as the diffusion coefficient of LSM is concerned, this is fully exchanged with oxygen isotopes (18O) under controlled condi- consistent electrochemical measurements by impedance spectro- tions of temperature and time. The sensitivity of APT toward scopy (D*impedance = 5.4 ∙ 10−17 cm2 ∙ s−1) and with literature oxygen isotopes has been demonstrated previously40. Unlike data (cf. following section and Supplementary Fig. 8). conventional Isotope Exchange Depth Profiling (IEDP) by ToF Secondary Ion Mass Spectrometry41, which is commonly employed in bulk and homogeneous materials, our approach is Electrochemical performance. Complementary electrochemical characterized by a greatly enhanced lateral resolution down to the characterization was carried out by impedance spectroscopy (EIS) <10 nm-level (vs > 100 nm for IEDP)42. Figure 2d represents a in conventional two-electrode supported electrolyte cells (Au- snapshot of the oxygen incorporation in a LSM-SDC VAN after LSM-SDC VAN/YSZ/Ag) and is summarized in Fig. 3. 6000 s of exchange with labelled oxygen at T = 550 °C, showing In Fig. 3a, exemplary Nyquist plots for LSM-SDC and concentration maps of 18O and 16O oxygen atoms, alongside the analogous LSM cells are reported (T = 750 °C, air atmosphere). different cations, as obtained from APT. In the top panels, one LSM exhibits the typical polarization arc of MIEC materials can immediately recognize the LSM pillars, indicated by the stemming from a co-limited process, i.e., ambipolar diffusion presence of high Mn concentration areas, surrounded by a ceria (≈45° straight element at high frequencies) and ORR (semicircle matrix. In the bottom panels, areas of maximum 18O concentra- at low frequencies)47. One can appreciate that in LSM-SDC both tion, corresponding to oxygen incorporated during the exchange, contributions are drastically reduced (cf. also Supplementary are clearly correlated to the SDC matrix while LSM pillars present Fig. 9 for comparison of the area-specific resistance—ASR). The much lower concentration. This provides direct display of the polarization arc exhibits an almost perfect impedance arc— presence of a preferential local oxygen diffusion pathway (along indicating that the oxygen incorporation is limited by the surface the nm-wide ceria regions), after the incorporation through the reduction reaction only—while the mass transport contribution is active surface (top of the APT tip in Fig. 2), proving the efficacy of completely suppressed. This is fully consistent with the APT the proposed VANs in combining the excellent ionic conductivity characterization of the oxygen kinetics highlighting fast vertical of SDC with the good oxygen affinity of the LSM (typically diffusion of oxygen through the SDC phase and is ascribed to the hindered by its poor oxygen diffusivity). The effect of materials intrinsic fast ionic conductivity of the fluorite, together with the combination is here maximized by the unique VAN micro- advantageous off-plane VAN geometry. Nyquist plot for single structure, which is characterized by no tortuosity for off-plane phase SDC cells, exhibiting much higher ASR, is reported for mass transport and by high triple-phase boundary density (>106 comparison in Supplementary Fig. 10. cm·cm−2, cf. Fig. 1) for oxygen incorporation. Notably, the By adjusting the EIS spectra of the cells with a physically reported direct observation of fast local oxygen pathways, which meaningful equivalent circuit (i.e., a modified Jamnik-Maier is demonstrated here in the case of VANs, has wide applicability circuit for LSM and a simple ZARC element for LSM-SDC— and represents a decisive step forward to the local analysis of please refer to Supplementary Note 4 for further details on oxygen incorporation in nanomaterials, for which singularities impedance spectroscopy analysis), the effective rate constant for are often found at the nm-level (e.g., fast grain boundary oxygen the ORR kq can be obtained: kq ¼kb T=Rs cO2  z 2i e2 , being Rs the diffusion)43. In order to quantitatively assess this picture, the surface resistance for the oxygen chemical reaction48. The results measured diffusion profiles (integrated over the area between the are reported in Fig. 3b, where LSM-SDC is put in direct dotted lines in Fig. 2d) were fitted using Finite Element Method comparison with (i) stoichiometric LSM, (ii) LSM-YSZ dense (FEM) simulations on a modelled 3D geometry consisting of LSM composites and (iii) state-of-the-art cobalt-based MIEC for columns (with slow oxygen diffusivity) embedded in an SDC intermediate-temperature electrode applications (La0.8Sr0.2- matrix (further details on the modelling are given in Supple- CoO3—LSCo82)45,46. One can see that the oxygen incorporation mentary Note 3). Please note that the experimental data do not kinetics are remarkably enhanced in LSM-SDC with respect to suggest any singularity in the oxygen diffusivity at the interface, LSM ( ≈ 100 fold enhancement) and to LSM-YSZ. Besides the either because not present or because of a very limited effect on activation energy Ea is decreased by more than 1 eV (≈2.9 eV for the behavior of the VAN thin film. For these reasons, the FEM stoichiometric LSM and ≈1.8 eV for LSM-SDC), becoming model does not include any nanoscopic parameters for describing comparable to LSCo (Ea ≈ 1.5 eV). Please note that, even though oxygen transport at the LSM/SDC interface, which is modeled as much lower values are expected for porous Co-based bulk an abrupt change of diffusivity in the two materials (continuous electrodes49, here the relevant comparison is between dense thin layer)44. The comparison between the measured and the fitted films with potential application as functional layers. In Fig. 3b, (18O/(18O + 16O)) profiles are represented in Fig. 2e, top panel. also the k value obtained by the FEM simulation of the APT The fitting of the experimental profiles allowed extraction of the oxygen isotope profiles is reported (red star), demonstrating the oxygen surface exchange coefficient of the composite (k* = 2.4 ∙ very good agreement with the impedance data. With knowledge 10−9 cm s−1) and the diffusion coefficient of oxygen in LSM (D* on the elementary processes of oxygen incorporation in LSM = 3.6 ∙ 10−17 cm2 ∙ s−1) at 550 °C (note that no information on (atomic adsorbate (O ad ) incorporation-limited ORR process) 47,50, the ion diffusivity is accessible by tracer method for SDC since it is also possible to provide a mechanistic interpretation of the these regions saturate owing to the fast transport kinetics). By synergistic LSM-SDC incorporation mechanism by analyzing the comparing the tracer fraction and the Mn concentration profiles oxygen partial pressure (pO2) dependence of kq. Figure 3c shows (Fig. 2e, bottom panel), one can also observe that an 18O that in the low-pO2 regime, both the single phase material and the concentration decay is visible in LSM (maximum concentration nanocomposite follow the expected ≈1⁄2 slope resulting from the NATURE COMMUNICATIONS | (2021)12:2660 | https://doi.org/10.1038/s41467-021-22916-4 | www.nature.com/naturecommunications 5 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-22916-4 Fig. 3 Electrochemical performance of LSM (black stars) and LSM-SDC (green bullets) nanocomposites. a Representative initial EIS Nyquist plots measured at 750 °C (air atmosphere) in an out-of-plane configuration and using a porous Ag layer as a low impedance counter-electrode for LSM and LSM-SDC. The inset shows the LSM-SDC impedance arc on a smaller scale. b Surface exchange coefficient kq as a function of inverse temperature, as compared to stoichiometric LSM (from our labs and from Ref. 46), state-of-the-art LSM-YSZ dense composites (from Ref. 45) and to state-of-the-art La0.8Sr0.2CoO3 (our lab). The red star represents the value resulting from the FEM simulation of the APT 18O fraction profile. c pO2 dependence for kq. The slopes for LSM and LSM-SDC are indicated. d ASR vs time for LSM and LSM-SDC, measured during ageing treatment. The lines are intended as a guide to the eye. final incorporation step (i.e., the encounter between O ad and a oxygen adsorbates can be readily incorporated. This is in lattice oxygen vacancy (V  O ), leading to a predicted reaction rate agreement with past theoretical and experimental works on 1 R (pO2) /θ ðO LSM bulk and grain boundaries, which assign to the approach of ad Þ / pO2 , where θ is the surface coverage) being 2 oxygen vacancies to the O ad the rate-determining step (RDS) for limiting in both cases. Under such conditions, also the absolute kq values are similar between the single-phase material and the ORR in LSM53. The LSM RDS is therefore alleviated in our nanocomposite, further corroborating this interpretation. How- VANs. Notably, these considerations are perfectly in line with ever, at high pO2, the limited availability of oxygen vacancies in previous works from the co-authors43, in which a decrease in LSM (which are mostly concentrated at the grain boundaries) ORR activation energy was measured for B-site deficient LSM. governs the saturation in the kinetics of the incorporation process There, it was concluded that lowering the Mn concentration (i.e., sluggish pO2-kq dependence), resulting in an excess of O generates a reduction of oxygen vacancy formation energy and a ad at consequent enhancement of the ORR at the LSM grain the film surface (cf. also Supplementary Fig. 11)51. Conversely, in boundaries. Interestingly, the final ORR activation energy value the case of LSM-SDC, the supply of oxygen vacancies from SDC for strongly B-site deficient LSM is similar to the value of our allows for readily incorporating oxygen also at high pO2, as VANs (≈1.8 eV):54 In both cases, introducing mobile oxygen schematically depicted in the sketches in Fig. 3c. This further vacancies in the system greatly enhances the ORR kinetics. In demonstrates the beneficial effect of the vertical phase alternation VANs, a better electrochemical performance is ensured by the and maximized triple-phase boundary density in the VAN higher mobility and concentration of oxygen vacancies in the structure for engineering electrochemical functional layers. SDC phase. Notably, oxygen incorporation in our VANs may be facilitated The formation of space-charge zones and especially of oxygen by the extension of the active area for oxygen incorporation also vacancies redistribution at the LSM-SDC interface could possibly to the SDC surface as a consequence of adsorbates “spill-over” act as an additional factor in favor of faster ORR kinetics as a from LSM to SDC52. Such considerations also allow explaining consequence of oxygen vacancy accumulation at the LSM side of the observed strong reduction of the ORR activation energy, the interface55. This however is not expected to be a dominant which passes from ≈2.89 eV for single phase LSM to ≈1.77 eV for effect since the space-charge width is very narrow in highly doped LSM-SDC: The fluorite phase—in intimate contact with LSM via systems. In our case, for SDC and LSM nominal acceptor the VAN architecture—introduces a large amount of highly concentration is 20% at.; From here, one can estimate the Debye mobile oxygen vacancies in the system, through which surface 6 NATURE COMMUNICATIONS | (2021)12:2660 | https://doi.org/10.1038/s41467-021-22916-4 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-22916-4 ARTICLE Fig. 4 AFM and SEM top-view micrographs for LSM and LSM-SDC. In (a), a full comparison between the surfaces of the two compounds in the as-grown and aged states is provided. The resulting root mean square roughness are: Rms = 0.8 (7.1) and Rms = 0.7(1.0) nm for LSM-SDC and single phase LSM as grown (after ageing), respectively. In (b –LSM) and (c—VAN), lower magnification SEM for the surfaces after thermal treatment (In-Lens detector – 3 kV acceleration voltage). length to be ≈1 nm56, i.e., at least one order of magnitude lower microstructural characterization by atomic force microscopy than the pillar width. (AFM) and scanning electron microscopy (SEM). The results are As far as the in-plane current percolation is concerned, shown in Fig. 4. In Fig. 4a, the comparison between the AFM although a perfect VAN structure should not, in principle, allow micrographs from LSM and LSM-SDC before and after thermal for electronic current transport, one can imagine that the (yet treatment (700 °C, 100 h) highlights that no surface evolution is typically negligible) electronic conductivity of the SDC under apparent for LSM-SDC, while secondary phase outgrows are working conditions may be sufficient to ensure a certain in-plane formed on the LSM surface. These are ascribed to the well-known percolation owing to the very small pillar width (≈10 nm). formation of Sr-based precipitates as a consequence of massive Current collection for the electrochemical experiments was in any dopant migration which is typical of doped MIECs, and which is case ensured by the utilization of porous Au paste (see also responsible for the severe degradation of the surface reactivity experimental section), which allows the electrode to be partially owing to a reduced availability of active sites60. The top-view in contact with the gas atmosphere while offering a closely spaced SEM images for the materials after ageing (LSM and LSM-SDC in percolating path thus limiting in-plane potential drops. Such an panels b and c, respectively), provide information at lower- approach has been extensively used in the past for dense and magnification and highlight the profound difference between the porous thin film electrodes57. two surfaces: Namely, deleterious Sr segregation is fully hindered Lastly, in Fig. 3d we report the results on the thermal stability in LSM-SDC. Please note that the bulk thermal stability of VANs of the SDC-LSM nanocomposites, evaluated by carrying out a had been already been proven by TEM (cf. Fig. 1). 100 h treatment at 700 °C in air, while in-situ measuring the polarization resistance. It is well-known that, under such conditions, LSM and doped perovskite materials in general Surface chemical analysis by LEIS. Complementarily to AFM undergo changes in the surface composition as a consequence of and SEM imaging, the stability of VANs surface upon thermal dopant migration from the bulk, causing the formation of ageing has been quantitatively assessed using state-of-the-art low insulating Sr-based secondary phases58,59. In our case, single- energy ion scattering (LEIS). LEIS is capable of probing the ele- phase LSM undergoes the expected degradation, which is mental composition of the very outer atomic surface layer of the reflected in a 50% ASR increase in 100 h. Surprisingly, the material and is therefore well-suited for analyzing surface che- LSM-SDC nanocomposites are characterized by stable thermal mical composition with the highest sensitivity42. The LEIS spectra behavior with virtually no degradation over the time span under for VAN as-grown and after thermal ageing (100 h at 700 °C) are consideration: Even a 70% decrease of the ASR is observed before reported in Fig. 5a, b, respectively. The stoichiometric ratios [Sr]/ stabilization (Please refer to Supplementary Fig. 12 for the time [La] can be obtained from the integration of the peak areas. It evolution of Dq and kq for LSM and VANs). becomes evident that not just Sr has a very low tendency to segregate in the VAN (the final ratio [Sr]/[La] = 0.36 ± 0.06 is close the bulk value after the ageing heat treatment—Fig. 5b) but Top-view microscopy highlights no dopant segregation in also that the Sr fraction is remarkably decreased in the final state VANs. In order to rationalize the observed long-term electro- with respect to the material as-grown ([Sr]/[La] = 0.86 ± 0.05— chemical stability, we first carried out complementary Fig. 5a). Such a reduction of the Sr content at the surface not only NATURE COMMUNICATIONS | (2021)12:2660 | https://doi.org/10.1038/s41467-021-22916-4 | www.nature.com/naturecommunications 7 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-22916-4 confirms the high thermal stability of the VANs, but also explains in the absence of precipitates61.) The initial high Sr content for the observed ASR evolution over time, which decreases during the the as-grown material also reveals that the thermodynamic con- first ≈40 h and is constant afterwards—cf. Fig. 3d. This indicates ditions (high T and low pO2) and fast kinetics of the film that Sr-La substitution is thermodynamically unfavorable at the synthesis process (PLD) are very far from the final equilibrium in VANs surface. (Please note the dominant Sr-termination in the air, while a bulk-like equilibrium situation is approached upon as-deposited state is expected to lead to slower ORR kinetics even annealing in the VANs structure. Such a behavior is in contrast with the common behavior of MIEC materials showing thermally-induced Sr segregation and represents an outstanding improvement with respect to state-of-the-art MIEC electrodes. DFT: effects of local high entropy on lattice stabilization. We performed theoretical calculations based on spin polarized Density Functional Theory (DFT—see further details in Supplementary Note 5 and elsewhere in ref. 62) of a cubic perovskite supercell containing 40 atoms (Fig. 6a) with the aim of obtaining further insights into the observed thermal stability of the LSM-SDC VAN. Our LSM columnar structures, as highlighted experimentally, are characterized by high levels of cation intermixing (Sm and Ce) and by isotropic compressive strain (cf. Fig. 1). By carrying out energy minimization (Fig. 6b), we found that the introduction of Sm cations into the LSM model (La0.75Sr0.125Sm0.125MnO3) leads to a pronounced cell contraction of around 0.3%. Ce doping, even to quite high levels, has a smaller effect upon the lattice parameter, yet a reduction of a cell parameter is predicted with a drop of ~0.1% for a composition of La0.625Sr0.125Ce0.25MnO3. Therefore, one can first conclude that the observed isotropic compressive strain of LSM in the VAN structure (cf. Fig. 1) is a consequence of the chemical incorporation of Sm from the SDC layers (with an additional but smaller influence of Ce at higher dopant contents)63, ruling out an important role of off-plane epitaxial Fig. 5 LEIS spectra for LSM-SDC VAN. Analysis of surface as deposited (a) elastic strain possibly deriving from the observed lattice match and after thermal treatment (100 h, 700 °C—b). The stoichiometric ratio is epitaxy (cf. Fig. 1e). Taking this into account, we directly evaluated obtained from integration of the peak areas. Absolute differences in peak the cation interdiffusion as a possible source of the Sr segregation intensities should be ascribed to the specific experimental conditions. suppression. As a consequence of the observed spontaneous cationic intermixing, here nominal LSM is, as a matter of fact, a Fig. 6 DFT calculations for high-entropy LSM. a Example of a 40-atom periodic supercell used in DFT calculations showing relaxed positions of the Mn (gray), Sr (blue), La (green), O (red) and Ce (purple) ion and (b) calculated values of the lattice strain due to chemical doping. c box plot showing the mean and distribution of site volumes around the Sr (blue) and La (green) sites. d bond valence mismatch at the ion positions for the different cation dopants, where for example SrLa-SmLa corresponds to a stoichiometry of La0.75Sr0.125Sm0.125MnO3. 8 NATURE COMMUNICATIONS | (2021)12:2660 | https://doi.org/10.1038/s41467-021-22916-4 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-22916-4 ARTICLE six element high entropy oxide having Sconfig,VAN = 1.21 R (Sconfig term durability at typical operational temperatures of solid oxide =R∑Ni¼1 xi lnxi where xi is the atomic fraction in the unit formula electrochemical cells. as resulting from Table 1). The same calculation for The well-defined geometry, together with the intimate phase La0.8Sr0.2MnO3 yields a much lower value (Sconfig,LSM = 0.56 R). alternation and the long-range interface coherence, makes the Entropy-induced lattice stabilization has been recently reported structure ideal for directly probing structural and functional for bulk LSM;64 Here, we directly evaluate the effect of homo- properties in great detail. Evidence of the presence of a fast geneously dispersed multiple atom species in the perovskite A-site oxygen diffusion pathway at the nanoscale along the SDC phase on the local strain field around the Sr atom, i.e., on the elastic was presented here using isotopic exchange depth profiling based driving force for atom surface segregation60. To this end, we on APT. By this approach we proved that VAN nanocomposites considered the Sr (La)-O polyhedral volume (Fig. 6c) and the bond can solve current kinetic limitations of existing electrodes by a valence sum at each lattice site (Fig. 6d) for different cation dis- synergistic combination of the functionalities of their individual tributions. In bulk LSM with a composition of La0.875Sr0.125MnO3, constituents. we estimate that the Sr-ion is significantly overbonded with a Overall, the described VAN architecture is an easy-to-implement, valence mismatch of ~0.15 valence units (v.u.), i.e., the surrounding superior strategy for oxide electrochemistry resulting in long-term oxygen ions are closer than required by the formal valence state of thermal stability and driving fast oxygen reduction kinetics. Our the Sr ion, while La ions are slightly underbonded, ~−0.04 v.u. This work represents a decisive step forward in the implementation of opposite valence mismatch is the likely driving force for the advanced thin film technology for high performance electrochemi- exchange of overbonded Sr ions with underbonded La ions on the cal energy conversion devices and opens up a path for the surface (in combination with electrostatic effects). This is con- development of next-generation functional oxides. ceptually similar to the Sr ion having a larger ionic radius than the La lattice ion it replaces59. Such an instability is strongly alleviated Methods by cationic intermixing (Fig. 6c, d): Particularly, the introduction of Film fabrication. LSM-SDC, LSM and LSCo thin films were deposited in a large- Ce leads to an increase in the Sr-O site volume, in turn reducing the area pulsed laser deposition chamber (PVD Systems—PLD 5000) equipped with a 248 nm KrF excimer laser (Lambda Physics—COMPex PRO 205) on YSZ (100) Sr-O overbonding relative to the La-O polyhedra and thereby substrate (Crystec GmbH) under the following conditions: temperature 800 °C decreasing the elastic driving force for Sr surface segregation. In (600 °C for LSCo), oxygen pressure 0.007 mbar, target-substrate distance 90 mm, other words, the increase of configurational entropy dictates the laser fluency 1.1 ≈ J·cm−2, laser frequency 2 Hz for LSM-SDC, 10 Hz for LSM and enhancement of the LSM stability by triggering a local lattice dis- LSCo. A thin (≈5 nm) buffer layer of Ce0.8Gd0.2O2, fabricated under the same conditions, was deposited between LSCo and YSZ in order to avoid the formation tortion. Besides, in agreement with recent theoretical works on high of secondary phases. Since z-resolved 18O labelling experiments are only feasible entropy oxides65, one may consider also a stabilizing effect of the for films grown on an inert substrate, a perovskite SrTiO3 substrate (100—Crystec interfaces, i.e., the contribution of surface energy to the final free GmbH) was employed for the sample that underwent APT analysis. energy, which is maximized here by the very high interface density. LSM-SDC films were fabricated using home-made targets obtained by mixing commercial powders of La0.8Sr0.2MnO3 and Ce0.8Sm0.2O2 (1:1 wt. %—Kceracell) Critically, the LSM in the VAN under consideration is able to via ball milling in ethanol solution. The dried powder mix was uniaxially pressed accept Ce concentrations well beyond the reported solubility limits (7 MPa, 30 s) to form a pellet (diameter ≈ 1 inch). Sintering was carried out at 1300 of Ce in single phase LSM (2% mol)39, which is crucial for the °C for 4 h (heating and cooling ramps ≈5 °C/min). A commercial target was used observed effects and which further demonstrates the stability of the for LSM and LSCo. structure. Within the experimental resolution of the techniques, no detectable elemental clustering can be observed by HR-TEM (cf. Microscopy and XRD. TEM and HAADF STEM imaging, EDS mapping and line Fig. 1) or by APT (cf. Fig. 2 and Supplementary Fig. 3), confirming scans were acquired by FEI Talos F200X. The TEM samples were prepared using conventional mechanical procedures including polishing, dimpling, and ion milling the single solid-solution nature of our highly doped LSM in VANs. (PIPS 691). Films were analyzed by XRD by using a lab diffractometer with 4-angle Unlike previously reported studies on entropy-stabilized LSM64, in goniometer (MRD X’Pert Pro from Malvern-Panalytical) and Cu Kα tube and 2 × which no increase in the electrochemical properties was reported, Ge(110) monochromator. SEM and AFM characterization were performed in a the relevance of our VAN strategy lies in combining high thermal Zeiss Auriga and a Park System, respectively. stability—owing to local high-entropy in the LSM phase—and strongly enhanced ORR kinetics due to the intimate phase alter- Atom probe tomography. Atom probe tomography (Cameca LEAP 4000X Si) was nation and off-plane geometry. Because the VAN structure is self- performed on focused ion beam prepared specimens (FEI Helios NanoLab 600i) that were mounted on TEM grids and hardware that allowed for TEM imaging assembled, both the composition and interfaces which form at the (FEI Talos F200X) and analysis of the APT specimens66,67. The 355 nm laser- growth temperature are the stable ones, i.e., they should not change pulsed APT analysis was performed at 57.3 K using either a 0.8 or 24 pJ laser with further annealing under conditions similar to the growth energy and 500 kHz pulse rate (cf. Supplementary Table 1). The flight path length conditions as they are the energetically preferred. Please note that was 90 mm and the ion detection rate was set to 3.5 ions per 1000 pulses, resulting in a bias range of 2200–5000 V during the data collection. Reconstructions were the TEM investigation presented in Fig. 1, showing very well- generated in Cameca’s IVAS 3.6.14 software using the TEM images of the speci- defined columnar structure and long-range order, was carried out mens as a guide for setting the reconstruction parameters. Atomic volumes of the after thermal ageing (100 h at 700 °C), confirming the structural species were adjusted based on the unit cell parameters of the oxides. stability of our VANs. This is very different to the case of artifi- cially- grown superlattice structures, where the artificially-formed Oxygen tracer annealing. 18O tracer annealing was carried out at a temperature of interfaces are unlikely to be the equilibrium ones (i.e., coherence is 550 °C for 6000 s in an atmosphere of 200 mbar of enriched oxygen (90% 18O lost above a critical thickness and chemical instabilities may occur concentration). Prior to the exchange, an annealing (12 h) in natural oxygen was carried out. during operation). In summary, we were able to combine, at the nanoscale, the excellent oxygen exchange properties of LSM with the good Low energy ion scattering. To explore the uppermost sample surface chemical composition, a Qtac100 LEIS instrument (ION-TOF GmbH) was employed. Neon diffusivity of SDC to fabricate nearly ideal electrode functional primary ion beam was chosen for its higher mass resolution with respect to the layers for SOCs by design. This was achieved using VANs. Most elements of LSM-SDC composite analyzed in this study. Ne+ 5 keV ion beam was importantly we highlighted that, by the spontaneous formation of rastered over a sample area of 500 μm × 500 μm. The energy range was maintained a heavily co-doped manganite phase in the VAN (high entropy in 850–3600 eV. The Ne+ dose was kept around 5 × 1014 ions/cm2 for each ana- lysis. Cerium mean isotope LEIS peak overlaps with the lanthanum one. For this oxide), a stable compound is formed which hinders Sr- reason, the cerium higher mass isotope was also fitted to separate in this way segregation to the surface. This resulted in an outstanding long- cerium from lanthanum contribution. NATURE COMMUNICATIONS | (2021)12:2660 | https://doi.org/10.1038/s41467-021-22916-4 | www.nature.com/naturecommunications 9 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-22916-4 Electrochemical impedance spectroscopy. The electrochemical characterization 17. Tsvetkov, N., Lu, Q., Sun, L., Crumlin, E. J. & Yildiz, B. Improved chemical of the films was carried out using a Novocontrol impedance spectrometer using a and electrochemical stability of perovskite oxides with less reducible cations at frequency range 106−0.1 Hz and a voltage amplitude of 0.05 V. A low-impedance the surface. Nat. Mater. 15, 1010–1016 (2016). Ag counter-electrode was applied by brushing Ag conductive paste (Sigma 18. Fondard, J. et al. Outstanding oxygen reduction kinetics of La0.6Sr0.4FeO3–δ Aldrich), while a porous Au paste was applied on top of the films to ensure a surfaces decorated with platinum nanoparticles. J. Electrochem. Soc. 167, homogeneous current distribution, limiting the in-plane potential drops while 070526 (2020). leaving a large fraction of the surface exposed to air. A fine gold mesh was used as a 19. Perry Murray, E. & Barnett, S. A. (La,Sr)MnO3-(Ce,Gd)O2-x composite current collector on both sides. Measurements were carried out in a ProboStat test cathodes for solid oxide fuel cells. Solid State Ion. 143, 265–273 (2001). station (NorECs) placed inside a vertical furnace and using synthetic air 20. Wang, B., Yi, J., Winnubst, L. & Chen, C. Stability and oxygen permeation atmosphere. behavior of Ce0.8Sm0.2O2−δ-La0.8Sr0.2CrO3−δ composite membrane under large oxygen partial pressure gradients. J. Memb. Sci. 286, 22–25 (2006). FEM and DFT simulations. FEM simulations were performed by COMSOL Mul- 21. Li, W., Tian, T. F., Shi, F. Y., Wang, Y. S. & Chen, C. S. Ce0.8Sm0.2O2-δ-La 0.8 tiphysics using the Transport of diluted species module. DFT theoretical calculations Sr0.2MnO3-δ dual-phase composite hollow fiber membrane for oxygen have been based on spin polarized DFT using plane wave CASTEP code68. These separation. Ind. Eng. Chem. Res. 48, 5789–5793 (2009). calculations used the Perdew, Burke, and Ernzerhof exchange correlation functional 22. Nielsen, J. & Hjelm, J. Impedance of SOFC electrodes: a review and a for solids (PBESol) in combination with ultrasoft pseudopotentials with a plane wave comprehensive case study on the impedance of LSM:YSZ cathodes. expansion to 600 eV and a Monkorst-Pack grid69 spacing no >0.0404 Å−1 70. The Electrochim. Acta 115, 31–45 (2014). presence of Mn within the LSM structure may cause a significant miscalculation of 23. Anelli, S. et al. Improved mesostructured oxygen electrodes for highly the exchange energy and so we use the DFT + U correction with a U-J value of performing solid oxide cells for co-electrolysis of steam and carbon dioxide. J. 4.0 eV on the Mn ions. Although largely empirically derived this value has been Mater. Chem. A 7, 27458–27468 (2019). validated against more complete hybrid Hartree-Fock DFT calculations for the LSM 24. Yun, C. et al. Achieving ferromagnetic insulating properties in La0.9Ba0.1MnO3 structure and also in other Mn oxides62. An initial ordering of cations was chosen thin films through nanoengineering. Nanoscale 12, 9255–9265 (2020). for study and the ion positions and cell parameters were relaxed according to a BFGS 25. Huang, J., MacManus-Driscoll, J. L. & Wang, H. New epitaxy paradigm in minimization. The initial spin polarization of the Mn ions was taken to be ferro- epitaxial self-assembled oxide vertically aligned nanocomposite thin films. J. magnetic and were free to refine during the progression of the minimization. Mater. Res. 32, 4054–4066 (2017). 26. Choi, E. M. et al. 3D strain-induced superconductivity in La2CuO4+δ using a Data availability simple vertically aligned nanocomposite approach. Sci. Adv. 5, eaav5532 The APT, impedance spectroscopy. AFM, SEM and LEIS data that support the findings (2019). of this study are available in Zenodo (link) with the identifier “https://doi.org/10.5281/ 27. Yang, S. M. et al. Strongly enhanced oxygen ion transport through samarium- zenodo.4544312”. doped CeO2 nanopillars in nanocomposite films. Nat. Commun. 6, 8588 (2015). 28. Cho, S. et al. Self-assembled oxide films with tailored nanoscale ionic and Received: 23 December 2020; Accepted: 29 March 2021; electronic channels for controlled resistive switching. Nat. Commun. 7, 1–10 (2016). 29. Cho, S. et al. Self-Assembled Heteroepitaxial Oxide Nanocomposite for Photoelectrochemical Solar Water Oxidation. Chem. Mater. 28, 3017–3023 (2016). 30. Chen, A., Su, Q., Han, H., Enriquez, E. & Jia, Q. Metal Oxide Nanocomposites: References a Perspective from Strain, Defect, and Interface. Adv. Mater. 1803241, 1–30 1. Wachsman, E. D. & Lee, K. T. Lowering the temperature of solid oxide fuel (2018). cells. Science 334, 935–939 (2011). 31. Yoon, J. et al. Vertically Aligned Nanocomposite Thin Films as a Cathode/ 2. Shin, S. S. et al. Multiscale structured low-temperature solid oxide fuel cells Electrolyte Interface Layer for Thin-Film Solid Oxide Fuel Cells. Adv. Funct. with 13 W power at 500 °C. Energy Environ. Sci. 11–16 (2020). Mater. 19, 3868–3873 (2009). 3. Maier, J. Nanoionics: ion transport and electrochemical storage in confined 32. Cho, S., Kim, Y. N., Lee, J., Manthiram, A. & Wang, H. Microstructure and systems. Nat. Mater. 4, 805–815 (2005). electrochemical properties of PrBaCo2O5+δ/Ce0.9Gd0.1O1.95 vertically aligned 4. Steele, B. C. H. & Heinzel, A. Materials for fuel-cell technologies. Nature 414, nanocomposite thin film as interlayer for thin film solid oxide fuel cells. 345–352 (2001). Electrochim. Acta 62, 147–152 (2012). 5. Gao, Z., Mogni, L. V., Miller, E. C., Railsback, J. G. & Barnett, S. A. A 33. Ma, W. et al. Vertically aligned nanocomposite La0.8Sr0.2CoO3/ perspective on low-temperature solid oxide fuel cells. Energy Environ. Sci. 9, (La0.5Sr0.5)2CoO4 cathodes—electronic structure, surface chemistry and 1602–1644 (2016). oxygen reduction kinetics. J. Mater. Chem. A 3, 207–219 (2015). 6. Jung, W. C. & Tuller, H. L. A new model describing solid oxide fuel cell 34. Dubbink, D., Koster, G. & Rijnders, G. Growth mechanism of epitaxial YSZ cathode kinetics: Model thin film SrTi1-xFexO3-δ mixed conducting oxides—A on Si by Pulsed Laser Deposition. Sci. Rep. 8, 1–10 (2018). case study. Adv. Energy Mater. 1, 1184–1191 (2011). 35. Kelly, T. F. & Larson, D. J. Atom Probe Tomography 2012. Annu. Rev. Mater. 7. Acosta, M., Baiutti, F., Tarancón, A. & MacManus-Driscoll, J. L. Res. 42, 1–31 (2012). Nanostructured Materials and Interfaces for Advanced Ionic Electronic 36. Marquis, E. A. & Vurpillot, F. Chromatic aberrations in the field evaporation Conducting Oxides. Adv. Mater. Interfaces 6, 1–15 (2019). behavior of small precipitates. Microsc. Microanal. 14, 561–570 (2008). 8. Bertei, A., Yufit, V., Tariq, F. & Brandon, N. P. A novel approach for the 37. Vurpillot, F., Bostel, A. & Blavette, D. Trajectory overlaps and local quantification of inhomogeneous 3D current distribution in fuel cell magnification in three-dimensional atom probe. Appl. Phys. Lett. 76, electrodes. J. Power Sources 396, 246–256 (2018). 3127–3129 (2000). 9. Irvine, J. T. S. et al. Evolution of the electrochemical interface in high- 38. Koettgen, J. et al. Understanding the ionic conductivity maximum in doped temperature fuel cells and electrolysers. Nat. Energy 1, 1–13 (2016). ceria: trapping and blocking. Phys. Chem. Chem. Phys. 20, 14291–14321 10. Khan, M. Z., Song, R. H., Mehran, M. T., Lee, S. B. & Lim, T. H. Controlling (2018). cation migration and inter-diffusion across cathode/interlayer/electrolyte 39. Konysheva, E., Francis, S. M. & Irvine, J. T. S. Crystal Structure, Oxygen interfaces of solid oxide fuel cells: a review. Ceram. Int. 47, 5839–5869 (2021). Nonstoichiometry, and Conductivity of Mixed Ionic–Electronic Conducting 11. Cai, Z., Kuru, Y., Han, J. W., Chen, Y. & Yildiz, B. Surface electronic structure Perovskite Composites with CeO[sub 2]. J. Electrochem. Soc. 157, B159 (2010). transitions at high temperature on perovskite oxides: The case of strained 40. Kinno, T., Tomita, M., Ohkubo, T., Takeno, S. & Hono, K. Laser-assisted atom La0.8Sr0.2CoO3 thin films. J. Am. Chem. Soc. 133, 17696–17704 (2011). probe tomography of 18O-enriched oxide thin film for quantitative analysis of 12. Kubicek, M. et al. Tensile lattice strain accelerates oxygen surface exchange oxygen. Appl. Surf. Sci. 290, 194–198 (2014). and diffusion in La1-xSrxCoO3-qδ thin films. ACS Nano 7, 3276–3286 (2013). 41. De Souza, R. A., Zehnpfenning, J., Martin, M. & Maier, J. Determining oxygen 13. Koo, B. et al. Enhanced oxygen exchange of perovskite oxide surfaces through isotope profiles in oxides with Time-of-Flight SIMS. Solid State Ion. 176, strain-driven chemical stabilization. Energy Environ. Sci. 11, 71–77 (2018). 1465–1471 (2005). 14. Navickas, E. et al. Dislocations Accelerate Oxygen Ion Diffusion in 42. Druce, J., Téllez, H., Ishihara, T. & Kilner, J. A. Oxygen exchange and La0.8Sr0.2MnO3 Epitaxial Thin Films. ACS Nano 11, 11475–11487 (2017). transport in dual phase ceramic composite electrodes. Faraday Discuss. 182, 15. Zhao, C. et al. Heterointerface engineering for enhancing the electrochemical 271–288 (2015). performance of solid oxide cells. Energy Environ. Sci. 13, 53–85 (2020). 43. Chiabrera, F. et al. Engineering Transport in Manganites by Tuning Local 16. Choi, M. et al. Engineering of Charged Defects at Perovskite Oxide Surfaces Nonstoichiometry in Grain Boundaries. Adv. Mater. 31, 1805360 (2019). for Exceptionally Stable Solid Oxide Fuel Cell Electrodes. ACS Appl. Mater. 44. Saranya, A. M. et al. Unveiling the Outstanding Oxygen Mass Transport Interfaces 12, 21494–21504 (2020). Properties of Mn-Rich Perovskites in Grain Boundary-Dominated 10 NATURE COMMUNICATIONS | (2021)12:2660 | https://doi.org/10.1038/s41467-021-22916-4 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-22916-4 ARTICLE La0.8Sr0.2(Mn1- xCox)0.85O3±δ Nanostructures. Chem. Mater. 30, 5621–5629 69. Pack, J. D. & Monkhorst, H. J. ‘special points for Brillouin-zone integrations’-a (2018). reply. Phys. Rev. B 16, 1748–1749 (1977). 45. Ji, Y., Kilner, J. A. & Carolan, M. F. Electrical properties and oxygen diffusion 70. Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized gradient approximation in yttria-stabilised zirconia (YSZ)-La0.8Sr0.2MnO3±δ (LSM) composites. Solid made simple. Phys. Rev. Lett. 77, 3865–3868 (1996). State Ion. 176, 937–943 (2005). 46. Usiskin, R. E., Maruyama, S., Kucharczyk, C. J., Takeuchi, I. & Haile, S. M. Probing the reaction pathway in (La0.8Sr0.2)0.95MnO3+δ using libraries of thin Acknowledgements film microelectrodes. J. Mater. Chem. A 3, 19330–19345 (2015). J.S. acknowledges the support of ICN2 (funded by the CERCA programme/Generalitat 47. Fleig, J., Kim, H. R., Jamnik, J. & Maier, J. Oxygen reduction kinetics of de Catalunya and by the Severo Ochoa programme SEV-2017-0706) for the XRD Lanthanum Manganite (LSM) model cathodes: Partial pressure dependence measurements. M.A. acknowledges the support from the Feodor Lynen Research Fel- and rate-limiting steps. Fuel Cells 8, 330–337 (2008). lowship Program of the Alexander von Humboldt Foundation and the Isaac Newton 48. Maier, J. On the correlation of macroscopic and microscopic rate constants in Trust, 17.25(a). M.A. and J.D. acknowledge the support from the EPSRC Centre of solid state chemistry. Solid State Ion. 112, 197–228 (1998). Advanced Materials for Integrated Energy Systems (CAM-IES) under EP/P007767/1. J.D. 49. Shen, F. & Lu, K. Comparison of Different Perovskite Cathodes in Solid also acknowledge support from EPSRC grants EP/N004272/1, EP/T012218/1, the Royal Oxide. Fuel Cells Fuel Cells 18, 457–465 (2018). Academy of Engineering- CIET1819_24, ERC POC grant 779444, Portapower. X.W. and 50. Kuklja, M. M., Kotomin, E. A., Merkle, R., Mastrikov, Y. A. & Maier, J. H.W. acknowledge the funding support from the U.S. National Science Foundation for Combined theoretical and experimental analysis of processes determining the TEM effort at Purdue University (DMR-1565822 and DMR-2016453). This project cathode performance in solid oxide fuel cells. Phys. Chem. Chem. Phys. 15, has received funding from the European Union’s Horizon 2020 research and innovation 5443–5471 (2013). program under grant agreement No 824072 (HARVESTORE), No 681146 (ULTRA- 51. Navickas, E. et al. Fast oxygen exchange and diffusion kinetics of grain SOFC) and No 101017709 (EPISTORE) and was supported by an STSM Grant from the boundaries in Sr-doped LaMnO3 thin films. Phys. Chem. Chem. Phys. 17, COST Action MP1308: Towards Oxide-Based Electronics (TO-BE), supported by COST 7659–7669 (2015). (European Cooperation in Science and Technology). The authors thank A. Aguadero for 52. Shen, Z., Skinner, S. J. & Kilner, J. A. Oxygen transport and surface exchange the fruitful discussions and A. Kuzyk for support on graphics. mechanisms in LSCrF-ScCeSZ dual-phase ceramics. Phys. Chem. Chem. Phys. 21, 13203–13215 (2019). Author contributions 53. Mastrikov, Y. A., Merkle, R., Heifets, E., Kotomin, E. A. & Maier, J. Pathways Experiments were conceived by F.B., F.C., M.A., J.D., A.M., A.T. Sample preparation was for oxygen incorporation in mixed conducting perovskites: A DFT-based carried out by F.B., M. A., electrochemical characterization by F.B. APT was performed Mechanistic analysis for (La, Sr)MnO3-δ. J. Phys. Chem. C. 114, 3017–3027 by D.D., FEM simulations by F.C. DFT calculations were done by D.P. and A.Ch. XRD (2010). analysis was carried out by J.S., TEM analysis by X.W. under the supervision of H.W. A. 54. Chiabrera, F. M. Interface Engineering in Mixed Ionic Electronic Conductor Ca. performed the oxygen exchange treatment. F.B. designed the research plan and Thin Films for Solid State Devices, PhD Thesis, University of Barcelona prepared the initial draft. All the authors contributed to the interpretation and discussion (2019). of the experimental data and to editing the initial draft. 55. Baiutti, F. et al. High-temperature superconductivity at the lanthanum cuprate/lanthanum-strontium nickelate interface. Nanoscale 10, 8712–8720 (2018). Competing interests 56. Kim, S., Kim, S. K., Khodorov, S., Maier, J. & Lubomirsky, I. On determining The authors declare no competing interests. the height of the potential barrier at grain boundaries in ion-conducting oxides. Phys. Chem. Chem. Phys. 18, 3023–3031 (2016). 57. Wells, M. P. et al. Route to High-Performance Micro-solid Oxide Fuel Cells on Additional information Supplementary information The online version contains supplementary material Metallic Substrates. ACS Appl. Mater. Interfaces 13, 4117–4125 (2021). available at https://doi.org/10.1038/s41467-021-22916-4. 58. Jung, W. & Tuller, H. L. Investigation of surface Sr segregation in model thin film solid oxide fuel cell perovskite electrodes. Energy Environ. Sci. 5, Correspondence and requests for materials should be addressed to F.B. or A.T. 5370–5378 (2012). 59. Lee, W., Han, J. W., Chen, Y., Cai, Z. & Yildiz, B. Cation size mismatch and Peer review information Nature Communications thanks the anonymous reviewer(s) for charge interactions drive dopant segregation at the surfaces of manganite their contribution to the peer review of this work. Peer reviewer reports are available. perovskites. J. Am. Chem. Soc. 135, 7909–7925 (2013). 60. Kim, D., Bliem, R., Hess, F., Gallet, J. J. & Yildiz, B. Electrochemical Polarization Reprints and permission information is available at http://www.nature.com/reprints Dependence of the Elastic and Electrostatic Driving Forces to Aliovalent Dopant Segregation on LaMnO3. J. Am. Chem. Soc. 142, 3548–3563 (2020). Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in 61. Cao, Y., Gadre, M. J., Ngo, A. T., Adler, S. B. & Morgan, D. D. Factors published maps and institutional affiliations. controlling surface oxygen exchange in oxides. Nat. Commun. 10, 1–15 (2019). 62. Pavone, M., Muñoz-García, A. B., Ritzmann, A. M. & Carter, E. A. First-principles study of lanthanum strontium manganite: Insights into electronic structure and oxygen vacancy formation. J. Phys. Chem. C. 118, 13346–13356 (2014). Open Access This article is licensed under a Creative Commons 63. Ansari, A. A. et al. Physico-chemical properties and catalytic activity of the Attribution 4.0 International License, which permits use, sharing, sol-gel prepared Ce-ion doped LaMnO3 perovskites. Sci. Rep. 9, 1–12 (2019). adaptation, distribution and reproduction in any medium or format, as long as you give 64. Yang, Y. et al. A novel facile strategy to suppress Sr segregation for high- appropriate credit to the original author(s) and the source, provide a link to the Creative entropy stabilized La0.8Sr0.2MnO3-δ cathode. J. Power Sources 482, 228959 Commons license, and indicate if changes were made. The images or other third party (2021). material in this article are included in the article’s Creative Commons license, unless 65. Ramesh, R. & Schlom, D. G. Creating emergent phenomena in oxide indicated otherwise in a credit line to the material. If material is not included in the superlattices. Nat. Rev. Mater. 4, 257–268 (2019). article’s Creative Commons license and your intended use is not permitted by statutory 66. Thompson, K. et al. In situ site-specific specimen preparation for atom probe regulation or exceeds the permitted use, you will need to obtain permission directly from tomography. Ultramicroscopy 107, 131–139 (2007). the copyright holder. To view a copy of this license, visit http://creativecommons.org/ 67. Gorman, B. P. et al. Hardware and Techniques for Cross- Correlative TEM licenses/by/4.0/. and Atom Probe Analysis. Micros. Today 16, 42–47 (2008). 68. Clark, S. J. et al. First principles methods using {CASTEP}. Z. Krist. 220, 567–570 (2005). © The Author(s) 2021 NATURE COMMUNICATIONS | (2021)12:2660 | https://doi.org/10.1038/s41467-021-22916-4 | www.nature.com/naturecommunications 11