betaFeedback

WorksPhysics and Astronomy

Displacive Jahn–Teller transition in NaNiO2_2

Abstract

Below its Jahn–Teller transition temperature, TJTT_\mathrm{JT}, NaNiO2_2 has a monoclinic layered structure consisting of alternating layers of edge-sharing NaO6_6 and Jahn-Teller-distorted NiO6_6 octahedra. Above TJTT_\mathrm{JT} where NaNiO2_2 is rhombohedral, diffraction measurements show the absence of a cooperative Jahn-Teller distortion, accompanied by an increase in the unit cell volume. Using neutron total scattering, solid-state Nuclear Magnetic Resonance (NMR), and extended X-ray absorption fine structure (EXAFS) experiments as local probes of the structure we find direct evidence for a displacive, as opposed to order-disorder Jahn-Teller transition at TJTT_\mathrm{JT}. This is supported by ab initio molecular dynamics (AIMD) simulations. To our knowledge this study is the first to show a displacive Jahn-Teller transition in any material using direct observations with local probe techniques.

Introduction

In an undistorted octahedrally-coordinated environment, partially-occupied ege_g orbitals in a transition metal cation would be degenerate and therefore unstable to symmetry-reducing distortions by the Jahn–Teller (JT) effect.[1] This lifts the degeneracy of both the t2gt_{2g} and partially-occupied ege_g orbitals, lowering the energy of the system.[1, 2, 3, 4] The resulting octahedral distortion is often a linear combination of two possible van Vleck modes:[5, 6] a planar rhombic Q2Q_2 distortion and a tetragonal Q3Q_3 elongation/compression [Figure S1]. Experimentally, the distortion is normally dominated by the Q3Q_3 tetragonal elongation. This JT distortion, and associated orbital ordering, is relevant to many phenomena including unconventional superconductivity in the cuprates,[7, 8] magnetic structure through coupling of spin and orbital ordering,[9] and ionic mobility,[10, 11] and can lead to structural transitions in a material when it is used in a battery electrode.[10, 11, 12]

In a crystalline material, the energy of a JT-distorted system is often minimised when the axes of elongation of neighbouring octahedra are correlated; this is termed a cooperative JT distortion, in contrast to a non-cooperative system in which axes of elongation are randomly distributed.[4] Transitions from low-temperature, cooperative JT distortions to a high-temperature state with an undistorted average structure can generally be classified as order-disorder or displacive.[13] In the former case, non-cooperative JT distortions persist locally but are averaged out in the bulk structure, whereas in the latter case the local structure is undistorted. This paradigm has been applied to several systems. [13, 14, 15, 16, 17] The most well-studied is in the JT-distorted d4d^4 Mn3+^{3+} perovskite LaMnO3_3 which shows evidence for an order-disorder JT transition; [14, 15, 18, 19] theoretical works [20, 21, 22] and total scattering experiments[16] indicate a transition to a Potts model [20] type of orbital disorder.

The order-disorder/displacive paradigm has also been applied to the layered transition metal oxides with formula AMO2AMO_2 (AA=alkali metal, MM=transition metal). [13, 17] This family of materials includes several battery materials such as LiNiO2_2, NaNiO2_2, LiMnO2_2, and NaMnO2_2, all of which feature transition metal ions with degenerate d7/d4d^7/d^4 ege_g orbitals which are liable to JT distortions. The AMO2AMO_2 materials have layers of edge-sharing MO6MO_6 octahedra forming a triangular network of MM cations, separated by a layer of octahedrally-coordinated alkali metal ions, in contrast to the perovskites such as LaMnO3_3 which have corner-sharing octahedra and an approximately cubic cation network. The aristotype of the structure has rhombohedral R3ˉmR\bar{3}m symmetry, but NaNiO2_2 and each of α\alpha- and β\beta-XXMnO2_2 (XX=Li, Na) show cooperative JT ordering[13] resulting in a macroscopic distortion. NaNiO2_2 exhibits a monoclinic C2/mC2/m distortion due to collinear JT-elongated octahedra, which disappears on heating during a first-order transition which onsets at 460 K[23, 24, 25, 26, 27] (see Figure 1) to the aristotype Rˉ3m\bar{R}3m structure. LiMnO2_2 and NaMnO2_2 exhibit polymorphism and more complex cooperative ordering.[13] LiNiO2_2 is a complicated case, with clear experimental evidence for multiple Ni–O bond lengths[28, 29, 30] but diffraction data are typically modelled with the aristotype structure which does not allow for cooperative JT ordering. There have been a broad array of proposed forms of JT ordering[30, 31] for LiNiO2_2, with the energetically most-favourable being a zigzag ordering with monoclinic P21/cP2_1/c symmetry.[17, 31, 32, 33, 34] Alternative phenomena for LiNiO2_2 involving spin- or even charge-disproportionation[32, 34, 35] have also been proposed; these are likely a feature of other nickelates such as AgNiO2_2[36] [37] or the nickelate perovskites.[38, 39, 40]

Image not extracted

Figure 1. Crystal structures of (a) C2/mC2/m monoclinic and (c) R3ˉmR\bar{3}m rhombohedral NaNiO2_2. (b) and (d) show the Ni-Ni distances (grey for non-elongated; green for elongated) and JT-elongated Ni-O distances (pink) in these two structures. In the rhombohedral structure there are no elongated Ni-Ni or Ni-O bonds.

In contrast to a recent theoretical study,[13] which concludes that the layered alkali metal transition metal oxides ABABO2_2 (AA=Li, Na; BB=Ni, Mn) should exhibit an order-disorder JT transition, we have recently found evidence for a displacive transition in LiNiO2_2 using ab initio Molecular Dynamics (AIMD) and variable-temperature X-ray diffraction (XRD).[17]

In this work, we have studied the variable-temperature properties of NaNiO2_2. We use variable-temperature synchrotron X-ray diffraction, neutron total scattering, Ni K edge extended X-ray absorption fine structure (EXAFS), 23^{23}Na magic-angle spinning (MAS) solid state nuclear magnetic resonance (NMR) spectroscopy, and AIMD to study the changes in NaNiO2_2 with a focus on the local Ni environment through the monoclinic→\torh*ombohedral transition at ∼\sim460 K where the cooperative JT distortion disappears. We find evidence of a displacive JT transition, in contrast to the majority of previous studies on Jahn–Teller transitions which report order-disorder transitions. Taken together with previous results,[17, 27] this suggests a broader conclusion that the JT transitions in layered triangular-lattice nickelates are different from those in the transition metal perovskites where local JT distortions are reported to persist in the high-temperature phase.

Results

Variable-temperature synchrotron X-ray diffraction.

Variable-temperature synchrotron X-ray powder diffraction was performed to observe temperature-dependent changes in the average crystal structure and analysed by Rietveld refinement. [41]

On heating, the crystal structure of NaNiO2\mathrm{NaNiO}_{2} is monoclinic (C2/mC2/m) until ∼460 K where Bragg peaks associated with the high-temperature rhombohedral (R3ˉmR\bar{3}m) phase begin to emerge, indicating this is approximately the onset of the cooperative transition temperature TJTT_{\mathrm{JT}}. A mixed-phase regime exists, in which the phase fraction of the monoclinic phase decreases with heating, until the sample becomes entirely rhombohedral around 505 K. These findings are consistent with previous variable-temperature diffraction studies of NaNiO2\mathrm{NaNiO}_{2}.[24, 25, 26]

Selected results of the Rietveld analysis are shown in Figure 2, with further information in SI Section 2. The volume per formula unit is larger in the rhombohedral than the monoclinic structure [Figure 2(b)]. There is positive thermal expansion in both the monoclinic and rhombohedral phases. In the monoclinic phase, NiO6\mathrm{NiO}_{6} octahedral volume is essentially temperature-independent [inset of Figure 2(b)] and the volume increase is entirely driven by NaO6\mathrm{NaO}_{6} octahedral expansion. Through the phase transition, there is a discontinuous increase in unit cell volume per formula unit, primarily due to a ∼1.2% jump in NiO6\mathrm{NiO}_{6} octahedral volume (compared with a ∼0.6% increase for NaO6\mathrm{NaO}_{6} octahedra) from the fully monoclinic to fully rhombohedral phases. In the rhombohedral phase both the NiO6\mathrm{NiO}_{6} and NaO6\mathrm{NaO}_{6} octahedra expand on heating at around the same rate.

Image not extracted

Figure 2. (a) Representative Rietveld refinement of synchrotron X-ray diffraction data for NaNiO2_2 at 476.1 K in the mixed-phase regime where the rhombohedral (R3ˉmR\bar{3}m) and monoclinic (C2/mC2/m) phases coexist. (b) Volume per formula unit as a function of temperature, showing the slight increase in unit cell volume of the rhombohedral phase compared with the monoclinic phase. Insets are NaO6_6 (top left) and NiO6_6 (top right) octahedral volume with temperature, calculated using VanVleckCalculator; only volumes in the single-phase regions are plotted. Temperature-dependence of the (c) inter-layer distances, c/3c/3 for the rhombohedral phase and csin⁡(β)c\sin(\beta) for the monoclinic phase, and (d) intra-layer distances, aa for the rhombohedral phase, and a/3a/\sqrt{3} and bb for the monoclinic phase. The intra-layer distances correspond to the Ni-Ni and Na-Na distances within the plane. For the monoclinic phase, closed circles are the individual distances, open circles are averaged. In (c), points are experimental values obtained by Rietveld refinement, and solid lines are a linear fit. In (b,c,d) error bars are smaller than data points.

Figure 2(c) shows that there is a significant increase in inter-layer distance on heating from the monoclinic to rhombohedral structures. Figure 2(d) shows that the two different intra-layer distances (corresponding to Ni-Ni and Na-Na distances within the plane of the layer) of the monoclinic cell equalise in the rhombohedral cell, due to the increase in cell symmetry, and the distance in the rhombohedral cell is slightly decreased compared with the average distance in the monoclinic cell.

23Na^{23}\mathrm{Na} nuclear magnetic resonance.

A variable-temperature 23Na^{23}\mathrm{Na} NMR (VT-NMR) experiment was carried out to investigate changes in the local structure with temperature using both magic angle spinning (MAS) and static measurements. Changes in both of the spectra are observed at TJTT_{\mathrm{JT}} [Figure S37]. At low temperatures a single Na+\mathrm{Na}^{+} environment is observed consistent with the average structure of monoclinic NaNiO2\mathrm{NaNiO}_{2}. A second higher chemical shift environment is observed on heating which we ascribe to Na+\mathrm{Na}^{+} in the high-temperature, rhombohedral phase of NaNiO2\mathrm{NaNiO}_{2}. There is a limited TT-range where both peaks co-exist until at higher TT only a single peak persists. Both environments have large hyperfine shifts due to the presence of paramagnetic Ni3+\mathrm{Ni}^{3+} ions, the peak shifting to lower values as the Ni3+\mathrm{Ni}^{3+} become less paramagnetic at higher temperatures. These results are consistent with the diffraction data and indicate a change in the local Na+\mathrm{Na}^{+} environment at TJTT_{\mathrm{JT}}.

Neutron pair distribution function.

Total scattering neutron experiments have been performed on NaNiO2\mathrm{NaNiO}_{2}. The Bragg data was published previously [27] and is consistent with the synchrotron XRD presented in the previous section.

Pair Distribution Functions (PDFs), presented as D(r)D(r) in Figure 3(a), [42] were obtained from the neutron total scattering data at 293 K and 450 K (both monoclinic average structure), and 500 K (rhombohedral average structure). There are two Ni–O peaks at 293 K and 450 K, consistent with the Ni–O bond length splitting due to a Jahn–Teller distortion, and a single Ni–O peak at 500 K, implying that NiO6\mathrm{NiO}_{6} octahedra are undistorted by the Jahn–Teller effect at this temperature. This is consistent with the picture from the average structure, as described from the variable-temperature synchrotron data. However, at 500 K the Ni–O peak is highly asymmetric with a tail on the high-rr side.

Initial analysis of the PDF data was performed using small-box analysis, also known as real-space Rietveld refinement. Results are presented in Section S3.1 of the SI. The

room-temperature data can be fit well with the JT-distorted, monoclinic structure used for Rietveld analysis of the reciprocal space data. The PDF data at 500 K can be fit adequately with both the monoclinic and the rhombohedral (JT-undistorted) structures, with a slightly higher RwpR_{\mathrm{wp}} for the rhombohedral structure (although this is to be expected given the smaller number of free parameters). In the monoclinic cell, the difference between the Ni-O bond lengths is finite but very small at 500 K (∼0.041 A˚\sim0.041\,\text{\AA}, compared with ∼0.243 A˚\sim0.243\,\text{\AA} at 293 K), which is not consistent with the presence of a local Jahn–Teller distortion.

Ultimately the small-box analysis did not fit all features of the PDF data, and in any case small-box analysis with a single unit cell would be insensitive to some types of local Jahn–Teller ordering. To test for the presence of local Jahn–Teller distortions obscured by the peak asymmetry, big-box analysis was performed on the PDF data, in conjunction with the Bragg scattering data.

To this end, a rhombohedral NaNiO2\mathrm{NaNiO}_{2} unit cell with centred orthorhombic setting was prepared, consisting of 6 formula units and obtained via transformation from the normal rhombohedral cell using Equation S5. A 16×9×316 \times9 \times3 supercell of this unit cell was used as the starting point for big box refinements, with lattice dimensions ∼45 A˚\sim45\,\text{\AA} and containing 10368 atoms; see Methods for further details. The atomic positions in this model were then refined against the experimental data. Here, empirical bond valence sum[43] restraints (discussed in Methods) were applied to the atomic positions to ensure physical behaviour, although qualitatively equivalent results are obtained without using these restraints [SI Section S3.9]. The big box refinements were run several times to ensure repeatability [SI Section S3.11].

An example fit at 500 K to both the reciprocal-space and real-space datasets can be seen in Figure 3(b,Figure 3c) and the resulting supercells are shown in Figure 4. Qualitatively, it can be seen by simple inspection of the O-Ni-O layers in the supercell that there is a cooperative Q3Q_{3} [Figure S1] elongation of the NiO6\mathrm{NiO}_{6} octahedra at 293 K (the Jahn–Teller distortion) which does not seem to be present at 500 K. These supercells are quantitatively analysed in terms of the Ni-O bond length distribution [Figure 5(a)] and it can be seen that below the monoclinic→\torhombohedral transition there are two clusters of Ni–O bond lengths; a shorter cluster representing four bonds and a longer cluster with two bonds. However, at 500 K there is just a single cluster, which is not consistent with Jahn–Teller-distorted octahedra. The model is consistent with the significant peak asymmetry in the Ni–O bond length observed in the PDF data, with the peak asymmetry occurring as a consequence of the broader bond length distribution for longer bonds.

Image not extracted

Figure 3. (a) Experimental neutron PDFs for NaNiO2NaNiO_2 at 293 K, 450 K, and 500 K, in the rr-range where the nearest-neighbour Ni-O peaks occur. Peaks are labelled for the 293 K dataset in blue text and 500 K dataset in green text. The 500 K data is vertically offset from the low-temperature data for ease of distinguishing the changes compared with the lower-TT datasets. Grey dashed horizontal lines occur at D(r)=0D(r) = 0 for the lower two temperatures (lower) and 500 K (upper). (b,c) Example fits to the PDF and Bragg diffraction data at 500 K using the supercell fitting (experimental data: blue; fitted data: orange; difference pattern; green).

Image not extracted

Figure 4. Cross sections of the abab plane for a single O-Ni-O layer of the supercell obtained by big box PDF analysis, showing the distribution of elongated (r>2.1 A˚r > 2.1\,\mathring{\mathrm{A}}) Ni-O bonds at (a) 293 K and (b) 500 K. Na ions are hidden for clarity. Ni: grey; O: red.

Other evidence for the absence of Jahn–Teller distortion in the local structure of rhombohedral NaNiO2_2 can be obtained by considering two metrics for the octahedral distortion. The bond length distortion index (BLDI), [44], Figure 5(b), quantifies the deviation of bond lengths from the average for a given polyhedron and is often used to quantify Jahn–Teller distortions. [27, 46, 47] For a big box model where thermal effects are modelled by deviation of atoms from their average positions (rather than using atomic displacement parameters, as is conventional for Rietveld refinement), we would expect the BLDI to increase with heating in the absence of any electronic or magnetostructural transitions. The average BLDI for all NiO6_6 octahedra is 0.0547 and 0.0575 (to 3 significant figures) at 293 K and 450 K, respectively, in the monoclinic phase, before decreasing to 0.0458 at 500 K. This is clearly shown by the decreasing distribution of bond lengths in Figure 5(a). In contrast, the average BLDI continuously increases with heating for the NaO6_6 octahedra, increasing from 0.0260 at 293 K, to 0.0352 at 450 K, and finally 0.0450 at 500 K (see Figure S10). This trend in the BLDI for NiO6_6 octahedra can be explained by the disappearance of a local Jahn–Teller distortion at TJTT_{\mathrm{JT}}. We can also quantify the distortion using the parameter ρ0\rho_0, where ρ02=Q22+Q32\rho_0^2 = Q_2^2 + Q_3^2. Figure 5(c) shows P(ρ0)P(\rho_0), the probability of octahedra having a given value of ρ0\rho_0, defined as:

Image not extracted

Figure 5. (a) Ni-O bond distributions at 293 K (top), 450 K (middle), and 500 K (bottom), as obtained via big box analysis of neutron PDF data, for the smallest to largest Ni-O bond in each octahedron; vertical lines are the bond lengths from Rietveld refinement of the neutron data reported in Ref. 27 (b) Histogram of bond length distortion index [44] for each NiO6_6 octahedron at 293 K, 450 K, and 500 K. (c) Probability function of distortion [Equation (1)] at 293 K, 450 K, and 500 K. (b) and (c) are obtained using VanVleckCalculator.[6, 45]

P(ρ0)=f(ρ0)2πρ0(1)P(\rho_0) = \frac{f(\rho_0)}{2\pi\rho_0} \tag*{(1)}

where f(ρ0)f(\rho_0) is a histogram of ρ0\rho_0 for all NiO6_6 octahedra. This parameterisation shows maximum probability at ρ0≈0\rho_0 \approx0 at 500 K, compared with large non-zero magnitudes at 293 K and 450 K, which also indicates the absence of a local Jahn–Teller distortion in the rhombohedral structure.

dral phase. This result that the most probable ρ0\rho_0 substantially decreases at 500 K, compared with lower temperatures, is resilient to other ways of running the big box refinement, for instance with different starting configurations [SI Section 3.7], changes in the BVS restraints [SI Section S3.9], and multiple runs with the same starting parameters [SI Section 3.11].

X-ray absorption spectroscopy.

XAS at the Ni K edge was performed, with analysis focusing on the extended X-ray absorption fine structure (EXAFS) in real-space. This data, χ(r)\chi(r), is analogous to a partial PDF of Ni-O bond lengths, but with the significant difference that there is a phase shift resulting

in peaks in χ(r)\chi(r) being down-shifted by around 0.5 Å. It is therefore a useful supplement to the PDF data. The room-temperature χ(r)\chi(r) data presented here resembles the EXAFS on NaNiO2_2 reported previously. [48]

Both the X-ray absorption near edge structure (XANES) and EXAFS change on heating through TJTT_{\mathrm{JT}} [Figure 6]. In the XANES, although there is no change in peak position (indicating that the Ni oxidation state does not change from Ni3+^{3+}) there is a pre-edge feature around 8335 eV in dμ/dEd\mu/dE which is more prominent for the T<TJTT < T_{\mathrm{JT}} regime than the T>TJTT > T_{\mathrm{JT}} regime. This pre-edge is likely due to quadrupole forbidden transitions, and will be less prominent in higher symmetry Ni environments. This observation is therefore consistent with the absence of a local Jahn–Teller distortion at T>TJTT > T_{\mathrm{JT}}.

The EXAFS data, Figure 6(b), also shows a significant change on heating. The two peaks corresponding to two Ni-Ni interatomic distances at ∼\sim2.5 Å when T<TJTT < T_{\mathrm{JT}} merge into a single peak in the T>TJTT > T_{\mathrm{JT}} regime, consistent with the single Ni-Ni interatomic distance in the rhombohedral structure. The peak at ∼\sim1.5 Å corresponds to the Ni-O bond, which exhibits a far more subtle change in shape through the Jahn–Teller transition than the Ni-Ni peak, and fitting to the data is required to examine the changes in local structure.

Image not extracted

Figure 6. (a) XANES at the Ni K edge at various temperatures, where each spectrum is offset by an arbitrary amount for clarity; inset shows dμ/dEd\mu/dE for the 298.15 K and 533.15 K, with an asterisk marking the feature on the rising edge. (b) EXAFS data at the Ni K edge, with heating, for NaNiO2_2 at various temperatures, with the solid line representing the fit with a {2,2} model consisting of two Ni-O bond lengths and two Ni-Ni interatomic distances. The feature around 3.4 Å is the Ni-Na atomic distance, which was not included in any of the fitting. (c) Short and long Ni-O bond lengths obtained from fitting the {2,2} and {2,1} models, to the EXAFS data; Ni-O lengths from fitting the {1,1} model are also shown. (d) Figures of merit for fits performed with the {2,0} and {1,0} models, reduced χ2\chi^2 and Bayesian Information Criteria (BIC), in the vicinity of the Ni-O bond only.

Models were obtained using FEFF[49] and refined against the χ(r)\chi(r) data over different ranges. Two different R(A˚)R(\mathrm{\AA}) ranges were considered; one with just the Ni-O shell R(A˚)=0.5R(\mathrm{\AA}) = 0.5–2.02.0 Å, and the other R(A˚)=0.5R(\mathrm{\AA}) = 0.5–3.13.1 Å to also include the Ni-Ni interatomic distances. Models with both 1 or 2 Ni-O and Ni-Ni distances were considered and we describe these using curly braces enclosing the comma-separated values for the number of lengths for each bond; for instance a model with a single Ni-O distance in the reduced range is described as {1,0}, whereas a model with two Ni-O bond lengths and two Ni-Ni bond lengths is described as {2,2}. Five models were used, over both ranges: {1,0}, {2,0}, {1,1}, {2,1}, and {2,2}. Models in which the first number is 1 correspond to JT-undistorted, and models where the first element is 2 are compatible with a JT distortion.

The data in the Ni-O range (0.5 Å to 2.0 Å) were fitted well with both the JT-distorted {2,0}\{2,0\} and JT-undistorted {1,0}\{1,0\} models Figure S32 at all studied temperatures, with the {2,0}\{2,0\} model consistently fitting better than the {1,0}\{1,0\} model. It is then necessary to determine whether the improved fit is due to a more realistic model or simply due to the additional number of refined parameters with the JT-distorted {2,0}\{2,0\} model. We consider two figures of merit for evaluating fitting quality, the reduced χ2\chi^2 (rχ2r\chi^2) and the Bayesian Information Criterion (BIC).[50] The TT-dependence of these parameters, Figure 6(d), indicates that, in the monoclinic regime, model {2,0}\{2,0\} better describes the data (consistent with the established picture of static cooperative JT-distorted NiO6_6 octahedra), but when T>TJTT > T_{\mathrm{JT}} the model {2,0}\{2,0\} only achieves a better fit as it has more refined parameters; i.e., the JT-undistorted model {1,0}\{1,0\} is the better description of the Ni-O shell. In addition, at T>TJTT > T_{\mathrm{JT}} the short and long Ni-O bond lengths converge within error for the {2,0}\{2,0\} model Figure S33.

Over the larger range (0.5 Å to 3.1 Å) fitting was performed using models {2,2}\{2,2\}, {2,1}\{2,1\}, and {1,1}\{1,1\}. When T<TJTT < T_{\mathrm{JT}}, the {1,1}\{1,1\} and {2,1}\{2,1\} models perform poorly as they cannot reproduce the Ni-Ni splitting observed experimentally, Figure 6(b). The temperature-dependence of the figures of merit, Figure S30, indicate that below TJTT_{\mathrm{JT}}, {2,2}\{2,2\} is the favoured model, consistent with diffraction. Above TJTT_{\mathrm{JT}}, all three models qualitatively fit the data well, but the two figures of merit support different models; BIC favours the undistorted model {1,1}\{1,1\} whereas rχ2r\chi^2 the {2,2}\{2,2\} model.

Does the EXAFS data support the possibility of a local Jahn–Teller distortion in NaNiO2_2 above TJTT_{\mathrm{JT}}? This seems unlikely for the following reasons. Firstly, when the fit is restricted solely to the Ni-O bonds the JT-undistorted model {1,0}\{1,0\} fits best according to both metrics. Secondly, the Ni-O bond length separation for the {2,2}\{2,2\} model (∼0.15 Å, Figure 6(c)), are significantly reduced compared to the separation at T<TJTT < T_{\mathrm{JT}} (around 0.27 Å), which differs from the order-disorder transition in LaMnO3_3 where there is little change in bond length from EXAFS.[15] Thirdly, when we plot histograms of the four shortest and two longest Ni-O bonds in the big box PDF data at 500 K, Figure S10, which is analogous to the {2,...}\{2,...\} models used for the EXAFS fitting, we also observe a bond length separation despite other forms of analysis showing the absence of Jahn–Teller distortions according to the PDF data; this suggests that the {2,2} EXAFS model is consistent with a thermally-disordered, JT-undistorted state above TJTT_{\mathrm{JT}}. For these reasons, even if the {2,2} model is favoured by EXAFS above TJTT_{\mathrm{JT}}, this could be explained without requiring a local Jahn–Teller distortion. We therefore interpret the EXAFS data as consistent with the absence of local Jahn–Teller distortions for T>TJTT > T_{\mathrm{JT}}.

Ab initio molecular dynamics.

Ab initio molecular dynamics simulations of NaNiO2_2 simulation cells starting with a collinear ordering of Jahn–Teller distortions, in a 3×3×33 \times3 \times3 supercell (216 ions) expanded from the monoclinic cell, were performed at temperatures between 34 and 1207 K [Figure 7(a,c)]. Both the cell shape and volume were allowed to change. At low temperatures, the AIMD trajectories show thermal vibrations but the JT distortions persist, and remain collinear. On heating, the distortions decrease in magnitude and are no longer aligned to a single axis.

Image not extracted

Figure 8. (a) Calculated Ni-O pair distribution functions from AIMD; these are convolved with experimental Qmax⁡Q_{\max} in Figure S36. (b) The probability function of ρ0\rho_0 [Equation (1)] from AIMD, for the cells with the C2/mC2/m starting structure. (c) Ni-O bond lengths at two temperatures for AIMD runs from the C2/mC2/m starting structure.

At high temperatures, the octahedra approach a JT-undistorted state.

The temperature-dependence of the Jahn–Teller distortions is quantified by looking at the Ni-O PDF [Figure 8(a)] and the probability function of ρ0\rho_0 [Figure 8(b)]. At low temperatures (34 K), the PDF shows distinct peaks corresponding to short and long Ni-O bonds with a large, positive ρ0\rho_0. On heating (972 K, 1207 K) the two peaks in the PDF merge to a single broad asymmetric peak. ρ0\rho_0 significantly decreases, with the highest probability at ρ0=0\rho_0 = 0. This can also be seen in the heat map distributions in Eg(Q2,Q3)E_g(Q_2, Q_3) shown in Figure 7.

Image not extracted

Figure 7. (a,c,d,f) Snapshots of a NiO6_6 layer as obtained from AIMD simulations with C2/mC2/m (a,c) and Rˉ3m\bar{R}3m (d,f) starting configurations, at various temperatures. Ni: grey; O: red. Ni-O bonds above 2.1 Å are shown. (b,e) Van Vleck Eg(Q2,Q3)E_g(Q_2,Q_3) diagrams as a function of temperature for both configurations, showing the displacive transition; note that a rotation of 120° from the Q2=0Q_2 = 0 line means a change in the axis of elongation. Analogous figures based on the PDF big box analysis can be seen in Figure S11.

However, at these high temperatures there remains a residual distortion evident from a small asymmetry in the Ni-O PDF distribution, with a preference for elongation along the collinear axis [Figure 7(b)] 1207 K. We attribute this to the finite simulation cell size enforcing a periodicity in the structure that may cause longer-range correlations of the NiO6_6 octahedra. To explore this, simulations were run from a similarly-sized starting configuration (4 × 4 × 1 expansion of the rhombohedral unit cell; 192 ions) without JT distortions, and hence without inter-octahedral correlations to begin with, cooling the sample from 1000 K. The same underlying behaviour is observed as seen in the simulations starting from JT distorted simulation cells, i.e. distortions are observed at T<TJTT < T_{\mathrm{JT}} that vanish at T>TJTT > T_{\mathrm{JT}} and (re)emerge at T<TJTT < T_{\mathrm{JT}}. At ca. 1000 K, the NiO6_6 octahedra are fully isotropic [Figure 7(e) 1000 K]. On cooling, Jahn–Teller distortions emerge. These are row-ordered [Figure 7(e) 300 K], i.e. arranged collinearly within a row, rather than across the simulation cell, as predicted previously. [33] The rows show a preference for two out of the three possible directions of the distortions (i.e. arctan⁡(Q2/Q3)=0\arctan(Q_2/Q_3) = 0 or 2π/32\pi/3 but never 4π/34\pi/3), resulting in two clusters of data points in the van Vleck plot at low temperatures [Figure 7(e) 50 K]. This is likely a consequence of the very small domain sizes, resulting presumably from the rapid cooling on the timescale of the AIMD simulations (a few picoseconds), not allowing the system to arrange into a macroscopically collinear ground state. We note that in the transition regime for both simulation cells, the octahedral distortions are elongated along two out of the three octahedral axes [Figure 7(b) 972 K, (e) 300 K], suggesting the transitions between the collinearly distorted phase with elongations along one axis and the displacive phase with random fluctuations along all axes may occur through an intermediate state with distortions occurring preferentially along two axes. If cooled slowly, the distortions are expected to order collinearly, whereas the rapid quenching of the AIMD simulation freezes the distortions in.

The calculated transition temperature between the JT-distorted state and the JT-undistorted state depends on the initial configuration of the simulation cell, and this does not change when the length of the AIMD runs is extended further. On heating the distorted simulation cell, the onset of the JT-transition TonsetT_{\mathrm{onset}} is found to occur at around 586 K, i.e. at a slightly higher temperature than the experimental onset temperature obtained from synchrotron X-ray diffraction, 460 K. On cooling the undistorted simulation cell, the onset of the transition TonsetT_{\mathrm{onset}} is below 300 K, i.e. slightly lower than the experimental onset temperature. A closer look at these differences in TonsetT_{\mathrm{onset}} reveals that there are two factors causing the deviations.

First, as the transition is a first-order transition, hysteric behaviour is expected, i.e. the transition temperature is expected to differ between heating and cooling, and, albeit to a smaller extent, hysteresis has been observed experimentally, too.[25, 26] Second, the AIMD simulations starting from the distorted and undistorted simulation cells result in different domain sizes, affecting the transition temperatures further. The cooperative collinear distortions of the distorted simulation cell correspond to a scenario of infinitely large domains which shift the transition temperature to its maximum value. When cooling the undistorted simulation cell, very small domains form [Figure 7(e) 300 K] resulting in short-range correlations between octahedral distortions. The transition temperature obtained from cooling the undistorted simulation cell therefore constitutes a lower cutoff to the transition temperature from the JT-undistorted phase to the distorted phase. The AIMD simulations thus predict the transition from the JT-distorted to the JT-undistorted phase to occur at T<586T < 586 K, and the transition from the JT-undistorted to the JT-distorted phase at temperatures T>300T > 300 K. The onset temperature derived for the transition to the JT-undistorted phase based on synchrotron XRD of samples of intermediate domain sizes in this study, 460 K, lies approximately in the middle of the predicted temperature window.

Discussion

Our studies of the local Ni3+\mathrm{Ni}^{3+} environment, using neutron PDF and Ni K edge EXAFS, suggest the absence of local static Jahn–Teller-distortions for T>TJTT > T_{\mathrm{JT}}. Further, while a dynamic Jahn–Teller distortion[4] has previously been suggested for LiNiO2\mathrm{LiNiO}_{2}[51, 52, 53], by analogy with similar proposals for LaMnO3\mathrm{LaMnO}_{3}[54], such a dynamic Jahn–Teller distortion would also be resolvable via our neutron PDF and EXAFS measurements. These observations suggest that the JT transition in NaNiO2\mathrm{NaNiO}_{2} is better characterised as displacive rather than order-disorder.[13] This is different to the observations of an order-disorder transition in LaMnO3\mathrm{LaMnO}_{3}[14, 15, 18, 19], and the proposed order-disorder transition for NaNiO2\mathrm{NaNiO_2}. [13] To our knowledge, only in LiNiO2\mathrm{LiNiO_2} [17] has prior evidence for a displacive Jahn–Teller transition been put forward, using AIMD and inferences from changes in strain.

As in previous reports we observe a first order phase transition in NaNiO2\mathrm{NaNiO_2} with the monoclinic and rhombohedral phases coexisting over a finite temperature range. The observed discontinuous increase in unit cell volume from the monoclinic to the rhombohedral structures [Figure 2] is also consistent with a first-order phase transition. The positive thermal expansion through the transition is consistent, via the Clausius-Clapeyron equation, with our previous finding[27] of dTJT/dP>0dT_{\mathrm{JT}}/dP > 0. We note that in the displacive transition in SrTiO3\mathrm{SrTiO_3}[55] a positive dTa/dPdT_{\mathrm{a}}/dP is also observed. In contrast, a negative thermal expansion is found at the transition temperature in LaMnO3\mathrm{LaMnO_3}[56] and ferroelectric PbTiO3\mathrm{PbTiO_3}[57, 58], both order-disorder transitions, with dTC/dP<0dT_{\mathrm{C}}/dP < 0.[59, 60] The displacive transition may therefore be related to the sign of dTJT/dPdT_{\mathrm{JT}}/dP.

In NaNiO2\mathrm{NaNiO_2}, the high-temperature phase will be enthalpically unfavourable and hence the phase transition will be entropically-driven, according to the free energy ΔG=ΔH−TΔS\Delta G = \Delta H - T\Delta S, where ΔH\Delta H is the enthalpy change, TT is temperature, and ΔS\Delta S is the entropy change. There are two relevant contributions to the entropy: configurational and vibrational. We find the configurational entropy of orbital disorder for LiNiO2\mathrm{LiNiO_2} and NaNiO2\mathrm{NaNiO_2} to be subextensive (see Appendix A), meaning it does not increase with increasing system size.[61] A subextensive configurational entropy has also been found for an Ising model on an elastic triangular lattice.[62] Consequently the entropy term is dominated by dynamic vibrational effects which disfavour local Jahn–Teller distortions. This is likely a contributing factor in the displacive JT transition we have observed, and is in contrast to high-temperature LaMnO3\mathrm{LaMnO_3} where, by analytic and Monte-Carlo calculations using the Potts model, there is extensive configurational entropy of orbital disorder[20, 21] and hence an order-disorder transition.

An alternative hypothesis we considered for the disappearance of Jahn–Teller distortions in high-temperature NaNiO2\mathrm{NaNiO_2} was the possibility for delocalisation of the electrons resulting in a change in the energy landscape of the system. This is by analogy to the insulator-to-metal transition (IMT) previously observed in the perovskite nickelates. [63, 64] However, while variable-temperature conductivity, σ(T)\sigma(T), data from Ref.[65] [Figure S6], shows an increase in conductivity by several orders of magnitude on heating, there is no change of sign of dσ(T)/dTd\sigma(T)/dT. This is similar to the conductivity data for LaMnO3_3, [54] which would suggest the electronic conductivity is not closely related to the order-disorder/displacive nature of the transition.

There are a number of other differences between NaNiO2_2 and LaMnO3_3 which could affect the Jahn–Teller transition. For instance, the edge-sharing octahedral connectivity in NaNiO2_2 results in much larger intra-octahedral angular distortion compared with the corner-sharing octahedra in LaMnO3_3. This difference in connectivity is a consequence of the triangular Ni network in NaNiO2_2. Given the constraint that axes of elongation of adjacent JT-distorted Ni atoms cannot point at the same O anion, there are no possible long-range JT-disordered configurations, [30] which means that this triangular Ni network cannot host a Potts-type Jahn–Teller disorder such as that proposed for LaMnO3_3. [16, 20, 21] Another difference is that d4d^4 Mn3+^{3+} is generally found to be JT-distorted, whereas nominally d7d^7 Ni3+^{3+} is susceptible to other types of electronic configuration such as charge disproportionation; [36, 38, 39, 40] [37] this may be due to the smaller electronegativity difference between Ni–O, compared with Mn–O, which has been argued to cause charge-transfer insulating [66] behaviour in many nickelates. [64, 67, 68] Attempting to deconvolute the role of these different factors will require further study on the transition in other Jahn–Teller-distorted materials, along with theoretical and computational work.

Whether or not a JT transition is displacive or order-disorder may be important in some battery cathode materials. This will be particularly apparent if the JT transition occurs below room-temperature, as it does in, for instance, the NaNix_xCo1−x_{1-x}O2_2 solid solution for x≤0.8x \le0.8 [65] and the spinel LiMn2_2O4_4.[69] In these cases, the presence or absence of local JT distortions will result in different ionic mobility, and will impact the crystallographic changes that occur with electrochemical cycling. The implications of this on electrochemistry should be studied in future works.

Conclusion

We have shown that the Jahn–Teller transition in NaNiO2\mathrm{NaNiO_2} is displacive in character, with the absence of local JT-distortions or orbital ordering in the NiO6\mathrm{NiO_6} octahedra at T>TJTT > T_{\mathrm{JT}} (TJTT_{\mathrm{JT}} being the cooperative Jahn–Teller transition temperature). To our knowledge this is the first time that local probes of structure have shown a displacive Jahn–Teller transition, though order-disorder transitions have been reported for other JT-distorted perovskite LaMnO3\mathrm{LaMnO_3}.[14, 15, 18, 19]

Our findings are not consistent with previous computational studies on NaNiO2\mathrm{NaNiO_2}, which predicted an order-disorder transition with persisting local JT distortions,[13] but our own ab initio molecular dynamics calculations support our experimental findings. This complementarity between local probe experiments and simulation provides further support for our earlier ab initio molecular dynamics study on LiNiO2\mathrm{LiNiO_2} which also predicted a displacive JT transition.[17]

We found that the configurational entropy of orbital disorder is subextensive in layered transition metal oxides such as NaNiO2\mathrm{NaNiO_2}, and contrast this to the extensive configurational entropy[20] in the case of perovskites such as LaMnO3\mathrm{LaMnO_3}. Further theoretical work is required to better understand these different behaviours.

Methods

Sample synthesis.

Samples were prepared by solid state synthesis. Na2O2\mathrm{Na_2O_2} (Alfa Aesar; 95%) and NiO (Alfa Aesar; 99.995%) were mixed and pelletised in a 1.05:1 molar ratio of Na:Ni, with excess Na to account for Na-loss during heating. The pelletised precursor mixture was placed in an alumina crucible and heated to 973 K973\ \mathrm{K} for 70 hrs in a tube furnace under a constant flow of O2\mathrm{O}_2. O2\mathrm{O}_2 was maintained throughout the heating and cooling processes. The sample showed a colour change from light green (the NiO/Na2O2\mathrm{NiO}/\mathrm{Na}_2\mathrm{O}_2 precursor mixture) to black, indicating successful synthesis. To prevent reaction with moisture, the sample was stored and handled in an inert Ar-atmosphere. The same sample was previously studied in Ref., [27] where electron microscopy and Rietveld[41] analysis of laboratory X-ray diffraction and neutron diffraction data are presented.

Synchrotron X-ray diffraction.

Variable-temperature, ambient-pressure synchrotron X-ray diffraction was performed using the I11 instrument[70, 71] at Diamond Light Source (λ=0.824110 A˚\lambda= 0.824110\ \mathring{\mathrm{A}}) using the Mythen-2 position-sensitive detector, with a data collection time of ∼10\sim10 seconds. The sample was contained in a 0.5 mm0.5\ \mathrm{mm} diameter glass capillary, sealed inside a glovebox with epoxy (Loctite Double Bubble). The sample was heated at a rate of 12 K/min12\ \mathrm{K}/\mathrm{min} from 322 K322\ \mathrm{K} to 796 K796\ \mathrm{K}, with periodic measurements.

Data were analysed by sequential Rietveld refinement[41] using the software package TOPAS[72]. A pseudo-Voigt peak function was used, and background was fitted using a Chebyshev polynomial with 20 terms. A small correction for preferred orientation was applied, using the March-Dollase model.[73, 74]

Nuclear magnetic resonance.

Samples for variable-temperature nuclear magnetic resonance (VT-NMR) were packed into 4 mm4\ \mathrm{mm} ZrO2\mathrm{ZrO}_2 magic angle spinning (MAS) rotors in an Ar filled glovebox, and fitted with ZrO2\mathrm{ZrO}_2 caps. 23Na{}^{23}\mathrm{Na} chemical shifts were calibrated using solid NaCl as an external secondary reference (7.21 ppm7.21\ \mathrm{ppm} relative to 1 M1\ \mathrm{M} NaCl(aq) at 0.0 ppm0.0\ \mathrm{ppm}). 23Na{}^{23}\mathrm{Na} MAS NMR spectra were acquired using a Bruker Avance IIIHD spectrometer (ν0[1H]=500.13 MHz\nu_0[{}^1\mathrm{H}] = 500.13\ \mathrm{MHz}, ν0[23Na]=132.46 MHz\nu_0[{}^{23}\mathrm{Na}] = 132.46\ \mathrm{MHz}, ν0[207Pb]=104.26 MHz\nu_0[{}^{207}\mathrm{Pb}] = 104.26\ \mathrm{MHz}) spectrometer, with a Bruker 4 mm4\ \mathrm{mm} HX probe and 14 kHz14\ \mathrm{kHz} MAS.

Projection magic-angle turning and a phase-adjusted sideband separation (pj-MATPASS) pulse sequence was used to obtain the isotropic 23Na{}^{23}\mathrm{Na} spectra.[75]

Static VT-NMR samples were measured in the same spectrometer using a Bruker HX static probe. A Hahn-echo pulse sequence with π/4=2.05 μs\pi/4 = 2.05\,\mu\mathrm{s} optimized on solid NaNiO2\mathrm{NaNiO}_{2} was used. Temperature was calibrated through a NMR shift thermometer compound Pb(NO3)2\mathrm{Pb}(\mathrm{NO}_{3})_{2}, based on the known temperature-dependence of the isotropic chemical shift of Pb(NO3)2\mathrm{Pb}(\mathrm{NO}_{3})_{2}.[76] 207Pb{}^{207}\mathrm{Pb} NMR spectra were acquired with the same probe at the same temperature values, using a Hahn-echo pulse sequence. The isotropic chemical shift values for Pb(NO3)2\mathrm{Pb}(\mathrm{NO}_{3})_{2} were obtained by fitting the data using the SOLA fitting program within TOPSPIN 4.1.4.

Neutron total scattering.

Variable-temperature total-scattering neutron diffraction was performed on the NOMAD instrument[77] at the Spallation Neutron Source, Oak Ridge National Laboratory, USA. NaNiO2\mathrm{NaNiO}_{2} was stored under Ar in a sealed NMR tube (2 cm sample height, 5 mm outer diameter) for the measurements. Heating was performed using a furnace, at a rate of 5∘K5^\circ\mathrm{K} per minute. The sample was measured on heating at 293 K, 450 K, 500 K, and after cooling at 316 K. The neutron diffraction data and the refined crystal structure was previously published in Ref.[27] In the Fourier transform to real space, a sliding Qmax⁡Q_{\max} as a function of rr is used, with Qmax⁡=40 A˚−1Q_{\max} = 40\,\text{\AA}^{-1} at low-rr and gradually decreasing Qmax⁡Q_{\max} beyond that, with Qmax⁡=20 A˚−1Q_{\max} = 20\,\text{\AA}^{-1} above r=25 A˚r = 25\,\text{\AA}.

Analysis of pair distribution function.

Small-box pair distribution function analysis, also known as real-space Rietveld refinement, was performed using TOPAS 6.[72] Data were fitted in the range 1.0 A˚\text{\AA} to 10 A˚\text{\AA} in real space only. The starting structures were the normal C2/mC2/m and R3ˉmR\bar{3}m structures used commonly in the literature for the low- and high-temperature phases of NaNiO2\mathrm{NaNiO}_{2}.

Pair distribution function analysis was performed via a “big box” approach used TOPAS [7, 72] following a modified version of the method introduced in Ref. 78 A supercell model was refined against both the Bragg data from NOMAD bank 4 (as defined in Ref. 77) and the experimental neutron pair distribution function. In this refinement, atomic coordinates were refined until convergence following the Rietveld algorithm. [41] During initial iterations, restraints were applied to prevent large atomic movements, but these were removed in later stages of the refinement. Upon convergence, each atom was randomly shifted by a distance less than or equal to 0.1%0.1\% of the unit cell size (∼0.045\sim0.045 Å) in each spatial dimension to help minimisation. This typically occurred several hundred times. The final configuration was the converged structure with the best fit quality. Our approach differs from the more commonly-used approach in RMCProfile. [79] Data were fitted in the range 0.50.5 Å to 2020 Å in real space.

A 16×9×316 \times9 \times3 supercell of a pseudo-orthorhombic cell (space group: P1P1; a≈2.846a \approx2.846 Å; b≈5.321b \approx5.321 Å; c≈15.703c \approx15.703 Å; β=γ=90∘\beta= \gamma= 90^\circ; α=90∘\alpha= 90^\circ for the 500500 K supercell and ∼88∘\sim88^\circ for the T<TJTT < T_{\mathrm{JT}} supercells) was used for the refinements. This NaNiO2_2 unit cell was obtained via transformation from the rhombohedral cell using Equation S5 in Supplementary Information. The supercell has edge lengths ∼45\sim45 Å and contains 1036810368 atoms. In this analysis, the thermal parameters for all atoms were fixed at low values (Biso=0.01B_{\mathrm{iso}} = 0.01 Å2^2) such that thermal effects are modelled by the deviation of each site from its ideal position. To ensure that chemically reasonable models were obtained, soft restraints were included. The relative weighting of restraints against data was set to ensure refinements were not dominated by restraints. The SI includes refinements performed in the absence of restraints to check their influence on the conclusions (Section S3.9). The main soft restraint was the calculated bond valence sum [43] (as an analog for oxidation state), VcalcV_{\mathrm{calc}}, of cations deviates from its expected value (Na: +1+1; Ni: +3+3), an established approach [80] in PDF analysis. [78, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103] Here, VcalcV_{\mathrm{calc}} is calculated as follows:

Vcalc=∑i=16exp⁡[r0−riB](2)V_{\mathrm{calc}} = \sum_{i=1}^{6} \exp\left[\frac{r_0-r_i}{B}\right] \tag*{(2)}

where the empirical parameters r0r_0 and BB are based on the cation and anion species, and rir_i are the bond lengths between the metal atom and the iith oxygen. In this work, B=0.37B = 0.37 Å and r0Na=1.672r_0^{\mathrm{Na}} = 1.672 Å and r0Ni=1.7335r_0^{\mathrm{Ni}} = 1.7335 Å; these values were selected to give average bond lengths of 2.335 Å and 1.99 Å for NaO6_6 and NiO6_6 octahedra respectively, to match the room-temperature average structure.

Since the P1P1 space group has a floating origin, a restraint was applied to keep the average shift of Ni sites close to zero.

The supercells obtained from refinement were then analysed using VanVleckCalculator, [6, 45] a Python 3[104] code which is based on PyMatGen.[105]

Ni K edge X-ray absorption spectroscopy.

X-ray absorption spectroscopy (XAS) measurements were performed on the BM23 instrument[106] of the European Synchrotron Radiation Facility (ESRF), France. Powdered NaNiO2_2 was mixed homogeneously with a powdered boron nitride binder (∼\sim16 mg:∼\sim70 mg NaNiO2_2:BN) and pressed into a 13 mm pellet in a dry, oxygen-free nitrogen environment. The pellet was cut and placed in a sealed sample holder under flowing helium. Temperature control was achieved using a resistive heater, and at each temperature the sample was left for 6 minutes to thermally equilibrate. At each temperature the sample was measured ten times to ensure reproducibility, and an average spectrum produced. X-ray absorption was measured in the vicinity of the Ni K edge (∼\sim8.3 keV).

X-ray absorption was calculated using the Beer-Lambert law[107] comparing X-ray intensity in ionisation chambers before and after transmission through the sample, and a third ionisation chamber measured the absorption through Ni foil as a reference. Data normalisation was performed using the Python-based[104] Larch[108] package. For the normalisation of the data, a pre-edge range between -350 eV and -45 eV from the edge centre was fit with a straight line, and the post-edge region between 200 eV and 1250 eV was fit with a quadratic polynomial. Data were transformed from EE-space to kk-space via the transform k=2meℏ2(ℏω−E0)k = \sqrt{\frac{2m_{\mathrm{e}}}{\hbar^2}(\hbar\omega- E_0)}, where mem_{\mathrm{e}} is the mass of the electron, E0E_0 is the edge energy, and ω\omega is the frequency of the measured photon. Finally, the EXAFS function χ(k)\chi(k) was obtained by fitting and subtracting a spline to the EXAFS data in kk-space, where RbkgR_{\mathrm{bkg}} was set to 1 Å and the spline was fit within a kk-range of 0 Å−1^{-1} to 18.25 Å−1^{-1}, with a k2k^2-weighting. For the Fourier transform of EXAFS from kk-space to RR-space, a kk-range of 2.5 Å−1^{-1} to 15 Å−1^{-1} was used, and to minimise Fourier ripples from the cutoffs in kk-space, a Kaiser-Bessel window with end width dk=7dk = 7 Å−1^{-1} was applied to the data in kk-space before the transformation.

Calculated EXAFS data were obtained by multiple-scattering path expansion using FEFF,[49] as implemented in Larch in Python. In this model, data is fit using the EXAFS equation over each set of equivalent paths jj, defined as follows:

χ(k)=S02∑jNj∣fj(k)∣kRj2sin⁡[2kRj+ϕj(k)]×exp⁡[−2Rjλj(k)]exp⁡(−σj2k2)(3)\chi(k) = S_0^2 \sum_j \frac{N_j |f_j(k)|}{kR_j^2} \sin[2kR_j + \phi_j(k)] \times\exp\left[-\frac{2R_j}{\lambda_j(k)}\right]\exp(-\sigma_j^2 k^2) \tag*{(3)}

where FEFF is used to calculate scattering amplitude ∣fj(k)∣|f_j(k)|, photoelectron mean free path λj(k)\lambda_j(k), and phase ϕj(k)\phi_j(k) for all paths. NjN_j is the number of equivalent paths for each jj, which is fixed. The amplitude reduction factor S02S_0^2 was kept fixed at a value obtained by fitting against the Ni foil reference, S02=0.79286024S_0^2 = 0.79286024. The path distance RjR_j and Debye-Waller factors σj\sigma_j for each path were refined during model fitting. Edge position E0E_0 was also freely refined.

Ab initio molecular dynamics.

AIMD simulations were performed according to the Generalized Gradient Approximation (GGA),[109] and the projector augmented wave method (PAW),[110] as implemented in the Vienna Ab Initio Simulation Package (VASP).[111, 112] The plane-wave energy cutoff was set to 500 eV. Supercells are expansions of the monoclinic (3 × 3 × 3; 216 ions) and rhombohedral (4 × 4 × 1; 192 ions) unit cells, respectively. Simulations were performed at the Γ\Gamma point and checked against calculations with a 2 × 2 × 2 Monkhorst-Pack[113] kk-point grid for consistency.

The convergence criteria for the electronic and ionic relaxations were set to 10−6 eV10^{-6}\,\mathrm{eV} and 5⋅10−3 eV/A˚5 \cdot10^{-3}\,\mathrm{eV}/\text{\AA}, respectively.

For Ni, the 4s23d84s^{2}3d^{8} electrons were treated as valence electrons. To account for the strongly correlated d electrons, a rotationally invariant Hubbard UU parameter[114] of Ueff=6 eVU_{\mathrm{eff}} = 6\,\mathrm{eV} was selected, which was used successfully in previous studies of layered oxide cathodes, both for 0 K DFT calculations and finite temperature AIMD simulations. [17, 115, 116] For oxygen, the 2s22p42s^{2}2p^{4} electrons were considered as valence electrons.

AIMD simulations were performed for the isothermal-isobaric ensemble (NpTNpT, constant pressure, particle number, and temperature) at zero pressure, allowing the cell shape and volume to change. A Langevin thermostat was used with friction coefficients set to zero to minimise impact on the lattice vibrations. Trajectories were run for 3 ps at timesteps of 1 fs. Following an equilibration period of 1 ps, snapshots were sampled every 2-5 fs, using Ovito,[117] for the analysis of the distortions.

Appendix A: proof of subextensive configurational entropy for orbital disorder

Here, a mathematical proof is presented that, given a set of assumptions, there is subextensive configurational entropy of orbital disorder in layered oxide materials based on a triangular array of JT-active cations. We define subextensive as meaning that the configurational entropy does not increase with increasing system size.

This relies on three basic axioms:

  1. The effect on orbital ordering of inter-layer interactions is negligible.

  1. Each oxygen anion participates in precisely one Jahn–Teller-elongated bond.

  1. Jahn–Teller distortions are purely of the Q3Q_{3} tetragonal elongation type.

Axiom (1) is justified because the O-AA-O (AA=Li,Na) layers prevent orbital interactions because the Na does not overly constrain the position of the O atom and there is not much correlation between the O atoms around an Na site. Axiom (2) because participation in more than one JT-elongated bond would make the O anion under-bonded. Axiom (3) because this is how octahedra are at room-temperature in NaMO2\mathrm{Na}MO_2 (MM=Ni,Mn) materials.

Configurational entropy per layer is given by:

Sconfiglayer=ln⁡[Wlayer](4)S^{\mathrm{layer}}_{\mathrm{config}} = \ln[W_{\mathrm{layer}}] \tag*{(4)}

where WlayerW_{\mathrm{layer}} is the number of configurations within the layer. Note we define entropy in units of Boltzmann’s constant kBk_{\mathrm{B}}.

Given Axiom 2, there are only two possible alignments of the elongated axes between neighbouring octahedra. Either the axes of elongation are parallel (“A”) or angled 120∘120^\circ apart (“B”). There are three rows of Ni sites within a layer, separated by an angle 120∘120^\circ. In a monoclinic cell such as in NaNiO2\mathrm{NaNiO}_2 at T<TJTT<T_{\mathrm{JT}}, all interactions would be of the “A” type. If there were true orbital disorder, each chain would be a random combination of these two possible configurations “A” and “B”, for instance ”AABBABAAABAABBABABB”. If LL is the length of such a chain (divided by Ni-Ni distance), then we have the following number of configurations:

Wlayer=3×2L=3×2A1/2(5)W_{\mathrm{layer}} = 3 \times2^L = 3 \times2^{A^{1/2}} \tag*{(5)}

where the 3 factor occurs because there are three directions where chains occur, and we relate LL to an arbitrary area AA via L=A1/2L=A^{1/2}.

Given that NN is proportional to AA, where NN is the number of Jahn–Teller-elongated axes within an area of the single layer, we can use N=αAN=\alpha A and substitute this into the equation for configurational entropy as follows:

Sconfiglayer=ln⁡[3×2α′N1/2]=C+N1/2ln⁡[2](6)S^{\mathrm{layer}}_{\mathrm{config}} = \ln\left[3 \times2^{\alpha' N^{1/2}}\right] = C + N^{1/2}\ln[2] \tag*{(6)}

where we have used α′=α1/2\alpha' = \alpha^{1/2} and used constant C=ln⁡[3]+α′ln⁡[2]C = \ln[3] + \alpha' \ln[2].

If we then calculate configurational entropy per NiO6\mathrm{NiO}_{6} octahedron for a large system, as N→∞N \to\infty, C/N→0C/N \to0 and so:

SconfiglayerN≈N−1/2ln⁡[2](7)\frac{S^{\mathrm{layer}}_{\mathrm{config}}}{N} \approx N^{-1/2}\ln[2] \tag*{(7)}

Therefore, as N→∞N \to\infty, so does SconfiglayerN→0\frac{S^{\mathrm{layer}}_{\mathrm{config}}}{N} \to0. This indicates that the configurational entropy of orbital disorder is subextensive.

Data availability

Data from the BM23 instrument at the European Synchrotron Radiation Facility is available at doi:10.15151/ESRF-ES-962076745. [118] All other data is available in the University of Cambridge repository at doi.org/10.17863/CAM.112349. [119]

Supplementary Information

A document containing Supplementary Information to this article is available. The Supplementary Information contains: A description of the Van Vleck modes. Further analysis of supercell outputs of big box PDF analysis, and results from various different starting configurations. Expanded AIMD results. NMR data. Tabulated results of all analysis techniques.

Acknowledgements

The authors acknowledge Oak Ridge National Laboratory, a United States Department of Energy Office of Science User Facility, for use of the NOMAD instrument at the Spallation Neutron Source (experiment IPTS25164). We acknowledge the European Synchrotron Radiation Facility for provision of beam time on BM23 (experiment CH6437). We acknowledge I11 beamline at the Diamond Light Source, UK, for the synchrotron XRD measurement done under BAG proposal (the data presented in this work under CY34243; essential preliminary data from CY28349). Calculations were performed using the Sulis Tier 2 HPC platform hosted by the Scientific Computing Research Technology Platform at the University of Warwick (EP/T022108/1).

We would like to thank Dr Euan N. Bassey, Lucy Haddad, Dr Anastasia Yu. Molokova, Dr Chloe C. Coates, Dr Farheen N. Sayed, Dr Gheorghe-Lucian Păscuț, and Dr João Elias F. S. Rodrigues for useful discussions.

Crystal structure figures were prepared using VESTA-3120 and graphs were prepared using MatPlotLib.121

Funding

This work was supported by the Faraday Institution (FIRG001, FIRG017, FIRG024, FIRG060).

L.A.V.N-C acknowledges a scholarship EP/R513180/1 to pursue doctoral research from the UK Engineering and Physical Sciences Research Council (EPSRC) and additional funding from the Cambridge Philosophical Society. J.M.A.S. acknowledges support from the EPSRC Cambridge NanoCDT, EP/L015978/1. A.L.G. acknowledges European Research Council (ERC) funding under grant 788144.

References

  1. [1]Jahn, H. A.; Teller, E. Stability of polyatomic molecules in degenerate electronic states—I—Orbital degeneracy. Proceedings of the Royal Society of London. Series A—Mathematical and Physical Sciences 1937, 161, 220–235.
  2. [2]Öpik, U.; Pryce, M. H. L. Studies of the Jahn–Teller effect. I. A survey of the static problem. Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences 1957, 238, 425–447.DOI
  3. [3]Longuet-Higgins, H. C.; Öpik, U.; Pryce, M. H. L.; Sack, R. A. Studies of the Jahn–Teller effect. II. The dynamical problem. Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences 1958, 244, 1–16.
  4. [4]Goodenough, J. B. Jahn–Teller phenomena in solids. Annual Review of Materials Science 1998, 28, 1–27.DOI
  5. [5]Van Vleck, J. H. The Jahn–Teller Effect and Crystalline Stark Splitting for Clusters of the Form XY₆. The Journal of Chemical Physics 1939, 7, 72–84.DOI
  6. [6]Nagle-Cocco, L. A. V.; Dutton, S. E. Van Vleck Analysis of Angularly Distorted Octahedra using Van VleckCalculator. IUCr Journal of Applied Crystallography 2024, 57, 20–33.
  7. [7]Fil, D. V.; Tokar, O. I.; She lankov, A. L.; Weber, W. Lattice-mediated interaction of Cu²⁺ Jahn–Teller ions in insulating cuprates. Physical Review B 1992, 45, 5633.pubmed.ncbi.nlm.nih.gov/10000282
  8. [8]Keller, H.; Bussmann-Holder, A.; Müller, K. A. Jahn–Teller physics and high-Tc superconductivity. Materials Today 2008, 11, 38–46.DOI
  9. [9]Khomskii, D. I.; Streltsov, S. V. Orbital effects in solids: basics, recent progress, and opportunities. Chemical Reviews 2020, 121, 2992–3030.
  10. [10]Kim, H.; Yoon, G.; Park, I.; Park, K.-Y.; Lee, B.; Kim, J.; Park, Y.-U.; Jung, S-K.; Lim, H.-D.; Ahn, D., et al. Anomalous Jahn–Teller behavior in a manganese-based mixed-phosphate cathode for sodium ion batteries. Energy & Environmental Science 2015, 8, 3325–3335.DOI
  11. [11]Li, X.; Wang, Y.; Wu, D.; Liu, L.; Bo, S.-H.; Ceder, G. Jahn–Teller assisted Na diffusion for high performance Na ion batteries. Chemistry of Materials 2016, 28, 6575–6583.DOI
  12. [12]Choi, J. U.; Kim, J.; Hwang, J.-Y.; Jo, J. H.; Sun, Y.-K.; Myung, S.-T. K₀.₅₄[Co₀.₅Mn₀.₅]O₂: New cathode with high power capability for potassium-ion batteries. Nano Energy 2019, 61, 284–294.
  13. [13]Radin, M. D.; Thomas, J. C.; Van der Ven, A. Order-disorder versus displacive transitions in Jahn–Teller active layered materials. Physical Review Materials 2020, 4, 043601.DOI
  14. [14]Qiu, X.; Proffen, T.; Mitchell, J. F.; Billinge, S. J. L. Orbital correlations in the pseudocubic O and rhombohedral R phases of LaMnO₃. Physical Review Letters 2005, 94, 177203.pubmed.ncbi.nlm.nih.gov/15904332
  15. [15]Souza, R. A.; Ramos, A. Y.; Tolentino, H. C. N.; Granado, E. Local structure in LaMnO₃ across the Jahn–Teller transition. Physica Scripta 2005, 2005, 428.
  16. [16]Thygesen, P. M. M.; Young, C. A.; Beake, E. O. R.; Romero, F. D.; Connor, L. D.; Proffen, T. E.; Phillips, A. E.; Tucker, M. G.; Hayward, M. A.; Keen, D. A.; Goodwin, A. L. Local structure study of the orbital order/disorder transition in LaMnO₃. Physical Review B 2017, 95, 174107.DOI
  17. [17]Genreith-Schriever, A. R.; Alexiu, A.; Phillips, G. S.; Coates, C. S.; Nagle-Cocco, L. A. V.; Bocarsly, J. D.; Sayed, F. N.; Dutton, S. E.; Grey, C. P. Jahn–Teller distortions and phase transitions in LiNiO₂: Insights from ab initio molecular dynamics and variable-temperature X-ray diffraction. Chemistry of Materials 2024, 36, 2289–2303.
  18. [18]Araya-Rodriguez, E.; Ramos, A. Y.; Tolentino, H. C. N.; Granado, E.; Oseroff, S. B. Local distortion in LaMnO₃ across the Jahn–Teller transition. Journal of Magnetism and Magnetic Materials 2001, 233, 88–90.
  19. [19]García, J.; Subías, G.; Sánchez, M.; Blasco, J. Jahn–Teller dynamic distortions in LaMnO₃ determined by EXAFS. Physica Scripta 2005, 2005, 594.
  20. [20]Ahmed, M. R.; Gehring, G. A. The phase diagram of an anisotropic Potts model. Journal of Physics A: Mathematical and General 2005, 38, 4047.DOI
  21. [21]Ahmed, M. R.; Gehring, G. A. Potts model for the distortion transition in LaMnO₃. Physical Review B 2006, 74, 014420.DOI
  22. [22]Ahmed, M. R.; Gehring, G. A. Volume collapse in LaMnO₃ studied using an anisotropic Potts model. Physical Review B 2009, 79, 174106.DOI
  23. [23]Dyer, L. D.; Borie Jr, B. S.; Smith, G. P. Alkali metal-nickel oxides of the type MNiO₂. Journal of the American Chemical Society 1954, 76, 1499–1503.DOI
  24. [24]Dick, S.; Müller, M.; Preissinger, F.; Zeiske, T. The structure of monoclinic NaNiO₂ as determined by powder X-ray and neutron scattering. Powder Diffraction 1997, 12, 239–241.DOI
  25. [25]Chappel, E.; Nunez-Regueiro, M. D.; Chouteau, G.; Isnard, O.; Darie, C. Study of the ferrodisorsive orbital ordering in NaNiO₂ by neutron diffraction and submillimeter wave ESR. The European Physical Journal B-Condensed Matter and Complex Systems 2000, 17, 615–622.
  26. [26]Sofin, M.; Jansen, M. New route of preparation and properties of NaNiO₂. Zeitschrift für Naturforschung B 2005, 60, 701–704.DOI
  27. [27]Nagle-Cocco, L. A. V.; Bull, C. L.; Ridley, C. J.; Dutton, S. E. Pressure Tuning the Jahn–Teller Transition Temperature in NaNiO₂. Inorganic Chemistry 2022, 61, 4312–4321.
  28. [28]Pickering, I. J.; George, G. N.; Lewandowski, J. T.; Jacobson, A. J. Nickel K-edge X-ray absorption fine structure of lithium nickel oxides. Journal of the American chemical society 1993, 115, 4137–4144.DOI
  29. [29]Rougier, A.; Delmas, C.; Chadwick, A. V. Non-cooperative Jahn–Teller effect in LiNiO₂: an EXAFS study. Solid State Communications 1995, 94, 123–127.DOI
  30. [30]Chung, J.-H.; Proffen, T.; Shamoto, S.; Ghorayeb, A. M.; Croguennec, L.; Tian, W.; Sales, B. C.; Jin, R.; Mandrus, D.; Egami, T. Local structure of LiNiO₂ studied by neutron diffraction. Physical Review B 2005, 71, 064410.DOI
  31. [31]Chen, Z.; Zou, H.; Zhu, X.; Zou, J.; Cao, J. First-principle investigation of Jahn–Teller distortion and topological analysis of chemical bonds in LiNiO₂. Journal of Solid State Chemistry 2011, 184, 1784–1790.DOI
  32. [32]Chen, H.; Freeman, C. L.; Harding, J. H. Charge disproportionation and Jahn–Teller distortion in LiNiO₂ and NaNiO₂: A density functional theory study. Physical Review B 2011, 84, 085108.DOI
  33. [33]Radin, M. D.; Van der Ven, A. Simulating charge, spin, and orbital ordering: application to Jahn–Teller distortions in layered transition-metal oxides. Chemistry of Materials 2018, 30, 607–618.DOI
  34. [34]Foyevtsova, K.; Elfimov, I.; Rottler, J.; Sawatzky, G. A. LiNiO₂ as a high-entropy charge-and bond-disproportionated glass. Physical Review B 2019, 100, 165104.DOI
  35. [35]Poletaev, A. D.; Cotton, J. P.; Morgan, B. J.; Islam, M. S. Temperature-Dependent Dynamic Disproportionation in LiNiO₂. arXiv preprint arXiv:2211.09047 2022,
  36. [36]Wawrzyńska, E.; Coldea, R.; Wheeler, E. M.; Mazin, I. I.; Johannes, M. D.; Sörgel, T.; Jansen, M.; Ibberson, R. M.; Radaelli, P. G. Orbital degeneracy removed by charge order in triangular antiferromagnet AgNiO₂. Physical Review Letters 2007, 99, 157204. Kang, J.-S.; Lee, S. S.; Kim, G.; Lee, H. J.; Song, H. K.; Shin, Y. J.; Han, S. W.; Hwang, C.; Jung, M. C.; Shin, H. J.; Kim, B. H.; Kwon, S. K.; Min, B. I. Valence and spin states in delafossite AgNiO₂ and the frustrated Jahn–Teller system ANiO₂ (A= Li, Na). Physical Review B 2007, 76, 195122.
  37. [38]García-Muñoz, J.; Rodríguez-Carvajal, J.; Lacorre, P. Neutron-diffraction study of the magnetic ordering in the insulating regime of the perovskites RNiO₃ (R= Pr and Nd). Phys. Rev. B 1994, 50, 978.pubmed.ncbi.nlm.nih.gov/9975764
  38. [39]Mizokawa, T.; Khomskii, D. I.; Sawatzky, G. A. Spin and charge ordering in self-doped Mott insulators. Phys. Rev. B 2000, 61, 11263.DOI
  39. [40]Garcia-Munoz, J. L.; Aranda, M. A. G.; Alonso, J. A.; Martinez-Lope, M. J. Structure and charge order in the antiferromagnetic band-insulating phase of NdNiO₃. Phys. Rev. B 2009, 79, 134432.
  40. [41]Rietveld, H. M. A profile refinement method for nuclear and magnetic structures. Journal of Applied Crystallography 1969, 2, 65–71.DOI
  41. [42]Keen, D. A. A comparison of various commonly used correlation functions for describing total scattering. Journal of Applied Crystallography 2001, 34, 172–177.DOI
  42. [43]Brown, I. D.; Altermatt, D. Bond-valence parameters obtained from a systematic analysis of the inorganic crystal structure database. Acta Crystallographica Section B: Structural Science 1985, 41, 244–247.DOI
  43. [44]Baur, W. H. The geometry of polyhedral distortions. Predictive relationships for the phosphate group. Acta Crystallographica Section B: Structural Crystallography and Crystal Chemistry 1974, 30, 1195–1215.DOI
  44. [45]Nagle-Cocco, L. A. V. GitHub: https://github.com/naglecocco/VanVleckCalculator (accessed 20th August 2024).
  45. [46]Kimber, S. A. J. Charge and orbital order in frustrated Pb₃Mn₇O₁₅. Journal of Physics: Condensed Matter 2012, 24, 186002.DOI
  46. [47]Lawler, K. V.; Smith, D.; Evans, S. R.; Dos Santos, A. M.; Molaison, J. J.; Bos, J.-W. G.; Mutka, H.; Henry, P. F.; Argyriou, D. N.; Salamat, A.; Kimber, S. A. J. Decoupling Lattice and Magnetic Instabilities in Frustrated CuMnO₂. Inorganic Chemistry 2021, 60, 6004–6015.DOI
  47. [48]Jacquet, Q. et al. A fundamental correlative spectroscopic study on LiₓNiO₂ and NaNiO₂. Advanced Energy Materials 2024, 2401413.DOI
  48. [49]Newville, M. EXAFS analysis using FEFF and FEFFIT. Journal of Synchrotron Radiation 2001, 8, 96–100.pubmed.ncbi.nlm.nih.gov/11512993
  49. [50]Neath, A. A.; Cavanaugh, J. E. The Bayesian information criterion: background, derivation, and applications. Wiley Interdisciplinary Reviews: Computational Statistics 2012, 4, 199–203.DOI
  50. [51]Stoyanova, R.; Zhecheva, E.; Friebel, C. Magnetic interactions in layered LiNiO₂ revealed by EPR of Ni³⁺. Journal of Physics and Chemistry of Solids 1993, 54, 9–13.DOI
  51. [52]Sugiyama, J.; Ikedo, Y.; Mukai, K.; Nozaki, H.; Månsson, M.; Ofer, O.; Harada, M.; Kamazawa, K.; Miyake, Y.; Brewer, J. H.; Ansaldo, E. J.; Chow, K. H.; Watanabe, I.; Ohzuku, T. Low-temperature magnetic properties and high-temperature diffusive behavior of LiNiO₂ investigated by muon-spin spectroscopy. Physical Review B 2010, 82, 224412.
  52. [53]Sicolo, S.; Mock, M.; Bianchini, M.; Albe, K. And yet it moves: LiNiO₂, a dynamic Jahn–Teller system. Chemistry of Materials 2020, 32, 10096–10103.DOI
  53. [54]Zhou, J.-S.; Goodenough, J. B. Paramagnetic phase in single-crystal LaMnO₃. Physical Review B 1999, 60, 15002.DOI
  54. [55]Fossheim, K.; Berre, B. Ultrasonic Propagation, Stress Effects, and Interaction Parameters at the Displacive Transition in SrTiO₃. Physical Review B 1972, 5, 3292.DOI
  55. [56]Chatterji, T.; Fauth, F.; Ouladdiaf, B.; Mandal, P.; Ghosh, B. Volume collapse in LaMnO₃ caused by an orbital order-disorder transition. Physical Review B 2003, 68, 052406.
  56. [57]Rossetti, G. A.; Cline, J. P.; Navrotsky, A. Phase transition energetics and thermodynamic properties of ferroelectric PbTiO₃. Journal of Materials Research 1998, 13, 3197–3206.DOI
  57. [58]Pan, Z.; Fang, Y.-W.; Nishikubo, T.; Hu, L.; Kawaguchi, S.; Azuma, M. Tolerance factor control of tetragonality and negative thermal expansion in PbTiO₃-based ferroelectrics. Chemistry of Materials 2022, 34, 2798–2803.DOI
  58. [59]Zhou, J.-S.; Uwatoko, Y.; Matsubayashi, K.; Goodenough, J. B. Breakdown of magnetic order in Mott insulators with frustrated superexchange interaction. Physical Review B 2008, 78, 220402.DOI
  59. [60]Sani, A.; Hanfland, M.; Levy, D. Pressure and temperature dependence of the ferroelectric–paraelectric phase transition in PbTiO₃. Journal of Solid State Chemistry 2002, 167, 446–452.DOI
  60. [61]Camp, P. J.; Fuertes, A.; Attfield, J. P. Subextensive entropies and open order in perovskite oxynitrides. Journal of the American Chemical Society 2012, 134, 6762–6766.DOI
  61. [62]Shokef, Y.; Souslov, A.; Lubensky, T. C. Order by disorder in the antiferromagnetic Ising model on an elastic triangular lattice. Proceedings of the National Academy of Sciences 2011, 108, 11804–11809.DOI
  62. [63]Alonso, J. A.; García-Muñoz, J. L.; Fernández-Díaz, M. T.; Aranda, M. A. G.; Martínez-Lope, M. J.; Casais, M. T. Charge disproportionation in RNiO₃ perovskites: simultaneous metal-insulator and structural transition in YNiO₃. Physical Review Letters 1999, 82, 3871.DOI
  63. [64]Bisogni, V.; Catalano, S.; Green, R. J.; Gibert, M.; Scherwitzl, R.; Huang, Y.; Strocov, V. N.; Zubko, P.; Balandeh, S.; Triscone, J.-M.; Sawatzky, G. A.; Schmitt, T. Ground-state oxygen holes and the metal–insulator transition in the negative charge-transfer rare-earth nickelates. Nature Communications 2016, 7, 1–8.DOI
  64. [65]Delmas, C.; Saadoune, I.; Dordor, P. Effect of cobalt substitution on the Jahn–Teller distortion of the NaNiO₂ layered oxide. Molecular Crystals and Liquid Crystals Science and Technology. Section A. Molecular Crystals and Liquid Crystals 1994, 244, 337–342.DOI
  65. [66]Zaanen, J.; Sawatzky, G. A.; Allen, J. W. Band gaps and electronic structure of transition-metal compounds. Physical Review Letters 1985, 55, 418.pubmed.ncbi.nlm.nih.gov/10032345
  66. [67]Ronda, C.; Arends, G. J.; Haas, C. Photoconductivity of the nickel dihalides and the nature of the energy gap. Physical Review B 1987, 35, 4038.pubmed.ncbi.nlm.nih.gov/9941930
  67. [68]Foyevtsova, K.; Sawatzky, G. A. A Band Theory Perspective on Molecular Orbitals in Complex Oxides. Journal of Modern Physics 2019, 10, 953.DOI
  68. [69]Yamada, A.; Tanaka, M. Jahn-Teller structural phase transition around 280 K in LiMn₂O₄. Materials Research Bulletin 1995, 30, 715–721.DOI
  69. [70]Thompson, S. P.; Parker, J. E.; Potter, J.; Hill, T. P.; Birt, A.; Cobb, T. M.; Yuan, F.; Tang, C. C. Beamline I11 at Diamond: A new instrument for high resolution powder diffraction. Review of Scientific Instruments 2009, 80, 075107.DOI
  70. [71]Thompson, S. P.; Parker, J. E.; Marchal, J.; Potter, J.; Birt, A.; Yuan, F.; Fearn, R. D.; Lennie, A. R.; Street, S. R.; Tang, C. C. Fast X-ray powder diffraction on I11 at Diamond. Journal of Synchrotron Radiation 2011, 18, 637–648.DOI
  71. [72]Coelho, A. A. TOPAS and TOPAS-Academic: an optimization program integrating computer algebra and crystallographic objects written in C++. Journal of Applied Crystallography 2018, 51, 210–218.DOI
  72. [73]March, A. Mathematische Theorie der Regelung nach der Korngestalt bei affiner Deformation. Zeitschrift für Kristallographie-Crystalline Materials 1932, 81, 285–297.DOI
  73. [74]Dollase, W. A. Correction of intensities for preferred orientation in powder diffractometry: application of the March model. Journal of Applied Crystallography 1986, 19, 267–272.DOI
  74. [75]Hung, I.; Zhou, L.; Pourpoint, F.; Grey, C. P.; Gan, Z. Isotropic high field NMR spectra of Li-ion battery materials with anisotropy> 1 MHz. Journal of the American Chemical Society 2012, 134, 1898–1901.DOI
  75. [76]Takahashi, T.; Kawashima, H.; Sugisawa, H.; Baba, T. ²⁰⁷Pb chemical shift thermometer at high temperature for magic angle spinning experiments. Solid State Nuclear Magnetic Resonance 1999, 15, 119–123.pubmed.ncbi.nlm.nih.gov/10670904
  76. [77]Neufeind, J.; Feygenson, M.; Carruth, J.; Hoffmann, R.; Chipley, K. K. The nanoscale ordered materials diffractometer NOMAD at the spallation neutron source SNS. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 2012, 287, 68–75.
  77. [78]Fuller, C. A.; Berrod, Q.; Frick, B.; Johnson, M. R.; Avdeev, M.; Evans, J. S. O.; Evans, I. R. Oxide ion and proton conductivity in highly oxygen-deficient cubic perovskite SrSc₀.₃Zn₀.₂Ga₀.₅O₂.₄. Chemistry of Materials 2020, 32, 4347–4357.DOI
  78. [79]Tucker, M. G.; Keen, D. A.; Dove, M. T.; Goodwin, A. L.; Hui, Q. RMCPprofile: reverse Monte Carlo for polycrystalline materials. Journal of Physics: Condensed Matter 2007, 19, 335218.DOI
  79. [80]Norberg, S. T.; Tucker, M. G.; Hull, S. Bond valence sum: a new soft chemical constraint for RMCPprofile. Journal of Applied Crystallography 2009, 42, 179–184.DOI
  80. [81]Marrocchelli, D.; Madden, P. A.; Norberg, S. T.; Hull, S. Cation composition effects on oxide conductivity in the Zr₂Y₂O₇−Y₃NbO₇ system. Journal of Physics: Condensed Matter 2009, 21, 405403.DOI
  81. [82]Abrahams, I.; Liu, X.; Hull, S.; Norberg, S. T.; Krok, F.; Kozanecka-Szmigiel, A.; Islam, M. S.; Stokes, S. J. A combined total scattering and simulation approach to analyzing defect structure in Bi₃YO₆. Chemistry of Materials 2010, 22, 4435–4445.DOI
  82. [83]Liu, X.; Abrahams, I.; Hull, S.; Norberg, S. T.; Holdynski, M.; Krok, F. A neutron total scattering study of defect structure in Bi₃Nb₀.₅Y₀.₅O₆.₅. Solid State Ionics 2011, 192, 176–180.DOI
  83. [84]Norberg, S. T.; Eriksson, S. G.; Hull, S. Comparison of short-range ion–ion correlations in the α, β and δ phases of Bi₂O₃. Solid State Ionics 2011, 192, 409–412.DOI
  84. [85]Norberg, S. T.; Hull, S.; Ahmed, I.; Eriksson, S. G.; Marrocchelli, D.; Madden, P. A.; Li, P.; Irvine, J. T. S. Structural Disorder in Doped Zirconias, Part I: The Zr₀.₈Sc₀.₂−ₓYₓO₁.₉ (0.0 ≤ x ≤ 0.2) System. Chemistry of Materials 2011, 23, 1356–1364.DOI
  85. [86]Norberg, S. T.; Thomas, P. A.; Tucker, M. G. A neutron total scattering study of local coordination in KTiOPO₄ from room temperature to 900°C. Journal of Physics: Condensed Matter 2011, 23, 175401.DOI
  86. [87]Burbano, M.; Norberg, S. T.; Hull, S.; Eriksson, S. G.; Marrocchelli, D.; Madden, P. A.; Watson, G. W. Oxygen vacancy ordering and the conductivity maximum in Y₂O₃-doped CeO₂. Chemistry of Materials 2012, 24, 222–229.DOI
  87. [88]Chong, S. Y.; Szczecinski, R. J.; Bridges, C. A.; Tucker, M. G.; Claridge, J. B.; Rosseinsky, M. J. Local structure of a pure Bi A site polar perovskite revealed by pair distribution function analysis and reverse Monte Carlo modeling: correlated off-Axis displacements in a rhombohedral material. Journal of the American Chemical Society 2012, 134, 5836–5849.DOI
  88. [89]Norberg, S. T.; Hull, S.; Eriksson, S. G.; Ahmed, I.; Kinyanjui, F.; Biendicho, J. J. Pyrochlore to fluorite transition: The Y₂(Ti₁−xZrx)₂O₇ (0.0 ≤ x ≤ 1.0) system. Chemistry of Materials 2012, 24, 4294–4300.DOI
  89. [90]Leszczynska, M.; Liu, X.; Wrobel, W.; Malys, M.; Krynski, M.; Norberg, S. T.; Hull, S.; Krok, F.; Abrahams, I. Thermal variation of structure and electrical conductivity in Bi₄YbO₇.₅. Chemistry of Materials 2013, 25, 326–336.DOI
  90. [91]Norberg, S. T.; Rahman, S. M. H.; Hull, S.; Knee, C. S.; Eriksson, S. G. The proton conducting electrolyte BaTi₀.₅In₀.₅O₂.₇₅: determination of the deuteron site and its local environment. Journal of Physics: Condensed Matter 2013, 25, 454214.DOI
  91. [92]Keeble, D. S.; Barney, E. R.; Keen, D. A.; Tucker, M. G.; Kreisel, J.; Thomas, P. A. Bifurcated polarization rotation in bismuth-based piezoelectrics. Advanced Functional Materials 2013, 23, 185–190.DOI
  92. [93]Payne, J. L.; Tucker, M. G.; Evans, I. R. From fluorite to pyrochlore: characterisation of local and average structure of neodymium zirconate, Nd₂Zr₂O₇. Journal of Solid State Chemistry 2013, 205, 29–34.DOI
  93. [94]Kalland, L.-E.; Norberg, S. T.; Kyrklund, J.; Hull, S.; Eriksson, S. G.; Norby, T.; Mohn, C. E.; Knee, C. S. C-type related order in the defective fluorites La₂Ce₂O₇ and Nd₂Ce₂O₇ studied by neutron scattering and ab initio MD simulations. Physical Chemistry Chemical Physics 2016, 18, 24070–24080.
  94. [95]Chen, D. P.; Neuefeind, J. C.; Koczkur, K. M.; Bish, D. L.; Skrabalak, S. E. Role of Short-Range Chemical Ordering in (GaN)₁−ₓ(ZnO)ₓ for Photodriven Oxygen Evolution. Chemistry of Materials 2017, 29, 6525–6535.DOI
  95. [96]Diaz-Lopez, M.; Freire, M.; Joly, Y.; Colin, C. V.; Fischer, H. E.; Blanc, N.; Boudet, N.; Pralong, V.; Bordet, P. Local structure and lithium diffusion pathways in Li₄Mn₂O₅ high capacity cathode probed by total scattering and XANES. Chemistry of Materials 2018, 30, 3060–3070.
  96. [97]Borowska-Centkowska, A.; Leszczynska, M.; Krok, F.; Malys, M.; Wrobel, W.; Hull, S.; Abrahams, I. Local structure and conductivity behaviour in Bi₇WO₁₃.₅. Journal of Materials Chemistry A 2018, 6, 5407–5418.
  97. [98]Kitamura, N.; Hayashi, N.; Ishida, N.; Idemoto, Y. Local structure in A-site-deficient perovskite Na₀.₅Bi₀.₅TiO₃ and its effect on electrical conduction. Chemistry Letters 2019, 48, 1398–1401.DOI
  98. [99]Kitamura, N.; Kubo, Y.; Ishida, N.; Idemoto, Y. Study of atomic ordering across the layer in lithium-rich layered positive electrode material towards preparation process optimization. Journal of Power Sources 2019, 437, 226905.DOI
  99. [100]Marlton, F. P.; Nayak, S.; Venkateshwarlu, S.; Chan, N. H.; Kong, J.; Zhang, Y.; Tucker, M. G.; Jørgensen, M. R. V.; Pramanick, A. Broad Distribution of Local Polar States Generates Large Electrothermal Properties in Pb-Free Relaxor Ferroelectrics. Chemistry of Materials 2021, 33, 8844–8853.DOI
  100. [101]Krayzman, V.; Bosak, A.; Playford, H. Y.; Ravel, B.; Levin, I. Incommensurate Modulation and Competing Ferroelectric/Antiferroelectric Modes in Tetragonal Tungsten Bronzes. Chemistry of Materials 2022, 34, 9989–10002.DOI
  101. [102]Miyazaki, R.; Ikeda, K.; Kitamura, N.; Takabayashi, Y.; Kimura, K.; Hayashi, K.; Hihara, T. Reverse Monte Carlo analysis of NaI-LiI solid electrolyte based on the neutron total scattering data. Materials Today Communications 2022, 32, 104014.DOI
  102. [103]Ming, J.; Leszczyńska-Redek, M.; Malys, M.; Wrobel, W.; Jamroz, J.; Struzik, M.; Hull, S.; Krok, F.; Abrahams, I. Dopant clustering and vacancy ordering in neodymium doped ceria. Journal of Materials Chemistry A 2024, 12, 10203–10215.
  103. [104]Van Rossum, G.; Drake, F. L. Python 3 Reference Manual; CreateSpace: Scotts Valley, CA, 2009.
  104. [105]Ong, S. P.; Richards, W. D.; Jain, A.; Hautier, G.; Kocher, M.; Cholia, S.; Gunter, D.; Chevrier, V. L.; Persson, K. A.; Ceder, G. Python Materials Genomics (pymatgen): A robust, open-source python library for materials analysis. Computational Materials Science 2013, 68, 314–319.DOI
  105. [106]Mathon, O.; Beteva, A.; Borrel, J.; Bugnazet, D.; Gatla, S.; Hino, R.; Kantor, I.; Mairs, T.; Munoz, M.; Pasternak, S.; Perrin, F.; Pascarelli, S. The time-resolved and extreme conditions XAS (TEXAS) facility at the European Synchrotron Radiation Facility: the general-purpose EXAFS bending-magnet beamline BM23. Journal of Synchrotron Radiation 2015, 22, 1548–1554.DOI
  106. [107]Kerr, B. V.; King, H. J.; Garibello, C. F.; Disanayake, P. R.; Simonov, A. N.; Johannessen, B.; Eldridge, D. S.; Hocking, R. K. Characterization of energy materials with X-ray absorption spectroscopy: advantages, challenges, and opportunities. Energy & Fuels 2022, 36, 2369–2389.DOI
  107. [108]Newville, M. Larch: an analysis package for XAFS and related spectroscopies. Journal of Physics: Conference Series. 2013; p 012007.DOI
  108. [109]Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized gradient approximation made simple. Physical Review Letters 1996, 77, 3865.pubmed.ncbi.nlm.nih.gov/10062328
  109. [110]Blöchl, P. E. Projector augmented-wave method. Physical Review B 1994, 50, 17953.pubmed.ncbi.nlm.nih.gov/9976227
  110. [111]Kresse, G.; Furthmüller, J. Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set. Physical Review B 1996, 54, 11169.pubmed.ncbi.nlm.nih.gov/9984901
  111. [112]Kresse, G.; Joubert, D. From ultrasoft pseudopotentials to the projector augmented-wave method. Physical Review B 1999, 59, 1758.DOI
  112. [113]Monkhorst, H. J.; Pack, J. D. Special points for Brillouin-zone integrations. Physical Review B 1976, 13, 5188.DOI
  113. [114]Dudarev, S. L.; Botton, G. A.; Savrasov, S. Y.; Humphreys, C. J.; Sutton, A. P. Electron-energy-loss spectra and the structural stability of nickel oxide: An LSDA+U study. Physical Review B 1998, 57, 1505.DOI
  114. [115]Das, H.; Urban, A.; Huang, W.; Ceder, G. First-principles simulation of the (Li-Ni-vacancy) O phase diagram and its relevance for the surface phases in Ni-Rich Li-Ion cathode materials. Chemistry of Materials 2017, 29, 7840–7851.DOI
  115. [116]Genreith-Schriever, A. R.; Banerjee, H.; Menon, A. S.; Bassey, E. N.; Piper, L. F. J.; Grey, C. P.; Morris, A. J. Oxygen hole formation controls stability in LiNiO₂ cathodes. Joule 2023, 7, 1623–1640.
  116. [117]Stukowski, A. Visualization and analysis of atomistic simulation data with OVITO–the Open Visualization Tool. Modelling and simulation in materials science and engineering 2009, 18, 015012.DOI
  117. [118]Nagle-Cocco, L. A. V.; Tacconis, C.; & Steele, J. M. A. DOI:10.15151/ESRF-ES-962076745. [Data set]. (Accessed 20th August 2024).
  118. [119]Nagle-Cocco, L. A. V.; Genreith-Schriever, A. R.; Steele, J. M. A.; Tacconis, C.; Bocarsly, J. D.; Mathon, O.; Neufeind, J. C.; Liu, J.; O’Keefe, C. A.; Goodwin, A.

Paper details

vlnc:W_j29kfbrk
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents